Lithium-Ion Capacitors: A Review of Design and Active Materials

Abstract

Lithium-ion capacitors (LICs) have gained significant attention in recent years for their increased energy density without altering their power density. LICs achieve higher capacitance than traditional supercapacitors due to their hybrid battery electrode and subsequent higher voltage. This is due to the asymmetric action of LICs, which serves as an enhancer of traditional supercapacitors. This culminates in the potential for pollution-free, long-lasting, and efficient energy-storing that is required to realise a renewable energy future. This review article offers an analysis of recent progress in the production of LIC electrode active materials, requirements and performance. In-situ hybridisation and ex-situ recombination of composite materials comprising a wide variety of active constituents is also addressed. The possible challenges and opportunities for future research based on LICs in energy applications are also discussed.

1. Introduction

Due to the rapid growth of renewable energy production recently [1], there is a significant requirement for electrical energy storage technologies. Energy storage offers the ability to moderate the variability of electrical energy [2]. This represents a rapidly emerging market for energy storage that is currently underutilised. The characteristics of the energy storage needs, in general, are electro-compatibility and will relate more specifically to cheap and highly efficient storage solutions for stationary purposes, and energy and power intensity for the transport and transmission sector (to which lithium-based batteries are the answer). In addition, there is a need for shifting the current battery production from fossil-based energy to renewables to reduce the embedded emissions of energy storage systems. Moreover, the materials required in the production of energy storage system should ideally originate from areas free of geopolitical conflicts, child labour, corruption and environment unfriendly excavation and extraction methods [3]. In this light, lithium-ion batteries (LIBs) utilising ethically mined materials and energy produced by renewables have huge international market advantages when considering environmental, social and corporate governance (ESG) aspects.

Lithium-ion capacitors (LICs) were first produced in 2001 by Amatucci et al. [4]. LICs are considered one of the most effective devices for storing energy and are often seen as an offspring from LIBs for several reasons. In addition, Sodium-ion and Potassium-ion capacitors (SIC and KIC, respectively), have also become of commercial interest as they similarly are a hybrid device combining an ion battery with a traditional capacitor. This review will focus on the LIC developments as the main example of a hybrid capacitor, but it must be noted that there are many similarities between LICs, NICs and KICs. LIBs normally have high energy density (>150 W h kg⁻¹) and have no memory impact as in conventional Ni-Cd/Ni-MH batteries [5,6]. Despite this, their low power density (<1 kW kg⁻¹), and lower cycle life and capacity loss [7] hinders their utilisation in some applications. By comparison, LICs can provide high power density (>10 kW kg⁻¹) and long cycle life (usually >5000 cycles); however, their comparatively low energy density (5–10 W h kg⁻¹) restricts their applications in certain fields (LIC applications have been thoroughly discussed in [8,9,10]). To close the performance gap between LIBs and traditional capacitors, LICs have been developed to incorporate the strengths of both LIBs and traditional capacitors [6,11,12,13,14].

By practice, LIBs consist of a metal oxide cathode, separator, electrolyte, and a Lithium-based anode. In contrast, non-aqueous liquid electrolyte LICs with high power densities (>10 kW kg⁻¹) and long-term cyclic durability (10,000–100,000 cycles), are ideal for large-scale energy storage [5,15]. Typical LIC designs use a high-capacity battery-type electrode and a high-rate capacitor-type electrode [5,16]. During the charge-discharge cycle, charges are deposited concurrently and asymmetrically in the LIC by surface ion adsorption/desorption on the capacitor-type electrode and Li⁺ intercalation/de-intercalation on the battery-type electrode, respectively [17,18]. It is worth noting that the battery- and capacitor-type electrodes of the LIC system work in various potential windows, which may expand the range of operating voltage of LICs and contribute to high energy density.

Supercapacitors can be classified into asymmetric supercapacitors, pseudo-capacitors (PCs) and electric double-layer capacitors (EDLCs) depending on the charging storage mechanisms [19]. PCs store energy using the fast redox reaction on the surface of the electrode components, such as metal oxides, metal sulphides and polymer conductors. Whereas energy storage on the surface of the porous carbon with a large specific surface area (SSA) is accomplished through strong ion absorption and desorption for EDLCs.

LIBs operate through Li⁺ that travels back and forth between the electrodes followed by the electrode materials’ bulk redox reaction. The combination of the positive EDLC electrode and the negative LIB electrode forms a LIC, reducing the deficiencies in both LICs and LIBs. LICs may typically be categorised as dual-carbon, non-carbon, and mixed forms, distinguished by whether or not the electrodes contain carbon materials. The densities of energy and power of LICs depend primarily on the configuration of electrode materials in the devices [17,20]. The most promising industrial prospects are for dual-carbon LICs (DC-LICs), in which both electrodes are composed of carbon materials. Owing to the surplus and low carbon levels, the average expense of DC-LICs may be effectively popular.

Until now, owing to their strong gravimetric functional potential and excellent electrochemical operation, a number of components, such as metal compounds, polyanions and metalloid/metal compounds, have been produced in LICs for battery-type electrodes [21]. However, their further production in LICs is constrained by the poor conductivity, wide volume variability and strong polarisation of these active materials [5]. Carbonaceous materials have been extensively researched and used in energy storage fields like LICs [22,23,24,25], benefiting from low expense, sufficient supplies, plenty of allotropes and transformations as well as superior physical/chemical stability. Because of their large specific surface region, strong conductivity and excellent usability to electrolytes, carbon materials were also inserted into the electrodes to solve such issues [26,27]. Furthermore, carbon compounds may also specifically function as the active compounds of battery-type electrodes in LICs, since they have active Li⁺ intercalation/de-intercalation sites [28,29].

Unlike battery-type electrodes, LICs counter-electrodes are capacitor-type. Consequently, numerous porous carbon materials with a wide specific surface region, such as activated carbon (AC) and graphene are potential candidates for LICs capacitor-type electrodes [30,31]. Their capacitances rely primarily on the surface of carbon-based electrodes [32] for the adsorption/desorption of ions. Therefore, porous structures with sufficient distribution of pore size play an important role in the electrochemical production of carbon materials in LICs [33,34]. Specific materials are typically inserted into the carbonaceous materials to have pseudo-capacity to further improve the capacitance of the capacitor-type electrode.

In this review, we will identify the fundamental active materials used for LICs and analyse electrode/capacitor-type carbon-based battery-based electrode materials over many decades. By concentrating on the most common cases with industrial potential, we demonstrate the main components of LICs. The knowledge gaps and future trends associated with LICs are also addressed, which provides a detailed insight into the potential areas of further research of LIC active materials.

2. Lithium-Ion Capacitor Fundamentals

Throughout the consumer electronics, automobile, aerospace and stationary industries, electrical energy-storage devices play a key function. There are several types of reversible electrochemical energy storage systems in the format of accumulators, flow cell systems, secondary batteries (rechargeable) and primary batteries (single-use). In terms of secondary and primary batteries, the former has a lower energy intensity while the latter offers better power; however, the cyclable nature of secondary batteries makes them ideal for use as an energy storage solution. The current prevailing technology is LIBs. Large LIBs have a gravimetric energy upwards of 200 Wh kg⁻¹, and with an overall effective power density of up to 350 W kg⁻¹. In comparison, most industrial electrochemical capacitors have average power density as high as 10 kW kg⁻¹, and with gravimetric energy of up to 7 Wh kg⁻¹ [2]. Figure 1 gives an overview of the power and energy densities of ultracapacitors and LIBs compared to other energy storage technologies. A common target for advanced electrical energy storage systems is to provide high energy as well as high power in a single system [35,36,37,38]. A LIC is a comparatively modern system, intermediate in energy between batteries and supercapacitors, though giving supercapacitor-like power and cyclability properties.

It is important to define the components of LICs before recognising the basic energy storage function of the LICs. As a kind of asymmetric supercapacitor, LICs usually consist of a battery-type electrode with the insertion/extraction of lithium ions and a pseudo-capacitance or ion adsorption/desorption capacitor-type electrode [39,40]. As the battery-type electrode does not only serve as an anode but also as a cathode, LICs have been previously separated into two types [41]:

• The battery-like electrode acts as the anode, and the capacitor-like electrode acts as the cathode. Anions are usually absorbed on the porous surface or defects that may be apparent on the cathode during the charge cycle, whereas Li⁺ ions are intercalated (as Li) in the active anode.
• The capacitor-type electrode functions as the anode and the battery-type electrode functions as the cathode. Li⁺ de-intercalates from the cathode during the charging phase. Li⁺ immediately migrates in the electrolyte and adsorbs on the anode. This requires some redox reaction at the cathode (e.g., cathode oxidation: LiCoO₂ → Li⁺ + e⁻ + Li⁻-CoO₂; and, anode reduction: Li⁺+C₆-graphite + e⁻ → LiC₆-graphite).

