A water-recyclable, robust, and self-healing sugar-based supramolecular network enabled by Maillard-analogous initialization of polymerization

Abstract

Crosslinked functional polymers exhibit exceptional mechanical and chemical properties critical for applications spanning biomedical engineering, advanced adhesives, and self-healing materials. However, challenges in recycling, either due to irreversible crosslinks or, in the case of covalent adaptable networks (CANs), limited solid-state plasticity that typically requires catalysts, significantly restrict sustainability. To address these limitations, we present a novel water-mediated polymerization strategy inspired by the radical-generating mechanism of the Maillard reaction, utilizing maltose as both an initiator and a functional side group in a simple, catalyst-free, aqueous reaction with acrylamide (AAm). This mild, one-pot reaction occurs below 100 °C, forming adaptively functionalized supramolecular networks (AFSNs) that form supramolecular networks through hydrogen bonding and display dynamic imine linkages to the maltose side chains supporting self-healing and re-shaping. These elastomers are characterized by impressive mechanical strength (up to 5 MPa tensile strength), high elongation (up to 1000%), notable fracture energy (36 kJ m⁻²), robust adhesive performance (up to 4.8 MPa), and rapid self-healing capability at room temperature. Crucially, the elastomer's supramolecular network can be fully and repeatedly dissolved and reprocessed using only water, preserving mechanical integrity without chemical degradation. This sustainable approach provides a practical solution for synthesizing and recycling high-performance crosslinked materials while eliminating environmental hazards, guiding the future development of green polymer chemistry and functional material design.

---

New concepts

This work introduces a transformative concept for green polymer chemistry by harnessing a Maillard-analogous, water-mediated polymerization process in which a natural reducing sugar, maltose, functions simultaneously as a radical initiator and a functional side group. This dual role enables the catalyst-free, initiator-free, and solvent-free polymerization of acrylamide in water under mild conditions, forming an adaptively functionalized supramolecular network stabilized by dynamic imine linkages and dense hydrogen bonding. This approach overcomes key limitations of current covalent adaptable networks, which rely on catalysts or elevated temperatures for recycling and reshaping. The resulting elastomers exhibit room-temperature self-healing, complete water-driven recyclability, and compound recycling while maintaining exceptional mechanical strength (∼5 MPa), high stretchability (∼1000%), and strong adhesion (∼4.8 MPa). Importantly, the polymer can be repeatedly dissolved, remolded, and reassembled using only water, eliminating the need for hazardous chemicals or complex processing. This concept provides new insight into materials design by demonstrating that edible sugars can serve as intrinsic molecular engines that drive polymerization, crosslinking, and recyclability. It opens a new paradigm for sustainable material development where renewable molecules govern both synthesis and lifecycle, advancing the convergence of green chemistry, high-performance polymers, and circular materials science.

1. Introduction

Functional polymers, particularly those engineered for advanced applications such as biomedical devices,¹ self-healing materials,^2,3 adhesives,^4,5 and responsive coatings,^6,7 frequently rely on reversible bond formation, which can be covalent (covalent adaptable networks (CANs)) or non-covalent. In addition, polymerization, network formation, and/or processing, including solution casting,⁸ electrospinning,^9,10 spin coating,¹¹ and printing technologies,^12,13 may require the use of organic solvents. In addition to the associated costs, conventional organic solvents such as toluene, acetone, dichloromethane (DCM), and tetrahydrofuran (THF) pose significant environmental and health risks, as these solvents are classified as volatile organic compounds (VOCs) with documented toxicological effects.^14,15 In terms of progressing sustainable polymers, the use of bio-based building blocks and enabling recycling without employing potentially hazardous organic solvents continue to be major challenges. While CANs represent a significant advancement in polymer technology due to their unique combination of robust mechanical properties derived from covalent crosslinking and their intrinsic ability to undergo dynamic reconfiguration through reversible covalent bonds¹⁶ enabling recycling and self-healing,^17,18 their catalyst and temperature-related solid-state plasticity limit their practical applicability and ease of use.^19–21

We envisioned a biobased polymer system that combines the ability to use water as a solvent for synthesis with recycling, self-healing and re-shaping capability. Our strategy to reach this goal is an adaptively functionalized supramolecular network (AFSN), i.e. a polymer whose side groups are attached via dynamic or exchangeable covalent bonds, while the side groups shall in addition form a supramolecular network through hydrogen bonds. When using a polymer backbone with sufficient water solubility, the supramolecular network shall be water soluble while the dynamic covalent bonds to the side chains and hydrogen bonding network should result in re-shapability and self-healing under dry or low-water containing conditions.

The concept for putting this strategy into practice was inspired by the Maillard reaction. The Maillard reaction involves the non-enzymatic conversiton of amine-containing compounds (e.g., amino acids, peptides, or proteins) and the reduction of compounds such as aldehydes or acyloins (e.g. of aldoses and ketoses) and has a special relevance in food chemistry for browning and formation of flavors. Mechanistically, in the first step, an imine is formed that is subsequently converted to more stable compounds in complex equilibria involving re-arrangement reactions (Amadori and Heyns rearrangement) as well as (hetero)cyclizations, partially via radical mechanisms. While the Maillard reaction is defined through reactions of amines with carbonyl compounds, amide nitrogen atoms can also react with carbonyl compounds,²² giving aminals or, depending on the structure of the substrates, potentially acylimines (also occurring as putative intermediates) or enamides. Hence, we proposed to use acrylamide as a polymerizable amide that can be sustainably sourced from renewable biomass feedstocks²³ and couple it in a Maillard-like reaction. Maltose, a naturally abundant disaccharide rich in hydroxyl groups, was strategically chosen due to its high water solubility and ability to form extensive hydrogen bonding. These attributes collectively enhance both the sustainability and performance of the resulting polymer. While maltose predominantly forms a stable cyclic pyranose form, a small fraction spontaneously converts into the open-chain (aldehyde) form, exposing reactive aldehyde groups required for the above described reaction.²⁴ Furthermore, beyond forming covalent bonds, maltose is known to generate radical species in situ via the Maillard reaction.^25–28

These radicals can initiate polymerization reactions in monomers containing vinyl (C=C) functionalities, such as AAm, and in the material described here, this will result in a fully bio-based hydrogel. This sustainable, catalyst-free method entirely eliminates the use of toxic solvents and external initiators, using water simultaneously as a benign reaction solvent and processing medium, facilitating not only the polymerization itself but also subsequent reshaping, recycling, and self-healing processes, entirely eliminating the need for toxic organic solvents or external additives. The approach presented herein leverages the intrinsic dynamic chemistry of imine formation, which, combined with extensive hydrogen bonding derived from sugar moieties, will yield polymers with tunable mechanical performance, robust adhesion, and excellent recyclability. The sustainability and practical advantages of this method arise from the combination of a fully biomass-based polymer network with self-healing ability and reshaping capability under mild conditions, a combination that, to our knowledge, has not been reported previously.^29–42 To date, imine-based networks have typically required acidic conditions to achieve efficient self-healing and chemical recycling, representing a major limitation for broader applicability. In contrast, the α-carbonyl imine (acylimine) is more electrophilic and hydrolytically labile than conventional imines, suggesting an easier exchange reaction important for the targeted self-healing and re-shaping. This water-mediated methodology significantly simplifies polymer synthesis and recycling while minimizing the environmental impact, providing a versatile platform that aligns closely with the principles of green chemistry and sustainable polymer design. In the following, the synthesis of the material with evidence of the proposed mechanism and the characterization of the thermomechanical properties of the P(Mal–AAm) network system are shown. Furthermore, the self-healing capacity, re-shaping and adhesive tests are presented.

