Ag@polydopamine-functionalized borate ester-linked chitosan hydrogel integrates monitoring with wound healing for epidermal sensor

Introduction

With changes in living standards, wearable sensors have been widely used in healthcare for physiological signals monitoring such as human activity, heart rate, respiration rate, and pulse^1,2,3,4. Despite significant progress, there remains a substantial demand for multifunctional sensors in clinical applications, as wearable sensors in clinical practice are often accompanied by complex pathological conditions, such as persistent inflammatory responses or bacterial infections^1,5,6. Therefore, in addition to track physiological signals on the skin, effective wound closure and prohealing performance are important for wearable sensors aimed at improving wound management outcomes and facilitating patient rehabilitation^7,8. Moreover, integrating real-time monitoring with tissue repair can provide more specific data to guide precise treatment.

Recently, significant efforts have been dedicated to the development of flexible sensors with high stability and sensitivity, where various additive fillers such as Ag nanowires^9,10,11, graphene oxide (GO)^12,13,14, carbon nanotubes (CNTs)^15,16,17, polypyrrole (PPy)^18,19,20, MXene nanosheet^21,22,23, are commonly utilized in the fabrication of diverse composites to enhance the sensing performance. However, most sensors are primarily built via incorporating conductive substances into elastomers (e.g., PDMS^24,25, Ecoflex^26,27,28), which encounters the shortcomings of multilayer assembly, poor biocompatibility, non-degradability, limited conformability and unsatisfactory adhesion^29,30,31,32. When applied to the scenarios of acute skin injuries (e.g., surgical wounds, accidental traumas) and chronic wounds (e.g., burns, bacterial infections, and diabetic ulcers), the potential issues of secondary injuries, bacterial infection, and immunogenicity remain serious concerns^33,34. In contrast, hydrogels offer significant advantages over traditional sensors due to their hydrophilic network, soft tissue-like flexibility, tunable physicochemical properties and excellent biocompatibility^2,29,35,36. Therefore, it is essential to develop advanced hydrogels that integrate both real-time sensing capabilities and wound healing functions to address the clinical demands for wound management or rehabilitation treatment.

Recently, a series of multifunctional hydrogel sensors that integrate real-time motion monitoring and wound healing have been developed and reported^37,38,39,40. For example, Li et al. have reported a magnet-oriented hydrogel (named GOH-MPG) by incorporating magnetoelectric nanosheets (Fe₃O₄-polydopamine (PDA)-reduced graphene oxide (rGO), MPG) into gelatin-oxidized dextran (Gel-ODex) with mechanical-electrical anisotropy and photothermal antibacterial properties for wound treatment and monitoring⁴¹. Ma et al. developed an ionic conductive hydrogel cross-linked by phenylboronic acid (PBA) modified dextran methacrylate, choline-based ionic liquid and crystallized polyvinyl alcohol to achieve the synergistic effect of ROS-scavenging and electroactivity for the treatment and monitoring of chronic diabetic wounds⁴². Dong et al. designed a bilayer composite hydrogel sensor, comprising a tough layer of cellulose nanofibers-incorporated poly-acrylamide network and a adhesion layer of TA coated CNFs for health monitoring and wound healing⁴³. Despite the promising achievements, practical hydrogels sensors that can ideally integrate comprehensive properties of excellent sensing performance, conformal adhesion, superior self-healing, antibacterial capabilities, pro-regenerative feature, economical synthesis and user-friendliness are generally difficult (Supplementary Table 1).

To address these, we present a multifunctional hydrogel sensor comprised of Ag-decorated polydopamine (Ag@PDA) nanoparticles with borate ester bond linked the gallic acid (GA)-modified and 3-carboxyphenylboronic acid (PBA)-modified chitosan matrix, namely Ag@PDA-(CSPBA/CSGA) (Fig. 1). Due to the abundant dynamic bonds formed between the phenylboronic acid groups and pyrogallol moieties present in CSGA/CSPBA matrix, the hydrogel was endowed with injectable and self-healing properties. The incorporation of Ag@PDA nanoparticles into the hydrogel network not only enhanced the conductivity, mechanical properties, and adhesion, but also imparted photothermal conversion capability along with robust antibacterial activity against both gram-negative (Escherichia coli, E. coli) and gram-positive (Staphylococcus aureus, S. aureus) bacteria. The versatility of the Ag@PDA-(CSPBA/CSGA) hydrogel sensor was demonstrated through its attachment to human body, facilitating real-time monitoring of subtle and large-scale movements. Furthermore, the Ag@PDA-(CSPBA/CSGA) hydrogel integrated the functionalities of its individual components, leading to significant therapeutic effects on skin wounds. This represents promising approach for personal health monitoring, and clinical treatment of tissue repair and wound closure. Furthermore, it offers a wide range of potential applications in smart artificial skin, rehabilitation training, and wearable human-machine interfaces.

