Xiao Wang,Jingchao Chai,Jianbing“Jimmy”Jiang

Department of Chemistry,University of Cincinnati,P.O.Box 210172,Cincinnati,OH,45221-0172,United States

Keywords:Membrane Redox flow battery Redox mediator Slurry

ABSTRACT The ever-increasing demand for energy has stimulated the development of economical non-fossil fuels.As representative of clean energy,solar and wind have been identified as the most promising energy sources due to their abundance,cost efficiency,and environmental friendliness.The intrinsic intermittent of the clean energy leads to the urgent requirements large-scale energy storage technique.Redox flow batteries(RFBs)are attractive technology due to their independent control over energy and power.Insoluble redox-active flow battery is a new type of electrochemical energy storage technology that disperses redox-active particles in the electrolyte.Compared with traditional flow batteries,insoluble flow batteries have advantages of large energy density and are very promising in the development of large-scale energy storage systems.At present,three types of insoluble flow batteries have been explored:slurry-based flow batteries,metal/slurry hybrid,and redox-mediator-assisted flow batteries.This Review summarizes the research progress of insoluble flow batteries,and analyzes the key challenges from the fundamental research and practical application perspectives.

1.Introduction

The rapidly increasing demanding for sustainable energy prompts the development of efficient and stable energy storage systems,such as lithium-ion battery and flow batteries,which alleviate the intermittent issues of renewable energy from wind and solar[1-6].Redox flow battery(RFB)is a promising technology due to its advantages in safety,cost,and long service life[7-10].Compared with conventional solid-state batteries wherein redox-active materials are confined in solid electrodes,the active materials in RFBs are stored in external reservoirs and circulate through compartments for energy storage and release.This unconventional construction endows the feature of decoupled energy and power,as well as superior scalability[11-15].The basic composition of the flow battery includes anolyte and catholyte reservoirs,electrodes,separators,and battery casings.Driven by peristaltic pumps,the anolyte and catholyte circulate between the reservoirs and the two half-cell compartments.The oxidation and reduction half reactions occur on the electrodes.

The earliest flow battery concept was proposed by Thaller in 1974[16].National Aeronautics and Space Administration,U.S.A.(NASA)also developed flow batteries using Fe/Cr electrolytes[17].In the following years,inorganic redox batteries developed rapidly,including Cr(II)/-Cr(III)redox couple,Ti(III)/Ti(IV)couple,Zn-Br couple,Sn2+/Sn4+couple[16].In 1984,Maria Skyllas-Kazacos et al.at the University of New South Wales utilized vanadium ions at different valence states in an all-vanadium flow battery.Since the early 1990s,industrial companies,led by Sumitomo Electric(SEI)and Kashima-kita Electric Power Company,have conducted in-depth exploration of key materials and engineering design issues of stacks,and gradually developed all-vanadium flow battery systems for commercialization.Organic redox flow batteries made considerable progress in the past decade as well.Organic compounds such as anthraquinone,viologen,2,2,6,6-tetramethylpiperidinyl-N-oxyl(TEMPO)are widely used in redox flow batteries[5,18,19].

The redox-active materials are critical components in RFBs.From the perspective of energy density,the most important characteristics are redox potential and solubility[14,20].Despite the remarkable progress made in the past decades,the large-scale application of RFBs are still limited by the low solubility of redox materials.Although the solubility of some system reached 5.2 M[21],the typically solubility 1-2 M for aqueous system and 0.1-0.5 M for non-aqueous system was far from enough to meet the energy requirements for flow batteries[22,23].With a battery operating voltage of~1.5 V,an aqueous RFB contributes about 40 Wh/L of energy density under the aforementioned solubility range(1-2 M)[24].The energy density in non-aqueous systems is even lower due to the limited solubility of redox-active materials in organic solvents,even though the battery voltage could be higher because of the wider electrochemical window of organic solvents.

