Synergizing network topology and solvent compatibility for gels with hyperelasticity and omniadhesion

Abstract

In polymeric gels, hyperelasticity and strong adhesion are often required but difficult to achieve simultaneously. Here we propose a principle of hyperelastic and omniadhesive gels composed of polymer networks with long dangling chains and sufficient good solvents. The molecules of good solvents screen off interchain interactions for hyperelasticity. The long dangling chains disentangle and adsorb to substrate for strong adhesion. We synthesized such gels by controlling the polymerization kinetics. When a monomer solution is partially cured, some monomers form a network and others form the solvent. The resulting gel, termed homogel, consists of a polymer network with enormous dangling chains and solvent of identical chemistries. An interval exists where the dangling chains are long and disentangled substantially, and the homogel exhibits both hyperelasticity and omniadhesion. We demonstrated such a gel with a hysteresis of 4.1% (stretch = 10) and adhesion energy of ~510 joules per square meter. The principle is generic and applicable to gels of different types.

A general principle of synergizing network topology and solvent compatibility is proposed for hyperelastic and omniadhesive gels.

INTRODUCTION

Polymeric gels are three-dimensional polymer networks swollen with solvents. The inclusion of solvents sets gels apart from elastomers and plastics by combining the advantages of solids (e.g., load-bearing capability) and liquids (e.g., mass transportation), enabling a wide range of applications such as tissue engineering (1), bioadhesives (2, 3), energy storage (4), stretchable ionotronics (5, 6), and stretchable human-machine interfaces (7, 8). In practical applications, gels need to be hyperelastic (i.e., to exhibit low hysteresis) and form reliable adhesion to various substrates for the sake of resisting fatigue loads (9–11) and interfacial cracks (2, 12), respectively. However, achieving both hyperelasticity and strong adhesion is inherently challenging (13, 14).

Typically, a hyperelastic polymer network minimizes interchain sliding or entanglements, whereas tough adhesion relies on significant energy dissipation through friction between polymer chains. Efforts have been made to balance hyperelasticity and adhesiveness in polymer networks but with trade-offs (13–15). In addition, forming tough adhesion often requires pretreatments or specific chemical modifications of the surfaces (12, 16). Achieving autonomous strong adhesion on different substrates would greatly enhance the flexibility and ease of implementation (15).

In this work, we propose a general strategy to synthesize hyperelastic and omniadhesive polymeric gels (Fig. 1A). The polymer network of such a gel contains long dangling chains and is infiltrated with sufficient molecules of a good solvent. Within the polymer matrix, the principle of “like plasticizes and lubricates like” prevails that the solvent molecules effectively separate polymer chains, reducing interchain friction. With an adequate amount of solvents, the polymer network is free from dangling chain entanglements, and the gel exhibits hyperelasticity (Fig. 1A, left). On the surface, the solvent molecules aid in the disentanglement of long dangling chains, allowing them to anchor to various substrates. This anchoring enables effective load transfer into the matrix to elicit energy dissipation for strong adhesion (Fig. 1A, right).

Fig. 1. Principles of hyperelastic and omniadhesive gels.

(A) Schematic of a hyperelastic and omniadhesive gel. The gel is hyperelastic because of the elimination of dangling chain entanglements in the bulk (left) and omniadhesive because of the disentanglements of long dangling chains on the surface (right). (B) Phase diagram of the polymerization of a monomer solution. During the polymerization, storage modulus and loss modulus increase over time, intersect at t_gelation, and plateau at t_plateau. The polymer content increases at the sacrifice of solvent. A homogel forms at t_gelation < t < t_plateau. An interval, t_HO, exists where the homogel exhibits both hyperelasticity and omniadhesion.

Essential to the formation of hyperelastic and omniadhesive gels is network topology, which should be rich in long dangling chains, and the presence of sufficient good solvents. There are many synthesis strategies to tailor the network topology of gels, using, e.g., slide-rings (17), dense entanglements (9), click chemistry (18), and phase separation (19). In stark contrast, pairing a given polymer network with a good solvent is challenging because of their often-disparate chemical structures, whereas their compatibility profoundly affects the mechanical properties of gels (20). Predictions based on the solubility theory frequently deviate from experimental results (20), and experimental compatibility assessments are inefficient (21). Moreover, elaborately formulated good solvents can be vulnerable as minor composition variations may induce substantial changes in compatibility (22).

Herein, we use a facile one-step method to synthesize gels containing abundant dangling chains and solvents that naturally share identical chemistries with the polymers by controlling the kinetics of polymerization (Fig. 1B). In a monomer solution, the storage modulus (G′) is lower than the loss modulus (G″) at the uncured liquid state. During curing, monomers link to form polymer chains, and the polymer chains cross-link into a polymer network. Throughout this process, G′ and G″ increase over time and intersect at the gelation point, t_gelation, which signifies the formation of a nascent polymer network. Eventually, both G′ and G″ plateau upon reaching a fully cured state at t_plateau, which indicates the transformation into a viscoelastic elastomer/plastic with immense entanglements.

