Mechanoresponsive Hydrogels Emerging from Dynamic and Non‐Covalent Interactions

Abstract

Mechanoresponsive hydrogels undergo changes in their physical and chemical properties in response to mechanical stimuli such as strain, force, or shear stress. These responses are often mediated by dynamic or non‐covalent intermolecular interactions. Unlike covalent bonds, which confer desirable mechanical strength but result in static networks, dynamic crosslinking motifs introduce reversibility that enables mechanically actuatable behaviors such as self‐healing, shear‐thinning or ‐thickening, and strain‐stiffening. This review highlights these four distinct mechanoresponsive behaviors in dynamic hydrogels, examining their underlying mechanisms, characterization methods, and emerging applications, with a focus on the critical role of dynamic interactions in enabling their mechanoresponsive properties.

Mechanoresponsive behaviors in hydrogels often arise from dynamic or non‐covalent interactions between polymer chains. This review highlights self‐healing, shear‐thinning, shear‐thickening, and strain‐stiffening hydrogels formed through dynamic crosslinking motifs, discussing their underlying mechanisms and emergent applications across biomedical and engineering contexts.

1. Introduction

Hydrogels are 3D networks of hydrated polymer chains stabilized by either chemical (covalent) or physical (non‐covalent) crosslinking interactions.^{[ 1 ]} These crosslinks render the material insoluble, resulting in hydrated, percolated networks capable of absorbing several times their dry weight in water or physiological fluids.^{[ 2} , ^{3 ]} While covalently crosslinked networks often exhibit desirable mechanical properties, such as high stiffness and strength, they typically lack responsiveness to environmental cues. In contrast, dynamic‐covalent bonds retain some of the mechanical advantages of covalent interactions while enabling equilibrium‐governed bond exchange, which can be modulated by environmental factors.^{[ 4} , ⁵ , ^{6 ]} Hydrogels formed using non‐covalent or dynamic‐covalent interactions thus exhibit dynamic behavior in response to various thermal, chemical, or mechanical stimuli.^{[ 7} , ⁸ , ⁹ , ¹⁰ , ¹¹ , ^{12 ]} Specifically, common mechanical stimuli include strain, force, pressure, and shear rate. Materials capable of sensing and responding to mechanical stimuli are ubiquitous in nature. For example, many biological materials—such as the extracellular matrix—exhibit strain‐stiffening behavior.^{[ 13} , ^{14 ]} Stiffening under applied strain enables biological communication and offers cells protection against large deformations.^{[ 15} , ^{16 ]} Similarly, shear‐thinning, characterized by a reduction in viscosity with increasing shear rate, is observed in blood as it flows through narrow vessels at high shear rates.^{[ 17 ]} Living organisms also possess the remarkable ability to self‐heal, restoring structure and function following damage. Inspired by these natural phenomena, numerous hydrogel technologies have been developed to emulate the mechanoresponsive features of biological systems using synthetic, bioinspired, or biomimetic design strategies.

With inspiration from biology, applications for mechanoresponsive hydrogels are extensive.^{[ 18 ]} Strain‐stiffening hydrogels have been used for tissue engineering, where they replicate mechanical cues from the extracellular matrix for cell communication and development.^{[ 19} , ^{20 ]} Strain‐stiffening hydrogels have also been used in the design of strain sensors, particularly for wearable and implantable flexible electronics.^{[ 21} , ^{22 ]} Force, pressure, and temperature sensors have likewise been reported from mechanically actuable hydrogels.^{[ 23} , ²⁴ , ^{25 ]} Self‐healing hydrogels offer key advantages in the context of greater device longevity and stability, especially for materials exposed to mechanical deformation.^{[ 26} , ^{27 ]} Hydrogels that self‐heal often also exhibit shear‐thinning, with the combination of both features enabling uses in applications like extrusion‐based printing and or minimally invasive injection for the delivery of therapeutics or cells, enabling materials to yield under shear and subsequently recover their properties.^{[ 28 ]}

Dynamic and non‐covalent crosslinking interactions are important for the realization of mechanoresponsive hydrogels (Figure 1A).^{[ 7 ]} Dynamic‐covalent chemistries used to crosslink hydrogels include several classical equilibrium reactions, such as reversible Diels‐Alder reactions between a diene and a conjugated alkene, Schiff‐base bonding between an aldehyde or ketone and a primary amine, thiol‐disulfide exchange reactions, and boronate esters formed by reaction of boronic acids with diols.^{[ 5} , ^{6 ]} The use of these bonds to crosslink polymer chains results in the formation of hydrogels with concomitant dynamic mechanical properties,^{[ 29 ]} with opportunities for use as dynamic biomaterials.^{[ 30 ]} Likewise, many non‐covalent and supramolecular interactions between molecules can be used to facilitate hydrogel crosslinking.^{[ 11 ]} Metal‐ligand coordination complexes facilitate crosslinking through chelation of a metal ion by pendant organic ligands,^{[ 31} , ^{32 ]} while carboxylic acid groups of alginate naturally coordinate with divalent ions for gelation.^{[ 33 ]} The field of peptide self‐assembly abounds with examples of dynamic hydrogels arising from hydrophobic interactions and hydrogen bonding.^{[ 34} , ³⁵ , ³⁶ , ^{37 ]} Supramolecular host–guest interactions are also commonly used in hydrogel crosslinking, leveraging ordered and shape‐complementary inclusion of small hydrophobic guests within macrocyclic cavitand hosts, such as cyclodextrins and cucurbit[n]urils.^{[ 38} , ³⁹ , ^{40 ]} Hydrogelation can also be driven by electrostatic attraction between oppositely charged groups,^{[ 41 ]} hydrophobic interactions,^{[ 42 ]} dipolar interactions,^{[ 43 ]} or ordered hydrogen bonding.^{[ 44 ]} Indeed, the diversity of dynamic bonding interactions for crosslinking of polymeric or supramolecular precursors offers many routes to create hydrogels with dynamic properties.

Figure 1.

Dynamic hydrogel networks. A) Dynamic and non‐covalent interactions are used to crosslink polymer chains in hydrogel networks. B) Example frequency sweep of a viscoelastic material behaving as a Maxwell element.

