Supercapacitors Based on Flexible Substrates: An Overview

Authors

  • Dr. Deepak P. Dubal,

    1. School of Materials Science and Engineering, Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    2. Technische Universität Chemnitz, Institut für Chemie, AG Elektrochemie, 09107 Chemnitz (Germany)
    Search for more papers by this author
    • These authors contributed equally to this work

  • Jong Guk Kim,

    1. School of Materials Science and Engineering, Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    2. Research Institute for Solar and Sustainable Energies (RISE), Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    Search for more papers by this author
    • These authors contributed equally to this work

  • Youngmin Kim,

    1. School of Materials Science and Engineering, Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    2. Research Institute for Solar and Sustainable Energies (RISE), Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    Search for more papers by this author
  • Prof. Rudolf Holze,

    1. Technische Universität Chemnitz, Institut für Chemie, AG Elektrochemie, 09107 Chemnitz (Germany)
    Search for more papers by this author
  • Chandrakant D. Lokhande,

    1. Thin Film Physics Laboratory, Department of Physics, Shivaji University, Kolhapur, 416004 (M.S) (India)
    Search for more papers by this author
  • Prof. Won Bae Kim

    Corresponding author
    1. School of Materials Science and Engineering, Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    2. Research Institute for Solar and Sustainable Energies (RISE), Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)
    • School of Materials Science and Engineering, Gwangju Institute of Science and Technology (GIST), 261 Cheomdan-gwagiro, Buk-gu, Gwangju 500-712 (South Korea)===

    Search for more papers by this author

Abstract

Recently our modern society is demanding flexible, low-cost, and lightweight electrochemical energy storage systems, which are very important in variety of applications ranging from portable consumer electronics to large industrial-scale power and energy management. Among different energy storage systems, flexible supercapacitors have been considered as one of the most promising candidates due to their significant merits such as high power density along with the unique properties of being flexible, lightweight, shape versatile, and eco-friendly in comparison to other energy storage systems. In this regard, this review article describes the principles of supercapacitors and the recent research progress on flexible supercapacitor electrodes, for which metal substrates, carbon-based paper, conventional paper, textiles, sponges, and cables are used as substrates to fabricate high-performance flexible supercapacitors. Finally, the future challenges and perspectives for the development of flexible supercapacitors based on bendable substrates and their applications are discussed.

1. Introduction

Production of renewable energy has been rapidly increasing from readily available resources such as solar and wind energy.1 However, due to the irregular nature of solar and wind energy, efficient energy storage systems are highly desirable to store the electrical energy generated from these sources. This will improve the effective use of the entire power system by storing energy when in excess and releasing it at times of high demand. In this regard, among various energy storage systems, the most promising for this ongoing new smart-storage model are the electrochemical energy storage (EES) systems, including batteries, supercapacitors (SCs), etc.25 Batteries with high energy density and SCs with high power density are considered to be recent outstanding candidates as low-cost, eco-friendly, high-performance energy storage devices. However, batteries have some shortcomings such as limited cycle life, high manufacturing cost, low power density, and the requirement of several hours of charging time. On the other hand, SCs can store energy at a high rate by forming electrochemical double layers of charge or through pseudocapacitive surface redox reactions.5 Therefore, SCs have attracted significant attention recently, mainly due to their high power density (one or two orders of magnitude higher than that of batteries), long cycle life (two or three orders of magnitude longer than that of batteries), and the ability to store and release the energy within the time frame of a few seconds. Also, SCs offer higher reliability and better safety than batteries, which leads to a much lower maintenance cost.2

During the past several years, we have witnessed a radical evolution of electronic devices in which one of the major trends has been increased portability. Due to the popularization of commercialized pocket electronic devices, such as wearable or foldable electronics, electronic paper, mobile electronic devices, and smart products, the modern world is strongly demanding flexible, inexpensive, lightweight, environment-friendly energy storage devices.3 Particularly, flexible SCs (FSCs) have all the significant advantages mentioned above, which could make them very promising energy storage devices. For the development of high-performance FSCs, special attention should be paid to the electrochemical, mechanical, and interfacial properties of the flexible electrodes.4 Typically, the flexible electrodes should be highly conductive and have high surface area in addition to their required flexibility. Nowadays, various strategies have been developed especially for the design of flexible electrodes for FSCs, such as (i) fabricating flexible free-standing films of the active materials; (ii) supporting active materials on flexible substrates. Among these strategies, the use of flexible substrates as support for active materials is attracting great attention due to their low cost, high flexibility, and smooth surfaces. Therefore, novel support structures for the loading of pseudocapacitive materials are expected to yield high capacitance and flexibility while also being inexpensive and environmentally friendly.5

To this point, a number of recent reviews have presented excellent overviews of the progress towards carbon-based SC materials, such as activated carbons (ACs), carbon nanotubes (CNTs),68 graphene,918 and templated carbons,1923 different structures from zero- to three- dimensional (0D–3D),2427 and hybrid composites.5a, 28, 29 However, very few reviews were focused on the FSCs.30 Also, due to the growing interest towards lightweight, flexible, wearable electronics for widespread portable applications, there is strong demand for the development of such devices. Therefore, in this review, we will provide a summary of the recent research activities focusing on the use of flexible substrates as shown in Figure 1. Recently, metal substrates, carbon-based electrodes, porous materials such as conventional paper, textiles and sponges, and cable type electrodes with thin conducting layers have been used as substrates to fabricate high performance FSCs.

Figure 1.

An overview of support substrates such as metal, carbon-based electrodes, cable-type electrodes, and porous materials such as conventional paper, textiles, and sponges to fabricate high-performance FSCs.

  • 1.Metal substrates: To make flexible devices of solar cells, SCs, and batteries, metal is one of the substrate materials under consideration. The various metal substrates such as stainless steel (SS), aluminum (Al), copper (Cu), nickel (Ni), and titanium (Ti) have many advantages, such as high strength, high conductivity, and ease of preparation. Recently several reports proved that metal substrates can be excellent supports for fabricating high-performance SCs.31, 32
  • 2.Carbon paper and carbon nanofoam: Carbon paper consists of carbon microfibers manufactured into flat sheets. It is used as an electrode that facilitates the diffusion of reactants across the catalyst layered portion of the membrane electrode assembly.33, 34 Carbon nanofoam consists of a cluster-assembly of carbon strings in the form of a loose 3D web. Due to their porosity, mechanical integrity, large surface area, and good conductivity, they have been widely used as a substrate in fuel cells, SCs, etc.35
  • 3.Conventional paper substrates: Various forms of paper are widely used in our daily life for wiping, packaging, and decorating. In addition, applications of paper have been expanded to flexible electronic devices, such as photodiodes, transistors, circuits, and displays.36 Recently, several studies demonstrated that paper can also be an excellent support for loading active materials to fabricate high performance SCs.37
  • 4.Textile substrates: A textile is a flexible, porous material made by weaving or pressing natural or synthetic fibers, such as cotton or polyester. Emerging research on cotton fabrics with versatile functionalities revolutionized the use of cotton fabrics as a wearable platform for a multitude of applications.38 Moreover, prototypes of textile SCs39 and fiber SCs40 were recently demonstrated for textile-based energy storage applications.41
  • 5.Sponge substrates: Synthetic sponges made of cellulose or polyester have a porous nature and are used commonly for packaging, household cleaning, and personal care. In particular, the commercially available synthetic sponge was employed as an excellent substrate for the fabrication of SC devices.4245
  • 6.Cable type substrates: Recently, as a new type of FSC, the cable-type stretchable SC has been reported, which can be applied to various wearable electronics because this platform can be interwoven into any shape or place. The cable-type capacitors could be deformed into any complex shape while maintaining high performance and integration.46, 47

Thus, in this review, first, we provide some introductory background on SCs that will facilitate our review and analysis of the literature. Then we provide a summary of the use of metal, carbon-based, porous electrodes and porous substrates, such as paper, textiles, sponges, and cable-type electrodes with thin conducting layers as substrates for the loading of pseudocapacitive materials to develop FSCs. Finally, we will discuss the future challenges and perspectives of SCs.

2. Principles of Supercapacitors

2.1. Energy storage mechanisms of SCs

SCs store the energy by using two operating mechanisms, distinctly for the electrochemical double-layer capacitor, the pseudocapacitor, and hybrid capacitors, which will each be discussed in turn.