For practical applications,a RFB should possess several features including excellent electrolyte stability,high energy density,and high electrolyte solubility[25].Given the fact that most of the redox materials are insoluble(both in aqueous and nonaqueous solvents)[14,26,27],it is challenging to design redox-active materials that satisfy all requirements.In contrary,directly utilizing insoluble materials in RFBs takes advantage of higher-energy-density of traditional solid-state battery and scalability in solution-based RFB[28,29].To this end,RFBs composed of insoluble redox materials have emerged as attractive way to overcome the solubility limitations in the traditional RFBs and several works have received increasingly attention[28,29].We noticed that most of the reviews focus on soluble redox substances[30,31].Only a few review articles summarize the insoluble RFB yet with a primary focus on inorganic redox materials[28,29].Here,we review three types of insoluble material-based flow battery with a primary focus on energy density and battery.

performance:slurry-based,metal/slurry hybrid,redox mediatorassisted flow batteries(Fig.1).

Fig.1.Schematics of(a)all-slurry flow battery,(b)metal/slurry flow battery,and(c)redox-mediator-assisted flow battery.

2.Slurry flow battery

The components of the slurry battery are almost the same as the traditional solution-based flow battery,namely electrolyte reservoirs,electrodes,separators,and battery framework.The difference is the dispersion state of the active material(from solution to suspension)and the conductive electrode(from the porous carbon felt in solution-based RFBs to carbon paper to avoid filtration and sedimentation of the electrode material).By replacing the solution-based electrolytes with stable,non-settling,dispersed redox-active materials.Slurry flow batteries combine the advantages of flow batteries and the high energy-density of solid redox-active materials[32-36].With this unique advantage,slurry flow batteries are expected to become one of the most promising large-scale energy storage technologies[37-40].

2.1 All-organic slurry flow battery

The structural diversity and abundant sources of organic molecules enable facile tunability of electrochemical properties by different functional substituents.However,the traditional solution-based organic flow battery encountered solubility limitations,therefore,applying insoluble organic redox materials in the form of suspension to the flow battery is promising to improve the energy density.Jin's research group designed a water-dispersed polymer particle“slurry”battery using polyhydroquinone and polyimide suspension as the anolyte and catholyte,respectively,which displayed multi-electron activity.(Fig.2a)[41].The high concentration(1 M)of polymer particles“slurry”overcome the solubility-Limitation of the polymer in the aqueous solution and effectively broadens the applicability of insoluble redox materials in flow batteries.However,the redox reaction kinetics of polymers occurred inside the particles was limited by the particle sizes(Fig.2b).Through detailed study of the effects of polymer particles sizes on electrochemical properties and battery performance,it was revealed that reduced polymer particles(from microns to nanometers)promotes the diffusion and charge transfer ability of active polymers,resulting in increased capacity utilization.In addition,due to the big size of polymer particles,an inexpensive dialysis membrane,commonly used in biochemical experiments,was used as the battery separator instead of expensive ion exchange membrane(Nafion?117),which could largely reduce the cost of battery.The crossover of catholyte and anolyte materials were effectively suppressed by size exclusion effect of the dialysis membrane and large molecule size of polymers.Benefited from the enhanced redox kinetics and compatibility of separator and slurry electrolytes,this slurry flow battery accomplishes a capacity of 4.95 Ah/L with 300 stable cycles at a current density of 20 mA/cm2(capacity retention rate of 70%).

In addition to polymers,porphyrin has shown promise in RFBs owing to its reversible redox property originating from theπ-conjugated macrocyclic ring.However,porphyrins typically have low solubility in both aqueous and non-aqueous electrolytes due to the tendency of aggregation of the planar porphyrin framework.Chen et al.constructed a symmetrical non-aqueous RFB using dispersed 5,10,15,20-tetraphenylporphyrin(H2TPP)[42].Porphyrin H2TPP displayed four reversible one-electron-transfer reactions in tetrabutylammonium perchlorate/dichloromethane(TBAP/CH2Cl2)electrolyte(Fig.3a),affording the H2TPP-based symmetrical RFB a voltage of 2.83 V,showing higher voltage.

