Challenges for Aqueous Organic Redox Flow Batteries: The Degradation of Electrolytes and the Design of Ion‐Selective Membranes

Abstract

Aqueous organic redox flow batteries (AORFBs) are promising candidate for the next‐generation large‐scale energy storage. To develop a stable and cost‐effective AORFB system, the organic active materials and ion‐selective membranes (ISMs) are the most significant components to take into account. However, the degradation and crossover of organic materials downgrade the stability of AORFBs and represent major challenges for their practical application. Therefore, mitigating the degradation of organic species and designing high‐performance ISMs are of highly desired. In this review, the degradation mechanisms of several representative organic species are first discussed in detail. And then some effective strategies for degradation mitigation are summarized. In addition, the updated advancements in ISMs research are analyzed, with particular emphasis on strategies for mitigating the crossover issue via the design and regulation of ion transport channels. Finally, current research hotspots and future research priorities are discussed, offering insights into the challenges and potential approaches for the optimizations of these two key components.

This review discusses the mechanisms and mitigation strategies for the degradation of organic active materials in aqueous organic redox flow batteries (AORFBs). In addition, the current advancements in membranes research are discussed, demonstrating their pivotal role in addressing the crossover issue. Finally, recent hotspots and future research priorities are provided, offering insights into the major challenges and potential approaches for advancing these critical components.

1. Introduction

The global energy transition is facing challenges: while worldwide energy demand continues to grow, fossil fuels, which have long served as the primary energy source, are being phased out due to stringent carbon neutrality targets.^{[ 1} , ² , ^{3 ]} Renewable and low‐carbon energy sources such as solar and wind power are emerging as promising alternatives. However, these energy forms exhibit inherent intermittency.^{[ 4} , ⁵ , ^{6 ]} Direct integration of such fluctuating power into the grid would reduce equipment utilization, introduce emissions associated with power redundancy, and increase the costs of safeguarding grid stability.^{[ 1} , ^{7 ]} To address this dilemma, there is an urgent need to develop a new energy storage technology system that offers both millisecond‐level responsiveness and long‐duration storage capacity. Geography‐independent secondary battery technology holds promise for enabling large‐scale storage of intermittent renewable energy. Among these, lithium‐ion batteries (LIBs), known for their high energy density and long cycling lifetime, have been successfully commercialized in small‐scale electronics and electric vehicles.^{[ 8} , ⁹ , ^{10 ]} However, their widespread application in large‐scale electrical energy storage is hindered by limited decoupling capability of energy and power, safety concerns associated with organic solvents, and scarcity of raw materials.^{[ 11} , ¹² , ^{13 ]}

Redox flow batteries (RFBs) store chemical energy utilizing redox‐active electrolytes.^{[ 14 ]} Under non‐operating conditions, the electrolytes are stored in external tanks. During charging and discharging, the electrolyte solutions are pumped into the cell where electrochemical reactions facilitate the conversion between electrical and chemical energy (Figure 1a). As a result, RFBs enable decoupling of energy and power and offer considerable scalability.^{[ 15} , ^{16 ]} Among them, aqueous redox flow batteries (ARFBs) have gained greater attention due to their enhanced safety.^{[ 17} , ^{18 ]} Currently, ARFBs employing inorganic ions as electrolytes (primarily vanadium redox flow batteries (VRFBs)) dominate the market.^{[ 19} , ^{20 ]} However, due to the scarcity of inorganic materials, high costs, and significant life‐cycle carbon footprint (≈450 g CO₂ e kg⁻¹), research interest has shifted toward earth‐abundant organic redox‐active materials.^{[ 21} , ^{22 ]}

Figure 1.

a) Schematic illustration of the basic structure of RFBs. b)The redox reaction of quinones, phenazines, viologens and TEMPOs.

Organic redox‐active materials employ earth‐abundant elements, and their structural diversity and tunability enable the use of molecular engineering strategies to modulate their physical and chemical properties, such as: (i) functional group modifications improve solubility, expand the electrochemical window, and enhance cycling stability; (ii) cost‐effective modifications based on natural products, coupled with high‐quality; and (iii) tuning of electronic effects enables precise redox potential control.^{[ 23} , ²⁴ , ^{25 ]} To date, the most extensively studied organic redox‐active species include quinones,^{[ 26} , ^{27 ]} phenazines,^{[ 28 ]} TEMPO,^{[ 29} , ^{30 ]} and viologen^{[ 30} , ^{31 ]} derivatives (Figure 1b).

Figure 2 summarizes the research progress and developmental milestones of these molecules in recent years. Coupled with low‐cost synthesis processes and appropriate pollution management methods, aqueous organic redox flow batteries (AORFBs) could serve as a more economical and environmentally friendly alternative for grid‐scale energy storage.^{[ 32} , ^{33 ]}

Figure 2.

The development of typical organic electrode materials in AORFBs.

Stability is a critical factor that influences the sustainable development and large‐scale application of AORFBs.^{[ 34 ]} Among the various factors affecting the stability of AORFBs, the degradation of electrolytes and crossover of redox‐active species stand out as two major challenges to be addressed. Electrolyte degradation is a well‐known cause of impaired stability of AORFBs. Due to the inherently high reactivity of organic materials, side reactions like the structural rearrangements (e.g., tautomerization reactions) or attacks by nucleophilic species (e.g., hydroxide ions and water molecules) are prone to take place.^{[ 35} , ³⁶ , ³⁷ , ^{38 ]} Crucially, these processes are often irreversible. Once occurs, the electrolyte will convert into electrochemically inactive species, resulting in irreversible loss of battery capacity.

Meanwhile, the crossover of redox‐active species is another major reason for the insufficient stability of AORFBs. Assuming that the anolyte serves as the capacity‐limiting side, the reduced anolyte penetrates the membrane and then migrates to the catholyte side during the charging process. Then rapid electron transfer takes place between the anolyte and the oxidized catholyte, which is known as self‐discharge.^{[ 28 ]} As time goes by, an increasing amount of the anolyte undergoes crossover, leading to an undesired decline in discharge capacity and thus impairing the Coulombic efficiency. To address this issue, the design of ion‐selective membranes is a promising and cost‐effective strategy.

Therefore, to tackle the current issue of insufficient stability in AORFBs, it is first essential to investigate the underlying degradation mechanism of the electrolytes so that more targeted strategies to inhibit the adverse degradation could be developed. Meanwhile, the development of AORFB‐oriented and ion‐selective membranes also represents an effective approach to enhance stability. Through the modification of ion transport channels, surface and the polymer backbone, these membranes can not only suppress the crossover of redox‐active species but also enable high‐efficiency transport of charge‐carrying ions. Additionally, in order to achieve the efficient and accurate evaluation of AORFBs performance, researchers have proposed and developed a series of in‐situ and in operando characterization techniques (e.g., in‐situ NMR and UV–Vis spectroscopy).^{[ 39 ]} These advanced characterization methods facilitate the real‐time and accurate monitoring of the working processes and degradation mechanisms of AORFBs.^{[ 40 ]} More importantly, since one of the primary and ultimate targets of AORFB development is to achieve green and low‐cost electrochemical energy storage, it is a crucial task to assess the overall cost of AORFBs.^{[ 41 ]} Currently, due to the insufficient stability of electrolytes, frequent replacement of electrolytes is required for long‐term operation, which increases the operating costs. Meanwhile, commercial ion‐exchange membranes (e.g., Nafion and Selemion) are so expensive that their large‐scale adoption would result in overwhelmingly high costs. Thus, the development of electrolytes with high stability and ion‐selective AORFB‐oriented membranes also represents an effective measure to reduce the overall cost of AORFBs.

In the following sections, we select several representative categories of organic active species and discuss their common degradation mechanisms as well as strategies for improving their stability. Meanwhile, we will present the current progress achieved in the modification of membranes and elucidate the crucial role of membranes in enhancing the stability of AORFBs, specifically by suppressing the capacity fade caused by the crossover of redox‐active species.

2. Redox‐Active Species Applied in AORFBs

Prior to formally discussing the degradation mechanisms of organic active species, it is essential to understand their fundamental properties such as redox potential and solubility. The redox potential serves as the criterion for distinguishing between catholytes and anolytes. More importantly, it acts as a key basis for determining whether an organic material is applicable for AORFBs at a given pH condition. Taking alkaline AORFBs (pH 14) as an example, the redox potential of organic materials should locate within the electrochemical stability window of water (−0.83 V to 0.4 V vs. SHE. Note that this is a theoretical value and the influence of polarization should be taken into account in practical operation.). Only in this way can the occurrence of hydrogen evolution reactions (HER) or oxygen evolution reactions (OER) be avoided and make sure a stable operation. Figure 3 displays the redox potential of several common organic active materials, including anthraquinones, phenazines, viologens, TEMPO‐based and other compounds. Meanwhile, solubility in water solution is also a crucial criterion for evaluating whether an organic active material is suitable for application in AORFBs. According to the like‐dissolves‐like principle, organic molecules are generally more soluble in organic solvents (e.g., ethyl acetate, acetonitrile and DMF) whereas their solubility in water is very low. To address this issue, the water solubility of organic molecules can be significantly enhanced by approach such as introducing appropriate solubilizing groups. For instance, pristine phenazine molecules are barely soluble in water, while their solubility could be successfully increased to as high as 1.8 m after the introduction of sulfonic acid group and hydroxyl groups. In particularly, these two properties have a direct and profound impact on the energy density of AORFBs. Generally, higher cell voltage and higher solubility of active materials facilitate higher energy density. Since AORFBs was first reported in 2014, elevating energy density and enhancing stability have remained the two most important tasks. Currently, a considerable number of literatures have thoroughly discussed methods for boosting the energy density of AORFBs.^{[ 42 ]} Therefore, this review will otherwise focus on the degradation mechanisms of several representative redox‐active species (quinones, phenazines, viologens and TEMPO‐based compounds) and strategies to improve the stability of AORFBs.

Figure 3.

Comparison of redox potential of representative electroactive materials for AORFBs. Each symbol represents an organic active molecule. The applicable pH conditions for various organic compounds are marked using three colors: red (acidic), green (near‐neutral) and blue (alkaline). From left to right: FcNCl,^{[ 43 ]} BTMAP‐Fc,^{[ 31 ]} BOH‐Fc,^{[ 44 ]} BMIP‐Fc;^{[ 45 ]} TEMPTMA,^{[ 29 ]} N+TEMPOD,^{[ 46 ]} 4‐OH‐TEMPO,^{[ 30 ]} TPP‐TEMPO,^{[ 47 ]} TMAAcNH‐TEMPO;^{[ 48 ]} DBEAQ,^{[ 27 ]} DHAQ,^{[ 26 ]} 4C7SFL,^{[ 49 ]} Cys‐AE,^{[ 50 ]} DPivOHAQ;^{[ 51 ]} BHPC,^{[ 28 ]} HSPC,^{[ 52 ]} DHPS,^{[ 53 ]} 2,3‐O‐DBAP,^{[ 54 ]} 1,8‐PFP;^{[ 55 ]} Vi‐OEG3,^{[ 56 ]} APBPy,^{[ 57 ]} MV,^{[ 30 ]} (NPr)₂SV,^{[ 58 ]} BTMAP‐Vi^{[ 31 ]}; PTO‐PTS,^{[ 59 ]} TANQ,^{[ 60 ]} HydBPyMeCl,^{[ 61 ]} CSFA‐Cl,^{[ 62 ]} dex‐NDI,^{[ 63 ]} BDAHP,^{[ 64 ]} ds‐BC.^{[ 65 ]}

The degradation of electroactive materials represents a pivotal factor contributing to the capacity decay of the AORFBs. A profound comprehension of the degradation mechanisms of electroactive species is of highly desired since only in this way more effective molecular engineering strategies could be devised to substantially enhance the stability and lifetime of the AORFBs. In general, typical decomposition mechanisms of organic compounds include nucleophilic attack, disproportionation, dimerization, and tautomerization. For example, quinones are prone to experience nucleophilic attack^{[ 66} , ^{67 ]} and disproportionation.^{[ 68 ]} Viologens deployed in neutral systems would undergo dimerization to form a stable charge‐transfer complex when being single‐electron reduced.^{[ 69 ]} In addition, tautomerization is susceptible to occur among aza‐substituted conjugated π‐systems such as phenazines.^{[ 70 ]} In the following section, we summarize the decomposition mechanisms of quinones, phenazines, viologens and TEMPO. Also, by reviewing several representative works in recent years, we demonstrate how rational molecular engineering and other strategies help to optimize the performance of AORFBs.

