Electrode Materials and Prediction of Cycle Stability and Remaining Service Life of Supercapacitors

Abstract

This paper reviews the research progress of supercapacitors (SCs), including the influence of electrode materials on energy storage mechanism and performance, and life prediction. Supercapacitors show application potential in many fields due to their high-power density, fast charge–discharge capability, long cycle life, and environmental protection characteristics. In this paper, the energy storage mechanism of the double-layer capacitor, pseudocapacitor, and asymmetric supercapacitor are discussed. New electrode materials, such as carbon-based materials, metal oxides, and conductive polymers, are reviewed based on the performance optimization measures that are involved in the microstructure design of electrode materials, and integrate the rule prediction of supercapacitors into comprehensive learning. When designing and using supercapacitors, we should not only pay attention to their life but also pay attention to their remaining service life in real time. The paper also mentions the progress of life prediction technology, which is of great significance to improve the reliability and maintenance efficiency of energy storage equipment, and ensure the long-term stable operation of energy storage systems. Future research directions include increasing energy density, extending life, adapting to extreme environments, developing flexible and wearable devices, intelligent management, and reducing costs.

1. Introduction

In the past half a century, the global economy has grown rapidly, various industries have developed vigorously [1,2,3], and people’s quality of life has gradually improved. Followed by the continuous expansion of the use of energy, the large-scale use of traditional energy has exposed its shortcomings to the public [4,5,6], and the energy crisis has gradually emerged, with the problem of environmental pollution becoming increasingly serious. The research of new energy resources and new energy storage methods has attracted attention, and new energy storage methods have emerged in an endless stream [7]. Compared with many problems of traditional storage methods, such as low energy efficiency, limited life, and environmental pollution, the research of capacitors, new batteries, and supercapacitors has been continuously making headway in new energy storage technologies, becoming an important part of new energy storage. The storage of the battery, as shown in Figure 1A, mainly uses the mutual conversion of electrical and chemical energy, and relies on the continuous chemical reaction of its cathode, anode, and electrolyte to give it a high energy density [8]. In the process of battery discharge, the anode oxidizes and releases electrons, while the cathode absorbs electrons in a reduction reaction. These electrons flow from the anode to the cathode through an external circuit, providing electrical energy. For example, when a lithium-ion battery is discharged, lithium ions are released from the anode and moved through the electrolyte to the cathode, while electrons flow through the external circuit to complete the release of electrical energy. Charging is the opposite, with an external power source forcing electrons from the cathode to the anode and lithium ions moving from the cathode back to the anode. However, battery energy storage also has its disadvantages; with the constant chemical reaction in the battery, it not only results in degradation of the battery, but also leads to its limited life. Furthermore, battery power density is small, the electrolyte is toxic and not environmentally friendly, and so on. The capacitor (Figure 1B) is an energy storage method based on charge storage and electric field forming two-point characteristics. Capacitors usually consist of two plates separated by an insulating material (also known as a medium or electrolyte) and common dielectric materials include ceramics, plastic film, paper, air, etc. The charging process of the capacitor is as follows: when the capacitor is connected to the power supply, the power supply will accumulate charge on the two plates of the capacitor, whereby one plate accumulates positive charge, and the other plate accumulates an equal amount of negative charge, and the accumulation of charge will form an electric field between the two plates. The discharge process is as follows: when the capacitor is connected to the circuit, the capacitor will discharge through the circuit and release the stored electrical energy. During the discharge process, the voltage at both ends of the capacitor is gradually reduced, and the charge is gradually reduced until the capacitor is completely discharged. The capacitor has fast response and high power, which means that the capacitor can be charged and discharged more quickly.

However, compared with the previous two supercapacitors, with their unique storage mechanism and excellent performance indicators, supercapacitors have become an important part of solving existing energy problems [9], featuring a wide temperature range, fast charge and discharge, high power density, and long life [10,11,12,13]. As shown in Figure 2, supercapacitors show great application potential and social value in many fields such as smart grid, new energy vehicles, aerospace, transportation, and portable electronic equipment [14,15,16,17], which rely on their high quality, high safety, and high reliability. The commercial application of supercapacitors is also gradually expanding. Among them, the life of supercapacitors has always been a hot spot of concern.

The prediction of remaining useful life (RUL) is essential to ensure the reliability of supercapacitors. Accurate estimation of the remaining effective life cycle of supercapacitors under actual operating conditions not only helps to optimize maintenance strategies and reduce the risk of unplanned downtime, but also has great significance for reducing the operating cost of the entire system and improving economic benefits. In addition, an in-depth understanding of the key factors that affect the life of supercapacitors, especially the performance and degradation mechanisms of plate materials, provides materials scientists and engineers with directions to optimize and improve product performance. Electrode material, which is the core component of supercapacitor, is directly related to its energy storage efficiency and service life. From activated carbon to advanced composites, from graphene to conductive polymers, the microstructure, surface properties, and chemical stability of materials have a profound impact on the performance of capacitors. The innovative research and development of electrode materials must not only pursue higher electrochemical performance, but also consider its durability and environmental adaptability. In this context, research on supercapacitor plate materials has become the key to improving the overall performance of supercapacitors and realizing their wide application. This paper aims to review the current technical progress of RUL prediction for supercapacitors and discuss in detail how plate materials affect the cyclic stability of supercapacitors. By analyzing the aging mechanism and monitoring technology of supercapacitors, combined with the latest research results, we hope to provide readers with a comprehensive and in-depth understanding to support the future development of supercapacitor technology and ultimately achieve long-term stable applications of this promising energy storage technology.