The intermediate state is partially cured, where some monomers form a connected network, while others are left as the solvent. The polymer network and the solvent naturally have like chemistries, and the resulting gel is termed homogel. The polymer network contains numerous dangling chains, and the solvent molecules reduce the friction between polymer chains. Within the context of homogel, an interval, t_HO, exists, where the dangling chains are long, yet dangling chain entanglements disappear. Homogels within this interval exhibit both hyperelasticity and omniadhesion. Below t_HO, the homogel is sparsely cross-linked and mechanically weak. Above t_HO, the homogel contains entanglements of dangling chains and exhibits viscoelasticity (23).

RESULTS

Synthesis of hyperelastic and omniadhesive homogels

The principle of hyperelastic and omniadhesive gels is independent of specific chemistries. Here, we illustrate the principle using an ionically conductive homogel, which contains electrostatic interactions between polymer chains (11). We synthesize an ionic monomer, 2-(acryloyloxy)-N, N, N-trimethylethan-1-aminium bis((trifluoromethyl)sulfonyl)amide, abbreviated as DT hereafter, with a viscosity of 0.19 ± 0.02 Pa·s (fig. S1). The composition and chemical structures are confirmed by the ¹H-nuclear magnetic resonance spectrum (fig. S2). We formulate a precursor with DT as the monomer, 1,6-hexanediol diacrylate (HDDA) as the covalent cross-linker, and 2-hydroxy-2-methylpropiophenone (1173) as the photoinitiator (Fig. 2A).

Fig. 2. Synthesis and mechanical characterizations of hyperelastic and omniadhesive PDT homogels.

(A) Chemical structures of the constituents. The light gray, dark gray, blue, yellow, green, and red spheres represent hydrogen, carbon, nitrogen, sulfur, fluorine, and oxygen atoms, respectively. (B) Dissipation factor varies with curing time. Cross-linker content: 0.1 mol %. Inset, the variations of G′ and G″ with curing time. (C) Polymer content (w_p) and solvent content (w_s) vary with curing time. The black dashed line indicates the gelation time, and the red indicates the synthesis time for hyperelastic and omniadhesive PDT homogels. (D) Representative stress-stretch curves. (E) Consecutive cyclic loading and unloading curves with increasing stretches. Cross-linker content: 0.05 mol %. Inset, the curves of PDT elastomer with a cross-linker content of 0.05 mol %. (F) Cyclic loading and unloading curves with a cross-linker content of 0.1 mol % under different stretch rates. (G) Ashby plot of hysteresis versus stretch, compared to ionogels (11, 26, 27), hydrogels (9, 10, 17), and polydimethylsiloxane (PDMS) (14). The brown arrow means the stretch is >17 (11). (H) 90°-peel force/width-displacement curves on various substrates. Cross-linker content: 0.05 mol %. The inset shows the peel of a sample, dyed with methylene blue, from a PMMA substrate. (I) Ashby plot of hysteresis versus adhesion energy, compared to ionogels (27, 35–39), hydrogels (40–48), and elastomers (49). The data of PAAm hydrogels and VHB 4905 are obtained through our measurements. Hysteresis is measured when the PDT homogel (cross-linker content: 0.05 mol %) is stretched near rupture. The error bars in (C), (G), and (I) represent SDs. n = 3.

We synthesize the homogel using one-pot free radical polymerization. We measure the gelation curves for precursors with various cross-linker contents. After gelation at ~764 s, the dissipation factor (tan δ) minimizes at ~866 s (Fig. 2B and fig. S3). The interval for hyperelastic and omniadhesive homogels is determined near this point (based on hysteresis and adhesion tests elucidated in following sections), with t_HO ranging from ~860 to ~920 s. Note that t_gelation is determined on the basis of the Winter criterion G′ = G″ (24), which only captures the relaxation modes faster than the gelation process.

The resulting poly(2-(acryloyloxy)-N, N, N-trimethylethan-1-aminium bis((trifluoromethyl)sulfonyl)amide) (PDT) homogels are optically transparent (fig. S4) and nonvolatile (fig. S5A). The cross-linker content has minor influences on t_gelation and t_HO (fig. S6), which could be due to the effective cross-links induced by topological entanglements (9, 25). The corresponding solvent content (w_s) is almost invariant with cross-linker content (fig. S7). The solvent content decreases while the polymer content (w_p) increases with curing time, both plateauing at ~1260 s (Fig. 2C). We set the synthesis time for hyperelastic and omniadhesive PDT homogels at t_synthesis = 900 s with w_s = 38 wt %. The average molecular weight of the solvent molecules, determined by gel permeation chromatography (GPC), is 865 g/mol, corresponding to an average degree of polymerization of ~2 (fig. S8).