2.1. Quinone‐Based Electroactive Materials

Quinone compounds are the most widely studied active materials for AORFBs, categorized into benzoquinone (BQ), naphthoquinone (NQ), and anthraquinone (AQ) derivatives.^{[ 71 ]} Since Huskinson et al. reported the application of 9,10‐anthraquinone‐2,7‐disulphonic acid (AQDS) in AORFB for the first time,^{[ 17 ]} a variety of quinone derivatives have been extensively designed and studied due to their high redox reversibility and structural tunability.

The simplest type of quinone derivatives are benzoquinones and they mainly serve as catholytes. However, AORFBs based on benzoquinones only demonstrate moderate capacity retention rate as these molecules are prone to undergo degradation process such as nucleophilic attack. Yang et al. investigated the performance and decomposition mechanism of the AORFB based on 4,5‐dihydroxybenzene‐1,3‐disulfonic acid (BQDS) catholyte.^{[ 72 ]} They found that during the charging process, the hydroxyl groups of BQDS first underwent oxidation to quinone and then was attacked by water to form reduced product, which is demonstrated in Figure 4a. From the prospective of molecular structure, the benzene ring is linked with two electron‐withdrawing sulfonic acid groups, which results in a highly electron‐deficient nature of the BQDS. Moreover, two unsubstituted aromatic ring carbons remain available as attack sites for nucleophilic reagents such as water molecules. Therefore, to suppress the nucleophilic reaction of benzoquinone, appropriate electron‐donating groups should be introduced while protecting the unsubstituted sites. Murali et al. designed a novel 3,6‐dihydroxy‐2,4‐dimethylbenzenesulfonic acid (DHDMBS) as a Michael reaction‐free catholyte.^{[ 66 ]} However, a capacity decay was still observed upon further cycling, which was attributed to the crossover of DHDMBS. Surprisingly, Murali et al. quantified the total amount of DHDMBS on both the catholyte and anolyte after cycling, which was lower than that before cycling. NMR analysis revealed the desulfonation can be considered as a substitution of the sulfonic acid group by a proton under acidic conditions (Figure 4b). Although reducing proton concentration does inhibit the rate of desulfonation, studies by Narayan et al. revealed that the increasing pH introduces new challenges for DHDMBS: hydroxide ions abstract the hydrogen atom from methyl group, leading to an irreversible structural transformation. Thus, under alkaline conditions, the catholyte commonly employs the more stable ferrocyanide/ferricyanide redox couple.

Figure 4.

a) The degradation of BQDS induced by nucleophilic attack. b) The desulfonation of DHDMBS. c) The possible decomposition mechanism of DHBQ. d) The tautomerization of bislawsone. e) The degradation pathway of 3‐AFNQ.

In fact, the employment of benzoquinone as a catholyte material presents a dilemma. On one hand, a reasonably high redox potential is favorable for enhancing the energy density of the battery. On the other hand, introducing electron‐donating groups to suppress nucleophilic reactions causes a decrease in the redox potential of benzoquinone. In the two studies above, benzoquinones were used as the catholyte in AORFBs. However, these batteries had the drawback of an insufficient open‐circuit voltage. To circumvent this issue, Yang et al. used the simple 2,5‐dihydroxy‐1,4‐benzoquinone (DHBQ) as the anolyte in alkaline AORFB.^{[ 73 ]} Although this cell exhibits a high voltage of 1.21 V, its capacity fade rate is undesirably high as 9% per day. In Yang et al. opinions, DHBQ undergoes decomposition, particularly due to nucleophilic attack at the unsubstituted C3 and C6 positions of DHBQ by hydroxide ions in alkaline solution (Figure 4c).

Due to the extended π‐conjugated structure of bis‐cyclic Naphthoquinones, they are usually applied as the anolyte of AORFBs and exhibit superior stability to benzoquinones. However, naphthoquinones remain susceptible to nucleophilic attack. For instance, Tong et al. synthesized bislawsone with four‐electron‐transfer capability by linking two Lawsone monomers at the 3‐position.^{[ 74 ]} Although Michael addition was suppressed, capacity decay was still observed. Tong et al. speculated that when bislawsone was half‐reduced to quinone‐hydroquinone, enol‐ketone tautomerization occurs and irreversibly transforms it into 2,3‐dihydrobislawsone (Figure 4d). According to theoretical predictions, several more stable bislawsone derivatives contain oxy‐alkyl substituents, which is a common strategy to enhance anthraquinone stability. Liu et al. synthesized a biomimetic 3‐(2‐chloro‐1,4‐naphthoquinon‐3‐ylamino)propanoic acid (3‐AFNQ) as the anolyte for alkaline AORFB.^{[ 75 ]} The high‐concentration battery demonstrated good stability with a capacity retention of 99.61% per day. The authors identified minor Cl detachment from 3‐AFNQ forming compound 2‐amino‐1,4‐naphthoquinone, which is susceptible to Michael addition reactions and causes slight capacity fade (Figure 4e).

2,6‐dihydroxyanthraquinone (2,6‐DHAQ) is a classic anthraquinone‐based electroactive material with cost‐effective advantages. However, AORFBs based on DHAQ exhibit significant capacity decay (5%–8% per day).^{[ 76 ]} Goulet et al. investigated the degradation mechanism through symmetrical cell measurements.^{[ 77 ]} Their findings revealed that the reduced form DHAHQ undergoes disproportionation to form DHA, which further undergoes oxidative coupling and irreversibly dimerizes into (DHA)₂ (Figure 5a). Such process was confirmed by in‐situ ¹H‐NMR test (Figure 5b).^{[ 39 ]} Although (DHA)₂ is electrochemically active, it is not an ideal negolyte material. First, it is unstable under pH 14. Meanwhile, it can only transfer two electrons, resulting in a specific capacity per molecular weight that is only half that of DHAQ. Lastly, its redox potential is not as low as DHAQ's, reducing the open‐circuit voltage and energy density of AORFB. The authors further pointed out that anthraquinones with lower redox potentials are more susceptible to disproportionation. Ideally, DHAQ's low‐potential advantage should be preserved while enhancing stability. To address this, Goulet et al. proposed two strategies. First, controlling the state of charge to avoid the formation of DHA. Second, introducing air into the negolyte to oxidize DHA back to DHAQ. By simply adjusting the charge cut‐off voltage from 1.6 to 1.25 V, the capacity decay rate of DHAQ‐based AOFBs was successfully reduced to lower than 1% per day. In addition to the two strategies mentioned above, Jing et al. proposed novel solutions in subsequent studies. They discovered that (DHA)₂ can be oxidized back to DHAQ at ≈0 V (vs. Fe(CN)₆ ^4‐/3−).^{[ 78 ]} By setting the discharge cut‐off voltage to 0.01 V, DHAQ regeneration was achieved at the end of each discharge step. This strategy effectively reduced the capacity decay rate to 0.39% per day even setting a high state of charge (Figure 5c). In addition, given that the disproportionation of reduced DHAQ generates hydroxide ions, Bahari et al. adjusted the concentration of OH⁻ in the electrolyte solution to 1.5 M, effectively suppressing the disproportionation. By simultaneously combining controlling states of charge and DHAQ regeneration strategies, they successfully reduced the capacity decay rate of DHAQ‐based AOFBs to 0.02% per day.^{[ 79 ]}

Figure 5.

a) The disproportionation of DHAQ and the subsequent dimerization. b) Schematic of the in‐situ NMR setup. Reproduced with permission.^{[ 39 ]} Copyright 2020, Springer. c) The performance of DHAQ‐based flow battery utilizing deep discharge strategy. Reproduced with permission.^{[ 78 ]} Copyright 2022, Springer. d) The degradation induced by side chain loss.

Implementing appropriate molecular engineering strategies can yield anthraquinone materials with superior stability. Kwabi et al. synthesized 4,4′‐((9,10‐anthraquinone‐2,6‐diyl)dioxy)dibutyrate (2,6‐DBEAQ) decorated with carboxylic side chains and found the decay rate of it only 0.0075% per day under pH 14.^{[ 27 ]} Compared to DHAQ, 2,6‐DBEAQ exhibits higher redox potentials and the negatively charged carboxyl groups were spaced farther apart, thus reducing the negative impact of intramolecular Coulombic repulsion on stability. However, side chain cleavage under alkaline conditions represents a decomposition pathway for this class of electroactive organics and elevated temperatures exacerbate the decomposition (Figure 5d). To address this issue, Wu et al. designed and synthesized 3,3′‐(9,10‐anthraquinone‐diyl)bis(3‐methylbutanoic acid) (DPivOHAQ) with carbon‐carbon bond‐linked side chains, achieving ultrastable operation under pH 14 (capacity decay rate as low as 0.0018% per day).^{[ 51 ]} Based on previous studies of DHAQ, the authors attributed the degradation of DPivOHAQ to disproportionation of the reduced state, which generates anthrone. Since the disproportionation reaction also generates hydroxide ions, increasing the pH of electrolyte effectively suppresses the disproportionation. Additionally, aeration can oxidize the anthrone back to anthraquinone, restoring most of the decayed capacity. Based on the carbon‐carbon bonded strategy, Jing et al. further synthesized a series of saturated or unsaturated side chain‐substituted anthraquinone materials, from which the most stable candidate 9,10‐anthraquinone‐2,6‐dipropionic acid (AQDP) was screened and identified (capacity decay rate of 0.0128% per day).^{[ 80 ]} The authors highlighted that a stable redox molecule should have at least two carbons between its redox core and the water‐solubilizing group. In addition to the degradation mechanisms mentioned above, researchers revealed that specific anthraquinone‐based AORFBs undergo desulfonation and bromination. Bromine substitution of aromatic ring hydrogens is a type of electrophilic attack. For example, Gerhardt et al. assembled DHAQDS‐bromide flow cells using a posolyte containing 0.5 m Br₂ and 3 m HBr. However, bromine is able to permeate the membrane to the negolyte containing DHAQDS, which leads to the irreversible bromination between DHAQDS and bromine. They observed that DHAQDS reacted with bromine to form an orange solid, as indicated by the disappearance of aromatic proton signals in NMR spectra.^{[ 81 ]} Furthermore, elemental composition analysis of the solid product found very little content of sulfur, confirming the occurrence of desulfonation. The authors supposed that the electron‐donating hydroxyl groups increase the electron density of the aromatic system, thereby facilitating electrophilic attack by bromine.

To summarize, regarding the degradation issue of anthraquinone‐based compounds, the feasible strategies are as follows: (1) Regulating the testing conditions of the full cell. Controlling an appropriately low state of charge (SOC) to reduce the generation of the disproportionation product. Simultaneously, setting a discharge cutoff voltage of ≈0 V vs. [Fe(CN)₆]^4−/3− to achieve the oxidative regeneration of oxidated anthraquinones; (2) Adjusting the composition of the electrolyte. Appropriately increasing the hydroxide ion concentration to inhibit the occurrence of the disproportionation of reduced anthraquinones; (3) Implementing molecular engineering strategies. Connecting saturated long alkyl side chains to the anthraquinone core via stable carbon‐carbon bonds.

In recent years, researchers have developed a series of quinone derivatives and optimized their performance indicators such as redox potential, solubility and stability. Some representative quinone‐based compounds are selected and their structures and properties are presented in Figure 6 and Table 1 .

Figure 6.

Structures of quinone derivatives and their solubilities reported in AORFBs.

Table 1.

The redox potential of quinone derivatives and their cycling stability in alkaline AORFBs.

