CNTs decorated elastomer nanofiber composite

Accepted Manuscript Highly stretchable, anti-corrosive and wearable strain sensors based on the PDMS/CNTs decorated elastomer nanofiber composite Ling Wang, Yang Chen, Liwei Lin, Hao Wang, Xuewu Huang, Huaiguo Xue, Jiefeng Gao PII: DOI: Reference:

S1385-8947(19)30013-0 https://doi.org/10.1016/j.cej.2019.01.014 CEJ 20735

To appear in:

Chemical Engineering Journal

Received Date: Revised Date: Accepted Date:

13 October 2018 18 December 2018 3 January 2019

Please cite this article as: L. Wang, Y. Chen, L. Lin, H. Wang, X. Huang, H. Xue, J. Gao, Highly stretchable, anticorrosive and wearable strain sensors based on the PDMS/CNTs decorated elastomer nanofiber composite, Chemical Engineering Journal (2019), doi: https://doi.org/10.1016/j.cej.2019.01.014

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Highly stretchable, anti-corrosive and wearable strain sensors based on the PDMS/CNTs decorated elastomer nanofiber composite Ling Wanga, Yang Chena, Liwei Lina, Hao Wanga, Xuewu Huanga, Huaiguo Xuea and Jiefeng Gao*a, b a

School of Chemistry and Chemical Engineering, Yangzhou University, Yangzhou,

Jiangsu, 225002, China b

State Key Laboratory of Polymer Materials Engineering Sichuan University,

Chengdu, Sichuan 610065, P. R. China *Corresponding authors: [email protected];

Abstract Conductive polymer composite based strain sensors have promising applications in the fields of artificial skin, wearable health-care device, etc. However, fabrication of strain sensors with good stretchability, anti-corrosion, excellent durability and reliability remains challenging. In this work, a superhydrophobic strain sensor based on conductive

thermoplastic

polyurethane/carbon

nanotubes/polydimethylsiloxane

(TPU/CNTs/PDMS) was prepared by ultrasonication induced CNTs decoration onto the electrospun TPU nanofiber surface, followed by the PDMS modification. Uniformly dispersed CNTs on the nanofiber surface with a hierarchical structure construct the conductive network. The PDMS layer with a low surface energy endows the nanofiber composite with superhydrophobicity thus anti-corrosion property. The introduction of CNTs/PDMS improves both the Young's modulus, tensile strength and the elongation at break. The superhydrophobicity and conductivity can be maintained after the cyclic stretching-releasing test, displaying excellent durability. When used as a wearable strain sensor, the nanofiber composite is capable of detecting body motion and could work even under harsh conditions (moisture, acid and alkaline environment), showing promising application in wearable electronics. Keywords: Carbon nanotubes; Conductive nanofiber composite; Strain sensor; Superhydrophobic wearable electronics

1 Introduction Electrospun polymer nanofibrous membranes have received great interest from both academia and industry because of their light weight, large aspect ratio and thus high specific surface area, and unique porous structure, and have wide applications in the field of energy, catalysis, environment protection, and so on [1-3]. When the nanofibrous membrane becomes electrically conductive, its application can be further extended. In particular, the conductive nanofiber composites (CNCs) with flexibility and stretchability have wide applications in flexible electronics, wearable sensors, due to their excellent breathability and controllable sensitivity [4-6]. Generally, the CNCs are prepared by incorporation of conductive nanofillers such as carbon nanotubes (CNTs) and graphene into the polymer solution, followed by electrospinning [7-9]. The conductive nanofillers are wrapped by a layer of insulating polymer, leading to a low conductivity [10, 11]. As a result, a high filler concentration is required in the polymer solution, so that the conductive nanofillers are squeezed onto nanofiber surface to construct the continuous conductive pathway. At relatively high filler loadings, the solution viscosity could be dramatically increased, making electrospinning much harder [12-14]. Also, it is difficult for these nanofillers to achieve uniform distribution inside the nanofibers even at a low loading, because the nanofillers possessing large surface energy tend to aggregate in the polymer matrix. Consequently, the nanofiller agglomeration can limit the increase in or even deteriorate the mechanical properties (e.g., lack of flexibility) [15, 16]. To tackle this issue, the conductive nanofillers are designed to be located on the nanofiber surface. For example, the acid treated CNTs or graphene oxide could be absorbed onto the nanofiber surface by hydrogen bonding between the oxygen containing groups on CNTs or graphene oxide surface and amino groups in the polymer chain [17-21]. It was also reported that ultrasonication could be served as the driving force to achieve CNTs or graphene decoration onto nanofiber surface [18, 22-25]. Some research has been done for fabricating the conductive elastomer fiber composite as the wearable strain sensors [4,6,26]. For example, TPE-wrapped SWCNT

fibers for high-performance strain sensors were prepared by combination of the coaxial wet-spinning approach and a post-treatment process [26]. However, little attention was paid to the use of the strain sensors in harsh conditions such as humid, acid and salt environment, which is quite important for the practical applications. Therefore, it is desirable to develop the CNC based wearable strain sensors that are flexible, stretchable, and anti-corrosive. Herein, we propose a facile method to prepare superhydrophobic and stretchable CNCs. The CNTs are first anchored onto the TPU nanofiber surface, followed by the PDMS modification. CNTs anchored nanofibers possessing large aspect ratio become conductive elements and form conductive network in the composite. Moreover, the CNTs decoration significantly improves the surface roughness, and the CNCs turn to superhydrophobic after the modification of PDMS that possesses a low surface energy. The introduction of CNTs and PDMS on the nanofiber surface also enhances the mechanical properties of the TPU nanofiber membrane including the tensile strength, Young’s modulus and elongation at break. The obtained CNC is not only water but also corrosive solution repellent, and the superhydrophobicity could be maintained at different strain and after numerous cyclic strain sensing tests. The elastic CNCs could be used to detect various body motions, and the strain sensing performance is stable even after the CNCs experienced acid and alkali treatment. The superhydrophobic and stretchable CNCs have promising applications in wearable electronics.