2. Results and discussion

Design and synthesis of P(Mal–AAm)

The development of sustainable and environmentally benign polymer synthesis processes remains a critical challenge in modern polymer chemistry. To address this, we synthesized a novel waterborne imine elastomer poly(maltose-acrylamide) (P(Mal–AAm)), using maltose and acrylamide (AAm) via a simple one-pot polymerization in water at 80 °C. The synthesis procedure itself was remarkably straightforward and entirely water-based, requiring no external catalysts, radical initiators, or hazardous organics solvents (Fig. 1a, Fig. S1 and Movie S1). Experimentally, aqueous solutions of maltose and AAm were simply heated to 80 °C for 30 minutes, during which rapid polymerization occurred. Initially, the reaction involves the formation of reversible imine linkages. While, as described in the Introduction section, potentially acylimine, aminal, or enamide structures may be formed between the aldehyde group of maltose and the amide functionality of AAm and are potentially in equilibrium, for the further discussion, we concentrate on the acylimine form (subsequently in the text referred to as “imine”) that is most relevant for the reactions described below and for which we have evidence that intermediates are formed.

Fig. 1 Free radicals generated by the Maillard reaction initiate the polymerization of P(Mal–AAm). (a) Schematic illustration of the P(Mal–AAm) synthesis. (b) EPR spectra of the mixture of AAm, water, and maltose heated for different times. The hyperfine splitting constants, A_N ≈ 15.9 and A_H ≈ 22.7, indicate the presence of the hydroxyethyl radical (CH₃–RC˙–OH). (c) Evolution of storage modulus (G′), loss modulus (G″), and viscosity during the polymerization process of P(Mal–AAm). The gel point is at approximately 18 min. (d) GPC results of P40–70(Mal–AAm), indicating the formation of high molecular weight species. (e) FTIR spectra of S40–70(Mal–AAm) (the solution of maltose and AAm, the number indicates the AAm mass ratio) and P40–70(Mal–AAm) (the as-prepared product). The dashed line indicates the comparison of the characteristic peaks of C=O and C=C bonds. (f) ¹H NMR spectra of S40(Mal–AAm) and P40(Mal–AAm). (g) ¹³C NMR spectra of AAm, maltose, and P40(Mal–AAm).

Following the formation of the imine intermediates, the reaction undergoes subsequent transformations resulting in Amadori rearrangement products, typical in a Maillard reaction pathway. These intermediate glycation products have been shown previously to generate stable radicals, including pyrazinium cationic radicals, primarily through Strecker degradation pathways.^43,44 These in situ generated radicals acted as initiators to effectively trigger polymerization of AAm's C=C bonds, forming extended poly(AAm) chains onto which maltose-derived glycation products were grafted (Fig. S2). This dual process of imine formation and radical-initiated polymerization proceeded simultaneously, resulting in the rapid formation of an interconnected supramolecular polymer network. The radical intermediate generated during polymerization was conclusively verified by EPR using 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) as a spin trapper. The EPR spectra in Fig. 1b displayed a prominent radical signal that intensified over time throughout the polymerization process, providing clear evidence of ongoing radical generation. The hyperfine splitting constants, A_N ≈ 15.9 (nitrogen hyperfine coupling constant) and A_H ≈ 22.7 (hydrogen hyperfine coupling constant), corresponded accurately to the DMPO adduct of the hydroxyethyl radical (CH₃–RC˙–OH). Critically, these radical species were observed only in the presence of AAm, confirming the participation of AAm in Maillard-analogous chemistry. It should also be noted that no radicals were detected in the absence of maltose, as no polymerization occurred without its presence (Fig. S6). The GPC and EPR results from the blank experiments further support this finding, showing no evidence of polymer chain growth or radical formation comparable to that observed in the complete reaction system.

The time-dependent increase in EPR intensity indicates progressive radical generation through the Maillard-analogous pathway. During the early stage, radicals form gradually until sufficient propagation occurs, explaining the induction period observed before gelation (Fig. 1c). Once the growing polymer chains reach a critical length and crosslink density, rapid network formation takes place, leading to macroscopic gelation. Real-time rheology monitoring of the polymerization further supports this process, revealing a marked increase in the storage modulus (G′) relative to the loss modulus (G″) recorded within 20 min of reaction initiation at 80 °C, confirming transition from a liquid to solid-like elastomer. The observed increase in modulus aligns closely with the proposed simultaneous processes of imine formation and radical-initiated polymer chain growth, generating PAAm chains functionalized with maltose units via dynamically exchangeable imine bonds. These maltose units engage in extensive hydrogen bonding, which contributes to the physical crosslinking of the network, allowing it to remain soluble in water and enabling its water-mediated recyclability (Fig. 4). This novel synthetic route represents a significant advancement toward environmentally benign polymer synthesis, circumventing the traditional reliance on volatile organic solvents, hazardous catalysts, or initiators. The dual-mechanism approach leveraging Maillard-analogous chemistry simultaneously exploits the formation of dynamic covalent bonds (imine linkages) and radical polymerization, resulting in robust, fully aqueous-synthesized functional materials.

Molecular weight distribution of the P(Mal–AAm) oligomer

To investigate the impact of the acrylamide (AAm) concentration on the structural and mechanical properties of the synthesized P(Mal–AAm), a series of samples with varying acrylamide-to-maltose weight ratios (40 : 60, 50 : 50, 60 : 40, and 70 : 30, denoted as P40–P70(Mal–AAm) were prepared. Gel permeation chromatography (GPC) was employed to characterize the molecular weight distribution of these polymers, as presented in Fig. 1d. All P(Mal–AAm) samples demonstrated significantly high molecular weights, with the weight-average molecular weight (M_w) ranging from approximately 550 000 to 684 000 g mol⁻¹ and the number-average molecular weight (M_n) spanning from 240 000 to 270 000 g mol⁻¹. A clear shift toward higher M_w values is observed with increasing AAm content, reflecting the changed monomer to initiator ratio. Notably, all samples exhibited consistent molecular weight dispersity (Đ) values of approximately ∼2.2–2.5 (Table S1). This consistent Đ value indicates the uniformity and reproducibility of the synthetic procedure, which are crucial parameters influencing the mechanical performance and processability of the resulting elastomers. This uniform polymer growth translates into reproducible and reliable mechanical and functional properties, critical for potential industrial and practical applications of these environmentally sustainable polymer materials.

Structural characterization of P(Mal–AAm)

The chemical structure of P(Mal–AAm) was comprehensively confirmed through FTIR and NMR analyses (Fig. 1e–g). FTIR spectra of both the uncured reaction mixtures and the cured P(Mal–AAm) samples exhibited several characteristic peaks (Fig. S3). A broad and intense peak at around 3350 cm⁻¹ corresponds to the characteristic N–H stretching vibration of the amide group from AAm. The intensity of the broad absorption band in the range of 3500–3200 cm⁻¹ progressively diminished as the AAm content increased, consistent with reduced maltose-induced hydrogen bonding interactions. The peak observed around 3170 cm⁻¹ was attributed to the methylene (=CH₂) stretching vibrations present in AAm. As the proportion of AAm increases from P40 to P70, a clear reduction in hydrogen-bonded N–H stretching intensity was accompanied by an increase in the intensity of the =CH₂ peak, reflecting the increased AAm content. In the fingerprint region, the characteristic absorption bands associated with N–H stretching and out-of-plane bending modes of the C=C groups of AAm were clearly visible at approximately 1130 cm⁻¹ and 815 cm⁻¹, respectively. The enlarged FTIR spectra in Fig. 1e comparing the relative peak areas for C=O (1670 cm⁻¹) and C=C (1600 cm⁻¹) before and after the curing process demonstrated a pronounced reduction in the intensity of the C=C bond peak relative to the C=O peak after curing. This observation strongly supports the selective radical-induced polymerization of the C=C bonds, aligning with the proposed Maillard-like reaction-driven radical polymerization mechanism. Also cured samples contained some unreacted AAm.

