Sustainable direct current powering a triboelectric nanogenerator via a novel asymmetrical design

Hanjun Ryu a, Jeong Hwan Lee a, Usman Khan a, Sung Soo Kwak a, Ronan Hinchet a and Sang-Woo Kim *ab
aSchool of Advanced Materials Science and Engineering, Sungkyunkwan University (SKKU), Suwon, 16419, Republic of Korea
bSKKU Advanced Institute of Nanotechnology (SAINT), Sungkyunkwan University (SKKU), Suwon, 16419, Republic of Korea. E-mail:

Received 19th January 2018 , Accepted 27th March 2018

First published on 28th March 2018

Triboelectric nanogenerators (TENGs) typically have very low outputs in terms of root-mean-square (RMS) values as the outputs are in the form of short duration pulses. In order to overcome this limitation, here we introduce multi-phase rotation-type TENGs (MP-TENGs). The rotator of the MP-TENGs consists of alternating components of PTFE and mono cast nylon, whereas the stator consists of PTFE with multiple back electrodes. Each electrode produces an output of the same frequency and amplitude, but of different regularly shifted phases. Consequently, the outputs from all the electrodes are full-wave rectified and superimposed, and the total output of the MP-TENGs is almost constant direct current (DC), unlike typical TENGs, with much higher RMS values. A 5-phase TENG has produced 380 V, 3.6 mA m−2, and 4.9 W m−2, and the current crest factor is dramatically decreased to 1.26. Finally, due to the almost constant DC output, the MP-TENGs successfully operate electronic devices and charge a battery.

Broader context

Recently, converting wasted mechanical energy into electric energy has received great attention as a new energy source of wireless sensor network systems for independent and permanent operation. Triboelectrification effect based nanogenerators have already been adapted to extremely small power consumption electronics, but it is greatly required to enhance the nanogenerator's performance for a variety of electronics. So far, typical approaches have been utilized to increase the power output performance, but there have been few reports on the generation of direct current output performance. Here, we report a new strategy of multi-phase rotation-type triboelectric nanogenerators (TENGs) to generate direct current output in order to enhance the root-mean-squre (RMS) output property. Each electrode generates regularly phase shifted alternative currents having the same frequency and amplitude. The full-wave rectified output currents are a superposition of the resulting direct current outputs which reach their peak value sequentially. Therefore, multi-phase TENGs are promising generators for self-powered electronics to achieve high RMS output performance.


Nanogenerators (NGs) have recently been introduced for the purpose of self-powering largely distributed wireless electronics by harvesting various mechanical energies from their working environment.1–4 Although various energy converting systems such as triboelectric,5–9 piezoelectric,10–12 and pyroelectric13–15 have been investigated, stable direct current (DC) generating systems need to be developed for practical purposes. Various strategies for triboelectric NGs (TENGs) have been proposed to improve high performance and high efficiency alternating current (AC)–DC transforming systems.16–19 Nonetheless, the output of TENGs is typically in the form of short duration pulses. Consequently, the root-mean square (RMS) value (or the effective value) of the output voltage and current is much lower than the peak value. The crest factor, which is defined as the ratio of the peak to the RMS value of output voltage is therefore much higher for TENGs. For example, for the TENGs,20 the crest factor was 6.7, which is 4.7 times higher than that of a sinusoidal waveform electromagnetic generator (EMG), which is a very popular mechanical energy harvesting technology.21 Such typical high crest factors of TENGs can seriously influence their performance in terms of energy storage using batteries for powering various applications. The conventional approach for overcoming this limitation includes increasing the frequency to bring the pulses near to, and enhance, the peak power, although this may not always be possible, such as in biomechanical applications where the working frequencies and powers are typically low.22–24 Therefore, the development of a proper TENG design which inherently has low crest factors is urgently needed.

In order to overcome the TENG's limitation of typically high crest factors, here, we introduce for the first time a multi-phase TENG (MP-TENG) design with a low current crest factor of 1.26. The MP-TENG is based on a typical rotation type TENG structure, but it has multiple electrodes. Each of the electrodes generates an AC of the same frequency and amplitude, but their phases differ and regularly shift. The multiple phase shifted AC outputs are full-wave rectified and the TENG's output is a superposition of the resulting DC outputs. Due to the regularly shifted phases, the TENG output reaches its peak value sequentially during each of the individual signals. Consequently, an almost constant DC output power is delivered to an external load. For the output performance, the MP-TENG demonstrated a current crest factor as low as 1.26, which is close to the DC crest factor of 1, as well as a charging efficiency of around 5 times higher than that of the single phase TENG. In addition, the dependence of the TENG's output on the number of phases, rotational speed (i.e., revolutions per minute (rpm)), and number of segments of the TENG is thoroughly investigated. As practical applications, it has been shown that the MP-TENG can periodically charge a wearable electronic device, activate a temperature sensor, and fully charge a commercial battery. This novel design of a MP-TENG ensures high efficiency and can impose on all rotation type TENGs for DC output performance. Also, the proposed MP-TENG will contribute to today's smart lifestyle as the technology mitigates the need for additional charging of the battery of each electronic device.