2 Experimental 2.1 Materials and methods Thermoplastic polyurethane (TPU) pellets (code Desmopan 385S) with a melting point of 220 °C were provided by Bayer (Hong Kong). Polydimethylsiloxane (PDMS) was purchased from Dow Corning Corporation Midland-Michigan USA. Multi-walled carbon nanotubes (CNTs, diameter: 20-40 nm; length: 10-30 μm) fabricated via the chemical vapor deposition were obtained from Chengdu Organic Chemicals, Ltd., Chinese Academy of Science. Dimethylformamide (DMF), tetrahydrofuran (THF) and hexane were purchased from Sinopharm Chemical Reagent Co., Ltd. All of the

materials were used as received without any further purification. 2.2 Preparation of TPU nanofibers Firstly, TPU pellets were dispersed into a DMF/THF mixture (3/1, v/v) with the aid of magnetic stirring at 60 °C for 12 h to get a homogeneous TPU solution with the concentration of 14 wt%. Then, the prepared TPU solution was filled in a syringe and then pushed out at a rate of 1.0 mL h-1 through the metal nozzle that was connected to a high voltage power. The applied voltage was 15 kV and the distance between the metallic needle and the rotating drum was 12 cm. The electrospun nanofibers were collected on an aluminum foil attached on a rotating drum. Finally, TPU nanofiber membrane with a thickness of 0.12±0.02 mm was obtained. 2.3 Preparation of conductive TPU nanofibers CNTs were firstly ultrasonicated in the ethanol (1mg/ml) by using a high intensity ultrasonic horn (20 kHz, JY98-IID, Ningbo Scientz Biotechnology Co., Ltd) for 30 min at 0 °C to get well dispersed CNTs solution. Then the as prepared TPU fiber membrane was placed into the CNTs suspension and then experienced different time of ultrasonication. The ultrasonic frequency was 20 kHz and the power was 190 W (20% of the maximum power 950 W). During the ultrasonication, the CNTs were gradually anchored onto the surface of nanofibers. Finally, the CNTs decorated nanofiber membrane was rinsed with deionized water for several times to remove the impurities and then dried in an oven for 6 hours at 60 °C. The prepared nanofiber composites are defined as TPU/CNTs-X, where X represents the ultrasonication time (min) the nanofiber membrane experienced. 2.4 Surface modification of conductive TPU nanofiber composite 2 wt% PDMS solution was obtained by mixing the PDMS and the curing agent with the mass ratio of 10:1 into hexane, which was then subjected to magnetic stirring for 30 min. The CNTs anchored nanofiber composite was immersed into the PDMS solution for different time. Finally, the PDMS modified nanofiber composite was placed into a vacuum oven at 80 °C for 2 hours, during which the solvent evaporated and PDMS was gradually cured. The obtained superhydrophobic and conductive

nanofiber composites are denoted as TPU/CNTs-X/PDMS-Y, where Y represents the soaking time (min) of the TPU/CNTs in the PDMS solution. 2.5 Characterization Fourier transform infrared spectroscopy (FT-IR) measurement was conducted on a Cary 610/670 instrument with attenuated total reflection (ATR) mode and a range from 400 to 4000 cm−1. The thermal stability was evaluated by using thermogravimetric analysis system (TGA/ Pyris 1 TGA, PerkinElmer Co. Ltd, USA). 1.3 mg sample was heated from ambient temperature to 800 °C at an accelerated heating rate of 10 °C/min under a nitrogen flow. The Raman spectra was performed by employing a Renishaw in Via Raman confocal microscope with 532 nm laser excitation at 1 cm-1 resolution in a range of 400 to 4000 cm-1. The surface morphology of nanofiber composites was examined by a field emission scanning electron microscopy (FE-SEM, Zeiss Supra55, Germany). The specimens were also cry-fractured after frozen in the liquid nitrogen for 10 min for the SEM observation of the sectional surface. All the sample surfaces were gold sputtered before the SEM test. The nanofiber composites located on a copper grid were used for Transmission electron microscopy (TEM, Philips FEG -CM200) measurement at an acceleration voltage of 20 kV. 2.6 Contact angle measurement The surface repellence of the nanofiber composite membranes was evaluated by the contact angle (CA), which was measured by using a video optical contact angle measuring instrument (OCA40, Germany). 5 μL deionized water was dropped onto the surface of the membrane and the CA was measured after 30 s. The CA was calculated by averaging three measured values on different positions of the nanofiber membrane surface. 2.7 Mechanical and electrical property test The mechanical properties of the samples were determined using a universal testing machine (Instron Co., Ltd., U.S.A.). The dumbbell-shape sample strips were cut from the fiber mat with dimensions of 20×4×0.12 mm3 for the measurement. The mechanical properties including elastic modulus, tensile strength and elongation at break were obtained by averaging the test results of three specimens. The conductivity of the fiber

composites was tested by a Four-Probe Resistance Meter (RTS-9, Guangzhou Four Probe Technology Co., Ltd.). Note that at least five positions of the mat were measured to get the average value of the conductivity. 2.8 Strain sensing test The copper wires were connected with two ends of the nanofiber composite (5 cm (length) × 1 cm (width) × 120 µm (thickness). Note that the conductive silver paste was used to guarantee the good contact between the electrodes (the copper wires) and nanofiber composite strip. To explore the strain sensing behaviors, the nanofiber composite was subjected to cyclic loading and unloading that was conducted using the universal testing machine, and the real time resistance was in-situ recorded by a resistivity meter (Changzhou Tonghui Electronics Co., Ltd. TH2684A). The sensing responsivity was defined as the ratio of R/R0, where R0 represented the original resistance of the fiber mat and R was the transient resistance of the sample during loading-unloading cycles.

3 Results and discussion 3.1 Preparation and characterization of TPU nanofiber composite Fig. 1a, b and c show the schematic preparation procedure of TPU/CNTs/PDMS nanofiber mats. The pristine TPU nanofiber is prepared through electrospinning and the white TPU fiber mat shows good flexibility as shown in Fig. S1a. Also, the SEM image in Fig. S1b and TEM image in Fig. S1c indicate that pristine TPU nanofibers possess a rather smooth surface. Once the fibers experience ultrasonication in CNTs suspension, CNTs are anchored onto the surface of nanofiber owning to the interfacial collisions between the polymer nanofiber and the CNTs [27, 28]. SEM image in Fig. S1d indicates that CNTs are firmly anchored on the nanofiber surface, forming a hierarchical structure and thus a rough nanofiber surface, which is also verified from the TEM image shown in Fig. S1e. After a layer of PDMS possessing a very low surface energy is coated onto the fiber surface, the nanofiber composite still remains flexible and becomes superhydrophobic as displayed in Fig.1d, and water droplets keep the spherical shape on the surface of the composite fiber mat, which can be seen in Fig. S1f. The SEM

image in Fig. 1e, cross sectional SEM image in Fig. 1f and TEM image in Fig. 1g unveil that the PDMS, which could be regarded as a kind of “glue”, sticks the CNTs together, and thus promotes the interfacial interaction between CNTs and the polymer. The uniform decoration of PDMS onto the TPU/CNTs nanofiber composite is also corroborated by the elemental mapping, as exhibited in Fig. h-k. Clearly, the carbon, nitrogen, oxygen and silicon are uniformly distributed on the whole nanofiber surface.