(i) The electrochemical double-layer capacitor (EDLC) stores electrical energy through the electrostatically reversible adsorption of ions of the electrolyte on the electrode surface, which is electrochemically stable with a high accessible surface area.48 As there is no chemical reaction involved in this charge storage mechanism, the process is highly reversible for millions of cycles and results in a long lifetime for the capacitors. The specific capacitance, C (F g−1) of a SC can be described as follows:(1)

equation image(1)

for which εr is the relative dielectric constant of electrolyte, ε0 is the dielectric constant of vacuum, d is the distance between electrolyte ions and the electrode (nanoscale charge separation distance), and A is the specific surface area of the electrodes. The operation mechanism of the EDLC is based on surface dissociation as well as ion adsorption from both the electrolyte and crystal lattice defects. As shown in Figure 2 a, this EDLC arises at the interface between the electrode material particles and electrolyte, where electric charges accumulate on the electrode surfaces and, to maintain electro-neutrality, electrolyte ions with a counterbalancing amount of charge build up on the electrolyte side. During the process of charging, the electrons travel from the positive electrode to the negative electrode through an external load. Within the electrolyte, cations move towards the negative electrode and anions move towards the positive electrode. In this way, energy is stored at the double-layer interface. During discharge, the reverse processes take place. Because there is no transfer of charge across the interface between the electrolyte and electrode, there are no chemical or composition changes in this EDLC process (i.e., non-Faradaic processes). Consequently, charge storage in EDLCs is highly reversible, leading to very high cycling stabilities (up to 106 cycles).49

Figure 2.

Schematic of two different charge-storage mechanisms: a) EDLC and b) redox reactions based pseudocapacitance.

The performance of an EDLC is dependent on several factors, such as the nature of the electrode surface and electrolyte. The materials for EDLCs can be optimized by choosing electronically conducting materials with high surface area, such as carbon-based nanomaterials, ACs, graphene, CNTs, or carbon aerogels (CAs). An EDLC can utilize either an aqueous or organic electrolyte. Aqueous electrolytes with common acids or bases such as H2SO4 and KOH show lower equivalent series resistances (ESRs) and are less subject to the minimum-pore-size requirements as compared to organic electrolytes with organic solvents such as acetonitrile. However, a lower breakdown voltage is one of the shortcomings of aqueous electrolytes. Therefore, the choice between an aqueous and organic electrolyte often depends on the intended application of the SCs,50 with the consideration of the tradeoffs between capacitance, ESR, and voltage.51, 52 Various forms of carbon nanomaterials have been investigated as effective materials in EDLC electrodes due to their advantageous features such as high conductivity, electrochemical stability, and proper porosity. ACs are the most widely used for the EDLC applications, but many other forms of nanostructured carbons, including aerogels,53 nanotubes,54 carbide-derived carbons,55 onion-like carbons,56 and graphene,57 have been also explored as alternatives to ACs for improving the energy density and power density for the EDLCs.58

(ii) In contrast to EDLCs, pseudocapacitors store electrical energy faradaically through the transfer of charge between the electrode and electrolyte as shown in Figure 2 b. This is accomplished through electrosorption, reduction–oxidation reactions, and doping–dedoping processes.48 These processes of pseudocapacitors can provide greater capacitances and energy densities than the EDLCs.59 There are different types of electrode materials used for pseudocapacitors: conducting polymers such as polyaniline (PANI) or polypyrrole (PPy) and transition metals oxides such as RuO2, MnO2, or Co3O4, etc.6066 It has been shown that the pseudocapacitive processes increase the working voltage as well as the specific capacitance of the SCs.67 As the electrochemical processes occur both on the surface and in the bulk near the surface of the solid electrode, a pseudocapacitor exhibits far larger capacitance values and energy densities than EDLCs (the capacitance of a pseudocapacitor can be 10–100 times higher than that of an EDLC). However, pseudocapacitors usually suffer from relatively low power densities and stabilities than the EDLCs.68 The most successful transition metal oxide for SCs is ruthenium oxide (RuO2) due to its high capacitance and stability. The capacitance of RuO2 is achieved through the insertion and removal of protons into its amorphous structure. In its hydrous form, the capacitance exceeds that of carbon-based and conducting polymeric materials.69 Furthermore, the ESR of hydrous ruthenium oxide is lower than that of other electrode materials. As a result, RuO2 pseudocapacitors have the ability to achieve higher energy and power densities than EDLCs and conducting polymer pseudocapacitors. The electrochemical reaction for the charge storage and delivery processes in the RuOxn H2O generally can be represented as follows:70(2)

equation image(2)

Four steps affecting the capacitive performance of the RuOxn H2O electrode were suggested: (i) electron hopping within the RuOxn H2O particles; (ii) electron hopping between the particles; (iii) electron hopping between the electrode materials and current collectors; and (iv) proton diffusion within the RuOxn H2O particles.71, 72 The advantages of amorphous RuO2 include high supercapacitance, high conductivity, and good electrochemical reversibility. The maximum supercapacitance of 788 F g−1 has been obtained from an amorphous sample of RuO2,73 which is at least two times greater than those obtained from other capacitive electrode materials. A tubular arrayed porous structure made of RuOxn H2O material was reported to yield a very high specific capacitance of approximately 1300 F g−1.74

Manganese oxide (MnO2) is another promising candidate for application to SCs due to its good supercapacitive performance, low cost, and environmentally benignity. The capacitive behavior of MnO2 was reported for electrochemical cycling in an aqueous KCl electrolyte.75 The charge storage mechanism for the pseudocapacitive MnO2 is based on surface adsorption of electrolyte cations A+ (e.g., K+, Na+ and Li+) as well as proton incorporation according to the reaction(3):58

equation image(3)

The theoretical capacitance of MnO2 is approximately 1300 F g−1, but this value has been rarely achieved in experiments mainly due to its poor electronic conductivity. In addition, two factors strongly influence the cycle life of the MnO2 electrode: (i) manganese dissolution and (ii) the oxygen evolution reaction. Unfortunately, most of the MnO2-based devices show approximately 20 % reduction in supercapacitance after about 1000 charge/discharge cycles, which suggests the difficulty in achieving long term cycling stability for a device.

Conducting polymers are the subject of intense investigation as SC electrode materials. The conducting polymers have electrochemical characteristics of reversible doping and de-doping ability so that they can store the charge throughout the entire volume. Several kinds of conducting polymers such as PPy,76 PANI,77 and polythiophene78 are used for SC applications. They can provide different properties through their molecular design such as various and flexible shapes, lightweight devices due to their low specific gravity, and an eco-friendly nature.79 However, if used as bulk materials, conducting polymers suffer from a limited cycling stability that leads to the decay of their electrochemical performance.

(iii) Hybrid SCs can be another type of SC combining the EDLC and pseudocapacitors. In hybrid SCs, both the EDLC effect and the pseudocapacitance can be simultaneously generated in a single SC. From the hybrid SC one can achieve higher energy and power densities with good cycling stability by utilizing both Faradaic and non-Faradaic processes to store the charges. In hybrid SCs the redox reaction takes place on one of the electrodes and the non-Faradaic charge/discharge process occurs on another electrode in a single SC. Such a SC is called an asymmetric hybrid SC. Other two types of hybrid SCs are battery-like hybrids and composite hybrids.29 In practice, various types of SCs have been investigated to combine different electrode materials to improve the SC performance.80

2.2. Energy density and power density of SCs

The energy density and power density are two major parameters for energy storage devices. Energy density represents the amount of energy stored per unit mass, whereas the power density describes how fast the energy is released. The energy density (E) and power density (P) are calculated by:(4), (5)

equation image(4)
equation image(5)

in which C is the capacitance, V is the voltage, R is the ESR, and m is the deposited mass. The equations indicate that the capacitance–voltage window must be high and the ESR must be low for an ideal energy storage system.

