Nanostructured materials for supercapacitor applications

In 2003, Nobel Laureate Richard E. Smalley delineated that energy is the topmost problem faced by human society.¹  Increasing global population has caused raising demands for technological gears, especially in portable devices and automobiles. This has led to the gradual depletion of fossil fuels and caused increasing environmental issues. According to the report of International Renewable Energy Agency (IRENA), International Energy Agency (IEA) and the Renewable Energy Policy Network for the 21st century (REN21), about 29% of total final energy is consumed by transportion.²  Considering this, several automotive companies are focusing to develop electric vehicles and hybrid electric vehicles for which efficient energy storage devices are needed. Batteries and fuel cells are widely used energy storage devices providing high energy density; however, they lack the ability to provide higher power, safer operations, longer life cycle, and faster charge-discharge cycles.³ 

Recently, serious safety issues related to Li-ion batteries caused due to fires or explosions in devices such as cell phones, laptops, electric vehicles, and airplanes have sparked worldwide attention.⁴  Few examples include fires or explosions in Samsung Note 7, Tesla electric car battery fire, and the Boeing 787 Dreamliner battery issues, remind us about the most urgent issues to be resolved before being proposed in the practical application. Moreover, charging time required for batteries especially in electric vehicles is another big challenge. A recent article in The New York Times addresses this issue as “Charging Time Trauma” which could be a limiting factor for the people to switch over from gas vehicles to electric vehicles.⁵  Such issues can be addressed through nanomaterial research and solving the current drawbacks of traditional devices. Researchers have provided significant attention over developing supercapacitors type batteries to address previous issues and create a more reliable and safe source of energy.⁶  Supercapacitors provide high power density, longer cycle life, ultra-fast charge–discharge, and safer operation, leading to appreciable attention as a device of interest. Use of nanostructured materials have shown promising capabilities for fabricating supercapacitors with higher efficiency to solve growing energy problems.⁷  Till date, thousands of reports and patents covering aspects of the electrical double layer (EDL), reversible redox reaction, and hybrid mechanisms have been developed. This technology has flourished into a multi-billion dollar industry covering a wide range of applications such as telecommunication devices, smart electronic devices, industrial actuators, automobiles, aerospace appliances, power tools, and many more.⁸ 

Currently, supercapacitor research focuses on developing nanostructured material arrays providing flexibility, transparency, smaller-size, higher energy and power densities, and cost-effectiveness. This could be governed by considering basic electrode materials of capacitor devices summarized in later sections. However, before debating on this topic, it is essential to understand the general charge-storage mechanism of the supercapacitors.

The first generation capacitor was called a “Leyden Jar” that stored the charge electrostatically over the surface. Later, the concept was properly defined as EDL by Helmholtz. Based on this concept the first patent was filed for an electrochemical capacitor device by General Electric.⁹  After a few years, a new patent was granted on an electrochemical capacitor where a new concept of reversible redox reactions over the interface of electrodes was observed providing higher capacitance compared to the previous devices.¹⁰  The concept seems to combine behavior of high surface area carbon along with reversible redox materials of batteries. Although the device performance was high, a concept study of such mechanisms was not clearly defined based on its capacitance type. In 1971, an electrochemical capacitor based on RuO₂ was discovered storing the charge based on faradic reactions following the nature as a capacitor.^11,12  The concept was termed as pseudocapacitance and the device was termed as a pseudocapacitor. EDL capacitors provided high power density while pseudocapacitors delivered higher energy density.³  Hence, today's capacitor research tends to combine the concepts of EDLCs and pseudocapacitors to obtain a hybrid device with high energy and power densities.³  Thus, the capacitor's charge storage mechanism can be categorized as an electric double layer, pseudocapacitance, and hybrid mechanism and capacitor device is termed as an electric double layer capacitor, pseudocapacitor, and hybrid capacitor, respectively.

The charge storage behavior of a supercapacitor is analyzed based on its energy density and power densities. Energy density (E) and power density (P) can be defined as:¹³ 

Where C is the capacitance in Farad, V is the operating potential in volt, R is the equivalent series resistance in Ohm. In order to achieve high energy density, it is essential to increase the capacitance and operating potential range of a device. While power density can be improved by reducing the series resistance from the system and improving potential range, the capacitance of the capacitor device is dependent upon the type of electrode material used and its charge storage mechanism. EDL and pseudocapacitance are two basic charge storage mechanism used for supercapacitor device.³  While, a hybrid mechanism mostly involves a combination of EDLC type electrode and battery type electrode as cathodes and anodes, with calculative material loading to obtain high energy and power density of the final cell.^14,15  The cyclic voltammetry (CV) and galvanostatic charge–discharge (CD) tests are used to analyze the capacitance of the electrode materials.¹⁶  Specific capacitance (C_sp) can be calculated from the cyclic voltammetry using an expression given below:¹³ 

Where Q is the absolute quantity of charges in Coulomb, m is the mass of the electrode in grams and ΔV is the potential window in volt. Higher amount of charge particles provides better capacitance. The specific capacitance can be calculated using charge–discharge curves:¹⁷ 

Where I is the discharge current in Ampere and Δt is the discharge time in seconds. Longer discharge time suggests better capacitance of the material. The behaviour of CV and CD curves shows distinct characteristics in EDLC, pseudocapacitors, and hybrid capacitors. The detailed analysis of the charge storage mechanism for the distinct type of capacitor electrode materials focusing on nano-structured materials is described.

