1. Introduction
With the rapid development of society, global textile production has doubled, resulting in enormous economic value. However, simultaneously, according to international reports, more than 150 million tonnes of waste textiles are generated globally each year, while the recycling rate is extremely low [
1]. The recycling rate of waste cotton textiles is less than
in developed countries such as the United States and the United Kingdom (Fig. S1(a) in Appendix A) [
2]. The actual recycling rate of waste textiles in China, a large producer of waste textiles, is less than
(Fig. S1(b) in Appendix A). Cotton textile waste accounts for
of the total waste, and approximately
of waste cotton textiles are mainly incinerated and buried as waste. However, incineration produces large amounts of greenhouse gases and air pollution, and burial releases chemical dyes from waste clothing into the soil, causing significant damage to water and soil resources [
3,
4]. An increasing number of countries are focusing on this problem and strongly advocating for the recycling of used textiles. For example, the Chinese government has issued a call to achieve a
recycling rate of waste textiles by 2030. Therefore, it is important to recycle waste textiles and turn them into high-value-added products, such as fiber-based electronics and smart wearables [
5⇓⇓⇓-
9].
Compared with devices loaded on polymer elastomers, fiber-based wearable electronics can be the ultimate platform for smart wearables by seamlessly integrating functional devices with everyday clothing [
8⇓⇓⇓-
12]. In addition, the three-dimensional porous structure constructed using the fiber material shows excellent heat and humidity exchange and provides a comfortable wearing experience, which adds value to wearable devices for long-term applications. Smart textiles have various capabilities; for example, integrating functionalized fibers enables energy-harvesting technologies such as triboelectric nanogenerators, thermoelectric generators, photovoltaic cells, and piezoelectric nanogenerators [
13⇓⇓⇓⇓-
18]. Fiber-based thermoelectric materials have been widely studied because they can directly convert waste heat from the human body into electricity and power wearable devices [
19⇓⇓⇓⇓⇓⇓⇓-
27]. Wen et al. [
28] reported high-performance stretchable thermoelectric fibers consisting of poly(3,4-ethylenedioxythiophene) (PEDOT): poly(styrenesulfonic acid) (PSS) and waterborne polyurethane by a simple one-step wet-spinning process. Temperature sensors based on thermoelectric fibers can precisely distinguish between hot and cold stimuli. Sun et al. [
29] reported n- and p-type stretchable thermoelectric fibers that can be used in wearable fabrics. Through rational design and integration, the fabric structure was made capable of harvesting human energy and sensing strain. However, most current fiber-based thermoelectric devices focus on temperature or tensile strain sensing, while the implementation of compression sensing functions, such as electronic skin and smart gloves, imposes higher requirements on the compressive strain sensing properties of materials. In addition, multi-signal decoupling is another goal of smart wearable systems, but only a few studies have been reported on the decoupling of temperature-strain sensing signals based on fiber-based thermoelectric materials.
Herein, we report scalable fiber-based thermoelectric aerogels with excellent elasticity and temperature-compression sensing decoupling by a one-step freeze-drying method through the reuse of cotton fibers extracted from waste cotton fabrics. The aerogels exhibited excellent mechanical properties that were attributed to the crosslinking of trimethoxy(methyl)silane (MTMS) and cellulose. The aerogels with a fast-compressive strain response time of and a Seebeck coefficient of approximately show significant potential for wide-range self-powered temperature sensing. Simultaneous temperature and compression monitoring was achieved by converting the external stimuli into separate electrical signals using independent thermoelectric and piezoresistive effects in a single thermoelectric aerogel. Remarkably, these signals decoupled and did not interfere with each other. Aerogels have been made into compressive strain sensors for guessing games and recognizing human facial expressions. Furthermore, a large-scale sensing array device is prepared and integrated with a wireless Bluetooth module. The integrated sensing system can perform hot and cold object recognition and high-temperature warnings on the mobile phone side without signal interference caused by compressive strain. Sustainable preparation of thermoelectric aerogels from waste cotton fibers achieves closed-loop recycling of waste textiles, reduces environmental pollution, and provides an effective path for the high-value-added use of waste textiles.