3. Use of Flexible Substrates for Loading of Pseudocapacitive Materials

3.1. Metal substrates

Due to the excellent properties of metal substrates, such as the high electric conductivity and good mechanical properties, they have been used extensively as electrode substrates for batteries and SCs.81, 82 Furthermore, the electrode configuration of electroactive materials that are directly synthesized on the substrate surface could lead to an increase in energy density and flexibility of SCs, whereas conventional powder-formed materials require additional binders, resulting in low energy density and mechanical durability. In this sense, branched MnO2 nanorods directly grown on a flexible stainless-steel (FSS) substrate were developed by using a one-step solution method as shown in Figure 3 a.83 Branched MnO2 nanorods on FSS substrates showed excellent electrochemical properties such as a high capacitance of 578 F g−1 and high electrochemical stability of 86 % after 5000 cycles, as shown in Figure 3 d. In addition, preparation of the branched MnO2 nanorods on FSS substrates with 30 cm2 areas (Figure 3 b) was demonstrated, in which the device fabrication and demonstration by using MnO2 FSCs were achieved with an application to light-emitting diodes (LEDs), as shown in Figure 3 e, f. It is interesting that after charging the cell, MnO2 FSCs could effectively power the LED device even though the substrate was rolled in the glass jacket (Figure 3 c and f). Moreover, vertically aligned MnO2 nanorod forests on large FSS substrates were synthesized by Santhanagopalan et al.84 by using a hydrothermal synthesis with high-voltage electrophoretic deposition technology. As-prepared vertically aligned MnO2 nanoforests revealed an increased surface area and reduced contact resistance, which led to the high electrochemical performance of the SCs. A symmetric cell was constructed by placing two MnO2-nanorod-array electrodes in parallel over a filter paper as the separator and Na2SO4 (0.1 M) as the electrolyte. The SCs showed a high power density of 340 kW kg−1 at an energy density of 4.7 Wh kg−1 and excellent cyclability (over 92 % capacitance retention over 2000 cycles). Gund et al.85 successfully demonstrated the temperature-dependent morphological evolution and its subsequent effect on the electrochemical properties of Ni(OH)2 on FSS substrates. Different nanostructures of Ni(OH)2 such as nanoplates, stacked nanoplates, nanobelts, and nanoribbons have been fabricated on FSS substrates. Among different nanostructures, Ni(OH)2 nanoplates showed the maximum specific capacitance of 357 F g−1 with good long-term cycling stability.

Figure 3.

a) Schematic representation of the optimized ion-diffusion path in branched nanorods and cross-sectional SEM image of branched MnO2 nanorods on the FSS substrate. b) Digital photographs of large-surface-area, flexible MnO2 on a FSS substrate. c) Illustration of the stacked MnO2 electrode with a separator to assembly the device. d) CV curves of the branched MnO2 nanorods at 1st, 1000th, and 5000th cycles. e) Illustration of the assembled cell of the MnO2 electrode in a Na2SO4 (1 m) aqueous electrolyte. f) Setup used for the actual demonstration of flexible MnO2 electrodes. Reprinted with permission from Ref. 83, Copyright 2013, Wiley-VCH Verlag GmbH & KGaA.

Chou et al.86 prepared flexible, porous Co(OH)2 nanoflake films by using an electrodeposition method on inexpensive FSS mesh and reported the maximum capacitance of 609.4 F g−1. The specific capacitance was maintained with only a slight decrease of less than 5 % even though the mass loading of Co(OH)2 increased by more than 340 % from 0.14 to 0.62 mg cm−2. The specific capacitance showed a decrease by less than 15 % for a current density increase up to 10 times from 714 to 7143 mA g−1, which indicates a good high-rate performance. Additionally, only 19 % specific capacitance loss appeared after 3000 cycles, showing the long-term electrochemical stability, and after 3000 cycles, the specific capacitance of the porous Co(OH)2 nanoflake electrode was 364 F g−1, which was still higher than that of carbon-based materials. Recently, uniformly stacked Mn3O4 nanosheets were prepared on a FSS mesh-like substrate by using a simple chemical bath deposition method.87 For the application of the prepared Mn3O4 nanosheets as an electrode in SCs, they exhibited excellent electrochemical performances due to the high utilization of active materials. The specific capacitances of stacked Mn3O4 nanosheets were found to be 398, 332, and 278 F g−1 for the mass loadings of 0.023, 0.027, and 0.029 g cm−2, respectively. Interestingly, the capacity retained after 2000 cycles was approximately 82 % of the maximum specific capacitance.

For practical applications, prototype FSCs have been developed by using FSS substrates with large areas. For example, Dhawale et al.88 reported a chemical strategy for the synthesis of nanostructured PANI electrodes for SC on large-area FSS foil (area 100 cm2). PANI electrodes showed outstanding electrochemical performance, including high specific capacitance and very stable cycle life. The specific capacitance reached a maximum value of 503 F g−1 with a specific energy and a specific power density of 96.2 Wh kg−1 and 8.9 kW kg−1, respectively. Furthermore, the practical demonstration for large-scale application of PANI electrodes was given by powering a toy fan. Nanostructured Mn3O4 thin films were prepared by using a green, successive ionic layer adsorption and reaction method on the FSS substrate.89 These films were further assembled to fabricate lightweight and portable Mn3O4–Mn3O4 symmetric SC cells. This cell achieved a maximum specific capacitance of 72 F g−1 with a high stability over 10 000 cycles. Thus, the FSS substrate could fulfill the recent as well as future demands for compact and mechanically flexible energy storage devices to improve the transportability of energy for energy management applications by using ultrathin devices.

Besides the FSS substrate, Al,90 Cu,91 Ni,92 and Ti93 substrates also have been used for preparing FSC electrodes. For example, Reit et al.90a reported vertically aligned CNTs (VACNTs) on flexible Al substrates by using a low-pressure chemical vapor deposition (CVD) process. The flexible VACNT electrode showed a specific capacitance of 79 F g−1, energy density of 1.1 Wh kg−1, and a power density of 8.6 MW kg−1 in an aqueous NaCl electrolyte. Zhang et al.91a fabricated flexible electrode by using CuO nanobelts and CNTs on a metallic Cu substrate. A CuO/CNTs material showed specific capacitance of 150 F g−1 at a current density of 1 A g−1 in a LiPF6/ethylene carbonate–diethyl carbonate electrolyte. Also, it showed an energy density of 130.2 Wh kg−1 and power density of 1.25 kW kg−1 with long-term cycling stability over 1000 cycles. Zhou et al.92a prepared an electrode of PPy-anchored CoO nanowire arrays on the flexible Ni-foam current collector. The PPy-anchored CoO nanowire array electrode showed a specific capacitance of 2223 F g−1 in a NaOH aqueous electrolyte with a high cycling stability maintained over 2000 cycles. Furthermore, asymmetric SCs composed of PPy-anchored CoO as a positive electrode and AC film as a negative electrode showed an energy density of 11.8 Wh kg−1 and power density of 5.5 kW kg−1 with excellent cyclability over 20 000 cycles. Xia et al.93a reported a hierarchical Co3O4@Pt@MnO2 core–shell–shell structured array electrode on the flexible Ti substrate. A high specific capacitance of 539 F g−1 was achieved by using the Co3O4@Pt@MnO2 electrode at a current density of 1 A g−1 in an aqueous Na2SO4 electrolyte. Moreover, this bendable electrode showed high specific energy and power densities of 39.6 Wh kg−1 and 19.6 kW kg−1, respectively, at a rate of 40 A g−1 with excellent cycling stability over 5000 cycles.

3.2. Carbon paper and carbon nanofoams

Carbon fiber paper (CFP) consists of a network of microsized carbon fibers that have been extensively employed in proton-exchange-membrane fuel cells as electrode substrate for the catalyst support, current collector, and gas transport layer.9497 From the similar needs in the electrochemical reactions of SCs and fuel cells, CFP can be a promising current collector and backbone for the conformal coating of transition metal oxides for SCs. The single carbon fibers in CFP are well connected to form a conductive network with appropriate pore channels, which can create an efficient electron transportation path as well as effective electrolyte access to the electrochemically active materials. To prepare the CFP with a high surface area, carbonization of the polymer nanofiber paper as the carbon source has been developed. For example, Ra et al.98 reported the synthesis of porous CFP by one-step carbonization/activation of polyacrylonitrile at 1000 °C in a CO2 atmosphere. The SCs made of such PAN-derived carbon paper were operated at high power density both in aqueous and in organic electrolytic media, keeping high energy density values. The PAN-based nanofiber papers contain nitrogenated and oxygenated functionalities that affect the redox reactions of the SCs, giving a high pseudo-faradaic contribution. The well-modified pore size distribution of carbon paper with nanometer-sized pores and mesopores, gives rise to a high capacitance and a high power capability in an organic medium.