Fig. 1 (a-c) Schematic illustration for the preparation of TPU/CNTs/PDMS nanofiber composites. (d) Photograph of the TPU/CNTs/PDMS nanofiber composite with good flexibility (Inset: one water droplet on the material surface). (e) The SEM image of the nanofiber composite surface. (f) cross sectional SEM image and (g) TEM image of the TPU/CNTs/PDMS nanofiber composite. (h-k) Scanning mapping images of green area in (e) for C, N, O, Si, respectively. The functional groups of the composite and the interactions between different chemical groups are investigated by the FT-IR spectrum (Fig. 2a). For the pure TPU nanofiber, the peaks at 3332 cm-1 and 1078 cm-1 are attributed to the N–H and C–O–C stretching bands of urethanes. Two additional absorption peaks at 1730 cm-1 and 1595 cm-1 are assigned to −H−N−COO− groups. Also, the peak at 2956 cm-1 represents the alkene –CH stretching vibrations. After CNTs are anchored into the fiber surface, the peaks at 3332 cm-1, 2956 cm-1, 1730 cm-1, 1078 cm-1 slightly downshift to 3330 cm-1, 2949 cm-1, 1726 cm-1, 1072 cm-1 respectively, showing certain interaction between TPU

and CNTs [27]. When the conductive nanofiber composite is modified with PDMS, some functional groups appear in the spectrum. The peaks at 1412 cm-1 and 1254 cm-1 are assigned to the asymmetric and symmetric deformation of –CH3, respectively. The peak at 1016 cm-1 is associated with the asymmetric stretching band of Si–O–Si. In addition, the peaks at 845 cm-1 and 795 cm-1 belong to Si–C and Si–(CH3)2, respectively [29]. Raman spectroscopy is also a good technique to detect the carbon structure and the possible interactions between carbon nanofillers and the polymer matrix. Fig. 2b shows the Raman spectrum of CNTs and the nanofiber composite. For CNTs, three characteristic peaks at around 1354.4 cm-1, 1579.9 cm-1, 2699.4 cm-1 are typical D, G and 2D band, reflecting its graphite structure. Normally, the D and 2D bands are related to disordered graphite structure and the second-order overtone of a different vibration mode in graphene-like structure, respectively, while the G band represents the stretching mode of the C–C bonds [30]. Generally, the peak intensity ratio of D to G, namely the ID/IG indicates the defect density in CNTs. It is calculated that the ID/IG is around 1.04, 0.81 and 0.99 for CNTs, TPU/CNTs and TPU/CNTs/PDMS respectively. These results indicate that sp3 defects might exist in the sp2 carbon network. It is found that the D and 2D bands downshift to 1347.5 cm-1 and 2691.2 cm-1 for the TPU/CNTs, also 1347.1 cm-1 and 2689.8 cm-1 for TPU/CNTs/PDMS, respectively. On the contrary, the G bands upshift from 1579.9 cm-1 to 1595.2 cm-1 and 1586.7 cm-1 for TPU/CNTs and TPU/CNTs/PDMS respectively. These peak shifts may be related with the interfacial stress transfer between CNTs and polymer nanofibers [31]. It is believed that the dispersion state of CNTs in the polymer matrix and the chemical environment surrounding CNTs are the potential factors to intrigue the shift of D and G band. Poor dispersion of CNTs would induce a subtle shift, while larger shift occurs for well dispersed CNTs [32]. In this work, CNTs are uniformly anchored into the nanofiber surface under the assistance of ultrasonication. The thermal stability of the nanofiber composites is investigated by thermogravimetric analysis (TGA). The TGA curves and corresponding differential thermogravimetric curves of TPU fiber and its composite are shown in Fig. 2c and Fig.

2d. The weight percentages of the residue for the pristine TPU fibers, TPU/CNTs-40 and TPU/CNTs-40/PDMS-240 after thermal degradation at 800 °C are about 4.05%, 6.40% and 6.68%, respectively. Based on the TGA results, the weight percentage of the CNTs in the TPU/CNTs-40 is calculated to be about 2.38%. It is also obvious that with the introduction of CNTs and PDMS, the TGA curves shift towards the high temperature zone. The temperature of maximum decomposition is 394.1 °C for TPU/CNTs-40 and 397.1 °C for TPU/CNTs-40/PDMS-240, which is around 20 °C higher than that of pure TPU. The good dispersion of the CNTs on the nanofiber surface and the strong interaction between the CNTs (or CNTs/PDMS) and the polymer nanofibers may be responsible for the improved thermal stability.

Fig. 2 (a) FT-IR spectrum of TPU nanofiber mat and its composite. (b)Raman spectra of the CNTs, TPU/CNTs-40 and TPU/CNTs-40/PDMS-240. (c) TGA and (d) differential thermogravimetric curves for TPU, TPU/CNTs-40 and TPU/CNTs40/PDMS-240. 3.2 Electrical conductivity, superhydrophobicity and mechanical properties As mentioned, ultrasonication is the driving force for the CNTs’ decoration onto the nanofibers. CNTs anchored nanofibers with a large aspect ratio becomes conductive

element, and form conductive network in the composite. Naturally, the extension of the ultrasonication time would lead to the increase of the CNTs concentration and hence the electrical conductivity of the nanofiber composite. As shown in Fig. S2a, the electrical conductivity significantly increases from 10-10 S/m to around 10-1 S/m at an ultrasonication time of 3 min, indicating the conducive network has been established. With increase of the ultrasonication time to 10 min, the conductivity slightly rises to 10 S/m, and then it fluctuates around this value with further increasing the ultrasonication time, which indicates the nanofiber surface has already been saturated by CNTs at the ultrasonication time of 10 min, losing the viscoelasticity. The hydrophobic CNTs anchored onto the nanofibers enhance the surface roughness and hence the hydrophobicity of the membrane. Base on the Cassie-Baxter theory, water droplets cannot enter inside the surface cavity but they are instead located on the asperities, i.e., the droplets are suspended on the air pocket of the rough material surface [33]. In Cassie-Baxter state, the apparent liquid contact angle θ* can be expressed using equation (1)

cos  *  f1 cos   f 2

(1)

