Introduction

Flexible electrodes used for supercapacitors have received great interest with the development of portable and wearable electronic devices in recent years [1,2,3]. However, for most flexible electrodes, especially for fiber-like [4,5,6] and paper-like [7,8,9] ones, the energy output for single electrode is limited and not suitable to use for large-scale devices [3, 10]. Therefore, three-dimensional (3D) flexible electrodes arouse an increasing attention [3, 11,12,13,14]. In general, 3D electrodes are composed of body materials and active fillers [15,16,17]. Body materials include cotton, carbon cloth (CC), sponge-like monoliths [18,19,20], etc. Fillers involve activated carbon (AC), metallic oxides, etc. [3, 21,22,23,24]. For most 3D electrode, the weight of electrochemical passive body material may be an obstacle to the specific capacitance of the whole electrode [3, 25].

Flexible supercapacitor electrodes based on activated carbon fiber felt (ACFF) were reported lately by our laboratory [26]. ACFF provides an areal capacitance of ~2000 mF cm−2. The porosity of ACFF is ~95%, which ensures the ion movements within the material and allows the introduction for additives into the textile [27, 28]. In the previous study, carbon nanotube (CNT) and graphene (GN) were introduced into ACFF using freeze-drying method and resulted in significant enhancements in electrochemical performances [26]. However, the CNT and GN only improve the electrical conductivity and possess nearly no capacitance, and the high porosity of ACFF may indicate a possibility for introducing more fillers into the matrix.

To further improve the electrochemical performances of composite textiles and make use of the porous matrix, MnO2 was considered to be introduced into ACFF, as MnO2 preserves high theoretical capacitance [29,30,31]. Different procedures were researched to fabricate a most promising composite textile using CNT and MnO2. It was found out that among all the prepared textiles, MnO2/CNT/ACFF composite textiles fabricated by dipping ACFF into CNT and MnO2/CNT suspensions process the best electrochemical properties. The comparison of different approaches may provide a train of thought to design a composite material, and the fabricated composite textile which achieved significant improvements in electrochemical performances may be a promising material for flexible supercapacitors.

Experimental

Preparation of MnO2/CNT suspensions

ACFF textiles used in the experiments were obtained from Shenzhou Carbon Fiber Co., Ltd, China. 5 wt% CNT paste (Model: NTP2021) was purchased from Shenzhen Nanotech Port Co., Ltd. In order to get MnO2/CNT suspensions, 12 g 5 wt% CNT pastes and 50 ml 0.4 mol L−1 KMnO4 aqueous solution were mixed at ambient temperature and stirred for 12 h to speed up reactions and disperse extra CNTs and MnO2 uniformly [32,33,34,35]. Filtration method was used to get reaction products, and during this period, products were washed by deionized water for several times to ensure no MnO4 − exists. Products were then dried at 80 °C and dispersed in deionized water to get suspensions with the concentration of 2 wt%. 1 wt% CNT suspensions were prepared by mixing 5 wt% pastes with deionized water.

Preparation of different textiles

CNTs and MnO2 were introduced into ACFF textiles by a “dipping and drying” method. The preparation procedures for different textiles are displayed in Fig. 1a. ACFF were dipped into prepared CNT suspension for 1 min and dried at 80 °C. The prepared CNT/ACFF textiles are denoted as “FC” (procedure 1). “FKM” was prepared by directly dipping ACFF into 0.1 M KMnO4 aqueous solution for 1 min, washed by deionized water thoroughly and dried at 80 °C (procedure 2). Textiles denoted as “FM” were prepared by dipping ACFF textiles into prepared MnO2/CNT suspensions (procedure 3). “FCM” textiles were fabricated by a two-step procedure, dipping ACFF into the two kinds of suspensions separately (procedure 4).

Figure 1
figure 1

Schematics of different textiles, including fabrication processes and microstructures (a). SEM images of CNT fillers (b) and MnO2/CNT fillers (c)

Full size image

Characterizations of prepared textiles

Morphologies of CNT, MnO2/CNT fillers, ACFF and prepared composite textiles were observed using scanning electron microscopy (SEM). Energy-dispersive spectroscopy (EDS) mapping was conducted on FKM, FM and FCM textiles to show the distribution of C, O, Mn and K elements. X-ray diffraction (XRD) was performed on MnO2/CNT powders and FKM textiles to determine the crystal structure of generated MnO2. An electronic balance was used to conduct the weights of textiles both before and after preparing processes. And areal densities (D) for different textiles were calculated baed on the results. Thermal gravimetric analyses (TG) were conducted on FM and FCM textiles. Electrical conductivities of all textiles were measured in a four-point probe instrument. The specific surface area and pore-size distribution of different textiles were described in a Brunauer–Emmett–Teller (BET) analyzer by N2 adsorption tests.