Than that in aqueous RFBs[23,43,44].They also proposed a Y-zeolite modified poly(vinylidene fluoride)(PVDF)porous membrane(Fig.3b)with a pore size of 0.74 nm,allowing the transference of perchlorate(ClO4-)(0.55 nm),yet suppressing the permeability of H2TPP(1.14 nm).To increase the electrical conductivity of the electrolyte,5 wt% Ketjen Black as the conductive material was employed.The constructed RFB presented a high capacity of 8.72 Ah/L at 20°C and excellent cycle stability in the temperature range of-40.

Fig.2.Schematic diagram of(a)polymer-based RFB and(b)reaction mechanism of polyhydroquinone/polyimide slurry battery.

Fig.3.(a)Structure and CV of H2TPP in TBAP/CH2Cl2.(b)Ion-selectivity principle of the Y-PVDF ion-selective membrane.

Other all-organic semi-solid flow battery systems using 10-methylphenothiazine, thioxanthone, tris(dialkylamino)cyclopropenium oligomer and polythiophene were also reported[35,45,46].In these all-organic slurry flow batteries,suspensions electrolytes with uniform dispersion and high electrochemical stability can be prepared by simple methods of mechanical stirring and blending of conductive agents and organics.Also due to the relatively large particle size of suspended particles,the choice of ion exchange membranes for flow batteries is less demanding.Based on these advantages,all-organic slurry flow batteries show high Coulombic efficiency and excellent cycle stability.

2.2.Inorganic slurry flow battery

The inorganic material-based slurry flow battery adopts the same working principle as depicted in the aforementioned organic systems.The stability and electronic conductivity of the dispersed redox-active particles determine the stability,lifetime,and charge and discharge energy efficiency of the battery system.Chiang and coworkers proposed semi-solid flow battery in carbonate using lithium cobalt oxide(LiCoO2)as the catholyte and Li4Ti5O12as the anolyte materials(Fig.4),which are typical electrode materials in Li-ion batteries[24].This way the high-energy-density feature of lithium ion batteries is combined with the advantage of decoupled output power and energy of flow batteries.With designer battery reservoir and reaction cell design,Chiang et al.estimated that the energy density of the slurry battery system can achieve as high as 300-500 Wh/L(130-250 Wh/kg).Compared with traditional redox flow battery and lithium-ion batteries,the materials and manufacturing costs of inorganic slurry flow battery are in the range of$40-80/kWh,far lower than that of $250/kWh for electric car power systems and$100/kWh for grid energy storage systems[47-49].

In addition to the research focus on redox-active materials in inorganic slurry flow battery[50-52],Chiang et al.optimized the ratio between different particles in the slurry battery[31],Craig Carter et al.and Dominguez-Benetton et al.studied and fluidity of electroactive suspensions in slurry redox flow batteries through model simulation research methods[33,50].Franco and coworkers conducted a comprehensive modeling-based study on the charge-discharge behavior of slurry flow battery[51].Other topics such as diffusion.

Dynamics of the electrode suspension[52],membrane preparation were also widely discussed for optimization of the slurry-based flow batteries[41,42].

Fig.5.(a)Scheme of the LiFePO4/LiTi2(PO4)3 inorganic semi-solid flow battery system(b)Cyclability,capacity retention,and Coulombic and energy efficiencies vs.cycle number.

In addition to non-aqueous inorganic slurry flow batteries,inorganic slurry flow battery in aqueous media was also investigate.Chiang et al.used the suspended lithium iron phosphate/lithium titanium phosphate(LiFePO4/LiTi2(PO4)3)catholyte in lithium nitrate(LiNO3)alkaline aqueous electrolyte(Fig.5a)[34].It was found that several parasitic reactions,such as the degradation of FePO4and hydrogen evolution of LiTi2(PO4)3,could result in reduced electrochemical efficiency.Though the obtained aqueous battery exhibited a poor cycle performance(10th cycle capacity retention of about 40%)due to poor compatibility of lithium ions with aqueous solutions,including cathode stability and anode-mediated hydrolysis(Fig.5b),the inorganic slurry flow battery provides a promising approach towards higher energy density,while keeping the promising feature of decoupled energy and power.