Though many mechanoresponsive materials developed with dynamic and non‐covalent bonding motifs are bioinspired, only a subset of dynamic interactions are suitable for use as biomaterials. Materials that interface with the body must be functional at physiological conditions (37 °C, pH 7.4, ≈150 mm salt). Additionally, the bonding mechanisms of biomaterials must be compatible with biological materials. Bioorthogonal reactions, which occur in biological systems without interfering with biochemical processes or biomolecules, are typically required.^{[ 45 ]} As a counterexample, ketones or aldehydes used to form dynamic‐covalent Schiff base or imine bonds may react promiscuously with the abundant amines in biological materials. In other cases, such as host–guest complexes and phenylboronic acid–diol complexes, interfering interactions may occur with native groups present in the body. Yet, the selection of interactions with affinities well in excess of these interfering interactions may still enable function in this case, in spite of ubiquitous biological competitors.^{[ 46 ]}

The dynamicity of crosslinks in materials prepared with dynamic interactions is most often quantified by the rate constant of the reverse crosslinking reaction or interaction, k_off . This quantity is often obtained through rheological experiments, either by a frequency sweep or a stress relaxation experiment. Viscoelastic materials behave as viscous liquids at timescales much longer than their characteristic network relaxation time and elastic solids at timescales much shorter than this relaxation time (Figure 1B). The relaxation time of a network behaving as a Maxwell element can be determined from a frequency sweep, as the inverse of the crossover frequency in Hz (τ_R = 1/ω_c ). The reverse reaction rate constant of the network may then be approximated as k_off ≅ 1/τ_R = ω_c (Figure 1B). For some mechanoresponses, the forward rate constant for the crosslinking reaction, k_on , is also important, and may be determined from kinetic experiments fitting a reactant or product concentration over time to a rate law. Alternatively, K_eq may be obtained, and the relationship K_eq = k_on /k_off utilized to calculate k_on . K_eq may be determined from the plateau modulus (G₀ ) for an ideal network hydrogel system using ideal network theory or in a small molecule system with isothermal titration calorimetry (ITC), surface plasmon resonance (SPR), nuclear magnetic resonance (NMR), or kinetic experiments. In contrast to dissociative hydrogel networks, which are governed by distinct kinetic parameters for the forward and reverse crosslinking reactions, associative dynamic networks are characterized by a single parameter: the crosslink exchange rate, k_ex . This rate has been quantified in hydrogels by monitoring the time‐dependent concentrations of model small‐molecule analogs and fitting the data to appropriate kinetic rate laws.^{[ 47 ]} Alternatively, the characteristic relaxation time constant (τ) of these networks has been determined through rheological stress relaxation measurements.^{[ 48} , ⁴⁹ , ^{50 ]}

This review explores the use of dynamic and non‐covalent crosslinking interactions in order to endow hydrogels with mechanoresponsive properties. The creation of bioinspired hydrogels capable of self‐healing, non‐Newtonian changes in viscosity dependent on shear rate, or strain‐stiffening will be described alongside the chemical origins of these mechanoresponsive behaviors, with an eye toward the preparation of functional materials. Dynamic and/or non‐covalent crosslinking motifs are essential to enabling these specific mechanoresponsive behaviors. Reorganization of dynamic bonds following mechanical disruption enables self‐healing of hydrogels on the bulk scale, typically resulting in full recovery of mechanical properties. These same dynamic bonds can be disrupted under elevated shear rates, leading to shear‐thinning behavior—a decrease in viscosity with increasing shear. In some dynamic polymer networks, sufficiently slow bond dissociation has also enabled a transient shear‐thickening regime preceding shear‐thinning, wherein viscosity increases with applied shear rate. This mechanoresponse arises from shear‐induced chain stretching or the conversion of intramolecular to intermolecular interactions, and depends on reversible associations between polymer chains. Strain‐stiffening has likewise been observed in hydrogel systems featuring a range of dynamic and non‐covalent intermolecular interactions, in contrast to covalently crosslinked hydrogels where this behavior is absent. Notably, hydrogel strain‐stiffening has been found to correlate strongly with the dynamicity of crosslinking interactions. While other mechanoresponses exist, including mechanochromism, self‐strengthening, and strain‐softening, those discussed here have been selected because of the critical role that dynamic or non‐covalent interactions play in enabling the specific mechanoresponse.

2. Self‐Healing Hydrogels

Self‐healing hydrogels offer an ability to repair damage arising from their fracture, extrusion, or mechanical perturbation.^{[ 51 ]} This property is advantageous for improving the longevity and stability of soft materials, enabling improved performance as sensors, wearable electronics, or wound dressings.^{[ 52} , ⁵³ , ⁵⁴ , ^{55 ]} In addition to improving the lifetime of these materials, dynamic bond reorganization and self‐healing, in combination with shear‐thinning, enable moldable and space‐filling character to maximize contact with surfaces, even those with irregular or microscale surface topology, and enable improved adhesion or void‐filling.^{[ 56} , ⁵⁷ , ⁵⁸ , ^{59 ]} The expanding use of 3D printing offers another important direction for shear‐thinning and self‐healing hydrogels, using these materials as inks that shear‐thin during extrusion and quickly solidify after printing.^{[ 60 ]} The use of self‐healing gels as cell‐laden “bio‐inks” for 3D printing has further demonstrated the ability of these hydrogels to mechanically protect cells for improved survival under shear.^{[ 61} , ⁶² , ⁶³ , ⁶⁴ , ^{65 ]} Self‐healing gels have also been used as support baths in freeform printing, guiding the printing of materials with complex or delicate features.^{[ 66} , ⁶⁷ , ^{68 ]} In this case, the support gel is disrupted as a material is extruded into it, yet quickly self‐heals to confine the extruded material and support the printed structure. Beyond 3D printing, self‐healing hydrogels enable applications based on extrusion or injection; self‐healing following injection provides minimally invasive delivery of therapeutics in vivo, including the formation of depots for sustained release.^{[ 69} , ⁷⁰ , ^{71 ]} Shear‐thinning is required alongside self‐healing for many of these applications, and will be discussed in greater detail in a proceeding section.

