Potassium-ion battery cathodes: Past, present, and prospects

Journal of Power Sources 484 (2021) 229307

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

Potassium-ion battery cathodes: Past, present, and prospects Zhenrui Wu a, b, Jian Zou a, Shulin Chen c, Xiaobin Niu a, Jian Liu b, **, Liping Wang a, d, * a

School of Materials and Energy, University of Electronic Science and Technology of China, Chengdu, 610054, China School of Engineering, The University of British Columbia, Kelowna, V1V 1V7, Canada c Electron Microscopy Laboratory, School of Physics, Peking University, Beijing, 100871, China d Tianmu Lake Institute of Advanced Energy Storage Technologies, Changzhou, 213300, China b

H I G H L I G H T S

G R A P H I C A L A B S T R A C T

• Summarized the technical parameters of all reported KIB cathodes and gave benefits/drawbacks of each category. • Brought novel cathodes onto the table and highlighted the most promising cathodes. • Clarified some stereotypes of KIB cath­ odes based on current development. • Justified the prospects of KIB systems.

A R T I C L E I N F O

A B S T R A C T

Keywords: Potassium-ion batteries Cathode materials Prussian blue Tunnel-structured manganese oxides Vanadium pentoxides

Since 2004, potassium-ion batteries (KIBs) have shown the merits of high energy densities and high power densities at low costs. To further improve their overall performance, it is essential to understand the re­ quirements for cathodes in KIBs and screen out structures targeting at accommodating large-sized K ions. In this review, reported KIB cathodes are classified according to their chemistries, crystal structures, and ionic storage mechanisms. For each category, electrochemical properties are compared in detail; advantages/disadvantages are given. Regarding the critical criteria for practical applications (low cost, long cycling lifespan, high energy density, and high power density), Prussian blue analogs (PBAs) are the most suitable KIB cathodes now with a specific energy density of 535 Wh kg− 1 in half cells. Other cathodes such as tunnel-type manganese oxide, bilayered vanadium pentoxide, and some organic compounds also exhibit a stable energy density of >500 Wh kg− 1. The development of KIB cathode materials has enriched the fundamental understanding of materials design, synthesis, and ionic storage mechanisms. However, it is still a challenge nowadays to develop competitive high energy density KIB cathodes with a stable and long cycling life.

1. Introduction Due to fossil fuels’ consumption, researchers have started to develop

novel stable energy storage techniques to store electricity from renew­ able sources, such as wind and solar power, to meet the increasingly growing energy demands. Since lithium-ion batteries (LIBs) were firstly

* Corresponding author. School of Materials and Energy, University of Electronic Science and Technology of China, Chengdu, 610054, China. ** Corresponding author. E-mail addresses: [email protected] (J. Liu), [email protected] (L. Wang). https://doi.org/10.1016/j.jpowsour.2020.229307 Received 17 October 2020; Accepted 29 November 2020 Available online 8 December 2020 0378-7753/© 2020 Elsevier B.V. All rights reserved.

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

per ton) is one order of magnitude lower than lithium metal. Moreover, the transitional metal compounds in cathode synthesis for KIBs (such as manganese, iron, vanadium) are cheaper than cobalt - a very often used element in commercial LIB cathodes. Similarly, K-salts in the electrolyte have a lower cost than Li-salts. For example, the cost of KPF6 (US$338 per kg) is only <5% of LiPF6 (US$7113 per kg) [9]. Besides, instead of overpriced copper, aluminum can be used in KIBs as the current col­ lector since potassium metal does not form an alloy with aluminum at low potentials, reducing both the cost and the weight. In a full KIB cell, cheaper hard carbon (HC) from a facile pyrolysis method is commonly used as the anode, while graphite synthesized with multi-steps is a commercial anode for LIBs with a generally higher expense. Further­ more, in the battery industry, safety and stability are also big parts of the cost-reduction strategy. Many studies have presented highly reversible charging/discharging (C/D) capabilities of KIBs with more than 500 cycles at 80% capacity retention [10–14]. Second, KIBs can generate high potentials because of the relatively lower redox potential of K+/K. The potential of K+/K redox pair (− 2.936 V versus the standard hydrogen electrode [SHE]) is comparable to that of Li+/Li (− 3.040 V), much lower than Na+/Na (− 2.714 V). The low redox potential of K+/K gives high operational voltages of KIBs and ensures even high energy density. Remarkably, the standard potential of K+/K (− 2.88 V versus SHE) is even 0.09 V lower than that of Li+/Li (− 2.79 V) in propylene carbonate (PC) [15]. Even in the mixed solvents of ethylene carbonate (EC) with diethyl carbonate (DEC), the redox potential of K+/K is also 0.15 V lower than that of Li+/Li; in this way, KIB systems realize higher operational voltages than LIBs, which is different from Na-ion batteries (NIBs) and Mg-ion batteries (MIBs) [16]. Taking K0FeSO4F cathode reaction as an example: K0FeSO4F is obtained via electrochemical oxidation from the K1FeSO4F in a Li//K1FeSO4F half cell. This cathode realized the highest intercalation/deintercalation potential of 4.0 V versus K+/K in K//K0FeSO4F half cell, higher than the 3.7 V of Li0FeSO4F (versus Li+/Li) and the 3.4 V of Na0FeSO4F (versus Na+/Na) [29]. Recently, we discovered that fluorinated carbons (CF0.88) as cathodes also have an even higher energy density of 805 Wh kg− 1 at 2C in K cells than that in Li cells (776 Wh kg− 1) [30]. It is worth noting that the lower standard reduction potential of K+/K (− 2.88 V) than − 2.56 V for Na+/Na in carbonate solvents allows for the formation of a stage-I intercalation compound (KC8) at an average operational poten­ tial of ca. 0.3 V versus K+/K, realizing a high specific capacity of 250 mAh g− 1 at the fully discharged state; however, Na-ion storage in graphite is still difficult to achieve so far [15,16]. Ji and co-workers made a breakthrough and reported a reversible K-ion inter­ calation/deintercalation in graphite, presenting the feasibility of trans­ ferring the well-developed LIBs to the low-cost KIBs at the industry level [31]. As a result, despite most KIB cathodes’ limited capacity, high en­ ergy densities of KIBs can be compensated by their cathode materials’ higher operational potentials. Third, fast ionic kinetics of K ions in the electrolytes ensures high rate performances of the C/D process at high current densities; many studies reported the high power densities of KIBs and indicated KIBs’ prospect in applications such as large storage system (EES) [32]. Studies indicate that although K+ has the largest ionic radius (1.38 Å) compared with Li+ (0.76 Å) and Na+ (1.02 Å), K+ has the smallest Stokes’ radius (3.6 Å) compared with Li+ (4.8 Å) and Na+ (4.6 Å) in PC solvents [17]. As a result, K+ has the highest ionic mobility (i.e., diffusion coefficient, D+ A) and ionic conductivity (σ+ A ), as shown in Table 1. The mechanism has been studied in detail: due to the lower charge density, larger-sized K ions have less Lewis acidity than Li ions and Na ions. After solvated with solvents, solvated-K ions have smaller Stokes’ radii in liquid electrolytes, which renders the higher ionic mobility, larger transference number, and lower desolvation energy [12,13]. In this way, K+ exhibits faster ionic diffusion kinetics than Li+ and Na+. The good ionic mobility of K ions can be further proved by the high ionic conductivity of K+ in both aqueous and non-aqueous electrolytes [33–35]. Based on the advantages above, replacing Li ions with K ions as the charge carriers will enable

Table 1 Physical and chemical properties of lithium, sodium, and potassium elements. Properties

Lithium

Sodium

Potassium

Atomic number Atomic mass (g mol− 1) Density at 25 ◦ C (g cm− 3) Ionic radius (Å) Crust abundance (wt.%) Crust abundance (mol.%) Cost of carbonate (US$ ton− 1) Cost of industrial grade metal (US$ ton− 1) Cost of APF6 (US$ kg− 1) [9] Cost of AClO4 (US$ kg− 1) [9] E0 (A+aq/A) vs. SHE (V) [15] E0 (A+PC/A) vs. SHE (V) [15] E0 (A+EC/DEC/A) vs. Li+EC/DEC/Li (V) [16] Strokes’ radius in EC (Å) [17] Strokes’ radius in PC (Å) [17] Solvation energy in EC (eV) [18] Desolvation energy in PC (eV) [19] Desolvation energy in EC (eV) [19] Desolvation energy in DEC (eV) [19] Desolvation energy in EMC (eV) [19] − 1 σ+ A in 0.8 M APF6-PC (mS cm ) [20] − 1 σ+ A in 1 M AFSI-EC/DEC (mS cm ) [16] − 1 σ+ A in 1 M APF6-EC/DMC (mS cm ) [21] − 1 σ+ A in 1 M AFSI-DME (mS cm ) − 9 D+ cm2 s− 1) A in graphite ( × 10 − 9 2 − 1 D+ in HC ( × 10 cm s ) A

3 6.941 0.534 0.76 0.0017 0.005 6500 100,000 7113 2119 − 3.040 − 2.79 0 1.95 4.8 5.85 2.24 2.17 2.14 2.07 4.38 9.3 9.8 16.9 [22] 9 [25] 2.1 [27]

11 22.9898 0.97 1.02 2.3 2.1 200 3000 2589 282 − 2.714 − 2.56 0.3 2.34 4.6 4.76 1.65 1.59 1.54 1.49 6.38 9.7 10 16 [23] – 1 [28]

19 39.0983 0.89 1.38 1.5 0.78 1000 13,000 338 248 − 2.936 − 2.88 − 0.15 2.36 3.6 4.12 1.24 1.19 1.09 1.06 6.55 10.7 11.5 13.1 [24] 0.6 [26] 1.2 [28]

commercialized in 1991, LIBs have become the optimal choice for the current market of rechargeable batteries due to their high energy den­ sities and long cycling lifespans [1]. However, LIBs are now facing a host of challenges: slow energy density improvement, insufficient cost re­ ductions, and unsecured resource supply [2]. Notwithstanding the dominance of LIBs in portable devices, electric vehicles (EVs), and hybrid electric vehicles (HEVs), LIBs are costly at present for applica­ tions in large-scale energy storage systems (ESSs) such as smart grids and public transportations [3]. 40% of LIB production cost nowadays comes from electrode materials, leading to surging demand for lithium and transition metal resources (such as cobalt) [4]. Recent studies have indicated that the ever-increasing economic prosperity in LIBs market­ place comes at a cost to the environment in the form of resource depletion [4]. What is more, the low crust abundance and massive consumption of lithium and cobalt resources also lift the cost of LIBs up, which may one day become an issue, especially considering the scarcity of lithium and some of the transition metals currently used in LIBs [5]. As a result, novel and low-cost energy storage systems are urged to be researched and developed. Essentially, LIBs are concentration cells following the ‘rocking-chair’ prototype: Li-ions deintercalate from the cathode, move through the electrolyte, and insert into the anode, resulting in a Li+-deficit cathode and a Li+-rich anode during the charging process, and vice versa during the discharging process. Considering the large radii and high valences of Mg2+ and Al3+, creating a strong solvation effect and hindering the insertion process, K+ with similar physicochemical properties as Li+ and richer abundance have drawn more attention. Since the 2000s, re­ searchers have intensively studied potassium-ion batteries (KIBs), rep­ resenting low-cost substitutes for LIBs. The necessity and rationality of KIB cathode development are discussed below based on the production cost, energy density, and power density. First, the cost of KIBs can be largely cut down, considering the abundant resources and cheap anodes. Potassium is the second most abundant element among alkali and alkaline earth elements in the earth’s crust (Ca > Na ≈ K > Mg>…>Li), bringing in a cost-benefit [6, 7]. As listed in Table 1, the crust abundance of potassium is 1.5 wt.%, close to sodium (2.3 wt.%) and 1000 times larger than that of lithium (17 ppm); potassium carbonate costs US$1000 per ton versus US$6500 of Li2CO3 [8]; the cost of industrial-grade potassium metal (US$13,000 2

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

Fig. 1. The working principle of KIBs (a), and the development trajectory of KIB cathode from 2004 (b).

batteries to achieve better rate performance at a higher loading of active materials without sacrificing specific capacity. Moreover, K ions have the lowest desolvation energy compared with Li ions and Na ions in many liquid organic solvents (e.g., PC, EC, DEC, DMC in Table 1), leading to good ionic kinetics and promising rate capability [32]. Due to the decent energy density and high power density at a low cost, the KIB strategy has appeared to be a promising substitute for largescale energy storage applications. Fig. 1a shows both the macroscopic and microscopic processes of KIB in the C/D process. The electricity produced from clean energy such as wind power and solar power can be stored as chemical energy in KIBs in the charging process. At the mi­ croscopy level, electrons move from the cathode (e.g., K2Mn[Fe(CN)6]) to the anode (e.g., Hard carbon) via an external circuit while K ions as internal charge carriers migrate to the anode through the electrolyte and separator, creating electromotive forces and a huge potential gap. In the discharging process, K ions escape from the anode, diffuse through the separator and eventually land in the cathode, while the electrons flow through the external circuit and power the electrical appliances. Based on the benefits, various components in KIBs such as the cathode, anode, electrolytes, binders, and separators have been inves­ tigated and achieve much progress in the past few decades. Regarding the anode in KIBs, in real applications, K metal’s use is not likely to be accepted due to its high chemical activity. Recently, soft carbon [16], hard carbon (HC) [31], and graphite [36] were evaluated as anode materials in KIBs: compared to soft carbon, graphite exhibits a more pronounced capacity fade and lower rate capability, but both deliver high capacities of about 260 mAh g− 1 [31]. HC’s use in KIBs is inter­ esting as the capacity is mostly delivered above 0.1 V, which reduces the risk of K-metal plating [36]. Till now, the anode materials including carbon (e.g., graphite [37,38], HC [39,40]), metal alloys (e.g., K–Sb) [41], intercalation compounds (e.g., K2Ti4O9 [42], K2Ti8O17 [43], NASICON-KTi2(PO4)3 [44], Ti3C2 MXene [45]), conversion anodes (e.g., Co3O4–Fe2O3) [46], and organic materials (e.g., K2TP [47], K2PC [48]) have been researched; these anodes exhibited good electrochemical performances and promising application prospects. In fact, the research on the anode materials is significantly ahead of the cathodes in many aspects, such as the specific capacity and cycling stability [49]. In this way, KIB cathode materials should be intensively studied to develop decent cathodes for full KIB cells considering their relatively limited capacity in current researches. In Fig. 1b, the energy density evolution of KIB cathodes are shown by year. In 2004, Eftekhari et al. proposed the first article about a suitable cathode in KIBs, i.e., Prussian blue (PB) [50]. Afterwards, various studies on other cathode materials for KIBs have started to thrive especially since 2012. As shown in Fig. 1b, five categories of cathode materials

