Introduction

The human somatosensory system includes a variety of sophisticated receptors to process diverse environmental information, enabling complex tasks and interactions1,2,3,4. Advances in computer vision and electronic olfaction have been significant, and wearable sensors are gaining attention for their potential to replicate human sensory functions in robotics and wearables5,6,7,8. Wearable temperature sensors are an important part of the wearable sensors. In recent years, research on temperature sensors has focused on the field of healthcare, leading to the development of a diverse array of wearable temperature sensors. These developed wearable sensors are mostly designed for real-time body temperature monitoring and offer unprecedented potential in terms of continuous disease diagnostic capabilities within healthcare applications9,10,11,12. However, there has been relatively little research into the development of temperature sensors for use to reflect the thermal sensations of the human body. Our thermal receptors can collect a great deal of temperature information delivered by various heat exchange processes on skin, such as feeling ambient temperature and a light breeze, which is attributed to their ultra-high temperature sensitivity (0.02 °C)13. The sensitivity mismatch between the temperature sensor and the human skin will result in a difference in its detection of skin temperature and human feelings, so that it fails to accurately reflect the changes in human delicate thermal feelings. Therefore, replicating exquisite thermal sensation places extremely high demands on the sensing characteristics of temperature sensors, requiring even greater sensitivity than human thermal receptors. The development of wearable temperature sensors with ultra-high sensitivity remains challenging due to limitations in both material’s thermal properties and device assembly techniques14,15.

Temperature sensing materials, as the core of sensors, are crucial to improve the sensing performance of temperature sensors. Significant efforts have been devoted by scientists to develop materials with high temperature coefficients of resistance (TCR) over the years. Semiconductor materials with electron and hole carriers, such as NiO and VO2, are typical materials with high TCR, which have been chosen in previous work to fabricate flexible thermal receptors16,17. The thermal excitation mechanism of carriers in semiconductor materials results in a significant temperature dependence of their conductivity. However, rigid semiconductor particles exhibit poor compatibility with flexible matrix materials, and the introduction of semiconductor particles tends to degrade the mechanical properties of flexible polymer matrixes, which restricts their applications in flexible electronics.

In comparison, ionic conductors such as hydrogel materials have attracted considerable attention as ideal materials for fabricating bionic artificial thermal receptors due to their superior flexibility and biocompatibility18,19. The researchers have created a range of transparent and stretchable temperature sensors, leveraging the temperature-dependent mobile ions migration mechanism of ionogels. These sensors were then utilized for real-time body temperature monitoring20. Nevertheless, it has been observed that the performance of these developed sensors falls short in terms of precisely synchronizing with the restoration process of human body temperature. Notably, the receptors of the human somatosensory system are essentially ionic conductors, all of which work on the basis of ionic dynamics21,22,23,24. For instance, the thermal receptors in our skin sense temperature with an ultra-high sensitivity of 0.02 °C through Ca2+ channels that belongs to the transient receptor potential (TRP) family25,26,27. Inspired by this, we have developed a silica-in-ionogel (SIG) sensor that utilizes an innovative ion capture and release mechanism. This design incorporates silica (SiO2) microspheres into a conventional ionogel (CIG) matrix, where hydrogen bonding (H-bond) tether ions to SiO2. Elevated temperatures break these bonds, releasing ions and enhancing conductivity. Thus, this unique ion dynamics mechanism endows the sensors with a sensitivity much superior to that of human thermal receptors (0.008 °C) over a wide temperature range (25–85 °C). In addition, the sensors demonstrate excellent linearity (R2 > 0.99) and stability. Moreover, the facile synthesis process permits scalable preparation of functionalized SIG, enabling future development of sensors for commercial production. In addition, the developed sensors are capable of synchronously reflecting the body’s perception of hot and cold sensations, and they accurately capture changes in our apparent temperature caused by clothes and environmental factors, such as air flow, temperature, etc. Thus, these sensors are expected to be used for smart temperature control systems in homes to guarantee thermal comfort for the human.