Anolyte	Catholyte	E1/2 [V vs. SHE]	pH	Capacity fade rate [demonstrated concentration]	Refs.
DHBQ	K4Fe(CN)6	−0.72	14	0.24%/cycle (0.5 m)	[73]
Bislawsone	K4Fe(CN)6	−0.551	14	0.74%/day (0.5 m)	[74]
3‐AFNQ	K4Fe(CN)6	−0.41	14	0.39%/day (0.5 m)	[75]
DHAQ	K4Fe(CN)6	−0.7	14	5.6%/day (0.5 m)	[26]
2,6‐DBEAQ	K4Fe(CN)6	−0.515	12	0.04%/day (0.5 m)	[27]
2,6‐DPPEAQ	K4Fe(CN)6	−0.49	12	0.014%/day (0.5 m)	[82]
DPivOHAQ	K4Fe(CN)6	−0.48	12	0.0018%/day (0.5 m)	[51]
2‐2PEAQ	K4Fe(CN)6	−0.477	14	0.03%‐0.05%/day (0.1 m)	[83]
AQDP	K4Fe(CN)6	−0.456	14	0.0128%/day (0.1 m)	[80]
PEGAQ	K4Fe(CN)6	−0.52	14	0.5%/day (1.5 m)	[68]
DAEAQ	K4Fe(CN)6	−0.61	14	0.14%/day (0.5 m)	[84]
Cys‐AE	K4Fe(CN)6	−0.51	14	0.0438%/day (0.5 m)	[85]
Cys‐DHAQ	K4Fe(CN)6	−0.492	14	0.011%/day (0.5 m)	[86]

2.2. Phenazine‐Based Electroactive Materials

In addition to quinone compounds, researchers have been actively developing novel anolyte materials suitable for alkaline and near‐neutral environments in recent years, such as phenazine derivatives. Phenazines and their derivatives of natural origin exhibit biological activities such as antibacterial, antiviral, and antitumor properties and are also commonly used in the dye industry. They possess the advantages of being naturally available and inexpensive. The core of phenazine compounds is a N‐heterocycle ring, which can undergo a two‐electron and two‐proton transfer process and exhibits good redox reversibility. The practical application of phenazine‐based materials began in 2018 when Hollas et al. synthesized 7,8‐dihydroxyphenazine‐2‐sulfonic acid (DHPS).^{[ 53 ]} Due to the strong solvation effects of sulfonate and hydroxyl groups, molecular aggregation was suppressed and endowing DHPS with an extremely high solubility of 1.8 M and an electron concentration of 3.6 M in 1 M NaOH solution. In addition, DHPS demonstrates an exceptionally low redox potential of ‐0.88 V (vs. SHE) due to the negatively charged sulfonate group with mitigated electron‐withdrawing effects and two strongly electron‐donating hydroxyl groups. Consequently, the DHPS||K₄Fe(CN)₆ AORFB achieves a cell voltage of 1.4 V, with the negative electrolyte delivering a capacity density of 67.4 Ah L⁻¹ under near‐saturated conditions. However, this battery exhibited a capacity decay rate of 0.68% per day (Figure 7a).

Figure 7.

a) The long‐term cycling performance of DHPS‐based flow battery. Reproduced with permission.^{[ 53 ]} Copyright 2018, Springer. b)The desulfonation, deoxygenation and tautomerization of DHPS. c) The tautomerization route of phenazine derivatives. d) The side chain loss of DAEP. e) The long‐term cycling performance of 2,3‐O‐DBAP‐based flow battery. Reproduced with permission.^{[ 54 ]} Copyright 2024, American Chemical Society.

Wellala et al. investigated the degradation mechanism of DHPS and pointed out that the reduced H₂DHPS underwent desulfonation to form H₂‐2,3‐DHP.^{[ 70 ]} Additionally, H₂DHPS also undergoes deoxygenation, converting into MHPS with a single hydroxyl group. MHPS then undergoes tautomerization and eventually transforms into an electrochemically inactive hydrogenated species (Figure 7b). In addition, Wellala et al. further indicated that hydroxyl groups positioned at the ortho to the central nitrogen can provide DHP with improved stability. For instance, 1,4‐DHP and 1,6‐DHP exhibited lower capacity decay rates (0.029% and 0.031% per day, respectively). Pang et al. systematically discussed the universal tautomerization mechanism of phenazine derivatives.^{[ 87 ]} They suggest that N‐position hydrogens in reduced phenazine migrates to the side ring to form electrochemically inactive 1,2‐dihydrophenazine, resulting in capacity decay (Figure 7c). Theoretically, phenazines substituted at positions 2 and 3 exhibit superior stability. However, the hypothesis is yet to be confirmed by experiments. Liu et al. synthesized two phenazine derivatives, Phenazine‐(2,3‐diyl)dioxy diacetic acid and Phenazine‐(2,3‐diyl)dioxy dibutyric acid (DAEP and DBEP), to validate this assumption.^{[ 88 ]} In Liu et al. opinions that side‐chain cleavage occurred, but it does not directly affect phenazine redox activity (Figure 7d). During long‐term cycling, the DBEP‐based AORFB showcased impressive stability with a capacity decay rate of only 0.024% per day. Kong et al. synthesized a series of phenazine isomers (O‐DBAP) bearing same substituents as DBEP.^{[ 54 ]} The authors primarily investigated the effects of substituent positions on the performance of phenazine derivatives. Results revealed that the energy of re‐2,3‐O‐DBAP is lower than its tautomers and indicated that tautomerization is thermodynamically unfavoured. Subsequent cycling tests demonstrated that the 2,3‐O‐DBAP‐based AORFB achieved a low capacity decay rate of 0.0127% per day (Figure 7e). In addition to degradation mechanisms such as tautomerization, desulfonation and side chain cleavage, researchers have identified specific decomposition pathways for phenazine derivatives with particular structure (e.g. hydroxyphenazine). Wang et al. suggested that the substituent adjacent to the hydroxyl group significantly influence properties of hydroxyphenazine such as solubility, stability, and redox potential.^{[ 28 ]} They found that dimerization occurred between oxidized and reduced states of 2‐hydroxyphenazine (HP) (Figure 8a), while 2‐amino‐3‐hydroxyphenazine (AHP) and Benzo[a]hydroxyphenazine‐7/8‐carboxylic acid (BHPC) did not. Subsequently, they performed a galvanostatic‐potentiostatic cycling test and reduced the decay rate to 0.04% per day.

Figure 8.

a) The dimerization of 2‐hydroxyphenazine. b) The structures of a series of PFPs. Reproduced with permission.^{[ 55 ]} Copyright 2021, Elsevier. c) The hybrid resonant structure of HSPC. Reproduced with permission.^{[ 52 ]} Copyright 2025, Wiley. d) The long‐term cycling performance of HSPC‐based flow battery. Reproduced with permission.^{[ 52 ]} Copyright 2025, Wiley.

Significant advancements have been made in studies on improving the stability of phenazine derivatives. However, environmental factors especially the temperature have frequently been overlooked. Given that renewable energy sources such as solar energy are typically generated in sunlight‐rich regions where ambient temperature is elevated, developing high‐temperature‐tolerated active materials is of great significance. To address this, Xu et al. synthesized a series of propionic‐acid‐functionalized phenazines (PFPs) adopting the carbon‐carbon bond linking strategy, which demonstrated exceptional stability at both ambient and elevated temperatures (Figure 8b).^{[ 55 ]} Owing to the chemically inert carbon‐carbon bonds, no tautomerization was observed in the reduced PFPs. Results indicated that 2,2’‐(phenazine‐1,2‐diyl)diacetic acid and 2,2’‐(phenazine‐1,4‐diyl)diacetic acid (1,2‐EFP and 1,4‐EFP) exhibited superior properties of high solubility, low reduction potentials, and unsusceptibility to tautomerization, providing critical insights for rational design of next‐generation electroactive materials.

The aforementioned researches have demonstrated that the main strategies for enhancing the stability of phenazine derivatives include: (1) Regulating the position of substituents. 1,4‐, 1,6‐, and 2,3‐substituted phenazine compounds can inhibit adverse tautomerization and thus exhibit higher stability; (2) Optimizing the bonding mode of substituents. For phenazine compounds where substituents are linked to the redox core via carbon‐carbon bonds, the substituents are less likely to detach compared with the traditional bonding mode such as carbon‐oxygen or carbon‐nitrogen bonds. Consequently, these compounds exhibit higher stability.

Recently, Li et al. have proposed a novel strategy based on a resonance hybrid design, which effectively enhances the stability of phenazine compounds.^{[ 52 ]} They pointed out that the reason for tautomerization of reduced phenazines lies in their unstable state with excessively high electron density, where the molecular structure tends to rearrange to reduce the overall energy. To not only ensure a sufficiently high electron density for inhibiting the attack of nucleophilic reagents but also avoid the occurrence of tautomerization, Li et al. synthesized 2,3‐dihydroxyl‐substituted phenazine dication (HSPC) via electrochemical methods and subsequently investigated the hybrid resonate structure of HSPC (Figure 8c). Its redox centers are distributed among different resonance forms between phenazine and quinone, thereby enabling efficient and reversible electron transfer during redox reactions. Finally, high‐concentration batteries (0.7 M) were assembled. The battery exhibited excellent stability, with capacity decay rates of 0.006% per day (Figure 8d).

Up to now, an increasing number of phenazine derivatives have been applied in alkaline AORFBs. We have selected a number of representative phenazine‐based molecules and presented their structures and properties in Figure 9 and Table 2 . Moreover, as the degradation mechanism of phenazine derivatives is further clarified, more strategies to enhance the stability will be supposed, which facilitates the realization of long‐life AORFBs.

Figure 9.

Structures of phenazine derivatives and their solubilities reported in AORFBs.

Table 2.

The redox potential of phenazine derivatives and their cycling stability in alkaline AORFBs.

Anolyte	Catholyte	E1/2 [V vs. SHE]	pH	Capacity fade rate [demonstrated concentration]	Refs.
DHPS	K4Fe(CN)6	−0.90	14	0.68%/day (1.4 M)	[53]
BHPC	K4Fe(CN)6	−0.78	14	0.08%/day (0.5 M)	[28]
1,6‐DPAP	K4Fe(CN)6	−0.56	12	0.0015%/day (0.5 M)	[87]
1,8‐PFP	K4Fe(CN)6	−0.588	14	0 (1.0 M)	[55]
DBEP	K4Fe(CN)6	−0.68	14	0.044%/day (0.6 M)	[88]
2,3‐O‐DBAP	K4Fe(CN)6	−0.705	14	0.0127%/day (0.1 M)	[54]
HDMPC	Li4[Fe(CN)6]	−0.708	14	0.38%/day (0.1 M)	[89]
HSPC	K4Fe(CN)6	−0.36	10	0.006%/day (0.7 M)	[52]

2.3. Fluorenone‐based electroactive materials

To expand the development space of aqueous organic flow batteries, researchers have been developing novel electroactive materials, such as fluorenone‐based compounds. Feng et al. successfully achieved the reversible hydrogenation and dehydrogenation of fluorenone compounds in aqueous solution and applied them in AORFBs.^{[ 49 ]} By introducing electron‐withdrawing groups (such as carboxyl and sulfonic acid groups), the authors successfully decreased the pKa value of fluorenol and thus effectively enhanced the redox reversibility of fluorenone compounds. In addition, by constructing an asymmetric structure, the authors increased the solubility of fluorenone compounds to exceeding 1 M and achieved an electron concentration close to 3 M. Feng et al. assembled a full battery with the screened 4‐carboxylic‐7‐sulfonate fluorenone (4C7SFL) at a concentration of 1.36 M. The results show that 4C7SFL exhibits excellent long‐term cycling stability (with a capacity decay rate of only 0.0209% per day). Different from previous reported AORFBs, fluorenone showed a higher capacity utilization rate at high concentrations. The authors investigated and discussed this phenomenon. They found that after 4C7SFL is reduced to a radical anion, a disproportionation reaction would occur to generate 4C7SFL and 4C7SFL‐OH, which is the rate‐determining step. A high concentration of 4C7SFL is beneficial for the occurrence of disproportionation reaction so that increasing the capacity utilization rate of the active materials.

2.4. TEMPO‐Based Electroactive Materials

TEMPO (2,2,6,6‐tetramethylpiperidine‐1‐oxyl) and its derivatives constitute a variety of stable, non‐conjugated redox‐active nitroxide radicals that exhibit reversible single‐electron transfer characteristics during electrochemical redox processes (Figure 1). These compounds reveal significant potential as catholyte materials for pH‐neutral AORFBs due to their superior standard redox potential and exceptional electrochemical reversibility. Given the poor water solubility of TEMPO, current research increasingly focuses on its derivatives modified with hydrophilic groups, particularly through γ‐position modifications.

4‐OH‐TEMPO demonstrates high redox reversibility and fast redox kinetics in pH‐neutral aqueous systems. Additionally, it offers several key advantages such as low molecular weight, relatively high redox potential (0.8 V vs. NHE), significant water solubility (2.1 m), and excellent electrochemical stability.^{[ 30 ]} These properties make it a promising cathode material for pH‐neutral AORFBs. However, as an electrically neutral molecule, 4‐OH‐TEMPO suffers from significant crossover during battery cycling, which limits its practical applications. To address this issue, two primary strategies can be employed: increasing the molecular volume or modifying TEMPO with charged functional groups.