have been reported these years including: first, potassium metal poly­ anionic compounds (KMPs) (FeSO4F [29], FePO4 [51], K3V2(PO4)3 [KVP] [52,53], KVPO4F [54–58], KVOPO4 [57,59], KVP2O7 [60], KFePO4 [KFP]/KMnPO4 [KMP] [61], K3V2(PO4)2F3 [KVP-F] [62], KxFe (SO4)2 [KFS] [63], KTiPO4F [64]); second, transitional metal oxides (TMOs) (K2FeO4 [65], KxMnO2 [KMO] [10,66], KxFe0.5Mn0.5O2 [FMO] [67,68], KxCoO2 [KCO] [69,70], KxNiyCozMn1-y-zO2 [NCM] [71], KxV2O5 [KVO] [11,72,73], KCrO2 [74], KMn8O16 [75–77], KxCo0.5Mn0.5O2 [CMO] [78,79], KxNiyMn1-yO2 [NMO] [80–82], KFeO2 [KFO] [83], KxMgyMn1-yO2 [MMO] [84]); third, organic materials (3,4, 9,10-perylene-tetracarboxylic acid-dianhydride [PTCDA] [85–89], poly (anthraquinonyl sulfide) [PAQS] [90], polytriphenylamine [PTPAn] [91], polyaniline [PAn] [92], copper-tetracyanoquinodimethane [CuTCNQ] [93], anthraquinone-1,5-disulfonic acid sodium [AQDS] [94], poly(pentacenetetrone sulfide) [PPTS] [95], poly(Nvinylcarba­ zole) [PVK] [96], poly(N,N′ -diphenyl-p-phenylenediamine) [PDPPD] [97], octahydroxytetraazapentacene [OHTAP] [98], poly(N-phenyl-5, 10-dihydrophenazine) [PDPPZ] [14], 5,15-bis(ethynyl)-10,20-diphe­ nylporphinato]copper(II) [CuDEPP] [99]); fourth, Prussian blue ana­ logs (PBAs) (K1Fe[Fe(CN)6] [PB] [12,50], K2Mn[Fe(CN)6] [Mn-HCF] [13,100–104], K0Fe[Fe(CN)6] [BG] [105,106], K2Fe[Fe(CN)6] [PW] [107–110], K4[Fe(CN)6] [111]); and fifth, other materials (S [112–115], Se [116,117], SeS2 [118,119], Te [120], O2 [121,122], CO2 [123], I2 [124–126], graphite intercalation compound [GIC] [127,128], KMnF3 [129–131]). At present, most reported reactions of cathode materials in KIBs are intercalation-type [132]. These cathode materials (i.e., TMOs, KMPs, PBAs, and organic materials) normally have open structures for K-ion intercalation/deintercalation. Besides, conversion-type cathodes are discussed in this article as well. Among these materials, the massive volume changes of TMOs in the continuous cycling process undoubtedly result in pulverization and create some “dead” areas, which are elec­ trically isolated from the conductive agents and cause capacity fading; organic materials achieve large specific capacity but with low working potential leading to limited energy density, let its multiple phase changes in charging/discharging process; conversion-type cathodes such as I2, S@CMK-III are able to realize incomparable capacity while they are extremely unstable in cycling at the same time. As a result, despite the complicated synthesis steps, KMPs and PBAs are promising candidates for future KIB cathodes in a general scope due to their minor volume changes, excellent ionic kinetics, and thus admirable stability. Primarily, Mn-modified Prussian blue (KxMA[MB(CN)6]1− δ⋅nH2O) has already achieved a decent energy density of above 500 Wh kg− 1 with admirable cycling stability of over 500 cycles before the specific ca­ pacity drops to 80% of the initial number. 3

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

Fig. 2. Electrochemical properties of TMO cathodes: a) evolution of TMO cathodes (solid points stand for good cycling performance of more than 300 cycles at 80% of capacity retention; hollow points stand for poor cycling performance of fewer than 300 cycles before retention of 80%), C/D profiles of b) K0.77MnO2⋅0.23H2O, c) KCo0.4Mn7.6O16, d) δ-K0.42V2O5⋅0.25H2O and structures (e–g) of best-performed cathodes. Three structures of TMO cathodes: e) monolayered, f) tunnel-type, and g) bilayered structures. b) Reproduced with permission [10]. Copyright 2019, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. c) Reproduced with permission [76]. Copyright 2019, The Royal Society of Chemistry. d) Reproduced with permission [73]. Copyright 2018, American Chemical Society. e) Reproduced with permission [74]. Copyright 2018, American Chemical Society. f) Reproduced with permission [77]. Copyright 2018, Elsevier Ltd. g) Reproduced with permission [139]. Copyright 2018, American Chemical Society.

In terms of studies on electrolytes and their additives, 0.5–1 M KPF6EC/DEC (1:1 vol.) remains the most popular collocation among all nonaqueous electrolytes used in KIB articles till now. Typically, 1–5 wt.% fluoroethylene carbonate (FEC) can be used as an electrolyte additive, especially for Prussian blue cathodes [13,104,110,133]. For the organic cathode, 0.5–3 M KFSI in ether-based solvents has shown compatibility as well [86]. Studies indicate that these collocations are beneficial for stable electrode-electrolyte interphase and smooth ionic kinetics [134]. For example, ether-based (e.g., DME) electrolytes normally generate better ionic conductivity than carbonate-based ones. Developing elec­ trolytes with proper addictives bearing higher potential, stabilizing the SEI layer, and depressing further side-reactions is a priority in K-ion battery up-gradation [135]. Komaba et al. demonstrated that in EC/DEC (1:1 vol) solvent, KFSI [potassium bis(fluorosulfonyl)imide] is the most suitable conductive salt in comparison with KClO4 and KPF6 due to its higher solubility and ionic conductivity [16]. To date, ‘solvent in salt’ as a novel electrolyte system was studied in Wang et al., which can improve

the cathode materials’ cycling life. Regarding the binders, poly­ vinylidene fluoride (PVDF) is the most commonly used one for the non-aqueous system [109]; other popular binders are polytetrafluoro­ ethylene (PTFE) [55,65,74,83,136–138], carboxyl methyl cellulose (CMC) [62,72,91,95,112,139,140], sodium alginate (SA), teflonized acetylene black (TAB) [78], and LA132 [96]. In this article, the electrochemical performance of the five categories of cathodes is thoroughly summarized. The reasons for their different kinetics in LIBs, NIBs, and KIBs are discussed in aspects of solvation effect, ionic transport in electrolytes, ionic diffusion, and electron transfer in solid electrodes and solid electrolyte interphase (SEI) layer. Further details on mechanisms of ionic diffusion trajectory are stated as well. 2. Layered transitional metal oxides (TMOs) The first reports on K+-based layered transitional-metal compounds 4

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

edge-sharing MnIIIO6 octahedra. A sieve-like structure in a × b plane is shown in Fig. 2f, which consists of double chains of edge-shared MnIIIO6 octahedra. Along the c axis, (2 × 2) and (1 × 1) tunnels are observed; therefore, this structure is denoted as tunnel-type. Fig. 2g illustrates the bilayered structure of δ- K0.42V2O5⋅0.25H2O: double layers of vanadium oxides are separated by interstitial water, in this structure with an expanded interplanar spacing of 9.65 Å. TMOs used as cathode materials generally have layered structures due to their short-pathed ionic diffusion in the solid-state. Concerns about the rigid structures of TMOs remain, which might hinder largesized K-ion intercalation. At the very beginning, KxMO2, such as K0.5MnO2 and K0.6CoO2, was firstly studied in this category with facile synthesis routes and decent energy densities for KIBs. However, it re­ mains a challenge to maintain structural stability during cycling. Compared with its counterpart in LIBs, KxMO2 exhibits a more slope-like C/D profile due to multiple phase transitions and structural rearrange­ ment. The research of K-TMOs began with potassium manganese oxides. Utilizing starch-derived HC as the anode and layered birnessite (K0.3MnO2) as the cathode, a full-cell KIB with excellent cycling per­ formance achieved a specific capacity of 136 mAh g− 1 (0.55 K+) at a potential of 1.5–4.0 V at a current rate of 0.1C (27.9 mA g− 1) with 1.5 M KFSI in EC: DMC (1:1 vol.) as the electrolyte [66]. However, in 1.5–3.5 V, this material realized only 0.26 K+ insertion with retention of 91% at the 50th cycle and 57% at the 685th cycle. After that, Gao et al. syn­ thesized P2–K0.31MnO2⋅0.36H2O and improved the discharge capacity to 125 mAh g− 1 at 0.2C (16 mA g− 1) from 1.5 V to 4.0 V while main­ taining the excellent rate performance of >60% capacity retention at 5C [146]. Among all the reported KxMO2 materials to date, a high K-content birnessite cathode (K0.77MnO2⋅0.23H2O, space group: C1) proposed by Lin et al. delivered the highest specific capacity of 134 mAh g− 1 (93% retention after 100 cycles) with an operational potential of 2.6 V at a current density of 100 mA g− 1 [10]. In previous studies in LIBs and NIBs, to improve the specific capacity and structural stability of birnessite materials, Mn3+ has been

(KxTiS2, KxCoO2) were published in 1969 and 1975 [66,141]. After that, layered KxMnO2 (0.3 ≤ x ≤ 0.6) has been reported as a cathode material in aqueous K-ion capacitors (KICs) in the 1990s [142,143]. With a two-dimensional diffusion plane, TMOs and analogs KxMA2 (M = Sc, Ti, V, Cr, Mn, Fe, Co, Ni; A = O, S, Te) have been intensively studied in LIBs and NIBs. For KIBs, these studies have firstly burst into boom since 2016. The close-packed structures for this type of cathodes allow topotactic insertion/extraction of alkali ion with high capacities, low over-potentials, and high volumetric density [144]. Fig. 2a shows the voltage versus reversible gravimetric capacity of typical TMOs. It is worth mention that bilayered vanadium pentoxide exhibited the highest energy density of 678 Wh kg− 1, contributed by its high operational potential of ~3 V [73]. Conventional monolayered TMOs such as KxMnO2 and KxCoO2 exhibit medium operational poten­ tials of ~2.7 V (versus K+/K) and low energy densities of <400 Wh kg− 1. Recently, novel materials have been reported and have shown unique performance. For example, tunnel-structured KMn8O16 achieved a high specific capacity of >200 mAh g− 1 [77]; honeycomb-like K2NiTeO6 exhibited a high operational potential of 3.6 V (versus K+/K) [145]. Amorphous KFeO2 has been reported as a high-voltage cathode (3.43 V) in 2019; however, it suffered from a minimal specific capacity of 60 mAh g− 1 and an evident capacity fading to half of the initial after 50 cycles, which indicated an irreversible structural transition at a mild voltage window of 1.5–3.7 V [83]. Most TMOs exhibit monolayered structures. However, tunnel-type and bilayered structures have also been observed. Here, from each structure, the best performing representatives have been screened out: monolayered K0.77MnO2⋅0.23H2O [10], tunnel-type KCo0.4Mn7.6O16 [76], and bilayered δ-K0.42V2O5⋅0.25H2O [73]. Their typical C/D pro­ files are shown in Fig. 2b–d. Compared with K0.77MnO2⋅0.23H2O, KCo0.4Mn7.6O16 achieved better energy density by accommodating more K ions. However, bilayered vanadium pentoxide exhibited the highest energy density due to its high operational potential of ~3 V. Fig. 2e shows that K ions are harvested between the monolayers built by

Fig. 3. Structural changes of layered TMOs during C/D process and energetically favourable structures of TMO materials. a) Ionic storage difference of AxCoO2, A = Li, Na, K. b) C/D profile and phase transition of P3–K0.5MnO2. c) Phase transitions of common TMOs. d) site preference energy of Cr, Mn, Fe, Co, and Ni in KMO2; crystal structure of TM in e) octahedral, f) tetrahedral, and g) pyramidal sites. h) plots of the distance between nearest-neighbor K ions vs. ionic radius of metals in KMO2. a) Reproduced with permission [159]. Copyright 2017, The Royal Society of Chemistry. b) Reproduced with permission [166]. Copyright 2017, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. c) Data derived from refs: O2-LixCoO2 [159], O3-LixCoO2 [167], P2-NaxCoO2 [159], O3-NaxCoO2 [168], P2-KxCoO2 [70], P2-NaxMnO2 [167], P2-KxMnO2 [169], P2-KxMn2/3Ni1/3O2 [80], K5/9Mn7/9Ti2/9O2 [158], P3-KxMnO2 [166], P3-KxMn0.9Ni0.1O2 [82], P3-KxMn0.8Fe0.1Ni0.1O2 [170], O3-KxCrO2 [74], O3-KxCrS2 [137]). d-h) Reproduced with permission [74]. Copyright 2018, American Chemical Society. 5