Ag@polydopamine-functionalized borate ester-linked chitosan hydrogel integrates monitoring with wound healing for epidermal sensor — **Fig. 1**

Results

Characterization of the Ag@PDA-(CSPBA/CSGA) hydrogel

To endow the hydrogel sensor with conformal adhesion, fast self-healing properties and excellent biocompatibility, a chitosan-based biopolymer was utilized in this work and modified with functional groups of pyrogallol moieties and phenylboronic acid through amide linkage via EDC/NHS chemistry (Fig. 2). The Fourier-transformed infrared spectroscopy (FTIR) spectra of chitosan showed characteristic peaks at 1587 and 1645 cm⁻¹ could be assigned to N-H bending and C = O stretching vibrations, respectively (Fig. 2a). After the modification with GA or PBA, we observed that the C = O stretching absorption peak of amide I in chitosan at 1645 cm⁻¹ slightly shifted to 1629 cm⁻¹. Meanwhile, a distinct sharp peak corresponding to the C = C bond of the benzene ring was found in the spectra of CSGA and CSPBA at 1520 cm⁻¹ and 1540 cm⁻¹, respectively (Fig. 2a). In accordance with this, the B 1 s peak at 191.3 eV in the X-ray photoelectron spectroscopy (XPS) spectra of CSPBA (Fig. 2b), along with the characteristic B-OH peak appearing at 29.848 ppm in the ¹¹B nuclear magnetic resonance (¹¹B NMR) spectra of CSPBA (Supplementary Fig. 1), further confirmed the presence of the phenylboronic acid group in CSPBA. Additionally, the chemical shift of the aromatic protons in the gallol group at 7.1 ppm was clearly observed in the ¹H nuclear magnetic resonance (¹H NMR) spectrum of CSGA, of which the substitution degree of GA was calculated to be 6.375% (Fig. 2c). For the CSPBA, the ¹H NMR spectrum of CSPBA displayed pronounced characteristic peaks corresponding to the aromatic protons of the phenylboronic acid group within the range of 7.4−8.3 ppm. The substitution degree of PBA on chitosan was calculated to be 22.875% (Fig. 2c).

**Fig. 2: Physicochemical properties of the Ag@PDA-(CSPBA/CSGA) hydrogel.**

To enhance conductivity and facilitate multiple biologic functions, a type of Ag@PDA nanoparticle was introduced into the dynamic CSGA/CSPBA network. Transmission electron microscopy (TEM) images revealed that the Ag@PDA nanoparticle had a spherical PDA core, with Ag dots approximately 25 nm in diameter uniformly distributed on the surface via in situ reduction of Ag⁺ into Ag⁰ (Fig. 2d and Supplementary Fig. 2). The Raman spectra of PDA and Ag@PDA nanoparticles showed two broad bands at 1345 and 1580 cm⁻¹, which corresponded to the stretching vibration of aromatic rings in the PDA (Supplementary Fig. 3). In addition to the peaks of C 1 s (284.8 eV), N 1 s (399.59 eV), O 1 s (532.28 eV) associated with PDA, the XPS spectra of Ag@PDA exhibited two prominent peaks for Ag 3 d at 368 eV and 374 eV (Fig. 2e). Furthermore, X-ray diffraction (XRD) data showed the amorphous characteristics of PDA, as indicated by a broad peak at 2θ = 24.3°. In contrast, distinct diffraction peaks corresponding to the (111), (200), (220), and (311) lattice plane of Ag were observed in the XRD pattern of Ag@PDA at 38.3°, 44.3°, 64.4° and 77.5°, respectively (Fig. 2f). Moreover, a size-dependent surface plasmon resonance band located at 440 nm was found in the UV spectra of Ag@PDA, further revealing the presence of metallic Ag on PDA (Supplementary Fig. 4).

Next, we prepared the Ag@PDA incorporated CSGA/CSPBA hydrogels at a CSPBA-to-CSGA weight ratio of 1:1 and evaluated their conductivity and strain sensing abilities at different concentrations of Ag@PDA. Scanning electron microscopy (SEM) images showed that all hydrogels featured an interconnected and microporous network, with Ag@PDA nanoparticles dispersed throughout the CSGA/CSPBA matrix (Fig. 2g and Supplementary Fig. 5). However, at a low concentration of Ag@PDA (0.15 mg/mL), the Ag@PDA nanoparticles were unable to form enough connections within the CSGA/CSPBA network. Additionally, unsatisfactory distribution of Ag@PDA nanoparticles with noticeable aggregation was found in the matrix of 0.90Ag@PDA-(CSPBA/CSGA) hydrogel. By comparison, the Ag@PDA-(CSPBA/CSGA) hydrogel prepared with 0.45 mg/mL of Ag@PDA displayed an optimal outcome. It was evident that the Ag@PDA nanoparticles were uniformly distributed throughout the hydrogel (Supplementary Fig. 5).

Since the substitution degree of GA and PBA on the chitosan were calculated to be 6.375% and 22.875% respectively, we adjusted the weight ratios of CSGA-to-CSPBA in hydrogels to make the molar ratio of pyrogallol group-to-phenylboronic acid group close to 1:1. The gelation capability of hydrogel decreased with an increase in CSGA content (Supplementary Fig. 6a). Notably, the hydrogel containing a CSPBA-to-CSGA weight ratio of 1:3 could not gel within 2 min. Even when the gelation time was extended to 24 h, it still failed to form a fully crosslinked hydrogel. In contrast, the 0.45Ag@PDA-(CSPBA/CSGA) at a CSPBA-to-CSGA weight ratio of 1:1 could achieve rapid gelation within 20 s (Supplementary Fig. 6b). As mentioned previously, Ag@PDA-(CSPBA/CSGA) hydrogels comprising 2% (w/v%) of CSGA, 2% (w/v%) of CSPBA and 0.45 mg/mL of Ag@PDA were chosen for the subsequent experiments and sensing applications.