2. Experimental section
2.1. Materials
PEDOT:PSS (1.3 wt% dispersion in , conductive grade) was obtained from Sigma-Aldrich (USA). Dimethyl sulfoxide (DMSO) was purchased from Sinopharm Chemical Reagent Co., Ltd. (China). Single-walled carbon nanotube (CNT, purity > 90 wt%, diameter 1- ) were purchased from Chengdu Organic Chemicals Co., Ltd. (China). MTMS was obtained from Shanghai Macklin Biochemical Co., Ltd. (China). The waste cotton fabric was obtained. All chemicals were used without further purification.
2.2. Preparation of cotton staple fibers
First, waste cotton fabric was washed and dried to remove impurities. The fabric was then cut into small pieces and soaked in deionized water. Thereafter, the small pieces were pulverized under high-speed mechanical conditions of to obtain dispersed cotton staple fibers. Finally, the staple fibers were collected for future use by filtering.
2.3. Fabrication of cotton fiber-based thermoelectric aerogel
Cotton staple fibers(0.4g)were dispersed in water and stirred for at , which ensured that the fibers were evenly dispersed. A mixture of CNT and PEDOT:PSS was added to the homogeneous dispersion in a ratio. In this system, the solid content of both CNT and PEDOT:PSS was . DMSO was then added to the solution at a PEDOT:PSS volume ratio of . Then, of silane coupling agent was added dropwise in the system with continuous stirring for . At last, the mixed system was put into a specific mold and freeze-dried for at to obtain a thermoelectric aerogel.
2.4. Fabrication of the dual mode sensor and the array device
The obtained thermoelectric aerogel was cut into small cubes by manual machining. The top and bottom surfaces where the electrodes needed to be mounted were trimmed to be smooth. The top and bottom sides of the aerogel were then connected to copper wires to prepare a dual-mode sensor, and silver glue was added to reduce the contact resistance.
The large-area array device consists of nine thermoelectric aerogels connected in series. The lower end of the aerogel was connected to the upper end of the aerogel via copper tape, and silver glue was added to enhance the contact.
2.5. Characterizations
The microscopic morphology of the aerogels and the distribution morphology of the CNT were measured using a field-emission scanning electron microscope (FE-SEM; Hitachi SU-8010, Hitachi Ltd., Japan). Thermal conductivity was measured using a Hot Disk TPS 2500S instrument (Hot Disk AB, Sweden). The mechanical performances of the aerogels were tested using a microcomputer-controlled electronic universal material testing machine (CTM2050; Xieqiang Instrument Manufacturing (Shanghai) Co., Ltd. China). The Seebeck coefficients were measured using a custom-built setup. A multimeter (Keithley 2450; Keithley Instruments Inc., USA) was used to measure the output voltage. Peltier elements were attached to both ends of the aerogel to generate a temperature gradient. The sensor signals (resistance and voltage) were recorded in real-time using a Keithley 2450 multimeter.
3. Results and Discussion
Fig. 1(a) shows a schematic of the preparation process of a thermoelectric aerogel. Waste cotton was first briefly treated to remove excess impurities and then cut into small pieces and placed in deionized water. A suspension with uniformly dispersed cotton fibers is then obtained in the presence of a high-speed shear. Subsequently, CNT and PEDOT:PSS were added to the suspension as a thermoelectric material for ultrasonic sonication. Subsequently, MTMS was added dropwise to the mixture under continuous stirring, and the mixture was stirred for an additional
. In particular, the cross-linking reaction between MTMS and cellulose contributes to the mechanical properties of the thermoelectric aerogel after formation (
Fig. 1(b)). Finally, shaped and dimensionally tunable thermoelectric aerogels were obtained by adding the prepared dispersions dropwise into customized molds and freeze-drying them (Fig. S2 in Appendix A). The thermoelectric aerogel exhibited excellent resilience and stability, maintaining good compression elasticity after repeated compression (
Fig. 1(c), Supplementary Video S1 in Appendix A). Overall, owing to their good elastic and electrical properties, the prepared thermoelectric aerogels show promise for applications in wearable electronics, such as hazard-aware warnings, human health management, and human-computer interactions (
Fig. 1(d)).
