■ INTRODUCTION

The applications of flexible sensors in electronic skin, 1,2 health detection, 3,4 and soft robotics 5,6 have spurred the pursuit of higher sensitivity and better scalability. 7 The frequent replacement of batteries in flexible devices also presents challenges for continuous monitoring in living systems or extreme environments. 8,9 Efforts to address this challenge have led to the development of flexible energy harvesting systems, including triboelectric nanogenerators (TENG), 10,11 thermoelectric generators (TEG), 12,13 and humidity-generating modules. 14 However, constant external loads are needed for TENG, whereas humidity-generating modules are slow to respond to changes in ambient humidity. By leveraging the temperature difference between the body and the environment, especially in extreme conditions, thermoelectric generators attached to the human body can often generate a power of up to 20 mW/cm 2 for driving sensors and devices. The thermoelectric performance hinges on the choice of materials. Electronic thermoelectric (e-TE) materials, such as bismuth telluride, lead-antimony alloys, and organic PEDOT:PSS, are characterized by the directional movement of electrons within the material from the hot to the cold end in the presence of a temperature difference, forming a potential difference between the two ends. 15 Despite excellent thermoelectric properties at high to medium temperatures, the high cost of bismuth telluride hinders its wide application. 16,17 The lead-antimony alloys, which exhibit good thermoelectric properties in the lowtemperature range, are not suitable for skin-interfaced applications due to the toxic lead. 18 Meanwhile, PEDOT:PSS, with high electrical conductivity, exhibits a low Seebeck coefficient due to its high thermal conductivity. 19 Unlike e-TE materials, ionic thermoelectric (i-TE) materials can provide an enhanced Seebeck coefficient due to the synergistic effect of the redox reaction and Soret effect, along with nontoxicity and low cost. 20,21 The thermoelectric properties can be further enhanced by combining e-TE with i-TE materials in composite films with ionic liquids. 22,23 However, these thermoelectric ionic gels still suffer from low elastic modulus, 24,25 poor stability, 26,27 and complicated assembly procedures. 28 Flexible thermoelectric arrays have been applied to monitor body movements. 29,30 However, most sensors nowadays are designed to detect target signal inputs, 31-,33 and sensors usually operate in complex environments where multiple input signals can interfere with each other and affect accuracy. Combining the sensors with an external real-time monitoring system can wirelessly transmit the measured signals 34,35 for disease prevention, diagnosis, and treatment. 36 This study reports a low-cost, scalable thermoelectric composite based on porous laser-induced graphene (LIG) foam and a P-type NKKC (NaCl, K 3 [Fe(CN) 6 ], K 4 [Fe-(CN) 6 ], and CMC Na) ion gel. By leveraging the ion thermoelectric effect, 37,38 the thermoelectric composite LIG/ NKKC, with an enhanced Seebeck coefficient of 189.90 ± 2.83 μV/K, enables temperature sensing in the range from -10 to 100 °C with a resolution of 0.1 °C. The self-powered strain sensor also exhibits a maximum sensitivity with a gauge factor (GF) of 105.9 ± 5.6 for strain up to 30%, a detection limit of 0.071%, and excellent stability over 5,000 cycles at 30% stretching. A flexible power generation system 39,40 can be further formed by pairing P-type LIG/NKKC with N-type PFF (PVA/FeCl 3 /FeCl 2 ) in series, along with a voltage amplification circuit, to achieve an output voltage of 104.18 mV at a temperature difference of 20 °C. The integration of the resulting sensor/array with wireless circuits provides an intelligent system to remotely detect human physiological conditions and monitor storage/bathroom.

■ RESULTS AND DISCUSSION

The fabrication of the PDMS/LIG thermoelectric ion gel sensor starts with laser scribing PI to prepare honeycombshaped three-dimensional laser-induced graphene (LIG) with a large number of randomly stacked 2D graphene sheets, followed by transferring with PDMS to obtain PDMS/LIG thin films. 41 When LIG is subjected to strain, the transmission path of electrons undergoes drastic changes, affecting the transmission speed of electrons and resulting in significant changes in resistance signals. 42 This paper introduces a thermoelectric composite with an ion hydrogel composed of NaCl, K 3 [Fe(CN) 6 ], K 4 [Fe(CN) 6 ], and CMC Na (NKKC), sandwiched between two LIG/PDMS layers (Figure 1a), to combine the Soret effect 43,44 and oxidation-reduction reaction. 45,46 The difference in the thermal mobility between anions and cations causes the accumulation of cations at the cold end to form an internal electric field. As a result, ferricyanide at the cold end receives electrons and gets reduced e Fe(CN) Fe(CN)