Carbon paper substrates using CNTs have also been fabricated. The CNTs, consisting of single or multilayer graphene layers, possess excellent anisotropic mechanical strength and high electrical conductivity, and thus have attracted interest in fabrication of macroscopic CNT films for FSC electrodes. Di et al.99 prepared ultrastrong, fold-crack-free CNT films by solid-state layer-by-layer assembly of CNT sheets without chemical treatment as shown in Figure 4 a. The CNT films have a maximal tensile strength of 2.0 GPa and Young’s modulus of 90 GPa, which is higher than metal substrates such as Al foil and high-strength steel as depicted in Figure 4 b. Furthermore, the electrical resistance obtained showed negligible change after the film was bent to a radius of 1.5 mm for 2000 times. Even after the film was folded and unfolded to 180° over 100 times, the resistance did not show an observable increase. By using this CNT film as a current collector, a flexible CNT/MnOx composite electrode was prepared by using an electrodeposition method. MnOx-deposited flexible CNT papers displayed a specific capacitance of 199.4 F g−1 at a current density of 1 A g−1 in a Na2SO4 aqueous electrolyte as demonstrated in Figure 4 c. Moreover, the composite electrode could even maintain 80 % of the capacitance value with the current density increasing from 1 to 20 A g−1, showing high stability and good rate performance as presented in Figure 4 d. Such high capacitance performance was attributed to the excellent electrical conductivity of the aligned CNT strip that can provide fast charge transport paths for MnOx coatings. To further improve the mechanical strength of CNT-based paper, hybridized CNT papers with other materials have been studied. For example, multiple conductive and mechanical networks were introduced to connect each component in the flexible electrodes. Zhang et al.100 recently reported all-paper-based SCs by using microfibrillated cellulose (MFC) as an skeleton, MWCNTs as an electrode material, and polyethylene oxide (PEO)–LiCl as the solid electrolyte. The electrode sheets were made by impregnating modified MWCNTs and solid electrolyte (PEO and LiCl) into the MFC sheets. To assemble a symmetric cell, two electrode sheets were used to sandwich the MFC separator sheet. A symmetric cell showed a specific capacitance of 154.5 mF cm−2. Moreover, the specific capacitance remained almost the same when the SCs were bent under different curvature, and the tensile stress and modulus of the electrode were measured to be 1 MPa and 123 MPa (from the longest straight line), respectively, which are much better than the conventional liquid-based or gel-based SCs, revealing the high flexibility of the symmetric SCs.

Figure 4.

a) Digital photographs of a flexible CNT film strip. b) Tensile strength and modulus comparison of as-prepared CNT films with buckypaper, graphite foil, and graphene paper. c) Cyclic voltammograms of a CNT strip and a CNT/MnOx composite strip at a scan rate of 100 mV s−1. d) Specific capacitances of the CNT/MnOx composite film at different current densities. Reprinted with permission from Ref. 99, Copyright 2012, American Chemical Society.

To enhance the energy density of FSCs, the pseudocapacitive materials were introduced into the carbon paper and carbon nanofoam substrate. For example, Fischer et al.101 demonstrated a very simple approach for incorporating homogeneous, nanoscale MnO2 deposits within the twisting interior of carbon nanofoam electrodes. Initially, the carbon nanofoam was prepared by pyrolysis of carbon-paper-supported resorcinol–formaldehyde nanofoam under an inert atmosphere. MnO2 was subsequently incorporated by a self-limiting electroless deposition method and controlled permanganate self-decomposition under neutral pH conditions. Highly porous MnO2–carbon nanofoam hybrids with pore sizes in the range of 10–60 nm were formed. The electrochemical parameters were evaluated in an aqueous Na2SO4 electrolyte and a specific capacitance was found to be 110 F g−1 with an areal capacitance of 1.5 F cm−2. MnO2-supported carbon-nanofoam hybrid electrodes showed good cycling stability at a high charge/discharge rate of 0.25 A g−1. Furthermore, Yang et al.102 prepared two different nanostructures such as nanocubes and nanonets of Co3O4 on CFP to investigate the suitable microstructure for SCs. Co3O4 nanonets embedded on carbon paper showed a high specific capacitance of 1190 F g−1 for 0.4 mg cm−2 mass loading of Co3O4 at a current density of 0.25 A g−1. The specific capacitance was retained as high as 1124 F g−1 even as the current density increased to 25.34 A g−1, indicating good rate capability of the Co3O4-nanonet-supported carbon paper. This was possibly due to the effective hierarchical design structure, which facilitated electron/ion transport and the redox reaction. Negligible capacitance losses were observed over 5000 charge/discharge cycles for the carbon-paper-supported Co3O4 composite. The nanoscale network structure of the Co3O4 claimed in this investigation was effective to release the stress caused by the volume change associated with reversible intercalation and/or adsorption of the charge carrier. Thus, Co3O4-nanonet thin films supported on a conductive CFP demonstrated the enhanced redox kinetics at high charge/discharge rates while maintaining electrochemical and structural stability.

Besides CNTs, graphene is the most intensively studied of the carbon allotropes in the FSC electrode area. In this regard, graphene coating has been used to improve the mechanical strength and conductivity of electrode materials. For instance, Jin et al.103 fabricated a ternary composite paper electrode composed of reduced graphene sheet (GR), CNTs, and MnO2. The paper was prepared by electrodeposition of MnO2 on flexible CNT paper and further adsorption of GR on its surface by a simple soaking method. The electrical conductivity of the composite paper remained stable even after bending 1000 times, which could indicate the high flexibility of the composite paper. The CNT/MnO2/GR paper showed a specific capacitance of 486.6 F g−1. The asymmetric FSC was prepared using CNT/MnO2/GR as the positive electrode, CNT/PANI as the negative electrode, and Na2SO4/PVP gel as the electrolyte and separator. An energy density of 24.8 Wh kg−1 was achieved and the Coulombic efficiency was near 100 % using the asymmetric SC.

3.3. Conventional paper substrates

As a cheap and recyclable material, various types of paper are widely used in our daily life for wiping, packaging, and decorating. Recently, conventional paper has been adopted in flexible electronic devices, such as photodiodes, transistors, circuits, and displays.104107 Besides, the SC is another application of the conventional paper that can be used as an excellent support for loading of active materials.37, 108111 As mentioned before, the CNTs have been extensively studied for SC electrodes due to their high mechanical strength and electrical conductivity. In this sense, Hu et al.37 reported an excellent example of the use of a CNT–paper composite for SCs. Initially, the CNT ink was prepared by mixing the appropriate amount of single-walled CNTs (SWCNTs) in surfactant-assisted aqueous solution. This CNT ink was subsequently applied onto a piece of commercial copy paper by the simple and scalable Meyer rod coating method as shown in Figure 5 a, b. Such conformal coating of CNTs makes the copy paper highly conducting with a sheet resistance of approximately 10 Ω sq−1. The paper electrode was flexible and did not lose its conductivity under bent/curled conditions as depicted in Figure 5 c, d. The strong binding properties between CNTs and paper are attributed to the large capillary effect and strong Van der Waals force.

Figure 5.

a) Meyer rod coating of CNTs ink on commercial Xerox paper. b) Conductive Xerox paper after CNT coating. c) Sheet resistance changes after bending the paper with different radii. d) Film adhesion test with commercial scotch tape: CNTs remain bound to the paper. e) Gravimetric capacitances at various currents measured in different electrolytes. f) Capacitance retention measured in different electrolytes. After 40 000 cycles, 97 % and 99.4 % of the initial capacitances are still maintained for sulfuric acid and organic electrolytes, respectively. Reprinted with permission from Ref. 37, Copyright 2009, National Academy of Sciences.

This CNTs-coated copy paper was directly applied as both active electrodes and current collectors in SCs for which a piece of plain paper functioned as a separator. The maximum specific capacitance of 200 F g−1 (based on CNT mass only) was achieved in H2SO4 electrolyte for 72 mg cm−2 CNTs loading, as shown in Figure 5 e. Additionally, they showed good specific power and energy density of 200 kW kg−1 and 47 Wh kg−1, respectively, in an organic electrolyte. Further, CNTs-coated paper showed excellent cycling performance with negligible capacitance losses (only 3 % in H2SO4 and 0.6 % in organic electrolyte) after 40 000 cycles as demonstrated in Figure 5 f.