Hydrogel self‐healing is often verified experimentally on the bulk scale, with cut or fractured gels placed in contact and allowed to heal along their joined edge (Figure 2A).^{[ 72} , ^{73 ]} The time required for repair of induced damage, which may entail cuts, holes, or slits, can be used to quantify the rate of self‐healing.^{[ 73 ]} Physical manipulation or more formal mechanical testing, such as by uniaxial tension testing or rheology, is then used to confirm successful healing and recovery of original properties. Tensile or compressive mechanical properties of a material can be monitored before damage and after self‐healing in order to quantify the extent of self‐healing under an applied uniaxial strain.^{[ 74} , ⁷⁵ , ^{76 ]} Alternatively, rheological experiments may monitor self‐healing through restoration of the storage modulus (G′) following bond rupture induced by high values of applied strain (Figure 2B).^{[ 77} , ⁷⁸ , ^{79 ]} A typical step‐strain experiment applies a low strain in order to determine the initial stiffness of the material, subsequently applying a strain in excess of the failure strain in order to induce bond rupture and reduce G′ below the loss modulus (G″). The applied strain is then decreased back to the low strain, and G′ recovery is monitored over time.

Figure 2.

Hydrogel self‐healing. A) Bulk hydrogel self‐healing assessment. Two cut hydrogels are placed together. Over time, the restoration of dynamic bonds at the interface allows for the union of the two halves to form a fully healed gel. B) Step‐strain rheological experiment used to monitor the recovery of the storage modulus (G′) over time following bond rupture induced by high values of applied strain. C) Example illustrating pH‐dependent dynamics of 4‐arm PEG hydrogels crosslinked with Thia‐Michael bonds and its impact on hydrogel self‐healing. Adapted from Reference^{[ 80 ]} (2021, FitzSimons, et al. under CC‐BY‐ND‐NC 4.0). D) Example illustrating 8‐arm PEG hydrogels prepared from supramolecular host–guest crosslink motifs spanning a range of affinities and corresponding bond dynamics, with healing shown that was dependent on the rate of dynamic bond exchange. Adapted with permission from Reference^{[ 38 ]} (2019, American Chemical Society).

Self‐healing materials can be categorized by the conditions under which repair occurs as well as their mechanism of repair. Autonomous self‐healing materials exhibit spontaneous and intrinsic self‐healing under ambient conditions, typically by leveraging dynamic and reversible intermolecular interactions or crosslinks to allow self‐healing without additional input.^{[ 81} , ^{82 ]} Following disruption of these interactions from a mechanical stimulus, they are able to reform due to their dynamic and reversible nature, physically repairing or restoring the mechanical properties of the material. Extrinsic self‐healing materials achieve repair though inclusion of structural features like capillaries, tubes, or capsules that contain chemical agents to drive self‐healing when ruptured, typically through the formation of new covalent bonds.^{[ 83} , ^{84 ]} As intrinsic and autonomous self‐healing hydrogels achieve their mechanoresponsive function through contributions from dynamic and non‐covalent interactions, these will be the primary focus of the current review.

Dynamic‐covalent bonds, including reversible Diels‐Alder reactions,^{[ 85} , ^{86 ]} Schiff‐base bonds,^{[ 87} , ⁸⁸ , ^{89 ]} disulfide bonds,^{[ 90} , ⁹¹ , ^{92 ]} boronate esters,^{[ 93} , ⁹⁴ , ⁹⁵ , ⁹⁶ , ^{97 ]} and thiol‐ene bonds,^{[ 80} , ^{98 ]} have all been integrated into self‐healing hydrogels. In addition, hydrogels prepared from non‐covalent interactions, including hydrogen bonding motifs,^{[ 99} , ¹⁰⁰ , ¹⁰¹ , ^{102 ]} electrostatic interactions,^{[ 103} , ^{104 ]} metal‐ligand coordination,^{[ 105} , ¹⁰⁶ , ^{107 ]} host–guest interactions,^{[ 108} , ¹⁰⁹ , ^{110 ]} peptide self‐assembly,^{[ 111} , ¹¹² , ¹¹³ , ^{114 ]} hydrophobic interactions,^{[ 115 ]} and protein‐ligand pairs,^{[ 116} , ^{117 ]} among others, have been shown to afford autonomous self‐healing. Due to the equilibrium‐governed nature of the interactions underlying such hydrogels, these materials typically exhibit autonomous self‐healing at ambient temperature and neutral pH. Others types of dynamic crosslinking interactions, such as those found in many vitrimers and covalent‐adaptable networks, often require elevated temperature to initiate bond rearrangement and self‐healing.^{[ 118} , ¹¹⁹ , ^{120 ]} The autonomous and intrinsic self‐healing exhibited by hydrogels prepared using dynamic and/or non‐covalent interactions therefore offers a distinct advantage over extrinsic or stimuli‐driven self‐healing materials.