Comparison between ¹H NMR spectra of the uncured reaction mixture (S40) and the cured polymer P40(Mal–AAm) in Fig. 1f reveals the emergence of new alkyl chain peaks at around 2.3 and 1.7 ppm, confirming the successful formation of the AAm polymer via free radical polymerization. The ¹H NMR spectra of all P(Mal–AAm) samples presented in Fig. S4 showed consistent spectral patterns at different AAm contents, indicating a uniform polymerization process. Fig. S5 provides detailed assignments of the chemical shifts (δ) for protons in both the uncured reaction blend of maltose–AAm and cured P(Mal–AAm) for 40 wt% AAm. Proton assignments were rigorously established based on integral values and splitting patterns arising from adjacent hydrogens. The integral value of each peak tallies with the total number of hydrogen atoms representing that particular chemical environment. The ‘roofing effect’ that arose from the second order splitting pattern was also taken into consideration to relate the multiplets of adjacent carbon's hydrogen when relevant. In the low field region, a complex multiplet at around 6.3 ppm is attributed to the vicinal hydrogen, H_c, on the centre carbon in AAm. The two complex doublets at 6.2 ppm and 5.8 ppm correspond to the geminal alkene hydrogen atoms, H_b and H_a, in AAm, respectively. H_b, being closer to the electron-withdrawing carbonyl group, is more de-shielded and appears downfield compared to H_a. The two doublets around 5.4 ppm and 5.2 ppm represent the characteristic peak of the hydrogen atom on anomeric carbon in both glucose units, H_d and H_j. Going upfield, the complex multiplets around 4.9 ppm to 3.3 ppm represent the remaining hydrogen atoms in the maltose unit, labelled H_{e,f,g,h,i,k,l,m,n,o}. The emergence of new peaks at approximately 2.3 and 1.7 ppm, corresponding to alkyl chain proton H_p,q, verified the formation of polyacrylamide via radical polymerization. The full ¹H NMR spectra of all four reaction blends and four P(Mal–AAm) samples reveal the absence of any impurity peaks in the entire ppm range, reflecting the purity of the synthesized polymer. The formation of imine linkage within the polymer networks was further substantiated by ¹³C NMR (Fig. 1g). The appearance of peaks at 180 ppm directly indicated the presence of C=N bonds, demonstrating the successful formation of imine bonds between maltose and AAm. The Amadori product and polyacrylamide were also confirmed by their respective peaks in ¹³C NMR spectra (35 ppm), as well as the proton signals of the hydrogen atom adjacent to the amine group (2.4 ppm) and the new alkyl chain (1.7 ppm) in the ¹H NMR spectra. The consistent presence of these peaks across all P(Mal–AAm) samples, with varying integration values reflecting the AAm content, is depicted in Fig. S4. These results collectively validate the polymerization mechanism involving free radical polymerization and imine formation, highlighting the successful synthesis of P(Mal–AAm).

This green polymerization process demonstrates remarkable versatility and applicability to a wide range of reducing sugars and unsaturated monomers. To further illustrate this broad applicability, acrylamide reacted individually with other reducing sugars such as glucose, fructose, and lactose under identical reaction conditions. Notably, AAm solution alone failed to cure upon heating (Fig. S6), but curing was effectively initiated upon the introduction of reducing sugars, yielding elastomers with distinct mechanical properties. This fits with the observation in the EPR experiment and with the proposed mechanism. Expanding the reaction scope, common monomers bearing C=C bonds, such as acrylic acid (AAc), (hydroxyethyl)methacrylate (HEMA), 2-hydroxyethly acrylate (HEAc), and N,N-dimethylacrylamide, were also investigated. Our findings show that monomers containing amide groups, particularly AAm and N,N-dimethylacrylamide, exhibited significantly shorter curing times in the presence of reducing sugars than AAc, HEMA and HEAc. Additionally, variations in pH of the reaction environment notably influenced polymerization efficiency and material properties of polymers. Changes in pH conditions affect the equilibrium distribution between the cyclic and open-chain (aldo) sugar forms, consequently altering the generation rate of free radicals essential for efficient polymerization. Acidic conditions (pH = 1) notably enhanced curing rates and improved mechanical properties, as reducing sugars are more prone to forming reactive intermediates in such environments. Furthermore, it has been shown that acidic conditions allow polymerization of acrylamides and acrylates without further addition of a radical initiator under certain circumstances, which might support the curing process studied here.⁴⁵ These observed enhancements were consistent across various sugar–monomer combinations, summarized comprehensively in Table S2.

Tensile properties of P(Mal–AAm)

The synthesized P(Mal–AAm) exhibited remarkable mechanical stretchability and toughness, solidifying its position as a high-performance elastomer. The tensile properties of P(Mal–AAm), tailored through varying AAm weight ratios, reveal a clear correlation between the monomer content and structural behavior, reflecting the versatility of the supramolecular polymer network. With increasing AAm proportion, polymer chain density correspondingly increased, significantly influencing the tensile characteristics and overall performance of the elastomer. At a low AAm content (up to 50 wt%), the resulting polymers P40(Mal–AAm) and P50(Mal–AAm) remained fully homogenous and transparent and exhibited remarkable elongation at break, reaching approximately 1000% (Fig. 2a and Movie S2). However, the tensile strength at these compositions remained relatively modest (below 0.6 MPa). As the AAm content increased to 60 wt% (P60(Mal–AAm)), AAm oversaturation in the matrix resulted in partial crystallization at room temperature. The formation of these crystalline domains introduced reinforcing phases within the elastomer matrix, effectively enhancing the tensile strength to approximately 1 MPa while maintaining impressive elongation capacity (Fig. 2b).

Fig. 2 Mechanical characterization of P(Mal–AAm). (a) Images of P40(Mal–AAm) stretched to 500% and 1000% elongation. (b) Tensile stress–strain curves of P40–70(Mal–AAm). (c) A comparison of fracture energy, Young's modulus and tensile strength of P40–70(Mal–AAm). (d) Elastomer with a low ratio of AAm (P20(Mal–AAm)) exhibited excellent ductility. It persists large deformation by inflation generating thin films of 3.4 µm. (e) Elongation at break and tensile strength of P(Mal–AAm) synthesized with varying water-to-maltose mass ratios of 1 : 4, 2 : 3, 1 : 1 and 3 : 2. (f) Curing time and tensile strength of P50(Mal–AAm) synthesized at various temperatures. (g) Stress–strain curves of P(Mal–AAm) bonding different metals by lap shear. (h) Adhesion strength of P(Mal–AAm) for different metals. (i) Demonstration of P(Mal–AAm) as a metal adhesive (adhesion area: 2 cm²) lifting a 75 kg weight.

At a further increased AAm content, extensive crystallization transformed the polymer into a white-colored material (P70(Mal–AAm)), markedly elevating tensile strength to around 5 MPa (Fig. S1 and S7). Despite the significant presence of rigid crystalline phases, P70(Mal–AAm) retained high elongation at break of approximately 1000%, underscoring the unique resilience of the polymer network. This balance between mechanical strength and flexibility is attributed to the dynamic interplay between imine cross-linking bonds and robust hydrogen bonding interactions within the matrix. The crystalline phases likely served as strong physical reinforcements, whereas the dynamic bonds provided effective energy dissipation and structural recovery during deformation.

