Composite Hydrogel With Ultra‐High Mechanical Anisotropy and Biotissue‐Like Softness by Confining Electrostatic and Entropic Repulsion Between Cofacially Aligned Nanosheets

ABSTRACT

Owing to their biotissue‐like nature, mechanically anisotropic hydrogels have attracted extensive attention. Recent studies have achieved high mechanical anisotropy by densely assembling oriented polymers and/or nanofillers into stiff networks. However, this strategy involves soft‐direction hardening to afford intrinsically hard hydrogels, limiting their applications, particularly in biomedical fields. Here, we developed the first hydrogel that simultaneously achieves excellent mechanical anisotropy and a low elastic modulus in the soft direction, using magnetically oriented, negatively charged nanosheets. The hydrogel is directionally reinforced by the electrostatic and entropic repulsions between nanosheets. Although increasing the nanosheet concentration enhances the anisotropy, it cannot exceed the threshold where the gel‐precursor becomes viscous and unalignable. We discovered that a hydrogel, prepared with nanosheets at an alignable concentration and a less‐crosslinked polymer network, spontaneously shrinks vertically toward the nanosheets with expelling ∼30% of water, driven by nanosheet–polymer adsorption and polymer reconfiguration. This densification concentrates the nanosheets while preserving their orientation, selectively amplifying the repulsion between nanosheets. Consequently, the hydrogel achieves an extraordinary anisotropy factor (hard modulus/soft modulus: E _hard/E _soft = 380), while maintaining a low modulus (E _soft = 3.1 kPa). This strategy produces anisotropic hydrogels that cover the modulus range of soft biotissues and exhibit excellent vibration insulation and rectification properties.

To pursue biomedical applications of anisotropic hydrogels, both high mechanical anisotropy and biotissue‐like softness are desirable. We developed the first hydrogel that simultaneously achieves excellent mechanical anisotropy and softness with an elastic modulus of 1–10 kPa, using unidirectionally oriented nanosheets embedded in a polymer network. The key was to enhance the repulsion between cofacial nanosheets through force‐induced training of the gel after its preparation.

1. Introduction

Hydrogels are soft, flexible, water‐rich three‐dimensional networks that have been extensively studied for their biocompatibility, responsiveness, and similarity to biological tissues [1, 2, 3]. Most of them are synthesized by polymerizing or assembling molecular components randomly oriented in water, which results in structurally isotropic networks. In contrast, many biotissues possess unidirectionally oriented structures, and derive their superb functions from their mechanical anisotropy [4, 5, 6, 7, 8, 9]. Therefore, mechanically anisotropic hydrogels are expected to provide attractive platforms for various applications, including tissue engineering, wearable devices, soft robotics, etc [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21]. So far, various anisotropic hydrogels have been developed by structuring polymers or nanofillers under anisotropic environments (Figure 1 and Table S1), using directional external fields (electric or magnetic fields, mechanical forces, etc.) or templates (ice nanopillars, bacterial cellulose, delignified wood, etc.) [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32].

FIGURE 1.

(a) Schematic representation of two types of mechanically anisotropic hydrogels: stiffness (left) and repulsion (right) types. (b) Overview of mechanically anisotropic hydrogels developed so far. Plots of mechanical anisotropy factor (E _hard/E _soft) vs. elastic modulus in the soft direction (E _soft).

To further develop the unique functions of anisotropic hydrogels, a high mechanical anisotropy factor is desirable. i.e., a high ratio between the elastic moduli in the hard and soft directions (E _hard/E _soft). For this aim, the simplest approach is to assemble oriented polymers and/or nanofillers at a high concentration (e.g. 30%–40%) through strong interactions, such as covalent bonds, hydrogen‐bonding, coordination, electrostatic, and hydrophobic interactions, as well as crystallization. This results in hydrogels reinforced by dense, stiff networks, which are referred to here as ‘stiffness gels’ (Figure 1a, left) [33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44]. Recent studies based on this approach have developed gels with an excellent mechanical anisotropy factor (Figure 1b, blue zone) [39, 40, 41, 42, 43, 44]. However, this approach inevitably strengthens gels in all directions, which compromises their intrinsic softness, i.e., increases their elastic modulus in the soft direction (E _soft) [33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44]. Consequently, anisotropy factors above ∼20 have been achieved only in hard gels with E _soft = 100–1000 kPa [39, 40, 41, 42, 43, 44], whereas in soft gels with E _soft = 1 to 10 kPa, an anisotropy factor has not yet exceeded the level of 10 (Figure 1b). This limitation severely restricts their biomedical applications. For effective integration with biotissues, their elastic modulus should match that of the target tissue, typically in the range of 1–100 kPa [11, 45, 46, 47, 48, 49]. Thus, developing hydrogels that combine pronounced mechanical anisotropy with biotissue‐like softness remains a critical yet unresolved challenge.