Sensing, rheological and adhesive properties of the Ag@PDA-(CSPBA/CSGA) hydrogel

Next, we tested the conductivity of Ag@PDA-(CSPBA/CSGA) hydrogel containing varying concentrations of Ag@PDA. Notably, a significant increase in conductivity exceeding 12.1-fold was observed for the Ag@PDA-(CSPBA/CSGA) hydrogel compared to the CSGA/CSPBA (Fig. 3a). Since a high concentration (0.90 mg/mL of Ag@PDA) would affect the dispersion of nanoparticles, the conductivity of Ag@PDA-(CSPBA/CSGA) hydrogel first increased and then decreased with the increasing concentration of Ag@PDA. Among them, the Ag@PDA-(CSPBA/CSGA) hydrogel containing 0.45 mg/mL of Ag@PDA nanoparticles exhibited the highest conductivity. Following this enhancement in conductivity, we next investigated the sensitivity of hydrogels by gauge factor (GF), which is calculated via division of the resistance change by the strain variation (Fig. 3b and Supplementary Fig. 7). Consistent with the conductive tests, the Ag@PDA-(CSPBA/CSGA) hydrogel containing 0.45 mg/mL of Ag@PDA achieved the highest sensitivity, and its average GF value was calculated to be 2.49 under 100% tensile strain. Also, we measured the relative resistance changes of Ag@PDA-(CSPBA/CSGA) hydrogel with various CSPBA-to-CSGA weight ratio. After complete gelation, the GF values of hydrogels with weight ratios of CSPBA-to-CSGA at 1:2 and 1:3 were calculated to be 1.95, 1.78, respectively. Among them, the highest GF was observed in the group of CSPBA-to-CSGA weight ratio of 1:1 (Fig. 3b and Supplementary Fig. 8).

**Fig. 3: The sensing and mechanical properties of the Ag@PDA-(CSPBA/CSGA) hydrogels.**

Taking advantage of the desirable sensitivity of Ag@PDA-(CSPBA/CSGA) hydrogel, we investigated its sensing properties, including fast responsiveness, electrical stability and repeatability. First, the response time was evaluated by applying a small strain of 1% at a tensile speed of 1100 mm/min. As shown in Fig. 3c, our Ag@PDA-(CSPBA/CSGA) hydrogel sensor could respond to the externally applied strain and recover within 263 ms, enabling real-time monitoring of human motions and subtle physiological activities. Moreover, Fig. 3d showed the relative resistance changes of Ag@PDA-(CSPBA/CSGA) hydrogel corresponding to various stretching speeds (20, 40, 60, 80, and 100 mm/min) at 30% strain. The responses with regular and repeatable pulse shape in ΔR/R₀ signal indicated that the Ag@PDA-(CSPBA/CSGA) hydrogel could effectively respond to different strain frequencies while maintaining stable output signal. Additionally, the strain-dependent resistance behavior was observed through continuous step strain experiment (Fig. 3e). The relative resistance of Ag@PDA-(CSPBA/CSGA) hydrogel changed stepwise with the applied strain and returned to its initial value upon release, indicating the reliability of the hydrogel sensor. To evaluate the long-term durability of hydrogel sensor, 100 repeated stretching-releasing cycles were conducted at 50% strain and 80 mm/min tensile speed. As depicted in Fig. 3f, although minor upward drift can be seen during cyclic test, the resistance changes between each strain cycle were maintained, demonstrating the stably dynamic response performance and durability of the Ag@PDA-(CSPBA/CSGA) hydrogel sensor. The observed upward drift during cyclic test may be attributed to the water evaporation and viscoelastic effects in hydrogels.

Furthermore, due to the plentiful reversible cross-links within Ag@PDA-(CSPBA/CSGA) hydrogel network, it exhibited injectability, tissue adhesiveness and self-healing performance (Supplementary Figs. 9, 10). The stretchability of Ag@PDA-(CSPBA/CSGA) hydrogel exceeded 600% strain at a 50 mm × 15 mm × 1 mm standard size on the Ecoflex substrate (Supplementary Fig. 11). Moreover, the rheological tests showed that the storage moduli (elastic behavior, G’) of Ag@PDA-(CSPBA/CSGA) hydrogel were higher than loss moduli (viscous behavior, G”) across the entire frequency range. This finding indicated that the Ag@PDA-(CSPBA/CSGA) hydrogel could withstand angular distortion and exhibited solid-like behavior. In addition, the strain amplitude sweeps showed that the G’ and G” curves intersected at 598% strain, demonstrating the solid-liquid transition due to the disruption of the hydrogel network occurred at 598% strain (Fig. 3g, h). Further, the lap-shear adhesive data confirmed a good tissue adhesive strength of Ag@PDA-(CSPBA/CSGA) with pork skin (10.95 kPa) as shown in Fig. 3i.

Biocompatibility, antioxidant and antibacterial of the Ag@PDA-(CSPBA/CSGA) hydrogel sensor

To ensure biocompatibility with minimal physiological reactions, the hemocompatibility and cytocompatibility of the Ag@PDA-(CSPBA/CSGA) sensor were investigated. The ex vivo red blood cell hemolysis test showed that Triton induced distinct hemolysis, as evidenced by a bright red supernatant resulting from the release of hemoglobin from red blood cells (RBCs) in the Triton group (Fig. 4a). While, all hydrogel groups did not cause obvious hemolysis and presented similar results comparable to those observed with PBS treatment. Even at a high concentration of hydrogel (6 mg/mL), the RBCs co-incubated with hydrogel displayed a normal morphology (Supplementary Fig. 12). The hemolysis ratios of all hydrogel groups were below 2.65%, demonstrating the hemocompatibility of Ag@PDA-(CSPBA/CSGA) sensor was in the acceptable biosafety range (<5%). For the cytocompatibility, cell viability of Ag@PDA-(CSPBA/CSGA) sensor was assessed by directly co-culturing with L929 cells through CCK-8 assay and Live/Dead staining. Compared with control, the hydrogel had negligible negative effects on cell proliferation (Fig. 4b), and no obvious cellular apoptosis was detected in the Live/Dead staining of hydrogel group (Fig. 4c).