Under the premise of decoupling power and energy,insoluble redox materials are introduced into the flow battery system,which overcomes the upper limitation of.

Solubility and expands the scope of the redox materials.The slurry battery has remarkable prospects in the application on large-scale energy storage.

3.Metal/slurry hybrid flow battery

The metal/slurry flow battery uses metal,instead of suspended materials,as anolyte in the inorganic semi-solid flow battery[53].This new flow concept offers more negative potential from the anode half-reactions,and alleviates interface issues between non-conductive redox-active materials and the conductive carbon additives.

Tarascon et al.studied a Li/Si flow battery system by varying several chemical/physical factors,including the design of flow battery,the composition of the suspension,and flow rate(Fig.6)[54].A static battery was prepared using suspended LiFePO4as the catholyte,affording a power density of 328 mW/cm2at 104 mA/cm2.In traditional lithium-ion batteries,morphology and volume changes of silicon during charge and discharge cause rupture of electrodes,which could result in capacity decay[55].because of the special conductive form of the suspension system and the suspension characteristics of the active particles,the volume change issue can be alleviated.Tarascon and coworkers used suspended nano silicon powder as the active material to explore the feasibility of metal flow batteries in carbonate with lithium hexafluorophosphate[56],and achieved high capacity(2800 mAh/g),high Coulombic efficiency(98%),and low polarization.Morante et al.discussed the electrochemical performance of lithium nickel manganese(LiNi1/3Mn1/3O2)suspension in flow battery,and investigated the effects of impedance variation under dynamic and static conditions on battery performance[53].

In addition to the inorganic electrode metal/slurry flow batteries,organic electro-active materials can also be applied in metal/slurry flow batteries.Liao et al.paired polyaniline suspension with zinc metal,and the polyaniline catholyte showed a maximum theoretical energy density of 66.5 Wh/L[57].Cao et al.adopted polypyrrole microparticle suspension and a manganese dioxide solid electrode as redox electrolyte,displaying a significant improvement on cycle performance[58].Lu et al.introduced the eutectic concept into the manufacture of suspension electrolytes(Fig.7)[59].Compound 10-methylphenothiazine(Fig.7b)was melted and recrystallized on the surface of conducting carbon Ketjen black,which reduced the amount of conductive carbon black and optimized the suspension system.This strategy successfully increased the energy density of 10-methylphenothiazine from 7 Ah/L(soluble)to 55 Ah/L(slurry).

Compared with other slurry flow batteries,the metal/slurry flow batteries have specially designed metal anolyte[56,60,61].Thanks to the high negative potential of metal itself,the metal/sulfur batteries with high voltage and high capacity can be achieved[61-65].The introduction of the negative metal electrode simplifies the composition of the battery.This.

Design effectively mitigates the crossover issue of traditional flow batteries and greatly improves Coulombic efficiency,making it possible to use inexpensive size exclusive membranes.One downside of the metal/slurry battery is metal dendrite generation,which could pose safety concerns.

Fig.6.Schematics of various cells parts(a)Li/Si slurry flow battery and(b)electrochemical cell configuration.

Fig.7.(a)Schematic illustration of metal/slurry flow battery.(b)Chemical structure and redox reaction of 10-methylphenothiazine(MPT).(c)Preparation process of MPT and Ketjen black composite active redox material.

4.Redox mediator-assisted redox flow battery

The suspension flow batteries overcome the solubility limitation of redox species by employing a flowable slurry of insoluble materials.While the concepts have been successfully demonstrated in several reports,the high viscosity and large quantity of conducting additives engender complex fluid dynamics and reduce volumetric energy density,compromising the energy efficiency and imposing challenges in scalability and maintenance.To address these issues,Wang et al.proposed redox-targeting-based flow batteries[66-69].Unlike aforementioned suspension-based flow batteries,the active material of redox-targeting-based flow batteries do not flow with the electrolyte;the transfer of charge is achieved by soluble redox mediators in the electrolytes.