where f1 and f2 refer to the fraction of solid protuberance and air that contact with the water droplet, respectively. Since f1 + f2 = 1, a larger f2 (more air pockets) corresponds to a higher θ*(CAs of water in air is regarded as 180). In this study, the CA for the TPU nanofiber membrane is around 125º, and it increases to 130º for both TPU/CNTs-3 and TPU/CNTs-10 because of the slight enhancement of the nanofiber surface roughness (see Fig. S2b). For TPU/CNTs-30, many CNTs are entangled together and form a fluffy network (as shown in Fig. S2c). Where a large amount of air pocket is trapped, corresponding to a high surface roughness and thus a large CA of 145º. After the ultrasonication time is longer than 30 min, excess CNTs anchored on the nanofiber surface reduce the surface roughness (see the surface morphology of TPU/CNTs-50 in Fig. S2d) and hence the f2. As a result, the CAs of TPU/CNTs-40 and TPU/CNTs-50 have a sustained decrease. As known, PDMS can be used as the matrix or the encapsulation materials for

preparation of flexible strain sensors [34-36]. In our study, the introduction of PDMS could decrease the surface energy of TPU/CNTs nanofiber composite, leading to the formation of superhydrophobic surface. On the other hand, the insulating PDMS wrapping the CNTs would hinder the electron transportation and hence increase the resistivity. Fig. 3a shows the influence of the dipping time in PDMS on the CAs and electrical conductivity of the nanofiber composite. It is found the CA greatly increases from 132º for the TPU/CNTs-50 to 151º for the TPU/CNTs-50-PDMS-360, while the conductivity slightly drops from 27.1 S/m to 19.7 S/m. With prolonging the dipping time from 60 to 360 min, the CAs and conductivity level off. Compared with CNTs distribution on the surface of TPU/CNTs-50 (Fig. 3b), the CNTs are buried into the PDMS leaving one end exposed outside (Fig. 3c for the TPU/CNTs-50/PDMS-180 and Fig. 3d for TPU/CNTs-50/PDMS-360).

Fig.3 (a) The electrical conductivity and water CAs for TPU/CNTs-50-PDMS-Y nanofiber composite. SEM images of the nanofiber surface after PDMS treatment for different time (b) 0 min, (c)180 min and (d) 360 min. As mentioned, it’s desirable that the strain sensor can work in corrosive environment like humid, acidic and alkaline conditions. The superhydrophobic surface possesses excellent water repellency, which can effectively prevent the moisture or even the corrosive solution diffusion inside the porous nanofiber composite. The inset in Fig. 4a shows the aqueous droplets with different pH (water, acid, base and salt) on

the surface of TPU/CNTs-50/PDMS-360. It’s obvious that these droplets show spherical shapes and the corresponding CAs are summarized in the histogram. All CAs are around 150º, displaying superhydrophobicity. The superhydrophobicity can ensure the stability of the electrical conductivity of the nanofiber composite in humid or even corrosive environment. As shown in Fig. S3, the resistance is almost unchanged during the test under a humidity of 85%, i.e., the resistance variation (R/R0), maintains a stable value of around 1.0. The nanofiber composite also exhibits excellent anti-corrosion property, which is shown in Fig. 4b. The nanofibrous membrane floats on the surface of the acid solution, and both CAs and conductivity are stable and fluctuate around their initial values during the 6 h immersion in the solution. As known, the strain sensors are generally used in cyclic mechanical deformation. Therefore, the stability of superhydrophobicity and conductivity for the nanofiber composite based wearable strain sensor is quite important. Fig. 4c shows the photographs displaying water droplets on the surface of the nanofiber composite under different strain. It’s evident that the water droplets stand on the material surface with a spherical shape even at ε=100%, exhibiting excellent water repellence. Generally, the strain would, to a large degree, damage the hierarchical structure of material surface, thereby decreasing the surface roughness and hence the CA [37, 38]. Surprisingly, the CA of the TPU/CNTs50/PDMS-360 first fluctuates at around 150º with the strain from 0 to 50%, and then slightly increases with strain, which is shown in Fig. 4d. The strain induced increase of the CA is related with morphology evolution under the tensile strain, which will be discussed in the following section. Fig. 4e shows the CAs and conductivity of the nanofiber composite mat during the cyclic stretching-releasing test with the strain of 50%. The conductivity decreases slightly at first ten cycles and is then almost unchanged during the next 90 cycles; the CAs fluctuate at 150º in the whole cyclic loading-unloading test. The surface morphology of the composite after 100 cyclic stretching-releasing tests can be found from the SEM images shown in Fig. S4a and Fig. S4b, and no visible damage or cracks are present on the composite surface, indicating the excellent recoverability of the conductive network and hence durability of the nanofiber composite strain sensors. Abrasion test is also conducted to assess the

wear-resist performance of the strain sensor. As shown in the insets of Fig. S4c, the composite strip with a weight of 20 g on its surface lies on a sand paper, and then the strip is pulled forward for 5 cm as one abrasion test. As displayed in Fig. S4c, the conductivity and CA keep almost unchanged even after 100 times abrasion test, and no obvious change for the microstructure is observed after multiple abrasion tests (Fig. S4d). The sensing behavior of the sample after the abrasion test is also discussed in the following section. The stable superhydrophobicity and electrical performance guarantee the reliability of the nanofiber composite as the wearable strain sensors in the practical applications. Also, the introduction of CNTs and PDMS could improve the mechanical properties of the nanofiber membrane, which is demonstrated in the stressstrain curves shown in Fig. 4f. The mechanical properties including the elongation at break, tensile strength and Yong’s modulus of pure TPU, TPU/CNTs and TPU/CNTs/PDMS are summarized in the table S1. The TPU/CNTs-40/PDMS-240 has a much higher Yong’s modulus, tensile strength and the elongation at break of 2.62 ±0.13 MPa, 9.41±0.94 MPa and 544.9±16.1% than TPU nanofibrous membrane (1.58±0.14 MPa, 5.72±0.54 MPa and 514.4±13.5%, respectively). The mechanical properties of the samples immersed in an acid solution with pH of 1 for 5 h (denoted as TPU/CNTs-40/PDMS-240 (acid)) are also tested. Note that the nanofiber composites can only float on the acid solution surface because of its superhydrophobicity. The Yong’s modulus, tensile strength and the elongation at break are 2.81±0.03 MPa, 10.93±1.27 MPa and 536.3±28.8%, respectively, which is almost the same with those of the pristine TPU/CNTs-40/PDMS-240 without the acid treatment. The uniform distribution of CNTs and good interfacial interaction between CNTs and TPU nanofibers promote the effective stress transfer, contributing to the improvement of both Young’s modulus and tensile strength [39-41]. After a thin layer of PDMS was coated onto the CNTs and nanofibers surface, the elastic PDMS layer can improve the component interaction through the interfacial binding. More importantly, the significantly increased interfacial area between CNTs and the PDMS effectively transfers the stress applied on the nanofiber composite, making contribution to the enhancement of mechanical properties of composites. Also, stretching is beneficial to

the nanofiber alignment, which increase the interfacial area between the nanofibers with rough surface hence leads to the improvement of the mechanical performance [42, 43].