Fig. 2c further compares the fracture energy and Young's modulus of various P(Mal–AAm) samples. The increase in the AAm content increased both modulus (up to 5.5 MPa) and the fracture toughness (36 kJ m⁻²). At a low AAm proportion of 20 wt%, insufficient polymer chain density prevented effective entanglement and hydrogen bonding, resulting in a viscous, fluid-like material rather than a solid elastomer. As demonstrated in Fig. 2d, this viscous P20(Mal–AAm) material exhibited exceptional ductility, demonstrated by its ability to be blown into hollow spheres with a wall thickness of 3.4 µm, as illustrated in the air gun blowing process depicted in Movie S3. Additionally, P20(Mal–AAm) exhibited ultra-high elongation properties, enabling it to be drawn into ultrafine fibers with a diameter of 5.4 µm as supported by the microscopy image in Fig. S8.

The crystalline nature of P70(Mal–AAm) was further elucidated by cyclic loading–unloading tests conducted under constant strain to assess the effect of stress-induced crystallite reorientation. As depicted in Fig. S9a, the cyclic tensile stress–strain curves for P70(Mal–AAm) exhibited a progressive upward shift with an increasing number of loading cycles (from N = 1 to N = 20). This observed trend indicates a gradual strengthening and enhancement of modulus upon repeated mechanical deformation. Such mechanical strengthening can be primarily attributed to the dynamic realignment of polymer chains and crystalline domains under repeated tensile loading. The underlying mechanism responsible for this progressive mechanical reinforcement involves the interplay of dynamic imine bonds and hydrogen-bonded cross-links. Under tensile deformation, the polymer network dynamically reorganizes, leading to improved chain alignment and crystallite orientation along the axis of applied stress. This molecular rearrangement results in reduced polymer chain slippage and greater stress-bearing efficiency, contributing to an increased modulus with successive loading cycles. Concurrently, the hysteresis area enclosed by the cyclic stress–strain curves progressively expand, signifying enhanced energy dissipation capacity. This enhanced energy absorption arises from the realignment and closer packing of crystalline domains, which increasingly engage in load distribution, thus improving overall mechanical resilience and fatigue resistance.

To gain further structural insight, small-angle X-ray scattering (SAXS) analyses were performed. Scattering patterns before and after cyclic loading (Fig. S9b and c) reveal a distinct transition from an initial isotropic (circular) scattering pattern, indicative of randomly oriented crystallites, to a more pronounced anisotropic (elliptical) scattering pattern after 20 loading cycles. This transition clearly demonstrates that repeated tensile deformation facilitates significant crystalline reorientation along the direction of applied strain, corroborating the mechanical observations. The synergistic interaction between the dynamically exchangeable imine bonds, reversible hydrogen bonding, and crystalline domains underpins the exceptional mechanical adaptability and self-strengthening properties of P(Mal–AAm) elastomers. Such distinctive features render this elastomer particularly suitable for applications demanding long-term durability and mechanical robustness, such as flexible electronics, load-bearing materials, and soft robotic components.

Additionally, the mechanical performance of P(Mal–AAm) is markedly influenced by the maltose-to-water ratio, as illustrated in Fig. 2e. At a fixed acrylamide content of 50 wt%, decreasing water content enhanced the material rigidity and improved tensile strength, peaking at approximately 0.8 MPa at a maltose : water ratio of 1 : 0.25. Slightly increasing the water content to a ratio of 1 : 0.67 transitioned the elastomer toward a more ductile regime, achieving an elongation at break of around 1000%. This enhanced elongation is attributed to optimized hydrogen bonding and physical cross-link density facilitated by maltose hydroxyl groups. However, further increasing water content (1 : 1 and 1 : 1.5) weakened intermolecular hydrogen bonding likely due to dilution effects, thus diminishing the structural integrity and reducing elongation at break to approximately 450%. Hence, fine-tuning the maltose-to-water ratio allows precise control over the mechanical strength and flexibility by modulating the density of hydrogen bonding, resulting in versatile materials suitable for diverse practical requirements. Furthermore, the curing temperature significantly impacted both reaction kinetics and resultant mechanical properties. As illustrated in Fig. 2f, elevating the curing temperature from 55 °C to 80 °C notably accelerated the polymerization process, significantly reducing curing time without adversely affecting the tensile properties of the material. However, further increasing the curing temperature to 90 °C substantially impaired mechanical properties, likely due to excessive water evaporation during polymerization, disrupting network formation. Consequently, 80 °C was identified as the optimal curing temperature, offering an ideal balance of rapid polymerization kinetics and desirable mechanical properties within a relatively short curing duration (30 min). Extending the curing time beyond 30 minutes at 80 °C (evaluated for periods of 1 h and 3 h) showed negligible improvements in mechanical performance (Fig. S10), confirming that the chosen curing protocol (80 °C, 30 min) efficiently achieves optimal mechanical outcomes without incurring unnecessary energy consumption.

Adhesive performance of P(Mal–AAm)

Beyond its excellent mechanical performance, the potential of P(Mal–AAm) as a sustainable adhesive was evaluated through adhesion tests involving various metal substrates. Adhesives are integral to numerous industrial applications and everyday life; however, conventional adhesives often require complex, high-temperature, and energy-intensive production processes, thus raising environmental and sustainability concerns.⁴⁶ In contrast, P(Mal–AAm) is a more sustainable and practical alternative, as its preparation involves a simple, rapid, catalyst-free, and energy-efficient polymerization conducted at a relatively mild temperature of 80 °C for merely 30 min. Experimental results demonstrate the robust adhesion performance of P50(Mal–AAm), achieving impressive adhesion strengths ranging from 2.5 MPa to 4.8 MPa across diverse metal substrates (Fig. 2g and h). This variation in adhesion performance is likely influenced by differences in surface characteristics such as roughness, surface energy, and chemical compatibility between the adhesive polymer and substrates. For instance, smoother substrates like stainless steel result in lower mechanical interlocking compared to rougher surfaces like aluminum, which exhibit more pronounced topography to facilitate superior mechanical anchoring. Remarkably, the highest adhesion strength of 5 MPa was observed for aluminum substrates. Highlighting the exceptional practical potential of P(Mal–AAm), Fig. 2i shows the material's remarkable load-bearing capability, with two aluminum substrates bonded over an adhesion area of merely 2 cm² successfully supporting a 75 kg load. Such extraordinary adhesion strength arises not only from effective mechanical interlocking but also from the strong interfacial hydrogen bonding between the polymer and the aluminum substrate, coupled with the cohesive strength of the elastomeric supramolecular polymer network itself. These combined interactions effectively distribute stress across the interface, significantly contributing to the adhesive's high-performance characteristics. Another notable advantage of P(Mal–AAm) is its complete recyclability through a simple water-mediated dissolution process, aligning with sustainable chemistry principles. The hydrogen bond network disaggregates in the presence of large amounts of water, enabling effortless detachment of substrates without damage, as demonstrated in Movie S4. There might also be some hydrolysis of imine bonds. This unique property enables the full recovery and reuse of both adhesive materials and substrates, thereby completely avoiding harsh solvents or energy-intensive recycling methods typically required for conventional adhesives. Thus, the intrinsic recyclability and easy reversibility of P(Mal–AAm) provide substantial environmental benefits, making it particularly suitable for applications where temporary or reversible adhesion is required, such as in reconfigurable manufacturing, sustainable packaging, and maintenance and repair applications.