Literature search and screening were conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The literature data were mainly collected from major academic databases including Web of Science, Scopus, ScienceDirect, and IEEE Xplore. The primary search period covered 2018–2025, with a small number of representative studies published before 2018 also included for reference. Inclusion criteria consisted of studies related to supercapacitor electrode materials, degradation mechanisms, or lifetime prediction; works containing experimental, modeling, or review data; and publications written in English and peer reviewed. Exclusion criteria included papers lacking electrochemical performance- or lifetime-related data, conference abstracts and non-peer-reviewed reports, duplicate records, and studies not directly related to the topic.

After systematic screening, a total of 4769 records were initially identified. After title and abstract screening, 714 irrelevant studies were excluded, and finally, 76 highly relevant papers were included for comprehensive analysis. The entire identification and screening process followed the PRISMA flow, as illustrated in Figure 3.

2. Scope and Objectives

This review aims to provide a comprehensive and integrated overview of the current progress in supercapacitor electrode materials and remaining useful life (RUL) prediction technologies, emphasizing the intrinsic connection between material properties, electrochemical behavior, and lifetime performance. The scope of this paper covers three major aspects. First, it summarizes the working principles and classification of supercapacitors, establishing a theoretical basis for understanding how electrode materials influence energy storage mechanisms. The discussion includes electrochemical double-layer capacitors, pseudocapacitors, and hybrid capacitors, highlighting the differences in charge storage processes and the corresponding material requirements. Second, the paper systematically reviews the development of electrode materials, including carbon-based materials, transition metal oxides, conductive polymers, and emerging materials such as MXenes and black phosphorus. Special attention is given to how their microstructure, porosity, conductivity, and surface chemistry affect charge transport efficiency, capacitance retention, and cyclic stability. This section links material degradation behavior—such as pore collapse, surface oxidation, and loss of conductivity—to the long-term performance decline in supercapacitors. Third, the review focuses on RUL prediction and life modeling, integrating the material degradation mechanisms with data-driven and physics-based approaches. The discussion includes the evolution from empirical and Arrhenius-based models to modern machine learning and deep learning frameworks such as LSTM, CNN, and hybrid networks. By correlating electrochemical aging data with model outputs, these approaches provide quantitative tools for predicting supercapacitor health and remaining life under real operating conditions.

The overarching objective of this review is to bridge the gap between material science and system-level reliability assessment. By linking electrode material optimization with lifetime prediction strategies, the paper aims to guide future research toward designing supercapacitors that are not only high-performing but also durable and predictable in service. This integrated perspective will support the intelligent management of energy storage systems and contribute to the realization of long-life, sustainable, and cost-effective supercapacitor technologies.

3. Working Principle and Types of Supercapacitors

3.1. Structure and Working Principle

A supercapacitor is an efficient energy storage device, usually consisting of two electrodes, an electrolyte, and a porous diaphragm [18,19,20], as shown in Figure 4. The design and material choice of these components is critical, making supercapacitors starkly contrast to conventional batteries and capacitors in terms of performance. The electrodes of supercapacitors generally use materials with high specific surface area, such as activated carbon, carbon nanotubes, or graphene [21,22,23]. These materials not only provide sufficient surface area to store the charge, but also ensure good electrical conductivity, which improves charge and discharge efficiency. The choice of decomposition properties has an important effect on the performance of supercapacitors. The electrolyte can be aqueous, non-aqueous or solid. The aqueous electrolyte is safe and environmentally friendly, but the operating voltage is relatively low. Non-aqueous electrolytes have higher operating voltage and energy density. The diaphragm is located between the positive and negative electrodes, and its main function is to prevent the direct contact of the electrodes from causing a short circuit, while allowing ions to move freely. Diaphragm materials typically have high resistance and good ion conductivity to ensure efficient transportation of ion.

The working principle of supercapacitors is mainly based on electric double layer capacitance and Faraday reaction. When the power supply is connected, an electric field forms between the positive and negative electrodes, causing ions to move through the electrolyte, forming an electric double layer to store charge. In addition, some supercapacitors use electrochemical reactions for energy storage, namely the Faraday reaction. Due to the high specific surface area of the electrode material and the good conductivity of the electrolyte, this design allows the supercapacitor to achieve rapid charge and discharge, combined with a high power density and energy density. According to different energy storage mechanisms, supercapacitors are divided into electrochemical double-layer capacitors, pseudocapacitors, and hybrid capacitors, as shown in

Table 1. Classification of supercapacitors.

Type	Charge Storage Mechanism	Disadvantages	Electrode Materials
Electrochemical Double-Layer Capacitor (EDLC)	Non-Faradaic process (Electrical Double-Layer, EDL)	Low specific capacitance, low energy density	Carbon materials
Pseudocapacitor	Faradaic process (Redox reactions)	Relatively low-rate performance	Metal oxides or polymers
Hybrid Capacitor	Pseudocapacitance + EDL	Structurally complex	Oxidation–reduction reactions of materials and carbon materials

. A electrochemical double-layer capacitor (EDLC) is a kind of energy storage device based on double-layer theory. They store energy through a double electric layer formed at the interface of the electrode and electrolyte and do not involve chemical reactions [24,25,26]. EDLCs usually use activated carbon with a high specific surface area as the electrode material [27] to maximize the double-layer capacity of the electrode surface.

