Addressing the Achilles' heel of pseudocapacitive materials: Long‐term stability

Funding information Merced nAnomaterials Center for Energy and Sensing (MACES), Grant/Award Number: NNX15AQ01 Abstract Electrode materials with high energy densities and long-lasting performances are crucial to durable and reliable electrochemical energy storage devices for modern information technologies (eg, Internet of things). In terms of supercapacitors, their low energy densities could be enhanced by using pseudocapacitive electrodes, but meanwhile, their ultralong lifetimes are compromised by the limited chargedischarge cycling stabilities of pseudocapacitive materials. This review article discusses on the cycling instability issues of five common pseudocapacitive materials: conjugated polymers (or conducting polymers), metal oxides, metal nitrides, metal carbides, and metal sulfides. Specifically, the article includes the fundamentals of the failure modes of these materials, as well as thoroughly surveys the design rationales and technical details of the cycling-stability-boosting tactics for pseudocapacitive materials that reported in the literature. Additionally, promising opportunities, future challenges, and possible solutions associated with pseudocapacitive materials are discussed.


| INTRODUCTION
Electrical energy is of paramount importance in the Information Age. Energy storage devices that store, manage, and distribute electrical energy are indispensable for the development of contemporary information technologies, such as smart electronics, the Internet of things (IoTs), and the Internet of everything. 1,2 Among various energy storage devices, supercapacitors, also known as ultracapacitors or electrochemical capacitors, stand out because of their ultrahigh power density and superior durability. Generally, supercapacitors' power densities are~100 times higher than those of conventional batteries, and supercapacitor devices can last millions of charge-discharge cycles without noticeable capacitance fade. 1,3 Driven by the need to boost the energy densities of supercapacitors, a diverse array of pseudocapacitive electrode materials have been investigated to replace conventional carbon-based electrical double layer capacitive materials, for example, activated carbon. Activated carbon exhibited limited capacitance, particularly at ultrafast charging and discharging rates, due to the dominance of micropores that are poorly ion accessible. 4,5 Pseudocapacitive electrodes, including conducting polymers (or more rigorously speaking, conjugated polymers), 6,7 metal oxides, 8 metal sulfides, 9,10 metal nitrides, 11 and metal carbides, 12,13 are promising materials that could simultaneously achieve high energy density and high power density. 14 The charge-storage mechanisms of pseudocapacitive materials are associated with surface-controlled (kinetically fast), faradaic redox reactions. [14][15][16] There are three types of pseudocapacitive reactions ( Figure 1): underpotential deposition, surface redox reactions, and fastrate ion intercalation or insertion. Underpotential deposition refers to electrodepositions occurring at potentials less negative than the Nernstian, or thermodynamically equilibrium, potentials (eg, Pb deposition on Au, Figure 1A). The underpotential deposition of hydrogen gas from hydrogen atoms adsorbed on the Pt surface led to the concept of pseudocapacitance in 1962. 17 In the catalysis field, these reactions could evaluate the electrochemically active surface areas of catalysts, 18 but have not yet been applied to energy storage due to their selflimiting nature and limited capacities. Surface redox reactions store charges via chemical reactions involving charge transfer across electrode-electrolyte interfaces ( Figure 1B). These processes change the valence states of electrode materials. Hydrous RuO 2 and MnO 2 are two model redox pseudocapacitive compounds. Pseudocapacitive materials storing electrical energy through fast-rate ion intercalation or insertion need to possess ionconducting channels or layers to host guest ions from electrolytes ( Figure 1C). These reactions, unlike those in Li-ion batteries, should not trigger phase transformations during charge-discharge processes. Orthorhombic phase Nb 2 O 5 (T-Nb 2 O 5 ), 19 bronze phase TiO 2 [TiO 2 (B)], 20 and hexagonal phase WO 3 (h-WO 3 ) 21 are archetypical intercalation pseudocapacitive materials. Unfortunately, these pseudocapacitive charge-storage processes induce the intrinsic instability of some pseudocapacitive materials, which undermines the durability of supercapacitors, making electrical energy storage unstable and unreliable.
Cycling stability, a metric quantifying how much capacitance retained after certain durations of operation or numbers of charge-discharge cycles, is a critical figure-of-merit for supercapacitor electrodes. Materials with exceptional capacitance and rate capability have almost no practicability if their lifetimes are limited. The widely accepted criteria for the end-of-life of a supercapacitor are defined as the time when 20% capacitance is lost or the internal resistance doubles. 22,23 In laboratories, cyclic voltammetry (CV, Figure 2A) and galvanostatic (or constant-current) chargedischarge (GCD) tests ( Figure 2B) are two common methods to evaluate cycling stability. The capacitances of single electrodes and supercapacitors are calculated at cycles with fixed intervals. The percentage of capacitance retained when compared to the begin-of-life (or initial) capacitance is coined capacitance retention and is often used to compare the cycling stability of different materials ( Figure 2D). In recent years, researchers proposed methods in addition to CV and GCD, such as voltage hold or potential floating 22,24,25 (Figure 2C), to assess the cycling stability of supercapacitor devices. In these methods, a supercapacitor undergoes three steps per testing cycle: first charged to its maximal potential, dwelled at the peak potential for certain periods, and fully discharged. As each cycle follows the actual working steps of supercapacitors, these methods can provide more realistic results than CV or galvanostatic charge-discharge alone. Industrial protocols on the evaluation of supercapacitor aging are mainly about F I G U R E 1 Schematic illustrations of three pseudocapacitive reactions: A, underpotential deposition (Pb deposition on Au), B, surface redox reactions (protonation and deprotonation of RuO 2 ), and C, fast-rate ion intercalation or insertion. Reproduced with permission from Reference 16. Copyright 2017, T.L electrical-double-layer capacitors and involve harsh testing conditions, such as extensively applying high voltages and large currents, to meet the stringent reliability requirements for large-scale, enduring commercial applications. 26,27 This review article will discuss the recent progress in stabilizing most common pseudocapacitive electrode materials, that is, conjugated polymers, metal oxides, metal nitrides, metal carbides, and metal sulfides. The article begins by summarizing the typical failure modes of these pseudocapacitive electrodes (Section 2), followed by diving into the strategies of curtailing the capacitance-degradation issues (Section 3). Each subsection in Section 3 will start with general descriptions about the latest progress of the corresponding method and then exemplified with one or two representative works. In the last section (Section 4), we will summarize the mechanisms and examples of all the strategies, as well as comment on the challenges and associated opportunities of stabilizing pseudocapacitive materials. We need to point out that the instability of electrode materials is only one of the reasons causing the poor stability of some pseudocapacitors. Other factors that are not directly related to electrodes, such as gel electrolyte solidification or dehydration, electrolyte decomposition, or unsuitable operation temperatures, would also lead to the demise of supercapacitors. However, to keep the article cohesive and coherent, this review will focus only on the stability issues of pseudocapacitive electrode materials, the most common reason for the failure of high-energy-density supercapacitors.