The determination of the interval for hyperelastic and omniadhesive homogels might depend on specific deployment scenarios, the requirements of applications, and the chemistry of the material systems. Nevertheless, the mechanical properties generally deteriorate with a shorter polymerization time; for instance, the PDT homogels formed at 780 s are squishy. On the other hand, viscoelasticity prevails at a longer polymerization time. The PDT homogels formed at 960 s exhibit a pronounced hysteresis of 38.7% (fig. S9).

Performance of hyperelastic and omniadhesive homogels

We then characterize the mechanical properties of PDT homogels, which are hyperelastic and omniadhesive unless otherwise specified. The homogels become stiffer and less stretchable as the cross-linker content increases (Fig. 2D and fig. S10). Subject to consecutive cyclic tests with increasing stretches at 0.05 s⁻¹, the homogels exhibit negligible hysteresis (Fig. 2E and fig. S11), while the PDT elastomer exhibits prominent hysteresis (Fig. 2E, inset). The homogels also exhibit negligible hysteresis under cyclic tests with various maximum stretches (fig. S5B). Subject to a stretch, the PDT homogel can fully recover from deformation, whereas the PDT elastomer cannot (movie S1). The homogels maintain low hysteresis at stretch rates over two orders of magnitude (Fig. 2F and fig. S12).

We collect the data of hysteresis and stretch of various polymeric materials and compare them (Fig. 2G). The hysteresis of PDT homogels is <10% over a wide stretch range, which is much lower than that of PDT elastomers and comparable to those of the widely used polyacrylamide (PAAm) hydrogels (fig. S13) and polydimethylsiloxane elastomers (14), and other recently developed hyperelastic gels (9–11, 17, 26, 27). The homogels exhibit excellent resistance to fatigue damage under unnotched cyclic loading, as evidenced by the stability of their stress-stretch curves over thousands of cycles (fig. S5C).

The PDT homogels exhibit autonomous and strong adhesion to various substrates (fig. S14). We perform 90°-peel tests on surfaces of various materials, including ceramic, glass, metal, elastomer, plastic, and wood (Fig. 2H). When the sample is backed with an inextensible film, the steady-state peel force, F_ss, gives the adhesion energy by Γ₉₀ = F_ss/B, where B is the width of the samples. The adhesion energy of the homogel with a cross-linker content of 0.05 mol % is on the order of 100 J/m² on all substrates, significantly higher than that of the corresponding PDT elastomer (figs. S15A and S16). The adhesion performance remains stable for more than 200 cycles (fig. S15B). The adhesion energy is comparable to that of a commercial adhesive (VHB 4905, 3M) and much higher than that of the ionic PAAm hydrogel (fig. S17, A and B). Depending on the specific application requirements, the adhesion energy can be further improved by, e.g., optimizing the chemical compositions (13). Because of the noncovalent nature of interfacial bonding, the adhesion energy increases with loading velocity (fig. S17, C and D) (28).

Strong adhesion is further validated by 180°-peel and lap-shear tests (figs. S18 and S19). Notably, the PDT homogels simultaneously have two seemingly incompatible properties: high adhesion energy and low hysteresis (Fig. 2I), whereas existing hyperelastic gels are deficient in adhesion. In addition, the PDT homogels are conductive, with a conductivity on the order of 10⁻² S/m (fig. S20). Hyperelasticity gives rise to a stable resistance more than 10,000 cycles (fig. S21). Such hyperelastic and omniadhesive soft conductors are ideal materials of choice for emerging applications such as flexible electronics and stretchable ionotronics (movie S2).

Characterizations of network structures

We design experiments to characterize network structures and probe the underlying mechanisms of the unusual combination of hyperelasticity and omniadhesion. Since the network topology, solvent content, and molecular weight all keep changing during polymerization, we prepare homogels with a cross-linker content of 0.1 mol %, remove their solvents, and then reswell them in monomer solution to various solvent contents. In this way, the physical and chemical properties of solvents are consistent, and the solvent content is controllable. Moreover, we control network structures by terminating the polymerization at 900 s, 960 s, and 4 hours, obtaining three network topologies, i.e., partially cured hyperelastic and omniadhesive (PC_HO) network, partially cured (PC) network, and fully cured (FC) network.