Results and discussion

Temperature sensitive mechanisms

To obtain the SIG temperature sensor, a facile and feasible method was employed as shown in Fig. 1a. The SIG sensors were fabricated by drop-coating SIG ionogel precursor solutions onto a flexible PET substrate with gold electrodes, followed by Teflon tape encapsulation. The obtained SIG sensors consist of three parts: an electrode substrate, an ionogel temperature-sensitive film, and an encapsulation layer. Notably, the SIG film has good transparency and flexibility, which provides a necessary prerequisite for developing wearable sensors (Fig. S1a, b). The simple mixing method permits the preparation of SIG film in large batches (10 × 10 cm). As a contrast, the sensor size is 2 × 1 cm, and the area covered by sensitive material is only 1 cm2 (Fig. S1c).

Fig. 1: Fabrication and sensing mechanism of SIG sensor.
figure 1

a Schematic diagram of the fabrication process of the SIG sensor. b Schematic diagram of the temperature sensitivity mechanism of the sensor. c FT-IR spectra of CIG and SIG

Full size image

Compared with conventional ionogels, a moderate amount of SiO2 nanospheres was introduced into the SIG ionogel network. And the hydroxyl-rich SiO2 surface promotes hydrogen bonding with ionic liquids, binding ions to SiO2 microspheres and creating a stable, bound ionic state, as shown in Fig. 1b. The formation of hydrogen bonding enables SIG films to retain mechanical properties that are nearly indistinguishable from those of CIG films. Excitingly, the SIG film exhibits favorable mechanical properties with an elongation at break exceeding 900%, and its Young’s modulus is calculated to be 1.06 MPa (Fig. S2).

When a thermal stimulus is applied, thermal activation releases ions from their H-bond bound state, transitioning them into a free state and increasing conductivity. Simultaneously, ion dissociation and ion mobility are also enhanced, further contributing to the increase in conductivity. This ion dynamics mechanism will empower SIG sensor with superior temperature sensing performance.

In the designed sensing mechanism, the binding effect of SiO2 particles on ions is critical. To confirm hydrogen bond formation, FT-IR spectra of both CIG and SIG were measured, as shown in Fig. 1c. The characteristic peak at 3340 cm−1 was associated with N-H stretching, and the peaks corresponding to C-H stretching and C=O stretching were observed at 2956 cm−1 and 1736 cm−1, respectively, all of these above characteristic peaks belong to the TPU matrix28. In addition, the characteristic Si-O peak of SiO2 at 1050 cm−1 was only observed in the SIG spectrum. Notably, the C-F stretching peak of ILs in CIG, initially at 1169 cm−1, shifted to a higher wavenumber (1186 cm−1) in the SIG spectrum, indicating hydrogen bonding between SiO2 and the ILs 29.

To determine the optimal loading content of ILs in the ionogel to obtain a SIG sensor with the highest temperature response, the relationship between the impedance modulus of the CIG sensor and the ILs content was investigated, as shown in Fig. 2a. The results show that the conductivity of the CIG sensor increases significantly with the ILs content when the ILs content in the ionic gel is low, and the conductivity changes slowly after the ionic liquid content is higher than 2 wt%, the change rate of impedance modulus change rate is close to zero. The observed phenomenon implies that the concentration of conductive ions in the ionogels has a significant effect on the conductivity of the sensor in the concentration interval of less than 2 wt%. For designed SIG temperature-sensitive ionogels, the introduction of ionic bound centers reduces the conductivity of ionogels, and the conductivity of SIG ionogels returns to the previous level when the ions are released by thermal activation. Combined with this temperature sensing mechanism, it can be expected that the SIG sensor will have the highest temperature response at an ILs content of 2 wt%.