Coin-cell supercapacitors were prepared to test electrochemical performances of the materials. CNT powders (dried from CNT paste) or MnO2/CNT powders were mixed with acetylene black and polyvinylidene fluoride (PVDF) by a weight ratio of 8:1:1, respectively. N-Methylpyrrolidone (NMP) was used to disperse the powder thoroughly. The mixtures were coated on stainless steel foil and dried at 110 °C for at least 10 h to form CNT or MnO2/CNT electrodes. Symmetric coin-cell supercapacitors with 1 M H2SO4 electrolyte were prepared to characterize the electrochemical performances of CNT electrodes, MnO2/CNT electrodes and various textile electrodes, and the area of each electrode is the same ~1.131 cm2 and each electrode weights 20–30 mg for different textiles. Cyclic voltammetry (CV) tests and galvanostatic charge–discharge measurements were performed at a VMP3 electrochemical station. Areal capacitances (C S), energy densities (E S) and power densities (P S) of different electrodes/supercapacitors were then calculated.

Results and discussion

Morphology of textiles

Microstructures of CNT and MnO2/CNT fillers are shown in Fig. 1b, c. MnO2 grains on the surface of CNTs are observed in MnO2/CNT fillers. X-ray diffraction patterns of MnO2/CNT fillers and FKM textile are shown in Fig. S2. From the XRD pattern, a type of birnessite MnO2 is found in both materials. SEM images of different textiles and surface morphologies of carbon fibers are presented in Fig. 2a–e and their insets. ACFF is constructed by many disordered carbon fibers and has high porosity. The surface of carbon fibers is smooth. For composite textiles, several fillers can be found on the surface of carbon fibers. Flaky joints between carbon fibers distributed by CNTs are found in FC, and these joints may help increase the electronic conductivity. In FKM, the main structure of ACFF keeps its status and almost MnO2 grew uniformly in the form of nanosheets on the surface of carbon fibers. However, this may cause a decrement in the electronic conductivity because of the low electronic conductivity of MnO2. For FM, MnO2/CNT fillers form irregular-shaped aggregations. In FCM textiles, both flaky joints between carbon fibers and irregular-shaped aggregations are observed. The detailed images of the two kinds of morphologies are presented in Fig. 2f, g. The distribution of MnO2 in FCM and comparison of the morphologies among FC, FCM and FKM are further discussed in supporting information using EDS results (Fig. S3) and detailed SEM images (Fig. S4).

Figure 2
figure 2

SEM images of ACFF (a), FC (b), FKM (c), FM (d), and FCM textiles (e), insets are magnified images of single fiber for each textile; f and g are detailed images of two main morphologies in Fig. 2e (FCM textile); weight fraction composition (h) and electrical conductivity of textiles (i)

Full size image

The percentage of fillers in composite textiles (Fig. 2h) ranges from ~15 to ~25%. Detailed compositions of different textiles were calculated by computing methods given in supporting information using TG results shown in Fig. S2. Electronic conductivity values of different textiles are given in Fig. 2i. Compared with ACFF, electronic conductivities of FKM suffered a significant decrement. For other composite textiles, the introduction of fillers into textiles results in an improvement in electronic conductivity because of the introduction of CNT fillers [36]. Table 1 presents the values of areal density (D), S BET, V p and d of different textiles. For all the composite textiles, areal density shows an increment compared with ACFF because of the introduction of fillers. ACFF preserves a high S BET of ~1700 m2 g−1, while S BET, V p and d values shows rather smaller values than ACFF. It can be seen from SEM images that some filler covers on the surface of carbon fibers, which may cause the decrement in S BET value. The aggregations occupy some holes in ACFF, which leads V p and d values to decrease in composite textiles.

Table 1 D, S BET, V p and d values of different textiles
Full size table

Electrochemical performance of textile electrodes

CV curves of CNT and MnO2/CNT fillers are displayed in Fig. 3a, b. The CV curves of CNT show nearly rectangular shape at different scan rates, indicating a double-layer capacitive behavior. Its capacitance is ~20 F g−1, similar to commercial CNTs reported [37,38,39]. Capacitance of MnO2/CNT, calculated from CV curve at 2 mV s−1, is ~150 F g−1, much higher than CNT.