Bond dynamicity is crucial to enabling effective self‐healing in hydrogel networks. Healing of polymeric materials has been theorized to occur in five stages: 1) rearrangement, 2) surface approach, 3) wetting, 4) diffusion, and 5) randomization.^{[ 121} , ^{122 ]} Chain diffusion or segmental mobility at the interface of the materials to be joined is thus essential and has been considered the primary driving force for repair.^{[ 123 ]} The reptation model has attempted to describe the motion of polymer chains trapped within a tube in order to quantify the effects of diffusion.^{[ 122 ]} A diffusion–reaction model developed to study self‐healing in dynamic polymer networks incorporated association and dissociation rates between polymer chains and showed that self‐healing time decreased with increasing association rate.^{[ 124 ]} Faster association reactions promoted more rapid chain reconnection. However, chain mobility in this model was treated separately via the Rouse friction coefficient and was not explicitly linked to reaction kinetics. Chain diffusion was defined through this friction coefficient, and the model concluded that when association is sufficiently fast, the self‐healing timescale becomes limited by chain diffusion rather than reaction rate. Notably, the potential dependence of chain diffusion on association or dissociation kinetics was not accounted for. An alternative model developed to describe probe mobility in fluorescence recovery after photobleaching (FRAP) experiments defined chain mobility using an effective diffusivity, D_eff .^{[ 125 ]} In this model, D_eff depended on the diffusivity of a free chain and the binding strength (k_on*/k_off ). Thus, chain mobility was governed not by the individual rate constants but by their ratio. This framework may likewise be applicable to hydrogel self‐healing. The rate of self‐healing in dynamic and non‐covalent hydrogel networks is primarily controlled by the forward rate constant (k_on ) of its bonding interactions, with higher values of k_on enabling more rapid bulk self‐healing. In contrast, networks wherein polymer diffusion is constrained or forward reaction rates are otherwise slowed exhibit slower bulk self‐healing.^{[ 121 ]} This may be a result of increased crosslink density or polymer concentration, decreasing the effective mesh size of the network, or polymer chains with high molecular weight or bulky architectures.^{[ 126 ]} Meanwhile, the reverse rate constant (k_off ) dictates the rate of crosslink exchange and molecular mobility, impacting the degree to which self‐healing may occur over experimentally relevant time scales.^{[ 28 ]} Low values of k_off may act to constrain diffusion and inhibit self‐healing. For example, the dynamics of 4‐arm PEG hydrogels crosslinked with Thia‐Michael bonds were varied with pH; although the hydrogel prepared at pH 3 had a sufficiently high k_on of ≈0.8 M⁻¹s⁻¹, its k_off was very low, ≈7 × 10⁻⁵ s⁻¹, resulting in insufficient bond exchange and no self‐healing in 90 min (Figure 2C).^{[ 80 ]} In contrast, the hydrogel prepared at pH 7 had k_on ≈16 M⁻¹s⁻¹ and k_off ≈0.01 s⁻¹, exhibiting rapid self‐healing. Similarly, 8‐arm PEG hydrogels with supramolecular host–guest crosslinking motifs were prepared spanning a range of affinities and concomitant dynamics, with lower affinity (K_eq = 1.3 x 10⁹ M⁻¹, faster k_off ) hydrogels demonstrating complete self‐healing within 30 min, while higher affinity (K_eq = 5.4 x 10¹² M⁻¹, slower k_off ) hydrogels did not self‐heal within a 12 h testing period (Figure 2D).^{[ 38 ]} In these examples, slow k_off limited the self‐healing capacity of the hydrogel and served to inhibit this mechanoresponse. Given sufficient dynamicity, however, non‐covalent and dynamic interactions in hydrogel networks enable self‐healing. The reversible nature of these interactions allows for bond reorganization and reformation after disruption in order to physically repair or restore material mechanical properties.

3. Non‐Newtonian Viscosity

Hydrogel viscosity is dependent on both the intermolecular interactions between polymer chains and the number of elastically active chains in the polymer network.^{[ 127 ]} Newtonian fluids do not exhibit viscosity changes as a function of applied shear rate. However, many hydrogels prepared from dynamic and non‐covalent interactions exhibit non‐Newtonian behavior. At very low shear rates, the material is essentially at rest, leading to a constant or nearly constant “zero‐shear” viscosity (η₀ ) regime indicative of inherent resistance to shear.^{[ 128 ]} However, as shear rate increases, the viscosity of non‐Newtonian materials becomes shear rate dependent, either increasing or decreasing with increasing shear rate (Figure 3A).^{[ 129 ]} Most often, hydrogels formed from dynamic and non‐covalent crosslinking motifs exhibit shear‐thinning, with viscosities decreasing as shear rate is increased, arising from dynamic or non‐covalent crosslinks being disrupted and translating to a decrease in the density of elastically active chains and thus a decrease in viscosity. However, certain hydrogels can exhibit shear‐thickening at intermediate shear rates, prior to shear‐thinning. Hydrogel materials of both shear‐dependent classes exhibit a second region of nearly constant and very low viscosity at high shear rates, termed the “infinite‐shear” viscosity (η_∞ ).^{[ 130 ]} This value is representative of the minimum possible resistance to flow after shear has disrupted the maximum possible number of intermolecular interactions.

Figure 3.

Non‐Newtonian shear‐thinning and shear‐thickening of hydrogels. A) Shear‐thinning hydrogels (top) exhibit Newtonian “zero‐shear” viscosity (η₀ ) plateaus at low shear rates (left) followed by shear‐thinning due to crosslink disruption at elevated shear rates, reaching “infinite‐shear” viscosity (η_∞ , right) plateaus at high shear rates. Shear‐thickening (center) hydrogels observe an increase in viscosity from η₀ under intermediate shear rates, most often followed by shear‐thinning. Shear‐thickening (bottom) is attributed to an increase in intermolecular interactions or chain stretching as chains are extended in the direction of flow at elevated shear rates. B) A load cell coupled with a syringe pump to perform injection force measurements. Adapted with permission from Reference^{[ 131 ]} (2017, American Chemical Society). C) Schematic of an associative telechelic polymer network model used to study divergent shear‐thinning and shear‐thickening behaviors. D) Thermogelator poly(alkylene oxide)‐modified carboxymethylcellulose demonstrating that interaction strength between polymer molecules, modulated here by temperature, affects shear‐thickening capacity of hydrogels. Used with permission of Royal Society of Chemistry; from Reference^{[ 137 ]} (2020; permission conveyed through Copyright Clearance Center, Inc).

Both shear‐thinning and shear‐thickening responses can be measured with shear rheometry under increasingly applied shear rate, measuring viscosity as a function of shear rate.^{[ 130} , ^{131 ]} Such an experiment enables visualization of the entire range of shear‐dependent behavior, including η₀ at low shear rates, shear‐thickening and/or shear‐thinning at intermediate shear rates, and η_∞ at very high shear rates (Figure 3A). However, a shear rate sweep may not provide a complete picture of the hydrogel properties. Discontinuities in the shear rate sweep may occur, indicative of artifacts or flow instabilities from wall slip, edge fracture, and shear banding rather than true material properties.^{[ 132} , ¹³³ , ¹³⁴ , ^{135 ]} Moreover, if the equilibrium behavior of the hydrogels is of interest, transient shear rate experiments may be performed.^{[ 132 ]} In such experiments, shear stress is measured as a function of time for discrete shear rates, and steady values of shear stress over time indicate equilibrium behavior. The injectability or extrudability of shear‐thinning hydrogels is often of interest, a property that can be characterized by performing injection force measurements, coupling a load cell or other force sensor with a syringe pump to measure injection force as a function of time and flow rate (Figure 3B).^{[ 132} , ^{136 ]} Alternatively, a mechanical tester fitted with a syringe may be used to perform these experiments.^{[ 131 ]}