6 3 6 4 [ ] + [ ] , whereas ferrocyanide at the hot end loses electrons and undergoes an oxidation reaction e Fe(CN) Fe(CN) 6 4 6 3 [ ] [

] with the electrons driven by the external transfer back to the cold end. The synergistic Soret effect and redox reaction increase the Seebeck coefficient of the resulting thermoelectric materials. 47 The resulting sensor can detect strain from the change in resistance and perceive temperature changes from the measured voltage at both ends, providing opportunities for remote human health monitoring (Figure 1b) and signal decoupling. Both holes in P-type materials and electrons in N-type materials move from the high-temperature region to the low-temperature region in the presence of a temperature difference, which contributes to the redox reaction for significantly increased output voltage. The demonstration of the integrated system includes intelligent warehousing and bathroom monitoring systems for the rational use of energy (Figure 1c).

The porous graphene foams with mesopores (∼2 μm) and an interconnected 3D structure allow the penetration of PDMS (Figures 2a and S1). The interconnected three-dimensional porous structure changes the conductive pathway upon stretching, with the varying resistance used to measure the tensile strain. A single peak at 24°in X-ray diffraction (XRD) of the laser-printed portion of the powder scraped from the PI proves the presence of LIG (Figure 2b). After being freezedried (necessary for SEM) to characterize the ion thermoelectric hydrogel, a multilayer lamellar structure is observed in the NKKC (Figure 2c). FTIR of the NKKC shows four groups of characteristic peaks: (1) 3450.24, 1607.45, 1421.57, and 1062.21 cm -1 from CMC-Na; (2) 2120.89.24 cm -1 from K 3 [Fe(CN)] 6 ; 3) 2059.15 cm -1 from K 4 [Fe(CN)] 6 , and (4) 588.67 cm -1 from NaCl (Figure 2d). The assembled sensor (Figure 2e) can maintain structural integration even upon stretching (Figure 2f) or bending (Figure 2g).

The increasing tensile strain reduces the contact points of porous graphene and increases the number of material defects, thereby reducing the mean free path (λ) of electrons. As a result, the resistivity (ρ) increases according to

n e m e e 2 2

=

(where m e is the mass of the electron, n e is the density of the electron, and e is the electronic charge), further leading to increased resistance. The strain-dependent electronic transmission characteristics of LIG (Figure 3a) result in enhanced sensitivity compared with conventional resistive strain gauges and thin-film strain sensors. 48 The presence of PDMS in the composite also allows complete recovery of the sensor, enabling it to exhibit elastic properties and low hysteresis. 49 As the strain increases from 0 to 30% (Figure S2a) (before fracture at 45.60% with a fracture stress of 36.97 kPa, Figure S2b), the resistance increases (Figures 3b, S3 and S4) in a piecewise linear manner, with GF of 62.3 ± 0.22 (R 2 =0.99) from 0 to 15% and 105.9 ± 5.6 (R 2 = 0.99) from 15 to 30% (Figure 3c). While large GF values are preferred, excessively large GF values may lead to a decrease in the accuracy of the measured strain. 50 The sensor response in the strain range from 0.2 to 0.8% also confirms the capability of the sensor to detect small deformations (Figure 3d). The good sensor performance in both large and small strain ranges is provided by the unique three-dimensional hierarchical porous structure of the LIG. Besides the static response, the sensor also exhibits good dynamic response to the frequency range from 0.2 to 1 Hz (for 5% stretching; Figure 3e), with consistently rapid response over a wide range of frequencies. In addition to a fast response/recovery time of 0.26 s/0.3 s to 5% stretching (Figure 3f), the sensor also exhibits a low limit of detection of 0.071% (Figure 3g) and high cycling stability over 5241 cycles (Figure 3h). The initial gradual decrease in the cycling test is likely because the interaction between the ion hydrogel and the polymer matrix has not yet reached a steady state at the beginning of the test. 51 The dual-mode strain sensor, with a low detection limit and high sensitivity, compares favorably with those in the previous literature reports (Table S1).