Design of the MP-TENG and electrical performance

The MP-TENG employs a rotation type TENG design with multiple electrodes that have regularly shifted phases. It consists of two parts, a rotator and a stator, as shown in Fig. 1a. The rotator is composed of 5 cm radius polytetrafluoroethylene (PTFE) and mono cast (MC) nylon components. The PTFE and MC nylon components are alternately mounted onto an acrylic substrate using an adhesion layer. In addition, the surface of the rotator is mechanically polished in order to ensure surface uniformity. The stator, on the other hand, consists of an electrification layer (PTFE) with back electrodes. The electrodes are formed on a printed circuit board (PCB) and the PTFE layer is attached to the electrodes using an adhesion layer; the adjacent electrodes have a 1 mm gap between them in order to avoid the possibility of short circuiting; the detailed fabrication process is described in Fig. S1 (ESI).
image file: c8ee00188j-f1.tif
Fig. 1 Design and output performance of the multi-phase triboelectric nanogenerator. (a) Schematic description of the 5-phase TENG. (b) Multi-phase management circuit diagram of the MP-TENG. (c) Output voltage of the MP-TENGs during 1 cycle. (d) Output current of the MP-TENGs during 1 cycle.

Since circular segmentation is proposed to enhance the output current of sliding-mode rotation-type TENGs,25 each electrode can be further divided into a number of segments. Therefore, for N number of segments per electrode and n number of total phases, the segment angle of each electrode, θs, can be determined by the following relation, eqn (1).

image file: c8ee00188j-t1.tif(1)

In order to have regularly shifted phases, every electrode needs to have a smaller area than the area of the rotator's components, and also needs to have a regularly different PTFE and MC nylon contact ratio. In principle, this can be guaranteed by ensuring the number of rotator components is one less than the number of phases. The angle of each component, θr can therefore be determined as follows, eqn (2).

image file: c8ee00188j-t2.tif(2)

Each electrode produces a separate AC output of the same frequency and amplitude, but with regularly shifted phases resulting from the regularly different contact ratio of the rotator's components. However, in order to determine the total output, the AC outputs from each electrode are rectified using full-wave bridge rectifiers and are superimposed as shown in Fig. 1b. The diodes had a high breakdown voltage and low reverse current diode to better adapt the TENG. Regularly shifted phases are crucial in order to have an almost constant DC-like output. For example, the rectified single-phase output significantly differs from a constant DC output. In contrast, the results obtained from the superposition of three signals having regularly shifted phases of 0, +120, and +240 degrees are close to a constant DC output. Therefore, the proposed MP-TENG, which has an output that is the superposition of various signals with shifted phases can potentially produce a constant DC output and is the subject of investigation in this article.

The performance of TENGs is strongly influenced by the choice of their friction materials. Conventional freestanding rotation-type TENGs utilize a rotator structure that consists of uniformly separated radially-arrayed strips of a single friction material.18 In contrast, here we utilize a rotator structure composed of strips of two friction materials, i.e. PTFE and MC nylon, which are radially interlaid between each other (see Fig. 1a). Upon triboelectrification, the two different parts of the rotator are therefore charged with opposite surface charges. Such a presence of both positive and negative triboelectric charges on the rotator surface can potentially enhance both the triboelectric potential and the charge/current transfer between the back electrodes, thereby enhancing the TENG's performance; Supplementary note 1 and Fig. S2 (ESI) discuss in detail the enhancement of the TENG's performance in a rotator composed of two different materials compared to that composed of a single material.