Regenerative braking and grid stabilisers are significant possible end-uses of the LICs. Regenerative braking energy recovery from cars, heavy-duty engines, and increasingly light-duty vehicles represents a major potential opportunity that is not completely explored due to the shortcomings of current secondary battery and supercapacitor technologies (electrochemical capacitor and ultracapacitor). On-board regenerative braking is a possible mechanism to restore a significant fraction of the energy [42]. LICs do not have the inherent density of energy-storage needed for economic and spatially effective implementation. In comparison, LIBs do not have enough capacity, which needs a charge rate of up to 200 C (1 C is 1 h charge, 200 C is 18 s charge), or around 100 times larger size than what is required for just breaking adsorption.

LICs form a new class of devices capable of bridging the output of commercial EDLC supercapacitors and traditional LIBs [5,15,20]. As previously discussed, LICs are capable of producing 4 to 5 times higher energy values than EDLC ultracapacitors. This superior capacity is obtained by depending on a carbonate-based battery electrolyte that generates a higher voltage of the unit than acetonitrile (4.2 V vs. 2.7 V). Although 5000 cycles is a substantial lifetime for a LIB, they cannot be implemented in place of an ultracapacitor. This is because ultracapacitors run continuously up to thousands of charges per day in some cases. Therefore, LICs appear ideal to replace ultracapacitor electrochemical technologies.

Typically, while the battery-type electrode acts as an anode (case 1 above), it needs to be pre-lithiated during production before being implemented in a LICs, which may bring the anode’s capacity closer to that of Li [43,44]. Pre-lithiation is a method used for LIBs and LICs to compensate for the potential loss of active lithium so higher reversible potentials and higher gravimetric energies are obtained. It is achieved by storage of a certain amount of active lithium in the anode prior to the cells first charge/discharge cycle. There are many methods available for pre-lithiation and can be performed to the active material or the electrode as a whole [45]. As a consequence, a LICs working voltage can raise and even exceed 4.5 V. The energy density is compared by their capacitance and the operating voltage [39,46], as typically chosen comparable specific energy units are V, kJ/mol and kWh/kg. The energy efficiency of LICs can be significantly enhanced due to the large working voltage of the battery-type electrodes. That is, with constant operating currents and resistances (rj-loss), this efficiency reduction becomes relatively less significant by increasing the open-circuit voltage (efficiency = (E_ocp-rj)/E_ocp). Nevertheless, a phase transition frequently follows the faradaic reaction of the battery-type electrode, which results in a weak rate efficiency, lower cycle life and slow dynamics [46,47,48,49]. To overcome these issues, the incorporation of highly conductive carbonaceous additives such as graphene, carbon nanotubes (CNTs) and AC are important in addition to the production of nanoscale-structured electrodes to obtain improved electronic conductivity [50,51,52].

Unlike the pure battery-type electrode, reversible ion adsorption or rapid redox reactions occurs on the capacitor-type electrode sheet, which provides the possibility that LICs have comparable power density to that of ultracapacitors by integrating an effective battery-type electrode with strong kinetics [53]. Also essential to their functional implementation is the cycling efficiency of LICs. Besides the inherent behaviour of the electrode of the battery type and the electrode of the capacitor-type, the mass ratio between the electrode of the battery type and the electrode of the capacitor-type often plays a crucial role in the cycle life of the LIC.

The charges for an asymmetric cell should be distributed at both electrodes (i.e., Q_anode = Q_cath). The charges deposited are aligned with the electrode’s basic potential (C) and mass (m) (Q $\propto$ C $\times \Delta E\times$ m) [54,55]. The optimal mass ratio between the battery-type electrode and the capacitor-type electrode can therefore be determined by the following equation [56]:

$$\frac{m_{cath}}{m_{anode}}=\frac{C_{anode}\times \Delta E_{anode}}{C_{cath}\times \Delta E_{cath}}$$

(1)

where the electrode mass, the basic capacitance and the voltage range in the load/discharge phase for the anode and the cathode are $m$, $C$ and $\Delta E$, respectively. Nonetheless, the capacitance of the capacitor-type electrode is significantly smaller compared to that of the battery-type electrode [23,31]. Therefore, the manufacturing of high-density carbon products (e.g., micro- and macro-carbon composites), is the primary aim in the production of advanced LICs.

Two techniques are typically used to improve the natural capacitance of capacitor-type electrodes: construction of capacitor-type electrodes with broad natural surface area and incorporation of pseudocapacitive or heteroatom doping materials [57,58]. As described above, carbon products, either as active products or as conductive additives, perform irreplaceable roles in the applications of LICs due to the excellent intra- or inter-particle conductivity and the outstanding electrolyte accessibility [59,60,61,62].

3. Electrode Materials

Nanostructured carbons are significant LIC materials used either independently or in conjunction with a second active step of Li such as TiO₂. Yao et al [59] published a review of the carbons used in LICs (Table 1). The carbons are often rich in heteroatoms (especially oxygen) and contain varying degrees of graphene ordering. For carbon allotropes, extremely deficient or heteroatomic carbons do not fall into the classic taxonomy, but instead reflect the similarity of structure and chemistry between pure graphite/graphene and completely amorphous activated carbon. Lithium deposition is typically poorly known in nongraphic carbons, except at fairly high concentrations typical of a battery. The charge levels expected by hybrid devices are far less known and require further research.

3.1. Anodes

3.1.1. Carbon Materials

Graphite, with a hexagonal mesh structure made from carbon atoms, has been used in consumer LIBs goods as a negative electrode material since 1991 [63]. Graphite will provide a large plateau power around 300 mAh g⁻¹ below 0.2 V and will guarantee a secure charge-discharge plateau. Alternatively, graphite suffers from a low rate capability induced by the strong crystallinity and anisotropy that further limits the power density [6]. To increase the capacity of graphite, ball milling, extremely conductive carbon surface grinding, graphisation degree regulation, and the application of defects and additives are all used [64,65,66].

Graphitised carbon, for which the space and microstructures of the interlayer can be conveniently modified, has the most interesting perspectives for use in anodes of LICs. The coexistence of graphical structures and amorphous structures allows graphised carbon to successfully combine a high plateau potential and superior rate efficiency. Catalytic graphisation with the aid of a metal-ions is an important process for preparing graphised carbon under fairly mild conditions (<1200 °C) [67]. The degree of graphitisation can be regulated by adjusting the temperature, precursor and catalyst structures.

Table 1. Carbonaceous material-based lithium-ion capacitor (LIC) summary.

Configuration (Anode//Cathode)	Voltage	Max Energy (Wh/kg) at Power (W/kg)	Energy (Wh/kg) at Max Power (W/kg)	Cyclability
N-doped carbon nanopipes//reduced graphene oxides [68]	0–4 V	262 at 450	78 at 9000	91% over 4000 cycles
graphene//armored graphene [69]	0–4.3 V	160 at 900	59 at 19,000	89% over 1000 cycles
microcrystalline graphite//mesoporous carbon nanospheres/graphene [70]	2.2–4.2 V	80 at 152	32 at 11,600	93% over 4000 cycles
reduced GO//resin-derived carbon combined with GO [71]	0–4 V	148.3 at 150	45 at 6500	79% over 3000 cycles
B&N-doped carbon nanofiber//B&N-doped carbon nanofiber [32]	0–4.5 V	220 at 225	104 at 22,500	81% over 5000 cycles
graphite//activated graphene [72]	2–4 V	147.8	Not reported	Not reported
graphite//functionalised graphene [73]	2–4 V	106 at 84	85 at 4200	100% over 1000 cycles
hard carbon//activated carbon [74]	1.4–4.3 V	80 at 150	65 at 2350	82% over 10,000 cycles
hard carbon//bioderived mesoporous carbon [75,76]	1.7–4.2 V	121 at 300	50 at 9000	81% over 8000 cycles
graphite//activated carbon [72]	1.5–5 V	145.8 at 65	18 at 18,000	65% over 10,000 cycles
hard carbon//activated carbon [77]	2–4 V	82 at 100	14 at 20,000	97% over 600 cycles
graphite//graphene [78]	2–4 V	135 at 50	105 at 1500	97% over 3500 cycles
N-doped hard carbon//activated carbon [79]	2–4 V	28.5 at 348	13.1 at 6940	97% over 5000 cycles
soft carbon//activated carbon [76]	0–4.4 V	115 at 25	16 at 15,000	63% over 15,000 cycles
graphene//activated carbon [80]	2–4 V	95 at 27	61.7 at 222.2	74% over 300 cycles
graphite//activated carbon [81]	2–4 V	103	Not reported	77% over 100 cycles

Table 1. Carbonaceous material-based lithium-ion capacitor (LIC) summary.

