Supercapacitors (SCs) are emerging as a new class of energy storage device because they can store more energy than conventional capacitors and provide higher power than batteries.1, 4 Therefore, SCs have potential applications in a number of electronic devices, including hybrid electric vehicles, laptops, cell phones and cameras.5, 6 However, to meet the increasing energy demands for next-generation portable and flexible devices, the energy density of SCs should be substantially increased without sacrificing the power density and cycle life.3, 7 Energy density (E) of a SC device can be calculated according to the equation E = 1/2 CV2,5, 8 and increased by maximizing the device capacitance (C) and the cell voltage (V). An effective approach to increase the cell voltage is to use organic (vs aqueous) electrolytes,9, 11 such as tetraethylammonium tetrafluoroborate in acetonitrile,10 or lithium perchlorate in propylene carbonate.11 While organic electrolytes can provide a wider voltage window (up to ≈3 V) than aqueous electrolytes, they usually suffer from high cost, poor ionic conductivity and high toxicity, which could limit their applications. A promising alternative is using aqueous electrolytes that have higher ionic conductivities and are more environmentally friendly, and to increase cell voltages by developing asymmetric supercapacitors (ASCs) with a battery-type Faradaic electrode as energy source and a capacitive electrode as a power source.8, 12, 15 In comparison to symmetric supercapacitors (SSCs), ASCs combine the voltage windows of two different electrodes to increase the maximum operation voltage in aqueous electrolyte (up to 2 V), and, in turn, the device energy and power density.15, 16 Intensive efforts have been devoted to explore various ASC systems,8, 12, 18 such as graphene-MnO2//activated carbon nanofiber,8 Ni(OH)2-graphene and graphene,16 V2O5//active carbon,17 and MnO2//FeOOH.18 While most ASC devices are operated in aqueous electrolytes, the device fabrication requires high-cost encapsulation techniques to avoid the possible leakage of electrolytes, and it is difficult to make small and flexible devices. In comparison to conventional SCs, solid-state SCs have advantages such as small-size, light-weight, ease of handing, excellent reliability, and a wider range of operating temperatures.19, 22 In this regard, solid-state SCs hold great promises for flexible and wearable electronics. However, the fabrication of solid-state ASCs has been barely explored.12 Recently, Lee and co-workers reported a solid-state ASC based on RuO2 and graphene films.12 The realization of low-cost, solid-state ASCs with high specific/volumetric energy and power density is still challenging.
In this work, we focused on the development of high- performance and flexible solid-state ASC devices based on one-dimensional core–shell nanowire (NW) electrodes. Core–shell NWs were grown directly on flexible conductive substrates, which not only provide good strain accommodation, but also enable the fabrication of flexible SC devices without the need of a binder.19, 23 In comparison to bulk materials, this NW branching structure provides a large interfacial area between electrode and electrolyte for charge transport, and a shortened diffusion path for intercalation/de-intercalation of active species.24, 25 Moreover, the unique core–shell structure could lead to new and/or enhanced function of the electrode. For instance, the core material could serve as a conducting scaffold for supporting electrochemically active materials.23 MnO2 is a promising pseudocapacitive material with high theoretical specific capacitance (≈1400 F g−1). However, MnO2 as an electrode material suffers from relatively low electrical conductivity (10−5 − 10−6 S cm−1).4, 21, 26 TiO2 is an inexpensive and electrochemically stable semiconductor commonly used as electrodes in electrochemical devices. It has higher electrical conductivity (10−5 − 10−2 S cm−1)7, 27, 28 than MnO2. More importantly, we have recently demonstrated that the carrier density of pristine TiO2 can be increased by 3 orders of magnitudes upon hydrogenation. Herein, we report the use of hydrogen-treated TiO2 (denoted as H-TiO2) NWs as the core (conducting scaffold) to support electrochemically active MnO2 and carbon shells. The solid-state ASC device based on H-TiO2@MnO2 (positive electrode)//H-TiO2@C (negative electrode) core–shell nanowire electrodes achieved a maximum energy density of 0.30 mWh cm−3, which is higher than most of reported solid-state SCs.5, 6, 19, 21, 29
The growth procedures of core–shell NW electrodes are illustrated in Figure 1a. H-TiO2 NWs were prepared on a carbon cloth by hydrothermal method (see the Experimental Section). Scanning electron microscopy (SEM) images show the entire surface of the carbon fibers was covered uniformly by H-TiO2 NWs with diameters of 100–200 nm and lengths of 1–2 μm (Figure S1, Supporting Information). X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy studies revealed the formation of oxygen vacancies (Ti3+ sites) on TiO2 NW surfaces after hydrogenation (Figure S2,S3). Electrochemical impedance spectroscopy data (Figure S4) showed that the charge transfer resistance (Rct) for H-TiO2 NWs (3.5 Ω) is smaller than pristine TiO2 NWs (21.7 Ω), indicating improved electrical conductivity. Amorphous MnO2 and carbon layers were conformably coated onto the H-TiO2 NW surface via anodic electrodeposition and the hydrothermal method, respectively (see the Experimental Section). The morphology and alignment of H-TiO2 NWs were preserved after shell coating (Figure 1b,d). Transmission electron microscopy (TEM) images collected for a representative H-TiO2@MnO2 core–shell NW showed that an amorphous layer was uniformly coated on the entire NW (Figure S5). Selected-area electron diffraction (SAED) analysis confirmed that the H-TiO2 NW is a single crystal and grows along the  axis. The lattice fringe spacing was measured to be 0.32 nm (Figure 1c, upper inset), which is consistent with the d-spacing of (110) planes of rutile TiO2 (JCPDF # 65-0192). TEM image also revealed the amorphous nature of the MnO2 shell, which has an average thickness of 7.0 nm. XPS analysis confirmed the composition of this shell is MnO2 (Figure S6). Likewise, TEM images collected from H-TiO2@C core–shell NWs showed that an amorphous carbon shell of about 18 nm in thickness was uniformly covered on the H-TiO2 NW surface (Figure 1e and Figure S7a). The successful deposition of carbon shell was also confirmed by Raman spectroscopy (Figure S7b).
