A Self‐Healing Magnetoelectric Sensor with Pain Sensing for Underwater Soft Electronics

ABSTRACT

Replicating the skin's ability to sense touch, feel pain, and heal itself is key to developing the next generation of durable soft electronics. These capabilities become more critical in underwater environments, where divers and underwater machines face severe challenges such as limited dexterity, device damage, and restricted power availability. Here, we develop a self‐healing magnetoelectric sensory system (SMES) that uniquely integrates self‐powered tactile and proximity sensing with damage detection and autonomous recovery for amphibious operation. The SMES features a multilayer architecture composed of a damage‐sensing layer and an underlying magnetoelectric sensing layer, both utilizing a self‐healing elastomer with patterned liquid‐metal conductors. The design enables the system to detect and recover from pricking, puncturing, and cutting damage while maintaining stable functionality. The SMES exhibits good sensitivity, rapid response, and robust durability in both air and water. Demonstrations with a smart diving glove and a soft robotic hand highlight its potential for noncontact communication and mechanoreception with damage feedback, paving the way toward next‐generation amphibious soft machines that can feel and heal like living skin.

This study introduces a self‐healing magnetoelectric sensory system designed for amphibious soft electronics. By utilizing liquid‐metal conductors and self‐healing elastomers, the system enables autonomous recovery from mechanical damage. Its multilayered architecture features dual self‐powered proximity and tactile sensing alongside pain sensing in both air and water, facilitating underwater noncontact communication and advanced mechanoreception with damage feedback.

1. Introduction

Nociception, a vital sensory function distinct from simple mechanoreception, serves as a physiological warning system inherent to biological skin, enabling organisms to feel painful stimuli and protect the body from ongoing or severe harm [1, 2, 3]. The core purpose of this system is to initiate appropriate responses, notably the skin's self‐regeneration ability following injury. Specifically, it triggers reflexive withdrawal for acute threats, instant healing for minor damage, and sustained healing along with avoidance behaviors for major tissue injuries. Replicating this pain‐sensing capability is essential for creating reliable and adaptable flexible electronics devices, leading recent studies to focus on sensory systems that integrate both pain and touch perception [4, 5, 6]. In most existing designs, however, pain and touch sensing are coupled within a single sensor, which limits their specificity for accurate stimulus distinction and crucially lacks the capability for self‐repair after damage.

Fortunately, advances in various self‐healing materials [7, 8, 9, 10] enable electronic devices to recover from structural and functional damage, thereby preventing premature failure. Inspired by biological skin, integrating the abilities to sense touch, detect damage or pain, and initiate self‐healing is key to developing robust, long‐lasting robotic and wearable technologies. These features become even more crucial when electronic devices are used underwater, where limited dexterity and the heightened risk of damage or power constraints necessitate self‐powered proximity sensing alongside touch sensing for safer operation. Integrating damage‐sensing and self‐healing capabilities facilitates minimizing external repair requirements and prevents minor structural failures from escalating into catastrophic system failures [11, 12]. Despite its importance, a self‐powered sensor with built‐in damage detection and self‐repair specifically designed for underwater conditions remains largely unexplored.

Herein, we address this gap by developing a self‐healing, self‐powered magnetoelectric sensor with integrated damage feedback. Constructed primarily from a self‐healing elastomer and patterned liquid metal conductors, the device operates reliably in both air and underwater environments, enabling self‐powered proximity and tactile sensing based on the electromagnetic induction (EMI) principle. A key feature of our design is the damage feedback functionality, which is physically separated from the sensing layer and enabled by a protective top layer. This layer facilitates both damage detection, analogous to pain sensing, and self‐repair. The device can detect and recover from various forms of mechanical damage, including pricking, puncturing, and cutting, in real‐time, thereby preserving its functional integrity. In addition, this multifunctional device demonstrates good sensitivity, rapid response, and robust durability. We further showcase its practical applicability by integrating it into a smart diving glove for underwater wireless proximity communication and a robotic hand for underwater mechanoreception with visual damage feedback.