3.1. Shear‐Thinning Hydrogels

As previously described, many hydrogels that exhibit shear‐thinning also self‐heal, as dynamic crosslinks enable both the rupture of these bonds under shear and the dynamic reorganization and healing of these bonds following damage. Such hydrogels are often utilized for applications that involve material injection or extrusion and subsequent restoration of material integrity. These applications include inks for 3D printing and injectable hydrogels for therapeutic delivery or regenerative medicine.^{[ 138} , ^{139 ]} As ink for printing, fluidization when extruded through the printer head is critical for smooth and homogeneous printing, with rapid self‐healing post‐extrusion ensuring high‐fidelity printed structures.^{[ 140 ]} Therapeutic compounds, such as proteins and small‐molecule drugs, can be injected easily as cargo within shear‐thinning hydrogels.^{[ 141} , ^{142 ]} The injectability of shear‐thinning hydrogels enables minimally invasive administration and localized delivery, including to areas that may otherwise be difficult to access.^{[ 143} , ¹⁴⁴ , ^{145 ]} Subsequent in situ self‐healing of the hydrogel affords depots for controlled release to local and/or systemic sites.^{[ 146} , ^{147 ]} Shear‐thinning is also an important feature for hydrogels used as wound dressings, where high water content aids in keeping the wound area hydrated, allows permeation of oxygen, and absorbs wound exudates.^{[ 148} , ¹⁴⁹ , ^{150 ]} Specifically, shear‐thinning gels allow for the filling of irregular or deep features, as found in wound beds, ensuring maximum surface contact for protection of the wound and topical delivery of therapeutic agents.^{[ 151} , ^{152 ]}

Shear‐thinning hydrogels can also be used as “bio‐inks” for the encapsulation and printing of cell‐laden structures,^{[ 28} , ^{153 ]} or for the delivery of therapeutic cell populations.^{[ 154} , ¹⁵⁵ , ^{156 ]} The ability for cells to remain viable throughout the high‐shear extrusion process is critical for these applications, and shear‐thinning hydrogels have been shown to protect encapsulated cells from shear‐induced damage.^{[ 154} , ^{157 ]} It has been hypothesized that this is due to “plug flow” of the hydrogel as it is extruded, wherein only the gel interface at the wall of the syringe experiences high shear forces.^{[ 62 ]} This layer of gel shear‐thins while the bulk hydrogel remains largely intact, protecting the vast majority of encapsulated cells from high shear forces. The extensional flow as the gel exits the syringe has been identified as the highest source of cell damage and death, with shear‐thinning hydrogel carriers also protecting cells from high shear forces at this critical location.^{[ 64 ]}

3.2. Shear‐Thickening Hydrogels

While the majority of hydrogels with dynamic or non‐covalent interactions exhibit shear‐thinning, certain dynamic materials can exhibit increasing viscosity under intermediate shear rates. Shear‐thickening has been well‐studied in colloidal systems, and is often attributed to stress‐induced jamming or hydrodynamic coupling of particles into hydroclusters.^{[ 158} , ¹⁵⁹ , ¹⁶⁰ , ^{161 ]} Interactions between colloids and polymers in solution have also led to shear‐thickening fluids due to shear‐induced aggregation and an increase in the number of polymer‐colloid bonds at elevated shear rates.^{[ 162 ]} The behavior of colloidal suspensions and gels under shear often results from physical interactions and hydrostatic forces. Conversely, the behavior of bulk hydrogels under shear results from intermolecular interactions between individual polymer chains, including both crosslinks and chain entanglements. Accordingly, the contributions of bond dynamics to such interactions have direct implications on the behavior of hydrogels under shear.

Shear‐thickening in associative polymers has primarily been studied in the solution state. Networks of telechelic hydrophilic polymer chains end‐modified with hydrophobic chains have most commonly been used to study the divergent behaviors of shear‐thinning and shear‐thickening (Figure 3C).^{[ 127} , ¹⁶³ , ¹⁶⁴ , ^{165 ]} These include poly(alkylene oxide) end‐capped with alkanes and hydrophobically ethoxylated urethanes (HEURs). Such polymer chains form dynamic micelle crosslinks arising from hydrophobic interactions between the hydrophobic end‐blocks. Likewise, solutions of poly(4‐vinylpyridine) (PVP) reversibly crosslinked by metal coordination complexes along the polymer chain have also been used extensively to study shear rate‐dependent changes in viscosity.^{[ 166} , ¹⁶⁷ , ^{168 ]} For these systems, shear‐thickening is attributed to non‐Gaussian chain stretching at elevated shear rates, shear‐induced transformation of intramolecular associations to intermolecular associations, or a combination of both.^{[ 137} , ¹⁶³ , ^{168 ]} All of these mechanisms require chain extension in the direction of shear, which can only happen given strong and stable crosslinks with slow rates of dissociation. If crosslinks dissociate too easily with shear, the network junctions are disrupted, chains are not able to extend in the direction of flow sufficiently, and the material shear‐thins. Slow crosslink dissociation and high association strength have thus been shown to be critical for shear‐thickening. Chain segment relaxation time, as it relates to the lifetime of a crosslink and the time that a crosslinker remains unbound, has also been identified as critical to viscosity increases at elevated shear rates.^{[ 127} , ^{166 ]} A model developed to study the nonlinear rheological phenomena of HEUR systems incorporates a single nonlinear parameter, G_m .^{[ 127 ]} G_m is defined as the ratio τ_s /τ_e , where τ_s is defined as the interaction time of a hydrophobic end group within a micelle, and 1/τ_e is the overall rate of dissociation of a hydrophobic end group from a micelle. G_m thus represents a ratio of the interaction time of the hydrophobic end group to the timescale of ejection of the hydrophobe from the micelle. Both τ_s and τ_e vary nonlinearly with increasing shear rate in a manner dependent on fluid and polymer characteristics, leading to nonlinearities in G_m . It has been demonstrated that G_m < 1 leads to decoupling of hydrophobic groups from micelles and shear‐thinning, while G_m > 1 results in an increasing number of elastically active chains bridging micelles and a period of shear‐thickening prior to shear‐thinning at elevated shear rates.^{[ 127 ]} Work studying divergent shear‐thinning and shear‐thickening in systems of PVP with metal‐cordination complexes defined similar parameters for relevant timescales of interactions.^{[ 166 ]} The average time that a crosslinker remains unbound after dissociating from one complex and prior to forming another is defined as τ₁ . A second parameter, τ_segment , was defined as the local relaxation time of a polymer chain segment. Shear‐thickening was observed when τ₁ > τ_segment , and τ₁ < τ_segment led to shear‐thinning.^{[ 166 ]} Accordingly, when the relaxation time of a chain segment is slower than the rebinding process of crosslinks, the chain is unable to reorganize or effectively form new intermolecular interactions, and thus shear‐thickening is not observed.^{[ 166 ]} In contrast, if chain relaxation time is faster than the rebinding time, chain segments are able to effectively reorient under shear, recombining to convert intrachain interactions to interchain interactions, increasing viscosity and resulting in shear‐thickening.