The electrochemical studies for the core–shell NWs were conducted in a three-electrode cell in 5 M LiCl aqueous electrolyte, with a Pt counter electrode and a Ag/AgCl reference electrode. Figure 2a shows the cyclic voltammograms (CVs) collected for a H-TiO2@MnO2 film directly coated on a carbon cloth substrate (MnO2), TiO2@MnO2 and H-TiO2@MnO2 electrodes, at a scan rate of 100 mV s−1. The loading amounts of MnO2 are comparable in these electrodes (about 0.19–0.23 mg cm−2). As expected, the MnO2 containing electrodes exhibited substantially larger current density than H-TiO2 electrode because MnO2 is more electrochemically active than TiO2. Significantly, the current density of the H-TiO2@MnO2 electrode is higher than the values obtained for the MnO2 and TiO2@MnO2 electrodes. This can be attributed to the increased surface area and improved charge transport from MnO2 materials to the carbon cloth substrate via the TiO2 NW core. Hydrogenation improved the electrical conductivity of TiO2, which explains the increased performance observed for the H-TiO2@MnO2 electrode compared to the TiO2@MnO2 electrode. Given that the H-TiO2 NW electrode has negligible capacitance compared to the capacitance of MnO2 based electrodes, the specific capacitances of these electrodes were calculated based on the mass of MnO2 (Figure 2b). H-TiO2@MnO2 yielded the highest specific capacitance of 449.6 F g−1 at the scan rate of 10 mV s−1, whereas the MnO2 and TiO2@MnO2 electrodes only achieved 325.8 and 359.7 F g−1 respectively at the same scan rate. When the scan rate changed from 10 to 200 mV s−1, the H-TiO2@MnO2 electrode retained 54.6% of its capacitance, which is also substantially higher than that obtained for the MnO2 (29.4%) and TiO2@MnO2 (43.4%) electrodes. Additionally, galvanostatic measurements showed that the H-TiO2@MnO2 electrode exhibits the smallest IR drop and longest discharge time (Figure S8) among these electrodes, again confirming the superior electrical conductivity and electrochemical performance of the H-TiO2@MnO2 electrode. These results convincingly show that the H-TiO2 NWs are good supports for electrochemically active MnO2. Furthermore, CV curves were collected for an amorphous carbon layer directly coated on the carbon cloth substrate (C), TiO2@C, and H-TiO2@C electrodes (Figure 2c). The specific capacitance of the H-TiO2@C electrode reached 253.4 F g−1 at the scan rate of 10 mV s−1, which is larger than the values obtained for the C (162.2 F g−1) and TiO2@C (197.1 F g−1) electrodes (Figure 3d). Moreover, the H-TiO2@C electrode showed excellent rate capacitance of 70.1% when the scan rate increased from 10 to 400 mV s−1. This rate capacitance is considerably larger than the values obtained for the C (44.4%) and TiO2@C (49.7%) electrodes. The enhanced performance of the H-TiO2@MnO2 and H-TiO2@C electrodes is due to the increased accessible surface area by the NW structure and improved electrical conductivity of TiO2 NWs after hydrogenation.