| FAILURE MODES OF PSEUDOCAPACITIVE MATERIALS
Their undesirable cycling stability often shadows the inherently large capacitance of pseudocapacitive materials. These drawbacks result from complex factors associated with the structural integrity, electronic properties, and electrochemical behaviors of pseudocapacitive materials.

| Structure disintegration
Materials store charges through incorporation, insertion, or intercalation of guest species often fail because of structural instability. Similar to battery materials, the uptake and release of guest species accompany expansion and contraction of the host materials, respectively. Such repeated volumetric deformations create internal osmotic stress that cracks, pulverizes, or detaches active materials.
Conjugated polymers are typical pseudocapacitive materials suffering from structure disintegration. Taking polyaniline (PANI) as an example, its charge-storage activity relies on the doping and de-doping of counterions during charge and discharge, respectively ( Figure 3). Starting from the undoped leucoemeraldine form, PANI impregnates anions in electrolytes upon oxidation and transits to the partial-doped emeraldine salt, and eventually to the full-doped pernigraniline salt. The neutrality of PANI is maintained by charge balance between the positively charged quaternary nitrogen sites on backbones and the negatively charged anions. The insertion of counterions drags in solvent molecules, which also F I G U R E 2 A-C, Schemes of cycling stability evaluation methods of, A, cyclic voltammetry, B, galvanostatic (or constant-current) charge-discharge test, and, C, voltage-hold (or potential floating) test. D, An illustration of a typical cycling stability result. The inset shows the mathematical equation of capacitance retention swells PANI. Volumetric expansion due to the incorporation of ions and solvents is called osmotic expansion. The degree of expansion depends strongly on types and concentrations of the doped ions and counter ions. [28][29][30] Bay et al quantified the volumetric change of a free-standing dodecylbenzene sulfonate-doped polypyrrole (PPy) thin film in 0.1 to 5.0 M NaCl aqueous electrolytes. 28 In 0.1 M NaCl, the maximal elongation in the lateral direction was 25 cm 3 mol −1 charge. This value decreased to 12 cm 3 mol −1 charge in 5.0 M NaCl due to the reduced water uptake from the aqueous electrolyte. Lizarraga et al studied the volumetric swelling of PANI in protic aqueous solvents including HClO 4 , H 2 SO 4 , and HCl. The largest volumetric expansion reached~6 vol% in 1 M HCl. 31 Although the volumetric expansions of conjugated polymers is less significant than those of some battery electrodes (eg, Si anodes in Li-ion batteries experience~300 vol% expansion upon full lithiation), 32 the long duration and the repetitive charge-discharge cycles would amplify the destructive force of osmotic expansion and notably degrade the cycling stability. Additionally, intrinsic expansion due to changes in bond lengths and molecular conformations upon ion doping also swells conjugated polymers, but their influence is marginal compared to that of osmotic expansion. 28 Volumetric deformation is also the direct cause to the structural instability of many metal oxides and metal sulfides, particularly those with layered structures, for example, birnessite-MnO 2 (δ-MnO 2 ) 33-36 and Co 9 S 8 . 37 For example, using atomic force microscopy (AFM), Penner and coworkers reported that the diameter of a δ-MnO 2coated Au nanowire increased from~120 to~140 nm ( Figure 4A-C). 34 This considerable volumetric expansion cracked the MnO 2 coatings and rapidly decayed the capacitance. For metal sulfides, Su et al used in situ transmission electron microscopy (TEM) technique to record the volumetric expansion of Co 9 S 8 during lithiation ( Figure 4D-F). 37 After lithiation for 130.75 seconds, the imaged CNT-confined Co 9 S 8 nanowire expanded 4.5% axially and 32.4% radially. Although the extent of the volumetric changes of pseudocapacitive metal oxides and sulfides is not as significant as those of battery electrodes, this relatively small internal stress induced by the volumetric deformation, if left undissipated, could accumulate during long-term cycling and become large enough to demolish their structural integrity.

| Loss of electrical conductivity
Decreasing electrical conductivity of a pseudocapacitive material is detrimental, as it impedes electron transport and deteriorates rate capability. Composition changes due to self-oxidation, overoxidation, and overreduction are the common reasons leading to reduced electrical conductivity of pseudocapacitive materials.

| Metal nitrides
Electrically conductive metal nitrides are promising pseudocapacitive electrodes, but self-oxidation rapidly decreases their capacitance, especially when tested in aqueous electrolytes. Taking TiN as an example, the Pourbaix diagram of Ti-H 2 O shows that the most stable Ti species in water with pH < 14 is Ti(III) or Ti(IV) oxide. 38 TiN oxidizes in alkaline electrolytes via 39 : F I G U R E 3 Scheme of volumetric change and concurrent molecular structure evolution of polyaniline (PANI) during charge and discharge processes Avasarala and Haldar reported that in acidic media, the oxidation process of TiN is potential dependent. 40 Between 0.5 and~0.9 V, the reaction is: And between~1.0 and 1.5 V: Besides electrooxidation, TiN can even spontaneously self-oxidize in air at room temperatue 40 : The presence of Ti─N─O and Ti─O peaks in the Ti 2p X-ray photoelectron spectrum (XPS) and the loss of N signal in the N 1s XPS spectrum ( Figure 5) 40-42 are indicators of TiN oxidation. Additionally, Fourier transform infrared spectroscopy (FTIR) can indicate oxidation based on the presence of the ─OH stretching peak at 3050 cm −1 . 39 Oxidation converts highly electrically conductive metal nitrides to semiconducting metal oxides, which increases electrical resistances and impedes fast electron transport. For example, Choi and Kumta observed that the electrical conductivity of TiN nanocrystals decreased from 2662 to 46 S m −1 after being charged and discharged for 400 times in a 6 M KOH aqueous electrolyte. 39 The dissolution of the oxidation products in acidic electrolytes drove oxidation throughout TiN and largely diminished its charge-storage capability.