2. Results and Discussion

2.1. Structural Design and Fabrication

We designed the amphibious, self‐healing magnetoelectric sensory system (SMES) organized in a multi‐layered configuration (Figure 1a). This architecture comprises a top damage‐feedback layer, an electrical coil layer (for magnetic coupling), a spacer layer, and a bottom magnetic layer. The top layer functions as a self‐healing damage sensor (Figure 1b), specifically engineered for both damage detection and self‐repair. The three underlying layers collectively form a self‐healing, self‐powered proximity and tactile sensor (Figure 1c).

FIGURE 1.

An amphibious, self‐healing magnetoelectric system with integrated damage sensing. (a) Schematic illustration of the multilayered SMES, encapsulated in a self‐healing elastomer (SHE), designed for underwater soft electronics. The SMES is applicable to both wearable smart gloves for wireless, noncontact interaction and robotic hands for mechanoreception with damage feedback. Optical photographs of (b) the damage sensor and (c) the self‐powered proximity and tactile sensor, both exhibiting self‐healing capability. (d) FESEM image and corresponding element analysis of EGaIn wires encapsulated in the SHE, with a cross‐sectional view of a single EGaIn wire shown in (e). (f) Mechanical self‐healing performance of the damage sensor at 50°C. (g) Comparison of the SMES with recent underwater soft sensors [16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30], highlighting its multifunctional advantages.

The sensory system primarily relies on an as‐prepared self‐healing elastomer (SHE) composed of poly(vinylidene fluoride‐co‐hexafluoropropylene (PVDF‐HFP), 1‐ethyl‐3‐methylimidazolium bis(trifluoromethylsulfonyl)imide ([EMIM][TFSI]), and hexamethylenediamine (HMDA). This fluorine‐rich material functions as a hydrophobic encapsulation layer, where the high electronegativity of fluorine and the electrostatic nature of C─F bonds minimize interactions with water, ensuring stability in aqueous environments. The self‐healing mechanism, as reported in our previous study [13], is driven by dynamic, reversible ion‐dipole interactions between [EMIM] cations and the fluorine atoms of the polymer backbone. While [EMIM][TFSI] enhances segmental mobility as a plasticizer, it significantly reduces mechanical elasticity. To counteract this, HMDA was incorporated as a crosslinker to restore the elasticity necessary for reversible deformation (Figure S1). Mechanical characterization confirmed that HMDA effectively suppresses viscoelasticity and increases the Young's modulus (Figure S2) [14, 15]. However, while higher HMDA concentrations yield superior elasticity and smaller hysteresis, they simultaneously diminish stretchability and self‐healing capability by restricting the polymer chain (Figure S3). Consequently, an optimal HMDA content of 0.5 wt.% was selected to balance healing performance and elastic recovery. This composition was further validated via Fourier transform infrared spectroscopy (FTIR‐ATR), which confirmed that the low‐density crosslinking does not impede the synergistic polymer‐ionic liquid interactions essential for the material's functionality (Figure S4).

Due to the inherent advantages of liquid metals, including excellent flexibility, high conductivity, and flowability, Eutectic gallium‐indium (EGaIn) was employed as the self‐healing conductor for the system's electrical functionalities. To create various EGaIn patterns encapsulated within the SHE, 3D‐printed Field's metal was utilized as a sacrificial template (Figure S5). This simple fabrication method ensures good repeatability for various electrical layers (Figure S6). Field‐emission scanning electron microscopy (FESEM) images revealed an EGaIn wire diameter of approximately 250 µm (Figure 1d,e; Figure S7). Furthermore, element analysis verified the homogeneous distribution of both the EGaIn and the SHE components throughout the structure. Figure 1f shows the excellent stretchability and self‐healing capability of the top damage sensor following a three‐day healing period at 50°C. Our SMES offers unprecedented, all‐in‐one multifunctionality compared to recently reported underwater proximity or tactile sensors (Figure 1g and Table S1) [16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. It uniquely integrates device self‐healing, stretchable elasticity, self‐powered tactile and proximity sensing, and damage feedback, distinguishing it from existing underwater soft electronics.