While stiffness gels face the above limitations, another type of anisotropic hydrogels, termed “repulsion gels,” have emerged to address this challenge (Figure 1a, right) [50, 51, 52, 53]. These gels contain ultrathin nanosheets (∼1 nm thick) that possess dense surface charges and interaction sites. These nanosheets in water are unidirectionally oriented by an external field (electric fields, magnetic fields, mechanical forces, etc.) and then fixed in a polymer network [50, 51, 52, 53, 54, 55, 56, 57, 58]. Despite the low contents of nanosheet (∼1 wt%) and polymer (5–10 wt%), strong electrostatic and entropic repulsions occur between cofacial nanosheets, owing to the high surface charge of nanosheets and the confinement of polymer chains in a narrow space [50, 51, 52, 53, 59]. Consequently, the gel is reinforced in the direction vertical to nanosheets (high E _⊥) while remaining soft in the parallel direction (low E _//), leading to high mechanical anisotropy (E _⊥/E _// = up to 85; Figure 1b, purple zone) [51]. If the nanosheet content were increased more, the vertical modulus (E _⊥) would rise sharply without affecting the parallel modulus (E _//), affording a gel with pronounced anisotropy and intrinsic softness with a sufficiently low elastic modulus. However, the content of nanosheets dopable into a hydrogel is limited, because when the nanosheet concentration exceeds the threshold, the gel precursor's viscosity rises significantly to prohibit their alignment by external fields [60, 61, 62, 63].

Here, we developed a series of hydrogels that simultaneously exhibit an excellent mechanical anisotropy factor (up to 380) and intrinsic softness with elastic moduli of 1–10 kPa, filling the missing profiles in previously reported anisotropic hydrogels (Figure 1b, red zone and Table S2). The key to this achievement is our unexpected finding of the spontaneous densification of a repulsion‐type gel after gel preparation. The gel is synthesized with a nanosheet content low enough for field‐induced alignment, while the crosslinking of the polymer network is exceptionally low. Owing to the reconfiguration of the flexible polymer network, the gel shrinks to ∼70% of its original size specifically along the vertical direction of nanosheet alignment, while preserving nanosheet orientation. This shrinkage expels ∼30% of the water, increasing the nanosheet concentration to 1.7 wt%, which cannot be achieved by direct synthesis of the gel. Since the distance between cofacial nanosheets is narrower, they become more repulsive, while their overall content remains low, enabling unprecedented properties: a mechanical anisotropy factor (E _hard/E _soft) of up to 380 with an intrinsic elastic modulus (E _soft) as low as 3 kPa. By tuning the initial gel composition, this approach yields a library of anisotropic hydrogels spanning the full modulus range of biotissues from brain to muscle, opening diverse biomedical applications. Moreover, the densified hydrogels exhibit excellent vibration insulation and rectification, suggesting additional potential in vibration engineering and soft robotics.