| Metal oxides
Oxygen-deficient metal oxides are attractive pseudocapacitive materials because of their enhanced electrical conductivity, augmented charge-carrier mobility, and improved redox activity compared to their stoichiometric counterparts. However, using nonstoichiometric metal oxides as charge storage media in an improper potential window may lose oxygen defects, reduce electrical conductivities, and inhibit the charge transport. For example, molybdenum oxide (MoO x ) with Mo 4+ and Mo 5+ has higher electrical conductivities than MoO 3 (Mo 6+ ). 48 Cycling MoO x in a potential window between −1.0 and 0 V vs SCE reduced the oxygen nonstoichiometry due to oxidation of Mo 4+ and Mo 5+ to Mo 6+ . This process augmented the electrical resistance and decreased the charge-carrier mobility of MoO x , which led to over 25% capacitance decay after only 500 charge-discharge cycles.

| Conjugated polymers
The counterion drain effect gradually degrades the electrical conductivity of conjugated polymers. This problem is related to the repeated swell-shrinkage of conjugated polymers that collapses ion diffusion channels and retards doping of ions from electrolytes. This consequence reduces the population of bipolarons, that is, positive charged sites on the backbones of conjugated polymers that conduct electrons, 49-51 and deactivates conjugated polymers.
Additionally, charging conjugated polymers to potentials higher than the upper limits of their potential windows could permanently damage the electrochemical activities of conjugated polymers. This process is called overoxidation and was extensively studied in the 1980s and 1990s. In 1982, Bard and coworkers reported that polypyrrole films "were unstable to oxidative conditions such as potentials >0.6 V vs SCE or the presence of halogens." 52 Others have observed the similar phenomena in the following years. [53][54][55][56][57] The mechanism of overoxidation and its influence on the structure-electrical conductivity relationship are complex and polymer dependent, 50 and calls for continuing investigations. The net outcome of overoxidation, however, is identical to the counterion drain effect: reducing the electrical conductivities of conjugated polymers and making them electrochemically inert.

| Dissolution of active components
Formation of highly soluble species in electrolytes during charge storage could break the reversibility of electrodes' redox reactions and consume active materials. For example, scanning MnO 2 at potentials below its stable potential window irreversibly reduces MnO 2 to low-valence products, for example, Mn 2 O 3 or MnOOH, which undergo disproportionation and yield Mn 2+ ions. 58,59 Specifically, the dissolution reaction in acidic electrolytes is: In alkaline electrolytes: Hsieh et al argued the validity of Equations (7) and (8) because they believed that the Mn 2+ concentration in electrolytes was too small to account for the capacitance loss. 60 This discrepancy was explained by Nam et al who utilized Mn K-edge X-ray near-edge absorption spectroscopy and concluded that the irreversible reduction of MnO 2 proceeded not only near its surface but also in bulk. 61 Additionally, Ataherian et al systematically investigated the interplays among the valence, morphology,  59 Their experimental results confirmed the relationship between Mn valence reduction and surface morphology reconstruction. The morphological change indicated that MnO 2 experienced a dissolution-recrystallization process during the reduction process. The dissolution, however, was self-limiting because the mixed-valence nature of MnO x surface halted the continuous dissolution of lowvalent manganese oxides and hydroxides.
For metal nitrides and carbides, dissolution could also be triggered by oxidation. For example, V 2 O 5 and V 2 O 3 , the main oxidation products of VN, are readily soluble in acidic aqueous media. 62 The combined loss of electrical conductivity and active materials of VN resulting from oxidation rapidly deteriorate its cycling stability. Besides metal nitrides, Messner et al demonstrated that various metal carbides are also electrochemically oxidizable at potentials higher than 1.0 V vs Ag/AgCl in aqueous electrolytes at various pH values. 63 The metal and carbon elements in these metal carbides were oxidized to corresponding water-soluble metal salts and carbide-derived carbon, respectively.

| Gas evolution
Gas evolution is a generic adverse event for the cycling stability of all the pseudocapacitor electrodes. When the operation potential exceeds the stable potential thresholds of electrolytes, gas bubbles (eg, hydrogen and oxygen gases) evolve on surfaces of positive and/or negative electrodes due to water electrolysis. Vigorous gas evolution could destruct electrode-current collector interfaces, delaminate active materials from substrates, create pinholes and gaps that increase contact resistance, raise IR drop, decrease coulombic efficiency, and impose safety concerns of joule heating. [64][65][66] Particularly, gas generation is detrimental to sealed pseudocapacitors because it not only consumes the limited electrolytes but also balloons device pouches that may eventually lead to explosions.
For MnO 2 , Ataherian and Wu observed that oxygen evolution at potentials more than 1.0 V vs Ag/AgCl (saturated KCl) reduced Mn 4+ to Mn 3+ and Mn 2+ via a pathway shown in Figure 6. 67 Such irreversible reaction decreased the population of active Mn 4+ species and lowered the capacitance of MnO 2 .
Determining the stable operating potential windows of electrolytes and electrodes before any electrochemical tests is an effective way to avoid gas evolution problems. In CV, steep increase or decrease in current at potentialwindow limits indicates gas evolution. To prevent such destructive processes, the potential windows must be narrowed. Technical details on deciding suitable potential windows are out of this review's scope, but are available in a recently published perspective article. 68

| STRATEGIES FOR ENHANCING THE CYCLING STABILITY OF PSEUDOCAPACITIVE MATERIALS
This section thoroughly surveys the current progress in the strategies aiming at enhancing the cycling stability of pseudocapacitive materials. The discussion is segmented according to the type of materials: conjugated polymers, metal oxides, metal nitrides, and carbides, as well as metal sulfides. Each subsection is elaborated with specific strategies and exemplified with one or two representative examples. Metal nitrides and carbides are combined because they suffer from the same problem of oxidation.