Fig. 4 (a) CAs of the aqueous droplets including acidic, alkaline and salt solutions on the surface of TPU/CNTs-50/PDMS-360 (inset is the photograph showing the solution droplets on the composite surface). (b) The variation of the conductivity and CA of TPU/CNTs-50/PDMS-360 as a function of immersion time in an acidic solution (pH=1) (Inset is the photograph showing the nanofiber composite membrane on the acid solution surface because of its superhydrophobicity). (c) photographs of TPU/CNTs50/PDMS-360 membrane under different strain with water droplets on the surface. (d) CA variation for TPU/CNTs-50/PDMS-360 under different strain. (e) The conductivity and CAs of TPU/CNTs-50/PDMS-360 experiencing cyclic stretching-releasing under the strain of 50%. (f) Representative stress-strain curves of pure TPU mat, TPU/CNTs40, TPU/CNTs-40/PDMS-240 and TPU/CNTs-40/PDMS-240 after immersion in an acid solution with pH of 1 for 5h. 3.3 Strain sensing performance Fig. 5a-c demonstrate the relative resistance change (R/R0) of TPU/CNTs-50/PDMS360 under cyclic stretching-releasing at ε=30%, 50%, 100%, respectively. Fig. 5a and 5c are the sensing signals for 20th-30th cycles extracted from the cyclic stretchingreleasing curves as shown in Fig. S5a and S5b, respectively. The sensing intensity decreases after the first cyclic test but gradually become stable. A permanent damage to the conductive network caused by the unrecoverable macromolecular segment movement is regarded as the main cause of the sensing phenomenon [44]. The R/R0 and intensity (the maximum value of R/R0) are almost identical for each cycle, demonstrating good stretchability and resilience. The sensing intensity for TPU/CNTs-

50/PDMS-360 is around 10.3% under 30% strain, and it increases to about 17% and 19.5% with increasing the strain to 50% and 100%, respectively. Interestingly, dual peaks (marked with red circles) are present under larger strain in the sensing curve. Generally, the shoulder peaks are generated by the competition between the destruction and recovery of the conductive network during releasing. The sensing curves for relatively small strains (30% and 50%) show no obvious signal of shoulder peak, while that is obvious for a larger strain (100%). It’s believed that small strain causes weak damage to the conductive pathway, making the recovery easier during releasing process; however, larger strain severally damages the conductive network, resulting in drastic competition between the destruction and reconstruction of the conductive pathway during the recovery process. It is reasonable that a larger strain intrigues greater intensity because of more severe damage to the conductive network [45]. For comparison, the sensing behavior of the nanofiber composite after 100 times abrasion is shown in Fig. S6. Fig. S6a-S6c are the cyclic sensing curves of TPU/CNTs50/PDMS-360 after the abrasion test at ε=30%, 50%, 100%, respectively. Fig. S6d-S6f show the ten cyclic tests extracted from the red rectangle regions in (a), (b) and (c), respectively, and the sensing intensity is around 10.0%, 15.3% and 20.5% under the strain of 30%, 50% and 100% respectively. The curves are almost identical with those of the pristine sample for the sensing test (Fig. 5a-5c). The results demonstrate outstanding stability and reliability of the sensing behavior of the sensor after experiencing multiple stretching and abrasion. To examine the reusability of the strain sensor, the nanofiber composite (TPU/CNTs-50/PDMS-360) undergoes cyclic tensile tests under a strain of 50% (Fig. 5d) for over 700 cycles and still outputs stable sensing signals, demonstrating the excellent durability of the strain sensor in practical application. The curve in Fig. 5b is the ten cycles extracted from the rectangle region of the cyclic test, and R/R0 shows good recoverability originating from the good resilience of the CNTs/PDMS decorated elastomer nanofiber. Figure 5e shows the change of relative resistance (R/R0) of the nanofiber composite based sensor with strains from 0% to 100%. Obviously, the R/R0 increases with the loading of strain and decreases with the unloading of strain. The sensitivity is usually evaluated by the gauge

factor (GF) (GF =(∆R/R0)/ε), corresponding to the slopes of the fitting curves in the linear regions in Fig. 5f. It’s clear that the sensing curve is divided to three linear regions with three different slopes. GFs are accordingly calculated to be 19.4%, 7.14% and 33.9% for the three regions, respectively. To evaluate the influence of the strain rates on the sensing response of the strain sensor under 50% strain, the stretching rates are set as 20, 50 and 100 mm/min, and the sensing signals are plotted in Fig. S7a. Notably, the adjustment of applied strain rates almost shows no impact on the sensing signals. As mentioned, the CNTs content controlled by the ultrasonication time can influence the establishment of conductive network and thus the strain sensing performance. Fig. S7b shows the resistance response of R/R0 at 100% strain for nanofiber composites prepared under different ultrasonication time with a strain rate of 20 mm/min. It is found that the TPU/CNTs-1/PDMS-360 with a low conductivity of 0.164 S/m shows the highest response intensity of 33.8% among the three nanofiber composites, because its vulnerable conductive network is easily damaged when subjected to the strain. By the contrary, the nanofiber composite with a high conductivity (e.g., TPU/CNTs30/PDMS-360) possesses a perfect conductive network that can, to a certain degree, resist the strain induced damage of the conductive pathway, corresponding to a relatively low sensing intensity.

