# ADVANCED MATERIALS FOR SUPERCAPACITORS

EDITED BY : Wenyao Li, Min Zeng, Yuanlong Shao and Xiulin Fan PUBLISHED IN : Frontiers in Chemistry and Frontiers in Materials

#### Frontiers eBook Copyright Statement

The copyright in the text of individual articles in this eBook is the property of their respective authors or their respective institutions or funders. The copyright in graphics and images within each article may be subject to copyright of other parties. In both cases this is subject to a license granted to Frontiers. The compilation of articles constituting this eBook is the property of Frontiers.

Each article within this eBook, and the eBook itself, are published under the most recent version of the Creative Commons CC-BY licence. The version current at the date of publication of this eBook is CC-BY 4.0. If the CC-BY licence is updated, the licence granted by Frontiers is automatically updated to the new version.

When exercising any right under the CC-BY licence, Frontiers must be attributed as the original publisher of the article or eBook, as applicable.

Authors have the responsibility of ensuring that any graphics or other materials which are the property of others may be included in the CC-BY licence, but this should be checked before relying on the CC-BY licence to reproduce those materials. Any copyright notices relating to those materials must be complied with.

Copyright and source acknowledgement notices may not be removed and must be displayed in any copy, derivative work or partial copy which includes the elements in question.

All copyright, and all rights therein, are protected by national and international copyright laws. The above represents a summary only. For further information please read Frontiers' Conditions for Website Use and Copyright Statement, and the applicable CC-BY licence.

ISSN 1664-8714 ISBN 978-2-88963-893-2 DOI 10.3389/978-2-88963-893-2

#### About Frontiers

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

#### Frontiers Journal Series

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

#### Dedication to Quality

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

#### What are Frontiers Research Topics?

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

# ADVANCED MATERIALS FOR SUPERCAPACITORS

Topic Editors:

Wenyao Li, University College London, United Kingdom Min Zeng, Lanzhou Institute of Chemical Physics (CAS), China Yuanlong Shao, Soochow University, China Xiulin Fan, University of Maryland, College Park, United States

Citation: Li, W., Zeng, M., Shao, Y., Fan, X., eds. (2020). Advanced Materials for Supercapacitors. Lausanne: Frontiers Media SA. doi: 10.3389/978-2-88963-893-2

# Table of Contents

*05 Performance of Na-ion Supercapacitors Under Non-ambient Conditions—From Temperature to Magnetic Field Dependent Variation in Specific Capacitance*

Sudipta Biswas, Ananya Chowdhury and Amreesh Chandra


Qilong Ren, Guangyu Wu, Weinan Xing, Jiangang Han, Pingping Li, Bo Li, Junye Cheng, Shuilin Wu, Rujia Zou and Junqing Hu


Xuan Zheng, Guangjin Wang, Fei Huang, Hai Liu, Chunli Gong, Sheng Wen, Yuanqiang Hu, Genwen Zheng and Dongchu Chen

*76 Facile Synthesis of Novel V0.13Mo0.87O2.935 Nanowires With High-Rate Supercapacitive Performance*

Haishun Jiang, Wenjing Sun, Wenyao Li, Zhe Wang, Xiying Zhou, Zexing Wu and Jinbo Bai

*83 Fabrication and Electrochemical Performance of Al-Doped ZnO Nanosheets on Graphene-Based Flexible Substrates*

Qi Yu, Ping Rong, Shuai Ren, Liyun Jiang and Yapeng Li

*89 Three-Dimensional Graphene-Based Composite Hydrogel Materials for Flexible Supercapacitor Electrodes*

Enping Lai, Xinxia Yue, Wan'e Ning, Jiwei Huang, Xinlong Ling and Haitao Lin

*94 Capacity Contribution Induced by Pseudo-Capacitance Adsorption Mechanism of Anode Carbonaceous Materials Applied in Potassium-ion Battery*

Jiahao Liu, Ziqiang Xu, Mengqiang Wu, Yuesheng Wang and Zaghib Karim

*100 N-Propyl-N-Methylpyrrolidinium Difluoro(oxalato)borate as a Novel Electrolyte for High-Voltage Supercapacitor*

Weili Zhang, Fuming Zhang, Peng Zhang, Shuo Liang and Zhiqiang Shi

*108 Hollow Co3O4@MnO2 Cubic Derived From ZIF-67@Mn-ZIF as Electrode Materials for Supercapacitors*

Jiani Xu, Chaoting Xu, Yanhong Zhao, Jianghong Wu and Junqing Hu

*114 Nitrogen and Phosphorus Co-doped Porous Carbon for High-Performance Supercapacitors*

Jiaming Zhou, Shewen Ye, Qinqin Zeng, Hui Yang, Jiahao Chen, Ziting Guo, Honghui Jiang and Karthikeyan Rajan

# Performance of Na-ion Supercapacitors Under Non-ambient Conditions—From Temperature to Magnetic Field Dependent Variation in Specific Capacitance

Sudipta Biswas, Ananya Chowdhury and Amreesh Chandra\*

*Department of Physics, Indian Institute of Technology Kharagpur, Kharagpur, India*

Single phase NaFePO<sup>4</sup> can works as economically viable cathode material for Na-systems similar to LiFePO4–a material that led to the commercialization of Li-ion based energy systems. The reported microstructures of hollow NaFePO<sup>4</sup> particles, with porous walls, establish their advantages over solid morphologies. The hollow structures deliver stable electrochemical specific capacitance of 115 F g−<sup>1</sup> in 2 M NaOH electrolyte, over a large number of cycles. This observation is directly attributed to the increased surface area, transport channels and redox sites, which become available in the porous-hollow particles. Hitherto unreported electrochemical performance under non-ambient environment is also discussed. In contrast to recently reported in Fe-based metal oxides, where significant change in specific capacitance has been reported as a function of magnetic field, it is observed that NaFePO<sup>4</sup> can protect itself and suppress modifications. More importantly, NaFePO<sup>4</sup> can work as an efficient electrode material in the temperature range RT to 65◦C, which makes it useful for automotive industry.

Keywords: electrochemical, non-ambient conditions, hollow structure, energy storage, supercapacitor

# INTRODUCTION

The rapidly expanding consumer market of mobile and wearable electronics is driving the research for supplementary energy storage systems, which can complement or even compete with Li-ion technologies. A recent statistical study has predicted that the energy storage capacity required only by mobile technologies like phone, tablets and laptops will be more than 25 GW h by 2025<sup>1</sup> . For the major part of the last 4 decades, Na-ion based energy storage systems remained under the shadow of their more illustrious counterparts based on Li-ion (Zhao et al., 2011). Infact, careful examination of literature shows that Na-ion energy storage systems (ESS) were actually investigated before Li-ion technologies. In 1960s, Kummer and Weber of Ford Motor Co extensively investigated β-Al2O<sup>3</sup> Na+ion conducting solid electrolyte based battery, which eventually led to the development of sodium/sulfur technologies<sup>2</sup> . Even after large-scale commercialization of Li-batteries, it was always known that sodium, with similar oxidation state and electronic configuration, can deliver characteristics similar Li<sup>+</sup> based energy storage devices. The re-emergence of interests in Na-ion energy storage devices is also linked to the constraints associated with most Li-ion based materials/technologies (Zhang et al., 2017). These range from

#### Edited by:

*Yuanlong Shao, King Abdullah University of Science and Technology, Saudi Arabia*

#### Reviewed by:

*Liang Zhou, Wuhan University of Technology, China Wei Kong Pang, University of Wollongong, Australia*

\*Correspondence:

*Amreesh Chandra achandra@phy.iitkgp.ac.in*

#### Specialty section:

*This article was submitted to Energy Materials, a section of the journal Frontiers in Materials*

Received: *15 January 2019* Accepted: *18 March 2019* Published: *09 April 2019*

#### Citation:

*Biswas S, Chowdhury A and Chandra A (2019) Performance of Na-ion Supercapacitors Under Non-ambient Conditions—From Temperature to Magnetic Field Dependent Variation in Specific Capacitance. Front. Mater. 6:54. doi: 10.3389/fmats.2019.00054*

**5**

<sup>1</sup>http://energystorage.org/system/files/attachments/esa\_vision\_2025\_final.pdf 06-09-2018

<sup>2</sup>https://patentimages.storage.googleapis.com/1a/51/27/63e4b6a1229db4/US3413150.pdf 06-09-2018

geo-economical/political, availability, increasing demand, environmental impacts, and countering the IP protection.

The primary reason, which restricted the use of Na-ion based technologies, was the limited number of materials that could handle the chemistry of bigger ion like Na+. The growing understanding and expertise to develop novel nanoparticles ranging from solid, hierarchical to hollow, has allowed the resurgence of Na-ion based energy storage systems (Yabuuchi et al., 2014; Fang et al., 2015; de la Llave et al., 2016; Guo et al., 2016; Wang et al., 2019; Zhu et al., 2019). Different sodium based materials viz. Na4Ti5O12, NaTiO2, Na2Ti3O7, Na3V2(PO4)3, Na4Fe(CN)6, NaVO2, etc., are now being investigated (Didier et al., 2011; Yin et al., 2012; Li et al., 2015; Guo et al., 2016),(Jiang et al., 2016).

It is well known that the discovery and fabrication of LiFePO<sup>4</sup> led to rapid commercialization of the Li-ion batteries (Liang et al., 2015). Similar to it, NaFePO<sup>4</sup> has been suggested as a promising cathode material for Na-ion batteries (Liu et al., 2018). The major problem associated with this material is related to the difficulty of synthesizing it in single phase. Recently, there have been reports that have presented synthesis protocols, which have led to single phase NaFePO<sup>4</sup> with simple morphologies and broad particle size

NaFePO4.

distributions (Minakshi et al., 2016; Rahman et al., 2017). The electrochemical response of these particles has been investigated only in batteries (Park et al., 2013; Slater et al., 2013). Typical energy densities that have been reported are ∼210–600 W h kg−<sup>1</sup> . Intriguingly, its capacitive behavior remains ignored or mostly limited (Lu et al., 2015; Lim et al., 2016; Zhang et al., 2017; Ramakrishnan et al., 2018). For application in supercapacitors, tuning of morphology, surface area, pore size/dimension, and redox sites becomes essential.

In this paper, it is established that NaFePO<sup>4</sup> can also be used as efficient electrode material in supercapacitors, provided the morphology is carefully tuned. The results clearly prove that the hollow particles of NaFePO<sup>4</sup> have much higher capacitive behavior, in comparison to the solid counterparts. Superior performance of hollow structures can be attributed to the availability of more channels/sites for redox activities, while the enhanced surface area enforces increased contribution from the pseudo-capacitance (Sharma et al., 2018c).

Unlike the common strategy of reporting the supercapacitor response only at ambient conditions, the paper also discusses the behavior under non-ambient conditions. It is observed that NaFePO<sup>4</sup> can be easily used at elevated temperatures, which is essential if they are to be utilized in hybrid vehicles. The specific capacitance remains nearly same, even when the operating temperature is increased from RT to 65◦C. Infact, slight increase in the specific capacitance is observed at elevated temperatures. The observation can be explained in terms of improved diffusion and accessibility of inner core for electrolyte ions insertion/de-insertion.

Very recently, few papers have indicated that Fe-based metal oxides used in supercapacitors will have to thoroughly re-characterized near magnetic field. Significant variation in specific capacitance can occur in such oxides as a function of varying dc-magnetic field, owing to modulated electron flow caused by the Lorentz force (Sharma et al., 2018a). It is demonstrated here that NaFePO<sup>4</sup> does show variation under magnetic field but the changes are much lower than that observed in pure Fe2O3. Therefore, NaFePO<sup>4</sup> based Na-ion capacitors can also be used near magnetic without disturbing the associated electronic circuitry.

#### EXPERIMENTAL

#### Material Used

Ferric nitrate nonahydrate (Fe(NO3)3, 9H2O), stearic acid, ammonium dihydrogen phosphate ((NH4)H2PO4), tri-sodium citrate di-hydrate (Na3C6H5O7, 2H2O), citric acid, ethylene glycol, and sodium nitrate (NaNO3) were procured from LobaChemie Pvt. Ltd. (India) and MerckSpecialities Pvt. Ltd. (India). All the analytical grade precursors were used directly without further purification.

# Material Synthesis

#### Synthesis of Hollow and Porous structures

Porous and hollow NaFePO<sup>4</sup> microstructures were synthesized using one pot facile hydrothermal route followed by calcination in air. In a typical experimental procedure, 25 ml of 0.1 M ferric nitrate solution was mixed with 25 ml of 0.1 M stearic acid

NaFePO4 in 2 M NaOH electrolyte.

solution. Subsequently, 245.1 mg trisodium citrate (Na3C6H5O7, 2H2O) was added and the solution was stirred for 2 h. Appropriate amount of ammonium dihydrogen phosphate ((NH4)H2PO4) was added to the precursor solution so as to ensure Na:Fe:PO<sup>4</sup> concentration ratio was 1:1:1. 50 ml of this yellow colored solution was then transferred to a teflonlined stainless steel (capacity 250 ml) autoclave. The autoclave was kept at 180◦C for 24 h, before allowing it to slowly cool down to the room temperature. The precipitate was collected by centrifugation at 3,200 rpm. The precipitate was subsequentially washed three times using de-ionized water and dried overnight in a vacuum oven at 70◦C. Dried sample was crushed and annealed at 600◦C for 4 h in air to obtain hollow NaFePO<sup>4</sup> powder. To compare the performance of these hollow microstructures, solid nanoparticles of NaFePO<sup>4</sup> were also prepared. The corresponding synthesis protocol is given in the **Supplementary Information**.

#### Material Characterization

The phase formation of NaFePO<sup>4</sup> was confirmed by analyzing powder X-ray diffraction (XRD) profile obtained using a Rigaku MiniFlex600 diffractometer with Cu-Kα (λ = 0.15406 nm) as the incident wavelength.

For surface and particle morphological studies, both scanning and transmission electron micrographs were collected. For SEM data, CARL ZEISS SUPRA 40 SEM was used, while the TEM data were obtained utilizing a TEM FEI-TECHNAI G220S-Twin microscope operated at 200 kV. Brunauer-Emmett-Teller (BET) surface area and porosity were determined with a Quantachrome Nova Touch surface area and pore size analyzer. Zeta potential and particle size distribution were inferred by analyzing the DLS data obtained from a Horiba Scientific Nano Particle Analyzer SZ-100.

TABLE 1 | Specific capacitance of hollow and porous NaFePO4 form CV curves in 2M NaOH electrolyte.


TABLE 2 | Specific capacitance of hollow and porous NaFePO4 form CD curves in 2M NaOH electrolyte.


FTIR measurements were performed using Nexus 870 instrument in the range of 500–1,400 cm−<sup>1</sup> . The spectrum was acquired in transmittance mode with a resolution of 1 cm−<sup>1</sup> .

For the electrochemical characterization of the synthesized materials and fabricated devices, typical cyclic voltammetry (CV), and galvanostatic charge-discharge (GCD) were performed using the MetrohmAutolab (PGSTAT302N) potentiostatgalvanostat. The measurements were undertaken both in three- and two-electrode configurations, using an aqueous electrolyte. For impedance data of the electrochemical systems, in the frequency range of 50 mHz to 1 MHz, a N4L-PSM 1735 impedance analyzer was used.

### Electrode Preparation and Electrochemical Characterization

A slurry was prepared by mixing 80 wt % active material, 10 wt % activated charcoal, and 10 wt % polyvinylidenefluorideco-hexafluoropropylene (PVDF- HMP) using N-Methyl-2-pyrrolidone as the mixing media. All electrochemical measurements were performed in 1 M Na2SO4, 1 M NaNO3, and 1 M NaOH electrolyte solutions, which allowed determination of an optimum electrolyte. Best performance was observed in NaOH and the corresponding results are presented in the paper. Platinum wire and Ag/AgCl/3.0 M KCl were used as the counter and reference electrodes, respectively.

#### RESULT AND DISCUSSION

It is now well established that, for useful electrochemically active materials, parameters such as: phase, particle morphology/size along with the nature of pores and pore-size/volume are critical (Singh and Chandra, 2015; Akhtar et al., 2016; Sharma et al., 2018b). The XRD plot obtained shown in **Figure S1** could be indexed using the Pnmb space group of NaFePO<sup>4</sup> following the JCPDS card no. 04-012-9665 (Kosova et al., 2014). XRD pattern for solid particle are shown in **Figure S2**.

**Figures 1a,b** shows the SEM micrographs of NaFePO<sup>4</sup> observed at various magnifications. It is clear that particles, with a cavity in the middle were stabilizing. Recently, hollow nanostructures of metal oxides have been suggested as electrode materials for next generation supercapacitors (Sharma et al., 2018b; Wei et al., 2018). Till date, there have been no reports, which have suggested formation of hollow NaFePO<sup>4</sup> particles/microspheres. The size of the synthesized NaFePO<sup>4</sup> microspheres varied in the range 1–3µm. The micrographs also indicated that the wall surface of the hollow microspheres was not solid but porous. This can be an additional advantage of such particles because all the desired parameters i.e., surface area, pores, and transport channels, will be able to contribute in the final electrochemical reactions and/or specific capacitance. The particle growth mechanism is discussed in the supporting evidence (see **Figure S3**). The growth process was found to be a convoluted picture of reaction as well as diffusion kinetics. Corresponding SEM micrographs for solid NaFePO<sup>4</sup> structures are shown in **Figure S4** (see Supplementary Information).

The corresponding TEM images of the microspheres are given in supporting evidence (**Figure S5**). The associated elemental mapping and EDAX data for the NaFePO<sup>4</sup> microspheres are also described in the supporting evidence (**Figure S6**). The atomic ratio of Na:Fe:P:O was found to be 1.05:1:1.08:3.57, which confirmed the formation of NaFePO<sup>4</sup> with nominal composition and homogeneous distribution of the elements throughout the sample.

The surface area, pore size/volume and particle size distribution values are shown in **Figures 2A–D**. The BET adsorption-desorption curves showed typical type IV isotherms, indicating slit shaped mesopores of ∼6 nm. BET surface area was 20 m<sup>2</sup> g −1 .

Zeta potential denotes the electro kinetic potential in colloidal dispersions. In any chemical reaction, electro-positivity of the surface is determined by the overall pH of the solution. In the present studies, hollow nanostructures returned higher electropositivity. This can only happen when there is an increased capacity to facilitate OH<sup>−</sup> accommodation. This phenomenon helps to attract ions toward surface and hence increases the capacitance of the material. The surface charge (zeta potential) was −41.49 mV, as shown in **Figure 2D**.

It has been demonstrated that vibrational spectroscopy is very useful for probing fundamental sodium ion intercalation in a variety of crystalline sodium based electrode materials (Sharma et al., 2018c). The FTIR spectra of NaFePO<sup>4</sup> showed symmetric, asymmetric vibration as well as bending modes. The absorption bands at 1,009, 1,068, and 1,133 cm−<sup>1</sup> , shown in the **Figure S7a**, could be assigned to the ν<sup>3</sup> type asymmetric stretching modes between P-O in PO3<sup>−</sup> 4 . The ν1modes near 980 and 947 cm−<sup>1</sup> are attributed to the PO3<sup>−</sup> 4 intramolecular symmetric stretching vibrations (Sharma et al., 2018b). The asymmetric bending modes of the PO3<sup>−</sup> 4 anion, assigned as v4, were observed at 629, 577, and 540 cm−<sup>1</sup> (see **Figure S7b**).

Before starting exhaustive electrochemical characterizations, it is critical to optimize the operating potential window. CV studies were performed at a scan rate of 50 mV s−<sup>1</sup> in three different aqueous electrolytes viz., 1 M Na2SO4, NaNO3, and NaOH. The mass of the electrode was kept at ∼1 mg. The observed CV curves in different voltage windows are shown in **Figures S8a**–**c**. The analysis clearly indicated that the electrolytes: Na2SO<sup>4</sup> and NaNO<sup>3</sup> provided an electrochemical window ranging from −0.2 to 0.7 V. By using NaOH as electrolyte, voltage window of −0.3 to 0.5 V could be obtained. The CV curves as a function of varying scan rates, in the respective stabilized operational potential windows for the three electrolytes are shown in **Figures S8d**–**f**. After studying the electrochemical behavior of the hollow NaFePO<sup>4</sup> in three different electrolytes, NaOH was found to return the highest specific capacitance.

The estimated specific capacitance, with varying scan rates, is tabulated in **Table S1**. As expected, the specific capacitance showed appreciable dependence of the scan rates. The maximum specific capacitance, at 5 mV s−<sup>1</sup> scan rate, was ∼75 F g−<sup>1</sup> when 1 M NaOH electrolyte was used. When scan rate increased to 150 mV s−<sup>1</sup> it was 20 F g−<sup>1</sup> . The capacitance retention was found to be only ∼29%, when the scan rate was enhanced by 30 times.

The capacity to deliver the power of a commercial supercapacitor is obtained by the values of specific capacitance measured in galvanostatic charge discharge (CD) curves. The CD curves for the working electrode, in three different electrolytes are shown in **Figures S8g–i**.

The specific capacitance values, at different current densities, in three different electrolytes are listed in **Table S2**. The maximum capacitance of 87 F g−<sup>1</sup> was obtained by using NaOH as electrolyte, which corroborated the results obtained from CV measurements. At higher specific currents, the specific capacitance decreased. This happens due to the underutilization of bulk capacitance, which is a normal behavior for supercapacitor electrode materials. At higher specific currents, the transfer of electrons toward the electrode is faster and hence the increase of potential would be higher. Consequently, the electrode gets reduced time to stay at a certain voltage and lower specific capacitance value is observed.

The electrolyte concentration has major effect on the electrochemical performance of an electrode material and is linked to the material's storage capacity. In the present case, the optimized concentration was found to be ∼ 2 M (NaOH). Operational window of the material was rechecked in 2 M NaOH and corresponding data is shown in **Figure 3A**. CV curves, at different scan rates, were collected in 2 M NaOH electrolyte and are shown in **Figure 3B**. Maximum specific capacitance was found to be 108 F g−<sup>1</sup> at a scan rate of 5 mV s−<sup>1</sup> . In the same potential window, CD profiles were also collected at different current densities. These are depicted in **Figure 3C**. The maximum specific capacitance in this case was found to be ∼115 F g−<sup>1</sup> at 1 A g−<sup>1</sup> . The specific capacitance values for the porous and hollow NaFePO4, utilizing 2 M NaOH as electrolyte, at different scan rates and current densities are listed in **Tables 1**, **2**, respectively and corresponding graphs are shown in **Figure 3D**. It was also observed that the hollow structure based electrodes showed higher capacitance retention ability.

The performance of hollow structures was compared with that of NaFePO<sup>4</sup> solid particles. From the CV analysis, the maximum specific capacitance of 48 F g−<sup>1</sup> was observed in the case of solid particles. The CD profiles gave the specific capacitance to be 81 F g−<sup>1</sup> . Clearly, the hollow structures had much higher performance. The advantages of both outer/inner surfaces, low diffusion length, high availability of surface adsorption sites due to the hollow cavity of the particles would drive this enhanced performance. The CV and CD curves for solid structures are shown in **Figure S9** and the corresponding specific capacitance values are given in **Tables S3**, **S4**, respectively.

Cyclic stability of an electrode is one of the important property that decides its industrial application. In the present case, 93.7% of capacity retention was found after 1,500 cycles for hollow microsphere [see **Figure 4A**]. Solid structures also showed similar capacity retention after 1,500 cycles. This can be expected because the redox reactions would be similar as the chemical formula or composition was not varying as a function of particle morphology. The advantages rendered by hollow particles was leading to higher specific capacitance. Hollow and solid structures had BET surface areas of 20 and 6 m<sup>2</sup> g −1 , respectively. Hollow structures, with pore radius

∼5.7 nm, would also allow better intercalation/ de-intercalation of electrolyte ions than the solid particles that showed pores of ∼1.7 nm radius. The surface area, pore size/volume, particle size distribution, and zeta potential of solid particles are depicted in **Figures S10a–d**.

The charge transport kinetics of the electrode was studied by the analysis of electrochemical impedance spectroscopy (EIS) data. Nyquist plots give the information about electrode-electrolyte interactions and equivalent series resistance (ESR). The ESR values for hollow microsphere and solid nanostructures were ∼4.0 (**Figure 4B**) and ∼6.6 (**Figure S11**), respectively. This indicated low charge transfer resistance at the working electrode and electrolyte interface in the case of hollow structures. Capacity retention for

TABLE 3 | Values of specific capacitance at different scan rates at different temperature (screw cell device).


hollow structures can be justified by the ESR measurements. The values of ESR increased from ∼4.0 to 5.4 , in case of hollow NaFePO4. This increment in the ESR values can be attributed to the electrode degradation and reduction in the available ion channels within the materials as a function of cycling. It is clear from the lower frequency regions in **Figure 4B**, that the capacitive behavior was lower in case of hollow structures.

Supercapacitor coin cell type device was fabricated using activated carbon as the negative and NaFePO<sup>4</sup> as the positive electrode. Activated carbon was tested in three-electrode TABLE 4 | Values of specific capacitance at different current densities at different temperature (screw cell device).



configuration before making the device. **Figures 5A,B** shows the cyclic voltametric curves at different scan rates from 5 to 200 mV s −1 and charge discharge at different current density from 0.5 to 5 A g−<sup>1</sup> . From the CV results, the working electrochemical window was found to be −1 to 0 V.

TABLE 6 | Values of specific capacitance at different scan rate under varying magnetic field (in three electrode configuration).


Highest specific capacitance observed from the CV measurement was 120 F g−<sup>1</sup> , at a scan rate 5 mV s−<sup>1</sup> . Using charge-discharge measurements, it was estimated as 126 F g −1 at current density of 1 A g−<sup>1</sup> . Specific capacitance values from CV and CD for activated carbon are shown in **Tables S5**, **S6**, respectively.

Fabrication of asymmetric supercapacitor device needs charge balancing of the electrode materials. The optimal charge balance condition was estimated using the mass balance formula:

$$\frac{m\_+}{m\_-} = \frac{V\_-C\_-}{V\_+C\_+} \tag{1}$$

where C<sup>−</sup> and C<sup>+</sup> are the capacitances (in F g−<sup>1</sup> ) measured at the same scan rate, using the three electrode system, for negative and positive electrodes, respectively while 1V<sup>+</sup> and 1V<sup>−</sup> denote the working potential window for the positive and negative electrodes, respectively. The required mass ratio for positive and negative electrode materials (m+/m−) was thus estimated as 1.5 at 5 mV s−<sup>1</sup> .

**Figure 6** shows the cyclic voltammetry and charge discharge curves for device. Whatman glass fiber paper was used as separator (pre-soaked in electrolyte) and 2 M NaOH was used as aqueous electrolyte. The electrochemical voltage window was found to be from 0 to 1.3 V. No H2/O<sup>2</sup> evolution was discernible in this voltage window. From the CV measurements, maximum specific capacitance for the NaFePO4//AC device was ∼22 F g−<sup>1</sup> . The device performance was found to decrease at higher scan rate, as a result of under-utilization bulk of the material. The maximum specific capacitance observed from CD was 23 F g−<sup>1</sup> . Nearly linear charge discharge curves showed the dominance of the electric double layer capacitance (EDLC) in the device.

In most of the applications, there are few additional parameters, which contribute in determining the final electrochemical performance. These include: temperature, external frequency, magnetic field, etc., (Ramakrishnan et al., 2018; Sharma et al., 2018a). Till date, such studies have remained ignored in the Na-ion based supercapacitors. Given below are the performance of the NaFePO<sup>4</sup> based supercapacitors under variable temperature and magnetic field.

It has been suggested that, at elevated temperatures, contributions from additional factors such as ion conductivity, solubility limits, viscosity, thermal stability, etc., cannot be ignored. For investigating the thermal stability, the device was slowly heated at increments of 10◦C. The device was equilibrated for 30 min at a given temperature before the electrochemical measurements were performed. As presented above, the specific capacitance in NaFePO<sup>4</sup> based screw cell type asymmetric device was ∼23 F g−<sup>1</sup> . The CV and CD curves, with increasing temperature, are shown in **Figures S12**, **S13**. CV curves with increasing temperature at a scan rate of 5 and 50 mV s−<sup>1</sup> are shown in **Figures 7A,B**, respectively. Comparison of charge discharge curve at 1 A g−<sup>1</sup> with increase in temperature and variation of specific capacitance with scan rate is shown in **Figures 7C,D**, respectively. Clear signature was obtained, which suggested increase in the specific capacitance with increasing temperature. At 65◦C, the specific capacitance increased to 29 F g −1 , which was nearly 25% higher than the value obtained at room temperature. Additionally, the device continued to show the characteristics expected from a capacitive system. The corresponding values of specific capacitance from CV and CD analysis are listed in **Tables 3**, **4**, respectively. The intrinsic internal resistance decreased with increasing temperature and the values are given in **Table 5**.

On comparing with literature dealing with elevated temperature behavior in supercapacitors fabricated using metal oxides, it was interesting to note that the Na-ion based device showed much lower variation as a function of temperature. This can be directly correlated with the fact that the olivine structures show much lower thermal expansion than the other relevant metal oxides. Therefore, change of ∼50◦C will not lead to appreaciable change in the surface area or pore structure in the NaFePO<sup>4</sup> particles. It is known that, with increasing temperature, the reaction kinetics increases. Therefore, in the Na-based olivines, the increase in specific capacitance with increasing temperature will only be dominated by the enhanced electrochemical reactions near the electrolyte-electrode interface, which will vary too much in the temperature window of ∼40◦C.

It has been very recently reported that electrode materials based on ferromagentic ions may have to revisited because devices fabricated using them can have appreciable effect under/near variable magnetic field. NaFePO<sup>4</sup> also has Fe, which is a well-known ferromagnetic atom. The magnetic field dependent studies in NaFePO4, while it is being used as an electrode material in storage device, has never been reported. Cyclic voltammetry (CV) and galvanostatic charge-discharge (CD) measurements were performed under varying magnetic fields (B = 0, 10, 20, and 30 gauss). CV curves at different scan rate and at different current densities under varying magnetic fields (B = 0, 10, 20, and 30 gauss) are shown in **Figure 8** and **Figure S14**. Schematic of experimental arrangement for performing the magnetic field dependent measurements is shown in **Figure S15** of the supplementary information. The calculated specific capacitance values are listed in **Table 5**. It was clear that the specific capacitance showed an increasing trend, with increasing magnetic field. On reaching the magnetic field of 30 gauss, the specific capacitance at 5 mV s−<sup>1</sup> increased to 124 F g −1 from 109 F g−<sup>1</sup> , which was observed at 0 Gauss. This corresponded to a mere 14% change. Additionally, no change was observed even by further increasing the magnetic field. In comparison, nearly 75% increase has been reported in Fe2O<sup>3</sup> or MnO2. So, NaFePO<sup>4</sup> clearly showed stability near magnetic field, which would make this material even more important for industrial use. The values of specific capacitance at different scan rates is shown in **Table 6** and corresponding specific capacitance values from charge discharge is shown in **Table S7**.

One of the reasons behind the increase in capacitance response in magnetic field is magneto hydrodynamic (MHD) effect. This increases the magnetic current in the system because of the feedback mechanism and increase in vigorous hydrodynamic stirring. Further, the change of electronic energy state due to the applied magnetic field may also improve the performance of electrode material. The suppressed magnetic field dependence can be easily explained if we examine the expected valence state of the elements in NaFePO4. Fe is expected to be in a +2 state, which is known to be non-magnetic. If magnetism has to be induced in it then large magnetic field has to be applied as per the Van Vleck paramagnets (Smolenski et al., 2016). Additional, Na (valency +1) is a known diamagnetic. Therefore, it will oppose the realignment in transport channels, which may

occur owing to structural reorientation driven by the magnetic iron (Fe3+, which can also stabilize during material synthesis). To confirm the presence of ionized state of Fe analysis, XPS data was collected (shown in **Figure 9**) and analyzed. XPS result showed that the percentage of Fe2<sup>+</sup> and Fe3<sup>+</sup> was ∼55 and 45%, respectively, in NaFePO4. As Fe2<sup>+</sup> is nonmagnetic, Fe3<sup>+</sup> drives the response under external magnetic field. The effect is more pronounced in Fe2O<sup>3</sup> that had much higher percentage of Fe3<sup>+</sup> (70%) with respect to the Fe2<sup>+</sup> state (30%). This makes NaFePO<sup>4</sup> useful for application near the magnetic environment.

#### CONCLUSION

Single phase NaFePO<sup>4</sup> particles, with hollow cavity having porous walls, are reported. This microstructure allows efficient utilization of active surface area and pore structure, leading to high specific capacitance. Bulk structures return specific capacitance of ∼48 F g−<sup>1</sup> at a scan rate 5 mV s−<sup>1</sup> and 81 F g−<sup>1</sup> at a current density of 1 A g−<sup>1</sup> . For hollow microstructures structure, the specific capacitance is found to be 108 F g−<sup>1</sup> at a scan rate 5 mV s−<sup>1</sup> and 115 F g−<sup>1</sup> at a current density 1 A g−<sup>1</sup> . NaFePO<sup>4</sup> also shows long term stability under non-ambient conditions, where parameters such as magnetic field or temperature are varied. NaFePO<sup>4</sup> has the capacity to deliver performance similar to LiFePO<sup>4</sup> and make Na-ion based energy systems industrially and economically viable.

#### REFERENCES


#### AUTHOR CONTRIBUTIONS

SB and AnC have contributed equally in sample preparation and data analysis under direct supervision of AmC as part of their Ph.D. program. AmC has also contribute in data analysis and interpretation.

#### FUNDING

MES program of the Department of Science and Technology, India.

#### ACKNOWLEDGMENTS

The authors acknowledge the financial support received from DST (India) under the MES scheme to pursue work under the project entitled: Hierarchically nanostructures energy materials for next generation Na-ion based energy storage technologies and their use in renewable energy systems.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmats. 2019.00054/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Biswas, Chowdhury and Chandra. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Biomass-Derived Porous Carbon Materials for Supercapacitor

#### Hui Yang\* † , Shewen Ye† , Jiaming Zhou and Tongxiang Liang\*

School of Materials Science and Engineering, Jiangxi University of Science and Technology, Ganzhou, China

The fast consumption of fossil energy accompanied by the ever-worsening environment urge the development of a clean and novel energy storage system. As one of the most promising candidates, the supercapacitor owns unique advantages, and numerous electrodes materials have been exploited. Hence, biomass-derived porous carbon materials (BDPCs), at low cost, abundant and sustainable, with adjustable dimension, superb electrical conductivity, satisfactory specific surface area (SSA) and superior electrochemical stability have been attracting intense attention and highly trusted to be a capable candidate for supercapacitors. This review will highlight the recent lab-scale methods for preparing BDPCs, and analyze their effects on BDPCs' microstructure, electrical conductivity, chemical composition and electrochemical properties. Future research trends in this field also will be provided.

#### Edited by:

Wenyao Li, Shanghai University of Engineering Sciences, China

#### Reviewed by:

Kaibing Xu, Donghua University, China Guangjin Wang, Hubei Engineering University, China

\*Correspondence:

Hui Yang yanghui\_2521@163.com Tongxiang Liang liang\_tx@126.com

†Co-first authors

#### Specialty section:

This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry

Received: 01 March 2019 Accepted: 03 April 2019 Published: 24 April 2019

#### Citation:

Yang H, Ye S, Zhou J and Liang T (2019) Biomass-Derived Porous Carbon Materials for Supercapacitor. Front. Chem. 7:274. doi: 10.3389/fchem.2019.00274 Keywords: biomass, porous carbon, supercapacitor, electrochemical performance, energy storage

# INTRODUCTION

The ever-growing population, crisis of energy shortage, and environmental pollution caused by burning fossil fuels have brought about different kinds of problems, and they stimulate the process of developing clean and sustainable energy, such as solar energy, wind power etc. (Dunn et al., 2011; Chu and Majumdar, 2012; Dubal et al., 2015; Kamat, 2015; Kazmerski, 2016; Peng et al., 2017). All these closely relate to the development of the advanced energy storage system. Until now, for electrochemical energy storage systems, for example, various rechargeable batteries and supercapacitorsare identified as the most promising energy storage devices (Chmiola et al., 2006; Miller and Simon, 2008; Kim et al., 2010; Cheng et al., 2011; Xu et al., 2014; Gogotsi, 2015). Although batteries have a relatively longer life cycle, (<1,000 cycles), high-energy density and satisfactory rate performance, distinct from rechargeable batteries, supercapacitors possess the following characteristics: high-power energy, extremely long cycle life (>100,000 cycles) without losing capacitance and safety, it is quite essential to bridge the gap between conventional electrochemical capacitors and rechargeable batteries, and meet the demand of electric vehicles which require high power (Liu et al., 2018a; Xu et al., 2018).

Basically, supercapacitors can be classified under two categories according to their charge storage mechanism, which further affects the power density of supercapacitors. The electrical double layer (EDL) capacitance and pseudo-capacitance are quite different. For the former, charges are stored through ions' adsorption by electro-static interaction at the near-surface of active materials, where EDL forms; while the energy storage/release mechanism for pseudo-capacitance is by means of redox reaction, ions insertion/extrusion and under-potential depositions (Augustyn et al., 2014; Mccloskey, 2015). Both the aforementioned supercapacitors require chemical and structure stability, outstanding electrical conductivity, and high SSA. Hence, to satisfy the requirement of high capacity and superior rate performance for supercapacitors, the exploitation of advanced

**16**

electrode materials with low cost, adjustable composition, and microstructure (especially pore structure) turns increasingly important.

In the past decades, thanks to the excellent conductivity property, high SSA, variable pore structure and porosity, carbon materials, such as commercial activated carbons (Chan et al., 2004; Jänes et al., 2007; Zhang and Zhao, 2009; Lv et al., 2011; Wang et al., 2014), graphene (Wang et al., 2009; Liu et al., 2010; Zhang et al., 2010; Cao et al., 2011), graphene oxide (GO) (Li et al., 2013; Li and Yang, 2014; Down et al., 2018), carbon nanotubes (CNTs) (An et al., 2001; Futaba et al., 2006; Pushparaj et al., 2007), porous carbon (PC) (Lee et al., 2000; Vix-Guterl et al., 2005; Hou et al., 2015) and their composites (Salunkhe et al., 2015) have achieved various kinds of essential accomplishments on their volumetric performance. However, their high production cost, complicated and unsustainable preparation process greatly limited their wide application. Besides, the gravimetric capacitance for CAs is keeping at a relatively low level for several decades, that should be ascribed to the rich micro-pores, which are hard accessible for ions, especially at high-current density condition (Li et al., 2014). Therefore, enlarging the ion-accessible area through creating a suitable hierarchical porous structure is a key factor for achieving high specific capacitance (Cs) and rate performance (Zhao et al., 2016). Biomass materials abundantly present across our surroundings, and their low cost, accessible, environmentally friendly, and recyclable properties ensure they are ideal candidates for resources of carbon materials. At the same time, their naturally hierarchical porous structure and various elements (N, S) facilitate electrolyte penetration and extra active sites' generation, respectively (Chen et al., 2016).

In this review, we provide a summary of the recent and significant advances in hierarchical porous carbon (HPC) derived from biomass materials, emphasizing on the relationship between interconnectivity of the pore structure and electrochemical efficiency, and synthetic strategies for preparing HPC as supercapacitors. Finally, the current challenges and future directions are briefly discussed.

#### RELATION BETWEEN PORE FEATURES AND CAPACITIES

In general, the capacitance of PC materials relies on the pore size distribution, connectivity and hydrophilicity; these factors will be further influencing the ion diffusion process and energy density of active materials. According to the findings by Dubinin (1960) and the classification by the International Union of Pure and Applied Chemistry (IUPAC) in 1985 (Sing, 1985), pores with different widths can be divided among three categories: macropores (>50 nm), mesopores (2–50 nm) and micropores (<2 nm). With the development of porous materials, three new pore types have been classified since 2015, they are nanopores (<100 nm), supermicropores (0.7–2 nm) and ultramicropores (<0.7 nm).