Self-healing properties of P(Mal–AAm)

The inherent ability of biological tissues to self-repair after experiencing mechanical deterioration has long inspired the development of self-healing materials. Imine-based networks are among the most extensively studied CANs for self-healing purposes; however, conventional imine-based polymers typically require external stimuli such as elevated temperatures or mechanical compression to achieve efficient healing.⁴⁷ In stark contrast, the P(Mal–AAm) developed herein exhibited exceptional self-healing capability, with two scissor-cut specimens rapidly merging into a single, cohesive material within just three minutes at room temperature, as visually shown in Movie S5. The outstanding self-healing performance arises primarily from the synergistic action of dual dynamic mechanisms: dynamic imine exchange and abundant intermolecular hydrogen bonding, as illustrated in Fig. 3a. The presence of maltose substantially increases the density of hydrogen-bonding sites, which function as dynamic and reversible cross-linkers within the polymer network. The self-healing process was further enhanced by hydrogen bond-assisted dynamic imine exchange. The presence of hydrogen bonding positions the amide group of AAm into close proximity to the imine exchange sites, thereby facilitating nucleophilic attack of the amide nitrogen onto the electrophilic carbon atom of the imine bond (Fig. S11). This dynamic exchange process is feasible due to the intrinsic susceptibility of imine bonds to nucleophilic nitrogen attack relative to carbonyl groups.⁴⁸ As a result, a transient aminal intermediate forms, subsequently undergoing rapid proton exchange to regenerate new imine and amide groups, thus completing the chemical healing process. Concurrently, the diffusion and entanglement of polymer chains across the damaged interface further strengthened the self-healing capability. This physical intertwining of chains improved the mechanical stability of the healed regions.

Fig. 3 Self-healing properties of P(Mal–AAm). (a) Schematic illustrating the self-healing mechanism of P(Mal–AAm) through hydrogen bonding and dynamic imine exchange. (b) Different ratios of P(Mal–AAm) could be assembled by the healing. The assembled product maintains excellent tensile properties. The interfacial strength is greater than the AAm P40 strength. At 700% elongation, the crack appeared within the P40(Mal–AAm) rather than at the healed interface. (c) The self-healing degree (defined as the ratio of the healed sample strength over the original sample strength) of P40–70(Mal–AAm) over time at 25 °C and 60 °C. (d) Tensile stress–strain curves of the original and self-healed P60(Mal–AAm).

In the self-healing experiment, two dog-bone-shaped specimens (one dyed red for visual contrast) were cut into halves and then combined together at room temperature, as depicted in Fig. S12. The inherent mobility of the polymer chains enables diffusion across the damaged interface, swiftly re-establishing disrupted hydrogen bonding and imine linkages. Remarkably, the cut fragments autonomously healed into a single cohesive piece within just one minute, without the need for external stimuli. Additionally, self-healing tests were also performed by combining P(Mal–AAm) specimens with varying AAm contents, specifically P40(Mal–AAm) and P70(Mal–AAm) at 60 °C for 12 h, resulting in fully integrated and structurally coherent samples (Fig. 3b and Movie S6). Upon cooling, the P70(Mal–AAm) segment reverted to its original opaque white appearance, visually revealing predefined patterns. When subjected to tensile stretching, distinct deformation behaviors emerged between the two joined segments due to differences in their moduli. Notably, at a stretch ratio λ of 7.5, cracks appeared within the softer P40(Mal–AAm) segment rather than at the healed interface, indicating the robustness and mechanical integrity of the healed boundary. In Fig. 3c, the efficiency of the self-healing process was influenced by both the weight ratio of AAm and the surrounding temperature. Healing efficiency (defined as the ratio of the tensile strength of healed samples to that of original, undamaged samples) was optimized by introducing a thin water layer at the interface to prevent water loss. At room temperature (25 °C), P40(Mal–AAm) and P50(Mal–AAm) exhibited nearly complete restoration of their original tensile strength within 48 h (Fig. S13). Conversely, samples with a higher AAm content (60 and 70 wt%) demonstrated relatively low room temperature healing efficiency (∼40% strength recovery), attributed to the presence of crystalline domains, which restricted chain mobility and hindered effective reconstruction of hydrogen bonding at the interface.

Increasing the temperature to 60 °C significantly improved the self-healing efficiency for all P(Mal–AAm) compositions. At higher temperature, the mobility and flexibility of the polymer chains increased, accelerating chain reorientation, interfacial diffusion, and the reformation of dynamic bonds. Fig. 3d presents the tensile stress–strain curves for P60(Mal–AAm) samples healed at 60 °C at varying times of healing. The progressive recovery in both tensile strength and elongation with increasing healing duration ultimately achieved nearly 100% restoration of original mechanical performance after 12 h. Similarly, P70(Mal–AAm) samples also attained complete recovery of mechanical properties under these elevated-temperature conditions, as corroborated by additional tensile test results shown in Fig. S13. Collectively, these results confirm that the robust and efficient self-healing behavior of P(Mal–AAm) arises from the well-orchestrated interplay between dynamic imine chemistry, strong hydrogen bonding, and enhanced chain mobility, demonstrating its immense potential for sustainable applications requiring high-performance self-healing capabilities.

Chemical recycling of P(Mal–AAm)

Previous studies have demonstrated the chemical recyclability of hydrogen-bonded supramolecule and imine networks, which often relies heavily on acidic solutions, organic solvents, or additional amine-based reagents (Tables S3 and S4). In this work, we achieve chemical recycling of P(Mal–AAm) solely through water, as larger amounts of water lead to cleavage of the inter-strand hydrogen bonding and hence dissolution of the maltose-functionalized P(Mal–AAm). In this process, the inherent susceptibility of imine bonds to hydrolysis may also play a role.⁴⁹ As illustrated in Fig. 4a and b, the waterborne imine network within P(Mal–AAm) readily underwent hydrolysis in water at ambient temperature, cleaving their original carbonyl and amide functional groups. Compared to imine, acylimine features an additional adjacent C=O group, which makes the carbon in the C=N bond more susceptible to nucleophilic attack by water. This increased reactivity explains why P(Mal–AAm) can undergo chemical recycling under milder conditions. As a result, the polymer network disassembles into its constituent water-soluble components. Simultaneously, the addition of excess water also disrupts intermolecular hydrogen bonding by diluting the interaction between maltose hydroxyl groups within the polymer matrix, significantly compromising structural integrity and facilitating dissolution. Upon subsequent remolding and controlled water evaporation, the polymer network spontaneously reassembles through reformation of dynamic imine linkages. This cycle of disassembly and reassembly highlights the dynamic and reversible nature of the imine chemistry employed, enabling straightforward, water-mediated recyclability without the need for additional chemicals, elevated temperatures, or energy-intensive procedures commonly associated with traditional recycling methods (Fig. 4c and Movie S7). The recycling process comprised three main steps: (i) dissolution of the polymer in water, (ii) remolding of the dissolved polymer solution into a defined shape, and (iii) evaporation of the excess water to attain a final moisture content of approximately 5% as confirmed by TGA in Fig. S14. Notably, this straightforward recycling protocol was reproducibly conducted over five consecutive cycles without any significant deterioration in visual appearance or structural integrity, as confirmed by comparative images taken after multiple recycling events. The recycled P40–70(Mal–AAm) materials exhibited tensile strengths of approximately 5, 6, 10, and 14 MPa, respectively, with elongation at break reaching up to 400% (Fig. S15). Owing to the reduced moisture content, the recycled P40–70(Mal–AAm) samples consistently showed higher tensile strength than their original counterparts, with strength recovery ranging from 300% to 1500% relative to the initial performance. This enhancement in tensile strength, accompanied by a moderate reduction in elongation, is likely due to gradual water loss during recycling, which diminishes hydrogen bonding and leads to a stiffer polymer network.

Fig. 4 Recyclability and adhesive performance of P(Mal–AAm). (a) Schematic illustration of the recycling mechanism of P(Mal–AAm), showing dissolution in water at 25 °C and reformation upon water evaporation. (b) Chemical scheme of the reversible imine-based network. (c) Photographs of the P(Mal–AAm) recycling process through water-assisted dissolution, remolding, and demolding over repeated cycles. (d) Schematic representation of the closed-loop compound recycling process, where equal proportions of P40–70(Mal–AAm) inputs yield predictable mechanical properties based on their composition. (e) Tensile stress–strain curves of P(Mal–AAm) after successive compound recycling cycles. (f) Comparison of tensile strength after each compound recycling cycle.