3.2. Classification of Supercapacitors

This type of supercapacitor has extremely high charge and discharge rates and excellent cyclic stability [28], but its energy density is relatively low because they do not involve the internal charge storage mechanism of the electrode material. Pseudocapacitors, also known as Faraday capacitors, have charge storage mechanisms that involve rapid, reversible REDOX reactions on the surface of the electrode material [29]. These reactions typically occur in materials such as transition metal oxides or conductive polymers with variable valence states. Unlike EDLC, charge storage for pseudocapacitors involves Faraday processes, but these reactions mainly occur in the surface or near-surface region of the electrode material. Typical battery-type pseudocapacitors include transition metal sulfides (tms), such as CoS, NiS, Ni₃S₂, Co₃S_4, and FeS₂ [30,31,32,33]. Pseudocapacitors have a high energy density because they use the REDOX activity of the electrode material to store charge but may have limitations in terms of cycle stability and power density. Hybrid capacitors combine the characteristics of EDLC and pseudocapacitors, using different types of electrode materials to achieve higher energy density and power density. One electrode of such a capacitor exhibits primarily electrostatic capacitance and the other primarily electrochemical capacitance, such as a lithium-ion capacitor [34,35]. Hybrid capacitors, with their advantages in terms of energy density, specific capacitance, and life cycle, are becoming the leading candidates for next-generation energy storage devices. They store more energy than conventional supercapacitors while maintaining their charging speed, while having a longer life, higher durability, faster charging and discharging speed, and more environmentally friendly and safe features. Hybrid capacitors can last twice as long as lithium-ion batteries.

4. Supercapacitor Electrode Material and Service Life

4.1. Materials and Life of Supercapacitors

The selection and preparation of electrode materials are crucial for improving the life of supercapacitors (SCs) [36]. The ideal SC electrode should have thermal stability, high specific surface area (SSA), excellent corrosion resistance, good electrical conductivity, stable chemical properties, and suitable surface wettability. In addition, these materials should be cost-effective and environmentally friendly. The charge transport capacity of the electrode material is also a key factor to improve the capacitive performance [37]. In this paper, cyclic steady states will be used to indirectly reflect its life and state. The cyclic stable state is not only affected by the surface area of the electrode, but also by other important parameters, such as pore size distribution, pore shape, pore size, and pore accessibility to the electrolyte [38,39,40]. Therefore, when designing supercapacitors, the two core requirements are as follows: the first is to select electrode materials with high SSA to enhance the electrochemical active site; and the second is to adjust the pore size and shape, such as the circle, vertical rectangle, horizontal rectangle, and square of graphene nanopore, to adapt to different electrolyte ion transport requirements [41,42]. The electrode materials are mainly divided into three categories as shown in Figure 5: carbon materials, transition metal oxides (TMOs), and conductive polymers (CPs) [43]. The specific properties and advantages of these materials will be discussed in detail in the following sections. Through careful design and selection of electrode materials, the performance of supercapacitors can be significantly improved to meet the needs of specific applications.

Carbon materials, especially porous carbon materials, are widely used as electrode materials for supercapacitors because of their high specific surface area and good electrochemical stability. Activated carbon is the most common EDLC electrode material [44], which has the advantages of high specific surface area, low cost, and high chemical and thermal stability [45]. However, activated carbon may undergo structural collapse and loss of surface functional groups during long-term cycling, leading to performance attenuation. To improve the performance of activated carbon, the researchers employed a variety of methods, such as physical or chemical activation, to increase its specific surface area and pore volume. In addition, the template method is also used to prepare porous carbon materials with controllable aperture distribution and interconnected pore networks, thereby improving their performance in supercapacitors.

Transition metal oxides play an important role in the electrode materials of supercapacitors, where they are favored for their high specific capacitance and conductivity. These materials include mono-metal oxides such as RuO₂, MnO₂, Co₃O₄ [46], Ni(OH)₂ [47], as well as polymetallic oxides, which generally exhibit better electrical conductivity and cycle stability [48]. Transition metal oxides have the advantages of large natural abundance, abundant valence states, and easy design and manufacture, which make them have a wide range of application potential in the field of energy storage. However, they also have some challenges, such as short cycle life and low electronic conductivity, which limit their performance in practical applications. To overcome these limitations, researchers are actively exploring a variety of strategies, including tuning electronic structures to optimize electrochemical performance, developing composites such as carbon materials to improve structural stability, designing hollow structures to increase surface area, and improving surface chemistry to enhance energy storage capacity. In addition, multiple metal oxides are explored to provide more complex valence states and reaction routes using multiple metal reaction sites [49]. These research directions are aimed at improving the performance of transition metal oxides in supercapacitors to achieve more efficient and stable energy storage solutions.

As electrode materials for supercapacitors, conductive polymers have attracted wide attention due to their excellent electrical conductivity, good flexibility, and easy processing. Common conductive polymers include polyaniline (PANi), polypyrrole (PPy), and polythiophene (PTh), which can significantly increase the specific capacitance and energy density of supercapacitors. In addition, conductive polymers have good chemical stability and mechanical strength, which can enhance the durability and reliability of supercapacitors to a certain extent.

4.2. Carbon-Based Materials

Carbon-based materials play an important role in supercapacitors because of their unique structure and excellent electrochemical properties.