2. Results and Discussion

2.1. Design Principles of the Anisotropic Hydrogel Embedded With Aligned TiNSs

We have previously developed a repulsion‐type gel by using titanate nanosheets (TiNSs; Figure 2a i) and a polymer network composed of a monomer N,N‐dimethylacrylamide (DMA) or acrylamide (AAm) and the crosslinker N,N’‐methylenebis‐(acrylamide) (BIS) (Figure 2a ii), denoted as TiNS‐gel (Figure 2a iii) [50, 51, 52, 53]. TiNS is a monolayer single crystal of TiO₆ octahedra, notable for its ultrathin thickness (0.75 nm), large lateral width (several µm), high negative surface charge (1.5 C m⁻²), and abundant hydrogen‐bonding Ti─O─Ti groups on the surface [59, 64, 65]. In water dispersion, TiNSs adopt a cofacial arrangement with a periodic spacing of ∼200 nm, governed by the balance between electrostatic repulsion and van der Waals attraction [66, 67, 68]. The formation mechanism of the long periodicity has been established theoretically (Section S16) [66, 68]. When a dispersion of TiNS (1.2 wt%) containing monomer (6 wt%) and crosslinker (0.06 wt%) was placed in a 10 T magnetic field, the TiNS aligned perpendicular to the field (Figure 2b, Step 1), and subsequent in situ photopolymerization afforded a gel with fixed TiNS orientation (Step 2’). Unlike other alignment methods, magnetic alignment can achieve high‐order, homogeneous nanosheet alignment regardless of sample thickness.

FIGURE 2.

(a) Components and structure of TiNS‐gel. (i) Schematic structure of TiNS. (ii) Molecular structures of monomers and a crosslinker. (iii) Schematic structure of TiNS‐gel, in which electrostatic and entropic repulsions occur between cofacially oriented TiNSs. (b) Procedures for the preparation and subsequent annealing/training of TiNS‐gel. (Step 1) Magnetic orientation of TiNSs in water containing monomer and crosslinker. (Step 2 and Step 2’) Hydrogelation of the mixture with aligned TiNSs by in situ polymerization at low (Step 2) and high (Step 2’) crosslinker contents. (Step 3) Shrinkage of TiNS‐gel with low crosslinker content (obtained from Step 2) by time‐induced annealing or force‐induced training, driven by further adsorption of nanosheets and polymer chains and reconfiguration of polymer chains. (c) Changes in the internal structure of TiNS‐gel upon compression in the parallel (left) and vertical (right) directions to the alignment of TiNSs.

Between TiNSs in the gel, strong electrostatic repulsion raised from their dense negative charges (Figure 2a iv, yellow). In addition, polymer chains were anchored on TiNS with a high Young's modulus estimated at ∼230 GPa [69], which restricted the movement of the polymer chains and thus generated entropic repulsion (Figure 2a iv, blue). When the structure was compressed vertical or parallel to the TiNS alignment (Figure 2c), the nanosheets responded differently in terms of nanosheet distance and buckling. Under vertical compression, TiNSs retained their cofacial alignment while the nanosheet distance decreased. In contrast, parallel compression induced the buckling of the lamellar structure of TiNSs (Figure S1). Thus, electrostatic and entropic repulsions were enhanced only under vertical compression. Consequently, the elastic modulus of TiNS‐gel in the vertical direction (E _⊥) becomes significantly higher than that in the parallel direction (E _//) [50, 51, 52, 53]. When TiNS‐gel is doped with external ions, its mechanical anisotropy decreases, indicating the presence of electrostatic repulsion between TiNSs [50, 51]. Because the tensile deformation parallel and vertical to TiNS is equivalent to the compression deformation vertical to and parallel to TiNS, respectively, TiNS‐gel also exhibits anisotropic elastic modulus in tensile tests (Figure S2) [53].

A promising strategy to further enhance the mechanical anisotropy of TiNS‐gel is to increase its TiNS content, thereby making cofacial nanosheets closer and amplifying electrostatic and entropic repulsions (Figure 2a iii). However, our previous studies revealed that the TiNS concentration in the precursor solution cannot exceed 1.2 wt% [50, 51, 52, 53]. Similar to the cases of other colloidal nanosheets, concentrations beyond this threshold cause strong interference between TiNSs and sharply increase the viscosity of their dispersion [60, 61, 62, 63], preventing their magnetic alignment (Figure S3a). Although adding ions can reduce viscosity at higher TiNS concentrations, this effect arises from ionic shielding of electrostatic repulsion and thus diminishes the anisotropy of TiNS‐gel [51, 66, 67, 68]. Densification of TiNS‐gel, once formed at a low TiNS concentration, is also ineffective, as water removal by heating or drying inevitably induces structural heterogeneity and orientational disorder, which lowers the mechanical anisotropy.