­

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

substituted by other elements such as Li [147,148], Ni [149,150], Zn [151], Cu [152], Al [153], Co [150], Fe [154], and Ti [155]. Nickel is previously studied as a dopant in KIBs because the ionic radius of Ni2+ (0.69 Å) is close to Mn3+ (0.645 Å) [156]. Cobalt doping and iron doping have also been proved to be an effective way to improve overall per­ formance [70]. For example, cobalt-doped P3–K0.45Co0.5Mn0.5O2 ob­ tained a high operational potential of 2.85 V [78]; iron-doped P2–K0.65Fe0.5Mn0.5O2 exhibited a high specific capacity of 151 mAh g− 1 and a high energy density of 377.5 Wh kg− 1, the second-highest energy density following vanadium oxides among layered materials [67]. Here, P2, P3, and O3 refer to different structures of KMOs: the alphabet P or O indicates the oxygen environment of K atoms to be prismatic or octahedral; the digit 2 or 3 refer to the number of critical interlayers between two different oxygen layers [157]. Very recently, K5/9Mn7/9Ti2/9O2, partially replacing Mn4+ with Ti4+, realized a stable P2-OP4 phase transition like NaxMnO2. The structural decay is mostly reduced. The existence of Ti4+ with the same valence state and close ionic radius to Mn4+ (0.605 Å versus 0.53 Å) combated the gliding tendency of MnO6 slabs in an electrochemical process, leading to a highly reversible P2-OP4 phase transition compared with P2–O2 and O3–P3 transitions [158]. As an extension of LIBs, KxCoO2 (KCO) was expected to maintain excellent performance in KIBs. However, studies show that cobalt oxides have minimal performance in KIBs due to the incompatibility of largesized K ions and rigid-arranged layers consisting of edge-sharing CoIIIO6 octahedra [69,70,159]. In 2016, Kim et al. proposed the first KCO material as the cathode in KIBs: solid-state method synthesized P2–K0.6CoO2, which demonstrated a specific capacity of 80 mAh g− 1 with an operational potential of 2.7 V at a gentle current of 2 mA g− 1 from 1.7 V to 4 V [70]. After that, Deng et al. improved this material’s cycling performance by using a hydrothermal method before a solid-state reaction, and their material maintains 87% of the initial ca­ pacity (82 mAh g− 1) after 300 cycles at a current of 40 mA g− 1 [69]. Yuya et al. compared the interlayer spacing changes of KxCoO2 with NaxCoO2 and LiCoO2 and illustrated the phase transform during different ions intercalating [159]. K+ intercalation in the P2-KxCoO2 differs from Li+ intercalation in O3-LixCoO2, one of the most commonly used LIB materials. Theoretically, the radius of K+ is greater than that of Li+; However, solvated K+ has a smaller Stokes’ radius. In P2-type KxCoO2, K ions move through shared planes at two adjacent triangular prism sites, while Li+ migrates through adjacent edges at near tetrahe­ dral sites in O3-type LiCoO2 [160,161]. In this way, K ions in triangular prismatic sites electrostatically bear much weakly screened K–K repul­ sion by further apart Co3+ in the CoIIIO6 layer, which drives a lower activation barrier compared with Li-ions and leads to a more productive interaction [70]. Now, researches on KxCoO2 reach a bottleneck. At the premise of a reversible phase transition of KxCoO2, this material can barely accom­ modate >0.5 K+ extractions (low specific capacity). Taking AxCoO2 as an example, x is 1, 0.74, 0.6 for A = Li, Na, K. Large K ions broaden the distance of oxygen layers, hindering their screening effect on K–K repulsion; these ions also enlarge the crystal lattice, leading to irre­ versible dimensional changes [162]. As a result, to avoid the structural collapse, multiple plateaux of K+/vacancy ordering are observed to release growing K–K repulsion while discharging [70]; at the same time, limited sites can be taken by these large K ions. Nevertheless, Ding et al. presented that the c axis of AxCoO2 changed by 3.5% for 0.35 K+ and 4% for 0.45 Na+, which indicated the feasibility of realizing stable K-storage by highly reversible topotactic reactions [163]. Studies conclude three evident differences of ionic intercalation be­ tween K+ and Na+ in O3-type AxCoO2. First, as shown in Fig. 3a, the C/D curves of KxCoO2 are generally steeper than those of LiCoO2 and Nax CoO2. Second, the lower plateau of the intercalation step and more slopes in the C/D profile of P2-KxCoO2 present narrower intercalation ranges of K ions (0.33–0.68 K+ at 1.7–4.0 V) than Na ions in P2-NaxCoO2 (0.35–0.87 Na+ at 2.0 V–3.8 V) [162]. Third, slightly less capacity

witnessed more plateaux identified as K+/vacancy ordering in KxCoO2 under almost identical voltage window that is caused by a larger ionic size and a more expansive slab space. Notably, while discharging, O′ 3-type KCrO2 utilizes stepwise inter-slab slippage (e.g., O3→O′ 3→P′ 3→P3) to release the increasing K–K repulsion as K ions intercalate; after that, a reversible phase transition (O3→P′ 3→P3→O′ 3) was observed if keeping discharging [164,165]. As a result, a stoichio­ metric structure of KCrO2 can be achieved under the same voltage window. Similarly, moderate voltage curves of metal sulfides and pol­ yanionic compounds as KIB cathode materials are deduced owing to their naturally more substantial screening effect on K–K compulsion. Here, the prime (’) indicates a phase distortion from the hexagonal symmetry [157]. Another example is P3-type K0.5MnO2 [166], which delivered a reversible specific capacity of ~100 mAh g− 1 with excellent capacity retention, referring to Kim et al. First-principle calculations are intro­ duced to illustrate its phase transition of ‘P3 (R3m)→O3→X (O3-like)’ as shown in Fig. 3b, driven by the relative phase stability of different ox­ ygen stacking concerning the K content [166]. Later, new modifications were introduced by Zhao et al.: their AlF3 coated K1.36Mn3O6 (empiri­ cally equals to K0.46MnO2) delivered a highly reversible discharge ca­ pacity of 112 mAh g− 1 with a remarkably enhanced capacity retention of 94.9% at 100th cycle at a current density of 10 mA g− 1 [171]. In 2020, Deng et al. first utilized triethyl phosphate (TEP) as the solvent of a non-flammable electrolyte, in which K0.5MnO2 exhibited a high specific capacity of 120 mAh g− 1 at a wide voltage window of 2–4.2 V and high capacity retention of 84% after 400 cycles [172]. It is worth noting that in KIBs, these transitional metal oxides experience more evident phase transitions with a lower capacity compared with that in LIBs. In most cases, K-TMOs stably exist in Kdeficiency structures, exhibiting a phase transition from prismatic to octahedral structures as K ions are extracted from layers. A slope voltage curve was observed due to multiple phase transitions and K+/vacancy ordering to accommodate large K ions [169]. To conclude, an illustra­ tion of the phase transition of potassium TMOs is presented in Fig. 3c, which indicates that the K content in TMOs dominates the specific phase. It is worth noting that potassium transitional metal oxides must work with less K-ion mobility to maintain reversible structures, pre­ senting more staged features in C/D curves ascribing to the K+/vacancy ordering. In a layered oxide crystal structure, the TM central ions and alkali ions segregate into alternating slabs, forming a 2D open framework for the fast migration of alkali ions. Jahn-Teller (J-T) distortion brings lat­ tice changes and terrible crystal stability. J-T effect occurs in MnIIIO6 octahedra (Mn3+: e2gt22g) when KxMnO2 is deeply discharged [169] and in FeⅣO4 tetrahedra (Fe4+: e2gt22g) when K0.7FeO2 is intensely charged [83]. Interestingly, Na2Mn3O7, a layered sodium manganese oxide, was used as a cathode material in KIBs. First reported by Krishnakanth et al., Na2Mn3O7 delivered a specific capacity of 155 mAh g− 1 at 0.05C with an evident capacity fading since the J-T distortion in MnIIIO6 octahedra triggers the dissolution of Mn3+, leading to structural instability and active material loss [144]. Despite the low cost and environmental safety of manganese-based oxides, optimizations are needed to solve the issues such as the J-T distortion of Mn3+ and the simultaneous disproportionation of 2Mn3+→Mn2++Mn4+ [169]. These problems lead to severe asymmetric volume change and introduce unnecessary dissolution of Mn2+ into the electrolyte. Except for surface coating and hybrid cation doping, Lei et al. proposed a novel strategy of dual interfacial structure built by a K-poor spinel interlayer and an organic/inorganic solid-electrolyte interphase (SEI) film [169]. Such a structure can enhance the stability of Mn-based cathode in NIBs [173,174]. Furthermore, the underlying K+ storage mechanism in K-birnessite materials was studied via DFT-based first principle calculations by Gao et al., who firstly introduced concerted ionic diffusion for K-ion diffusion. Intrigued by abundant K vacancies, multiple K ions migrate together along the group diffusion 6

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

pathways [146]. Till now, oxides (or sulfides) of transition metals (including Mn, Co, Ti, V, Cr, Fe) and binary/trinary TMOs have been reported as feasible cathodes for KIBs. However, only KCrO2 rigidly realized K+-stoichio­ metric stability, while the others worked under a K-deficient structure for milder and more reversible phase transitions [74]. Most potassium transitional metal oxides presented K-deficient properties because the larger radius of K ions leads to a severer electrostatic repulsion. Even though some TMOs such as KFeO2 can be synthesized via a conventional solid-state method, their K+-stoichiometric structure is unstable in the C/D process, which presented capacity retention of 50% of the initial 60 mAh g− 1 after 50 cycles [83]. In this case, the crystal structure of KFeO2 (KAlO2-type, orthorhombic, Pbca) is distinctive from LiFeO2 and NaFeO2, both of which have stable layered structures [83]. TMOs with isostructure to KAlO2 (e.g., KFeO2) has long been considered as a solid-state electrolyte for K+ transportation because the ample space formed by corner-shared FeIIIO4 tetrahedra provides multiple diffusion channels [175–178]. These non-stoichiometric layered materials mentioned above such as KxMnO2 (x = 0.3 and 0.5), KxCoO2 (x = 0.4 and 0.6), and K0.7Fe0.5Mn0.5O2 are all K-deficient compounds. On the one hand, their stoichiometric compounds were generally crystallized into non-layered forms. On the other hand, the strong K–K electrostatic repulsion in a stoichiometric composition of KMO2 has a detrimental effect on the structural stability. However, K-deficiency (x ≤ 0.7 in KxMO2) limits the practical use of TMOs in rocking-chair batteries due to the possible low specific capacity. After 2018 when the first Cr-based TMO, P3–K0.69CrO2 delivered a high discharging capacity of >100 mAh g− 1 and excellent capacity retention of ~65% at the 1000th cycle at 1C [179], Kim et al. reported O3-type KCrO2, the first K-stoichiometric TMO cathode [74]: As shown in Fig. 3d, KCrO2 exhibited the most stable structure when Cr3+ occupied the octahedral sites; for other transitional metal ions such as Mn3+, Fe3+, Co3+, Ni3+, pyramidal sites are preferred. The detailed structures of octahedral (O), tetrahedral (T), and pyramidal (P) lattices of KCrO2 are given in Fig. 3e–g. However, despite the K-stoichiometry in KCrO2, the initial low Coulombic efficiency (50%), severe capacity loss, and the involvement of toxic chromium elements all prevent its practical application as an energy storage device. To further explain the K-defi­ ciency in many other TMOs and the K-stoichiometry in KCrO2, Fig. 3h compared the distance between nearest-neighbor K ions with the ionic radii of TM-ions in layered TMOs. Layers built by large TM-ions, such as Sc, In, Er, Tl, Y, Pr, and La, can easily accommodate K-ions; however, most small TM-ions cannot effectively screen out the strong K–K repul­ sion, destabilizing the layered structures. Ab-initio computation proved the above-hull energy (eV/f.u.) of transitional metals ordering as Fe > Co > Ni > Ti > Mn > V > 0>Cr. Therefore, the strong octahedral field of Cr3+ can resist K–K repulsion and maintain stable stoichiometric structures. Vanadium oxides are also promising TMO cathodes due to their multiple valence states, diverse phases, and high specific capacity in LIBs and NIBs [180–182]. As calculated via the first principle by Koch et al., metastable β-V2O5 showed the lowest insertion energies and the lowest diffusion barriers for K-ions, compared with orthorhombic α-V2O5, bronze-type VO2(B), and rutile-type VO2(R) [183]; while α-V2O5, built by tighter-arranged VO5 square pyramids, is a represen­ tative intercalation host for Li-ions and Na-ions due to the multiple oxidation states from V2+ to V5+ [184,185]. A typical vanadium oxide without structural water exhibited a high operating potential of 3.2 V and a medium capacity of 131 mAh g− 1 at 30 mA g− 1 [139]. After assembling a full cell with pre-potassiated graphite as anode, this cathode delivered a reversible capacity of 94 mAh g− 1 (with a mean operational potential of 2.53 V) and showed a high power density of 5480 W kg− 1. In 2018, Clites et al. reported δ-K0.42V2O5∙0.25H2O, exhibiting a highly reversible capacity of 226 mAh g− 1 at 20 mA g− 1 with an operational potential of ~3 V [73]. Notably, an extended age process is essential to synthesize a uniformed single-phase material of

bilayered δ-V2O5⋅nH2O xerogel [73,186]. To further improve the elec­ tronic conductivity, Vishnuprakash et al. proposed the strategy of V2O5@rGO composite: depositing vanadium oxide nanorods onto reduced graphene oxide nanosheets, which realized a high specific ca­ pacity of 222 mAh g− 1 at 0.5 C (147 mA g− 1) with high retention of 80% after 500 cycles under 2–4.5 V [11]. Although TMO materials have many advantages such as facile syn­ thesis and decent specific capacity compared with other KIB cathode materials, TMO cathodes’ theoretical discharge capacities in KIBs are always lower than in NIBs and LIBs. More specifically, K+ intercalation always presents the lowest in a single C/D plateau after normalization compared with Li+ and Na+ because multiple phase transitions occur in KIBs. Simultaneously, among AxMO2 cathodes, K-TMOs generally deliver the most limited capacity in such a given voltage window due to its generally steeper C/D profile. Besides, except for bilayered δ-K0.51V2O5 and honeycomb-like K2Ni2TeO6, this type of KIB cathodes’ operational potential usually is lower than 3 V, which limits the popu­ larity of TMOs in KIB research. Notably, bilayered δ-K0.51V2O5 owns high operational potential since its double oxygen layers vastly diminish the K–K repulsion; similarly, the high operational potential of K2Ni2TeO6 is attributed to TeⅥO6 stacking layers. 3. Prussian blue analogous (PBAs) Since the 1980s, Prussian blue (KFeIIIFeII(CN)6) and other hex­ acyanoferrates have been intensively studied as active materials for aqueous batteries. Until 2004, the pioneering work of Eftekhari was published, wherein PB was first utilized as a cathode in non-aqueous KIBs in academia [50]. With 1 M KBF4 in EC/EMC (3:7 wt.) as the electrolyte, this PB cathode realized 0.9 K+ intercalations (78 mAh g− 1) at an operational potential of 3.7 V with a capacity retention of 88% after 500 cycles [106,187–190]. It is regarded as a milestone in the development of Prussian blue analogous, starting a new era for PBAs: to be used as cathode materials in KIBs. The insight is to balance the restrained large-sized K-ion diffusion with loose three-dimensional structures. As expected, the diffusion co-efficient PB was 5 × 10− 11–4 × 10− 10 cm2 s− 1, more than one order of magnitude greater than most layered TMOs in KIBs [50]. In theory, KxMA[MB(CN)6]1− δ⋅nH2O (MA = Ti [191], Mn [13,100, 102,103], Fe [12,50,106–110,192–194], Co [108], Ni [133,187,195], Cu [196], Zn [197], Ba [198]; MB = Fe; 0≤x ≤ 1) (PBAs) can achieve 2 K+ (f.u.) embedding, corresponding to a specific capacity of ~156 mAh g− 1. With two FeIII sites as the active redox center, Fe-HCF delivered a higher specific capacity than the previously reported Fe4[Fe(CN)6]3 electrodes due to one-electron transfer per formula unit. Denoting Fe-HCF with 0, 1, 2 K+ as Berlin Green (BG)/Prussian Green (PG), Prussian Blue (PB), and Prussian White (PW), the entire redox reaction process for the insertion of K+ into Fe-HCF can be expressed by the following two steps [105]: i. Low-spin (LS) FeIII/FeII: BG/PG → PB III + − III II FeIII HS [FeLS (CN)6 ] + K + e →KFeHS [FeLS (CN)6 ]

ii. High-spin (HS) FeIII/FeII: PB → PW II + − II II KFeIII HS [FeLS (CN)6 ] + K + e →K2 FeHS [FeLS (CN)6 ]