The MP-TENG design is intended to produce a constant DC-like output, i.e. having a low crest factor, for high performance energy storage using batteries/capacitors. In order to demonstrate the practical effectiveness of the design concept, the MP-TENG was operated at 600 rpm and its performance was characterized. Fig. 1c, d and Fig. S3 (ESI) show the corresponding output voltage, and output current, and capacitor charging performance of the MP-TENG for single, 3, and 5 phases. Fig. S4 (ESI) describes the characterization setup of MP-TENGs with single, 3, and 5 phases. The output voltage and current of the single phase TENG is simply a rectified AC signal. However, with an increase in the number of phases from one to three, both the peak values and minimum values of the output current and voltage increase; the increase in the output performance from single to multiple phases is due to the segmentation resulting in the superposition of the multiple individual electrical output signals. For example, the 3-phase TENG has a maximum voltage and current of ∼310 V and ∼91 μA m−2, respectively, and a minimum voltage and current of ∼105 V and ∼34 μA m−2, respectively. On the other hand, the 5-phase TENG has a maximum voltage and current of ∼410 V and ∼135 μA m−2, respectively, and a minimum voltage and current of ∼318 V and ∼84 μA m−2, respectively (see Fig. 1c and d). Most importantly, the current crest factor reduced from 2.01 for the single phase TENG to 1.43 for the 3-phase TENG and to 1.26 for the 5-phase TENG. In order to verify the corresponding enhancement in charging efficiency, the MP-TENGs were utilized to charge a 10 μF capacitor (see Fig. S3, ESI). The 5-phase TENG charged the capacitor to a 4.8 times higher voltage than the single phase TENG; the voltages on the capacitor for the single, 3-phase, and 5-phase TENGs are 2.4 V, 4.1 V, and 11.6 V, respectively. In summary, due to a constant DC-like output voltage, the proposed MP-TENG can greatly enhance the charging performance of the TENGs.

Working mechanism of the MP-TENG

Typically, the working mechanism of the MP-TENG is based on the coupling of triboelectrification and electrostatic induction. The triboelectric charges on the surface of the rotator electrostatically induce the opposite charges on the electrodes, thereby causing charges among the electrodes. However, different phases occur because every electrode has a regularly different PTFE and MC nylon contact ratio. Consequently, each electrode undergoes a regularly different electrostatic induction and the corresponding electrical current therefore has regularly different phases. An example is shown in Fig. 2, which schematically describes the working mechanism of a 5-phase TENG, where a single cycle of rotation is divided into five parts in order to describe the 5-phases. Fig. 2a–e show the charge distribution at the five stages and Fig. 2f shows the induced charge on each electrode as a function of the angle of rotation. Initially, it is assumed that the PTFE, MC nylon, and PTFE freestanding layers are fully charged due to triboelectrification; while the triboelectric charge on the PTFE freestanding layer is omitted here, Fig. S2 (ESI) and Supplementary note 1 describe it in detail. In the initial state (see Fig. 2a), only electrode 1 is fully facing the PTFE and has positive induced charges on its surface in order to screen the negative surface charges on PTFE. After a rotation of θs (θs = 72°), electrode 2 fully faces the PTFE (see Fig. 2b), which induces positive charges on its surface, thereby producing a peak of the charge on the surface, as shown in Fig. 2f. Furthermore, as the rotator rotates by 2θs, 3θs, and 4θs, electrodes 3, 4, and 5, respectively, begin to fully face the PTFE (see Fig. 2c–e, respectively) and the peaks of the induced charge are produced on the electrodes as shown in Fig. 2f. Therefore, although the electrodes have different induced charge distributions at any single time, every electrode shows the same periodic behavior. Besides, one cycle of rotation generates two cycles of alternate charges on each electrode. Ideally, the alternating peaks of the charge distributions should be identical as shown in Fig. S5 (ESI). However, the second peak is considered to be smaller than the first peak due to factors such as axial mismatch, surface flatness, etc. The axial mismatch and flatness of the surface indeed cause a decrease in the friction/contact areas, which results in a decrease of the triboelectric surface charges. Nonetheless, the electrical output currents, which are the time derivatives of the induced charge on the electrodes, with different phases, produce an almost DC-like output of the MP-TENG. To emphasize the importance of DC rather than AC, the MP-TENG and conventional rotation type AC-TENG are used to demonstrate the continuous operation of LEDs. The MP-TENG turns on the LEDs that are aligned in the ‘DC’ shape and the AC-TENG turns on the LEDs that are aligned in the ‘AC’ shape (see the ESI). Both TENGs operate at 200 rpm at the same time, but only the ‘AC’ shape LEDs flicker and the MP-TENG continuously provides energy to the ‘DC’ shape LEDs (see Fig. 2g).
image file: c8ee00188j-f2.tif
Fig. 2 Working mechanism of the multi-phase triboelectric nanogenerator. (a) Initial and final states of the rotator in which electrode 1 fully faces the PTFE. (b) Second state of the rotator in which electrode 2 fully faces the PTFE. (c) Third state of the rotator in which electrode 3 fully faces the PTFE. (d) Fourth state of the rotator in which electrode 4 fully faces the PTFE. (e) Fifth state of the rotator in which electrode 5 fully faces the PTFE. (f) Schematic graph of the triboelectric induced charge depending on the rotating angle of the rotator. (g) Video snapshots of the DC and AC TENGs’ performance comparison.