2.2. Mechanical Performance and Self‐Healing

The self‐healing elastomer containing 0.5 wt.% HMDA exhibited exceptional mechanical elasticity under varying strain conditions, from 50% to 300% (Figure 2a; Figure S8a). The inclusion of the HMDA crosslinker resulted in a nearly 30% boost in the elastic recovery ratio, reaching up to 92% compared to the elastomer without HMDA (Figure 2b). As depicted in Figure S9, the SHE demonstrated rapid recovery, achieving complete restoration within approximately 3 s after 500% strain release, which was captured by a high‐speed camera. Crucially, this high elastic recovery was maintained even in underwater environments (Figure S8b). In addition to its mechanical resilience, such a SHE possesses self‐healing capability in both air and underwater conditions at room temperature (RT). In both media, the self‐healing performance progressively improved with longer healing times. While mild heat accelerates the self‐healing process (Figure S10), it also significantly enhances the overall self‐healing efficiency of the SHE. Under heated conditions, the healing efficiency reached approximately 82.4% in air after seven days and nearly 100% underwater after 10 days (Figure 2c,d). Additionally, peel tests were performed to evaluate the interfacial strength of the bilayer SHE (Figure S11). High peel strength was observed after healing for either 3 days at RT or one day at 50°C. These results indicate that the SHE's self‐healing mechanism facilitates robust adhesion, ensuring the device's structural integrity.

FIGURE 2.

Mechanical performance, damage sensing, and self‐repair capabilities. (a) First‐cycle loading‐unloading curves of the SHE under various stretch ratios (λ), defined as the ratio of the current length (L) to the original length (L ₀). σ _stress is engineering stress. (b) Elastic recovery for uncrosslinked SHE (without HMDA) and crosslinked SHE in air and water. (c,d) Mechanical performances of initial and self‐healed SHE samples under various conditions in air and underwater over a duration of 1 to 10 days (10 day). (e) Resistance changes of the damage sensor after syringe needle pricking using different ways: a gentle prick (∼2 N), and a heavy prick (∼5 N) applied at two distinct speeds (1 mm/s for rapid pricking and 0.01 mm/s for slow pricking, and subsequent autonomous self‐repair. (f) Resistance changes during damage (tweezer puncture and knife cut) and pressure‐assisted self‐repair. (g) Comparison of resistance changes of the cut‐damage sensor during stretching after instantaneous and long‐term self‐healing. (h) Resistance changes of the damage sensor subjected to various damages underwater from different sharp objects.

2.3. Damage Feedback with Detection and Self‐Repair

Given the material's mechanical self‐healing and elasticity, we investigated the electrical self‐repair capability of the top damage‐feedback layer to demonstrate the sensor's function in damage detection and electrical self‐restoration. As shown in Figure 2e, when the sensor was subjected to pricking using a syringe needle at various speeds (quick versus slow) and degrees (gentle versus heavy), it consistently exhibited a prompt increase in electrical resistance, followed by autonomous self‐repair back to its initial performance without external intervention. A heavy “pricking pain” led to a larger resistance change (on the kiloohm level) than the gentle counterpart (an ohm‐level change). Interestingly, the pricking speed did not influence the magnitude of the resistance change, only its self‐recovery time. The sensor's capacity for autonomous electrical self‐repair was further validated by subjecting it to damage from other sharp objects, such as tweezers and a blade tip (Figure S12). In these cases, the resistance similarly and quickly recovered after autonomous repair without any external triggers.

Moreover, when the sensor sustained puncture and minor cut damage, it also achieved effective self‐repair within a short period with the assistance of external mechanical pressure. Following this instantaneous self‐healing, the cut‐damaged sensor retained strain‐sensing functionality, withstanding small stretching cycles without failure (Figure 2g). Although a second fracture occurred under extensive stretching, the sensor was capable of re‐repairing and quickly recovering its initial electrical performance. To enable the sensor to withstand cyclic large‐scale deformation, long‐term self‐healing is essential. After a seven‐day self‐healing period at RT, the sensor was successfully able to endure such cyclic large‐scale deformation (Movie S1).