2.2. Effects of the Adsorbability and Flexibility of Gel Network on TiNS‐Gel

The above issue of TiNS‐gel was unexpectedly addressed through our systematic study on how network interactions and flexibility influence the nanostructure and macroscopic shape of TiNS‐gel. For this aim, AAm and DMA were selected as adsorptive and less‐adsorptive monomers, respectively, based on the number of hydrogen‐bond donor sites (Figure 2a ii). Network flexibility was tuned by varying the BIS crosslinker ratio relative to the monomer (Figure 2a ii) [70, 71, 72]. A decrease in the BIS ratio lowers the number of crosslinking points in the polymer network and increases the length of the polymer chain between neighboring crosslinking points. Consequently, the polymer network gains conformational freedom and reconfigurability. This may lower the hydrogel's elastic modulus and allow for the polymer network's reconfiguration after gelation. By using the procedure we reported previously [50, 51, 52, 53], a series of TiNS‐gels were prepared at various compositions of TiNS (0.8 wt%), monomer (AAm or DMA, 3–12 wt%), and BIS (0–10⁻² relative to monomer). The nanostructure of the as‐prepared gel was studied by small‐angle X‐ray scattering (SAXS; Figure 3a). For the less‐adsorptive monomer DMA, both the orientation of TiNS (Figure 3a i) and the cofacial between TiNSs (Figure 3a ii, iii and Figure S4) remained nearly unchanged from precursor to gel, regardless of DMA or BIS concentration. We also monitored the macroscopic profiles after taking out the as‐prepared gel from the container. When the BIS content exceeded 10⁻³ relative to DMA, the gels were elastic and maintained their shape, whereas at BIS levels below 10⁻⁴, they gradually lost shape upon annealing at room temperature (Figure 3c i). In both cases, the gels retained their internal water during annealing (Figure 3c ii). These results indicate that the network based on DMA does not strongly interact with TiNSs and simply fixes their location and orientation.

FIGURE 3.

Effects of the adsorbability and flexibility of gel network on the microscopic structure and macroscopic shape of TiNS‐gel. (a,b) SAXS profiles of the gel precursors and the as‐prepared gels prepared from TiNS (0.8 wt%), monomer (a, DMA; b, AAm), and crosslinker (BIS). (i) 2D SAXS images. (ii) Intensity–q plots at various monomer concentrations of 0–12 wt% and a fixed crosslinker concentration of 0 wt%. (iii) Intensity–q plots at a fixed monomer concentration of 5 wt% and various crosslinker/monomer ratios of 0, 10⁻⁴, 10⁻³, and 10⁻². (c,d) Time‐course conversion from the as‐prepared gels into the annealed gels, which were prepared from TiNS (1.2 wt%), monomer (c, DMA; d, AAm), and crosslinker (BIS) at a fixed monomer concentration of 5 wt% and at various crosslinker/monomer ratios of 0, 10⁻⁴, 10⁻³, and 10⁻². (i) Polarized optical microscopy images. (ii) Time‐course changes in the lengths of gels in the direction vertical to the alignment of TiNSs

In contrast, for the adsorptive monomer AAm, in situ polymerization followed by room‐temperature annealing caused an unusual decrease in the distance between cofacial TiNSs. During the conversion from precursor to the as‐prepared gel, TiNS orientation was preserved (Figure 3b i). However, the distance between cofacial TiNSs markedly shrank from ∼150 to ∼60 nm (Figure 3b ii, iii and Figure S4). This shrinkage occurred regardless of BIS content but became more pronounced at higher AAm concentrations. Despite the reduced nanosheet distance, the as‐prepared gel retained all the water from the precursor, indicating structural inhomogeneity of the AAm gel. Indeed, the AAm gel with shrunk cofacial TiNSs exhibited notably lower transparency compared to the DMA gel with unshrunk cofacial TiNSs (Figure S5).

Interestingly, even without BIS, AAm produced an elastic gel that maintained its shape without a container (Figure 3d i), suggesting that TiNS served as a physical crosslinker through multivalent interactions with poly(AAm) [59]. Furthermore, when the as‐prepared gel with [BIS]/[AAm] lower than 10⁻³ was annealed, they gradually shrank in the vertical direction to TiNS while expelling water (Figure 3d ii). Shrinkage reached equilibrium after ∼500 h, reducing the gel length to 65–75% of its original value. Given [BIS]/[AAm] was lower than 10⁻³, this shrinkage occurred regardless of TiNS content (Figure S6). Overall, TiNS‐gels spontaneously densified when the network was sufficiently adsorptive to TiNS and less crosslinked. As discussed later, this shrinkage provided a clue to address the limitation in the nanosheet concentration of TiNS‐gel.