To combine the Soret effect and redox reaction in thermoelectric ion gel, NaCl, based on the Soret effect, is combined with potassium ferricyanide (K 3 [Fe(CN) 6 ]) and potassium ferrocyanide (K 4 [Fe(CN) 6 ]) based on an oxidation-reduction reaction. Due to the Soret effect, the temperature difference moves Na + in NaCl toward the cold end to form a positive potential due to the larger thermal mobility of Na + than Cl -. Meanwhile, Cl -aggregates at the cold end, forming a negative potential. At the cold end, the plate loses electrons to form a positive potential, and [Fe(CN) 6 ] 3-converts into [Fe(CN) 6 ] 4-after receiving electrons due to the redox reaction. Similarly, at the hot end, [Fe(CN) 6 ] 4-loses electrons to become [Fe(CN) 6 ] 3-and the electrode plate receives electrons to form a negative potential, generating a voltage difference between the positive and negative plates. The ionic conductivity of the thermoelectric material is enhanced by directing the ions to migrate directionally down the temperature gradient due to the Soret effect. The efficient migration of ions promotes the thermoelectric potential difference in the thermocouple pair and increases the efficiency of the thermoelectric effect. At the same time, redox reactions lead to changes in carrier concentration, which in turn affects the thermoelectric properties. The Seebeck coefficient of thermoelectric devices is increased due to the synergistic effect of the Soret effect and redox reaction 52,53 (Figure 4a). The performance optimization of the thermoelectric ion hydrogel first starts with the concentration (0.1, 0.2, 0.3, and 0.4 mol/L) of potassium ferricyanide (K 3 [Fe(CN)] 6 ) and potassium ferrocyanide (K 4 [Fe(CN) 6 ]) with a ratio of 1:1 to yield the optimal concentration of 0.2 mol/L (Figure 4b). Further optimization of the NaCl concentrations of 1, 2, 3, and 4 mol/L shows the largest Seebeck coefficient at the concentration of 3 mol/L (Figure 4c). As the temperature difference increases from 5 to 25 °C, the voltage increases (Figure 4d) with a slope of -189.902 ± 2.833 μV/K (R 2 = 0.99) to represent the Seebeck coefficient in the linear fit (Figure 4e), which proves that the NKKC has excellent performance in temperature difference power generation. Meanwhile, the thermoelectric device also has good linearity over a very small temperature range from 0.2 to 0.8 °C (Figures 4f and S5a). Besides the capability to detect a small temperature difference of 0.1 °C (Figure 4g), the sensor can also operate in the temperature difference range from -10 to 100 °C (Figure 4h) with a relatively fast response/recovery time of 20.74/77.80 s for a temperature difference of 6 °C (Figure S5b). However, increased water evaporation at elevated temperatures, especially during extended use, would cause swelling, which affects sensor performance (Figure S6). Here, the temperature difference is created by a Peltier, an ice bag, and a water-cooled plate (Figure S7a), with 15.4 °C for the water-cooled plate (Figure S7b), 100 °C on the Peltier surface (Figure S7c), and -10 °C for the ice bag (Figure S7d). With a temperature difference of 85 °C between the Peltier and water-cooled plate, the voltage across the ends of the sensor is -15.96 mV. The reported temperature sensor demonstrates comparable or even superior sensing performance over the others reported previously (Table S1). Increasing the resistance of the external resistor in an environment with a temperature difference of 5 °C leads to increased voltage but decreased current density, yielding a maximum power density of 27.5 mW/m 2 (Figure 4i), indicating the potential for self-powered sensing.

As a self-powered strain sensor, 54 the strain can be detected from the current signal (Figure 5a). As the temperature difference (ΔT) increases from 0 to 18 °C (5% stretching), the current signal gradually increases from 0 to 2500 nA (Figure 5b), with a zoomed-in response curve shown in Figure 5c for ΔT of 5 °C. Complete recovery and good linearity are also observed for both small (0.2-0.8% in Figure S8) and large (5-30% in Figure 5d) strain ranges (ΔT = 4 °C). A linear fit of the current signal and strain provides sensitivities of 12.73 (R 2 = 0.998), 2.55 (R 2 = 0.999), and 0.902 (R 2 = 0.984) in the tensile strain ranges of 0-5%, 5-10%, and 10-30%, respectively (Figure 5e). The self-powered strain sensor can also detect a tiny strain of 0.05% with a large relative current response of 1.9% (ΔT = 4 °C) (Figure 5f). Despite the initial differences in current responses for varied temperature differences, they saturate as time increases (Figure 5g), which leads to the relationship between the current response and strain (Figure 5h) with little effect from the temperature difference. The selfpowered strain sensor can measure human respiratory rates (Figure 5i).