**Fig. 4: Biocompatibility and bioactivity of the Ag@PDA-(CSPBA/CSGA) hydrogel sensor.**

Since oxidative stress is closely associated with inflammation and various pathological processes⁴⁴. Gallic acid and dopamine, the phenolic compounds, have been extensively reported as potent antioxidant and anti-inflammatory agents^45,46. Therefore, we proposed that the Ag@PDA-(CSPBA/CSGA) hydrogel might exert beneficial effects in antioxidation. To confirm this, the antioxidative capability of the Ag@PDA-(CSPBA/CSGA) sensor was evaluated by 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radical scavenging test. As indicated by the UV-vis spectra (Fig. 4d), the absorption peak of DPPH at 517 nm diminished upon the addition of Ag@PDA-(CSPBA/CSGA). And the DPPH scavenging efficiency showed a concentration dependent manner with respect to Ag@PDA-(CSPBA/CSGA) (Fig. 4e). Notably, 93.44% of DPPH free radicals were eliminated at a low concentration of Ag@PDA-(CSPBA/CSGA) hydrogel (4 mg/mL), revealing an excellent antioxidative capability of the Ag@PDA-(CSPBA/CSGA) sensor.

In addition, the phenolic compounds are also associated with diverse functions, such as photothermal conversion and antibacterial capability etc⁴⁷. Moreover, due to the surface plasmon resonance from metallic materials, the Ag@PDA nanoparticles could exhibit a synergistic effect on photothermal property and antibacterial ability. As shown in Fig. 4f, the Ag@PDA-(CSPBA/CSGA) sensor enabled fast and stable photothermal conversion. The temperature of sensor increased with the increase of power density of near-infrared (NIR) laser (Supplementary Fig. 13). When exposed to 1 W/cm² of NIR irradiation for 5 min, the temperature of Ag@PDA-(CSPBA/CSGA) sensor could reach 60.4 °C, and exhibited excellent photothermal stability during 4 on/off irradiation cycles.

To evaluate the antibacterial capability of Ag@PDA-(CSPBA/CSGA) sensor, bacterial colony-forming units (CFUs) counting tests were conducted using E. coli (gram-negative bacteria) and S. aureus (Gram-positive bacteria) (Fig. 4g). Notably, the CFUs treated with the Ag@PDA-(CSPBA/CSGA) hydrogel were significantly less than those in the PBS group. This phenomena could be attributed to following aspects: the phenolic hydroxyl groups of gallic acid and dopamine facilitate bacterial capture, while the electrostatic interactions from chitosan (positive amino groups) and silver ions lead to membrane damage and cytoplasmic leakage⁴⁸. Moreover, when combined with a photothermal effect, the bacterial survival of E. coli and S. aureus exposed to the Ag@PDA-(CSPBA/CSGA) sensor were only 7.24% and 1.92%, respectively, implying a potent antibacterial activity of Ag@PDA-(CSPBA/CSGA) hydrogel sensor. All above data confirm that the Ag@PDA-(CSPBA/CSGA) hydrogel exhibits excellent biocompatible along with remarkable antioxidative and antibacterial properties, making it an attractive alternative as wearable sensor in further healthcare applications.

In vivo wound healing performance of the Ag@PDA-(CSPBA/CSGA) hydrogel

Due to the excellent bioadhesive, antibacterial and antioxidative features of the Ag@PDA-(CSPBA/CSGA) hydrogel, we next examined its wound closure and healing performance in a rat full thickness skin incision model (2 cm). Following the creation of skin incisions, surgical suture, PDMS and Ag@PDA-(CSPBA/CSGA) hydrogels were immediately applied to the wound sites. The wound treated with PBS was set as the control group. For the PDMS treatment, its inadequate tissue adhesive ability made them insufficient to bridge the defect (Fig. 5a). This substantially increased the risk of microbial infection and delayed healing. At day 7 post-surgery, poor wound healing was observed for both the control and PDMS groups, in which large unhealed areas existed due to movement by the rats. In contrast, the suture and Ag@PDA-(CSPBA/CSGA) hydrogel facilitated closure of the incision. Notably, the Ag@PDA-(CSPBA/CSGA) hydrogel exhibited greater advantages than other treatments. Moreover, as shown in Supplementary Fig. 9a, the Ag@PDA-(CSPBA/CSGA) hydrogel could be continuously extruded from a 25-gauge steel needle and injected onto the wound site with rapid gelation time while allowing for arbitrary shape (Supplementary Fig. 9b). After treatment, the Ag@PDA-(CSPBA/CSGA) hydrogel displayed instant tissue adhesiveness, and the incised wound was tightly sealed (Supplementary Fig. 9b). At day 7, the incisions treated by suture and hydrogel showed improved wound healing performance relative to other groups. Although the wounds treated with sutures healed, obvious scabs were observed on the skin. Among all the experimental groups, the injured tissue treated with Ag@PDA-(CSPBA/CSGA) hydrogels revealed superior recovery.