In earlier reports, the CNTs were coated individually on different anode and cathode substrates, and they were assembled with a separator for the SC devices. For low-cost and lightweight paper-based FSCs, however, Hu et al.108 demonstrated an integrated structure which consists of anode, cathode, and separator on a single sheet of paper. This integrated structure could allow low-cost and high-speed printing for lightweight, paper-based SCs. In the first step, poly(vinyliedene fluoride) (PVDF) was coated on both sides of paper by Meyer rod coating method and the paper was dried at 65 °C in an oven for 20 min. On this PVDF-coated paper, CNT ink was subsequently applied on both sides. Thus, CNTs on both sides act as the cathode and anode whereas the PVDF-treated paper itself functioned as an electrolyte membrane and separator. Apart from Meyer rod coating, Hu et al. also investigated an even-more-scalable method in the form of ink-jet printing to prepare the CNT-paper composite.108

As the introduction of pseudocapacitive materials into electrodes could further improve the performance of SCs, Kang et al.109 fabricated flexible, paper-based SCs using CNTs and pseudocapacitive MnO2 materials. First, CNTs were synthesized by using a CVD process and then dispersed in a surfactant (sodium dodecylbenzenesulfonate) mediated aqueous solution to prepare CNT ink. This CNT ink was further applied on the paper by using a simple drop and dry method. Subsequently, flower-like nanostructures of MnO2 were electrochemically deposited on the conducting CNT-supported paper. Mass loading of MnO2 was controlled by the number of the cyclic voltammogram (CV) cycles. MnO2/CNT/paper-based SCs showed a high specific capacitance of 540 F g−1 considering only the mass of MnO2 (0.38 mg cm−2) in Na2SO4 electrolyte. At a current density of 5 A g−1, the specific energy and power densities were found to be 20 Wh kg−1 and 1.5 kW kg−1, respectively. However, at low current density (1 A g−1), a specific energy density of 43.3 Wh kg−1 was achieved at the expense of the specific power density (0.4 kW kg−1). Over 1000 charge/discharge cycles, only 5 % capacitance loss was observed even at high current density of approximately 7 A g−1.

Recently, many processes were involved in solution-based synthesis methods, and thus they could lead to environmental unfriendliness and high cost. Zheng et al. demonstrated the fabrication of SCs by directly drawing multilayer graphene onto a commercial Xerox paper by using graphite rods or pencils.107 Electrochemical testing of the SCs showed capacitances of 12 F g−1 and 23 F g−1 for the copy paper and home-made paper, respectively, in H2SO4 electrolyte. The improved electrochemical performance of the homemade paper was attributed to the more uniform surface texture and the resulting smaller and thinner graphitic flakes. The SCs based on graphene paper retained approximately 90 % capacity after 15 000 cycles, showing the excellent cycling performance.

3.4. Sponge substrates

Commercial, cost-effective sponges are composed of interconnected polyester fibers and can be used as excellent support for SC devices due to their various properties as follows: 1) the sponge possesses hierarchical macroporous network structure and high absorbance to liquid media with high internal surface area, 2) the interconnected and junction free network structures enable uniform and continuous coating without any interruption, 3) the macroporous structures with high absorbance are helpful for the access of the electrolyte to the surface of the active electrode. These unique properties of the sponge structure suggest an excellent supporting substrate for loading active materials.112114 Based on these ideas, Li et al.115 prepared CNT sponges consisting of highly porous conductive networks by using a CVD process. The original CNT sponge and its heavily compressed thin sheets also showed high flexibility and was able to be further folded without significant fractures. A coin-sized symmetrical SC cell was assembled with the CNT sponge in LiPF6/ethylene carbonate–dimethyl carbonate–ethyl methyl carbonate (w/w/w=1) electrolytes. The devices delivered a maximum specific capacitance of 28.5 F g−1 at 1 mV s−1 and showed excellent stability over 15 000 charging cycles with negligible degradation. The good performance could be originated from the robust sponge structure capable of maintaining the conductive network and surface area under cyclic compression. To further improve the energy density, pseudocapacitive Fe3O4 materials were composited with carbonaceous sponge materials. For instance, Wu et al.116 fabricated 3D flexible carbonaceous gels (CGs) from crude biomass watermelon as a carbon source, which were then used as a support of Fe3O4/CGs composites. The Fe3O4/CGs composites further transformed into magnetite CAs (MCAs) by further calcination. The MCAs were able to keep the porous structure of the original CGs, which allows the sustained and stable transport of both electrolyte ions and electrons to the electrode surface. The MCAs exhibited an excellent capacitance of 333.1 F g−1 at 1 A g−1 in KOH aqueous solution. Moreover, the MCAs also showed high cycling stability with 96 % of the capacitance retention after 1000 cycles.

As another pseudocapacitive material, PPy is one of the most important materials in SC electrodes due to its high electrical conductivity, facile synthesis, low cost, and high stability. Zhao et al.117 developed a unique strategy for in situ formation of PPy-graphene (PPy-G) foam by using PPy as a mediator. The PPy could effectively prevent the self-stacked behavior of graphene oxide (GO) during the hydrothermal process and it subsequently increased the surface area for forming large volume of 3D graphene with thin connection walls. A symmetrical two-electrode capacitor was assembled with a controlled compression state of PPy-G in a NaClO4 aqueous electrolyte. The electrode delivered a specific capacitance of 350 F g−1 at 1.5 A g−1 without significant capacitance degradation, even up to 1000 cycles.

Chen et al.42 synthesized a CNT–commercial sponge composite by the “dip and dry” method. The conformal CNT coating layer was continuous in all three dimensions and achieved a conductivity of approximately 1 S cm−1 for 200 nm thickness of CNT coating. Symmetrical SCs based on CNT-sponge were fabricated in an aqueous Na2SO4 (1 M) electrolyte and subjected to the CV test in a wide range scan rate of 0.001–200 V s−1. The resulting CV curves exhibited a rectangular shape at a scan rate up to 20 V s−1. The results also revealed a linear dependence of discharge current on the scan rate up to 8 V s−1. Over 100 000 cycles at a scan rate of 10 V s−1, the CNT-sponge exhibited only 2 % degradation of specific capacitance. To improve the energy density of device, flower-like MnO2 was further loaded onto the CNT-sponge by electrochemical deposition as shown in Figure 6 a–c, showing its flexible features. The mass loading of the MnO2 on the CNT-sponge was controlled by using the deposition time. A specific capacitance of 1230 F g−1 was achieved for CNT/sponge/MnO2 at a scan rate of 1 mV s−1 for a 0.03 mg cm−2 mass loading, which is very close to the theoretical value of MnO2 (1370 F g−1). However, the specific capacitance of the CNT/sponge/MnO2 further decreased with increasing loading mass of MnO2, as shown in Figure 6 d. The CNT/sponge/MnO2 SCs exhibited specific power and energy densities of 63 kW kg−1 and 31 Wh kg−1, respectively, with only 4 % capacitance degradation after 10 000 cycles at a charge/discharge current of 5 A g−1 (Figure 6 e–f). Moreover, by using kitchen sponges, Chen et al.43 further investigated the electrochemical properties of the CNT/sponge/MnO2 in two organic electrolytes [1 M of tetraethylammonium tetrafluoroborate (Et4NBF4) in propylene carbonate (PC) and 1 M of LiClO4 in PC] to improve the energy density of the symmetrical SCs compared with that of aqueous electrolyte. It was observed that by replacing the aqueous electrolyte with organic electrolytes, the working voltages of the SCs can be steadily increased to 2.5 V, almost three times higher than the voltage in the aqueous electrolyte, which can further increase the energy density of SCs tripled in Et4NBF4–PC and by six times in LiClO4–PC electrolytes. The maximal energy density obtained was 152 Wh kg−1 in 1 M LiClO4–PC electrolyte. The excellent electrochemical stability of SCs in Na2SO4 under 10 A g−1 showed approximately 91.8 % of the initial capacitance was retained after 10 000 cycles of charge/discharge. The cycling performance of SCs in organic electrolytes showed stability of 75.5 % and 61.6 % after 10 000 cycles in Et4NBF4–PC and LiClO4–PC electrolytes, respectively, but the device retained a significant energy density advantage even after 10 000 cycles. Furthermore, Ge et al.118 synthesized sponge-structured graphene–MnO2 hybrid SCs by using a commercial sponge. The bare sponge was first coated with GO nanosheets by using a dipping procedure into the GO nanosheets aqueous solution and subsequently removing unwanted GO and drying. The GO-coated sponge was then dipped into hot hydroiodic acid for 1 min to reduce the GO nanosheets, forming three dimensional conductive frameworks as carbon materials. Subsequently, reduced GO (RGO) coated sponge was quickly dipped in Mn(CH3COO)2 and KMnO4 aqueous solutions, respectively, to deposit pseudocapacitive MnO2 on conductive sponge@RGO electrodes. The uniform and conformal coating of flower-like MnO2 was achieved due to the high absorption ability and macroporous surface of the sponge. Further, the mass loading of MnO2 on a skeleton of sponge@RGO was controlled by dipping times into Mn(CH3COO)2 and KMnO4 aqueous solution. The conformal coating of MnO2 flowers on the RGO-coated sponge increased the specific capacitance by 3–4 times as compared with sponge@RGO due to the pseudocapacitance brought about by the MnO2 nanostructure. The sponge@RGO@MnO2 exhibited an exactly rectangular CV, indicating ideal capacitive behavior. However, the specific capacitance of the sponge@RGO@MnO2 further decreased with increasing mass loading of MnO2. The highest specific capacitance achieved was 450 F g−1 by sponge@RGO@MnO2 with three times dip deposition at a scan rate of 2 mV s−1. Additionally, sponge@RGO@MnO2 provided the maximum energy density of 8.34 Wh kg−1 and the highest power density of 47 kW kg−1 at an operating voltage of 0.8 V. Furthermore, the sponge@RGO@MnO2 hybrid provided an outstanding cycle performance with 10 % degradation after 10 000 cycles at a charge/discharge current density of 10 A g−1.

