Internal catalysis significantly promotes the bond exchange of covalent adaptable polyurethane networks

Significance

There is a trade-off between the mobility and cross-linking structure of covalent adaptable networks (CANs), making it challenging to develop CANs with excellent mechanical properties and high self-healing efficiency. This work presents a pioneering approach by introducing internal catalysis into the cross-linking units to fabricate elastomers. The incorporation of neighboring groups significantly enhances the reversibility of dynamic bonds, resulting in a remarkable improvement in the dynamic behavior of the polymer network. By synergistically combining this strategy with the multiple-cross-linking-in-one site topological design, the resulting i-Canogel exhibits outstanding mechanical strength and self-healing properties at room temperature.

Abstract

Self-healing covalent adaptable networks (CANs) are not only of fundamental interest but also of practical importance for achieving carbon neutrality and sustainable development. However, there is a trade-off between the mobility and cross-linking structure of CANs, making it challenging to develop CANs with excellent mechanical properties and high self-healing efficiency. Here, we report the utilization of a highly dynamic four-arm cross-linking unit with an internally catalyzed oxime-urethane group to obtain CAN-based ionogel with both high self-healing efficiency (>92.1%) at room temperature and superior mechanical properties (tensile strength 4.55 MPa and toughness 13.49 MJ m⁻³). This work demonstrates the significant potential of utilizing the synergistic electronic, spatial, and topological effects as a design strategy for developing high-performance materials.

The sustainability of polymers has become a critical challenge for the polymer industry in the current century, driven by the principles of the circular economy (1). To address this issue, cross-linked polymers with dynamic covalent bonds, known as covalent adaptable networks (CANs), have emerged as a new type of sustainable polymers that combine the benefits of both thermosets and thermoplastics (2). Self-healing CANs provide extended lifespan, easier recyclability, and enhanced reliability, contributing to the attainment of carbon neutrality and sustainable development (3, 4).

However, most CANs require external stimuli such as light, heat, or pressure to initiate the self-healing process. The inconvenient process limits their applications. Thus, the development of CANs with spontaneously self-healing ability at room temperature is highly sought after. To achieve fast dynamic bond exchange reactions in CANs and allow for self-healing at low temperatures, various mechanisms are often employed to reduce the bond energy, including steric hindrance (5), coordination (6, 7), and other electronic effects (8, 9), external catalysts, such as acids (10), transition metals (11, 12), or nucleophilic catalysts (13). However, the aging of catalysts significantly reduces the lifespan of these materials. On the other hand, the presence of covalent cross-linking units enhances the rigidity and stability of CANs, while hindering the mobility and rearrangement ability of the network. In order to enhance self-healing capability, it is common to reduce the cross-linking density. However, this often comes at the expense of sacrificing the mechanical properties of the material. Therefore, approaches to simultaneously enhance the self-healing ability and mechanical properties of CANs are highly needed.

In this study, we designed four-arm dynamic units with internal catalysis as cross-linking points for CANs. The essence of internal catalysis or neighboring group participation (NGP) includes both electronic regulation, which refers to the influence of neighboring groups on the properties of adjacent groups through electron sharing or transfer (14, 15), and spatial effects, which involve the impact of the position of neighboring groups on their properties (16–18). Recently, NGP has been introduced to improve the recyclability (17, 19–21) and healing ability at high temperatures (22). Despite its great potential, the ability of NGP to enhance room temperature self-healing of CANs has not been realized. Sijbesma et al. reported polyamide dynamic covalent networks formed by four-arm cross-linking units to have excellent elasticity and high-temperature processing stability (23, 24). Herein, we designed a four-arm cross-linking unit that effectively integrated the electronic, spatial, and topological factors, synergistically contributing to both the cross-linking density and the enhanced mobility of the network. At the same time, dynamic bonds were incorporated into this unique cross-linking unit, where NGP was integrated to facilitate the exchange of dynamic bonds. In short, a highly cross-linked but highly dynamic network was designed to simultaneously achieve excellent mechanical properties and self-healing ability at room temperature. For detailed design, tetrafunctional diaminoglyoxime (AMG) reacted with isophorone diisocyanate (IPDI) to form neighboring urea bonds and oxime-urethane bonds in the four-arm cross-linking unit. The nucleophilic nitrogen atoms in urea bonds can attack oxime-urethane bonds to facilitate the exchange of oxime-urethane bonds at cross-linking units (Fig. 1). The essence of the oxime-urethane bond exchange is transcarbamoylation, typically accompanied by proton transfer and molecular rearrangement. Oxime-urethane bonds combine urethane groups (–NH–COO–) and imine bonds (C=N). The strong electron-withdrawing effect of the imine bond makes the urethane group unstable and prone to cleavage, generating oxime and isocyanate groups (25, 26). These groups can subsequently react to reform the oxime-urethane bond, giving the bond a dynamic nature. However, at low temperatures, the reversible cleavage and reformation reaction rate of the oxime-urethane bond is relatively slow (6, 27, 28). The capacity of oxime-urethane bonds to undergo exchange reactions is facilitated by the neighboring urea bond’s nitrogen, which acts as a nucleophile to attack the carbonyl carbon within the oxime-urethane bond, leading to the formation of a five-membered cyclic intermediate transition state at room temperature (19, 28, 29). To validate our concept, we implemented this design in a CAN-based ionogel system (i-Canogel), which demonstrated outstanding self-healing efficiency at room temperature, as well as excellent mechanical and electrical properties.