**Fig. 5: Wound healing assessment via an in vivo 2 cm long full-thickness skin incision model (the ruler’s smallest scale is 1 mm).**

Hematoxylin and eosin (HE) staining and Masson’s trichrome staining were performed to confirm the healing performance (Fig. 5b). HE results revealed significant unrecovered dermis in both the control and PDMS groups, where notable fibrous tissues accompanied by a substantial infiltration of inflammatory cells (green triangle) were observed in the injured region. Moreover, we found that the regeneration of epidermal layer was unsatisfactory, characterized by a considerable gap of unhealed epidermal. Masson trichrome staining also indicated that the collagen in the defect region of control and PDMS appeared thin and irregular. Clearly, the PDMS had no positive effects on the wound healing. In contrast, a better closure and significantly enhanced healing rate were observed for the suture and Ag@PDA-(CSPBA/CSGA) groups. Although the skin incision could be well-bridged by suture, the uneven tension caused by the stitching might impede tissue regeneration. Notably, healing at the center of suture group was relatively superior compared to that at the surface and deep regions. Additionally, a certain degree of inflammation was detected in the suture group. By contrast, the Ag@PDA-(CSPBA/CSGA) hydrogels showed the best healing outcome (Fig. 5c). In addition to exhibiting the smallest unhealed wound area, the newly formed epidermis displayed a similar architecture to the adjacent normal skin. Additionally, a large amount of hair follicles and sebaceous glands were observed near the incision site. Furthermore, collagen in the Ag@PDA-(CSPBA/CSGA) group was dense and well-organized. These results demonstrate that the Ag@PDA-(CSPBA/CSGA) hydrogel can accelerate the wound closure and promote the tissue regeneration, exhibiting significant potential for in vivo application.

Human and robotic motion monitoring by the Ag@PDA-(CSPBA/CSGA) hydrogel sensor

Based on its good electrical conductivity and stretchability, the Ag@PDA-(CSPBA/CSGA) hydrogel exhibits great potential as a strain sensor for motion monitoring. In addition, its excellent adhesion and biocompatibility enable the Ag@PDA-(CSPBA/CSGA) hydrogel to directly adhere to human skin for signal detection. In the exercise monitoring test, the hydrogel sensor was directly attached to the skin without any external adhesive media. This study showed that the Ag@PDA-(CSPBA/CSGA) hydrogel sensors are capable of monitoring large movements (joint movements and running/walking status) and small movements (swallow and breath), as well as gesture recognition through an integrated sensing glove. The Ag@PDA-(CSPBA/CSGA) hydrogel sensors were also able to detect large movements of human finger, wrist, elbow (Supplementary Fig. 14a−c), and knee joints (Fig. 6a) by measuring real-time relative resistive changes. These results suggeste promising applications of the Ag@PDA-(CSPBA/CSGA) hydrogel sensor in monitoring human motion and joint disease. As shown in Fig. 6b, human running was monitored in real-time as walking at 1 km/h and running at 3 km/h, 6 km/h, 9 km/h and 12 km/h alternately. Notably, even after numerous cycles of high-speed running, the hydrogel sensor maintained a firm attachment to the skin (Supplementary Fig. 15). These findings demonstrate that the Ag@PDA-(CSPBA/CSGA) hydrogel sensors can effectively distinguish different individual motion states. Furthermore, perspiration from physical activity did not cause the Ag@PDA-(CSPBA/CSGA) hydrogel sensor to detach from the skin.

**Fig. 6: Practical applications of the Ag@PDA-(CSPBA/CSGA) hydrogel sensor in real-time motion monitoring.**

Subtle movements, such as pharyngeal swallowing, could be detected by the Ag@PDA-(CSPBA/CSGA) hydrogel sensor. As shown in Fig. 6c, the subtle skin contraction resulting from the movement of Adam’s apple swallowing was monitored by the relative resistance changes of our hydrogel sensor, demonstrating its potential for monitoring swallowing disorders in patients. Besides, the normal, rapid, and deep breathes at a rate of 0.5−4s/breath could be tracked, as shown in Fig. 6d. The breath rate and depth could be reflected by the frequency and the amplitude of the output resistance signal, respectively. The results indicate that the Ag@PDA-(CSPBA/CSGA) hydrogel sensor is capable of distinguishing the subtle physiological activities.

As shown in Fig. 6e, the Ag@PDA-(CSPBA/CSGA) hydrogel sensors were integrated onto a rubber glove, with each finger serving as an independent data channel for recognizing hand gestures. The gestures depicted in Fig. 6e i−viii correspond to the initial state, thumb up, OK, clenched fist, pointing gesture, number six, victory sign, and love symbol, respectively. The five horizontal plots in Fig. 6f represented the relative resistance responses curves of the thumb, index finger, middle finger, ring finger, and little finger during gesturing activities. Obviously, the resistance response values (ΔR/R₀) of the fingers increased with greater bending angles. This characteristic facilitates analysis of both the bending state and bending angle of each finger. Gestures could be distinguished by examining variations in rapidly changing signal details across all channels as shown vertically in the five plots of Fig. 6f. In addition, the hydrogel sensors were also attached onto the robot hand joints to recognize fully open-close and two stepwise motions of the robot hand (Supplementary Fig. 14d). The capability of Ag@PDA-(CSPBA/CSGA) hydrogel sensor in distinguishing hand gestures and the robotic movements endows it with promising potential for application in sign language communication and robotics.