| Spatial confinement
Compounding conjugated polymers with mechanically soft materials is the most widely practiced method to address the structural disintegration of conjugated polymers. These materials function as buffer or cushion substrates that release the stress built in conjugated polymers. Even if the incorporated conjugated polymers crack, the broken pieces could still anchor on the substrates without complete disintegration. Among various host materials, carbon materials are most popular because of their abundance, soft nature, and versatile morphologies. Specifically, porous carbon particles, 69,70 carbon nanotubes, 71-79 reduced graphene oxide (rGO) sheets, 77,78,80-91 graphene oxide, 92 graphite foils, 93 etched carbon fibers, 94 and carbon aerogels 95 are reported carbon hosts. Graphene analogs with layered structures that can accommodate large volumetric expansion, such as molybdenum disulfide (MoS 2 ), 96,97 black phosphorus, 98 and two-dimensional transition metal nitrides and carbides (MXenes) 99 are also potent candidates for stabilizing conjugated polymers. Surface modifications, including sulfonation, 79 are performed to introduce hydrophilic functionalities that can form covalent bonds and/or hydrogen bonds with conjugated polymers and strengthen their interfaces.
Gogotsi and coworkers demonstrated that Ti 3 C 2 T x (T x represents the terminal functional groups on the Ti─C layers), a compound of MXene, appreciably improved the cycling stability of polypyrrole (PPy). 99 Utilizing the hydrophilic functional groups on the surface of Ti 3 C 2 T x , the researchers in situ polymerized pyrrole intercalating in the Ti 3 C 2 T x layers ( Figure 7A). The laminated structure of Ti 3 C 2 T x was well preserved after the polymerization, as evidenced by scanning electron microscopy (SEM, Figure 7B) and the periodic pattern under TEM ( Figure 7C). Due to the electrochemical activity of PPy, the PPy/Ti 3 C 2 T x composite film exhibited capacitances 1.5 to 2.5 times higher than those of bare Ti 3 C 2 T x electrodes across a wide range of scan rates ( Figure 7D) and 92% capacitance retention after 25 000 charge-discharge cycles ( Figure 7E).

| Surface coating
Compositing conjugated polymers might offer limited protection for nanostructured conjugated polymers with irregularly shaped surfaces, because achieving conformal coverage is challenging. Therefore, conformal coating or encapsulation are strategies to protect nanostructured conjugated polymers. In general, most coating materials are rigid, for example, metal oxides (RuO 2  100 and   TiO 2   101 ) and Nafion, 102 to withstand the large volumetric expansions of conjugated polymers. Soft and elastic carbon-based materials have also been used as binders to hold the structure of conjugated polymers. 103,104 One representative example of soft coating materials is glucose-derived carbonaceous shell. 103 Glucose was charred hydrothermally and conformally deposited 5-nm-thick carbonaceous shells onto PANI and PPy nanowires (PANI@C and PPy@C) ( Figure 8A,B). With this coating to strengthen their structural integrity, PANI@C and PPy@C retained~95% and~85% of their initial capacitance after 10 000 charge-discharge cycles in 1 M H 2 SO 4 aqueous electrolytes, respectively ( Figure 8C, D). These stability performances were among the best of conjugated polymers. In comparison, the capacitance of the unprotected PANI and PPy nanowires both dropped to~20% capacitance at the end of the cycling test. Moreover, the cycling stability enhancement was at little cost of capacitance, as evident from the nearly overlapped cyclic voltammograms between the uncoated and coated nanowires (insets of Figure 8C,D). SEM proved that both PANI@C and PPy@C nanowires were preserved after cycling, while all the bare PANI and PPy nanowires disappeared. This discrepancy suggested that the carbon layer effectively suppressed the structure disintegration of PANI and PPy.

| Engineering structures
Engineering structures stabilize conjugated polymers without compounding with other materials. The voids or empty spaces, such as inter-wire gaps of nanowire arrays, 105 hollow chambers of nanotubes, 87,106,107 and pores within three-dimensional architectures 108 buffer volumetric expansions. Sometimes, gel electrolytes are filled in these spaces to provide additional  109 Conjugated polymers are mainly produced by electropolymerization or chemical polymerization with limited controllability on their morphologies. Therefore, nanostructured conjugated polymers usually need hard templates for synthesis.
Smoukov and coworkers demonstrated an excellent example of an interpenetrating nanowire network to stabilize poly(3,4-ethylenedioxythiopene) (PEDOT) nanowires. 109 Using sequential polymerization, the researchers synthesized self-standing poly(ethylene oxide) (PEO)/PEDOT films consisting of interpenetrated PEO and PEDOT nanowires ( Figure 9A-C). PEO was percolated with LiClO 4 to serve as an ion reservoir, and PEDOT conducted electrons ( Figure 9D). The configuration of soft PEO-wrapped PEDOT was helpful in uniformly dissipating the internal stress built in the charged PEDOT nanowires to enhance the cycling stability of PEDOT. The capacitance of PEO/PEDOT retained 97.5% after 3000 cycles at 10 A g −1 , whereas that of neat PEDOT continuously decreased within the first 1200 cycles under identical testing conditions ( Figure 9E).

| Ordering molecular structures
Configuring molecular structure is effective in minimizing the structural breakdown of PPy. For example, Zhi and coworkers showed that PPy films with a planar molecular configuration displayed outstanding cycling stability. 32 Specifically, the authors compared two types of PPy films synthesized by different methods. One was e-PPy films produced via constant-current electrodeposition in aqueous solutions containing 0.1 M ptoluenesulfonic acid, 0.3 M sodium toluenesulfate, and 0.5 vol% pyrrole at 0 C. The aligned electric field favored the deprotonation of pyrrole at its α-position, leading to PPy with a planar configuration ( Figure 10A). Another one is c-PPy synthesized by chemical polymerization in  Figure 10B).
Constant-current cycling stability test revealed that e-PPy and c-PPy exhibited drastically different behaviors. The capacitance of c-PPy lost more than 50% after only 1000 cycles ( Figure 10C), while that of e-PPy retained >97% after 15 000 cycles. More significantly, the latter retained~50% capacitance after more than 230 000 cycles.
To understand the origin of the different cycling stability, the authors characterized the morphologies of the two electrodes after the test. Raman spectroscopy showed that the peak at 1573 cm −1 , corresponding to C─C  Figure 10D). This result indicated that the structure of c-PPy was deconstructed at the end of cycling. The C─C stretching Raman peak of e-PPy, in contrast, was still discernible after cycling, meaning its molecular arrangement was preserved. The SEM images showed that the surface morphology of e-PPy changed little after cycling ( Figure 10E, F), while the nanospheres of c-PPy disappeared ( Figure 10G,H). The enhanced structural integrity of e-PPy was attributed to its planar molecular structure, which uniformly distributed osmotic stress across the entire PPy planes and prevented local-stress-induced damage. The effect of molecular ordering in stabilizing PPy was later acknowledged by Zhou and coworkers. 110