Configuration (Anode//Cathode)	Voltage	Max Energy (Wh/kg) at Power (W/kg)	Energy (Wh/kg) at Max Power (W/kg)	Cyclability
N-doped carbon nanopipes//reduced graphene oxides [68]	0–4 V	262 at 450	78 at 9000	91% over 4000 cycles
graphene//armored graphene [69]	0–4.3 V	160 at 900	59 at 19,000	89% over 1000 cycles
microcrystalline graphite//mesoporous carbon nanospheres/graphene [70]	2.2–4.2 V	80 at 152	32 at 11,600	93% over 4000 cycles
reduced GO//resin-derived carbon combined with GO [71]	0–4 V	148.3 at 150	45 at 6500	79% over 3000 cycles
B&N-doped carbon nanofiber//B&N-doped carbon nanofiber [32]	0–4.5 V	220 at 225	104 at 22,500	81% over 5000 cycles
graphite//activated graphene [72]	2–4 V	147.8	Not reported	Not reported
graphite//functionalised graphene [73]	2–4 V	106 at 84	85 at 4200	100% over 1000 cycles
hard carbon//activated carbon [74]	1.4–4.3 V	80 at 150	65 at 2350	82% over 10,000 cycles
hard carbon//bioderived mesoporous carbon [75,76]	1.7–4.2 V	121 at 300	50 at 9000	81% over 8000 cycles
graphite//activated carbon [72]	1.5–5 V	145.8 at 65	18 at 18,000	65% over 10,000 cycles
hard carbon//activated carbon [77]	2–4 V	82 at 100	14 at 20,000	97% over 600 cycles
graphite//graphene [78]	2–4 V	135 at 50	105 at 1500	97% over 3500 cycles
N-doped hard carbon//activated carbon [79]	2–4 V	28.5 at 348	13.1 at 6940	97% over 5000 cycles
soft carbon//activated carbon [76]	0–4.4 V	115 at 25	16 at 15,000	63% over 15,000 cycles
graphene//activated carbon [80]	2–4 V	95 at 27	61.7 at 222.2	74% over 300 cycles
graphite//activated carbon [81]	2–4 V	103	Not reported	77% over 100 cycles

Hard carbon is a suitable choice for high power LICs [64,77,82], with better kinetics and a greater space distance between the carbon layers than graphite. Zhang et al. [83] contrasted the impact of LIC output on two specific hard carbon materials (spherical and irregular hard carbon) with an AC cathode, observing that the irregular hard carbon exhibited optimum electrochemical efficiency with a high energy density of up to 85.7 W h kg⁻¹ and a power density of up to 7.6 kW kg⁻¹ centred on the active material mass of two electrodes in the voltage range of 2–4 V. Furthermore, preparation of a hard carbon anode extracted from a carbonaceous source, such as a nitrogen-doped carbonised polyimide microsphere (CPIMS) [79], can also be a feasible solution. Despite this, the voltage hysteresis and sloping characteristics during hard carbon charging/discharge are unfavourable to the cycling stability of LICs. Sun et al. [84] extensively examined the electrochemical efficiency and power fading behaviours of hard carbon LICs and noticed that power fading of LICs during cycling was induced by a rise in internal resistance and a depletion of lithium deposited on the anode.

Benefiting from the superior mobility of electron carriers and strong lithium-ion transport kinetics, graphene is also a promising candidate as an anode material, rather than being used exclusively as a cathode material [15]. Ren et al. [80] fabricated pre-lithiated graphene nanosheets in LICs as an anode, providing a cumulative power density of 220 W kg⁻¹ at an energy density of 62 W h kg⁻¹ with a capacity retention of 74 % at 400 mA g⁻¹ after 300 cycles.

Ahn et al. [85] provided a highly oriented graphene sponge (HOG) with an ultra-high energy density as an anode. The AC/HOG LICs demonstrated 3.6 times greater diffusivity of the lithium-ion than the AC/graphitized carbon LICs. As a result, they obtained large energy densities of 232 at 57 W kg⁻¹ and 131.9 W h kg⁻¹ at 2.8 kW kg⁻¹. In a DC-LICs device with AC as the cathode, Phattharasupakun et al. [86] documented a nitrogen-doped reduced graphene oxide (N-rGO) aerogel anode. They exhibited a maximum specific energy of 170.28 W h kg⁻¹ in the voltage range of 2.0–4.0 V, and the average power density exceeded 25.75 kW kg⁻¹ after 2000 cycles with almost no decay in efficiency.

Composites processing is an efficient way to combine multiple forms of carbon materials (e.g., graphite, hard carbon and graphene.) as an active substrate for producing high-performance LIC materials. Lim et al. [66] recorded high energy intensity and high power intensity DC-LICs, using natural graphite-coated hard carbon as anode materials. The DC-LIC achieved improved densities in energy and strength, as well as enhanced cycling efficiency.

3.1.2. Transition Metal Materials

Standard transition metal oxides, including Li₄Ti₅O₁₂, TiO₂, Nb₂O₅, Fe₂O₃ and SnO₂, have been extensively studied as anodes of LICs due to their large abundance and strong basic gravimetric potential [87,88,89,90,91]. Despite this, most transition metal oxides typically experience a phase change during the charge/discharge cycle, which is seen by the plateaus in the galvanostatic charge–discharge curves and strong peaks in the CV profiles. These phase transitions can result in broad volume growth, resulting in a negative impact on the integrity of electrodes, resulting in low cycling efficiency. Nevertheless, the composition of TiO₂ and Li₄Ti₅O₁₂ during reversible lithium intercalation/extraction is more robust. The integration of such transition metal oxides into porous carbon materials would solve or mitigate the above issues to achieve hybrid architectures. A recent review by Liu et al. [92] gives a comprehensive overview of transition metal LICs.

Table 2. Transition metal-based LIC summary.

Configuration (Anode//Cathode)	Voltage	Max Energy (Wh/kg) at Power (W/kg)	Energy (Wh/kg) at Max Power (W/kg)	Cyclability
TiO2 hollow spheres at graphene//graphene [93]	0–3 V	72 at 303	10 at 2000	65% over 1000 cycles
TiO2 at mesoporous carbon//AC [94]	0–3 V	67.4 at 75	27.5 at 5000	80.5% over 10,000 cycles
TiO2 nanobelt arrays//graphene hydrogels [95]	0–3.8 V	82 at 570	21 at 19,000	73% over 600 cycles
TiO2 at rGO//AC [87]	1–3 V	42 at 800	8.9 at 8000	Not reported
TiO2–CNT//active carbon [96]	1–3 V	59.6 at 120	31.2 at 13,900	Not reported
Li4Ti5O12–CNT//graphene foam [97]	1–3.6 V	101.8 at 436.1	12.7 at 12,300	84.8% over 5000 cycles
Li4Ti5O12//reduced graphene oxide [98]	1–3 V	45 at 400	30 at 3300	100% over 5000 cycles
nanocrystalline Li4Ti5O12//active carbon [99]	1.5–3 V	55 at 64.6	28.8 at 10,300	Not reported
TiO2-coated Li4Ti5O12//active carbon [100]	0.5–2.5 V	74.85 at 300	36 at 7500	83.3% over 5000 cycles
Li4Ti5O12//N-doped porous carbon [101]	1–3 V	63 at 200	16 at 5000	Not reported
graphene-Li4Ti5O12//graphene-sucrose [102]	0–3 V	95 at 45	32 at 3000	94% over 500 cycles
spheres Li4Ti5O12//active carbon [103]	1–3.5 V	74.3 at 156.26	41.7 at 468.7	93% over 500 cycles
graphene-wrapped Li4Ti5O12//active carbon [104]	1–2.5 V	50 at 16	15 at 2500	75% over 1000 cycles

gives an overview of transition metals used in LIC electrode design.

3.1.3. Polyanion and Carbon Composites

Typically, the Li₃V₂(PO₄)₃, developed with VO₆ octahedra and PO₄ tetrahedra corner-sharing, crystallises in a monoclinical configuration. The comparatively broad gap space of these crystals allows for the quick diffusion and reaction kinetics of different lithium sites. Because Li₃V₂(PO₄)₃ content is an amphoteric powder that can be either decreased by lithium injection or oxidised by lithium elimination, the electrochemical efficiency of both low (1.0–3.0 V) and high (3.0–4.3 V) voltage applications for Li₃V₂(PO₄)₃ has been studied [105]. These were observed to provide maximum energy densities between 27 and 25 W h kg⁻¹, respectively. Unlike Li₃V₂(PO₄)₃, LiTi₂(PO₄)₃ has a NASICON-type frame structure, which consists of PO₄ tetrahedra bound by octahedral unit corners of TiO₆. Each of the PO₄ tetrahedrons are connected to four octahedral TiO₆ units, and in effect a TiO₆ unit is connected to six PO₄ tetrahedrons, allowing for multiple ionic replacements at various lattices [41]. Cyclic voltammetry calculation was observed with a two-phase reaction process at 2.38 V during Li-insertion and an extraction at 2.60 V. As a result, the LICs based on LiTi₂(PO₄)₃ carbon-coated anodes display ultra-high energy and power densities of 14 W h kg⁻¹ and 180 W kg⁻¹, respectively [106].