Generally, HS-MII/III presented lower redox potential due to its lower t2g energy. Although the electron falls into t2g orbital for both HS FeIII and LS FeIII, the energy of LS t2g is always lower than that of HS one due to the larger splitting energy of the low-spin state. Splitting energy is mainly decided by the strength of the ligand field and is also influenced by the valence and size of the center TM atom. In most cases, ligands with a strong field introduce great splitting energy and vise versa. When splitting energy is greater than electron-pairing energy (Δ0>PE), a new 7

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

Fig. 4. Electrochemical performances and crystal structures of Prussian blue analogs: a) development of PBAs in electrochemical performance (Qdis refers to discharge capacity); C/D curves of b) Fe–HCFe, c) Mn–HCFe, and d) Mn–HCFe@Graphene; e) crystal structure of Mn–HCFe; f) phase transition and Jahn-Teller distortion in K-ion storage. b) Reproduced with permission [109]. Copyright 2017, The Royal Society of Chemistry. c,d) Reproduced with permission [13]. Copy­ right 2019, The Royal Society of Chemistry and the Centre National de la Recherche Scientifique. f) Reproduced with permission [104]. Copyright 2017, The Royal Society of Chemistry. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

hosted electron tends to pair with an unpaired electron in t2g orbital. After t2g orbital is filled with electrons, new electrons start occupying eg orbital, with higher energy. Therefore, strong-field ligands always lead to a low-spin state of the center atom. As K–PBAs discharging from 4 to 1.5 V, the electrons firstly fall into LS site with lower energy at the reductive potential of 3.4 V followed by HS site at 3.2 V; as charging, it falls into HS site and then LS site at the oxidative potential of 3.4 and 3.6 V respectively. One of the PBA materials’ competitive benefits is the minor volume change due to their loose 3-D structures. Large K+ seems tiny compared to the abundant space during the charging process, so ionic insertion does not lead to volume change. In this case, the minor changes are mainly due to the slight expansion of Fe ions as electrons insert into t2g orbitals. PBAs’ development as non-aqueous KIB cathodes is a history of replacing the Fe in N-octahedral with other transitional metals such as Cu, Co, Ni, Zn, Ti, Mn. Among all TM-HCF substitutes, Mn-HCF as the cathode in KIBs exhibits s higher operational potential of ~535 Wh kg− 1 that is close to commercial LiFePO4 (LFP, 578 Wh kg− 1) cathode in LIB, as shown in Fig. 4a.

Compared with the conventional K1.92Fe[Fe(CN)6]0.94⋅0.5H2O (Fig. 4b), Mn-HCF can achieve a more stable cycling performance and a higher operational potential; as shown in Fig. 4c, Mn-HCF also exhibited a high specific capacity of 133 mAh g− 1 at 0.1C and high capacity retention of 92.8% after 200 cycles [109]. To further improve the energy density, Liao et al. proposed K1.85Mn[Fe(CN)6]0.98⋅0.7H2O@graphene [13]. After Mn-HCF was synthesized by the same co-precipitation method, it was then mixed with graphene via a facile ball-milling method. Their Mn-HCF@graphene cathode exhibited an excellent life­ span of over 500 cycles before the capacity fading to 80%, and more importantly, the highest operational potential of ~4 V at 0.1C (Fig. 4d). The crystal structure of Mn-HCF is given in Fig. 4e. KxMA[MB(CN)6]1− δ⋅nH2O, where MA/MB denotes the N-coordinated/Ccoordinated TM ions and δ/n presents the contents of [MB(CN)6] va­ cancies/residual water, consists of a three-dimensional network of MA–N–C–MB chains. These chains form an ample cubic inner space allowing for facile insertion and extraction of alkaline cations where MA/MB ions are coordinated to six nitrogen/carbon atoms. The CN li­ gands connect these two octahedral units, noting as face-centered cubic 8

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

Fig. 5. Electrochemical performance of polyanion materials and crystal structure of layered KVOPO4: a) Polyanionic compounds have different active redox pairs and present different voltage and capacity; C/D profiles of b) KxFeSO4F, c) K1-xVOPO4, d) K1-xVPO4F; e) crystal structure illustration of the layered KVOPO4 synthesis process. b) Reproduced with permission [29]. Copyright 2012, American Chemical Society. c-d) Reproduced with permission [57]. Copyright 2017, The Royal Society of Chemistry. e) Reproduced with permission [213]. Copyright 2019, The Royal Society of Chemistry.

structure (FCC, space group Fm3m, a = 10.18 Å for FeFe(CN)6). MnN6 and FeC6 have different spin states. due to the difference of ligand field (C and N), Fe coordinated with six carbon atoms is in a low-spin (LS) state, and alternative transitional metals (Mn in this case) coordinated with six nitrogen atoms are in a high-spin (HS) state, denoted as octa­ hedral FeIII-C6 and MnII-N6 respectively. Weak ligand field of N atoms leads to narrower splitting energy (Δ0) of center metal atoms, which means electrons tend to fall into eg orbitals with higher energy after three t2g orbitals half-full inserted. In this way, a more stable structure with half-full orbitals denoted as HS-MII might be achieved at a cost to slight deformation from standard octahedral (Oh) to z-axis-enlarged octahedral (D4h). On the periodic table of elements, d-orbital is the most exterior electron layer for all transitional metal elements in the 4th period. With the electric field’s interference, an ideal sphere field with five degenerated d-orbital splits into two eg orbitals (doubly degenerate

electronic ground state) at a higher energy level and three t2g orbitals in lower energy level. The energy gap between the two energy levels is named splitting energy. To further reduce the overall energy, eg and t2g orbitals are split into b1g, a1g, eg and b2g orbitals, which decreases the orbital and electronic degeneracies. This also changes the electron cloud’s shape and eventually leads to the geometrical distortion (e.g., tetragonal elongation or compression in z-axis in octahedral high-spin d4 complexes) and turns the symmetric group of Oh into D4h. Bie et al. analyzed the phase transitions of K2Mn-HCF during K+ extraction/insertion from/into PBAs by operando XRD measurements [104]. During the initial charging process, two two-phase reactions were observed: as seen in Fig. 4f, the transition started from monoclinic (P21/n), where K ions occupied four sites in the lattice built by inter-connected MnIIN6 and FeIIC6 octahedra with cooperatively rotated cyanide anions. After that, it changed to cubic (Fm3m), tetragonal 9

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

(I4m2) structures, and vice versa. Interestingly, this structural evolution of K2Mn[Fe(CN)6] is similar to Na2-xMn[Fe(CN)6], where the phase transition from cubic to tetragonal can be ascribed to the Mn3+-triggered Jahn–Teller distortion in MnIIIN6 octahedra [199]. Mn2+ and Fe2+, the two transitional metal ions in K2Mn[Fe(CN)6], have a ground electron arrangement of 3d5 and 3d6 in d-orbitals. However, influenced by the CN ligands in the chain of Fe–CN–Mn, Mn2+ and Fe2+ are in high-spin and low-spin state. Thus, their detailed electron arrangement is t32ge2g and t62ge0g. Only MnN6 exhibited a Jahn-Teller distortion with the elon­ gating in the z-axis. This transformation reduces the overall energy and brings a larger space for K-ion diffusion. It is worth mentioning that the electrochemical performance of PBAs depends on the synthesis method with poor repeatability. The products obtained by the rapid precipitation method have high water content and many [Fe(CN)6] defects; thus, they often face various problems such as less K+ embedding, poor kinetics, low Coulombic efficiency, and short cycling life. These problems are likely caused by structural imperfection and abnormal cationic arrangement of the PB lattices: the bonded or interstitial water may affect the thermal stability of PBAs, bringing in concerns about low Coulombic efficiency and safety issues [200,201]. Therefore, slow crystallization at an atomic level, as well as subsequent drying and water removal steps, is of great importance to achieve better performance [202]. Currently, there is no reasonable explanation for this; thus, further development of scientific theories and meticulous experiments is needed to clarify the effects of synthesis conditions on crystallization, defects, and water content to improve repeatability and commercialize PBAs as KIB cathode earlier. The open-framework crystal structures of PBAs contain such large interstitial sites, allowing for fast kinetics of alkali ions and minor lattice strain; therefore, PBAs have been viewed as suitable hosts for Li+ [203] in non-aqueous LIBs, Na+ [204] in NIBs, and K+ [105] in KIBs. Comparing PBAs’ performance in NIBs and KIBs, it is found that K–PBAs achieve high energy densities with higher operational potentials but lower specific capacities. According to density functional theory (DFT), larger ions can form more reliable interactions with the metalorganic framework; thus, K+-embedded products after discharging will be more stable; in this process, much energy is consumed so that K–PBAs have to work under the highest potentials [205]. However, restrained by the decomposition voltage of non-aqueous electrolytes, the capacity of K–PBAs cannot be fully realized because the second K+ embedding is achieved at high potentials beyond safety, very likely leading to elec­ trolyte decomposition. Therefore, although CoIII/II substitution in Na–PBAs has been proved to be useful to generate extra capacity in non-aqueous NIBs, this strategy cannot be applied in KIBs. Because both iron-based and manganese-based PBAs have already exhibited such high second plateau potentials to realize second K+ embedding.

KMPs exhibited less specific capacity in KIBs than LIBs and NIBs, but more benefits, such as high operational potentials, anion-inductive diffusion, small K–K compulsion, and smooth 3-D tunnels, guarantee the relatively high energy densities. The research towards layered polyanion materials in KIBs can be dated back to 2018 when Vanadyl-phosphate di-hydrate VOPO4⋅2H2O was first explored as a cathode material for KIBs [59]. This material was built by VO(OH2)PO4 layers stacking along the c axis with K ions and interstitial water occupying the ample interlayer space [208]. VO(OH2) PO4 layers were built by VO6 octahedra and PO4 tetrahedra arranging like a two-dimensional rock-salt, where V atoms and four planar O atoms in a VⅤO6 octahedron were connected to PO4 tetrahedra while two apex O atoms in a VⅤO6 octahedron were dangling in interplanar space. The P–O covalent inductive effect was observed since the strong P–O bonds thin the electron density of V center and weaken the V–O covalent bond, which makes VOPO4⋅2H2O have the highest potential and makes KVPO4F (a fluorinated analog to VOPO4) present the highest energy density (455 Wh kg− 1 in K half cells) among all reported KMP cathode [59]. As shown in Fig. 5a, KMPs exhibited electrochemical advantages of high potentials or high specific capacity determined by the active redox pairs: V5+/V4+, V4+/V3+, and Fe3+/Fe2+: V5+/V4+ compounds overall exhibits both high operational potentials and high specific capacity, while V4+/V3+ and Fe3+/Fe2+ compounds correspondingly show more limited capacity and lower potentials. Moreover, KMP materials exhibit ample structures; thus, a secondary classification can be done regarding their structural features. In details, KTiPO4F [64], KVPO4F [54–57], α-KVOPO4 [57,209], and KFeSO4F [29] share the same orthorhombic structure (space group: Pna21) as KTiOPO4, denoted as KTP-type, in which TiⅣO6 octahedra and PO4 tetrahedra were connected through vertexes to form a robust framework with an ample 3D K-ion storage space [64]. K3V2(PO4)3 [52,53], K2.95Rb2(PO4)3 [210], and K3V2(PO4)2F3 [62] are NASICON-type compounds, while Fe2(MoO4)3 belongs to the anti-NASICON compounds [211]. Pyrophosphate com­ pounds consist of KVP2O7, KMoP2O7, KTiP2O7 [60], K2FeP2O7, K2MnP2O7 [61], K4Fe3(PO4)2P2O7 [212]; K ion insertion/extraction will trigger a conformational change of the two tetrahedral PO4 units in pyrophosphates [60]. Also, amorphous α-FePO4 has been reported in 2014 [51]. Interestingly, some KMPs (e.g., VOPO4⋅2H2O [59], KVOPO4 [213], KxFe(SO4)2 [63]) presented layered structures. According to the reported KTP-type materials to date, these cathodes regularly exhibit an operation potential of >3.5 V and a specific capacity of 90–120 mAh g− 1. Three typical KTP materials are KFeSO4F [29], α-KVOPO4, and KVPO4F [57]. The Ti atoms in KTiOPO4 can be replaced by other transitional metal atoms such as Fe and V; For V-replaced KTP materials, VⅣO6 orthorhombic can be substituted with VIIIO4F2 so that KVPO4F lattices can be constructed. KFeSO4F was reported by Recham et al. (2012), who starts the research of polyanion cathodes in KIBs [29]. As shown in Fig. 5b, with the redox pair of Fe2+/Fe3+, this material exhibited a decent specific capacity of 102 mAh g− 1 in K half-cells with a high operational potential of 3.5 V. This work shed light on the appli­ cation prospects of polyanion compounds as cathode materials in KIBs. In 2017, Chihara et al. firstly proposed vanadium KTP-type cathodes for KIBs with much higher operation potentials [57]. In Fig. 5c–d, with the redox pairs of V4+/V5+ and V3+/V4+, α-KVOPO4 and KVPO4F exhibited extremely high operational potentials of 4.02 V and 4.13 V, respectively, when discharging from 4.8 V to 2 V in 0.7 M KPF6 – EC/DEC (1:1 vol) electrolyte. KVPO4F in this work realized a specific capacity of 92 mAh g− 1 (theoretical capacity: 131 mAh g− 1); more importantly, this material showed excellent cycling stability with no capacity fading at the initial 50 cycles with a small volume change of 5.8% at fully discharged state (5.0 V). Very recently, Liao et al. proposed a full KIB using KVPO4F as the cathode and VPO4 as the anode, which achieved a stable capacity of 103 mAh g− 1 at 0.2C with a high operational potential of 3.1 V and excellent cycling stability of 86.8% capacity retention after 2000 cycles at 10C [54]. The superb rate performance (up to 50C) of this K-ion full