Performance optimization and applications

Circular segmentation has previously been proposed in order to enhance the performance of rotation-type TENGs.25 Therefore, the segmentation of each electrode into a number of segments can influence the output of the TENG. Also, the speed of rotation is another crucial parameter that can strongly influence the output performance of the TENG. Here, we thoroughly investigate the effect of both the electrode segmentation and the speed of rotation on the output performance of the MP-TENG. To confirm the segmentation and rotation speed effect of the 5-phase TENG performance, output voltage measurement at a load resistance of 10 MΩ and current measurement at short circuit were performed with various segments from 1 to 18 at 240, 600, and 920 rpm. As demonstrated in Fig. S6–S8 (ESI), the output current and power increase as the speed of revolution is increased from 240 rpm to 920 rpm. Also, in terms of the effect of the electrode segmentation, both the output voltage and output current were increased from 191 V and 255 μA m−2, respectively in the 1 segment device to 380 V and 3.6 mA m−2, respectively in the 18 segment device at 920 rpm (see Fig. 3a). Fig. 3b shows the output current and power of the 18 segment 5-phase TENG at 920 rpm as a function of load resistance. The TENG had a maximum power of 4.9 W m−2 at a matched load resistance of 1 MΩ. Compared with the other segmentation MP-TENGs, the matched load resistance had a reverse proportional relationship with the number of segments of the TENG of 50 MΩ to 1 MΩ (see Fig. S6–S8, ESI).
image file: c8ee00188j-f3.tif
Fig. 3 5-phase TENG output performance and demonstration. (a) Output performance of the 5-phase TENG in terms of output voltage and output current. (b) Output current and power performance of the 18 segmentation 5-phase TENG as a function of the external load resistance. (c) Demonstration of the MP-TENG to periodically charge a wearable electronic device.

To demonstrate the possibility of the MP-TENG as a power source of portable electronics, it was not directly connected to the battery of an electronic device in order to prevent electrical damage. We utilized a commercial power management product, the EH-301 module (Advanced Linear Devices, Inc.), to manage the MP-TENG. Fig. S9 (ESI) describes the setup for charging a wearable electronic device. The EH-301 module, which has an integrated 1 mF capacitor and an automatic switch, manages the capacitor voltage by discharging the capacitor from 5.2 V to 3.1 V and charging from 3.1 V to 5.2 V. Using the EH-301 module, the TENG required 75 s to charge a 1 mF capacitor to 5.2 V and transferred energy to the wearable electronic device, until the capacitor voltage reduced to a level of 3.1 V (see Fig. 3c). The equivalent galvanostatic current (Ieg) can be calculated to be 69.76 μA (see the ESI).16 After the initial process, this system needed 29 s to be ready to periodically charge the wearable electronic device for 3 s. Finally, this system successfully demonstrated periodic charging of the wearable electronic device and was able to turn on the battery charging indicator of the wearable electronic device.

A turbine is efficient in converting linear motion into rotary motion, which can extend the availability of the MP-TENG to the power of sensor network systems. A building's air circulation system, which always produces one way airflow, is one of the practical places to realize self-powered sensors (see Fig. 4a). The always-running MP-TENG can supply enough energy to drive the sensors, and the controller controls the indoor environment based on periodically communicated sensor information. A temperature sensor (MCP 9800, Microchip Technology Inc.) requires a lot of current value in active mode, so capacitors and a commercial power management integrated circuit (PMIC) (DC2151A, Linear technology) are utilized for providing boosted output current up to at 4.5 V (see Fig. 4b). For practical usage, nine miniaturized MP-TENGs with a radius of 3 cm and a thickness of 1 mm are stacked and are operated at 200 rpm. After the initial process of PMIC operation, this system needs 800 s to be ready to periodically measure and display the temperature, 22.87 °C (see Fig. 4c and Fig. S10 ESI, see the Supplementary movie 2, ESI). In addition, as temporary energy storage for the MP-TENG, a lithium (Li)-ion battery that can supply energy to commercial electronic devices is suitable. To demonstrate the possibility of recharging a commercial Li-ion battery, a thin-film Li-ion battery (EFL700A39, STMicroelectronics) is fully recharged from 3.7 V to 4.2 V by the MP-TENGs within 4.2 h at 400 rpm (see Fig. 4d). It is also demonstrated that a portable battery is charged in Supplementary movie 3 (ESI). The successful charging of the wearable electronic device, operating the temperature sensor, and charging the battery demonstrate that, due to its DC-like output, the MP-TENG provides an excellent opportunity for powering commercial electronics for self-powered applications.

