Smart hydrogels inspired from natural principles and structures have been widely utilized in artificial joint or tissue[1-6], drug release carrier[7-9], shape memory material[10-12], actuators[13-15] and sensors[16, 17]. For example, inspired by biological metabolism of living muscle self-growing hydrogel was fabricated for soft robots and intelligent devices;[18] by mimicing the dynamic lubricating mechanism of articular cartilage, a shear-responsive supramolecular lubricating hydrogel was fabricated for cartilage substitution.[19] During the dynamic process of practical application, these kinds of hydrogels might be damaged inevitably. To solve such problem, the strategy of self-healing is introduced in the bio-inspired hydrogel construction.[20-22] For example, a self-healing supramolecular poly(N-acryloyl glycinamide) hydrogel was fabricated based on the amplified hydrogen bonding interactions between amino acid residues in polymer hydrogel that mechanics could be recovered in static recovering condition.[23] By utilizing the noncovalent interaction of the biphasic heteronetwork, a gel materials with switchable mechanical property could be synthesized that exhibited self-healing capacity. The recovering process was also performed in a static environment.[24] These self-healing researches on hydrogel implemented to recover the physical and chemical characters under a static condition. However, the bio-functions of the living tissue in body generally occurs in a dynamic shearing environment, like the lubrication of cartilage [25-28] or cornea [29, 30]. The healing of soft materials under shearing would become more difficult due to the additional disturbation. Therefore, it is a great challenge to develop self-healing hydrogels that could recover the living bio-function in a dynamic mechanical environment.
Generally the strategy of lubricating recovery is mainly considered: i) in macro scale the lubricants is infused in structures of the surface for healing; ii) in micro scale the microcapsules with healing agents are encapsulated in the materials, which would be released for healing after being damaged (Scheme 1).[31] However, the recovery of living lubricating bio-functions is more likely to be performed in molecular scale concerning of the molecular assembly-disassembly. To construct a self-healing hydrogel for recovering the living lubricating bio-function in molecular scale, two points should be considered: i) the healing process should happen under a dynamic mechanical stimulus; ii) the original living functions should be healed or kept during the dynamic stimulated process. Semi-convertible hydrogel (SCH) was developed as a new type responsive lubricating materials, which combined reversible noncovalent supramolecular network into the covalent polymer network.[32] The responsive gel-sol transition of supramolecular network afforded the dynamic function; the covalent polymer network endowed high mechanical framework. Herein, we demonstrated a semi-convertible hydrogel shear-assisted self-healing ability for recovering the active lubricating function in molecular scale. This semi-convertible hydrogel was integrated by triple networks including the self-healing polyhydroxyethyl acrylamide (PHEAA) network, shear responsive N-fluorenylmethoxycarbonyl-L-tryptophan (FT) supramolecualr noncovalent network, and rigid polyvinyl alcohol (PVA) covalent network (Figure 1a). PHEAA is a typical self-healing material owing to tight intermolecular hydrogen bonding.[33-35] The highly cooperative hydrogen bonding could serve as reversible sacrificial interactions that was expected to be recovered after being broken. When PHEAA network is introduced into the interpenetrating network of shear force responsive supramolecaular FT and covalent PVA for forming the semi-convertible hydrgel, it might offer a possible way to heal the damage and the dynamic lubricating function. First the sol-state FT monomers would appear onto the surface as lubricanting layer due to the disassembly of FT supramolecular network in the hydrogel under shearing, which endowed the hydrogel the shear-responsive lubricating ability (Figure 1b). If the hydrogel was scratched (Figure 1c), some disassembled FT monomers sol would flow into the scratched gap with less shearing stress (Figure 1d). Then two synergic aspects would be involved in the self-healing process under dynamic shearing process: the re-formation of hydrogen bonding of hydroxide groups between the two adjacent PHEAA chains; the re-assembly of sol-state FT molecules flowed in the scratched gap (Figure 1e). [36, 37] It could predict that both the responsive lubricating ability and mechanical property of the SHSCH might be self-healed under dynamic shearing condition.