In the field of biomass-derived carbon materials, Lin [50] employed water spinach as a precursor to synthesize bio-inspired hierarchical porous carbon. The natural multiscale channels of the plant were preserved after high-temperature carbonization, forming interconnected micro–mesoporous networks that significantly enhanced ion diffusion and charge storage efficiency, demonstrating the inherent advantages of plant templates in constructing hierarchical pore structures. Zhang [51] further combined wood fibers with the metal–organic framework ZIF-8 to fabricate N-doped hierarchical porous carbon fibers (CWFZ2). The material exhibited a high specific surface area of 593.52 m² g⁻¹ and delivered a specific capacitance of 270.74 F g⁻¹ at 0.5 A g⁻¹ with 98.4% retention after 10,000 cycles, which was attributed to the synergistic effects between the conductive network and abundant active sites. Similarly, Ou et al. prepared N,O co-doped porous carbon (BCPC-3) [52] from buckwheat husks via one-step KOH activation, achieving a high surface area of 805.91 m² g⁻¹ and pore volume of 0.60 cm³ g⁻¹. The material delivered 330 F g⁻¹ at 0.5 A g⁻¹ and maintained 140 F g⁻¹ even at 100 A g⁻¹, while a symmetric supercapacitor assembled with BCPC-3 achieved an energy density of 6.1 Wh kg⁻¹. These results indicate that rational pore architecture and heteroatom-doping can effectively enhance ion transport and charge storage capability. In addition to biomass-derived carbons, Pati [53] synthesized high-porosity-activated carbon (PAC) from petroleum coke through KOH activation, achieving a specific surface area of 2105.6 m² g⁻¹ and an average pore size of 1.78 nm. The PAC electrode exhibited a capacitance of 470 F g⁻¹ at 0.5 A g⁻¹ with 98% retention at 10 A g⁻¹, benefiting from the abundant micropores and graphitic layers that facilitated both conductivity and ion diffusion. Overall, the selection of carbon precursors, regulation of pore structures, and heteroatom-doping are crucial strategies for optimizing carbon-based electrode performance, offering valuable guidance for the design of green and high-performance supercapacitor materials.

4.3. Metal Oxides

Metal oxides are considered as promising electrode materials for supercapacitors due to their high theoretical capacitance and environmental friendliness.

In the Nd-doped AlFeO₃ electrode prepared by the hydrothermal method, Somaily [54] found that the introduction of rare-earth ions effectively regulated the crystal structure and electronic distribution, significantly enhancing ion transport and electrical conductivity, achieving a high specific capacitance of 1326 F·g⁻¹ with excellent cycling stability. Liu [55] constructed a P-doped Ni₀.₅Cu₀.₅Co₂O₄ hollow-sphere structure that combines multi-metal synergy with non-metal doping, which improved electron mobility and shortened ion diffusion paths, resulting in a high energy density of 118.6 Wh·kg⁻¹ in the assembled device. Furthermore, Pawar [56] employed atomic layer deposition to coat ultrathin MoS₂ layers on NiCo₂O₄ nanosheets, forming a highly conductive oxide/sulfide heterostructure that effectively enhanced electron transport and rate capability. Meanwhile, Wu [57] synthesized Mn–Co binary metal oxides by tuning the Mn/Co ratio, achieving excellent electrochemical activity and flexible stability, with the quasi-solid-state device maintaining high capacity even at elevated current densities. Imran [58] incorporated MnO₂ into a Zn/Ni–MOF framework, where the porous channels of the MOF promoted ion diffusion, and the redox behavior of MnO₂ further improved the overall energy storage performance.

MnO₂ is a transition metal oxide with a variety of valence states, with a variety of crystal structures, such as α-MnO₂, β-MnO_2, and γ-MnO₂. MnO₂, which can be prepared by hydrothermal synthesis, sol–gel method, or electrochemical deposition. Lu et al. prepared Mn₂O₃@MnO₂ composite nanofibers through hydrothermal and electrospinning methods [59]. The MnO₂ electrode showed a capacitance retention of 86.5% in KOH electrolyte, indicating good cyclic stability after 5000 cycles.

4.4. Conductive Polymers

Conducting polymers such as polyaniline (PANI), polypyrrole (PPy), and polythiophene (PTh) have shown great potential in supercapacitors due to their high electrical conductivity and pseudocapacitive behavior. Peringath [60] chemically synthesized PANI and PTh films, where the PANI electrode exhibited a specific capacitance of 13.22 mF·cm⁻² in H₃PO₄ electrolyte, significantly outperforming PTh. Liu [61] employed a frozen interfacial polymerization method to grow PANI nanoarrays on holey graphene, constructing a binder-free flexible gel electrode with a high capacitance of 793.7 F·g⁻¹ and 90.5% retention after 5000 cycles. Tawade [62] developed a 2D reduced graphene oxide–polythiophene (rGO–PTs) nanocomposite via a swollen liquid crystalline lamellar mesophase, achieving an excellent capacitance of 1412 F·g⁻¹ and remarkable durability. Hong [63] further optimized the α–α coupling configuration of PPy through frozen interfacial polymerization, obtaining a highly conjugated film that maintained over 200 F·g⁻¹ even after 60,000 cycles. Overall, structural regulation and carbon-based hybridization of polymer electrodes significantly enhance their electrochemical performance and cycling stability, offering new design strategies for high-performance and flexible energy storage devices.

4.5. New Materials and Structures

The development of new materials and structures provides new possibilities for the performance improvement of supercapacitors.