The initial solvent contents of the three network topologies are 38, 18, and 0 wt %. After removing the initial solvents and reswelling the networks in monomer solution to equilibrium, the solvent contents are ~90 wt % (PC_HO network), ~85 wt % (PC network), and ~83 wt % (FC network) (fig. S22, A to C), respectively. The corresponding average chain lengths are 19.25 nm (PC_HO network), 9.99 nm (PC network), and 7.86 nm (FC network) (text S1), indicating that the average chain length decreases with the extent of polymerization (fig. S22D).

We perform uniaxial tension and rheological tests at temperatures above the glass transition temperature (T_g) of the FC network (fig. S23). Analyzing the tensile data according to the Rubinstein-Panyukov relation reveals that the PC_HO network has the longest strands between cross-links and the highest portion of entanglements (fig. S24 and text S2) (29, 30). Constructing the master curves through the time-temperature superposition treatment of rheological data and invoking the Curro-Pincus method indicate that the PC_HO network has the longest dangling chains (fig. S25 and text S3) (31, 32). The tendencies are consistent with molecular dynamics simulations (fig. S24 and table S1).

Mechanisms of hyperelasticity

To study the effects of solvent on hyperelasticity, we swell the PC_HO networks in monomer solution to various solvent contents. The identical chemistry of the polymer network and solvent renders excellent compatibility, with a Flory-Huggins interaction parameter χ < 0.5 (fig. S26 and text S4) and a negative mixing enthalpy (fig. S27 and text S5). After swelling to a prescribed solvent content, the samples are relaxed for 120 hours to ensure equilibrium before further tests (fig. S28). Hysteresis decreases monotonically by 30 wt % and becomes constantly <10% afterward (Fig. 3A and fig. S29A). Simulations also show that the mean distance between polymer chains increases while hysteresis decreases with w_s (fig. S30). The T_g exhibits a similar trend (inset of Fig. 3A and fig. S29B). A transition occurs at w_s = 30 wt %, where the scaling of T_g with w_s deviates. This transition is caused by the suppression of dangling chain entanglements.

Fig. 3. Mechanisms of hyperelasticity and omniadhesion.

(A) Hysteresis of the PC_HO network varies with w_s. The inset shows the variation of T_g with w_s, and the dashed lines are linear fittings. (B) Shear modulus (μ) of the PC_HO network varies with φ_p. The inset illustrates the change of network topology, from an entangled one to an unentangled one, after swelling to φ_p ≤ 71% (i.e., w_s ≥ 30 wt %). (C) Storage modulus (G′) and loss modulus (G″) of PC_HO networks with different solvent contents as a function of angular frequency at 60°C. The slopes are obtained via linearly fitting the data at angular frequency > 1 rad/s. (D) The variations of hysteresis with different solvent contents. (E) Adhesion energy, measured via 90°-peel tests on PMMA substrates, as a function of relaxation time with w_s = 30 wt %. The solid lines are fitted by $\mathrm{\Gamma }={\mathrm{\Gamma }}_p(1-e^{-t/\mathrm{\tau }})$, where ${\mathrm{\Gamma }}_p$ is the plateau and $\mathrm{\tau }$ is the characteristic relaxation time. (F) The scaling exponent α as a function of relaxation time. The inset schematizes the relaxation of dangling chains. The error bars in (D) to (F) represent SDs. n = 3. (G) Schematics showing structural evolutions with w_s for networks with different topologies. h, hours.

To validate the hypothesis macroscopically, we plot shear modulus (μ) as a function of polymer volume fraction (φ_p) (Fig. 3B) and obtain a transition at φ_p = 71% (corresponding to w_p = 70 wt %). When φ_p < 71%, μ scales with φ_p exponentially with a power of 0.56, which agrees well with the prediction of affine swelling model (text S6) (25) and suggests insignificant dangling chain entanglements (Fig. 3B, inset). When φ_p > 71%, the power index increases drastically to 4.71 because of the presence of dangling chain entanglements (33). Therefore, hyperelasticity is attributed to the elimination of dangling chain entanglements with enough good solvents.

We then validate the hypothesis microscopically. The PC_HO network with a solvent content exceeding 30 wt % exhibits a frequency-independent storage modulus, varying by less than 20% over the range of 0.1 and 10 rad/s (Fig. 3C). The loss modulus follows a power-law frequency dependence: $G^{″}\sim {\mathrm{\omega }}^{\mathrm{\alpha }}$, where the exponent α correlates with mesh size and the effective length of dangling chains (31, 32), and is inversely proportional to the average number of constraints per dangling chain. α increases with w_s and then plateaus at w_s > 30 wt %. The tan δ decreases and becomes less dependent on frequency as w_s increases (fig. S31). When w_s ≥ 30 wt %, the Mooney stress becomes mostly constant (fig. S32A and text S7) (28). The ratio of the Mooney-Rivlin coefficients C₂/C₁ displays a two-stage dependence on w_s and approaches zero when w_s ≥ 30 wt % (fig. S32B and text S7) (33). All these results, again, suggest the disappearance of dangling chain entanglements at w_s ≥ 30 wt %.

