Ultra‐Tough Self‐Healing Hydrogel via Hierarchical Energy Associative Dissipation

Abstract Owing to high water content and homogeneous texture, conventional hydrogels hardly reach satisfactory mechanical performance. Tensile‐resistant groups and structural heterogeneity are employed to fabricate tough hydrogels. However, those techniques significantly increase the complexity and cost of material synthesis, and have only limited applicability. Here, it is shown that ultra‐tough hydrogels can be obtained via a unique hierarchical architecture composed of chemically coupled self‐assembly units. The associative energy dissipation among them may be rationally engineered to yield libraries of tough gels with self‐healing capability. Tunable tensile strength, fracture strain, and toughness of up to 19.6 MPa, 20 000%, and 135.7 MJ cm⁻3 are achieved, all of which exceed the best known records. The results demonstrate a universal strategy to prepare desired ultra‐tough hydrogels in predictable and controllable manners.

atoms in carboxyl groups are essentially identical, each can form hydrogen bonding with N (Figure S2D).In MD simulations, each N-O distance was therefore the average distance between N and the two carboxylic O atoms.The first N-O peak in the r-g(r) plot belonged to directly bonded N•••H-O.When stable bidentate hydrogen bonding in Figure 2A formed, the two carboxyl groups being involved were fixed diagonally (Figure S2D).In this case, when the first N-O distance was 2.5 Å, the diagonal N-O distance was 4.2 Å.Therefore, the proposed core structure was verified.

Additional discussions on the IR spectra
As the amine concentration increased, the intensity of carboxyl C=O stretching (1698 cm -1 ) got reduced while the asymmetric (1562 cm -1 ) and symmetric (1398 cm -1 ) COO -stretching became stronger (Figure 3E), suggesting the gradual deprotonation of carboxyl groups due to elevated pH (Figure S5).It was found that the shift of -CH 2 and -CH 3 bands was the steepest when the volume of added amine changed from 60 μL to 80 μL, indicating a proper amine concentration was the key to trigger the majority of hydrophobic interactions (Figure S4G).
When the added amine exceeded 100 μL, the peak position reached a constant value and the corresponding hydrogels appeared to be transparent (Figure S6E).Those observations had verified the scheme in Figure S3E that at high amine concentrations, although hydrophobic interaction still presented, the lack of free hydrophilic groups and the high pH prevented the formation of hydrophobic domains.

The influence of substitutional group on phase separation
As demonstrated by experiments, having electron withdrawing groups or hydrophobic groups on the polyacid chain helped broaden the phase separation window while hydrophobic groups on the amine narrowed it down (Figure S6A-S6C).This discrepancy in the effect of substitutional group was likely caused by the structural character of hydrophobic domains.
When hydrophobic domain formed, the z-core acted as a seed to integrate a few hydrophobic segments, rendering a hydrophobic inner sphere.The remaining hydrophilic moieties on the polymer then intertwined the inner sphere to create a hydrophilic shell as a stabilization layer in aqueous media (Figure S4D).
The substitutional group on the polyacid chain enhanced interactions in the inner sphere, therefore reduced the number of required acidic monomers to form a single stable hydrophobic domain.Consequently, phase separation happened across a broader acid:amine range.In contrast, enhancing the hydrophobicity of amines made them stronger seeds so phase separation could happen at low amine concentrations and each inner sphere contained more hydrophobic segments.This led to a quick consumption of polyacid chains.No sufficient building blocks would be available at high amine concentrations to enable phase separation.

A spring model of HEAD gel
The role of each energy dissipation structure during mechanical deformation could be expressed by a spring model, in which the spring constant k 1 <<k 2 <k 3 (Figure S9A).The zcore featured the largest elastic range but the smallest spring constant.Hydrophobic domains had moderate elasticity and strength while the advanced structure was strong but rigid.To survive in large stretching, multiple hydrophobic domains and advanced structures had to be tightly connected to share the geometric deformation.
When the system was in lack of hydrophobic interactions, z-cores dominated the mechanical behavior.Although they could withstand large deformation, their low strength made the gel very soft.In contrast, if hydrophobic interaction was too overwhelming, molecular interactions tended to be localized within individual unit so connections among units were lost.In this case, the gel was hard but had very poor ductility.To reach superior comprehensive properties, a chemically balanced network was desired in which the elastic range of united hydrophobic domains and advanced structures was roughly the same as that of z-core (Figure S9B).