### Pore Size Distribution

Although intense efforts are focused on improving the surface area and redox active sites of PC materials, optimization of pore size distribution is also an effective mode to enhance capacitance, especially for EDL capacitors. Generally, nanoporous carbons consist of pores ranging from micropores to mesopores, and most of them cannot maintain long-range ordering, or be accessible to ions from the electrolyte, though we know the electrochemical performance of PC is dependent on the interface of carbon and electrolyte. The macropore, mesopore and micropore have different functions in the electrochemical charge/discharge process. Macropores serve as ion-buffing reservoirs for meso- and micropores; mesopores provide abundant transport channels for ions' diffusion; in spite of some regions of micropores are inaccessible to adsorb ions, it still affects the charge status through controlling the diffusion of ions and molecular sieve effects, and then changing the capacitance of PC materials (Liu et al., 2018c); in addition, as the size of pores decreases toward <1 nm, an anomalous capacitance increase occurred in the organic electrolyte due to ion dissolution. Chmiola and co-workers reported that PC could achieve the highest capacitance as its average pore size matched the size of desolvated ions (Chmiola et al., 2006). Kondrat (Kondrat et al., 2012) found that different carbon materials with the same average pore size will display very different capacitive properties, because of the difference in pore size distribution; besides, for the monodisperse porous electrode, its energy density and pore size do not fit a non-monotonic function (**Figure 1**). Moreover, Cheng (Wang et al., 2008) believed that high micropore volume and the micropore-to-totalpore ratio are crucial for gaining a high-rate electrochemical performance of PC.

# Pore Connectivity

interconnected pores at different size dimensions are qualified as a hierarchical porous structure, which facilitates electrolyte infiltration and ion diffusion through a different pore canal (Borchardt et al., 2013). Until now, it is still a big challenge to construct carbon materials with 3D-interconnected, long-range order macroporous and mesoporous structures. These structures, combined with decent SSA and proper pore size distribution, allow efficient diffusion of any substance, such as electrolytes, ions etc., to the interior space of all channels. The strategies for fabrication HPC mainly includes templating (hard/soft) (Yuan et al., 2016; Xiong et al., 2017; Zhu et al., 2017) and nontemplating ways (Lv et al., 2012), while these methods are timeconsuming and highly costly. As a promising renewable resource, Biomass always possess a naturally interconnected, multichannel and porous structure, hence, they are the excellent candidates for preparing HPC. Wang and co-workers used pomelo peel as the carbon sources, in virtue of its foamy fibrous layer and abundant oxygen-containing functional groups, HPC was obtained after KOH activation, it gained a capacitance of 222.6 F/g at relatively low discharge current density (0.5 A/g), at the same time, it

maintained a good rate of performance and a suitable cycling stability (Li et al., 2017a).

#### Pores' Hydrophilicity

Until recently, the charging mechanism for supercapacitors is derived from ions into the PC network by applied potential driving. Hence, the pore wettability, also named as surface hydrophobic/hydrophilic balance, closely relates to pores' surface functional groups, and has a great impact on the penetration of guest species into the pore systems and transfer of electrons (Xiao et al., 2016). Therefore, the functionalization of PC materials through surface modification gradually becomes an important issue to achieve appropriate wettability (**Figure 2**). The modification techniques always involve post-treatment of PC materials in oxidizing media, or doping carbons with oxygen, nitrogen and other elements (Lin et al., 2015; Chen et al., 2017b). Generally, N- and O-doping, introduced by grafting different functional groups, facilitate adsorption of ions and further improve the hydrophilicity of the carbon matrix. Therefore, moderate oxidation accompanied by heteroatom doping provide faradic pseudocapacitance, and then increase the capacitance value of electrode materials. However, overoxidation results in collapse of pore structure and large interface resistance (Wei et al., 2016). Moreover, most natural biomass contains nitrogen, boron, sulfur and other trace elements, which would be doped into the carbon framework as heteroatoms, and that will generate a more active site, and decrease the hydrophobicity of PC. Li utilized corncob as a precursor, and obtained nitrogen-doped activated carbons with N content up to 4 wt%, the corresponding hybrid-type supercapacitor achieved high-energy density and rate performance (Li et al., 2015a).

# BRIEF OVERVIEW OF FABRICATION STRATEGIES FOR PC

Up to now, various synthetic methods have been developed for preparing biomass-derived PC. The carbonization methods, such as pyrolysis (Dubal et al., 2016) and hydrothermal carbonization (Enterría et al., 2016), are used earliest to obtain PC. The pyrolysis is a dry-carbonization reaction, which usually takes place in an inert atmosphere or low oxygen environment at an elevated temperature condition (300–900◦C), the main components of biomass are gradually transferred into biochar through a series of reaction, such as cross-linking depolymerization, fragmentation reactions etc. The performance of biochar depends on the reaction temperature, time and catalyst (Li et al., 2016b); For the hydrothermal carbonization, it is carried out in aqueous environment at elevated temperature (<300◦C) and autogenous high pressure, it is a chemical process for conversion of biomass to carbonaceous materials, whose properties are determined by reaction temperature, time, pressure, and water/biomass ratio. Furthermore, Compared with the pyrolysis reaction, hydrothermal carbonization results in higher biochar yield. However, biochars obtained through the aforementioned methods have low SSA and porosity. Hence, as the most common strategies for increasing SSA of carbon materials, activation, including physical and chemical activation, have been wildly used.

Physical activation is carried out at a high temperature (>700◦C) in the presence of gases like CO2, H2O, air and ozone (Abioye and Ani, 2015; Chang et al., 2016; Lota et al., 2016). The process of physical activation has two steps: firstly, carbonization happens at a low temperature in an inert atmosphere, and during this process, volatile matters are eliminated and biochar is formed. Subsequently, gasification reaction leads to the formation of abundant open pores because of the introduction of oxidizing gas at a high temperature. It should be noteworthy that increasing reaction temperature and prolonging treating time are helpful to improve the porosity of carbon materials, while the pore size distribution will be broadened.

Prior to carbonization, biochar is pre-mixed with certain chemicals, such as an acid (Sun et al., 2015), strong base (Qu et al., 2015), or a salt (Sevilla and Fuertes, 2016), then the mixture is carbonized at relatively low temperatures (450–900◦C). Although there are drawbacks for chemical activation, for example, high cost, apparatus corrosion, and non-recoverable chemicals, it is still preferred over physical activation owing to its lower reaction temperature, shorter reaction time and larger SSA. Among

various kinds of activating reagents, KOH (Lv et al., 2011) is the most wildly used chemical. The activation mechanism of KOH activation (**Figure 3**) can be summarized and shown as the following reaction equations.

$$KOH + C \leftrightarrow K + CO\_2 + H\_2O \tag{1}$$

$$\rm KOH + CO\_2 \Leftrightarrow K + CO\_2 + H\_2O \tag{2}$$

$$KOH + C \leftrightarrow K + H\_2 + K\_2CO\_3 \tag{3}$$

$$\text{C} + \text{CO}\_2 \Leftrightarrow \text{CO} \tag{4}$$

#### Cellulose-Derived PC

As the most abundant and sustainable natural polymer, cellulose is the main component of green plants' primary cell wall and consists of a linear chain of D-glucose. Cellulose fibers exhibit high surface area and aspect ratios, excellent mechanics and flexibility, broad chemical-modification capacity, so it has attracted intense attention in the past several years for supercapacitor (Yang et al., 2015; Yu et al., 2016). Generally, according to the origin of cellulose, it can be classified in two ways: commercial cellulose and isolated cellulose from lignocellulose for PC precursor. After carbonization and

activation through a physical or chemical process at elevated temperature, then PC was gained. Although they have extremely high SSA, compared with HPC, it exhibited limited C<sup>s</sup> because of low wettability, slow ions' diffusion rate and relatively small amount of the effective active site. Hence, recently, many efforts have been focusing on the fabrication of HPC derived from cellulose.

The unique structure of carbon aerogels provides high SSA, hierarchical porous structure, interconnected macro-/meso- /micropores, efficient ions diffusion rate, and abundant active sites. Nanocellulose microcrystalline would be gelling at the base condition, after freeze-drying and CO<sup>2</sup> activation process, carbon aerogel with interconnected 3D nanostructure could be prepared. The final products showed high SSA (1,873 m<sup>2</sup> /g) and high pore volume (2.65 cm<sup>3</sup> /g); besides, the C<sup>s</sup> reached 302 F/g and 205 F/g at 0.5 A/g and 20 A/g, respectively (Zu et al., 2016). "leavening" strategy (**Figure 4**) is applicable to various biomass and their derivatives, the corresponding electrode materials exhibited a C<sup>s</sup> of 253 F/g with superior cycling stability (Deng et al., 2015). Zhuo et al. synthesized HPC through a dissolving-gelling process, after carbonization process, carbon aerogel with abundant macropores, mesopores and micropores was obtained, and gained high C<sup>s</sup> of 328 F/g at 0.5 A/g with 96% of the capacitance retention after 5,000 cycles (Zhuo et al., 2016). Sodium carboxymethyl cellulose aerogels derived PC could be obtained through sol-gel and KOH activation, its relative low C<sup>s</sup> (152 F/g at 0.5 A/g) might be attributed to the limited SSA (<500 m2/g) (Yu et al., 2016). What's more, Long et al. cellulose nanofibrils and short cellulose nanofibrils were assembled into macro/micro/mesoporous structure by Li and co-workers, who found the hierarchical structure maintained high surface area (1,244 cm<sup>2</sup> /g), good C<sup>s</sup> (170 F/g) at high current densities (Li et al., 2017b). Waste paper is also a promising candidate for preparing PC. Kraft pulp was mixed with KOH solution and freeze-dried, then the mixture was calcined at an elevated temperature under an inert atmosphere. Dense graphene-like HPC was fabricated, and its gravimetric and volumetric C<sup>s</sup> is 309 F/g and 309 F/cm<sup>3</sup> at 1 A/g, respectively (Mo et al., 2018).

It is believed that increasing surface functional groups into the carbon framework contribute to the enhancement of the capacitance of supercapacitors. Hence, introducing of N, B, P, and O, which act as electron donor or electron acceptor, have the following effects: improving wettability between PC and electrolyte; facilitate the binding between carbon materials and ions; enhancing the active sites in the carbon framework (Chen et al., 2013, 2017d; Wu et al., 2015; An et al., 2017). Contrary to oxygenic functional groups, N-containing functional groups own basic characters, which leads to donor-acceptor properties, and improves the final electrochemical properties of electrode materials (Wang et al., 2008). As the carbon sources and as a template, cellulose nanocrystals were used to control the growth of melamine-formaldehyde, then rod-like porous Ndoped carbon particles were fabricated; they achieved 352 F/g at 5 A/g and maintained over 95% C<sup>s</sup> retention after 2000 cycles (Wu et al., 2015). Through pyrolysis and activation procedure, PC nanosheets (PCN) were doped by N and S, the co-doped PCN had a C<sup>s</sup> of 298 and 233 F/g at charge/discharge current density of 0.5 and 50 A/g. Besides, only 2% capacitance lost after 10,000 cycles (Li et al., 2016b). Phosphorus (P) doped HPC possessed outstanding rate capability, and it showed a C<sup>s</sup> of 133 F/g (146 mF/cm<sup>2</sup> ) at 10 A/g with ∼98% capacitance retention after 10,000 cycles (Yi et al., 2017).

# Lignin-Derived PC

As the most abundant natural aromatic polymer and the second raw material from plants, lignin, with a production ∼ 50 million tons per year, is known as one of the most promising candidates for carbon sources considering environmental and economic aspects (Suhas and Carrott, 2007). Recently, the lignin-based PC nanocomposite gradually has attracted much attention because of its excellent electrochemical performance (Kai et al., 2016; Ma et al., 2016; Wang et al., 2016a). Generally, it is quite difficult to isolate the lignin because of the potential oxidation and condensation reaction during the isolation process; besides, the mass production and purity of lignin, isolated through hydrolysis or solubilization of plants, also limit the wide application of lignin-based PC for supercapacitor. However, series of works have been published and some progress has been made. Physical and chemical activations are the most common strategy to prepare lignin-based HPC. Zhang (Zhang et al., 2015a) utilized KOH as the template and activating agent and successfully prepared lignin-derived HPC with an interconnected 3D network. Its abundant oxygen-containing group ensures a relatively good pseudocapacitance performance (165 F/g in 1 M H2SO<sup>4</sup> at 50 mA/g). Increasing SSA of carbon materials is a good choice for promoting the C<sup>s</sup> of electrode materials; hence, a higher HPC derived from lignin was obtained by KOH activation,

the corresponding products provided a C<sup>s</sup> of 286.7 F/g at 0.2 A/g (Zhang et al., 2015b). In addition, a hard template, such as zeolite, was used to control the assembly of lignin, and then control the pore structure of the final products, this lignin-based templated PC showed a high C<sup>s</sup> (250 F/g at 50 mA/g) (Ruiz-Rosas et al., 2014). One dimension of PC fibers derived from lignin not only contributed to faster electron conduction, but also facilitated the electrolyte infiltration by providing abundant pore structure, Hu et al. reported that porous fiber activated by KOH achieved a total capacitance of 1 F at 10 mg loading (Hu et al., 2014). Besides, lignin-based HPC film was developed by Chang and co-workers, as shown in **Figure 5**, a flexible film consists of lignin, PVP and Mg (NO3)<sup>2</sup> was synthesized through electrospinning, then the precursor was carbonized and pickling to remove the MgO, the final PC film exhibited C<sup>s</sup> of 248 F/g and outstanding rate and cyclic performance (Ma et al., 2018).

(2015). Copyright 2015 Royal Society of Chemistry.

To enhance the performance of lignin-derived PC, surfactant, organic solvent and silica were added into the mixture of lignin, after carbonization and activation by 2 M NaOH, the carbon film (∼310µm) was used as an electrode, because of its hierarchical pore structure and superb electrical conductivity, this novel carbon electrode achieved an ultrahigh areal capacitance of 3 F/cm<sup>2</sup> and a high volumetric capacitance of 97.1 F/cm<sup>3</sup> , high mass loading and excellent electrochemical performance ensure it is a promising candidate for supercapacitor (Li et al., 2016a). Furthermore, lignin-derived byproducts were used as carbon resources by hydrothermal treatment and activation, the as-prepared N-doped PC with hierarchical bowl-like pore structure exhibited a high conductivity (4.8 S/cm), favorable C<sup>s</sup> (312 F/g at 1 A/g in 6 M KOH) and excellent rate capability (81% retention at 80 A/g) (Wang et al., 2017). Moreover, Wang reported that N-doped PC derived by KOH activated urea-modified lignin could be obtained; it possessed a welldeveloped porous structure and extremely high SSA (3,130 m2 /g). The corresponding supercapacitors achieved C<sup>s</sup> of 273 and 306 F/g in 6 M KOH and KOH-PVA solid electrolytes, respectively (Wang et al., 2016b).

# Alginate-Derived PC

Except for cellulose, other kinds of natural polysaccharides obtained by artificial extracting have been wildly used as PC precursor because of their low cost, accessibility and component stability.

Alginate is a polysaccharide composed by covalently linked mannuronate and guluronate, and widely distributed to the cell walls of brown algae. Furthermore, the chelation between metal ions and alginate results in gelling and forming an "egg-box" structure (Davis et al., 2003), which further being calcinated at 800◦C, three-dimensional macro-meso-microporous HPC aerogels were prepared. The final products showed an excellent rate performance (65% capacity retention at 100 A/g) and decent capacity (188 F/g at 1 A/g) (Wang et al., 2018a). Li and co-workers synthesized alginate gels with interconnected macropore structure utilizing freeze-drying process, then after activation by KOH and removal of Ca, interconnected HPC was obtained; it showed a high-rate capability with 222 F/g and long cycling life at 10 A/g (Li et al., 2015c). On account of the limited C<sup>s</sup> of an EDLC capacitor, O- and N- enriched PC was prepared by calcining kelp in NH<sup>3</sup> atmosphere, the products acquired high volumetric C<sup>s</sup> (>360 F/cm<sup>3</sup> ) and superb cycling performance (Li et al., 2015b). In addition, Nitrogen-doped PC fibers derived from cobalt alginate with egg-box structure, which led into the formation of abundant large mesopores, and the product presented an excellent capacitive behavior of 197 F/g at 1 A/g and superb cycling ability (Tang et al., 2018). Geng (Geng et al., 2016) reported a facile method to synthesis HPC derived from sodium alginate by pre-carbonization and NaOH activation, and it exhibited a capacity of 451 F/g in 2 M KOH solution; while its rate performance, which greatly depends on the electrical conductivity of the electrode materials, was not satisfied.

# Starch-Derived PC

As the most common carbohydrate in the human diet, starch, a polymeric carbohydrate consisting of glucose units jointed by glycosidic bonds, is a promising candidate for supercapacitors as its low cost, accessibility (Lei et al., 2016; **Figure 6**). Direct carbonization of starch with the aid of activated agents is easy and commonly used by researchers for preparing PC. Guo (Guo et al., 2018) synthesized starch-derived PC through one-pot carbonization, the as-prepared PC achieved a C<sup>s</sup> of 385 F/g at 1 A/g. Zhang (Zhang et al., 2015b) proposed a simple strategy to fabricate hierarchical hollow porous spherical carbon using starch as raw materials and KHCO<sup>3</sup> as the activation agent. The relatively large SSA and rich porous structure endowed the electrode materials with a high capacitance (265.4 F/g at 1 A/g) and excellent rate capacitance (137 F/g at 100 A/g). Hence, carbon derived by different botanical origin would influence the electrochemical performance of the products, and this was verified by Bakierska (Bakierska et al., 2017). In addition, hydrothermal treatment of starch is useful for improving the hydrophilic of carbon materials by introducing the oxygencontaining functional group. Therefore, before carbonization and activation at elevated temperature, hydrothermal carbonization was performed, the potato-derived PC spheres displayed a high C<sup>s</sup> (245 F/g) and good rate performance (61% capacitance retention at 10 A/g) (Qiang et al., 2015); similar processing procedure was used by Hong with sweet potato starch as the carbon source, a C<sup>s</sup> of 208 F/g at 1 A/g was gained (Hong et al., 2018).

The foaming process (Wang et al., 2016b) through physical or chemical blowing with the aid of carbonate of urea leads to the formation of abundant macropores. For this reason, inspired by bread leavening, Deng (Deng et al., 2015) proposed a noncasting and template-free method to fabricate HPC by mixing the starch with KHCO<sup>3</sup> followed by carbonization at a hightemperature condition. The as-prepared samples owned a C<sup>s</sup> of 253 F/g and with no distict capacitance loss after 10,000 cycles. At the same time, Chang (Chang et al., 2017) employed a chemical blowing strategy to synthesis N-doping sheet-like PC by graphitization and chemical activation. The samples gained extremely high surface area (2,129 m<sup>2</sup> /g), large pore volume (0.97 cm<sup>3</sup> /g), good C<sup>s</sup> (337 F/g at 0.5 A/g) and reasonable cycle stability. Similar preparation strategies were employed to treat starch by other researchers to investigate the electrochemical properties of the PC (Pang et al., 2016a; Yang et al., 2016; Du et al., 2017). Calcium acetate was used as the hard template and mixed with starch before carbonization, the PC with tunable pore size by adjusting the ratio of starch/calcium acetate was prepared. When the current density increased from 0.1 A/g to 10 A/g, the C<sup>s</sup> of the electrode materials transformed from 277 F/g to 182 F/g, and the cycling results indicated that the PC owned an extremely outstanding cycle stability even after 20,000 cycles (Zhang et al., 2016b). Compared with other kinds of biomass, corn starch is a relatively pure resource as its high-yield character. PC derived corn starch by hydrothermal carbonization, and activation acquired high SSA (1,239 cm<sup>2</sup> /g), good capacitance (144 F/g) and energy density (19.9 Wh/kg), that were performed slightly better than commercial PC (Pang et al., 2016b). The rate performance of the electrode can be improved by combining PC and carbon cloth together. Zhong (Zhong et al., 2018) synthesized a binder-free activated carbon electrode via sol-gel and KOH activation; the corresponding electrodes achieved 272 F/g (1 A/g) and C<sup>s</sup> retention of 75.9% at 50 A/g.

# Chitin-Derived PC

As one of the most abundant natural polymers, chitins, with stiff chain conformation and considerable nitrogen concentration (∼6.9 wt%), can be completely dissolved into the mixture solution of NaOH and urea. Generally, the addition of PETF hampers the electron transport and restrains the rate capacity of electrode materials. Zhang (Zhang et al., 2016a) employed a one-step synthesis strategy for preparing a binder-free PC electrode by sol-gel method. After KOH activation, the electrode achieved gravimetric C<sup>s</sup> of 272 F/g and 75.9% C<sup>s</sup> retention at 50 A/g. By means of emulsification and carbonization at high temperature, N-doped microsphere with ample interconnected porous structure, high SSA, unique elasticity and outstanding rate performance were obtained (Suhas and Carrott, 2007). Gao employed prawn shells as a carbon precursor, the obtained Ndoped activated carbon exhibited very fine C<sup>s</sup> of 357 F/g (6 M KOH) and 695 F/g (1 M H2SO4) (Zhang et al., 2015b). The fungus was also used as PC sources, Long (Ruiz-Rosas et al., 2014) found the corresponding graphene-like carbon possessed high specific surface are (1,103 m<sup>2</sup> /g), outstanding volumetric (360 F/cm<sup>3</sup> ) and cycle stability (99% capacitance retention after 10,000 cycles). At the same time, N-doped PC derived from shrimp shells with high surface area (1,271 m<sup>2</sup> /g) by KOH activation also got a high C<sup>s</sup> (239 F/g at 0.5 A/g). Moreover, as a renewable biomass mainly composed of chitin, cicada slough derived PC was obtained through carbonization in air and KOH etching in inert atmosphere, results showed that the products exhibited fairly high oxygen content (∼30%), moderate nitrogen content (∼4%) and high C<sup>s</sup> (266 F/g at 0.5 A/g) (Hu et al., 2014).

## Gelatin-Derived PC

As a renewable biomass resource, gelatin owns abundant rich – NH<sup>3</sup> groups, which ensure a good wettability after carbonization, and facilitates nitrogen atoms doped during the activation process. Hence, it is believed to be a promising candidate for the raw material of PC for supercapacitor. Recently, it is still a big challenge for preparing bio-sources derived carbon materials with 2D planar architectures, which promote ions transportation and electrons conduction, and further affecting the capacitance and rate performance of the electrode materials. Therefore, utilizing gelatin and dopamine as carbon and N sources, respectively, Fan and colleagues (Fan et al., 2015) developed a general strategy for synthesizing 2D PC nanosheets through a series of treatment, including intercalation, thermal treatment and chemical etching; meanwhile, layered montmorillonite was added to control the microstructure of gelatin, and brought about the formation of two-dimensional nanosheet-like PC. The final products exhibited an enhanced rate capability and a high C<sup>s</sup> of 246 F/g; besides, the decent capacitance retention (81% at 100 A/g) indicates the electrode material is a good candidate of electrode material for supercapacitor. Furthermore, a free-standing N-doped PC film derived from gelatin/copper hydroxide nanostrands composite films were synthesized by Hu et al. (2016); the binder-free mesoporous N-doped carbon film possessed promoted specific energy (28.1 Wh/kg) and high specific capacity (316 F/g at 0.5 A/g), while its rate performance was not satisfied. Doping with multi-elements is another strategy to improve the electrochemical performance of PC because of the synergistic effect. Therefore, a facile yet sustainable approach was used to produce B/N co-doped PC. Boric acid was added into the gelatin solution, and it acted as hard templates after crystallization during the evaporation of aqueous solution at an elevated temperature. The platelike shaped boric acid regulating the assembly of gelatin, after carbonization and activation, 2D plate-like PC formed. As the voltage windows is 0.8 and 1.0 V, the C<sup>s</sup> of the PC is 240 and 230 F/g, respectively. It also showed improved capacitance retention (Zheng et al., 2016b). Except for boric acid, graphene oxide was applied as a regulator to modify the pore structure, composition, and microstructure of PC derived from gelatin, and results showed that the carbon nanosheets with thickness range from 10 to 30 nm, owned high C<sup>s</sup> , decent rate capability and high capacitance retention (76% at 20 A/g). At the same time, Zhang applied a similar approach to preparing layer-like PC with the thickness ∼100 nm, which delivered a high discharge C<sup>s</sup> (366 F/g at 1 A/g), good rate capability (221 F/g at 30 A/g) and suitable cycling performance (Zhang et al., 2015c).

## Plant-Tissue-Derived PC

It is well-known that a thick electrode with high areal active materials loading is urgent in supercapacitor designs as it is closely related to the energy density of supercapacitor devices, and it is useful to reduce the cost of manufacturing through maximizing the packing density of electrode materials and decreasing the layers of inactive materials. What is noteworthy is that the transportation rates of the ion and electron are inversely proportional to the thickness of the electrode in the real-world application. However, natural wood possesses a unique anisotropic structure and plenty of open channel along the growth direction, which facilitates alleviating the electrode materials' tortuosity, promoting the ions transfer, and further increasing the rate capacity of the electrode materials (**Figure 7**). Hence, a surface modified porous wood carbon derived from poplar wood was prepared by Liu (Liu et al., 2012), and the products achieved a maximum gravimetric and volumetric capacitance of 234 F/g and 36 F/cm<sup>3</sup> , respectively. Basswood was carbonized by multistep thermal treatment and CO<sup>2</sup> activation; the obtained HPC with extremely high mass loading of MnO2 gained high-energy density (1.6 mWh/cm<sup>2</sup> ) and power density

(24 W/cm<sup>2</sup> ) (Chen et al., 2017a). In addition, highly anisotropic, multichannel wood carbon doped by N and S were also exhibited well electrochemical properties, the PC showed a C<sup>s</sup> of 704 F/g at 0.2 A/g, and still maintained a competitive C<sup>s</sup> of 349 F/g at 4 A/g (Tang et al., 2018). Pinecone tree activated carbon was prepared by KOH activation at 800◦C, and the products showed a C<sup>s</sup> and energy density of 69 F/g at 0.5 A/g and 24.6 Wh/kg, respectively, besides, it exhibited excellent voltage stability after holding 110 h (Barzegar et al., 2017).

Similar to the wood's microstructure, bamboo-derived carbon was obtained by carbonization with the aid of KOH under N<sup>2</sup> atmosphere. After being doped by boron and nitrogen, the corresponding HPC exhibited a C<sup>s</sup> of 281 F/g (1 M, KOH) and energy density (37.8 Wh/kg) (Chen et al., 2015). As N-enrich biomass, the bamboo shoot was carbonized by Chen (Chen et al., 2017c) to prepare N-doped PC materials, which was endowed with high SSA, highly interconnected pores and uniform nitrogen dopant distribution. Its outstanding C<sup>s</sup> and rate performance ensured it is a promising candidate for supercapacitors. Wei employed broussonetia papyrifera as a biomass source, after hydrothermal treatment in KOH solution, then pyrolysis and activated at high temperature in Ar atmosphere, N-doped HPC with high SSA (1,212 m<sup>2</sup> /g) and outstanding C<sup>s</sup> (320 F/g at 0.5 A/g) were fabricated (Wei et al., 2015). Thubsuang reported that rubberwood waste treated with H3PO<sup>4</sup> or NaOH led into the formation of carbon monoliths, which exhibited a maximum gravimetric capacitance, volumetric capacitance and energy density of 129 F/g, 104 F/cm<sup>3</sup> , and 14.2 Wh/kg, respectively (Thubsuang et al., 2017). What's more, wood waste was used as raw material, and the N-doped PC-PANI composite possessed a decent C<sup>s</sup> (347 F/g at 2 A/g) and high energy density (44.4 Wh/kg) (Yu et al., 2015). Compared with conventional KOH activation, impregnate-activation method resulted in PC derived from pine tree sawdust with higher surface area and richer interconnected pores, and achieved good C<sup>s</sup> in the organic electrolyte (146 F/g) and IL electrolyte (224 F/g) (Wang et al., 2017). Except for these woody materials aforementioned, other woody wastes, such as spruce bark (Sun et al., 2018), Java Kapok tree (Kumar et al., 2018), plane tree (Yao et al., 2017), Melia azedarach (Morenocastilla et al., 2017), auriculiformis tree bark (Momodu et al., 2017), ginkgo leaves (Hao et al., 2017) (**Figure 8**) were developed as carbon for fabricating PC, whose C<sup>s</sup> was all better than commercial activated carbon. Musa basjoo, which appears like a tree and have three layers of structure involves nano-/micro-/millimeter level, that endowed the PC with abundant interconnected pores after KOH activation. The products displayed high C<sup>s</sup> and good cycling performance (Zheng et al., 2016a).

In addition to wood, plants with a similar structure like shrimp shell, which preserve its layered structure after carbonization and washing with an acid solution, should be a kind of promising candidate for HPC. The microstructure of these plants, such as the flower petals andtree leaves, could be kept by shape fixing via a salt recrystallization strategy, which involves the addition of salt crystals with a high melting point (Ding et al., 2015). Hence, the salt sealing strategy was utilized to fabricate graphenelike PC nanosheets (GPCN) derived from salvia splendens' flower petals. The obtained PC consisted of 10.26% O and 2.3% N through elemental analysis, and it owns high SSA (∼1,051 m<sup>2</sup> /g). Thanks to these characteristics, the GPCN-based carbon materials exhibited ∼220 F/g at 20 A/g, and showed a high-rate capability (86.3% capacity retention from 1 to 100 A/g in 6 M KOH solution) (Liu et al., 2018b). In addition, PC derived from paulownia flower by pyrolysis carbonization and chemical activation also showed high SSA, suitable size distribution, superb wettability and partial graphitization phase, which endowed the electrode with a C<sup>s</sup> of 297 F/g at 1 A/g in 1 M H2SO<sup>4</sup> solution, and the corresponding electrode material exhibited high-energy densities of ∼22 Wh/kg as the power output is 3,781 W/kg (Chang et al., 2015). Moreover, other kinds of flower-derived carbon have been prepared by carbonization and activation process, such as camellia and pine cone flower

(Nagaraju et al., 2016; Ma et al., 2017), Borassus flabellifer flower (Sivachidambaram et al., 2017), Osmanthu's flower (Zou et al., 2018), cornstalk (Wang et al., 2018b), rice husk (Dong et al., 2018), grape seeds (Guardia et al., 2019), rice straw (Liu et al., 2018d), walnut shell (Wang et al., 2019), wood sawdust (Sevilla et al., 2019) all of these PC electrode materials delivered a relative high C<sup>s</sup> and good rate capacity owing to their fine microstructure and abundant N, O elements.

What's more, leaves with a delicate hierarchical structure, which promote the diffusion of electrolytes in living organisms, are also a good choice for transferring it into HPC materials. Li proposed a facile strategy to fabricate a free-standing bio-carbon supercapacitor derived from sisal leaves by being carbonized at 1,000◦C for two times and by chemical activation afterwards, was removed and freeze-dried; the developed free-standing electrode exhibited a high C<sup>s</sup> of 204 F/g at 1 A/g and good rate capacity; at the same time, it also delivered a relative steady capacity in an organic electrolyte (Wang et al., 2016c). It is worth noting that the combination of the physical and chemical activation method might bring about unexpected results. Li (Li et al., 2015d) proposed a novel strategy to treat the fallen leaves by the activations of mixed KOH and KHCO3. The combination of KOH and KHCO<sup>3</sup> led to the enlargement of pore size, which benefited the diffusion of ions and is helpful to enhance the specific capacity and rate performance of carbon materials, the electrochemical characterizations' results revealed that this novel strategy endows the PC with a capacity of 242 F/g at 0.3 A/g in 6 M KOH solution, and it also showed a relative good cyclic performance. Although the activation process is a common method to fabricate biomass-derived PC materials, while the physical or chemical activates, reaction always accompanied with the collapse of the fine microstructure of the biomass, and further affecting the ion/electron diffusion during energy storage. Hence, a possible way should be developed to replace the activation process while endows the PC with high SSA and abundant pore structure. Huang (Huang et al., 2017) proposed a facile technique, which involves indicalamus leaves and polytetrafluoroethylene as carbon precursor and silica-in-situ-remover, respectively, the obtained PC owns a SSA of ∼1,800 m<sup>2</sup> /g, and exhibited high capacitance (326 F/g in 1 V supercapacitor), high-energy density (23.7 Wh/kg at power density of 224.5 W/kg). Furthermore, hydrothermal treatment is a benefit to introducing rich oxygengroup into the framework of carbon, and dual-doped by N, S can promote the electron conduction and increase the active sites of PC materials. Hao and colleagues reported that HPC derived from gingko leaves by hydrothermal treatment in H2SO<sup>4</sup> and then activated by KOH at elevated temperature in inert atmosphere, the final products contained small amount of N and S according to the EDS analysis; and the PC maintained a high C<sup>s</sup> of 364 F/g at 0.5 A/g and excellent cycling capability (98% capacity retention after 30,000 cycles) (Hao et al., 2016). In addition the leaves aforementioned, other leaves, such as tea (Inal et al., 2015; Ma et al., 2017), euonymus japonicas (Zhu et al., 2015), corn (Yang et al., 2017), cabbage (Wang et al., 2016c), have been used as carbon precursors for synthesizing PC materials for supercapacitor electrodes.

#### CONCLUSION AND PROSPECTIVE

In the new era of energy storage, compared with traditional activated carbon materials, biomass-derived HPC has achieved superior performance because of their natural fine structure and rich race elements in the organic tissue. However, it is worth noting that their rate capability performance is not satisfied especially at a high charge/discharge electric current condition. That should be contributed to the high concentration of oxygen groups in carbon framework, which is helpful for enhancing the hydrophilic of carbon materials, and further facilitating the infiltration of electrolyte, while these oxygen groups hamper the electric conductivity of PC. Hence, it is urgent to develop an efficient strategy to balance the wettability and electric conductivity of the carbon matrix, and that calls for the joint from theoretical and experimental researches.

Till now, the chemical activation process through KOH, NaOH etc. is a common tool for preparing porous carbon with a rich porous structure, especially for micropores. It should be noteworthy that a large amount of volatile-gases will be formed because of the reaction between KOH and carbon matrix at elevated temperature, these volatile gases, including metal K, CO, CO2, H2, and H2O, not only lead to the formation of abundant micropores, they also threaten the safety of the instrument because of the strong corrosiveness of metal K. Therefore, other activation strategies should be developed to meet the demand of industrialization.

Ion diffusion is a key issue in supercapacitor systems. Although molecular simulation provides different viewpoints for us to design electrode materials with high efficiency; recently, it mainly focuses on the ion-diffusion within ultra-small pores, and it is still a big challenge for us to investigate the diffusion process of ions in hierarchical pore structure, which is extremely important to enhance the electrochemical performance of PC. Moreover, from the perspective of the environment, security and cost, aqueous electrolyte, but not ionic liquids, should be a better choice for the wide application of supercapacitors in energy storage. However, most of the previous studies have been focused on simulating ion behavior in ionic liquids, which is quite different from the aqueous electrolytes, including ion size, solvation shell and diffusion coefficient. Hence, it is necessary to carry out related studies, and investigate the ion diffusion behavior in aqueous electrolytes, then push the development of supercapacitor.

The impact of carbon doping is still unclear, although doped PC materials provide great opportunities to meet the energy density gap between supercapacitor and battery. As we know, introducing nitrogen into the framework of carbon is always accompanied with the enhancement of electrical conductivity and increasing of electrochemically-active sites compared with commercial carbon, which further results in huge improvement of C<sup>s</sup> because of an additional faradaic contribution, while the enhancement is quite limited in an aqueous electrolyte. Meanwhile, although the promising prospect of aqueous electrolyte, till now, ions liquids offer better capacitive performance, while it still needs to investigate numerous cations, anions and solvent molecules to achieve more excellent performance to meet the requirement of the industry. Therefore, high-throughput techniques, already applied in battery materials and electrolytes, is a powerful tool for the exploitation of promising systems among thousands of candidates for supercapacitors, and the systems-screening process is closely related to the results of in-situ experiments and simulation in molecular level.

Unveiling the ion adsorption and charge storage in PC materials for supercapacitor is essential for its applications; in the meantime, it also promotes the development of other energy storage system, for example, redox flow batteries, biofuel cells, flow capacitors. Hence, it is of central importance to combine the experimental and theoretical tools gained from supercapacitors to promote the development of other current/future technologies.

At last, it is noteworthy that previously reported electrode materials with high gravimetric/areal/volumetric capacitance is meaningless if we ignore the corresponding mass loading/working area/ total volume of the electrode. Hence, effective metrics are required to evaluate the performance of numerous materials for supercapacitors, including biomassderived PC. As for gravimetric capacitance, ultra-small mass loading means the storage of a limited amount of charges. So,

#### REFERENCES


the mass loading of electrodes is a key parameter in comparing the capacitance variation, and it is still urgent to improve electrochemical performance of supercapacitor at high mass loading, but not the C<sup>s</sup> of the electrode with small loading active materials.

#### AUTHOR CONTRIBUTIONS

SY and JZ were responsible for literature searching and drafting. All authors contributed equally to the final writing of the paper.

#### FUNDING

This research was supported by National Nature Science Foundation of China (51702139) and Youth Science Foundation (20151BAB216007, GJJ150637, 20161BAB216122).

performance supercapacitor electrodes. Electrochim. Acta 157, 290–298. doi: 10.1016/j.electacta.2014.12.169


carbon nanosheets for high-performance supercapacitors. J. Mater. Chem. A 33, 15954–15960. doi: 10.1039/C8TA04080J


and hydrothermally treated melia azedarach stones. Materials. 10:747. doi: 10.3390/ma10070747


potential as supercapacitor electrode materials. J. Mater. Sci. 52, 6837–6855. doi: 10.1007/s10853-017-0922-z


materials for high-performance supercapacitor. J. Environ. Chem. Eng. 6, 258–265. doi: 10.1016/j.jece.2017.11.080

Zu, G., Shen, J., Zou, L., Wang, F., Wang, X., Zhang, Y., et al. (2016). Nanocellulose-derived highly porous carbon aerogels for supercapacitors. Carbon 99, 203–211. doi: 10.1016/j.carbon.2015. 11.079

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Yang, Ye, Zhou and Liang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Preparation of Flexible Substrate Electrode for Supercapacitor With High-Performance MnO<sup>2</sup> Stalagmite Nanorod Arrays

Yuanyu Ge, Xianfeng Wang and Tao Zhao\*

*College of Chemistry, Chemical Engineering and Biotechnology, Donghua University, Shanghai, China*

A large-area MnO<sup>2</sup> stalagmite nanorod arrays (SNAs) growing vertically on flexible substrates were successfully fabricated by an easy heat-electrodeposition method. The large specific capacitance (646.4 F g−<sup>1</sup> at 500 mA g−<sup>1</sup> ) and excellent rate capability (42.3% retention with 40 times of increase) indicate that the prepared MnO<sup>2</sup> SNAs flexible electrode has outstanding electrochemical performance. Furthermore, after 5,000 repetitions of CV tests, the overall specific capacitance could retain ∼101.2% compared with the initial value meant a long cycling life. These outstanding properties could be ascribed to the effective conductive transport path between Ni substrate and MnO<sup>2</sup> nanorods, and owing to the stalagmite like structure of MnO<sup>2</sup> nanorods, the exposed sufficient active sites are beneficial to the electrolyte infiltration.