The capacitor assembly consists of two active electrodes, a separator which prevents short-circuit of these electrodes and electrolyte.¹³  As the capacitor is charged, ions within the electrolyte solution notice an ionic pull creating diffusion and accumulation of ions over the surface of the electrodes. This phenomenon of charge storage mechanism is called an electric double layer. EDL allows almost instantaneous charging and discharging behavior (about 10⁻⁸ s).¹⁸  Hence, EDLCs are capable of responding rapidly towards change in potential within the required time frame. Ultrafast charging and discharging cycles resulted due to charge accumulation without any chemical reactions allow EDLCs to provide high power density and longer life cycle. Their applications include power drills, heavy load vehicles, starting engines, power steering, quick acceleration, energy regeneration while braking, energy harvesting, and many more.¹³ 

The charge storage mechanism of EDL was first proposed by Helmholtz.¹⁹  Fig. 1a gives an insight to the Helmholtz model where two layers of the opposite charges are formed at the electrolyte and electrode interface. The accumulated ionic layer over the electrode–electrolyte interface had thickness H. Later, it was believed that due to the mobility of ions, the distribution of ions was larger than H, and was more diffused in electrolyte than on electrode surface illustrated schematically in Fig. 1b.^13,20,21  This arrangement is called the Gouy-Chapman model. The Helmholtz model considered charge accumulation near electrodes’ interface, while the Gouy-Chapman model proposed charge distribution more diffused towards electrolyte. Considering both possibilities, Stern combined both models and designed the Gouy-Chapman–Stern model (Fig. 1c). The charge regions were accounted as a stern layer and a diffusion layer.^13,22,23  This model was widely accepted for EDLC.

Electrochemical testing provides the detailed correlation between chemical and electrical effects for a supercapacitor device.²⁵  Fig. 2 provides a brief schematic for CV and CD curves of an ideal EDLC device. In general, EDL based capacitors show CV curves close to a rectangular shape with constant current and CD curves with linear and symmetrical charging–discharging cycle.¹⁶  However, during practice, current in CV curves shows little variation and CD curves display deviation from linearity.

Carbon nanotubes (CNTs) are one of the important carbon-based materials having sp² hybridized wrapped tubular graphene sheet-like structure. Fig. 3 shows the atomic configuration of different types of carbon nanotubes.²⁶  CNTs can be synthesized using arc discharge, laser ablation or chemical vapor deposition methods.²⁷  Depending upon layers of tubular sheets, CNTs can be a single-walled carbon nanotube (SWCNT), a double-walled carbon nanotube (DWCNT) or a multi-walled carbon nanotube (MWCNT). The diameter of SWCNT and MWCNT range between 0.8–2 nm and 5–20 nm (sometimes more than 100 nm), respectively.²⁸  Length of CNTs could be from less than 100 nm to several centimeters, making their scalability for the molecular and microscopic level. Moreover, excellent mechanical (tensile strength ∼150 GPa), thermal (thermal conductivity ∼3000 W m⁻¹ K), electronic properties (electrical conductivity ∼10⁷ S m⁻¹) and high surface area (20–500 m² g⁻¹) of CNTs make them useful for wide range of applications such as in supercapacitors, batteries, automotive parts, sensors, hydrogen storage, and nanometer-sized semiconductor devices.^27,29–33  Additionally, CNTs have higher charge transport capability, high mesoporosity, and high electrolyte accessibility making them favorable electrode materials for EDLCs.

The International Union of Pure and Applied Chemistry (IUPAC) classifies pore size as micropores (<2 nm), mesopores (2–50 nm), and macropores (>50 nm).³⁴  The study suggests that ions within the aqueous electrolytes are in hydrated form within the size range of 6–7.6 nm.^35,36  Thus, mesopores with a size range of 30–50 nm could provide maximum capacitance for EDLCs. An et al. synthesized SWCNTs using the arc discharge technique.³⁵  Heat treatment at various temperature resulted in SWCNTs with different surface areas. As heat treatment temperature was raised from 500 to 1000 °C, the specific surface area increased to 357 m² g⁻¹ and the average pore diameter reduced from about 70 nm to 30 nm. This resulted in higher capacitance close to the estimated theoretical capacitance (71–178 F g⁻¹).

Along with higher surface area, it is essential to have appropriate structure, composition, and diameter of CNTs to allow excellent charge storage properties. Frackowiak et al. fabricated a supercapacitor using MWCNTs to study microtexture and elemental composition.³⁷  Surface area of the nanotubes ranged from 130 to 410 m² g⁻¹. Lowest capacitance value was observed for CNT synthesized using the NaY–zeolite catalyst, which produced CNTs with closed tips causing limited mesopore volume and lower active surface area. The CNTs with numerous edge planes resembling nanofilament morphology synthesized using Co/Si catalyst followed by acid treatment showed the most efficient charge storage behavior. Post-treatment improved the surface area from 410 to 475 m² g⁻¹ and added surface groups improving charge storage properties. However, CNTs synthesized using chemical vapor deposition of propylene on the alumina template with high surface area showed moderate results due to hydrophobic rigid structure and larger canal diameter. It was concluded that post-treatments could further improve the capacitive properties of CNTs.

