Advanced Hub Main Navigation Menu

Volume 35, Issue 1 2407759
Research Article
Open Access

Highly Stretchable Thermoelectric Fiber with Embedded Copper(I) Iodide Nanoparticles for a Multimodal Temperature, Strain, and Pressure Sensor in Wearable Electronics

Kukro Yoon

Kukro Yoon

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Sanghyeon Lee

Sanghyeon Lee

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Chaebeen Kwon

Chaebeen Kwon

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Chihyeong Won

Chihyeong Won

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Andrew and Peggy Cherng Department of Medical Engineering, Division of Engineering and Applied Science, California Institute of Technology, Pasadena, CA, 91125 USA

Search for more papers by this author
Sungjoon Cho

Sungjoon Cho

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Seungmin Lee

Seungmin Lee

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Minkyu Lee

Minkyu Lee

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Jinhan Lee

Jinhan Lee

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

Search for more papers by this author
Hyeokjun Lee

Hyeokjun Lee

Department of Robotics and Mechatronics Engineering, Daegu Gyeongbuk Institute of Science and Technology (DGIST), 333 Techno Jungang-daero, Hyeonpung-eup, Dalseong-gun, Daegu, 42988 Republic of Korea

Search for more papers by this author
Kyung-In Jang

Kyung-In Jang

Department of Robotics and Mechatronics Engineering, Daegu Gyeongbuk Institute of Science and Technology (DGIST), 333 Techno Jungang-daero, Hyeonpung-eup, Dalseong-gun, Daegu, 42988 Republic of Korea

Search for more papers by this author
Byeonggwan Kim

Corresponding Author

Byeonggwan Kim

Department of Chemical Engineering and Applied Chemistry, Chungnam National University, 99 Daehak-ro, Yuseong-gu, Daejeon, 34134 Republic of Korea

E-mail: [email protected]; [email protected]

Search for more papers by this author
Taeyoon Lee

Corresponding Author

Taeyoon Lee

School of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722 Republic of Korea

E-mail: [email protected]; [email protected]

Search for more papers by this author
First published: 09 July 2024
Citations: 1

Abstract

Thermoelectric (TE) fibers have excellent potential for multimodal sensor, which can detect mechanical and thermal stimuli, used in advanced wearable electronics for personalized healthcare system. However, previously reported TE fibers have limitations for use in wearable multimodal sensors due to the following reasons: 1) TE fibers composed of carbon or organic materials have low TE performance to detect thermal variations effectively; 2) TE fibers composed of rigid inorganic materials are not stretchable, limiting their ability to detect mechanical deformation. Herein, the first stretchable TE fiber-based multimodal sensor is developed using copper(I) iodide (CuI), an inorganic TE material, through a novel fabrication method. The dense CuI nanoparticle networks embedded in the fiber allow the sensor to achieve excellent stretchability (maximum tensile strain of ≈835%) and superior TE performance (Seebeck coefficient of ≈203.6 µV K−1) simultaneously. The sensor exhibits remarkable performances in strain sensing (gauge factor of ≈3.89 with tensile strain range of ≈200%) and pressure sensing (pressure resolution of ≈250 Pa with pressure range of ≈84 kPa). Additionally, the sensor enables independent and simultaneous temperature change, tensile strain, and pressure sensing by measuring distinct parameters. It is seamlessly integrated into a smart glove, demonstrating its practical application in wearable technology.

1 Introduction

With the development of smart wearable electronic devices and IoT technologies, mobile medical fields have been remarkably advanced in recent years. In particular, wearable electronic devices for personalized healthcare systems based on long-term and real-time human body monitoring have been significantly spotlighted.[1, 2] These intelligent devices consistently record human biosignals and track their metabolic conditions in real time. To detect accurate human body data such as movement,[3-7] body temperature,[8-10] and human-object contact,[11-14] smart sensors inspired by human skin somatosensory systems are needed.[15-17] The detection of mechanical (tensile strain and pressure) and thermal (temperature change) stimuli is significant. Sensors must be flexible and stretchable to detect mechanical stimuli, while using thermoelectric (TE) materials is essential for detecting temperature changes.[18-23]