#### Edited by:

*Wenyao Li, Shanghai University of Engineering Sciences, China*

#### Reviewed by:

*Bo Li, Shanghai Jiao Tong University, China Guangjin Wang, Hubei Engineering University, China*

> \*Correspondence: *Tao Zhao tzhao@dhu.edu.cn*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *09 April 2019* Accepted: *25 April 2019* Published: *14 May 2019*

#### Citation:

*Ge Y, Wang X and Zhao T (2019) Preparation of Flexible Substrate Electrode for Supercapacitor With High-Performance MnO*2 *Stalagmite Nanorod Arrays. Front. Chem. 7:338. doi: 10.3389/fchem.2019.00338* Keywords: large-area, manganese dioxide, stalagmite nanorod arrays, supercapacitors, flexible electrode

# INTRODUCTION

The expanding requirement for energy consumption has stimulated the development of electrochemical energy storage devices (Simon and Gogotsi, 2008). Supercapacitors (SCs) occupy an important position in the field of energy storage devices due to their high power density, highspeed of charging and discharging and long cycle life (Burke, 2000). Nowadays, the development of supercapacitors to higher mechanical flexibility has become one of the important trends. It is shown that materials, such as graphene, transition metal dichalcogenides, MXenes (Han et al., 2018), and nanocellulose-graphene composites (Xing et al., 2019) could be used to construct nanomaterialsbased flexible electrodes with high performance for supercapacitors, and there were relationships between structures and properties.

There are many different SCs electrode materials, such as carbon-based material (Li G. et al., 2018; Takeuchi et al., 2018) for the electric double layer capacitors (EDLCs), transition metal oxides, nitrides (Yi et al., 2018) and nickel based materials (Feng et al., 2014; Li et al., 2018) for the pseudocapacitors (PCs). Manganese dioxide (MnO2) is one of the rapid developed metal oxide electrode materials in recent years (Wang et al., 2015). α-MnO<sup>2</sup> has a high specific capacitance among the various crystallographic structures of MnO2, which is mainly due to its largest tunnel (Sanger et al., 2016) that can store more foreign cations for charge balance. More importantly, the electrolytes used for MnO<sup>2</sup> electrodes are non-corrosive neutral solutions, which is green and fit for environmental protection (Li et al., 2012).

As we know, the effects of active materials on the performance of SCsinclude morphology, structure, contact with collector plate and active sites, etc. Unfortunately, during the redox reaction, the electrons transport of MnO<sup>2</sup> electrodes often restricted by its high electrical resistivity. In addition, the defect of the conventional powder electrode preparation is that the generated 'dead volume' by the bonding process of the active material to substrate could make a deterioration in the electrochemical properties of the electrodes (Liu et al., 2016). To solve the defects noted above, it could be a feasible way to manufacture binder-free MnO<sup>2</sup> supporting on 3D conductive substrates. The benefits of the 3D binder-free electrode are the enhanced electron transport and sufficient free space between the nanostructures, which contribute to the generation of more active sites for Faradaic reaction. Li et al. prepared the Ni foam-based ultrafine MnO<sup>2</sup> nanobelts with capacitance of 509 F g−<sup>1</sup> at 0.2 A g−<sup>1</sup> (Li et al., 2013). Davoglio et al. (2018) synthetized the α-MnO<sup>2</sup> particles at the aqueous-organic interface with capacitance of 289 F g−<sup>1</sup> at 0.5 A g−<sup>1</sup> . Although some achievements have been made in the preparation of MnO<sup>2</sup> electrodes on conductive substrates, there is still great potential to develop easy methods for the preparation of high performance MnO<sup>2</sup> nanowires with various forms.

In this paper, the heat-electrodeposition method was adopted to formulate a nanoarrays of stalagmite like MnO<sup>2</sup> on the flexible substrates. The prepared electrode gets capacitance of 646.4 F g−<sup>1</sup> (500 mA g−<sup>1</sup> ) and 42.3% retention (current density increased 40 times) for a remarkable rate capability. And the total capacitance retention rate after 5,000 cycles is ∼101.2%. Furthermore, to verify the generality of the synthesis method, another flexible activated carbon fiber (ACF) was also used as the substrate for the growth of MnO<sup>2</sup> nanoarrays.

#### EXPERIMENTAL

#### Material Preparation

The preparing method in detail was: the heat-electrodeposition process carried on a 3D porous Ni foam. Before electrodeposition, the Ni foam was cut into ∼ 3 × 1 cm<sup>2</sup> , and then immersed into a 5 mol/L HCl solutions along with supersonic wave treatment for 10 min to dissolve the NiO layer on the surface. The Ni foam obtained from the previous step was rinsed to neutral with distilled water, and then subjected to vacuum drying (60◦C, 4 h). The heat-electrodeposition occurred in a cell with the water bath. The composition of the electrolyte was as follows: Mn(CH3COO)<sup>2</sup> (0.01 M), CH3COONH<sup>4</sup> (0.02 M) and dimethylsulfoxide (DMSO, 10 vol.%). The corresponding working electrode, the counter electrode and the reference electrode were the treated Ni foam, the Pt plate (1.5 × 1.5 cm<sup>2</sup> ) and saturated calomel electrode (SCE), respectively. The heat-electrodeposition condition was applied at a constant current (0.5 mA cm−<sup>2</sup> ) by the Autolab electrochemical workstation at ∼ 80◦C for 60 min. After that rinsed the obtained sample to neutral and placed it in a 60◦C vacuum dryer for 4 h. Finally, the sample was calcined in N<sup>2</sup> atmosphere (heating-up 0.5◦C min−<sup>1</sup> , 250◦C, 2 h). The weight gain of the sample after the deposition was the active matter weight.

#### Material Characterizations

In order to analyze the samples qualitatively, the X-ray diffractometer (XRD; Rigaku D/max-2550 PC, Cu-Kα radiation) spectrum was utilized. To observe the microstructures, scanning electron microscope (SEM, Hitachi S-4800) and transmission electron microscope (TEM, JEOL JEM-2100F) were adopted. Mass weighing (Mettler Toledo XS105DU, δ = 0.01 mg).

#### Electrochemical Measurements

The Electrochemical Workstation (Autolab PGSTAT302N, electrolyte 0.5 M Na2SO4) was used to measure electrochemical performance. In the test, the obtained MnO<sup>2</sup> electrode (∼1 cm<sup>2</sup> ) was used as the working electrode. The counter and reference electrode were the same as mentioned above. Specific capacitance calculation (Guan et al., 2017):

$$C = I \cdot \Delta t / (\Delta V \cdot m) \tag{1}$$

where C (F g−<sup>1</sup> ) is the specific capacitance, I (A) is the discharge current, 1t (s) is the discharge time consumed in the potential range of 1V, 1V (V) is the potential window, m (g) is the mass of the active materials.

The weight of the 1 cm<sup>2</sup> MnO<sup>2</sup> electrode was ∼ 1.32 mg. The cyclic voltammetry (CV) potential window was −0.1 to 0.9 V. The scan rates increased from 1 to 100 mV s−<sup>1</sup> . The galvanostatic charge-discharge (GCD) curves were measured under current densities from 0.5 to 20 A g−<sup>1</sup> . The cycle life was obtained by CV test (50 mV s−<sup>1</sup> ) with repetitions of 5,000.

#### RESULTS AND DISCUSSIONS

Except for the two strong peaks of 3D Ni foam substrate, the XRD diffraction pattern in **Figure 1** shows that other peaks at 12.8, 28.8, 36.7, 37.5, 56.4, 60.3, 65.1, and 69.7◦ , which are characteristic (110), (310), (400), (211), (600), (521), (002), and

(541) reflections of α-MnO<sup>2</sup> (JCPDS 44-0141, a = b = 9.785 Å and c = 2.863 Å), respectively.

An aligned and dense MnO<sup>2</sup> nanoarray is presented in **Figure 2a** and there are no macroscopic defects among them. The enlarged SEM image of **Figure 2b** shows a highly open structure was formed by a vertical MnO<sup>2</sup> arrays on the substrate, which is conducive to the full entry of electrolytes. The TEM characterization was performed in order to get more structural information of the MnO<sup>2</sup> arrays. **Figure 2c** is a single stalagmite like MnO<sup>2</sup> nanorod which appears to be truncated cone-shaped with burrs on its surface. The typical diameters are ∼30 nm of the root and ∼80 nm of the top, and the length is up to ∼180 nm. HRTEM image of the single MnO<sup>2</sup> nanorod edge (**Figure 2d**) suggests that the surface is clearly with uniform single crystal and the interplanar spacing of 0.28 nm matches the (001) lattice plane of α-MnO<sup>2</sup> crystal. In addition, FFT diffraction pattern can also be associated with [100] zone axis of α-MnO<sup>2</sup> crystal (inset of **Figure 2d**).

In order to understand how the unique structure of stalagmite MnO<sup>2</sup> nanorod arrays (MnO<sup>2</sup> SNAs) was formed, the timedependent electrodepositing experiments were carried out through controlling the electrodeposit reaction time. **Figure 3a** shows that after deposition of 2 min, nanoparticles can be seen on the surface of the substrate with the diameters of 10–20 nm. When the reaction time increases to 20 min, lots of well-distributed cylindrical nanorods began to appear, as displayed in **Figure 3b**. After doubling the reaction time, some irregular burrs were observed around the increased nanorods (**Figure 3c**). As shown in **Figure 3d**, after 60 min of the electrodeposition, it is found that large-scale and uniform nanorod arrays were formed. The change of the morphology of the samples after different electrochemical deposition time could help us estimate the formation process of the MnO<sup>2</sup> nanorod arrays. In the early stage of the reaction, the formation of the nucleus in the precursor solution and the 1D growth behavior of Mn2<sup>+</sup> (Ding et al., 2011) occurs successively, which result in the deposition of some tiny nanorods randomly and fast all over the surface. As the deposition intensifies, nanoparticles develop into nanorods with burrs, similar to stalagmite. When the reaction time reached to 1 h, the nanorods presented a truncated cone-shaped and formed a high-density array. The formation process of MnO<sup>2</sup> SNAs was elucidated in **Figure 3e**.

Furthermore, to verify the generality of the synthesis method, another flexible ACF substrate was chosen to replace the Ni foam. **Figure S1** shows that the dense needle-like MnO<sup>2</sup> nanorod arrays were grown on the ACF, which proves that the heatelectrodeposition method could be used to fabricate electrodes with different flexible substrates.

**Figure 4A** shows that the MnO<sup>2</sup> SNAs electrode CV with scanning rate from 1 to 100 mV s−<sup>1</sup> were approximate to rectangle and symmetrical. All CV curves feature exhibits the pseudocapacitive nature (**Figure S3**) and a high-speed charge and discharge process (Hu et al., 2016). The possible reason for this phenomenon could be that the active sites in the microstructures of the stalagmite MnO<sup>2</sup> SNAs were fully in contact with electrolytes. The symmetrical GCD curves of the MnO<sup>2</sup> SNAs shown in **Figure 4B** revealed that the reversible redox reaction

and good electrochemical capacitance characteristics of the system. In addition, by comparing the specific capacitance of the Ni foam substrate and the prepared MnO<sup>2</sup> SNAs electrode in **Figure S2**, it is found that the substrate has little effect on the capacitance value.

In **Figure 4C**, the calculated specific capacitance of the electrode were 646.4, 587.1, 538.1, 463.6, 387.6, and 273.6 F g −1 (Red Star), respectively, which were higher than reported works with similar MnO<sup>2</sup> structure (Li et al., 2013; Wang et al., 2017; Davoglio et al., 2018). **Figure 4C** also shows that the capacitance retention increased with the increase of current density. When it increased 40 times from 0.5 to 20 A g−<sup>1</sup> , the corresponding specific capacitance reduced from 646.4 to 273.6 F g−<sup>1</sup> with the retention of 42.3%. The excellent rate capability indicates the potential of the MnO<sup>2</sup> SNAs electrode in high-power applications, and it could be caused by the unique open structure. The advantages of the open structure are that it is favorable for electrolyte infiltration and it owns large electrolytic accessible area, which not only promote redox reaction but also facilitates intercalation and de-intercalation of active species.

**Figure 4D** shows that a beneficial cycle life of the electrode and the capacitances retention enlarged during the initial 3,000 loops. The possible explanation is that the initially inactive part of the MnO<sup>2</sup> SNAs electrode was activated by the permeated electrolyte and then the specific capacitance is increased (Xia et al., 2012).

After 5,000 cycles, a high specific capacitance retention of 101.2% indicated that the cyclical stability of the MnO<sup>2</sup> SNAs electrode was good. The strong contact between the active matters and the substrate facilitating the collection and enhancement of the electron participation reaction could be an important reason for the effective cycle stability. Thus, it is concluded that the electrode material of the assynthesized MnO<sup>2</sup> SNAs demonstrates an excellent cycle life.

#### CONCLUSION

In conclusion, the stalagmite MnO<sup>2</sup> nanorod arrays successfully grew on the flexible substrate by heat-electrochemical deposition method. The prepared MnO<sup>2</sup> SNAs electrode has the high specific capacitance, the outstanding rate capability and the long cycle life, all of which all suggest its excellent electrochemical performance. In addition, this approach could pave the way for a facile low-temperature heat synthetic route for generating a variety of metal oxides arrays flexible substrate electrode.

# DATA AVAILABILITY

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

# AUTHOR CONTRIBUTIONS

YG did the experiments and described the images of figures. XW helping with writing. TZ was the supervisor of this research work. All authors participated in the analysis of experimental data and manuscript preparation.

#### REFERENCES


## FUNDING

This work was supported by National Key R&D Program (2017YFB0309100) from Ministry of Science and Technology of the People's Republic of China.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00338/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Ge, Wang and Zhao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# In situ Growth of Cu2O/CuO Nanosheets on Cu Coating Carbon Cloths as a Binder-Free Electrode for Asymmetric Supercapacitors

Lina Xu<sup>1</sup> , Jiao Li <sup>1</sup> \*, Haibin Sun<sup>1</sup> \*, Xue Guo<sup>1</sup> , Jiakun Xu<sup>2</sup> , Hua Zhang<sup>1</sup> and Xiaojiao Zhang<sup>1</sup>

*<sup>1</sup> School of Materials Science and Engineering, Shandong University of Technology, Zibo, China, <sup>2</sup> Yellow Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Qingdao, China*

Cu2O/CuO nanosheets *in-situ* grown on Cu-Carbon cloths (Cu-CCs), namely Cu2O/CuO@Cu-CCs, are constructed by a simple strategy with electroless copper plating, chemical etching, and thermal dehydration. The as-prepared material is directly used as binder-free electrodes for supercapacitors (SCs). CCs coated with Cu, as the current collector, can effectively promote the charge collection and electron transfer, while the hierarchical Cu2O/CuO nanosheets provide massive active sites for fast faradic reactions. The composite electrode exhibits high specific capacitance [1.71 F cm−<sup>2</sup> , equivalent to 835.2 F g−<sup>1</sup> , at the current density of 10 mA cm−<sup>2</sup> (3.57 A g−<sup>1</sup> )]. The asymmetric supercapacitor device using Cu2O/CuO@Cu-CCs as the positive electrode and activated carbon as the negative electrode, achieves a superior energy density up to 60.26 Wh kg−<sup>1</sup> at a power density of 299.73 W kg−<sup>1</sup> and an excellent long-term cycling stability (9.65% loss of its initial capacitance after 5,000 cycles). The excellent electrochemical performance is mainly ascribed to the unique hierarchical structure of Cu2O/CuO@Cu-CCs, making it attractive as a potential electrode material for high performance SCs.

Keywords: copper oxide, nanostructures, electrode, carbon cloth, asymmetric supercapacitor

# INTRODUCTION

Supercapacitors (SCs), one of the most promising energy storage devices, have received extensive attention owning to their high power density, fast charge/discharge speed, long cycling life span, and low-cost (Lu et al., 2014; Xiong et al., 2015; Sami et al., 2017; Dai et al., 2018). According to the reaction mechanisms, SCs can be classified into electrical double layer capacitors (EDLCs) and pseudocapacitors (PCs) (Wei et al., 2012). For EDLCs, the charges are stored electrostatically at the electrode/electrolyte interface while typically taking carbon materials as active materials (Surendran et al., 2018). For PCs, the energy is stored within the electrode through the faradic redox reaction while taking transition metal oxides/hydroxides and conducting polymers as the electrode materials, thus the PCs provide much higher energy density and specific capacitance than EDLCs. Nevertheless, there are many of problems scarcely understood which attract large numbers of investigator devote oneself to resolve, such as inadequate energy density and capacitance, poor electrochemical stability for practical applications.

#### Edited by:

*Wenyao Li, Shanghai University of Engineering Sciences, China*

#### Reviewed by:

*Pankaj Madhukar Koinkar, Tokushima University, Japan Edward Gillan, The University of Iowa, United States*

#### \*Correspondence:

*Jiao Li haiyan9943@163.com Haibin Sun hbsun@sdut.edu.cn*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *07 April 2019* Accepted: *21 May 2019* Published: *06 June 2019*

#### Citation:

*Xu L, Li J, Sun H, Guo X, Xu J, Zhang H and Zhang X (2019) In situ Growth of Cu2O/CuO Nanosheets on Cu Coating Carbon Cloths as a Binder-Free Electrode for Asymmetric Supercapacitors. Front. Chem. 7:420. doi: 10.3389/fchem.2019.00420*

In recent years, various transition metal oxides/hydroxides, such as RuO<sup>2</sup> (Wang et al., 2014), NiO (Ouyang et al., 2019), Ni(OH)<sup>2</sup> (Kim et al., 2017), MnO<sup>2</sup> (Huang et al., 2015), Co3O<sup>4</sup> (Liu T. et al., 2018), Co(OH)<sup>2</sup> (Yang et al., 2018), V2O<sup>5</sup> (Foo et al., 2014), CuO (Bu and Huang, 2017; Li et al., 2017; Liu Y. et al., 2018), Cu2O (Zhang W. et al., 2016; Ji et al., 2017), have been applied to achieve excellent capacitive performance for PCs. Among these materials, CuO, Cu2O, or Cu2O/CuO nanostructures with different configurations including nanoneedle, nanoflowers, nanowires (Dong et al., 2014; Wang et al., 2015; Chen et al., 2016; Xu et al., 2016; Yang et al., 2016), are attracting considerable interest due to their environmental friendliness, numerous reserve, lowcost, chemical stability, and excellent electrochemical properties [theoretical capacitance of CuO up to 1,800 F g−<sup>1</sup> (Liu Y. et al., 2018) and Cu2O is up to 2,247 F g−<sup>1</sup> (Wu et al., 2017)]. However, most of metal oxides/hydroxides were poor in electrical conductivity, making it difficult to achieve high specific capacitance (Xu et al., 2016). To resolve this issue, oxides/hydroxides are typically mixed with ancillary carbon black or binder and then bonded to current collector, but leading to a significant decrease of the overall specific capacitance (Yuan et al., 2017). An effective approach is that, nanostructured electrode materials directly grow on current collectors, forming binder-free electrodes, thus achieving higher energy density (Dong et al., 2014).

Carbon material containing carbon nanotube, graphene and carbon fiber is one kind of the preferred current collectors due to their excellent electrical conductivity and electrochemical stability (Prasad et al., 2011; Moosavifard et al., 2014; Bu and Huang, 2017). Among various carbon materials, carbon cloths (CCs) with low-cost, chemical stability and desirable conductivity, are regarded as novel carbonaceous materials, which are consist of numerous uniform carbon fibers with three-dimensional (3D) structure (Guo et al., 2014; Zhang Y. et al., 2016). The 3D network structure is conducive to shorten the diffusion pathway of ions and accelerate the flow of ions during the electrochemical process. Numerous electrode materials of PCs taking CCs as current collectors have been developed, such as NiCo-LDH@NiOOH (622 F g−<sup>1</sup> at 1 A g −1 ) (Liang et al., 2018), MnNiCoO4@MnO<sup>2</sup> (1931 F g −1 at 0.8 A g−<sup>1</sup> ) (Saray and Hosseini, 2016), MnO<sup>2</sup> nanosheet arrays (2.16 F cm−<sup>2</sup> , at 5 mA cm−<sup>2</sup> ) (Guo et al., 2014). Currently, copper oxide and its composite materials are mainly grown on copper foam and copper foil (Zhang et al., 2015; Singh and Sarkar, 2017), and the combination of CuO or Cu2O with CCs is also in the developing situation. For example, Xu et al. (2016) fabricated CuO nanoflower arrays on CCs, the energy density and power density are 10.05 Wh kg−<sup>1</sup> and 1,798.5 W kg−<sup>1</sup> , respectively. Wan et al. (2017) developed forestlike cuprous oxide/copper with the energy density of 24.0 Wh kg−<sup>1</sup> at 0.625 kW kg−<sup>1</sup> . However, it is still challenging to evolve the commercially viable Cu oxides/hydroxides with high energy/power density, specific capacitance, and excellent cycling stability (Dong et al., 2014). Therefore, it will be worthy to make a thorough research on CuO or Cu2O electrodes grown on CCs.

In order to improve the kinetics and electrochemical performance of electrodes, two typical methods are usually employed. One straightforward approach is to increase the specific surface area of electrodes to provide more active sites for faradaic redox reaction (Daoping et al., 2014). The other method is to improve the conductivity of electrode material to accelerate electron conduction (Lu et al., 2013). Herein, we firstly synthesized Cu2O/CuO nanosheets directly grown on CCs which is coated with Cu film by a simple strategy with electroless copper plating, chemical etching and thermal dehydration. The uniform Cu film on carbon microfiber cloth has a strong binding force. In addition, Cu2O/CuO nanosheets in situ grown on CCs provide sufficient active sites for charge/discharge electronic, which is important for energy storage of supercapacitor. Finally, it is worth mentioning that there are still Cu films between CCs and Cu2O/CuO nanosheets after chemical etching, which is important for promoting electronic conduction.

# EXPERIMENTAL

# Materials Synthesis

CCs (WOS1002) were purchased from CeTech. (NH4)2S2O<sup>8</sup> (Tianjin Huachen Company) and all other reagents (from Aladdin) were of analytical grade without further treatment. In a typical electroless copper plating process, CCs, cut into squares (25 × 25 mm), were firstly heated to 400◦C at a heating rate of 10◦C min−<sup>1</sup> and hold for 30 min in muffle furnace under air atmosphere to remove a part of impurities. And then, the CCs were immersed into concentrated nitric acid to make the surface rough, followed by the sensitization and activation treatment. Stannous chloride/hydrochloric acid and silver nitrate/ammonium hydroxide solutions were used as the sensitizer and activator, respectively (Yuan et al., 2017). The composition of the sensitizing and activating solution are shown in **Supplementary Tables 1, 2**. The sensitization and activator treatment adsorbs a layer of active silver particles on the surface of the carbon cloth as active metal particles, and copper ions were first reduced on the active metal particles, so that the reduction reaction of copper proceeds on the surface of the carbon cloth. Catalyzed CCs with a number of active sites were obtained after in NaOH (10%) for 3 min. Subsequently, the catalyzed CCs were immersed into plating solutions and stirred at a rotating speed of 200 r min−<sup>1</sup> for 60 min at 25◦C, during which Cu films were coated on CCs, thus obtaining Cu-CCs samples. The amount of copper retained is about 0.009 g cm−<sup>2</sup> on the carbon cloth. The composition of the electroless copper plating solution is shown in **Supplementary Table 3**. Formaldehyde is used as a reducing agent, and the main chemical reactions in electroless copper plating solutions are as follows:

Cu2<sup>+</sup> <sup>+</sup> 2HCHO <sup>+</sup> 4OH<sup>−</sup> <sup>→</sup> Cu <sup>+</sup> 2HCOO<sup>−</sup> <sup>+</sup> <sup>H</sup>2↑ + <sup>H</sup>2<sup>O</sup> 2Cu2<sup>+</sup> <sup>+</sup> HCHO <sup>+</sup> 5OH<sup>−</sup> <sup>→</sup> Cu2<sup>O</sup> <sup>+</sup> HCOO<sup>−</sup> <sup>+</sup> 3H2<sup>O</sup> Cu2<sup>O</sup> <sup>+</sup> 2HCHO <sup>+</sup> 2OH<sup>−</sup> <sup>→</sup> 2Cu <sup>+</sup> 2HCOO<sup>−</sup> <sup>+</sup> <sup>H</sup>2↑ + <sup>H</sup>2<sup>O</sup>

In the chemical etching process, the Cu-CCs were dipped into 100 mL mixed solutions with 2.5 mol L−<sup>1</sup> NaOH and 0.1 mol L−<sup>1</sup> (NH4)2S2O<sup>8</sup> at 25◦C for a while, Cu(OH)<sup>2</sup> arrays were in situ grown on Cu-CCs. After being washed, Cu(OH)<sup>2</sup> arrays were decomposed into Cu2O/CuO arrays through a thermal dehydration at 120◦C in air for 3 h, thus obtaining Cu2O/CuO@Cu-CCs electrodes.

#### Materials Characterization and Electrochemical Measurements

The phase compositions of the products were identified by X-ray diffraction analysis (XRD, Rigaku-Dmax 2500 diffractometer). The microstructure and morphology were observed by scanning electron microscopy (SEM, HITACHI S4800) and highresolution transmission electron microscopy (HRTEM, Tecnai G2 F20 STWIN, FEI, USA). X-ray photoelectron spectroscopy (XPS, Kratos Axis Ultra DLD, Britain) was performed using Mg Ka as the exciting source.

Cyclic voltammetry (CV) and galvanostatic charge-discharge (GCD) tests of Cu2O/CuO@Cu-CCs electrodes were tested on a CHI 660E electrochemical workstation (Shanghai Chenhua Instrument Company, China) in a three-electrode electrochemical cell using a 6 M KOH aqueous solution as the electrolyte at room temperature. The Cu2O/CuO@Cu-CCs electrodes were used as the working electrode, while a platinum wire and an Ag/AgCl electrode as the counter and reference electrode, respectively. Electrochemical impedance spectroscopy (EIS) tests were performed in the frequency ranging from 106 to 0.01 Hz. The specific capacitances were calculated from the discharge part of the GCD curves using the following equation.

$$\mathcal{C} = (\mathcal{I} \int \mathcal{V} \text{dt}) / (\mathcal{S} \mathcal{V}^2) \tag{1}$$

where C represents the specific capacitance (F cm−<sup>2</sup> ), I represents the discharge current (A), 1t is the total discharging time (s), S is the area of the sample (cm<sup>2</sup> ), and 1V is the potential change (V) within the discharge time 1t.

#### Fabrication and Electrochemical Measurements of Asymmetric Supercapacitor

Active carbon, acetylene black, and poly tetra fluoroethylene (PTFE) with a mass ratio of 80:10:10 were mixed with moderate amount of ethanol. The resulting mixture was brushed on carbon cloth and dried at 80◦C for 10 h in a vacuum oven. Acetylene black and PTFE are acted as conductive agents and binders, respectively. The asymmetric supercapacitor (ASC) device was assembled by using Cu2O/CuO@Cu-CCs electrode (with a diameter of 1 cm) and active carbon electrode as the positive and negative electrode, respectively. The filter papers soaked with 6 M KOH solution were taken as separators. As a electrochemical property's asymmetric supercapacitor, the charge stored between the two electrodes should keep the balance relationship (q<sup>+</sup> = q <sup>−</sup>), which could be calculated by equation (Liu Y. et al., 2018).

$$\mathbf{q} = \mathbf{C} \bullet \mathbf{m} \bullet \Delta \mathbf{V} \tag{2}$$

where C represents the specific capacitance (F g−<sup>1</sup> ), m is the mass of active materials on both electrodes (g), 1V is the potential window (V). Therefore, the mass ratio of electroactive material between the two electrodes could be calculated by equation (Li et al., 2019).

$$\text{m}^+/\text{m}^- = \text{C}^- \Delta \text{V}^- / \text{C}^+ \Delta \text{V}^+ \tag{3}$$

where C <sup>−</sup> (F g−<sup>1</sup> ) and 1V <sup>−</sup> (V) are the specific capacitance and the voltage range of scanning segment of the AC electrode, respectively. C<sup>+</sup> (F g−<sup>1</sup> ) and 1V <sup>+</sup> (V) are the specific capacitance and the voltage range of scanning segment of the Cu2O/CuO@Cu-CCs electrode. The specific capacitance, energy density and power density of the ACS device were calculated using the following equations (Ensafi et al., 2018; Liu Y. et al., 2018).

$$\text{Cs} = \text{(2I} \int \text{Vdt})/(\text{mV}^2) \tag{4}$$

$$\mathcal{E} = 1/2\mathcal{C}\Delta\mathcal{V}^2\tag{5}$$

$$\mathbf{P} = \mathbb{E}/\Delta \mathbf{t} \tag{6}$$

where Cs represents the specific capacitance (F g−<sup>1</sup> ), I is the discharge current (A), 1V is the potential window (V), 1t is the discharge time (s), m is the mass of active materials on both electrodes (g), E and P correspond to the energy density (Wh kg−<sup>1</sup> ) and power density (W kg−<sup>1</sup> ), respectively (Guan et al., 2017).

#### RESULTS AND DISCUSSION

The schematic illustration of the growth process of Cu2O/CuO@Cu-CCs electrodes is shown in **Figure 1**. Firstly, the Cu film is uniformly coated on the CCs through electroless copper plating, forming Cu-CCs samples. Subsequently, Cu(OH)<sup>2</sup> nanosheet arrays are in situ grown on Cu film by alkaline oxidative enchanting in NaOH and (NH4)2S2O<sup>8</sup> solution, during which the oxidative S2O 2− 8 is attached on the surface of Cu-CCs, and partial CuO are oxidized to Cu2+(Chen et al., 2016). With the reaction of Cu2<sup>+</sup> and OH−, Cu(OH)<sup>2</sup> nanosheet arrays are formed and then are decomposed into Cu2O/CuO nanosheets by thermal dehydration, thus obtaining Cu2O/CuO@Cu-CCs electrodes

XRD patterns of CCs, Cu-CCs and Cu2O/CuO@Cu-CCs are shown in **Figure 2a**. As can be seen by comparing peaks of CCs and Cu-CCs, the Cu films on CCs leads to the decrease of the characteristic peaks of carbon fiber at 2θ = 26.4◦ . For Cu-CCs samples, there are two strong diffraction peaks at 2θ = 43.5 and 50.6◦ , corresponding to the (111) and (200) planes of the metallic copper (JCPDS no. 04-0836), respectively (Chen et al., 2016). After the heat treatment at 120◦C, the sample exhibits four peaks at 2θ of 35.6, 36.4, 39.1, 42.3◦ , in which 2θ = 35.6 and 39.1◦ correspond to (−111) and (200) planes of the CuO substrates (JCPDS no. 48-1548), while the else two peaks (2θ = 36.4 and 42.3◦ ) are attributed to the (111) and (200) reflections of Cu2O (JCPDS no. 05-0667). It is worthy to note that Cu and CCs peaks are still observed, therefore, the composition is confirmed to be Cu2O/CuO@Cu-CCs.

Shown in **Figure 2b** is the SEM image of bare CCs, it can be observed that the surface of the carbon fibers is smooth and

FIGURE 1 | Schematic illustration of the growth process of Cu2O/CuO nanosheets on Cu coating carbon cloth to prepare Cu2O/CuO@Cu-CCs electrodes.

FIGURE 2 | (a) XRD patterns of bare CCs, Cu-CCs, and Cu2O/CuO@Cu-CCs samples. SEM images of bare (b) CCs, (c) Cu-CCs, and (d) Cu2O/CuO@Cu-CCs samples. (e) TEM image and (f) HRTEM image of Cu2O/CuO nanosheets.

the diameter is around 8–10µm. In **Figure 2c**, the carbon fibers are uniformly coated with copper films. After being etched in alkaline solution, the morphologies of Cu2O/CuO nanosheets vary with the change of etching time (**Supplementary Figure 1**). With an etching time of 25 min, the sample exhibits a highly porous cross-linked structure with abundant thin Cu2O/CuO nanosheets (**Figure 2d**). The lamellar nanosheets can effectively increase the number of active sites, which may be beneficial for promoting charge transfer and redox reaction (Liu Y. et al., 2018). As shown in **Figure 2e**, the porous Cu2O/CuO nanosheets are ultra-thin, which may enlarge the specific surface area to accelerate the intercalation and de-intercalation of ions (Chen et al., 2016). In addition, the HRTEM image in **Figure 2f** shows that the measured interplanar spacing of 0.250 and 0.226 nm for the well- defined lattice fringes are consistent well with the (−111) and (200) plane of CuO (JCPDS no. 48-1548), and there is a part of interplanar distances calculated to be 0.214 nm, which can be directed as the (200) plane of Cu2O (JCPDS no. 05-0667)

The XPS spectrums of the surface atomic composition and chemical state of the Cu2O/CuO@Cu-CCs samples are obtained by Gaussian curve-fitting. As illustrated in **Figure 3A**, the complete spectrum indicates the existence of C, Cu, and O elements in the sample. As shown in **Figure 3B**, there are two sharp peaks located at 932.43 and 952.5 eV, which are correspond to Cu 2p3/2 and Cu 2p1/2, respectively, illustrating the coexistence of Cu<sup>+</sup> and Cu<sup>0</sup> species (Wan et al., 2017). At the same time, there are three satellite peaks with binding energies of 934.2, 943.1, and 953.9 eV indicated the existence of CuO in the samples. Therefore, it can be concluded that the copper is mainly Cu2+, Cu+, and Cu<sup>0</sup> (Liu Y. et al., 2018). The existence of Cu<sup>0</sup> can be beneficial for improving the electronic conduction of electrodes. As shown in **Figure 3C**, the O 1s XPS spectrum can be deconvoluted into two peaks, one is the peak at 530.3 eV, which represents the oxygen in Cu2O lattice. Another is the high intensity peak at 531.1 eV, which is attributed to the CuO. This result further confirms the coexistence of CuO and Cu2O (Singh and Sarkar, 2017). In the **Figure 3D**, C 1s spectrum shows a high intensity peak at 284.5 eV, demonstrating that the intensity of C-C functional group peak is notably strong, further illustrating carbon fiber is stable in Cu2O/CuO@Cu-CCs.

The CV curves of CCs, Cu-CCs and Cu2O/CuO@Cu-CCs at a scan rate of 30 mV s−<sup>1</sup> are shown in **Figure 4A**. It is obvious

that strong pair of anodic and cathodic peaks is clearly visible for Cu2O/CuO@Cu-CCs and Cu-CCs samples, mainly due to the Faradaic redox reaction (Dong et al., 2014). For Cu-CCs, copper ions mainly derive from the reaction of Cu and KOH electrolyte solution during electrochemical measurement. The pronounced pseudocapacitive characteristic of Cu2O/CuO@Cu-CCs is mainly attributed to the porous cross-linked Cu2O/CuO nanosheets while the contribution of capacitance for CCs can be negligible. **Figure 4B** shows the GCD curves of different electrodes at a constant current density of 10 mA cm−<sup>2</sup> . The non-linear behavior of GCD curves further verifies that the main sources for charge storage originate from Faradaic reactions. The Cu2O/CuO@Cu-CCs electrode discussed above is the sample etched for 25 min (CV and GCD curves of other samples are shown in **Supplementary Figures 2A,B**), and this sample shows the best pseudocapacitive characteristic with a specific capacitance of 1.71 F cm−<sup>2</sup> (835.2 F g−<sup>1</sup> ) at 10 mA cm−<sup>2</sup> (3.57 A g −1 ) (**Figure 4C**), which is outperform the previously published values of Cu2O/CuO-based electrodes (1.674 F cm−<sup>2</sup> , equivalent to 594.27 F g−<sup>1</sup> , at 2 mA cm−<sup>2</sup> ; 839.9 Fg−<sup>1</sup> , at 1 mVs−<sup>1</sup> ; 357 F g −1 , at 10 A g−<sup>1</sup> ) and more exhaustive data were displayed in **Supplementary Table 4**. The EIS analysis was studied to further clarify the electrochemical behaviors of different electrodes. The Nyquist diagrams are shown in **Figure 4D**, which consist of an approximate semicircle in the high-frequency region and a line in the low-frequency region. All real-axis intercepts are as low as approximately 0.5 , illustrating all the samples have excellent electronic conduction due to the CCs and Cu-CCs current collectors. The depressed semicircle at the high frequency region corresponds to charge transfer resistance (Rct) caused by

Faradaic reactions (Ensafi et al., 2018). The Cu2O/CuO@Cu-CCs electrode has the smallest semicircle, illustrating an enhanced charge transfer. Also, the straight line in low-frequency region can be ascribed to Warburg impedance related to the fast charge diffusion in the electrolyte (Ensafi et al., 2018).

The electrochemical performances of Cu2O/CuO@Cu-CCs at various scan rates and current densities (**Figures 4E,F**) demonstrate a perfect reversibility during the charge-discharge process. Clearly, the slope of GCD curves decline suddenly at 0.18–0.25 V in charge part and the same as discharge part, corresponding the pseudocapacitance behavior in the CV scans, which is associated with the Faradaic redox reactions of Cu2+/Cu<sup>+</sup> redox pairs related to OH<sup>−</sup> as bellows (Guan et al., 2017; Sami et al., 2017).

$$\begin{aligned} \text{CuO} + \text{H}\_2\text{O} + 2\text{e}^- &\leftrightarrow \text{Cu}\_2\text{O} + 2\text{OH}^- \\ \text{Cu}\_2\text{O} + \text{H}\_2\text{O} + 2\text{OH}^- &\leftrightarrow 2\text{Cu(OH)}\_2 + 2\text{e}^- \\ \text{CuOH} + \text{OH}^- &\leftrightarrow \text{Cu(OH)}\_2 + \text{e}^- \\ \text{CuOH} + \text{OH}^- &\leftrightarrow \text{CuO} + \text{H}\_2\text{O} + \text{e}^- \end{aligned}$$

Remarkable, with the current density increases from 5 to 30 mA cm−<sup>2</sup> , the GCD curves present a gradually decreased discharge time but tends to preserve similar shape (**Figure 4F**) and the electrode retains 68.5% of its capacitance (**Figure 4G**), suggesting an excellent rate capability. Furthermore, the Cu2O/CuO@Cu-CCs electrode delivers excellent cycling stability with only 14.4% loss in specific capacitance after 5,000 cycles at 5 mA cm−<sup>2</sup> (**Figure 4H**), which can be explained by the stable structure of electrodes after cycling (the inset in **Figure 4H**).

For further exploring of the application, the electrochemical performances of the ASC device are investigated. As shown in **Figure 5A**, the device is sandwiched with the Cu2O/CuO@Cu-CCs positive electrode, active carbon negative electrode, and diaphragm separator soaked with 6 M KOH aqueous solution. **Figure 5B** shows the exactly complementary potential windows range of simple AC and Cu2O/CuO@Cu-CCs electrode, which suggest the high potential window of the ACS device. Furthermore, the calculated mass ratio of the electroactive materials of negative and positive electrodes according to Equation (3) is about 1:20. **Figures 5C**,**D** show the CV curves at a scan rate of 30 mV s−<sup>1</sup> and GCD curves at a current density of 1 A g−<sup>1</sup> with different potential windows, respectively. It is obvious that the shapes of the CV curves stay nearly same at different potential windows and the maximum potential window is extended to 1.6 V. The perfect symmetry and nearly unchanged shapes at different potential windows of GCD curves

(the inset shows the SEM images before and after 5,000 cycles) (H) of Cu2O/CuO@Cu-CCs electrode.

also contribute to the outstanding capacitive performance of this ASC device.

**Figure 5E** shows the CV curves of the ASC device at a scan rate ranging from 10 to 100 mV s−<sup>1</sup> . Apparently, the excellent synergy effect of the two electrodes leads to the high operation voltage of 1.6 V, which is three times as wide as the potential window of Cu2O/CuO@Cu-CCs electrode in the threeelectrode system. Meanwhile, the curve shape retains the same at different scan rates. The GCD curves at current densities from 0.5 to 10 A g−<sup>1</sup> are shown in **Figure 5F**. It is obvious that very low voltage drops are visible compared with threeelectrode test even at high current densities. And the symmetrical shape indicates high reversibility of the device. Thus, the device shows excellent rate capability (**Supplementary Figure 3**). In addition, owing to the broad potential window and huge specific capacitance, the ASC device shows a high energy density of 60.26 Wh kg−<sup>1</sup> at a power density of 299.73 W kg−<sup>1</sup> , higher

than some other literatures (**Figure 5G**). In order to investigate the long-term cycling stability and durability of the device, we performed 5,000 continuous GCD cycles at a current density of 2 A g−<sup>1</sup> . The ASC device exhibits an excellent cycling stability with keeping 90.35% in its specific capacitance after 5,000 GCD cycles (**Figure 5H**). This kind of electrode material will be a promising electrode for further engineering all-solid-state high-performance supercapacitor due to its excellent capacitor performance and flexibility characteristic.