As described in Section 2.1, energy and power densities are greatly affected by the working potential of the electrolyte used. Organic and ionic electrolytes allow broader working potential compared to aqueous, thereby improving the characteristic supercapacitor behavior. Gruner and co-workers fabricated thin film supercapacitors using SWCNTs as an active electrode material and charge collector.³⁸  Supercapacitor performance was studied by using aqueous gel (PVA/H₃PO₄) and organic electrolyte (LiPF₆/EC: DEC). Both electrolytes showed energy density of 6 Wh kg⁻¹, while the power density of 23 and 70 kW kg⁻¹ for an aqueous gel and an organic electrolyte, respectively. Another flexible supercapacitor device using CNTs coated on paper as an electrode and ionic liquid gel electrolyte was reported.³⁹  The energy density of 41 Wh kg⁻¹ and power density of 164 kW kg⁻¹ was observed. Ali et al. fabricated supercapacitor electrode using SWCNTs and 1M Et₄NBF₄/propylene carbonate as an electrolyte.⁴⁰  Comparing the ability for operating potential range using CV test for activated carbon electrodes and SWCNT electrodes, SWCNT electrodes showed identical symmetric CV curves, while activated carbon electrodes showed parasitic chemical reaction at a higher potential with strong peak current. Hence, by utilizing the full potential of 4 V, significantly high energy (94 Wh kg⁻¹) and power densities (210 kW kg⁻¹) were achieved surpassing most of the previous reports. Although electrolytes provide a wide operating potential range, however, the electrochemical properties depend upon material capabilities to comply with supporting and performing well within that potential range. Thus, a major research focuses on improving properties of CNTs by a heteroatom doping or forming composites with pseudocapacitive materials which could improve the capacitance.^27,41,42 

Graphene is a 2-dimensional sheet of an sp²-hybridized polycyclic aromatic hydrocarbon of quasi-infinite size.^3,43  It can be produced in two ways: top down and bottom up techniques. The top-down approach consists of mechanical, wet chemical, and electrochemical exfoliation of mono-layer frames from graphite.^43–46  Graphite is hexagonally arranged layers of a carbon allotrope in a planer condensed ring system.⁴³  First graphene sheets were produced by mechanically exfoliating monoatomic sheets of few-layer graphene from a graphite crystal.⁴⁷  Another way to obtain graphene is by unzipping CNTs as shown in Fig. 4.^48–50  The bottom-up approach consists of chemical vapor deposition growth of graphene sheets over metal substrates such as Ni and Cu, using hydrocarbon precursor such as methane or ethylene at high temperature.^44,51–53  Suspended single-layer graphene possesses significantly higher charge mobility of ∼230 000 cm² V s⁻¹, compared to few-layer graphene (10 000 cm² V s⁻¹) and multi-layer graphene sheets (15 000 cm² V s⁻¹ at 300 K and 60 000 cm² V s⁻¹ at 4 K).^47,54  Presence of π-electrons in graphene results in overlap of conduction and valence band making zero band gap system.⁵⁵  Thus, it is also called as Dirac solid. Monolayer graphene exhibits extremely high tensile strength and Young's modulus of ∼130 GPa and ∼1 TPa, respectively.⁵⁶  Moreover, single layer graphene exhibits higher thermal conductivity ∼4840–5300 W mK⁻¹ compared to CNTs.⁵⁷  These excellent properties along with the high specific surface area of 2630 m² g⁻¹ make graphene an ideal material for the EDLCs.⁵⁸  The intrinsic capacitance of graphene is 21 μF cm⁻², considering this area, the theoretical capacitance of graphene is estimated as 21 μF cm⁻²×2630 m² g⁻¹=550 F g⁻¹.^58,59  However, aromatic structured sheets of graphene, exhibits van der Waal interactions causing aggregated stacks, thereby preventing dispersion and wide range applications. Hence, chemical exfoliation of graphene using Hummer's method (strong acids) to create epoxy, carboxylic, and hydroxyl groups on the surface of graphene sheets makes them hydrophilic, preventing their aggregation and forming a soluble dispersion in water to allow easy processing.³  These sheets are termed as graphene oxide sheets. Although a large number of graphene sheets can be obtained using this method,⁶⁰  the presence of oxygen-containing functional groups produced by replacing double-bonded conjugation, significantly reduces the conductivity. Hence, in order to regain its conductivity, graphene oxide is further reduced using different methods such as thermal treatment, microwave irradiation, chemical reduction, controlled laser reduction.^61–64  For example, Guex et al. reported an experimental review over the chemical reduction of graphene oxide to form reduced graphene oxide (rGO) and observed significant improvement in the conductivity from 0.0003 S m⁻¹ (GO) to 1500 S m⁻¹ (rGO).⁶³  Much work has been reported on the use of graphene-based materials for supercapacitor applications.