Fig. 1.

Internal catalysis mechanism for the oxime-urethane exchange.

Results and Discussion

Model Study of Neighboring Urea Bonds Promoting Oxime-Urethane Exchange.

For the synthesis of model small molecules, N-hydroxyacetamidine was reacted with phenethyl isocyanate to obtain the neighboring bond model compounds DBD (Fig. 2A and SI Appendix, Scheme S1). The NMR spectra (¹H NMR and ¹³C NMR) were utilized to confirm the structure of DBD (SI Appendix, Figs. S1 and S2). Acetone oxime was reacted with phenethyl isocyanate or propyl isocyanate to obtain the DC or AC (Fig. 2A and SI Appendix, Schemes S2 and S3), which lack a neighboring urea bond. The structure of DC and AC was confirmed by ¹H NMR and ¹³C NMR (SI Appendix, Figs. S3–S6). At room temperature, the production of ABD or AC and D obtained through the reaction of compound DBD or DC with A was monitored in real-time using ¹H NMR (Fig. 2 B and C). In ¹H NMR spectra, the signals at 1.915 or 1.929 ppm of DC, 1.918 or 1.933 ppm of AC, and 1.737 ppm of DBD corresponded to the protons of the methyl group on the oxime. In the control group, the intensity of 1.915 or 1.929 ppm proton peaks did not change with time, while the proton peaks at 1.918 or 1.933 ppm almost did not appear, indicating that compound AC was not significantly produced (Fig. 2B). Significant signal changes were observed in the [DBD]+[A] experimental group with time. Fig. 2C shows that the appearance and gradual intensification of the 1.750 ppm proton peaks, along with the decrease in the 1.737 ppm peaks, are indicative of the formation of compound ABD. To quantitatively evaluate the reaction equilibrium rate, the peak area of the methyl group in AC and ABD was integrated. The conversion rate of AC and ABD can be determined using the equations [AC]/([DC]+[AC]) and [ABD]/([ABD]+[DBD]), respectively, where [DC], [AC], [DBD], and [ABD] represent the concentrations of DC, AC, DBD, and ABD at a given time. At 25 °C after 96 h, the conversion rate of AC was nearly negligible, whereas that of an internal catalyst was approximately 35.38%. Fig. 2D shows that the rate of exchange reaction in the presence of an internal catalyst is significantly greater compared to that without the internal catalyst. To better understand the effect of temperature on the exchange reaction of oxime-urethane bonds, the exchange reaction of the small-molecule model was conducted at 50 °C and 100 °C. At 50 °C, the DC peak in the control group remained stable (SI Appendix, Fig. S7A), and the product AC peak barely appeared, consistent with the results at room temperature. When the temperature increased to 100 °C, the AC peak gradually emerged over time (SI Appendix, Fig. S7B). In contrast, the experimental group responded quickly to temperature changes. At 50 °C, the ABD product peak appeared shortly (SI Appendix, Fig. S7C). At 100 °C, the exchange rate significantly increased (SI Appendix, Fig. S7D). SI Appendix, Fig. S7E showed the conversion rate in the experimental group at 50 °C and 100 °C, indicating the reaction reached equilibrium after 60 min at 100 °C. These results demonstrated that the reversibility of the oxime-urethane bond was substantially enhanced by internal catalysis and the exchange rate increased as the temperature rose. These results indicated that the exchange reaction rate of the experimental group was accelerated in comparison to the control group, thereby demonstrating that the presence of neighboring urea bonds played a facilitating role in the dynamic exchange reactions of oxime-urethane groups. This provides clear evidence of the effect of NGP.

Fig. 2.

Model study of neighboring urea bonds promoting oxime-urethane exchange. (A) Exchange reaction between model compounds DBD or DC and A produced ABD or AC and D at 25 °C. (B) Real-time ¹H NMR spectra of the mixture of DC and A. (C) Real-time ¹H NMR spectra of the mixture of DBD and A. (D) Conversion rate of the molecular model with or without internal catalyst.

Design of Ionic Covalent Adaptable Networks with Superior Self-Healing Properties.