According to the CV studies, we confirmed that H-TiO2@MnO2 and H-TiO2@C electrodes have stable voltage windows between 0.0 and 0.8 V and between –1.0 and 0 V (vs Ag/AgCl), respectively. Therefore, it is expected that the operating cell voltage can be extended to 1.8 V when they are assembled into ASCs. Figure 3a and Figure S9a show the CV and discharge-charge curves collected at different voltage windows for the H-TiO2@MnO2//H-TiO2@C solid-state ASC in a LiCl/PVA gel electrolyte. Indeed, the stable electrochemical windows of the ASC can be extended to 1.8 V. Significantly, the calculated volumetric capacitance (based on the volume of the entire device) increases from 0.37 to 0.68 F cm−3 when the operation voltage increases from 0.8 to 1.8 V (Figure S9b). According to the equation of E = 1/2 CV2, the energy density of the ASC was improved by 900%.8
CV curves of the H-TiO2@MnO2//H-TiO2@C ASC device measured at various scan rates with voltage windows ranging from 0 to 1.8 V exhibit rectangular-like shapes (Figure S10), revealing the ideal capacitive behavior and a fast charge/discharge property. The electrochemical performance of the ASC device was also measured in 5 M LiCl aqueous electrolyte for comparison. Figure 3b shows the calculated volumetric capacitance and specific capacitance (based on the mass of MnO2 and carbon shell) based on the CV curves. Significantly, the volumetric and specific capacitance of the solid-state device are comparable to the aqueous device. The solid-state ASC device achieved the maximum volumetric and specific capacitance of 0.71 F cm−3 and 141.8 F g−1 at 10 mV s−1, which are substantially higher than the values recently reported for graphene/MnO2//activated carbon nanofiber (78 F g−1),8 MnO2 NW/graphene//graphene (40 F g−1),15 and MnO2 nanofibers/graphitic hollow carbon spheres//graphitic hollow carbon spheres (50 F g−1)13 at the same scan rate. Moreover, the solid-state ASC device shows good rate capacitance, with 56% of volumetric capacitance retained when the scan rate increased from 10 to 400 mV s−1.
The superior performance of the solid-state ASC device was further confirmed by galvanostatic charge/discharge measurements. As shown in Figure 3c, the charging and discharge curves of the solid-state ASC device are reasonably symmetric, with a good linear relation of discharge/charge voltage versus time. This again demonstrates the ideal capacitive characteristic and rapid charge/discharge property of the ASC device. The volumetric capacitance calculated based on the discharging curves was calculated to be 0.70 F cm−3 at 0.5 mA cm−2 (139.6 F g−1 at 1.1 A g−1), which is substantially larger than the values obtained from recent reports for other solid-state SCs.5, 19, 21, 29 The excellent performance of solid-state ASC devices can be ascribed to the following reasons: (1) NW electrodes exhibit a large accessible area that allows efficient ion intercalation/de-intercalation; 2) the highly conducting H-TiO2 NW core provides an effective pathway for charge transport; and (3) the binder-free device fabrication enables a low interfacial resistance and fast electrochemical reaction rate. Moreover, the electrode materials fabricated on carbon cloth substrate hold great potentials for developing flexible energy storage devices. Significantly, the CV curves collected for the ASC device under flat, bent, and twisted conditions are similar (Figure 3d), indicating their excellent mechanical stability as flexible energy storage devices.
The long-term cycling performance of the ASC devices were evaluated in the 1.8 V voltage window at a scan rate of 100 mV s−1 for 5000 cycles (Figure 4a). The solid-state ASC device exhibits excellent stability with 91.2% retention of the initial capacitance after 5000 cycles. This retention rate is slightly better than the values reported for other aqueous and solid-state ASCs, such as MnO2 NW/graphene//graphene (79% after 1000 cycles),15 MnO2//graphene hydrogel (83.4% after 5000 cycles),14 PANI/CNT//PANI/CNT (88.6% after 1000 cycles),22 and RuO2/graphene//graphene (95% after 2000 cycles).12 In comparison to aqueous electrolyte, solid-state electrolyte suppressed the dissolution of MnO2, and thus, improved the cycling stability.
Figure 4b compares the volumetric power (P) and energy densities (W) of the ASC devices to the values reported for other solid-state symmetric SCs (SSCs). The maximum volumetric energy of the flexible ASC device was calculated to be 0.30 mWh cm−3, which is substantially higher than values reported for other solid-state SCs, such as multiple walled carbon nanotubes-based SSCs (0.008 mWh cm−3, PVA/H2SO4),29 single-walled carbon nanotubes-based SSCs (0.01 mWh cm−3, PVA/H3PO4),6 graphene-based SSCs (0.06 mWh cm−3, PVA/H3PO4),5 and MnO2/carbon particle (CNP)-based SSCs.21 Furthermore, we also compared the specific energy and power density of the ASC devices (based on the mass of MnO2 and C) to the previously reported devices (Figure S11). Importantly, the maximum energy densities obtained for H-TiO2@MnO2//H-TiO2@C devices with a cell voltage of 1.8 V in both aqueous and solid-state electrolyte are considerably higher than the values reported for other MnO2-based ASCs.8, 13, 15, 30, 31 To demonstrate the potential application of these flexible solid-state ASCs, a ASC device was employed to power a red light-emitting-diode (LED). The ASC device can power a red LED (1.5 V) for about 2 min after charging at 2 mA cm−2 for 15 s.
In summary, we have demonstrated a flexible solid-state ASC device with H-TiO2@MnO2 core–shell NWs as the positive electrode and the H-TiO2@C core–shell NWs as the negative electrode. This device operates in a 1.8 V voltage window and is able to deliver a high specific capacitance of 139.6 F g−1, maximum volumetric energy density of 0.30 mWh cm−3 (59 Wh kg−1) and volumetric power density of 0.23 W cm−3 (45 kW kg−1). Additionally, the device exhibits excellent cycling performance (8.8% capacitance loss after 5000 cycles) and good flexibility. The capability of developing complex nanostuctured electrodes could advance the design and fabrication of high-performance and flexible ASCs.