Fig. 2: Electrical properties of CIG and SIG sensors.
figure 2

a Curve of impedance modulus of CIG sensor with ILs content. b–d Impedance modulus-temperature curves, Arrhenius curves, and parameter comparison histogram of CIG sensor, SIG sensor with varying SiO2 amounts. e–f Curves of dielectric relaxation time of SIG-3 sensor and CIG sensor as a function of temperature

Full size image

To meet the needs of daily temperature monitoring, the temperature sensing performance of SIG sensors was evaluated over a specific temperature range (25–85 °C). And to optimize the temperature response, SIG sensors with different concentrations of SiO2 microspheres (0.05, 0.10, 0.15, and 0.20 mol/L) were obtained, designated as SIG-1 sensor, SIG-2 sensor, SIG-3 sensor, and SIG-4 sensor, respectively. In contrast, a CIG-based sensor was also obtained and named CIG sensor. In AC testing, the electrical signal of the sensor is significantly correlated with the AC voltage frequency. The curves of impedance modulus versus AC voltage frequency for CIG sensor and SIG sensors at different temperatures were recorded as shown in Fig. S3a–e. Since ions cannot fully migrate in time at high AC voltages frequency, ionic charge accumulation is diminished, and the impedance modulus of the sensor decreases dramatically when the AC voltage frequency exceeds 10 kHz. Subsequently, the temperature response of the SIG sensor was measured at different AC voltage frequencies (Fig. S4). The results show that the temperature response of the sensor remains unchanged at AC voltage frequencies less than 1 kHz. Considering the smaller signal noise at high frequencies, the AC voltage frequency was set to 1 kHz in the subsequent tests. As shown in Fig. S5, SIG sensors exhibit lower conductivity compared to CIG sensor. The conductivity of the SIG sensors decreases as the SiO2 content in SIG increases due to the ion-binding effect of SiO2. As shown in Fig. 2b, SIG sensors exhibit negative temperature coefficient (NTC) behavior similar to conventional thermistors due to the specific ion release and conduction mechanisms. Notably, SIG sensors exhibit conductivity levels approaching that of CIG sensor in the high-temperature range, which is attributed to the release of bound-state ions in SIG. Clearly, the conductivity of the SIG sensors varies nonlinearly with temperature, and the nonlinear correlation of them can be described by the Arrhenius empirical equation, consistent with our previous findings (Fig. 2c)30. Furthermore, due to the temperature-controlled hydrogen bonding interaction present in the SIG network, the thermal activation energy (\({E}_{a}\)) is hypothesized to correlate with both ion migration and hydrogen bonding dynamics (Supplementary Text). The SIG sensors demonstrate a higher \({E}_{a}\) due to the ions are bound, resulting in their migration being hindered (Table S1). As shown in Fig. 2d, the temperature sensing parameters of SIG sensors outperform those of CIG sensor. In contrast, SIG-4 sensor shows inferior performance compared to SIG-3 sensor, as the excessive SiO2 content reduces conductivity at high temperature (Fig. S6). The temperature sensing performances of the optimal sensor (SIG-3 sensor) were investigated in detail as follows.

Ion relaxation in ionic conductors is related to the ion content; therefore, in order to further validate the designed temperature-sensitive mechanism, the dielectric relaxation at different temperatures of the SIG-3 sensor was investigated, as shown in Fig. 2e31. As the temperature increases, the dielectric relaxation frequency shifts to a higher frequency, and the relaxation time decreases, which implies an increase in the content of conducting ions. Similarly, the ion relaxation frequency of the CIG sensor was recorded (Fig. S7). As shown in Fig. 2f, temperature causes a greater amount of variation in the ion relaxation time of the SIG sensors compared to the CIG sensors. In addition, impedance analysis was also carried out to verify the designed temperature sensitivity mechanism of the sensor. The complex impedance plots exhibit the typical arc-shaped curve with the arc diameter decreasing as temperature increases, corresponding to the resistance decrease, which indicates that the temperature leads to a significant change in the ion concentration (Fig. S8). As a conclusion, temperature sensitivity of the sensor is improved due to temperature-induced greater conductive ion content changes in SIG ionogels.