To illustrate the working principle of redox-targeting-based flow batteries,LiFePO4is used as the example of stationary redox electrolyte.The insoluble catholyte LiFePO4and two soluble redox mediators,ferrocene(Fc)and dibromoferrocene(FcBr2),are stored in the catholyte reservoir.During the charging process,Fc and FcBr2are initially oxidized to give Fc+and FcBr2+on the electrode in the cell(Fig.8).The FcBr2+,which has higher redox potential than LiFePO4,oxidizes LiFePO4to Li1-xFePO4in the reservoir and FcBr2+itself is reduced to FcBr2.The reduced FcBr2flows back to the cell and gets re-oxidized to FcBr2+,thus to convert all LiFePO4to FePO4.In the discharge process,FcBr2+and Fc+are first reduced to FcBr2and Fc on the electrode in the cell.Fc,with higher reducing power.

Fig.8.Reaction mechanism of(a)charging and(b)discharging processes in redox-targeting-based flow batteries.

Is then circulated to the reservoir to reduce FePO4to LiFePO4,and Fc itself is oxidized to Fc+.The re-oxidized Fc+flows back into cell and is rereduced to Fc.This process repeats until all FePO4is reduced to LiFePO4to complete the discharge process[67,68].

Fig.9.(a)Schematic illustration.(b)Photograph of a redox-mediator redox flow battery.(c)CV of the redox mediators,LiFePO4,and TiO2.(d)Charge-discharge profiles of the redox flow battery at different current densities.

Wang et al.proposed a redox-mediator-assisted redox flow battery with titanium dioxide(TiO2)as anolyte,and LiFePO4as catholyte(Fig.9a and b).The FcBr2/Fc and cobaltocene[Co(Cp)2]/bis(pentamethylcyclopentadienyl)cobalt[Co(Cp＊)2]served as the redox mediators in the catholyte and anolyte,respectively(Fig.9c)[66].The reversible chemical delithiation/lithiation of redox materials provides an elegant means for advanced large-scale energy storage with decoupled energy and power.The obtained redox-mediator-assisted flow battery displayed a volumetric capacity of 243 Ah/L(Fig.9d),which was higher than most of traditional RFBs[31].

Using this strategy,polyaniline/mesocarbon microbeads flow battery with Fe(III/II)and V(IV/III)as redox mediators represents a 3-fold improvement in volume capacity,compared to flow battery assembled with soluble redox materials[70].With the same strategy of redox mediator in aqueous redox flow batteries,a low-cost,high-capacity RFB was constructed based on ferrocyanide/ferricyanide anolyte and Prussian blue catholyte,exhibiting a high capacity retention rate per cycle 99.991%and capacity 61.6 Ah/L[10].

Similar to metal/slurry flow battery,metals can also serve as negative materials of redox-targeting flow batteries.Wang et al.proposed the redox-targeting metal flow batteries with lithium or zinc as the anolyte[71-73].For the lithium-based redox-targeting flow batteries,LiFePO4was used as the stationary material,and dual functional(2,3,5,6-tetramethyl-p-phenylenediamine)(TMPD)was used as the redox mediator(Fig.10a)[74].Compound TMPD delivered two pairs of peaks at 3.20 and 3.60 V vs.Li/Li+(Fig.10b),which straddle that of LiFePO4(3.45 V vs.Li/Li+).The TMPD functioned as the reductant,while the dictation radical TMPD?2+as the oxidant.The dual-functional TMPD simplifies the composition of catholyte.