Moreover, we swell the PC networks and FC networks in monomer solutions to various solvent contents, relax them for 120 hours (fig. S33), and measure their hysteresis (fig. S34). Figure 3D summarizes hysteresis of different network topologies and different solvent contents. In general, hysteresis decreases with w_s and becomes constantly <10% at w_s ≥ 30 wt %, regardless of network topology. This is because the polymer chains are separated by the same relative distance with the same solvent content, which also explains the independence of the critical w_s on cross-linker content in fig. S7. At w_s < 30 wt %, dangling chain entanglements cause prominent hysteresis. The abnormally high hysteresis at w_s = 15 wt % of the FC network might be due to residual glassy domains since the T_g is above zero (fig. S35).

Mechanisms of omniadhesion

The marked onmiadhesion is attributed to the long dangling chains on the surface. The dangling chains, after disentangling sufficiently from the network, can adsorb to various substrates, yielding strong interlinks for strong adhesion. We swell the three networks to w_s = 30 wt %, relax them for a period of time, and then measure their adhesion energies against poly (methyl methacrylate) (PMMA) substrates. All adhesion energies increase with relaxation time (Fig. 3E and fig. S36). Specifically, the PC_HO network exhibits the highest adhesion energy owing to the longest dangling chains. In addition, we swell the PC_HO network to different solvent contents and the adhesion energy also increases with relaxation time (fig. S37).

Microscopically, we probe the variations of loss modulus with frequency for the PC_HO network at different relaxation times using the Curro-Pincus method (31, 32). The increase of α with relaxation time manifests the gradual disentanglement of dangling chains (Fig. 3F and fig. S38). Besides, the elastic moduli of the three networks with a solvent content of 30 wt % are comparable, whereas the adhesion energy with PC_HO network is significantly higher than that with FC network (fig. S39A). Hence, the disentangled long dangling chains are crucial for strong adhesion.

We further measure adhesion energy of the three networks after swelling to various solvent contents. In rubbery state, the adhesion energy decreases with w_s because of the dilution of polymer chains (fig. S39B). Note that the number density of dangling chains is comparable to the strands between cross-links for one-pot synthesis (29). In this sense, the PC_HO network is expected to have the lowest number density of dangling chains among the three networks. Nevertheless, it has the longest dangling chains, which prevail such that it exhibits the highest adhesion energy (fig. S40). Thus, we conclude that a network with long dangling chains exhibits low hysteresis and high omniadhesion simultaneously above a critical content of good solvent, where the network is free from dangling chain entanglements in the bulk and rich in disentangled long dangling chains on the surface (Fig. 3G).

Generality of hyperelastic and omniadhesive gels

The principle of hyperelastic and omniadhesive gels is generic. We further illustrate the principle using polyanions and neutral polymers. For polyanions, we use 1-ethyl-3-methyl imidazolium (3-sulfopropyl) acrylate (ES) as the monomer (34), which exhibits a viscosity of 0.31 ± 0.03 Pa·s (fig. S1), a t_gelation of ~56 s, and a t_HO ranging from ~59 to ~61 s (Fig. 4A and fig. S41A). The narrow t_HO is due to the fast polymerization dynamics under the reaction conditions. The synthesis time for hyperelastic and omniadhesive poly(1-ethyl-3-methyl imidazolium (3-sulfopropyl) acrylate) (PES) homogels is ~60 s, and the solvent content is ~40 wt % (fig. S41B). Subject to consecutive cyclic tests with increasing stretches, the PES homogels exhibit negligible hysteresis while the corresponding PES elastomers are hysteretic (Fig. 4B). The adhesion energy of PES homogels is significantly improved compared to those of PES elastomers on various substrates (Fig. 4C and fig. S41, C and D).

Fig. 4. Implementation of the principle to other hyperelastic and omniadhesive homogels.

(A) The variations of G′ and G″ with curing time for PES. The insets show the chemical structures of ES and the variation of dissipation factor with curing time, with the minimum being attained at ~59 s. (B) Consecutive cyclic loading and unloading curves with increasing stretches of PES homogel and PES elastomer. (C) Adhesion energy of PES homogel and PES elastomer on various substrates. (D) The variations of G′ and G′ with curing time for PTMSPMA. The insets show the chemical structures of TMSPMA and the variation of dissipation factor with curing time, with the minimum being attained at ~2146 s. (E) Consecutive cyclic loading and unloading curves with increasing stretches of PTMSPMA homogel and PTMSPMA elastomer. (F) Adhesion energy of PTMSPMA homogel and PTMSPMA elastomer on various substrates. The error bars in (C) and (F) represent SDs. n = 3.