Owing to the three-dimensional porous structure inside the aerogel, both the thermoelectric aerogel and the cotton fiber aerogel exhibit ultra-lightweight properties, such as the ability to stand on flower petals (
Fig. 2(a), Fig. S3 in Appendix A). As can be seen from the SEM, the fibers inside the aerogel are lapped together, which directly confirms the cross-linking effect (
Fig. 2(b), Fig. S4 in Appendix A). The prepared thermoelectric aerogel showed an ultralow thermal conductivity and density close to those of the cotton fiber aerogel, which is favorable for the construction of devices with reasonable temperature differences for temperature sensing (
Fig. 2(c)). Both cotton-fiber-based and thermoelectric aerogels were tested for compression rebound (by finger pressure) to demonstrate their superiority (
Fig. 2(d)). As shown in
Fig. 2(e), the thermoelectric aerogel exhibited excellent resilience at compressive strains of
, and
, owing to the excellent mechanical properties resulting from the cross-linking action of MTMS. The resistance and compressive strain exhibited a good linear relationship when the thermoelectric aerogel was pressed (
Fig. 2(f)). The corresponding decrease in resistance with increasing compressive strain is owing to the increase in conductive pathways caused by more laps in the fiber network within the aerogel when pressed. We present an equivalent schematic model to enhance our conclusions (Fig. S5(a) in Appendix A). In addition, the electrical conductivities of the aerogels were measured under different compressive strains (Fig. S5(b) in Appendix A). The cyclic resistance response of the aerogels at
, and
compressive strains was evaluated to demonstrate the stability of their sensing (Fig. S6 in Appendix A). The mechanical and thermoelectric properties of the aerogels remained stable after 500 compression cycles (Fig. S7 in Appendix A). In addition, the compression response-rebound time of the device exhibited a fast response time of
and recovery time of
at
(Fig. S8 in Appendix A). In addition to their excellent compression properties, thermoelectric aerogels also exhibited fascinating temperature-sensing properties based on the thermoelectric effect. As shown in
Fig. 2(g), the assembled devices were connected to a self-built test platform for testing, with pelletized elements placed at the ends of the devices to create temperature differences. At room temperature
, the temperature difference
between the two ends of the aerogel shows a good linear relationship with the output thermal voltage. A Seebeck coefficient(S)of
for this aerogel was obtained by fitting, corresponding to the slope of the data in
Fig. 2(h). In addition, the minimum detectable temperature difference of our devices is
, and can distinguish between different temperature differences (Figs. S9 and S10 in Appendix A). The response time of the device at a temperature difference of
is shown in Fig. S11 in Appendix A. Interestingly, when we further measured the Seebeck coefficient at different compressive strains, we found that it was not affected by the compressive deformation of the device (
Fig. 2(i)). This also indicates that the resistance and thermal voltage behaviors of the device were decoupled from each other.
Therefore, we devised a reasonable scheme to demonstrate the decoupling characteristics of the device. According to a typical thermoelectric mechanism, the voltage generated by the device
is defined as
. As shown in
Figs. 3(a) and
(b), when one end of the device is heated by a Peltier element, the temperature difference between the two sides of the device is detected according to the thermoelectric effect, and is expressed as a shift in the current(I)-voltage(V)curve. When further compression was applied, the device deformed owing to the strain applied from the outside, resulting in a change in the intrinsic resistance (
Fig. 3(c)). At this point, the
curve showed an upward shift. When the applied temperature difference was withdrawn, the
curve returned to its origin (
Fig. 3(d)). As shown in
Fig. 3(e), a typical
curve was observed in the absence of any stimulus, indicating that the device exhibited clear thermoelectric characteristics. However, because of the thermoelectric effect of the PEDOT:PSS/CNT, the
curve shifted significantly to the right when different temperature differences were applied. For comparison, the
curves of the devices at different compressive strains are shown in
Fig. 3(f). This suggests that the temperature stimulus has a limited effect on the resistance and that the pressure signal has a negligible effect on the
. These characteristics allow the devices to be tested for decoupled temperature and compressive stimulation using voltage and resistance variations as the output signals.
The reliability of the signal decoupling was further illustrated by applying a series of stimuli to the device, including the initial state, applying pressure, and finally applying a temperature difference. As shown in
Fig. 4(a), when the device is in Step 1, it shows a height of
. When pressure is applied, the height of the device changes to
, as shown in Step 2. A further temperature difference was applied to the device, and the state was recorded in Step 3.