Fiber-type sensors are ideal for wearable electronic devices because they are lightweight, soft, flexible, durable, and easy to integrate into clothing.[24-27] Several studies have reported on stretchable fiber-type sensors made from TE materials that can simultaneously sense strain and temperature.[28-31] Using the wet-spinning method, Zhang et al. fabricated a stretchable TE fiber composed of carbon nanotube (CNT) and polymer (WPU/PVA).[28] The fiber exhibited a Seebeck coefficient (S) of 44 µV K−1 and a power factor (PF) of 1.78 µW m−1 K−2 at room temperature, with a maximum tensile strain of 27%. Wan et al. developed a fiber-type temperature–strain sensor by coating PEDOT:PSS on stretchable yarns.[29] This fiber-type sensor showed an S of 15 µV K−1 and a PF of 0.68 µW m−1 K−2 at room temperature, with a maximum working strain of 200%. Although these aforementioned sensors show considerable potential for simultaneously detecting tensile strain and temperature, their utilization is limited by the inherently low TE performances of organic or carbon-based materials. Inorganic materials typically exhibit higher TE performance; however, their rigidity and fragility can endow fiber-type sensors made from these materials brittle.[32-35] Thus, developing stretchable sensors using inorganic materials remains a significant challenge. To circumvent this issue, Kwon et al. designed a novel approach, wherein Bi2Te3 nanoparticles were densely generated inside and outside a cotton fabric.[19] This Bi2Te3 TE fabric exhibited an S of −83.79 µV K−1 and a PF of 25.77 µW m−1 K−2 at room temperature, with pressure and strain sensing of up to ≈32 kPa and 200%, respectively. Although this fabric functions well as a multifunctional sensor, it is woven in such a form that only non-stretchable cotton textiles are structurally allowed to be stretchable, which limits its practical use as a strain sensor.

In this study, we developed a stretchable TE fiber-based multimodal sensor using copper(I) iodide (CuI), inspired by the human skin's somatosensory system. This sensor can detect changes in tensile strain, pressure, and temperature. CuI is a p-type semiconductor material with a wide bandgap (Eg) of 3.1 eV, attributed to the presence of copper vacancies. It has a single type of carrier (hole) with a high mobility (µ > 40 cm2 V−1 s−1 in bulk material), which is advantageous for achieving a large S. Hence, CuI is regarded as a promising inorganic TE material.[36-38] Moreover, CuI is an environmentally friendly material composed of nontoxic and naturally abundant elements compared with other representative TE materials such as Bi2Te3, PbTe, SnSe, and Sb2Te3. Herein, we used a novel solution-based chemical synthesis method to embed CuI nanoparticles (CuINPs) into polyurethane (PU) polymer fibers. The embedded CuINPs form dense networks inside the PU polymer, enabling the stretchable TE fiber-based multimodal sensor to possess high electrical conductivity (σ = 2.965 S cm−1) and TE performances (S = ≈203.6 µV K−1 and PF = ≈12.29 µW m−1 K−2). When the CuI fiber sensor is stretched, CuINP networks resist deformation within the PU polymer, maintaining electrical connection and TE properties. This design enables the CuI fiber sensor to achieve extensive stretchability (maximum tensile strain of ≈835%) and high durability, maintaining stable performance across over 10 000 stretching cycles. This represents an unprecedented advancement over previously reported state-of-the-art stretchable TE fibers (Table S1, Supporting Information).[28-31, 39-48] The CuI fiber sensor demonstrates a high gauge factor (GF) of ≈3.89 across a strain sensing of up to 200%. In addition, a capacitive pressure sensor was fabricated by stacking two fibers coated with a PU dielectric layer. This sensor exhibits a high-pressure sensing range of ≈84 kPa with a precise resolution of ≈250 Pa. Thus, the fabricated CuI fiber sensor can simultaneously and independently detect the temperature change, tensile strain, and pressure. Furthermore, sensors were integrated into a wearable smart glove as a proof of concept to demonstrate their practical application, displaying their capability to sense temperature, strain, and pressure simultaneously.