Briefly, the self-healing semi-convertible hydrogel was synthesized by a one-pot method. The supramolecular FT network assembled by π-π interaction was interpenetrated into the covalent PHEAA and PVA polymer network after preparation. According to FT content (0, 1.0, 1.5, 2.0 and 2.5 mg/ml) in the hydrogel, the prepared SHSCH was marked as SHSCH0, SHSCH1, SHSCH2, SHSCH3 and SHSCH4, respectively. The mechanical property of prepared hydrogels was measured by using a materials testing system. The fracture strength of SHSCHs was gradually decreased from 6.96 ± 0.51 to 0.92 ± 0.05 MPa following the increase of FT concentration (Figure 2a). The reason might be attributed that the weak noncovalent bondings of FT supramolecular chains make the hydrogel network more flexible and unstable compared to the covalent PHEAA and PVA polymer network. Although the strength of hydrogels decreased with the increase of supramolecular contents, the minimum value of SHSCH4 was still higher than 0.9 MPa, which would ensure the requirement on the following friction and self-healing test. The thixotropy of SHSCHs and FT supramolecular hydrogel was investigated by the rheology method (Figure 2b). For FT supramolecular hydrogel, the storage modulus G’ and loss modulus G’’ conjoined together at a low oscillation of around 27 Pa. The FT gel-sol transition presented the disassembly of FT network. For the SHSCH network, the storage modulus G’ decreased with the increased FT content in hydrogel network; meanwhile G’ and G’’ of SHSCHs (concentration of FT was lower than 1.5 mg/ml) would not intersect together following the increasing shear oscillation, no obvious gel-sol transition for the whole hydrogel occurred in testing oscillation stress range. As increasing content of supramolecular FT in the SHSCHs (higher than 2.0 mg/ml), the difference between G’ and G’’ of SHSCH4 decreased, and intersected together at high oscillation stress range. The hydrogel would suffer gel-sol transition at this high shear oscillation due to the increase of supramolecular FT content in SHSCH network. Therefore, the SHSCH2 was mainly used in following experiment for demonstrating the self-healing ability under static and dynamic shearing condition. The bonding interactions of noncovalent and covalent in SHSCH was analyzed by the differential scanning calorimeter (Figure 2c). The stronger bonding interactions corresponds higher melt point (Tmelt) and endothermic enthalpy (∆H). It can be seen that the ∆H and Tmelt of SHSCH2 was 3.51 J/g and 60.40 °C, which was lower than those (∆H = 7.44 J/g and Tmelt = 66.83 °C) of covalent PHEAA and PVA interpenetrating network (SHSCH0). It might evidence that the introduction of the supramolecular FT contents into the covalent network system would result in the weaker interactions compared to the the covalent network system. It was also consistent with the mechanical and rheological results in Figure 2a and b.
The shear-responsive lubrication of SHSCH was investigated on a Bruker UMT Tribo Lab system, which was tested by ball-on-disk reciprocating tribometer. A ceramic Al2O3 contact ball (Diameter of 5.0 mm) was used to perform linear motion on the prepared hydrogel surface. When shear force was acted on the hydrogel, thixotropic FT section in the hydrogel network disassembled, and simultaneously occurred gel-sol transition. The sol-state disassembled FT monomeric units appeared onto the hydrogel surface and formed a lubricating layer (Figure 1b). The friction coefficient on SHSCH2 during 30 shearing cycles gradually decreased from 0.46 ± 0.01 to 0.34 ± 0.04 in the experimental condition of load : 10 mN, shear velocity : 0.1 mm/s, and humidity: 80% (Figure 2d).
The results might be ascribed to the appearance of sol-state supramolecular lubricating layer on the hydrogel surface originated from the disassembled FT units and segments under shearing. Compared to SHSCH2, the friction coefficient on SHSCH0 surface, namely the hydrogel without FT, would increase following the shear cycles as shown in Figure S1. It evidenced that FT played key role on the shear-responsive lubricating function of SHSCH. Here after 15 shear cycles, the friction coefficient would not change significantly and became relative stable, which meant the disassembled process of FT on the hydrogel surface and superficial region was almost finished. Therefore, the friction results in the subsequent experiment were mainly adopted after 15 shear cycles. Moreover, the friction on SHSCH was also observed by controlled humidity ranged from 20 to 80%. As shown in Figure S2, we found that the shear-responsive lubrication on the SHSCH was performed well at the humidity of 80%. Regarding to the humidity less than 80%, the friction coefficient would fall down following the first several shear cycles, and then increase graduially under continously shearing. It demonstrated that the shear-responsive lubrication would work on the prepared SHSCHs, and lost at low humidity range due to the possible incease of roughness originated from the water evaporation on hydrogel surface.