Fig. 4(b) shows the real-time output voltage and relative resistance of the device subjected to temperature differences and compressive strains. First, the device was monitored in a relaxed state and no temperature difference was applied in Step 1. In this state, the
and the relative resistance of the device did not change. In Step 2, a certain amount of pressure is applied to the device, causing it to undergo a pressure deformation. At this point, the resistance decreased significantly, and the voltage remained essentially unchanged. In Step 3, one end of the device is heated and the other end is cooled (kept at room temperature), significantly increasing the thermal voltage of the device, while keeping the resistance unaffected. This value remained stable for the next
. We also show the actual resistance of the device under different compressive-strain states when subjected to different temperatures (Fig. S12 in Appendix A). These results not only provide further evidence that the compressive strain and thermal voltage of the device are decoupled from each other but also that our device maintains the stability of its self-powered temperature sensing even when subjected to compressive strain. Similar decoupling phenomena have been reported previously [
30,
31].
To demonstrate the potential practical applications of the prepared devices, we monitored their ability to recognize fingers (heat sources) and plastic bars (cold sources). This is mainly reflected in the resistance changes and output voltage response of the temperature-compression load-unloading cycle. As shown in
Fig. 4(c), when the device was pressed repeatedly with a plastic bar, the resistance of the device changed regularly, whereas the voltage remained essentially constant. When the device was pressed repeatedly with a finger, the resistance of the device showed a regular change, whereas the thermal voltage also showed a corresponding response (
Fig. 4(d)). This can be attributed to the thermoelectric effect caused by the difference between the finger and device temperatures. With excellent elasticity, dual-mode sensing characteristics, and unique decoupling capabilities, the presented thermoelectric aerogels demonstrated superior comprehensive properties among recently reported fiber-based thermoelectric materials and devices (
Fig. 4(e)) [
32⇓⇓⇓-
36].
Wearable devices often suffer from heat transfer from human waste heat owing to direct contact with the skin or wear, leading to disturbances in the transmitted signal. Based on the superior decoupling, our thermoelectric aerogel can be prepared as a compressive strain sensor for compressive strain sensing without the fear of skin-temperature interference. The compressive strain sensor was prepared by connecting the top and bottom of the aerogel to a copper wire and adding silver glue to reduce contact resistance. We mounted it on the wrist to distinguish between different gestures (Fig. S13 in Appendix A). The most representative example is the guessing game, where the computer port determines the winner by recognizing hand gestures, which has positive implications for human-computer interactions (
Fig. 5 (a)). As shown in
Fig. 5(b), when we show the "Paper" action in the guessing game, the resistance shows a large positive response. When we show "Rock", the resistance shows a negative response (
Fig. 5(c)). When the gesture changes to "Scissor", the resistance shows a weakly positive response (
Fig. 5(d)). Notably, when mounting the device, it was subjected to compressive strain. This is when the sensor primarily monitors two trends of movement on the body surface: protrusion of the muscle under the skin and subsidence of the muscle during localized exertion. When a particular behavior leads to muscle protrusion, it can be equated to a pressing state for the sensor, where the electrical signal is negative. However, when certain behaviors cause the muscle to sink, the compressive strain to which the sensor is subjected at the time of installation disappears, at which point the electrical signal will be positive. In addition, the device was worn directly in the throat to distinguish the facial movements of the human body (
Fig. 5(e)). When the user performs a drinking action, the device shows a negative response in terms of resistance change (
Fig. 5(f)). Conversely, when the user says "Hello," the resistance of the device tends to become smaller (
Fig. 5(g)). We have also added the change in resistance when coughing and saying "OK" to illustrate the advanced application of our device for facial movement recognition (Fig. S14 in Appendix A). Similarly, the resistance signal circumvents the effects of the temperature.
Based on scalable sample preparation, nine thermoelectric aerogels were prepared in tandem to form an array of thermoelectric device regions for practical applications in self-powered temperature-sensing. The device was integrated into a flexible fabric that can be worn in everyday life (Fig. S15 in Appendix A). As shown in Fig. S16 in Appendix A, the device exhibited an increased thermal voltage when the heat source was brought close to the array device, whereas it exhibited a negative thermal voltage response when the cold source was closed. This indicates that our array device can accurately differentiate between hot and cold objects.
