Guest-dependent solid-state superstructure plasticity and its application to adsorptive benzene sieving Authors: Ming Li, Qian Zhang, Hongyan Zhang, Serhii Vasylevskyi, Min-Sung Ko, Juan S. Cervantes, Feihe Huang, Jonathan L. Sessler Published: 2026 ABSTRACT Conformational plasticity offers a powerful route to adaptive molecular materials, yet its translation into guest-dependent superstructures with controllable function remains rare in the case of synthetic macrocycles. Here we show that two conformationally adaptive macrocycles, MeC3 and MeC4, provide a well-defined macrocyclic platform for investigating guest-dependent conformational plasticity in the solid-state. Specifically, supramolecular co-assembly of MeC3 with molecular guests that differ in geometry, polarity and electronic character affords a diverse set of guest-dependent macrocyclic polymorphs and higher-order solid-state host–guest superstructures. Combined spectroscopic, computational, and crystallographic analyses correlate guest identity with the stabilization of distinct macrocyclic conformations that then propagate into divergent packing modes and superstructures. As a presumed consequence of the adaptive solid-state behavior, MeC3 crystals selectively sieve benzene vapor from benzene/cyclohexane mixtures with >95% selectivity. The MeC3 sieve material can be readily regenerated through a reversible guest-loaded to guest-free crystal-to-crystal transformation. In contrast to MeC3, MeC4 proved effective as a solid-state receptor for fullerenes C60 and C70. This work highlights solid-state conformational plasticity as a promising design principle for the creation of functional solids with emergent recognition, assembly, and separation properties. Introduction Conformational adaptability is a hallmark of biological macromolecules and underpins a range of essential processes, including molecular recognition,1 signal transduction,2 and allosteric regulation.3 Supramolecular chemists have long sought to create synthetic hosts that not only mimic the structural complexity of biomolecular systems but also exhibit similar adaptability and selective recognition.4–10 Macrocyclic hosts have played a central role in this effort on account of their well-defined cavities, rich noncovalent interaction profiles, and structural tunability.11–16 This has rendered them attractive platforms for biomimetic applications.17–22 A critical determinant of the performance and functionality of synthetic macrocyclic hosts is conformational plasticity, which is generally used to describe the ability to access distinct conformational states under external stimuli, such as exposure to guests or changes in the bulk environment.23–27 To date, this concept has been exploited in the construction of adaptive materials with responsive recognition features and in efforts to mimic naturally occurring induced fit and conformational selection mechanisms (Fig. 1a and b).28–30 However, this guiding principle remains relatively unexplored in the solid-state.31–33 This paucity of effort reflects the scarcity of suitable macrocyclic platforms that combine structural preorganization with guest-responsive conformational adaptability. It is compounded by the difficulty of correlating local conformational changes with higher-order packing and superstructure morphologies. Schematic illustration of molecular recognition pathways involving (a) induced fit, in which binding precedes conformational adaptation, and (b) conformational selection, in which conformational equilibrium precedes binding. (c) Cartoon representation of the guest-dependent solid-state superstructure plasticity observed for MeC3 and MeC4 in the presence of different guests. The spheres at the vertices of the triangle and square denote the orientations of the N-substituents; different colors indicate alignment in opposite directions. In contrast to solution-phase systems, crystalline host–guest materials offer several distinct advantages including (i) robustness and stability that support functional applications, such as molecular separations and solid-state sensing,34–40 (ii) well-defined structures that enable direct structure–function correlations,41–43 and (iii) an ability to undergo guest-induced conformational transformations allowing for rational materials design.44–46 Typical examples include single-crystal-to-single-crystal (SCSC) transformations, wherein the crystalline lattice remains intact, an attribute that has enabled applications in molecular recognition and reversible switching in the solid-state.47–49 Much less studied are systems where the crystalline materials display guest-dependent adaptive behavior. In particular, relatively little effort has been devoted to crystalline systems in which different guest-inclusion environments