| Tuning composition
As discussed in Section 2.2.3, the counterion drain effect gradually depletes the dopants in conjugated polymers, making them electrically insulating. While stabilizing the structures of conjugated polymers is useful to maintain their ion diffusion channels, a more direct method to minimize the counterion drain effect is to dope conjugated polymers with immobile anions, such as β-naphthalene sulfonate, 49 sulfanilic acid azochromotrop, 111 p-toluenesulfonate, 112 perchlorate, 113 and Tiron, 114 or with bulky redox-active molecules including hydroquinone-benzoquinone. 115,116 The substantial steric hindrance among the polymer chains immobilizes these ionic dopants inside the polymer networks. During charge and discharge, these anions could stay near the polarons without leaking out. Moreover, these bulky anions support ion-diffusion channels and allow for reversible doping and de-doping of small anions from electrolytes, which maintains the charge-storage capacities of conjugated polymers.
A typical example of bulky counterion is β-naphthalene sulfonate anion (NS − ). 49 Compared to sulfate anion, the large naphthalene ring of NS − significantly restricted its movement among PPy chains. The cycling stability test showed that sulfate anion-doped PPy films on graphite foils exhibited the poorest stability: its capacitance decayed to~50% after 2000 cycles ( Figure 11A, PPy-S). Sulfatecontaining PPy films on exfoliated graphite foils exhibited capacitance retention of~90% after 10 000 cycles, because the graphene layers atop the exfoliated graphite foils served as mechanical buffers to accommodate the volumetric deformation of PPy ( Figure 11A, FEG/PPy-S). Remarkably, when doping NS − into PPy films on exfoliated graphite foils, its cycling stability further improved to 97% after 10 000 cycles ( Figure 11A, FEG/PPy-NS). SEM showed that the film morphology of the NS − -doped PPy was intact after cycling ( Figure 11B). More importantly, EDS showed that the S signal of NS − almost unchanged after the stability test, whereas that of sulfonate anion completely disappeared ( Figure 11C,D). This observation unequivocally proved that the NS − stayed in PPy during cycling, which suppressed the counterion drain effect and promoted long lifetimes of PPy electrodes.

| Constructing carbon-based composites
Combining metal oxides with carbon materials simultaneously addresses the oxides' low electrical conductivity and structural instability. First, carbon matrices provide electrons fast diffusion pathways in the metal oxide/carbon composites. Second, the softness of carbon substrates can buffer the volumetric change of metal oxides during charge and discharge. To date, carbon materials of various dimensions, including zero-dimensional carbon dots, 117,118 one-dimensional carbon fibers 119 and carbon nanotubes, 120-126 two-dimensional graphene or rGO sheets, [127][128][129][130][131][132][133][134][135][136] three-dimensional porous carbon networks 137 have been combined with metal oxides. Additionally, carbon 138,139 and TiN 140 coatings on metal oxides also stabilize the structure of metal oxides.
Guan et al demonstrated three-dimensional, graphene foam-supported carbon nanotube forests as soft substrates for loading Fe 2 O 3 nanoparticles. 120 The CNTs and Fe 2 O 3 particles were grown by chemical vapor deposition and atomic layer deposition (ALD), respectively ( Figure 12A). SEM ( Figure 12B) and TEM images ( Figure 12C) both showed that Fe 2 O 3 nanoparticles distributed uniformly on the outer surfaces of CNTs. Specifically, CNTs with Fe 2 O 3 deposited by 600 ALD cycles (GF-CNT@600-Fe 2 O 3 ) showed cycling stability of 86.2% over 50 000 charge-discharge cycles in 2 M aqueous KOH electrolytes ( Figure 12D). This stability outperformed most other Fe 2 O 3 -based composite electrodes, which was ascribed to the strong bonding between Fe 2 O 3 nanoparticles and CNTs, the highly reversible redox reactions of the ultra-small Fe 2 O 3 crystals, and the cushion space among the Fe 2 O 3 nanocrystals. The initial capacitance decay might correlate to the partial aggregation of the Fe 2 O 3 nanoparticles ( Figure 12E) that reduced the electrolyte-accessible surface area for redox reactions.

| Tailoring morphology
The structural instability of metal oxides has been partially addressed by building unique nanostructures. The stabilization mechanism is identical to that of the nanostructured conjugated polymers (Section 3.1.4). The excellent structural tunability of metal oxides enabled the preparation of nanostructures without hard templates. Nanowires, 141-143 nano-needles, 144 nanosheets, [143][144][145][146][147] and nanorods 148 are morphologies showing improved cycling stability compared to films. Nevertheless, nanostructures alone cannot resolve the dissolution of some metal oxides.
One representative example of how morphology influences the cycling stability of metal oxides was demonstrated by Ouyang et al. 144 Using cetyl trimethyl ammonium bromide and NH 4 F as structure-directing agents, the authors synthesized

| Gel electrolytes
Polymer-based gel electrolytes offer twofolded protections to metal oxides that liquid electrolytes cannot: Their elasticity and softness curtail the structural instability of metal oxides, and the low water content of gel electrolytes limits dissolution of water-soluble metal ions. 34 Poly(vinyl alcohol) (PVA) 62 and poly(methyl methacrylate) (PMMA) 33 are common polymers used to prepare gel electrolytes.
Penner and coworkers achieved extraordinary cycling stability of MnO 2 nanowires in PMMA gel electrolytes ( Figure 14). 33 Using photolithography and electrodeposition, the researchers made interdigitate gold nanowires coated with δ-MnO 2 nanosheets (Au@MnO 2 ) ( Figure 14A Figure 14D). Au@MnO 2 nanowires with 143-nmthick MnO 2 shells achieved the best cycling stability, experiencing no capacitance loss after 200 000 chargedischarge cycles. Post-mortem analysis using SEM revealed that the MnO 2 shells in the liquid electrolytes partially detached due to structure breakdown, whereas the MnO 2 shells were well preserved in the PMMA gel electrolytes. These results signified that gel electrolytes prevented the dissolution of δ-MnO 2 ( Figure 14E).
Wang et al demonstrated that PVA/LiCl gel electrolytes effectively suppressed the dissolution of vanadium oxide nanowires. 62 The unsatisfactory cycling stability of vanadium oxide in aqueous electrolytes mainly results from the formation of water-soluble vanadium species during charge and discharge (Section 2.3). Substituting aqueous electrolytes with gel electrolytes reduces water contents but still retains the ion diffusion capability of electrolytes, thus minimizing vanadium oxide dissolution. Experimental results showed that vanadium oxide nanowires in the PVA/LiCl gel electrolyte displayed approximately nine times higher capacitance retention than those in a 5 M LiCl aqueous electrolyte over 5000 charge-discharge cycles ( Figure 15A). Compared to the vanadium oxide nanowires before cycling ( Figure 15B), nanowires cycled in aqueous electrolytes almost disappeared (dissolution) and cracked (volumetric expansion) ( Figure 15C), while those covered with gel electrolytes were preserved and visible under SEM ( Figure 15D).