Figure 6.

a) Photograph showing flexibility of a MnO2/CNT/sponge electrode. b) SEM image of a 3D macroporous hierarchical MnO2/CNT/sponge electrode. c) SEM image of a MnO2/CNT/sponge showing conformal coating of MnO2 onto the skeleton of a CNT/sponge. d) Variation of specific capacitance with scan rate for devices having different MnO2 deposition times, ranging from 3 to 40 min. e) Galvanostatic charge/discharge of MnO2 electrode for 5 min deposition time under different current densities. f) Plots of capacity retention vs. cycle number for a CNT/sponge device up to 100 000 cycles at a high scan rate of 10 V s−1 and 40 min MnO2/CNT/sponge SCs under 10 000 cycles at a specific current of 5 A g−1, respectively. Reprinted with permission from Ref. 42, Copyright 2011, American Chemical Society.

3.5. Textile substrates

Textile, or cloth, is a highly flexible and porous material produced from natural or synthetic fibers by pressing, weaving, knitting, or felting. Textiles exhibit various mechanical properties such as flexibility, light weight and high stretchability, depending on the different fabrication processes, which makes it a promising substrate in energy storage devices.119121 In addition to this, textiles possess 3D open-pore structures that can be helpful for the conformal coating throughout the entire textile network, which results in a much higher areal mass loading of active materials and therefore higher areal power and energy density.122125 For example, Jost et al.126 studied the electrochemical properties of porous carbon materials impregnated into woven cotton and polyester fabric textile substrates using a printmaking technique (i.e., screen printing). In particular, flexible electrodes of commercial ACs coated on both cotton lawns and polyester microfibers achieved high gravimetric and areal capacitances of an average of 85 F g−1 and 0.43 F cm−2, respectively, at 0.25 A g−1. Also, the symmetric SCs showed a capacity retention of 92 % over the 10 000 cycles. Liu et al.127 fabricated flexible graphene sheets (GNSs)–cotton cloth (CC) composite textiles by using a simple brush-coating technique followed by a heat-treatment process. As-prepared GNS–CC composite textiles exhibited 3D network structures, good electrical conductivity, high flexibility, and strong adhesion between GNSs and the cotton fibers. A symmetric SC was prepared by using the two GNS–CC composite fabrics and pure CC as a separator. Finally, the assembled layers were rolled up, put into a cylindrical bottle, dipped into the electrolyte, and covered tightly with the epoxy resin. The symmetric SC showed a specific capacitance of 81.7 F g−1 at a scan rate of 10 mV s−1 in a KOH aqueous electrolyte. The power density of 1.5 kW kg−1 was obtained at a scan rate of 3.0 A g−1 (with energy density of 7.13 Wh kg−1 at this point) at room temperature. It was observed that the specific capacitance showed a decrease of only 6.2 % after 1500 cycles, revealing that the symmetric SC has high cycling durability for repetitive cycling. Wang et al.128 developed a GO-assisted electrophoretic deposition (EPD) method to fabricate the porous hybrid graphene–CNT (G–CNT) layer on the carbon cloth textile. In the EPD process, GO acts as charging additives to co-deposit graphene and CNTs under an applied electrical field. CNTs could prevent restacking of graphene sheets after being reduced from GO. The porous structure of hybrid G–CNTs on carbon cloth could enhance the electrolyte accessibility when used as electrodes in SCs. The symmetric FSCs were assembled using two G–CNT/carbon cloths as positive and negative electrodes, glass fiber paper as the separator, and H2SO4 solution (1.0 M) as the electrolyte. At 1 A g−1, the G–CNT/carbon cloth electrode showed a specific capacitance of 151.0 F g−1, and 99.1 % of the initial capacitance still remained at 2 A g−1 after 2000 cycles. The G–CNT/carbon cloth FSCs showed high flexibility and they could be repeatedly bent without significant change of performance, demonstrating the performance of the FSCs were not influenced by mechanical bending stress. Furthermore, Hu et al.129 reported a CNT–textile composite by a simple dipping-and-drying method. To improve the conductivity of the CNT–textile composite, the dip-and-dry process was repeated several times. With a 200 nm thick CNT coating, the textile fiber achieved a conductivity of 125 S cm−1 with the sheet resistance of less than 1 Ω sq−1. The electrochemical properties of the CNT–textile composite with 0.24 mg cm−2 CNT loading in an organic electrolyte with LiPF6 (1 m) were demonstrated by galvanostatic cycling at current densities from 20 A cm−2 to 20 mA cm−2. The maximum specific capacitances achieved at 20 A cm−2 and 20 mA cm−2 were 140 F g−1 and 80 F g−1, respectively. Higher CNT mass loading up to 8 mg cm−2 was investigated, resulting in an elevated areal capacitance of 0.48 F cm−2. Due to the light weight of the textile (≈12 mg cm−2) the overall mass of the CNT–textile composite device was only 48 mg cm−2, for which the active material CNTs accounted for more than 30 %. Based on the overall mass, CNT–textile SCs showed an outstanding specific power density of 10 kW kg−1 and specific energy density of 20 Wh kg−1, compared to the commercial SCs. Concerning long-term cycling performance, CNT–textile SCs showed excellent stability with only 2 % loss after 130 000 cycles. As mentioned above, open-porous CNT–textiles provided an effective 3D support for deposition of other active functional materials.

To further enhance the energy density of SCs, by taking advantage of the textile support, Hu et al. electrochemically deposited pseudocapacitive MnO2 onto the CNT–textile composite with loadings in the range of 0.06–8.3 mg cm−2.130 A uniform coating of nanoflower-like MnO2 particles was achieved on the CNT–textile fibers. For application as an electrode in SCs, the highest specific capacitance achieved was 410 F g−1 for 0.06 mg cm−2 MnO2 loading at a scan rate of 5 mV s−1, and the highest areal capacitance obtained was 2.8 F cm−2 for a mass loading of 8.3 mg cm−2 with the low scan rate of 0.05 mV s−1. To reveal the advantages of the CNT–textile as a 3D support, 0.8 mg cm−2 MnO2 was deposited on both the CNT–textile and a flat Pt foil. It was observed that the specific capacitance of MnO2 supported on the Pt foil was 18 F g−1 and that on CNT–textile composite was 185 F g−1 at a scan rate of 5 mV s−1. Additionally, the pouch cell assembled with the MnO2/CNT/textile as the positive electrode and the reduced MnO2/CNT/textile as the negative electrode achieved a specific power of 13 kW kg−1 with a specific energy of 5–20 Wh kg−1. However, this MnO2/CNT/textile SC exhibited a relatively poor cycling performance (20 % capacitance loss after 200 cycles and 40 % loss after 10 000 cycles). The failure was possibly due to the loss of active materials and the mismatched capacity between the two electrodes.

As mentioned above, graphene as new class of carbonaceous materials possesses a high electric conductivity and high electrochemical stability, which could be further used to improve textile-based SCs for large-scale energy storage systems by Yu et al.131 Initially, graphene ink was prepared through a simple, low-cost and high yield, solution-based exfoliation technique. The graphene ink was applied by using a dip-and-dry method on the textile to fabricate a graphene–textile composite and finally the MnO2/graphene/textile was fabricated by subsequent electrodeposition of MnO2. Electrochemical performance of the MnO2/graphene/textile showed an excellent specific capacitance of 315 F g−1 in Na2SO4 at a scan rate of 2 mV s−1 for MnO2 loading of approximately 0.3 mg cm−2 (≈5 times higher capacitance than that of graphene–textile). With MnO2/graphene/textile as the positive electrode and CNT–textile as the negative electrode, an asymmetric SC was assembled. An asymmetric SC was achieved with a specific power density of 110 kW kg−1 and specific energy density of 12.5 Wh kg−1 with good cycling performance of approximately 95 % capacitance retention over 5000 cycles.