4. Potassium metal polyanionic compounds and pyrophosphates (KMPs) We denote potassium metal polyanionic compounds as KMPs. Large polyanions of (XO4)3 (X = S, P, Si, As, Mo, W) in the lattice can enhance the redox potentials as the cathodes and stabilize the structures [206]. Polyanionic compounds show similar 3-D tunnels to PBAs with less interstitial water and fewer transitional metal vacancies. Multiple an­ ions exhibited synergetic inductive effects to boost their extremely high operational potentials, contributing to their high energy density [64]. Also, these large-sized anions provide sufficient interlayer space for K-ion accommodation, ensuring structural stability in the C/D process. KFeC2O4F with a 3-D open framework structure exhibited excellent cycling stability: it realized a high specific capacity of ~112 mAh g− 1 at 0.2 A g− 1 with an excellent capacity retention of 94% at the 2000th cycle [207]. However, the high operational potentials of most KMPs introduce new demands for further development of electrolytes that can work at >4.2 V. At the same time, due to the heavy polyanion framework, the specific capacities of many KMPs are limited to <100 mAh g− 1. Till now, 10

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

Fig. 6. Electrochemical performance of organic cathode in K-ion batteries: a) organic cathodes electrochemical characteristics of reported organic cathodes; b) redox reaction mechanism of anthraquinone-1,5-disulfonic acid sodium (C14H6Na2O8S2, AQDS) salt as conjugated carbonyl type reacting with K ions and of PVK as ni­ trogen radical type reacting with anions such as PF−6 ; c) C/D profiles, d) rate performance, and e) cycling performance of PDPPZ cathode with 2.2 M KPF6 in diglyme as electrolyte cycling from 2.5 V to 4.5 V. c-e) Reproduced with permission [14]. Copyright 2019, American Chemical Society.

battery makes it one of the most promising alternatives to present large-scale energy storage systems. NASICON-K3V2(PO4)3 as KIB cathode exhibited a limited electro­ chemical performance, such as the low specific capacity of ~80 mAh g− 1 at an operational potential of 3.2 V. However, it can be used as both the cathode and the anode in symmetric full cells [52]. K3V2(PO4)3 trans­ formed into KV2(PO4)3 at ~ 4 V and K5V2(PO4)3 at <1 V. Based on this, the symmetric full K-ion cell exhibited a specific capacity of 90 mAh g− 1 from 0.01 V to 3 V at 25 mA g− 1 with a high capacity retention of 88.6% after 500 cycles. This full symmetric cell can effectively buffer the sig­ nificant volume changes because cathode expansion is always accom­ panied by anode shrinkage, and vice versa, leading to the consistency of the full-cell volume. To further improve the electrochemical perfor­ mance of NASICON-type cathodes, a similar fluorination strategy of changing VⅣO6 orthorhombic into VIIIO4F2 is applied. Taking advan­ tages of high electronegativity of F for high operational potentials, proposed by Lin et al., K3V2(PO4)2F3 realized a much higher average operational potential of 3.7 V than K3V2(PO4)3 (3.2 V) [52,62]. In this way, the energy density is much enhanced to ~400 Wh kg− 1.

F-substituted NASICON-KVP also exhibited excellent cycling perfor­ mance of 95% capacity retention after 180 cycles at 25 mA g− 1 and minor volume change of 6.2%. It is worth noting that via a delicate structure design, layered KVOPO4 (L-KVOP) exhibited a competitive energy density of >400 Wh kg− 1 [213]. As shown in Fig. 5e, firstly, K0.5VOPO4∙1.5H2O nanosheets were synthesized via a facile hydrothermal method; after chemical potassiation (by adding KI-methanol solution) and dehydration, inter­ stitial/coordinated water molecules were removed, leading to a rigid layered structure. Computational material science has a profound effect on experi­ mental research on metal-organic open framework conjunct with mul­ tiple anions. Park et al. combined the depth-first search (DFS) algorithm and DFT calculation to screen out compounds with a 3 Å of channel radius at a number density of 3 per 100 Å2 [60]. After considering the strong inductive effect in polyanionic ligands, the use of Hubbard U corrections between 3.1 and 4.2 eV provides more realistic values [214–216]. The results are seven pyrophosphates (KTiP2O7, KVP2O7, KCrP2O7, KFeP2O7, KMoP2O7, K2(VO)3(P2O7)2, and K2MnP2O7), two 11

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

Fig. 7. Electrochemical performance of K–S/Se/Te batteries: a) K–S cell model of K//Celgard/SWCNT//S/CNF with SWCNTs coating the Celgard separator on the cathode side; b) rate performance of S, Se, Te in LIBs, NIBs, and KIBs; c) schematic of the reaction mechanism in K–S batteries; C/D profiles of d) S2,3, e) Se@NPCS, and f) Te@CNT/rGO at multiple current densities; g) theoretical volume change and lattice transition of Te cathode in KIBs. a) Reproduce with permission [221]. Copyright 2018, Elsevier B.V. b,f,g) Reproduced with permission [222]. Copyright 2020, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. c) Reproduced with permission [223]. Copyright 2018, Published by Elsevier B.V. d) Reproduced with permission [112]. Copyright 2019, American Chemical Society. e) Reproduced with permission [224]. Copyright 2020, The Royal Society of Chemistry.

metaphosphates (KNi(PO3)3, KCo(PO3)3), and one vanadate (KMnVO4) [60]. The molecular weight limits the specific capacity of this type of material. Till now, only KVP2O7, KMoP2O7, and KTiP2O7 provide min­ imal specific capacities of 53 mAh g− 1, 25 mAh g− 1, and 22 mAh g− 1 cycling from 2 V to 5 V at 1C with an operational potential of 4.2 V [60]. Interestingly, based on DFT calculation, Fe4+/Fe3+ in KFeP2O7 is ex­ pected to have a redox potential of >5 V. In this way, high-voltage-tolerant electrolyte needs to be further developed. TMOs exhibit a medium operational potential of ~2.7 V and a me­ dium specific capacity of 90–150 mAh g− 1. PBAs achieve an improve­ ment of the operational potential of ~3.5 V while maintaining a highly reversible capacity of ~140 mAh g− 1. Some KMPs show extraordinarily high potential of ~4.3 V at the cost of specific capacity. Among the three

of them, PBAs undoubtedly exhibit the best overall electrochemical performance, representing the most promising KIB cathodes. 5. Organic compounds Organic materials show high reversible capacities, long cycling life, and good physical flexibility. Many of them exhibited excellent energy density of >500 Wh kg− 1, such as poly(pentacenetetrone sulfide) (PPTS) [95], poly(N-phenyl-5,10-dihydrophenazine) (PDPPZ) [14], and copper-tetracyanoquinodimethane (CuTCNQ) [93], representing the best-performed cathode in different electron-transfer mechanisms. However, for KIBs, organic materials are still underdeveloped in their infancies. Moreover, since there is no potassium source in organic 12

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

cathodes, a pre-potassiation process is required. Similar to NIBs, organic cathodes for KIBs are mainly divided into cation-insertion type and anion-insertion type by insertion/extraction ions, as shown in Fig. 6a. For those containing conjugated carbonyls, the electrophiles of K-salt (K+ in this case) attack conjugated carbonyls as unsaturated bonds and carry out electrophilic additions to transfer electrons. In this case, K-ion capturing happened since the carbon-tooxygen double bonds can attract and bind alkali metal ions. This cate­ gory includes traditional 3,4,9,10-perylene-tetracarboxylic acid dia­ nhydride (PTCDA), poly(anthraquinonyl sulfide) (PAQS), anthraquinone-1,5-disulfonic acid sodium salt (AQDS), poly(pentace­ netetrone sulphide) (PPTS), Mx(CO)n (n = 4, 5, 6), octahydroxyte­ traazapentacene (OHTAP), etc. Electrochemically, these materials have low operational potentials of <2.5 V, leading to limited energy densities of ~400 Wh kg− 1. For example, PPTS showed a low operational po­ tential of 1.6 V; however, it exhibited a high specific capacity of 260 mAh g− 1 at 100 mA g− 1 and good cycling performance of 74% retention after 3000 cycles at 5 A g− 1, with excellent rate performances of capacity retention of 84% at 2 A g− 1, 71% at 5 A g− 1, 65% at 10 A g− 1 [95]. Organics containing nitrogen radicals often exhibit much higher operational potentials of >3 V. Poly(N-vinyl carbazole) (PVK) even realized a very high operational potential of 4 V. Polytriphenylamine (PTPAn), Polyaniline (PAn), poly(N, N′ -diphenyl-p-phenylenediamine) (PDPPD) also belong to this category. Their positive charge centers electrostatically draw nucleophiles (such as FSI− , PF−6 ); in this way, electron transfer is accomplished. This process is denoted as nucleo­ philic addition. The two different mechanisms are presented in Fig. 6b. Till 2018 when Ma et al. proposed a novel composition compound CuTCNQ, researchers started realizing that the two mechanisms can be simultaneously achieved at a proper voltage window if the organic cathodes contain both functional groups [14]. In CuTCNQ, TCNQ2contains conjugated carbonyls; Cu2+ represents the positive charge center. Discharging from 4.1 V to 2 V, this material exhibited a moderate operational potential of 2.7 V and a reversible capacity of 180 mAh g− 1 after 15 cycles at 50 mA g− 1 without an evident decay before 50 cycles. At present, PDPPZ is the most promising organic cathode for KIBs. As shown in Fig. 6c–e, PDPPZ exhibited the highest energy density of 583.2 Wh kg− 1 and power density of above 104 W kg− 1 with excellent cycling performance of 79% capacity retention after 1000 cycles at 2 A g− 1 [14]. Furthermore, using dipotassium terephthalate (K2TP, ~0.4 V versus K+/K) as the anode, the performance of organic cathodes (such as PPTS [95], PDCDA [89]) in all-organic full KIBs are tested. Tang et al. re­ ported that PPTS//K2TP full cell exhibited a high reversible specific capacity of 254 mAh g− 1 with 66% retention after 100 cycles at 0.1 A g− 1 [95]. Based on the total mass of active materials on both sides, this full cell realized an energy density of 85 Wh kg− 1 with an operational potential of 1.4 V.

Fig. 8. A summary of promising high energy density cathode materials in KIBs.

chalcogen cathodes, they are usually paired with K metal (a much more active alkali metal than Li) as the anode in KIBs, which brings new is­ sues, such as internal stress and metal dendrites. The theoretical specific capacities are 1672 mAh g− 1 for S, 679 mAh g− 1 for Se, and 420 mAh g− 1 for Te. The electrical conductivities are 5 × 10− 28 S m− 1 for S [218], 1 × 10 − 3 S m− 1 for Se [219], and 2 × 102 S m− 1 for Te [220]. Chalcogen cathodes usually suffer from soluble intermediates (such as polysulfides and polyselenides) into the electrolyte, which not only causes the loss of active materials and low efficiency but also leads to the corrosion of metal anodes. For addressing the problem of active material dissolution, a practical solution is filling chalcogen powders into nano­ structured carbon foam and adding interlayers between the cathode and separator, as shown in Fig. 7a. Especially for sulfur, due to its extremely low electrical conductivity, nanoporous carbon products (e.g., CMK-3) are generally involved as a second conductive phase to better transfer electrons at the sacrifice of gravimetric energy density and volumetric energy density. To date, compared with LIBs and NIBs, chalcogen cathodes exhibited better rate performance in KIBs at a high current density of up to 15C (Fig. 7b). Cyclo-S8 and small molecular sulfur S2-4 as two main sulfur cathodes in K–S batteries present different C/D profiles [115]. The reaction mechanism of K–S cells is illustrated in Fig. 7c. Despite the lower operational potential, S2-4 often exhibits better rate performance and cycling stability than Cyclo-S8. In 2019, Xiong et al. reported a small molecule sulfur cathode (S2,3) with a high reversible specific capacity of 1198.3 mAh g− 1 and a moderate Coulombic efficiency of 97% (Fig. 7d). Later studies on selenium and tellurium with higher electrical conduc­ tivity show a similar two-electron conversion-type reaction to K–S bat­ teries. Very recently, Huang et al. reported high-content selenium cathode (60%) supported by a nitrogen-doped porous carbon sponge (NPCS), which presented a reversible specific capacity of 654 mAh g− 1 at 0.1C. However, the capacity retention is only ~50% at the 100th cycle. Also, as shown in Fig. 7e, this material exhibited evident capacity decay at higher current densities up to 1C. Te powder with glucose-CNTs matrix exhibited a high specific ca­ pacity of 419 mAh g− 1 (volumetric specific capacity: 2619 mAh cm− 3) without evident fading at the initial 100 cycles. It showed excellent rate performance at up to 15C (Fig. 7f) [222]. In this case, the carbon matrix, as a generally used strategy, was applied. This matrix effectively alle­ viated the volume expansion, which should be +398% from pristine Te to fully discharged K2Te (Fig. 7g). This strategy also suppressed the shuttle effect during the potassiation process. Furthermore, DFT was applied to calculate the formation energies with a certain number of K+ ions intercalating in pristine Te lattice, where K2Te presented the most thermodynamically stable structure with the lowest formation energy of

6. Conversion-type materials Most reported cathodes for KIBs are topotactic-type materials, which means electron transferring is realized by K-ion intercalation/dein­ tercalation in a specific lattice framework. On the one hand, this inert lattice framework provides diffusion channels for K ions, leading to excellent kinetic performance. On the other hand, it adds extra weight to the cathode in KIBs, limiting the theoretical capacity. Conversion-type composites, however, are much lighter than intercalation-type mate­ rials such as TMOs, PBA, and KMPs. The mechanism for conversion-type cathodes involves host-formation reaction or surface-conversion reac­ tion. Especially for nanocomposites without bulk diffusion they often exhibit excellent rate performance. KF-MnO nanocomposite realized 160 mAh g− 1 (0.8 K+/Mn) in the first discharging process with an operational potential of 2.4 V versus K+/K, presenting a high energy density of 390 Wh kg− 1 [217]. A significant group of conversion-type cathode materials is the chalcogens (oxygen family in group 16). Due to the medium potential of 13

Z. Wu et al.

Table 2 Electrochemical performance of reported high-energy-density cathode materials for KIBs. Materials

K0.77MnO2 ∙0.23H2O KCo0.4Mn7.6O16 δ-K0.42V2O5 ∙0.25H2O V2O5@rGO

14

K1.92Fe[Fe (CN)6] 0.94∙0.5H2O K1.75Mn[Fe (CN)6] 0.93∙0.16H2O KVOPO4

Space group

C2/m I4/m – Pmnm Fm3m

Electrolyte

0.8 M KPF6 - EC/DEC (1:1 vol) 1 M KPF6 - EC/DEC/PC (1:1:1 vol) 0.8 M KPF6 - EC/DEC (1:1 vol) 1 M KPF6 - EC/DEC (1:1 vol) 0.8 M KPF6 - EC/DEC +1v. % FEC

Loading

Current

1st Vave

Window

1st CE

1st Qdis

Capacity retention

Rate

D+ K

σe

(mg cm− 2)

(mA g− 1 @1C)

(V vs. K+/K)

(V vs. K+/ K)

(%)

(mAh g− 1)

(%@cycle)

(%@mA g− 1)

(cm2 s− 1)

(S cm− 1)

2–3

100

2.6

1.5–4

100

134

93%@100

57%@1000

1.86E−

1.33

10

2.6

1.5–4

96

205

66%@30

85%@100

–

10

− 14~− 18

0.37

20

3

2–4.3

98

226

74%@50

58%@300

E

–

0.5C

2.3

1.5–4.3

92

222

80%@500

2.8E−

2

0.1C @130

3.6

2–4.2

99

133

93%@200

91%@1C, 54% @2C 84%@5C, 75% @10C

13

8~− 9

E−

4.1E−

Volume change

6

Em

Ref.