# CONCLUSIONS

In short, we constructed Cu2O/CuO@Cu-CCs electrodes by a simple process with electroless copper plating, chemical etching and thermal dehydration. The ASC device with Cu2O/CuO@Cu-CCs positive electrode and AC negative electrode showed high energy density of 60.26 Wh kg−<sup>1</sup> at a power density of 299.73 W kg−<sup>1</sup> using 6 M KOH aqueous solution as the electrolyte. Also, the ASC device express an excellent cycling stability with keeping 90.35% in its specific capacitance after 5,000 GCD cycles. Also, this kind of electrode material will be a promising electrode for further engineering all-solid-state high-performance supercapacitor.

## DATA AVAILABILITY

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

#### REFERENCES


## AUTHOR CONTRIBUTIONS

LX synthesized Cu2O/CuO@Cu-CCs samples and analyzed part of characterizations. JL was the supervisor of this research work. HS and XG helped with the data analysis. JX analyzed XPS measurements. HZ organized a part of the data. XZ supplemented a part of the experiment.

#### FUNDING

This work was supported by A Project of Shandong Province Higher Educational Science and Technology Program (Grant No. J18KA002), National Natural Science Foundation of China (Grant No. 51702189), and Natural Science Foundation of Shandong Province (Grant No. ZR2017BEM033).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00420/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Xu, Li, Sun, Guo, Xu, Zhang and Zhang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Construction of Ultrathin Nitrogen-Doped Porous Carbon Nanospheres Coated With Polyaniline Nanorods for Asymmetric Supercapacitors

#### Pingping Yu<sup>1</sup> , Qunliang Wang<sup>1</sup> , Lingxia Zheng<sup>2</sup> and Yanfeng Jiang<sup>1</sup> \*

*<sup>1</sup> Department of Electronic Engineering, College of Internet-of-Things, Jiangnan University, Wuxi, China, <sup>2</sup> Department of Applied Chemistry, Zhejiang University of Technology, Hangzhou, China*

#### Edited by:

*Wenyao Li, Shanghai University of Engineering Sciences, China*

#### Reviewed by:

*Linfeng Hu, Fudan University, China Anqi Ju, Donghua University, China*

\*Correspondence: *Yanfeng Jiang jiangyf@jiangnan.edu.cn*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *14 May 2019* Accepted: *07 June 2019* Published: *26 June 2019*

#### Citation:

*Yu P, Wang Q, Zheng L and Jiang Y (2019) Construction of Ultrathin Nitrogen-Doped Porous Carbon Nanospheres Coated With Polyaniline Nanorods for Asymmetric Supercapacitors. Front. Chem. 7:455. doi: 10.3389/fchem.2019.00455* Porous carbon materials produced by biomass have been widely studied for high performance supercapacitor due to their abundance, low price, and renewable. In this paper, the series of nitrogen-doped hierarchical porous carbon nanospheres (HPCN)/polyaniline (HPCN/PANI) nanocomposites is reported, which is prepared via *in-situ* polymerization. A novel approach with one-step pyrolysis of wheat flour mixed with urea and ZnCl<sup>2</sup> is proposed to prepare the HPCN with surface area of 930 m<sup>2</sup> /g. Ultrathin HPCN pyrolysised at 900◦C (∼3 nm in thickness) electrode displays a gravimetric capacitance of 168 F/g and remarkable cyclability with losing 5% of the maximum capacitance after 5,000 cycles. The interconnected porous texture permits depositing of well-ordered polyaniline nanorods and allows a fast absorption/desorption of electrolyte. HPCN/PANI with short diffusion pathway possesses high gravimetric capacitance of 783 F/g. It can qualify HPCN/PANI to be used as cathode in assembling asymmetric supercapacitor with HPCN as anode, and which displays an exceptional specific capacitance of 81.2 F/g. Moreover, HPCN/PANI//HPCN device presents excellent cyclability with 88.4% retention of initial capacity over 10,000 cycles. This work will provide a simple and economical protocol to prepare the sustainable biomass materials based electrodes for energy storage applications.

Keywords: sustainable sources, ultrathin porous nanospheres, polyaniline nanorods, nitrogen-doped, supercapacitors

# INTRODUCTION

Carbon materials as the developing anodes have played pivotal roles in the energy storage area owing to its abundant sources, cost-effective, high chemical stability, and good conductivity (Gao and Fang, 2015; Benzigar et al., 2018). The excellent properties endow carbon materials (carbon nanotubes (CNT), activated carbon (AC) and graphene) with promising potential applications in the electric double-layer capacitance (EDLC) supercapacitors (Wu et al., 2017; Jiang et al., 2018). The electrochemical performances of the supercapacitors largely depend on characteristics of these adopted carbon materials, including morphologies, surface properties, specific surface area, and porous channels (Peng et al., 2014; Liu et al., 2016; Goldfarb et al., 2017; Li et al., 2018; Yang et al., 2018). Therefore, carbon materials with accessible surface area as well as high electron transport efficiency are essential for preparing the carbon-based supercapacitors. The materials with hierarchical porous nanostructure are intensively investigated recently (Badwal et al., 2014; Xu et al., 2018), which could be the promising candidate to satisfy the multiple requirements for the supercapacitors.

Activated carbon is generally prepared by pyrolyzing versatile carbonaceous precursors with physical or chemical activation to introduce the hierarchical pores. However, when the surface area increases to 3,000 m<sup>2</sup> g −1 , its electric capacity is still low (100–140 F/g) in the organic electrolyte (Barbieri et al., 2005). The interplayed aspects indicate that the capacity depends on the pore size distribution, accessibility of the electrolyte and the electrical conductivity (Zhang H. et al., 2015). Therefore, numerous biomass-based carbon combining hierarchical pores and heteroatom doping have gained considerable attentions, resulting in pesudocapacitance and electrical double layer capacitance. Carbonization of natural biomass materials, including seaweeds (Kang et al., 2015), peanut shell (Ding et al., 2015), rice bran (Hou et al., 2014), plant leaves (Liu B. et al., 2017; Zhang et al., 2018; Zhao et al., 2018), fruit (Wu et al., 2014), and wheat flour (Wu et al., 2015; Yu et al., 2016), has been investigated. It is demonstrated as a feasible approach, which could result in a low cost and eco-friendly way to prepare hierarchical porous carbon electrodes.

Among all the kinds of the investigated biomass, wheat flour is considered as the most promising one. Wheat flour can be well dispersed in aqueous solution via magnetic stirring to form suspension because it contains starch (72–80%) and protein (8– 10%). Thus, wheat flour is a green carbon source to prepare the hierarchical porous carbon. Chemical activation is a preferable approach to prepare the high performance porous carbon using the chemical agents (KOH, NaOH, ZnCl<sup>2</sup> and H3PO4) (Duan et al., 2016; Lei et al., 2016; Pang et al., 2016a,b; Goldfarb et al., 2017; Yang et al., 2019). The strong activation effect of KOH and NaOH can degrade the wheat flour into small molecular and lose a large proportion of carbon atoms, making the low yield of carbon materials (Wang and Liu, 2017). ZnCl<sup>2</sup> as a milder activation reagent can react with precursor by dehydrating and cross-linking along the temperature increase, creating hierarchical pores, high specific surface area and high yield.

Polyaniline (PANI) is considered as a good developing supercapacitor electrode due to its large pseudocapacitance, intrinsic redox states and simple synthesis (Yu et al., 2013a, 2014, 2017). However, because of the degradation of PANI electrodes in the process of repeated charging and discharging, the cycle life of PANI electrodes is usually poor (Liu et al., 2014). In order to overcome this shortcoming, combination of the wellordered PANI nanostructure and carbon materials has been explored to prevent the collapse. Wang et al. reported that the ordered whiskerlike PANI/mesoporous carbon composites formed "V-type" channels which shorten the ion transport pathway and increase the contacted area for electrolyte (Wang et al., 2006). Liu et al. synthesized hierarchical graphene/PANI hollow microspheres hybrid electrode with supercapacitor capacitance of 446.2 F/g, which exhibits an outstanding long cycle life (Liu L. et al., 2017). Sulfur-encapsulated porous carbon nanospheres/polyaniline composites were synthesized to improve chemical stability and electronic conductivity (Li et al., 2016). However, these PANI-based binary composites exhibit single-scale pores, complex synthetic procedure, and high cost, which is not suitable for the high-performancesupercapacitor.

In this study, the one-step carbonization for wheat flour is employed by pyrolysis of urea and ZnCl<sup>2</sup> to prepare interconnected hierarchical porous N-doped carbon nanospheres (HPCN). **Figure 1** presents the approach for the fabrication of the HPCN/PANI composites by chemical oxidative polymerization. The highest specific capacitance of HPCN with surface area of 930 m<sup>2</sup> /g is 168 F/g, and then PANI nanorods are vertically coated on the HPCN by chemical bonding interaction with nitrogen groups, providing the significant enhancement of supercapacitor performances. Moreover, this assembled asymmetric supercapacitor (ASC) combining HPCN and HPCN/PANI can work in 1 M H2SO<sup>4</sup> electrolyte and its electrochemical performances are investigated in this paper.

# EXPERIMENTAL SECTIONS

## Synthesis of Hierarchical Porous Hpcn

HPCN were formed using wheat flour as renewable biomass resource through one-step carbonization. Typically, the waste wheat flour, urea and ZnCl<sup>2</sup> (1:1:2, w/w) were added to 50 mL distilled water with magnetic stirring, followed by carbonization at 700–900◦C for 1 h under N<sup>2</sup> flow. As-carbonized samples were stirred in 10% HCl and DI water successively for 15 h. After purification, the HPCN was dried under 80◦C vacuum oven. The obtained products were termed as HPCN7, HPCN8, and HPCN9 standing for the pyrolysis temperature at 700, 800, and 900 ◦C, respectively.

# Fabrication of HPCN/PANI

The HPCN (100 mg) were dispersed in 1 M H2SO<sup>4</sup> solution containing aniline monomers (AN, 45.65 µL) by strong stirring. 40 mL 1 M H2SO<sup>4</sup> of ammonium persulfate (APS, 114.12 mg, [AN]/[APS] = 1:1) solution was rapidly added with stirring for more than 5 min. The oxidative reaction was conducted at −5 ◦C overnight, and the obtained composites were filtered, rinsed with DI water and overnight dried at 50◦C. The content of PANI is 26.21% in HPCN7/PANI, 29.34% in HPCN8/PANI and 32.05% in HPCN9/PANI, respectively.

# Material Characterization

The composites were conducted with a field-emission scanning electron microscope (FESEM, Zeiss Sigma), a transmission electron microscope (TEM, TECNAI G<sup>2</sup> S-TWIN), X-ray diffraction (XRD, Bruker D8-A25, λ = 1.5405 Å, Cu Kα radiation, 10-50◦ ), Raman spectra (LabRam-1B, 632.8 nm laser), X-ray photoelectron spectroscopy (XPS, PHI 5000C ESCA) and ASAP 3000.

#### Electrochemical Measurements

The mixture was blended with ratio of 90/5/5 for as-prepared samples/carbon black/ polytetrafluoroethylene. The uniform paste was coated on a titanium mesh (1 cm in diameter) as working electrode. The HPCN and HPCN/PANI were tested in three-electrode cell. The asymmetric supercapacitors were fabricated with negative electrode of HPCN9 and positive electrode of HPCN9/PANI. The filter paper were placed between two electrodes to be the separator. The mass of active materials deposited on Ti mesh was approximated to be 2.0 mg. Electrochemical performances of electrodes were conducted with an electrochemical workstation of CHI 660D in electrolyte of 1 M H2SO<sup>4</sup> aqueous.

According to the GCD plots, the specific capacitance of the HPCN and HPAN/PANI electrodes can be calculated depending on equation (1):

$$C\_{sp} = \frac{I\Delta t}{m\Delta V} \tag{1}$$

where I (A) is the current, 1t (s) stands for the time required for full discharge, m (g) is the weight of the electrodes, and 1V (V) is the voltage difference. In order to keep the charge balance, the mass ratio of positive and negative electrode was calculated using the following formula (2):

$$\frac{m\_+}{m\_-} = \frac{C\_- V\_-}{C\_+ V\_+} \tag{2}$$

The specific capacitance (Casy), energy density (E, Wh/kg) and power density (P, W/kg) of ASC device were obtained by the following equations:

$$C\_{\rm asy} = \frac{I \Delta t}{M \Delta V} \tag{3}$$

$$\mathcal{E} = \frac{\mathcal{C}\_{\text{asy}} \Delta V^2}{2 \times 3.6} \tag{4}$$

$$P = 3600 \times \frac{E}{\Delta t} \tag{5}$$

where M (M = m++m−) is the total mass of electrode materials.

#### RESULTS AND DISCUSSION

The morphologies of HPCN7, HPCN8, and HPCN9 (carbonized at 700, 800 and 900◦C) indicate that the samples consist of a large amount of uniformly interconnected carbon nanospheres (**Figures 2A–C**). The diameter of the HPCN7 (**Figure 2A**) is 70 nm, which is higher than that of HPCN8 (61 nm, **Figure 2B**) and that of HPCN9 (50 nm, **Figure 2C**) because carbonization temperature increases from 700 to 900◦C. In this chemical activation process, ZnCl<sup>2</sup> as dehydrating agent can promote the condensation reaction of wheat flour, reduce the gasification of carbon atoms and lead to high carbon yields (>35%). ZnCl<sup>2</sup> is also as an effective activation-graphitization agent that promotes aromatic condensation reactions at high temperature due to its Lewis acid nature and introduce the porous structure to improve the surface area. The nanospheres are interconnected to form the porous network. This structure has benefit for fast electrode/electrolyte interface kinetics.

The porous contour of the HPCN can be used as a buffer boost for expansion and shrinking of PANI nanorods. PANI nanorods are vertically deposited on the HPCN7, HPCN8, and HPCN9 to improve the specific capacitance, as shown in **Figures 2D–F**. Compared to HPCN9/PANI (**Figure 2F**), there are obvious denser and longer PANI nanorods stacked on HPCN7/PANI (∼50 nm, **Figure 2D**) and HPCN8/PANI (∼40 nm, **Figure 2E**) than those in HPCN9/PANI. This shows that the thicknesses of the PANI layers of HPCN7/PANI and HPCN8/PANI are higher than that of the HPCN9/PANI. This may be due to the HPCN7 and HPCN8 with more nitrogen-groups and higher specific surface area to

be bonded with much more aniline monomers to form the PANI nanorods.

The above analysis shows that HPCN9/PANI has thinner PANI layer. The average diameter of carbon nanospheres in HPCN9 is 50 nm. Their thin carbonaceous walls are connected to form the reticular morphology (**Figure 3A**). Plenty of nanopores can be observed in the HPCN9. The thickness of these carbonaceous walls (**Figure 3B**) is about 3 nm equal to 6–8th layer of graphene. For the sample HPCN9/PANI (**Figure 3C**), the PANI nanorods are grown on the porous carbon nanospheres, maintaining the interconnected network structure. **Figure 3D** presents that the diameter and the length of the PANI nanorods are about 5–10 and 15–17 nm, respectively. The well-ordered and much smaller sized PANI in HPCN9/PANI can largely short the diffusion pathway by enhancing utilization of the active materials, improving the electrochemical performances.

**Figure 4A** shows XRD profiles of the HPCN and HPCN/PANI composites. As-prepared HPCN7, HPCN8, and HPCN9 all show two characteristic diffraction peaks (002) and (101) at around 24◦ and 44◦ , respectively. The (002) diffraction with broad width and low intensity of (101) plane indicate the amorphous carbon structure (Yu et al., 2016). Three crystalline peaks located at 2θ = 14.8◦ , 25.2◦ , and 20.3◦ for the three-HPCN/PANI composites are assigned to be the crystal planes of (011), (200), and (020) for emeraldine salt polyaniline, indicating the periodicity of the perpendicular and the parallel of the polymer chains, respectively, (Yu et al., 2013a).

The Raman spectra of HPCN7, HPCN8, and HPCN9 (**Figure 4B**) exhibit the strong D-peak at 1,344 cm−<sup>1</sup> , representing disordered carbon confirmed with XRD results (Kudin et al., 2008). G-band at 1,597 cm−<sup>1</sup> indicates graphitic carbon with vibration of sp<sup>2</sup> C. The value of ID/I<sup>G</sup> for HPCN7, HPCN8, and HPCN9 is 1.04, 1.03, and 1.01, respectively. It indicates that the HPCN9 has relatively higher graphitized degree, offering good electric conductivity. There are other additional four characterization peaks for HPCN7/PANI, HPCN8/PANI, and HPCN9/PANI. The peak at 1,175 and 1,482 cm−<sup>1</sup> is assigned to in-plane C-H bending and C = N stretching of quinoid ring, respectively. At 1356 and 1,593 cm−<sup>1</sup> , the peaks indicate protonated C-N stretching vibration and C-C

stretching of benzenoid ring. It shows PANI nanorods in the emeraldine form (Yu et al., 2013a,b; Chen et al., 2016; Zheng et al., 2016).

The XPS spectra demonstrate the fact that the nitrogen atoms exist in the HPCN9 and HPCN9/PANI in **Figure 4C** observed the C, N, and O 1s peaks. The deconvoluted C 1s peaks of HPCN9 reveal 284.7 eV of the sp<sup>2</sup> C=C bond of graphitic carbon and 285.8 eV of sp<sup>3</sup> C-C/C-H bonds while the one at 287.1 eV attributes to C=O/N-C=N bonds (**Figure S1a**). The concentration of nitrogen species is 2.4 at.% and 6.7 at.% for HPCN9 and HPCN9/PANI, respectively (**Tables S1, S2**). The N 1s spectra (**Figures S1b,c**) of HPCN9 are deconvoluted into three bands: 401.2, 399.8 and 398.6 eV, which is corresponding to quaternary N, pyrrole N and pyridine N, respectively (Li et al., 2015; Wang et al., 2015). It indicates the O/N functional groups originated form wheat flour and urea during the carbonization process. For HPCN9/PANI, three peaks centered at 401.3, 399.6, and 398.2 eV indicate the positively charged nitrogen atoms (-NH<sup>+</sup> =), benzenoid amine (-NH-) and quinoid amine (-N=) in the deconvoluted N 1s spectra (**Figure S1d**), respectively (Yu et al., 2013a; Zhang Y. et al., 2015; Wang and Liu, 2017).

**Figure 4D** is the N<sup>2</sup> adsorption/desorption isotherms of the HPCN and HPCN/PANI. It can be seen that the isotherms of HPCN and HPCN/PANI are I/IV type adsorption/desorption. At the relative low pressure, there is a fast and distinct adsorption, while it shows slight hysteresis loop at P/P<sup>0</sup> of 0.3–1.0. This suggests that HPCN and HPCN/PANI contains micropores and mesopores. In addition, the increment in adsorption quantity at P/P<sup>0</sup> of 1.0 is caused by the small amount of macropores. The specific surface area (**Figure S2**) follow the trend: HPCN9 (978 m<sup>2</sup> /g) > HPCN8 (965 m<sup>2</sup> /g) > HPCN7 (930 m<sup>2</sup> /g). HPCN9/PANI has superior surface area of 639 m<sup>2</sup> /g to HPCN7/PANI (580 m<sup>2</sup> /g) and HPCN8/PANI (602 m<sup>2</sup> /g), which is higher than the reported carbon materials and PANI hybrids (Liu L. et al., 2017; Yu et al., 2017), promising to achieve favorable capacitor performance. The decreased pore volumes and specific surface area suffer from pore blockage after coating PANI nanorods by the in-situ polymerization, while the average pore diameter of HPCN9/PANI is still 1.0 nm (inset in **Figure 3D**).

The electrochemical characterizations of the HPCN and HPCN/PANI electrodes are demonstrated by CV curves, GCD and EIS in a three-electrode setup based on the −0.2–0.8 V. **Figure 5A** presents the nearly rectangular CV curves of HPCN9 electrode with scanning rate of 5–200 mV/s. There is a slight distortion because of the interconnected hierarchical porous nanospheres and nitrogen atoms doping from urea, indicating the good rate performance. It exhibits the electrochemical double layer capacitive features. The CV curve of HPCN9 possesses the higher integrated areas than those of HPCN7 and HPCN8 at 100 mV/s (**Figure 5B**), revealing the higher charge storage capabilities of HPCN9. Therefore, the hierarchical porous HPCN can be the good scaffold for the PANI nanorods loading. The enhanced area of CV curves is attributed to

the well-ordered porous HPCN/PANI structure. Two new pairs of intense peaks in HPCN7/PANI, HPCN8/PANI, and HPCN9/PANI are assigned to the redox reactions of PANI, implying the main chain change of PANI as semiconducting leucoemeraldine/conducting polaronic emeraldine and faradaic transformation of emeraldine/pernigraniline, respectively (Yu et al., 2013b). The HPCN9/PANI shows the higher specific capacitance than HPCN7/PANI and HPCN8/PANI based on the areas of the CV curves, consistent with the GCD plots (**Figure 5C**). Different from the symmetric triangle plot for HPCN, there are voltage plateau and longer discharge time in the GCD plots of HPCN/PANI at 1 A/g, suggesting the pseudocapacitive characteristics. Obviously, the HPCN9/PANI possesses the optimal capacitive performance owing to the longer discharge time. Therefore, capacity performance of electrodes increases with the addition of PANI nanorods confirming by the GCD and CV tests. The maximum specific capacitance (Csp) of HPCN9, HPCN8, and HPCN7 obtained in **Figure 5D** is 168 F/g, 133 F/g, and 152 F/g, superior to the capacitance of previous published biomass-carbon electrodes (Zhang H. et al., 2015; Zhang Y. et al., 2015; Lei et al., 2016; Pang et al., 2016a,b). The HPCN9 shows the interconnected porous structure to help reducing the diffusion pathway of electrolyte ions and providing low-resistance to enhanced capability performances. The calculated Csp of HPCN9/PANI, HPCN8/PANI, and HPCN7/PANI (**Figure 5D**) at 1 A/g is 783

F/g, 751 F/g and 710 F/g, respectively. **Table 1** is the comparison of specific capacitances of HPCN/PANI and previously reported carbon/PANI composites. The exceptional capacitance mainly arises from the synergic effect of pseudocapacitance and doublelayer capacitance, due to the well-ordered PANI nanorods on the interconnected porous network structure. The nanometer size of PANI with "V-type" channels facilitates the large accessible surface area between electrolyte and active species for fast faradaic reactions and shorten diffusion pathway to ensure the effective utilization of PANI nanorods.

**Figure 6A** shows the Nyquist plots for HPCN and HPCN/PANI electrodes. HPCN7/PANI, HPCN8/PANI, and HPCN9/PANI electrodes exhibit small charge transfer resistance (Rct) corresponding to equivalent circuit (**Figure S3**), which is 0.75, 0.68, and 0.51 , respectively, higher than that of HPCN7 (0.42 ), HPCN8 (0.32 ), and HPCN9 (0.17 ), due to the PANI nanorods grown on the porous carbon nanospheres. The decreased Rct from HPCN7/PANI to HPCN9/PANI is attributed to the increased electrode/electrolyte interface area, facilitating the fast redox reactions in the as-prepared electrodes. Moreover, the almost straight line at low frequency of HPCN9/PANI plot shows ideal capacitive characteristic, indicating to the good rate performance. Apart from the high specific capacitance and small resistances, the HPCN, and HPCN/PANI electrodes also exhibit good cycling durability (**Figure 6B**). HPCN electrodes exhibits outstanding cycling behavior with losing



5% of initial capacitance over 5,000 cycles. HPCN9/PANI electrode remains 698 F/g and holds 89.3% of initial value after 5,000 cycles, while HPCN7/PANI and HPCN8/PANI show the capacitance retention of 85.1 and 86.5%, respectively. The fast capacitance decay may be caused by the collapse and swelling of PANI structure because of the doping/dedoping of electrolyte ions.

Furthermore, the HPCN9//HPCN9/PANI asymmetric supercapacitor (ASC) was fabricated by sandwiching a H2SO4/filter paper as the separator between negative (HPCN9) and positive electrode (HPCN9/PANI) (**Figure 7A**). The voltage window of HPCN9//HPCN9/PANI ASC device is 1.6 V. CV curves of HPCN9//HPCN9/PANI ASC device (**Figure 7B**) maintain the good roughly rectangular shape at 200 mV/s. The GCD curves (**Figure 7C**) exhibit a slight non-linearity and almost symmetric characteristic especially at the low current density for ASC device, indicating the contribution of the redox reaction from PANI. The equivalent series resistance (ESR) of ASC device is 1.4 (**Figure S4**), indicating a small contact resistance. The Csp of HPCN9//HPCN9/PANI asymmetric supercapacitor is 81.2 F/g at 1 A/g (**Figure 7D**), and remains 69.5% retention at 8 A/g.

To examine the long-term cycling stability of HPCN9//HPCN9/PANI ASC, GCD tests are measured to

the change of specific capacitance at 1 A/g (**Figure 8a**). HPCN9//HPCN9/PANI ASC shows 88.4% of initial capacitance after 10,000 cycling, lower than that of honeycomb-like porous carbon (HPC)//HPC/PANI (91.6% retention after 5,000 cycles) (Yu et al., 2016), due to the mechanical degradation of PANI in the acid electrolyte. However, the value is higher than graphene/PANI//graphene/RuO<sup>2</sup> ASC (30% decay over 2500 cycles) (Zhang et al., 2011) and activated graphene//PANI/MnO2/carbon cloth ASC (30% loss of initial value after 5,000 cycles) (Zhao et al., 2015). **Figure 8b** displays

HPCN and HPCN/PANI electrodes over 5,000 cycles at 1 A/g (B).

the Ragone plot of HPCN9//HPCN9/PANI ASC, which has the highest energy density of 31.2 Wh/kg at 860 W/kg and retains that value of 25.1 Wh/kg at 6.88 kW/kg. As far as authors know, the results are higher than recently reported AC//AC fiber/PANI (20 Wh/kg, 2.1 kW/kg) (Salinas-Torres et al., 2013), AC//SiC-N-MnO<sup>2</sup> (30.06 Wh/kg, 0.11 kW/kg) (Kim and Kim, 2014), mesoporous carbon/PANI (23.8 Wh/kg, 0.21 kW/kg) (Cai et al., 2010), AC/CNT//graphene/MnO2/CNT (27 Wh/kg, 7.8 kW/kg) (Cheng et al., 2013), graphene//MnO<sup>2</sup> (25.2 Wh/kg, 0.1 kW/kg) (Cao et al., 2013), and MnO2/graphene// grapheme

(30.4 Wh/kg, 5 kW/kg) (Wu et al., 2010). Furthermore, the morphology of ASC device is investigated after 10,000 cycles. The nanospheres of HPCN9 (**Figure S5**) still stack with each other and form a porous network structure after 10,000 cycles. Obviously, HPCN9/PANI (**Figure 8c**) remains well-constructed porous network structure with advantage for high efficient ion diffusion. The length of PANI nanorods decreases to be ∼8 nm owing to the degration during the long cycling process (**Figure 8d**). The excellent electrochemical performances of HPCN9//HPCN9/PANI ASC arise from the unique structure of well-ordered PANI nanorods grown on the hierarchical interconnected porous HPCN. Nitrogen heteroatom in HPCN enhance the surface wettability and active sites for electrolyte ions. The micropores in HPCN9 can offer high available interfacial areas; the mesopores make ions with high mobility to drift into the internal regions by efficient diffusion channels, while the interconnected network provides the macropores as the ion-buffering reservoirs, contributing to the excellent rate capacity. Moreover, the nanoscale PANI with high pseudocapacitance enhances supercapacitor performances of the integrated asymmetric device. Therefore, the synergistic effect of PANI and HPCN results in high electrochemical performances.

# CONCLUSIONS

In summary, the interconnected hierarchical porous N-doped HPCN with high conductivity is successfully synthesized in this paper, starting from the easy accessible raw materials, such as sustainable wheat flour, urea and ZnCl2. The prepared HPCN exhibits high specific surface area of 930 m<sup>2</sup> /g with rational micro-mesopore distribution, which also delivers a specific capacitance of 168 F/g with remarkable stability (5% decay of initial value over 5,000 cycles). The well-ordered PANI nanorods deposited on HPCN9 displays high specific capacitance of 783 F/g. As-fabricated HPCN9//HPCN9/PANI ASC device delivers specific capacitance of 81.2 F/g, maximum energy density and power density, and 11.6% loss of original capacitance over 10,000 cycles. This work will provide a possibility of sustainable biomass materials to be the supercapacitor electrodes using the facile and low-cost process in the modern society.

# DATA AVAILABILITY

All datasets generated for this study are included in the manuscript/**Supplementary Files**.

# AUTHOR CONTRIBUTIONS

PY designed the work and wrote the paper. QW performance the experiments. LZ polished the english of this manuscript. YJ was responsible for the drafting.

# FUNDING

This work was funded by the National Natural Science Foundation of China (No. 51802124 and No. 51702287) and Jiangsu Province (BK 20180626) and the Fundamental Research Funds for the Central Universities (JUSRP11858).

#### ACKNOWLEDGMENTS

The authors would like to thank the Engineering Research Center of Internet of Things Technology Applications Ministry of Education for kind support.

#### REFERENCES


# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00455/full#supplementary-material


high-performance supercapacitors. J. Power Sources 299, 347–355. doi: 10.1016/j.jpowsour.2015.09.018


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Yu, Wang, Zheng and Jiang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Highly Ordered Mesoporous NiCo2O<sup>4</sup> as a High Performance Anode Material for Li-Ion Batteries

Qilong Ren1†, Guangyu Wu<sup>2</sup> \* † , Weinan Xing<sup>2</sup> \*, Jiangang Han<sup>2</sup> , Pingping Li <sup>2</sup> , Bo Li <sup>3</sup> \*, Junye Cheng<sup>4</sup> , Shuilin Wu<sup>4</sup> , Rujia Zou<sup>1</sup> and Junqing Hu<sup>1</sup> \*

*<sup>1</sup> State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai, China, <sup>2</sup> College of Biology and the Environment, Nanjing Forestry University, Nanjing, China, <sup>3</sup> Department of Vascular Surgery, Shanghai Ninth People's Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China, <sup>4</sup> Center of Super-Diamond and Advanced Films, Department of Materials Science and Engineering, City University of Hong Kong, Hong Kong, China*

#### Edited by:

*Min Zeng, Lanzhou Institute of Chemical Physics (CAS), China*

#### Reviewed by:

*Hui Yang, Jiangxi University of Science and Technology, China Ming Ma, Shanghai Institute of Ceramics (CAS), China*

#### \*Correspondence:

*Guangyu Wu gywuchem@163.com Weinan Xing xingweinan@126.com Bo Li boli@shsmu.edu.cn Junqing Hu hu.junqing@dhu.edu.cn*

*†These authors have contributed equally to this work*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *20 June 2019* Accepted: *08 July 2019* Published: *23 July 2019*

#### Citation:

*Ren Q, Wu G, Xing W, Han J, Li P, Li B, Cheng J, Wu S, Zou R and Hu J (2019) Highly Ordered Mesoporous NiCo2O<sup>4</sup> as a High Performance Anode Material for Li-Ion Batteries. Front. Chem. 7:521. doi: 10.3389/fchem.2019.00521* The controlled synthesis of highly ordered mesoporous structure has attracted considerable attention in the field of electrochemistry because of its high specific surface area which can contribute the transportation of ions. Herein, a general nano-casting approach is proposed for synthesizing highly ordered mesoporous NiCo2O<sup>4</sup> microspheres. The as-synthesized mesoporous NiCo2O<sup>4</sup> microsphere materials with high Brunner-Emmett-Teller (BET) surface area (∼97.77 m2g −1 ) and uniform pore size distribution around 4 nm exhibited a high initial discharge capacity of ∼1,467 mAhg−<sup>1</sup> , a good rate capability as well as cycling stability. The superior electrochemical performance was mainly because of the highly porous nature of NiCo2O4, which rendered volume expansion during the process of cycling and shortened lithium-ions transport pathways. These properties showcase the inherent potential for use of highly ordered mesoporous NiCo2O<sup>4</sup> microspheres as a potential anode material for lithium-ion batteries in the future.

Keywords: highly ordered, mesoporous, NiCo2O4, lithium-ion batteries, nano-casting

# INTRODUCTION

Lithium-ion batteries (Yoo et al., 2008; Pan et al., 2013) have been proven to be viable alternatives to traditional energy storage devices and are vital tools when coupled with emerging renewable energy sources (Lewis and Nocera, 2006; Song, 2006; Chheda et al., 2007) (i.e., wind, solar, etc.). However, current energy demands outpace what is commercial systems are capable of, leading to the development of next-generation lithium-ion batteries (Liu et al., 2012; Lu et al., 2018; Shen et al., 2018). Particularly, the electrode is vital for optimal electrochemical performance (Toupin et al., 2004; Jiang et al., 2012; Wang et al., 2012; Yang et al., 2019a); however, conventional anode materials (Courtel et al., 2011; Chen et al., 2013; Chang et al., 2014) such as graphite have low theoretical specific capacities (372 mAh g−<sup>1</sup> ) and fail to satisfy energy storage demands. Consequently, transition metal oxides (TMOs) (Yuan et al., 2014; Tabassum et al., 2018) have been proposed as viable alternatives, attributed to high energy densities, which result in capacities more than double of those observed with graphite. However, this capacity relies on the morphology and structure of the TMO, which can suffer from undesired volume expansion and collapse of the structure as the battery cycles, leading to catastrophic failure.

Recently, space designed nanostructures have been acknowledged as an effective strategy to remit the volume expansion as well as shorten the lithium-ion transport pathways by providing a larger specific surface area, including implementation of a porous network (Wen et al., 2007; Lou et al., 2008; Shen et al., 2012, 2014; Yuan et al., 2012; Qie et al., 2013; Yang et al., 2019b), mesopores structures. For example, mesoporous Co3O<sup>4</sup> (Li et al., 2008), NiO (Yin et al., 2018), NiCo2O<sup>4</sup> (Li et al., 2018), and SnO<sup>2</sup> (Han et al., 2019) have been fabricated and have shown good electrochemical performance. However, limited attention has been focused to study ternary systems such as NiCo2O<sup>4</sup> despite a higher electrical conductivity and specific capacity owing to its enhanced chemical kinetics. Inspired by ammonium hydrogen carbonate-assisted solvothermal route to prepare Ni0.33Co0.67CO<sup>3</sup> microspheres (Li et al., 2013), they led to a recent breakthrough technique to form mesoporous microspheres, we present a modified structure design containing ordered mesopores to further improve the electrochemical performance of NiCo2O4.

Herein, we developed a template-assisted method to synthesize novel NiCo2O<sup>4</sup> microspheres containing highly ordered mesoporous structures and nanoparticles to effectively alleviate the huge volume expansion and enhance the electrical conductivity. The synthesis process is schematically shown in **Figure 1**. Firstly, mesoporous silica (KIT-6) is synthesized as a template by a typical approach which is illustrated in the experiment section in detail. Then the template is immersed in the mixed solution of Ni(NO3)<sup>2</sup> and Co(NO3)<sup>2</sup> to introduce the Ni2<sup>+</sup> and Co2<sup>+</sup> to be filled into KIT-6. After a subsequent calcination step at 450◦C under N<sup>2</sup> atmosphere, the NiCo2O4@KIT-6 is successfully prepared. Finally, NaOH solution is utilized to remove the template of KIT-6 in the NiCo2O4@KIT-6 and obtain mesoporous NiCo2O<sup>4</sup> microspheres. When acted as an anode material for Li-ion batteries, mesoporous NiCo2O<sup>4</sup> microsphere electrode exhibits the superior electrochemical performance, whose stable specific capacity was 430 mAhg−<sup>1</sup> after 100 cycles, which is better than that of the non-porous NiCo2O<sup>4</sup> (270 mAhg−<sup>1</sup> after 100 cycles). The improved lithium storage performance mainly benefits from the rationally designed mesoporous structures of NiCo2O4. We believe that this versatile strategy could be extended to more ternary TMO materials for the development of high property electrode in LIBs.

# EXPERIMENTAL SECTION

#### Materials

All chemicals (analytical reagent grade) used in this work, including polyethylene oxide-polypropylene oxidepolyethylene oxide (PEO-PPO-PEO, P123, MW = 5.8 K), hydrochloric acid, n-butanol, tetraethyl orthosilicate (TEOS), Co(NO3)2•6H2O, Ni(NO3)2•6H2O and ethanol were purchased from Sigma-Aldrich.

## Material Synthesis

#### Synthesis of Mesoporous SiO<sup>2</sup> (KIT-6)

The typical synthetic process was as follows (Zhou et al., 2018): 4.75 g of P123 were dissolved in a mixed solution containing 163 mL of deionized (DI) water and 7.5 ml of concentrated hydrochloric acid under stirring at 35◦C. Once the P123 was fully dissolved, 4.53 g of n-butanol were added. After 1 h, TEOS (9.675 g) were added into the above mixed solution. The solution was then stirred for an additional 24 h at 35◦C, followed by another 24 h incubation at 35◦C. Cooling the mixed solution to room temperature, the precipitates could be collected by centrifugation for 3 times and moved to a vacuum oven at 90◦C for 12 h. Then, after calcining the collected deposit at 550◦C for 6 h, white mesoporous SiO<sup>2</sup> was obtained.

#### Synthesis of Mesoporous NiCo2O<sup>4</sup> Microspheres

In a simple process, 0.4 g of KIT-6 were added to a solution containing 4 ml of 1 M Co(NO3)2·6H2O and 2 ml of 1 M Ni(NO3)2·6H2O in ethanol under stirring for 1.5 h at room temperature. Then, the solution was heated at 70◦C until the ethanol was completely evaporated, and the solid was calcined at 200◦C for 4 h. Immediately following, 2 mL of 1 M Co(NO3)2·6H2O and 1 mL of Ni(NO3)2·6H2O in ethanol were added following the previous steps and an additional calcination 450◦C for 6 h. To obtain mesoporous NiCo2O<sup>4</sup> microspheres, the powder was immersed in a 2 M NaOH solution to etch away the KIT-6 templates. Then, the samples were collected by centrifugation, washed for 3 times and moved to a vacuum oven at 60◦C for 12 h. As a control, conventional NiCo2O<sup>4</sup> microspheres were produced through the above-mentioned process without KIT-6 templates.

#### Materials Characterization

The mesoporous NiCo2O<sup>4</sup> microspheres were characterized by using a PANalytical X' Pert X-ray diffractometer (Holland), with Cu-Kα radiation at 40 kV and 40 mA, selected-area electron diffraction (SAED), scanning electron microscope (SEM, S-4800) and transmission electron microscope (TEM, JEM-2100F). The N<sup>2</sup> adsorption/desorption isotherms were used to calculate the specific surface area and Barrett-Joyner-Halenda (BJH) equation was used to calculate the pore size distribution and average pore diameter.