2.3. Mechanism of the Changes in Nanosheet Distance and Gel's Whole Length Caused by Adsorptive and Flexible Gel Network

As for the mechanism underlying the unusual shrinkage of both TiNS distance and macroscopic gel length, three key steps are hypothesized (Figure 2b, Steps 1–3).

• Step 1: Hydrogen‐bonding interaction of the monomer to TiNS. On the surface TiNS, abundant of Ti‐bridging O atoms (Ti─O─Ti) are exposed [64, 65], which are known to serve as hydrogen‐bonding acceptors. Therefore, monomer molecules with hydrogen‐bonding donor sites can interact with the TiNS surface via hydrogen‐bonding interaction. This was supported by the fact that shrinkage was suppressed when the NH group of the monomer was blocked by a CH₃ group, i.e., by replacing from AAm to DMA (Figure 3). ¹H NMR spectra further confirmed this interaction (Figure S7) [59]: upon adding TiNS to AAm in water, both NH and vinyl peaks broaden, whereas in DMA, which lacked NH groups, the vinyl peaks were unaffected by the addition of TiNS.

• Step 2: Bridging between cofacial nanosheets by multivalent polymer chains during gelation [73, 74, 75]. Through the gelation, the monomer was converted into the polymer that can adhere to the TiNS surface more strongly via the multivalent hydrogen‐bonding interaction with Ti‐bridging O atoms on the TiNS surface. Evidence for this mechanism was the pronounced decrease in TiNS distance during in situ polymerization of AAm, despite the overall gel volume remaining constant (Figure 3b). The role of multivalency was further confirmed by comparing AAm with pre‐synthesized poly(AAm). When AAm (2.5 wt%) was mixed with TiNS (0.8 wt%), the dispersion remained unchanged (Figure S8b). However, adding multivalent poly(AAm) to a TiNS dispersion induced efficient chain bridging, leading to partial gelation at 0.5 wt% and complete gelation at 2.5 wt%. Even at 0.1 wt%, poly(AAm) dramatically increased the viscosity of TiNS dispersions, underscoring the strong bridging interaction (Figure S9b). In contrast, poly(DMA), which lacked hydrogen‐bond donor groups, did not induce these effects (Figures S8a and S9a).

• Step 3: Further adsorption and reconfiguration of polymer chains after gelation. When the gel network was both adsorptive and flexible, the resulting gel gradually shrank (Figure 3d), suggesting that the as‐prepared gel was a kinetically trapped metastable state that tended to transform into a thermodynamically stable state [76, 77]. In situ polymerization was essential for this trapping. A reference gel prepared by directly mixing TiNS (1.2 wt%) with pre‐synthesized poly(AAm) (5.0 wt%) in water showed no shrinkage over time (Figure S10), despite having the same composition as the in situ polymerized gel (Figure 3d, [BIS]/[AAm] = 0). The origin of this instability raised from the poly(AAm) chains formed under low monomer concentration: they were relatively loose, and unnaturally stretched to bridge cofacial TiNSs, preventing full interaction with TiNS during gelation (Figure 2b, Step 2). To relieve this instability, poly(AAm) chains undergo further adsorption onto TiNS and reconfigure into more compact conformations, driving spontaneous gel shrinkage (Figure 2b, Step 3). Consistent with this mechanism, when AAm was polymerized at higher concentrations (>9 wt%; Figure S6), the resulting network achieved adsorption saturation and bridging at natural conformations, yielding a thermodynamically stable gel that no longer shrank, even at a low content of BIS.

To further validate the proposed mechanism, we prepared TiNS‐gels with flexible networks using various monomers containing different numbers of hydrogen‐bond donors. When the monomer possesses multiple hydrogen‐bond donors, they can cooperatively interact with the Ti bridging O atoms on the TiNS surface. Consistent with our hypothesis, greater numbers of hydrogen‐bond donors led to more pronounced shrinkage, both in TiNS distance during gelation and shrinkage in the macroscopic gel length after gelation (Figure S11a,b). We also observed the gel shrinkage when other inorganic nanosheets were used instead of TiNS (Figure S11c), indicating the generality of this phenomenon.