(Wh kg− 1) d: 3.1%

349

[10]

–

–

533

[76]

–

d: -4.68%

678

[73]

–

–

511

[11]

*2.23E3 [225]

1.3%

479

[109]

–

c: 3%

536

[104]

–

9.4%

419

[213]

0.7 M KPF6 - EC/DEC (1:1 vol) + 2 v.% FEC

–

30

3.8

2–4.5

80

141

91%@100

76%@1000

–

P42mc

2

3.65

2–4.6

90

115

4.2

5-Feb

78

103

86%@5C, 71% @10C 85%@5A/g

9E

–

–

433

[54]

2

50

4.05

2–4.7

25

114

87%@100 @0.5C 83%@100 @50 mA/g 93%@100

–

3

0.2C @120 20

–

–

–

462

[96]

0.5

200

3.6

2.5–4.5

90

162

80%@500, 73%@1000 81%@2A/g

–

–

–

583

[14]

2

50

2.7

2–4.1

94

244

66%@200, 51%@1000

–

0.25

huge

659

[93]

KVPO4F

Pna21

PVK

–

0.5 M KPF6 - PC/FEC (1:1 vol) 0.5 M KPF6 - PC/FEC (1:1 vol) 1 M KPF6 - PC +2 v.% FEC

PDPPZ

–

2.2

CuTCNQ

Pn

–

M

KPF6 - Diglyme

96%@100, 79% @1000@2A/g 70%@50

− 11

Journal of Power Sources 484 (2021) 229307

P21/n

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

− 1.028 eV. Since K–S/Se/Te batteries are quite new cell models proposed very recently, there are issues on both chalcogen cathode and K metal anode sides. For example, the enormous theoretical volume change and the shuttle effect of polychalcogens in K-salt based electrolytes are still unclear. Despite tellurium’s limited capacity compared with sulfur and selenium, it still exhibited the best electrical conductivity and much improved high-rate performance in the chalcogen group.

the work reported in this paper. Acknowledgments Dr. Wang’s work was supported by the Fundamental Research Funds for the Central Universities, China [ZYGX2019Z008]. Dr. Liu’s work was supported by Nature Sciences and Engineering Research Council of Canada (NSERC), Canada Foundation for Innovation (CFI), BC Knowl­ edge Development Fund (BCKDF), and the University of British Columbia (UBC).

7. Conclusions and perspectives

References

We summarize the electrochemical performance distributions of typical KIB cathodes in Fig. 8 and Table 2. Among KIB cathodes, K2MnHCF (PBAs) exhibits a high energy density of 535 Wh kg− 1 and excellent rate performance of 75% capacity retention at 10C [104]. Despite the incomparable energy density of commercial LIB cathodes such as LiFePO4 (170 mAh g− 1, 3.4 V, 578 Wh kg− 1) and LiCoO2 (190 mAh g− 1, 4 V, 780 Wh kg− 1), K2Mn-HCF have realized excellent power density with decent cycling performance at a low cost. Some KMPs such as KVPO4F exhibit a high operational potential of 4.3 V. Organic materials such as p-DDPZ maintain a high energy density of >400 Wh kg− 1 even at 10 A g− 1. These all indicate possible applications of KIB in large-scale energy storage systems (ESS) and high-power uninterruptible power supply (UPS). Some conversion-type materials, such as sulfur, achieve the highest specific capacity among all cathodes. However, similar to Li–S cells, further studies on electrolyte development are needed to prevent the dissolution of polysulphides. The ionic diffusion of K ions in sulfur strongly depends on the carbon conductive additives, reducing its specific capacity advantages. Organic materials exhibit very different electrochemical perfor­ mance dependant on their functional groups. The most significant dif­ ference lies in the operational potentials. Organic materials with nitrogen radicals generally show high operational potentials of >3 V, while others with conjugated carbonyls exhibit meager operational potentials of ~1.7 V on average. PDPPZ is the most promising one regrading to its high energy density of 583 Wh kg− 1 (slightly higher than LFP, 578 Wh kg− 1) and excellent cycling performance of 81% capacity retention after 1000 cycles at 2 A g− 1. A detailed electrochemical comparison of the most promising cathode representatives is given in Table 2. Among them, K2Mn[Fe(CN)6] and PDPPZ exhibit bright application prospects. Most KIB cathodes are derivatives from LIB cathodes. However, the large ionic radius of K+ brings in differences in chemical compositions, phases, voltage profiles, and cycling lifespan. Overall, their electro­ chemical performances, especially energy density, are not competitive to LIB cathodes and NIB cathodes. The development of KIB cathodes is still at an early stage. Different from LIBs and NIBs, the conventional layer transition metal oxides do not yet present superiority in KIBs. It is always a challenge to obtain a high stoichiometry with a control phase for high specific capacity. Strategies to optimize the composition by tuning the synergist effects of doping different transition metals (such as Cu2+, Fe2+, Ti4+) are suggested. At present, most of the studies on KIBs cathodes are conducted to explore their specific capacity and cycling life. At the same time, it is vital to evaluate their compatibility with the electrolyte, the thermal stability, the volume expansion/contraction, the rate performance, and the high/low-temperature performances. Meanwhile, the electrolytes and binders for KIBs cathodes inheriting from the traditional LIBs have barely been studied yet; highly-concentrated electrolytes and multi-salts electrolytes should be further researched. All these parameters are essential to the full cell design in practical applications.

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39]

Declaration of competing interest

[40] [41]

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence 15

L. Wang, Z. Wu, J. Zou, P. Gao, X. Niu, H. Li, L. Chen, Joule 3 (2019) 2086–2102. H. Kim, H. Ji, J. Wang, G. Ceder, Trends in Chemistry, 2019. E.A. Olivetti, G. Ceder, G.G. Gaustad, X. Fu, Joule 1 (2017) 229–243. D. Chung, E. Elgqvist, S. Santhanagopalan, Clean Energy Manufacturing Analysis Center, CEMAC), 2015. B. Wang, E.H. Ang, Y. Yang, Y. Zhang, M. Ye, Q. Liu, C. Li, Chem. Eur J. (2020), https://doi.org/10.1002/chem.202001811. A. Yaroshevsky, Geochem. Int. 44 (2006) 48–55. A. Eftekhari, Z. Jian, X. Ji, ACS Appl. Mater. Interfaces 9 (2016) 4404–4419. J.E. Tilton, J.I. Guzm´ an, Mineral Economics and Policy, Routledge, 2016. J.C. Pramudita, D. Sehrawat, D. Goonetilleke, N. Sharma, Adv. Energy Mater. 7 (2017). B. Lin, X. Zhu, L. Fang, X. Liu, S. Li, T. Zhai, L. Xue, Q. Guo, J. Xu, H. Xia, Adv. Mater. 31 (2019), e1900060. P. Vishnuprakash, C. Nithya, M. Premalatha, Electrochim. Acta 309 (2019) 234–241. S. Chong, Y. Chen, Y. Zheng, Q. Tan, C. Shu, Y. Liu, Z. Guo, J. Mater. Chem. 5 (2017) 22465–22471. Y. Sun, C. Liu, J. Xie, D. Zhuang, W. Zheng, X. Zhao, New J. Chem. 43 (2019) 11618–11625, https://doi.org/10.1039/C9NJ02085C. F.A. Obrezkov, V. Ramezankhani, I. Zhidkov, V.F. Traven, E.Z. Kurmaev, K. J. Stevenson, P.A. Troshin, J. Phys. Chem. Lett. 10 (2019) 5440–5445. Y. Marcus, Pure Appl. Chem. 57 (1985) 1129–1132. S. Komaba, T. Hasegawa, M. Dahbi, K. Kubota, Electrochem. Commun. 60 (2015) 172–175. Y. Matsuda, H. Nakashima, M. Morita, Y. Takasu, J. Electrochem. Soc. 128 (1981) 2552. T.A. Pham, K.E. Kweon, A. Samanta, V. Lordi, J.E. Pask, J. Phys. Chem. C 121 (2017) 21913–21920. M. Okoshi, Y. Yamada, S. Komaba, A. Yamada, H. Nakai, J. Electrochem. Soc. 164 (2016) A54–A60. T. Thomberg, R. V¨ ali, J. Eskusson, T. Romann, A. J¨ anes, J. Electrochem. Soc. 165 (2018) A3862–A3870. S. Amara, J. Toulc’Hoat, L. Timperman, A. Biller, H. Galiano, C. Marcel, M. Ledigabel, M. Anouti, ChemPhysChem 20 (2019) 581–594. J. Qian, W.A. Henderson, W. Xu, P. Bhattacharya, M. Engelhard, O. Borodin, J.G. Zhang, Nat. Commun. 6 (2015) 1–9. J. Zheng, S. Chen, W. Zhao, J. Song, M.H. Engelhard, J.-G. Zhang, ACS Energy Lett. 3 (2018) 315–321. T. Hosaka, K. Kubota, H. Kojima, S. Komaba, Chem. Commun. 54 (2018) 8387–8390. M. Levi, E. Markevich, D. Aurbach, J. Phys. Chem. B 109 (2005) 7420–7427. L. Wang, J. Yang, J. Li, T. Chen, S. Chen, Z. Wu, J. Qiu, B. Wang, P. Gao, X. Niu, J. Power Sources 409 (2019) 24–30. H. Guo, X. Li, X. Zhang, H. Wang, Z. Wang, W. Peng, N. Carbon Mater. 22 (2007) 7–10. Z. Jian, Z. Xing, C. Bommier, Z. Li, X. Ji, Adv. Energy Mater. 6 (2016) 1501874. N. Recham, G. Rousse, M.T. Sougrati, J.-N. Chotard, C. Frayret, S. Mariyappan, B. C. Melot, J.-C. Jumas, J.-M. Tarascon, Chem. Mater. 24 (2012) 4363–4370. C. Jiang, B. Wang, Z. Wu, J. Qiu, Z. Ding, J. Zou, S. Chen, P. Gao, X. Niu, L. Wang, Nano Energy 70 (2020) 104552. Z. Jian, W. Luo, X. Ji, J. Am. Chem. Soc. 137 (2015) 11566–11569. A. Eftekhari, Z. Jian, X. Ji, ACS Appl. Mater. Interfaces 9 (2017) 4404–4419. M. Okoshi, Y. Yamada, S. Komaba, A. Yamada, H. Nakai, J. Electrochem. Soc. 164 (2017) A54–A60. D. Su, A. McDonagh, S.Z. Qiao, G. Wang, Adv. Mater. 29 (2017) 1604007. C.D. Wessells, S.V. Peddada, M.T. McDowell, R.A. Huggins, Y. Cui, J. Electrochem. Soc. 159 (2011) A98–A103. Z. Jian, Z. Xing, C. Bommier, Z. Li, X. Ji, Adv. Energy Mater. 6 (2016). H. Liang, B. Hou, W. Li, Q. Ning, X. Yang, Z. Gu, X. Nie, G. Wang, X. Wu, Energy Environ. Sci. 12 (2019) 3575–3584. Y. uLi, Y. Lu, P. Adelhelm, M.-M. Titirici, Y.-S. Hu, Chem. Soc. Rev. 48 (2019) 4655–4687. Z. Wu, L. Wang, J. Huang, J. Zou, S. Chen, H. Cheng, C. Jiang, P. Gao, X. Niu, Electrochim. Acta 306 (2019) 446–453. L. Tao, L. Liu, R. Chang, H. He, P. Zhao, J. Liu, J. Power Sources 463 (2020) 228172. Y. Xu, J. Zhang, D. Li, Chem. Asian J. 15 (11) (2020 June 2) 1648–1659.