For LIC anodes, TiNb₂O₇ will serve as an alternate nominee for Li₄Ti₅O₁₂. TiNb₂O₇’s monoclinical crystal structure comprises of disordered Nb and Ti atoms, which may have two-dimensional interstitial spaces for rapid Li-ion injection and show an energy density of 110.4 W h kg⁻¹ at 99.58 W kg⁻¹ [107]. Importantly, the TiNb₂O₇ on C electrode’s Li injection activity was analysed in detail and the pseudo-capacitive reaction mechanism for intercalation was achieved [41].

3.1.4. Metalloid/Carbon and Metal Materials

The production and materials used for sustainable LICs manufacturing is becoming increasingly important. The use of silicon within carbon electrodes provide a promising route for sustainable LIC production in the future. Silicon is particularly of interest due to it being the second most abundant element in the Earth’s crust. Due to its low lithiation ability (<0.5 V) and strong real theoretical efficiency (4200 mAh g⁻¹), Si is a promising material for high-performance LIC anodes [108,109]. Despite its excellent load-discharge platforms and extremely high specific capacity, its cycling and rate performance may be poor due to its severe volume expansion and low electronic and ionic conductivity [110]. Furthermore, due to its inherent semi-conductive design, the low conductivity may also restrict its performance in charge/discharge at high current density [111,112,113].

Composite processing is an efficient way to combine multiple forms of active materials as an active substrate for producing high-performance LIC materials. Soft carbon is a frequently used commercial anode material with high conductivity, fast lithium-ion transport and long cycling performance, but its specific capacity and operating platform still require improvement. Therefore, using a small amount of silicon-carbon composite in a soft carbon anode could ameliorate the anode’s charge/discharge kinetics and also provide surplus lithium to slow the rate of active lithium consumption in long-term cycling after anode pre-lithiation. Using this approach, it has been observed that such a LIC has over 95% capacitance retention after 10,000 cycles at 20 °C [110].

Based on 3-electrode hybrid configuration [77], other types of lithium, such as lithium silicide, can be used for the anodes [114]. Pre-lithiation has traditionally been done on the anode; however, pre-lithiation on the cathode has been observed, which in turn is contradictory to how LIBs are produced [115]. That would be potentially better because the system is instead constructed in a thermodynamically stable manner (which may not be the case with a symmetric unit, e.g., a carbon–carbon LIC).

Sn has a greater electrical conductivity relative to Si, which will contribute to Sn anodes showing a higher rate capacity. Nevertheless, during the charge/discharge phase, Sn also suffers from a significant volume shift which can result in extreme polarisation of the electrode material [116,117,118]. The construction of tiny Sn nanoparticles with a binding carbon substratum is currently a successful method for overcoming these problems.

3.2. Cathodes

In the early research into LICs, activated carbon was dominantly used as the cathode material with a focus on the energy-storage process of surface adsorption. This was because it shows a large surface area (33,000 m₂ g⁻¹), excellent conductivity (almost 60 S m⁻¹) and strong chemical stability [41]. An AC cathode’s energy storage capability also provides a power of approximately 50 mAh g⁻¹ [41]. Compared with anode electrode performance, it is comparatively smaller. For load-balance, the cathode’s mass load is typically two to four times that of the anode, depending on the cathode and anode’s different capacitances [41]. Amatucci et al. [4] produced a LIC system using AC as the cathode, and nanostructured Li₄Ti₅O₁₂ (LTO) as the anode, the first use of AC in LICs. The voltage window of LICs in an organic electrolyte dependent on AC cathodes has since been designed in the range of 1.5–4.5 V to ensure high energy density and preserve the long cycle stability of LICs [72].

Additionally, carbon derived from metal-organic frameworks (MOF) with various architectures has also been widely researched in LICs. For example, large surface area (2714 m₂ g⁻¹) carbon cuboids were synthesised by pyrolysing the zinc-based MOF-5, which exhibits a peculiar crumpled-sheet porous morphology assembled with the required micro and mesoporosity values [119]. The MOF-dependent LIC provides a maximum effective energy density of 65 W h kg⁻¹ with excellent power capacity, dependent on the advanced structure. Likewise, polyhedral hollow carbon derived from MOF was also produced and used in LICs [120]. It is worth noting that the removal of metal ions during the preparation of MOF-derived carbon is a required step due to the presence of metal ions in the MOF precursor.

Graphene’s unique 2D structure and physical properties and its derivatives make them distinctive building blocks for producing various porous 3D architectures. Thanks to the combination of porous structures and the excellent intrinsic properties of graphene and its derivatives, these 3D architectures exhibit excellent chemical stability and strong specific surface area as well as fast electron transport kinetics. Therefore, these high-performance 3D architectures are strong candidates for use as the cathodes of LICs. Various techniques for constructing graphene-based architectures for LICs have been comprehensively documented in recent years [13,41,63,98]. Importantly, integrating microporous carbon with surface connectivity into graphene structures is an important technique for further increasing the LIC’s strength and energy densities.

4. Design of Lithium-Ion Capacitors

In terms of LIC design, the process of pre-lithiation, the working voltage and the mass ratio of the cathode to the anode allow a difference in energy capacity, power efficiency and cyclic stability. An ideal working capacity can usually be accomplished by intercalating Li⁺ into the interlayer of graphite. In this way, it is possible to achieve reduced electrode resistance, increased energy density and stable cycling efficiency [63]. A very critical element in maintaining power equilibrium is the mass ratio of the cathode to the anode (m+/m−). Taking into account the aforementioned criteria, finding a compromise between the correct degree of pre-lithiation, an effective functioning voltage window and the optimum m+/m− is of considerable importance. Although some researchers perform pre-lithiation on one or both of the electrodes, others depend entirely on Li in the electrolyte (e.g., from LiPF₆), resulting in reduced performance due to the loss of active Li [59,121,122]. The need for a solid electrolyte interface (SEI), its creation, continuous growth and cracking is a major problem correlated with the usage of carbon anodes. During cycling, one will need to hold the SEI steady and prevent Li plating that involves careful regulation of the negative electrode voltage, temperature and effective current density. Therefore, a carbon–carbon design requires an asymmetric mass loading with the weight of the lower-capacity adsorption carbon cathode being up to 5 times higher than that for the lithium-containing anode.

4.1. Pre-Lithiation Strategy

During the production of LICs, a pre-lithiation cycle is performed to the active material or whole electrode of the anode guarantee that it operates at fairly low potential and to supply lithium ions for the anode-side insertion/extraction reaction. In turn, the added lithium through pre-lithiation reduces the electrode resistance and accounts for the permanent active lithium loss during cycles induced by the creation of solid electrolyte interface films (SEI) on the anode. The following can be performed using distinctive pre-lithiation approaches [13,30,45,63]: the electrochemical process (ECP); the short-circuit pre-lithiation through external short-circuit (ESC) and internal short-circuit (ISC) methods; the introduction of permanent lithium transition metal oxides (LTMOs) on the cathode side; and, the application of sacrificial lithium chloride to the electrolyte. Pre-lithiation methods can also be adapted between different capacitor chemistries (e.g., LIC, NIC and KIC) [123]. Jin et al. (2020) give a thorough review of pre-lithiation technologies in LICs, outlining the current progress and perspectives [124].

An important step in creating high-performance LIC systems is the pre-lithiation of the electrodes. According to the restricted solubility of salts such as LiPF₆ or NaClO₄ in carbonate solvents, permanent ion loss results in ionic conductivity degradation [63]. During prolonged cycling, devices focused on non-lithiated electrodes have been shown to have decreased lithium ion content in the electrolyte, significantly deteriorating the overall LIC potential [81,121]. Besides SEI, Li can be trapped in the majority of electrode materials (e.g., hard carbons) [125]. Pre-charging at least one electrode is a key element in evaluating a hybrid device’s output in terms of initial energy, power conservation and cycling conservation [63]. A study of Ragone properties typically shows that, in addition to the electrolyte [80,81,126,127], systems at the peak of the energy–power range are those that utilise a secondary supply of ions.