To further evaluate its versatility, we explored the compound recycling capability of P(Mal–AAm) by mixing equal portions of P40, P50, P60, and P70 formulations, as illustrated in Fig. 4d. Compound recycling is a strategy in which multiple polymer variants are dissolved, blended, and remolded to yield a recycled material whose mechanical properties reflect the weighted average of the input formulations. This approach enables tunable mechanical performance in the recycled product, offering tailored properties that depend on the composition of the input mixture. The resulting compound-recycled P(Mal–AAm) demonstrated tensile properties closely matching theoretical tensile strength predictions of ∼8 MPa, confirming that the final material accurately reflects the proportional contributions of its constituents (Fig. 4e and f). Although minor variations in mechanical properties were observed over successive recycling cycles, these fluctuations were negligible, indicating excellent retention of performance. These slight differences are attributed to residual moisture within the matrix, which subtly affects hydrogen bonding and, in turn, the mechanical response of the network. The demonstrated compound recyclability of P(Mal–AAm) highlights its potential as a sustainable and environmentally friendly polymer, offering energy-efficient processing without the need for waste sorting.

In the recycling process, we observed that the loss of water content during evaporation led to an improvement in tensile strength. To investigate this phenomenon further, we prepared a series of P50 samples with varying water contents, denoted as the P50-W series (Table S5). In these samples, the maltose : AAm ratio was kept constant, while the water content was varied from 10 to 30 wt%, resulting in samples labeled P50-W10 to P50-W30. Mechanical testing (Fig. S16) revealed a decrease in tensile strength with increasing water content. Notably, P50-W20 exhibited the highest elongation, which can be attributed to a moderate concentration of hydrogen bonding, consistent with our previous observations. Lap shear testing of the P50-W series further demonstrated that increasing water content led to a reduction in adhesion strength, likely due to the dilution of hydrogen bonds. The self-healing properties of the P50-W series were also evaluated at 25 °C over different durations (Fig. S17). P50-W10 demonstrated approximately 40% tensile recovery and 75% elongation recovery, while P50-W30 displayed significantly improved recovery, with about 81% tensile recovery and 85% elongation recovery. These results suggest that the higher water content enhances the material's self-healing ability under the same recovery conditions. Chemical recycling of the P50-W series, accompanied by a decrease in the water content (confirmed by TGA), resulted in enhanced tensile strength in all recycled samples, which can be attributed to the increased concentration of hydrogen bonds (Fig. S18 and S19). Additionally, we examined the tensile properties of the P50-W series in different moisture environments, with all samples showing reduced tensile strength but increased elongation at break when exposed to high humidity (RH = 90%), as shown in Fig. S20. This indicates that external moisture affects the intrinsic hydrogen bond concentration of the network, which is expected since the network is water-recyclable. These findings underscore the significant influence of water and moisture content on both the mechanical performance and self-healing properties of the material. Specifically, the higher water content enhances the self-healing recovery, while exposure to moisture promotes elongation at break but reduces tensile strength due to the dynamic nature of hydrogen bonding.

3. Conclusions

In summary, this study demonstrates that natural reducing sugars can effectively serve as sustainable organic initiators for the green synthesis of polymers functionalized with maltose through dynamic imine bonds. Due to the ability of extensive hydrogen bonding especially between the carbohydrate units, a supramolecular elastomer is formed. Utilizing edible, biocompatible sugars such as maltose alongside acrylamide, we have developed a simple, water-mediated polymerization process that operates under mild conditions (below 100 °C), avoiding toxic solvents, catalysts, and external initiators. The resultant polymers exhibit remarkable mechanical properties—including high tensile strength (up to 5 MPa), extraordinary elongation (up to 1000%), and excellent fracture toughness (36 kJ m⁻²), originating from abundant hydrogen bonding and dynamic imine linkages provided inherently by the sugar initiators. Moreover, these sugar-derived linkages enable excellent self-healing capability and recyclability, as demonstrated by rapid autonomous healing at room temperature and complete chemical recycling through simple dissolution and reprocessing in water. Additionally, the compound recycling capability of P(Mal–AAm) was explored by mixing equal portions of P40–70 formulations. This approach yielded a recycled material with tunable mechanical properties, reflecting the weighted average of the input formulations. The compound, recycled P(Mal–AAm), demonstrated tensile properties closely matching theoretical predictions (∼8 MPa), highlighting its versatility and sustainability. Despite minor fluctuations in mechanical properties over successive recycling cycles, excellent performance retention was observed, further emphasizing the material's potential for environmentally friendly and energy-efficient processing without the need for waste sorting. Overall, this innovative approach not only addresses environmental and health concerns posed by conventional polymer synthesis methods but also introduces a versatile mechanism to impart robust mechanical performance and dynamic healing capabilities into polymeric materials. Our findings establish a strong foundation and inspire future development of diverse sustainable polymers, highlighting the unique capabilities of natural sugar initiators to advance both materials science and green chemistry.

4. Experimental

Synthesis of P(Mal–AAm)

P(Mal–AAm) samples were prepared by simply mixing water, maltose (Macklin, M874782) and acrylamide (AAm, Aladdin, M128783) in various ratios via a one-pot melt reaction. Initially, the monomers were dissolved by stirring at 80° for 10 min. The mixture was then allowed to further react under the same conditions for an additional 30 min, enabling in situ curing without the need for external catalysts. For the synthesis of all samples, a water-to-maltose weight ratio of 2 : 3 was employed to ensure complete solubility of the monomers. AAm was added in 40, 50, 60, and 70 wt% relative to maltose to produce various imine elastomer samples, designated as P40–70(Mal–AAm). Portions of the mixture were taken out prior to the curing process for further analysis and are referred to as S(Mal–AAm). The resulting P(Mal–AAm) samples exhibited varying physical appearances, ranging from a colorless liquid to a creamy white liquid as the AAm weight ratio increased. ¹H NMR (600 MHz, D₂O): δ 6.10–6.20 (m, 4H), 5.71 (d, 1H), 5.30 (d, 1H), 5.12 (d, 1H), 4.65 (d, 1H), 4.55 (d, 1H), 3.85–3.17 (m, 10H), 2.24–2.09 (m, 2H), 1.93 (m, 1H), 1.67 (m, 1H), 1.55 (m, 1H) ppm.

Maillard-analogous reaction on other reducing sugars and monomers

In this part, maltose was replaced with other reducing sugars, including glucose (Macklin, G6172), fructose (Macklin, F875004), and lactose (Aladdin, L103493), while AAm was replaced with acrylic acid (Sigma-Aldrich, 8001810100), (hydroxyethyl)methacrylate (Sigma-Aldrich, 8005880250), 2-hydroxyethly acrylate (Aladdin, H104535) and N,N-dimethylacrylamide (Macklin, N828383). In each case, the mass ratio of monomer : reducing sugar : water was kept at 5 : 3 : 2. The mixture was dissolved at 80 °C and then placed in an oven at the same temperature. For some formulations, curing was only successful at pH = 1, whereas others cured at pH = 7. The curing time, mechanical properties, and images of the products are presented in Table S2.

Rheological test

The viscosity, storage modulus (G′) and loss modulus (G″) of S40(Mal–AAm) during the curing process were measured using an Anton-Paar MCR302 rheometer. An amplitude sweep was performed at a frequency of 1 Hz to evaluate the material's viscoelastic behavior. The sample was carefully prepared to ensure uniformity and to be free of air bubbles, and the temperature was maintained constant at 80 °C throughout the test.