By regulating the load and shear velocity, the shear-responsive lubrication on the SHSCH was further investigated. By keeping the load at 10 mN constantly and changing the shear velocity of the contact ball from 0.1 mm/s to 0.5 mm/s, the friction force on SHSCH2 before being sheared was 4.64 ± 0.32, 5.17 ± 0.15, 5.53 ± 0.82, 6.06 ± 0.23 and 7.60 ± 0.45 mN, respectively. After being sheared all of them at these five velocity setpoints would reduce, which were 3.42 ± 0.38, 3.75 ± 0.18, 4.09 ± 0.37, 4.89 ± 0.64 and 4.96 ± 0.66 mN correspondingly. (Figure S3a). When keeping the shear velocity constant at 0.1 mm/s, the load was controlled from 6 mN to 14 mN, the friction force (2.38 ± 0.23, 3.28 ± 0.39, 4.64 ± 0.32, 6.50 ± 0.98 and 8.86 ± 1.24 mN) at each load setpoint would reduce to 1.85 ± 0.15, 2.45 ± 0.45, 3.42 ± 0.38, 4.71 ± 0.50 and 6.84 ± 0.75 mN, correspondingly (Figure S3b). All these results evidenced the prepared SHSCHs presented well shear-responsive lubricating capibility. Due to aggregation induced emission property of FT supramolecular hydrogel the fluorescence intensity would decrease following the disassembled process when applying shear cycles.[15] The fluorescence spectrum on the SHSCH surface in solid mode was detected with 365 nm excited wavelength following shear cycles (Figure 2e). The reduction of the fluorescence intensity proved there was more and more FT supramolecular disassembly on the surface as the increase of shear cycles.
To investigate self-healing ability on mechanical property of SHSCH in dynamic shearing condition, a direct observation on the hydrogel under static and shearing condition was shown in Figure 3. A scratch on hydrogel surface with the width of about 100 μm was made by single side blade. In our experiment SHSCH0 used as a contact pair moved reciprocately and continuously on the scratch of SHSCH with the load of 300 mN and velocity of 80 mm/s in humidity of 80% at room temperature (Figure 3a). It was marked as the dynamic condition. Regarding to the static self-healing condition, SHSCH was kept in the same humidity and temperature environment of dynamic condition, but there is no shear acted on the SHSCH. As shown in Figure 3b, the surface was self-healed after 240 min in static condition. When the contact hydrogel was shearing along the scratch, the scratch on the surface can be also self-healed (Figure 3c). Here, we further marked the angle between shearing direction (red arrow) and the scratch as the shearing angle (SA). By increasing the SA from 0o to 90o as shown in Figure 3c-f, the recovery of the scratch become more and more difficult, and could not be healed when the SA was higher than 60o (Figure 3e and f).
Beside the direct observation on the self-healing behaviors of SHSCHs, the mechanical property is utilized to evaluate the self-healing ability of the hydrogel under shearing. In the experiment the tensile stress of SHSCH would decrease from 376.12 ± 42.20 to 74.70 ± 2.79 kPa before and after being scratched. After 240 min dynamic healing process at SA of 0o the tensile stress would increase to 221.78 ± 5.89 kPa (Figure S4). Compared to this, the tensile stress in the static condition was about 155.26 ± 3.28 kPa after self-healing (Figure S4). In order to evaluate this recovery efficiency of SHSCHs, it was calculated by comparing the healed tensile strength and the original tensile strength of hydrogel.The healing efficiency of 58.96% in dynamic shearing condition was higher than that of 41.28% in static condition (Figure 3g). It indicated that the dynamic shearing process could improve the self-healing behavior more effectively than that in the static condition. The reason could be ascribed that there was more FT monomeric sol flowed into the scratched gap under shearing compared to the static condition. The the π-π assembly of FT to gel would associate the hydrogen bonding assembly of polyhydroxyethyl acrylamide in the scratch gap. When the shear angle was managed to increase from 0o to 90o, the recovery efficiency would reduce from 58.96% to 13.30% (Figure 3h). It might be attributed that the shear would tear the scratch position when SA was not 0o; and the tear force perpendicular to the scratch increased with SA that would prevent the self-healing of scratch. Based on the above observation and analysis, the following characterization on the self-healing ability of SHSCH under dynamic shearing process was performed at SA of 0o.