As illustrated in Figure S13, the self‐healing capability offers a distinct advantage over a sensor relying on non‐self‐healing materials. For instance, following minor damage induced by a knife, a silicone (Ecoflex)‐encapsulated sensor (a non‐self‐healing analogue) can restore its electrical function but lacks the structural integrity to withstand subsequent deformation or external mechanical disturbance. This mechanical vulnerability, due to the non‐self‐healing nature of the Ecoflex, significantly facilitates the leakage of the internal EGaIn (Movie S2). More importantly, the self‐healing damage sensor retains its damage detection and self‐repair functionalities even when fully submerged in water (Figure 2h). Regardless of damage type—be it needle prickles, tweezer punctures, or knife cuts—the sensor can self‐heal underwater, recovering both its electronic and mechanical performance (Movie S3). This conclusively demonstrates the suitability of the damage sensor for potential application in aquatic environments.

2.4. Self‐Powered Proximity and Tactile Sensing

The SMES leverages the self‐powered advantages based on the EMI principle, enabling combined proximity and tactile sensing in both aerial and aqueous environments. The viability of this EMI‐based mechano‐electric conversion for soft electronics has been established in our prior work [31, 32, 33]. The primary advantage of the EMI strategy lies in its non‐contact reliability within underwater or high‐humidity environments. This approach offers a self‐powered and non‐contact solution specifically optimized for aquatic tactile feedback. The operational principle relies on the change in magnetic flux across the electrical coil, induced by the relative motion between the magnetic and electrical components. In accordance with Faraday's law of induction [34, 35], we explored key parameters using a proximity sensing setup where two devices were brought into close each other without physical contact, thereby generating an induced voltage (Figure 3a). Notably, the device configuration featuring a built‐in magnet either maintained or yielded a slightly higher voltage output than a magnet‐free configuration under the identical test conditions (Figure S14). Electromagnetic simulations further supported these experimental results. For the magnet‐free device, proximity induces an overall increase in the magnetic field passing through the EGaIn helix coil (Figure S15). In contrast, the configuration with the built‐in magnet experiences a noticeable alteration in the distribution of the magnetic field vectors normal to its electrical component (the EGaIn helix), and a localized region with a reduced change in the magnetic field is observed (Figure 3b). This occurs because the original magnetic field vectors normal to this region, originating from the device's built‐in magnet, are counteracted or subsequently reversed by the approaching external magnetic source (Movie S4).

FIGURE 3.

Self‐powered proximity and tactile sensing capabilities of the SMES. (a) Schematic illustration of the proximity‐sensing experimental setup, depicting the key parameters influencing output performance: noncontact distance (d), motion displacement (L), and motion frequency (f). (b) Calculated magnetic intensity distributions on the EGaIn coil surface during proximity sensing. (c–e) Voltage outputs of the proximity sensor tested across varying noncontact distances (at constant L and f), motion displacements (at constant d and f), and motion frequencies (at constant d and L), respectively. (f) Effect of the sensors fabricated with different coil turns on the output voltage. (g) Calculated change in the magnetic field vector intensity normal to the coil surface under tactile sensing. (h) Output voltage of the tactile sensor under various applied pressures. (i) Normalized output voltage as a function of applied pressure, illustrating the SMES tactile sensitivity.

The proximity sensor's output performance was tuned by varying the noncontact distance (d) while all other parameters were held constant. The built‐in magnet produces a surface magnetic field intensity of approximately 180 mT (Figure S16). We observed that the voltage output decreases remarkably as d increases from 1 to 20 mm (Figure 3c; Figure S17). This dependence is mainly attributed to the decline in the magnetic field change over a constant interaction time. Given the exponential decay of the magnetic field with distance, d, a larger d results in a smaller change in the normal magnetic flux through the coil during movement, leading to a lower induced voltage. Conversely, at a fixed d of 3 mm, varying the motion displacement (L) from 5 to 30 mm (at a constant speed) had a negligible effect on the voltage output (Figure S18). This suggests that the influence of the external magnetic field on the coil is minimal when the initial separation distance exceeds 5 mm from the sensor (Figure S19).