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307

[42] B. Kishore, V. Gopal, N. Munichandraiah, J. Electrochem. Soc. 163 (2016) A2551–A2554. [43] J. Han, M. Xu, Y. Niu, G.-N. Li, M. Wang, Y. Zhang, M. Jia, C. Li, Chem. Commun. 52 (2016) 11274–11276. [44] J. Han, Y. Niu, S. Bao, Y. Yu, S. Lu, M. Xu, Chem. Commun. 52 (2016) 11661–11664. [45] P. Lian, Y. Dong, Z.-S. Wu, S. Zheng, X. Wang, W. Sen, C. Sun, J. Qin, X. Shi, X. Bao, Nano Energy 40 (2017) 1–8. [46] I. Sultana, M.M. Rahman, S. Mateti, V.G. Ahmadabadi, A.M. Glushenkov, Y. Chen, Nanoscale 9 (2017) 3646–3654. [47] K. Lei, F. Li, C. Mu, J. Wang, Q. Zhao, C. Chen, J. Chen, Energy Environ. Sci. 10 (2017) 552–557. [48] Q. Deng, J. Pei, C. Fan, J. Ma, B. Cao, C. Li, Y. Jin, L. Wang, J. Li, Nano Energy 33 (2017) 350–355. [49] H. Kim, J.C. Kim, M. Bianchini, D.H. Seo, J. Rodriguez-Garcia, G. Ceder, Adv. Energy Mater. 8 (2018) 1702384. [50] A. Eftekhari, J. Power Sources 126 (2004) 221–228. [51] V. Mathew, S. Kim, J. Kang, J. Gim, J. Song, J.P. Baboo, W. Park, D. Ahn, J. Han, L. Gu, Y. Wang, Y.-S. Hu, Y.-K. Sun, J. Kim, NPG Asia Mater. 6 (2014) e138-e138. [52] L. Zhang, B. Zhang, C. Wang, Y. Dou, Q. Zhang, Y. Liu, H. Gao, M. Al-Mamun, W. K. Pang, Z. Guo, S.X. Dou, H.K. Liu, Nano Energy 60 (2019) 432–439. [53] J. Han, G.-N. Li, F. Liu, M. Wang, Y. Zhang, L. Hu, C. Dai, M. Xu, Chem. Commun. 53 (2017) 1805–1808. [54] J. Liao, Q. Hu, X. He, J. Mu, J. Wang, C. Chen, J. Power Sources (2020) 451. [55] H. Kim, D.-H. Seo, M. Bianchini, R.J. Cl´ement, H. Kim, J.C. Kim, Y. Tian, T. Shi, W.-S. Yoon, G. Ceder, Adv. Energy Mater. 8 (2018). [56] V.A. Nikitina, S.M. Kuzovchikov, S.S. Fedotov, N.R. Khasanova, A.M. Abakumov, E.V. Antipov, Electrochim. Acta 258 (2017) 814–824. [57] K. Chihara, A. Katogi, K. Kubota, S. Komaba, Chem. Commun. 53 (2017) 5208–5211. [58] Z. Liu, J. Wang, B. Lu, Sci. Bull. 65 (15) (2020 August 15) 1242–1251. [59] J. Hyoung, J.W. Heo, M.S. Chae, S.-T. Hong, ChemSusChem (2018). [60] W.B. Park, S.C. Han, C. Park, S.U. Hong, U. Han, S.P. Singh, Y.H. Jung, D. Ahn, K.S. Sohn, M. Pyo, Adv. Energy Mater. 8 (2018). [61] T. Hosaka, T. Shimamura, K. Kubota, S. Komaba, The Chemical Record, 2018. [62] X. Lin, J. Huang, H. Tan, J. Huang, B. Zhang, Energy Storage Mater. 16 (2019) 97–101. [63] W. Ko, H. Park, J.H. Jo, Y. Lee, J. Kang, Y.H. Jung, T.-Y. Jeon, S.-T. Myung, J. Kim, Nano Energy 66 (2019 December), 104184. [64] S.S. Fedotov, N.D. Luchinin, D.A. Aksyonov, A.V. Morozov, S.V. Ryazantsev, M. Gaboardi, J.R. Plaisier, K.J. Stevenson, A.M. Abakumov, E.V. Antipov, Nat. Commun. 11 (2020) 1484. [65] G.X. Do, B.J. Paul, V. Mathew, J. Kim, J. Mater. Chem. 1 (2013). [66] C. Vaalma, G.A. Giffin, D. Buchholz, S. Passerini, J. Electrochem. Soc. 163 (2016) A1295–A1299. [67] T. Deng, X. Fan, J. Chen, L. Chen, C. Luo, X. Zhou, J. Yang, S. Zheng, C. Wang, Adv. Funct. Mater. (2018) 28. [68] X. Wang, X. Xu, C. Niu, J. Meng, M. Huang, X. Liu, Z. Liu, L. Mai, Nano Lett. 17 (2016) 544–550. [69] T. Deng, X. Fan, C. Luo, J. Chen, L. Chen, S. Hou, N. Eidson, X. Zhou, C. Wang, Nano Lett. 18 (2018) 1522–1529. [70] H. Kim, J.C. Kim, S.-H. Bo, T. Shi, D.-H. Kwon, G. Ceder, Adv. Energy Mater. 7 (2017). [71] C. Liu, S. Luo, H. Huang, Z. Wang, A. Hao, Y. Zhai, Z. Wang, Electrochem. Commun. 82 (2017) 150–154. [72] L. Deng, X. Niu, G. Ma, Z. Yang, L. Zeng, Y. Zhu, L. Guo, Adv. Funct. Mater. (2018) 28. [73] M. Clites, J.L. Hart, M.L. Taheri, E. Pomerantseva, ACS Energy Lett. 3 (2018) 562–567. [74] H. Kim, D.-H. Seo, A. Urban, J. Lee, D.-H. Kwon, S.-H. Bo, T. Shi, J.K. Papp, B. D. McCloskey, G. Ceder, Chem. Mater. 30 (2018) 6532–6539. [75] S.-Z. Yang, K.R. Tallman, P. Liu, D.M. Lutz, B. Zhang, S.J. Kim, L. Wu, A. C. Marschilok, E.S. Takeuchi, K.J. Takeuchi, Y. Zhu, Nano Energy 71 (2020). [76] Z. Tai, M. Shi, S. Chong, Y. Chen, Q. Tan, C. Shu, Y. Liu, Sustain. Energy Fuels 3 (2019) 736–743. [77] S. Chong, Y. Wu, C. Liu, Y. Chen, S. Guo, Y. Liu, G. Cao, Nano Energy 54 (2018) 106–115. [78] H.V. Ramasamy, B. Senthilkumar, P. Barpanda, Y.-S. Lee, Chem. Eng. J. 368 (2019) 235–243. [79] J.U. Choi, J. Kim, J.-Y. Hwang, J.H. Jo, Y.-K. Sun, S.-T. Myung, Nano Energy 61 (2019) 284–294. [80] J.H. Jo, J.U. Choi, Y.J. Park, Y.H. Jung, D. Ahn, T.Y. Jeon, H. Kim, J. Kim, S. T. Myung, Adv. Energy Mater. 10 (2020). [81] X. Zhang, Y. Yang, X. Qu, Z. Wei, G. Sun, K. Zheng, H. Yu, F. Du, Advanced Functional Materials, 2019. [82] M.K. Cho, J.H. Jo, J. Choi, S.-T. Myung, ACS Appl. Mater. Interfaces 11 (31) (2019) 27770–27779. [83] S.C. Han, W.B. Park, K.-S. Sohn, M. Pyo, J. Solid State Electrochem. 23 (11) (2019) 3135–3143. [84] J. Weng, J. Duan, C. Sun, P. Liu, A. Li, P. Zhou, J. Zhou, Chem. Eng. J. 392 (2020 July 15), 123649. [85] H. Fei, Y. Liu, Y. An, X. Xu, G. Zeng, Y. Tian, L. Ci, B. Xi, S. Xiong, J. Feng, J. Power Sources 399 (2018) 294–298. [86] L. Fan, R. Ma, J.W. Wang, H. Yang, B. Lu, Adv. Mater. 30 (2018), e1805486. [87] Z. Xing, Z. Jian, W. Luo, Y. Qi, C. Bommier, E.S. Chong, Z. Li, L. Hu, X. Ji, Energy Storage Mater. 2 (2016) 63–68.

[88] Y. Chen, W. Luo, M. Carter, L. Zhou, J. Dai, K. Fu, S. Lacey, T. Li, J. Wan, X. Han, Y. Bao, L. Hu, Nano Energy 18 (2015) 205–211. [89] Z. Tong, S. Tian, H. Wang, D. Shen, R. Yang, C.S. Lee, Advanced Functional Materials, 2019. [90] Z. Jian, Y. Liang, I.A. Rodríguez-P´erez, Y. Yao, X. Ji, Electrochem. Commun. 71 (2016) 5–8. [91] L. Fan, Q. Liu, Z. Xu, B. Lu, ACS Energy Lett. 2 (2017) 1614–1620. [92] H. Gao, L. Xue, S. Xin, J.B. Goodenough, Angew. Chem. Int. Ed. 57 (2018) 5449–5453. [93] J. Ma, E. Zhou, C. Fan, B. Wu, C. Li, Z.-H. Lu, J. Li, Chem. Commun. 54 (2018) 5578–5581. [94] J. Zhao, J. Yang, P. Sun, Y. Xu, Electrochem. Commun. 86 (2018) 34–37. [95] M. Tang, Y. Wu, Y. Chen, C. Jiang, S. Zhu, S. Zhuo, C. Wang, J. Mater. Chem. 7 (2019) 486–492. [96] C. Li, J. Xue, A. Huang, J. Ma, F. Qing, A. Zhou, Z. Wang, Y. Wang, J. Li, Electrochim. Acta 297 (2019) 850–855. [97] F.A. Obrezkov, A.F. Shestakov, V.F. Traven, K.J. Stevenson, P.A. Troshin, J. Mater. Chem. 7 (2019) 11430–11437. [98] A. Slesarenko, I.K. Yakuschenko, V. Ramezankhani, V. Sivasankaran, O. Romanyuk, A.V. Mumyatov, I. Zhidkov, S. Tsarev, E.Z. Kurmaev, A. F. Shestakov, O.V. Yarmolenko, K.J. Stevenson, P.A. Troshin, J. Power Sources 435 (2019). [99] S. Lv, J. Yuan, Z. Chen, P. Gao, H. Shu, X. Yang, E. Liu, S. Tan, M. Ruben, Z. ZhaoKarger, M. Fichtner, ChemSusChem (2020). [100] Y. Tang, W. Li, P. Feng, M. Zhou, K. Wang, K. Jiang, Chem. Eng. J. 396 (2020 September 15), 125269. [101] M. Ye, J. Hwang, Y. Sun, ACS Nano 13 (2019) 9306–9314. [102] L. Xue, Y. Li, H. Gao, W. Zhou, X. Lü, W. Kaveevivitchai, A. Manthiram, J. B. Goodenough, J. Am. Chem. Soc. 139 (2017) 2164–2167. [103] X. Jiang, T. Zhang, L. Yang, G. Li, J.Y. Lee, Chemelectrochem 4 (2017) 2237–2242. [104] X. Bie, K. Kubota, T. Hosaka, K. Chihara, S. Komaba, J. Mater. Chem. 5 (2017) 4325–4330. [105] Z. Shadike, D.-R. Shi, T.-W. Tian-Wang, M.-H. Cao, S.-F. Yang, J. Chen, Z.-W. Fu, J. Mater. Chem. 5 (2017) 6393–6398. [106] P. Padigi, J. Thiebes, M. Swan, G. Goncher, D. Evans, R. Solanki, Electrochim. Acta 166 (2015) 32–39. [107] Q. Xue, L. Li, Y. Huang, R. Huang, F. Wu, R. Chen, ACS Appl. Mater. Interfaces 11 (2019) 22339–22345. [108] X. Wu, Z. Jian, Z. Li, X. Ji, Electrochem. Commun. 77 (2017) 54–57. [109] J. Liao, Q. Hu, Y. Yu, H. Wang, Z. Tang, Z. Wen, C. Chen, J. Mater. Chem. 5 (2017) 19017–19024. [110] G. He, L.F. Nazar, ACS Energy Lett. 2 (2017) 1122–1127. [111] Y. Pei, C. Mu, H. Li, F. Li, J. Chen, ChemSusChem 11 (2018) 1285–1289. [112] P. Xiong, X. Han, X. Zhao, P. Bai, Y. Liu, J. Sun, Y. Xu, ACS Nano 13 (2019) 2536–2543. [113] Y. Liu, W. Wang, J. Wang, Y. Zhang, Y. Zhu, Y. Chen, L. Fu, Y. Wu, Chem. Commun. 54 (2018) 2288–2291. [114] X. Lu, M.E. Bowden, V.L. Sprenkle, J. Liu, Adv. Mater. 27 (2015) 5915–5922. [115] J. Ding, H. Zhang, W. Fan, C. Zhong, W. Hu, D. Mitlin, Advanced Materials, 2020, e1908007. [116] X. Huang, Q. Xu, W. Gao, T. Yang, R. Zhan, J. Deng, B. Guo, M. Tao, H. Liu, M. Xu, J. Colloid Interface Sci. 539 (2019) 326–331. [117] Y. Liu, Z. Tai, Q. Zhang, H. Wang, W.K. Pang, H.K. Liu, K. Konstantinov, Z. Guo, Nano Energy 35 (2017) 36–43. [118] W. Zhang, H. Wang, N. Zhang, H. Liu, Z. Chen, L. Zhang, S. Guo, D. Li, J. Xu, ACS Appl. Mater. Interfaces 11 (2019) 29807–29813. [119] Y. Liu, D. Yang, W. Wang, K. Hu, Q. Huang, Y. Zhang, Y. Miao, L. Fu, M. Wu, Y. Wu, J. Mater. Chem. 8 (2020) 4544–4551. [120] Q. Liu, W. Deng, C.-F. Sun, Energy Storage Mater. 28 (2020) 10–16. [121] W.D. McCulloch, X. Ren, M. Yu, Z. Huang, Y. Wu, ACS Appl. Mater. Interfaces 7 (2015) 26158–26166. [122] X. Ren, K.C. Lau, M. Yu, X. Bi, E. Kreidler, L.A. Curtiss, Y. Wu, ACS Appl. Mater. Interfaces 6 (2014) 19299–19307. [123] W. Zhang, C. Hu, Z. Guo, L. Dai, Angew. Chem. Int. Ed. 59 (2020) 3470–3474. [124] M. Qian, M. Tang, J. Yang, W. Wei, M. Chen, J. Chen, J. Xu, Q. Liu, H. Wang, J. Colloid Interface Sci. 551 (2019) 177–183. [125] S. Ma, P. Zuo, H. Zhang, Z. Yu, C. Cui, M. He, G. Yin, Chem. Commun. 55 (2019) 5267–5270. [126] K. Lu, H. Zhang, F. Ye, W. Luo, H. Ma, Y. Huang, Energy Storage Mater. 16 (2019) 1–5. [127] B. Ji, F. Zhang, X. Song, Y. Tang, Adv. Mater. (2017) 29. [128] L. Fan, Q. Liu, S. Chen, K. Lin, Z. Xu, B. Lu, Small 13 (2017). [129] S. Wang, Y. Chen, B. Cui, B. Li, S. Wang, Y. Cui, Z. Ju, Q. Zhuang, Appl. Surf. Sci. 514 (2020). [130] S. Wang, Y. Cui, B. Cui, Q. Zhuang, B. Li, H. Zheng, Z. Ju, Advanced Materials Interfaces, 2019. [131] S. Wang, B. Cui, Q. Zhuang, Y. Shi, H. Zheng, Nanoscale Res. Lett. 14 (2019) 238. [132] J. Zhang, T. Liu, X. Cheng, M. Xia, R. Zheng, N. Peng, H. Yu, M. Shui, J. Shu, Nano Energy 60 (2019) 340–361. [133] B. Huang, Y. Shao, Y. Liu, Z. Lu, X. Lu, S. Liao, ACS Appl. Energy Mater. 2 (2019) 6528–6535. [134] L. Qin, N. Xiao, J. Zheng, Y. Lei, D. Zhai, Y. Wu, Advanced Energy Materials, 2019. [135] W. Zhang, Y. Liu, Z. Guo, Sci. Adv. 5 (2019) 7412.