To investigate the practical applications of the self-powered sensing and decoupling properties of the device, the user wore the fabric with the device directly on their hand as a glove for hot and cold object detection. As shown in the infrared diagram in
Fig. 6(a), the device exhibited a regular positively correlated thermal voltage response when the user repeatedly approached and moved away from the bottle containing hot water, independent of the change in resistance caused by the compressive strain (
Fig. 6(b), Fig. S17 in Appendix A). Similarly, when the user approached and moved away from the ice water, the thermal voltage was negatively correlated to avoid interference from resistance (Fig. S18 in Appendix A). In addition, owing to the low thermal conductivity of the aerogel, the device maintained a constant thermal voltage over a long period during gripping (
Fig. 6(c)). The compression deformation due to gripping did not interfere with the thermal voltage output of the device, which is further evidence of the advanced nature of our strategy.
We further integrated the device with a wireless circuit to build a wireless temperature differentiation system using Bluetooth transmission (
Fig. 6(d)). The system operates as follows. The voltage signal generated by the thermoelectric effect inside the sensor is collected by an analog-to-digital converter and passed through an operational amplifier. The signal was then analyzed using a microcontroller unit and the data were sent in real-time to a mobile terminal via Bluetooth (Fig. S19 in Appendix A). Using this feature, the system is expected to be used for the remote control of intelligent robots for the recognition of hot and cold objects and to grasp the thermal voltage differences generated by hot and cold objects (
Fig. 6(e)). Importantly, owing to the unique decoupling of the device, interference with the output thermal voltage signal due to the resistance generated by deformation is avoided during the gripping process. The detailed process is shown in
Fig. 6(f). When the robotic hand wearing the device grasped hot water, the application on the mobile phone showed a rapidly increasing thermal voltage response, which could then be remotely judged as hot water using a mobile phone terminal. Similarly, the repeated crawling of hot water responded to the mobile side (Fig. S20 in Appendix A). In summary, the temperature-pres sure-stimulated decoupling properties exhibited by the integrated device have important practical applications and are expected to be widely used as flexible smart components in intelligent robots and health-monitoring products.
4. Conclusions
We developed a waste cotton-derived three-dimensional elastic fiber-based thermoelectric aerogel using scalable, green, and sustainable preparation strategies. The aerogel has a Seebeck coefficient of at room temperature, a thermal conductivity of less than , and a fast response time, which can be used for both compressive strain sensing and temperature sensing. Notably, these sensing signals were decoupled from each other. Therefore, it can be used for human facial expression recognition based on the compressive strain sensing performance while circumventing the interference of human body temperature. Furthermore, owing to the scalability and ease of use of the aerogel, it can be fabricated into an array device and integrated with a wireless module for real-time temperature sensing and high-temperature warnings on mobile devices. Similarly, during the operation of the array device, changes in resistance due to body movement do not affect the thermal voltage output signal of the device, ensuring its stable operation. Overall, the green preparation process of thermoelectric aerogels from waste cotton fibers effectively solves the environmental problems caused by waste textiles and provides an effective method for transforming waste textiles into high-value-added smart wearable devices.
Acknowledgments
This work was partly supported by the grants (51973027 and 52003044) from the National Natural Science Foundation of China, the Fundamental Research Funds for the Central Universities (2232023A-05), the International Cooperation Fund of Science and Technology Commission of Shanghai Municipality (21130750100), and Major Scientific and Technological Innovation Projects of Shandong Province (2021CXGC011004). This work has also been supported by the State Key Laboratory for Modification of Chemical Fibers and Polymer Materials (KF2216), the Donghua University Distinguished Young Professor Program to Prof. Liming Wang, and the Fundamental Research Funds for the Central Universities and Graduate Student Innovation Fund of Donghua University (CUSF-DH-D-2022040) to Xinyang He.
Compliance with ethics guidelines
Xinyang He, Mingyuan Liu, Jiaxin Cai, Zhen Li, Zhilin Teng, Yunna Hao, Yifan Cui, Jianyong Yu, Liming Wang, and Xiaohong Qin declare that they have no conflict of interest or financial conflicts to disclose.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.strusafe.2024.102455.
PDF
(3902KB)