This training compensates for the depletion of the Li ion induced by initial SEI forming and permanent bulk trapping. As long as there is no substantial increase in SEI and bulk trapping during cycling, the concentration of electrolytes stays relatively constant. Pre-lithiation has also been successfully used to expand the voltage window and extend LIC device cyclability by regulating the anode’s voltage oscillation, thereby reducing the load depth and related SEI production, volumetric expansion and ion trapping [72,81,121].

The most prevalent pre-lithiation procedure is electrochemical deposition. A closed electrical circuit develops bonding between the anode and the lithium metal [128]. This uses an electrically insulating and ionically conducting separator, similar to what is found in LIBs. Depending on the pre-doping process controllability [72,128,129,130], the external circuit can contain an electrical capacitor or resistor, or even a short circuit. The charging cycle can be operated with a programmable system added to the terminals [74,131]. The main drawback to this approach is extra cell assembly/disassembly procedures, where the unit ends up being assembled twice. While scientifically useful and to some extent convenient, it is not obvious how the industry can make such an approach cost-effective.

A theoretically more efficient solution could be based on chemical doping in the presence of an electrolyte [132,133,134]. This would occur through the direct reaction between the electrode substrate and lithium. The main advantage of this approach is its sleek simplicity and scale-up potential for industries; however, lithiation is in effect unobservable, meaning the state of charge (i.e., lithium content in the electrode), cannot be controlled or monitored until after the doping is complete. Despite this, research has been performed to alleviate this by constructing LICs with multielectrode structures to escape cell-reassembly and track the electrical signals during doping [74,135,136].

Cao and Zheng [77] demonstrated a three-electrode hybrid configuration that used stable lithium energy. Based on this theory, other types of lithium, such as lithium silicide, lithium-rich transition metal oxides, were used for different anodes (e.g., steel, silicone, etc.) [114,127,132,137]. Pre-lithiation has traditionally been done on the anode, which in turn is contradictory to how LIBs are produced; however, pre-lithiation on the cathode has been observed [115]. That would be potentially better because the system is instead constructed in a thermodynamically stable manner (which may not be the case with a symmetric unit, e.g., carbon–carbon).

During LIB production, lithium-rich ion-donating materials show a strong hysteresis during charging and discharge. This in turn causes a cycle of permanent de-lithiation during the LIBs first charge. The ions are then inserted between the main storage processes in the electrolyte and transfer back and forth during subsequent cycling. In this way materials such as Li3N [129,138] and Li6CoO4 [128,139] were able to be used as cathode additives to minimise failure. It may be possible to directly transfer this method to LICs, for example, Lukatskaya et al. [56] reported an example of the concept of a chemical additive ion source that could advance LIC technology due to its efficacy and scalability [56].

In a LIC, the activated carbon can contain sacrificial synthetic lithium salt (3,4-dihydroxybenzonitrile dilithium) [13]. This compound is irreversibly decomposed after the first cycle, with the residues dissolving in the electrolyte. It acts as a lithium-ion source inside the graphite anode without adversely impacting the electrochemical output of the LIC. Due to the ease of converting this principle into industrial production of various LIC electrodes, the sacrificial additive method seems to have the greatest potential for practical implementations among all pre-charging approaches. In addition, the recent development of a cascade-based methodology for pre-lithiation may lead to advances in the stability of pre-lithiation of anodes.

4.1.1. Electrochemical Pre-Lithiation

Lithium metal is used as the counter-electrode for the EC method, with graphite being used as a functional electrode. Lithium ions migrate from the metal to the graphite by cycling at a fairly small current, during which the lithium electrode and graphite electrode become physically separated. Lee et al. [73] obtained pre-lithiated graphite by cycling graphite half-cells 2 times at 0.1 °C followed by a discharge of the LIC to 0.05 V during the third cycle. Similarly, by combining a mesocarbon microbead working electrode with a lithium reference electrode, Zhang et al. [140] formed pre-lithiated mesocarbon microbeads. The merit of this method is that, by setting the cut-off capacity and producing a stable SEI layer, the lithiated content can be precisely controlled. Nevertheless, the reassembly phase of the pre-lithiated anode is a time-consuming procedure unfeasible currently for use in industrial production.

4.1.2. Short-Circuit Pre-Lithiation

Often pre-lithiation of short circuits may be split into ESC (external short-circuit) and ISC (internal short-circuit) techniques. For the ESC process, a sacrificial lithium metal electrode and a targeting electrode (e.g., graphite) are separated using a porous polymer separator in a non-aqueous electrolyte and are connected to wire to naturally facilitate the penetration of lithium ions into the electrode [81,141]. Kim et al. [142] compared the different pre-lithiation approaches and showed that the ISC method was straight-forward and effective for obtaining high energy intensity DC-LICs. In the presence of an electrolyte media, the ISC technique is performed by solid metallic lithium (e.g., Li metal or Li powder) touching a carbon anode directly. Cao and Zheng [77] described a network of hard carbon/activated carbon DC-LICs by utilising a mixture of stable lithium metal powder (SLMP). Once the hard carbon—SLMP mass ratio reached between 5:1–8:1, a healthy life cycle was obtained. The extra Li powder-covered hydrogen fluoride (HF) against etching of the SEI layers. The biggest benefit of short circuit pre-lithiation is its straight-forward usability and scale-up. Despite this, the key disadvantage is that lithium utilisation cannot be precisely regulated [132].

4.1.3. Irreversible Transition Metal Oxides at the Cathode

Compared to the usage of metallic lithium, the use of permanent electrochemical lithium-transition metal oxides as a pre-lithiation agent will significantly enhance the health of LICs [63]. Usually, the chosen lithium transition metal oxides only operate on the first charge cycle to facilitate the incorporation of lithium cations into the graphite anode. Li₆CoO₄, Li₅FeO₄, Li₂CuO₂, Li₅ReO₆ and Li₂RuO₃ have been developed to be used as pre-lithiation agents [115,143,144,145,146]. Through this method, the degree of pre-lithiation can be managed by regulating the volume of lithium metal oxides in the cathode.

4.1.4. Adding Sacrificial Organic Lithium Salt

Jeżowski et al. [56] published a pre-lithiation process for sacrificial organic lithium salt, using 3,4-dihydroxybenzonitrile dilithium salt (Li₂DHBN) as a sacrificial salt in 1 M LiPF₆/EC–DMC. The insoluble Li₂DHBN was combined with activated carbon and released during the first charge cycle to form the soluble 3,4-dioxobenzonitrile (DOBN). The Lithium ions intercalated into the graphite to achieve pre-lithiation. Of such approaches, stable metallic lithium powder and the inclusion of permanent lithium transition metal oxides could have the most promising prospects for industrial production; however, monitoring the lithiation degree and health issues are still relevant factors. Recently, Li₃N was employed as a sacrificial lithium compound for metal-free pre-lithiation by adding it to the electrode mixture during active material synthesis. Using this additive, stable cyclability of 91% after 10,000 cycles was observed [147].

4.2. Electrolytes

The electrolyte often serves as an integral part of the pairing LIC electrodes, greatly influencing the energy capacity, power efficiency, and cycling stability. To obtain sufficient energy and power efficiency, multiple electrolyte specifications should be met [6]: outstanding ionic conductivity; strong electrical insulation; large voltage operating window; exceptional stability; and, ideally low toxicity. Electrolytes can be classified into 4 types: aqueous electrolyte; organic electrolyte; ionic liquid electrolyte; and inorganic solid-state electrolyte. Aqueous electrolytes have high ionic conductivity and low resistance, favourable for ion transfer; however, the working voltage range is generally less than 1.2 V and is constrained by water decomposition [148]. Due to their marginal instability, good thermal, chemical and electrochemical resilience, low flammability and superior conductivity, ionic liquid-based electrolytes have been examined [149,150]; however, their use is not ideal industrially due to their high economic costs. A solid-state polymer may be an acceptable alternative electrolyte because the ionic conducting medium has electronic separators that preserve the LICs health through high cycle numbers. While in recent years solid-state polymer electrolytes have flourished and considerable improvement has been made, their ionic conductivity needs to be significantly enhanced for room temperature use.

Organic electrolytes may have the best prospects for commercially available LICs. They have a larger voltage window than aqueous electrolytes, and their operating condition is milder than that of solid-state polymer electrolytes and ionic liquid electrolytes. Organic electrolytes are commonly used in LIBs and LICs; however, their high economic cost, low potential power, low conductivity and health concerns linked to flammability, variability and toxicity are problematic [149,150]. Despite this, a modern polyethylene glycol-functionalised polysilsesquioxane has been developed for the manufacture of hybrid ionogel electrolytes for LICs [151]. This gave the electrolyte excellent thermal stability and superior ionic conductivity.