stabilize not only distinct macrocyclic conformations, but also diverse packing arrangements and higher-order superstructures. This is a phenomenon we refer to as solid-state superstructure plasticity. Appropriately chosen macrocyclic systems may allow a better understanding of how guest-dependent local adaptation can be translated into collective solid-state behavior and functions. The present study was undertaken in an effort to explore this possibility. Macrocyclic arenes derived from seven-membered-ring-fused monomers are particularly attractive as chemical tools for studying guest-dependent solid-state responsive effects. This is because they offer a potentially useful balance between structural definition and adaptive character. Subsequent to our initiating the present study, a direct precedent for the methyl-substituted series was provided by Ogoshi and co-workers, who in 2025 identified the cyclic trimer MeC3 as an expanded-cavity host and detailed its basic host–guest behavior; however, its larger methyl-substituted congeners were not obtained in isolable form.50 Related studies by Jia and co-workers, also in 2025, established an ethyl-substituted iminodibenzyl platform in which both trimeric and tetrameric homologues proved accessible; these systems exhibited distinct conformational, structural, and recognition characteristics, with the trimeric host supporting guest-modulated room-temperature phosphorescence.51 These recent reports notwithstanding, the broader potential of these frameworks, of which MeC3 is representative, and its higher methyl-substituted homologues (e.g., MeC4) as adaptive solid-state platforms remain largely unexplored. Here, we report the conformationally adaptive behavior of two electronically rich macrocyclic hosts, namely the known cyclic trimer MeC3 and its cyclic tetrameric homologue MeC4. In our hands, both macrocycles proved accessible in a one-pot manner. Moreover, as detailed below, both systems were found to exhibit pronounced guest-dependent conformational plasticity in the solid-state. Specifically, co-assembly of macrocycles MeC3 and MeC4 with a series of guest molecules gives rise to a diverse family of solid-state host–guest superstructures (Fig. 1c). Guest binding, in turn, stabilizes a range of molecular conformations in the crystalline-state, that run the gamut from symmetric cone-like structures to highly distorted or twisted geometries. After activation to produce crystalline MeC3α, the trimer was found to capture benzene from benzene/cyclohexane mixtures with >95% selectivity. It could then be regenerated through a reversible crystal-to-crystal transformation, making it a promising crystalline separation material. Results and discussion One-pot access to the homologous MeC3 and MeC4 platforms and solution-phase analyses of the conformational features To establish a homologous macrocyclic platform for investigating guest-dependent solid-state superstructure plasticity, we sought direct access to discrete macrocycles of different ring sizes from a common precursor. As noted above, the trimeric system MeC3 was recently reported.50 In this study, larger methyl-substituted macrocyclic congeners, including tetrameric species, were detected by mass spectrometry; however, they were not obtained in isolable quantities. On the other hand, this finding led us to consider that formation and isolation of the tetrameric framework might be possible. In fact, under the present condensation conditions (Fig. 2a), both MeC3 and its tetrameric homologue MeC4 can be obtained and isolated via a one-pot procedure involving the condensation of precursor M with paraformaldehyde in CH2Cl2 under TFA catalysis at 35 °C. These two homologues could be isolated in analytically pure form, albeit in modest yields (4.7% and 6.1%, respectively), after chromatographic separation and recrystallization. The isolation of MeC4 enabled direct examination of its solid-state conformational features and permitted comparisons with MeC3. Both macrocycles were characterized by 1D/2D NMR spectroscopy and Q-TOF mass spectrometry (Fig. S2–S9). Notable differences between MeC4 and MeC3 were seen in their respective 1H NMR spectra (Fig. 2b). For instance, slight downfield shifts were seen for most proton peaks in MeC4 relative to MeC3 (i.e., Δδ = +0.18 ppm for HB (vs. Hb), +0.09 ppm for HC (vs. Hc), +0.12 ppm for HD (vs. Hd), +0.06 ppm for HE (vs. He), and +0.07 ppm for HF (vs. Hf)), where capital and lower case letters are used to denote the protons of MeC4 and MeC3, respectively. These differences could reflect variations in the relative conformational preferences of the two macrocyclic systems. (a) One-pot synthesis of MeC3 and MeC4. (b) Comparison of the truncated 1H NMR spectra (500 MHz, CDCl3, 298 K) of (I) MeC3 and (II) MeC4. (c) Partial 