The present invention relates to a redox battery comprising a polyoxometallate as a redox couple. Also provided is the use of a polyoxometallate as a redox couple in a redox battery, a polyoxometallate for use as a redox couple in a redox battery and a redox battery stack comprising the redox batteries described herein.

Redox batteries are well known in the art and have a wide range of applications. Generally, a redox battery comprises an energy converter, comprising an electrochemical cell with two electrodes separated by an ion exchange membrane or other separator, as well as a chemical store of energy comprising two redox couples located on either side of the membrane, one at a low redox potential and the other at a high redox potential. The chemical store and the converter can be sized separately and the two redox couples are stored in vessels away from the converter.

A commonly known redox battery is the vanadium redox battery. The present form of vanadium redox battery, which uses sulphuric acid electrolytes, was patented in 1988. This battery exploits the ability of vanadium to exist in solution in four different oxidation states. On one side of the membrane, νθ^ is converted to VC½ ⁺ as electrons are removed from the positive electrode during charging. On the other side, V ^3÷ is converted into V ²⁺ on the introduction of t electrons from the negative electrode. During discharge, this process is reversed, resulting in an open circuit voltage of 141V at 25°C. The electrodes are generally an inert material, such as graphite, or porous metal electrodes such as titanium, which are coated with carbon. Both half cells are additionally connected to storage tanks and pumps so that very large volumes of the electrolytes can be circulated through the ceil.

The circulation of the electrolytes and the pumps required to implement it mean that vanadium batteries are generally restricted to large fixed installations and are not generally used in mobile applications. However, such batteries can offer an almost unlimited capacity by increasing the storage tank volume. Further, if the electrolytes are accidentally mixed, the battery suffers no permanent damage. Vanadium batteries can also be left discharged for a prolonged period of time with little ill-effects and can easily be recharged by replacing the electrolyte or by electrochemically recharging the electrolyte. They can also respond quickly to changes in power requirements and can be discharged and recharged many times without damaging the electrode.

The large capacity and size of vanadium redox batteries therefore means that they are used in applications such as averaging out the production of highly variable energy generation sources such as wind or solar power, or to help generators cope with large surges in demand. They are also suitable for applications in which the batteries must be stored for long periods of time with little maintenance, such as in some military electronics applications. The key features of redox batteries are the relative costs and efficiencies. Typically, the charge/discharge efficiency is required to be above 75%. The main costs are due to the converter (which scales w th the power input/output) and the couple chemistries (which scale with the energy storage).

The cost per power output/input is dictated by the current density of the redox battery. Losses in the cell dictate the magnitude of the current density and need to be kept as small as possible in order to maintain an acceptable overall charge/discharge efficiency.

It is desirable to have the potentials of the couples as far apart as possible to maximise the energy store. Further, it Is also desirable to have the redox couple concentrations as high as possible so as to reduce losses from pumping.

However, this requires the couples to have rapid electrode kinetics, as well as requiring that losses such as those from resistances in the cell components and the electrolyte are as low as possible. In order to do this, current densities are typically kept low, thereby reducing the power density so that, by comparison with a fuel ceil, a large (and therefore expensive) converter unit is required. Vanadium redox batteries therefore have a relatively low energy-to-volume ratio, as weli as a more complex structure than other batteries known in the art. Vanadium batteries are also limited in concentration and temperature range by the insolubility of the various species used. Again, this limits the efficiency of the operation of the system, meaning that a larger and more costly energy converter is required,

Typically, the membranes used in redox batteries of the prior art are cationic exchange membranes, which allow hydrogen ions to pass through them, thereby completing the circuit. Such membranes are not totally selective and there is a discharge of ions across the membrane which must be corrected on an occasional basis. The vanadium redox battery does have an advantage over other redox batteries, as the ions are similar on both sides of the membrane, However, the crossover of vanadium species through the membrane represents a self-discharge with the associated loss of efficiency.

It would therefore be advantageous to have a redox couple where the ions are large and negatively charged, Such ions would therefore be excluded from the membrane by the Donan principle and strongly inhibited from crossing the membrane. A further requirement would be for redox couples with electrode kinetics that are as rapid as possible and therefore are as reversible as possible, as well as for redox couples that are able to operate at high temperatures and concentrations.

Accordingly, a first aspect of the present invention involves a redox battery comprising a poiyoxometallate as at least one of the redox couples. The redox battery in accordance with the invention preferably comprises two electrodes separated by an ion exchange membrane or other separator; means for supplying a first redox couple to the first electrode region of the cell; means for supplying a second redox couple to the second electrode region of the cell, the potential of the first redox couple being higher than that of the second redox couple, with at least the higher potential redox couple comprising polyoxometallate.

In an embodiment of the present invention, the polyoxometallate is in the form of a large structure, wherein the polyoxometallate is substantially rejected from the membrane or other separator by means of charge and/or size, so as to be substantially prevented from flowing through the ion exchange membrane or other separator. Preferably, the large structure is a Keggin ion, Keggin ions are large, anionic structures that are repelled by the membranes used in redox batteries. The use of such structures therefore helps to prevent the discharge of ions across the membrane. Another such structure is the bi-nuclear Dawson- Wells structure, typically containing two central atoms and 18 metal centres. Several other more complex structures are also known in the art.