The most significant or distinct required components in electrolytes are lithium salts, solvents and additives. Solvents such as polycarbonates, ethers, ethylene carbonates, diethyl carbonates, dimethyl carbonates, propylmethyl carbonates and ethyl methyl carbonates can dissolve lithium salts and transport lithium ions [72,81,152,153]; however, the SEI shape and structure can be highly affected by the use of solvents. Additives are connected to other compounds that are applied to the electrolyte and can effectively boost efficiency. According to the different effects of the additives, they can be classified as improving ion conductivity; improving device properties (e.g., the SEI); improving low-temperature efficiency and thermal stability; preventing overloading; and, reducing electrolyte acid and water content [6,154,155].

4.3. Modelling and Simulation of LICs

LICs have a clear advantage in offering high power and energy density within a single energy storage system while maintaining a longer cycle life. Although there has been substantial research into materials and chemistries for the production of LICs, there is a requirement for research on the effect of electrode balancing and pre-lithiation on the LICs usable energy. Computational approaches have been able to provide aid in determining these effects using physics-based models based on experimental data.

When designing an LIC, many factors need to be taken into account to ensure accurate hybridisation of the LIB components with the traditional capacitor components. Using theoretical and experimental data, computational modelling has been able to be used to assist this process. Moreover, such modelling has allowed theoretical guidelines to aid the design of LICs to obtain optimal operation. Choi and Park (2014) developed a thorough theoretical analysis that aids the design of high-performance LICs, and also developed guidelines to improve the development phase of new LICs [156]. In summary, these guidelines state that:

• battery and capacitor components must be homogeneously hybridised to operate as a single electrode (see [157] for in-depth hybridisation approaches and principles);
• the capacitor component must be highly electronically-conductive and able to store electrochemical energy in the organic electrolyte by electrostatic force, and,
• electrode material must have a high surface area to achieve high energy density and an open porous structure to achieve optimal ionic conduction.

Ghossein et al. (2018) noted that there were limitations in using LIB and supercapacitor models based on electrochemical impedance spectroscopy for LICs [158]. They stated that these models were not suitable for accurately describing the impedance of LICs at low frequencies. They concluded that this was due to the size of the pores in the cathode being identical to the anode in the LIC and supercapacitor models. Therefore, they introduced a new model that was specific to LICs and includes accurate pore sizes and distributions of the cathode. This allowed accurate modelling of impedance values at all frequencies when compared to experimental data. Madabattula el al. (2019) developed a model to study the relationship between usable energy at a variety of effective C rates and mass ratios of the electrodes in a LIC [159]. By extending the model to analyse the pre-lithiation requirement, they were able to determine the limits of pre-lithiation in Lithium Titanium oxide anodes, and how negative polarisation of the activated carbon in the cathode can improve cell capacity. In addition, in a LIC cell with a higher mass ratio, the authors were able to relate the effects of electrolyte depletion with poor power performance.

5. Knowledge and Research Gaps

Low Coulombic-related decreases in available charge in batteries are usually correlated with the irreversible and continuous creation of an SEI and certain cathode electrolyte interlay (CEI) [160,161]. This can also be the case with LICs. LICs usually have a smaller Lithium pool than traditional LIBs, have more strict criteria for reducing internal resistance losses attributable to high voltage and undergo more cycling. All three of these aspects render SEI forming a substantial challenge for LICs, and due to the wide voltage window, CEI formation can result in plugging of surface pores and, therefore, a loss of active surface area [13].

SEI products are expected to be similar to what is observed with other types of carbons used in LIBs, due to carbonate electrolyte reduction in LICs [162,163]. The electrolyte solutions solvents and salts are both thermodynamically unstable and undergo reduction on the anode during charging (when the anode is temporarily a cathode), which may operate at a potential near (or below) that of metallic lithium [164,165]. Therefore, a substance with a wide surface-to-volume ratio can induce higher irreversible losses in terms of performance [166,167]. Such reduction films may make the anode surface unreactive, protecting the electrolyte solution from further decomposition; however, any volumetric changes experienced during electrochemical cycling may weaken and fracture the SEI layer, with each cycle exposing fresh material to the electrolyte [109,168,169,170], creating a new layer of the SEI.

The SEI comprises primarily of electrolyte reducing materials. The instability of the SEI will inevitably trigger the LIC unit to lose overall power and fail [168,171,172,173]. Apart from solvent-reduction products such as Li₂CO₃ and alkyl carbonates, the SEI anode also partially consists of LiF, which is a decomposition product of the LiPF₆ salt but can also be produced by reaction with trace amounts of water to HF and eventually LiF [174,175]. Research on this points to radial compositional and structural gradients inside the SEI sheet, as well as complex growth-shrinking characteristics with cycling [176,177,178,179].

For a variety of carbon-supported nanomaterials (e.g., oxides and sulphides), an additional aspect relating to SEI formation in hybrid ion capacitor anode materials is linked to the cycling-induced capability gain sometimes mentioned in the literature [13]. An increase in the performance induced by cycling is not uncommon for Li-based anodes, especially for oxides [180,181,182,183,184]. This phenomenon has been attributed to a contribution to load-storage through surface adsorption as a result of the extra surface area created by conversion compound cycling [184,185]; however, recent discussions have suggested this is not the case as when a carbonate electrolyte is used, the surface area of both electrodes during cycling decreases considerably [13]. They further claim that the increase in power is due to the internally parallel nanostructured anodes rather is the cycling-induced reversible forming of a polymer gel on the surface of the nanostructured electrode during lithium charging. This Faradaic cycle can be mechanistically identical to reversible and high Coulombic quality redox reactions in polymer-based electrochemical capacitors [186,187,188].

In LIC devices a parasitic oxidation substance (i.e., CEI), is most likely formed on the cathode. The most intuitive evidence for this oxidation layer is the commonly observed loss of the first-cycle capacity for carbon cathodes in carbonate electrolytes [98,189,190]. A secure passivation sheet could prevent any further oxidation of electrolytes on the electrode surface [191,192,193,194]. The chemical structure of the passivation substrate (e.g., size or type) on the cathodes has been observed to differ greatly from material to material [193]. Surface characterisation methods (e.g., X-ray photoelectron spectroscopy and atomic force microscopy), have been used to study passivation layers in cathodes [193,194,195,196]; however, the CEI chemistry, composition or voltage impact on its development are yet to be analysed.

In terms of its chemistry and composition, the CEI in LICs can be somewhat different from that developed on the classic LIB intercalation cathodes [13]. CEI creation is catalysed by a faradic mechanism. The sum of irreversible ability due to CEI should be directly linked to the surface area and the functionality of the surface heteroatoms, as well as the electrolyte carbonate types. Electrochemical impedance analysis has been used to observe the development and evolution of the passivation layer on a carbon cathode in an indirect way [77,197,198]; however, more in-depth analysis of the CEI in LIC devices is required, because its growth may be important to performance.

Besides the SEI and CEI materials used in LICs, carbon-based electrodes can suffer from Li plating to a greater degree than as observed in LIBs. This specifically applies to graphite, which continues to plate Lithium at various charging rates [199,200,201,202,203,204,205,206].

Nanostructured lithium titanate could greatly reduce the magnitude of lithium metal plating [121]; however, Li₄Ti₅O₁₂ alone is not adequately electrically conductive to reach the high levels required in LICs. The ideal method to prevent metal plating is to ensure that the anode stays far enough from 0 V vs. Li/Li⁺, especially at high charging levels; however, this would require the quantification of a three-to-four-electrode cell (including reference electrodes), something that is rarely done. It should also be pointed out that the voltage swing of the individual electrodes can only be estimated from their mass-to-capacity ratio in a true two-electrode cell. This causes more problems in specifically regulating the composition of both the metal and SEI.

6. Next-Generation Lithium-Ion Capacitors

Much research in recent years has revolved around developing the electrodes used in LICs. As a result, many types have been observed to have varied performance in LIC designs. Table 3 gives an overview of the different materials researched recently.

6.1. Pseudocapacitive Oxides

Apart from carbons and titanium compounds, a variety of new anode materials is available that are either specifically designed for LICs or are added to specific systems that have original uses in LIBs. Those nanomaterials are primarily used for anodes, not for cathodes. Conversely, even the best carbon-based adsorption cathodes offer a fraction of the capacity of existing LIC anodes. Cathode work in the field is even more important, as LIB cathodes are not specifically transferable.