Self-repair behavior of various HEAD gels
The optimal self-repair condition varied among different HEAD gels (table S2).In general, softer and elastic gels healed at a faster speed under RT owing to their z-core dominated structure.Tougher gels involving heavy hydrophobic interactions, on the other hand, required a longer healing time at elevated temperatures.For example, A55 and A56 gels could completely recover from damage after a 4 h incubation at RT (Figure 4K).A53 and A54 gels required a higher self-healing temperature (40 °C) and elongated incubation time (Figure S13A).Harder gels such as M47 had very slow self-repair dynamic.To speed up the process, the wound was first transiently heated at about 240 °C for 3 s by a hot plate or iron.The heated area quickly turned clear, indicating the loss of phase separation.The samples were then incubated at 40 °C for 1 day during which new hierarchical structures would form.The as-obtain self-repaired samples possessed partially recovered properties, as shown in Figure S13B.

Figure S1 .
Figure S1.Reactive species tested in the preparation of HEAD gels.The green and red color indicate preferred and undesirable reactants, respectively.

Figure S3 .
Figure S3.Additional principles to design z-core.(A) Having hydrophobic groups and carboxyl groups on different carbons results in an energetically unfavorable configuration (dash circled region) with regularly aligned hydrophobic groups facing towards aqueous media.(B) Pictures of MAAc and CAc based precursors after incubation, demonstrating the selectivity on position of hydrophobic groups.(C) Pictures reflecting the effect of the number of 1,2-diamine unit on gelation.

Figure S4 .
Figure S4.Amine-concentration dependent phase separation.(A1)-(A2) Additional fluorescent image of 30 and 200 gel in Figure 3A.(B1)-(B2) Additional SEM image of 30 and 200 gel in Figure 3A.(C)-(E) Schemes showing the phase separation process at low amine concentration (C), moderate amine concentration (D) and high amine concentration (E).(F) IR spectra of MAAc gels with various amount (μL) of EDA.(G) Peak shift of -CH 2 and -CH 3 stretching in (F) as a function of added EDA.

Figure S8 .
Figure S8.Scheme showing the formation dynamics of different structural components.

Figure S9 .
Figure S9.A spring model of HEAD gels.(A) Spring model of each structural component under static condition.Black circles indicate strong connections between different units.(B) Mechanical behavior of various HEAD gels explained by the spring model.Black circles indicate strong connections between different units.

Figure S10 .
Figure S10.Strategy to tune the property of HEAD gels.(A)-(B) Effect of matching water affinity.(C) Effect of introducing matched and unmatched hydrophobicity.The scheme on the right showing the alternating hydrophobic (red) and hydrophilic (blue) segments in TEDETA and cis-/trans-CHDA.(D) Tensile behavior of HEAD gels as a function of the length of amine (A51, A53 and A54).(E) Longer amines enhanced the elasticity in relatively hydrophobic reaction systems, but showed little correlation with the strength.(F) Effect of concentration of reactants.

Figure S11 .
Figure S11.A flow chart describing the design strategy of HEAD gels.

Figure S12 .
Figure S12.Additional characterizations of HEAD gels.(A)-(B) Additional tensile results.Note the maximum measurable strain was 15000% due to instrumental limit.(C) Pictures of a stretched A56 gel.Material appeared to be sliding out of the shoulder and extended into extremely thin thread, demonstrating the role of z-core.(D) Tensile curves of M47 gel before and after swelling.(E) Tensile curves of A54 gel before and after swelling.(F) Pictures showing the load-bearing capacity of the scaffold in Figure 5L in both air and water.

Figure S13 .
Figure S13.Additional self-healing tests.(A) Tensile test of pristine A56 gel and selfrepaired (SR) A56 gels incubated for 4 h and 24 h.(B) Tensile results of pristine and SR M47 gel.

Figure S14 .
Figure S14.Pictures of the M53 scaffold before and after load tests.No obvious damaged was observed.

Table S1 .
A list of all the recipes of precursor.