Fig. 5 Strain sensing behavior of the superhydrophobic and conductive nanofiber composite (TPU/CNTs-50/PDMS-360) at different strain, (a) ε=30%, (b) ε=50%, and (c) ε=100%. (d) The cyclic strain sensing performance of TPU/CNTs-50/PDMS-6h

under 50% strain for over 700 cycles. Fig. 5b is extracted from the red rectangle area of the cyclic test. Note that the strain rate is fixed at 20 mm/min. (e) The sensing performance of the strain sensor with the incremental strain under loading-unloading tests. (f) The relative resistance change (ΔR/R0) of the sensor as a function of strain (blue line and red lines are the experimental curve and linear fitting curves). 3.4 Body motion detection As mentioned, the superhydrophobic and conductive TPU/CNTs/PDMS nanofiber composite with high flexibility, stretchability and durability has potential applications in wearable strain sensors that can be used for real-time healthcare monitoring. As a proof of concept, the nanofiber composite is attached onto the human joints to collect the signal to detect the body motion. As demonstrated in Fig. 6a, when the tester folds and unfolds his finger repeatedly (see the photograph in the insets), the resistance changed simultaneously. Even after a dozen times of folding and relaxing, the sensing signals keep almost the same, displaying good stability and repeatability. Moreover, the sensor can also work in a harsh environment, and the sensing behavior keeps unchanged when the water and acid droplets are dripped onto the material surface (Fig. S8), due to the excellent water proof and corrosion resistance of the material surface. It is desired that the nanofiber composite has a reliable sensing performance even in a harsh condition. To examine its anti-corrosion property, the nanofiber composite membrane is immersed in the acidic and alkaline solution for 5 h as shown in the insets, followed by the sensing test. Interestingly, the resistance response is almost the same with that of the pristine nanofiber composite. The excellent anti-corrosion and strain sensing performance originate from the strong water repellence, i.e., the superhydrophobicity. Despite its porous structure, the superhydrophobic membrane can only float on the solution surface (see the photograph in the inset in Fig. 6a), making it impossible for the acid or alkaline solution to diffuse into the nanofiber composite. Apart from the finger movement, the integrated composite device could also be used for detection of elbow and knee motion (Fig. 6b and 6c). For example, different responsivity could be achieved when the tester bends his elbow with angles of 45º, 90º, and 135º, respectively (Fig. 6b). The response intensity is enhanced with the increase of the bending angle of the elbow, because a larger bending angle corresponds to a higher strain, which would cause greater damage to the conductive network, giving rise to a higher R/R0. The

wearable strain sensor can also be attached onto the neck of the volunteer to monitor his sitting position as shown in Fig. 6d, and the signal of resistance variation is collected when the volunteer nods periodically.

Fig.6 Nanofiber composite as a wearable strain sensor for detection of body joint motion. (a) Sensing behavior of finger joint bending before and after the immersion of the sample in acid/base solution. (b) Sensing signals of the elbow motion with different bending angles (insets are the pictures of strain sensor on the elbow under different bending angles). (c) Resistance responsive curves of nanofiber composite on the knee for monitoring squating (insets are the pictures of sample attached on the knee under different state). (d) Responsive curves of the strain sensor on the neck under cyclic nodding (insets are the photographs of the sensor attached on the neck of the volunteer). 3.5 Strain sensing mechanism The surface morphology of conductive TPU nanofiber composite under different strain is observed to reveal the evolution of the conductive network and thus the strain sensing mechanism, as shown in Fig.7. Before stretching, the CNTs are tightly entangled with each other on fiber surface, forming a perfect conductive network just as shown in Fig. 7a. However, the network begins to be damaged under the stretching. At a relatively small strain, the separation between the conductive nanofibers causes a slightly increased resistance. This is in accordance with the result of body motion test.

When the composite is further stretched to 30%, many ultrathin crazing and wrinkles are present (see the red circle in Fig. 7b). The orientation of fibers is observed along the stretching direction with the strain of 50% as shown in Fig. 7c. The nanofiber alignment becomes more evident with further increase of the strain to 100%, which can be observed from the blue arrows in Fig. 7d. It is also worth noting that many cracks are generated at a relatively high strain (the red circles in Fig. 7c and 7d). The fiber alignment reduces the contact point among the conductive nanofibers while the cracks cut the conductive pathway, leading to the large increase of resistance especially at high strains. Although the cracks on the nanofiber surface can undermine the conductive network, they can, to some extent, increase nanofiber surface roughness, resulting in the strain induced enhancement of superhydrophobicity as exhibited in Fig. 4d. On the other hand, when the strain is released from 100% to 50%, the gap between the cracks becomes narrower and the fiber orientation declines as shown in Fig. 7e. As the strain is further released to 30%, the cracks on the surface become vague and the fibers tend to be distributed randomly (Fig. 7f). After the nanofiber composite comes back to its original state, the conductive network and hence the resistance can return to its initial values, displaying outstanding resilience and reproducibility. The surface morphology evolution of the conductive nanofiber composite without PDMS modification under stretching and releasing is also observed in Fig. S9. Compared with the morphology of PDMS modified composite, no obvious cracks and damage to the conductive network on the fiber surface are observed, corresponding to a smaller response intensity when served as the strain sensor (see Fig. S10).

Fig.7 SEM images of TPU/CNTs-50/PDMS-6h under different stretching and releasing state. (a) ε=0. (b) ε=30%. (c) ε=50%. (d) ε=100%. (e) ε=50%. (f) ε=30%.

4. Conclusion In this work, a stretchable and superhydrophobic nanofiber composite is prepared by decorating the CNTs onto the TPU nanofiber under the assistance of ultrasonication, followed by the PDMS modification. The electrical conductivity could be controlled by controlling the ultrasonication time. The well dispersed CNTs anchored on the TPU nanofiber surface entangled with each other, forming the conductive pathway. The introduction of CNTs and PDMS layer could not only promote the water repellence but also improve the mechanical properties of the TPU nanofibrous mats including the Young's modulus, tensile strength and elongation at break. The superhydrophobic nanofiber composite could repel the corrosive alkaline and salt solution, and the membrane remains superhydrophobic even at a large strain. When used as a strain sensor, the nanofiber composite exhibits good durability (showing the same signal even after immersion in acidic/alkaline solution for 5 h) and repeatability (700 cycles of stretching-releasing test). The sensor is also applicable for the full-range human motion detection (like finger bending, elbow movement, squating and nodding), displaying a promising application in wearable electronics. The damage and recovery of the conductive network inside the nanofiber composite during stretching and releasing is responsible for the strain sensing performance.

Conflicts of interest The authors declare no conflict of interest.