2.4. Time‐Induced Annealing and Force‐Induced Training of TiNS‐Gel With Adsorptive and Flexible Gel Network

We then performed a detailed characterization of TiNS‐gel after time‐induced annealing. When the as‐prepared gel with the composition of TiNS / AAm / BIS = 1.2 / 5.0/ 5.0 × 10⁻⁴ wt% was annealed for 500 h to equilibrium, it shrank to ∼71 ± 6% of its original length while expelling ∼29 ± 6% of water (Figure 4a). The expelled water contained no TiNS but contained polymer at ∼3.5 wt% (Figure S12), yielding final TiNS and polymer contents of ∼1.7 and ∼5.6 wt%, respectively. Increasing the initial nanosheet content in the gel enhanced the amount of adsorption polymer chains, leading to a reduction of the polymer content in the expelled water. Notably, TiNS‐gels with such a high TiNS content (∼1.7 wt%) cannot be synthesized directly because TiNS loses magnetic orientability at this concentration (Figure S3b).

FIGURE 4.

Preparation and characterization of the annealed and trained TiNS‐gels. The as‐prepared gel was made of TiNS (1.2 wt%) and a polymer network with adsorptive monomer (AAm, 5.0 wt%) and low crosslinking density ([BIS] = 5 × 10⁻⁴ wt%). (a,b) Preparation of the annealed (a) and trained (b) gels. Schematic representation (upper) and time‐course changes in gel length in the direction vertical to the alignment of TiNSs (lower). (c,d) SAXS profiles of the as‐prepared gel (c,d, upper) and the annealed gel (c, lower) and trained gel (d, lower). (i) 2D SAXS images. (ii) Intensity–q plots. (iii) Intensity–azimuthal angle plots.

The nanostructure of the annealed gel was examined by SAXS (Figure 4c). During annealing, the distance between cofacial TiNSs decreased from 58 to 35 nm (Figure 4c ii), suggesting that the macroscopic gel shrinkage and the nanoscale distance decrease were coupled. In fact, highly crosslinked TiNS‐gel ([BIS]/[AAm] = 10⁻²) showed no change in either macroscopic length or TiNS distance (Figure 3d and Figure S13). Remarkably, despite the drastic structural changes during annealing, the orientational order of TiNS was fully preserved (Figure 4c iii). It should be noted that 2D SAXS confirmed alignment only within the X‐ray beam region (∼100 µm), but polarized optical microscopy demonstrated that the orientation propagated throughout the entire gel (Figure S14a).

Time‐induced annealing was a reliable method to condense TiNS in gels up to ∼1.7 wt%, but costing long equilibration time (>500 h). Because this transition depended on the relaxation from a kinetically trapped metastable state to a thermodynamically stable state, external stimuli may accelerate it. However, heating the as‐prepared gel caused deformation and orientational disorder (Figure S15), while air‐drying led to disorder and structural inhomogeneity (Figure S16). In contrast, compression applied vertically to the TiNS alignment, denoted as force‐induced training, accelerated relaxation and shrinkage by expelling water. When a stress was applied and the resulting strain was kept constant, the stress gradually decreased as the polymer network underwent relaxation to dissipate the stress within a short period (Figure S17). The compression vertical to the nanosheet likely deforms the metastable polymer network, bringing it closer to the most stable structure in terms of distance between cofacial nanosheets and the mode of polymer–nanosheet interaction. Notably, time‐induced annealing and force‐induced training appear to be analogous phenomena, as the variation of monomer and crosslinker contents in the gels yielded synchronous changes in both annealing kinetics and training efficiency (Figures S6 and S18).