The development of lithium metal-based flow batteries in organic medium is primarily limited by the inferior lithium ion conductivity of ion-exchange membrane.These batteries can only be operated at a low current density,rendering the power density much lower than that of aqueous flow batteries.To address this problem,Wang et al.developed a pH-neutral aqueous flow battery assembled with zinc[73].The LiFePO4was also utilized as the main energy-storage electrode,while anionic sulfonated ferrocene derivatives,ferrocene bis(propyl sodiumsulfite)(Fc-SO3Na) and bromoferrocene bis(propyl sodiumsulfite)(BrFc-SO3Na),served as the redox mediators.Benefitting from the unique design,this battery presented an energy density of about 9 times higher than the battery without LiFePO4in catholyte tank.Lu et al.used soluble and inherent polychalcogenides as redox media in sulfur-selenium/lithium flow batteries,which facilely reduced the number of redox media involved in the reaction and simplified the composition of the electrolyte.Polychalcogenides compounds not only participate in battery reactions as reactants,the intermediate products in the battery reaction act as redox mediators,through which stargate,this battery achieves a high material utilization rate(≤99%)and high capacity(1096-1268 Ah L-1)[75].In addition,Wang et al.proposed and applied an optimization strategy based on single-molecule redox-targeting reaction driven by Nernstian potential,which simplified the composition of the electrolyte and improved the voltage efficiency of redox mediator-assisted RFB.Focusing on the main problems of redox mediator-assisted RFBs,this strategy combines stable redox mediators with excellent kinetic properties and material particles with optimized quantity to achieve enhanced battery performance[76].Compared with the slurry batteries,the redox materials are stationary during charge/-discharge in a redox-mediator assisted redox flow battery,which greatly reduces the viscosity of the solution and lowers pump loss.The redox mediator-assisted redox reactions are core of this type of battery.Through the target potential of the mediator molecules,in theory,all high-energy materials in traditional solid-state batteries can be applied in the redox mediator-assisted flow battery,affording high-density energy storage.Therefore,the search and development of suitable redox mediators has become the most challenging and important part of the redox mediator-assisted battery.

Fig.10.(a)Schematic structure of Li/LiFePO4 redox-mediator-assisted flow battery.(b)CV of TMPD and LiFePO4,inset is the chemical structure of TMPD.

5.Opportunities,challenges and outlooks

Flow batteries have attracted remarkable attention for large-scale energy storage due to their special configuration.Aqueous RFBs,represented by all-vanadium RFBs,have shortcomings of high cost and narrow operating temperature range.The non-aqueous flow batteries,though have wider operation temperature and battery voltage,also suffer from low solubility of the active material and the lack of suitable ion conductive membrane.

The insoluble RFBs overcome the limitation of low solubilities of redox-active materials in aqueous and non-aqueous media.While the working principle and feasibility of slurry battery have been verified,there are still many challenges in the development of insoluble flow batteries.For slurry batteries,the use of a large amount of conductive materials to alleviate the non-conductivity of redox-active materials causes the increased viscosity of the suspension system and reduced energy density.The poor performance at high current rate also hinders its large-scale application.The redox-targeting-based flow batteries combine the advantages of traditional flow batteries and semi-solid flow batteries,providing a new way for the development of high energy density flow batteries.Despite advantages,there are still limitation to be solved including,complexity of the electrolyte,sacrifice of voltage efficiency caused by the presence of the redox medium,large number of SEI layers generated at low potential,and difficulty of matching the redox medium and the electrode material.Therefore,the wise choice of suitable redox mediators and matching redox materials and the design optimization of energy storage reservoirs are critical to enhance performance of the insoluble material-based flow batteries.

Novel battery designs have the potential to enhance energy density by loweringthe high viscosity ofelectrolytes.Aspects forimprovementinclude the control of the flow of the suspension and electron communication at the interface of electrode and slurry electrode[8,36].The development on mesoscale modeling provides additional momentum to improve overall energy efficiency and system reliability[51].On the other hand,the optimization of slurry property,such as enhanced electronic conductivity,is expected to reduce the content of conductive additives and improve energy density.Carbon coating methods commonly used in electrode materials for lithium-ion batteries have shown improvement of electrochemical performance of low-conductivity active materials[77].For redox mediator-assisted flow batter,the compromised voltage efficiency can be alleviated by judicious choice of appropriate redox mediators[73,75].

Declaration of competing interest

None.