#### RESULTS AND DISCUSSION

The mesoporous NiCo2O<sup>4</sup> microspheres were prepared via nano-casting, with the crystalline structure and phase purity

FIGURE 2 | (a) XRD patterns of mesoporous and non-porous NiCo2O4. (b,c) high magnification SEM images of mesoporous NiCo2O4. (d) TEM image and (e,f) high-resolution TEM image of mesoporous NiCo2O4. Inset in (f) shows the corresponding SAED pattern.

characterized by X-ray diffraction (XRD). **Figure 2a** compares the diffraction patterns of both mesoporous and non-porous NiCo2O<sup>4</sup> microspheres, showing eight obvious peaks at 2θ values of 18.9, 31.1, 36.7, 38.4, 44.6, 55.4, 59.1, and 64.9 for the (111), (220), (311), (222), (400), (422), (511), and (440) planes, respectively, which consist with the cubic spinel NiCo2O<sup>4</sup> (JCPDS No.20-0781)without any apparent impurities.

The morphology and structure of obtained NiCo2O4are elucidated by SEM and TEM, as shown in **Figures 2b,c**, the surface morphology of the mesoporous microspheres, when compared to nonporous structures (**Figure S1**), is significantly rougher with highly ordered, uniform pores. This suggests that the polymeric precursor was fully injected within the KIT-6 microspheres, allowing the continuity of the mesoporous structure in subsequent deposition reactions. The highly ordered pores of the NiCo2O<sup>4</sup> microspheres are clearly revealed with a pore size of ∼4 nm as shown by high-resolution TEM in **Figures 2d,e**. Additionally, in **Figure 2f**, the measured interplanar distance was found to be 0.29 nm, which aligns well with the (220) planes of spinel NiCo2O4. It is worth noting that well-defined diffraction rings were presented by the SAED pattern (Inset in **Figure 2f**), which correspond to the (440), (422), (400), (311), and (200) planes. The polycrystalline diffraction rings are in accordance with the result from the XRD pattern.

The pore diameter distribution and specific surface area of mesoporous NiCo2O<sup>4</sup> samples were determined via N<sup>2</sup> adsorption-desorption measurements. The result of specific surface area was calculated from the isotherms (**Figure 3**) was 97.77 m<sup>2</sup> g −1 for mesoporous NiCo2O<sup>4</sup> microspheres, while the non-porous NiCo2O<sup>4</sup> microspheres were 26.63 m<sup>2</sup> g −1 . Additionally, the mesoporous structure was further analyzed by pore diameter distribution in the inset of **Figure 3**. NiCo2O<sup>4</sup> microspheres displayed a pore volume (0.494 cm<sup>3</sup> g −1 ) with an average pore diameter of 3.416 nm. Ascribed to this special microsphere structure, the mesoporous structure could shorten the diffusion paths for lithium ions and provided buffering space to adapt

NiCo2O<sup>4</sup> (e). The typical plots for Z' vs. ω <sup>−</sup>0.5 for the mesoporous NiCo2O<sup>4</sup> (f).

the volume expansion during the process of Li<sup>+</sup> insertion and extraction.

To confirm whether the mesoporous NiCo2O<sup>4</sup> microspheres would be applicable as the anode materials in lithium-ion batteries, the as-prepared products were further investigated using cyclic voltammetry (CV), where **Figure 4a** exhibits the first three cycles of the mesoporous NiCo2O<sup>4</sup> microspheres. In the 1st cycle, one dominant peak at 0.8 V could be assigned to the decomposition and reduction of Ni and Co ions. Meanwhile, the anodic peaks at approximately 2.1 V could be owing to the oxidation reaction of metallic Ni and Co to NiO and CoO. In subsequent tests, the cathodic peak broadened and shifted to 1.0 V. According to the analysis of the CV curves and previous literature reports, the Li reactions for mesoporous NiCo2O<sup>4</sup> materials are as follows:

$$\rm NiCo\_2O\_4 + \rm 8Li^+ + 8e^- = Ni + 2Co + 4Li\_2O \tag{1}$$

$$\text{Ni} + \text{Li}\_2\text{O} = \text{NiO} + 2\text{Li}^+ + 2\text{e}^- \tag{2}$$

$$2\text{Co} + 2\text{Li}\_2\text{O} = 2\text{CoO} + 4\text{Li}^+ + 4\text{e}^- \tag{3}$$

$$2\text{CoO} + 2/3\text{Li}\_2\text{O} = 2/3\text{Co}\_3\text{O}\_4 + 4/3\text{Li}^+ + 4/3\text{e}^- \quad \text{(4)}$$

Galvanostatic tests were executed to ensure the influence of highly ordered mesoporous NiCo2O<sup>4</sup> samples on the specific capacity values and capacity retention. The 1st, 2nd, 10th, and 100th cycles of the galvanostatic charge and discharge curves for mesoporous samples, at a current of 100 mA.g−<sup>1</sup> , are shown in **Figure 4b**. A long voltage plateau was showed in the initial discharge curve between 0.7 and 1.2 V, corresponding to the strong reduction peak that appeared during the first cathodic CV scan. The initial charge capacity for the mesoporous NiCo2O<sup>4</sup> microspheres was ∼1467 mAhg−<sup>1</sup> , which is markedly higher than the non-porous NiCo2O<sup>4</sup> microspheres (∼820 mAhg−<sup>1</sup> , **Figure S2**); however, this decreased to 1,060 mAhg−<sup>1</sup> after the 2nd discharge. The irreversible capacity increase during the 1st charge may be due to the generation of a solid electrolyte interface (SEI) layer.

A comparison of the specific capacity values obtained for non-porous NiCo2O<sup>4</sup> and mesoporous NiCo2O<sup>4</sup> over 100 cycles plotted is shown in **Figure 4c**. The non-porous NiCo2O<sup>4</sup> suffers the most severe capacity fading with capacity values decreasing to 270 mAhg−<sup>1</sup> after 100 cycles, while the capacity retention was improved for mesoporous NiCo2O<sup>4</sup> microspheres, obtaining capacity values of 430 mAhg−<sup>1</sup> after 100 cycles. The significant increase in the capacities for mesoporous NiCo2O<sup>4</sup> compared to ordinary NiCo2O<sup>4</sup> is attributed to the inherent properties of the mesoporous structure. Nevertheless, the final capacity is relatively low compared to the initial capacity, which may be attributed to the aperture being too small to relieve the expansion of the active material completely (Bhaway et al., 2017). For conversion-mode materials, the mesoporous NiCo2O<sup>4</sup> provides particularly high buffering space, to adapt the large volume expansion during the process of Li<sup>+</sup> insertion/extraction and shorten the diffusion paths for lithium ions due to the interconnected architectures (**Figure S3**). Moreover, the coulombic efficiency for mesoporous NiCo2O<sup>4</sup> over 100 cycles is also shown in **Figure 4c**. The initial coulombic efficiency is quite low (∼70%); however, the efficiency remains >95% after the 10th cycle. The rate performance of ordinary NiCo2O<sup>4</sup> microspheres and mesoporous NiCo2O<sup>4</sup> microspheres at various current densities were then compared in **Figure 4d**. The mesoporous NiCo2O<sup>4</sup> microsphere electrode exhibits high initial discharge capacity and the capacity was recession with the increase of current density. However, the mesoporous NiCo2O<sup>4</sup> microsphere electrode exhibited a higher capacity and rate of lithium-ion storage than the non-porous NiCo2O<sup>4</sup> microsphere electrode.

To further investigate the mechanism for the improved electrochemical performance of the mesoporous NiCo2O<sup>4</sup> microsphere, electrochemical impedance measurements (**Figure 4e**) and a fitting process (**Figure S4**) were measured containing the mesoporous NiCo2O<sup>4</sup> electrode and ordinary NiCo2O4. As shown in **Figure 4e**, a semicircle and a straight line were acquired in the high frequency part and low frequency part, respectively. The intersection point from the curve in the high frequency part is on behalf of the electrolyte resistance. Moreover, the high frequency semicircle corresponds to the charge transfer resistance and the low frequency region with an inclined line represents the process of Li-ion diffusion. The initial value for charge transfer resistance is 66 , which is far less than the ordinary NiCo2O<sup>4</sup> (134 , **Table S1**), which could be related to the more effective ion and electron transfer in the interface of electrolyte and active material, so that the cell has improved electrode reaction kinetics and a better cell cycling results.

TABLE 1 | Warburg factor (σ) and diffusion coefficient (D<sup>+</sup> Li) of sample ordinary NiCo2O<sup>4</sup> and mesoporous NiCo2O4.


Moreover, the diffusion coefficient of the lithium ions (D<sup>+</sup> Li) can be confirmed in accordance with the following equation:

$$\mathcal{D}\_{\rm Li^{+}} = \frac{0.5 \mathcal{R}^{2} \mathcal{T}^{2}}{\mathcal{A}^{2} \mathcal{F}^{4} \mathcal{C}^{2} \sigma\_{\alpha}^{2}}$$

Where R = gas constant, T = absolute temperature, A = surface area of the electrode, F = Faraday constant, c = concentration of Li+ in the material, and σ = Warburg factor obeying the following relationship:

$$\mathbf{Z}\_{\rm Re} = \mathbf{R}\_{\rm S} + \mathbf{R}\_{\rm ct} + \sigma \,\omega^{-0.5}$$

Where ZRe is the real part of impedance, thus ω is the angular frequency at low frequency. The linear relationship of ZRe and ω <sup>−</sup>0.5 is shown in **Figure 4f**, with the slope of the fitted straight line indicating the value of σ. **Table 1** revealed the calculated values of σ and D<sup>+</sup> Li, where it can be clearly seen that the mesoporous NiCo2O<sup>4</sup> exhibits a higher value of D<sup>+</sup> Li than nonporous NiCo2O4.

#### CONCLUSION

Highly ordered mesoporous NiCo2O<sup>4</sup> microspheres with a honeycomb-like structure, were synthesized via a nano-casting method. The mesoporous NiCo2O<sup>4</sup> electrode possesses a high initial discharge capacity of ∼1,467 mAh.g−<sup>1</sup> at 100 mAg−<sup>1</sup> and it exhibited both a high surface area and good rate capability. Such high electrochemical performance is due to its excellent surface area of mesoporous NiCo2O<sup>4</sup> and the rapid ion transport in the electrolyte/electrode interface. This work may open a new sight in the synthesis of excellent Li-storage electrode materials and show an application in a promising candidate for Li-ion batteries.

#### DATA AVAILABILITY

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher. Requests to access the datasets should be directed to hu.junqing@dhu.edu.cn.

#### AUTHOR CONTRIBUTIONS

GW, BL, and JHu designed the project. QR, WX, JHa, PL, JC, SW, and RZ carried out the experiment and performed the experimental data analysis. QR, GW, and WX wrote the paper. BL revised the manuscript. All authors contributed to discussion of the results.

#### FUNDING

This work was financially supported by the National Key Research and Development of China (2017YFC0505803), the National Natural Science Foundation of China (41471191), Qing Lan Project of Jiangsu Province (Qinglan2016-15), Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), and the Fundamental

#### REFERENCES


Research Funds for the Central Universities (grant no. CUSF-DH-D-2017044).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00521/full#supplementary-material


advanced electrodes for supercapacitors. Adv. Funct. Mater. 22, 4592–4597. doi: 10.1002/adfm.201200994


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Ren, Wu, Xing, Han, Li, Li, Cheng, Wu, Zou and Hu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Corrigendum: Highly Ordered Mesoporous NiCo2O<sup>4</sup> as a High Performance Anode Material for Li-Ion Batteries

Qilong Ren1†, Guangyu Wu<sup>2</sup> \* † , Weinan Xing<sup>2</sup> \*, Jiangang Han<sup>2</sup> , Pingping Li <sup>2</sup> , Bo Li <sup>3</sup> \*, Junye Cheng<sup>4</sup> , Shuilin Wu<sup>4</sup> , Rujia Zou<sup>1</sup> and Junqing Hu<sup>1</sup> \*

*<sup>1</sup> State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai, China, <sup>2</sup> College of Biology and the Environment, Nanjing Forestry University, Nanjing, China, <sup>3</sup> Department of Vascular Surgery, Shanghai Ninth People's Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China, <sup>4</sup> Center of Super-Diamond and Advanced Films, Department of Materials Science and Engineering, City University of Hong Kong, Hong Kong, China*

Keywords: highly ordered, mesoporous, NiCo2O4, lithium-ion batteries, nano-casting

#### **A Corrigendum on**

#### Approved by:

*Frontiers Editorial Office, Frontiers Media SA, Switzerland*

#### \*Correspondence:

*Guangyu Wu gywuchem@163.com Weinan Xing xingweinan@126.com Bo Li boli@shsmu.edu.cn Junqing Hu hu.junqing@dhu.edu.cn*

*†These authors have contributed equally to this work*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *13 August 2019* Accepted: *14 August 2019* Published: *28 August 2019*

#### Citation:

*Ren Q, Wu G, Xing W, Han J, Li P, Li B, Cheng J, Wu S, Zou R and Hu J (2019) Corrigendum: Highly Ordered Mesoporous NiCo2O<sup>4</sup> as a High Performance Anode Material for Li-Ion Batteries. Front. Chem. 7:600. doi: 10.3389/fchem.2019.00600*

#### **Highly Ordered Mesoporous NiCo2O<sup>4</sup> as a High Performance Anode Material for Li-Ion Batteries**

by Ren, Q., Wu, G., Xing, W., Han, J., Li, P., Li, B., et al. (2019). Front. Chem. 7:521. doi: 10.3389/fchem.2019.00521

In the original article, the author order list was incorrectly showed as Guangyu Wu1∗† , Qilong Ren2† , Weinan Xing2<sup>∗</sup> , Jiangang Han<sup>2</sup> , Pingping Li<sup>2</sup> , Bo Li3<sup>∗</sup> , Junye Cheng<sup>4</sup> , Shuilin Wu<sup>4</sup> , Rujia Zou<sup>1</sup> , Junqing Hu1<sup>∗</sup> . The correct author order's list is Qilong Ren1† , Guangyu Wu2∗† , Weinan Xing2<sup>∗</sup> , Jiangang Han<sup>2</sup> , Pingping Li<sup>2</sup> , Bo Li3<sup>∗</sup> , Junye Cheng<sup>4</sup> , Shuilin Wu<sup>4</sup> , Rujia Zou<sup>1</sup> , Junqing Hu1<sup>∗</sup> .

In the published article, reflecting the above correction, there was an error with affiliations "1" and "2". Instead of appearing as "1 College of Biology and the Environment, Nanjing Forestry University, Nanjing, China" and "2 State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai, China," they should be reversed to appear as "1 State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai, China" and "2 College of Biology and the Environment, Nanjing Forestry University, Nanjing, China".

Furthermore, there was an error in affiliation "4". Instead of "Department of Materials Science and Engineering, Center of Super-Diamond and Advanced Films, City University of Hong Kong, Hong Kong, China," it should be "Center of Super-Diamond and Advanced Films, Department of Materials Science and Engineering, City University of Hong Kong, Hong Kong, China".

The authors apologize for this error and state that this does not change the scientific conclusions of the article in any way. The original article has been updated.

Copyright © 2019 Ren, Wu, Xing, Han, Li, Li, Cheng, Wu, Zou and Hu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Liquid Phase Exfoliated Hexagonal Boron Nitride/Graphene Heterostructure Based Electrode Toward Asymmetric Supercapacitor Application

Xuan Zheng<sup>1</sup> , Guangjin Wang2,3 \*, Fei Huang<sup>4</sup> , Hai Liu<sup>3</sup> , Chunli Gong<sup>3</sup> , Sheng Wen<sup>3</sup> , Yuanqiang Hu<sup>3</sup> , Genwen Zheng<sup>3</sup> and Dongchu Chen<sup>2</sup> \*

*<sup>1</sup> Hubei Provincial Key Laboratory of Green Materials for Light Industry, School of Materials and Chemical Engineering, Hubei University of Technology, Wuhan, China, <sup>2</sup> School of Materials Science and Energy Engineering, Foshan University, Foshan, China, <sup>3</sup> College of Chemistry and Materials Science, Hubei Engineering University, Xiaogan, China, <sup>4</sup> Key Laboratory of Functional Foods, Ministry of Agriculture, Guangdong Key Laboratory of Agricultural Products Processing, Sericultural & Agri-Food Research Institute Guangdong Academy of Agricultural Sciences, Guangzhou, China*

#### Edited by:

*Wenyao Li, Shanghai University of Engineering Sciences, China*

#### Reviewed by:

*Shijie Li, Zhejiang Ocean University, China Jiajun Wang, Tianjin University, China*

#### \*Correspondence:

*Guangjin Wang wgj501@163.com Dongchu Chen Chendc@fosu.edu.cn*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *20 June 2019* Accepted: *16 July 2019* Published: *02 August 2019*

#### Citation:

*Zheng X, Wang G, Huang F, Liu H, Gong C, Wen S, Hu Y, Zheng G and Chen D (2019) Liquid Phase Exfoliated Hexagonal Boron Nitride/Graphene Heterostructure Based Electrode Toward Asymmetric Supercapacitor Application. Front. Chem. 7:544. doi: 10.3389/fchem.2019.00544* In this paper, owing to the electrostatic interaction between graphene and h-BN, a facile liquid phase exfoliation method was carried out to fabricate h-BN/graphene based van der Waals heterostructure nanocomposites without additional chemical cross-linkers. The physicochemical properties of as-prepared composites were characterized by several electron microscopic and spectroscopic measurements. The h-BN/graphene heterostructure composites were employed to use as the anodes of asymmetric supercapacitor, and exhibited exceptional capacitive performance due to their synergistic effects. It is expected that the as-prepared h-BN/graphene materials can boost scalable heterostructure electrodes in supercapacitors, and our liquid phase exfoliation method can be used for the construction of the other energy storage and electronics.

Keywords: heterostructure, h-BN, graphene, asymmetric supercapacitor, liquid phase exfoliation

# INTRODUCTION

Graphene, as one of the most important representative of 2D nanomaterials, has been a research hotspot in recent years (Kong et al., 2017). It has special Dirac electronic properties (Loan et al., 2014), high carrier migration rate (Morozov et al., 2008), excellent thermal conductivity and mechanical properties (Balandin et al., 2008; Kong et al., 2017), and also been favored by the industry, academia and research institutes. Nevertheless, graphene is a zero band gap material, and its conducting and valence band interlace at the Dirac point. Therefore, the major dilemma to promote the application of graphene in electronic devices is the expansion of band gap. To process with this challenge, researchers have come up with many techniques to expand the bandgap of graphene, including the preparation of graphene nanoribbons (Jiao et al., 2009; Wang and Dai, 2010), nanomesh (Jingwei et al., 2010), and the chemical modification (Li et al., 2011). However, physical etching or chemical reaction would inevitably lead impurities into the boundary or surface of graphene, and greatly reduce the carrier mobility of grapheme (Liao et al., 2014).

Hexagonal Boron Nitride (h-BN) is defined as "white graphene" or "graphene-like" Boron Nitride, with an approximate honeycomb lattice formed by sp2 hybridization and a band gap of 5.9 eV (Wang et al., 2016). Because of its broad band gap, h-BN can be applied in prospects of spintronics, energy storage and composite materials. Because the h-BN and graphene have a rare low mismatched lattice constant, researchers have also revealed new features that the h-BN/graphene heterostructure can regulate the intrinsic electronic structure (Tran et al., 2016). Compared with the single-layer graphene, the carrier mobility (140,000 cm<sup>2</sup> /Vs) of the h-BN/graphene heterostructure film prepared by chemical vapor deposition method is 3.5 times higher than that of the single-layer graphene film (40,000 cm<sup>2</sup> /Vs) (Wang et al., 2013). Besides, the h-BN/graphene heterostructure was proved to have great potential value for application in energy storage devices, like Li-ion battery (Pomerantseva and Gogotsi, 2017; Wu et al., 2017; Li et al., 2018). Researchers also added the h-BN/Graphene heterostructure material into PVA fiber, greatly enhancing the mechanical properties of the composite fiber and increasing the conductivity to 3 S/m, so as to obtain the high-strength conductive composite fiber (Boland et al., 2016). Thus, combining graphene with other 2D materials (such as h-BN) to form heterostructure is bound to greatly expand the research scope of this field like a "snowball."

At present, large-scale preparation of high-quality h-BN/graphene heterostructure materials is still a recognized problem. From the preparation methods for h-BN/graphene heterostructure, it can be divided into two main modes: one is to grow graphene and h-BN on the substrate surface through CVD (chemical vapor deposition) method, and then transfer and mechanical superposition; the other is to directly use CVD method to grow graphene on substrate and then continue to grow h-BN on the surface of graphene, while this method needs to investigate the lattice mismatch problem. It is noted that the CVD method can ensure integrality of h-BN/graphene heterostructure, but the reaction condition is extremely harsh (usually requires high temperature vacuum environment), the heterostructure size is limited and the cost is high, so it's insufficient to meet the needs of practical application. Therefore, an effective preparation method is urgently needed to make up for the shortcomings of the CVD method to prepare the h-BN/graphene heterostructure materials.

As our best knowledge, liquid phase exfoliation method has been rarely carried out to prepare h-BN/graphene heterostructure. Nevertheless, some investigators have attempted to use this chemical method to create heterostructures, like graphene-black phosphorous-graphene sandwich heterostructure (Sun et al., 2015), metal oxide heterostructures (Xu et al., 2016, 2019; Mahmood et al., 2018; Wan et al., 2018), and (layered double hydroxide) LDH/rGO heterostructure (Ge et al., 2016) which are served as the electrodes in energy storage component. Compared with the heterostructure prepared by CVD method, the liquid phase exfoliation process is simpler and cheaper, and has a broader development prospect.

In this paper, we used the glycerol/urea system in graphene and h-BN exfoliation. The smaller h-BN nanosheets were attached on bigger graphene through ultrasonic assistant. Based on a certain amount of space size differences, the phenomenon of reunification in h-BN or graphene can be effectively prevented under the action of electrostatic force. Furthermore, the h-BN/graphene heterostructure system prepared in this work can be stably dispersed in organic solution, and the system can maintain long-term stability. Moreover, the h-BN/graphene heterostructure materials with different mass ratios in application of supercapacitors is studied in detail. As a result, the h-BN/graphene heterostructure materials show the maximum capacitance of 134 F/g and good cycling stability (96 % of the initial capacitance after 10,000 cycles at 10 A/g). Meanwhile, the assembled asymmertic supercapacitor (ASc) exhibits maximum energy density of 2.05 Wh/kg at high power density of 1998.5 W/kg.

# EXPERIMENTAL SECTION

# Materials Preparation

Graphite powder (300 mesh, purity>95%, XFNANO Materials) or h-BN powder (1 um, purity>98%, Sigma-Aldrich) were dissolved in the urea/glycerol (molar ratio = 2:1) dispersion. After that, 200 mL of the graphite or h-BN dispersion were transferred to a 800 mL flat bottom beaker, under which graphite or h-BN powder were exfoliated and dispersed through mechanical stirring at 800 rpm for 24 h. Then, the obtained products were evenly transferred to a 50 ml centrifuge tube, then centrifuged for 25 min at 5,000 rpm. The top half of the centrifuged graphite or h-BN dispersion were collected and redispersed in DMF, followed by filtration and ultrasonic washing with large amounts of DMF and ethanol, and drying in vacuum oven at 60◦C. The yield of graphene or h-BN was decided by taking off the mass of the residual solid (Zheng et al., 2018).

The graphene/DMF and h-BN/DMF dispersion solution were mixed together with a certain mass ratio (the mass content ratio of graphene and h-BN was 1:2, 1:1, 2:1, respectively). The resulting mixed solution was ultrasonic for 30 min, then stirred at room temperature for 24 h and centrifuged at 1,000 rpm for 30 min. The upper liquid was discarded, and the solid precipitation was finally obtained, namely, h-BN/grapheme (BN/G) heterostructure materials.

# Characterization

Transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HRTEM) images were obtained using a JEM-2001F (JEOL, 200 kV primary beam) equipped with a Gatan CCD camera. The morphology of the samples was observed by using S-4800 (Hitachi, accelerating voltage of 25 KV) cold field emission scanning electron microscope (FESEM). The crystal structure of heterostructure was analyzed by D8 Advance X-ray diffraction systems (Bruker, λ = 0.154056 nm). Raman spectra of the samples were tested by an InVia (Renishaw, UK) spectrometer to reflect the composition equipped by a 633 nm laser. The thickness of all samples was also analyzed by MFP-3D-SA AFM (AsylumResearch, USA) using a tapping mode at a scan rate of 1 Hz.

#### Electrode Preparation and Electrochemical Tests

The BN/G working electrodes were prepared by mixing 90 wt.% active material (BN/G heterostructure materials) and 10 wt.% polyvinylidene fluoride (PVDF) in NMP solvent to form a slurry. Then the slurry was coated onto a clean Ni mesh (current collector) and dried in a vacuum oven at 60 ◦C for 24 h. The counter electrode was the activated carbon (AC, SSA = 1,800(±100) m<sup>2</sup> /g, purity>94.2%, XFNANO Materials). The counter electrodes were prepared via the same method as the above working electrodes. The loading mass of each electrode was about 0.15 mg/cm<sup>2</sup> . Microporous polypropylene and 2 M KOH solution were used as the separator and electrolyte, respectively.

The electrochemical performance of samples was measured on the CHI660E (Chenhua, China) electrochemical workstation. Cyclic voltammetry (CV) tests were carried out at the scan rate of 10∼200 mV/s. Galvanostatic charge-discharge (GCD) tests were performed at various current densities from 0.5 to 10 A/g. Electrochemical impedance spectroscopy (EIS) was employed in a frequency from 0.01 Hz to 100 kHz at an amplitude of 5 mV. All the related electrochemical tests were performed at room temperature. Specific capacity of the BN/G in two-electrode was calculated from Equations (1) and (2) (Wang et al., 2014):

$$C = \frac{\int I\_1 dV}{\text{sm}\Delta V} \tag{1}$$

$$C = \frac{\text{I}\_2\text{t}}{\text{m}\Delta\text{V}}\tag{2}$$

Where I represents the discharged current (A), s is the scanning rate (V/s), 1t and 1V are the discharged time (s) and the voltage drop upon discharging, respectively, and m is the mass of the electroactive materials (g). The energy density (E, Wh/kg) and the power density (P, W/kg) of the BN/G//AC were calculated based on the following Equations (3) and (4) (Balogun et al., 2016):

$$E = 0.5CV^2/3.6\tag{3}$$

$$P = \frac{E \times 3600}{t} \tag{4}$$

Where C represents specific capacity of the capacitor (F/g) calculated according to Equation (2), 1t and 1V are the discharged time (s) and the voltage drop upon discharging, respectively.

# RESULTS AND DISCUSSION

**Figure 1** illustrates the preparation of the BN/G heterostructure. The graphene/DMF and h-BN/DMF dilute solution were added into the flat bottom beaker, respectively. Then the mixture was stirred to force the exfoliated h-BN nanosheets entering into the graphene layers to form heterostructure through π-π accumulation. Firstly, the morphology and structure of as-prepared graphene were characterized to verify the effectiveness of glycerol/urea liquid phase exfoliation system. In this work, due to the effectiveness of the exfoliation system, graphene/DMF dispersion solution was randomly diluted to a certain concentration to observe its micromorphology through TEM. As shown in **Figure 2a**, the graphene exhibited a dispersive lamellar distribution under a low magnification, whilst the unique wavy lattice fringes of graphene were further observed in the high magnification (**Figure 2b**). and these small lattice fringe is peculiar to the crystal of graphene material. At the same time, dotted impurities can be observed, which is most likely due to the non-covalently modified π-π accumulation effect formed by the interaction of urea or solvent with graphene in the glycerol/urea system (Chen et al., 2018). The selective fast fourier transform (FFT) analysis of the partial amplification of graphene in **Figure 2b** revealed the unique hexagonal pattern of electron diffraction in the few-layer of graphene. This result can clearly show the successful exfoliation of graphite (Paton et al., 2014).

In order to further observe the graphene prepared by the glycerol/urea system mentioned above, the graphene/DMF dispersion samples were analyzed with the SEM morphology. **Figure S1a** was the dispersion diagram of graphene after exfoliation, comparative measurements on scale bar of 1 um suggested that the size of graphene was close to 6 um. In addition, according to the morphology at high magnification (**Figure S1b**), the edges of graphene sheets were corrugated, and there were obvious upwarp, verifing the successful exfoliation of graphite.

We also used AFM to conduct random sample analysis of the exfoliated graphene, which was compared with the sampling method of TEM: the sample to be tested was dripped with pipette gun to the newly prepared mica film after vacuum dust removal. It can be observed the 2D morphology of typical graphene in the tapping mode from **Figure S2**. The size of the graphene lamella

was also more than 5 um, which as similar with the observed size of TEM and SEM. Based on the theoretical thickness of graphene (0.34 nm), the thickness of two adjacent areas in view of **Figure S2** is about 0.68 nm (Pattammattel and Kumar, 2015). This value indicated that the as-prepared graphene was close to double-layer, showing the exfoliated graphene by glycerol/urea system was few-layer.

To test the universality of glycerol/urea in the above system, we used the same process to treat other 2D material, such as h-BN, and also obtained h-BN/DMF dispersion solution.

FIGURE 4 | The TEM (a–e) and selected area electron diffraction (SAED) (f) images of h-BN/graphene heterostructure.

Then the h-BN samples were observed by TEM. As shown in **Figure 3**, it can be seen that the h-BN nanosheets were stacked on top of each other, and their sizes were below 1 um, much smaller than the size of graphene. From the HRTEM images (**Figures 3b,c**), it can be preliminarily concluded that h-BN powders were exfoliated into nanosheets. Combined with the FFT image in **Figure 3d**, the obtained h-BN lamination belongs to few-layer, which also proves the availability of glycerol/urea system (Varrla et al., 2015).

**Figure S3** is the SEM images of h-BN nanosheets. The white hexagonal circular lamella in **Figure S3b** is h-BN nanosheet with a size of <1 um, which is consistent with the intuitive size of TEM photographs. **Figure S3c** shows that h-BN nanosheet have obvious vertical stratification, indicating that h-BN raw powder is peeled into h-BN nanosheet. **Figure S4** shows the AFM diagram of h-BN sample, the thickness of the two adjacent white crystals is about 0.7 nm, which is in line with the theoretical value of the double-layer h-BN nanosheet (Kim et al., 2011; Tran et al., 2016).

To understand the crystal structure characteristics of the as-prepared BN/G heterostructure samples, XRD analysis was performed in **Figure S5**. It can be concluded that the graphene has extremely sharp diffraction peak at 26.5◦ , and h-BN has obvious diffraction peak at 26.7◦ , which is mainly reflected by the relatively small layer spacing difference between h-BN and graphene. The (002) crystal faces of graphene and h-BN still show no obvious deviation, implying that the formation of heterostructure is only a physical stacking process.

**Figure S6** shows the typical Raman absorption peaks of h-BN/graphene, h-BN and graphene, such as D peak (∼1,331 cm−<sup>1</sup> ) and G peak (∼1,578 cm−<sup>1</sup> ), which are shown in Graphene and BN/G, respectively, demonstrating the presence of graphene in BN/G heterostructure. Due to the strong photoluminescence background of h-BN, the photoluminescence background shown in the BN/G region can be attributed to the defect states of h-BN, including defects along grain boundaries (Li et al., 2012). The heterostructure of BN/G was determined by HRTEM and SAED. **Figures 4a,b** depicted two BN/G heterostructure regions, in which the larger graphene sheet was covered by the smaller h-BN nanosheets. This structure size is in accordance with the previous TEM results of h-BN and graphene. The HRTEM image in **Figures 4c–e** further proved a certain difference between exfoliated h-BN and graphene on the interface edge of the BN/G heterostructure, and the lattice fringe of the two structures was particularly obvious. **Figure 4f** is the SAED result of lattice fringe corresponding to BN/G heterostructure. From its diffraction pattern, a set of clear hexagonal diffraction spots can be observed. The close view of the area indicated by the blue circle shows two separate diffraction points along the radial direction. These two points correspond to (100) plane diffraction of h-BN and graphene, respectively. The calculated plane spacing of the two points was 2.13 and 2.06 Å, which were in good agreement with the (100) crystal plane spacing values obtained by the XRD tests. The AFM height of BN/G heterostructure on mica substrate is shown in **Figure S7**. The h-BN is small in size and

stacked on top of the graphene layers in different thicknesses. However, the average thickness of graphene is larger than that of the previous graphene sheet, attributing to the alternating superposition of h-BN and graphene in the heterostructural materials (Yang et al., 2013a).

Inhibiting the stack of graphene is the key factor to realize high the performance of the electrochemical capacitor based on graphene materials. As for rGO, the capacitance is significantly reduced when re-stacking between graphene sheets, mainly due to the irreversible stacking of single rGO sheet during the reduction and drying process (Yang et al., 2013b). We expect that the h-BN nanosheet in the BN/G can be used as an effective electrolyte channel to increase the overall electroactive surface area (**Figure 1**). It is hoped that charges can be stored in graphene or h-BN electrodes through electrostatic interaction, adsorption and desorption of interface ions (**Figure S8**). Since graphene can be assumed as a zero-gap semiconductor with Fermi energy level located at Dirac point, the presence of B and N in h-BN makes it form h-BN/graphene superlattice above the Fermi energy level (Ge et al., 2016; Lee et al., 2016). Therefore, we prepared BN/G with different mass loads of h-BN and studied the potential application of h-BN as a 2D structural carrier and electrolyte channel in graphene-based membrane.

**Figures 5A–C** show the CV curves of three mass ratios BN/G measured by the two-electrode method at different scanning rates (10 ∼200 mV/s) in 2 M KOH. As illustrated in **Figures 5A–C**, we can notice the prominent redox peak, which ascribing to the change of oxidation state of N atoms by electrolyte insertion. **Figure 5D** is the change of capacitance with the increase of scan rates, the CV curve still shows a larger redox peak even when scan rate increases to 200 mV/s, this is because the Fermi level of BN/G electrode materials reached a higher potential compared to the redox potential of electrolyte, electrons transferred from the electrode to the interface of the electrolyte. Since h-BN is usually used as dielectric material, its capacitance is very low, while the capacitance of BN/G is relatively large, it possibly attributed to BN/G heterostructure shortening the ion transport path and increasing the specific surface area for charge storage.

However, due to the differences in the contributions of h-BN and graphene to the overall capacitance in BN/G, the accumulation of h-BN multilayer and the increase of Faraday resistance may be the reason for the lower capacitance of BN/G(2:1; 1:1). It is found that the graphene is inserted into the interlayer space of h-BN to play the role of interval layering, preventing the re-stacking of h-BN nanosheets. Therefore, the charge storage mechanism in BN/G may be caused by the synergistic effect of graphene and h-BN. **Figures 6A–C** shows the charge and discharge curves of BN/G samples with different mass ratios at different current densities. The specific capacitance value is calculated according to the above Equation (2). Compared

FIGURE 6 | Galvanostatic charge–discharge curves for all BN/G at different current densities: (A) BN/G(2:1), (B) BN/G(1:1), (C) BN/G(1:2), and (D) specific capacitance of all BN/G as a function of discharge current.

with BN/G(2:1) and BN/G(1:1), the discharge time of BN:G(1:2) is longer at the same current density. Furthermore, BN:G(1:2) has a specific capacitance of 134 F/g when the current density is 0.5 A/g. By contrast, BN/G(2:1) and BN/G(1:1) have specific capacitance of 71 and 106 F/g, respectively, at the same current density.

The variation trend of BN/G samples with different mass ratios is shown in **Figure 6D**. Driven by the increase of current density, the capacitance of BN/G declined, but it tended to be flat at current densities ranged from 2 to 10 A/g, indicating the BN/G heterostructural materials have good rate performance. When the current density is increased to 10 A/g, the capacitance retentions of BN/G(2:1), BN/G(1:1), and BN/G(1:2) are 55, 56, and 69 %, respectively. These capacitance retentions match well with the trend calculated by the CV method in **Figure 5D**, and the optimal ratio is BN:G = 1:2.

BN:G(1:2) was further taken as a representative to study its cycle stability for charging and discharging 10,000 times under the current density of 10 A/g. **Figure 7A** shows that after 10,000 times of rapid charge and discharge tests, BN:G(1:2) electrode has only 4% loss, exhibiting its excellent electrochemical stability. This is because h-BN has extremely high chemical and thermal stability, providing synergies with high electrochemical stability when it exists in conductive graphene systems (Kumar et al., 2015). **Figure 7B** shows the AC impedance diagram of BN/G samples with different mass ratios. The lattice in the figure shows that the BN/G sample exhibits excellent capacitive characteristics, especially the approximate vertical line in the lower frequency range. And the charge transfer resistances (Rct) of BN/G(2:1), BN/G(1:1), and BN/G(2:1) are small, and the equivalent series resistance (ESRs) is 6.7, 6.0, and 5.3 , respectively. The results can be concluded that h-BN in heterostructure creates electrolyte channel, and as an effective ion channel to boost and expedite diffusion of electrolyte ions such as K<sup>+</sup> and OH<sup>−</sup> (Hu et al., 2014).

Toward studying the electrochemical properties of our assembled BN/G//AC ASc, we further reflected its power density and energy density. **Figure 7C** showed the power-energy density relation curve of BN/G(1:2). The calculated values of the energy density and power density of BN/G(1:2) were obtained according to Equations (3) and (4). It can be seen from **Figure 7C** that the energy density of BN/G(1:2)//AC attenuates very little with the increase of power density. Even at the power density of 1998.5 W/kg, its energy density still retains at 2.05 Wh/kg, showing the bright prospects for the application of h-BN/grahenebased supercapacitors.

#### REFERENCES


## CONCLUSIONS

In summary, we reported a well-constructed heterostructure materials, in which the h-BN nanosheets were assembled with graphene through electrostatic interaction based on solution method. This scalable stacked heterostructure are well-aligned in vertical, but the layers are randomly stacked in horizontal. The synthesis method adopted in this work has extensive use, low cost and facile process, which can be applied to other 2D materials. Meanwhile, we further developed BN/G heterogstructure electrode material, and explored the feasibility of BN/G heterostructure electrode material in the application of asymmertic supercapacitors. The incorporation of h-BN not only constructs the electrolyte channel for the graphene layer, but also improves the electrochemical performance. Moreover, the assembled BN/G//AC ASc exhibits maximum energy density of 2.05 Wh/kg at high power density of 1998.5 W/kg, and excellent long cycling stability with 96% of initial specific capacitance after 10,000 cycles at 10 A/g, indicating its promising application in energy storage component.

## DATA AVAILABILITY

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

## AUTHOR CONTRIBUTIONS

XZ, GW, FH, HL, and CG were responsible for all the experiments and the analysis of data. FH, SW, YH, GZ, and DC were responsible for the drafting. All authors contributed equally to the final writing of the paper.