Because of their fast charging and discharging kinetics, the pseudocapacitive materials emerge as a significant subset of materials for LICs, together with higher gravimetric and volumetric efficiency relative to true EDLC electrodes [5]. Augustyn et al. [54] demonstrated that the Li⁺ intercalation into the orderly channels of bulk orthorhombic T-Nb₂O₅ was straight-forward, making the charging behaviour capacitor-like. These materials were referred to as pseudocapacitive intercalation compounds since, although the Li storage mechanism was not considered EDLC, EDLC-like triangular galvanostatic curves and box-like CVs were shown. Despite this, these LICs produce at fairly low power a maximum of 76 Wh kg⁻¹ rendering their total effective energy around a factor of 2 lower than that for carbon-based electrodes [207].

Vanadium oxide V₂O₅ undergoes a bulk ion intercalation reaction during reversible charging, also producing a sloping profile similar to a capacitor when used as an anode. Bulk V₂O₅ can accommodate electrolyte cations (e.g., H⁺, Li⁺ and K⁺) in aqueous systems [208,209,210]. They also mixed elemental analysis and X-ray diffraction to analyse the charging-storage process of V₂O₅ in an aqueous environment. They also discovered that K⁺ ions inject gaps between the (00l) Miller index planes into the interlayer, and an average energy of 42 Wh kg⁻¹ can be achieved with a loss of <5% over 10,000 cycles [208]. A V₂O₅-based hybrid solution can also operate in a far wider voltage window in Li+ organic systems. The authors tested the same materials V₂O₅ on CNT in lithium structures [210]. The LIC cell energy value was in the region of 40 Wh kg⁻¹, similar to aqueous electrolyte systems.

6.2. MXenes

MXene is a large family of metal carbides and carbonitrides in the two-dimensional form [211]. It has been observed that several cations (Na⁺, K⁺, Mg²⁺ and Al³⁺) can intercalate reversibly through the bulk of exfoliated multilayer T₃C₂T_x MXene in an aqueous electrolyte. It was also stated that Li+ ions would intercalate reversibly between layers of MXene in the organic electrolytes [212]. Because there is a significant intercalation reaction, the MXene shows pseudocapacitor CVs and galvanostatic profiles, and could therefore be grouped into the pseudocapacitive substance family [211,212,213]. In addition, the electrochemical analysis revealed that storage of Li was a reaction rather than a diffusion-controlled Faradic cycle, similar to the other pseudocapacitive materials.

An LIC system based on Ti₂C coupled with Kuraray YP17 activated carbon has also been proposed [214]. Here, the Ti₂C anode showed moderate working voltage with the LIC operating at an upper cut-off voltage of 3.5 V and delivering maximum energy of 50 Wh kg⁻¹. Wang et al. [214] also synthesised 3D TiC nanoparticle chains operating in a lower voltage (0.7 V vs. Li/Li⁺) area. The resulting LIC system, together with a high-capacity, nitrogen-doped porous carbon cathode, provided 101.5 Wh kg⁻¹ of energy and 23.4 Wh kg⁻¹ energy at an intense strength of 67,500 W kg⁻¹. Additionally, Luo et al. [215] developed a CTAB-Sn pillared Ti₃C₂ MXene-based LIC system that provided energy of 105.6 Wh kg⁻¹. An important issue worth pursuing is the cyclability of these materials. Whether the existing intercalation compounds can survive such extended service and, if not, whether their structure and/or electrolyte chemistry can be tuned to enhance performance is still unknown.

6.3. Conversion Compounds

Reversible materials dependent on reaction conversion are another type that was considered pseudocapacitive and has gained popularity as a substrate for the LIC electrodes. A conversion electrode is defined initially in LIB literature as the crystalline or amorphous A_xB_y material, which then decomposes reversibly into one or two Lithium compounds after charging [216]. Some examples of the structures used for LICs are MoS2, NbN, VN, MnO, Fe₂O₃/Fe₃O₄, NiCo₂O₄ and various alloys. Although MoS₂ undergoes a Lithium intercalation reaction down to 1.1 V, most reversible capability stems from reversible Mo and Li₂S conversion reaction down to 0 V [217]. The reversible capability of MoS₂ is stated to be as high as 1000 mAhg⁻¹, which actually is higher than the theoretical figure. The discrepancy may be due to a capacitive contribution and a reversible SEI growth contribution [218]. Wang et al. [219] used MoS₂-graphene composites in LICs, using sloping charge and discharge profiles, with the LIC delivering energy density as high as 188 Wh kg⁻¹ at 200 W kg⁻¹ and 45.3 Wh kg⁻¹ at 40,000 W kg⁻¹. MoS₂’s functional efficiency was around 600 mAh per gram and helps balance the sloping voltage plateau with fast energy delivery. The outstanding rate capability and the retention of cycling power possibly emerged in the anode from the engineered secondary carbon-based process.

6.4. Battery-Related Intercalation Ceramics

Research has been performed using ceramic LIB cathode materials coupled against carbon counter electrodes, developing a high-power LIB with a more sloping tension profile. Aravindan et al. [220] provided an excellent description of intercalation-type materials for configurations of LICs. The overwhelming majority of such architectures exhibit cyclability like batteries, lasting many thousands of cycles. For example, layered oxides, spinel oxide, olivine, NASICON and silicates are specific forms of ceramic battery cathodes used in LICs. Olivine LiFePO₄ is a well-established LIB commercial cathode which has also seen applications in LIC systems. During lithium intercalation/deintercalation it undergoes a two-phase reaction and shows a flat plateau at 3.4 V. Ping et al. [221] constructed a hybrid activated carbon + LiFePO₄ composite as the cathode and mesocarbon microbeads as the anode. The unit was cycled between 2 and 4 V with an average energy density of 69 Wh kg⁻¹ and a lifespan of 100 cycles with marginal decay.

NASICON cathodes have also recently gained recognition for applications in hybrid systems. The most frequently studied NASICONs are the phases Li₃V₂(PO₄)₃ (LVP) and Na₃V₂(PO₄)₃ (NVP), with vanadium as the active metal transfer part. Satish et al. [105] incorporated an LVP-C cathode with activated carbon as an anode in a LIC device, obtaining 25 Wh kg⁻¹. For an activated carbon/LVP configuration an energy density of 28 Wh kg⁻¹ was observed [222]; however, the energy efficiency becomes greatly increased (125 Wh kg⁻¹ at 300 W kg⁻¹) where activated carbon is used as a cathode when LVP is the anode.

Table 3. Next-generation electrode materials for LICs.

Configuration (Anode//Cathode)	Voltage	Max Energy (Wh/kg) at Power (W/kg)	Energy (Wh/kg) at Max Power (W/kg)	Cyclability
T-Nb2O5-graphene//activated carbon [223]	0.8–3 V	47 at 393	15 at 18,000	93% over 2000 cycles
mesoporous Nb2O5–C//activated carbon [224]	1–3.5 V	74 at 120	20 at 12,137	Not reported
V2O5 on CNT//activated carbon [210]	0–2.7 V	40 at 210	6.9 at 6300	78% over 10,000 cycles
γ-LixV2O5-BM50//activated carbon [225]	0–4.5 V	54.59 at 230	Not reported	100% over 400 cycles
CTAB-Sn on Ti3C2 MXene//activated carbon [215]	1–4 V	105.6 at 495	45.3 at 10,800	70% over 4000 cycles
Ti2C MXene//YP17 active carbon [212]	1–3.5 V	50 at 190	15 at 600	85% over 1000 cycles
TiC MXene//N-doped porous carbon [214]	0–4.5 V	101.5 at 450	23.4 at 67,500	82% over 5000 cycles
T-Nb2O5 on C//MSP-20 activated carbon [226]	1–3.5 V	63 at 70	10 at 6500	75% over 1000 cycles
Nb2O5-carbide-derived carbon//YP-50F AC [227]	1–2.8 V	30 at 220	18 at 5000	Not reported
Nb2O5–CNT//activated carbon [228]	0.5–3 V	33.5 at 82	4 at 4000	Not reported
LiNbO3 on graphene aerogel//boron carbonitride nanotube [229]	1–4 V	148 at 200	69.4 at 9900	82% over 7000 cycles
MoS2–C-RGO//PANI-derived porous carbon [219]	0–4 V	188 at 200	45.3 at 40,000	80% over 10,000 cycles
NbN//activated PANI-derived carbon [230]	0–4 V	149 at 200	5 at 45,000	95% over 15,000 cycles
VN-rGO//activated carbon [231]	0–4 V	162 at 200	64 at 10,000	83% over 1000 cycles
MnO cubes//activated carbon [232]	0–4 V	227 at 55	21 at 2952	93% over 3000 cycles
3D MnO array//activated carbon nanosheets [233]	1–4 V	184 at 83	83 at 18,000	83% over 5000 cycles
MnO nanoparticles//activated carbon [234]	0–4 V	220 at 100	35 at 2608	95.3% over 3600 cycles
MnO on C//trisodium citrate-derived carbon [235]	0–3.9 V	235 at 120	61 at 25,000	85.69% over 10,000 cycles
MnNCN//activated carbon [236]	0.1–4 V	103 at 150	22 at 4500	100% over 5000 cycles
FexO on graphene//porous graphene [237]	0–3.5 V	129.6 at 19	45 at 3500	75% over 3000 cycles
Fe2O3//activated carbon [238]	0–3.5 V	90	Not reported	55% over 2500 cycles
Fe3O4 in graphene//3D graphene [239]	1–4 V	204 at 55	85 at 2650	70% over 1000 cycles
NiCo2O4//activated carbon [240]	0–4.5 V	39.4 at 120	10 at 554	100% over 2000 cycles
activated carbon//LiMn1/3Ni1/3Fe1/3O2–PANI [241]	0–3 V	49 at 900	19 at 3000	100% over 3000 cycles
activated carbon//LiMn2O4 [242]	0.7–3 V	45 at 60	10 at 800	Not reported
LiNi0.5Mn1.5O4//activated carbon [243]	1.5–3.25 V	19 at 120	8 at 3500	81% over 1000 cycles
mesocarbon microbeads//LiFePO4 [221]	2–4 V	69	Not reported	100% over 100 cycles
Li3V2(PO4)3//activated carbon [222]	0–4 V	125 at 300	65 at 6000	80% over 200 cycles
activated carbon//Li3V2(PO4)3 [222]	0–2.5 V	28 at 35	14 at 1500	87% over 1000 cycle
activated carbon//Li3V2(PO4)3 [105]	0.5–2.7 V	25 at 88	13 at 320	Not reported
Li2MnSiO4//activated carbon [244]	0–3 V	54 at 150	37 at 1500	85% over 1000 cycles
LiMnBO3//PANI [245]	0–3 V	42 at 1500	15 at 5350	91% over 1000 cycles
CoNiP2O7//activated carbon [246]	0–4 V	116.3 at 200	66.7 at 6486.5	86.5 over 500 cycles