After observing the enhanced dynamic exchange reactions of oxime-urethane bonds in the model study, we proceeded to incorporate a reversible topological molecular structure of tetrafunctional diaminoglyoxime into polymer networks as the cross-linking unit, aiming to enable self-healing (Fig. 3A). The procedure for synthesizing i-Canogel was detailed in SI Appendix. The synthesis of i-Canogel n:m was accomplished through a one-step polycondensation process using commercially available materials: poly(butylene glycol adipate) diols (PBGAD), fluorinated tetraethylene glycol (FTEG), IPDI, AMG, and the ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl) imide ([EMI][TFSI]). The reaction was catalyzed by dibutyltin dilaurate. PBGAD was chosen as the soft segment due to its high compatibility with [EMI][TFSI] (30). FTEG was selected as the hard and hydrophobic segment as it primarily contains carbon-fluorine (C-F) bonds which are poor hydrogen donors and acceptors, leading to strong dipole-dipole interactions between FTEG segments (31). IPDI was picked as the hard segment because of its unique structure, which not only prevents crystallization but also promotes the exchange of dynamic structures through its steric hindrance effect (32). [EMI][TFSI] was selected as the functional component because it not only enhances the chain mobility through its physical lubrication effect but also imparts ionic conductivity to the material, resulting in the formation of an ionogel. The design and chemical structure of the proposed i-Canogel, denoted by the molar ratio of AMG, PBGAD, and FTEG, is illustrated schematically in Fig. 3B and SI Appendix, Scheme S4. We conducted a multidimensional comparison of the formulations, with a constant IPDI/AMG ratio of 10:3, changed the ratio of soft and hard segments: PBGAD/FTEG ratio of 2.5:1.5, 2:2, and 1.5:2.5 (i-Canogel 2.5:1.5, i-Canogel 2:2, and i-Canogel 1.5:2.5). Constant molar ratio of PBGAD/FTEG (2:2) was employed along with changing the content of cross-linker AMG: IPDI/AMG ratio of 10:4 (i-Canogel AMG 4), 10:3 [i-Canogel AMG 3 (i-Canogel 2:2)], and 10:2 (i-Canogel AMG 2).

Fig. 3.

The multifunctional integrated cross-linking structure design of i-Canogel with outstanding mechanical and self-healing properties. (A) Chemical structure of i-Canogel. (B) Schematic molecular structure of i-Canogel.

The structure of i-Canogel was analyzed using attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) (SI Appendix, Fig. S8). The peaks observed at 3,375 cm⁻¹ and 1,725 cm⁻¹ were attributed to the stretching vibrations of N–H and C=O, respectively, which suggested the formation of urethane groups. Notably, no peaks were observed at 2,264 cm⁻¹ corresponding to N=C=O skeleton stretching vibrations, indicating the complete reaction of the IPDI monomer (6). Accordingly, the self-healing ability was not caused by residual monomers. The formation of the cross-linked structure was demonstrated by immersion of i-Canogel 2:2 in tetrahydrofuran for 24 h without dissolution (SI Appendix, Fig. S9). Compared with CAN 2:2, which lacks [EMI][TFSI], the N–H peak of i-Canogel 2:2 shifted from 3,375 cm⁻¹ to 3,352 cm⁻¹, indicating that the ionic liquid forms hydrogen bonds with the N–H in the polymer chain. This result suggests a partial replacement of the hydrogen bonds formed by N–H and C=O in the network (30) (SI Appendix, Fig. S10). Additionally, the four vibrational bands corresponding to the symmetric stretch of S–N–S and O=S=O, as well as the asymmetric stretch of CF₃ and O=S=O from the [TFSI] anion, which are located at 1,049, 1,132, 1,176, and 1,346 cm⁻¹, respectively, exhibited a higher wavenumber shift in i-Canogel 2:2 (SI Appendix, Fig. S11). Therefore, the ionic liquid served as a lubricant, which promoted chain mobility and reduced the energy required to form five-membered rings by the NGP effect. Moreover, owing to the presence of FTEG chain segments, the [EMI][TFSI] could interact with the polymer chains via ion–dipole interactions. This meant the ionic liquid could disperse evenly in the CAN and would not leak out of the network (31). Elemental mapping of the cross-section of i-Canogel 2:2 revealed a uniform distribution of the elements C, O, N, F, and S, indicating the even distribution of [EMI][TFSI] in the CAN, without any noticeable phase separation (SI Appendix, Fig. S12). The excellent miscibility of the ionic liquid with the polymer contributed to the good transparency of the i-Canogel composite (SI Appendix, Fig. S13) (31).

Neighboring Bond Participation in Covalent Adaptable Networks.