We use 3-(trimethoxysilyl)propyl methacrylate (TMSPMA) as the neutral monomer, with a viscosity of 0.07 ± 0.02 Pa·s (fig. S1), a t_gelation of ~977 s, and a t_HO ranging from ~2000 to ~2645 s (Fig. 4D and fig. S42A). The synthesis time for hyperelastic and omniadhesive poly(3-(trimethoxysilyl)propyl methacrylate) (PTMSPMA) homogel is ~2400 s, and the solvent content is ~28 wt % (fig. S42B). Likewise, PTMSPMA homogels exhibit lower hysteresis and higher adhesion energy compared to PTMSPMA elastomers (Fig. 4, E and F, and fig. S42, C and D).

We have synthesized different homogels using the same protocol through a facile one-step synthesis process, while different homogels have different critical intervals of t_HO for hyperelasticity and omniadhesion, and the associated solvent contents are different as well. These differences are ascribed to the specific chemistry of different homogels, which profoundly affects physical properties such as solvent viscosity and network topology. The interval t_HO can be tuned by modulating the kinetics of polymerization and the chemistry of material systems. In general, a slower kinetics of polymerization gives a wider range of t_HO. Moreover, the criteria for low hysteresis and strong adhesion vary for different applications, and so does the interval t_HO.

The principle of hyperelastic and omniadhesive gels can be applied other than homogels. By taking advantage of the versatile approaches of tailoring network topology, one can make a polymer network with long dangling chains and then swell it with sufficient good solvents. For example, we remove the solvent of hyperelastic and omniadhesive PES homogels to obtain a hydrophilic PC_HO PES network and then reswell the network in water to a solvent content of 27 wt %. The resulting PES hydrogels exhibit low hysteresis and high adhesion (fig. S43). As a counterexample, swelling a hydrophobic PC_HO PDT network in water results in a phase-separated hydrogel with low adhesion energy (fig. S44).

DISCUSSION

We propose a general principle of hyperelastic and omniadhesive gels. The formation of such gels dictates that the polymer network contains dangling chains and enough good solvents. We demonstrate the principle by synthesizing homogels with hyperelasticity and omniadhesion through a facile one-step method. Whereas a regular gel consists of a polymer network and a solvent with different chemistries, a homogel naturally consists of a polymer network and a solvent with identical chemistries. We have exemplified hyperelastic and omniadhesive gels using different types of polymers, including polycations, polyanions, and neutral polymers. They are stretchable, low-hysteretic, transparent, and omniadhesive. Gels of hyperelasticity and omniadhesion promise broad applications including fatigue-resistant optically clear adhesives, soft robotics, human-machine interfaces, and stretchable ionotronics.

MATERIALS AND METHODS

Materials

Acryloyloxyethyltrimethyl ammonium chloride (DAC), lithium bis(trifluoromethane)sulfonimide (LiTFSI), 1-ethyl-3-methylimidazolium chloride ([EMIM]Cl), 3-sulfopropyl acrylate potassium salt (K[SPA]), TMSPMA, methoxyphenol, dichloromethane (DCM), HDDA, 1173, acrylamide (AAm, A108465), α-ketoglutaric acid (K105571), and N, N′-methylenebis(acrylamide) (MBAA, M104022) were purchased from Aladdin. All chemicals were used as received without further purification.

Synthesis of DT

To synthesize DT, 1 mol of DAC with a water content of 20 wt % was uniformly mixed with 1 mol of LiTFSI. This reaction yielded the water-insoluble monomer DT and the water-soluble salt lithium chloride. After removing the upper water layer, anhydrous sodium sulfate was added to remove the remaining water. The mixture was then filtered to obtain the purified monomer DT.

Synthesis of ES

The monomer ES was synthesized according to the following procedure (34). One mole of [EMIM]Cl and 1 mol of K[SPA] were dissolved in 700 ml of acetonitrile. To prevent decomposition, 10 mg of methoxyphenol was added as an inhibitor. The mixture was stirred vigorously overnight at room temperature. The precipitated potassium chloride was removed by filtration, and the solvent was evaporated using a rotary evaporator. The resulting crude product was dried under vacuum (10 torr) to remove residual acetonitrile. The viscous yellow oil was then redissolved in fresh DCM and cooled to below 0°C overnight. Precipitated salts were again removed by filtration. After completely removing DCM using a rotary evaporator, the final product, a viscous yellow oil, was obtained.