For further investigating the recovery of mechanical property in this dynamic shearing process, the load and shear velocity of the contact hydrogel pair were regulated as shown in Figure 3i and j. When the load changed from 100 mN to 300 mN, and the shear velocity was 80 mm/s, the tensile stress after self-healed increased from 161.39 ± 8.41 to 221.78 ± 5.89 kPa, and then decreased to 152.88 ± 2.02 kPa when load increased to 500 mN (Figure 3i). The reason might be attributed to a dynamic equilibrium between the load and the self-healing process. Under the relative low-loaded shearing condition, the reassembly of the FT sol might synergize the H-bonding assembly of PHEAA chains in the hydrogel, which would be benefited to the self-healing process. On the other hand, the scratch would be destroyed under high-loaded shearing condition, which dynamic instability during the shearing process weakened the self-healing ability. When the load was fixed at 300 mN, the recovered tensile stress after 240 min would increase from 160.68 ± 6.16 to 221.78 ± 5.89 kPa following the shear velocity changed from 10 to 80 mm/s (Figure 3j).
As exhibited in Figure S3a, the friction force would increase with the shear velocity. In another words, the shear force would increase with the velocity, which brought more disassembly of FT network. Therefore, the self-healing process participated by more FT sol might result in better recovery on mechanical property. Based on these results the optimized shearing condition for self-healing is in condition of the load of 300 mN and shear velocity of 80 mm/s, which would be adopted in the subsequent friction test for self-healing investigation.
The self-healing ability on the lubrication of SHSCHs under shearing and static condition was shown in Figure 4a, the friction coefficient for scratched area was 1.18 ± 0.08, and it was higher than 0.77 ± 0.05 of the original smooth surface. Following the healing time ranged from 0 to 360 min it decreased gradually no matter in dynamic shearing or static condition. The friction coefficient of 0.47 ± 0.01 in dynamic shearing condition was lower than that of 0.63 ± 0.05 in static condition. It closed to the friction coefficient of 0.49 ± 0.07 on original smooth hydrogel surface after being sheared. On one hand, the shear-responsive lubricating ability of the SHSCH was kept even after being scratched. The reason might be attributed that the homogeneous supramolecular FT networks on the hydrogel surface/superficial layer would disassembly triggered by the shearing action, the disassembled FT sol layer would appear on the surface no matter if the surface being scratched or not. On the other hand, this shear-responsive lubricating effect on the prepared SHSCH could be self-healed during the shearing process. This might be ascribed to the cooperating reassembly of π-π interaction in FT supramolecular network and hydrogen bonding in PHEAA inside the scratched gap as described in Figure 1.