16

Z. Wu et al.

Journal of Power Sources 484 (2021) 229307 [179] J.-Y. Hwang, J. Kim, T.-Y. Yu, S.-T. Myung, Y.-K. Sun, Energy Environ. Sci. 11 (2018) 2821–2827. [180] N.A. Chernova, M. Roppolo, A.C. Dillon, M.S. Whittingham, J. Mater. Chem. 19 (2009) 2526–2552. [181] H. Liu, W. Yang, Energy Environ. Sci. 4 (2011) 4000–4008. [182] K. West, B. Zachau-Christiansen, T. Jacobsen, S. Skaarup, Solid State Ionics 40 (1990) 585–588. [183] D. Koch, V.V. Kulish, S. Manzhos, MRS Commun. 7 (2017) 819–825. [184] J. Liu, H. Xia, D. Xue, L. Lu, J. Am. Chem. Soc. 131 (2009) 12086–12087. [185] V. Raju, J. Rains, C. Gates, W. Luo, X. Wang, W.F. Stickle, G.D. Stucky, X. Ji, Nano Lett. 14 (2014) 4119–4124. [186] B. Tian, W. Tang, C. Su, Y. Li, ACS Appl. Mater. Interfaces 10 (2017) 642–650. [187] C.D. Wessells, S.V. Peddada, R.A. Huggins, Y. Cui, Nano Lett. 11 (2011) 5421–5425. [188] K. Itaya, T. Ataka, S. Toshima, J. Am. Chem. Soc. 104 (1982) 4767–4772. [189] M. Pasta, C.D. Wessells, R.A. Huggins, Y. Cui, Nat. Commun. 3 (2012) 1149. [190] M. Jayalakshmi, F. Scholz, J. Power Sources 91 (2000) 217–223. [191] Y. Luo, B. Shen, B. Guo, L. Hu, Q. Xu, R. Zhan, Y. Zhang, S. Bao, M. Xu, J. Phys. Chem. Solid. 122 (2018) 31–35. [192] Y.H. Zhu, X. Yang, D. Bao, X.F. Bie, T. Sun, S. Wang, Y.S. Jiang, X.B. Zhang, J. M. Yan, Q. Jiang, Joule 2 (2018) 736–746. [193] A. Baioun, H. Kellawi, A. Falah, Curr. Nanosci. 14 (2018) 227–233. [194] C. Zhang, Y. Xu, M. Zhou, L. Liang, H. Dong, M. Wu, Y. Yang, Y. Lei, Adv. Funct. Mater. 27 (2017) 1604307. [195] S. Chong, Y. Wu, S. Guo, Y. Liu, G. Cao, Energy Storage Materials, 2019. [196] C.D. Wessells, R.A. Huggins, Y. Cui, Nat. Commun. 2 (2011) 550. [197] J. Heo, M. Chae, J. Hyoung, S.-T. Hong, Inorganic Chemistry, 2019. [198] P. Padigi, G. Goncher, D. Evans, R. Solanki, J. Power Sources 273 (2015) 460–464. [199] J. Song, L. Wang, Y. Lu, J. Liu, B. Guo, P. Xiao, J.-J. Lee, X.-Q. Yang, G. Henkelman, J.B. Goodenough, J. Am. Chem. Soc. 137 (2015) 2658–2664. [200] D. Asakura, M. Okubo, Y. Mizuno, T. Kudo, H. Zhou, K. Ikedo, T. Mizokawa, A. Okazawa, N. Kojima, J. Phys. Chem. C 116 (2012) 8364–8369. [201] T. Matsuda, M. Takachi, Y. Moritomo, Chem. Commun. 49 (2013) 2750–2752. [202] B. Wang, Y. Han, X. Wang, N. Bahlawane, H. Pan, M. Yan, Y. Jiang, iScience 3 (2018) 110–133. [203] L. Shen, Z. Wang, L. Chen, Chem. A Eur. J. 20 (2014) 12559–12562. [204] Y. You, H. Yao, S. Xin, Y. Yin, T. Zuo, C. Yang, Y. Guo, Y. Cui, L. Wan, J. B. Goodenough, Adv. Mater. 28 (2016) 7243–7248. [205] C. Ling, J. Chen, F. Mizuno, J. Phys. Chem. C 117 (2013) 21158–21165. [206] N. Nitta, F. Wu, J.T. Lee, G. Yushin, Mater. Today 18 (2015) 252–264. [207] B. Ji, W. Yao, Y. Zheng, P. Kidkhunthod, X. Zhou, S. Tunmee, S. Sattayaporn, H. M. Cheng, H. He, Y. Tang, Nat. Commun. 11 (2020) 1225. [208] H.R. Tietze, Aust. J. Chem. 34 (1981) 2035–2038. [209] R. Lian, D. Wang, X. Ming, R. Zhang, Y. Wei, J. Feng, X. Meng, G. Chen, J. Mater. Chem. 6 (2018) 16228–16234. [210] S. Zheng, S. Cheng, S. Xiao, L. Hu, Z. Chen, B. Huang, Q. Liu, J. Yang, Q. Chen, J. Alloys Compd. 815 (2020 January 30), 152379. [211] B. Senthilkumar, R. Kalai Selvan, P. Barpanda, Electrochemistry Communications, 2019. [212] H. Park, H. Kim, W. Ko, J.H. Jo, Y. Lee, J. Kang, I. Park, S.-T. Myung, J. Kim, Energy Storage Mater. 28 (2020) 47–54. [213] J. Liao, Q. Hu, B. Che, X. Ding, F. Chen, C. Chen, J. Mater. Chem. 7 (2019) 15244–15251. [214] T. Mueller, G. Hautier, A. Jain, G. Ceder, Chem. Mater. 23 (2011) 3854–3862. [215] A. Jain, G. Hautier, C. Moore, B. Kang, J. Lee, H. Chen, N. Twu, G. Ceder, J. Electrochem. Soc. 159 (2012) A622–A633. [216] S.Y. Lim, H. Kim, R.A. Shakoor, Y. Jung, J.W. Choi, J. Electrochem. Soc. 159 (2012) A1393–A1397. [217] H. Insang, Creative Commons, 2019. [218] X. Li, J. Liu, B. Wang, M.N. Banis, B. Xiao, R. Li, T.-K. Sham, X. Sun, RSC Adv. 4 (2014) 27126–27129. [219] S.K. Park, J.S. Park, Y.C. Kang, J. Mater. Chem. 6 (2018) 1028–1036. [220] J. He, W. Lv, Y. Chen, K. Wen, C. Xu, W. Zhang, Y. Li, W. Qin, W. He, ACS Nano 11 (2017) 8144–8152. [221] X. Yu, A. Manthiram, Energy Storage Mater. 15 (2018) 368–373. [222] S. Dong, D. Yu, J. Yang, L. Jiang, J. Wang, L. Cheng, Y. Zhou, H. Yue, H. Wang, L. Guo, Advanced Materials, 2020, e1908027. [223] L. Wang, J. Bao, Q. Liu, C.-F. Sun, Energy Storage Mater. 18 (2019) 470–475. [224] X. Huang, J. Deng, Y. Qi, D. Liu, Y. Wu, W. Gao, W. Zhong, F. Zhang, S. Bao, M. Xu, Inorg. Chem. Front. 7 (2020) 1182–1189. [225] Y.-H. Zhu, Y.-B. Yin, X. Yang, T. Sun, S. Wang, Y.-S. Jiang, J.-M. Yan, X.-B. Zhang, Angew. Chem. Int. Ed. 56 (2017) 7881–7885.

[136] N. Naveen, S.C. Han, S.P. Singh, D. Ahn, K.-S. Sohn, M. Pyo, J. Power Sources 430 (2019) 137–144. [137] N. Naveen, W.B. Park, S.P. Singh, S.C. Han, D. Ahn, K.-S. Sohn, M. Pyo, Small 14 (2018), e1803495. [138] Q. Zhao, Y. Hu, K. Zhang, J. Chen, Inorg. Chem. 53 (2014) 9000–9005. [139] Y.-H. Zhu, Q. Zhang, X. Yang, E.-Y. Zhao, T. Sun, X.-B. Zhang, S. Wang, X.-Q. Yu, J.-M. Yan, Q. Jiang, Inside Chem. 5 (2019) 168–179. [140] R. Ma, L. Fan, J. Wang, B. Lu, Electrochim. Acta 293 (2019) 191–198. [141] M. Danot, A. LeBlanc, J. Rouxel, Bull. Soc. Chim. Fr. (1969) 2670. [142] Q. Qu, L. Li, S. Tian, W. Guo, Y. Wu, R. Holze, J. Power Sources 195 (2010) 2789–2794. [143] H.Y. Lee, V. Manivannan, J.B. Goodenough, Comptes Rendus Acad. Sci. - Ser. IIC Chem. 2 (1999) 565–577. [144] K. Sada, B. Senthilkumar, P. Barpanda, ACS Appl. Energy Mater. 1 (2018) 5410–5416. [145] T. Masese, K. Yoshii, Y. Yamaguchi, T. Okumura, Z.-D. Huang, M. Kato, K. Kubota, J. Furutani, Y. Orikasa, H. Senoh, Nat. Commun. 9 (2018) 3823. [146] A. Gao, M. Li, N. Guo, D. Qiu, Y. Li, S. Wang, X. Lu, F. Wang, R. Yang, Adv. Energy Mater. 9 (2019). [147] N. Yabuuchi, R. Hara, M. Kajiyama, K. Kubota, T. Ishigaki, A. Hoshikawa, S. Komaba, Adv. Energy Mater. 4 (2014) 1301453. [148] E. de la Llave, E. Talaie, E. Levi, P.K. Nayak, M. Dixit, P.T. Rao, P. Hartmann, F. Chesneau, D.T. Major, M. Greenstein, Chem. Mater. 28 (2016) 9064–9076. [149] J.H. Jo, J.U. Choi, A. Konarov, H. Yashiro, S. Yuan, L. Shi, Y.K. Sun, S.T. Myung, Adv. Funct. Mater. 28 (2018) 1705968. [150] J.U. Choi, C.S. Yoon, Q. Zhang, P. Kaghazchi, Y.H. Jung, K.-S. Lee, D.-C. Ahn, Y.K. Sun, S.-T. Myung, J. Mater. Chem. 7 (2019) 202–211. [151] A. Konarov, J.H. Jo, J.U. Choi, Z. Bakenov, H. Yashiro, J. Kim, S.-T. Myung, Nano Energy 59 (2019) 197–206. [152] W. Kang, Z. Zhang, P.-K. Lee, T.-W. Ng, W. Li, Y. Tang, W. Zhang, C.-S. Lee, D.Y. W. Yu, J. Mater. Chem. 3 (2015) 22846–22852. [153] W.-L. Pang, X.-H. Zhang, J.-Z. Guo, J.-Y. Li, X. Yan, B.-H. Hou, H.-Y. Guan, X.L. Wu, J. Power Sources 356 (2017) 80–88. [154] J.U. Choi, Y.J. Park, J.H. Jo, L.-Y. Kuo, P. Kaghazchi, S.-T. Myung, ACS Appl. Mater. Interfaces 10 (2018) 40978–40984. [155] Y.J. Park, J.U. Choi, J.H. Jo, C.H. Jo, J. Kim, S.T. Myung, Advanced Functional Materials, 2019, p. 1901912. [156] A. Konarov, J.U. Choi, Z. Bakenov, S.-T. Myung, J. Mater. Chem. 6 (2018) 8558–8567. [157] X. Zhang, Z. Wei, K.N. Dinh, N. Chen, G. Chen, F. Du, Q. Yan, Small (2020), e2002700. [158] Y.-S. Xu, Q.-H. Zhang, D. Wang, J.-C. Gao, X.-S. Tao, Y. Liu, Y.-G. Sun, L. Gu, B.B. Chang, C.-T. Liu, S.-Q. Shi, A.-M. Cao, Energy Storage Mater. 31 (2020) 20–26. [159] Y. Hironaka, K. Kubota, S. Komaba, Chem. Commun. 53 (2017) 3693–3696. [160] R.J. Cl´ement, P.G. Bruce, C.P. Grey, J. Electrochem. Soc. 162 (2015) A2589–A2604. [161] S. Guo, P. Liu, H. Yu, Y. Zhu, M. Chen, M. Ishida, H. Zhou, Angew. Chem. Int. Ed. 54 (2015) 5894–5899. [162] R. Berthelot, D. Carlier, C. Delmas, Nat. Mater. 10 (2011) 74–80. [163] J. Ding, Y. Zhou, Q. Sun, X. Yu, X. Yang, Z. Fu, Electrochim. Acta 87 (2013) 388–393. [164] A. Kraytsberg, Y. Ein-Eli, Adv. Energy Mater. 2 (2012) 922–939. [165] C. Delmas, C. Fouassier, P. Hagenmuller, J. Ackerman, Inorganic Syntheses, 1984, pp. 56–61. [166] H. Kim, D.H. Seo, J.C. Kim, S.H. Bo, L. Liu, T. Shi, G. Ceder, Adv. Mater. (2017) 29. [167] P.K. Nayak, L. Yang, W. Brehm, P. Adelhelm, Angew. Chem. Int. Ed. 57 (2018) 102–120. [168] Y. Lei, X. Li, L. Liu, G. Ceder, Chem. Mater. 26 (2014) 5288–5296. [169] K. Lei, Z. Zhu, Z. Yin, P. Yan, F. Li, Chem 5 (12) (2019 December 12) 3220–3231. [170] J.U. Choi, J. Kim, J.H. Jo, H.J. Kim, Y.H. Jung, D.-C. Ahn, Y.-K. Sun, S.-T. Myung, Energy Storage Materials, 2019. [171] S. Zhao, K. Yan, P. Munroe, B. Sun, G. Wang, Adv. Energy Mater. 9 (2019). [172] L. Deng, T. Wang, Y. Hong, M. Feng, R. Wang, J. Zhang, Q. Zhang, J. Wang, L. Zeng, Y. Zhu, L. Guo, ACS Energy Lett. 5 (2020) 1916–1922. [173] K. Kubota, S. Kumakura, Y. Yoda, K. Kuroki, S. Komaba, Adv. Energy Mater. 8 (2018) 1703415. [174] K. Zhang, D. Kim, Z. Hu, M. Park, G. Noh, Y. Yang, J. Zhang, V.W.-h. Lau, S.L. Chou, M. Cho, Nat. Commun. 10 (2019) 1–12. [175] V.I. Voronin, G.S. Shekhtman, V.A. Blatov, Acta Crystallogr. Sect. B Struct. Sci. 68 (2012) 356–363. [176] M.V. Peskov, U. Schwingenschl¨ ogl, J. Phys. Chem. C 119 (2015) 9092–9098. [177] Z. Tomkowicz, A. Szytuła, J. Phys. Chem. Solid. 38 (1977) 1117–1123. [178] N.V. Proskurnina, V.I. Voronin, G.S. Shekhtman, L.N. Maskaeva, N.A. Kabanova, A.A. Kabanov, V.A. Blatov, J. Phys. Chem. C 121 (2017) 21128–21135.

17