Table 3. Next-generation electrode materials for LICs.

Configuration (Anode//Cathode)	Voltage	Max Energy (Wh/kg) at Power (W/kg)	Energy (Wh/kg) at Max Power (W/kg)	Cyclability
T-Nb2O5-graphene//activated carbon [223]	0.8–3 V	47 at 393	15 at 18,000	93% over 2000 cycles
mesoporous Nb2O5–C//activated carbon [224]	1–3.5 V	74 at 120	20 at 12,137	Not reported
V2O5 on CNT//activated carbon [210]	0–2.7 V	40 at 210	6.9 at 6300	78% over 10,000 cycles
γ-LixV2O5-BM50//activated carbon [225]	0–4.5 V	54.59 at 230	Not reported	100% over 400 cycles
CTAB-Sn on Ti3C2 MXene//activated carbon [215]	1–4 V	105.6 at 495	45.3 at 10,800	70% over 4000 cycles
Ti2C MXene//YP17 active carbon [212]	1–3.5 V	50 at 190	15 at 600	85% over 1000 cycles
TiC MXene//N-doped porous carbon [214]	0–4.5 V	101.5 at 450	23.4 at 67,500	82% over 5000 cycles
T-Nb2O5 on C//MSP-20 activated carbon [226]	1–3.5 V	63 at 70	10 at 6500	75% over 1000 cycles
Nb2O5-carbide-derived carbon//YP-50F AC [227]	1–2.8 V	30 at 220	18 at 5000	Not reported
Nb2O5–CNT//activated carbon [228]	0.5–3 V	33.5 at 82	4 at 4000	Not reported
LiNbO3 on graphene aerogel//boron carbonitride nanotube [229]	1–4 V	148 at 200	69.4 at 9900	82% over 7000 cycles
MoS2–C-RGO//PANI-derived porous carbon [219]	0–4 V	188 at 200	45.3 at 40,000	80% over 10,000 cycles
NbN//activated PANI-derived carbon [230]	0–4 V	149 at 200	5 at 45,000	95% over 15,000 cycles
VN-rGO//activated carbon [231]	0–4 V	162 at 200	64 at 10,000	83% over 1000 cycles
MnO cubes//activated carbon [232]	0–4 V	227 at 55	21 at 2952	93% over 3000 cycles
3D MnO array//activated carbon nanosheets [233]	1–4 V	184 at 83	83 at 18,000	83% over 5000 cycles
MnO nanoparticles//activated carbon [234]	0–4 V	220 at 100	35 at 2608	95.3% over 3600 cycles
MnO on C//trisodium citrate-derived carbon [235]	0–3.9 V	235 at 120	61 at 25,000	85.69% over 10,000 cycles
MnNCN//activated carbon [236]	0.1–4 V	103 at 150	22 at 4500	100% over 5000 cycles
FexO on graphene//porous graphene [237]	0–3.5 V	129.6 at 19	45 at 3500	75% over 3000 cycles
Fe2O3//activated carbon [238]	0–3.5 V	90	Not reported	55% over 2500 cycles
Fe3O4 in graphene//3D graphene [239]	1–4 V	204 at 55	85 at 2650	70% over 1000 cycles
NiCo2O4//activated carbon [240]	0–4.5 V	39.4 at 120	10 at 554	100% over 2000 cycles
activated carbon//LiMn1/3Ni1/3Fe1/3O2–PANI [241]	0–3 V	49 at 900	19 at 3000	100% over 3000 cycles
activated carbon//LiMn2O4 [242]	0.7–3 V	45 at 60	10 at 800	Not reported
LiNi0.5Mn1.5O4//activated carbon [243]	1.5–3.25 V	19 at 120	8 at 3500	81% over 1000 cycles
mesocarbon microbeads//LiFePO4 [221]	2–4 V	69	Not reported	100% over 100 cycles
Li3V2(PO4)3//activated carbon [222]	0–4 V	125 at 300	65 at 6000	80% over 200 cycles
activated carbon//Li3V2(PO4)3 [222]	0–2.5 V	28 at 35	14 at 1500	87% over 1000 cycle
activated carbon//Li3V2(PO4)3 [105]	0.5–2.7 V	25 at 88	13 at 320	Not reported
Li2MnSiO4//activated carbon [244]	0–3 V	54 at 150	37 at 1500	85% over 1000 cycles
LiMnBO3//PANI [245]	0–3 V	42 at 1500	15 at 5350	91% over 1000 cycles
CoNiP2O7//activated carbon [246]	0–4 V	116.3 at 200	66.7 at 6486.5	86.5 over 500 cycles

7. Concluding Remarks

The final power and energy of LIC systems are based primarily on design, charging-storage structure and materials used in electrodes. Overall, designs with the smoothest anode and cathode voltage, and the highest overall voltage gap are ideal. Although existing LIC designs are reasonably straightforward in material terms, the associated high-rate charging and storage mechanisms in the electrodes (especially in the anode) remain poorly understood and require further research and development. Key unresolved issues for LICs include electrode design, energy-to-power limitations, fast charging mechanisms in anodes that vary from LIB activity, SEI formation, and LIC cycling (although superior to LIBs, not yet at traditional EDLC level).

Because of SEI development, the risk of low-voltage metal plating and volume expansion, it would seem as though the anode is the bottle-neck for cycling life. Although a high-surface-area cathode may function at a voltage where a CEI is formed, the overall chemistry would be less harmful due to reduced volume expansion (CEI does not accumulate through cycling). Despite this, the creation of the CEI for LICs is almost entirely unexplored. In terms of energy density, there currently does not appear to be an ideal cathode material that could operate at comparable capacity as the anode while maintaining cycling stability. Activated carbon is stable but offers up to one-fifth of a hard carbon anode reversible capacity. More specifically, many features of carbon-based electrodes in LICs still need further investigation. One is cycling stability during high current density testing. As long-term cycling stability is essential for LICs, it appears important to evaluate the advantages of using carbon-based electrodes in LICs.

The developments in LICs are mainly due to the production of advanced carbon-based electrodes. Although significant strides have been made in the manufacture of carbon-based electrodes, more research remains to be done. The inherent electrochemical performance of the electrode materials and the growth of active materials on the electrode depends on the porous structure and surface properties of the materials used. Although many materials have been designed to either serve as the electrodes or to assist other active materials, their microstructures are disordered and random in most situations, their surface properties are not well regulated and some pores of these materials are not electrochemically available for the electrolyte, restricting their electrochemical efficiency. Furthermore, the physical properties of some materials provide additional areas of concern through their deformation during charge and discharge cycles. Although ambitious, attention should be given to the design of electrodes with precisely controllable microstructures.