By repeating the stress relaxation cycle in an appropriate number, TiNS‐gel can reach to the state similar to the equilibrium of the time‐induced annealing. This process becomes more efficient as the applied stress is larger. However, to prevent damage to the polymer network caused by polymer chain scission, the stress should be lower than the threshold. Considering these factors, we optimized the stress intensity and the number of cycles (Figures S17 and S18). A typical training cycle involved (i) compressing the gel under 5 kPa stress and (ii) holding the strain for 1000 s to relaxation (Figure 4b). After ∼40 cycles, the gel length stabilized at ∼70 ± 5% of its initial value, comparable to the equilibrium state reached by annealing. Based on the volume of expelled water and polymer concentration in it (Figure S12), TiNS and polymer contents in the trained gel were estimated to be ∼1.7 and ∼5.6 wt%, respectively. SAXS results confirmed that TiNS distance decreased from 63 to 41 nm (Figure 4d ii), while the orientational order of TiNS was preserved or even enhanced (Figure 4d). Polarized optical microscopy further demonstrated that the highly oriented structure was maintained across macroscopic scales (Figure S14b). Overall, force‐induced training can produce a gel similar to that obtained by time‐induced annealing (∼500 h) much faster (∼10 h).

2.5. Enhancement of TiNS‐Gel's Mechanical Anisotropy and Asymmetry by Annealing and Training

Having established force‐induced training as an efficient densification method, we next investigated its effect on the mechanical anisotropy of TiNS‐gel (Figure 5a). Although various TiNS, AAm, and BIS contents can be used, we selected a representative composition of TiNS/ AAm/BIS = 1.2/5.0/5.0 × 10⁻⁴ wt%, based on our systematic studies of composition–property relationships (Figures S19 and S20). After the force‐induced training, the mechanical anisotropy factor (E _⊥/E _//) increased dramatically to 387 (Figure 5b), as densification reduced the distance between cofacial TiNSs and amplified electrostatic and entropic repulsions. Time‐induced annealing also enhanced anisotropy (E _⊥/E _// = 266), but to a lesser degree (Figure S21). Notably, force‐induced training specifically increased the vertical modulus (E _⊥) from ∼314 to ∼1200 kPa (Figure 5b,c, blue), while maintaining the parallel modulus (E _//) at ∼3 kPa (Figure 5b,c, magenta). This preservation of softness reflected the low TiNS content even after densification (∼1.7 wt%) and retained nanosheet alignment. Noteworthy, E _// after training remained comparable to that of a gel without TiNS (Figure S22). By applying the same training method to as‐prepared gels of various compositions, we obtained a series of gels that consistently exhibited excellent mechanical anisotropy of E _⊥/E _// = ∼10² while retaining intrinsic softness with E _// of 1–10 kPa (Table S2). Overall, densification of TiNS‐gels, a typical type of repulsion‐type hydrogels (Figure 1a ii), provides a versatile strategy to simultaneously achieve high mechanical anisotropy and intrinsic softness with elastic moduli of 1–10 kPa.

FIGURE 5.

(a) Schematics for the anisotropic compression response of TiNS‐gel with horizontal TiNSs prepared from TiNS (1.2 wt%), AAm (5.0 wt%), and BIS (5 × 10⁻⁴ wt%). (b) Strain–stress curves upon compression in the parallel (magenta) and vertical (blue) directions of the as‐prepared gel (b, left) and the trained gel (b, right). (c) Magnification of the curves in (b). The elastic moduli were calculated with a strain range of around 5%. (d) Schematic for the asymmetric shear response of TiNS‐gel with tilted TiNSs prepared from TiNS (1.2 wt%), AAm (5.0 wt%), and BIS (5 × 10⁻³ wt%). (e,f) Strain–stress curves upon shear to the left (magenta) and right (blue) directions of the as‐prepared gel (e, left) and the trained gel (e, right), and their magnifications (f). The elastic moduli in the left shear were calculated with the strain range from 5 to 20%, while those in the right shear were calculated with the strain range from 0 to <10%.

Recently, we developed a gel that exhibits asymmetric elastic moduli (E _R ≠ E _L) in response to shear forces [78], which was obtained by embedding TiNS in a tilted manner into a gel matrix (Figure 5d). Such mechanically asymmetric gels have attracted significant attention for their ability to induce unidirectional energy transport [79, 80, 81]. Considering that shear forces can be decomposed into compressive and tensile components along the diagonal of the gel, the mechanism for the asymmetric gel to exhibit mechanical asymmetry is the same as the anisotropic gel (Figure 5d) [78]. Therefore, the force‐induced training should also be able to enhance mechanical asymmetry. In fact, the mechanical polarity (E _R / E _L) of the as‐prepared gel was 19, while that of the trained gel was improved to 27 (Figure 5e). Again, the force‐induced training specifically increased the right‐shear modulus (E _R), with maintaining the left‐shear modulus (E _L) low (Figure 5f).