By replacing the ‐OH at the γ‐position of 4‐OH‐TEMPO with a positively charged ‐N⁺(CH₃)₃, Janoschka et al. successfully synthesized a TEMPO derivative called TEMPTMA.^{[ 29 ]} Xu et al. utilized 4‐OH‐TEMPO as a precursor to prepare TMAP‐TEMPO, which also carries a positively charged functional group.^{[ 90 ]} Liu et al. engineered TMAAcNH‐TEMPO by introducing 3‐(trimethylammonium)acetylamino groups into the γ‐position of TEMPO through molecular engineering.^{[ 48 ]} TEMPTMA, TMAP‐TEMPO, and TMAAcNH‐TEMPO exhibit significantly reduced crossover through cation‐exchange membranes due to the Donnan exclusion effect. Moreover, the charge repulsion between TEMPO molecules suppresses dimerization, further enhancing their stability.

With the mitigation of molecular transmembrane crossover achieved, let we shift focus to the impact of environmental factors on the structural degradation of TEMPO derivatives and the consequent battery capacity deterioration. For instance, OH⁻ can nucleophilically attack the N atom in the oxidized TEMPO⁺ species (Figure 10 ), ultimately reducing it to TEMPO while generating reactive hydroxyl radical (OH·).^{[ 91 ]} Furthermore, the binding of OH⁻ to the β‐hydrogen of TEMPO⁺ induces ring‐opening reactions, and the resulting fragmented species may undergo subsequent decomposition or parasitic side reactions (Figure 10).^{[ 92 ]} Additionally, under conditions of elevated proton (H⁺) concentration, TEMPO derivatives are prone to disproportionation (Figure 10). Alternatively, over‐reduction of TEMPO in acidic environments facilitates its reaction with H⁺ to form hydroxylamine derivatives, further destabilizing the redox‐active framework (Figure 10).^{[ 93 ]} Hydroxylamine derivatives exhibit difficulty in being oxidized and regenerated during battery cycling. Furthermore, it has been proposed that hydroxylamine derivatives may form hydrogen‐bonded dimers with other active molecules, thereby limiting the availability of the remaining molecules (Figure 10). However, Nolte et al. demonstrated that the hydroxylamine generated from TEMPTMA reduction can be reoxidized back to its original TEMPTMA form through an oxidative regeneration process. Although this regeneration occurs at a relatively slow kinetic rate, it partially compensates for the capacity fading observed in battery systems.^{[ 94 ]}

Figure 10.

Side Reactions of TEMPO derivatives.

Both reactions between TEMPO derivatives and OH⁻ or H⁺ ultimately induce pH variations in the electrolyte solution. Such pH imbalance subsequently accelerates the decomposition of TEMPO derivatives through a self‐accelerating and irreversible decomposition pathway. This mechanistic understanding underscores the critical importance of maintaining stable pH conditions for system preservation. Furthermore, TEMPO derivatives exhibit enhanced thermal decomposition tendencies at elevated temperatures, which may limit their practical application under high‐temperature operating conditions.^{[ 93 ]} Considering that TEMPO derivatives are highly sensitive to pH changes, a critical question arises: do pH fluctuations occur during battery cycling, and if so, could they compromise the molecular stability? Unfortunately, there are currently few researches investigating this specific aspect.

Tang et al. investigated the correlation between the Hirshfeld charges of TEMPO derivatives and the Gibbs free energy (ΔG) associated with ring opening or hydrolyzation side reactions triggered by OH⁻ attack.^{[ 47 ]} Comparative analysis revealed that β‐hydrogen and the N atom of piperidine in TEMPTMA exhibit significantly higher Hirshfeld charges and lower ΔG values than those in TMAP‐TEMPO and TMAAcNH‐TEMPO. This property rationalizes the faster capacity decay observed in batteries employing TEMPTMA as the redox‐active material. Concurrently, they engineered a novel TEMPO derivative, TPP‐TEMPO, featuring reduced Hirshfeld charges on critical atomic sites. A combined analysis of these four derivatives revealed a potential linear correlation between Hirshfeld charges and the ΔG values of side reactions. As shown in Figure 11 , the results of battery operation established that TPP‐TEMPO exhibited a substantially lower capacity fading rate (0.010% per hour) compared to TEMPTMA, which degraded at a rate an order of magnitude higher (0.16% per hour). Beyond Hirshfeld charges, whether other fundamental physicochemical properties of molecules are correlated with molecular degradation remains an open question worthy of further investigation.

Figure 11.

Flow cells assembled with 5 mL of 0.1 m TPP‐TEMPO (or TMA‐TEMPO) and 7.5 mL of 0.1 m BTMAP‐Vi in 1 m NaCl solution at 20 mA cm⁻². Insets: The discharge curves of TPP‐TEMPO and TMA‐TEMPO at the 300th cycle. Reproduced with permission.^{[ 47 ]} Copyright 2025, Springer.

Crucially, the dominant decomposition mechanism and reaction kinetics correlate strongly with the electronic and steric properties of the γ‐position substituents. TMAAcNH‐TEMPO, featuring a bulkier γ‐position substituent with localized positive charges, exhibits enhanced stability relative to 4‐NH₂‐TEMPO. The steric hindrance generated by the extended γ‐group suppresses transmembrane migration through electrostatic repulsion, while the charge redistribution within the molecule reduces susceptibility to both electrophilic and nucleophilic attacks. These synergistic effects collectively stabilize the oxidized state, thereby significantly diminishing the likelihood of ring‐opening decomposition pathways.^{[ 48 ]}

Schubert et al. postulated three plausible decomposition pathways for oxidized TEMPO derivatives functionalized with quaternary ammonium groups.^{[ 94} , ^{95 ]} including ring opening via cleavage of the C‐N bond on piperidine, Hofmann elimination of quaternary ammonium groups and releasing chloromethane (Figure 12 ). The active site of TEMPO derivatives is stabilized through steric shielding by four adjacent methyl groups. Substitution of these methyl groups with bulkier substituents (e.g., tetrahydropyran rings) can enhance this shielding effect, thereby providing superior protection for both the TEMPO radical and its oxidized states. Compared with TEMPTMA bearing a strongly electron‐withdrawing quaternary ammonium group at the γ‐position, TMAAcNH‐TEMPO and TMAP‐TEMPO involve the incorporation of spacer groups between the quaternary ammonium group and the TEMPO backbone. This design strategy not only optimizes intramolecular charge distribution but also significantly suppresses Hofmann elimination pathways that generate cyclic double bonds. Substitution of conventional Cl⁻ counterions with BF₄ ⁻ effectively mitigates the undesirable the elimination of chloromethane, though concomitant impacts on ionic conductivity and solubility parameters must be considered.

Figure 12.

Possible side reactions of TEMPO derivatives and corresponding solutions. Reproduced with permission.^{[ 95 ]} Copyright 2022, Royal Society of Chemistry.

Xu et al. engineered a morpholine‐functionalized TEMPO derivative (Mor‐TEMPO), enhancing molecular stability by substituting the N‐methyl morpholine group.^{[ 96 ]} The steric shielding of the stable cyclic architecture suppresses Hofmann elimination side reactions. AORFBs employing 2.5 m Mor‐TEMPO demonstrated exceptional cycling stability, achieving 99.96% capacity retention per cycle (0.48% daily fade rate) over 200 cycles at 25 °C. Remarkably, symmetric cells retained high capacity retention (≈100%) after 200 cycles at 40 °C, with no detectable decomposition byproducts via post‐cycling analysis.

The inherent limitation of TEMPO derivatives lies in their single‐electron redox mechanism, which potentially constrains the enhancement of energy density of AORFBs. To address this challenge, Lv et al. developed a building‐block assembly platform that facilitated the synthesis and systematic investigation of 21 i‐TEMPOD dimeric derivatives (a molecule with two TEMPO‐based redox centers).^{[ 46 ]} These dimers effectively minimize redox‐inactive structural components while optimizing the spatial configuration of piperidine rings. This molecular engineering strategy establishes a robust aqueous coordination network that induces a distinctive “water‐in‐salt” state. Notably, N+TEMPOD demonstrates exceptional electrochemical performance, achieving a volumetric capacity of 101 Ah L⁻¹ at 2 m concentration while achieving no measurable capacity decay after 96 days (Figure 13 ). This breakthrough simultaneously addresses the dual challenges of energy density improvement and cycling stability in organic battery systems. The presented molecular design paradigm and assembly methodology provide valuable insights for developing next‐generation energy storage materials.

Figure 13.

Long‐term RFB cycling profile using 2.0 m N+TEMPOD as catholyte (capacity‐limiting side) and 1.0 m Dex‐Vi as anolyte at room temperature with an AMVN membrane. Reproduced with permission.^{[ 46 ]} Copyright 2023, Springer.

In summary, the degradation mechanism of TEMPO derivatives is exceedingly complex, and there is no clear consensus on its explanation. It is possible that in the near future, a novel degradation pathway will be discovered, or a more holistic understanding that integrates various mechanisms will be achieved, thereby providing renewed and systematic guidance for molecular design. Some representative TEMPO‐based compounds are selected and their structures and properties are presented in Figure 14 and Table 3 .

Figure 14.

Structures of a subset of TEMPO derivatives and their solubilities utilized in AORFBs. (The solubility of TEMPTMA was measured in a 0.3 m NaCl solution, while the solubilities of other molecules were determined in pure water. Under these predefined conditions, the pH is assumed to be 7, though actual pH values may vary).

Table 3.

The redox potential, pH and cycling stability of a subset of TEMPO derivatives in AORFBs. (If it is not explicitly stated whether the electrolyte contains substances that may alter the pH, the pH is assumed to be 7 by default. Although actual pH levels may fluctuate due to factors such as dissolution of CO₂.).

Nitroxide radicals	Anolyte	E1/2 [vs. SHE] a)	pH	Capacity fade rate [demonstrated concentration]	Refs.
4‐OH‐TEMPO	MV	0.80 V	7	0.11%/cycle (0.5 m)	[30]
TEMPTMA	MV	0.99 V	7	No significant fade (2.0 m)	[29]
TMAP‐TEMPO	BTMAP‐Vi	0.81 V	7	0.023%/hour (1.5m)	[90]
TMAAcNH‐TEMPO	(NPr)2V	0.84 V	7	0.0144%/hour (0.5m)	[48]
TPP‐TEMPO	BTMAP‐Vi	0.77 V	≈4 b)	0.0067%/hour (0.5 m)	[47]
Mor‐TEMPO	BTMAP‐Vi	0.98 V	7	0.48%/day (2.5 m)	[96]
N2‐TEMPO	(NPr)2V	1.0 V	7	0.025%/cycle (1.0 m)	[97]
TMP‐TEMPO	PyrTMAV	0.98 V	7	3.08%/day or 0.34%/cycle (1.5 m)	[98]
TPABPy	TPABPy	0.96 V	7	0.01%/cycle (1.0 m)	[99]
4‐CO2Na‐TEMPO	(SPr)2V	0.80 V	7.2	0.2%/cycle (0.4 m)	[92]
TEMPO‐4‐sulfate	Zn	0.81 V	c)	0.0058%/cycle (0.035 m)	[100]
MIAcNH‐TEMPO	Zn	0.83 V	7	0.0456%/day (1.5 m)	[101]
N+TEMPOD	Dex‐Vi	0.962 V	7	No measurable capacity decay (2.0 m)	[46]

^a) Partial values of E_1/2 are converted to the values versus SHE. (E_Ag/AgCl = 0.2 V, E_NHE = E_SHE = 0 V);

^b) This value is an approximate reading obtained from the literature figure;

^c) Since NH₄Cl was used as the supporting electrolyte, the exact pH cannot be determined.

2.5. Ferrocene Based Electroactive Materials

Ferrocene and its derivatives are capable of undergoing reversible single‐electron redox transfer, making them promising cathode candidate materials. However, due to the hydrophobic nature of ferrocene molecules, most ferrocene derivatives currently employed in AORFBs are modified with hydrophilic functional groups. Hu et al. introduced hydrophilic quaternary ammonium groups into the ferrocene structure, synthesizing FcNCl with a solubility of up to 4.0 m.^{[ 43 ]} The FcNCl‖MV AORFB system demonstrated a capacity decay rate of 0.04% per cycle at a concentration of 0.7 M.