The properties of various i-Canogel formulations were evaluated by adjusting the monomer ratio. The molar ratios of 2.5:1.5, 2:2, and 1.5:2.5 PBGAD/FTEG were used, and due to the lower molecular weight of FTEG, the content of the hard segment was found to have the most significant influence on the material properties. This accounts for the wide range of observed properties. To compare the impact of the PBGAD/FTEG molar ratio on the material properties, the molar ratio of IPDI and AMG was kept constant at 10:3 and the ionic liquid content was fixed at 50% for all formulations. The uniaxial tensile tests were performed to investigate the mechanical properties of the i-Canogel. As the content of the hard segment in the i-Canogel increased, the mechanical properties consequently improved. The tensile strength of the i-Canogel n:m (2.5:1.5, 2:2, and 1.5:2.5) at a stretching rate of 50 mm min⁻¹ increased from 3.49 ± 0.3 MPa to 6.48 ± 0.14 MPa, and toughness rose from 11.2 ± 1.58 MJ m⁻³ to 20.88 ± 1.8 MJ m⁻³ (Fig. 4A). Correspondingly, Young's moduli of the materials were also greatly improved from 0.45 ± 0.03 MPa to 0.96 ± 0.44 MPa. Moreover, because hard segments hindered the movement of molecular chains, the elongation at break decreased from 908.91 ± 35.47% to 714.87 ± 37.14%. These desired properties of i-Canogel were achieved through the incorporation of noncovalent interactions and a reversible four-arm cross-linking unit, which resulted in high toughness and good elasticity, respectively. Herein, the cyclic tensile tests were performed to demonstrate these performances. First, the cyclic tensile tests with different strain multiples of i-Canogel showed good stretch resilience (SI Appendix, Figs. S14–S16), and after 100 times tensile cycle test with a strain of 200%, i-Canogel 2:2 still maintained good stability (SI Appendix, Fig. S17). Second, the cyclic tensile tests with gradually larger strains (Fig. 4B), which performed with no waiting time between two consecutive loadings, showed great elasticity of i-Canogel 2:2. The dynamic covalent network and unbroken noncovalent interaction, including hydrogen and dipole–dipole interaction, entropically drove the stretched networks almost back to the initial state, after the i-Canogel 2:2 was stretched 100% strain on the first cycle. The result proved that i-Canogel 2:2 was a good elastic conductor within the strain up to 100% (6, 33). In the subsequent cyclic increasing strain (>100%), the residual strain and hysteresis loop area of the obtained stress–strain curve gradually increased, which indicated that a large number of noncovalent bonds break as sacrifices in the process of continuous stretching, successfully dissipating energy (6). Moreover, to demonstrate the elastic properties, repeated cyclic tensile tests at a large strain of 300% were performed on i-Canogel 2:2, with no waiting time between two consecutive cycle tests for ten cycles (Fig. 4C). The large hysteresis loop in the first cycle indicated significant energy dissipation due to the breaking of noncovalent bonds. The dissipation of energy during the second cycle of cyclic tensile testing was significantly lower than the first cycle due to insufficient time for the noncovalent bond to return to its original state. In subsequent tests, the hysteresis decreased slightly with increasing cycles, indicating continuous reorganization of sacrificial bonds. After relaxing for 2 h at 25 °C, the i-Canogel 2:2 showed good elastic recovery during the second cycle test, exhibiting a similar loading-unloading curve to the first cycle.

Fig. 4.

Mechanical properties of the i-Canogel. (A) The tensile stress–strain curves of pristine i-Canogel. (B) The tensile stress–strain curves of cyclic tensile tests with gradually larger strains of the prepared i-Canogel 2:2. (C) Repeated cyclic tensile curves of i-Canogel 2:2 at 300% strain. The samples were then left to rest for 2 h at 25 °C to allow for relaxation before the second cyclic tensile test. (D) The determined relaxation times of i-Canogel fitted to the Arrhenius equation.

To gain insights into the dynamic behavior of i-Canogel, stress–relaxation tests were conducted on all samples at different temperatures to determine the network relaxation times accurately. These tests involved applying a 5% torsional strain and recording the relaxation modulus over time. The stress relaxation test results demonstrated that the i-Canogel exhibited significant relaxation between 25 and 55 °C, which indicated the dynamic dissociation of the oxime-urethane bonds (6) (SI Appendix, Figs. S18–S20). Furthermore, the Maxwell model for viscoelastic fluids was used to determine the relaxation time (τ*), which was found to be 37% (G/G₀ = 1/e = 37%) of the normalized relaxation modulus. In addition, the temperature-dependent behavior of the relaxation time can be described by the Arrhenius equation τ (T) = τ₀ exp(E_a/RT), where τ represents the characteristic relaxation time, τ₀ represents the pre-exponential factor, and E_a represents the activation energy for stress relaxation (34). The activation energy (E_a) of stress relaxation for i-Canogel 1.5:2.5 was found to be significantly higher, at 77.92 ± 3.99 kJ mol^–1, compared to the E_a values of i-Canogel 2:2 and i-Canogel 2.5:1.5, which were 55.56 ± 2.51 kJ mol^–1 and 50.57 ± 1.99 kJ mol^–1, respectively, indicating that the material with the lower molar ratio of PBGAD to FTEG had a stronger network structure and higher thermal stability (Fig. 4D). These results indicated that the molar ratio of PBGAD to FTEG had a significant impact on the thermal stability and network relaxation of i-Canogel. To assess the thermal stability of i-Canogel, thermogravimetric analysis (TGA) was performed, as the characterization and processing of i-Canogel involve multiple heating steps at high temperatures (SI Appendix, Fig. S21). The results of TGA tests showed that all three networks had a similar temperature at 5% weight loss (T_deg,5%) in the nitrogen atmosphere, with values around 260 °C. Temperature-rising infrared testing was used to evaluate the generation of isocynanate group (SI Appendix, Fig. S22). At room temperature, the characteristic peak of N=C=O was negligible, but as the temperature increased, the characteristic peak of N=C=O became increasingly higher. This indicated that the oxime-urethane bond had dissociated the N=C=O group, so it was a dissociative dynamic bond (30). When the temperature reached a certain degree, the dissociation of the oxime-urethane bond far exceeded the restructuring, and the thermosetting material could exhibit a melting point like a thermoplastic material. It was proved by the rheological experiments that the melting point of the i-Canogel increased gradually with the decrease in the molar ratio of PBGAD to FTEG (SI Appendix, Figs. S23–S25). The melting points of i-Canogel n:m (1.5:2.5, 2:2, and 2.5:1.5) were found to be 138.9, 135.9, and 129.5 °C, respectively. Different contents of the cross-linking agent significantly impact the performance of material. ATR-FTIR confirmed the successful synthesis of i-Canogel AMG 2 and i-Canogel AMG 4 (SI Appendix, Fig. S26A). TGA results indicated that both i-Canogel AMG 2 and i-Canogel AMG 4 exhibited excellent thermal stability (SI Appendix, Fig. S26B). The dynamic mechanical analysis (DMA) showed that the T_g of i-Canogel AMG 2 and i-Canogel AMG 4 was lower than −60 °C (SI Appendix, Fig. S26 C and D). The rheological test revealed that the melting point increased with higher AMG content (SI Appendix, Fig. S26 E and F). Uniaxial tensile test results revealed that as the AMG content increased, the material’s tensile strength is enhanced (SI Appendix, Fig. S26 G–I). However, due to the high cross-linking agent content, the elongation at break for i-Canogel AMG 4 was limited to approximately 300%. i-Canogel AMG 2 with relatively low cross-linking agent content unsatisfied elasticity and lower tensile strength. Accordingly, we selected i-Canogel AMG 3 (i-Canogel 2:2) with a moderate IPDI/AMG ratio of 10:3 for the following studies.