Pairing the P-type LIG/NKKC (negative) with the N-type PFF (PVA/FeCl 3 /FeCl 2 ) (positive) in a sealed configuration (Figure S9) explores the synergistic thermoelectric effects and also prevents the loss of water and thermoelectric ions for enhanced stability. In the weakly acidic Fe 2+ /Fe 3+ solution, the PVA hydrogel, cross-linked through the reaction between its hydroxyl groups and the aldehyde groups in glutaraldehyde (GA), forms acetal or hemiacetal linkages between polymer chains. 55 For the N-type PFF, Fe 3+ gains electrons at the hot end to become Fe 2+ , whereas Fe 2+ loses electrons at the cold end to become Fe 3+ , with the lost electrons transferred to the P-type material through an external circuit. For the P-type thermoelectric device NKKC, [Fe(CN) 6 ] 3-gets electrons from the external circuit at the cold end to become [Fe(CN) 6 ] 4-, while [Fe(CN) 6 ] 4-loses electrons at the hot end to become [Fe(CN) 6 ] 3-. Connecting multiple thermoelectric pairs of Ptype and N-type devices in series forms a P-N thermoelectric array. The lost electrons migrate through the external circuit to the next pair when there are multiple units in the connection (Figure 6a). Analysis of the FTIR spectrum of the PFF material reveals characteristic peaks of PVA at 1433.01, 1090.20, and 833.32 cm -1 , along with characteristic peaks of FeCl 2 /FeCl 3 at 3413.52 and 1600.12 cm -1 (Figure 6b). SEM analysis of the freeze-dried PFF further reveals a porous structure (Figure 6c). As the temperature difference increases from 5 to 25 °C, the thermoelectric voltage from the N-type PFF increases from 0.5 to 2.3 mV (Figure 6d), with a Seebeck coefficient of 94.85 ± 5.8 μV/K (R 2 = 0.997) calculated from the linear fit (Figure 6e). The N-type PFF can also capture a small temperature difference of 0.1 °C with a reasonably large Experimental demonstration of the sensor to detect (g) a small temperature difference of 0.1 °C and (h) in the temperature difference range from -10 to 100 °C. (i) Changes of the current density and power density as a function of the voltage of load resistance for a temperature difference of 5 °C. Data are presented as means ± SD, with error bars denoting the SD of five independent samples. thermoelectric voltage of 9.4 μV (Figure 6f). As the external load resistance increases, a maximum power density of 0.59 mW/m 2 occurs at 0.3 mV (Figure 6g). The P-N type thermoelectric pair also exhibits a Seebeck coefficient as the sum of the absolute values of the two legs, which is attributed to the generated voltage of 283.8 μV/K from the P-N pair as the sum of the two legs (Figure 6h). Increasing the external load resistance for the P-N type pair identifies the maximum power density of 1.24 mW/m 2 (Figure 6i).

Our flexible temperature difference power generation array features a three-layer structure (Figure 7a), consisting of an upper electrode of PDMS-LIG, a middle layer of a P-N-type thermoelectric material, and a lower electrode of PDMS-LIG (Figure S10). Connecting multiple thermoelectric modules, such as 5 in series (Figure 7b), in a kirigami layout provides enhanced stretchability of the thermoelectric device and increases the thermoelectric voltage to 56.6 mV (ΔT = 40 °C) (Figure 7c). The thermoelectric voltage of 7.91 mV from the flexible array, 56 even under a temperature difference of 5 °C, allows the charging of capacitors with varying capacitances (10, 470, and 1000 μF) to above 7.5 mV (Figure 7d). With a short-circuit current of 12 μA, the Seebeck coefficient of the resulting array is calculated as 1.39 mV/K, resulting in 283.8 μV/K for each P-N junction. The generated thermoelectric voltage linearly increases from 14.8 to 52.3 mV as the number of units in the array increases from one to four (ΔT = 10 °C) (Figure 7e). An increased temperature difference from 27 to 47 °C can generate a higher thermoelectric voltage of 104.18 mV, even in the array with four serially connected units (Figure 7f and Video S1), charging the 10 μF capacitor to 100.45 mV (Figure 7g and Video S2).