Synthesis of PDT homogel and PDT elastomer

Various contents of cross-linker HDDA (1, 0.5, 0.1, 0.05, and 0.01 mol %) and photoinitiator 1173 (0.01 mol %) were mixed with DT. The molar ratios were in respect to the monomer unless otherwise specified. The precursor was injected into a mold made by sandwiching a 0.5-mm-thick poly(tetrafluoroethylene) (PTFE) spacer with two glass plates and irradiated with ultraviolet (UV) light (365 nm) for ~900 s to obtain PDT homogels with various properties. The precursor with a cross-linker content of 0.05 mol % was injected into the same glass mold and irradiated with UV light (365 nm) for 4 hours to obtain the PDT elastomer. The intensity of the UV light was 157 W/m².

Synthesis of three network topologies

The precursor with a cross-linker content of 0.1 mol % and a photoinitiator density of 0.01 mol % was irradiated for 900 s, 960 s, and 4 hours, respectively, to obtain hyperelastic and omniadhesive PDT homogels with an initial solvent content of ~38 wt %, partially cured homogels with an initial solvent content of ~18 wt %, and fully cured PDT elastomers with an initial solvent content of ~0.47 wt %. The initial solvents are removed by multiple swelling and deswelling process to obtain three network topologies: PC_HO network, PC network, and FC network.

Synthesis of PES homogel and PES elastomer

HDDA (0.05 mol %) and 0.01 mol % 1173 were mixed with ES to form a precursor. The precursor was injected into a glass mold and irradiated with UV light (365 nm) for 1 min and 4 hours to obtain hyperelastic and omniadhesive PES homogel and PES elastomer, respectively. The intensity of the UV light was 157 W/m².

Synthesis of PTMSPMA homogel and PTMSPMA elastomer

HDDA (0.01 mol %) and 0.05 mol % 1173 were mixed with TMSPMA and to form a precursor. The precursor was injected into a glass mold and irradiated with UV (365 nm) for 40 min and 4 hours to obtain hyperelastic and omniadhesive PTMSPMA homogel and PTMSPMA elastomer, respectively. The intensity of the UV light was 10 W/m².

Synthesis of PAAm hydrogel

A precursor was prepared by dissolving AAm (2.2 M), NaCl (2 M), MBAA (0.0006 the weight of AAm), and α-ketoglutaric acid (0.0017 the weight of AAm) in deionized water. The precursor was injected into a glass mold separated by a 0.5-mm-thick PTFE gasket and irradiated with UV (365 nm) for 1 hours. The intensity of the UV light was 157 W/m².

Synthesis of PES hydrogel

The solvents in hyperelastic and omniadhesive PES homogels were removed, and then the dry samples were swollen in water to a water content of 27 wt % to obtain the PES hydrogel samples.

Synthesis of PDT polymer for DSC tests

A total of 0.01 mol% 1173 was mixed with DT to form a precursor. The precursor was injected into a glass mold and irradiated with UV light (365 nm) for 4 hours to obtain dry PDT polymer.

Preparation of samples for GPC tests

The PDT homogel samples of 1 g were subjected to three cycles of swelling and deswelling in acetonitrile of 100 g. The acetonitrile containing the solvent molecules was collected and evaporate in a fume hood for 24 hours. To ensure complete removal of residual acetonitrile, deionized water was added to the residue. The mixture was frozen in liquid nitrogen and subsequently lyophilized for 2 days to obtain pure solvent molecules. The pure solvent molecules were then dissolved in tetrahydrofuran (THF) at a volume ratio of 1:1 and mixed thoroughly for 24 hours.

Gelation time measurement

The photorheological properties of PDT, PES, and PTMSPMA were measured in situ using a rheometer (TA DHR3) with a parallel plate geometry in a nitrogen atmosphere at room temperature. A UV light source (365 nm) was attached to the rheometer to illuminate UV light during curing. The intensity of the UV light was 142 W/m² for PDT and PES, and 23 W/m² for PTMSPMA.

Transmittance measurement

A 0.5-mm-thick PDT homogel with a cross-linker content of 0.1 mol % was attached to a polyethylene terephthalate (PET) film and fixed on the sample holder of a UV-vis spectrophotometer (Metasha, UV-8000). The wavelength was scanned from 350 to 820 nm. The transmittance of the PET film was measured for baseline calibration.

Rheology measurements

The viscosities of monomers DT, ES, and TMSPMA were measured by a rheometer (TA DHR-3) at room temperature. The oscillation mode was used with a frequency of 1 Hz and a shear strain of 0.1% at room temperature. The circular plate diameter was 20 mm, and the gap was set to 150 μm. The dissipation factor (tan δ) was measured in an oscillation mode with a shear strain of 0.1% at room temperature. The frequency range was 0.01 to 10 Hz. A circular plate with a diameter of 20 mm was used and the normal force was set as 0.1 N. The storage and loss moduli versus frequency curves were measured in the oscillation mode with a shear strain of 0.5% at different temperatures. A circular plate with a diameter of 20 mm was used, and the normal force was set as 0.1 N.