Characterization of P(Mal–AAm)

P40–70(Mal–AAm) and AAm were dissolved in water for gel permeation chromatography (GPC) analysis, which was conducted using a GPC system (Agilent 1260 Infinity system) with two ResiPore columns (250 mm × 4.6 mm, model PLgel5 µm MiniMIX-C) at a flow rate of 0.6 mL min⁻¹. The Fourier Transform Infrared (FTIR) spectra of S40–70(Mal–AAm) and P40–70(Mal–AAm) were recorded on a Thermo Fisher Scientific Nicolet iS20 in attenuated total reflection (ATR) mode, over the range of 4000 to 400 cm⁻¹. ¹H NMR and ¹³C NMR spectra were obtained using a Bruker 600 MHz. P40–70(Mal–AAm) samples were dissolved in D₂O for NMR analysis. For electron paramagnetic resonance (EPR) measurements, a P40(Mal–AAm) solution was heated for 0, 10, and 20 min and analyzed using a Bruker EMXplus-6/1 with DMPO as a free radical trapping agent. The X-ray diffraction (XRD) patterns of P40–70(Mal–AAm) were recorded over a 2θ range of 5° to 80° using a Rigaku Ultima IV diffractometer. Changes in the crystalline orientation of P70(Mal–AAm) before and after cyclic stretching were examined using small-angle X-ray scattering (SAXS) with a Xenocs Xeuss 2.0. Scanning electron microscopy (SEM) of gold sputtered fractured slices conducted on a ZEISS Sigma 300 revealed morphological changes in P40(Mal–AAm) and P70(Mal–AAm).

Mechanical testing

The mechanical properties of P(Mal–AAm) were tested using a Zwick/Roell ProLine 10 kN static testing system with a 200 N load cell. Samples for the uniaxial tensile test and the cyclic tensile test were dog-bone shaped with a length of 50 mm, a width of 4 mm, and a thickness of 1 mm. The initial gauge length was 12 mm, and the tensile rate was 100 mm min⁻¹. The tensile rate was adjusted to 1 mm min⁻¹ to calculate the Young's modulus. The strain range of 0.5% to 1% was selected to calculate the Young's modulus. To assess toughness, samples were cut into rectangular shapes with dimensions of 100 mm × 60 mm × 1 mm. The initial gauge length was 12 mm, and the tensile rate was 50 mm min⁻¹. To get the elastic energy stretch density function w(λ), the intact sample was monotonically stretched, and the stress–strain curve was recorded. A 2 mm crack was then introduced at the edge of the sample, and it was stretched to measure the maximum stretch λ_max. The toughness was calculated by w(λ_max) × H, where H is 12 mm.

Ductility demonstration of P20(Mal–AAm)

P20(Mal–AAm) has excellent ductility. It was sealed at the muzzle of an air gun and then inflated. P20(Mal–AAm) expanded into a spherical shape, and the thin wall of the sphere could reach a thickness of less than 5 µm.

Self-healing of P(Mal–AAm)

P40–70(Mal–AAm) were each shaped into dog-bone specimens with dimensions of 50 mm in length, 4 mm in width, and 1 mm in thickness. The samples were cut in the middle and the cut surfaces were placed together for healing. P40(Mal–AAm) and P50(Mal–AAm) required 48 h at room temperature for full self-healing, while P60(Mal–AAm) and P70(Mal–AAm) required 12 h at 60 °C for full self-healing. Since the healing time was relatively long, to prevent dehydration of the material, a thin layer of water was applied to the cut surfaces with a cotton swab, and the material was fully sealed during storage. Samples with different healing times and healing temperatures were subjected to mechanical tests.

Chemical recycling of P(Mal–AAm)

Equal portions of the four formulations (P40, P50, P60, and P70) were first dissolved together in deionized water at room temperature under constant stirring to achieve a homogeneous solution. The weight ratio of P(Mal–AAm) to water was 1 : 5. The mixed polymer solution was then poured into Teflon molds of desired geometry and placed in an oven at 40 °C to allow gradual water evaporation. Evaporation was continued until the residual water content reached approximately 5%, as confirmed by weight loss analysis. The resulting dried film was carefully demolded and subjected to mechanical testing. After each cycle, the used material was redissolved in water, remixed, and remolded following the same procedure.

Adhesion strength test

The single-lap shear tests were conducted to measure the adhesive strength of the samples. 0.2 g of the Mal–AAm oligomer was applied in between two surface-treated stainless steel substrates with dimensions of 15 mm × 50 mm × 2 mm in size, and the bonded area is 15 mm × 10 mm. The adhesive is formed by heating the overlapped specimen at 80 °C for 30 min. The samples were clamped in the tensile testing machine, and a constant extension rate of 50 mm min⁻¹ was applied. The stress–strain curve was recorded during the test, and shear strength was calculated by dividing the maximum stress observed at the point of debonding by the bonded area. The single-lap shear tests were repeated on different substrates, zinc, iron, titanium, and aluminum, to assess adhesion on different metals.

Author contributions

S. L. and T.-J. L.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, software, validation, visualization, writing – original draft, and writing – review and editing. S.-H. N. and Y. T.: investigation, methodology, and writing – original draft. Z. M.: investigation, software, and writing – original draft. T. F.: data curation, software, and writing – original draft. L. W.: investigation, methodology, and writing – original draft. J.-W. W. and J. L.: investigation, visualization, and writing – original draft. K.-Y. L. and M.-R. W.: validation, writing – original draft, and writing – review and editing. A. N.: conceptualization, validation, and writing – original draft. T.-W. W.: conceptualization, investigation, project administration, supervision, writing – original draft, and writing – review and editing. T. L.: funding acquisition, project administration, resources, supervision, writing – original draft, and writing – review and editing. X. Y.: conceptualization, methodology, project administration, funding acquisition, resources, supervision, writing – original draft, and writing – review and editing. W. Y.: conceptualization, project administration, supervision, writing – original draft, and writing – review and editing.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: GPC results. Alternative experiment using various sugars and monomers. Synthesis process illustration of P(Mal-AAm). Mechanisms of Amadori product and acylimine formation via Maillard-analogous pathway. FTIR, ¹H NMR, and ¹³C NMR spectra. Analysis of blank experiment. XRD results and SEM images. Microscopic images of P20 thin film. Cyclic stress–strain curves of P70. Mechanism of dynamic imine exchange. Self-healing demonstration of P40. Stress–strain curves of self-healed and recycled samples. Comparison of H-bonded supramolecular and imine-based networks. TGA results. Stress–strain curves of P50-W with different water content. The effect of environmental moisture on tensile properties. See DOI: https://doi.org/10.1039/d5mh01828e.

Acknowledgements

We would like to acknowledge the financial support from the following programs: National Natural Science Foundation of China (no. 12102388, 12472167, T2125009, and 92048302), the Key Research and Development Project of Zhejiang Province (2022C01022) and Fundamental Research Funds for the Central Universities (no. 226-2022-00141).