The polyoxometallate system of the present invention is also less corrosive than the standard vanadium systems, which include acids such as sulphuric acid. For example, one commonly used vanadium system includes a mixture of sulphuric and hydrochloric acids which is much more corrosive than the poiyoxometallate solution of the present invention.

In a further embodiment, at ieast the poiyoxometallate redox couple is provided in aqueous solution.

It has surprisingly been found that the electrode kinetics resulting from the use of a poiyoxometallate in a redox battery are much more reversible than the corresponding vanadium species. Preferably, the current density of operation is significantly higher than the standard for a vanadium redox battery.

In one embodiment, the low potential redox couple comprises a low potential poiyoxometallate such as a tungsten poiyoxometallate. Preferably, the structure of the tungsten poiyoxometallate is H _X I2KQ ₄ Q, where K is for example, P, Si, B or Co, preferably Si, B or Co. In another embodiment, the low potential redox couple comprises vanadium (II) and vanadium (III) ions.

The poiyoxometallate solution for use in the redox battery may be non stoichiometric. In one embodiment, the high potential poiyoxometallate is a vanadium poiyoxometallate. Accordingly, in another embodiment of the invention, vanadium ions can be added to the poiyoxometallate solution in the form of at Ieast one vanadium compound selected from V0 ₂ , V2O4, VOS0 ₄ , VO(acac)2, νθ(ΟΙ0 ₄ )2, VO(BF ₄ )2, hydrated versions thereof, and combinations of two or more thereof. In a further embodiment of the present invention, the high potential polyoxometallate is present at Sower concentrations, in combination with another source of vanadium, as outlined above. Preferably, the POM is at concentrations of less than 0.075 by the central atom. The vanadium concentration can be 0.075 M or greater, greater than 0.5 , greater than or equal to 1 M or equal to 2M.

Preferably, the molar ratio between the high potential polyoxometallate and the vanadium compound is at least about 1 :10 or at least about 1.5:10 or at least about 2:10 or at least about 2.5:10 or at least about 3:10, In a specific embodiment, VOS0 ₄ is added in the concentration range of 0,5 to 2M.

The polyoxometallate may be represented by the formula:

X _a [Z _b M _c O _d ]

Wherein X is selected from hydrogen, alkali metals, alkaline earth metals, ammonium and combinations of two or more thereof;

Z is selected from B, P, S, As, Si, Ge, Ni, Rh, Sn, AI, Cu, I, Br, F, Fe, Co, Cr, Zn, H _2l Te, Mn and Se and combinations of two or more thereof;

M is a metal selected from Mo, W, V, Mb, Ta, Mn, Fe, Co, Cr, Ni, Zn Rh, Ru, Tl, AI, Ga, In and other metals selected from the 1 ^st , 2 ^nd and 3 ^rd transition metal series and the ianthanide series, and combinations of two or more thereof; a is a number of X necessary to charge balance the [M _c O _d ] anion; b is from 0 to 20;

c is from 1 to 40; and

d is from 1 to 180,

St is to be understood that such formulae used herein are generic formulae and that a distribution of related species may exist in solution,

Generally, if does not include vanadium then the polyoxometallate is a low potential redox couple. If M comprises at least one vanadium species then the polyoxometallate is a high potential redox couple, It should also be understood that any numerical values are not necessarily integers.

Preferred ranges for b are from 0 to 15, more preferably 0 to 10, still more preferably 0 to 5, even more preferably 0 to 3, and most preferably 0 to 2.

Preferred ranges for o are from 5 to 20, more preferably from 10 to 18, and most preferably 12,

Preferred ranges for d are from 30 to 70, more preferably 34 to 82, and most preferably 34 to 40,

Vanadium or combinations of vanadium and molybdenum are particularly preferred for of the high potential poiyoxometaliate, Phosphorus is particularly preferred for the central atom, Z.

A combination of hydrogen and an alkali metal and/or alkaline earth metal is particularly preferred for X. One such preferred combination is hydrogen and sodium.

Specific examples of polyoxometallates include molybdovanadophosphosphoric acid (H5P 010V2O40), which would act as the high potential polyoxometallate and molybdophosphoric acid (H3PM0 ₁ 2O40 ₎ .

In a preferred embodiment of the present invention, the high potential polyoxometallate comprises vanadium, more preferably vanadium and molybdenum. Preferably the polyoxometallate comprises from 1 to 6 vanadium centres. Thus, particularly preferred polyoxometallates include H3Na ₂ P o-io 204o, H3 a ₃ PMogV3G4o, or H ₃ a4P o ₈ V ₄ 04o and compounds of intermediate composition. In addition, a mixture of these or other polyoxometallate species is also envisaged. For this embodiment, preferably, at least one X is hydrogen. For a further embodiment, X comprising at least one hydrogen and at least one other material selected from alkali metals, alkaline earth metals, ammonium and combinations of two or more thereof is preferred. Alternatively, the high and low potential redox couples can have the same formulation but just be in different redox states, thereby preventing contamination from cross-over.