Two-dimensional transition metal carbides (MXenes) are a new class of two-dimensional materials consisting of transition metal carbides, nitrides, or carbon nitrides. These materials are prepared by selective etching of element A in the MAX phase. Syamsai et al. synthesized the tantalum carbide MXene and tested its electrochemical properties in the H₂SO₄ electrolyte [64]. MXenes’ high conductivity and large specific surface area make it exhibit good cycle stability in supercapacitors.

Black phosphorus (BP) is a two-dimensional material with a layered structure that can be prepared from black phosphorus crystals by liquid phase stripping. Hao et al. used this method to prepare BP nanosheets and flexible asymmetric supercapacitors in a PVa-based gel electrolyte [65]. This BP-based supercapacitor has a capacitance retention rate of 96.5% after 30,000 cycles, showing excellent cycle stability.

Hybrid electrode materials have demonstrated remarkable potential in enhancing the energy density and cycling stability of supercapacitors through multi-component synergy. Choudhary developed a flexible and conductive nanocellulose/carbon nanotube (MACNC/CNT) composite electrode derived from agricultural waste, achieving an area capacitance of 1389.2 mF cm⁻² with 74.6% retention after 12,000 cycles, highlighting its green and sustainable advantages [66]. Deng [67] designed an MoO₃/MWCNT-COOH/P5ICA ternary hybrid that integrates the pseudocapacitance of metal oxides, the conductivity of carbon nanotubes, and the redox reversibility of conductive polymers, delivering a high specific capacitance of 166.7 mF cm⁻² and excellent cycling stability. Similarly, a CuO/C composite electrode prepared from cellulose aerogel exhibited a specific capacitance of 1001 F g⁻¹ at 2 A g⁻¹ and an energy density of 139 Wh kg⁻¹, demonstrating the effectiveness of its 3D porous framework in facilitating ion and electron transport [68]. In another study, potassium-gluconate-activated bacterial cellulose-derived porous carbon (APC-700) achieved a capacitance of 437 F g⁻¹ at 0.5 A g⁻¹ and an energy density of 23.5 Wh kg⁻¹, benefiting from its hierarchical porosity and oxygen-rich surface functionalities [69]. Moreover, Shan et al. utilized the natural vertical channels of wood to fabricate an N-doped carbon/MnO₂ composite electrode with a capacitance of 162.4 F g⁻¹ and superior rate capability [70]. Overall, hybrid materials effectively integrate the advantages of metal oxides, conductive polymers, and carbon frameworks, achieving synergistic enhancements in electrochemical performance and providing a promising route toward high-performance and sustainable supercapacitors.

In addition to the materials listed in

Table 2. Comparison table of supercapacitor performance for different electrode materials.

Electrode Material/Material System	Specific Capacitance	Cycling Stability	Cycle Number	Electrolyte	Ref.
Nd-AlFeO3	1326 F·g−1 @ 1 A·g−1	91%	5000	6 M KOH	[54]
P-Ni0.5Cu0.5Co2O4	1570.2 F·g−1 @ 1 A·g−1	78%	5000	3 M KOH	[55]
NiCo2O4-MoS2	2445 mF·cm−2	—	—	2 M KOH	[56]
Mn-Co Oxide	4.48 F·cm−2 @ 10 mA·cm−2	85%	4000	PVA-KOH Gel	[57]
MnO2@Zn/Ni-MOF	1537 F·g−1 @ 2 A·g−1	89%	4000	6 M KOH	[58]
MACNC/CNT	1389.2 mF·cm−2 @ 0.02 A·cm−2	74.6%	12,000	—	[66]
PANI/HG-10	793.7 @ 1 A·g−1	90.5%	5000	1 M H2SO4 (Gel State)	[60]
rGO-PTs nanocomposite	293 @ 0.5 A·g−1	99.4%	3000	2 M HCl + PVA Gel	[60]
N-doped wood fiber carbon	270.7 @ 0.5 A·g−1	98.4%	10,000	6 M KOH	[69]
Buckwheat-derived porous carbon	660 mF·cm−2 (≈300 F·g−1)	Excellent	>5000	6 M KOH	[52]
Petroleum coke-activated carbon	470 @ 0.5 A·g−1	98%	15,000	3–6 M KOH	[53]
ZIF-8-derived porous carbon	130–250	Excellent	—	6 M KOH	[51]

, there are other carbon-based superstructural materials, such as carbon aerogel, carbon nanofibers, and template carbon, which have also been widely studied for supercapacitors. These materials typically have a high specific surface area and a porous structure that helps improve the capacitive properties of the electrode. For example, carbon materials prepared by the template method can be precisely controlled to optimize their performance in supercapacitors by controlling their pore size and structure.

5. Remaining Service Life of a Supercapacitor

5.1. Remaining Service Life of Supercapacitors

Although supercapacitors have a longer life expectancy than other energy devices due to their unique energy storage mechanism, their actual life is still significantly affected by external conditions. Key aging factors include electrical loads (such as voltage and current levels) and thermal loads (such as operating temperatures). Supercapacitors are constructed with electrodes, electrolytes, diaphragms, and fluid collectors, and signs of aging typically include damage to the housing, breakdown of the electrolyte, and a decline in the performance of the electrode material. In some cases, the gas produced by the decomposition of the electrolyte may lead to an increase in internal pressure, which may trigger the rupture of the housing under prolonged use or extreme conditions. These aging effects can be effectively mitigated by choosing a stronger housing material or introducing a pressure release mechanism.