Consequently, the electrical output is predominantly controlled by the speed of relative motion. As shown in Figure 3d, maintaining motion frequency and noncontact distance constant, a larger L value corresponded to a higher motion speed (i.e., a shorter interaction time), which in turn generated a greater voltage output. Likewise, a positive correlation between electrical output and motion frequency was consistently observed, irrespective of the direction of relative movement (approaching or receding) (Figure 3e). The number of turns in the EGaIn helix coil poses a direct impact on the effective area (A) and thus positively scales the voltage output, as defined by Equation (2) in the Method section (Figure 3f). Furthermore, a comparison of different EGaIn pattern shapes (circle, hexagon, and square) revealed that the square‐shaped helix maximizes the coil area for a given maximum width, resulting in the highest voltage output (Figure S20).

Simulations provided a theoretical explanation for the self‐powered tactile sensing mechanism (Figure 3g). When external pressure is applied, the internal air layer, created by a middle spacer, is compressed. This compression decreases the distance between the EGaIn helix and the built‐in magnet, thereby altering the magnetic field vector normal to the coil surface and inducing a significant change in magnetic flux. The magnitude of the electrical output exhibited an upward trend with rising external force intensity (Figure 3h). The sensor's sensitivity can be tuned for specific applications by adjusting the width and height of the spacer layer (Figure S21). Spacer height primarily governs sensitivity in the low‐pressure region; specifically, a taller spacer raises sensitivity from 0.26 to 0.39 KPa⁻¹ below the 20 kPa region, which is comparable to other similar sensors [27]. In contrast, a wider spacer reduces sensitivity but enhances the device's tolerance to higher pressures. To balance overall thickness and performance, we chose intermediate design parameters for the spacer: a height of 2.5 mm and a width of 1.5 mm. Crucially, the inclusion of the damage‐feedback layer at the top of the sensor did not weaken sensitivity in the low‐pressure range. Instead, it substantially improved tolerance to considerable pressures up to 250 kPa, effectively serving as a protective layer for the underlying self‐powered sensor (Figure 3i and Figure S22). Attributed to its air‐gap architecture, the device exhibits non‐linear sensitivity typical of elastomeric sensors. To achieve precise force control in practical applications, this non‐linearity is mitigated through multi‐segment calibration, polynomial‐based software linearization, and closed‐loop gain compensation, ensuring high‐fidelity, real‐time performance across the entire operating range. Meanwhile, the sensor's rapid response (∼41 ms) and recovery time (∼64 ms) were essentially unaffected by damage to the feedback layer (Figure S23).

Notably, the self‐healing, self‐powered sensor demonstrated outstanding long‐term durability and underwater reliability. The sensor maintained a stable electrical output after 10 000 continuous loading‐unloading cycles (Figure S24). Besides, its proximity sensing performance remained effective and consistent even following a 10‐day underwater immersion period (Figure S25). Even when tested in simulated seawater, the device exhibited no significant performance degradation. The SHE material did not undergo observable shrinkage or swelling, thereby ensuring the structural integrity and reliability of the sensor in harsh aquatic environments (Figure S26).

2.5. Underwater Applications

To demonstrate the self‐powered, noncontact underwater sensory utility, we developed a smart diving glove for wireless proximity communication (Figure 4a). The glove integrates three core components: SMESs, onboard lighting‐emitting devices (LEDs), and a printed circuit board (PCB) (Figure S27). The SMES is mounted on the palmar side of each finger, while corresponding LEDs and the PCB are situated on the dorsal side of the glove (Figure 4b,c). All electronic components are encapsulated within a waterproof, compliant silicone enclosure to ensure functionality in aquatic environments (Figure S28).

FIGURE 4.

A SMES‐based smart diving glove for underwater wireless proximity communication. (a) Schematic illustration of the wearable smart glove integrating self‐powered proximity sensing, wireless communication, and real‐time severe damage detection for underwater operation. (b,c) Optical photographs showing the front and rear views of the glove. (d) Voltage outputs from the finger‐mounted SMES corresponding to distinct hand gestures. (e) Smartphone display of wireless signals from the SMES during two underwater scenarios: (i) thumb‐index finger proximity and (ii) thumb sliding sequentially across the index and middle fingers. (f) Real‐time risk warning process triggered by severe underwater damage to the SMES.