Mechanical Self-Healing Properties of the Prepared i-Canogel.

The results of the stress relaxation experiment indicated that the activation energy of the polymer network could significantly vary depending on the content of hard and soft segments. This could result in variations in the polymer network’s dynamic properties and affect its ability to recover from stress, with higher activation energies indicating a slower rate of network rearrangements and a decreased ability to self-heal. On the other hand, lower activation energies correspond to more facile network rearrangements and improved self-healing capabilities. Theoretically, the low E_a in i-Canogel 2:2 and i-Canogel 2.5:1.5 resulted in more effortless network rearrangements, enabling enhanced self-healing properties in these i-Canogel formulations. The mobility of the segments in the prepared i-Canogel was evaluated using a differential scanning calorimeter (DSC) (SI Appendix, Fig. S27). The DSC test results revealed that the temperature range of −70 to 200 °C did not show any glass transition point (T_g). However, the melting point was in agreement with the results obtained from the rheometer. To determine the T_g of the i-Canogel, DMA tests were performed (SI Appendix, Figs. S28–S30). The data indicated a decrease in the T_g of i-Canogel with an increase in the molar ratio of PBGAD to FTEG. The T_g values for i-Canogel n:m (1.5:2.5, 2:2, and 2.5:1.5) were −59.6, −63.9, and −67.8 °C, respectively. The T_g characterized the mobility of polymer chains at room temperature. At room temperature, i-Canogel exhibited enhanced elasticity and self-healing ability, as a result of its distinctive structure that involves the participation of neighboring bonds. The self-healing ability of i-Canogel was evaluated using a uniaxial tensile experiment (Fig. 5A). Prior to the experiment, a strip of i-Canogel was cut into two pieces and then reconnected under ambient conditions. Large number of hydrogen bonds in the polymer network were expected to promote the self-healing of the i-Canogel (35, 36). The self-healing ability was assessed after 24 h and the results showed that the tensile strength, elongation, toughness, and Young’s moduli of i-Canogel 1.5:2.5 and i-Canogel 2:2 had recovered to 1.58 ± 0.2 MPa, 228.89 ± 52.26%, 2.52 ± 0.77 MJ m⁻³, and 0.93 ± 0.06 MPa, and 1.09 ± 0.15 MPa, 328.89 ± 60.05%, 2.29 ± 0.67 MJ m⁻³, and 0.53 ± 0.05 MPa, respectively (Fig. 5 A and C and SI Appendix, Fig. S31). The self-healing efficiencies of the i-Canogel 1.5:2.5 and i-Canogel 2:2 were determined and found to be lower than 50% for all tests except for Young’s modulus, where it was found to be greater than 90%. However, the i-Canogel 2.5:1.5 showed outstanding self-healing capabilities. The self-healing efficiencies were remarkable, with a value of 98.9% for tensile strength, 97.4% for elongation, 94.1% for toughness, and 96.6% for Young’s modulus, respectively (Fig. 5A and SI Appendix, Fig. S32). This indicated that the i-Canogel 2.5:1.5 had a high ability to recover its mechanical properties after being damaged. This was likely due to the good compatibility of the [EMI][TFSI] with the polymer and the special molecular topological cross-linking structure that allows for neighboring bond participation, which contributes to the excellent self-healing properties of the material. The mechanical properties of i-Canogel 2:2 exhibited excellent self-healing properties when the healing time was extended to 48 h. The self-healing efficiency of tensile strength, elongation, toughness, and Young’s moduli were 92.1%, 89.5%, 85.7%, and 98.8%, respectively (Fig. 5 B and C). The self-healing efficiency of i-Canogel 1.5:2.5 was only slightly higher after 24 h, which could be attributed to its high content of hard segments (SI Appendix, Figs. S31 and S33). This conclusion was supported by the results of the stress relaxation experiment, which is consistent with the activation energy data. The self-healing behavior of polymers is often closely related to the properties of their surfaces. Additionally, the hydroxyl groups at the polymer chain ends likely accelerated the exchange of oxime-urethane bonds. We employed X-ray photoelectron spectroscopy (XPS) to quantify the hydroxyl groups on the material surfaces (SI Appendix, Fig. S34). In the high-resolution O 1s XPS spectra, the hydroxyl content for i-Canogel n:m (1.5:2.5, 2:2, and 2.5:1.5) was 8.54%, 10.65%, and 19.94%, respectively. The test results were consistent with the trends observed in self-healing performance evaluated by mechanical tests.