Graphene can be fabricated into fibers, films and/or foam structure depending on applications.⁶⁵  Meng et al. fabricated all-graphene core-sheath microfiber based supercapacitor and obtained a capacitance of 40 F g⁻¹.⁶⁶  This flexible fiber capacitor showed stable performance after 500 cycles of straight to bending tests and maintained the performance even after incorporation into textile using conventional weaving method. Flexible pillared graphene paper (GP) electrodes were fabricated by Wang et al. using a simple vacuum filtration method.⁶⁷  The pillared GP showed energy density of 26 Wh kg⁻¹ using an organic electrolyte (1M LiPF₆ in EC) providing a potential window of 3 V. Moreover, only 4.35% loss was observed in capacitance after 2000 cycles. Yang et al. synthesized highly porous electrode for efficient energy storage application.⁶⁸  The synthesis was carried out using exfoliation and reduction of GO using hydrogen. Hydrogen reduces the oxygen-containing functional groups in GO to water vapor and CO₂, generating high pressure which breaks the agglomeration in GO and converts into hydrogen annealed graphene (HAG). The supercapacitors using HAG were studied using ionic liquids EMIMBF₄ and LiPF₆ as electrolytes which provided a potential range of 4 and 3 V, respectively. HAG based supercapacitor showed ultra-high energy density of 148.75 Wh kg⁻¹ and power density of 30.95 kW kg⁻¹. Furthermore, the device showed stable performance over 7000 of cyclic testing. Although graphene provides an excellent option for EDLCs, several factors such as the source of raw material, cost of synthesis, cost of supplementary processes such as reduction of GO and complexity in synthesis, limits the wide-range applications of graphene as well as CNTs. Hazardous chemicals such as hydrazine used for reduction of graphene are challenging to be used in actual practice and finding other cost-effective alternative methods for graphene reduction for industrial applications in supercapacitors is much needed.

Activated carbons are widely used as electrode materials in EDLCs due to their high surface area, good electrical properties, and relatively cheap fabrication cost. Growing environmental issues and depleting petroleum reserves have caused the urgent need for alternative resources for energy storage devices. Major research activities for activated carbon focus on bio-based alternates that could result in highly porous carbon providing higher charge storage capacity.^69–72  A comprehensive review presented by Saidur et al. shows major issues related to waste management of bio-wastes obtained from woody crops and agricultural crops such as rice husk, corn waste, palm waste, sugarcane, wooden chips, bamboo fibres, cotton glen, sunflower shell, almond shell, walnut shell, wheat straw and papers, which could be a source of alternate fuels.⁷³  Fortunately, these carbonaceous materials could also serve as bright source for developing electrode materials for supercapacitors.⁷⁴  The wastes obtained from bio-sources are pre-carbonized followed by activation to create high surface area carbon called activated carbon. Several factors such as the source of carbon, type of activation, graphitic/diamond structure within carbon, pore size distribution, and elemental doping play an important role in obtaining carbon which can provide high charge storage capacity.⁷⁴  Activation of carbon can be carried out using the physical, chemical, physicochemical, and microwave-induced techniques.

Physical activation involves partial etching of carbon during the carbonization process using oxidizing agents mostly CO₂ or steam or a mixture of both.⁷⁵  In a typical process, precursors are pre-carbonized at 350–400 °C followed by final carbonization along with activating gasses at a temperature range of 600–900 °C. Following reactions occur during physical activation using CO₂ and H₂O (Steam):⁷⁵ 

Steam activation of coconut shell at a water flow rate of 0.12 mL min⁻¹, the temperature of 800 °C, and 60 min of activation time produced carbon with a high surface area of 1532 m² g⁻¹.⁷⁶  The study showed improved mesopores (∼3.9 nm) due to activation can also improve the ion-transport and thus charge storage capacity. The device showed a high specific capacitance of 192 F g⁻¹ and an energy density of 38.5 Wh kg⁻¹ at a current density of 1 A g⁻¹. Qu et al. proposed a direct one-step steam activation of corncob residue to achieve carbon with a high surface area (1043 m² g⁻¹) which was similar to the surface area of carbon (1018 m² g⁻¹) synthesized in a two-step process: pre-activation carbonization and activation after pre-carbonization.⁷⁷  One step activated carbon showed a high specific capacitance of 314 F g⁻¹ at a scan rate of 5 mV s⁻¹. Moreover, no capacitance decay was observed even after testing the electrode for over 100 000 cycles of charge–discharge study. Kumagai et al. synthesized micro and mesoporous carbon derived from a mixture of rice husk and beet sugar using CO₂ activation.⁷⁸  Longer activation time from 30 min to 60 min, decreased final yield from 46.7% to 27.7% but improved the surface area from 1103 to 1357 m² g⁻¹. The reactions involving physical activation of carbon using CO₂ and steam are endothermic in nature and thus require long activation time and result in lower yield.⁷⁵ 

Chemical activation involves one-pot activation of carbon using pre-carbonized carbon precursor and reaction with chemical agents such as KOH, NaOH, H₂SO₄ and ZnCl₂ at elevated temperature under inert atmosphere.^75,79  Compared to physical activation, chemical activation provides uniform pore size distribution and higher yield. KOH is among the widely used activating agents which creates porous structure within the carbon using following reaction:⁸⁰ 