Temperature sensing properties

We comprehensively evaluated the temperature response properties of the obtained sensor in the temperature range of 25–85 °C. As shown in Fig. 3a, the developed sensor exhibits classical Arrhenius characteristics, and the Arrhenius plots of the sensor demonstrate a favorable linear relationship between the logarithmic value of the impedance modulus and the inverse of the temperature within the temperature range of 25–85 °C (R2 > 0.99). The thermal index (B) of the sensor was calculated to be as high as 6113 K, and the temperature coefficient of resistance (TCR) at 300 K was approximately −6.79%/°C (Supplementary Text). To evaluate the dynamic response characteristics of the sensors, the dynamic response curve of SIG sensor was measured over heating cycles with various temperature intervals. As shown in Fig. 3b, the SIG sensor exhibits obvious and repeatable temperature response across heating and cooling cycles. Surprisingly, it was measured that the response of the SIG sensor is almost identical for both heating and cooling processes, showing negligible temperature hysteresis (about 0.66 °C), as shown in Fig. 3c. Notably, the symmetric temperature response behavior of the SIG sensor confirms the reproducibility of the designed thermally activated ion dynamics mechanism.

Fig. 3: Temperature sensing performance of SIG sensor.
figure 3

a Arrhenius plots of SIG sensor. b Dynamic response curve of the SIG sensor during various heating cycles. c Impedance modulus-temperature curves as the temperature rise and fall between 25 °C and 65 °C. d Dynamic response curve of the SIG sensor during faint heating processes. e Dynamic curve of SIG sensor response to 6 faint heating cycles. f Scatterplot of temperature measurement of the SIG sensor at constant temperature. g Dynamic response and recovery curve of SIG sensor. h Comparison of TCR and temperature sensitivity range of various wearable temperature sensors38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54

Full size image

Further experiments were carried out to demonstrate the performance of SIG-3 sensor in response to slight temperature variations. First, the impedance modulus change curve of the sensor was recorded when the temperature was increased in temperature increments of 0.05 °C from 36.10 °C to 36.35 °C. As shown in Fig. 3d, a temperature rises of 0.05 °C led to a significant change in impedance modulus of SIG sensor (about 0.33%). Unlike the nonlinear behavior observed in the 25–85 °C temperature range, the impedance modulus of the sensors varies linearly with temperature in the 36.10–36.35 °C temperature interval (R2 > 0.99). The enormous temperature response of the sensor enables the detection of even smaller temperature changes. Impressively, SIG sensor demonstrated clear response signals (about 0.18%) to even finer temperature cycling (0.03 °C) as shown in Fig. 3e, indicating that the sensor has the ability to stably recognize temperature changes of 0.03 °C. And the accuracy of the SIG sensor was determined to be ±0.004 °C, as calculated by comparing its measured temperature values with those from the reference temperature. The sensitivity of temperature sensor is considered as the most critical parameter in determining its performance. The variation in the temperature measurement of the SIG sensor at a constant temperature was recorded to assess its temperature sensitivity. As shown in Fig. 3f, a theoretical temperature sensitivity of ~0.008 °C was obtained near body temperature (35 °C). At higher frequencies, the sensor shows repeatable temperature variation distributions but captures more output temperature fluctuations, increasing the temperature standard deviation from 0.008 °C to 0.010 °C (Fig. S9).

Additionally, SIG sensor demonstrates rapid response and recovery behavior to various temperature shift (Fig. S10). As shown in Fig. 3g, the response time and the recovery time of the sensor to a temperature shift of 40 °C were calculated as 15 s and 225 s, respectively. As a comparison, the performance parameters of wearable temperature sensors and detectors based on hydrogel, ionogel, metal, and carbon materials that have been reported in the literatures were summarized and presented in Fig. 3h and Table S2 the SIG sensor exhibits the highest TCR over a wide temperature range. In summary, the temperature sensitivity performance of the developed SIG sensors outperforms those of state-of-the-art temperature sensors.