Discussion

In summary, we present a bioactive Ag@PDA-(CSPBA/CSGA) hydrogel sensor that consists of Ag@PDA nanoparticles embedded within a chitosan-based matrix, interconnected by pyrogallol-borate ester bonds. This design facilitates the integration of real-time monitoring and wound healing functions. The advantages of our approach are highlighted as follows:

Potent antibacterial & antioxidant efficiency: Due to the synergistic effects of Ag-induced bacterial metabolism inhibition and membranes disruption, coupled with photothermal effect from PDA, the hydrogel exhibited strong antibacterial effect. The antibacterial efficacy against E. coli and S. aureus were 92.76% and 98.08% respectively. Further, gallic acid, a natural polyphenol, possesses intrinsic antioxidant and anti-inflammatory effects. Therefore, our hydrogel can effectively protect the infection-risk individuals from further injury and subsequent infections.

Desirable conformal adhesion & self-healing capability: Benefiting from its abundant dynamic bonds (borate ester bonds, hydrogen bonds, metal coordination interactions), the hydrogel can conformal contact with skin even under conditions of intense and random movement, and achieve rapid self-healing within 5 s. These properties ensure the hydrogel can adapt to arbitrary interface with complex movements, thereby facilitating the collection of accurate and meaningful signal.

Excellent biocompatibility & in vivo prohealing capability: Our in vivo data have confirmed the significant therapeutic effects of Ag@PDA-(CSPBA/CSGA) hydrogel in a rat 2 cm-long full thickness skin incision model by reducing inflammation, accelerating wound healing, promoting re-epithelialization and collagen deposition. Moreover, the components of the hydrogel are FDA-approved biomaterials, and the hydrogel also exhibits comprehensive advantages of low cost, convenience and user-friendliness.

Good sensing properties: In comparison to previously reported hydrogel sensors derived from natural polymers (Supplementary Table 1), the Ag@PDA-(CSPBA/CSGA) hydrogel sensor possesses superior sensing performance, including good sensitivity (gauge factor = 2.49), rapid response and recovery time (263 ms), and high linearity (0.996 within 100% strain). In human motion monitoring tests, this sensor is capable of real-time monitoring of both large-scale movements (e.g., elbow, knee, finger, and wrist bending) and subtle human behaviors (e.g., swallowing and breathing).

Although the proposed “all-in-one” hydrogel sensor here presents great application potential for wound healing, personal healthcare monitoring and rehabilitation training, this study acknowledges certain limitations that warrant further improvement. First, the water evaporation, a common phenomenon observed in hydrogel sensor during their practical applications, can result in an upward drift of the signal curve. To mitigate this effect, we have incorporated 10% of glycerol into the hydrogel. Although the resistance changes between each strain cycle can be maintained, some degree of upward drift can still be seen during the cyclic test. Second, the physical cross-links and dynamic covalent cross-links endow the hydrogel with excellent injectability and self-healing capability, while the resilient of hydrogel would be compromised due to the dissipation of the reversible cross-links. Therefore, achieving a balance among these characteristics to meet clinical application demands remains challenging at present. Collectively, considering the above characteristics, the developed Ag@PDA-(CSPBA/CSGA) hydrogel sensors exhibit promising prospects for clinical applications related to tissue repair, wound closure, and personal health monitoring, and could be beneficial to offer new insights for the next generation of epidermal electronics.

Methods

Synthesis and characterization of the gallic acid modified chitosan (CSGA) and 3-carboxyphenylboronic acid modified chitosan (CSPBA)

For the preparation of CSGA, 1.0% chitosan solution (w/v%) was obtained by dissolving 1.0 g of chitosan (Sigma, 448877) in 100 mL of pH 5.0 ultrapure water, 1.37 g of gallic acid (GA, Aladdin, G131992), 1.54 g of 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride (EDC, TCI, D1601) and 0.93 g of N-hydroxysuccinimide (NHS, Sigma, 130672) were dissolved in 20 mL of ethanol, 30 mL of ultrapure water, and 10 mL of ultrapure water, respectively. Next, the GA solution, EDC solution and NHS solution were mixed and dropwisely added into 1.0% chitosan solution at room temperature (RT) under oxygen-free atmosphere. The pH of reaction was maintained at 5.1 by 3 M NaOH. After stirring for 12 h, the product was purified by dialysis (MWCO 8−14 kDa, Viskase, USA) and the final CSGA product was collected through lyophilization and stored at −20 °C until use.

For the preparation of CSPBA, 1.0 g of chitosan was dissolved in 200 mL of 0.5% acetic acid solution (v/v%, in ultrapure water). Next, 0.89 g of 3-carboxyphenylboronic acid (PBA, Aladdin, C103261), 1.03 g of EDC and 0.62 g of NHS were successively dissolved in 50 mL of methanol, followed by reacting with chitosan solution under vigorous stirring at pH 6.0 in dark and oxygen-free atmosphere. After reacting for 12 h, the product was purified by dialysis (MWCO 8−14 kDa) and the CSPBA was collected by lyophilization and stored at −20 °C until use.

The chemical structure of the CS, CSGA and CSPBA was characterized by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Scientific), Fourier transform infrared spectroscopy (FTIR, Nicolet iS50, Thermo Scientific). The chemical signal of PBA in the CSPBA was characterized by ¹¹B nuclear magnetic resonance (¹¹B NMR, AVANCE III HD 600 MHz, Bruker).