image file: c8ee00188j-f4.tif
Fig. 4 Applications of the MP-TENGs at low rpm. (a) Schematic image of the self-powered smart feedback system. (b) Circuit diagram of the self-powered sensor system. (c) Demonstration of a self-powered temperature sensor. (d) Voltage profile of Li-ion battery charging by the MP-TENG.


We demonstrated a new solution for the outputs of TENGs which are typically in the form of short duration pulses. We developed a MP-TENG that produces an almost constant DC-like output with a low crest factor. In the design, the rotator of the MP-TENG consists of alternating components of PTFE and MC nylon. On the other hand, the stator consists of PTFE as the complementary friction layer with multiple back electrodes. The output currents from the multiple electrodes have regularly shifted phases, but have the same frequency and amplitude. Consequently, once the outputs are superimposed by the full-wave bridge rectifier, the total output of the MP-TENG is almost a constant DC-like output. Therefore, the 5-phase TENG effectively converted rotation energy into electric energy with a significantly low current crest factor of 1.26; it therefore demonstrated a significantly higher average power output than the typical TENG modes. Also, due to the almost constant DC output, the charging performance of the MP-TENG can be 4.7 times higher than that of a single-phase TENG. The greatly enhanced output of the TENG can reach 380 V, 3.6 mA m−2, and 4.9 W m−2 by increasing the segmentation of each electrode and the speed of rotation. The effectiveness of the MP-TENG for powering commercial electronics and charging a battery was successfully demonstrated. In summary, our results suggest a new type of TENG for self-powered portable electronics and smart devices.

Conflicts of interest

H. R., J. H. L., S. S. K. and S.-W. K. are the co-inventors of the patent entitled ‘Energy harvester for generating frictional electric of polyphase’ patent number: 10-1705974, Republic of Korea. The remaining authors declare no competing financial interests.


This work was financially supported by the Basic Science Research Program (2018R1A2A1A19021947) through the National Research Foundation (NRF) of Korea Grant funded by the Ministry of Science and ICT and by the Technology Innovation Program (10052668, Development of wearable self-powered energy source and low-power wireless communication system for a pacemaker) funded by the Ministry of Trade, Industry and Energy (MOTIE, Korea).