Fig. 5.

Self-healing properties of the i-Canogel. (A) The stress–strain curves of pristine i-Canogel and healed samples for 24 h at room temperature. (B) The typical stress–strain curves of pristine and healed i-Canogel 2:2, confirming its self-healing ability for 48 h at room temperature. (C) The statistics of mechanical properties of the prepared i-Canogel 2:2. (D) Scatter plot of “Toughness,” “Strength,” and “Self-healing time” of i-Canogel 2:2 and other room temperature self-healing ionogels reported in the literature (37–44).

To compare the effects of with or without neighboring group catalyzed dynamic bonds on the self-healing property of materials, we designed the following CANs (SI Appendix, Scheme S5). We replaced the four-arm cross-linking unit AMG with dynamic chain extender without neighboring bonds participation (diacetylmonoxime, DMG) and nondynamic cross-linkers (glycerol) (6, 30). Consequently, we synthesized counterpart ionogel (i-Canogel DMG 2:2) containing the same content of ionic liquid as in i-Canogel 2:2. ATR-FTIR results indicated successful synthesis of the compounds (SI Appendix, Fig. S35A). Additionally, Fig. 5B and SI Appendix, Fig. S35B showed that i-Canogel DMG 2:2 swelled but did not dissolve after being soaked in THF for 24 h, proving the cross-linked structure of i-Canogel DMG 2:2. Furthermore, using TGA, DMA, and rheological tests, the T_{deg, 5%}, T_g, and melting point of i-Canogel DMG 2:2 were determined to be 245 °C, −65 °C, and 154 °C, respectively (SI Appendix, Fig. S35 C–E). The mechanical evaluations were also conducted (SI Appendix, Fig. S35F). The tensile strength, elongation, toughness, and Young’s moduli of i-Canogel DMG 2:2 were 0.56 ± 0.04 MPa, 801.1 ± 136.3%, 1.47 ± 0.09 MJ m⁻³, and 32.7 ± 7.9 kPa, respectively. The tensile strength, toughness, and Young’s moduli of i-Canogel 2:2 were 8.1, 9.2, and 19.3 times greater than those of i-Canogel DMG 2:2 (Fig. 3A). The results revealed that i-Canogel DMG 2:2 exhibited significantly lower tensile strength and room-temperature self-healing efficiency than i-Canogel 2:2 (SI Appendix, Fig. S35G). After 48 h, the self-healing efficiencies of i-Canogel DMG 2:2 based on tensile strength, elongation, toughness, and Young's modulus were 62.5%, 76.3%, 53.3%, and 78.8%, respectively. All results demonstrated that i-Canogel 2:2 with internally catalyzed four-arm dynamic cross-linking points has good mechanical properties and room-temperature self-healing capability superior to couterpart i-Canogel DMG 2:2. The temperature sweep curves from the rheological experiments clearly showed that as the temperature increased, the storage modulus of i-Canogel rapidly decreased (SI Appendix, Figs. S23–S25 and S35E). Compared to exchangeable dynamic bonds, dissociative oxime-urethane bonds were more conducive to extrusion and injection molding. When the temperature decreased, the recombination rate exceeded the dissociation rate, leading to a liquid–solid transition of the polymer network. The neighboring group participation enhanced the dissociation rate of oxime-urethane bonds, resulting in significantly lower melting point of i-Canogel 2:2 than that of i-Canogel DMG 2:2. At the same time, the four-arm cross-linked structure endowed i-Canogel 2:2 with excellent mechanical strength much higher than i-Canogel DMG 2:2. The mechanical self-healing performance of the ionogels, prepared using the neighboring bond participation effect, surpassed most of room temperature self-healing ionogels (37–44) (Fig. 5D).

Electrical Properties of the Prepared i-Canogel.