In recent years, a few dynamic hydrogels have been shown to shear‐thicken at intermediate shear rates prior to shear‐thinning. While the explicit contributions of crosslink junction stability and relative timescales of relaxation and binding have yet to be explored for shear‐thickening hydrogels, it is likely that findings from associative polymer solutions translate to these hydrogel systems. Similar to associative polymer solutions, the shear‐thickening mechanoresponse has been attributed to shear‐induced increases in intermolecular interactions or chain stretching. Shear‐thickening has been observed in near‐ideal hydrogel networks prepared from 4‐arm PEG macromers crosslinked with dynamic‐covalent Thia‐Michael bonds (Figure 2C).^{[ 132 ]} Varying macromer concentration and conjugate acceptor substituents in this system revealed changes in bond dynamics, with relaxation times ranging from ≈2–8 s.^{[ 132 ]} As concentration was increased for each chemistry, relaxation time increased while the degree of shear‐thickening decreased. However, flow instability prevented any conclusions from being drawn about the behavior of these gels following the thickening regime. The PEG macromers of these hydrogels do not participate in intrachain interactions apart from their dynamic‐covalent crosslinks, suggesting a role for dynamic crosslinks in the shear‐thickening response of these gels. Specifically, the reversibility and equilibrium nature of the thickening suggests that chain stretching is driving their shear‐thickening.^{[ 169 ]}

Other hydrogel systems attribute shear‐thickening to an increase in intermolecular interactions as the shear rate is increased. As polymer chains are extended in the direction of flow, chain segments that were inaccessible to intermolecular interactions may be exposed, resulting in an increase in these interactions. Shear‐thickening has been exhibited from poly(alkylene oxide)‐grafted biopolymers, which exhibit thermoresponsive gelation (Figure 3D).^{[ 137 ]} These materials behaved as Newtonian liquids at low temperatures, with shear‐thickening emerging when the temperature was increased; both stiffness and the magnitude of shear‐thickening increased with increasing temperature. This result supports the theory that the strength and stability of intermolecular interactions are important factors in enabling shear‐thickening. Specifically, the authors propose that upon shearing these gels, extension of the biopolymer backbone allows grafts that previously participated in intrachain interactions to instead participate in interchain crosslinking, contributing to an increase in viscosity before being disrupted at high shear rates. A shear‐induced transition from intrachain to interchain coordination was similarly suggested as the mechanism underlying gelation and shear‐thickening of a pseudo‐polyrotaxane system.^{[ 170 ]} Cellulose hydrogels crosslinked with dynamic‐covalent boronate esters also exhibited a brief region of shear‐thickening prior to shear‐thinning, attributed to a shear‐induced increase in interchain hydrogen bonding.^{[ 94 ]} This behavior was supported by transient network theory, which predicts an increase in the number of elastically active chains with elongational flow.^{[ 171 ]} However, when the applied shear rate is greater than the exchange rate of the dynamic interactions, these interactions will be disrupted, leading to shear‐thinning.^{[ 137} , ^{171 ]} The dynamic nature of the intra‐ and inter‐molecular interactions is a common theme among all of these systems, and is expected to be important to the realization of shear‐thickening in dynamic hydrogel networks. It is also possible that polymers with dynamic and non‐covalent intermolecular interactions are able to maintain elastically active chains under increasing shear because intermolecular interactions are reformed after being disrupted as the network is sheared, allowing chains to extend in the direction of flow. Such a mechanism for hydrogel shear‐thickening and subsequent shear‐thinning both relies on and provides a potential explanation for the role of dynamic intermolecular interactions in this mechanoresponse.

4. Strain‐Stiffening Hydrogels

Non‐linear viscoelasticity and strain‐stiffening are common mechanoresponses in many natural biological tissues and materials.^{[ 172 ]} The cytoskeleton and extracellular matrix comprise biopolymer networks that support cells and tissues,^{[ 15} , ¹⁷³ , ^{174 ]} offering mechanosensing and mechanotransduction between cells and their matrix and controlling a number of biological outcomes.^{[ 175} , ^{176 ]} Due to the strain‐stiffening nature of these networks, forces from cell adhesion result in changes to network mechanical properties.^{[ 14 ]} Two‐way communication between cells and their environments is thus dependent on the nonlinear elastic responses and dynamic nature of both intracellular and extracellular biopolymer networks.^{[ 14} , ¹⁹ , ^{176 ]} Study of the cytoskeleton and extracellular matrix offers insights into the mechanisms of mechanoresponsive strain‐stiffening in biological systems, attributed primarily to entropic elasticity.^{[ 15} , ¹⁷⁷ , ^{178 ]} As flexible biopolymer chains are elongated in the direction of applied strain, their conformational entropy decreases with increasing end‐to‐end distance (Figure 4A).^{[ 172 ]} The magnitude of strain‐stiffening can be described by the stiffening index (m), which offers a quantitative measure of the magnitude of the strain‐stiffening response and deviation from linear viscoelastic behavior. Hydrogels based on semiflexible biopolymers often display a theoretical maximum m value of 1.5.^{[ 179} , ^{180 ]} Stiff polymers, with persistence lengths much longer than their contour length, instead stiffen through an enthalpic mechanism arising from network connectivity, resulting in a much lower value of m; in the limit of purely stiff polymers, there is no entropic elasticity upon being strained as these are fully extended.^{[ 172 ]} Upon the application of macroscopic strain, the network crosslinks enable chains to bend (Figure 4B) and subsequently align and stretch (Figure 4C) in the direction of strain.^{[ 15} , ^{173 ]} For both entropic and enthalpic mechanisms of strain‐stiffening, the hierarchical nature of the hydrogel impacts nonlinear viscoelasticity, contributing to polymer bending stiffness and extensibility.^{[ 181 ]} Hierarchical organization of individual biopolymer chains into larger fibrils results in increased persistence length, altering the relative impacts of entropic and enthalpic stiffening in response to strain.^{[ 182} , ^{183 ]}

Figure 4.