1H NMR spectra (CD2Cl2, 500 MHz) of MeC3 and MeC4 at 188 K. (d) DFT-optimized low-energy conformers of MeC3 and their relative free energies: cone conformation (MeC3-I) and partial cone conformation (MeC3-II). (e) DFT-optimized low-energy conformers of MeC4 and their relative free energies: cone conformation (MeC4-I), partial cone conformation (MeC4-II), 1,2-alternate conformation (MeC4-III) and 1,3-alternate conformation (MeC4-IV). ΔGrel is referenced to the global minimum. Note that lower-case letters are used to denote protons on MeC3 and upper-case letters are used to show those on MeC4. To gain deeper insights into the solution-phase conformational behavior of MeC3 and MeC4, variable-temperature (VT) 1H NMR spectroscopic analyses were carried out in CD2Cl2. These experiments were designed to probe and compare their structural flexibility in solution and their ability to undergo adaptive changes in the solid state. Upon cooling, pronounced changes were seen in the spectrum of MeC3, indicating reduced intramolecular motion and restricted conformational dynamics (Fig. 2c and S10). DFT calculations carried out at the B3LYP/6-311+G(d,p) level revealed two low-energy conformers for MeC3, namely a symmetric cone (MeC3-I) and a partial cone (MeC3-II) form. The latter was found to be the thermodynamic minimum; however, DFT calculations revealed that the two forms are separated by less than 1 kcal mol−1 in relative free energy (ΔGrel, Fig. 2d). By comparison, MeC4 remained spectroscopically averaged over the full course of the temperature range examined. Even at 188 K, the aromatic HC and methylene-bridge HF resonances persisted as single, unsplit signals (Fig. 2c and S11). This finding is consistent with rapid conformational exchange on the NMR timescale. In this case, DFT calculations revealed a flat conformational landscape for MeC4 (Fig. 2e), with the cone (MeC4-I), partial cone (MeC4-II), 1,2-alternate (MeC4-III), and 1,3-alternate (MeC4-IV) conformers all lying within ca. 2.0 kcal mol−1 of one another. Tetramer MeC4-I was predicted to be the global minimum, closely followed by MeC4-IV and MeC4-III (ΔGrel = +0.11 and +0.47 kcal mol−1, respectively). This set of low-lying states provides a straightforward rationale for the persistent VT-NMR averaging observed for MeC4 in CD2Cl2 solution. On the other hand, this flexibility led us to consider whether guest binding in the solid-state would act to bias the system toward distinct conformational outcomes. We thus sought to determine the extent, if any, to which conformational plasticity is manifest in the solid-state and how it might be controlled through guest recognition. Guest-dependent cone-state superstructures of MeC3 The geometric characteristics of macrocyclic hosts are critical in dictating their guest selectivity and supramolecular assembly behavior. In the case of MeC3, which is composed of three conformationally preorganized building blocks linked via methylene bridges, the overall molecular geometry closely resembles a distorted hexagon. This structural resemblance suggested to us that benzene (C6H6, Bz), a planar aromatic molecule with a hexagonal shape, could serve as an effective binding partner, given its inferred size and shape complementarity with the central cavity of MeC3. In accord with this thinking, co-assembly of MeC3 with Bz resulted in the formation of a well-defined 1 : 1 host–guest complex (Bz@MeC3), in which the Bz molecule was embedded in near-coplanar fashion within the central cavity of MeC3, aligning well with the hexagonal topology of the macrocycle (Fig. 3a). A single crystal X-ray diffraction (SCXRD) analysis revealed that all MeC3 units in the crystal lattice adopt a uniform bowl-shaped cone conformation (Fig. S12). Moreover, molecules of MeC3 adopt a layer-like supramolecular arrangement stabilized by intermolecular [C–H⋯π] interactions between adjacent MeC3 units inferred based on the metric parameters (Fig. 3b and S13). The MeC3 units within each layer adopt the same cone orientation, while adjacent layers display alternating cone directions (Fig. 3c and S14). Based on these combined observations, we conclude that Bz fits well within the MeC3 cavity and is associated with the stabilization of a bowl-shaped cone-like arrangement of the resulting solid-state superstructure. Solid-state superstructures of (a) Bz@MeC3 and (d) Cy@MeC3, showing that the macrocyclic host MeC3 adopts a consistent cone-like conformation. Cavity-based 1 : 1 host–guest complexation modes are shown for both Bz@MeC3 and Cy@MeC3. The packing arrangements exhibit honeycomb-like superstructures as illustrated in (b) and (e) as viewed along the a-axis and c-axis, respectively. Schematic representation of the different macrocycle cone orientations seen in Bz@MeC3 (c) and Cy@MeC3 (f) within the overall packing superstructure. The hydrogen atoms in certain superstructures are omitted for clarity. To evaluate further the influence of guest geometry on the conformational behavior of MeC3, we tested cyclohexane (C6H12, Cy) as a potential guest molecule (Table S2). SCXRD analysis of the cocrystallized complex (Cy@MeC3) confirmed the formation of a discrete 1 : 1 host–guest adduct between MeC3 and Cy (Fig. 3d). The Cy molecule is accommodated within the macrocyclic cavity of MeC3. As observed for Bz, the encapsulation of Cy is accompanied by stabilization of a well-defined bowl-shaped cone conformation of MeC3 in the crystalline state (Fig. S16). In the solid-state superstructure of Cy@MeC3, the MeC3 macrocycles pack into layered arrays (Fig. 3e and S17). However, the packing motif of the host in the cone orientation differs from that observed in Bz@MeC3. For instance, whereas the cone orientations of MeC3 molecules alternate within a single layer, the molecules aligned along the c-axis exhibited uniform cone orientations (Fig. 3f and S18). These findings are considered to reflect the conformational adaptability of MeC3 and how subtle differences in guest structure (e.g., Bzvs.Cy) can be translated into distinct solid-state packing superstructures. Guest-dependent partial cone superstructures of MeC3 To probe further the conformational plasticity of MeC3 in the crystalline state, we examined its host–guest chemistry with substituted aromatic guests that differ subtly in shape and polarity, focusing on p-xylene (p-XL) and o-dichlorobenzene (o-DCB) (Table S3). Our objective was to determine whether these variations in the guest translate into distinct binding modes and, in turn, different solid-state conformations and crystalline superstructures for the host, MeC3. Single crystals of p-XL@MeC3 were obtained by slow evaporation. A SCXRD analysis revealed a 1 : 1 in-cavity complex in which p-XL is encapsulated within the macrocyclic pocket in a tilted orientation (Fig. 4a and b). The p-XL inclusion complex features MeC3 in a partial cone conformation, characterized by alternating orientations of the N-substituted arms (Fig. S24). Two crystallographically independent host–guest complexes are present in the asymmetric unit (type-1 and type-2), and the two MeC3 hosts adopt opposite partial cone geometries. This leads to formation of an anti-aligned dimeric pair. The two complexes display different [C–H⋯π] contact patterns (Fig. 4a): Type-1 features two contacts between p-XL and the aromatic walls of MeC3 (A and B), whereas type-2 exhibits two complementary [C–H⋯π] interactions involving one hydrogen from p-XL and one from MeC3 (C and D). A honeycomb-like packing motif is observed in the bc-plane, with the MeC3 macrocycles propagating along the a-axis to generate well-defined one-dimensional channels occupied by the encapsulated p-XL guests (Fig. S25). Within each column, the hosts adopt a periodically varying partial cone orientation, while an overall ABAB stacking pattern emerges in the ac-plane. These findings are consistent with a highly ordered and efficiently packed superstructure persisting in the solid state (Fig. 4c). (a–c) Views of the solid-state superstructures of p-XL@MeC3: (a) Black dashed lines indicate inferred [C–H⋯π] interactions (with A–D distances of 2.82, 2.74, 2.96, 2.82 Å, respectively) between MeC3 and p-XL; (b) view of two adjacent crystallographically distinct MeC3 molecules with distinct conformational orientations within an in-cavity complexation mode; (c) conformational features and packing of MeC3 in p-XL@MeC3: (1) 1D channels seen along the a-axis for the p-XL molecules in p-XL@MeC3; (2) alternating orientations of the partial cone conformers observed along the a-axis channel. (d–f) Crystallographic views of the solid-state superstructures of o-DCB@MeC3: (d and e) Planar and lateral views showing a portal-associated binding motif; (f) conformational features and packing of MeC3 in o-DCB@MeC3: (1) guest-mediated changes propagated along the b-axis; (2) identical partial cone conformations with uniform orientation within each column. Color-coded components denote symmetrical equivalence. We next turned to o-DCB, a sterically more demanding guest. A SCXRD analysis revealed the formation of a 1 : 1 host–guest complex (o-DCB@MeC3), where the MeC3 macrocycle again adopts a distorted partial cone conformation (Fig. 4d and S27). In contrast to p-XL in p-XL@MeC3, the o-DCB guest is not fully encapsulated in o-DCB@MeC3, instead, it resides near the portal region and engages primarily with the rim and exterior surface of the host (Fig. 4e). Analysis of the SCXRD data reveals the presence of two crystallographically distinct host–guest complexes, in which the MeC3 hosts are designated as type-3 and