Interference immunity and durability of SIG sensors

In practical applications, wearable temperature sensors frequently endure bending, mechanical pressure, and exposure to sweat (Fig. 4a)32,33. Therefore, cycle reliability, mechanical resistance, and environmental stability were evaluated to validate the application potential of the SIG sensor. First, the durability and reliability of temperature sensors are critical, especially for scenarios where trace thermal stimuli need to be sensed. Response attenuation caused by repeated thermal cycling of SIG sensor is a critical consideration that undermines the precision of sensor measurements. As shown in Fig. 4b, the SIG sensor was exposed to over 120 heating and cooling cycles, during which the SIG sensor exhibited temperature response error within 1.55%. The obtained sensors show excellent durability with a measurement precision of 0.23 °C for over 120 heating cycles. Remarkably, the dynamic response curves for the first and last 10 cycles of these 120 cycles were nearly identical, highlighting the sensor’s robustness (Fig. 4c).

Fig. 4: Reliability and crosstalk of SIG sensor.
figure 4

a Schematic diagram of the cross-talk signal generated during the temperature monitoring process. b The dynamic response curves of the SIG sensor to over 120 heating cycles. c First ten and last ten cycles in these heating cycles. d–e Temperature errors caused by pressure crosstalk and bending crosstalk. f Impedance modulus versus humidity curves for SIG sensor before and after packaging

Full size image

In addition, crosstalk caused by mechanical forces is also regarded. Temperature error, defined as the temperature measurement difference of the sensor between the disturbed state and the initial state, is employed to quantify mechanical disturbances. The effect of pressure on the measured value of the sensor temperature was first investigated. As shown in Fig. 4d, the SIG sensor is endured high pressure (400 kPa), resulting in a minimal temperature error of less than 0.1 °C. With applied pressures exceeding 400 kPa, the conductive network of the sensor suffered a large disruption, and its temperature error increased significantly. To investigate the temperature measurement errors of sensors induced by bending, the Arc-Chord Ratio was chosen as a quantitative parameter to characterize the bending extent of the sensor, which is defined as the ratio of the length of an arc covered by the sensor to its corresponding chord as shown in Fig. 4e. The SIG sensors exhibit good insensitivity to bending strains, and extreme bending of the sensor causes a temperature error of less than 0.1 °C. The thin and compact sensitive film of the sensor permits a significant resistance to mechanical interference. As shown in Fig. 4f, conductivity of unencapsulated sensors varies significantly with humidity due to the large effect of moisture on the ion conduction process of ionogel materials. In contrast, the conductivity of the encapsulated sensor is virtually unaffected by humidity, which demonstrates that moisture is well resisted by using waterproof Teflon membrane as an encapsulation layer.

Wearable apparent temperature monitoring applications

Thermal sensation refers to the subjective sensation through which the human body perceives ambient temperature and its fluctuations via cutaneous receptors. A close and complex relationship exists between thermal sensation and apparent temperature, making the latter a valid indicator for assessing thermal comfort34. As we know, the apparent temperature of human is influenced by a variety of factors, such as air flow speed, clothing, and sunlight conditions35. As shown in Fig. 5a, we tend to experience thermal discomfort when these environmental factors change, and an unsuitable temperature environment will increase the risk of illness, necessitating manual adjustments of the air conditioner to restore thermal comfort. The high sensitivity of the obtained sensors, which exceeds human thermal receptors, enables them to capture subtle variations in the apparent temperature. Therefore, automatic dynamic adjustment of ambient temperature based on sensor measurements could significantly improve living comfort and minimize additional user intervention.

Fig. 5: Apparent temperature perception and smart control system.
figure 5

a Factors influencing thermal comfort. b Curve of apparent temperature versus ambient temperature. c Curve of apparent temperature variation caused by incandescent lamp irradiation. d Curve of apparent temperature variation caused by air flow rate. e Curves of apparent temperatures attributed to clothing conditions. f Schematic diagram of smart temperature control system

Full size image

In order to demonstrate the ability of the sensor to reflect the thermal feeling of the human body, the sensor was attached to the arm to measure the apparent temperature of the subject. First, an experiment was carried out to measure the change in the subject’s apparent temperature as the ambient temperature varied. As shown in Fig. 5b, when the ambient temperature increased by 4.0 °C, the sensor measured an increase in our apparent temperature (about 2.0 °C), which is attributed to the fact that we lose less heat when we are in higher ambient temperatures. Similarly, under sufficient sunlight exposure, the skin absorbs the heat from sunlight, causing an increase in apparent temperature. In order to quantitatively investigate the apparent temperature changes induced by sunlight irradiation, fluorescent lamps were used to simulate sunlight irradiation. When a fluorescent light (power: 3 W, distance: 20 cm) illuminated the subject’s arm, the sensor at the arm exhibited a pronounced temperature response, with its temperature measurements increasing by ~0.3 °C. When the two fluorescent lights were turned on, the sensor measured an increase in the subject’s apparent temperature from about 33.4 °C to about 33.8 °C (Fig. 5c).

Conversely, faster air flow accelerates heat loss, reducing apparent temperature. As shown in Fig. 5d, when the air flow speed was increased, the temperature measurement of the sensor showed a significant decrease. The results show that an air flow speed of 6.0 m/s resulted in a temperature decrease at the arm of ~2.5 °C compared to no air flow passing, requiring higher air-conditioning temperatures to maintain comfort. In addition, clothing status also affects apparent temperature36. The sensor was employed to monitor changes in the apparent temperature of the arm when the sleeves of the shirt were pulled up and down, respectively. Wearing clothes raises the apparent temperature compared to bare skin, and thicker clothes keep us warmer (Fig. 5e). Therefore, the setting of comfortable ambient temperature should also take into effect the clothing status.

According to the above results, it is clear that simply setting the air conditioning temperature does not ensure lasting thermal comfort, highlighting the importance of a dynamic smart home temperature control system based on apparent temperature feedback (Fig. 5f)37. A simple proof-of-concept smart temperature control system has been developed to enable closed-loop regulation of environmental conditions to maintain appropriate human body temperature (Figs. S11–S12). In summary, it is a feasible strategy to maintain human thermal comfort by establishing a temperature control system guided by apparent temperature detected via wearable temperature sensors.

Although the present study did not include a systematic human-subject investigation, it clearly demonstrates the potential of high-sensitivity temperature sensing for wearable thermal-management applications. Future integration of artificial-intelligence algorithms with tailored big-data training models is expected to provide adaptive learning capabilities, enabling intelligent control systems that autonomously regulate built environments on the basis of the precise apparent-temperature data captured by these sensors, thereby delivering truly personalized thermal comfort.

Conclusion

In summary, a wearable temperature sensor based on the ion capture and release mechanism of silica-in-ionogel film was developed. The synergistic mechanism combining thermally activated ion release with thermally enhanced ion migration significantly improves sensing performance. The obtained sensor features a high thermal index (B = 6113 K), an ultra-high TCR (−6.79%/ °C), excellent durability (precision as 0.23 °C in over 120 thermal cycles), and exciting linearity across a wide range (25–85 °C). Notably, the optimal SIG sensor exhibits extremely high temperature sensitivity (0.008 °C) and can recognize temperature variations as small as 0.03 °C. Moreover, the developed sensors demonstrate excellent stability and reliability in temperature sensing. Breakthroughs in the essential properties of the sensors have expanded their potential applications. The developed sensors can measure fluctuations in apparent temperature caused by changes in sunlight intensity, air flow speed, and clothing status. This work highlights novel applications of wearable sensors in accurately reflecting human thermal feelings and proposes a new strategy for building a dynamic smart temperature control system to guarantee human thermal comfort.

Experimental sections

Materials

Thermoplastic polyurethane (TPU) particles were purchased from Wanthane Chemical Group Co., Ltd. The 1-Ethyl-3-methylimidazolium Trifluoromethanesulfonate ([EMIM] [TFSI]) was purchased from Aladdin Industrial Inc. N,N-Dimethylformamide (DMF) was acquired from Sinopharm Chemical Reagent Co. Ltd. Silica (SiO2) nanospheres were acquired from Macklin Inc (particle size: 15 nm). All reagents were used without further treatment. Teflon packaging tape and flexible polyethylene terephthalate (PET) electrode substrates were obtained from Alibaba.