| Valence state engineering
Tuning valence states of metal oxides directly curbs dissolution. This method requires comprehension of the interplays among valence state, electrical conductivity, thermodynamic stability, and electrochemical activities. Typically, increasing the concentrations of active and stable valence states in metal oxides promotes longlasting performance. Reported valence-tuning methods include electrooxidation, 149 electroreduction, 119 elemental doping, 150 building oxide-based heterostructures, 151 and changing potential window. 48 Despite its versatility, valence tuning must be meticulously performed to avoid stability enhancement at the cost of capacitance.
Song et al demonstrated an example of stabilizing amorphous vanadium oxide nanowires by electrically reducing some V 5+ to V 4+ in vanadium oxide. 119 The rationale of introducing low valent vanadium species was that VO 2 is thermodynamically stable and waterinsoluble at negative potentials in neutral aqueous electrolytes. 152 Vanadium oxide nanowires were first deposited via CV on electrochemically exfoliated carbon fibers, followed by electroreduction at −1.5 V (vs saturated calomel electrode) for 1 minute in 0.1 M vanadyl sulfate and 0.2 M ammonium acetate aqueous solutions at room temperature. This process shifted the V 4+ /V 5+ ratio from~40% to~50%. Cycling test showed that the reduced vanadium oxide on exfoliated carbon fiber (ECC/RVO x ) exhibited the best stability performance, exhibiting no capacitance loss after 100 000 charge-discharge cycles at 60 mA cm −2 ( Figure 16A). In contrast, the as-deposited vanadium oxide on exfoliated carbon fibers (ECC/VO x ) experienced constant capacitance degradation starting from the beginning of the cycling test, and only maintained~70% capacitance after 10 000 cycles. The initial capacitance loss of reduced vanadium oxide on the untreated carbon fibers (CC/RVO x ) was due to the structural collapse of RVO x , and it steadily raised back to~80% retention after 20 000 cycles. SEM images of the cycled electrodes revealed that the structure of ECC/RVO x was well preserved ( Figure 16B), but the nanowires of ECC/VO x and CC/RVO x dissolved ( Figure 16C) and detached ( Figure 16D), respectively. The discrepancy in cycling stability highlighted two advantages of ECC/RVO x : The exfoliated carbon shells of ECC avoided the volumetric-deformation-induced structural collapse of vanadium oxide, and the increased concentration of V 4 + suppressed nanowire dissolution. Additionally, XPS elucidated that V 4+ /V 5+ ratios of the three electrodes all increased during the cycling, and that of ECC/VO x plateaued at~60% ( Figure 16E). The observation suggested that the valence state of V underwent a dynamic transition that enriched tetravalence. Additionally, O 1s XPS spectrum of ECC/RVO x showed that the peak intensities of structural water and V─OH gradually increased ( Figure 16F). The continuous hydration of vanadium oxide was possibly a result of water uptake during the activation process at the beginning of the stability test. Nyquist plots unveiled that the water incorporation facilitated ion diffusion within the reduced vanadium oxide nanowires, as evident from

| Ion preinsertion
Inserting immobile guest ions is a potent structurestabilization method for metal oxides possessing tunneled or layered structures. The intentionally introduced ions need to bind strongly with the channel walls or layers of the host metal oxides, which prevents these ions from leaking out at high potentials. The inserted ions, therefore, function as structure scaffolds to minimize the structure expansion of metal oxides and impede pulverization. It is noteworthy that the effectiveness of this strategy is material dependent. One needs to consider the compatibility of the guest ions with their hosts. For example, Yao et al observed that introducing hydrated zinc and gallium ions in between δ-MnO 2 layers stabilized capacitance (no capacitance loss after 20 000 chargedischarge cycles), but Fe 3+ -intercalated δ-MnO 2 showed rapid capacitance decay (~10% capacitance retention) under identical testing conditions. 36 The discrepancy in the cycling stability performance was attributed to the different interaction strength between the cations and the δ-MnO 2 layers.

.1 | Protective encapsulation
Encapsulation provides metal nitrides and carbides physical barriers to prevent their contact with oxidants (eg, water and oxygen gas) and, thus, to halt oxidation. Most coating materials are electrically conductive to maintain the high electrical conductivity of the entire materials. Demonstrated coatings are composed of amorphous carbon shells derived from glucose, 41 polydopamine, 153 poly (acrylonitrile), 154 polymeric nano-micelles, 155 rGO, [156][157][158][159] and conjugated polymers. [160][161][162] Occasionally, ultrathin metal oxide layers are used to stabilize metal nitrides, 163 but the main purpose is to facilitate electron flow into the metal oxide coatings from the coated metal nitrides. The thickness of coatings is a critical parameter that needs optimization to avoid compromising the capacitance and rate capability of the encapsulated materials: Too thin coatings might be incomplete and provide limited protection; too-thick coverage could block ion accessibility into the coated materials and reduce capacitance. In situ or operando characterizations are valuable in acquiring fundamental understandings on the ion diffusion mechanisms across coatings, which is currently scarce but will aid the rational design of the thickness, porosity, and composition of the coating materials.
To mitigate the oxidation problem of TiN, Lu et al developed a carbon-coating method. 41 The carbon layers were first produced by decomposing glucose in a hydrothermal reaction, followed by thermal annealing at 800 C under N 2 atmosphere ( Figure 17A) to convert the glucose-decomposed products into carbon. The lattice-resolved TEM image ( Figure 17B) showed that a 1.5-nm-thick, amorphous carbon layer was uniformly deposited on a single-crystalline TiN nanowire (TiN@C). This carbon layer substantially elevated the capacitance retention of the TiN nanowires to~92% after 15 000 charge-discharge cycles in 1 M aqueous KOH electrolytes ( Figure 17C). In sharp contrast, the uncoated TiN nanowires retained only~9% capacitance under the same testing conditions. SEM ( Figure 17D) and XPS ( Figure 17E) of the cycled TiN@C nanowires proved that the structure and surface functionalities remained unchanged, respectively.