type-4, respectively. The associated complexes are found to display opposite partial cone geometries, thereby forming an anti-aligned partial cone host pair analogous to that seen in p-XL@MeC3, but with distinct guest positioning and noncovalent interaction motifs (Fig. 4d). In the crystal superstructure, o-DCB acts as a supramolecular linker that connects neighboring MeC3 units through a combination of various noncovalent contacts involving the chlorinated guest (Fig. S28). This connectivity yields an extended channel-like architecture in which MeC3 propagates along the crystallographic b-axis with a uniform partial cone orientation, giving rise to a linear stacking sequence (Fig. 4f and S29). Taken together, these results establish (i) the partial cone conformation as an accessible adaptive state of MeC3 and (ii) that disparate guest-binding modes within this state can support different crystalline superstructures. Electron-deficient guests stabilize charge-transfer solid-state superstructures of MeC3 As the next step in this study, we sought to explore whether electronic complementarity could likewise influence the adaptive host–guest behavior of MeC3. As a first step in testing this possibility, we analyzed the electrostatic potential (ESP) surface of MeC3. The resulting calculated ESP map revealed a negative potential mainly located around the macrocyclic cavity (Fig. 5a). On this basis, we suggest that MeC3 should serve as an electron-rich receptor that stabilizes donor–acceptor interactions with electron-deficient aromatics. Support for this supposition came from studies involving two representative electron-deficient guests (Fig. 5b and Table S4), namely 4-nitrobenzonitrile (NBN, para-disubstituted) and 1,5-difluoro-2,4-dinitrobenzene (DFN, tetra-substituted). The solution-phase interactions with these putative guests were probed by NMR and UV-Vis spectroscopy. Using MeC3·NBN as a representative pair, addition of MeC3 induced pronounced upfield shifts in the NBN aromatic resonances (Δδ = 0.65 and 0.31 ppm for H1 and H2, respectively). These changes were accompanied by downfield shifts in the signals for the MeC3 protons Ha and Hb (Fig. S31a). A corresponding UV-Vis spectral analysis revealed the presence of a new, broad CT band in the mixture that is absent from the spectrum of either component alone. These spectral changes were accompanied by a visible color change from colorless to pale yellow (Fig. 5c). Similar changes were observed when MeC3 was titrated with DFN (Fig. S31b and c). We thus conclude that MeC3 can form charge-transfer host–guest assemblies with the electron-deficient guests NBN and DFN in chloroform solution. ESP surfaces: (a) electron-rich MeC3; (b) electron-deficient guests NBN and DFN. (c) UV-Vis spectra (CHCl3) of (G) NBN, (HG) a host + guest mixture of NBN and MeC3, and (H) MeC3. Inset: color changes in a chloroform solution ascribed to presumed CT interactions between NBN and MeC3. (d) Solid-state superstructures of the CT complex NBN@MeC3. Here black dashed lines indicate inferred [C–H⋯π] interactions (with A–D distances of 2.88, 2.70, 2.69, 2.76 Å, respectively) between the MeC3 host (H) and the NBN guest (G). (e) Independent gradient model (IGM) analyses supporting the presence of multiple weak intermolecular interactions. (f) Schematic illustration of an antiparallel partial cone pair formed between adjacent MeC3 macrocycles in NBN@MeC3. (g) ESP map of NBN@MeC3. (h) Solid-state superstructures of the CT complex DFN@MeC3, where black dashed lines indicate inferred [C–H⋯π] interactions (with distances E–G = 2.38, 2.71, 2.43 Å, respectively) and [C–H⋯O] interactions (with distance H = 2.43 Å) between the MeC3 units and the DFN guests. (i) IGM analyses providing support for the presence of multiple weak intermolecular interactions in DFN@MeC3. (j) Schematic illustration of an antiparallel partial cone pair resulting from contacts between adjacent MeC3 macrocycles in DFN@MeC3. (k) ESP map of DFN@MeC3. Next, diffraction-quality crystals of NBN@MeC3 were obtained by slow vapor diffusion of MeOH into a CHCl3 solution of the complex. A SCXRD analysis revealed a 1 : 1 inclusion complex in which planar NBN is encapsulated within the MeC3 cavity to afford a [2]pseudorotaxane-like architecture (Fig. 5d). Two crystallographically independent host–guest complexes are present in the asymmetric unit. The MeC3 hosts in these structures are designated as type-5 and type-6, respectively. In both cases, the MeC3 macrocycles adopt opposite partial cone geometries with complementary orientations of the N-substituted arms (“two-down/one-up” vs. “two-up/one-down”). This results in reversed alignments of the included NBN guests (Fig. 5f, S32a and b). Although both complexes feature cavity inclusion, their local contact patterns differ (Fig. 5d): in the type-5 MeC3 host–guest complex, the NBN guest interacts with the inward-facing aromatic walls of MeC3via two [C–H⋯π] contacts (H⋯π = 2.70 and 2.88 Å; A and B), whereas in the complex involving the type-6 MeC3 host, the inferred interactions involve a guest-to-host [C–H⋯π] contact (H⋯π = 2.69 Å; C) and a complementary host-to-guest contact in which a MeC3 C–H group approaches the electron-deficient arene face of NBN (H⋯π = 2.76 Å; D). At the packing level, the complexes assemble into ordered one-dimensional channel arrays in which adjacent macrocycles adopt alternating partial cone orientations (Fig. S32c), enforcing an antiparallel arrangement of the encapsulated NBN molecules (Fig. S33). These noncovalent contacts were corroborated by an independent gradient model (IGM) analysis, which highlighted both host–guest interactions and inter-host contacts within the lattice (Fig. 5e). Consistent with the proposed donor–acceptor interactions, ESP mapping of the NBN@MeC3 complex revealed attenuation of the electron-rich character around the macrocyclic core upon guest inclusion (Fig. 5g). This finding provides support for the notion that electrostatic complementarity and charge-transfer (CT) contributions are key driving forces in the solid-state. The solid-state structure of DFN@MeC3 was likewise solved. Despite the greater steric demand of DFN relative to NBN, the superstructure and host conformational features closely resemble those of NBN@MeC3 (Fig. 5j, S34 and S35). DFN@MeC3 crystallizes in the triclinic P1̄ space group with two crystallographically distinct MeC3 macrocycles (referred to as type-7 and type-8, respectively) (Fig. 5h). In the type-7 DFN@MeC3 complex, the DFN guest is accommodated within the central cavity of MeC3, primarily stabilized by two [C–H⋯π] interactions (denoted E and F) as inferred from the metric parameters. In the type-8 complex, a slightly different binding pattern is seen wherein the DFN molecule is anchored within the macrocyclic cavity via the synergistic contributions of inferred [C–H⋯π] and [C–H⋯O] hydrogen bond interactions (cf. separations G and H). IGM analyses and ESP mapping again indicate a reduction in the electron-rich character of MeC3 upon complexation, mirroring the CT signatures observed for NBN@MeC3 (Fig. 5i and k). Despite their marked differences in steric demand and substitution pattern, both NBN and DFN lead MeC3 to adopt closely related partial cone conformations that are characterized by similar channel-type solid-state superstructures. This convergence leads us to suggest that donor–acceptor interactions within the adaptive electron-rich cavity of MeC3 can reduce the influence of guest-specific steric differences and favor a supramolecular assembly mode in the case of the resulting complex. Solid-state recognition of fullerenes C60 and C70 by MeC4 We next sought to explore whether the larger macrocycle MeC4 could exhibit analogous or enhanced guest-dependent conformational versatility. Relative to MeC3, MeC4 contains four methylene-bridged monomeric units and hence a larger central cavity. ESP analysis revealed that MeC4 possesses an electron-rich inner surface in analogy to what was seen for MeC3 (Fig. S36). We thus considered it might prove effective in encapsulating fullerenes C60 and C70 (Table S5). To test this possibility, MeC4 was cocrystallized with C60 and C70, followed by SCXRD analysis of the resulting crystalline complexes. Slow evaporation of toluene solutions containing MeC4 and excess C60 yielded diffraction-grade single crystals of MeC4·C60 with a host–guest–solvent ratio of 1 : 5 : 2. In the solid-state, the bowl-shaped cavity of MeC4 adopts a cone conformation (MeC4-I). One C60 guest is accommodated within its central cavity. Each 1 : 1 host–guest unit is surrounded by eight peripheral C60 molecules (Fig. 6a and S37). These interact with the core complex via a combination of inferred multiple π–π donor–acceptor interactions and appear further stabilized via van der Waals contacts with interstitial toluene molecules. As a result, a seemingly unique crystal superstructure is seen wherein units of MeC4·C60 assemble into a regular supramolecular array propagating along the b-axis (Fig. 6b and S38), with the orientation of the MeC4 macrocycles (in their cone conformations) alternating within each layer. (a and b) Crystallographic views of the solid-state superstructures of a MeC4·C60 cocrystal: (a) view showing the symmetric peripheral arrangement of eight C60 molecules around MeC4 that is accompanied by an endohedral encapsulation. Note the bowl-shaped cone conformation of the MeC4 macrocyclic scaffold; (b) view showing the overall stacking in MeC4·C60 highlighting the alternating orientations of the cone conformation adopted by the MeC4 within each