2 Results and Discussion

2.1 Fabrication of CuINPs Embedded Stretchable TE Fiber-Based Multimodal Sensor

The stretchable multimodal sensors are constructed with low-modulus polymer fibers containing density-controllable CuINPs. Figure 1a illustrates the facile fabrication process for producing the CuINPs embedded stretchable TE fiber through a novel solution process-based chemical synthesis method. To embed CuINPs uniformly into the fiber, we utilized a two-step solution process: 1) absorbing Cu ions into the fiber matrix and 2) embedding CuINPs into the fiber through iodization of the absorbed Cu ions. Commercial pristine-stretchable fibers were immersed in a Cu precursor solution, followed by hydroiodic acid (HI) injection as an iodization agent. The HI solution was quickly absorbed into the Cu precursor-absorbed fiber, causing the Cu ions and I ions to react and form CuI, which is embedded into the fibers as nanoparticles. The fibers were rinsed in ethanol several times to remove residual iodine. The density of CuINPs could be controlled by repeating the preceding synthesis processes. While previous methods for fabricating stretchable conductive fibers were confined to embedding metal nanoparticles into a polymer matrix through the reduction of metal ions,[49, 50] our novel solution process-based chemical synthesis method stands out for its capability to embed compound semiconductor nanoparticles within a polymer matrix. Figure 1b shows the representative optical image of the fibers before and after the synthesis of CuINPs. The light purple coloration and slight volumetric increase of the fiber are due to the thoroughly embedded CuINPs inside the fiber. Figure 1c shows a typical top-view of field emission scanning electron microscopy (FE-SEM) images of the fibers. Fully covered CuINPs on the surface of the fibers were observed in the highly magnified image. The average diameter size of CuINPs was ≈165 nm, as detailed in Figure S1 (Supporting Information). The cross-sectional FE-SEM image in Figure 1d demonstrates that CuINPs are uniformly embedded inside the fibers, confirmed through its energy-dispersive X-ray spectroscopy (EDS) mapping analysis. The highly magnified cross-sectional FE-SEM images are provided in Figure S2 (Supporting Information), showing that CuINPs are fully embedded in the polymer matrix and physically well bonded with the PU fiber. The focused ion beam transmission electron microscope (FIB-TEM) image of Figure 1e reveals the densely embedded CuINPs within the PU fiber, capable of forming robust electrical connections. The combination of selected area electron diffraction (SAED) pattern (Figure 1f) and X-ray diffraction (XRD) spectrums (Figure 1g) depicts the crystal structure of the synthesized CuINPs. From Figure 1f,g, the patterns and peaks can be indexed as (111), (200), (220), (311), (400), (331), and (422) planes of the zinc-blende structure of γ-CuI (JCPDS card No. 82-2111).

Details are in the caption following the image
Overview of the fabrication and characterization of the CuINPs embedded stretchable TE fiber. a) Schematic of the fabrication process of the CuINPs embedded stretchable TE fiber. b) Optical images of the pristine PU fiber (top) and CuINPs embedded stretchable TE fiber (bottom). c) Top-view FE-SEM images of the CuINPs embedded stretchable TE fiber. The highly magnified image shows fully covered CuINPs on the fiber surface. d) Cross-sectional FE-SEM and EDS mapping images of the CuINPs embedded stretchable TE fiber. e) FIB-TEM image of CuINP networks forming electrical connections inside the fiber. f) SAED pattern analysis of CuINPs embedded in the fiber. g) XRD spectrums of the CuINPs embedded stretchable TE fiber.

2.2 Physical Analysis of CuINPs Embedded Stretchable TE Fiber Based on Repetitive CuI Synthesis