GPC measurements

The number-average molecular weight (M_n), weight-average molecular weight (M_w), and molecular weight distribution (i.e. M_w/M_n) of the polymers were detected by GPC (waters GPC, USA) using THF as an eluent at room temperature. The columns were calibrated against standard polystyrene samples.

Tensile tests

All specimens for tensile tests were cut into a dumbbell shape with a width of 2 mm and a gauge length of 10 mm using a laser cutter (Epilog, Fusion Pro 36). The samples were loaded onto a tensile machine (Instron 5966, 100 N load cell) and stretched at a loading velocity of 30 mm/min. For the measurements of hysteresis, the samples were loaded to prescribed stretches and then unloaded to the original position. Hysteresis was calculated as the area enclosed by the loading and unloading curves divided by the area underneath the loading curve. At least three samples were tested for every characterization. For tensile tests at different stretch rates, the samples were cut into a dumbbell shape with a width of 2 mm and a gauge length of 6 mm, and the loading velocities were 3, 15, 30, 150, and 300 mm/min, respectively. The above tests were all conducted at room temperature. For tensile tests at 70°C, the samples were cut into a dumbbell shape with a width of 2 mm and a gauge length of 6 mm, and the loading velocities were 30 mm/min. The samples were loaded onto a tensile machine (Instron 5966, 100 N load cell) with a thermostatic chamber. The elastic modulus E was obtained by linear fitting of the tensile data with strain less than 5%. Assuming incompressibility, shear modulus is calculated by μ = E/3.

Fatigue tests

Cyclic fatigue tests were performed on a fatigue testing machine (Instron E3000, 250 N load cell) using a displacement-controlled mode with a triangular loading profile at room temperature. The PDT homogels with 1 and 0.5 mol % of cross-linker contents were in dumbbell shape with a width of 2 mm and a gauge length of 10 mm. The PDT homogels with 0.1, 0.05, and 0.01 mol % of cross-linker contents were in dumbbell shape with a width of 2 mm and a gauge length of 6 mm. The loading velocity and the cycle number were 30 mm/min and 1000 for the PDT homogels with 0.01 mol % of cross-linker content, and 600 mm/min and 10,000 for the PDT homogels with 1, 0.5, 0.1, and 0.05 mol % of cross-linker contents.

Peeling tests

To measure adhesion energy, specimens were cut into a rectangular shape with a width of 5 mm and a length of 30 mm. For 90° peel, the top surface of the sample was glued to a layer of 25-μm-thick PET backing and the bottom surface was directly attached to a substrate of various materials (wood, Al₂O₃ ceramic, copper block, silica glass, PMMA, and acetonitrile elastomer). For 180° peel, two pieces of PDT homogels were glued to a 25-μm-thick PET backing on one surface and then attached to each other on another surface. The samples were all loaded onto a tensile machine (Instron 5966, 100 N load cell) and peeled at 30 mm/min at room temperature. During all peel tests, cohesive failure occurred for the PDT homogels with a cross-linker content of 1 mol % and adhesive failure occurred for others.

Lap shear tests

Samples were cut into square shapes with a side length of 20 mm. The samples were sandwiched between two pieces of 1-mm-thick acrylate sheets, which were fixed on a tensile machine (Instron 5966, 100 N load cell) and pulled at a velocity of 30 mm/min at room temperature.

Electrochemical impedance spectroscopy

The impedance tests were performed using an electrochemical workstation (CHI6000E, CH Instruments Inc.) at room temperature. The sizes of sample were 10 mm by 15 mm with a thickness of 0.5 mm. Both the top and bottom surfaces of the samples were directly adhered to indium tin oxide electrodes. The amplitude of the testing voltage was set at 0.5 V. The frequency range is from 20 Hz to 100 MHz.

Stability tests

The PDT homogels with a cross-linker content of 0.1 mol % were cut into a dumbbell shape with a width of 2 mm and a gauge length of 6 mm. The two ends of the samples were fixed on a mechanical testing machine, meanwhile connected to an inductance, capacitance and resistance (LCR) meter via copper sheets. The samples were stretched and released by the machine at a velocity of 600 mm/min for 10,000 cycles at room temperature. The resistance was recorded accordingly by the LCR meter.

Supplementary Materials

The PDF file includes:

Supplementary Methods

Supplementary Text

Figs. S1 to S44

Table S1

Legends for movies S1 and S2

References

Other Supplementary Material for this manuscript includes the following:

Movies S1 and S2