Activated carbon produced from waste tea leaves using KOH activation showed significant improvement in the surface area form 3.55 m² g⁻¹ (unactivated) to 2532 m² g⁻¹ (activated).⁶⁹  Electrochemical properties of the high surface area carbon were studied in different aqueous electrolytes (3M KOH, NaOH, and LiOH) to understand the effect of electrolyte ions size on the charge storage capacity of tea leaves-derived carbon. The highest specific capacitance of 292 F g⁻¹ was obtained in 3M KOH electrolyte and the lowest specific capacitance of 246 F g⁻¹ was observed in 3M LiOH electrolyte. Similarly, electrochemical properties of activated carbons from bamboo fiber, jute fiber, corn straws, waste coffee, orange peel, and paper were studied for their applications as electrode materials in supercapacitors.^{69–72,80–82}  In the conventional heating method of activation, the thermal gradient from a hot source to internal carbon bed cause distortion and non-homogenous heating resulting in a longer process time and wastage of energy. Hence, the microwave assisted activation process was introduced, where thermal gradient is in opposite manner leading to shorter activation period and improved energy efficiency.⁷⁵  Ramasahayam et al. synthesized activated carbon from waste coffee grounds using microwave assisted activation in just 30 min.⁸³  Resulting carbon exhibited a high surface area of 999.64 m² g⁻¹ and a capacitance of 286 F g⁻¹ at 5 mV s⁻¹.

Other nanostructured carbons such as activated carbon nanofiber, carbon aerogel, onion-shaped carbon, and templated carbon have been studied. However, they suffer the disadvantage of higher process cost and complex preparation techniques.^26,84–87  Activated carbon-based materials show limited capacitance performance which could be further enhanced by doping different redox entities such as nitrogen, sulfur, and phosphorus to improve capacitive performance and can be studied from other reports.^83,88–91 

Pseudocapacitance is a charge storage mechanism which consists of fast and reversible faradic-redox reactions over or near the electrode-electrolyte interface at relatively slower speed (typically 10⁻²–10⁻⁴ s).^3,12,18  Pseudocapacitive materials could be intrinsic or extrinsic in nature depending upon materials’ size and structure.¹²  Intrinsic behaviour results when materials show pseudocapacitive charge storage characteristics over a broad range of particle size and structural morphology. While extrinsic behavior results for the materials which do not exhibit pseudocapacitive charge storage mechanism in the bulky state due to phase transformations resulted during storage of ions. The capacitance resulted due to faradic reactions in pseudocapacitive materials could be 10–100 times higher than EDL type materials providing higher energy density.⁸⁵  While slower reaction speeds result in inferior power density. Unlike EDLCs, the redox reactions in pseudocapacitors cause electrode materials to expand and contract while charging and discharging, resulting in the limited life cycle of the device.

The CV and CD curves in Fig. 5 provide the generalized behavior observed for the pseudocapacitive materials. The faradic reactions involved in charge storage of pseudocapacitive materials show a peak for the oxidation and reduction during CV test, along with EDL response.^92–94  Similarly, CD curves exhibit slight distortion within linear charge–discharge profiles corresponding to redox reactions linked during that particular transition.⁹⁵  In batteries, such behaviors are longer (non-linear discharge potential) and result in higher energy densities, however, in pseudocapacitors, such behavior is considerably suppressed.⁶  As discussed in previous sections, the high specific surface area is an essential factor for enhancing supercapacitor performance. Although porous carbon materials provide a high surface area, good conductivity, and easy accessibility for the electrolyte, they suffer from poor capacitance resulting in poor energy density. Hence, researchers have focused towards nanostructured pseudocapacitive materials such as transition metal oxides/hydroxides, spinals, and conducting polymers to achieve higher capacitance and therefore higher energy density, while maintaining considerable power density.⁹⁶  Also, due to higher capacitance, pseudocapacitors could result in 3–7 times smaller device size.

Metal oxides/hydroxides and conductive polymers are widely used as pseudocapacitive materials. Series of metal oxides/hydroxides such as RuO₂, MnO₂, NiO/Ni(OH)₂, Co₃O₄/Co(OH)₂, Fe₂O₃/Fe₃O₄, Cu₂O/CuO, V₂O₅, and SnO₂ are used as electrode materials for supercapacitors. On the other hand, polyaniline, polypyrrole, and polythiophene are well-known conductive polymers for supercapacitor electrodes. Although metal oxides and conductive polymers show higher theoretical charge storage values, practically, their capacitance values are significantly low and depend upon several factors such as phase crystallinity, crystal structure, surface area, morphology, conductivity, and mass loading over the electrode substrates.^96,97  Nanostructured metal oxide/hydroxide materials are extensively studied due to their unique morphologies and properties that can yield a high specific capacitance closer to their theoretical values. Morphology of nanostructured material dependents upon the synthesis approach used. Several physical and chemical synthesis techniques are used to develop nanostructured metal oxides and conductive polymers.^3,98–106  Some of the widely used techniques for metal oxides are hydrothermal, solvothermal, electrodeposition, electrophoretic deposition, a sol–gel, direct chemical precipitation, template-assisted synthesis and plasma-assisted synthesis.^98,107  While, conductive polymers are synthesized using in situ polymerization, electro-polymerization, interfacial polymerization and photo-polymerization.^3,17 