The high-temperature environment accelerates the chemical reaction, which accelerates the aging process of the activated carbon electrode. Spontaneous temperature increases and temperature differences associated with root-mean-square current also affect the aging of supercapacitors. Performance decline is mainly due to the degradation of activated carbon electrode, whereby oxidation and reactivation reaction may occur on its surface, destroying its original structure. Over time, the pores of the electrode can become clogged with by-products, leading to pore shrinkage, and a reduction in specific surface area, which reduces its energy storage capacity. In addition, the decomposition voltage of the electrolyte limits the upper limit of the operating voltage of the supercapacitor and affects parameters such as current density and operating temperature. Impurities from the decomposition of the electrolyte also reduce the ability of ions to migrate to the pores of the electrode, increasing the equivalent series resistance (ESR).

Meanwhile, performance parameters such as power density, charge–discharge capability, and cycling stability have a significant influence on the lifetime estimation of supercapacitor-based hybrid systems. Power density and charge–discharge ability not only reflect the ion and electron transport efficiency within the electrode–electrolyte interface but also represent the material’s capacity to maintain structural integrity and electrochemical reversibility under high-rate or high-current conditions. Materials with higher power density and faster charge–discharge response tend to exhibit lower internal polarization and heat accumulation, thereby reducing electrochemical stress and mechanical fatigue during repeated cycling. In contrast, poor charge–discharge kinetics can lead to uneven current distribution and localized overpotential, accelerating active material degradation and shortening the overall service life of the device.

Cycling stability, often expressed as capacity or capacitance retention after thousands of charges–discharge cycles, serves as a direct indicator of the durability and reliability of the electrode material. It provides crucial information for evaluating the long-term degradation trends of hybrid energy storage systems. From a modeling perspective, these parameters collectively determine both the transient dynamic performance and the long-term degradation kinetics of the device. They are therefore essential input features for lifetime prediction models—whether empirical, physics-based, or data-driven—significantly improving the accuracy and robustness of remaining useful life (RUL) estimation. By quantitatively correlating electrochemical performance data with degradation behavior, researchers can establish a more comprehensive understanding of the intrinsic relationship between material characteristics, operational stresses, and lifecycle evolution in supercapacitor-based hybrid systems.

Therefore, these aging factors work together to cause the remaining useful life (RUL) of the supercapacitor to gradually decrease along a non-linear trajectory until it reaches its useful life. It is important to study and predict the RUL of supercapacitors to ensure their long-term safety and reliability. A deep understanding of these aging mechanisms can provide guidance for the design, use, and maintenance of supercapacitors to extend their service life and optimize their performance in a variety of applications.

5.2. Prediction of Remaining Useful Life

In the health management of supercapacitors, accurate estimation of their remaining useful life (RUL) is essential for preventive maintenance and improved system reliability. In recent years, researchers have developed a variety of RUL estimation methods based on different theories and techniques. These approaches can be broadly divided into two categories: physical model-based approaches and data-driven approaches.

Xu et al. [71] proposed an RUL estimation method based on the Arrhenius model, which takes into account key parameters such as temperature, current intensity, and cycle number. The Arrhenius model is a classical chemical reaction kinetic model, which de-scribes the attenuation function of supercapacitor capacity with temperature change. The model parameters are determined by the experimental data, and the model can predict the capacity attenuation of supercapacitors under different conditions. The results show that the curve predicted by the model is in good agreement with the experimental data, and the error range is less than 3%. However, the applicability of this model is limited to the temperature range of 25–55 °C, which limits its application in a wider range of environmental conditions. But the dataset is relatively small and lacks diversity, and no clear cross-validation is reported, which may affect model generalization.

Weigert et al. [72] developed a three-layer BPNN model to estimate the cycle life of battery–supercapacitor systems. The RUL estimation method is based on short charge and discharge curve data, and the model hyperparameters are selected by the trial and error method. By adding the voltage on the discharge curve as an input, the error of the estimated result is reduced to less than 4%. This method is simple and easy to implement, but its accuracy depends on the quality and quantity of data. The study lacks detailed information about the dataset scale and stress diversity, such as variations in temperature, current rate, or cycling protocols. Moreover, there is no explicit mention of using cross-validation or independent testing, which may raise concerns about data leakage and model overfitting under different operating conditions.

Zhou et al. [73] developed a RUL estimation method using the long short-term memory (LSTM) network, as shown in Figure 6. LSTM networks are particularly well suited for working with time series data and capturing long-term dependencies. The method considers the number of charge and discharge cycles as inputs to the LSTM model and improves the prediction accuracy by optimizing the model hyperparameters. Compared with other RNN models, the LSTM model has achieved satisfactory results in RUL prediction, showing its effectiveness in processing supercapacitor degradation data. However, the computational cost of the model is relatively high, and the lack of uncertainty quantification, such as confidence intervals or model calibration, limits its interpretability. In addition, the performance of the model under domain shift—such as different temperature ranges or supercapacitor types—has not been fully verified.

Zhou et al. [74] further proposed a hybrid technique combining genetic algorithm (GA) and LSTM to estimate RUL for supercapacitors. GA optimization technology is used to improve the global search ability and quickly find the optimal solution. This hybrid model achieves an error of less than 1.61% in both steady state and hybrid pulse power characteristic modes, showing high prediction accuracy. However, this hybrid model has higher computational complexity and requires more computational resources.