2.6. Application of Trained TiNS‐Gel for Vibration Insulation and Rectification

Finally, the trained TiNS‐gel, realizing enhanced mechanical anisotropy and well‐preserved softness with an elastic modulus of <10 kPa, was applied to vibration insulation (Figure 6a). In the test, the upper face of a cylinder‐shaped TiNS‐gel with horizontal TiNSs was fixed to a lateral force sensor, while the lower face was attached to a vibrator oscillating horizontally. The frequency was varied from 0.1 to 10 Hz with a fixed strain amplitude of ± 20%, and results were compared with the as‐prepared gel (Figure 6a). In this configuration, higher anisotropy and low elastic modulus correspond to better insulation efficiency, reflected by a smaller lateral force amplitude (ΔF). Indeed, the trained gel exhibited at least twice the insulation efficiency of the as‐prepared gel across the entire frequency range, reaching a maximum of ∼4.5‐fold improvement at 0.1 Hz.

FIGURE 6.

(a) Vibration insulation profiles of the as‐prepared gel (navy) and the trained gel (red) with horizontal TiNSs. Vibration (amplitude = ±20% strain, frequency = 0.1 to 10 Hz) was applied from the bottom of the gel, and lateral force (F) was measured at its top. The insulation efficiency is estimated by how small ΔF is. (b) Vibration rectification of the as‐prepared gel (navy) and the trained gel (red) containing TiNSs with tilted alignment. Vibration (amplitude = ±20% strain, frequency = 1 Hz) was applied from the bottom of the gel, and lateral force (F) was measured at its top. The rectification efficiency is estimated by how large ΔF _L/ΔF _R is.

We also applied TiNS‐gel with tilted TiNSs to vibration rectification (Figure 6b), evaluating the ratio of lateral force amplitudes to the right (ΔF _R) and left (ΔF _L) as the rectification degree (ΔF _R/ΔF _L). The as‐prepared gel showed a rectification degree of 2.5, whereas the trained gel improved to 4.2. These results demonstrated that force‐induced training expanded the potential of mechanically anisotropic and polar gels for both vibration damping and directional control.

3. Conclusion

We developed a series of hydrogels that simultaneously achieve outstanding mechanical anisotropy and softness, with anisotropy factors of ∼10² and intrinsic elastic moduli of 1–10 kPa. In previous reports, hydrogels with anisotropy factors exceeding 10² were limited to modulus values above 100 kPa (Figure 1b, blue zone), whereas this study fills the previously unexplored regime (Figure 1b, red zone). The anisotropy factor of 387 in this work (Figure 5a) would be a new benchmark for soft hydrogels. The key to this achievement lay in the use of repulsion‐type gels (Figure 1a, right), where the low contents of nanosheet (<2 wt%) and polymer (<10 wt.%) preserved intrinsic softness with elastic moduli of 1–10 kPa. The content of nanosheet dopable into such gels was generally restricted by poor orientability above the threshold concentration, while we developed a practical method to condense nanosheets within the hydrogel while fully preserving their orientation (Figure 4), thereby dramatically enhancing anisotropy (Figure 5).

From an application perspective, our hydrogels span the full modulus range of soft biotissues, from brain to muscle (1–10 kPa), enabling potential biomedical uses. Their unique mechanical properties were also advantageous for vibration engineering (Figure 6). From a fundamental standpoint, this work highlights an important caution in polymer composite science. Traditionally, internal nanostructures of polymer–nanosheet composites were considered stable after formation. In contrast, our hydrogels underwent drastic structural evolution spontaneously after preparation (Figures 3 and 4). We also confirmed that similar structural evolution occurs for various polymers and nanosheets (Figure S11). Similar overlooked phenomena may occur in other polymer composites due to structural complexity. This finding opens a new concept for designing next‐generation polymer composites.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll74363‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.