The first-generation studied electrode material found to possess pseudocapacitance involving faradic-charge transfer reactions was RuO₂.¹²  The charge storage mechanism for hydrated RuO₂ is shown below:¹⁰⁸ 

Based on this equation, the maximum theoretical capacitance for RuO₂ and RuO₂·0.5H₂O can be estimated to be 1450 F g⁻¹ and 1360 F g⁻¹, respectively. Several RuO₂ nano-structured morphologies such as nanotubular structures, nanorods, nanoflowers, nanotubes, and nanosheet structures are studied.^109–117  For better charge storage performance, the insertion and extraction of protons within the structure of RuO₂ play a vital role. Hydrous RuO₂ provides improve proton conductivity and higher capacitance (850 F g⁻¹) compared to anhydrous structure (125 F g⁻¹).¹⁰⁸  Surface area is another key factor for exhibiting higher capacitance. Mesoporous RuO₂ with a high surface area (140 m² g⁻¹) showed a capacitance of 202 F g⁻¹ compared to the low surface area (39 m² g⁻¹) with 146 F g⁻¹.¹¹¹  Higher surface area allows more active sites for redox reactions and thus improves charge storage capacity. 3D nanoflowers of RuO₂ with interconnected spheres sized 250–300 nm showed high specific capacitance of 545 F g⁻¹ at 0.5 A g⁻¹ (Fig. 6).¹¹⁷  Hierarchical porous structure allows greater access of electrolyte to the active electrode surface and thereby providing excellent rate capability with only 8.6% decrease with increasing current density from 0.5 to 50 A g⁻¹. Similarly, a nanotubular array of RuO₂ nH₂O with a uniform wall thickness of ∼40 nm show extremely high specific capacitance of 1300 F g⁻¹ at 10 mV s⁻¹.¹¹⁴  Ordered 3D porous architecture allows favorable penetration of electrolyte reducing proton diffusion distance and providing high energy and power density of 7.5 Wh kg⁻¹ and 4320 kW kg⁻¹, respectively. Although RuO₂ shows promising performance, several factors such as cost and lack of availability limit its use for commercial applications. Hence, earth-abundant materials such as transition metal oxides based materials are considered as a cheap and effective alternative to RuO₂.

Manganese oxide (MnO₂) provides a cost-effective and less toxic alternative to rare-earth ruthenium oxide. MnO₂ exhibits a high theoretical specific capacitance of 1370 F g⁻¹.^96,97  The crystal structures (such as α-, β-, γ-, δ- and λ-) of MnO₂ possess different tunnel structures which affect the cation intercalation and thereby, resulting pseudocapacitive performance.^98,118  The charge storage mechanism of MnO₂ can be summarized as follows:¹¹⁹ 

Temperature and time during material synthesis significantly affect the final morphology. Nanostructured MnO₂ was synthesized by decomposition of KMnO₄ in HCl solution using the rapid microwave-assisted hydrothermal technique.¹²⁰  Lower synthesis temperature of 140 °C resulted in cauliflower-like δ-MnO₂ particles with the higher specific surface area (115–201 m² g⁻¹) exhibiting a higher capacitance of 176–202 F g⁻¹. While higher synthesis temperature leads to the formation of α-MnO₂ particles with the low specific surface area (25–33 m² g⁻¹) resulting in a lower capacitance of 46–61 F g⁻¹. The morphology of MnO₂ can also be adjusted by synthesis time and annealing process. Hierarchical tubular MnO₂ nanotubes were synthesized through a hydrothermal process using a sacrificial carbon nanofibers based template.¹²¹  By adjusting reaction time and annealing process, polycrystalline MnO₂ nanotubes were transformed to sponge-like nanowires and porous nanobelt morphology. Synthesis with increasing reaction time increased the diameter of tubular nanotubes (10, 45 and 180 min resulted in a diameter of ∼200 nm, ∼350 nm, and ∼470 nm, respectively). Fig. 7 shows the effect of reaction time on diameter, porous structure and surface area of nanotubes (time of 10, 45 and 180 min resulted in a surface area of 40 m² g⁻¹, 70.1 m² g⁻¹, and 90.1 m² g⁻¹, respectively). Highest specific capacitance was observed for MnO₂ nanotubes synthesized for 180 min, showing capacitance of 461 F g⁻¹ at 5 mV s⁻¹. Although several nanostructures of MnO₂ were studied, the capacitance for unmodified MnO₂ showed inferior practical results compared to RuO₂ limiting its wide-range applications.¹²² 

Nickel oxide and nickel hydroxide are other promising pseudocapacitive materials because of their cheap availability, thermo-chemical stability, environmental friendliness, and high theoretical capacitance (2082–2584 F g⁻¹).^123–129  NiO and Ni(OH)₂ undergoes the following set of redox reactions:^123,130,131 