To better elucidate the degradation mechanisms of ferrocene derivatives, Chen et al. investigated six ferrocene derivatives bearing different substituents.^{[ 44 ]} They proposed that the degradation of ferrocene primarily initiates from the nucleophilic attack on the metal center of the oxidized ferrocene species. Consequently, a higher density of the Lowest Unoccupied Molecular Orbital (LUMO) on the metal center of oxidized ferrocene makes it more susceptible to nucleophilic attack and subsequent degradation. Among these six compounds, BQH‐Fc whose metal center possesses the lowest LUMO density, exhibited the highest stability. The BQH‐Fc‖BTMAP‐Vi cell demonstrated a capacity retention rate of 99.993% per hour at a concentration of 1.5 m. They also proposed a decomposition pathway (Figure 15a). The Fe atom in the ferrocene derivative undergoes nucleophilic attack by a water molecule, leading to the detachment of a cyclopentadienyl (Cp) anion. This anion then reduces other oxidized ferrocene molecules, resulting in capacity loss. It has also been reported that the resulting [Fe^II(H₂O)₆] formed after nucleophilic attack by water is readily oxidized by oxygen, ultimately generating more potent nucleophiles, namely OH⁻ and OH· (Figure 15b).^{[ 102 ]} This suggests that this degradation mechanism is likely irreversible and self‐accelerating. Similarly, due to the strong nucleophilicity of OH⁻, ferrocene derivatives are generally not employed as electrolyte materials in alkaline systems.

Figure 15.

a) proposed mechanism explaining the degradation of ferrocene catholytes in water. b) Possible Reactions of By‐Products from Ferrocene Degradation. Reproduced with permission.^{[ 44 ]} Copyright 2021, Wiley.

Luo et al. demonstrated the mechanisms of thermal‐induced and photo‐induced dissociation of the Cp ligand (Figure 16 ).^{[ 103 ]} When the substituent exhibits an excessively strong electron‐withdrawing characteristic, it weakens the coordination interaction between the Fe center and the Cp ligand, readily leading to Cp dissociation or substitution by nucleophiles. Therefore, compared to FcNCl, the BTAMP‐Fc derivative designed by Beh et al., which features a 3‐(trimethylammonium)propyl substituent, more effectively suppresses ligand dissociation.^{[ 31 ]} However, this modification also results in a lower redox potential for BTAMP‐Fc compared to FcNCl. Thus, a trade‐off exists between the redox potential and the stability of the complex.

Figure 16.

Proposed thermal degradation mechanism of [C ₁‐FcN]²⁺, the charged state of FcNCl. Reproduced with permission.^{[ 103 ]} Copyright 2022, Royal Society of Chemistry.

Liu et al. designed an imidazolium‐decorated ferrocene derivative, BMIP‐Fc.^{[ 45 ]} By leveraging the H atoms on the imidazolium ring, which can interact with H₂O via hydrogen bonding, the dissociation of the Cp ligand is effectively suppressed (Figure 17 ). This design highlights the importance of hydrogen‐bond interactions in aqueous environments and suggests that strategically utilizing such interactions to inhibit molecular degradation may become a crucial design principle in the future. Some representative ferrocene‐based compounds are selected and their structures and properties are presented in Figure 18 and Table 4 .

Figure 17.

The proposed mechanism illustrates the beneficial effects of hydrogen bonds from the imidazole pendant in preventing H₂O from attacking BMIP‐Fc⁺ during cycling and IRI showing the surface map of BMIP‐Fc. Reproduced with permission.^{[ 45 ]} Copyright 2025, American Chemical Society.

Figure 18.

Structures of a subset of ferrocene derivatives and their solubilities utilized in AORFBs.

Table 4.

The redox potential, pH and cycling stability of a subset of ferrocene derivatives in AORFBs.

Ferrocene derivatives	anolyte	E1/2 [vs. SHE] a)	pH b)	Capacity fade rate [demonstrated concentration]	Refs.
FcNCl	MV	0.61 V	7	0.04%/cycle (0.7 m)	[43]
BTMAP‐Fc	BTMAP‐Vi	0.39 V	7	0.1%/day (1.3 m)	[31]
BOH‐Fc	BTMAP‐Vi	0.28 V	7	0.168%/day (1.5 m)	[44]
BMIP‐Fc	Dex‐Vi	0.38 V	7	Without capacity fading (0.5 m)	[45]

^a) Partial values of E_1/2 are converted to the values versus SHE. (E_Ag/AgCl = 0.2 V, E_NHE = E_SHE = 0 V);

^b) If it is not explicitly stated whether the electrolyte contains substances that may alter the pH, the pH is assumed to be 7 by default. Although actual pH levels may fluctuate due to issues such as dissolution of CO₂.

2.6. Viologen‐Based Electroactive Materials

Viologens (N,N′‐disubstituted bipyridinium derivatives) store and release electrical charges through pyridine rings. This enables a two‐electron redox process accompanied by distinct color changes. Combined with their relatively low reduction potential and high solubility in neutral aqueous electrolytes, they have become one of the most promising anode materials for pH‐neutral AORFBs. When employed as negative electrodes, capacity degradation at the anode side primarily results from molecular crossover through membrane and decomposition side reactions. Represented by the simplest viologen methyl viologen (MV),^{[ 30 ]} viologens typically possess cationic charges, making crossover relatively difficult. Furthermore, this phenomenon can also be mitigated by increasing molecular volume and enhancing molecular charge density. Therefore, decomposition side reactions of viologens constitute the primary contributor to capacity fade, outweighing crossover‐induced losses in most viologen‐based AORFBs.

In aqueous solutions, the electron transfer between MV²⁺ and MV⁺ demonstrates high reversibility. However, further reduction from MV⁺ to MV⁰ species presents critical challenges: The uncharged MV⁰ exhibits significantly reduced aqueous solubility, potentially leading to precipitation and consequent capacity loss through active material deposition.^{[ 29} , ^{104 ]} This solubility limitation fundamentally constrains achievable energy densities. It should be additionally noted that MV⁺ undergoes spontaneous dimerization via π–π stacking interactions, followed by disproportionation into MV²⁺ and MV⁰, thereby generating additional capacity fade (Figure 19 ).^{[ 104} , ^{105 ]} This side reaction exhibits strong concentration dependence. Such inherent chemical instability creates a fundamental challenge for implementing high‐concentration MV²⁺ electrolytes, as the increased active material loading inevitably exacerbates these degradation pathways. There are also reports that σ‐dimers can also form between MV⁺ species. Although this side reaction may not be predominant in static solutions, its occurrence during battery cycling should be given attention (Figure 19).^{[ 106 ]}

Figure 19.

Side Reactions of viologen derivatives.

Although MV can theoretically regenerate MV⁺ through protonation reactions,^{[ 107 ]} this process induces undesirable pH elevation in the electrolyte system (Figure 19). As illustrated in relevant scholarship, increased OH⁻ concentration facilitates nucleophilic attack on MV species, resulting in the detachment of substituents from the N atoms of bipyridines. This is an essential degradation mechanism that adversely affects both solubility and electrochemical activity of MV compounds (Figure 19).^{[ 104 ]} Notably, oxygen infiltration into the aqueous electrolyte initiates two parallel degradation pathways: (1) oxidation of MV⁺ to MV²⁺,^{[ 108} , ^{109 ]} which diminishes the anolyte's reducing capacity, and (2) generation of additional OH⁻ that exacerbate MV species decomposition through the aforementioned mechanism (Figure 19). These two processes collectively contribute to progressive capacity loss through both chemical and electrochemical degradation pathways. The application of viologens in aqueous alkaline systems necessitates cautious implementation, requiring maintenance of low‐oxygen operational environments. Research into oxygen‐tolerant viologens should also be explored.

The introduction of quaternary ammonium groups (with positive charge) into the side chains of viologen derivatives can effectively suppress intermolecular interactions and thus enhance solubility. Beh et al. integrated both as a viologen (BTMAP‐Vi) for AORFBs, featuring elongated side chains terminated with quaternary ammonium moieties.^{[ 31 ]} Its architecture synergistically employs steric hindrance and intensified coulombic repulsion to suppress dimerization of mono‐electron‐reduced radicals. Enhanced cationic density compared to methyl viologen improves aqueous solubility and mitigates transmembrane crossover through intensified electrostatic confinement. The BTMAP‐Vi||BTMAP‐Fc AORFB configuration established high cycling stability, achieving 99.9943% capacity retention per cycle over 250 cycles (Figure 20 ).

Figure 20.

a) Evolution of the capacity of the BTMAP‐Vi/BTMAP‐Fc RFB at a concentration of 1.3 m/1.3 m during extended cell cycling at 50 mA cm⁻². b) Representative voltage vs time traces of selected cycles. Reproduced with permission.^{[ 31 ]} Copyright 2017, American Chemical Society.

Incorporating charged functional groups into the side chains has become the dominant approach in viologen design. However, this does not represent the sole strategy available. Yao et al. synthesized a series of oligo(ethylene glycol) (OEG) substituted viologen derivatives (Vi‐OEGn, n = 2, 3, 4) to address solubility and stability challenges.^{[ 56 ]} The hydrophilic OEG side chains not only enhanced the stability of viologens but also suppressed permeation and dimerization of the singly reduced radical species (Vi⁺). The optimized OEG side chains improved the solubility of the doubly reduced state (Vi⁰) and facilitated reversible two‐electron redox behavior. The Vi‐OEG3||FcNEBr battery demonstrated exceptional cycling stability with a capacity fade rate of 0.00253%/day over 138 cycles.

Carrington et al. suggested that rapid capacity fade might fundamentally correlate with elevated radical concentrations during electrochemical cycling.^{[ 109 ]} Exhibiting either strong intermolecular association between reduced radical species or closed‐shell structure through intramolecular stabilization proves critical. This molecular engineering approach effectively minimizes deleterious side reactions such as oxidation pathways, thereby achieving significant mitigation of capacity fade. Geraskina et al. indicated that when side chains carried no charge, the dimerization between radicals was particularly pronounced and shows little variation with alterations in the chain length.^{[ 106 ]} This suggests that the performance enhancement of Vi‐OEG3 over MV requires further rigorous investigation. A hypothesis is that the polymerized radicals can re‐engage in reactions rather than undergoing disproportionation. If validated, this would indicate that radical dimerization is not the fundamental cause of capacity loss. Instead, potentially side reactions triggered by dimerization may be responsible. Consequently, beyond suppressing radical dimerization, inhibiting side reactions among dimerized radicals could also emerge as a critical design strategy.

Steric hindrance modulation is a common molecular design strategy. Properly designed steric hindrance can control molecular configurations or reducing the probability of being attacked by other molecules in the environment. DeBruler et al. developed a sulfonic acid‐functionalized viologen derivative, (SPr)₂V, leveraging coulombic repulsion forces between negatively charged sulfonate groups to inhibit dimerization.^{[ 110 ]} However, the zwitterionic nature of (SPr)₂V induces an intramolecular electrostatic attraction between sulfonate anions and the positively charged bipyridinium cores, enabling undesirable transmembrane shuttling. In response to this limitation, Li et al. developed a viologen analogue (R‐Vi) through 2,2’,6,6’‐tetramethyl‐4,4’‐bipyridine functionalization.^{[ 111 ]} The methyl groups enforce rigid rod‐like molecular conformations via steric hindrance, simultaneously expanding molecular volume and spatially separating charge distributions, thereby effectively mitigating crossover phenomena. Wang et al. synthesized a 4,4’‐dimethylpiperidinium‐substituted viologen (DBPPy).^{[ 112 ]} This molecular architecture employs the bulky piperidinium substituents to create a protective steric shield against OH⁻ attack. Electrochemical characterization revealed remarkable pH tolerance. When employed in a DBPPyCl₄||TPABPyCl₃ redox flow battery, the system achieved 99.99% capacity retention per cycle at 60 mA cm⁻².

Structural modification through spacer insertion between pyridinium rings represents a promising strategy for enhancing viologen derivatives’ electrochemical performance. Hu et al. synthesized a benzene‐bridged viologen derivative (APBPy).^{[ 57 ]} Unlike other conventional viologen derivatives that undergo a two‐step single‐electron transfer redox process, APBPy could achieve a one‐step two‐electron transfer redox process with a low potential of −0.76 V (vs. SHE). In the reduced state, molecular planarization occurs through enhanced conjugation, effectively redistributing electron density and improving the stability. In the study by Carrington et al., this mechanism was further elucidated. Specifically, APBPy possesses wide singlet–triplet energy gaps, facilitating intramolecular electron pairing, thereby reducing the concentration of free radicals within the system and suppressing the occurrence of side reactions. Liu et al. engineered a meta‐substituted thienoviologen (NPr)₂SV.^{[ 58 ]} This structural evolution strengthens intermolecular π–π stacking interactions, stabilizing the [(NPr)₂SV]^3+• dimeric state over monomeric radical species. Such stable dimer effectively inhibits oxygen‐mediated radical oxidation, enabling a 0.5 m (NPr)₂SV ‐based battery to achieve 200 cycles with minimal daily capacity fade (0.07%). Compared to conventional bipyridine derivatives, the development of synthesis methods compatible with industrial production represents a crucial research direction for π‐extended bipyridine derivatives, alongside their overall performance enhancement. Some representative viologen‐based compounds are selected and their structures and properties are presented in Figure 21 and Table 5 .