| Tuning electrolyte composition
As moisture is one of the culprits oxidizing the surfaces of metal nitrides and carbides, reducing water contents in electrolytes is beneficial to slow down or stop the oxidation of metal nitrides and carbides. Therefore, gel electrolytes (mainly PVA-based) 42,159,164 and organic electrolytes 165 with little or no water could stabilize metal nitrides and carbides. Recently, researchers have discovered that neutral electrolytes also prolong the lifetimes of metal nitride electrodes, owing to an equilibrium between dissolved metal ions and solid-state metal nitrides that decelerates surface oxidation. 166,167 One representative PVA-based gel electrolyte that stabilized TiN nanowires was reported by Lu et al. 42 The researchers added PVA powder into KOH aqueous solutions to convert the liquid solutions into PVA/KOH gel electrolytes. These electrolyte thin films were then sandwiched between TiN nanowires grown on carbon cloth, followed by being hot-pressed to infiltrate the inter-wire space among the TiN nanowires ( Figure 18A). As a result of surface oxidation, TiN nanowires cycled in aqueous KOH electrolytes showed apparent capacitance decay, with less than 10% capacitance remained after the first 3000 charge-discharge cycles. In contrast, TiN nanowires incorporated with PVA/KOH held 83% of its initial capacitance at the same cycle number ( Figure 18B). TiN nanowires cycled in PVA/KOH also retained their morphology ( Figure 18C) and only mildly oxidized, as evidenced by the slight reduction in the XPS peak area of Ti─N bond ( Figure 18D).

| Cation intercalation
For metal nitrides and carbides with layered structures, for example, MXenes, guest species in between their layers could change their intrinsic electronic properties and improve their oxidation resistance. For example, VahidMohammadi et al reported that intercalating Li + or Na + into V 2 CT x (T x stands for surface functional groups of ─OH, ─O, and ─F), a layered metal carbide, could make it air-stable and electrochemically robust. 168 The researchers first dispersed V 2 CT x nanosheets in water and mixed with concentrated aqueous solutions of LiCl and NaCl. Since V 2 CT x nanosheets were negatively charged, these cations preferably adsorbed onto V 2 CT x nanosheets. After filtration and drying to stack the ion-adsorbed V 2 CT x nanosheets together, self-standing V 2 CT x films intercalated with Li + or Na + were prepared ( Figure 19A). Outstandingly, in various aqueous electrolytes, Na + -intercalated V 2 CT x films showed almost no loss of capacitance after cycling for 10 000 cycles in various aqueous electrolytes ( Figure 19B), in sharp contrast to intercalant-free V 2 CT x films which lost more than 30% capacitance after merely 1500 cycles in 5 M aqueous LiCl electrolytes ( Figure 19C). Additionally, the difference in the film integrity between the cycled Li + -intercalated V 2 CT x and intercalant-free V 2 CT x films corroborated the cycling performance ( Figure 19D). Additionally, XPS showed that compared to as-prepared V 2 CT x ( Figure 19E), intercalant-free V 2 CT x contained much more oxidation products V 5+ after exposing to air in 1 month ( Figure 19F); however, the concentrations of all V-cations in Li + -intercalated, air-exposed V 2 CT x were comparable to those of as-prepared V 2 CT x ( Figure 19G), suggesting no appreciable oxidation happened. Density functional theory further found that electrons could transfer from the intercalants to the surfaces of V 2 CT x and primarily reside near the ─F surface groups. This electron redistribution made V 2 CT x more negatively charged and oxidation resistant.

| Mechanical buffering
Similar to metal oxides, metal sulfides, especially those with layered structures that can host guest molecules between adjacent layers, suffer from capacitance loss due to structure breakdown. To remedy these problems, incorporation metal  177,181,182 and three-dimensional carbon foams or aerogels, [183][184][185][186] are the most adopted materials because of their structural flexibility and high electrical conductivity. Structurally stable metal oxides (eg, NiO, 187 ) or amorphous metal sulfides (eg, amorphous Co 3 S 4 188 ) are occasionally selected to stabilize metal sulfides. Besides functioning as mechanical buffers, these substrates and wrappings could also anchor metal sulfide nanoparticles and prevent the small particles from coalescing during charge and discharge, promoting long lifetimes.
Lee, Kim, and coworkers demonstrated one excellent example of graphene wrapping to stabilize metal sulfides. 170 Specifically, the authors covered Co 8 FeS 8 and FeS nanoparticles with nitrogen-doped graphene (NG) sheets via a hydrothermal reaction ( Figure 20A). High-resolution TEM image presented that Co 8 FeS 8 nanoparticles were uniformly covered with graphene sheets without severe local aggregation ( Figure 20B). The mechanical robustness and high cycling stability of the N-doped graphene nearly doubled the cycling stability of Co 8 FeS 8 , compared to that of bare Co 8 FeS 8 , after 10 000 charge-discharge cycles in 2 M aqueous KOH electrolytes ( Figure 20C). SEM images of the cycled nitrogen-graphene wrapped Co 8 FeS 8 (Co 8 FeS 8 @NG) showed no particle degradation or aggregation before and after cycling ( Figure 20D,E), proving that NG wrapping was beneficial to stabilize and prevent nanoparticle aggregation.

| Structure engineering
Besides wrapping or supporting metal sulfides with flexible materials, nanostructured metal sulfides showed improved structural integrity. The rationale of this strategy is to accommodate the volumetric deformation of metal sulfides using the empty spaces among nanostructures. To date, a plethora of nanostructured metal sulfides, including hollow particles, 179,189,190 nanosheets/nanoplates, [191][192][193][194][195] [196][197][198][199] nanowires, 200 nanotubes, 201 and hierarchical architectures integrating multiple nanostructures have displayed improved cycling stability. 202,203 Huang et al reported a hierarchical structure involving nanotubes and nanosheets for stabilizing Co 9 S 8 . 203 They first grew ZnO nanowires onto commercial carbon  Figure 21A). SEM images demonstrated that Ni/Co 9 S 8 nanotubes grew uniformly on each carbon fiber, and Co 9 S 8 nanosheets covered the Ni    Figure 21B-D). TEM images illustrated the hollow tubes and the coated polycrystalline Co 9 S 8 nanosheets ( Figure 21E-G). This hierarchical architecture not only facilitated ion diffusion to render excellent rate capability ( Figure 21H) but also effectively suppressed the structure disintegration of Co 9 S 8 to bestow the high cycling stability in 2 M aqueous KOH electrolytes ( Figure 21I). The hierarchical structure remained intact after cycling (Figure 21I inset).
Lou and coworkers reported a general strategy of synthesizing box-in-box shaped metal sulfide nanoparticles with excellent cycling performance. 189 They first synthesized SiO 2 nano-boxes using Fe 2 O 3 nanocubes as sacrificial templates, followed by dispersing these SiO 2 nano-boxes into aqueous solutions containing nickel acetate, ammonia, and ammonia chloride. The weak alkaline environment provided by the buffer system of ammonia and ammonia chloride slowly dissolved SiO 2 into silicate anions, which reacted with Ni 2+ and formed nickel silicate. The diffusion-limited kinetics of Ni 2+ resulted in the double-walled nickel silicate hollow cubes ( Figure 22). Afterward, the double-walled nickel silicate particles were converted to nickel sulfide by sulfurization. The synthesized NiS double-walled boxes were coated with dense and uniform nanosheets both on the outer and inner surfaces of the boxes ( Figure 22B,C). Having the unique box-in-box configuration that alleviated the destruction power of volumetric swelling and shrinkage, the double-walled, hollow NiS particles retained 93.4% capacitance after 3000 charge-discharge cycles at 4 A g −1 ( Figure 22D). In comparison, the capacitance of solid NiS particles continually decreased during 1500 chargedischarge cycles at 4 A g −1 .