To evaluate the self-healing effect on the shear-responsive lubricating ability of SHSCH, a shear-responsive lubricating parameter (SRLP) is introduced as (μs-μ0) / μ0. Here the μs is the measured friction coefficient following the healing time; μ0 is the friction coefficient on original smooth SHSCH surface. As shown in Figure 4b, compared to SRLP ~ 0 of original hydrogel without being scratched, it was 0.39 on the scratched SHSCH. It presented that the surface friction coefficient increased after being scratched. The SRLP would decrease to negative level following the healing time in both static and dynamic shearing condition. It indicated that the shear-responsive lubricating behavior on SHSCH was existence and recovered gradually following the healing time no matter in static or dynamic self-healing condition. More interestingly, the SRLP value (-0.32) after being self-healed in shearing condition was lower than that (-0.17) in corresponding static condition. It showed that the shear responsive lubricating effect in dynamic shearing condition was better than that in static healing condition, which was consistent with the self-healing results on mechanical property. Meanwhile, This SRLP was closed to the value (-0.30) on the original smooth hydrogel after being sheared, which showed that the self-healed hydrogel in shearing condition could reach similar shear-responsive lubricating level of the original unscratched hydrogel. Figure 4c shows a comparison of the healing lubricating capability on different materials, such as alloy[38-42], ceramics[43], metal[44], silicon[45-49], epoxy[50-55], nylon[56] and hydrogel[57] (details in Supplementary Table S1). Among these materials the realization on self-healing of lubrication always accompanied with the bulk destruction or abrasion no matter through the method on microcapsules lubricants encapsulation (grey) or on infusing the lubricants directly (dark blue). The mechanic property would lose gradually with the occurrence of shear-responsive lubrication. Compared with these materials, the as-developed semi-convertible hydrogel displays a unique advantage in self-healing property on both responsive lubrication and mechanic under shearing but not based on the destruction to the material itself.
For identifying the shear-responsive lubricating layer of SHSCHs, a paraffin rotary microtome was used to prepare microscope section slice with the thickness of about 625 μm. The section slices of scratched hydrogel were shown in Figure 4d. After being scratched, there was subtle disassembled FT monomeric sol appeared on the scratched gap surface. The scratch was almost self-healed in a dynamic condition after 360 min, and left a slight trace. A sol layer could be observed on the healed hydrogel surface. As comparison, the section slice of original hydrogels before and after being sheared was shown in Figure 4e. A sol layer could also be observed on the surface that evidenced the shear-responsive lubricating function on the SHSCH. It confirmed that the shear-responsive lubricating ability of SHSCH was performed no matter if the surface was scratched or not.
As we know, the abrasion on artificial devices cannot be evitable that will strongly affect the device function. For demonstrating the self-healing of lubrication on SHSCH, a medical cartilage model was used to conduct abraded experiment of SHSCH (Figure 5a). Both surfaces of model articular head and fossa were covered with SHSCH2. A joint-like flexion and extension movement was applied between covered articular head and fossa with over load of 5 N. The abrasion cycle based on the joint flexion and extension was illustrated in Figure 5a. With the process of relative motion, the surface of the hydrogel would be abraded. After 30 flexion and extension cycles, compared to the smooth original hydrogel surface, the surface became very rough (Figure 5b). A dynamic self-healing process was applied on the abraded hydrogel at room temperature under shearing of 300 mN load and 80 mm/s shear velocity. After 6 hours, the roughness almost diminished (Figure 5b). The friction force during this process was also recorded as shown in Figure 5c. Compared to the original surface friction force (8.72±1.04 mN, black line), the friction force of hydrogel significantly increased to 15.89±1.07 mN after abrasion (blue line). After 6 h the friction force would self-healed that was 9.11±0.75 mN (orange line). It was closed to original surface friction force. These results showed that both the morphology and friction force could self-healed even after an over loaded condition of simulating the joint-like flexion and extension movement.
In summary, a self-healing semi-convertible hydrogel was developed by shear-responsive N-fluorenylmethoxycarbonyl-L-tryptophan supramolecular network, self-healing polyhydroxyethyl acrylamide network and rigid polyvinyl alcohol covalent network. On one hand, the dynamic shear-responsive lubricating function existed no matter if the hydrogel surface was scratched or not, which was attributed to the appearance of disassembled FT supramolecular sol layer on the surface triggered by shear force. On the other hand, the responsive lubrication and the mechanical property of SHSCH could be self-healed under shearing because of the synergistic interaction of the noncovalent hydrogen bonding assembly of PHEAA and the π-π assembly of FT. More interestingly, the recovery of shear-responsive lubrication and mechanical property of SHSCH in shearing condition was better than those in static condition. We anticipated that the prepared self-healing semi-convertible hydrogel would offer a possible way to fabricate the self-healing soft lubricating devices under dynamic environment and beyond. The future research work will actively pursue for the high recovering efficiency on more complex functions integration under shearing, especially the dynamic multi-function response system.