Haris et al. [75] developed a hybrid technique combining deep belief networks (DBNs) with Bayesian optimization and HyperBand (BOHB) to estimate the RUL of supercapacitors. The BOHB algorithm is used to optimize the model training process and reduce the amount of training data required. This hybrid model not only increases the speed of the prediction network, but also reduces the need for training data. Compared to Bayesian optimization and HyperBand techniques, the BOHB algorithm has a 77% increase in predictive network speed and uses only 6% of the data as a training dataset, which is especially valuable when data is scarce. However, the study provides limited information on data diversity and validation, which may affect the assessment of model robustness.

Wangkai et al. [76], in this innovative study, propose a new method that combines one-dimensional convolutional neural networks (1DCNNs) and an improved Informer model for improving the prediction accuracy and robustness of the remaining useful life (RUL) of supercapacitors in Figure 7.

In the data pre-processing stage, the researchers adopted the minimum–maximum feature scaling method to normalize the data and ensure that the model can effectively process data from different sources and structures. The formula is as follows:

$$X_{\mathrm{normalized}}={\displaystyle \frac{X-X_{\min }}{X_{\max }-X_{\min }}}$$

(1)

In the preceding command, X is the original data; X_normalized, X_min, and X_max are data normalization and maximum values in the dataset, respectively. This method scales the data to the range of zero to one through the following formula, which helps to improve the stability and efficiency of model training. In addition, the researchers also introduced the cross-entropy-loss function to couple the 1DCNN and Informer models to ensure effective cooperation between the two. The formula is as follows:

$$Cross-EntropyLoss=-{\sum }_i{y}_{i}log\left(p_i\right)$$

(2)

where Cross is cross entropy, EntropyLoss refers to cross entropy loss, and y_i is the distribution of real labels. In the improvement of the model, the relative position coding algorithm is introduced to improve the Informer model, so that it can capture the relationship between data series more effectively and reduce the uncertainty of prediction. The formula for relative position coding is as follows:

$$PE(p,2i)=\mathrm{s}\mathrm{i}\mathrm{n}({\displaystyle \frac{p}{{\mathrm{10,000}}^{2i}/d_{\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l}}}})$$

(3)

$$PE(p,2i+1)=\mathrm{c}\mathrm{o}\mathrm{s}({\displaystyle \frac{p}{{\mathrm{10,000}}^{2i}/d_{\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l}}}})$$

(4)

where p is the location of the input data, i is the dimension, and d_model is the dimension of the model. The 1DCNN model extracts local features through the convolutional layer, and its formula can be expressed as follows:

$$\mathrm{c}\mathrm{n}\mathrm{n}\left(\mathrm{k}\right)=\mathrm{\sigma }(\mathrm{W}\ast \mathrm{x}+\mathrm{b})$$

(5)

where cnn(k) is the output of the convolutional neural network at layer k, W is the weight matrix of the convolutional kernel, x is the input data, b is the bias term, ∗ is the convolution operation, and σ is the activation function. Informer model captures global dependency through self-attention mechanism, and its self-attention is calculated as follows:

$$Attention(W,K,V)=\mathrm{softmax} {\displaystyle \frac{Q{K}^T}{\sqrt{d_k}}} V$$

(6)

where the softmax function is used to normalize the weights, K is the key matrix, and V is the value matrix. In the process of model training and optimization, the Adam optimizer is used to minimize the loss function, and its update rule can be expressed as follows:

$${\theta }_{t+1}={\theta }_t-\alpha \cdot {\nabla }_{\theta }\cdot Loss$$

(7)

where θ represents the model parameter, α is the learning rate, ∇_θ represents the gradient to the parameter, and Loss is the loss function. The combination of these key steps and formulas results in significant improvements over prior art in several evaluation indicators, including a 32.71% reduction in the maximum root mean square error (RMSE), a 28.50% reduction in the mean absolute error (MAE), and a 4.79% increase in the coefficient of determination (R²). These results demonstrate the advantages of the proposed model in combining the local feature extraction capability of 1DCNN with the global sequence understanding capability of the Informer model, which can provide more accurate and stable RUL prediction.

Shen et al. [77] proposed a prediction model based on BO-BiLSTM, as shown in Figure 8. The model uses a long sliding window to process historical capacity data to enhance the model’s ability to capture capacity decay trend. The hyperparameters of the model were optimized by Bayesian optimization algorithm, and the accuracy of prediction was further improved. In the data pre-processing stage, the researchers adopted an effective data adjustment method to make it suitable for model input.

As an efficient global optimization algorithm, Bayesian optimization is used to optimize the hyperparameters of the BiLSTM model in this study. By establishing a probabilistic model of hyperparameters, Bayesian optimization can efficiently search for the optimal parameter combination and avoid the inefficiency of traditional grid search or random search. Its core formula is as follows:

$$P(f|D)~\varsigma \rho (m(k),k(x,x^{\prime }))$$

(8)

where P(f∣D) is the posterior probability distribution of the function f given data D, and GP is a Gaussian process, which is a probability distribution. In terms of model structure, the BO-BiLSTM model uses a bidirectional long short-term memory neural network (BiLSTM) to capture the back-and-forth correlation of time series data. The cell status update formula of BiLSTM is as follows:

$$C_t=f_t{C}_{t-1}+i_t\hspace{0.33em}\widetilde{C_t}$$

(9)

C_t is the cell state of the time step t; ft is the forgetting gate, which determines the previous cell state to be retained or forgotten; and C_t−1 is the cell state of the previous time step. The final output is determined by the output gate and the cell state, and the formula is as follows:

$$h_t=o_{t}anh(C_t)$$

(10)

where h_t is the hidden layer state of time step t. O_t is the output gate that determines which part of the cell state will be exported.