layer. (c and d) Crystallographic views of the solid-state superstructures of a MeC4·C70 cocrystal: (c) peripheral C70 units are associated with distortion of the MeC4 framework. Also shown is a view of the solid-state structure showing how the host MeC4 adopts a partial cone conformation upon binding the elongated guest C70. (d) View of the overall compact supramolecular array observed in crystals of MeC4·C70 as viewed along the a-axis. Color-coded components denote symmetrical equivalence. To probe the solid-state conformational adaptability of MeC4 toward larger, ellipsoidal guests, we subjected it to cocrystallization with C70 under high-temperature solvent evaporation conditions. In contrast to the symmetric bowl-shaped geometry observed in crystalline MeC4·C60, the MeC4 host in MeC4·C70 adopts a distinct “three-up-one-down” partial cone conformation (MeC4-II) (Fig. 6c and S39). Notably, no endohedral binding of C70 is observed. Instead, the elongated geometry of C70 apparently promotes π–π donor–acceptor interactions with the sidewalls of MeC4. This, in turn, leads to asymmetric host deformations that serve to optimize surface complementarity. As inferred from the metric parameters, the C70 molecules act as bridging elements linking adjacent MeC4 units via multiple contact points. The net result is a tightly packed supramolecular network (Fig. 6d). This packing motif is distinct from the layered architecture seen for MeC4·C60 (Fig. S40). Collectively, these observations lead us to suggest that MeC4 is a geometrically adaptive platform wherein the choice of guest determines whether cavity inclusion occurs. It also influences the host conformation and controls the extent to which higher-order architectures are observed in the crystalline state. To compare these guest-dependent solid-state structures more systematically, Hirshfeld surface analyses and two-dimensional (2D) fingerprint plots were performed for representative MeC3 and MeC4 host–guest crystals (Fig. S41–S46). For the MeC3 cone-state structures, Bz@MeC3 and Cy@MeC3 are dominated by H⋯H and C⋯H/H⋯C contacts, reflecting compact in-cavity guest inclusion and close packing. In the case of the MeC3 partial-cone structures, p-XL@MeC3 remains dominated by related H/C-based close contacts, whereas o-DCB@MeC3 exhibits additional H⋯Cl/Cl⋯H and C⋯Cl/Cl⋯C contacts arising from the chlorinated guest positioned near the host portal region. By contrast, the fullerene-containing MeC4 assemblies show markedly larger C⋯H/H⋯C and C⋯C contributions than the MeC3 host–guest structures; this is expected given the carbon-rich surfaces of C60 and C70. These results provide a contact-based comparison of the different crystal structures and support the view that guest geometry, binding mode, local contact profiles, and lattice packing collectively contribute to the observed conformational and solid-state superstructural diversity. Adaptive MeC3 crystals for selective benzene-over-cyclohexane discrimination Aromatic hydrocarbons and their cyclic aliphatic congeners often exhibit similar physicochemical properties, such as boiling point, volatility, and density. Bz and Cy are prototypical representatives of these two classes (Fig. 7a and S47). They exhibit nearly identical boiling points (Bz: 353.25 K; Cy: 353.85 K) and closely matched molecular weights and densities (Bz: 0.88 g cm3; Cy: 0.78 g cm−3). Consequently, their separation by traditional techniques such as fractional distillation is often inefficient and energetically demanding. This challenge has motivated the development of supramolecular strategies for their selective recognition and separation. While a number of sieving-based approaches have demonstrated success,44,52–58 there remains a need for easy-to-access systems that support Bzvs.Cy separations. Prompted by the distinct solid-state superstructures observed for Bz@MeC3 and Cy@MeC3 (vide supra), we explored the potential of MeC3-based materials for vapor-phase discrimination between these guests. Activated MeC3 (denoted MeC3α as noted above) was prepared by thermal activation under vacuum. It was found to be essentially solvent free as confirmed by 1H NMR spectroscopy, thermogravimetric analysis (TGA), and powder X-ray diffraction (PXRD) analyses (Fig. S48–S50). Moreover, Brunauer–Emmett–Teller (BET) analysis revealed a low surface area (12.55 m2 g−1), indicating a nonporous and densely packed structure (Fig. 7b). (a) Chemical structures and ESP surfaces of Bz and Cy. (b) N2 sorption isotherm of MeC3α. Adsorption, black symbols; desorption, red symbols. (c) Time-dependent MeC3α solid–vapor sorption plots for single-component Bz vapor. (d) Time-dependent MeC3α solid–vapor sorption plots for single-component Cy vapor. (e) Time-dependent solid–vapor sorption plots of MeC3α for an equi-volume