# FUNDING

This work was supported by the National Natural Science Foundation of China (No. 21802037) and the Natural Science Foundation of Hubei Province (No. 2018CFB669).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00544/full#supplementary-material


hydroxide (LDH) and reduced graphene oxide. Nano Energy 20, 185–193. doi: 10.1016/j.nanoen.2015.12.020


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Zheng, Wang, Huang, Liu, Gong, Wen, Hu, Zheng and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Facile Synthesis of Novel V0.13Mo0.87O2.935 Nanowires With High-Rate Supercapacitive Performance

Haishun Jiang<sup>1</sup> , Wenjing Sun<sup>1</sup> , Wenyao Li <sup>1</sup> \*, Zhe Wang<sup>1</sup> , Xiying Zhou<sup>1</sup> \*, Zexing Wu<sup>2</sup> and Jinbo Bai <sup>3</sup> \*

<sup>1</sup> School of Material Engineering, Shanghai University of Engineering Science, Shanghai, China, <sup>2</sup> College of Chemistry and Molecular Engineering, Qingdao University of Science and Technology, Qingdao, China, <sup>3</sup> Laboratoire Mécanique des Sols, Structures et Matériaux, CNRS UMR 8579, Ecole Centrale Supelec, Université Paris Saclay, Châtenay-Malabry, France

#### Edited by:

Elizabeth J. Podlaha, Clarkson University, United States

#### Reviewed by:

Dipankar Roy, Clarkson University, United States John Zhanhu Guo, University of Tennessee, Knoxville, United States

#### \*Correspondence:

Wenyao Li liwenyao314@gmail.com Xiying Zhou zhouxiying@sues.edu.cn Jinbo Bai jinbo.bai@centralesupelec.fr

#### Specialty section:

This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry

Received: 30 May 2019 Accepted: 12 August 2019 Published: 04 September 2019

#### Citation:

Jiang H, Sun W, Li W, Wang Z, Zhou X, Wu Z and Bai J (2019) Facile Synthesis of Novel V0.13Mo0.87O2.935 Nanowires With High-Rate Supercapacitive Performance. Front. Chem. 7:595. doi: 10.3389/fchem.2019.00595 Binary metal oxides composed of molybdenum–vanadium oxides are promising candidates for supercapacitors. Here, we report the synthesis of one-dimensional V0.13Mo0.87O2.935 nanowires through a facile one-step hydrothermal method. This nanowire presented a high specific capacitance of 394.6 F g−<sup>1</sup> (1 mV s−<sup>1</sup> ) as an electrode applied to the supercapacitor. Importantly, this electrode showed a perfect rate capability of 91.5% (2 to 10 A g−<sup>1</sup> ) and a continuous verified outstanding cyclic voltammetry of 97.6% after 10,000 cycles. These superior electrochemical properties make the synthesized V0.13Mo0.87O2.935 nanowires a prospective candidate for high-performance supercapacitors.

Keywords: molybdenum–vanadium oxides, nanowires, hydrothermal, high rate, supercapacitors

# INTRODUCTION

Due to overconsumption of non-renewable resources and the growing threat of global warming, reliable and clean energy supplies, such as the secondary battery and supercapacitor (SC) science and technology, are in urgent need of a breakthrough (Liu et al., 2016; Salanne et al., 2016; Liu M. et al., 2018; Liang et al., 2019). SCs are becoming more appealing than ever because of their rapid recharge capabilities, high power density, and durable life cycles (Salanne et al., 2016; Du et al., 2018; Kirubasankar et al., 2018; Ho and Lin, 2019; Le et al., 2019; Ma et al., 2019; Yang L. et al., 2019). It is well-established that three main electrode materials include conducting polymer, transition metal oxide, and carbon materials (Jabeen et al., 2016a,b; Chen et al., 2017; Li et al., 2018; Idrees et al., 2019). In this regard, transition metal oxides can increase the efficiency and improve the specific capacitances compared to conducting polymers and carbon materials (Yang et al., 2015; Fu et al., 2016; Qin et al., 2016a,b; Meng et al., 2017; An and Cheng, 2018). Unfortunately, it has either insufficient electrochemical stability or low conductivity, which still greatly hampers their widespread applications in SCs (Jiang et al., 2012). Therefore, an innovative material that can be applied as a significant electrode material in the field of SCs is still needed.

In the last few years, binary metal oxides with stoichiometric or even nonstoichiometric composition such as NiCo2O<sup>4</sup> (Ma et al., 2016), NiFe2O<sup>4</sup> (Yu et al., 2014), and MnCo2O4.5 (Hu et al., 2019) have achieved efficient energy storage. It stems from its defect–effect mechanisms (Ellis et al., 2007; Wang et al., 2017) or possible jump processes (Hu et al., 2012; Li et al., 2018; Yang Y. et al., 2019) that provided the needed efficient electron conductivity. Also, the electrochemical behavior of these binary metal oxides is different to simple metal oxides attributed to their composition, including the species and ratios of elements. In particular, binary metal oxides based on molybdenum oxides or vanadium oxides are also regarded as a potential candidate for SCs. Many binary metals– molybdenum oxides, such as NiMoO<sup>4</sup> (Cheng et al., 2015), a-MnMoO<sup>4</sup> (Purushothaman et al., 2012), CoMoO4•0.9H2O (Liu et al., 2014), and NiMoO<sup>4</sup> (Mehrez et al., 2019), and binary metal–vanadium oxides, such as β-Na0.33V2O<sup>5</sup> (Hong Trang et al., 2014), Li3VO<sup>4</sup> (Iwama et al., 2016), and BiVO<sup>4</sup> (Patil et al., 2016; Guo et al., 2019), have been prepared for highperformance SCs. Despite the tremendous efforts that have been made on the electrode materials for these binary metal oxides, researchers continue to explore the performance of the electrode material for sustainable, low-cost, and clean energy storage and conversion technologies. Especially, binary metal oxides composed of molybdenum–vanadium oxide are also expected to be of favorable potential as SCs. However, such reports are rare.

Herein, we report a simple preparation of onedimensional V0.13Mo0.87O2.935 nanowires through a one-step hydrothermal method. This nanowire electrode exhibits a high specific capacitance of 394.6 F g−<sup>1</sup> (1 mV s−<sup>1</sup> ) as an electrode material in SC. Additionally, this electrode showed a rate capability of 91.5% (2 to 10 A g−<sup>1</sup> ) and an outstanding cycle stability (97.6% after 10,000 cycles). Therefore, one-dimensional V0.13Mo0.87O2.935 nanowires have been prepared and applied as a high-performance SC electrode material.

### EXPERIMENTAL

#### Preparation

Firstly, the molybdenum powder (Mo, 0.192 g, 2 mmol) was mixed with 37 ml of deionized H2O and 3 ml of hydrogen peroxide at room temperature and then continuous stirred till the solution became light yellow. After that, 0.088 g of ammonium vanadate (NH4VO3, 0.75 mmol) was added to the solution until the solid powder was completely dissolved. Then, the resulting solution was decanted into a Teflon reaction kettle and heated in oven at 200◦C for 48 h. After cooling to room temperature, the obtained crude products were treated with 2 M nitric acid. Finally, the nanowires were collected through washing with distilled H2O till neutral and then dried under air at 60◦C for 18 h.

# Material Characterizations

The X-ray diffractometer (XRD; with Cu-Kα radiation) presented the structure and phase of one-dimensional V0.13Mo0.87O2.935 nanowires. The nanowires' morphological feature was studied by a scanning electron microscope (SEM; S-4800) and a transmission electron microscope (TEM; JEM-2100F). Compositions of the samples were tested by X-ray photoelectron spectroscopy (Thermo ESCALAB 250XI). An automated nitrogen adsorption analyzer (ASAP 2020, Micromeritics, America) presented N<sup>2</sup> adsorption–desorption isotherm under the 77 K conditions.

## Electrochemical Characterizations

Electrochemistry performances were tested in three electrode systems with 1 M Na2SO<sup>4</sup> electrolyte using Autolab potentiostat (PGSTAT302N). A saturated calomel electrode (SCE) was used as the reference electrode and a platinum (Pt) foil was used as the counter electrode. The working electrode was a mixture of one-dimensional V0.13Mo0.87O2.935 nanowires, acetylene black, and polyvinylidene fluoride (PVDF) according to a certain mass ratio (80:15:5) in a few N-methyl pyrrolidinone (NMP). After the mixture was stirred for 24 h, the formed slurry was dripped on graphite paper and then vacuum dried at 60◦C for 15 h. Cyclic voltammetry (CV) measurement was carried out in a voltage range of 0–1.0 V at different sweeping rates (1, 5, 10, 25, 50, 75, and 100 mV s−<sup>1</sup> ), and galvanostatic charge–discharge (GCD) was tested at different current densities (2, 4, 6, 8, and 10 A g−<sup>1</sup> ). EIS data are obtained at a frequency from 10−<sup>2</sup> to 10<sup>5</sup> Hz with an AC amplitude of 5 mV.

# RESULTS AND DISCUSSIONS

In the present work, the phase for one-dimensional V0.13Mo0.87O2.935 nanowire was first characterized. The XRD spectrum for the prepared product is indicated in **Figure 1** in that all diffraction peaks matched a hexagonal phase of onedimensional V0.13Mo0.87O2.935 nanowires (JCPDS card No. 48- 0766). No characteristic peaks from impurity have been detected, suggesting that the pure one-dimensional V0.13Mo0.87O2.935 nanowires were prepared. Furthermore, the diffraction peaks were sharp and intense, showing their high degree of crystallinity.

The SEM image in **Figure 2a** depicts the typical morphology of the one-dimensional V0.13Mo0.87O2.935 nanowires, which consists of a number of uniform nanowires with an edge length of more than 10µm. For more detail, the samples were examined by TEM as indicated in **Figure 2b** in that the diameters of the

nanowires are 20–30 nm with uniform nanostructures. The HR-TEM image is indicated in **Figure 2c**; those one-dimensional V0.13Mo0.87O2.935 nanowires have a similar crystal structure and no amorphous phase on the surface. It could be deduced from the lattice fringes that the lattice spacing is 0.26 nm, agreeing to the (220) plane of one-dimensional V0.13Mo0.87O2.935 nanowires. In further studying the details, the brighter spots in the FFT pattern (illustration in **Figure 2c**) pointed out an excellent crystal. Besides, **Figure 2d** confirmed that the lattice spacing of 0.26 nm in **Figure 2c** belongs to the (220) plane. These results closely matched the data obtained from the XRD analysis, further confirming the crystal structure of V0.13Mo0.87O2.935 nanowires.

The X-ray photoelectron spectroscopy (XPS) shows that the one-dimensional V0.13Mo0.87O2.935 nanowires are composed of three elements: V, Mo, and O (**Figure S1** of the Supporting Information). The XPS peak of V 2p in **Figure 3A** was determined to be a peak of V 2p3/<sup>2</sup> of 517.1 eV, and the V 2p1/<sup>2</sup> peak of V5<sup>+</sup> was not included because the low mole percentage of vanadium in the compound was the smallest (Geert et al., 2004; Liu X. et al., 2018). **Figure 3B** shows the Mo 3d spectrum composed of two peaks, the Mo 3d3/<sup>2</sup> from the peak at 236.0 eV indicates Mo6+, and another peak at 232.9 eV could be due to the superposition of Mo 3d5/<sup>2</sup> and Mo 3d3/2, which indicates Mo6<sup>+</sup> and Mo5<sup>+</sup> (Bica de Moraes et al., 2004). Meanwhile, in **Figure 3C**, the XPS peak of the O 1s was observed at 530.8 eV. In addition, the existence of Mo5<sup>+</sup> was ascribed to the oxygen anion vacancy in the framework of the compound structure, so that molybdenum is only coordinated by five oxygen species.

The one-dimensional V0.13Mo0.87O2.935 nanowires were further investigated by the N<sup>2</sup> adsorption–desorption isotherms

as indicated in **Figure 4**. According to IUPAC, the N<sup>2</sup> adsorption–desorption isotherms of the V0.13Mo0.87O2.935 nanowires are a typical type IV adsorption isotherm with the H3 hysteresis loop, exhibiting a mesoporous structure with slitshaped pores. The BET-specific surface area and pore diameters

(illustration in **Figure 4**) of the V0.13Mo0.87O2.935 nanowires are about 54.2 m<sup>2</sup> g −1 and 80 nm, respectively, which may be attributed to the assembly of the nanowires in space. This porous structure contributes to the diffusion of electrolyte ions and transport during the charge and discharge process of the SC electrodes (Hou et al., 2018, 2019).

The as-prepared one-dimensional V0.13Mo0.87O2.935 nanowires were applied to SC electrode materials. **Figure 5A** depicts the CV curves tested in the voltage from 0 to 1.0 V. Approximate rectangle-shaped and symmetrical CV curves were viewed without redox peaks, showing an EDLCdominated capacitance behavior of the one-dimensional V0.13Mo0.87O2.935 nanowires (Hung et al., 2011; Lokhande et al., 2011; Pujari et al., 2016). Besides, the specific capacitance (**Table S1** of the Supporting Information) of one-dimensional V0.13Mo0.87O2.935 nanowires was very high and was 394.6 F g −1 at 1 mV s−<sup>1</sup> . Notably, it can be seen that the CV curve mostly remains in an approximately rectangle-like shape with a sweeping rate between 1 and 100 mV s−<sup>1</sup> , which confirmed good electrochemical reversibility and outstanding high-energy storage performance; the CV plot tilt increases with increasing scan rates owing to the fact that the electrons do not migrate from the inside of the material to the surface of the electrode in time. **Figure 5B** shows the GCD curves of the one-dimensional V0.13Mo0.87O2.935 nanowire electrode at different current densities. It displayed

respectively. (C) Rate performance and (D) cycle stability of the V0.13Mo0.87O2.935 nanowire electrodes, in 1 M Na2SO4 electrolyte.

proximate central symmetry voltage profiles, which were consistent compared to the CV results, pointing to the onedimensional V0.13Mo0.87O2.935 nanowires having an excellent reversibility across the whole potential region. Furthermore, one-dimensional V0.13Mo0.87O2.935 nanowire electrodes presented high specific capacitances from 385.2 to 352.5 F g <sup>−</sup><sup>1</sup> while discharge current density was enhanced to 2, 4, 6, 8, and 10 A g−<sup>1</sup> (**Table S2** of the Supporting Information). Compared with other binary metal oxide electrodes, onedimensional V0.13Mo0.87O2.935 nanowire electrodes also indicated a strengthened specific capacitance as reported in the literature, such as CoMoO<sup>4</sup> (384 F g−<sup>1</sup> ) (Li et al., 2018), BiVO<sup>4</sup> (116.3 F g −1 ) (Patil et al., 2016), and MnMoO<sup>4</sup> (168.32 F g −1 ) (Veerasubramani et al., 2014).

The specific capacitances of the V0.13Mo0.87O2.935 electrodes with different current densities are indicated in **Figure 5C**. It maintained a remarkable rate performance of 91.5% from 2 to 10 A g−<sup>1</sup> . This result may be attributed to the active materials to form porous channels through intertwined networks, enabling efficient electrolyte transport and accessibility of active sites (Jiang et al., 2011). Therefore, it is possible to maintain a high specific capacitance even at higher current densities. **Figure 5D** indicates the long-term cycle stability of the one-dimensional V0.13Mo0.87O2.935 nanowire electrode, which was tested through CV tests repeating 10,000 cycles at 50 mV s−<sup>1</sup> . It can be observed that its specific capacitance retention showed outstanding stability, with the increase in some cycles fluctuating only a little. After 10,000 cycles, the retention rate value was found to be 97.6% of the initial value.

The V0.13Mo0.87O2.935 electrodes were subjected to electrochemical impedance spectroscopy (EIS) to explore relevant charge transfer resistance. **Figure 6** shows the Nyquist plot before and after 10,000 cycles of the one-dimensional V0.13Mo0.87O2.935 nanowire electrodes. The inset shows the corresponding equivalent circuit by its corresponding fitting curve (**Figure S2** in Supporting Information), which was fitted by an equivalent circuit consisting of a bulk solution resistance Rs , a charge-transfer Rct, and constant phase element (CPE). The R<sup>s</sup> values of the one-dimensional V0.13Mo0.87O2.935 nanowire electrode before and after 10,000 cycles are 2.02 and 2.10 Ω, respectively. Also, the value of Rct was connected with charge transfer after 10,000 cycles and is only slightly higher than before (68.6 vs. 50.1 Ω), manifesting superior conductivity and stability of the one-dimensional V0.13Mo0.87O2.935 nanowire microstructure owing to good ion conductivity of the interface between electrolyte and electrodes.

#### CONCLUSIONS

In summary, one-dimensional V0.13Mo0.87O2.935 nanowires were synthesized under a facile one-step hydrothermal condition. For application in a SC electrode, it was found to present a high specific capacitance of 394.6 F g−<sup>1</sup> (1 mV s−<sup>1</sup> ). Besides, this electrode showed a perfect rate capability of 91.5% at the current density that was enhanced five times and outstanding long-term cyclic stability (97.6% after 10,000 cycles). This study offers a common preparation method of binary molybdenum–vanadium oxide used in SCs with a superior electrochemical property.

#### DATA AVAILABILITY

All datasets generated for this study are included in the manuscript/**Supplementary Files**.

#### AUTHOR CONTRIBUTIONS

WL conceived and designed the experiments. HJ, WS, and ZWa performed the experiments and analyzed the data. HJ and WS wrote and revised the manuscript. WL, ZWu, XZ, and JB discussed and supervised the whole project. All the authors revised and checked draft.

#### FUNDING

The research was supported by the National Natural Science Foundation of China (51602193), Shanghai Chen Guang project (16CG63), the Fundamental Research Funds for the Central Universities (WD1817002) and the Talent Program of Shanghai University of Engineering Science, the Natural Science Foundation of Shandong Province of China (ZR2019BB002), and Shanghai University of Engineering Science Innovation Fund (18KY0503).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00595/full#supplementary-material

# REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Jiang, Sun, Li, Wang, Zhou, Wu and Bai. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Fabrication and Electrochemical Performance of Al-Doped ZnO Nanosheets on Graphene-Based Flexible Substrates

Qi Yu<sup>1</sup> \*, Ping Rong<sup>1</sup> , Shuai Ren<sup>1</sup> , Liyun Jiang<sup>2</sup> and Yapeng Li <sup>1</sup>

*<sup>1</sup> School of Materials Science and Engineering, Institute of Graphene at Shaanxi Key Laboratory of Catalysis, Shaanxi University of Technology, Hanzhong, China, <sup>2</sup> School of Physics and Telecommunication Engineering, Shaanxi University of Technology, Hanzhong, China*

In this work, Al-doped ZnO (AZO) nanosheets (NSs) were successfully synthesized on graphene-coated polyethylene terephthalate (GPET) flexible substrate via hydrothermal method. Studies have indicated that with the addition of Al3+, the nanostructure of ZnO gradually grows from nanorods (NRs) to NSs, and the (100), (002), and (101) diffraction peak strength of ZnO that grows perpendicularly to the substrate along the c-axis weakened. The mechanism of hydrothermal growth of AZO/GPET was also studied. The electrochemical properties of the samples were investigated by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS), and it was concluded that AZO NSs grown on GPET substrates has better capacitance performance than undoped ZnO NRs.

Keywords: ZnO, Al-doped, hydrothermal method, graphene-based flexible substrates, electrochemical performance

### INTRODUCTION

Up to now, energy reserves and environmental contamination are still the focus of extensive attention, especially the problems of air pollution, water pollution, global warming, and renewable energy, which are closely linked with our lives. In order to solve these problems, batteries and supercapacitors have become research hotspots of electrochemical energy storage systems. Among them, supercapacitor (SC), is of great attention in the fields of automobiles (Cao and Emadi, 2011; Biplab et al., 2017), wind power systems (Abbey and Joos, 2007), solar cells (Narayanan et al., 2015; Xu et al., 2015), and so on, because of its great power density, lack of required maintenance, wide operating temperature range, green environmental protection, long cycling life, etc. (Zhao et al., 2011). In addition to those, SCs can provide high power pulses in a short period of time compared to conventional capacitors or storage batteries. SCs, for example, are often used for intelligent start-stop control systems (lightweight hybrid power system), which are particularly prominent in plug-in hybrid electric vehicles (Cao and Emadi, 2011). Due to the different energy storage mechanism, SCs can be divided into electric double-layer capacitors (EDLCs) and faraday pseudo-capacitors. The former generates and stores energy by adsorption of a pure electrostatic charge on the electrode surface and the latter uses redox reaction to store electrical energy in an electrochemical manner (Liu et al., 2014). The properties of SCs are closely related to the electrode materials used, and examples of materials used in current research are: carbon materials

Edited by:

*Wenyao Li, Shanghai University of Engineering Sciences, China*

#### Reviewed by:

*Xuenian Chen, Henan Normal University, China Jun Li, Tsinghua University, China*

> \*Correspondence: *Qi Yu kukukoko2004@163.com*

#### Specialty section:

*This article was submitted to Energy Materials, a section of the journal Frontiers in Materials*

Received: *22 June 2019* Accepted: *12 August 2019* Published: *04 September 2019*

#### Citation:

*Yu Q, Rong P, Ren S, Jiang L and Li Y (2019) Fabrication and Electrochemical Performance of Al-Doped ZnO Nanosheets on Graphene-Based Flexible Substrates. Front. Mater. 6:208. doi: 10.3389/fmats.2019.00208* (Salinas-Torres et al., 2019), metal oxides (Wu et al., 2018) and conductive polymers. Of all electrode materials, carbon materials with high specific surface area and low internal resistance receive more attention, including activated carbon fibers (Ren et al., 2013), carbon aerogel (Liu et al., 2018), carbon nanotubes (Futaba et al., 2006), activated carbon (Wang et al., 2014; Isabel et al., 2016), porous carbon (Yang et al., 2019), and graphene (Zhang et al., 2016; Ren et al., 2018).

The theoretical specific capacity of semiconductor oxides such as ZnO, SnO2, and TiO<sup>2</sup> is 2,3 times that of graphite, which has attracted enormous attention. Among them, ZnO, as a ntype semiconductor, has tremendous senses for the fields of chemicals, electronics, and optics owing to its superior properties [i.e., a large exciton binding energy (60 meV) and a wide bandgap of about 3.3 eV at room temperature] (Klingshirn, 2010). In addition, as the electrode material of supercapacitor, ZnO has been paid more and more attention because of its advantages of high chemical stability and thermal stability, low cost, environment friendly, and easy doping. However, ZnO has the disadvantages of poor conductivity and large volume effect in the process of charging and discharging, which affects the practical application of ZnO as an electrode material. Graphene, a two-dimensional carbon nanomaterial with zero bandgap, is attracted much attention that as a prospective candidate electrode material for EDLCs due to its high carrier mobility, great chemical resistance, large surface area, high conductivity, and transparency (Han et al., 2014; Ren et al., 2018). However, the presence of Van der Waals makes graphene easy to reunite, thus reducing the specific surface area and specific capacity of graphene. Therefore, ZnO and graphene materials composite and doped, can achieve the complementary advantages of material properties.

Bhirud et al. prepared N-doped ZnO/graphene (NZO/GR) by situ wet chemical method and studied their electrochemical properties. It was observed, the specific capacitance of NZO/GR was 555 Fg−<sup>1</sup> , which was 529 Fg−<sup>1</sup> and 20% higher than pure ZnO/GR (Bhirud et al., 2015). Cu/ZnO doped graphene nanocomposites was investigated by Jacob et al. (2018). Electrochemical analysis showed that the material has a specific capacity of 630 mAhg−<sup>1</sup> and retains around 95% of this capacity after 100 cycles. Faraji and Ani (2014) reviewed the application of microwave-assisted metal oxide thin film electrodes in supercapacitors. And they noted that ZnO/GR composites have high specific capacitance and good reversible chargedischarge performance. Many previous studies have used the hummer method to prepare graphene to prepare ZnO/graphene nanoparticles (Wang et al., 2011; Bu and Huang, 2015; Zhang et al., 2015). Therefore, it is necessary to explore the preparation of ZnO nanofilms based on transparent conductive flexible graphene-coated polyethylene terephthalate (GPET) substrates.

In this paper, ZnO nanosheets (NSs) with different Al doped concentration on GPET substrates were fabricated by a simple-green hydrothermal method, and their electrochemical properties were studied. The effect of adding different concentrations of Al on electrochemical properties of ZnO composite nanostructures was compared.

# EXPERIMENTAL

# Synthesis of ZnO Nanosheets

The Al-doped ZnO (AZO) NSs with different concentrations were prepared on GPET substrates. The ZnO seed layer (about 30 nm thickness) was sputtered by radio frequency magnetron sputtering on the surface of GPET substrates, which used acetone (10 min), methanol (10 min), and deionized (DI) water to clean in turn by ultrasonic cleaning machines. In the hydrothermal growth process of AZO NSs, zinc nitrate hexahydrate (Zn (NO3)2·6H2O), and hexamethylenetetramine (C6H12N4) were mixed in DI water to prepare precursor solutions (30 ml). Then added aluminum oxide (Al2O3) as dopant to the solutions with the concentration of 0.1 and 0.05 mol/L, and kept stirring for 30 min under mild magnets. The precursor solutions were transferred to a Teflon-lined stainless-steel autoclave, and then the GPET substrates were immersed in it. After that, the autoclave was sealed and put into an oven, and heated at a temperature of 95◦C for 6 h. The products on the substrates were washed with DI water and dried naturally at room temperature.

# Structural Characteristics

Field emission scanning electron microscope (FESEM, by FEI Magellan 400) and X-ray diffraction (XRD, by Rigaku D/MAX-Ultima with Cu Kα radiation) were used to characterize the microscopic morphology and crystal structure of the samples, respectively.

# Electrochemical Measurement

The cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) of the samples have used electrochemical workstation (CHI760E) to test. AZO NSs on GPET substrates as a working electrode, Ag/AgCl as a reference electrode, and platinum foil as counter electrode comprise the three-electrode test system. 1 mol/L Na2SO<sup>4</sup> solution be used as electrolyte in this process of electrochemical measurement.

# RESULTS AND DISCUSSION

The SEM and XRD images of ZnO with different Al doped concentration (doping concentrations of 0, 0.05, and 0.1 mol/L) prepared on GPET substrates are seen in **Figure 1**. As shown in **Figure 1A**, in the case of the undoped Al elements, ZnO has a vertically arranged NR structure with hexagon of its top and a uniformly dense cover on the surface of GPET substrate, suggesting that undoped ZnO has a good degree of orientation. With the addition of Al elements, the structure of ZnO is gradually changed from NR to NS, which clearly observed in **Figures 1B,C**. It is not difficult to see that AZO NSs, with its smooth surface, still grow perpendicularly to the GPET substrate and are connected together to form a network structure. Compared with the ZnO NRs, the conductivity of obtained electrodes of the AZO NSs can be improved due to the fact that the AZO NSs array can develope the branched network. And the pseudo-capacitance of ZnO nanostructure may improve its capacitance value and thus obtain an excellent

electrochemical property (Zhang et al., 2015). As can be seen from the XRD image (**Figure 1D**), except for the characteristic peaks belonging to graphene and GPET substrate appearing at 26◦ and 54◦ , the other peaks are the diffraction peaks of ZnO, which are basically in agreement with the standard PDF card (JCPDS 89-1397) of ZnO. Moreover, the (002) diffraction peak strength is higher than (100) and (101), which indicates that AZO grow preferentially perpendicular to the substrate along the c-axis. The doping of Al generates stress during crystallization, and the crystal structure of ZnO changes accordingly. Further, the intensity of the diffraction peaks of (002) may become weak due to the incorporation of Al elements.

Schematic diagram of hydrothermal growth mechanism of ZnO NSs in **Figure 2** revealed that incorporation of Al inhibits the growth of ZnO NRs, thereby forming AZO NSs. The ZnO crystal has a (0001) plane and a (0001) plane, that is, a Zn positive polar surface and an O negative polar surface and six non-polar surfaces. Al2O<sup>3</sup> dissolved in the solution to produce complexing ions, and the positive polar surface (0001) of the ZnO lattice is more likely to adsorb the Al (OH)<sup>−</sup> 4 complexing ions with negative charges, which can hinder the growth of ZnO along the [0001] direction. The growth of NRs was inhibited along c-axis, which promoted the lateral growth of ZnO, and then formed ZnO NSs (**Figure 2C**; Koh et al., 2004). The main chemical reactions occurring in the solution during the formation of the ZnO NSs were involved in the following Equations (1) and (4):

$$\text{C}\_6\text{H}\_{12}\text{N}\_4 + 10\text{H}\_2\text{O} \rightleftharpoons \text{ } 6\text{HCHO} + 4\text{NH}\_3\cdot\text{H}\_2\text{O} \tag{1}$$

$$\text{Zn}^{2+} + \text{NH}\_3\cdot\text{H}\_2\text{O} \rightleftharpoons \text{Zn(NH}\_3\text{)}\_4^{2+} \tag{2}$$

$$\text{Zn(NH}\_3\text{)}\_4^{2+} + \text{OH}^- \rightleftharpoons \text{ZnO} + \text{NH}\_3\text{H}\_2\text{O} \tag{3}$$

$$\text{Al} + 4\text{OH}^- \rightleftharpoons \text{Al(OH)}\_4^- \tag{4}$$

In order to study the effects of different concentrations of Al doping on the electrochemical characteristics of ZnO nanostructures, the CV curves of three different samples were analyzed under a potential range of −0.8 to 0.9 V at scanning rates of 100 mV/s.

The CV curves, which has been clearly observed by **Figure 3A**, revealed that redox peaks of 0.1 mol/L AZO NSs can be significant observed, which indicated that the synthesized active substances are beneficial for rapid redox reactions (Pu et al., 2014). Comparing the integrated area of the samples on the current-potential axis (**Figure 3A**), it is well-known that the integrated area of the 0.1 mol/L AZO NSs is larger, indicating that the 0.1 mol/L AZO NSs has a stronger charge storage capacity. Further study on the of different scanning rates of 5, 10, 20, 30, 50, and 100 mV/s on the electrochemical characteristics of 0.1 mol/L AZO NSs nanostructures (**Figure 3B**), and the results showed that the electrical current density increases with the increases of scan rates, which confirmed that 0.1 mol/L AZO NSs nanomaterials have excellent scanning ability.

The Nyquist plots of AZO/GPET electrodes are shown in **Figure 4**. The impedance curves of all obtained samples consisted of high frequency zones (shown as semicircle), which reflects the charge transfer resistance (Rct) of the electrode, and low frequency zones (shown as slash),

FIGURE 3 | (A) CV curves of all samples at a scan rate of 50 mV/s. (B) CV curves of 0.1 mol/L AZO nanostructures grown on GPET substrates with different scan rates in the range of 5–100 mV/s.

which mirrors the diffusion resistance of the ions of the electrode. The semicircular diameter of the high-frequency zone is basically the same, indicating that the addition of Al does not enhance the charge transfer ability of ZnO/GPET electrode. In the low frequency zones the diffusion rate of electrode is proportional to the slope of impedance curve. The diffusion rate of AZO/GPET electrode is significantly greater than that of ZnO/GPET electrode, which indicated that AZO/GPET electrode has better electrochemical properties.

#### CONCLUSIONS

In summary, AZO NSs, which are evenly grown perpendicular to the GPET substrate, were successfully prepared using hydrothermal method assisted by ion sputtering. The structure, microscopic morphology and growth mechanism of the samples were analyzed, and it was concluded that the incorporation of Al3<sup>+</sup> inhibited the growth of ZnO NRs, but promoted the formation of ZnO NSs, and weakened the characteristic diffraction peak intensity of ZnO growing perpendicular to the substrate along the c-axis. The electrochemical performance test of the samples concluded that the AZO NSs have

#### REFERENCES


better electrochemical performance than the undoped ZnO NRs, which have broad application prospects in the field of capacitors.

# DATA AVAILABILITY

All datasets generated for this study are included in the manuscript/supplementary files.

# AUTHOR CONTRIBUTIONS

QY contributed to the research and result analysis and discussion. SR, PR, YL, and LJ assisted in the synthesis and analysis of the materials. All authors contributed to the general discussion.

#### FUNDING

This work was financially supported by the National Natural Science Foundation of China (Grant no. 51502166), Scientific Research Program Funded by Shaanxi Provincial Department (Grant no. 17JK0130), and the Industrial Field of Key Research and Development Plan of Shaanxi Province (Grant no. 2018GY-040).


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Yu, Rong, Ren, Jiang and Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Three-Dimensional Graphene-Based Composite Hydrogel Materials for Flexible Supercapacitor Electrodes

Enping Lai\*, Xinxia Yue, Wan'e Ning, Jiwei Huang, Xinlong Ling and Haitao Lin\*

*Guangxi Key Laboratory of Green Processing of Sugar Resources, College of Biological and Chemical Engineering, Guangxi University of Science and Technology, Liuzhou, China*

Three-dimensional (3D) graphene-based hydrogels have attracted great interest for applying in supercapcacitors electrodes, owing to their intriguing properties that combine the structural interconnectivities and the outstanding properties of graphene. However, the pristine graphene hydrogel can not satisfy the high-performance demands, especial in high specific capacitance. Consequently, novel graphene-based composite hydrogels with increased electrochemical properties have been developed. In this mini review, a brief summary of recent progress in the research of the three-dimensional graphene-based composite hydrogel for flexible supercapacitors electrodes materials is presented. The latest progress in the graphene-based composite hydrogel consisting of graphene/metal, graphene/polymer, and atoms doped graphene is discussed. Furthermore, future perspectives and challenges in graphene-based composite hydrogel

#### Edited by:

*Wenyao Li, Shanghai University of Engineering Sciences, China*

#### Reviewed by:

*Shijie Li, Zhejiang Ocean University, China Chengyi Hou, Donghua University, China*

#### \*Correspondence:

*Enping Lai nemodhu@163.com Haitao Lin lhthost@163.com*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *15 August 2019* Accepted: *17 September 2019* Published: *01 October 2019*

#### Citation:

*Lai E, Yue X, Ning W, Huang J, Ling X and Lin H (2019) Three-Dimensional Graphene-Based Composite Hydrogel Materials for Flexible Supercapacitor Electrodes. Front. Chem. 7:660. doi: 10.3389/fchem.2019.00660* for supercapacitor electrodes are also expressed. Keywords: flexible supercapacitor, electrode materials, grahpene-based hydrogel, three-dimensional

# INTRODUCTION

architecture, composite materials

In recent years, supercapacitors have been attracted intensive investigation for sustainable energy application, because of their advantages in excellent power density and high charge/discharge rates (Li et al., 2018). With the rapid growth of wearable electronics, flexible supercapacitors that can work under consecutive bending or stretching are urgently needed. However, it remains a great challenge to obtain supercapacitors with excellent electrochemical performance and good flexibility. Since the supercapacitors mainly consist of two electrodes and a separated membrane, the electrodes are often considered to be a key factor (Xu et al., 2018). Thus, it is critically important to develop innovative materials for flexible supercapacitor electrodes.

As a unique two-dimensional (2D) carbon material, graphene has gained tremendous attention in various application aspects due to its fast-charged carrier mobility, excellent conductivity and largely tunable surface area (Sattar, 2019). Based on the extraordinary physico-chemical properties, graphene and its functionalized derivatives (graphene oxide, GO) can be applied in supercapacitors as the electrode materials. Previous studies have demonstrated that the supercapacitors based on graphene can possess excellent specific capacitance (Horn et al., 2019). Unfortunately, the restacking or irreversible agglomeration of graphene sheets can suppress the high conductivity and decrease accessible surface area, which limit the improvement of capacitive performance. To tackle these challenges, three-dimensional (3D) graphene-based architectures including foam, hydrogels, sponges are developed.

Lai et al. Graphene Hydrogel for Supercapacitor Electrodes

Among various 3D macroscopic structures, graphene hydrogels consist of interconnected porous networks with large specific surface areas have received particular attention (Lu et al., 2017). These hydrogels provide multidimensional ion/electron transport pathways with the intrinsic properties of graphene, which makes them promising candidates for supercapacitor electrodes. Great achievements of supercapacitors based on graphene hydrogels have been obtained, while the pristine graphene hydrogel cannot meet the requirements in practical application (Ma et al., 2018). Functional materials or dopants such as metal oxides or hydroxide, conducting polymers, and so on have been introduced to the graphene hydrogels to further improve the electrochemical performance. In this mini-review, a brief retrospect on graphene-based composite hydrogel materials for flexible supercapacitor electrodes will be provided.

## GRAPHENE-BASED COMPOSITE HYDROGEL ELECTRODE MATERIALS

#### Graphene-Metal Composite Hydrogels Electrode Materials

Usually, metal oxides or hydroxides exhibit high pseudocapacitance mainly due to their faradic reaction beyond formation of electrical double-layers in the charge-discharge processes. Thus, metal oxides or hydroxides including MnO2, Ni(OH)2, and etc. have been widely incorporated into graphene to form composite materials to obtain enhanced performance.

Researches have been focused on coupling MnO<sup>2</sup> with other materials (Xu et al., 2019) for supercapacitor electrodes, and different structures of MnO<sup>2</sup> have also been incorporated in graphene hydrogel. Zhang et al. (2016) prepared micronanostructured pompon-like MnO2/graphene hydrogel composites, and the hydrogel with a MnO<sup>2</sup> content of 50% displayed a good capacitive behavior (445.7 F/g at 0.5 A/g). Tran et al. (2017) synthesized graphene/α-MnO<sup>2</sup> nanowire hydrogel with a high specific capacitance (397 F/g) at 1.0 A/g. Meng et al. (2018) designed a glucose and ammonia reduction system to synthesize δ-MnO2/graphene hydrogel with a capacity of 200.6 F/g.

As a transition metal hydroxide, Ni(OH)<sup>2</sup> have been commonly used. In view of the structure, the Ni(OH)<sup>2</sup> nanoplate (Mao et al., 2016) and nanoflower structure (Wang et al., 2016) have been designed to form 3D graphene-based frameworks, and the capacitance can be achieve to 782 F/g at 0.2 A/g and 1,632 F/g at 1 A/g, respectively. Recently, Li et al. (2019) coupled Ni(OH)<sup>2</sup> nanosheets with nitrogen-enriched graphene hydrogel, which featured a specific capacitance of 896 F/g at 0.5 A/g.

Besides, in consideration of the advantage of Co and Ni, the strategy of combining Ni with Co was developed. Hwang et al. (2018) embedded Ni-Co hydroxide nanoneedles in graphene hydrogel, and the nanocomposite exhibited a specific capacitance of 544 C/g at 2 A/g. Tiruneh et al. (2018) designed a binder-free hybrid graphene hydrogel with nickel cobalt sulfide embedded, which can exhibit a capacity of about 1,000 F/g at 0.75 A/g with outstanding stability.

In addition, as one of the most promising nanomaterials, TiO<sup>2</sup> has also been focused. Liu et al. (2017) obtained rice-like TiO2/graphene composite hydrogel, and the interaction between TiO<sup>2</sup> nanoparticles and graphene hydrogel can endow the composite hydrogel with superior physicochemical properties, resulting in an excellent capacity of 372.3 F/g at 0.2 A/g. In addition, to realize synergic effect of organic and inorganic materials, Zhang et al. (2018b) synthesized RuO2/graphene hydrogel, and then adsorbed 1,4-naphthoquinone (NQ) molecules onto the hydrogel to form an hybrid graphene hydrogel, and the hydrogel containing ∼14.6% RuO<sup>2</sup> also can show a superior specific capacitance of 450.8 F/g.

### Graphene-Polymer Composite Hydrogels Electrode Materials

Conductive polymers with the reverse doping-dedoping behavior have been extensively incorporated into 3D graphene hydrogels to fabricate high-performance electrodes. One of the successful conducting polymers is polyaniline (PANI), which has good environmental stability and high pseudocapcacitance. Various types of PANI including nanorods and nanowires have been used to form graphene/PANI composite hydrogel, and these hydrogels can show improved specific capacitance (Chen et al., 2017; Xu et al., 2017). To retain the essential features of the native hydrogel, the thin PANI layer wrapped on the graphene hydrogel by Gao et al. (2016) using an electrodeposition method, and the hydrogel displayed a good specific capacitance (710 F/g at 2 A/g). In addition, the hydrogel structure and the preparation temperature are also important. Wu et al. (2016) found that the phaseseparated structure can produce much channels for electrolyte diffusion in PANI/graphene hydrogel, leading to a specific capacitance as high as 783 F/g at 27.3 A/g. Zou et al. (2018b) used m-phenylenediamine (mPD) to preserve the conjugated structure of PANI, and the composite hydrogel showed an improved capacity of 514.3 F/g. They also developed doublecrosslinked network functionalized graphene/PANI hydrogel with high specific capacitance and mechanical strength (Zou et al., 2018a). For the preparation temperature, Ates et al. (2018) showed that the high capacitance can be kept at ∼99% of its pristine value for the hydrogel prepared at 25◦C, in comparison of that at 0◦C.