Preparation of the CIG and SIG material

First, to obtain the homogeneous TPU solution, 2.0 g of TPU particles were dissolved in 7.8 mL of DMF with continuous stirring for 6 h at 90 °C in an oil bath device. Then, 0.2 mL of ionic liquids (ILs, [EMIM][TFSI]) was introduced as conductive phases into the TPU solution with continued magnetic stirring for 2 h at 90 °C to obtain the CIG precursors solution. Later, different amounts of commercial SiO2 nanospheres were added to the above mixture and then stirred for another 2 h at 90 °C to prepare the SIG precursors solution. Finally, the CIG precursor solution and SIG precursor solution were transferred to polytetrafluoroethylene (PTFE) molds, and the CIG and SIG ionogel films were obtained after evaporation of the DMF solvent in a 60 °C oven for 4 h.

Design of CIG and SIG sensors

Similarly, an appropriate amount of the CIG/SIG precursor solution was applied dropwise to the flexible substrate, and the CIG/SIG ionogel film was firmly bonded to the PET electrode substrate after the solvent evaporated. Then, the CIG/SIG sensors were obtained by encapsulating the PET electrode substrate with the ionogel attached using PTFE tape.

Characterization

Fourier transform infrared (FT-IR) spectra of the CIG and SIG ionogel films were recorded on a WQF-510A FTIR spectrometer, covering the range from 400 to 4400 cm−1. The temperature sensing properties of sensors were recorded by the Keysight E4990A impedance analyzer, scanning AC voltage frequency range of 100 Hz to 10 MHz, and the sinusoidal voltage was set to 1 V. The temperature response of the sensors is defined as the change in the relative impedance modulus (ΔZ/Z0 · 100%) of the sensor caused by a variation in temperature. The response and recovery time of the sensor is defined as the time required for the response signal to reach 90% of the maximum response. The accuracy of the sensor is defined as the differences between the temperature values measured by the sensor and the corresponding reference temperature values provided by a temperature calibrator. The precision of the temperature sensor is defined as the standard deviation of temperature measurements obtained from multiple repeated tests.

The temperature environment required for testing the sensing characteristics of the sensors was generated by a home-built temperature control and calibration system, which included heating pad and Peltier element for temperature control and thermometer (YET-720L, KAIPUSEN, China) for temperature calibration. The heat generated by the heating pad is proportional to its operating voltage, and adjusting the operating voltage allows precise control over the heat output, thereby inducing slight changes in the sensor’s temperature. To comprehensively assess the sensor’s temperature sensitivity performance, the output temperature measurements of the sensor were recorded over a period of time in a constant temperature environment at a sampling frequency of 10 Hz, and the standard deviation of temperature measurements was then taken as the theoretical sensitivity of the sensor. To apply a uniform and quantitative pressure to the sensor, the sensor was sandwiched between square pieces of glass (1 × 1 cm), and the F305-IMT mechanical test system, ranging from 0.5 N to 500 N (MARK-10, Beijing), was utilized to apply pressure to the sensor. To investigate the impact of bending deformation on the sensor’s sensitive performance, the sensor was secured to a stepper motor (EB 1204 and CL-01A, HAIJIE Technology, Beijing), enabling it to undergo bending to various degrees. The degree of bending of the sensor can be accurately characterized by utilizing the ratio of the arc lengths to the chord lengths (Arc-Chord Ratio) of the arcs that align with the sensor’s shape. To evaluate the effect of humidity on the sensor, various humidity environments were created by different saturated salt solutions at their equilibrium states, including LiCl for 11% RH, MgCl2 for 33% RH, Mg(NO3)2 for 54% RH, NaCl for 75% RH, and KNO3 for 95% RH. Anemometers (UNI-T, UT363) were used to measure the speed of air flow in the environment.