The system generates self‐powered electrical signals when two fingers are brought into noncontact proximity. These signals are amplified, processed, and wirelessly transmitted to a smartphone via the integrated Bluetooth module on the PCB (Figure S29). Five distinct gestures (e.g., finger‐to‐finger proximity or sliding) were tested, each yielding a unique voltage output profile (Figure 4d). These profiles are mapped to a sequence of specific commands displayed on the smartphone: “Normal”, “Holding”, “Going up”, “Going down”, and “Help”. For example, thumb‐to‐index‐finger proximity signals the “Normal” condition. Conversely, sequentially sliding the thumb across the index and middle fingers transmits a “Help” command, signaling an assistance request (Figure 4e and Movie S5). Additionally, red LEDs serve as a real‐time visual warning, activating when the top‐layer damage sensor detects a severe risk of damage to each finger device. As an illustration, accidental damage to the thumb device (e.g., when grasping a sharp shell) instantly illuminates the corresponding red LED (Figure 4f).

Furthermore, an SMES‐equipped robotic hand was engineered to provide underwater mechanoreception and damage feedback. The hand achieved self‐powered tactile sensing while offering real‐time damage status during object delivery (Figure 5a and Figure S30). Three LEDs (green, yellow, and red) indicated normal, minor, and severe damage, respectively, based on specific resistance thresholds. The damage‐feedback mechanism (Figure 5b) is as follows: grasping a smooth object induces an ohm‐range resistance change (normal status), whereas injury from a sharp object elevates the resistance to the kiloohm range. A resistance change below 150 kΩ triggers a yellow LED, indicating minor damage capable of autonomous restoration within a short duration. Conversely, resistance exceeding 150 kΩ (or entering the megaohm range) activates a red LED, signaling severe structural damage that impedes unassisted functional recovery. This 150 kΩ threshold was empirically established as the critical benchmark for rapid self‐repair. Damage beyond this limit necessitates manual intervention or component replacement.

FIGURE 5.

Underwater mechanoreception and visual damage feedback using an SMES‐equipped robotic hand. (a) Schematic of an underwater robot executing delivery tasks equipped with the SMES and LEDs, enabling self‐powered mechanoreception and visual damage feedback. (b) Illustration of the SMES damage mechanism. (c) Underwater operations of the robotic hand: (i, ii) Grabbing and transporting a smooth stone and conch; (iii, iv) Minor damage inflicted by a sharp shell, followed by rapid self‐repair; and (v) Severe damage inflicted by the shell. (d,e) Corresponding changes in resistance and voltage output of the SMES for the cases shown in (c), respectively.

Delivery examples (smooth stone and conch) demonstrated normal‐status operation (Figure 5c‐i,ii). Enhanced mechanical pressure led to a proportional increase in strain‐sensing‐induced resistance change; accordingly, improved voltage output (Figure 5d‐i,ii, e‐i,ii; Figure S31a and Movie S6). Simulation of minor underwater damage (pricking or puncturing by a sharp shell) successfully triggered the yellow LED, followed by rapid self‐repair (Figure 5c‐(iii,vi),d‐(iii,iv); Movie S7). While operating normally, the SMES as a mechanoreceptor generated the expected two electrical signal peaks during distinct loading and unloading phases (Figure S31b–d). Critically, the occurrence of severe damage illuminated the red LED (Figure 5c‐vii) and was accompanied by the disappearance of the unloading peak, leaving only a single loading peak. This indicates a failure in short‐term recovery, signaling a critical risk of structural damage to the underlying self‐powered mechanoreceptor (Figure 5d–(v),e(v)).

3. Conclusions

This study presents a multi‐layered, self‐healing magnetoelectric sensory system with damage feedback for both air and underwater environments. The architecture integrates two main functional components: a top damage sensor and an underlying self‐powered proximity and tactile sensor. Both components utilize a highly elastic and stretchable self‐healing elastomer for encapsulation and patterned EGaIn for conductivity. The pain detection layer effectively handles diverse mechanical damages (pricking, puncturing, and cutting), enabling real‐time damage detection and self‐repair. Moreover, the SMES exhibits comparable sensitivity, rapid response, and robust long‐term durability, with its self‐powered sensor improving the reliability of underwater noncontact and contact interactions. This multifunctional design is well‐suited for practical underwater applications, including a smart diving glove for wireless proximity communication and a robotic hand featuring mechanoreception and damage feedback. We anticipate that the SMES system will significantly advance underwater human‐machine interactions, soft robotics, and electronic skins.