Due to the ultrahigh theoretical capacitance and mechanical flexibility, polypyrrole (PPy) has also attracted considerable attention. Wu and Lian (2017) synthesized graphene/PPy hydrogel with a specific capacitance of 363 F/cm<sup>3</sup> at 1.0 mA/cm<sup>3</sup> using hydroquinone as a functionalized molecule. Since the ion transportation may be restricted by the compact morphology of PPy/graphene composite, the strategy of PPy wrapped graphene was developed. Pattananuwat and Aht-ong (2017) controlled nanoporous structure for PPy coated on graphene hydrogel surface with aiding of the surfactant, and a high specific capacitance (640.8 F/g at 1 A/g) can be achieved. They also used poly(3,4-ethylenedioxythiophene) (PEDOT) with PPy to form the synergistic effect, and the resultant hydrogel exhibited a specific capcacitance of 342 F/g at 0.5 A/g (Pattananuwat and Aht-ong, 2016). Recently, a hybrid PPy/rGO hydrogel in-situ electropolymerization preparation of PPy on the outside layer of graphene was reported by Zhang et al. (2019), and the specific capacitance is 340 F/g at 1 A/g.

Except for PANI and PPy, biopolymer such as lignosulfonate that a derivate of lignin can be regarded as a candidate for electrode material due to its electroactive components. Xiong et al. (2016) prepared lignosulfonate/graphene hydrogel with a maximum capacity of 549.5 F/g at 1 A/g. Li et al. (2017a) also reported lignosulfonate functionalized graphene hydrogels, which can present a specific capacitance of 432 F/g. These renewable composite hydrogels exhibit great potential in supercapacitor as electrodes materials.

### Doped Graphene Composite Hydrogels Electrode Materials

As known, chemical doping is an effective strategy to modify the intrinsic properties of the materials. For graphene materials, nitrogen doping has been found to be an important method to provide graphene composite with high capacitance (Jin et al., 2019). For example, Liao et al. (2016) used urea and a small amount of ammonia to prepare N-doped graphene hydrogel, and the different N-types in graphene can resulted in an excellent specific capacitance (387.2 F/g at 1 A/g). Except for urea and ammonia, other materials containing N also have been used. Jiang et al. (2016) designed N-doped graphene hydrogel with a good specific capacitance (167.7 F/g at 1 A/g) by using carbohydrazide. Liu et al. (2016) used ammonium bicarbonate to prepared porous N-doped graphene hydrogel, and the hydrogel with high N content (10.8 at%) showed a high specific capacitance of 194.4 F/g. Gao et al. (2019) synthetized nitrogen-doped graphene-based hydrogels with concentrated sulfuric acid and o-phenylenediamine (oPDA), and the optimal hydrogel showed specific capacitance about 519.8 F/g.

Compared to single-atom doping, multiple doping can exhibit a synergetic effect and further improve the capacitive behavior of the materials. Typically, nitrogen and sulfur can be doped into graphene concurrently. After co-doping, the pseudo capacitance of the graphene will be increase, because of the redox faradic reactions existed at nitrogen-containing groups and sulf-containing species. Tran et al. (2016) outlined that N and S co-doped graphene hydrogel with holy defect can show a wonderful specific capacitance of 538 F/g at 0.5 mV/s, and the electrochemical property of the hydrogel can be modulated by the level of N and S doping. Li et al. (2017b) synthesized N/S co-doped graphene hydrogels with hierarchical pores, which can demonstrate a very good specific capacitance (251 F/g at 0.5 A/g) surprisingly. Zhang et al. (2018a) developed a one-step method to synthesize N/S co-doped graphene hydrogel, and the as-prepared hydrogel can show a capacity of 1,063 C/g at 1 A/g. Kong et al. (2018) constructed N,S-codoped graphene hydrogel with 3D hole, and the abundant in-plane pores leading to an outstanding specific capacitance of 320.0 F/g at 1 A/g.

#### CONCLUDING REMARKS

In this work, the recent advances in three-dimensional graphenebased composite hydrogel were reviewed in term of their use TABLE 1 | Characteristics of typical graphene-based composite hydrogel for supercapacitor electrodes.


as flexible supercapacitors electrode materials. As can be seen, much effort has been devoted in the field of 3D graphene-based composite hydrogel electrode materials. A variety of materials including metals, polymers, dopants have been reported to construct high performance graphene-based composite hydrogel, and the excellent capacitive behavior of the typical composite hydrogel can be achieved (**Table 1**). Nevertheless, a massive effort is still needed before real practical applications become possible. For one, the exhibited capacitance values of the hydrogel still need to be further enhanced. In particular, large scale production of graphene-based composite hydrogel with high capacitance and quality is still a challenge. More importantly, for industry-level applications, technical movements should focus on more cost-effective and straightforward approaches for the fabrication of 3D graphene-based composite hydrogel rather than the design of complicated nanomaterials. It is believed that continuous breakthroughs in the graphene-based composite hydrogel will be made with the further research and development in this exciting field, and the hydrogel can play a great role in flexible capacitive devices in the near future.

## AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

#### FUNDING

This work was supported by the Natural Science Foundation of Guangxi Province (No. 2018GXNSFBA281084) and the Middleaged and Young Teachers Basic Ability Promotion Project of Guangxi (No.2019KY0371).

### REFERENCES


graphene/polyaniline hydrogels for high performance flexible supercapacitors. J. Mater. Chem. A 6, 9245–9256. doi: 10.1039/C8TA01366G

**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Lai, Yue, Ning, Huang, Ling and Lin. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Capacity Contribution Induced by Pseudo-Capacitance Adsorption Mechanism of Anode Carbonaceous Materials Applied in Potassium-ion Battery

Jiahao Liu1†, Ziqiang Xu1†, Mengqiang Wu<sup>1</sup> \*, Yuesheng Wang<sup>2</sup> \* and Zaghib Karim<sup>2</sup> \*

<sup>1</sup> School of Materials and Energy, University of Electronic Science and Technology of China, Chengdu, China, <sup>2</sup> Center of Excellence in Transportation Electrification and Energy Storage, Hydro-Québec, Varennes, QC, Canada

#### Edited by:

Min Zeng, Lanzhou Institute of Chemical Physics (CAS), China

#### Reviewed by:

Dongliang Chao, University of Adelaide, Australia Chao Wang, Massachusetts Institute of Technology, United States

#### \*Correspondence:

Mengqiang Wu mwu@uestc.edu.cn Yuesheng Wang wang.yuesheng@ireq.ca Zaghib Karim Zaghib.Karim@hydro.qc.ca

†These authors have contributed equally to this work

#### Specialty section:

This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry

Received: 08 July 2019 Accepted: 05 September 2019 Published: 02 October 2019

#### Citation:

Liu J, Xu Z, Wu M, Wang Y and Karim Z (2019) Capacity Contribution Induced by Pseudo-Capacitance Adsorption Mechanism of Anode Carbonaceous Materials Applied in Potassium-ion Battery. Front. Chem. 7:640. doi: 10.3389/fchem.2019.00640 The intrinsic bottleneck of graphite intercalation compound mechanism in potassium-ion batteries necessitates the exploitation of novel potassium storage strategies. Hence, utmost efforts have been made to efficiently utilize the extrinsic pseudo-capacitance, which offers facile routes by employing low-cost carbonaceous anodes to improve the performance of electrochemical kinetics, notably facilitating the rate and power characteristics for batteries. This mini-review investigates the methods to maximize the pseudo-capacitance contribution based on the size control and surface activation in recent papers. These methods employ the use of cyclic voltammetry for kinetics analysis, which allows the quantitative determination on the proportion of diffusion-dominated vs. pseudo-capacitance by verifying a representative pseudo-capacitive material of single-walled carbon nanotubes. Synergistically, additional schemes such as establishing matched binder–electrolyte systems are in favor of the ultimate purpose of high-performance industrialized potassium-ion batteries.

Keywords: potassium-ion batteries, carbonaceous anodes, pseudo-capacitance adsorption, surface doping activation, kinetic analysis

# INTRODUCTION

Suffering from the geopolitical maldistribution of lithium resources, sodium-ion batteries (SIBs) and potassium-ion batteries (PIBs) reach a hotspot in view of wider resource reserves compared with lithium-ion batteries (LIBs) (2.09 wt% of K vs. 2.36 wt% of Na vs. 0.0017 wt% of Li) (Carmichael, 1989; Larcher and Tarascon, 2015). Significantly, the PIB system has the lowest negative potential (0.15 V below the Li/Li+) (Komaba et al., 2015; Eftekhari et al., 2016; Wang et al., 2018a,b) and satisfactory electrochemical kinetics in ionic diffusion kinetics and conductivity theoretically (Okoshi et al., 2013, 2017; Komaba et al., 2015; Eftekhari et al., 2016; Su et al., 2016) in non-aqueous electrolytes, ascribed to low de-solvation due to its weak Lewis acid character (Okoshi et al., 2017; Lei et al., 2018). Similar to the behavior of LIBs in graphite (Wang, 2017), the intercalation mechanism of PIBs involves three potassiation stages, generating the KC<sup>36</sup> in Stage III, KC<sup>24</sup> in Stage II, and finally the KC<sup>8</sup> in Stage I (Jian et al., 2017) with 270 mA h g −1 , which is far more stable than SIBs (Wang et al., 2013; Zheng et al., 2017). Nevertheless, large Shannon ionic radius (K<sup>+</sup> = 1.38 Å, Na<sup>+</sup> = 1.02 Å, Li<sup>+</sup> = 0.76 Å) and atomic mass (K = 39.10, Na = 22.99, Li = 6.94) (Shannon, 1976) have decreased the theoretical capacity and induced high volume expansion of 61% (Wen et al., 2015; Eftekhari et al., 2016; Zou et al., 2017). Although some strategies enter into consideration such as adopting expand graphite (An et al., 2018), implementing solvent co-intercalation (SCI) (Wang et al., 2013, 2019; David and Singh, 2014), and developing dual-carbon batteries (DCBs) (Carlin et al., 1996, 2010; Beltrop et al., 2017; Fan et al., 2017; Ji et al., 2017), essential kinetics deficiency is hard to surmount.

To address the irreversible expansion induced by equilibrium graphite, amorphous carbons come into the focused sight (Xing et al., 2017), which are admittedly classified as hard carbon (HC) and soft carbon (SC). HC is proven to have a prolonged cycling ability for its randomly oriented bend graphitic layers along the c axis without observed expansion after a thorough potassiation (Jian et al., 2016, 2017). On the contrary, SC is easily graphitized with turbostratic domains although far from the commercial graphite. SC presents obvious expansion to HC if undergoing complete potassiation (Luo et al., 2015a; Wang et al., 2017). Nonetheless, SC has a better rate performance than HC for more aligned domains (Jian et al., 2017), regarded as the reason for the better rate performance.

Pseudo-capacitance is the middle part of the battery and electrical double-layer capacitors (EDLCs) (Jiang and Liu, 2019) as shown in **Figure 1a**. In **Figure 1a**, Liu points out that the current generation of batteries depends on the Faradic electron transfer from the surface to the metal center based on the chargecompensating ions by intercalation or adsorption. In contrast, a pseudo-capacitor is different from EDLCs because it is not electrostatic-induced and the transfer process of surface electrons distinguishes the behavior from batteries. Pseudo-capacitance can be classified into two categories—intrinsic and extrinsic. The former (Chao et al., 2016) describes an inherent feature of specific materials such as RuO<sup>2</sup> and MnO2, which is on the strength of Faradaic electron transfer. However, the latter emphasizes the technological means of low dimension, nanoscale size, and high surface area (Wang et al., 2007; Brezesinski et al., 2009; Muller et al., 2015; Cook et al., 2016) among a majority of materials, for instance, the single-walled carbon nanotube (SWCNT) with surface-enriched potassium ions in **Figure 1b** (Hersam, 2009; Kang et al., 2013), attributed to the regular hexagonal arrangement of carbon atoms on the surface. This non-Faradaic pathway provides a possibility to utilize inexpensive carbonaceous anodes (Gogotsi and Penner, 2018) if proper surface treatments such as activization, doping, and plasma processing have been undergone (Chao et al., 2018).

These surface-dominated anodes uptake and absorb potassium ions with fast reaction kinetics during the electrochemistry process, avoiding the hindrance in the intercalation mechanism, which is regarded as the primary cause for high rate property and large capacity of PIBs undergoing charging and discharging processes (Shin et al., 2011; Shao et al., 2013; Chen et al., 2016; Long et al., 2016). As a consequence, this mini-review analyzes the methods for distinguishing the proportion of capacity contribution and summarizes the application of pseudo-capacitance to PIBs very recently, aiming to design a practical performance improvement approach.

# ANALYSES AND APPLICATIONS OF PSEUDO-CAPACITANCE

Generally, it is widely admitted that pseudo-capacitance is not a pure Faradaic progress but a rapid reversible surface redox reaction involved in EDLCs. The charge storage mechanism of complex PIBs behaviors is composed of two typical contribution progresses: surface-induced pseudo-capacitor process and diffusion-dominated process (Brezesinski et al., 2009; Wen et al., 2015; Xu et al., 2019b). Nonetheless, Faradaic and non-Faradaic reactions are electro indistinguishable for jointly contributing to the current parameter (Gogotsi and Penner, 2018). Most researchers employ cyclic voltammetry (CV) to determine the relative proportion of contribution from pseudo-capacitor and diffusion-dominated processes. The peak current is proportional to the square root of sweep rate describing the reversible diffusion-limited state (i-v 1 <sup>2</sup> ), whereas it is proportional to the sweep rate (i-v) describing the capacitive state. A representative power law relationship between the current and scan rate reveals the charge storage mechanism in PIBs (Wang et al., 2007; Torsten et al., 2010; Veronica et al., 2013):

$$\mathbf{i} = \mathbf{a}\boldsymbol{\nu}^b$$

where a and b are constants. The b value can be figured out by profiling the log(i)–log(v) curve. If b = 0.5, the Faradic diffusion is predominant; while b = 1, the pseudo-capacitance assumes the primary contribution (Sathiya et al., 2011; Lijun et al., 2014; Zou et al., 2017). Furthermore, as for a fixed sweep rate, the specific pseudo-capacitance contribution can be given in detail by the following formula (Torsten et al., 2010; Wang Y. et al., 2016):

$$\mathbf{i} = k\_1 \boldsymbol{\nu} + k\_2 \boldsymbol{\nu}^{\frac{1}{2}}$$

where the parameter k1v represents the capacitive process while the k2v 1 <sup>2</sup> is in favor of the diffusion process as stated earlier.

Furthermore, Marveh maintains that compared with the CV method, the step potential electrochemical spectroscopy (SPECS) has wider adaptive range with prominent advantages. In high sweep rates, SPECS presents precise characterization to depict the process of electrical double layer on the surface of electrodes (Forghani and Donne, 2018).

Recent works validate the validity of the pseudo-capacitance algorithm based on the surface-dominated pseudo-capacitance mechanism, which has been extensively applied in carbonaceous anodes in PIBs by constructing high surface area or activating.

Doping and activating are highly feasible methods that introduce abundant defects, expand specific surface area, and promote the conductivity, meanwhile adding charge storage for PIBs (Lijun et al., 2014; Share et al., 2016; Chen et al., 2017; Lei et al., 2018; Xu et al., 2018).

Nitrogen-doped strategy has a practical significance eliciting satisfying performance enhancements. According to the X-ray photoelectron spectroscopy (XPS) results, pyrrole nitrogen (N-5), pyridine nitrogen (N-6), and quaternary N (N-Q) are three Ndoping forms presented in **Figure 2A**, where N-5 and N-6 possess high electrochemical activity and generate additional defects in the surface of the graphene layer, hence promoting the adsorption

quantity of potassium ions, accelerating the kinetic process (Li et al., 2013; Wang et al., 2014; Xu et al., 2018). This differentiation of N forms is ascribed to their constructions of respective bonding electrons, resulting in different chemical activities. However, N-Q, located in the internal surface of graphene layer, bonding with three sp2 carbon atoms, is beneficial to improve electrical conductivity (Yang et al., 2018). Notably, N-6 is regarded as the most effective doping precursor because it replaces the carbon atom with a nitrogen atom at the defect or the edge of the graphite plane and occupies abundant active centers to adsorb potassium ions (Ma et al., 2012; Ding et al., 2014; Xie et al., 2017). Consistently, recent researches demonstrated that N-6 defects decreases with temperature increasing; meanwhile, the degree of graphitization rises, accompanied by the generation of N-Q. Xu concludes that among three temperature-controlled materials NCNF-650, NCNF-950, and NCNF-1100 derived from poly-pyrrole nanofibers, the pseudo-capacitance contribution of NCNF-650 occupies 90% at 1 mV s−<sup>1</sup> for abundant N-6 defects, while the others occupy 73 and 84% at 1 mV s−<sup>1</sup> as displayed in **Figure 1c** with their b value (Xu et al., 2018). The b value increases with the temperature dropping, revealing the degree deepening of pseudo-capacity, in accordance with the quantity N-6 defects. Similar results are obtained in Xie's report; nonetheless, Xie indicates that enhancement of electrochemical performance is a comprehensive result associated with N-6 defects, electrical conductivity, and transfer resistance. Three sets of temperature-controlled experiments point out that PNCM-700 is equipped with the best comprehensive performance compared with PNCM-500 and PNCM-900. As for the function of defects, Clement states that the D-bond from the Raman spectrum is employed to describe the sp<sup>3</sup> defect distribution, which intensifies a six-fold rate performance to the un-doped material (Clement et al., 2015; Share et al., 2016). The prominent significance of N-doping is to spread out the interlayer spacing and provide huge specific surface area to promote the pseudocapacitive effect.

Doping some other elements also achieves fair results. Oxidation functional groups on the carbon surface polishes up the wettability of carbon-based materials and advances the pseudo-capacitance behavior (Tarun et al., 2010; Shao et al., 2013; Wang X. et al., 2016; Wu et al., 2016; Xie et al., 2017; Wang et al., 2018). Adams reported that oxidation groups increase obviously on the surface capacitive storage while inducing the capacity reduction contributed by the intercalation mechanism. As a consequence, there is no significant enhancement to the total capacity (Adams et al., 2017). In addition, mixed-doping P and O doping (Ma et al., 2017) based on the triphenylphosphine precursor obtains a satisfactory capacity of 474 mA h g−<sup>1</sup> , benefiting from expanding the interlayer spacing; N and F doping immensely adds the conductivity distinctly (Ju et al., 2016; Share et al., 2016; Adams et al., 2017).

Activated hollow carbon nanospheres (HCS) underwent HF etching from C@SiO2 nanospheres in Wang's work. Wang emphasizes on the sharp increase on the surface area from 481.4 to 757.8 m<sup>2</sup> g −1 after activating utilizing KOH as the activator. Capacitive contribution occupies 71.2% at a sweep rate of 1 mV s −1 , leading to 192.7 mA h g−<sup>1</sup> at 2 A g−<sup>1</sup> after 5,000 cycles with a retention of 99.5% (Wang et al., 2018). Aforesaid data support the rule that the pseudo-capacity contribution has the tendency of positive correlation with sweep rate as summarized in **Figure 2B**. This work claimed that activated hollow carbon expands the layer spacing of the carbon anode and shortens the diffusion distance of K-ions.

Nevertheless, surface-modified strategies, whether doping or activating, may give rise to the decrease of initial Coulombic

#### REFERENCES


efficiency (ICE) unsatisfactorily. Compensatory methods work well in LIBs and SIBs (Suo et al., 2017) to establish appropriate binder–electrolyte systems (BESs), which directly impact the formation of solid electrolyte interphase (SEI), especially the morphology features such as thickness, pore, and wrinkle. Tailored BESs shape the SEI into a smooth and thin layer, hence improving the transfer efficiency of ions on the phase interface (Xu et al., 2019a). Similarly, employing KFSI and KTFSI electrolyte (Eftekhari et al., 2016; Jin et al., 2016), adding electrolyte additives (Wu et al., 2017), selecting hydrophilic binders such as CMC, PANa, and SA (Komaba et al., 2015; Luo et al., 2015b; Jin et al., 2016; Xu et al., 2019a), and utilizing prepotassiation technique (Yang et al., 2018) serve the same purpose for superb PIBs.

## CONCLUSIONS AND PERSPECTIVES

PIBs with carbonaceous anodes provide the possibility for industrialization under controlled price. Facilely, surface modifications such as doping and activating obviously enhance the pseudo-capacitance contribution, speeding up the rate and power performance based on a rapid electrochemical kinetics.

This non-insertion charge storage (pseudo-capacitance absorption) integrates with both battery-type and capacitortype characteristics, exhibiting distinct redox separation peaks including analogous linear capacitive voltage response. However, the relationship between capacitive and sweep rate is only authentic limited in a low and narrow sweep rate under the CV separation method. Precisely, the SPECS is suitable for a wider range of sweep rates, inducing detailed contribution information for each potential point (Forghani and Donne, 2018). In addition, matching binder–electrolyte with anodes accurately can synergistically promote specific capacity and rate properties, deriving high-performance PIBs.

# AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

# FUNDING

This work was supported by the Sichuan Science and Technology Program (Grant Nos. 2018GZ0006 and 2018GZ0134).


at graphitic and nongraphitic electrodes. J. Appl. Electrochem. 26, 1147–1160. doi: 10.1007/BF00243740


**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Liu, Xu, Wu, Wang and Karim. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# N-Propyl-N-Methylpyrrolidinium Difluoro(oxalato)borate as a Novel Electrolyte for High-Voltage Supercapacitor

#### Weili Zhang, Fuming Zhang, Peng Zhang, Shuo Liang and Zhiqiang Shi\*

*Tianjin Key Laboratory of Advanced Fibers and Energy Storage, College of Materials Science and Engineering, Tianjin Polytechnic University, Tianjin, China*

Development of high voltage electrolyte is one of the effective ways to improve the performance of supercapacitor. The new ionic liquid N-propyl-N-methylpyrrolidinium difluoro(oxalato)borate (Py13DFOB) was designed and mixed with propylene carbonate (PC) as electrolyte for supercapacitor. The operating voltage of the new electrolyte system has been proven to be up to 3.0 V by a series of electrochemical techniques. Surprisingly, the new salt exhibits nearly symmetric capacitance contribution in the positive and negative electrodes, leading to a high capacitance value of 130 F g−<sup>1</sup> . The energy and power density of EDLCs using Py13DFOB in the PC electrolyte reach 39.06 Wh kg−<sup>1</sup> (100 mA g−<sup>1</sup> ) and 8.03 kW kg−<sup>1</sup> (5,000 mA g−<sup>1</sup> ), respectively, at the working voltage of 3.0 V, significantly exceeding the performance of commercial electrolyte tetraethylammonium tetrafluoroborate (TEABF4). The results indicate that Py13DFOB can be a promising electrolyte salt for supercapacitor.

#### Edited by:

*Yuanlong Shao, King Abdullah University of Science and Technology, Saudi Arabia*

#### Reviewed by:

*Xingbin Yan, Lanzhou Institute of Chemical Physics (CAS), China Jun Jin, Shanghai Institute of Ceramics (CAS), China*

#### \*Correspondence:

*Zhiqiang Shi shizhiqiang@tjpu.edu.cn*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *12 July 2019* Accepted: *18 September 2019* Published: *09 October 2019*

#### Citation:

*Zhang W, Zhang F, Zhang P, Liang S and Shi Z (2019) N-Propyl-N-Methylpyrrolidinium Difluoro(oxalato)borate as a Novel Electrolyte for High-Voltage Supercapacitor. Front. Chem. 7:664. doi: 10.3389/fchem.2019.00664* Keywords: N-propyl-N-methylpyrrolidinium difluoro(oxalato)borate, supercapacitors, ionic liquid, electrolyte, high voltage

# INTRODUCTION

Supercapacitors (SCs), a kind of electrochemical energy storage device with high power density, long cycle life, and excellent reliability, has been widely used in many fields such as hybrid electric vehicle and high-power output equipment. However, the low energy density is the main factor hindering their further applications (Wang et al., 2012; Simon et al., 2014). It can be obtained from the calculation formula of energy density that expanding the voltage window of the cell is the most effective way for achieving high energy supercapacitor (Li et al., 2007; Snook et al., 2011; Boukhalfa et al., 2012; Díaz et al., 2012; Kato et al., 2012; Okashy et al., 2013; Xiang et al., 2013; Borenstein et al., 2014, 2015; Choi et al., 2014; Kumar et al., 2016). As a matter of fact, the working voltage V depends to a great extent on the stability of the electrolyte. Unfortunately, tetraethylammonium tetrafluoroborate (TEABF4) as the state-of-the-art electrolyte material can only withstand a working voltage of 2.5–2.7 V, which is usually limited by the oxidation and reduction stability of electrolyte ions.

Recently, Ionic liquids (ILs) have been intensively studied and viewed as potentially ideal electrolytes for increasing the operating voltage of EDLCs due to their relatively wide electrochemical stability (Ue et al., 2003; Zhu et al., 2007; Kim et al., 2014). In addition, ILs are attracted to supercapacitors owing to several other excellent properties in terms of non-volatile, non-flammable, and high thermal stability. However, most ILs are trapped in their low ionic conductivity and high viscosity compared to aqueous electrolytes and even organic electrolytes. Considering this situation, mixing ILs with organic solvents is a promising alternative strategy to enlarge the working voltage without sacrificing the power density and cycle life of the EDLCs (Guerfi et al., 2010; Kühnel et al., 2011). The introduction of organic solvent not only reduces the viscosity and increases the conductivity of the pure ionic liquid, but also maintains a large electrochemical stability window, which greatly improves the capacitive performance of the device.

The most conventional families of ionic liquids, which have been evaluated as the most prospective electrolytes for supercapacitors, are based on pyrrolidinium and imidazolium cations (Mousavi et al., 2016; Watanabe et al., 2017). In general, pyrrolidinium based ILs could deliver noticeably enhanced electrochemical stability than the ones based on imidazolium owing to its superior ability to resist oxidation and reduction, which are suitable for realizing novel symmetric supercapacitor with high working voltage (Lin et al., 2011; Brandt et al., 2013; Zhang et al., 2016; Martins et al., 2018). Moreover, the mixing of the ionic liquid with the organic solvent can substantially ignore the high viscosity of the pyrrolidinium-based ionic liquid. The choice of anions also has a great effect on the properties of the ionic liquid. Difluoro(oxalate)borate (DFOB) have received particular interest in recent years due to its high asymmetry compared to conventional anions, resulting in higher solubility of electrolyte salt in ester solvents. Also, the electron-withdrawing fluorine atom on DFOB leads to more delocalization charges, causing lower affinity of binding cation and higher conductivity of electrolyte salt. More importantly, DFOB possesses high electrochemical stability and non-corrosive to aluminum current collector, rendering it an ideal choice for high-voltage electrolytes (Lai et al., 2011; Allen et al., 2013; Tian et al., 2014; Wu et al., 2015). In this manuscript, we designed a new ionic liquid Py13DFOB mixed with an organic solvent PC as electrolyte for supercapacitor. The physicochemical properties (conductivity, melting point, thermal stability, etc.) of the new electrolyte salt were characterized for the first time. Afterwards, we evaluated the performance of supercapacitors containing Py13DFOB/PC electrolyte from the aspects of withstand voltage, cycle stability, energy density, power density, etc.

#### EXPERIMENTAL

#### Materials

N-Methyl pyrrole (>99%), 1-Bromopropane (>98%), acetonitrile (>99.5%) were obtained from Aldrich and used without further purification. The PC solvent (battery grade, extra dry <20 ppm of water) and Lithium difluoro(oxalato)borate (>99%) were purchased from Jiangsu Guotai Super Power New Materials Co. Ltd. (China). The prepared electrolyte salt (Py13DFOB) was placed in a glove box filled with high pure argon (<1 ppm O<sup>2</sup> and <1 ppm H2O), then dissolved in the PC with 1 mol L−<sup>1</sup> concentration, added 3Å molecular sieves to remove trace moisture for 1 week. The final water content is <20 ppm testing by Karl Fischer titration method (Mettler-Toledo C20, Switzerland). The impurities of halide and alkali metal ions were <2 ppm confirmed by Inductive Coupled Plasma Emission Spectrometer (ICP) test.

The activated carbon electrodes were prepared by mixing 82% activated carbon (Kuraray YP-50), 10% carbon black (VXC72) as the conductor, 4% carboxymethylcellulose sodium (CMC), and 4% styrene butadiene rubber (SBR) as the binder. The mixture was stirred to a sticky state followed by coating on aluminum foils. The electrodes were punched into disks with a diameter of 18 mm, then dried under vacuum at 120◦C for more than 12 h prior to be used. The mass of each electrode is about 6 mg and the thickness is ∼60 um (including 20 um aluminum foil).

## Electrolyte Synthesis/Purification

N-Methyl pyrrole (8.11 g, 0.1 mol) and Lithium difluoro(oxalato)borate (14.38 g, 0.1 mol) were dissolved in 100 mL acetonitrile. To this solution, 1-Bromopropane (12.99 g, 0.1 mol) was added dropwise for 12 h. The crude products were obtained by filtration and rotary evaporation of the reaction solution and then purified by extraction with ethyl acetate/deionized water. Finally, the target product of high purity clear liquid was obtained after decoloring by activated carbon. The resulting electrolyte salts are dried for 48 h at 80◦C and stored in a sealed container in a glove box with high pure argon (<1 ppm O<sup>2</sup> and <1 ppm H2O).

#### Characterization and Measurements

<sup>1</sup>H-NMR and <sup>13</sup>C-NMR spectra were performed on a Bruker AVANCE 400M spectrometer. Thermal gravimetric analysis was tested by a Netzsch thermogravimetric analyzer at a heating rate of 5◦C min−<sup>1</sup> from 20 to 600◦C under nitrogen conditions. Differential scanning calorimetry (DSC) analysis was performed on a TA Instruments Q2000 differential scanning calorimeter at a heating rate of 10◦C min−<sup>1</sup> under nitrogen atmosphere. The relationship between conductivity and concentration of Py13DFOB/PC was determined by using a conductivity meter (Mettler-Toledo S30, Switzerland). The viscosity of the electrolyte is tested by a viscosity testing device (A&D SV-10, Japan). Activated Carbon/Activated carbon symmetrical coin cells were prepared in an argon-filled glove box for electrochemical measurements of electrolytes. The galvanostatic charge/discharge tests (GCD) and cycling performance were tested at Arbin battery test system. Cyclic voltammetry (CV) was tested in the same range by Autolab electrochemical workstation (PGSTAT302N, Switzerland). The gravity specific capacitance of single electrode was obtained from equation C<sup>m</sup> = 2I1t/m1V, in which I is the current value, 1t is the discharge time, 1V is the potential difference between the end of the voltage drop and the end of the discharge, and m is the mass of the active material of single electrode. The energy density is calculated from the formula E = 1/2C1V 2 , where C is the specific discharge capacitance. The power density was calculated by the equation P = I1V/2m.

## RESULTS AND DISCUSSIONS

# Chemical Structure and Physical Properties Characterization of Synthesized Sample

<sup>1</sup>H NMR and <sup>13</sup>C NMR analysis were used to verify the purity of our synthesized sample. The results are as follows: <sup>1</sup>H NMR (D2O) δ: 3.42∼3.34 (m,4H), 3.17∼3.13 (m,2H), 2.90 (s,3H), 2.12∼2.06 (m,4H), 1.74∼1.64 (m,2H), 0.87∼0.83 (t,3H), <sup>13</sup>C-NMR (D2O): δ = 9.99, 16.75, 21.24, 47.97, 64.17, 65.71, 163.76 ppm, which confirm that Py13-DFOB is synthesized successfully. The original NMR spectrum is shown in **Figure S1**.

The chemical structure and ionic size of the synthesized Py13DFOB are shown in **Figure 1** and several basic physical properties (conductivity, viscosity, thermal stability, etc.) of the new electrolyte were summarized in **Table 1**. The phase transformation behavior of Py13DFOB investigated by DSC is illustrated in **Figure 2A**. Py13DFOB shows a low melting point value of 3.21◦C, which is due to the unfavorable packing of ions and the decrease of lattice energy of IL materials caused by the highly spatial asymmetry of Py<sup>+</sup> <sup>13</sup> and the large Vander Waals volume of DFOB−. TG analysis is used to observe the thermal stability of electrolytes. As can be seen in **Figure 2B**, Py13DFOB is subjected to two-step degradation, where the initial degradation temperature of 290◦C is considered to meet the thermal stability requirements of supercapacitors. The result is similar with the decomposition curve of the LiDFOB mentioned in the previous study (Allen et al., 2011).

Obviously, **Table 1** shows that the pure ionic liquid exhibits lower conductivity and higher viscosity than the organic electrolyte. Therefore, the mobility of the ions is enhanced by


the addition of the organic solvent PC, thereby making it easier for the electrolyte to access the pores of the activated carbon. As shown in **Figure 3**, the maximum conductivity of Py13DFOB/PC was observed to reach 14.5 mS cm−<sup>1</sup> at 2.0 mol L−<sup>1</sup> , and the conductivity drops sharply as the concentration exceeds or <2.0 mol L−<sup>1</sup> . Besides, the addition of the solvent significantly reduces the viscosity of the pure ionic liquid, taking a 1 mol L−<sup>1</sup> Py13DFOB/PC as an example, the viscosity value of which is only 3.8% of the pure ionic liquid. It should be emphasized that 1 mol L <sup>−</sup><sup>1</sup> Py13DFOB/PC was selected as the most preferred one to be applied later in the study based on the data in supporting documentation. Therefore, the electrochemical performance of the electrolyte applied to supercapacitors is not only determined by the conductivity, but viscosity is also a factor that cannot be ignored.

# Potential Window Opening: Stability of the Electrolytes

Compared with TEABF4/PC, the supercapacitor using 1 mol L <sup>−</sup><sup>1</sup> Py13DFOB/PC displayed higher energy density and power density under conventional working voltage (2.7 V) while maintaining excellent rate performance and long cycle stability (**Figure S3**). Further, the high voltage characteristics of the new

electrolyte system need to be clarified. Since the electrochemical window cannot directly reflect the working voltage of the full cell due to the difference in the structure of anions and cations, it is necessary to identify the real positive and negative stability limits of the new electrolyte in activated carbon electrodes (Fic et al., 2012). For this purpose, we initially evaluated the maximum operating voltage possible of 1 mol L−<sup>1</sup> Py13DFOB/PC via cyclic voltammetry investigations on three electrode cells with an AC-based working electrode, a largely oversized ACbased counter electrode with the same composition and a Li wire reference electrode. An efficiency threshold value of 98% was chosen, and further efficiency declines were considered to be a series of irreversible reactions involving electrolytes in the cell system. The obtained single cell electrochemical window limits are presented in **Figure 4**. With this condition, the positive and negative potential limits occur at +1.6 V vs. Li and −1.5 V vs. Li, respectively, resulting in a maximum operating voltage of 3.1 V.

We further measured the GCD curve of a symmetric full-cell at an operation voltage of 3.0 V while monitoring the capacitive behavior of positive and negative electrodes. It can be seen from **Figure 5** that the GCD curves exhibits superior capacitive behavior, and the nearly symmetric potential window of positive and negative electrodes shows that the capacitance contributions of DFOB<sup>−</sup> anion and Py<sup>+</sup> <sup>13</sup> cation are almost the same. More importantly, the cutoff potential of the positive and negative electrodes vs. Li is within the safe potential range when the operating voltage of the full cell reaches 3.0 V.

In previous studies, mass balancing is currently the preferred method to make full use of the maximum working voltage of electrolyte, but it will have a negative impact on the specific capacitance of the full cell due to the extra increase in the mass of one of the electrodes (Weingarth et al., 2013; Van Aken et al., 2015; Hu et al., 2016). Besides, due to the change of ion transfer channel and electronic response time of high-quality electrode, resulting in inconsistent polarity of the two electrodes, thus affecting the cycle life of the cell. Excitingly, the Py13DFOB/PC can almost fully exploit its maximum working potential window

without requiring additional mass balance, eliminating the time consumption and technical problems of electrode matching, and harvesting higher energy density.

## Electrochemical Study in EDLCs Configuration

the half-cell using 1 mol L−<sup>1</sup> Py13DFOB/PC.

To evaluate the withstand voltage of EDLCs configuration, the GCD curves of EDLCs based on 1 mol L−<sup>1</sup> Py13DFOB/PC operating at 500 mA g−<sup>1</sup> under several applied voltages are shown in **Figure 6A**. The GCD curves maintain a typical triangular shape, and the variation of potential with time shows an approximate linear relationship as the voltage rising to 3.0 V, which demonstrates excellent electrochemical reversibility and stability of EDLCs. However, the charging curve shifts rather than overlaps, and the linearity and symmetry of the GCD curves gradually deteriorated when the voltage exceeds 3.0 V, indicating detrimental processes may be occurring between electrodes and electrolytes. We obtained the same result from the relationship between specific capacitance (**Figure 6B**), IR drop (**Figure 6C**) with voltage obtained from the GCD test in **Figure 6A**, that is, the abrupt change of the curve all occurs when the voltage exceeds 3.0 V. The IR drop is closely related to the equivalent series resistance of the cell, reflecting the state of the working environment inside the EDLCs. When the voltage is raised from

FIGURE 5 | Potential profiles of the positive and negative electrodes during galvanostatic charge/discharge of symmetric supercapacitor with 1 mol L−<sup>1</sup> Py13DFOB/PC.

3.0 to 3.5 V, the IR drop abruptly increases from 0.074 to 0.152 V, indicating some irreversible changes in the internal environment of the cell. The rate behavior under different voltages (2.7, 3.0, 3.2, 3.5 V) was given by the GCD test with a current density ranging from 100 to 10,000 mA g−<sup>1</sup> . As shown in **Figure 6D**, the discharge capacitance of the cell decreases with increasing current density due to the rise in internal resistance (IR drop) caused by the kinetic limitation at the electrolyte/electrode interface. However, the cell with Py13DFOB/PC still shows the most superior capacitance performance and rate performance at the working voltage of 3.0 V. In general, it is proved that the new electrolyte system can work stably at 3.0 V by electrochemical evaluation of single electrode and assembled supercapacitor.

Next, we evaluated the electrochemical performance of EDLCs containing Py13DFOB/PC at 3.0 V operating voltage. **Figure 7A** showed the CV curves of the cell at different scan rates. The rectangular shape of the CV curves is deteriorated with the increase of the scan rate, but a good rectangular shape is still maintained at the high scan rate, demonstrating the outstanding transfer characteristic of the electrolyte ion. Similarly, GCD curves (**Figure 7B**) exhibit good linearity, symmetry, and negligible instantaneous

FIGURE 6 | (A) GCD curves of the EDLCs with 1 mol L−<sup>1</sup> Py13DFOB/PC at current density of 500 mA g−<sup>1</sup> with different voltage ranges. (B) Specific capacitance and (C) IR drop vs. voltage obtained from (A) charge-discharge curves. (D) The charge/discharge rates performance of EDLCs with 1.0 mol L−<sup>1</sup> Py13DFOB/PC under different voltage ranges.

voltage drop at different current densities, indicating superior double-layer characteristics. The specific capacitance obtained from the discharge time of the GCD curves can reach a high value of 130 F/g at a current density of 100 mA g −1 and still maintain 96 F g−<sup>1</sup> at 5,000 mA g−<sup>1</sup> under a working voltage of 3.0 V, which was about 78% of its initial capacitance.

With the aim to determine whether Py13DFOB/PC can operate steadily for a long time under the working voltage of 3.0 V, the charge-discharge performance of 10,000 cycles were recorded at current density of 1 and 10 A g−<sup>1</sup> . **Figure 7C** conveys a clear message that Py13DFOB/PC presents excellent long cycle performance at different current densities, delivering capacity retention of 92% at 1 A g−<sup>1</sup> and 87% at 10 A g−<sup>1</sup> after 10,000 cycles with nearly 100% coulombic efficiency, confirming the long-term electrochemical stability of Py13DFOB /PC at high working voltage.