Figure 3
figure 3

CV curves at different scan rates of CNT (a), MnO2/CNT (b), ACFF (c), FC (d), FKM (e), FM (f), and FCM (g). Areal capacitances of each electrode calculated from CV curves (h)

Full size image

CV scans were performed on different textiles at scan rates from 2 to 20 mV s−1, and results are shown in Fig. 3c–g. Capacitance of ACFF is ~2300 mF cm−2 at 2 mV s−1, and obvious deviations of CV curves are observed at higher scan rates, indicating poor rate performance of ACFF. As shown in Fig. 2d, CNT fillers can optimize the electrochemical performance of ACFF, including areal capacitance and rate performance, by enhancing the electronic conductivity of ACFF [26]. The capacitances of FKM are much lower compared with ACFF. It shows that introducing MnO2 into ACFF by this method cannot improve the electrochemical performance (Fig. 2e). When MnO2 is introduced in the form of MnO2/CNT fillers, improvements in areal capacitance are displayed. However, at higher scan rates, the capacitance of FM decreases faster than FC. Among all the textiles, FCM shows most inspiring electrochemical performance. An areal capacitance result of ~3717 mF cm−2 at 2 mV s−1 is achieved. CV curves of FCM still remain nearly rectangular at a scan rate of 20 mV s−1, demonstrating that the fillers provide crucial improvements on both areal capacitance and rate performance.

The results may be explained from Fig. 2 given before. To obtain high areal capacitance, introducing MnO2 into the textile may be sufficient. However, FKM, which possesses the highest percentage of MnO2, presents lowest areal capacitance among the textiles, even lower than ACFF. This is because the direct growth of MnO2 on the surface of carbon fibers results in poor electronic conductivity of the composite textile. For FM and FCM, MnO2 was introduced into ACFF in the form of MnO2/CNT fillers. This results in less MnO2 concentration in composite textiles but an improvement on electronic conductivity. For FC, FM and FCM, CNTs are all introduced into the textiles, but in two different forms: CNT or MnO2/CNT fillers. Because of the existence of MnO2, MnO2/CNT fillers achieves higher capacitance but worse rate performance than CNT fillers, and this causes similar results for FC and FM textiles. And over-all improvements in the composite textiles are realized by introducing the two kinds of fillers together.

Galvanostatic charge–discharge measurements of different textile supercapacitors were taken to obtain the relationship curves of voltage vs. time, presented in Fig. 4a–e. Compared with ACFF, charge–discharge curves of FC, FM and FCM are more ideally V-shaped. Longer discharge times and smaller IR drops are attained. At 5 mA cm−2, IR drop of ACFF and FKM is ~0.3 and ~0.7 V, respectively, while for supercapacitors fabricated with FC, FM and FCM composite textile electrodes, the IR drops are all ~0.1 V. The tendencies of the areal capacitances calculated by charge–discharge curves shown in Fig. 4f are similar with Fig. 3h and may be explained by the results of equivalent series resistance (ESR), which stands for the internal resistance of supercapacitors [40]. ESR is affected by many factors, including resistance of the electrode, ion diffusion resistances and interfacial resistances between electrodes and electrolyte. At 5 mA cm−2, ESR values of ACFF, FC, FM, FKM and FCM supercapacitors are about 27, 7, 11, 59 and 4 Ω cm−2, respectively. As a result, except for FKM, all composite textiles show better performances than ACFF. For FM, the ESR value is higher than FC and FCM, which is associated with the lower electronic conductivity of FM textile (Fig. 2i). As a result, the IR drops are larger and the rate performances are not so desirable at larger current densities. FC and FCM are similar with each other on rate performances, and the differences of capacitances are mainly caused by the contribution of MnO2 in MnO2/CNT fillers. A capacitance of 4148 mF cm−2 is achieved for FCM at the scan rate of 2 mA cm−2.

Figure 4
figure 4

Voltage versus time curves from galvanostatic charge–discharge tests of different textiles, ACFF (a), FC (b), FKM (c), FM (d), and FCM (e). Areal capacitance of each supercapacitor calculated form galvanostatic charge–discharge tests (f); Ragone plots of the supercapacitors (g); cycle performances of FCM supercapacitor at a current density of 30 mA cm−2 for 10000 cycles (h); capacitance retention and CV curves at 2 mV s−1 of FCM electrodes after bending for 0, 50 and 100 cycles (i)

Full size image

Because of the large specific capacitances, high energy densities and power densities are obtained. Ragone plots of prepared supercapacitors are shown in Fig. 4g. Compared with ACFF, FC and FCM show considerably well performances. For instance, the energy density of ACFF dropped sharply from ~85 to less than 5 μWh cm−2 when power density increases from ~450 to ~1600 μW cm−2, while FCM ranges from 141 to 103 μWh cm−2 with power density from 494 to 4466 μW cm−2.