4. Experimental Section

4.1. Materials Synthesis

Self‐healing elastomers (SHEs) were synthesized via a simple solution‐casting method. Initially, PVDF‐HFP (3 m Dyneon Fluoroelastomer) was fully dissolved in acetone. To this solution, [EMIM][TFSI] (Solvionic) was added at a 40 wt.% mass ratio, and the mixture was stirred at RT for 4 h. Subsequently, varying contents (0, 0.5, 1, 2 wt.%) of HMDA (Sigma–Aldrich) were introduced, and stirring continued for an additional hour to facilitate cross‐linking. The resulting yellowish, transparent, homogeneous solution was then cast into a petri dish and dried by evaporating the acetone at 50°C for 48 h, yielding the final SHE film.

For comparative studies, a non‐self‐healing silicone elastomer was prepared by mixing the two‐component liquid silicone (Ecoflex‐30, Smooth‐On) at a 1:1 mass ratio and curing the mixture at 60°C for 20 min. Polydimethylsiloxane (PDMS, Smooth‐On) and a silicon potting compound (GN‐7028, Chongqing Guining Technology) were used to fabricate the waterproof enclosure for encapsulating all electronic components. Both are also two‐component liquid silicones, mixed at a 10:1 mass ratio and cured at 75°C for 24 h.

4.2. Device Fabrication

The required patterns of Fields' metal (Roto144F, RotoMetals) were fabricated as a template on a glass substrate using tension‐driven three‐dimensional printing [36]. The SHE solution was then cast onto the patterned substrate and dried, yielding an SHE film with defined channels. Another SHE film was subsequently bonded atop the patterned film at 50°C for 24 h, forming completely sealed channels. Next, EGaIn (Sigma–Aldrich) was injected into these channels using a syringe, creating both the self‐healing damage‐feedback layer and the electrical layer of the self‐powered sensor. For the self‐healing magnetic layer, a commercial NdFeB permanent magnetic plate (4 mm in diameter, 1 mm in thickness; G52, Shenzhen LALACI) was encapsulated within the SHE. The SHE itself was also engineered into a spacer layer with specified parameters to create an internal air gap. Finally, the individual components—the damage‐feedback, electrical, magnetic, and spacer layers—were assembled into an intact SMES through the material's inherent self‐healing capability.

4.3. Materials Characterization and Mechanical Testing

Thermogravimetric analysis (TGA) of the SHE was performed using a TA Instruments Q500 under a nitrogen atmosphere. The samples were heated from 25°C to 750°C at a rate of 15°C/min. The distribution of the EGaIn within the SHE's patterned channels was observed using FESEM (Hitachi S‐4300). Elemental analysis of their components was subsequently conducted via energy‐dispersive X‐ray spectroscopy (EDX, Oxford Instruments).

Mechanical characterization involved both tensile and compression testing. Tensile tests were carried out using a Zwick/Roell Z2.5 instrument. Samples were prepared as standard Type V tensile bars (3.18 mm in width and 7.62 mm in gauge length) in accordance with ASTM D638. Both monotonic and cyclic tensile experiments were conducted at a strain rate of 1 mm/s and an ambient temperature. The methods used to determine SHE's elastic recovery and resilience are defined in reference to our previous work [10]. Standard compression tests were performed separately on a JSV‐H1000 compressive test apparatus (Japan Instrumentation System) at a rate of 600 mm/min.

4.4. Electrical Measurements and Self‐Healing

Electrical resistance changes and voltage outputs for the damage and self‐powered sensors, respectively, were measured using digital multimeters (Keithley, DMM7510). The standard tests for proximity sensing were achieved by using a linear stage (RXP45, Ruixin Technology) to execute approaching and receding cycles at varying frequencies and distances. The sensitivity of the tactile sensor was determined based on the approach described previously [27].

The self‐healing mechanical performance was evaluated by sectioning the sample into two pieces and immediately rejoining the fresh‐cut surfaces. The mechanical self‐healing efficiency (%) was quantified as the ratio of the material's post‐healing elongation to its initial elongation.