The performance of the model was evaluated by root mean square error (RMSE) and percentage absolute error (AEP). The experimental results show that the optimized BO-BiLSTM model has achieved significant improvement in prediction accuracy, and RMSE and AEP are reduced to 2.16% and 0.59%, respectively, showing excellent prediction performance.

As shown in

Table 3. Comparison table of input features and performance metrics for supercapacitor performance prediction models (↑ indicates an increase, ↓ indicates a reduction).

Model	Input Features	Dataset Source	Performance Metrics	Ref.
Arrhenius Model	Temperature, current intensity, cycle number	Experimental data (25–55 °C)	Error < 3%	[71]
3-layer BPNN	Short-term charge–discharge curve, voltage	Battery–supercapacitor hybrid experimental data	Error < 4%	[72]
LSTM Network	Charge–discharge cycle number	Supercapacitor degradation experimental data	—	[73]
GA–LSTM	Temperature, voltage, cycle number	Steady-state and HPPC experimental data	RMSE = 0.0161–0.0264	[74]
DBN + BOHB	—	—	Prediction speed ↑ 77%, training data ↓ to 6%	[75]
1D CNN + Improved Informer	Voltage, current, temperature (normalized data)	—	RMSE ↓ 32.71%, MAE ↓ 28.50%, R2 ↑ 4.79%	[76]
BO–Bist	Historical capacity data	Supercapacitor experimental data	RMSE = 2.16%, AEP = 0.59%	[77]

, the development of these methods not only improves prediction accuracy, but also enhances the adaptability and robustness of the model to the data under different working conditions, which has important practical application value for the maintenance strategy of supercapacitors and system reliability management. Future research will continue to explore more factors affecting the life of supercapacitors and develop more efficient algorithms to process larger datasets, providing more comprehensive solutions for the health management of supercapacitors and ensuring the efficient and stable operation of energy storage systems. Future research could further explore more factors that affect the life of supercapacitors, such as ambient temperature, operating current, etc., and develop more efficient algorithms to handle larger datasets.

6. Future Prospects

As an energy storage device between traditional capacitors and rechargeable batteries, supercapacitors (SCs) show great application potential in new energy vehicles, smart grids, renewable energy storage, and other fields due to their unique high power density, fast charge–discharge capability, long cycle life, and environmental protection characteristics. As a key factor in determining the performance of supercapacitors, the research progress of electrode materials has promoted the development of technology, especially the application of carbon-based materials, metal oxides, conductive polymers, etc. With the development of machine learning technology, the progress of life prediction technology has improved the reliability and maintenance efficiency of energy storage equipment.

In the future, supercapacitor research will focus on improving energy density, extending life, adapting to extreme environments, the development of flexible and wearable devices, intelligent management, and reducing costs. The development of new materials and the improvement of existing materials will increase the energy density and bring supercapacitors closer to the level of batteries. In-depth research on the aging mechanism of electrode materials and factors affecting performance decline will help to develop supercapacitors with longer life, reduce replacement frequency, and reduce maintenance costs. At the same time, the research of composite materials, such as combining conductive materials with flexible polymer substrates, will promote the development of flexible supercapacitors.

The study of environmental adaptability will enable supercapacitors to work under extreme conditions such as high temperature, low temperature, and high humidity, and expand their application range. Intelligence management combined with the Internet of Things and artificial intelligence technology will achieve real-time monitoring, fault prediction, and automatic maintenance of supercapacitors, improving the convenience and reliability of use. With large-scale production and material innovation, the cost of supercapacitors will gradually decrease, further promoting their wide application in various fields.

In general, the research and application of supercapacitors have broad prospects, and with the development of new materials and the application of new technologies, supercapacitors will play a more important role in the future energy storage field, providing strong support for sustainable development and energy transformation. With the rapid development of science and technology, the high efficiency, lightweight, and miniaturization of energy storage modules will become the constant theme, and the functional integration of supercapacitors will show great application potential in flexible energy storage, wearable devices, micro and nano devices, etc.

7. Conclusions

The research of supercapacitors (SCs) has made significant progress in the past few decades, especially in electrode materials, energy storage mechanisms, performance optimization, and lifetime prediction. Researchers are constantly exploring new electrode materials, such as carbon-based materials, metal oxides, and conductive polymers, to improve energy density and power density. A deeper understanding of the energy storage mechanisms of double-layer capacitors, pseudocapacitors, and asymmetric supercapacitors has enabled scientists to design more efficient electrode structures and more stable electrolyte systems. The performance optimization measures include the microstructure design of the electrode material, the selection of electrolyte and the interface optimization between the electrode and the electrolyte. These measures enable the supercapacitor to improve the energy density while maintaining a high power density.

With advances in machine learning and data science, life prediction technology has become more accurate, helping to predict the remaining useful life of supercapacitors, providing a scientific basis for maintenance and replacement. The research on the performance of supercapacitors in extreme environment provides the possibility for their application in a wide range of fields. The expansion of production scale and the development of new materials have led to a gradual reduction in the cost of supercapacitors, increasing their viability in commercial applications. The application of the Internet of Things and artificial intelligence technology makes the management and maintenance of supercapacitors more intelligent, improving the reliability and maintenance efficiency of the system. Overall, the research of supercapacitors covers innovation and optimization at the application level from basic material science to electrochemistry, laying a solid foundation for the wide application of supercapacitors in the field of energy storage and conversion, and indicating that they will play a more critical role in the future energy system.