Acknowledgements This work was financially supported by Natural Science Foundation of China (No. 51873178, 51503179, No.21673203), the Opening Project of State Key Laboratory of Polymer Materials Engineering (Sichuan University) (No. sklpme2018-4-31), Qing Lan Project of Jiangsu province, the China Postdoctoral Science Foundation (No. 2016M600446), the Jiangsu Province Postdoctoral Science Foundation (No. 1601024A), the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References [1] Y.Z. Zhang, Z. Zhang, S. Liu, G.R. Li, X.P. Gao, Free-Standing Porous Carbon Nanofiber/Carbon Nanotube Film as Sulfur Immobilizer with High Areal Capacity for Lithium-Sulfur Battery, ACS. Appl. Mater. Interfaces 10 (2018) 8749-8757. [2] X.-Q. Wu, J.-S. Shen, F. Zhao, Z.-D. Shao, L.-B. Zhong, Y.-M. Zheng, Flexible electrospun MWCNTs/Ag3PO4/PAN ternary composite fiber membranes with enhanced photocatalytic activity and stability under visible-light irradiation, J. Mater. Sci. 53 (2018) 10147-10159. [3] J. Gu, H. Gu, Q. Zhang, Y. Zhao, N. Li, J. Xiong, Sandwich-structured composite fibrous membranes with tunable porous structure for waterproof, breathable, and oilwater separation applications, J. Colloid Interface Sci. 514 (2018) 386-395. [4] N. Wang, Z.Y. Xu, P.F. Zhan, K. Dai, G.Q. Zheng, C.T. Liu, et al., A tunable strain sensor based on a carbon nanotubes/electrospun polyamide 6 conductive nanofibrous network embedded into poly(vinyl alcohol) with self-diagnosis capabilities, J. Mater. Chem. C 5 (2017) 4408-4418. [5] T. Wang, D.F. Song, H. Zhao, J.Y. Chen, C.H. Zhao, L.L. Chen, et al., Facilitated transport channels in carbon nanotube/carbon nanofiber hierarchical composites decorated with manganese dioxide for flexible supercapacitors, J. Power Sources 274

(2015) 709-717. [6] H. Yang, X.F. Yao, Z. Zheng, L.H. Gong, L. Yuan, Y.N. Yuan, et al., Highly sensitive and stretchable graphene-silicone rubber composites for strain sensing, Compos. Sci. Technol. 167 (2018) 371-378. [7] B.C. Weng, F.H. Xu, A. Salinas, K. Lozano, Mass production of carbon nanotube reinforced poly(methyl methacrylate) nonwoven nanofiber mats, Carbon 75 (2014) 217-226. [8] Q. Pan, N.J. Tong, N.F. He, Y.X. Liu, E. Shim, B. Pourdeyhimi, et al., Electrospun Mat of Poly(vinyl alcohol)/Graphene Oxide for Superior Electrolyte Performance, ACS. Appl. Mater. Interfaces 10 (2018) 7927-7934. [9] J.Y. Wu, A.K. An, J.X. Guo, E.-J. Lee, M.U. Farid, S. Jeong, CNTs reinforced super-hydrophobic-oleophilic electrospun polystyrene oil sorbent for enhanced sorption capacity and reusability, Chem. Eng. J. 314 (2017) 526-536. [10] S. Talaeemashhadi, M. Sansotera, C. Gambarotti, A. Famulari, C.L. Bianchi, P. Antonio Guarda, et al., Functionalization of multi-walled carbon nanotubes with perfluoropolyether peroxide to produce superhydrophobic properties, Carbon 59 (2013) 150-159. [11] F. Rivadulla, C. Mateo-Mateo, M. A. Correa-Duarte, Layer-by-Layer Polymer Coating of Carbon Nanotubes: Tuning of Electrical Conductivity in Random Networks, J. Am. Chem. Soc. 132 (2010) 3751-3755. [12] X.Y. Zhao, J.J. Luo, C.J. Fang, J. Xiong, Investigation of polylactide/poly(εcaprolactone)/multi-walled carbon nanotubes electrospun nanofibers with surface texture, RSC ADV. 5 (2015) 99179-99187. [13] L.Y. Mei, P. Song, Y.Q. Liu, Magnetic-field-assisted electrospinning highly aligned composite nanofibers containing well-aligned multiwalled carbon nanotubes, J. Appl. Polym. Sci. 132 (2015) 41995. [14] Y.H. Song, Z.Y. Sun, L. Xu, Z.B. Shao, Preparation and Characterization of Highly Aligned Carbon Nanotubes/Polyacrylonitrile Composite Nanofibers, Polymers 9 (2017) 1-13. [15] K. Nasouri, A.M. Shoushtari, A. Kaflou, H. Bahrambeygi, A. Rabbi, Single-wall

carbon nanotubes dispersion behavior and its effects on the morphological and mechanical properties of the electrospun nanofibers, Polym. Compos. 33 (2012) 19511959. [16] D.F. Wu, T.J. Shi, T. Yang, Y.R. Sun, L.F. Zhai, W.D. Zhou, et al., Electrospinning of poly(trimethylene terephthalate)/carbon nanotube composites, Eur. Polym. J. 47 (2011) 284-293. [17] L.A. Mercante, A. Pavinatto, L.E. Iwaki, V.P. Scagion, V. Zucolotto, O.N. Jr., et al., Electrospun polyamide 6/poly(allylamine hydrochloride) nanofibers functionalized with carbon nanotubes for electrochemical detection of dopamine, ACS Appl. Mater. Interfaces 7 (2015) 4784-4790. [18] X.Y. Guan, G.Q. Zheng, K. Dai, C.T. Liu, X.R. Yan, C.Y. Shen, et al., Carbon Nanotubes-Adsorbed

Electrospun

PA66

Nanofiber

Bundles

with

Improved

Conductivity and Robust Flexibility, ACS Appl. Mater. Interfaces 8 (2016) 1415014159. [19] K. Saetia, J.M. Schnorr, M.M. Mannarino, S.Y. Kim, G.C. Rutledge, T.M. Swager, et al., Spray-Layer-by-Layer Carbon Nanotube/Electrospun Fiber Electrodes for Flexible Chemiresistive Sensor Applications, Adv. Funct. Mater. 24 (2014) 492-502. [20] Y.P. Huang, L.S. Zhang, H.Y. Lu, F.L. Lai, Y.-E. Miao, T.X. Liu, A highly flexible and conductive graphene-wrapped carbon nanofiber membrane for high-performance electrocatalytic applications, Inorg. Chem. Front. 3 (2016) 969-976. [21] A.F. de Faria, F. Perreault, E. Shaulsky, L.H. Arias Chavez, M. Elimelech, Antimicrobial Electrospun Biopolymer Nanofiber Mats Functionalized with Graphene Oxide-Silver Nanocomposites, ACS Appl. Mater. Interfaces 7 (2015) 12751-12759. [22] J.F. Gao, M.J. Hu, Y.C. Dong, R.K.Y. Li, Graphite-nanoplatelet-decorated polymer nanofiber with improved thermal, electrical, and mechanical properties, ACS Appl. Mater. Interfaces 5 (2013) 7758-7764. [23] Y.L. Wang, J. Hao, Z.Q. Huang, G.Q. Zheng, K. Dai, C.T. Liu, et al., Flexible electrically resistive-type strain sensors based on reduced graphene oxide-decorated electrospun polymer fibrous mats for human motion monitoring, Carbon 126 (2018) 360-371.