Applying the LIG/NKKC dual-parameter sensor on the skin measures the heart rate 56,57 (83 beats per minute) (Figure 8a) and the temperature rise in the finger from 36.4 to 37.6 °C after exercise (Figure 8b). The differentiation among breathing patterns (i.e., slow breathing, hold breathing, and rapid breathing) can also be achieved based on the varied frequencies (Figure 8c). The integration of the sensor with wireless transmission modules further allows wireless monitoring of finger bending and finger temperature (Figure 8d,e and Video S3).

The small voltage generated from the flexible array under low-temperature differences can also be amplified by amplifiers to produce a stable and substantial thermoelectric voltage for intelligent control. For instance, the placement of the hand on the array creates a temperature difference larger than the threshold of 2 °C (Figure 9a) and can control and close the relay to start appliances such as fans and LED lights, which will be switched off when the user leaves the location (Video S4) for energy saving. The flexible thermoelectric power generation array, as a switching control device, can be exploited in intelligent storage systems (Figures 9b and S11, and Video S5) and intelligent bathroom systems (Figure 9c and Video S6), with photographs of the model system shown in Figure 9d,e. Placing a hand on the flexible thermoelectric power generation array triggers the activation of fans and LED lights in the circuit, while removing the hand for 10 s automatically turns off the appliances (Video S4), which is similar to the operation of intelligent bathroom systems (Figure 9e).

■ CONCLUSIONS

This work presents a thermoelectric composite based on NKKC/LIG for dual-parameter strain/temperature sensing. The resulting strain sensor exhibits a maximum GF of 105.9 for strain up to 30%, a low detection limit of 0.071%, a rapid response/recovery time of 0.26/0.3 s, and excellent cycling stability over 5000 cycles at 30% stretching. The high Seebeck coefficient of -189.90 μV/K also allows it to detect the temperature in a wide range from -10 to 100 °C, with a resolution of 0.1 °C. Besides the self-powered sensing, the NKKC/LIG dual-parameter sensor can also be integrated with the wireless circuit to remotely measure heart rate, respiratory rate, and body temperature. Combining the P-type NKKC/ LIG with the N-type PFF further creates a P-N pair, which can be connected in parallel or in series for enhanced thermoelectric properties. The demonstration of the flexible thermoelectric arrays includes intelligent storage and smart bathroom systems for energy savings. The thermoelectric composite could be further explored in temperature-strain decoupled sensing 58 for battery monitoring, smart garments/ clothing, and human health monitoring. ■ EXPERIMENTAL SECTION Materials. The polyimide (PI) film with a thickness of 75 μm was purchased from Suzhou Dongxuan Plastic Products Co. Ltd. (Jiangsu, China). The water-soluble tape was obtained from Yongri Adhesive Co. Ltd. (Shanghai, China). The PDMS kit (Sylgard 184 Silicone Elastomer) was purchased from Dow Corning Corporation (USA). K 3 [Fe(CN) 6 ] and FeCl 3 were purchased from Yien Chemical Technology Co. Ltd. (Shanghai, China). K 4 [Fe(CN) 6 ], FeCl 2 •4H 2 0, GA (G849973), and PVA (Mw 89,000-98,000) were obtained from McLean Biochemical Technology Co. Ltd. (Shanghai, China). NaCl and CMC-Na (800-1200 mPa•S) were purchased from Sinopharm Chemical Reagent Co. Ltd. (Qingdao, China). HCl (0.5 mol/L) was provided by Aimiani Biotechnology Co. Ltd. (Xiamen, China).

Fabrication of Honeycomb-Shaped LIG on PDMS. After fixing the PI film on the hydrogel onto a glass slide, a CO 2 laser (power of 9 W and scanning speed of 317.5 mm/s) was used to create a conductive three-dimensional porous laser-induced graphene (LIG) foam with a rectangular pattern of 3 × 15 mm. To transfer the LIG, the transparent PDMS solution, mixed at a 10:1 ratio, was applied onto the prepared PI-LIG, followed by degassing and curing (in a drying oven at 100 °C for 30 min). Rinsing the hydrogel off with water separated the flexible PDMS-LIG structure from the glass slide.

Fabrication of NKKC and PFF Thermoelectric Ionic Hydrogels. The potassium ferricyanide (K 3 [Fe(CN) 6 ]) and potassium ferrocyanide (K 4 [Fe(CN) 6 ]) at a ratio of 1:1 were added to 10 mL of water at a concentrations of 0.1, 0.2, 0.3, or 0.4 mol/L. Next, NaCl was added to the solution at a concentrations of 1, 2, or 3 mol/L, followed by the addition and stirring of 0.4 g of CMC-Na. The mixture was then stirred in a water bath at a speed of 1000 rpm for half an hour to obtain an ionogel (Figure S12).