Figure 2a shows the thermogravimetric analysis (TGA) results for CuINPs embedded stretchable TE fibers with an increasing number of synthesis cycles (Nsc). The weight percentages of embedded CuINPs in the fibers increase from 32.09 to 91.33 wt.% with repetitive synthesis cycles. Figure 2b presents the changes in the cross-sectional area of fibers as Nsc increases, which were obtained from corresponding cross-sectional FE-SEM images. The cross-sectional area merely increases up to the seventh synthesis cycle and significantly increases by more than 1.44 times after the ninth synthesis cycle. The schematics in Figure 2c show how the cross-sectional area of fibers increased in three phases: pristine, 1st–7th, and 9th–13th synthesis cycles. During the fabrication of the fibers, the ethanol solvent with Cu ions penetrates the PU chains, causing the polymer matrix to swells. After the CuINPs are synthesized and embedded, the polymer matrix shrinks sequentially as the ethanol solvent evaporates.[49, 50] As the CuINPs are partially embedded within the porous PU matrix, the cross-sectional area of the fiber marginally increased. The fiber volume, calculated from the cross-sectional area, is estimated to have increased by ≈20% at the seventh synthesis cycle. After the ninth synthesis cycle, the porous PU matrix seems to be entirely filled with CuINPs with a volume expansion of ≈72%. This volume expansion can be attributed to excessive CuINP formation from repetitive synthesis cycles accompanied by volumetric strain. The swelling and shrinking of the PU matrix in the fiber are visible to the naked eye and are also presented through the optical images in Figure S3 (Supporting Information). Figure 2d shows the stress of CuINPs embedded stretchable TE fibers with increasing Nsc against applied tensile strain (εa) up to 200%. Note that the slopes of the stress curves indicate the Young's modulus of the fibers. CuINPs embedded TE fibers were stretched to up to an εa of 200% without fracture. The changes in Young's modulus of each fiber with increasing Nsc are summarized, as shown in Figure 2e. The Young's modulus of the fiber at εa = 5% gradually increased up to 18.95 MPa (Nsc = 7). The increase in Young's modulus at εa of 5% with increasing Nsc can be attributed to the densely formed CuINP networks within the porous PU matrix, which become more rigid as the amount of CuINPs increases. Young's modulus seems saturated beyond the seventh synthesis cycle as sufficient CuINPs are formed. The inset in Figure 2e shows that Young's moduli at εa of 100% and 200% with Nsc increasing up to 7 are similar to those of the pristine fiber. Since crack propagation occurs in the CuINP networks when the fibers are stretched, Young's modulus of the fiber relies on the PU matrix as εa increases. Beyond the ninth synthesis cycle, Young's modulus gradually decreases due to the aforementioned volumetric strain that accumulates initial tensile stress in the PU matrix. Figure 2f demonstrates the initial σ of CuINPs embedded stretchable TE fibers with increasing Nsc. The σ gradually increases to 2.965 S cm−1 up to Nsc = 7 due to CuINPs filling the porous space of the PU matrix. As Nsc increases further, the σ reaches saturation due to the formation of sufficient electrical connections among the accumulated CuINPs. Figure 2g displays the σ of the CuINPs embedded fibers under stretched conditions up to εa of 200%. When the tensile strain is applied to the fibers, the CuINP networks deform along with the PU matrix, gradually decreasing σ. The σ of the fibers with more than nine synthesis cycles degrades more rapidly, attributed to the rigidness of the fibers. The higher content of CuINPs makes the fiber more rigid and brittle. Stretching rigid fibers causes crack propagation in the densely embedded CuINP networks, leading to partial electrical disconnections. Figure 2h indicates the representative TE performances of the fibers with increasing Nsc. The Seebeck coefficient (S), the ratio of an electrical potential gradient to an applied temperature gradient, showed similar levels (S = ≈203.6 µV K−1 at the seventh synthesis cycle) regardless of Nsc, due to the inherent thermal property of CuI. The PF, defined as PF = σS2, exhibited a curve similar to that of the σ (Figure 2g). It gradually increased up to the Nsc = 7 (PF = ≈12.29 µW m−1 K−2 at the seventh synthesis cycle) before reaching saturation.