The substitution degree of GA and PBA on chitosan were analyzed by ¹H NMR (Bruker, AVANCE III). The substitution degree of GA was calculated using the following equation:

$${rm{D}}{{rm{S}}}_{{rm{CSGA}}}( % )=frac{{{rm{S}}}_{6.8-7.4}/2}{{{rm{S}}}_{1.5-2.5}/3}* (1-{rm{DD}})* 100 %$$

(1)

S_6.8-7.4 is the peak area of the aromatic proton of gallol group. S_1.5-2.5 is the peak area of the acetyl proton of chitosan. DD is 75% deacetylation degree of the chitosan.

The substitution degree of PBA was calculated using the following equation:

$${rm{D}}{{rm{S}}}_{{rm{CSPBA}}}( % )=frac{{{rm{S}}}_{7.4-8.3}/4}{{{rm{S}}}_{1.5-2.5}/3}* (1-{rm{DD}})* 100 %$$

(2)

S_7.4-8.3 is the peak area of the aromatic proton of phenylboric group. S_1.5-2.5 is the peak area of the acetyl proton of chitosan. DD is 75% deacetylation degree of the chitosan.

Synthesis and characterization of the polydopamine (PDA) and Ag@PDA nanoparticles

For the fabrication of PDA nanoparticles, 2 mL NH₄OH (Sigma, 221228) was mixed with 40 mL ethanol and 90 mL ultrapure water under mild stirring in dark. 0.5 g dopamine hydrochloride (Sigma, H8502) was dissolved in 10 mL ultrapure water and then slowly injected into the above NH₄OH solution. After 24 h, PDA nanoparticles were collected by centrifugation at 12,000 rpm for 20 min. After washing with ethanol and water, the resultant PDA nanoparticles were stored in 4 °C until use. For the preparation of Ag@PDA nanoparticles, NH₄OH solution was added into 5% AgNO₃ (Aladdin, S116265, in ultrapure water, w/v%) solution until the brown precipitates were dissolved. Then, 0.6 mL of 5 mg/mL PDA nanoparticles (in ultrapure water) were added into above solution under mild stirring at RT in dark. After reaction for 1 h, the Ag@PDA nanoparticles were collected by centrifugation at 12,000 rpm for 20 min. After washing with ethanol and water, the resultant Ag@PDA nanoparticles were stored in 4 °C until use. The morphologies of PDA and Ag@PDA nanoparticles were investigated by transmission electron microscopy (TEM, JEM-3200FS, JEOL). The chemical structure of the PDA and Ag@PDA was characterized by XPS, X-ray diffraction (XRD, D8 discover, Bruker), Raman spectroscopy (LabRam HR, HORIBA JOBIN YVON), and UV-vis spectrophotometer (UV-2600, Shimadzu).

Preparation and characterization of the Ag@PDA-(CSPBA/CSGA) hydrogel sensors

For the preparation of Ag@PDA-(CSPBA/CSGA) hydrogel sensors, 9 mg of Ag@PDA nanoparticles were dispersed in 10 mL of 10% glycerol solution (in ultrapure water) and sonicated for 8 min at 600 W. Next, 0.2 g of CSPBA and 0.2 g of CSGA were dissolved into above Ag@PDA dispersion and 10 mL of 10% glycerol solution by magnetically stirring (1000 rpm, 25 min), respectively. After removing air bubbles by centrifugation at 2500 rpm, equal volumes of Ag@PDA/CSPBA solution and CSGA solution were vortex mixed, centrifuged (2000 rpm for 25 s) and transferred into Ecoflex mold, and the Ag@PDA-(CSPBA/CSGA) hydrogel-based sensors were obtained. The microstructure of Ag@PDA-(CSPBA/CSGA) hydrogel sensor was analyzed by a field emission scanning electron microscopy (FESEM, SUPRA55, ZEISS). The rheological property of the Ag@PDA-(CSPBA/CSGA) hydrogel was measured at 25 °C on a rotational rheometer (Anton Paar, MCR 302). For frequency-sweep and strain amplitude sweep test, measurements were conducted at a frequency of 1 rad/s and constant 1% strain. The photothermal conversion of Ag@PDA-(CSPBA/CSGA) sensor was measured using digital thermometer under 808 nm NIR exposure (0.5, 0.75 and 1 W/cm²) for 5 min or four 5 min on/10 min off NIR exposure cycles. For tissue adhesion, 300 μL of Ag@PDA-(CSPBA/CSGA) was injected between two pieces of the fresh porcine skin (50 mm × 25 mm × 2 mm). The lap-shear test of Ag@PDA-(CSPBA/CSGA) sensor on porcine skin was measured by a universal testing machine (ESM303, Mark-10, USA) with strain rate of 60 mm/min under 25 N tensile loading, and the lap-shear strength was determined at the point of detachment. The resistance changes of Ag@PDA-(CSPBA/CSGA) sensor (50 mm × 15 mm × 1 mm) were evaluated using a motorized force testers (Mark-10, 23-1030-09) and a vertically moving motorized test stand (Mark-10, ESM303) by an impedance analyzer (2400, Keithley) on the Ecoflex substrate. To investigate the composites of the CSGA/CSPBA on the gauge factor, the amount of CSPBA plus CSGA in each hydrogel was kept the same, only the weight ratios of CSPBA-to-CSGA were varied. The conductivity of Ag@PDA-(CSPBA/CSGA) sensor (Ag@PDA concentrations of 0, 0.15, 0.45, and 0.90 mg/mL) was analyzed using a four-probe resistivity meter (Helpass, HPS2663).