References

1. M.-S. Hosseini, I. Amjadi, M. Mohajeri, M. Zubair Iqbal, A. Wu and M. Mozafari, in Advanced Functional Polymers for Biomedical Applications, ed. M. Mozafari and N. P. Singh Chauhan, Elsevier, 2019, pp. 1–20.
2. M. Wang, P. Zhang, M. Shamsi, J. L. Thelen, W. Qian, V. K. Truong, J. Ma, J. Hu and M. D. Dickey, Nat. Mater., 2022, 21, 359–365.
3. Y. Wang, X. Huang and X. Zhang, Nat. Commun., 2021, 12, 1291.
4. F. Awaja, M. Gilbert, G. Kelly, B. Fox and P. J. Pigram, Prog. Polym. Sci., 2009, 34, 948–968.
5. Y. Zhao, Y. Wu, L. Wang, M. Zhang, X. Chen, M. Liu, J. Fan, J. Liu, F. Zhou and Z. Wang, Nat. Commun., 2017, 8, 2218.
6. T. Yimyai, D. Crespy and M. Rohwerder, Adv. Mater., 2023, 35, 2300101.
7. S. Nagappan, M. S. Moorthy, K. M. Rao and C.-S. Ha, in Industrial Applications for Intelligent Polymers and Coatings, ed. M. Hosseini and A. S. H. Makhlouf, Springer International Publishing, Cham, 2016, pp. 27–49.
8. U. Siemann, in Scattering Methods and the Properties of Polymer Materials, ed. N. Stribeck and B. Smarsly, Springer Berlin Heidelberg, Berlin, Heidelberg, 2005, pp. 1–14.
9. C. J. Luo, M. Nangrejo and M. Edirisinghe, Polymer, 2010, 51, 1654–1662.
10. J. Lasprilla-Botero, M. Álvarez-Láinez and J. M. Lagaron, Mater. Today Commun., 2018, 14, 1–9.
11. D. P. Birnie, in Chemical Solution Deposition of Functional Oxide Thin Films, ed. T. Schneller, R. Waser, M. Kosec and D. Payne, Springer, Vienna, 2013, pp. 263–274.
12. Z. Miao, J. Seo and M. A. Hickner, Polymer, 2018, 152, 18–24.
13. B. Philip, E. Jewell and D. Worsley, J. Coat. Technol. Res., 2016, 13, 911–920.
14. Z. Li, K. H. Smith and G. W. Stevens, Chin. J. Chem. Eng., 2016, 24, 215–220.
15. E. David and V.-C. Niculescu, Int. J. Environ. Res. Public Health, 2021, 18, 13147.
16. D. Montarnal, M. Capelot, F. Tournilhac and L. Leibler, Science, 2011, 334, 965–968.
17. C. J. Kloxin and C. N. Bowman, Chem. Soc. Rev., 2013, 42, 7161–7173.
18. C. N. Bowman and C. J. Kloxin, Angew. Chem., Int. Ed., 2012, 51, 4272–4274.
19. X. Chen, J.-W. Wong, J. T. Low, T.-J. Lok, Y. Xue, Z. Zeng, K. Zhang, Y. Shen, S. Li, H. Zhou, Q. Zhao, T.-W. Wong, T. Li and W. Yang, ACS Macro Lett., 2024, 13, 1037–1042.
20. S.-Y. Chien, J. Wang and Y.-L. Liu, ACS Appl. Polym. Mater., 2024, 6, 9191–9199.
21. S. Debnath, S. Kaushal and U. Ojha, ACS Appl. Polym. Mater., 2020, 2, 1006–1013.
22. A. Bayer and M. E. Maier, Tetrahedron, 2004, 60, 6665–6677.
23. J. Keramat, A. LeBail, C. Prost and N. Soltanizadeh, Food Bioprocess Technol., 2011, 4, 340–363.
24. B. Kramhoeller, M. Pischetsrieder and T. Severin, J. Agric. Food Chem., 1993, 41, 347–351.
25. G. P. Rizzi, Food Rev. Int., 2003, 19, 375–395.
26. R. L. Roberts and R. V. Lloyd, J. Agric. Food Chem., 1997, 45, 2413–2418.
27. R. V. Lloyd, A. J. Fong and R. M. Sayre, J. Invest. Dermatol., 2001, 117, 740–742.
28. T. Hayashi and M. Namiki, Agric. Biol. Chem., 1981, 45, 933–939.
29. Ö. Dağlar, F. Eisenreich and Ž. Tomović, Adv. Funct. Mater., 2024, 34, 2408299.
30. C. Bao, J. Liu, Y. Yin, J. Liu and Z. Xin, Polym. Chem., 2024, 15, 172–180.
31. H. Ding, J. Wang, P. Yu, H. He, H. Wang, W. Zhang, L. Wang, Y. Lei and B. Yu, Chem. Eng. J., 2024, 481, 148024.
32. X.-M. Ding, L. Chen, Y.-J. Xu, X.-H. Shi, X. Luo, X. Song and Y.-Z. Wang, ACS Sustainable Chem. Eng., 2023, 11, 14445–14456.
33. J. Zhang, C. Jiang, G. Deng, M. Luo, B. Ye, H. Zhang, M. Miao, T. Li and D. Zhang, Nat. Commun., 2024, 15, 4869.
34. Q. Zhao, M. Zhang, S. Gao, Z. Pan, Y. Xue, P. Jia, C. Bo, Z. Luo, F. Song and Y. Zhou, J. Mater. Chem. A, 2022, 10, 11363–11374.
35. Y. Liu, Z. Yu, G. Lu, W. Chen, Z. Ye, Y. He, Z. Tang and J. Zhu, Chem. Eng. J., 2023, 451, 139053.
36. P. Li, J. Zhang, J. Ma, C.-A. Xu, X. Liang, T. Yuan, Y. Hu and Z. Yang, Ind. Crops Prod., 2023, 204, 117288.
37. N. Zhang, X. Pan, A. Xi, W. Chen, T. Huang and Y. Zeng, Polym. Chem., 2024, 15, 3287–3299.
38. P. Yu, H. Wang, T. Li, G. Wang, Z. Jia, X. Dong, Y. Xu, Q. Ma, D. Zhang, H. Ding and B. Yu, Adv. Sci., 2023, 10, 2300958.
39. X. Liu, E. Zhang, J. Liu, J. Qin, M. Wu, C. Yang and L. Liang, Chem. Eng. J., 2023, 454, 139992.
40. G. Li, X. Zhang, J. Huang, T. Li, S. Yang, Y. Wang, J. Jiang, B. Xia, M. Chen and W. Dong, Chem. Eng. J., 2022, 435, 134766.
41. Z. Jia, H. Wang, P. Yu, H. He, Q. Huang, W. Hong, C. Liu, Y. Shi, J. Wang, Y. Xin, X. Jia, J. Ma and B. Yu, Small, 2024, 20, 2406821.
42. K. Hong, Q. Sun, X. Zhang, L. Fan, T. Wu, J. Du and Y. Zhu, ACS Sustainable Chem. Eng., 2022, 10, 1036–1046.
43. H. Yu, R. Zhang, F. Yang, Y. Xie, Y. Guo, W. Yao and W. Zhou, Trends Food Sci. Technol., 2021, 112, 795–807.
44. M. Namiki and T. Hayashi, ACS Symp. Ser. Am. Chem. Soc., 1983, 21–46.
45. M.-N. Antonopoulou, G. R. Jones, A. A. Kroeger, Z. Pei, M. L. Coote, N. P. Truong and A. Anastasaki, Nat. Synth., 2024, 3, 347–356.
46. D. E. Packham, in Eco-efficient Construction and Building Materials, ed. F. Pacheco-Torgal, L. F. Cabeza, J. Labrincha and A. de Magalhães, Woodhead Publishing, 2014, pp. 338–367.
47. M. Ciaccia and S. D. Stefano, Org. Biomol. Chem., 2014, 13, 646–654.
48. F. Schaufelberger, K. Seigel and O. Ramström, Chem. – Eur. J., 2020, 26, 15581–15588.
49. T. D. Martins, M. T. Viciosa, M. B. Oliveira, A. Fernandes, J. F. Mano, C. Baleizão and J. P. S. Farinha, Prog. Org. Coat., 2023, 180, 107552.

---

Footnote

† Siyang Li and Tow-Jie Lok contributed equally to this work.

---