After FeCl 3 /FeCl 2 was dissolved in 20 mL of water at a concentration of 1 mol/L, PVA was added at a mass fraction of 5%. Next, 0.5 mL of 0.5 mol/L diluted hydrochloric acid was added to the mixed solution, followed by stirring at 95 °C and 1000 rpm in a water bath for half an hour to obtain a uniform solution of FeCl 3 / FeCl 2 /PVA/HCl. Subsequently, a diluted glutaraldehyde (GA) solution (5 wt %) was added to the solution with rapid and vigorous stirring. The mixture was then allowed to stand for cross-linking for half an hour, resulting in the formation of PVA/FeCl 3 /FeCl 2 (PFF) thermoionic gel (Figure S13).

Preparation of the LIG/NKKC-Based Strain-Temperature Dual Parameter Sensor. After cutting the fabricated PDMS-LIG into 7 × 33 mm films, two identical films were used as the top and bottom layers of the device. Next, the PDMS solution, mixed in a 10:1 ratio, was poured into a glass mold to obtain a fully cured PDMS film with a thickness of 1 mm. The PDMS middle layer, with a slot of 3 × 15 mm in the middle, was cut using a laser, which was then bonded with the two PDMS-LIG films using the PDMS solution, followed by curing. After injecting 0.045 mL of NKKC thermoelectric ionic gel into the cavity with a syringe, the top and bottom layers were connected to the external circuit through conductive silver paste and copper wires (Figure S14).

Characterization and Measurements. The morphology and composition of the honeycomb-shaped LIG were analyzed using scanning electron microscopy (SEM, Nova Nano SEM450) and X-ray diffraction (XRD, D8 Advance), respectively. The morphology and composition of the NKKC and PFF thermoelectric ionic hydrogel were analyzed by SEM and Fourier Transform Infrared Spectroscopy (FTIR, DTGS KBr). Strain and thermoelectric sensing performance were evaluated using a custom-built setup with a tensile tester (JSV-H1000 and HF-1), a temperature logger (UNI-T UT3200+), a source meter (Keithley 2400), an electrochemical workstation (Ivium Vertex. C. EIS), Peltier elements (20 × 20 mm), water cooling plate (40 × 80 mm), a thermal imaging camera (ONE PRO), and a patch-type temperature sensor (WD3703) (Figures S2, S15, and S16). Voltage generated from the thermoelectric power array (Video S1) (MP4) Capacitor charged by the thermoelectric power generation array (Video S2) (MP4) Remote wireless monitoring system to detect finger strain and temperature changes(Video S3) (MP4) Switching control based on the thermoelectric power generation array (Video S4) (MP4) Intelligent storage system based on the thermoelectric power generation array (Video S5) (MP4) Intelligent bathroom systems based on the thermoelectric power generation array (Video S6) (MP4) Cross-sectional SEM images of the composite; schematic and IR image of the testing setup; strain and temperature responses of the strain/temperature dualparameter sensor; thermoelectric power generation arrays at 130 °C for extended periods of time, schematic of P-N type ionic thermoelectric pair, a three-layer thermoelectric array, and the connection between the device and electrical circuits; fabrication processes of the NKKC and PFF thermoelectric ionic hydrogels; fabrication and assembly of the LIG/NKKC-based strain-temperature dual-parameter sensor; performance comparison between the LIG/NKKC-based straintemperature dual-parameter sensor and other dualparameter sensors; and videos showing the use of thermoelectric sensors and power arrays for voltage output, charging the capacitor, monitoring finger strain/ temperature, control switch, and the intelligent storage/ bathroom systems (PDF)

■ ASSOCIATED CONTENT

■ AUTHOR INFORMATION Corresponding Authors Huanyu Cheng -Department of Engineering Science and Mechanics, The Pennsylvania State University, University Park, Pennsylvania 16802, United State; orcid.org/0000-0001-6075-4208; Email: huanyu.cheng@psu.edu Li Yang -State Key Laboratory of Reliability and Intelligence of Electrical Equipment, School of Health Sciences and Biomedical Engineering, Hebei University of Technology, Tianjin 300130, China; Tianzhong Yimai Technology Development Co., Ltd, Tianjin 300392, China; orcid.org/ 0000-0001-6798-3177; Email: yangli5781@126.com