Stretchable conductors play key roles in flexibility electronics, which have significant advantages in terms of flexibility, transparency, and elasticity (45). Compared to most existing electron-conductive materials, ion conductors (hydrogels and ionogels) exhibit softness, high stretchability, and transparency, leading to increasing attention. Furthermore, compared to hydrogels, ionogels are more stable as they are not susceptible to the issues caused by water freezing and evaporation. Therefore, ionogels have better application potential. The presence of [EMI][TFSI] in the i-Canogel imparted both physical lubrication, which improved the mobility of the polymer chains, and ionic conductivity, thus yielding an ionogel (46). The [EMI][TFSI] was stable and did not evaporate with time owing to anions being rich in C-F bonds and their negligible vapor pressure. The electrochemical impedance spectra revealed that as the hard segment content increased in the system, the resistance decreased (Fig. 6A). The ionic conductivities of i-Canogel 1.5:2.5, i-Canogel 2:2, and i-Canogel 2.5:1.5 at 25 °C were 4.9 × 10⁻², 5.1 × 10⁻², and 5.7 × 10⁻² S m⁻¹, respectively (Fig. 6B). The increase in hard segment content resulted in reduced flexibility of the chain segments, which in turn hindered the mobility of ions within the network. The decrease in ion mobility directly contributed to a reduction in electrical conductivity (47). The addition of the four-arm cross-linking units and ionic liquid in the i-Canogel led to its remarkable elasticity, enabling it to function as a strain sensor. The ionic conductivity of i-Canogel was temperature-dependent. As the temperature increased, the ionic conductivity of the material also increased. This can be attributed to the enhanced mobility and easier transport of ions at higher temperatures (Fig. 6C). To assess the temperature-sensing capability of i-Canogel, i-Canogel 2:2 was chosen for evaluation (Fig. 6D). The temperature-sensing performance of i-Canogel 2:2 showed a linear region from 25 to 60 °C. In addition, the sensitivity of the i-Canogel 2:2 was determined by the gauge factor (GF), which was calculated as GF = (ΔR/R₀)/ε, where ΔR is the change in resistance relative to the nominal resistance R₀ at zero strain, and ε represents the tensile strain applied to the sensor. The resistance of the ion-conductive i-Canogel 2:2 exhibited a linear relationship with the applied strain. Linear equations were fitted to the data for practical applications of the i-Canogel 2:2. The relative change in resistance (ΔR/R₀) was observed to have a proportional relationship with the applied small strain, with a GF of 2.91 within the strain range of 0-75%. For larger strains, the GF increased to 4.57 within the strain range of 100 to 300% (Fig. 6E). Strain sensors typically require conductors with exceptional elasticity and resistive responsiveness (48, 49). By employing such stretchable and conductive materials, an expanding array of strain sensors has been developed for real-time monitoring of human motions. We utilized i-Canogel 2:2 to fabricate a strain sensor. The variation in sensor resistance was observed during cyclic tensile and release cycles at various strain levels, encompassing both small strains (5 to 10%) (SI Appendix, Fig. S36A) and large strains (50 to 100%) (SI Appendix, Fig. S36B). In addition to excellent elasticity and resistive responsiveness, instant self-healing properties were critical for the strain sensor. This capability was demonstrated in SI Appendix, Fig. S37A, where i-Canogel 2:2 was incorporated into a circuit featuring a green LED connected to a power source. Cutting i-Canogel 2:2 into two segments caused the LED to turn off. Upon rejoining the two pieces of i-Canogel 2:2, the LED immediately illuminated again. Real-time monitoring of the resistance of the cut and healed i-Canogel 2:2 was conducted using a multimeter (SI Appendix, Fig. S37B). Upon cutting, the resistance sharply increased to infinity, indicating a breakdown in the conductive path. Rejoining the two cut sections of i-Canogel 2:2 rapidly restored the conductive path, resulting in a corresponding decrease in resistance. After the initial cut-healing cycle, the electrical resistance recovered to its original value. Due to the nonvolatile and hydrophobic properties of [EMI][TFSI] and the hydrophobic nature of the hard segment in the polymer chain (50, 51), the weight of i-Canogel remained stable even after being stored for 43 d under ambient conditions with varying relative humidity (33 to 82%) and temperatures (−1 to 30 °C) (Fig. 6F). This observation suggested that i-Canogel exhibited environmental insensitivity, making it suitable for use in open-air and moist environments.

Fig. 6.

Electrical performance of the prepared i-Canogel. (A) Electrochemical impedance spectra of i-Canogel. (B) Ionic conductivity results of i-Canogel. (C) Temperature dependence of the ionic conductivities of i-Canogel in the temperature range of 25 to 60 °C. (D) Change in the resistance of i-Canogel 2:2 vs. temperature. (E) The resistance-strain curve of i-Canogel 2:2 strain sensor. (F) Weight retention of i-Canogel vs. time without specific encapsulation in ambient conditions.

Conclusion

This work represents an instance where internal catalysis has been incorporated into the cross-linking units to develop a generation of elastomer with both high self-healing performance and superior mechanical properties. The presence of neighboring groups significantly enhances the reversibility of dynamic bonds, leading to a substantial improvement in the network’s dynamic behavior. Combined with the multiple-cross-linking-in-one site topological design, the resulting i-Canogel exhibits both remarkable strength and self-healing properties at room temperature. The incorporation of active neighboring bonds provides an efficient approach toward regulating emerging dynamic oxime-urethane and other groups. Considering the extensive application of polyurethanes, this work holds widespread significance. The essence of the four-arm cross-linking unit combining electronic effects, spatial effects, and topological structures in a synergistic manner provides a powerful chemical pathway for tuning material properties.