[24] H.C. Shi, H.Y. Liu, S.F. Luan, D. Shi, S.J. Yan, C.M. Liu, et al., Effect of polyethylene glycol on the antibacterial properties of polyurethane/carbon nanotube electrospun nanofibers, RSC ADV. 6 (2016) 19238-19244. [25] J.F. Gao, M.J. Hu, R.K.Y. Li, Ultrasonication induced adsorption of carbon nanotubes onto electrospun nanofibers with improved thermal and electrical performances, J. Mater. Chem. 22 (2012) 10867-10872. [26] J. Zhou, X.Z. Xu, Y.Y. Xin, G. Lubineau, Coaxial Thermoplastic Elastomer‐ Wrapped Carbon Nanotube Fibers for Deformable and Wearable Strain Sensors, Adv. Funct. Mater. 28 (2018) 1705591. [27] J.F. Gao, H. Wang, X.W. Huang, M.J. Hu, H.G. Xue, R.K.Y. Li, Electrically conductive polymer nanofiber composite with an ultralow percolation threshold for chemical vapour sensing, Compos. Sci. Technol. 161 (2018) 135-142. [28] J.F. Gao, H. Wang, X.W Huang, M.J. Hu, H.G. Xue, R.K.Y. Li, A superhydrophobic and electrically conductive nanofibrous membrane for a chemical vapor sensor, J. Mater. Chem. A 6 (2018) 10036-10047. [29] D. Bodas, J.Y. Rauch, C. Khan-Malek, Surface modification and aging studies of addition-curing silicone rubbers by oxygen plasma, Eur. Polym. J. 44 (2008) 2130-2139. [30] V. Datsyuk, M. Kalyva, K. Papagelis, J. Parthenios, D. Tasis, A. Siokou, et al., Chemical oxidation of multiwalled carbon nanotubes, Carbon 46 (2008) 833-840. [31] Y.Y. Huang, S.V. Ahir, E.M. Terentjev, Dispersion rheology of carbon nanotubes in a polymer matrix, Phys. Rev. B 73 (2006) 125422. [32] H.S. Xia, M. Song, Preparation and characterization of polyurethane–carbon nanotube composites, Soft Matter 1 (2005) 386-394. [33] X.L. Wang, Y.M. Pan, C.Y. Shen, C.T. Liu, X.H. Liu, Facile Thermally Impacted Water‐ Induced Phase Separation Approach for the Fabrication of Skin‐ Free Thermoplastic Polyurethane Foam and Its Recyclable Counterpart for Oil–Water Separation, Macromol. Rapid Comm. 39 (2018) 1800635. [34] J. Zhou, H. Yu, X.Z. Xu, F. Han, G. Lubineau, Ultrasensitive, stretchable strain sensors based on fragmented carbon nanotube papers, ACS Appl. Mater. Interfaces 9 (2017) 4835-4842.

[35] Y.J. Zheng, Y.L. Li, Z.Y. Li, Y.L. Wang, K. Dai, G.Q. Zheng, et al., The effect of filler dimensionality on the electromechanical performance of polydimethylsiloxane based conductive nanocomposites for flexible strain sensors, Compos. Sci. Technol., 139 (2017) 64-73. [36] Y. Wang, L. Wang, T.T. Yang, X. Li, X.B. Zang, M. Zhu, et al. Wearable and highly sensitive graphene strain sensors for human motion monitoring, Adv. Funct. Mater., 24 (2014) 4666-4670. [37] X.J. Su, H.Q. Li, X.J. Lai, Z.H. Chen, X.R. Zeng, Highly Stretchable and Conductive Superhydrophobic Coating for Flexible Electronics, ACS Appl. Mater. Interfaces 10 (2018) 10587-10597. [38] L.H. Li, Y.Y. Bai, L.L. Li, S.Q. Wang, T. Zhang, A Superhydrophobic Smart Coating for Flexible and Wearable Sensing Electronics, Adv. Mater. 29 (2017) 1702517. [39] N. George, P.K. Bipinbal, B. Bhadran, A. Mathiazhagan, R. Joseph, Segregated network formation of multiwalled carbon nanotubes in natural rubber through surfactant assisted latex compounding: A novel technique for multifunctional properties, Polymer 112 (2017) 264-277. [40] X.R. Wang, Y. Si, X.F. Wang, J.M. Yang, B. Ding, L. Chen, et al., Tuning hierarchically aligned structures for high-strength PMIA-MWCNT hybrid nanofibers, Nanoscale 5 (2013) 886-889. [41] A. Baji, Y.W. Mai, M. Abtahi, S.-C. Wong, Y. Liu, Q. Li, Microstructure development in electrospun carbon nanotube reinforced polyvinylidene fluoride fibers and its influence on tensile strength and dielectric permittivity, Compos. Sci. Technol. 88 (2013) 1-8. [42] Y.L. Zhao, S.G. Wang, Q.S. Guo, M.W. Shen, X.Y. Shi, Hemocompatibility of electrospun halloysite

nanotube and carbon nanotube-doped composite poly (lactic-

co-glycolic acid) nanofibers, J. Appl. Polym. Sci. 127 (2013) 4825-4832. [43] J.W. Zha, Y.Gao, D.L. Zhang, Y. Wen, R.K.Y. Li, C. Y. Shi, et al., Flexible electrospun polyvinylidene fluoride nanofibrous composites with high electrical conductivity and good mechanical properties by employing ultrasonication induced

dispersion of multi-walled carbon nanotubes, Compos. Sci. Technol. 128 (2016) 201206. [44] J.H. Kong, N.S. Jang, S.H. Kim, J.M. Kim, Simple and rapid micropatterning of conductive carbon composites and its application to elastic strain sensors, Carbon 77 (2014) 199-207. [45] H. Liu, J.C. Gao, W.J. Huang, K. Dai, G.Q. Zheng, C.T. Liu, et al., Electrically conductive strain sensing polyurethane nanocomposites with synergistic carbon nanotubes and graphene bifillers, Nanoscale 8 (2016) 12977-12989. Highlights: 

A flexible, superhydrophobic and conductive nanofiber composite was prepared.



The nanofiber composite exhibited excellent durability and corrosion resistance.



The nanofiber composite could be used for wearable strain sensors.



The sensing performance could be maintained after cyclic abrasion and stretching.