In vitro biocompatibility

For proliferation, L929 fibroblast cells (CL-0137, Procell) at a density of 4 × 10⁴ cells/well plate were incubated with Ag@PDA-(CSPBA/CSGA) hydrogel sensor in humidified atmosphere with 5% CO₂ at 37 °C. After 1, 3, 5 days of culture, cell viability was determined by CCK-8 assay (C0043, Beyotime) using a microplate reader (TECAN, Infinite200 PRO). For cytotoxicity assays, L929 cells were seeded with Ag@PDA-(CSPBA/CSGA) at a density of 5 × 10⁴ cells/well plate, assessed by Live/Dead staining (Thermo, L3224) and observed using confocal laser-scanning microscopy (CLSM, LSM880, ZEISS). Cells incubated on blank well were used as a control.

For hemolytic activity assay, erythrocytes were obtained from the whole blood of male Sprague-Dawley (SD) rats (4 weeks, Guangdong Vital River Laboratory Animal Technology Co., Ltd.). Briefly, the erythrocytes were collected by centrifugating the whole blood at 1500 rpm for 10 min, washed with PBS, and diluted to a final concentration of 5% (v/v%). Then, 500 μL of 5% erythrocytes were mixed with 500 μL of 2, 4, and 6 mg/mL Ag@PDA-(CSPBA/CSGA) hydrogel. After incubation at 37 °C for 1 h, the mixture was centrifugated at 1500 rpm for 10 min. Then, the supernatant was collected and the absorbance values were determined by microplate reader at 540 nm. The erythrocytes incubated with PBS and 0.1% Triton X-100 were served as negative and positive control, respectively. The hemolysis of hydrogel was calculated using the following equation:

$${rm{Hemolysis}},( % )=({rm{OD}}_{rm{hydrogel}}-{rm{OD}}_{{rm{negative}}; {rm{control}}})/({rm{OD}}_{{rm{positive}}; {rm{control}}}-{rm{OD}}_{{rm{negative}}; {rm{control}}})times 100 %$$

(3)

In vitro antioxidant evaluation

The antioxidant effect of Ag@PDA-(CSPBA/CSGA) sensor was performed using 2,2-diphenyl-1-picrylhydrazyl (DPPH, Alfa Aesar, 44150) assay. 3 mL of 100 μM DPPH (in ethanol) was incubated with 1, 2, 4, 6, and 8 mg/mL of Ag@PDA-(CSPBA/CSGA) at 37 °C in dark. After 1 h, the absorption of DPPH was measured at 517 nm. The antioxidant efficiency was calculated by following equation:

$${rm{Scavenging; efficiency}}=(1-{{rm{A}}}_{{rm{h}}}/{{rm{A}}}_{0})times 100 %$$

(4)

A₀ and A_h were the A₅₁₇ values of DPPH before and after incubation, respectively.

Antibacterial ability assay

Antibacterial efficiency of PBS, Ag@PDA-(CSPBA/CSGA) with or without 808 nm NIR irradiation (n = 3) was tested by standard plate count assay against E. coli (ATCC25922) and S.aureus (ATCC29213). Briefly, 10 μL of E. coli or S. aureus suspensions (1 × 10⁸ CFUs/mL) was added in each group. After incubation for 2 h at 37 °C, 1 mL of PBS was added into each group, of which 100 μL of bacterial suspension was transferred on separate nutrient agar plate and incubated at 37 °C for another 12 h under gently shaking. The colony-forming units (CFUs) on the petri dish could be counted and photographed. The bacterial survival ratio of each group was calculated as follows (5):

$${rm{Bacteria}}; {rm{survival}}; {rm{ratio}},( % )=frac{{rm{Survivor}}; {rm{bacterial}}; {rm{counts}}; {rm{of}}; {rm{experimental}}; {rm{group}}}{{rm{Bacterial}}; {rm{counts}}; {rm{of}}; {rm{control}}}times 100 % .$$

In vivo wound closure

Male SD rats (8−11 weeks old) were anesthetized (isoflurane, R510-22, RWD), shaved and disinfected, followed by a 2 cm long full-thickness skin incision on dorsal. Then, PBS, PDMS, suture and Ag@PDA-(CSPBA/CSGA) hydrogel were used to treat wound (n = 3). For histological analysis, animals were sacrificed, and the skin tissue was collected, fixed (4% neutral paraformaldehyde for 3 d), dehydrated, embedded, and sectioned (Leica, RM2265) at a 4.5 μm thickness. Then, sections were stained with hematoxylin and eosin and Masson’s trichrome staining. All animal experiments in this study were evaluated and approved by Institutional Animal Care and Use Committee (YSB-20220316-TW-A0526) of Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences.

Sensing application test

The conductive copper wires were introduced into both ends of the Ag@PDA-(CSPBA/CSGA) hydrogel sensors, which were then adhered to various human body parts (such as fingers, wrists, elbows, knees, and throats), as well as the joints of a robotic arm. The copper wires were secured with adhesive tape. The other end of each wire was connected to the positive and negative terminals of an impedance analyzer. Once activated, the impedance analyzer was used to conduct application tests, with the readings monitored via a host computer. For the gesture recognition application, five Ag@PDA-(CSPBA/CSGA) hydrogel sensors were attached to rubber gloves, each connected to two copper wires. These wires were then linked to the impedance analyzer to record resistance changes during different gesture actions.

Statistical analyses

Statistical analysis was performed using GraphPad Prism 9 and presented as the mean ± SD. The p < 0.05 was considered statistically significant. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, ns not significant.