Figure 21.

Structures of a subset of viologen derivatives and their solubilities utilized in AORFBs. (The solubility of R‐Vi was measured in a 1 m NaCl solution, while the solubilities of other molecules were determined in pure water. Under these predefined conditions, the pH is assumed to be 7, though actual pH values may vary).

Table 5.

The redox potential, pH and cycling stability of a subset of viologen derivatives in AORFBs. (If it is not explicitly stated whether the electrolyte contains substances that may alter the pH, the pH is assumed to be 7 by default. Although actual pH levels may fluctuate due to factors such as dissolution of CO₂).

Viologen derivatives	Catholyte	E1/2 [vs. SHE] a)	pH	Capacity fade rate [demonstrated concentration]	Refs.
MV	4‐OH‐TEMPO	−0.45 V	7	0.11%/cycle (0.5 m)	[30]
Vi‐OEG3	FcNEBr	−0.305 V	7	0.00253%/day (0.5 m)	[56]
PEG2‐BPy‐PEG2	BTMAPr‐Fc	−0.41 V	7	1.2%/day (1.95 m)	[113]
(SPr)2V	KI	−0.43 V	7	0.01%/cycle (0.5 m)	[110]
R‐Vi	K4Fe(CN)6	−0.55 V	7	1.11%/day (0.5 m)	[111]
DBPPy	TPABPy	−0.32 V	7	1.5%/day (1.0 m)	[112]
BTMAP‐Vi	BTMAP‐Fc	−0.358 V	7	0.1%/day (1.3 m)	[31]
Dex‐Vi	BTMAP‐Fc	−0.322 V	7	No discernible decay (1.5 V)	[114]
APBPy	TBABPy	−0.763 V	7	2.88/day (0.5 m)	[57]
(NPr)2SV	FcNCl	−0.29 V	7	0.0966%/cycle (1.5 m)	[58]
ATBPy	TPABPy	−0.55 V	7	0.445%/day (1.0 m)	[115]
3,4‐S2V	(NH4)4Fe(CN)6	−0.41 V	7	0.045%/day (1.1 m)	[116]
BPP‐Vi	K4Fe(CN)6	−0.46 V	9	0.016%/day (1.0 m)	[117]
MBPE‐Vi	Na4Fe(CN)6	−0.503 V	13.5	0.229%/day (0.5 m)	[118]
PPBPy	PSS‐TEMPO	−0.801 V	7	Similarly zero capacity fading (0.45 v)	[119]
(CBu)2V	(NH4)4Fe(CN)6	−0.43 V	9	Did not show capacity loss (0.9 m)	[120]
(SPr)34TpyTz	K4Fe(CN)6	−0.27, ‐0.42, and −0.62 V	7	0.44%/day (0.2 m)	[121]

^a) Partial values of E_1/2 are converted to the values versus SHE. (E_Ag/AgCl = 0.2 V, E_SCE = 0.245 V, E_NHE = E_SHE = 0 V).

3. The Design of Ion‐Selective Membranes

In addition to capacity fading caused by electrolyte decomposition, preventing cross‐membrane permeation between different electrolytes is another critical factor for maintaining long‐cycle stability of AORFBs^{[ 122} , ^{123 ]} Therefore, an ideal ion‐selective membrane (ISM) that can block redox‐active molecules while enabling rapid conduction of charge carriers is the key to addressing this issue^{[ 36} , ^{124 ]} As one of the key components of AORFBs, ISMs could conduct specific ions (e.g., H⁺, K⁺, Cl⁻, OH⁻) and prevent the permeation of redox‐active species between anolyte and catholyte^{[ 125} , ^{126 ]} These dual functionalities critically influence the key performance metrics of AORFBs, including coulombic efficiency (CE), voltage efficiency (VE), energy efficiency (EE), capacity retention, and cycling stability^{[ 127 ]} The development of ideal ISMs requires the concurrent achievement of high ionic conductivity, outstanding selectivity, mechanical durability, chemical stability, and cost‐effectiveness^{[ 128} , ^{129 ]} For ISMs used in AORFBs, the high ion‐selective transport property is crucial for enhancing battery performance^{[ 130 ]} The ion selectivity of the membrane is determined by the size sieving effect and electrostatic repulsion between ion transport channels and nanoscale redox‐active molecules^{[ 131 ]} Therefore, the design of ion transport channels is key to improving membrane performance. Unfortunately, in AORFBs applications, a fundamental trade‐off between ionic conductivity and selectivity arises from the intrinsic ion transport mechanisms within membrane.^{[ 76} , ^{123 ]} The following table showcases several common membranes along with their critical properties (ion conductivity and permeability) (Table 6 ).

Table 6.

Several common membranes for aqueous organic redox flow batteries and their key metrics (ion conductivity and permeability).

Mem.	Ion conductivity [mS cm−1]	Permeability [Fe(CN)6 4−, cm2 s−1]	Refs.
Nafion	6.61–11 (in 1 m KOH or NaOH, 30 °C)	2.2–46 × 10−9 (pH=14)	[132, 133, 134]
PIM	7.8–15 (in 1 M KOH, 30 °C)	1.0–16.8 × 10−10 (pH=14)	[132, 133]
sPEEK	14.4–43 (in 0.5 m NaCl)	1.6–15 × 10−9 (3 m NaOH)	[135]
PBI	3.03 (in 3 m NaOH)	8.19 × 10−6 (in 3m NaOH)	[136]

ISMs typically consist of hydrophobic polymer backbone, functional charged hydrophilic groups, and movable counterions.^{[ 137} , ¹³⁸ , ¹³⁹ , ¹⁴⁰ , ¹⁴¹ , ^{142 ]} Depending on the type of functional charged hydrophilic groups, ISMs are broadly classified into cation‐exchange membranes (CEMs)^{[ 143} , ^{144 ]} which contain cation‐conducting groups (e.g. sulfonic acid, phosphoric acid and carboxylic acid groups), and anion‐exchange membranes (AEMs)^{[ 145} , ^{146 ]} which carry anion‐conducting groups (e.g. quaternary ammonium, imidazole, and guanidinium cation groups). Since CEMs could selectively transport ions such as K⁺, NH₄ ⁺, or Na⁺, while AEMs could allow the selective passage of ions like Cl⁻, the appropriate ion‐selective membrane is often selected based on the properties of the electrolyte and supporting electrolyte when assembling AORFBs.

Perfluorosulfonic acid membranes (such as DuPont's Nafion® series) are the most representative commercial ISMs, featuring microphase separation morphologies, and are widely used in AORFB systems.^{[ 147 ]} Due to the characteristics of their flexible fluorinated backbones combined with polymeric sulfonated side‐chains, perfluorosulfonic acid membranes form unique microphase separation structures.^{[ 134 ]} Currently, commercial membranes represented by microphase‐separated membranes used in AORFBs include CEMs such as Nafion (117, 212, and so on) and Fumasep® E620 membranes, as well as AEMs such as Selemion (AMV or AME series) membranes.^{[ 27} , ^{90 ]} A typical feature of microphase‐separated membranes is the inherent randomness of their microphase‐separated structures. This randomness arises during membrane formation, as the microphase‐separated morphologies result from the aggregation of charged groups along molecular chains, a process that occurs stochastically.^{[ 148 ]} Consequently, in these membranes, achieving precise control over ion transport channels induced by microphase separation remains challenging. For instance, in aqueous environments, Nafion membranes develop hydrophilic domains with widths ranging from 1.8 to 3.5 nm, occupying ≈20% of the total volume.^{[ 132} , ^{133 ]}

Another critical issue for microphase‐separated membranes is the effective control of swelling.^{[ 151 ]} These membranes depend on the uptake of water molecules into the polymer chains to generate ion transport channels of suitable size for ion conduction. Although a high swelling could enhance the ion conductivity of the membranes, excessive swelling inevitably undermines its ion selectivity and structural stability.^{[ 136 ]} Xiao et al. constructed a composite structure comprising a hydrophobic fluorinated poly(arylene ether) backbone and hydrophilic piperazine side chains (QPFPAE), which is considered as AEMs with optimized microstructural morphology (Figure 22a).^{[ 149 ]} The QPFPAE membranes exhibit higher water uptake and swelling compared to the DSV membrane. As a result, when applied in neutral 4‐OH‐TEMPO||MV AORFBs, these membranes achieve an exceptional Cl⁻ conductivity of 55.9 mS cm⁻¹, surpassing that of the DSV membrane. However, at the same current density, the CE of AORFBs assembled with different QPFPAE membranes decreases with increasing membrane swelling, and all values remain lower than those of AORFBs assembled with DSV membranes. This research highlights the intrinsic trade‐off between ion conductivity and active species selectivity in ISMs.

Figure 22.

a) The AEM composed of hydrophobic fluorinated poly(arylene ether) backbone with hydrophilic piperazinium branch. Reproduced with permission.^{[ 149 ]} Copyright 2021, Elsevier. b) The schematic representation of the simple hypercrosslinking process to prepare self‐standing HC‐QPPO membrane. Reproduced with permission.^{[ 150 ]} Copyright 2024, Wiley.

Numerous efforts have been devoted to enhancing the ion selectivity conductivity of microphase‐separated membranes. Zhang et al. developed adamantane‐incorporated polyetherketone‐based AEMs exhibiting suppressed swelling ratios (<16%) coupled with significant enhancement in IEC.^{[ 146 ]} Alternatively, Cross‐linking technology has emerged as an effective approach to mitigate membrane swelling. Peng et al. fabricated hyper‐cross‐linked ISMs (HC‐QPPO) by employing QPPO as the matrix and FDA as the cross‐linker via a solution‐casting process (Figure 22b).^{[ 150 ]} The HC‐QPPO membrane exhibits a swelling ratio of merely 1/2.6 relative to pristine QPPO membranes across a temperature range of 30–80 °C, highlighting exceptional anti‐swelling properties. The 0.5 M BTMAP‐Vi||TEMPTMA AROFBs equipped with HC‐QPPO membrane exhibits stable performance under 150 mA cm⁻² current density, sustaining over 1800 charge‐discharge cycles with EE maintained at ∼70.5% and an ultralow capacity decay rate of 0.0017% per cycle.

In AORFBs, organic redox‐active electrolytes typically exhibit molecular sizes on the nanometer‐sized, whereas charge‐carrying ions are usually at the subnanometer‐sized.^{[ 155 ]} To overcome the inherent trade‐off effect of microphase‐separated membranes in balancing ion selectivity and conductivity, the construction of stable subnanometer‐sized ion channels within membranes has been proposed as an effective design strategy. In this context, the development of ISMs based on microporous materials, such as polymers of intrinsic microporosity (PIMs), covalent triazine frameworks (CTFs), polymer aromatic framework (PAFs), covalent organic frameworks (COFs), and metal‐organic frameworks (MOFs), have emerged as a promising research frontier.^{[ 156} , ¹⁵⁷ , ^{158 ]}

To address the dual challenges of membrane processability and subnanometer‐sized microstructure control, PIMs membranes, recognized as a promising solution, have emerged as a research hotspot in recent years.^{[ 159 ]} PIMs represent a class of linear polymers characterized by rigid, contorted molecular chains that stack to form interconnected sub‐nanometer porous networks, a concept first proposed by the McKeown group at the University of Edinburgh.^{[ 160 ]} The inherent microporosity (pore size < 2 nm) arises from the steric hindrance of these rigid macromolecular architectures, which prevents efficient chain packing. Based on synthetic mechanisms, PIMs are primarily categorized into dibenzodioxin‐based, Tröger's base‐derived, polyimide‐type, and polyoxazinthrone architectures.^{[ 152} , ¹⁵⁴ , ¹⁵⁵ , ^{161 ]} The remarkable solubility of PIMs in organic solvents endows them with superior membrane‐forming processability, enabling the fabrication of free‐standing membranes through solution‐casting techniques. These membranes contain internal sub‐nanometer channels filled with electrolyte, providing efficient ion transport pathways for AORFBs.^{[ 162 ]}

In recent years, significant progress has been achieved in designing high‐performance membranes based on PIMs used in AORFBs. In 2016, Yang et al. first reported Tröger's base‐derived PIMs membranes with regularly distributed cationic groups (Figure 23a), demonstrating unique ion transport characteristics through sub‐nanoscale channels (0.8 nm pore size).^{[ 152 ]} Remarkably, these membranes exhibit exceptional OH⁻ conductivity of 164.4 mS cm⁻¹ at a low IEC (0.82 mmol g⁻¹). In cation transport research, Baran et al. developed Aqua‐PIM membranes functionalized with amidoxime groups, which maintain stability under strongly alkaline conditions (pH >13) and enabled rapid selective K⁺ conduction in their deprotonated state (Figure 23b).^{[ 153 ]} The optimized membrane achieved 21.5 mS cm⁻¹ conductivity in 5.0 M KOH solution.