Details are in the caption following the image
Analysis of the physical properties of CuINPs embedded stretchable TE fibers due to Nsc. a) Weight percentages of CuINPs embedded in fibers due to Nsc investigated via TGA. b) Changes in the cross-sectional area of CuINPs embedded stretchable TE fibers due to Nsc. Scale bars in inset FE-SEM images = 100 µm. c) Schematic of the volumetric increase in fibers during repeated CuINP synthesis cycles. d) Stress–strain curves of CuINPs embedded stretchable TE fibers with Nsc = 1–13 under increasing εa. e) Changes in the Young's modulus of CuINPs embedded stretchable TE fibers at εa of 5%, 100%, and 200% due to Nsc. f) Changes in σ of CuINPs embedded stretchable TE fibers due to Nsc. g) Dynamic changes in σ of CuINPs embedded stretchable TE fibers with Nsc = 1–13 under increasing εa. h) Changes in TE properties (Seebeck coefficient and power factor) of CuINPs embedded stretchable TE fibers due to Nsc.

2.3 Strain–Temperature Sensing Performance of the CuI Fiber Sensor

The CuINPs embedded stretchable TE fiber, with Nsc = 7, proved effective as a fiber-based multimodal sensor due to its remarkable mechanical/electrical properties and TE performances, as confirmed in Section 2.2. Figure 3a shows relative changes in the electrical resistance (R) of the CuI fiber sensor against an εa of up to 200%. This increase in relative changes in the R of the sensor with increasing εa can be attributed to the loss of electrical connections resulting from the propagation of cracks within CuINP networks (refer to inset FE-SEM image in Figure 3a). The sensor demonstrated excellent GF, the ratio between relative changes in R and ɛa. The GF values, derived from the slope of the graph in Figure 3a, were ≈0.36, 1.48, and 3.89 for εa ranges of 0–40%, 80–120%, and 160–200%, respectively. The elastic hysteresis of the CuI fiber sensor was investigated, as shown in Figure S4 (Supporting Information), measuring at εa increment of 20% up to a maximum of 200%. The time-dependent behavior of the viscoelastic polymer induced mechanical hysteresis in the sensor during stretching–releasing cycles.[51, 52] The sensor exhibited an electrical hysteresis during the stretching–releasing cycle, caused by the viscoelastic polymer. Figure 3b shows that the CuI fiber sensor can be maximally stretched up to an εa of ≈835%, exhibiting a Young's modulus of 963.2 kPa before mechanical fracture. Furthermore, the relative changes in the R of the sensor increased up to the maximum εa of ≈835%, indicating a GF of ≈22.01 before mechanical fracture. The analysis of the mechanical and electrical properties of the CuI fiber sensor after 500 stretching-releasing cycles with εa of 100% is shown in Figure S5 (Supporting Information). Despite hundreds of tensile deformations, the maximum tensile strain range, Young's modulus, and GF exhibited no substantial performance degradation, demonstrating the high stability of the sensor. The FE-SEM images in Figure S6 (Supporting Information) confirmed that the CuINPs remained well without delamination even after repeated tensile deformations. As shown in Figure 3c, dynamic strain-sensing responses were measured with repeated loading–unloading of εa at 20%, 60%, 100%, 150%, and 200%. The CuI fiber sensor showed consistent and repeatable responses at all εa levels. The high durability of the sensor was evaluated by subjecting it to an intensive cyclic test of 10 000 stretching–releasing cycles at an εa of 100% (Figure S7, Supporting Information). Figure 3d shows the electrical response of the CuI fiber sensor during a single cycle of increasing and decreasing ΔT. The S of the sensor, derived from the slope of the graph in Figure 3d, was ≈203.6 µV K−1, with no electrical hysteresis observed. Figure 3e shows the representative TE performance of the CuI fiber sensor under various temperature differences (ΔT), output voltage (Vout), and output power (Pout). The maximum Pout (Pmax) was achieved when the internal R of the sensor equals external R.[53, 54] Pmax increased in proportion to the square of ΔT (Pmax = S2ΔT2/2R), with values of 0.59, 1.72, 3.27, 5.42, 9.37, 14.13, 23.55, 48.21, 72.51, 105.184, and 148.13 nW at ΔT = 5, 10, 15, 20, 25, 30, 40, 50, 60, 70, and 80 K, respectively (Figure S8, Supporting Information). In Figure 3f and Table S1 (Supporting Information), the TE and stretching performances of the CuI fiber sensor were compared with those of previously reported stretchable TE fibers.[28-31, 39-48] The CuI fiber sensor exhibited significantly high values for S and maximum εa, marking a notable advancement beyond state-of-the-art technologies. Figure 3g demonstrates that the CuI fiber sensor consistently maintained Vout values under an εa of 0–200% for each ΔT. This suggests that the sensor can detect temperature changes even under stretched conditions. Moreover, the capability of the sensor to detect temperature and strain simultaneously and independently was validated through current–voltage (IV) curves presented in Figure 3h–j. The x-intercepts of IV curves indicate the Vout of the sensor generated by the Seebeck effect, while the reciprocals of the slopes represent its internal R. The IV curve of the sensor with no ΔT or εa passes through the origin and has an internal R of 2.543 kΩ (black line in Figure 3h,i). Figure 3h shows the IV curves of the CuI fiber sensor under various ΔTs without εa. As ΔT increased to 10, 20, 30, and 40 K, the IV curve shifted in the positive direction, causing x-intercepts to increase to 1.83, 3.84, 5.34, and 7.55 mV, respectively while the slope of the curve remained constant. This indicates that increasing Vout does not affect the internal R of the sensor. Figure 3i shows the IV curves of the CuI fiber sensor under various εa values without ΔT. As εa increased to 10%, 20%, 30%, and 40%, the slope of the IV curve decreased without shifts in the IV curve. This decrease in slope indicates an increase in the internal R of the sensor to 5.50, 9.30, 17.18, and 30.43 kΩ, respectively. Figure 3j shows the IV curves of the CuI fiber sensor under various εa values at a constant ΔT of 30 K. All IV curves exhibited the same x-intercept of 5.28 mV. In contrast, the internal R of the sensor increased to 2.57, 5.05, 8.80, 18.53, and 29.75 kΩ as εa increased to 0%, 10%, 20%, 30, and 40%, respectively. These characteristics enable the CuI fiber sensor to detect temperature and strain independently and simultaneously by analyzing changes in the x-intercept and slope of the IV curve.

Details are in the caption following the image
Strain–Temperature sensing performance of the CuI fiber sensor. a) Relative changes in R of the CuI fiber sensor under changing εa. The inset FE-SEM shows the crack propagation of CuINP networks. b) Stress–strain curve and relative changes in the R of the CuI fiber sensor against εa, extending to the point of mechanical fracture. c) Dynamic relative changes in R of the CuI fiber sensor under repeated stretching–releasing cycles at various εa. d) Dynamic changes in the Vout of the CuI fiber sensor under changing ΔT. e) Vout and Pout curves of the CuI fiber sensor as a function of the load current under various ΔTs. g) Vout of the CuI fiber sensor under stretched conditions (εa = 0–200%) at each ΔT. h) IV curves of the CuI fiber sensor under various ΔTs with no εa. i) IV curves of the CuI fiber sensor under various εa with no ΔT. j) IV curves of the CuI fiber sensor under various εa with fixed ΔT of 30 K.