Previous reports suggest that SCs were constructed either by coating the textiles with functional thin film layers or by modifying the fibers with chemicals. However, the direct conversion of cotton textiles into electrically active textiles for constructing SCs remains a great challenge. In this regard, Bao et al.41 fabricated highly conductive and flexible activated carbon textiles (ACTs) by direct conversion of cotton t-shirt textiles through a traditional dipping, drying, and curing process. First, cotton textile was dipped in a NaF solution (1 M) and then dried at 120 °C for 3 h in a preheated oven, which was finally annealed in a horizontal tube furnace at 800–1000 °C for 1 h under vacuum and inert atmosphere conditions (Figure 7 a–c). The ACT was mechanically flexible even under folding conditions (Figure 7 c) and exhibited high conductance with a sheet resistance of 10–20 Ω sq−1. CV tests suggested that the ACT electrode achieved a specific capacitance of 70.2 F g−1 at a scan rate of 2 mV s−1, which is comparable to that of solution-processed textiles. High porosity, high conductivity, and superior capacitive characteristic make the ACTs unique supporting backbones for the deposition of pseudocapacitive MnO2 to construct MnO2/ACT hybrid composites for enhanced electrochemical performance.132 Electrochemical synthesis was performed by using ACTs as support and a mixed solution of Mn(CH3COO)2 and Na2SO4 as electrolyte for the deposition of MnO2. The mass loading of MnO2 was controlled by adjusting both the deposition current and time. Nanostructured MnO2 was uniformly coated onto the surface of ACT microfibers over almost the entire network of porous ACTs and exhibited a nanoflower-like morphology for deposition time of 120 min, as seen from Figure 7 d, e. The CV curves of the MnO2/ACTs hybrid composite showed a quasi-rectangular shape up to 20 mV s−1 scan rate, which further distorted from ideal capacitive behavior at a scan rate of 50 mV s−1, as shown in Figure 7 f. At a scan rate of 2 mV s−1, the specific capacitance achieved was 269.5 F g−1, which is 3 times more than that for ACTs. To further improve the energy density (i.e., potential window), Bao et al. assembled asymmetric SCs comprised of MnO2/ACT as the positive electrode, ACT as the negative electrode, and Na2SO4 (1 M) aqueous solution as electrolyte. Figure 7 g shows the CV curves of the asymmetric SC measured at different scan rates with a potential window ranging from 0 to 2 V. These CV curves show ideal capacitive behavior with nearly rectangular shape even at high scan rates up to 100 mV s−1. As seen from the Ragone plots (Figure 7 h) for the MnO2/ACT//ACT asymmetric SC, an energy density of 66.7 Wh kg−1 with power density of 0.8 kW kg−1 was achieved. Galvanostatic cycling tests over 1000 cycles for the MnO2/ACT//ACT asymmetric SC showed only 2.5 % loss in specific capacitance at a current density of 1 A g−1.

Figure 7.

Optical photograph of a) a commercial cotton t-shirt, b) a piece of ACT and c) a piece of ACT under folding conditions, showing its highly flexible nature. d) SEM image of a MnO2/ACT hybrid composite fabricated by electrochemical deposition MnO2 for 120 min, which shows that all the ACT fibers are uniformly coated by MnO2. e) SEM image of an individual ACT fiber coated with a thin film of MnO2 nanoflowers. f) Cyclic voltammatry curves of the MnO2/ACT hybrid composite at different scan rates in aqueous Na2SO4 (1 M) electrolyte. g) Cyclic voltammatry curves of the MnO2/ACT//ACT asymmetric SC at different scan rates with a potential window of 2 V in aqueous Na2SO4 (1 M) electrolyte. h) Ragone plots of the MnO2/ACT electrode, ACT electrode, and as-assembled MnO2/ACT//ACT asymmetric SC. Reprinted with permission from Ref. 41, Copyright 2012, Wiley-VCH Verlag GmbH & KGaA.

3.6. Cable-type substrates

Recently, novel cable-type SCs have been reported with the advantages that they could maximize the mechanical flexibility and be applied to various wearable electronics. These cable-type structured SCs can alleviate the shape restriction of conventional SCs, and therefore, such device can be woven into any shape and placed anywhere.133 For example, Fu et al.46 fabricated cable type FSCs that consist of two fiber electrodes, a helical spacer wire separator, and a Na2SO4 (1 m) electrolyte. Commercial pen ink as the active material was coated on the surface of the Ni wire substrate by using a simple dip-coating process. The FSC did not show any performance degradation after the bending test, which indicates the high flexibility of the devices. An areal capacitance value of 9.5 mF cm−2 was obtained from CV profiles at a scan rate of 1.0 V s−1. The capacitance was stable over 15 000 cycles with slight capacitance increases, which could demonstrate its good electrochemical stability. From charge/discharge cycles, during which the induced current increased from 0.083 to 16.7 mA cm−2, the power density improved from 0.042 to 9.07 mW cm−2, and the energy density of the FSC decreased from 2.70×10−6 to 1.76×10−6 Wh cm−2. For the electrochemical performances of two fiber electrodes, recently, Le et al.134 prepared a coaxial-fiber FSC, consisting of multiwalled CNTs (MWCNTs) coated onto carbon microfiber (CMF) bundles as the core electrode, with carbon nanofiber paper as an outer electrode, and with polymer gel as the electrolyte (Figure 8 a). The devices showed a length capacitance of 6.3 mF cm−1 at 230 μm of core electrode and energy density of 0.7 μWh cm−1 at power density of 13.7 μW cm−1, as depicted in Figure 8 b and c. Moreover, Figure 8 d demonstrated that 88 % of the capacitance of the coaxial-fiber SCs was retained after 1000 charge/discharge cycles. The change in the CV characteristics was negligible after 180° bending of devices, revealing that the components of devices such as CMF, MWCNTs, and CNF soaked with polymer gel electrolyte are highly flexible. Interestingly, this result implies that fiber SCs can be applied to textiles for flexible electronic devices.

Figure 8.

a) Schematic and digital photo of a coaxial-fiber SC. b) Cyclic voltammetry profiles of a coaxial fiber SC at different scan rates. Inset shows the high flexibility of the SC. c) Ragone plot with energy and power density per length or area of the core electrode. d) Cycle performance of a coaxial-fiber SC and coin-cell SC at a current density of 275 μA cm−1. Inset is a light-emitting diode powered using four coaxial fiber SCs connected in series. Reprinted with permission from Ref. 134, Copyright 2013, American Chemical Society.

Similarly, to improve the electrochemical properties, pseudocapacitive materials have been incorporated into carbon nanofibers. For instance, Wang et al.135 reported thread-like SCs constructed from two-ply CNT and PANI nanowire composite yarns with a polyvinyl alcohol (PVA)–H2SO4 electrolyte, in which two CNT@PANI@PVA yarns were twisted together to form the flexible two-ply yarn SC. The CNT@PANI yarn-based SC showed a high capacitance of 38 mF cm−2 compared with 2.3 mF cm−2 of the pure CNT yarn-based SC at a current density of 0.01 mA cm−2. Its capacitance was maintained at 91 % of its initial value after 800 charge/discharge cycles. Furthermore, the capacitance of the CNT@PANI@PVA SC changed very little even as the bending angle approached 180°, and maintained its capacitance almost fully after having been bent for 150 cycles at a constant scanning rate of 20 mV s−1. Consequently, as-prepared thread-like CNT@PANI@PVA SC can be knitted into a textile architecture alone or together with other smart devices and therefore it can be used in wearable electronics. All-solid-state FSCs based on a carbon/MnO2 core–shell fiber structure were reported by Xiao et al.136 The carbon fiber served as a scaffold and current collector, and MnO2 was deposited by using a low-cost redox process. All-solid-state SCs were prepared with two carbon/MnO2 core–shell fibers, polyester separator, and PVA/H3PO4 solid-state electrolyte. The device showed a high volumetric capacitance of 2.5 F cm−3 at a current density of 0.02 A cm−3, and an energy density of 2.2×10−4 Wh cm−3 at a scan rate of 20 V s−1. The capacitance also remained at 84 % of the initial capacitance after 10 000 charge/discharge cycles, showing high stability of electrochemical reaction. From CV profiles of the devices under bending states, variation of the CV curves of the SCs were not observed, revealing that the electrochemical properties were not significantly affected by the bending angle. Cai et al.137 prepared aligned PANI-coated MWCNTs composite fibers by using an electropolymerization method. The aligned MWCNT–PANI composite fibers were then coated with a layer of PVA/H3PO4 gel electrolyte on the composite fiber surface. Two modified composite fibers were finally twisted to fabricate the cable-type SC, and the twisted micro-SC showed a specific capacitance of 274 F g−1 and length capacitance of 263 mF cm−1. Furthermore, the device showed slightly decreased capacitance in the first 50 cycles, and then showed improved stability in the following 950 cycles. The device itself was also highly flexible and can be bent without an obvious destruction. Due to their high flexibility, these cable type SCs can be easily woven into clothes or other portable devices and serve as self-powered electric generators.