Several nanostructures of NiO/NiOH such as porous films, hollow nanospheres, nanobelt nanoplatelets, nanorod arrays, nanowires, nanoflakes, nanosheets, nanowhiskers, flower-like microspheres, 3D dendrites, and nanoflower were explored to exhibiting higher surface area for faradic reactions and high specific capacitance.^{125,130–153}  Oswald ripening mechanism, a well-known phenomenon causing the growth of small crystals through redeposition of dissolved species over the surface, is widely used to develop nanostructured metal oxides.¹⁵⁴  One of the example for NiO and Ni(OH)₂ consists reaction of nickel nitrate with urea to precipitate Ni(OH)₂ followed by dehydration using high-temperature calcination to form NiO.¹²⁹  A set of reactions is summarized as follows:¹⁴⁴ 

Cobalt oxide/hydroxide is one of the favored pseudocapacitive materials due to their higher theoretical capacitance of 3560–3700 F g⁻¹, cyclic performance, and corrosion resistance.^96–98  The charge storage mechanism and nanostructure synthesis are similar to NiO/Ni(OH)₂. Also, Co₃O₄/Co(OH)₂ can be designed in several sophisticated nanostructures such as thin films, nanowire arrays, nanotubes, nanosheets, nanoflowers, nanoflakes, ordered mesoporous structure, hollow nanowires, and monolayer hollow-sphere arrays.^156–173  Synthesis time and temperature play an important role in obtaining nanostructures of Co₃O₄. For example, Yang et al. synthesized Co₃O₄ using a hydrothermal technique where one reaction was carried out at 100 °C for 6 h, another for 100 °C for 9 h, and the third one at 120 °C for 9 h producing nanosheet arrays, nanosheet@nanowire arrays, and nanowire arrays, respectively (Fig. 9).¹⁶⁵ 

As observed in Fig. 10, the porous free-standing nanowires over nanosheets yielded the highest specific capacitance of 715 F g⁻¹ and maintained 100% retention for more than 1000 cycles. The conductive pathways, hierarchical porous channels, and large surface area play an important role for better charge storage performance. Novel bottom-up design for synthesizing ultrathin mesoporous Co₃O₄ nanosheet arrays using a facile electrodeposition of Co(OH)₂ over Ni-foam and converting it to mesoporous nanosheets of Co₃O₄ using calcination resulted in a high specific surface area of 118 m² g⁻¹.¹⁵⁶  Crystalline nanoparticles with a size of ∼5 nm and mesopores of 2–5 nm resulted in significantly higher specific capacitance of 2735 F g⁻¹ at a current density of 2 A g⁻¹, claiming one of the highest reported value for the Co₃O₄ system. Similar to Ni(OH)₂, Co(OH)₂ also exists in α- and β- phase.^158,172  α-Co(OH)₂ is amorphous and unstable disordered state. Due to its unstable nature α-Co(OH)₂ easily gets converted to more stable, compact, and regularly ordered hexagonal lattice-shaped β-Co(OH)₂ in alkaline electrolyte during electrochemical testing.¹⁷⁴  Several reports suggest that high capacitance values for Co(OH)₂ are due to its high theoretic capacitance.^158,172,173  3D porous nanoflake composite film of Co(OH)₂ was electrodeposited over Ni foam.¹⁵⁹  The electrochemical measurements showed a high specific capacitance of 2028 F g⁻¹ at a low current density of 2A g⁻¹ and maintained 1920 F g⁻¹ even at a higher current density of 40 A g⁻¹. Co(OH)₂ nanoflakes showed impressively high energy and power densities of 80 Wh kg⁻¹ and 11 kW Kg⁻¹, respectively. Another report over electrodeposited Co(OH)₂ in mesoporous film structure showed high specific capacitance of 2646 and 2274 F g⁻¹ at current densities of 8 and 48 A g⁻¹, respectively.¹⁶⁰  Although, capacitance results for synthesized Co(OH)₂ was reported to be high, the working potential of the material was around 0.5–0.6 V. This could create limited practical applications of metal oxide/hydroxide-based systems.

Scheme of the possible formation process of the Co₃O₄ hierarchically structure. Reproduced from ref. 165 with permission from the Royal Society of Chemistry.

Apart from previously mentioned metal oxides, Fe₂O₃, Fe₃O₄, V₂O₅, Bi₂O₃, SnO₂, WO₃·2H₂O and Ti oxide-based systems are also studied for energy storage applications. However, individual metal oxides suffer from several limitations and yield inferior energy storage capacity.^95,175–183  Therefore, spinal-based metal oxides have been considered as advanced materials for energy applications due to the presence of two or more different types of metal ions which can participate in the redox process.