Materials and Methods

Materials.

The dimethyl sulfoxide-d₆ (DMSO-d₆) and phenethyl isocyanate were purchased from Adamas®beta. Propyl isocyanate was obtained from 9dingchem. Acetone oxime was purchased from J&K Scientific. Glycerol, tetrahydrofuran, isophorone diisocyanate, and ditin butyl dilaurate were provided by Sigma-Aldrich. N-hydroxyacetamidine and diaminoglyoxime were obtained from Bidepharmatech. Fluorinated tetraethylene glycol was purchased from Fliorochem. Poly(butylene glycol adipate) diols (PBGAD, Mw ∼2,000) were purchased from Jining Baichuan Chemical Co., Ltd. (Jining, China) and dried under vacuum at 110 °C for 1 h before use. [EMI][TFSI] (≥99%) was obtained from the Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences. Dimethylglyoxime (DMG) was supplied by Sinopharm Chemical Reagent. All the reagents were used as received without further purification unless otherwise noted.

Synthesis of the i-Canogel and CAN 2:2.

PBGAD (n mmol), FTEG (m mmol), and AMG (j mmol) were dissolved in 4 mL of THF in a glass vessel equipped with a magnetic stirrer at 60 °C, where the n:m:j= 1:1:4, 3:3:2, 2.5:1.5:3, 2:2:3, and 1.5:2.5:3. IPDI (10 mmol) and DBTDL (0.2 wt%) were then added, and the mixture was allowed to react for 15 h under a nitrogen atmosphere. Subsequently, the required equivalent of [EMI][TFSI] (50 wt%) was added. The reaction mixture was poured into a polytetra fluoroethylene mold and reacted at 60 °C for 24 h before further curing under a vacuum at 60 °C for another 24 h to produce i-Canogel. CAN 2:2 was prepared in the same way as i-Canogel 2:2 but without the addition of [EMI][TFSI].

Synthesis of the i-Canogel DMG 2:2.

PBGAD (2 mmol), FTEG (2 mmol), DMG (3 mmol), glycerol (2 mmol), DBTDL (0.2 wt%), and IPDI (10 mmol) were dissolved in 4 mL of THF in a glass vessel equipped with a magnetic stirrer at 60 ^oC for 1 h under a nitrogen atmosphere. Then, the required equivalent of [EMI][TFSI] (50 wt%) was added. The reaction mixture was then poured into a polytetra fluoroethylene mold and reacted at 60 °C for 24 h before further curing under a vacuum at 60 °C for another 24 h to produce i-Canogel DMG 2:2.

NMR Measurements.

All NMR spectra were recorded on a Bruker AVANCE 600 NMR spectrometer using DMSO-d₆ as solvent.

Attenuated Total Reflectance FTIR Measurements.

The attenuated total reflectance FTIR (ATR-FTIR) spectra and temperature-rising infrared spectra were recorded on a Thermo Scientific Nicolet 8700 spectrometer.

Transparency Measurements.

Optical transmittance was measured using a Jasco V-630 UV–Visible spectrophotometer.

Morphology and Elemental Distribution Measurements.

The morphology and elemental distribution of the samples were characterized using a field emission scanning electron microscope (FE-SEM, HITACHI SU-8010, Japan) and energy-dispersive X-ray spectroscopy (Oxford Inca X-Max, UK).

Stress–Relaxation Measurements.

Stress-relaxation tests were conducted using an Anton Paar MCR702 with a 25 mm solid-mode rotor. Samples with a thickness of 1 mm were tested by applying a 5% strain at various temperatures.

Temperature Sweep Measurements.

Temperature sweep tests were performed on an Anton Paar MCR702 rheometer equipped with a 25 mm solid-mode rotor. The tests applied a 0.5% strain at a frequency of 1 Hz to samples that were 1 mm thick.

Mechanical Properties Measurements.

The mechanical properties were investigated using an MTS E42 tensile machine with a 100 N load cell. Uniaxial tensile tests and cyclic tensile tests were performed at a stretching speed of 50 mm min⁻¹ unless otherwise noted.

Conductivity Measurements.

Impedance spectroscopy was recorded on a CHI670E electrochemical analyzer. The conductivity of the i-Canogel was determined using the formula σ = L/AR, where L represents the length of the i-Canogel, A is the cross-sectional area of the i-Canogel, and R is the bulk resistance.

TGA Measurements.

TGA tests were performed on a TG 8209 F1 thermogravimetric analyzer (NETZSCH, Germany) from 40 to 600 °C at a heating rate of 10 °C min⁻¹ under a nitrogen atmosphere.

DSC Measurements.

DSC tests were carried out on a DSC-822 DSC (Mettler Toledo, Switzerland) at a heating rate of 10 °C min⁻¹ under a nitrogen atmosphere.

DMA Measurements.

Furthermore, DMA tests were conducted on a DMA1 dynamic mechanical analysis (Mettler Toledo, Switzerland) at a heating rate of 5 °C min⁻¹ and 1Hz frequency under an air atmosphere.

Resistance Measurements.

The electrical resistance of the samples was monitored using a Keithley DMM7510 multimeter.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.