Figure 23.

a) The synthesis of the TB‐Polymer‐based AEM materials. Reproduced with permission.^{[ 152 ]} Copyright 2016, Wiley. b) Aqua‐PIM membranes prepared by the functionalizing of ionizable amidoxime groups within the pores of microporous ladder polymer membranes. Reproduced with permission.^{[ 153 ]} Copyright 2019, Elsevier. c) Macromolecular structures of microporous polymers with a size‐selective ion separation function, and typical structures of hydrophilic microporous PIM polymers and the normalized pore size distribution derived from molecular simulation of AO‐PIM‐1. Reproduced with permission.^{[ 154 ]} Copyright 2020, Springer. d) Schematic of the architecture of hydrated micropores and the chemical structures of cPIMs containing pendant groups with varying hydrophobicity. Reproduced with permission.^{[ 130 ]} Copyright 2024, Springer.

The strategic design of chain rigid contorted architectures constitutes the cornerstone for developing high‐performance PIMs membranes. Ye et al. utilize spiral‐ring(SBI, SBF) and bridged bicyclic(BTrip, DBMP) structural units to precisely regulate the topological configuration and pore architecture of AO‐PIM membranes (Figure 23c).^{[ 154 ]} In K₄Fe(CN)₆||2,6‐DPPAQ AORFBs, AO‐PIM‐1 and AO‐PIM‐DBMP membranes demonstrate exceptional capacity retention rates of 98.08% and 98.78%, respectively, after 1000 cycles at 80 mA cm⁻². The study also reveales that 2,6‐DHAQ adsorption within membrane pores caused channel blockage, leading to progressive increases in ASR during cycling. Notably, through precise modulation of lateral aromatic rings to control local hydrophobicity, Wang et al. achieved sub‐nanometer pore size regulation (Figure 23d).^{[ 130 ]} The engineered c‐PIM‐Ph membrane demonstrate exceptional durability in neutral 2,6‐D2PEAQ||K₄Fe(CN)₆ systems, maintaining a remarkably low daily capacity decay rate of 0.014% over 60‐day cycling.

Scalable manufacturing represents a critical pathway for transitioning PIMs membranes from laboratory prototypes to industrial applications. Song et al. synthesized ether‐free mainchain ultramicroporous membrane via Friedel‐Crafts polymerization, with roll‐to‐roll processing enabling the fabrication of MTCP‐50 membranes exceeding 1000 mm in width (Figure 24 ).^{[ 163 ]} Through precise modulation of meta/para‐terphenyl ratios, these membranes achieved optimized sub‐nanometer pore architecture. In concentrated neutral TEMPTMA||MV systems, a 50 cm² single‐cell configuration demonstrate stable operation for 1140 h at 100 mA cm⁻² current density, achieving 77.8% EE.

Figure 24.

The synthesis procedure and pilot‐scale manufacturing of MTCP‐x AEMs. Reproduced with permission.^{[ 163 ]} Copyright 2023, Springer.

Some studies suggest that the preparation of PIMs requires precise post‐functionalization of linear polymer chains.^{[ 161} , ^{164 ]} Especially, when a high content of charged functional groups is introduced, the linear polymer chains may undergo aging and severe swelling, ultimately leading to a decline in membrane performance. In response to this issue, Zuo et al. innovatively fabricated sulfur‐containing triazine framework membranes (SCTF‐BP) via organo‐sol‐gel synthesis.^{[ 161 ]} These membranes employ an “ion coordination” mechanism within hydrophobic frameworks, coupled with multiple weak interactions, to realize near‐frictionless ion transport (Figure 25 ). The optimized SCTF‐BP membrane exhibits ultra‐low area‐specific resistance (0.17 Ω·cm²) in K₄Fe(CN)₆||DHAQ flow batteries, enabling stable operation at unprecedented current densities up to 500 mA cm⁻².

Figure 25.

A series of freestanding triazine framework membranes with hydrophobic functional side chains. Reproduced with permission.^{[ 161 ]} Copyright 2023, Springer.

Due to the complex synthesis procedures of PIMs, their performance nearing newly defined limits, and the presence of vulnerable sites (e.g., nitrile groups) in sulfonated PIM polymers that undergo hydrolysis under alkaline AORFB conditions, the design of novel ISMs that combine the high free volume of intrinsic microporous materials with chemically robust polymer backbones, such as poly(ether ketone), poly(ether ether ketone), and polybenzimidazoles, has emerged as a promising research hotspot for AORFB applications.^{[ 165} , ¹⁶⁶ , ^{167 ]} Wong et al. started from polyether ether ketone (PEEK) with high chemical stability and prepared intrinsic microporous sulfonated PEEK (sPEEK) membranes with high ionic conductivity by fully replacing the phenyl groups in the PEEK backbone with twisted triptycene structures, and employing a post‐sulfonation method using silyl‐protected sulfonating agent (Figure 26a).^{[ 135 ]} The sulfonate groups allow rapid exchange of K⁺ through electrostatic interactions, while the formation of continuous hydrogen‐bonding networks also facilitate the transport of OH⁻ through the Grotthuss (hopping) mechanism, enabling dual ion transport in alkaline electrolytes. In K₄[Fe(CN)₆]||2,6‐DHAQ AORFBs assembled with sPEEK membrane exhibited a peak power density of 560 mW cm⁻² at 100% SOC, and during long‐term cycling at 400 mA cm⁻², the energy efficiency reached 71%.

Figure 26.

a) Chemical structure of sPEEK‐Trip containing triptycene in the backbones with internal free volume. Reproduced with permission.^{[ 135 ]} Copyright 2025, Elsevier. b) Fabrication route and structure of the 3‐µm‐thick polymeric membrane. Reproduced with permission.^{[ 136 ]} Copyright 2025, Springer.

In addition, tailoring the charge properties of functional groups at selective sites within the membrane and enhancing the transport resistance to redox‐active species of the same charge via the Donnan effect represents an effective strategy to improve ion selectivity.^{[ 168} , ¹⁶⁹ , ^{170 ]} Li et al. realized crosslinking reactions between polymers and crosslinking agents at the interface of two immiscible solvents, leading to the formation of a nanoscale separation layer on the surface of polybenzimidazole membranes (PT membrane, Figure 26b).^{[ 136 ]} This layer exhibits a quasi‐ordered crosslinked network structure with a narrow pore size distribution. Based on the Donnan exclusion effect, the positively or negatively charged ion‐selective sites within membrane, hinder the crossover of redox‐active species across a wide pH range while allowing efficient passage of charge carriers. When applied in aqueous organic vanadium‐toluidine blue flow batteries (VTFBs), the PT membrane demonstrated the higher VE (88.51%) and EE (86.41%) than Nafion 212 membrane at 200 mA cm⁻². Furthermore, both VE and EE of VTFBs equipped with the PT membrane remained high and stable over 1000 charge–discharge cycles. The performance of AORFBs employing some membranes is summarized in Table 7 .

Table 7.

Physicochemical parameters and battery performance of some membranes in AORFBs.

Mem.	Catholyte/anolyte	Permeability [cm2 s−1]	Area Resistance [Ω cm2]	EE [%]	Decay rate [%]	Refs.
Nafion 212	2,6‐D2PEAQ/ Fe(CN)6 3‐/4−	2.2 × 10−8 (Fe(CN)6 4−)	1.5	≈50% @100 mA cm2	1.36 (per day)	[130]
Selemion AMV	MV2+/ 4‐HO‐TEMPO	0.21 × 10−9 (MV2+)	≈3.2	37.4% @160 mA cm2	0.11 (per cycle)	[149]
QPFPAE −100	MV2+/ 4‐HO‐TEMPO	8.63 × 10−9 (MV2+)	≈1.1	71.4% @160 mA cm2	0.16 (per cycle)	[149]
HC‐QPPO	BTMAP‐Vi/ TMAP‐TEMPO	7.54 × 10−10 BTMAP‐Vi	0.33	70.5% @150 mA cm2	0.0017 (per cycle)	[150]
AO‐PIM‐1	2,6‐DHAQ/ Fe(CN)6 3‐/4−	3.1 × 10−9 (Fe(CN)6 4−)	0.39	84.6% @80 mA cm2	0.5 (per day)	[171]
AO‐PIM‐DBMP	2,6‐DHAQ/ Fe(CN)6 3‐/4−	1.3 × 10−9 (Fe(CN)6 4−)	0.43	80.8% @80 mA cm2	0.05 (per day)	[171]
cPIM‐Ph	2,6‐D2PEAQ/ Fe(CN)6 3‐/4−	4.0 × 10−13 (Fe(CN)6 4−)	≈1.25	≈70% @100 mA cm2	0.014 (per day)	[130]
SCTF‐BP	2,6‐DHAQ/ Fe(CN)6 3‐/4−	9.35 × 10−11 (Fe(CN)6 4−)	0.17	50.4% @500 mA cm2	/	[161]
sPEEK ‐Trips‐1.55	2,6‐DHAQ/ Fe(CN)6 3‐/4−	1.88 × 10−10 (Fe(CN)6 4−)	0.59	71% @400 mA cm2	/	[135]
PT@1min	TB/V2+	3.72 × 10−13 (Fe(CN)6 4−)	0.009 (in 3M H2SO4)	81.14% @120 mA cm2	/	[136]

4. Conclusions and Prospects

AORFBs exhibit strong promises for large‐scale energy storage applications owing to their intrinsic safety, cost‐effectiveness, and environmental compatibility. Although significant advances have been made in the development of functional materials in recent years, most research efforts remain at the laboratory scale, with limited progress toward practical engineering applications. Therefore, it is imperative to systematically overcome key performance bottlenecks, such as capacity and cycling stability, through the integrated design of electrolytes and membranes to meet the growing performance demands of next‐generation AORFBs. In parallel, the economic viability and scalability of energy storage systems must be optimized in concert with material innovation.

Currently, commercial flow batteries still primarily utilize inorganic redox species. For AORFBs, to achieve commercialization and enhance market competitiveness, a key challenge lies in identifying long‐lived electroactive redox molecules. We summarize recent research progress on quinones, phenazines, viologens, and TEMPO derivatives, focusing on the degradation mechanisms of these molecules. The main findings are as follows:

• For quinone compounds: 1) The reduced form of anthraquinone is prone to disproportionation, yielding anthrone compounds. These anthrone compounds further undergo oxidative coupling to form anthrone dimers. 2) The reduced form of naphthoquinone tends to undergo tautomerization (enol‐keto tautomerism), generating electrochemically inactive keto isomers. 3) Nucleophilic group attacks lead to the introduction of new groups (e.g., hydroxyl, bromo) or the departure of existing groups (e.g., sulfonate).

• For phenazine compounds: 1) The reduced form of phenazine has a tendency to undergo tautomeric side reactions, transforming to redox‐inactive isomers. (2) Side chains, particularly connected to the phenazine core via C─O or C─N bonds, can readily cleave from the parent molecule.

• For viologen compounds: 1) The single‐electron reduced form of viologen undergo irreversible dimerization, making it difficult to dissociate and participate in oxidation reactions or undergoing disproportionation reactions to produce poorly soluble two‐electron reduced viologen. 2) Nucleophilic group attack leads to cleavage of the side chain connected in the pyridinium nitrogen, resulting in loss of redox activity.

• For TEMPO‐type compounds: 1) Nucleophilic group attack can trigger ring‐opening reactions of TEMPO or reduction of the oxidized TEMPO⁺ species with the generation of destructive OH·. 2) Both over‐reduction of TEMPO and disproportionation between TEMPO molecules can yield redox‐inactive hydroxylamine derivatives. 3) Hydrogen bonding between such hydroxylamine derivatives and TEMPO molecules impedes the normal participation of TEMPO in electrochemical redox reactions.