Binary metal spinals are metal composites with AB₂O₄ composition (A and B corresponds to different metal ions), which possessing high electrical conductivity compared to individual metal oxides and exhibit properties contributed by both metals ions.¹⁸⁴  As described in previous sections, the charge storage capacity of metal oxides could be significantly improved by increasing redox active surface area. Imparting nanostructures to transition metal oxides, significantly improved redox active surface area and thus higher capacitance could be obtained. Several spinal metal oxides such as MnCo₂O₄, CoMn₂O₄, MnNi₂O₄, NiCo₂O₄, NiMn₂O_4, and ZnCo₂O₄ are studied for energy applications.^185–201  Compared to individual manganese oxide, its spinal composite with metals ion of Ni or Co showed improved performance. Porous nanowires of MnCo₂O₄ (surface area of 106.6 m² g⁻¹) and CoMn₂O₄ (surface area of 112.7 m² g⁻¹) showed high specific capacitance of 2108 and 1342 F g⁻¹ at a current density of 1 A g⁻¹, respectively.¹⁹⁹  A study on electrochemical impedance spectroscopy suggests that CoMn₂O₄ nanowires showed ease in electrolyte diffusion and higher capacitor response compared to MnCo₂O₄ nanowires. Spinals can be also grown in various morphologies such as nanoneedles, yolk-shelled nanospheres, nanosheets, nanowires, and nanoflowers which could affect their energy storage capacity.^{187,188,193,194,196–198}  For example, nickel–cobalt hydroxide nanosheets were electrodeposited on NiCo₂O₄ nanowires over carbon fiber paper (CFP) using a hydrothermal process to obtain a 3D hybrid nanostructure (Fig. 11) which exhibit high energy and power density of ∼33 Wh kg⁻¹ and ∼41.25 kW kg⁻¹, respectively.¹⁸⁷  Due to the conductive nature of NiCo₂O₄, both Ni–Co hydroxide sheets and NiCo₂O₄ nanowires contribute to charge storage process. Another novel 3D flower-like hierarchitectures of NiCo₂O₄ can be observed in Fig. 12.¹⁸⁵  The structure showed a high specific surface area of 212.6 m² g⁻¹ exhibiting a high specific capacitance of 1191 F g⁻¹ at a current density of 1 A g⁻¹. A synergistic effect of porous structure and small diffusion lengths of nanosheet building blocks resulted in higher electrochemical performance. Moreover, the interconnected conductive structure of spinals supports each other to alleviate the structural damage caused by volume expansion during charge–discharge and results in excellent stability compared to individual metal oxides observed in previous studies.^191,193,194 

Conductive polymers (CPs) provide pseudocapacitance via fast reversible-redox reactions related to π-conjugated polymer chains.²⁰²  They offer several advantages such as flexibility, higher conductivity, capable of being formed into thin films, and easy to process.²⁰³  Polyaniline, polypyrrole, and polythiophene are widely studied conductive polymers for supercapacitor applications. However, compared to metal oxides conducting polymers show limited nanostructural morphology (such as porous nanoparticle films, nanowires, nanotubes, and nanofibers) and poor capacitive performance.^204–210  A typical mechanism for conjugated polymers is given as:¹⁵ 

Where, P_m is a conjugated polymer, A⁻ is anion and M⁺ is a cation.

Both metal oxides and conductive polymers exhibit the serious disadvantage of poor cyclability due to volumetric expansion and contraction. To overcome the structural instability in these compounds, nanostructural composites of metal oxides or conductive polymers with some stable carbon materials such as carbon nanotubes, graphene, carbon dots and activated carbon are preferred.³ 

In the case of metal oxides, faradic processes involving reversible surface absorption of protons from the electrolyte and redox reactions via ion exchange from electrolyte over electrode surface are primarily surface reactions.²¹¹  Such reactions take place within a potential window and actively participating specific surface area of the electrode. Hence, 5–10% contribution of EDL capacitance could be observed in typical pseudocapacitors.¹⁵  In case of metal oxides such as RuO₂, MnO₂, NiO, Ni(OH)₂, Co₃O₄ and, Co(OH)₂, nanostructured materials exhibiting higher specific surface area, smaller size, and higher surface energy have a significant effect over improved capacitance compared to bulk materials.^3,15,211  Although nanostructured materials exhibit better performance, its coagulation leads limited surface exposure affecting the cell performance. From the previous section, it can be understood that singular metal oxides have shown inferior performance compared to their theoretical abilities. Thus, it is highly desired that active materials are well dispersed with rigid adherence to conductive support that would significantly improve the supercapacitor performance. Considering this, several nanocomposites of metal oxides with graphene, mesoporous or activated carbon and CNTs have been studies.^14,211–232  Introducing composite structure with high surface area carbon materials, improved the specific capacitance and cyclic stability of metal oxides.

Unlike metal oxides, the charge storage mechanism of conductive polymers is induced by reversible oxidation and reduction of conjugated bonds. However, percentage doping, its mechanism, and stability of the redox switch affect the pseudocapacitance.²⁰³  Moreover, anion exchange between electrolyte and polymer adversely affect the charge storage density, reaction switch speed, and its overall cyclic stability. Thus, conductive polymer-carbon composites are synthesized to overcome these issues. Functionalized carbon materials such as graphene oxide and functionalized CNTs allows growth of conducting polymers forming advantageous composite structure allowing self-supporting, free-standing, and flexible electrodes to be used for supercapacitor application. Moreover, conductive carbon within the composite structure eliminates the use of conventional conducting additives or metal foils used as current collectors, which could significantly reduce the weight of the final device. Conventional capacitors are heavy, rigid, and show limited use for portable electronic devices.^203,204  Hence, use of conducting polymers-carbon composite could open the possibilities of future electronic devices such as electronic textiles. Detailed studies over the conductive polymer and metal oxides composed with graphene and other carbon-based hybrids can be understood from other reports.^{3,225,233,234} 

Dr Ram K. Gupta expresses his sincere acknowledgment to the Polymer Chemistry Program and Kansas Polymer Research Center, Pittsburg State University for providing financial and research support.