  1. W. Wu, L. Wang, Y. Li, F. Zhang, L. Lin, S. Niu, D. Chenet, X. Zhang, Y. Hao, T. F. Heinz, J. Hone and Z. L. Wang, Nature, 2014, 514, 470–474 CrossRef CAS PubMed .
  2. Y. Qin, X. Wang and Z. L. Wang, Nature, 2008, 451, 809–813 CrossRef CAS PubMed .
  3. U. Khan, T.-H. Kim, K. H. Lee, J.-H. Lee, H.-J. Yoon, R. Bhatia, I. Sameera, W. Seung, H. Ryu, C. Falconi and S.-W. Kim, Nano Energy, 2015, 17, 356–365 CrossRef CAS .
  4. J. Chen, Y. Huang, N. Zhang, H. Zou, R. Liu, C. Tao, X. Fan and Z. L. Wang, Nat. Energy, 2016, 1, 16138 CrossRef CAS .
  5. J. H. Lee, R. Hinchet, T. Y. Kim, H. Ryu, W. Seung, H. J. Yoon and S. W. Kim, Adv. Mater., 2015, 27, 5553–5558 CrossRef CAS PubMed .
  6. S. S. Kwak, S. Lin, J. H. Lee, H. Ryu, T. Y. Kim, H. Zhong, H. Chen and S. W. Kim, ACS Nano, 2016, 10, 7297–7302 CrossRef CAS PubMed .
  7. S. Kim, M. K. Gupta, K. Y. Lee, A. Sohn, T. Y. Kim, K. S. Shin, D. Kim, S. K. Kim, K. H. Lee, H. J. Shin, D. W. Kim and S. W. Kim, Adv. Mater., 2014, 26, 3918–3925 CrossRef CAS PubMed .
  8. H. Ryu, J.-H. Lee, T.-Y. Kim, U. Khan, J. H. Lee, S. S. Kwak, H.-J. Yoon and S.-W. Kim, Adv. Energy Mater., 2017, 7, 1700289 CrossRef .
  9. J. Wang, C. Wu, Y. Dai, Z. Zhao, A. Wang, T. Zhang and Z. L. Wang, Nat. Commun., 2017, 8, 88 CrossRef PubMed .
  10. R. Yang, Y. Qin, L. Dai and Z. L. Wang, Nat. Nanotechnol., 2009, 4, 34–39 CrossRef CAS PubMed .
  11. J.-H. Lee, H.-J. Yoon, T. Y. Kim, M. K. Gupta, J. H. Lee, W. Seung, H. Ryu and S.-W. Kim, Adv. Funct. Mater., 2015, 25, 3203–3209 CrossRef CAS .
  12. J. H. Lee, J. Y. Park, E. B. Cho, T. Y. Kim, S. A. Han, T. H. Kim, Y. Liu, S. K. Kim, C. J. Roh, H. J. Yoon, H. Ryu, W. Seung, J. S. Lee, J. Lee and S. W. Kim, Adv. Mater., 2017, 29, 1606667 CrossRef PubMed .
  13. S. Wang, Z. L. Wang and Y. Yang, Adv. Mater., 2016, 28, 2881–2887 CrossRef CAS PubMed .
  14. J.-H. Lee, H. Ryu, T.-Y. Kim, S.-S. Kwak, H.-J. Yoon, T.-H. Kim, W. Seung and S.-W. Kim, Adv. Energy Mater., 2015, 5, 1500704 CrossRef .
  15. J. H. Lee, K. Y. Lee, M. K. Gupta, T. Y. Kim, D. Y. Lee, J. Oh, C. Ryu, W. J. Yoo, C. Y. Kang, S. J. Yoon, J. B. Yoo and S. W. Kim, Adv. Mater., 2014, 26, 765–769 CrossRef CAS PubMed .
  16. S. Niu, X. Wang, F. Yi, Y. S. Zhou and Z. L. Wang, Nat. Commun., 2015, 6, 8975 CrossRef CAS PubMed .
  17. Y. Zi, J. Wang, S. Wang, S. Li, Z. Wen, H. Guo and Z. L. Wang, Nat. Commun., 2016, 7, 10987 CrossRef CAS PubMed .
  18. J. Chun, B. U. Ye, J. W. Lee, D. Choi, C. Y. Kang, S. W. Kim, Z. L. Wang and J. M. Baik, Nat. Commun., 2016, 7, 12985 CrossRef CAS PubMed .
  19. W. Seung, H.-J. Yoon, T. Y. Kim, H. Ryu, J. Kim, J.-H. Lee, J. H. Lee, S. Kim, Y. K. Park, Y. J. Park and S.-W. Kim, Adv. Energy Mater., 2017, 7, 1600988 CrossRef .
  20. J. H. Lee, R. Hinchet, S. K. Kim, S. Kim and S.-W. Kim, Energy Environ. Sci., 2015, 8, 3605–3613 CAS .
  21. A. Mansoor, W. M. Grady, R. S. Thallam, M. T. Doyle, S. D. Krein and M. J. Samotyj, IEEE Trans. Power Delivery, 1995, 10, 1416 CrossRef .
  22. G. Zhu, J. Chen, T. Zhang, Q. Jing and Z. L. Wang, Nat. Commun., 2014, 5, 3426 Search PubMed .
  23. Y. Yang, H. Zhang and Z. L. Wang, Adv. Funct. Mater., 2014, 24, 3745–3750 CrossRef CAS .
  24. C. Zhang, T. Zhou, W. Tang, C. Han, L. Zhang and Z. L. Wang, Adv. Energy Mater., 2014, 4, 1301798 CrossRef .
  25. T. Jiang, X. Chen, C. B. Han, W. Tang and Z. L. Wang, Adv. Funct. Mater., 2015, 25, 2928–2938 CrossRef CAS .


Electronic supplementary information (ESI) available. See DOI: 10.1039/c8ee00188j
Hanjun Ryu and Jeong Hwan Lee contributed equally to this work.

This journal is © The Royal Society of Chemistry 2018