Bz/Cy vapor mixture. (f) PXRD patterns of MeC3: (I) pristine MeC3α; (II) after exposure to Cy vapor; (III) after exposure to Bz vapor; (IV) after exposure to an equi-volume Bz/Cy liquid mixture vapor. (g) Relative uptakes of Bz and Cy adsorbed by MeC3α as measured by gas chromatography. (h) PXRD patterns of MeC3: (I) pristine MeC3α; (II) after adsorption of Bz vapor; (III) after desorption of (II). (i) Calculated binding energies of Bz@MeC3 and Cy@MeC3. Single-component solid–vapor sorption experiments revealed comparable uptake capacities for both Bz and Cy (ca. 0.8 mol mol−1MeC3α), but with distinct sorption kinetics: Bz equilibrated within 2 h, while Cy required ∼5 h (Fig. 7c and d). The PXRD patterns revealed substantial structural changes upon exposure to both Bz and Cy vapors, indicative of vapor-induced transformations giving rise to new crystalline forms (Fig. S51 and S52). As depicted in Fig. 7e, competitive sorption experiments using an equi-volume Bz/Cy liquid mixture showed strong selectivity toward Bz. While Bz was readily adsorbed to saturation, Cy uptake was negligible, demonstrating a pronounced molecular discrimination effect. The phase selectivity was further supported by PXRD analyses. The diffraction pattern of MeC3α after exposure to the mixed Bz/Cy vapor matched well with that of MeC3α exposed to Bz vapor alone and was also consistent with the simulated PXRD pattern for Bz@MeC3 derived from the SCXRD data (Fig. 7f and S53). Quantitative analysis by gas chromatography (GC) revealed that Bz accounted for over 95% of the total adsorbed vapor under these test conditions (Fig. 7g and S54). Since the adsorption of Bz by MeC3α is mainly ascribed to the weak interactions between the host and guest, we anticipated that the guest-free MeC3α form could be readily regenerated by heating at 130 °C under vacuum for 2 h. As shown in Fig. 7h, the PXRD pattern of the material treated in this way was essentially identical to that of pristine MeC3α. This finding leads us to suggest that MeC3α could be recycled for further use; in fact, it proved possible to do so over five cycles (Fig. S55). Mechanistic insights into the preferential adsorption and selective separation were obtained from DFT calculations and ESP surface analysis. Binding energy calculations (Fig. 7i) revealed a more favorable thermodynamic profile for Bz@MeC3 (Eb = −18.08 kcal mol−1) compared to Cy@MeC3 (Eb = −16.99 kcal mol−1). These calculations thus support the conclusion that MeC3 interacts more favorably with Bz than Cy. ESP mapping (Fig. S56) proved consistent with this preference: Bz is characterized by an electron-deficient periphery that is complementary to the electron-rich MeC3 cavity. In contrast, Cy exhibits a uniform, nonpolar ESP surface. The selective uptake of Bz by MeC3α thus appears to arise from the combined effects of a more optimal geometric fit, host–guest electrostatic complementarity, and the differing solid-state response of MeC3 toward Bz and Cy. Conclusions Reported here are studies of two conformationally adaptive macrocycles, MeC3 and MeC4, that support guest-dependent solid-state conformational plasticity. Changes in form were seen upon co-assembly with a diverse set of guest molecules. Moreover, these macrocycles were found to support a broad spectrum of crystalline host–guest superstructures upon exposure to various guests. Specifically, MeC3 adopts distinct conformations (bowl-shaped cone and partial cone) while the corresponding host–guest ensembles exhibit diverse higher-order crystalline superstructures that depend on guest identity. This adaptive behavior is thought to underlie the selective capture of Bz from Bz/Cy mixtures seen in the case of crystalline MeC3. The larger and more flexible cavity of MeC4 allows it to act as a receptor for fullerenes C60 and C70. Depending on the fullerene in question either symmetric cone-like geometries or highly distorted structures are observed in the resulting crystalline assemblies. Considered in concert, the present study not only underscores the pivotal role that conformational plasticity can play in solid-state recognition but also offers insights into how guest-adaptive functional crystalline materials may be exploited in separations. Author contributions M. L. conducted most of the experiments, analysed the data, and drafted the manuscript. Q. Z. contributed to selected computational studies. H. Z. contributed to the experimental work. S. V. contributed to the analysis of crystallographic data. M.-S. K. and J. S. C. assisted in the experimental studies. F. H. and J. L. S. supervised the project and revised the manuscript. All authors discussed the results and approved the final version of the manuscript. Conflicts of interest There are no conflicts to declare. Supplementary Material