Besides carbon-fiber backbones, graphene-based cable-type SCs have also been successfully prepared. Recently, Meng et al.138 reported all-graphene core–sheath fibers, in which a core of graphene fiber was covered with 3D porous network-like graphene (GF@3D–G). The cable SC was comprised of two intertwined electrodes, both of which were solidified in the PVA/H2SO4 gel electrolyte for fabrication of an all-solid-state fiber SC. The GF@3D–G SC showed a specific capacitance of 25–40 F g−1 with cycling stability over 500 cycles. Almost overlapped CV curves were observed for the GF@3D–G SC, demonstrating the high mechanical flexibility and cycling stability of the GF@3D–G SC. The device delivered energy density of 0.4–1.7×10−7 Wh cm−2 and power density of 6–100×10−6 W cm−2.

4. Conclusions and Perspectives

In this review, we have summarized the recent progresses of the use of low-cost and highly flexible substrates, mainly focused on metal substrates, carbon-based paper, conventional paper, sponges, textiles, and cable-type electrodes for FSCs. These substrates are highly flexible and mechanically stable after severe bending, which enables the FSCs to be used in flexible, wearable, and lightweight electronics for widespread portable applications. However, each type of flexible substrate shows both advantages and weaknesses for application to FSCs, which are summarized in Table 1. For instance, the metal substrate shows high electrical conductivity, mechanical stability, and the possibility for large-area production for FSC electrodes; therefore, this type of substrate has been used widely for practical devices although it also shows relatively high cost, heavy weight, and low surface area. The carbon-based substrate has the advantages of high conductivity and low weight, however, it also has the disadvantages of relatively low flexibility and high cost, which could significantly depend upon the fabrication process. The conventional paper substrate shows high flexibility, low cost, and low weight, but it also shows low conductivity and relatively low surface area. The sponge type substrate has the advantages of high surface area and low cost, although it has drawbacks such as low electrical conductivity and high volume or low density. Furthermore, the textile substrate possesses merits of high flexibility, low cost, and high surface area, but properties such as relatively low conductivity are a problem for practical application to FSCs. The cable-type substrates have attracted interest recently because of their high flexibility, shape versatility, and high conductivity, though they also have some problems such as relatively high fabrication cost and low energy density. To solve these issues related to each substrate, several approaches have been reported by using some techniques, such as inducing a porous structure in the substrate, coating a conductive layer onto the substrate, thermal annealing for carbonization of cotton cellulose, etc. From this comprehensive review, it is ultimately realized that novel support structures such as metal substrates, carbon nanofoam/carbon paper, conventional paper, sponges, textiles, and cable-type substrates developed for high mass loading of active materials can offer great promise for developing large-scale energy storage systems in the near future.

Table 1. Comparison of various flexible substrates for the FSC electrodes.

FSC substrate

Cost

Conductivity

Flexibility

Surface area

Weight

metal

moderate

high

high

low

high

carbon-based paper

moderate

high

moderate

moderate

low

conventional paper

low

low

high

moderate

low

textile type

low

low

high

high

low

sponge type

low

low

high

high

low

cable type

moderate

high

high

moderate

low

Additionally, despite the merit of high power performance of SCs, the primary challenges for SCs are the limited energy density and the relatively high cost, which hinders them from being the energy storage platform for many applications. In this sense, to fabricate low-cost FSCs with high energy density, long cycle life, and high rate performance, the following ways could be considered with using the flexible substrates: 1) Designing novel hybrid-structured electrodes that combine the advantages of both earth-abundant EDLC capacitance materials and low-cost pseudocapacitance materials with use of cost-effective flexible substrates. To develop a cost-effective hybrid material system for FSCs electrodes, two key aspects could be considered: i) improvement of the interactions between EDLC materials and pseudocapacitive materials because the interactions could not only enhance the Faradic processes across the interface to achieve efficient pseudocapacitance, but also could sustain the strain caused by phase/volume change for long cycle life, and ii) more detailed study of the interfaces, such as the interface between electrolyte and pseudocapacitive materials and the interface within the hybrid between the EDLC materials and the pseudocapacitive materials. 2) Highly scalable, facile, solution-based processes and printing techniques will be a solution for making large-scale and low-cost FSC devices. 3) Asymmetric configuration of FSCs would also be an effective solution due to the potential for enhanced energy density, reduced costs, and safer cell chemistries compared to recent EDLCs. Also, such FSCs system can further bridge the energy/power performance gap between conventional batteries and EDLCs in a form that offers significant advantages in terms of cost and safety. 4) The effective strategy to increase the energy and power densities of FSCs is the use of electrolytes that can be operated at higher voltages to enhance the working voltage window, as presented in Equation (4).

In conclusion, SC technology, especially for FSC technology, will play an important role in energy storage and harvesting to help to reduce the environmental pollution problems and minimize the use of hydrocarbon fuels. Thus, the above-mentioned primary challenges show that there is still significant room for future development in FSCs. Future work will be expected to further demonstrate FSCs as one of the most powerful and highly efficient energy and power storage devices to meet the specific needs of future lightweight, flexible, and wearable electronics. Moreover, FSCs are likely to be integrated into smart clothing, sensors, and other wearable electronics. At some cases FSCs will replace batteries, but in many cases they will either supplement batteries, increasing their efficiency and lifetime, or serve as energy solutions when an extremely large number of cycles, long lifetime, and fast power delivery are required.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2013029776 (Mid-career Researcher Program)) and by the Global Frontier R&D Program (0420-20120126) at the Center for Multiscale Energy System through NRF. We also appreciate the financial support by the Core Technology Development Program from the Research Institute for Solar and Sustainable Energies (RISE/GIST). One of the authors (D. P. D.) appreciates the award of a Humboldt Fellowship of the Alexander von Humboldt Foundation (AvH), Germany.

Biographical Information

Dr. Deepak Dubal is currently working as “Alexander von Humboldt Fellow” at the Technische Universität Chemnitz, Germany. He joined the Gwangju Institute of Science and Technology (GIST), South Korea as postdoctoral fellow in 2011. Dr. Dubal received his M.Sc. in 2008 and then Ph.D. from Shivaji University, Kolhapur, India in 2011 in solid state physics, under the direction of Prof. C. D. Lokhande. His research interest is focused on the chemical synthesis of nanostructured materials and their application in energy storage devices, such as Li-ion batteries and supercapacitors.

original image

Biographical Information

Jong Guk Kim received his B.S. degree from Chonnam National University in 2008 and his M.S. degree from the School of Materials Science and Engineering at Gwangju Institute of Science and Technology (GIST) in 2010. Now, he is a Ph.D. student under the supervision of Prof. Won Bae Kim at GIST. His current research interests are focusing on synthesis and characterization of nanomaterials for energy storage devices.

original image

Biographical Information

Prof. Rudolf Holze received his Ph.D. in 1983 from University of Bonn, Germany. Afterwards, he joined the Case Center for Electrochemical Sciences, Case Western Reserve University, Cleveland, USA, as postdoctoral fellow and then Oldenburg University in 1987 as associate professor in Physical Chemistry. Since 1993, he is full professor of Physical Chemistry and Electrochemistry in the Chemistry department at the Technische Universität Chemnitz, Germany. His research is focused on spectroelectrochemistry, self-assembled monolayers, lithium-ion batteries, electrochemical materials science, and corrosion. He is member of the Saxonian Academy of Science and of several editorial boards. He is the author of more than 315 articles in refereed international journals and eleven books.

original image

Biographical Information

Prof. Won Bae Kim received his Ph.D. in 2002 from Pohang University of Science and Technology (POSTECH), South Korea. He joined University of Wisconsin-Madison, USA in 2003 as a post-doctoral associate, and then became an assistant professor in 2005 and an associate professor in 2008 in the School of Materials Science and Engineering at Gwangju Institute of Science and Technology (GIST). He is currently a full professor of Materials Science and Engineering and the Research Institute for Solar and Sustainable Energies at GIST, South Korea. His current research is focused on synthesis of nanomaterials and their application to the electrochemical energy transfer devices of fuel cells, batteries, supercapacitors, and photoelectrochemical cells.

original image