2.4 Pressure Sensing Performance of the CuI Fiber Sensor

Figure 4a illustrates the fabrication procedure of the capacitive pressure sensor constructed using CuINPs embedded stretchable TE fiber. The pure PU solution was applied to coat the PU dielectric layer uniformly on the fiber surface. Subsequently, a couple of PU-coated fibers were perpendicularly stacked to each other, forming a capacitor structure at the cross point of the fibers. Figure 4b explains the working principle of the CuI fiber-based pressure sensor. The capacitance (C) generated at the cross point of the PU-coated fibers is defined as C = κϵ0(A/d), where κ is the dielectric constant of the PU, ϵ0 is the vacuum permittivity, A is the contact area between the PU-coated fibers, and d is the thickness of the PU dielectric layers. When pressure is applied to the sensor, the elastomeric PU dielectric layers are compressed, which 1) reduces the thickness of the dielectric layer between fibers and 2) increases the contact area between fibers. These changes consequently increase the C of the sensor. Figure 4c shows the relative changes in C of the pressure sensor against the applied pressure (pa) up to ≈84 kPa, with no capacitive hysteresis. The sensitivity of the pressure sensor (sp) is defined as the ratio between relative changes in C and pa, which can be derived from the slope of the graph. A high sp of ≈8.529 k Pa−1 was obtained in the low pa range under 4 kPa. The sp decreased to ≈0.838 and 0.194 k Pa−1 for the pa ranges of 10–30 and 45–84 kPa, respectively, due to the elastomeric behavior of the PU dielectric layer.[55-57] The inset photograph in Figure 4c represents the actual optical image of the pressure sensor. Based on the high sp of the pressure sensor, the resolution of the sensor was investigated by loading and unloading a pa of 250 Pa (Figure 4d). The C of the sensor approximately increased by 6.39% from 4.38 to 4.66 pF, before and after loading a pa of 250 Pa. Figure 4e shows the dynamic pressure sensing responses of the pressure sensor, with repeated loading–unloading of pa of 4, 20, 40, and 60 kPa. The sensor showed consistent and repeatable responses at all pa levels. Figure 4f,g demonstrates the relative changes in the R and C of the CuI fiber-based pressure sensor when εa and pa are simultaneously applied. As shown in Figure 4f, relative changes in the R of the sensor shown in Figure 4g remain constant regardless of pa at all εa levels. Conversely, relative changes in the C of the sensor decrease as εa increases across all pa levels. This phenomenon can be attributed to the reduction in d and A of the sensor due to the Poisson's ratio, as detailed in the Supporting Information (Figures S9–S11, Supporting Information). Figure 4h,i demonstrates Vout and relative changes in the C of the pressure sensor when ΔT and pa are simultaneously applied. The Vout of the sensor remains constant regardless of pa at all ΔT, as shown in Figure 4i. Meanwhile, relative changes in the C of the sensor seem to slightly decrease as ΔT increases across all pa levels, as shown in Figure S12 (Supporting Information). Since Vout due to ΔT affects the measurement of C, smaller C values are recorded compared with the actual C values, as detailed in the Supporting Information (Figure S13, Supporting Information). Thus, relative changes in the C of the sensor data are effectively calibrated based on Vout values and depicted in Figure 4i. The results shown in Figure 4f–i demonstrate that the sensor can simultaneously detect temperature and strain, completely independent of pressure influences. In addition, the measured C can be effectively calibrated based on the measured values of relative changes in R and Vout, enabling the simultaneous measurement of pressure along with temperature and strain.

Details are in the caption following the image
Capacitive pressure sensing performance of the CuI fiber sensor. a) Schematic of the fabrication process of the capacitive pressure sensor using PU-coated CuINPs embedded stretchable TE fibers. b) Schematic of the working principle of the capacitive CuI fiber-based pressure sensor. c) Relative changes in C of the CuI fiber-based pressure sensor under changing pa. The inset image shows the photograph of the CuI fiber-based pressure sensor. d) Capacitive response of loading and unloading pa of 250 Pa, exhibiting precise resolution of the CuI fiber-based pressure sensor. e) Dynamic relative changes in the C of the CuI fiber-based pressure sensor under repeated loading–unloading cycles at various pa. f,g) Changes in the relative R and relative C of the CuI fiber-based pressure sensor under simultaneous stimuli of various εa and pa. h,i) Changes in the Vout and relative C of the CuI fiber-based pressure sensor under simultaneous stimuli of various ΔT and pa.