Hydrogel strain‐stiffening. A) Entropic elasticity (left) results from a loss of conformational entropy when polymer chains are extended under strain. B) Enthalpic elasticity originates from unfavorable bending or C) stretching of polymer chains under strain. D) Storage modulus (G’) and loss modulus (G”) of a strain‐stiffening hydrogel as a function of applied strain. In the linear viscoelastic region, the storage modulus is approximately constant (G₀ ) before increasing in the stiffening regime and ultimately yielding at high strain. E) Stress‐strain curve of a strain‐stiffening hydrogel. The differential modulus (K’) is the derivative of stress with respect to strain. F) A plot of K’ as a function of stress is used to determine the critical stress (σ_c ) and stiffening index (m). G) Polyisocyanopeptides modified with oligo(ethylene glycol) units form hydrogels from hydrogen bond‐stabilized β‐helical structures and strain‐stiffen with stiffening indices near 1.5. Adapted from Reference^{[ 182 ]} (2014, Jaspers, et al. under CC‐BY) H) 4‐arm PEG hydrogels crosslinked with dynamic Thia‐Michael bonds exhibit strain‐stiffening, while gels crosslinked with covalent thiol‐maleimide bonds do not. The stiffening index (m) was found to be well‐correlated with the reverse reaction rate constant (k_off ) for dynamic Thia‐Michael hydrogels. Reprinted with permission from Reference^{[ 184 ]} (2025, American Chemical Society).

Strain‐stiffening hydrogels that exhibit deviations from linear viscoelasticity at high strains are typically characterized with shear rheology. An oscillatory amplitude or strain sweep of such a hydrogel reveals a linear viscoelastic region at low strains, in which the storage modulus (G’) is independent of strain, quantified as the low‐strain modulus (G₀ ) (Figure 4D). This is followed by a strain‐stiffening region, in which G’ increases with increasing strain. A stress‐strain curve will similarly show the linear viscoelastic region in which stress and strain are linearly related, followed by the strain‐stiffening region, which is marked by an increase in the slope (Figure 4E). The derivative of stress with respect to strain is termed the differential modulus (K’), and is constant in the linear viscoelastic regime and increases exponentially with stress in the strain‐stiffening regime. When K’ is plotted as a function of stress on log‐log axes (Figure 4F), the critical stress (σ_c ) may be determined from the intersection of a line fit to the low‐strain region, which has a slope of ≈0, and the linear portion of the strain‐stiffening region, the slope of which is the stiffening index (m).

Hydrogels prepared from biopolymers usually form through physical entanglements coupled with hydrogen bonding, hydrophobic interactions, and/or electrostatic interactions, resulting in the formation of networks with transient crosslinking between fibers or within bundles of fibers.^{[ 15 ]} These non‐covalent interactions enable strain‐stiffening in biopolymer hydrogels. Agarose is a polysaccharide that forms semiflexible fibrils through both inter‐ and intra‐chain hydrogen bonding, leading to strain‐stiffening hydrogels driven by enthalpic elasticity.^{[ 185} , ¹⁸⁶ , ^{187 ]} Biopolymers modified with alkylene oxide chains, which form hydrogels through hydrophobic interactions between alkylene oxide groups that form microphase‐separated crosslinking regimes at elevated temperatures, have also been shown to strain‐stiffen (Figure 3D).^{[ 137 ]} As the temperature was elevated, the magnitude of the strain‐stiffening response increased. This may be a result of the increase in interaction strength between the hydrophobic groups at elevated temperatures, enabling polymer chains to be extended and driving stiffening responses instead of rupturing these non‐covalent associations. Hydrophobic associations have also enabled the strain‐stiffening of triblock copolymer hydrogels.^{[ 188 ]} As temperature decreased, hydrophobic end‐blocks aggregated to form networks connected by hydrophilic midblocks. The strain at which stiffening began increased with the length of the hydrophilic midblock, attributed to the additional flexibility of the polymer chains. Gels prepared from polyisocyanopeptides modified with oligo(ethylene glycol) units have also demonstrated strain‐stiffening (Figure 4G).^{[ 182} , ^{189 ]} As temperature increases beyond their lower critical solution temperature, these polymers form β‐helical structures and hydrogels stabilized by hydrogen bonding along the polymer backbone.^{[ 190 ]} The stiffening indices of these hydrogels are near the theoretical maximum of 1.5, indicating entropically driven strain‐stiffening.

Biopolymers capable of participating in hydrogen bonding alongside reversible dynamic‐covalent bonds have also been shown to strain‐stiffen. One such hydrogel system was prepared by crosslinking phenylboronic acid‐modified carboxymethylcellulose with tannic acid, a small molecule polyphenol, with both dynamic‐covalent boronate ester crosslinking and hydrogen bonding of the polymer backbone contributing to the mechanical properties of the resulting hydrogels.^{[ 94 ]} Increasing the effective crosslink density decreased the strain‐stiffening response of the hydrogel, while increasing the polymer concentration increased the magnitude of strain‐stiffening. As a result, it was hypothesized that hydrogen bonding and the resulting formation of chain entanglements or bundles served as the primary driving force for strain‐stiffening, while dynamic‐covalent crosslinks may inhibit such interactions by constraining chain position and mobility. Dynamic‐covalent boronate ester crosslinks have also been used to crosslink other polymers prone to hydrogen bonding, giving rise to similar strain‐stiffening behavior.^{[ 93} , ^{191 ]} Carboxymethylchitosan has been similarly used to prepare hydrogels by crosslinking the polymer chains with dibenzaldehyde‐terminated telechelic PEG;^{[ 192 ]} a related report used this same molecule to crosslink polyethyleneimine to form hydrogels.^{[ 193 ]} Both of these gels combine hydrogen bonding with dynamic‐covalent imine bonds between amine groups on their polymer backbones and aldehyde groups on the PEG crosslinker and exhibit strain‐stiffening. Increasing the concentration of crosslinker resulted in a decrease in the stiffening index for the chitosan‐based gels. Accordingly, the use of dynamic‐covalent crosslinking interactions alongside the inherent entanglement and hydrogen bonding of certain polymers offers a route to prepare strain‐stiffening materials.