As evidenced, the possible side reactions of these organic molecules primarily include nucleophilic attack, tautomerization, and intermolecular disproportionation. To address these issues, molecular engineering strategies, such as changing substituents, designing conjugated structures, utilizing Coulombic interactions and hydrogen bonding, and developing stable radical or non‐radical systems, can be employed to mitigate side reactions of organic electroactive materials in AORFBs at the fundamental level. To enable more targeted design, it is also necessary to further explore and integrate systematic molecular degradation mechanisms. However, factors such as atom economy in synthesis, environmental impact of waste solutions, as well as post‐treatment and recycling issues must also be carefully considered from the initial stages of molecular design. Furthermore, only by addressing these issues and establishing appropriate environmental assessment methods can AORFBs truly be regarded as environmentally friendly.

In addition to the decomposition of organic molecules, irreversible capacity decay in batteries can also be caused by the retention of electrolyte on the walls of storage tanks, crossover of active species across the membrane and the leakage of electrolyte. For batteries with initially balanced anode and cathode capacities, excessive decomposition of molecules on one side can lead to overall capacity loss. These phenomena must not be overlooked when investigating molecular degradation mechanisms. To exclude the influence of molecular crossover, conducting symmetric cell cycling is essential.^{[ 27} , ^{172 ]} Such cyclings may even be more informative than full‐cell configurations in revealing the mechanisms of molecular side reactions.

The temporal decay rate is one of the key metrics currently used to evaluate capacity fade of batteries. Factors such as molecular concentration, pH changes during cycling, cross‐membrane water migration, and interactions between supporting electrolytes and organic molecules may all influence the temporal decay rate. The potential correlation between these factors and side reactions of organic molecules remains poorly understood.

Historically, ex‐situ chemical analysis techniques have played a significant role in elucidating capacity decay mechanisms. However, factors introduced during sample handling (such as oxygen‐induced oxidation of the electrode) may lead to discrepancies between analytical results and actual conditions. Moreover, an increasing number of molecules with low decay rates are designed. During battery cycling, the electrolyte containing these molecules exhibits minimal variations in composition. Even minor disturbances can significantly affect measurement outcomes. Therefore, it is essential to develop highly sensitive in‐situ analytical techniques, such as in‐situ spectroscopy, in‐situ EPR, and in‐situ NMR.^{[ 40 ]} Molecular dynamics (MD) simulations and density functional theory (DFT) calculations can also be effective in molecule screening and prediction of decomposition pathways.

TEMPO‐based compounds and ferrocene derivatives are among the most commonly used cathode materials, though they are primarily applied in pH‐neutral environments. Some quinone‐based molecules can serve as cathode materials under non‐neutral conditions, but they are highly susceptible to decomposition. Currently, the redox pair K₄[Fe(CN)₆]/K₃[Fe(CN)₆] is widely employed as the cathode under alkaline conditions. The development of highly stable organic cathode materials for non‐neutral environments remains an ongoing challenge. Although organic anode materials have been studied more extensively, very few oxygen‐tolerant anode materials are currently available. Developing oxygen‐resistant anodes or designing feasible oxygen balance strategies would greatly facilitate the practical application of AORFBs.

Ideal AORFBs should possess not only long‐lived cycling life but also high energy density. However, the solubility of organic molecules in aqueous electrolytes is limited. At high concentrations, side reactions during electrochemical processes may become more severe, and the viscosity of the electrolyte increases.^{[ 18} , ^{173 ]} Designing cathode molecules with higher redox potentials and anode molecules with lower redox potentials can increase the battery voltage, thereby enhancing energy density. However, higher voltage may trigger water splitting side reactions and accelerate molecular degradation. Therefore, it is essential to balance the redox potential, solubility, and decomposition rate of the molecules. Another feasible approach to improving energy density is to design organic molecules capable of transferring multiple electrons. For example, although viologens can undergo two‐electron reduction, their two‐electron reduced state often suffers from poor stability and low solubility in water. Enabling full utilization of their two‐electron storage capability while maintaining low decay rates could significantly enhance energy density. Furthermore, developing flow field and electrode designs better suited to organic molecules will also contribute to improving the performance of AORFBs.

In laboratory research, cost considerations are often overlooked. In industrial settings, challenges such as the expenses associated with large‐scale synthesis of molecules, electrolyte storage and flow through piping, thermal management of battery stacks, and equipment maintenance due to issues like electrolyte corrosion cannot be ignored. Some researchers have already begun to analyze these aspects.^{[ 33 ]} There is no doubt that AORFBs must be evaluated as integrated systems, where the various abovementioned costs need to be weighed against economic losses caused by electrolyte replacement (due to molecular degradation) and capacity fade. A low‐cost but short‐lived molecule is not necessarily preferable to a more expensive yet highly stable one, and vice versa. Ultimately, the priority should be to design molecules that offer both economic viability and excellent electrochemical performance.^{[ 174 ]} In this regard, electrochemical synthesis may present a promising alternative. Several researchers have already demonstrated the integration of electrochemical synthesis with AORFB systems,^{[ 175 ]} which will undoubtedly encourage further exploration into this broad and promising field.

It should be noted that current research and test methods for materials used in AORFBs lack uniformity, which hinders the advancement of AORFBs technology. For instance, studies on the solubility, potential, and stability of redox materials exhibit significant inconsistencies. The solubility of these molecules varies considerably depending on environmental temperature, pH, and the presence of additives.^{[ 176} , ^{177 ]} However, most studies only measure solubility under specific conditions according to their self‐assembled AORFBs. But the dissolution behavior and solubility levels of these materials in other solvent environments are still unknown. This obviously poses challenges for researchers seeking to further utilize or investigate these molecules. A similar issue arises in the testing of ion‐exchange membranes for AORFBs, such as the measurement of permeability coefficients for certain molecules across membranes. Even for the same electrolyte molecule, tested permeability can vary by orders of magnitude under different pH conditions or with different additives (e.g., varying salts or salt concentrations).^{[ 130} , ¹⁵⁴ , ^{178 ]} Regarding battery testing methodologies, the galvanostatic method is currently the most widely adopted. However, this approach is highly susceptible to influences from membrane resistance and temperature.^{[ 34 ]} Consequently, whether it is necessary to switch to potentiostatic tests or incorporate potentiostatic methods at the end of galvanostatic tests remains an open question without consensus. Therefore, standardizing the research and testing conditions for materials used in AORFBs could have profound implications for the development of this technology.

To date, the selection of low‐cost, high‐performance membrane materials for AORFBs remains limited. Although commercial ISMs are relatively mature, the persistent “trade‐off” effect between ionic conductivity and selectivity of the membranes remains unresolved.^{[ 131 ]} This trade‐off arises from the ill‐defined ion channel structures caused by random microphase separation in amorphous polymers during membrane formation.^{[ 148 ]} In contrast, membranes fabricated from high free‐volume polymers, such as PIMs, possess intrinsic micropores that can serve as ion channels. By leveraging the size‐sieving effect of these channels, the ion selectivity of the membranes can be significantly enhanced. However, both PIMs and emerging CTF membranes typically require costly monomers or complex preparation processes, which hinders their industrial application in AORFBs.

Another critical factor for ensuring the high stability of AORFBs is the compatibility between membranes and organic electrolytes. To enhance ion selective conductivity, it is essential to tailor ion transport channel sizes to match the dimensions of different redox‐active molecules. Nevertheless, weak interactions between redox‐active species and polymer membranes, such as electrostatic attraction, hydrogen bonding, dipole–dipole interactions, and π–π stacking, may cause undesired adsorption and accumulation of redox‐active molecules within the membrane or at the membrane/electrolyte interface.^{[ 130} , ¹⁶¹ , ¹⁶⁴ , ¹⁷¹ , ^{178 ]} Such excessive adsorption not only impedes ion transport efficiency but contaminates the membrane and degrades battery performance. Hence, these interactions must be carefully considered when designing both electrolyte molecules and ion‐selective membranes.

Long‐term operational stability of membranes under realistic conditions also requires further investigation. At present, most studies conduct cell cycling experiments for only a short duration, typically lasting just a few days or weeks. Meanwhile, the chemical and mechanical stability of membranes in stacked‐cell configurations over extended operating times has not been thoroughly verified. Under prolonged electrolyte immersion and electric field environments, strongly oxidizing electrolyte molecules may attack polymer chains, leading to detachment of functional groups (like some electron‐deficient groups) and gradual degradation of the polymer backbone (electron‐rich aromatic rings).^{[ 125} , ^{179 ]} Such methods ultimately deteriorate membrane functionality or even result in rupture. In addition, the membranes must endure AORFB assembly pressure as well as electrolyte pressure and temperature fluctuations during operation, which further challenge mechanical stability. Therefore, comprehensive evaluation of membrane stability is a prerequisite for advancing AORFBs from laboratory‐scale research to industrial deployment.

Meanwhile, significant progress has been achieved in the design strategies, synthesis methods, and performance optimization of membranes for AORFBs. Nevertheless, the transition from fundamental research to large‐scale commercialization still faces major challenges.^{[ 38} , ¹⁸⁰ , ¹⁸¹ , ¹⁸² , ^{183 ]} Beyond the difficulties in scaling up membrane materials synthesis reactions and implementing roll‐to‐roll (R2R) manufacturing, additional considerations include establishing the relationship between monomer structure and membrane stability, elucidating polymerization mechanisms during scale‐up, and conducting environmental impact and cost‐effectiveness assessments. Addressing these issues will be crucial to realizing the industrial potential of membranes in AORFB technologies. Moreover, an excellent review has provided a detailed discussion on the bottlenecks of membrane commercialization for energy devices, offering valuable guidance for bridging the gap between academic innovation and industrial implementation, and thereby accelerating the commercialization of membrane‐based energy technologies.^{[ 184 ]}

In summary, achieving performance breakthroughs in AORFBs requires multi‐scale collaborative innovation. At the molecular level, high‐throughput computational methods such as DFT, MD, and machine learning should be employed to accelerate the discovery and optimization of functional materials. At the experimental level, the establishment of a comprehensive structure–performance relationship database is essential. At the engineering level, efforts should focus on optimizing material synthesis protocols and the integrated design of battery systems. Ultimately, the successful commercialization of AORFBs hinges on the synergistic matching of transport characteristics among active materials, electrode interfaces, and membranes. Such advancements will enable the transition of AORFBs from laboratory‐scale research to energy storage applications, providing a robust technological foundation for the large‐scale integration of renewable energy sources.

Conflict of Interest

The authors declare no conflict of interest.

Biographies

Xuanyu Xie received his B.E. degree in Chemistry from Sun Yat‐sen University in 2023. He is currently pursuing a Ph.D. degree under the supervision of Prof. Yonggang Wang in Physical Chemistry at Fudan University. His research focuses on developing redox‐active organic molecules for aqueous organic redox flow batteries.

Jiaming Gao received his Bachelor of Science degree in Chemistry from Fudan University in 2024. He is currently pursuing his Ph.D. degree under the supervision of Professor Yonggang Wang. His research primarily focuses on the molecular design of organic electrolytes for aqueous organic redox flow batteries.

Zhaoqi Wang received his B.E. and Ph.D. degrees in the Department of Polymer Science and Engineering from the Dalian University of Technology of China in 2016 and 2022, respectively. He is now conducting postdoctoral research at the Fudan University (with Prof. Yonggang Wang). His research interests focus on electrolytes and membranes of aqueous organic redox flow batteries.

Ruiyang Li received his bachelor's degree from Fudan University in 2025. He is currently pursuing a Ph.D. degree in Fudan University. His research is mainly focused on aqueous redox flow batteries employing metal coordination complexes as redox‐active species.

Taoyi Kong is currently an Electrolyte Engineer at AESC, where he focuses on the development of high‐energy‐density lithium‐ion battery electrolytes. He previously conducted his Ph.D. research on the molecular design and synthesis of phenazine derivatives for aqueous organic redox flow batteries.

Yonggang Wang received his Ph.D. in Physical Chemistry from Fudan University in 2007. From 2007 to 2011, he worked as a Post‐doctoral Research Associate at the National Institute of Advanced Industrial Science and Technology (AIST), Japan. He is now a full professor at Fudan University, China. His research interests include electrochemical functional materials and their application in lithium‐ion batteries, All‐solid‐state batteries, aqueous batteries, Flow Batteries and Supercapacitors.