2.5 Wearable Application Using the CuI Fiber Sensor

To demonstrate the practical application of the CuINPs embedded stretchable TE fiber-based multimodal sensor, we developed a multiple-sensing system by seamlessly integrating the CuI fiber sensor into a smart glove. Figure 5a demonstrates the schematic illustration of the wearable smart glove for temperature–strain–pressure multiple sensing. The CuI fiber sensor was woven into the finger part of the smart glove by winding it in a perpendicular direction to the finger joint. Then, another fiber sensor was cross-woven at the fingerprint part of the fingertip, as shown in the schematic of the finger front view. The PU dielectric layers were coated on the cross point of two fiber sensors, enabling the fibers to be used as a capacitive pressure sensor. The fiber sensor wound on the finger can be used as temperature and strain sensors by measuring the Vout and R, as shown in the schematic of the finger side view. The estimated parameters are converted into digital signals through the analog signal processor and transmitted to light-emitting diodes (LEDs). The real images of the wearable smart glove and CuI fiber sensors are presented in Figure S14 (Supporting Information). Figure 5b explains how the smart glove is operated with a CuI fiber sensor. The multiple stimuli, including i) temperature change, ii) tensile strain, and iii) pressure, are simultaneously and independently monitored in real-time via distinct parameters: Vout, relative changes in R, and relative changes in C of the CuI fiber sensor integrated into the smart glove, respectively. Each parameter triggers the activation of red, yellow, and green LEDs, allowing for the visual confirmation of each stimulus detected. The circuit block diagram presented in Figure S15 (Supporting Information) shows how the smart glove and analog signal processor read parameters for each stimulus and control the LEDs. Figure 5c and Movie S1 (Supporting Information) show the continuous single-mode sensing results for the individual stimuli of i) temperature change, ii) tensile strain, and iii) pressure. Each red, yellow, and green LED on the smart glove lights up when the finger i) is exposed to hot air, ii) is bent, and iii) presses on a sponge. Figure 5d and Movie S2 (Supporting Information) show the multiple-sensing results for complex temperature–strain–pressure stimuli. When a hot water-filled beaker was grabbed and lifted, all LEDs on the smart glove light up simultaneously. The real-time signal data for the single-mode and multiple sensing demonstrations corresponding to Figure 5c,d, respectively, are presented in Figure S16 (Supporting Information). Thus, it has been proved that CuINPs embedded stretchable TE fibers can be utilized as multimodal sensors in practical wearable electronic devices, enabling the simultaneous and independent measurement of mechanical and thermal stimuli.

Details are in the caption following the image
Wearable smart glove application using the CuI fiber sensor. a) Schematic of a wearable smart glove for temperature–strain–pressure multiple sensing. The finger front and side views reveal the locations where each sensing parameter (Vout, R, and C) can be measured. b) Operating principle of the smart glove with CuI fiber sensor for detecting i) temperature change, ii) tensile strain, and iii) pressure. c) Continuous single-mode sensing results for each stimulus of i) temperature change, ii) tensile strain, and iii) pressure. Each red, yellow, and green LED was turned on. d) Multiple-sensing results for simultaneous stimuli of temperature change, tensile strain, and pressure. All the red, yellow, and green LEDs were turned on.

3 Conclusion

In this study, we developed a highly stretchable TE fiber-based multimodal sensor using an inorganic material for the first time by embedding CuINPs within a stretchable polymer fiber through our novel solution-based chemical synthesis method. The fabricated CuI fiber sensor showed the capability to independently and simultaneously detect both mechanical (tensile strain and pressure) and thermal (temperature change) stimuli by measuring distinct parameters (the relative changes in R, the relative changes in C, and Vout) for each stimulus. The CuI fiber sensor exhibited excellent stretchability with a maximum tensile strain of ≈835% and showed a high GF of ≈3.89 at a strain sensing range of ≈200%. Due to the dense networks of CuINPs, the sensor showed high durability under a stretching–releasing cyclic test of 10 000 cycles. In addition, the CuI fiber sensor achieved remarkable TE performance with a high σ of ≈2.965 S cm−1, S of ≈203.6 µV K−1, and PF of ≈12.29 µW m−1 K−2. The pressure sensor constructed by stacking two CuI fiber sensors coated with a PU dielectric layer showed a high-pressure sensing range of ≈84 kPa with a precise pressure resolution of ≈250 Pa. We seamlessly integrated the sensors into a smart glove to demonstrate the wearable application of the CuI fiber sensor. The smart glove showed impressive multiple sensing capabilities by independently and simultaneously detecting temperature change, tensile strain, and pressure. This study establishes the potential for future advancements in personal healthcare systems and human-machine interfaces by introducing a wearable multiple-sensing system that mimics the human skin somatosensory system.