4.5. Electromagnetic Simulations

The ANSYS Maxwell finite element analysis software was employed to investigate the three‐dimensional magnetic field distribution across a patterned EGaIn coil in various configurations (built‐in magnet and magnet‐free) induced by a moving NdFeB magnetic plate under different scenarios. The magnets involved were modeled as 3D cylinders with a diameter of 4 mm and a height of 1 mm, magnetized along the positive z‐axis using G52 NdFeB parameters. Specifically, the magnetic coercivity (H_c ) was set to −836 000 A/m, and the residual flux density (Br) was approximately 1.42 T. For the EGaIn coil, its equivalent circuit was modeled in 3D using a bulk electrical conductivity of 3.4 × 10⁶S/m . The velocity of the movable magnet was set to 75 mm/s for proximity sensing and 10 mm/s for tactile sensing.

The induced electromotive force (ε) was calculated based on Faraday's law of induction, expressed mathematically as:

$$\epsilon =-d{\mathrm{\Phi }}_B/dt$$ (1)

where Φ _B is the total magnetic flux through the circuit surface. The magnetic flux [35] is defined as the surface integral of the magnetic field B over a surface A:

$${\mathrm{\Phi }}_B=\int {\int }_{A}B·dA$$ (2)

where dA represents an infinitesimal area vector normal to the surface.

4.6. Smart Diving Glove for Underwater Wireless Communication

The smart glove was developed by integrating five SMESs, five surface‐mount device (SMD) LEDs, and a PCB into a diving glove base. To ensure waterproofing, the LEDs and the PCB were housed within compliant silicone enclosures. The PCB, custom‐manufactured via the JLCPCB platform using an EasyEDA layout, primarily incorporated amplifiers, a microcontroller unit (MCU, STM32F104), a Bluetooth module (HC‐04), and a 3.7 V, 180 mAh rechargeable lithium battery. This circuit was designed to process the SMES signals, control the corresponding LED activation, and facilitate the transmission of a wireless control signal to a paired smartphone via Bluetooth. The operation of the smartphone display is as follows: The MCU on the onboard PCB collects the sensor readings and transmits them to a remote computer via Bluetooth. This remote system runs a virtual smartphone environment in Python. Upon receiving the sensor data, the virtual machine evaluates the activation amplitude and sequence of each sensor and dynamically updates the display on the smartphone screen accordingly.

4.7. SMES‐Equipped Robotic Hand for Underwater Mechanoreception and Damage Feedback

Fabrication of the robotic hand was performed using a 3D printer (Bambulab X1C) with polylactic acid (PLA; eSun) via the fused deposition modeling (FDM) technique. Resistance signals from the integrated damage sensor of the SMES system were acquired by digital multimeters and streamed to a computer. A Python program was used to process this data and transmit corresponding control signals to an Arduino board. This board ultimately regulated the operational states of the three indicator LEDs (red, yellow, and green).

Author Contributions

X.Z., J. Z., and Y.J.T. conceived the idea and designed the experiments. X.Z. carried out experiments and collected the overall data. X.Z. and E.P. contributed to materials characterization. P. C., B. S., and X.Z. contributed to the electromagnetic simulation. J.Z., X.W., and X.Z. performed circuit designs and worked on demonstrations. Y.J.T. supervised the project. Y.J.T. and X.Z. analyzed all the data and co‐wrote the paper. All authors discussed the results and commented on the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: adma73104‐sup‐0001‐SuppMat.pdf.

Supplemental Movie 1: adma73104‐sup‐0002‐MovieS1.mp4.

Supplemental Movie 2: adma73104‐sup‐0003‐MovieS2.mp4.

Supplemental Movie 3: adma73104‐sup‐0004‐MovieS3.mp4.

Supplemental Movie 4: adma73104‐sup‐0005‐MovieS4.mp4.

Supplemental Movie 5: adma73104‐sup‐0006‐MovieS5.mp4.

Supplemental Movie 6: adma73104‐sup‐0006‐MovieS6.mp4.

Supplemental Movie 7: adma73104‐sup‐0006‐MovieS7.mp4.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.;