Ragone plots (energy density vs. power density) has been widely used to evaluate the overall performance of a supercapacitor device. From the results summarized in **Figure 7D**, Py13DFOB/PC obtained a considerably higher energy density and power density than TEABF4/PC at each current density due to the simultaneous increase in voltage and specific capacitance. It should be noted that the maximum energy density and power density of the supercapacitor based on Py13DFOB /PC can reach 39.06 Wh kg−<sup>1</sup> (100 mA g−<sup>1</sup> ) and 8.03 kW kg−<sup>1</sup> (5,000 mA g−<sup>1</sup> ), respectively, when the voltage goes up to 3.0 V.

#### CONCLUSION

In this work, we successfully synthesized a novel ionic liquid Py13DFOB as electrolyte salt for supercapacitor. The 1 mol L −1 solution formed by the mixture of Py13-DFOB and PC shows the closely transport properties as the commercial electrolyte (TEABF4/PC), which solves the problems of high viscosity and low conductivity of pure ionic liquids. More importantly, it is proved that Py13DFOB/PC can exhibit outstanding capacitance behavior at high operating voltage of 3 V confirmed by several electrochemical testing techniques. Moreover, we found that the nearly symmetric capacity contributions of positive and negative electrodes convey a high specific capacitance value of 130 F g−<sup>1</sup> . The energy density and power density of supercapacitor with Py13DFOB can reach 39.06 Wh kg−<sup>1</sup> (100 mA g−<sup>1</sup> ) and 8.03 kW kg−<sup>1</sup> (5,000 mA g −1 ), respectively. Based on these results, the new electrolyte system is considered to be a promising electrolyte for highvoltage supercapacitor.

#### DATA AVAILABILITY STATEMENT

All datasets generated for this study are included in the manuscript/**Supplementary Files**.

# AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

## REFERENCES


#### FUNDING

This work was supported by the National Science Foundation of China (51603147), Tianjin application foundation and advanced technology research plan project (15ZCZDGX00270, 14RCHZGX00859).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00664/full#supplementary-material


**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Zhang, Zhang, Zhang, Liang and Shi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Hollow Co3O4@MnO<sup>2</sup> Cubic Derived From ZIF-67@Mn-ZIF as Electrode Materials for Supercapacitors

Jiani Xu<sup>2</sup> , Chaoting Xu<sup>2</sup> , Yanhong Zhao<sup>2</sup> , Jianghong Wu1,3 \* and Junqing Hu<sup>1</sup>

*<sup>1</sup> College of Health Science and Environmental Engineering, Shenzhen Technology University, Shenzhen, China, <sup>2</sup> State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai, China, <sup>3</sup> College of Materials Science and Engineering, Changsha University of Science & Technology, Changsha, China*

Hollow Co3O4@MnO<sup>2</sup> cubic nanomaterials are synthesized by ZIF-67@Mn-ZIF sacrificial precursor through a facile thermal treatment. As a kind of supercapacitor electrode material, it demonstrates high performances, such as specific capacitance of 413 F g−<sup>1</sup> at the current density of 0.5 A g−<sup>1</sup> ; as the current densities raised from 0.5 to 10 A g−<sup>1</sup> (20 times increasing), there is still ∼41% retention of its initial capacitance. These satisfactory electrochemical properties should be put down to the hollow and porous structure and the relative higher BET surface area, which supplies more reactive sites for charge and discharge processes.

#### Edited by:

*Min Zeng, Lanzhou Institute of Chemical Physics (CAS), China*

#### Reviewed by:

*Shijie Li, Zhejiang Ocean University, China Hui Yang, Jiangxi University of Science and Technology, China*

> \*Correspondence: *Jianghong Wu wujianghong@sztu.edu.cn*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *27 October 2019* Accepted: *14 November 2019* Published: *13 December 2019*

#### Citation:

*Xu J, Xu C, Zhao Y, Wu J and Hu J (2019) Hollow Co3O4@MnO<sup>2</sup> Cubic Derived From ZIF-67@Mn-ZIF as Electrode Materials for Supercapacitors. Front. Chem. 7:831. doi: 10.3389/fchem.2019.00831* Keywords: metal-organic frameworks, ZIF-67, Mn-ZIF, Co3O4@MnO2, supercapacitors

# INTRODUCTION

As a new and efficient energy storage device, supercapacitors qualified the benefits of high-power density, high security, long service life, and fast short time storage and release (El-Kady et al., 2016; Shao et al., 2018; Xu et al., 2018). As a result, supercapacitors attracted wide attention in the application on portable consumer electrical products and electric automobiles, and so on (Qu et al., 2016; Li et al., 2018). The performance of the used electrode materials is the main factor affecting the performance of supercapacitors. Currently, the most studied materials are carbon-based materials (Zhang and Zhao, 2009; Zhang et al., 2017), transition metal oxides (TMOs) (Liu et al., 2011; Li et al., 2014; Yu and Lou, 2018; Xu et al., 2019), and conductive polymer materials (Snook et al., 2011; Du et al., 2017). In recent years, metal-organic frameworks (MOFs) are developed as a new type of porous materials ascribed to their great specific surface area, porosity and regulatory pores, functional and special optical and electrical properties (Yue et al., 2015; Salunkhe et al., 2017). So they have great potential in the high-performance supercapacitor after thermal treatment as TMOs' sacrificial precursor.

Up to the present, numerous TMO nanomaterials have been synthesized as supercapacitor electrodes from many kinds of MOF precursors. For instance, high surface area Co3O<sup>4</sup> nanoparticles have been obtained from the pyrolysis of ZIF-67 with an appreciable 190 F g−<sup>1</sup> specific capacitance value at 5 A g−<sup>1</sup> (Saraf et al., 2019), NiO architecture with porous structure was constructed by thermal treatment Ni-MOF under the air flow and demonstrated 324 F g−<sup>1</sup> at 1 A g−<sup>1</sup> (Han et al., 2017), and porous hollow α-Fe2O<sup>3</sup> microboxes synthesized by using MOF as precursor and self-template can reach 380 F g−<sup>1</sup> at 0.1 A g−<sup>1</sup> as supercapacitor electrode (Yu et al., 2019). Except for these single metal oxides, some mixed metal oxides, and metal oxide composites can also be obtained by MOF precursors. Chen and coworkers have fabricated porous small size ZnCo2O<sup>4</sup> nanoparticles (<20 nm) from a mixed zinc and cobalt-MOF, which exhibited an unexpected specific capacitance of 451 F g−<sup>1</sup> at 0.5 mV s−<sup>1</sup> (Chen et al., 2015). Hierarchical NiO/ZnO double-shell hollow spheres are obtained by Li and coworkers through calcining the bimetallic organic frameworks, which delivered 497 F g−<sup>1</sup> at current density of 1.3 A g −1 (Li et al., 2016). Xu and coworkers developed a Co3O4/ZnO nano-heterostructure via a solid-solid conversion process, the synthesized core-shell MOFs@MOFs were used as a template with cobalt and zinc as metal sources, which demonstrated 415 F g −1 specific capacitance value at 0.5 A g−<sup>1</sup> (Xu et al., 2016). The mixed metal oxides and the metal oxide composites as electrodes exhibit superior electrochemical performance compared with single ones. Despite these achievements, there are still large spaces to explore other metal oxide composites based on MOF precursors.

Herein, we have prepared single ZIF-67 nanocrystals first, combined it with Mn-ZIF to form ZIF-67@Mn-ZIF composite, and finally obtained Co3O4@MnO<sup>2</sup> electrode material by thermal treatment. After evaluating the electrochemical performance of Co3O4@MnO<sup>2</sup> electrode, we found that it exhibited excellent electrochemical properties. When the current density is 0.5 A g−<sup>1</sup> , the specific capacitance could achieve 413 F g−<sup>1</sup> , with 20 times current density increasing, it kept 41% retention of initial capacitance and good long-term cycling stability, which is a very promising electrode for use in a supercapacitor.

## EXPERIMENTAL

#### Preparation of ZIF-67

First, 1.455g Co(NO3)2·6H2O and 1.642g 2-methylimidazole were separately dissolved in 40 ml methanol. Second, the two different solutions were mixed and vigorously stirred for 60 s and reacted for 24 h to complete reaction at room temperature after 24 h. Third, the purple precipitates in the bottom were collected by centrifugation with ethanol as washing solution for several times. The collected purple precipitates were dried at 80◦C overnight in a vacuum drying chamber.

#### Preparation of ZIF-67@Mn-ZIF

First, 0.25 g Mn(NO3)2·6H2O was dissolved in 50 ml ethanol. ZIF-67 obtained in the first step was well-dispersed in the above solution. Then, the mixture was transferred into a beaker flask after 20 min of continuous stirring, and the reaction temperature was 50◦C and kept for 3 h in an oil bath.

## Thermal Treatment of ZIF-67@Mn-ZIF Crystals

The obtained ZIF-67@Mn-ZIF crystals could be converted to Co3O4@MnO<sup>2</sup> nanomaterials through a thermal treatment in a tube furnace with air flow at 300◦C for 0.5 h; the heating rate was controlled at 0.5◦C·min−<sup>1</sup> . As a contrast experiment, the single precursors (ZIF-67) were calcined under the same thermal conditions, and the final product is Co3O<sup>4</sup> nanomaterial.

## Material Characterizations

X-ray diffraction (XRD) patterns were measured by using monochromator Cu Kα radiation at a scanning rate of 2◦•min−<sup>1</sup> (PA-Nalytical X′Pert PRO). Binding energies were detected by the X-ray photoelectron spectroscopy (XPS; ESCALab250). The morphologies were obtained by scanning electron microscope (SEM) (Hitachi, SU-8000). The more detailed structures were investigated by transmission electron microscope (TEM) (JEOL, JEM-2100F), and the elements were detected by its equipped energy dispersive X-ray spectrometer (EDS). The BET surface area and pore size distribution are tested on Accelerated Surface Area & Porosimetry System (ASAP 2020, Micromeritics). XS analytical balance (Mettler Toledo; δ = 0.01 mg) is used to weigh the mass of the electrode materials.

#### Electrochemical Characterizations

The electrochemical performances of the final products were accomplished by the AUTOLAB PGSTAT302N electrochemical workstation in a standard three-electrode test cell at ∼25◦C with 1.0 M LiOH solution as electrolyte. The Ag/AgCl (3M KCl) electrode and platinum (Pt) plate (2.5 cm × 2.5 cm × 0.2 mm) directly served as the reference electrode and counter electrode, respectively. The fabricating processes of working electrode were as follows: Co3O4@MnO<sup>2</sup> materials (active electrode material, 80%) derived from ZIF-67@Mn-ZIF crystals were mixed with acetylene black (5%) and polyvinylidene difluoride (15%), which was mixed with appropriate volume N-methyl pyrrolidone solvent. The mixture was treated by ultrasonication to form a homogeneous slurry and dropped onto the graphite substrate current collector, the covered surface area is ∼1 × 1 cm<sup>2</sup> , and then dried under vacuum condition at 120◦C for 4 h to form the electrodes. For comparison, the Co3O<sup>4</sup> materials prepared from single ZIF-67 crystals were also fabricated into electrode with the same processes.

The electrochemical performances of the fabricated electrodes were evaluated from the galvanostatic chargedischarge (GCD) and cyclic voltammetry (CV) measurements. The equation of C = [(I × ∆t)/(m × ∆V)] is applied to calculate the specific capacitance values of Co3O4@MnO<sup>2</sup> and Co3O<sup>4</sup> electrodes, where the I (A), ∆t(s), ∆V (V), and m (g) represent the discharge current, the discharge time, the potential window, the mass of active materials in the electrodes, respectively.

#### RESULTS AND DISCUSSIONS

The synthesized products were analyzed by X-ray diffraction (XRD) first. The result is shown in **Figure 1a**. As can be seen from the obtained pattern, there are some strong diffraction peaks that appeared in 2θ = 7.3◦ , 10.4◦ , 12.8◦ , 14.8◦ , 16.5◦ , 18.1◦ , which can be confirmed with the sample ZIF-67 and highly consistent with reported literature (Qin et al., 2017). **Figures 1b,c** are the low to high magnification SEM images. The particles' morphology is uniform rhombic dodecahedral nanocrystals which were clearly synthesized ZIF-67 sample.

monodispersed, with a diameter of about 300–500 nm. **Figure 1d** shows a single ZIF-67 nanocrystal with dodecahedron and solid construction. After thermal treatment, the structure collapsed (**Figure S1**).

The calcined ZIF-67@Mn-ZIF products were detected by XRD, and the result is shown in **Figure 2a**. The main diffraction peaks consisted of cubic phase Co3O<sup>4</sup> (JCPDS card No. 074- 2120), which is obtained by calcinating the ZIF-67 nanocrystal (**Figures S2, S3**). In addition, there are some other small peaks that also appeared in the pattern (marked with blue star), which could be MnO<sup>2</sup> formed by Mn-ZIF (the exact components were detected by XPS, which is detailed later). **Figure 2b** is a low-resolution SEM image of the obtained Co3O4@MnO<sup>2</sup> products, which indicated that the products can be synthesized in large scale. In the enlarged SEM image of **Figure 2c**, the diameter of obtained Co3O4@MnO<sup>2</sup> products increased to about 800 nm; interestingly, the obtained Co3O4@MnO<sup>2</sup> products are with hollow structure (**Figure 2d**), and the corresponding EDS result in **Figure 2e** is consistent with our designed concept. Co, Mn, and O elements are from Co3O4@MnO<sup>2</sup> products, the existence of Cu and C signals is because the TEM grid is made of Cu substrate and carbon membrane, while the peak of Si could be an impurity that brings in the sample preparation process. Inset shows the HRTEM image of the Co3O4@MnO2, the d-spacing of 0.24 nm corresponding to the (311) lattice plane of the Co3O<sup>4</sup> crystal, and the d-spacing of 0.22 nm corresponding to the (200) lattice plane of the MnO<sup>2</sup> crystal (JCPDS No. 12-0716).

The obtained calcinated products were further detected by XPS to confirm the metal oxidation states and the chemical compositions. **Figure 3A** is the survey spectrum of the products, which shows the core levels of Co 2p, Mn 2p, and O 1s, respectively. To get clearer information, the high-resolution XPS spectra analysis was carried out. The Co 2p's high-resolution XPS spectrum is shown in **Figure 3B**. The main two peaks

centered at 780.3 and 796.2 eV can be appointed to the binding energies of 2p3/2 and 2p1/2 of Co(II), whereas the other two lower peaks centered at 786.1 and 802.5 eV can be appointed to the binding energies of 2p3/2 and 2p1/2 of Co(III). These results imply the Co3O<sup>4</sup> phase in our sample and agreement with the XRD result (Yan et al., 2012; Li et al., 2013). **Figure 3C** is the high-resolution XPS spectrum extracted from Mn 2p. The main two peaks are centered at 641.1 and 652.7 eV; therefore, the spin-orbital splitting calculated is 11.6 eV. These results well refer to the electronic orbits of Mn 2p3/2 and 2p1/2, pointing to Mn(IV) state of the products (Sui et al., 2009). As can be seen from the high-resolution spectrum of O 1s in **Figure 3D**, there are two distinct components, except for the binding energy of 531.2 eV assigned to the oxygen atoms in the hydroxyl groups, the strong peak of 529.6 eV should belong to the oxygen atoms in the chemical compositions of Co3O<sup>4</sup> and MnO<sup>2</sup> (Wei et al., 2008; Xia et al., 2010). These results further proved that the chemical component of as-fabricated products is Co3O4@MnO2.

HRTEM image of Co3O4@MnO2.

A typical IV type adsorption behavior was observed in the prepared Co3O<sup>4</sup> and Co3O4@MnO<sup>2</sup> products by the N<sup>2</sup>

FIGURE 3 | X-ray photoelectron spectroscopy (XPS) spectra of the products after calcinating the ZIF-67@Mn-ZIF. (A) Survey XPS spectrum and (B–D) high-resolution XPS spectra of Co 2p, Mn 2p, and O 1s.

adsorption-desorption isotherms (**Figures 4A,B**), which exhibit a mesoporous structure with slit type pores. The BET surface areas for the Co3O<sup>4</sup> and Co3O4@MnO<sup>2</sup> are 72.214 and 148.407 m<sup>2</sup> g −1 , a high BET surface area might be beneficial for the electrons and ions' storage and shuttle in the electrode because it provides more active sites, hence could lead to enhanced electrochemical capacity (Jiang et al., 2012). From the corresponding pore size distributions of the inset image, it can be found that the pore sizes are concentrated in 3–10 nm for Co3O<sup>4</sup> sample and 3–6 nm for Co3O4@MnO<sup>2</sup> sample. The porous structure is facilitating the electrolyte ion diffusion and transference in the course of charge and discharge processes.

The electrochemical properties of the Co3O4@MnO<sup>2</sup> electrode materials were evaluated, and the results are summarized and shown in **Figure 5**. Enclosed loops in **Figure 5A** show the electrode's CV performance at increasing scan rates from 1 to 100 mV s−<sup>1</sup> , unlike the pseudocapacitance behavior of single Co3O<sup>4</sup> nanocrystals (**Figure S4**). The shapes of the CV curves of Co3O4@MnO<sup>2</sup> indicate a typical electrical double layer capacitance (EDLC) behavior, and it retains well as the scan rate

of the specific capacitance against different current densities. (D) Long-term

cycling stability.

upscales to 20 mV s−<sup>1</sup> , demonstrating its good rate capability (Wu et al., 2015). The EDLC behavior of Co3O4@MnO<sup>2</sup> ascribes to the MnO<sup>2</sup> outer layer (Li et al., 2012). The slight shape deformation was observed when the scan rate achieves 50 and 100 mV s−<sup>1</sup> ; this could be ascribed to the polarization phenomenon at high scan rate (Salanne et al., 2016). The GCD properties were evaluated at current densities from 0.5 A g−<sup>1</sup> and extended to 10 A g−<sup>1</sup> in the voltage range from 0 to 0.6 V vs. Ag/AgCl (3M KCl). In **Figure 5B**, it is clear to observe a series of good symmetric triangle shape GCD curves, revealing its good EDLC behavior; this result is consistent with CV performance. Under a series of current densities, that is, 0.5, 1, 2, 4, 6, 8, and 10 A g−<sup>1</sup> , the specific capacitances were calculated to be 413, 370, 324, 273, 233, 200, and 168 F g−<sup>1</sup> , respectively. On the contrary, single Co3O<sup>4</sup> nanocrystal only delivers 187, 155, 108, 76, 57, and 45 F g−<sup>1</sup> at 1, 2, 4, 6, 8, and 10 A g−<sup>1</sup> , respectively (**Figure S5**). Obviously, the Co3O4@MnO<sup>2</sup> electrode presents better capacitance values than single Co3O<sup>4</sup> electrode, the reason could be owing to the multicomponent and higher BET surface of Co3O4@MnO<sup>2</sup> electrode that endows the more charge storage (Jiang et al., 2012). Under the extending current densities from 0.5 to 10 A g−<sup>1</sup> , the specific capacitance decreased from 413 to 168 F g−<sup>1</sup> , retaining ∼41% of its initial capacitance (shown in **Figure 5C**). While for single Co3O<sup>4</sup> nanocrystal, the rate capability is only 25% from 1 to 10 A g−<sup>1</sup> (187 vs. 45 F g−<sup>1</sup> , **Figure S6**). After 2,000 times cycles of CV test at 20 mV s−<sup>1</sup> , the capacitance retention remains at 110% (shown in **Figure 5D**), while only 80% of single Co3O<sup>4</sup> nanocrystal (**Figure S7**), indicating a good stability of Co3O4@MnO<sup>2</sup> electrode. It is clear to conclude that the performance of Co3O4@MnO<sup>2</sup> in connection with capacitance retention and cycling ability is much improved compared with single Co3O4.

# CONCLUSIONS

In conclusion, hollow Co3O4@MnO<sup>2</sup> cubic nanomaterials were synthesized by sacrificing the ZIF-67@Mn-ZIF precursor through an uncomplicated controlled thermal treatment. The porous structure and high BET surface area endow its excellent properties as supercapacitor electrode, it presented a high specific capacitance of 413 F g−<sup>1</sup> (0.5 A g −1 ) and showed a rate capability of 41% at the current density enhanced to 20 times with excellent stability, giving the impression that this hollow cubic nanomaterial possesses considerable potential as a supercapacitor electrode material.

# DATA AVAILABILITY STATEMENT

The XRD datasets generated for this study can be found in the repository of ICDD, with the accession numbers of 74-2120 and 12-0716 in JCPDS card.

# REFERENCES


# AUTHOR CONTRIBUTIONS

JW and JH conceived and designed the experiments. JX, YZ, and CX performed the experiments and analyzed the data. All authors revised and checked the draft.

# FUNDING

This research was supported by the National Natural Science Foundation of China (51701022, 51972055), the Shenzhen Science and Technology Research Project (Grant No. JCYJ20180508152903208), and the Shenzhen Pengcheng Scholar Program.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2019.00831/full#supplementary-material


Zhang, W., Xu, C., Ma, C., Li, G., Wang, Y., Zhang, K., et al. (2017). Nitrogensuperdoped 3D graphene networks for high-performance supercapacitors. Adv. Mater. 29:1701677. doi: 10.1002/adma.201701677

**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Xu, Xu, Zhao, Wu and Hu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Nitrogen and Phosphorus Co-doped Porous Carbon for High-Performance Supercapacitors

Jiaming Zhou, Shewen Ye, Qinqin Zeng, Hui Yang, Jiahao Chen, Ziting Guo, Honghui Jiang\* and Karthikeyan Rajan\*

*School of Materials Science and Engineering, Jiangxi University of Science and Technology, Ganzhou, China*

As one of the most promising fast energy storage devices, supercapacitor has been attracting intense attention for many emerging applications. However, how to enhance the electrochemical performance of electrode materials is still the main issue among various researches. In this paper, hierarchical porous carbons derived from *Eleocharis dulcis* has been prepared by chemical activation process with the aid of KOH at elevated temperature. Results show that the N, P co-doped porous carbon exhibits excellent electrochemical performance, it owns a specific capacitance of 340.2 F/g at 1 A/g, and obtains outstanding cycling stability of 96.9% of capacitance retention at 10 A/g after 5,000 cycles in a three-electrode system. Moreover, in the two-electrode system, the product still maintains a high specific capacitance of 227.2 F/g at 1 A/g, and achieves good electrochemical cycle stability (94.2% of capacitance retention at 10 A/g after 10,000 cycles); besides, its power/energy density are 3694.084 and 26.289 Wh/kg, respectively. Therefore, the combination of facile synthesis strategy and excellent electrochemical performance makes *Eleocharis dulcis*-based porous carbon as a promising electrode material for supercapacitor.

#### Edited by:

*Yuanlong Shao, King Abdullah University of Science and Technology, Saudi Arabia*

#### Reviewed by:

*Bingjie Zhang, Chapman University, United States Shijie Li, Zhejiang Ocean University, China*

#### \*Correspondence:

*Honghui Jiang jhonghui@163.com Karthikeyan Rajan karthikeyan148@gmail.com*

#### Specialty section:

*This article was submitted to Electrochemistry, a section of the journal Frontiers in Chemistry*

Received: *20 December 2019* Accepted: *04 February 2020* Published: *20 February 2020*

#### Citation:

*Zhou J, Ye S, Zeng Q, Yang H, Chen J, Guo Z, Jiang H and Rajan K (2020) Nitrogen and Phosphorus Co-doped Porous Carbon for High-Performance Supercapacitors. Front. Chem. 8:105. doi: 10.3389/fchem.2020.00105* Keywords: biomass, porous carbon, supercapacitor, Eleocharis dulcis, N/P co-doped

# INTRODUCTION

Rapid development of global economy, the depletion of chemical fuels and the ever-worsening environment are intensified with the continuous growth of the population, which increases the demand for clean sustainable energy. Thence it requires the development of efficient and clean energy storage devices (Wang et al., 2016; Liu et al., 2017b; Yang et al., 2019). Among them, the traditional Lithium-ion batteries will generate quantum and form lithium dendrites under high-power operation, the supercapacitors have distinctive properties such as excellent power density, rapid charging and discharging speed and superior cycle stability, is considered the best substitute for lithium-ion batteries (Zhao et al., 2013; Shao et al., 2018; Li et al., 2019). Although supercapacitors exhibit excellent properties, low specific capacity and energy density (typically <10 Wh/kg) toward large scale commercial devices are still major constraints (Winter and Brodd, 2004; Liu et al., 2017b).

Electrode materials are the important constituent which affect the properties of supercapacitor. Traditionally, different allotropes of carbon materials are used as an electrode in energy storage applications. Among them, sustainable biomass derived carbons are individual class of materials, with the advantage of low-cost, abundant and sustainable in nature, excellent electrical conductivity

**114**

and specific surface area (SSA) (Pandolfo and Hollenkamp, 2006; Jiang et al., 2013; Titirici et al., 2015; Gong et al., 2017; He et al., 2018). Number of researchers have derived carbon from different biomass sources such as Perilla frutescens (Liu et al., 2017a), Rice straw (Liu et al., 2018), Peanut shells (Xiao et al., 2018), Buckwheat flour (Huang et al., 2019), Peach gum (Lin et al., 2019), and Bamboo (Zhang et al., 2018). By using KOH through chemical activation Cheng et al. (2016) have prepared flexible carbon fiber aerogel from natural cotton and achieved specific capacitance of 283 F/g at 1 A/g. Similarly, the graded porous carbon material derived from walnut shells resulted the capacitance of 462 F/g at 1 A/g (Wang et al., 2019). Besides, the specific surface area of carbon materials derived from seaweed microspheres show as high as 1337.9 m<sup>2</sup> /g, the capacitance led to 309 F/g at 1 A/g, with the capacitance retention rate of 92% at 20 A/g with 10,000 cycles (Zhu et al., 2018). Therefore, it is necessary to reveal the relationship between different biomass sources and their relation to specific surface area and specific capacitance.

Due to their limited number of active sites on microporous carbon, it is important to investigate the improvement of electrical conductivity and electronegativity properties. In this work we have fabricated the nitrogen and phosphorus co-doped microporous derived carbon for supercapacitor application. The reason why biomass derived carbon materials can show excellent capacitance performance, is closely related to the incorporation of trace elements such as nitrogen and phosphorus into carbon materials in the carbonization process (Shen and Fan, 2013; Chen et al., 2014; Zhao et al., 2017). The presence of nitrogen in carbon materials is expected to improve the electron conductivity of the materials (Chen et al., 2013). Phosphorus contained materials could help to improve the electronegativity of carbon through the combination of lone pair electron nitrogen and carbon, thereby enhance the hydrophilicity of carbon materials. The nitrogencontaining groups on the surface are alkaline, which is conducive to ion adsorption (Shen and Fan, 2013). Nitrogen and phosphorus are belong to the same group in periodic table, however phosphorous possess higher electron-donor capacity, which is useful to achieve stable capacitive property (Zhao et al., 2017). Therefore, we believe that presence of nitrogen and phosphorus in carbon would greatly improve the capacitive performance.

China cultivates the largest quantities of Eleocharis dulcis (ED) in the world with the annual output of up to 1.75 million tons. Eleocharis dulcis is rich in trace elements such as nitrogen and phosphorous, and the phosphorus element is higher in root vegetables (Bao et al., 2018). Therefore, deriving carbon from Eleocharis dulcis without affecting the existing N and P elements could help to enhance the specific capacitance (Panja et al., 2015). In this work, we have derived the microporous activated carbon material from ED by chemical activation process and carbonized. To the results of N<sup>2</sup> adsorption-desorption isothermal analysis describes the highest, specific surface area of 2,454 m<sup>2</sup> /g with the specific capacitance of 340.2 F/g at 1 A/g.

# EXPERIMENTAL

# Materials Synthesis

In general, ED were freeze-dried and grinded, after screening through a 70-mesh sieve, the obtained powders were mixed with certain amount of KOH solution, then dried at 80◦C for 12 h. The as-prepared precursors were transferred into a tube furnace, after that, the powders were treated at an elevated temperature (800◦C for 1 h) under N<sup>2</sup> atmosphere. The obtained carbon material was washed by 2 M HNO<sup>3</sup> to remove the impurities, subsequently, the final products were washed with deionized water several times, then dried at 60◦C for 24 h. The samples treated with different KOH/ED mass ratio (1:1, 2:1, 3:1) were referred to referred to as NPC-1, NPC-2, and NPC-3, respectively.

# Materials Characterization

Scanning electron microscope (SEM, ZEISS Sigma) and transmission electron microscope (TEM, JEOL, JEM-2010, Japan) are used to analyze the morphology and microstructure of the samples. The phase and results of the samples are analyzed by X-ray diffraction (XRD, Empyrean). Raman spectroscopy (excitation beam wavelength 532 nm) is used to analyze the graphitization degree of materials. The nitrogen adsorption and desorption isotherms and the specific surface area, pore diameter distribution and pore volume of the samples are measured by N<sup>2</sup> adsorption-desorption experiment (Micromeritics, ASAP 2010M, USA).

## Electrochemical Measurements

All electrochemical measurements were performed at CHI760E electrochemical workstation (Chen Hua Shanghai). The electrochemical test used a standard three-electrode system, in which Hg/HgO is used as the reference electrode, Pt electrode was used as the counter electrode, the prepared NPCs were used as the working electrodes, and 6 M KOH aqueous solution as the electrolyte. The working electrodes were prepared according to the following procedure: 80 wt.% NPCs, 10 wt.% acetylene black, and 10 wt.% PVDF (binder) were thoroughly mixed in the N-methyl-2-pyrrolidone (NMP) solvent to obtain a uniform semi-fluid slurry; the prepared slurry was coated onto carbon cloth, and then dried in a vacuum oven at 60◦C for 12 h. The coating mass of active material in each working electrode is about 2 mg/cm<sup>2</sup> . The cyclic voltammetry and galvanostatic charge/discharge (GCD) curve had been performed at various scanning rate/current density, and the corresponding electrochemical impedance spectroscopy (EIS) was tested at an open circuit voltage (frequency range: 0.01–100 kHz, amplitude: 5 mV). The calculation details were provided in **Supplementary Material**.

# RESULTS AND DISCUSSION

The scanning electron microscopy (SEM) analysis was used to analyze the morphological significance of synthesized carbon materials. **Figure 1** shows the low and high magnification images of NPC-3. The microstructure images reveal the presence

of large hierarchical porous on the derived carbon The diameter of these large pore structures varies from few to several micrometers, Compared with the NPC-1 (**Figures S2A,B**) and NPC-2 (**Figures S2C,D**), the NPC-3 has more uniform hole distribution with macropores structure. The structure of NPC-3 was further studied by TEM, according to the TEM results (**Figures 1C–E**), there are many randomly distributed mesopores, which connect with the macropores and form a hierarchical porous structure. It is well-stated that KOH activation leads to the formation of abundant micropores due to its corrosive nature (Bleda-Martínez et al., 2005; Guan et al., 2009), the reaction mechanism is described as follows:

$$\text{KOH} + \text{C} \leftrightarrow \text{K} + \text{CO}\_2 + \text{H}\_2\text{O} \tag{1}$$

The electrochemical performance of carbon-based materials are greatly depends upon its solid/electrolyte interface, and its porous structure (Salanne et al., 2016). The contribution of micropores and mesopores to the specific capacitance were not discussed in detail. Previous reports indicated that neither micropores nor mesopores influenced the energy/power density (Kim et al., 2013). Besides, Lei et al. found that hierarchical micropores with wider size distribution led to high energy storage, which provided a fast transportation pathway for ions (Lei et al., 2011). The nitrogen adsorption-desorption analysis was conducted for the synthesized carbon and the results are shown in **Table 1**. **Figures 2A,B** represents the N<sup>2</sup> adsorptiondesorption isothermal and pore size distributions of NPC-1,

TABLE 1 | Adsorption parameters of different samples calculated from N<sup>2</sup> adsorption isotherms.


NPC-2, and NPC-3. Results reveal that all samples are composed of type I and type IV isotherms. The sharp adsorption of N<sup>2</sup> at low relative pressure (0∼0.1) indicates the presence of micropores in the porous carbon structure. The hysteresis loops at higher relative pressure represents the presence of mesopores structure in NPC-3; the curve shape near high relative pressure region indicates there are small amount of macropores structure (Lv et al., 2012). The pore size distribution of the samples has been tested. As can be seen from **Table 1**, with the increase of KOH, the BET surface area (SBET) and total pore volumes (Vtotal) of samples increased from 1,063 to 2,454 m<sup>2</sup> /g and from 0.352 to 1.345 cm<sup>3</sup> /g, respectively. The electrode material with high specific surface area can provide abundant electrochemical active sites and enhance the effective charge storage area between electrode and electrolyte, thus improving the performance of supercapacitor (Xu et al., 2018).

**Figure 3A** shows the X-ray diffraction (XRD) pattern and Raman spectra (**Figure 3B**) of synthesized carbon samples

derived from ED. The broad characteristic peaks at 25.9 and 43.1◦C are corresponding to the (002) and (100) crystal planes of carbon materials, respectively (Wan et al., 2015). Increasing the mass ratio of KOH to ED leads to the broadening of diffraction peak (002) and (100). This indicates that the KOH can significantly affect the orderings of the crystal planes. Raman spectroscopy of **Figure 3B** describes the D-band (at 1,333 cm−<sup>1</sup> ) G-band (at 1,589 cm−<sup>1</sup> ). The D-band is attributed to disordered nature of graphitic planes and G-band ascribed to ordered planes due to sp<sup>2</sup> hybrid carbon stretching vibration. More importantly, ID/I<sup>G</sup> reflects the degree of graphitization of the material (Ferrari et al., 2006; Zhou et al., 2014), and the ID/I<sup>G</sup> of NPC-1, NPC-2, and NPC-3 are 1.019, 1.010, and 1.002, respectively. The result indicates that the higher KOH ratio inhibits the graphitization of the material, raises the disorder of the microstructure of the material. This is consistent with the XRD results.

The surface chemical properties of NPCs are investigated by X-ray photoelectron spectroscopy (XPS) measurements. The characteristic peak for C1s (∼284.60 eV), N1s (∼400.45 eV), O1s (∼532.64 eV), and P2p (∼134.40 eV) were observed in the spectrum (**Figure 4A**). The C1s spectra (**Figure 4B**) of the NPC-3 display three distinct characteristic peaks at 284.70, 286.03, and 288.82 eV, they are corresponding to different carbon functional groups of C-C or C=C, O-C-O, and O-C=O, respectively (Li et al., 2016). N1s spectra contains three peaks located at 399.67, 400.44, and 401.68 eV, corresponding to pyridinic-N, pyrrolic-N, and quaternary-N (**Figure 4C**). Pyridinic-N and pyrrolic-N species have positive charge, they can enhance the electron transfer at high current density, while quaternary-nitrogen can increase the conductivity of materials (Yang et al., 2019). In addition, **Figure 4D** shows a P2p spectrum with a peak value of 134.53 eV, representing P-O functional group. According to **Table 2**, the doping amount of P is about 0.18∼0.25%. P has a higher electron delivery capacity than N, which can significantly improve the charge storage and transport capacity of carbon materials. Therefore, N and P doping are beneficial to the electrochemical performance of supercapacitors. As we all know, the N content of NPCs decreases with the increase of KOH mass, while it won't affect the P content. Interested, O at% is negatively correlated with P at%, this should be attributed to the part of P atoms, which are directly bonded to C atoms, and do not bind the edges of the carbon lattice by P-O.

The electrochemical characterizations were carried out for the synthesized carbon materials as shown in **Figure 5**. All the samples are tested under 6 M KOH electrolyte through the threeelectrode system. As shown in **Figure 5A**, all the CV curves of NPCs represent the quasi rectangular shape at 10 mV/s sweeping potential. This indicates that the charge can be reassembled

TABLE 2 | The contents of C, N, P and O in NPCs from XPS analysis.


quickly when the voltage is turned, it reveals the material has good rate capability and cycle performance (Xu et al., 2019). It can be seen that the CV curve of NPC-3 has the largest area which depicts the highest specific capacitance of, which represents 340.2 F/g at 1 A/g. **Figure 5C** shows the galvanostatic charge and discharge (GCD) curves of NPC-3 under different current densities. All curves show symmetrical triangular shape without any voltage drop, indicating that the material has good rate capability and cycle performance. However, the GCD curve is not strictly symmetric due to pseudocapacitive effect caused by the presence of N and P. **Figure 5B** plotted against capacitance with respect to different current densities. All the NPCs signify the decrease in its specific capacitance with respect to the increase in current density. Besides, the cycling stability of the NPC-3 has been resulted to 96.9% of the initial specific capacitance after 5,000 cycles at 10 A/g, this could be due to the 3D structure of carbon materials and the contribution of phosphorus and nitrogen functional groups, NPC-3 has good cycle stability.

In summary we summarize the reasons for the superior electrochemical performance of NPC-3: (1) The co-doped of N and P atoms produces more active sites, which leads to the increase of conductivity and electronegativity of porous carbon, increases the hydrophilicity of the material, and then increases the effective specific surface area, leading to the common effects of capacitance and pseudo capacitance (Chen et al., 2015). (2) The micropores in the hierarchical porous carbon can be used to store charge, and mesoporous and macropores materials can accelerate the migration rate of ions in electrolytes, improve multiplier performance and circle performance of the NPCs.

The NPCs have been assembled into symmetric supercapacitor to investigate their electrochemical performance. **Figure 6A** shows the NPC-3 cyclic voltammetry curves, which is approximately rectangular, indicating that the material has good capacitance performance. When the scanning speed reaches 100 mV/s, the curve slightly changes but still maintains the shape of rectangle, indicating that NPC-3 has good capacitance retention. Because of the NPCs' unique hierarchical porous structure and high specific surface area, the GCD curve (**Figure 6B**) presents the shape of a nearly symmetrical triangle, and its current density ranges from 0.5 to 50 A/g, representing highly reversible charge-discharge behavior. The capacitance reaches 227.2 F/g at 1 A/g. In addition, there is still a specific capacitance of 170.0 F/g at 10 A/g with high capacitance retention (73.9%) (**Figure 6C**). As recorded in **Figure 6D**, the capacitance of NPC-3 has been reduced after 10,000 cycles at 10 A/g, but it also can maintain 94.2% of the initial specific capacitance. It is well-known that power density/energy density is an important parameter to evaluate the quality of supercapacitors (Tang et al., 2015). Due to the incorporation of N and P elements and the design of graded porous structure, the power/energy density of NPCs is up to 3694.084 W/kg and 26.289 Wh/kg, respectively.

# CONCLUSION

In this work, ED-derived porous carbon has been prepared through carbonization and activation at elevated temperature. The introducing of trace elements N and P influences the electron conductivity, ions adsorption and capacitive stability of the matrix, and endows the products with excellent electrochemical performance. Besides, the as-prepared samples show high specific surface area because of the abundant hierarchical porous structure. In three-electrode testing system, NPC-3 exhibits a high specific capacitance (340 F/g at 1 A/g) and excellent rate capacity (190 F/g at 100 A/g). Furthermore, in two-electrode configuration, the corresponding materials also maintains superb electrochemical performance (227.2 F/g at 1 A/g and 170.0 F/g at 10 A/g). Its high energy/powder density (26.289 Wh/kg at a power density of 3694.084 W/kg) and good cycling stability ensure NPC holds great application promise for highperformance supercapacitor.

#### DATA AVAILABILITY STATEMENT

All datasets generated for this study are included in the article/**Supplementary Material**.

#### REFERENCES


#### AUTHOR CONTRIBUTIONS

JZ and SY were responsible for literature searching and drafting. All authors contributed equally to the final writing of the paper.

#### FUNDING

This research was supported by National Natural Science Foundation of China (51702139) and Youth Science Foundation (20151BAB216007, GJJ150637, and 20161BAB216122).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2020.00105/full#supplementary-material


for selectively functionalized nanoporous hybrid carbon. J. Am. Chem. Soc. 137, 1572–1580. doi: 10.1021/ja511539a


**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2020 Zhou, Ye, Zeng, Yang, Chen, Guo, Jiang and Rajan. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.