Another category of strain‐stiffening hydrogels is crosslinked with dynamic‐covalent bonds with polymer backbones that do not interact via hydrogen bonding or other non‐covalent interaction motifs. The limited examples for such materials are prepared with PEG macromer backbones, with hydrogelation arising exclusively from crosslinkable endgroups. Such materials are advantageous for isolating the impact of dynamic‐covalent crosslinks on strain‐stiffening. For example, the strain‐stiffening behavior of hydrogels prepared from combining phenylboronic acid‐ and diol‐terminated 4‐arm PEG macromers was studied, revealing a tunable strain‐stiffening response when varying temperature, macromer concentration, and pH.^{[ 194 ]} Across all variables, the stiffening index decreased with increasing gel stiffness. Given the flexibility of PEG, entropic elasticity is expected to play a significant role in the strain‐stiffening response. However, with a stiffening index significantly below 1.5, enthalpic elasticity likewise contributes to the strain‐stiffening response. Strain‐stiffening has also been observed in 4‐arm PEG gels crosslinked with dynamic‐covalent Thia‐Michael bonds (Figure 2C).^{[ 80} , ¹³² , ^{184 ]} Recently, the rheological behavior of hydrogels crosslinked with dynamic‐covalent Thia‐Michael bonds was compared to those crosslinked with covalent thiol‐maleimide bonds (Figure 4H). The covalent hydrogels exhibit no strain‐stiffening, while the dynamic gels strain‐stiffened, revealing a requirement for dynamic bonds to enable strain‐stiffening in synthetic hydrogel networks.^{[ 184 ]} Additionally, temperature and pH were varied to alter the dynamics of crosslink exchange, with k_off ranging from ≈0.003 to ≈0.4 s⁻¹. The stiffening index of the dynamic‐covalent hydrogels was found to be well‐correlated with the reverse rate constant (k_off ) of the crosslinking reaction. This work supports a direct relationship between crosslink dynamicity and the magnitude of hydrogel strain‐stiffening.

A variety of synthetic hydrogels that realize gelation through physical entanglements, hydrogen bonding, hydrophobic interactions, and dynamic‐covalent crosslinking have shown strain‐stiffening mechanoresponse. Recent work has revealed that dynamic and reversible crosslinks and other inter‐chain interactions are critical to realizing strain‐stiffening in hydrogel systems. Moreover, strain‐stiffening has not typically been reported for covalent gels. The vast majority of reports of synthetic hydrogels exhibiting strain‐stiffening have been studied in the context of fundamental insights, often inspired by the strain‐stiffening behavior of biological materials. In a few cases, synthetic strain‐stiffening gels have been utilized for tissue engineering, wound healing, or as sensors.^{[ 21} , ²² , ¹⁹² , ^{195 ]} The continual development and exploration of strain‐stiffening materials, particularly with regard to the tunability of this mechanoresponse, will enable further use in practical settings.

5. Conclusion

Hydrogel crosslinking and gelation can be driven by diverse intermolecular interactions, including covalent bonds, dynamic‐covalent bonds, and non‐covalent interactions such as metal‐ligand coordination, host–guest recognition, hydrophobic interactions, electrostatic interactions, and hydrogen bonding. While covalently crosslinked hydrogels often exhibit excellent static mechanical properties, they generally lack responsiveness to environmental stimuli and are incapable of self‐healing. In contrast, hydrogels incorporating dynamic or non‐covalent crosslinking motifs display mechanoresponsive behaviors, including self‐healing, shear‐thinning, shear‐thickening, and strain‐stiffening. These behaviors are governed by the kinetic parameters of bond formation (k_on ) and dissociation (k_off ). Rapid self‐healing requires high k_on , while excessively low k_off yields quasi‐covalent behavior that suppresses healing. Shear‐thinning arises from reversible bond dynamics, whereas shear‐thickening has been observed in systems with slow bond dissociation; low k_off allows for polymer chain extension under stress, while high k_off leads to rapid bond exchange and shear‐thinning. Strain‐stiffening, likewise, has been reported in dynamically crosslinked—but not covalently crosslinked—hydrogels, with increased dynamicity (higher k_off ) correlating with enhanced stiffening. Thus, dynamic crosslinking motifs offer powerful tools for engineering mechanoresponsive hydrogels. By tuning the kinetics of these interactions, it is possible to design materials with customized mechanical responses for applications in 3D printing, tissue engineering, drug delivery, and sensing.

Conflict of Interest

The authors declare no conflict of interest.

Biographies

Rachel C. Ollier is a recent graduate of the University of Notre Dame, where she conducted research in the lab of Professor Matthew Webber and obtained a Ph.D. in Chemical Engineering in 2025. She obtained a B.S. in Chemical Engineering from Miami University in 2020. Her research at Notre Dame focused on utilizing dynamic intermolecular interactions to develop mechanoresponsive hydrogels.

Matthew J. Webber is the Keating‐Crawford Collegiate Professor of Engineering and Professor of Chemical & Biomolecular Engineering at the University of Notre Dame. He earned a BS in Chemical Engineering from Notre Dame (2006), Ph.D. in Biomedical Engineering from Northwestern (2011), and completed postdoctoral training at MIT. He joined the Notre Dame faculty in 2016. His research focuses on dynamic and supramolecular biomaterials for drug delivery. Prof. Webber is a fellow of the American Institute of Medical and Biological Engineering and a Senior Member of the National Academy of Inventors.

Ollier R. C. and Webber M. J., “Mechanoresponsive Hydrogels Emerging from Dynamic and Non‐Covalent Interactions.” Adv. Mater. 37, no. 40 (2025): 2507397. 10.1002/adma.202507397