Cycling stability of FCM supercapacitor at a current density of 30 mA cm−2 is shown in Fig. 4h. For the first 500 cycles, a slight capacitance loss of ~1% is observed. After that, an increment in capacitance occurs from ~1000 to ~3000 cycles. After 4000 cycles, the capacitance retention stabilized at ~110%. As is shown in the inset graph, curves of the 5000th and the 10000th cycle are nearly coincident. This result indicates good cycling stability of composite textiles.

A comparison was made among previous reports on flexible supercapacitors and this work and the results are shown in Table 2. Because the surface area for human body is limited, it is necessary to discuss the electrochemical performances of supercapacitors for per unit area [3]. For this work, although E S and P S results are not the best, the areal capacitance result and cyclic performance are inspiring. Considering the composite textiles are fabricated by such simple processes and industrial raw materials, these results are more promising and competitive. Besides, the ability to withstand deformations such as many rounds of bending is an important characteristic for flexible electrodes. The FCM electrodes were repeatedly bent from 0° to 180° as shown in Fig. 4i. 96% capacitance retained after bending for 50 cycles, and bending 100 cycles brings only about 8% capacitance degradations. After these deformations, FCM electrode keeps its original appearances, and CV curves remain an ideal rectangular shape. These results indicate the excellent flexibility of prepared textiles and suggest that the fillers adhere strongly to ACFF textiles, which are essential for roll-to-roll fabrications: a method often used for mass production of flexible energy storage devices [42,43,44]. Devices based on activated carbon textiles (ACT) have already been realized [45]. So, it can be inferred that with a similar 3D flexible structure and good durability when suffering deformation, the prepared composite textile electrodes are practical to be fabricated into energy storage devices by roll-to-roll manufacturing.

Table 2 Comparison of prepared FCM supercapacitor with those in published papers
Full size table

Solid-state flexible supercapacitors were fabricated with FCM textiles and gel electrolyte. Schematic of fabrication processes is shown in Fig. 5a, and detailed procedures are found in supporting information [41]. Digital images and CV curves of supercapacitors at different bending angles are displayed in Fig. 5b. CV curves for prepared flexible supercapacitors show great deviation from symmetric coin-cell supercapacitors mentioned above. Because the textile electrodes are thick (~2 mm), it is difficult for electrodes to contact well with gel electrolytes. Also, pressure was added when preparing coin-cell supercapacitors, and this may cause a significant difference to the performances of textile electrodes. These problems may be easily solved by industrial fabrication methods, for example, roll-to-roll manufacturing mentioned above. CV curves of supercapacitors at different statuses are similar with each other. The capacitance retentions of supercapacitors at the bending angle of 90° and 180° are 99 and 93% compared with flat status, respectively. The flexible supercapacitor also shows good capability of bending deformation. After bending for 50 or 100 cycles, the capacitance almost keeps constant, confirming the flexibility of prepared supercapacitors. Using simple fabrication methods and structure design, MnO2/CNT/ACFF composite textiles were prepared and ready to use as flexible supercapacitor electrodes. Considering the needs of portable and wearable devices, flexible batteries are also of great significance [46,47,48]. Researches have been carried out on composite textile electrodes for flexible batteries, and promising results have been achieved [49, 50]. Using the thought of structure design explained above, different fillers can be combined with the electrochemical active flexible substrates. It is possible to extend the usage of this kind of composite textiles to flexible batteries.

Figure 5
figure 5

Fabrication schematic of flexible supercapacitor with FCM textile electrodes and PVA/KOH gel electrolyte (a); digital images, capacitance retention and CV curves of supercapacitors at flat, bent and rolled-up statuses (b); capacitance retention and CV curves of the flexible supercapacitors after bending for 0, 50 and 100 cycles

Full size image

Conclusions

Various fabrication procedures are designed to introducing CNT and MnO2 fillers into ACFF textiles. And the comparison of characteristics among prepared textiles was made. Electronic conductivity is found to increase with the introduction of CNTs, and introducing MnO2/CNT fillers rather than growing MnO2 directly on the surface of carbon fibers can improve electrochemical capacitances of composite textiles. Inspiring electrochemical performances, including areal capacitance, rate performance and cycle stability, are achieved for MnO2/CNT/ACFF composite textiles composed by ACFF, CNTs and MnO2/CNT fillers. Flexible supercapacitors prepared using composite textiles electrodes can be bent at different angles and endure repeated bending with little capacitance loss. As a whole, the comparison of textiles fabricated by different method may provide an idea of how to design a composite material. And composite textiles with excellent electrochemical performances and flexibility are promising for applications of high-performance flexible supercapacitor electrodes.