| Doping
Doping has dual effects in enhancing the cycling stability of metal sulfides. It reduces electrical resistance to minimize joule heating, as well as tunes crystal structures to metal sulfides with inherently good cycling stability. Inspired by previous experimental observations that Co 3 S 4 showed superior cycling stability, 204,205 Li et al, conceived the idea of doping Co 2+ to stabilize Ni 3 S 4 nanosheets. 206 The doping was realized by a solid-state reaction where precursors of Ni 3 S 4 along with cobalt nitrate hexahydrate (the Co 2+ source) were ground and mixed thoroughly, followed by aging at 150 C for 5 hours ( Figure 23A). XRD showed that all the diffraction peaks of Ni 3 S 4 consistently shifted to slightly larger diffraction angels ( Figure 23B), indicating the successful doping of Co 2+ . Since the crystallographic structure of Co 3 S 4 is identical to that of Ni 3 S 4 , there were no new XRD peaks after Co 2+ -doping. Co 2+ -doped Ni 3 S 4 and pristine Ni 3 S 4 have quite different cycling stability: The former kept 65.7% of its initial capacitance after 10 000 cycles, while the latter's capacitance dropped to 10.2% after only 2500 cycles ( Figure 23C). The stability enhancement was attributed to the partial conversion of Ni 3 S 4 to Co 3 S 4 with inherently better cycling stability.

| SUMMARY AND OUTLOOK
This review has described the typical failure modes and possible resolutions of five most commonly used pseudocapacitive materials: conjugated polymers, metal oxides, metal nitrides, metal carbides, and metal sulfides. This information, along with selected examples, is summarized in Table 1.
Cycling stability, as a crucial performance metric for pseudocapacitive materials, will continue to be extensively characterized for emerging electrode materials. Herein we would like to highlight four issues associated with cycling stability that could serve as useful guidelines for future researches.
First, researchers must avoid confusions between battery-like materials and pseudocapacitive materials. Pseudocapacitive materials store charges via Faradic redox reactions with fast kinetics. Their electrochemical signatures are similar to traditional electrical double layer capacitive materials (eg, activated carbon): rectangularshaped cyclic voltammograms and sloping potential-time GCD profiles ( Figure 24A,B). A pseudocapacitive material could show some deviations but should behave similarly to electrical double layer capacitive materials ( Figure 24C, D). Differently, battery-like materials that undergo diffusion-controlled reactions often have sluggish kinetics, and their cyclic voltammograms contain sharp, welldefined peaks whose peak potentials are nonoverlapping ( Figure 24E). Correspondingly, GCD profiles of batterylike materials have clear plateaus at different voltages in both charging and discharging ( Figure 24F). Protocols on the performance evaluation, result interpretation, and data presentation between pseudocapacitive and battery-like materials are drastically different. This information, as well as nontypical cases of pseudocapacitive processes and detailed differentiation criteria, are available in several excellent articles. 68,207,208 Thoroughly discussing the details are beyond the scope of this review, but we strongly encourage researchers, particularly new practitioners in electrochemical energy storage, to read these papers. Correctly identifying pseudocapacitive materials is the prerequisite for scientifically valid performance evaluations and comparisons, as the mechanisms on performance degradation and corresponding resolutions could differ fundamentally based on the nature of electrode materials.
Second, supplying cycling stability performance with sufficient experimental conditions is helpful to facilitate performance comparison. These details include cycle number, potential window, scan rate (for CV), current density (for galvanostatic charge-discharge test), potential holding duration (for voltage-hold or potential-T A B L E 1 Failure modes and corresponding solutions of the pseudocapacitive materials presented in this article floating method), mass loading, electrolyte composition, and testing temperature. The cycling stability of material could deviate profoundly under different testing conditions. Cycling a material with short durations (eg, <1000 cycles for capacitive materials) could overestimate cycling stability, thus, must be avoided. We are aware that extending cycling tests to the end-of-life for materials of interest might be inapplicable for laboratory evaluations. Therefore, we want to recommend herein the minimum cycles needed for evaluating the cycling stability of supercapacitor electrodes, as proposed by Balducci et al 209 :10 000 cycles and 5000 cycles for electrical double layer capacitive and pseudocapacitive materials, respectively. Extended stability tests are always recommended. Third, postmortem investigations of the electrodes after cycling are indispensable. Electron microscopy images, XRD, and XPS are useful tools to reveal the evolution in morphology, crystallinity, and composition of the cycled electrode materials, respectively. The obtained information is particularly critical in understanding the origins of capacitance fading, selfactivation processes, or stable cycling performance, which offers much more insights than values of capacitance retentions alone, and sometimes might even lead to ground-breaking discoveries. Readers could refer to the postmortem analyses on nanostructured MnO 2 electrodes performed by Chen et al as an excellent example. 143 Forth, the progress of developing pseudocapacitive materials with outstanding stability calls for interdisciplinary engagements. For example, methods and techniques practiced in battery fields (eg, operando electron microscopy) 177,210,211 could also be employed to study F I G U R E 2 4 Schemes of representative, A,C,E, cyclic voltammograms and, B,D,F, galvanostatic charge-discharge profiles of, A,B, electrical-double-layer capacitive, C,D, pseudocapacitive, and, E,F, battery-like materials pseudocapacitive materials. Mechanistic insights into intrinsic oxidation resistances of metal-based materials may come from experiences gained in fuel cell and catalysis fields. 212 Industrial developers could guide the stability evaluations of application-oriented devices to facilitate their conversion into commercial products. Close collaborations and conscientious communications across a wide range of disciplines are undoubtedly beneficial to the healthy development of all fields involved.
To sum up, exceptional long-term cycling stability is an essential quality of electrode materials in robust and reliable electrochemical energy storage devices. Achieving this goal is challenging, but consistent, collaborative, and comprehensive efforts will surely make it attainable.