4 Experimental Section

Fabrication of the CuINPs Embedded Stretchable TE Fiber

The commercial PU-based 840 denier spandex fibers (ELAFIT, Taekwang) were used as pristine-stretchable fibers. Copper (II) trifluoroacetate acetate (Cu(CF3COO)2, Thermo Fisher Scientific Inc.) dissolved in ethanol (40 wt.%) was used as a Cu precursor solution, and the 1:2 volume mixture of 57% hydroiodic acid solution (HI, YAKURI pure chemicals co.) and ethanol was used as an iodization agent. The pristine fibers were immersed in Cu precursor solution for 30 min to absorb Cu ions, and the iodization agent was subsequently injected into fiber-immersed Cu precursor solution by a 2.5:1 volume mixture to synthesize CuI. After 5 min, the fibers were repeatedly rinsed in ethanol to remove residual iodine and dried on a 70 °C hot plate. The above absorption and iodization processes were repeated to increase the content of CuINPs, forming dense CuINP networks inside the fibers.

Fabrication of CuINPs Embedded Stretchable TE Fiber-Based Capacitive Pressure Sensor

The pure PU solution was prepared by dissolving PU beads (Sigma–Aldrich) in 5 wt.% in an organic solvent, which consisted of a 3:1 weight mixture of tetrahydrofuran (THF, Sigma–Aldrich) and N, N- dimethylformamide (DMF, Sigma–Aldrich). The pure PU solution was partially applied on the vertically fixed CuINPs embedded stretchable TE fibers, and the organic solvent was evaporated for 5 min in air. The above process of applying the PU solution and evaporating the organic solvent was repeated ten times to form a PU dielectric layer uniformly on the surface of fibers. A couple of PU-coated fibers were stacked perpendicular to each other to construct a capacitor structure at the cross point.

Characterizations of CuINPs Embedded Stretchable TE Fiber

The top-view and cross-sectional images and EDS mapping analysis of the CuINPs embedded stretchable TE fiber were obtained via FE-SEM (JSM-IT500HR, JEOL). The CuINPs embedded stretchable TE fibers were hardened in resin to get clear cross-sectional images and cut using an ultramicrotome (Powertome, RMC). The highly magnified image of CuINP networks and SAED pattern of CuINPs were obtained via FIB-TEM (JEM-ARM200F, JEOL). XRD spectrum analysis of the CuINPs embedded stretchable TE fiber was performed by X-ray diffractometer (Ultima IV, Rigaku). The thermogravimetric analyzer (SDT Q600, TA Instruments) performed the TGA analyses to investigate the weight percentages of CuINPs in the fibers.

Multiple-Sensing Performance Evaluation of the CuI Fiber Sensor

The multiple-sensing properties of the CuI fiber sensors were examined using a homemade shielded setup. The stretchability test used a 4-axis motion controller (STM-4-SMS, ST1 Co., Ltd.) and a force transducer (S2M, HBM) with two Peltier modules. The electrical resistance and temperature gradient of the CuI fiber sensors were measured using a digital multimeter (Agilent 34410, Keysight), a data acquisition unit (Agilent 34970A, Keysight), and a source meter (2400 SMU, Keithley). The synchronized Seebeck voltage and temperature gradient were measured by two T-type thermocouples and the temperature gradient was confirmed by a high-resolution infrared (IR) camera (FLIR E40). Six points of ΔT and ΔV were obtained three times for each sample and linearly fitted to calculate the Seebeck coefficient. The pressure sensing test was performed using a custom-made motorized vertical stage (KTINC Co., Ltd) with a force transducer (U1A, HBM). The capacitance of the CuI fiber-based pressure sensor was measured through a capacitance meter (4280 1 MHz C Meter, HP).

Wearable Smart Glove Applications for Multiple Sensing

CuINPs embedded stretchable TE fiber-based multimodal sensors seamlessly integrated into a commercial glove through weaving. The Arduino Uno (R3) was used as an analog signal processor to convert analog signals generated from CuI fiber sensors into digital signals to control LEDs.

Acknowledgements

K.Y. and S.L. contributed equally to this work. This work was supported by the Korea Medical Device Development Fund grant funded by the Korea government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, the Ministry of Food and Drug Safety) (Project Number:1711194307, RS-2020-KD000093). This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government(MSIT) (No. RS-2023-00234581, No. RS-2024-00336147, and No. RS-2024-00460364). This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2022R1I1A1A01068817). This study was supported by the NAVER Digital Bio Innovation Research Fund, funded by NAVER Corporation (Grant No. 3720230070). [Correction added on December 31, 2024, after first online publication: The acknowledgement has been updated in this version.]

    Conflict of Interest

    The authors declare no conflict of interest.

    Data Availability Statement

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