Other polyoxometallates useful in the redox battery of the invention may be represented by the formula:

X _a [Z _{b 0} Od]

Wherein X is selected from hydrogen, alkali metals, alkaline earth metals, ammonium and combinations of two or more thereof;

Z is selected from B, P, S, As, Si, Ge, Ni, Rh, Sn, Al, Cu, I, Br, F, Fe, Co, Cr, Zn,

H ₂ , Te, Mn and Se and combinations of two or more thereof;

M comprises W and optionally one or more of Mo, V, Nb, Ta, Mn, Fe, Co, Cr, Ni,

Zn Rh, Ru, Tl, Al, Ga, In and other metals selected from the 1 ^st , 2 ^nd and 3 ^rd transition metal series and the ianthanide series;

a is a number of X necessary to charge balance the [Z _b M _c Q _d f ^" anion;

b is from 0 to 5;

c is from 5 to 30; and

d is from 1 to 180.

The use of tungsten in the polyoxometallate compound compared to the use of the other compounds disclosed in the prior art has numerous benefits. It has been found that the tungsten polyoxometallates are more stable at low pH and can be synthetically manipulated to a greater extent than molybdenum analogues. It is therefore possible to create a variety of different structures that are not available when using other polyoxometallate compounds, such as those containing molybdenum, and to use a wider range of materials.

Further, some compositions of polyoxometallates known in the prior art can have a lower solubility than is desired for maximum redox battery performance. It has surprisingly been found that solubility can be improved by using tungsten polyoxometallate redox couples. The tungsten polyoxometallates of the present invention also provide excellent electrochemical performance.

Preferred ranges for b are from 0 to 5, more preferably 0 to 2.

Preferred ranges for c are from 5 to 30, preferably from 10 to 18 and most preferably 12.

Preferred ranges for d are from 1 to 180, preferably from 30 to 70, more preferably 34 to 62 and most preferably 34 to 40,

The high potential polyoxometal!ate according to the above formula useful as the redox couple in the redox battery of the present invention preferably contains from 1 to 8 vanadium centres. Example formulae therefore include x xC½] where x ~ 1 to 6. in one embodiment of the present invention, the po!yoxometa!iate has the formula in another embodiment, the poiyoxometallate has the formula Xa[ZiW ₁ V 0 ₄ o3.

8, P, S, As, Si, Ge, Al, Co, Mn or Se are particularly preferred for Z, with P, S, Si, A! or Co being most preferred. The successful use of such a range of atoms would not be possible with a polyoxometallate that contains, for example, molybdenum, as outlined in the prior art. in particular, the use of silicon and aluminium in combination with tungsten in the polyoxometallates of the present invention has surprisingly been shown to significantly improve the performance of the fuel cells. For example, tungsten polyoxometallates with aluminium or silicon demonstrate more reversible electrochemical properties at a higher potential compared to certain other polyoxometallates.

M may consist of 1 to 3 different elements. In one embodiment, M is a combination of tungsten, vanadium and/or molybdenum. The polyoxometallate may be absent of molybdenum and further may be absent of any metals other than tungsten or vanadium. The polyoxometallate may alternatively consist of tungsten. M preferably includes more than two, more than four or more than six tungsten atoms.

Hydrogen, or a combination of hydrogen and an alkali metal and/or alkaline earth metal are particularly preferred examples for X. X preferably comprises a hydrogen ion or a combination of a hydrogen ion and an alkali metal ion, and more preferably comprises one or more of H ⁺ , Na ⁺ , K ⁺ or Li ⁺ . Preferred combinations include hydrogen, hydrogen with sodium and hydrogen with potassium.

In a preferred embodiment, the high potential po!yoxometallate may be H6[AIWiiVi0 ₄ o]. Alternatively the high potential poiyoxometallate may be X7 S W9V ₃ 04o] where, as an example, X can give rise to the general formula KaHs^i gVsG^] . Further, a mixture of these or other poiyoxometallate catalysts is also envisaged.

The high potential poiyoxometallate for use in the present invention may also be represented by the formula:

X _a VyQl2-yZ0 i) wherein X is selected from hydrogen, alkali metals, alkaline earth metals, ammonium and combinations of two or more thereof;

Z is selected from B, P, S, As, Si, Ge, Ni, Rh, Sn, Al, Cu, I, Br, F, Fe, Co, Cr, Zn, H ₂ , Te, n and Se and combinations of two or more thereof;

Q is selected Mo or W;

a is a number of X necessary to charge balance the jy _y Qi2- 04o] ^a" anion; and y is between 1 and 8, more preferably between 2 and 4, even more preferably between 3 and 4. The high potential polyoxometallate for use in the present invention may also be represented by the formula:

H3 ₊ xMy-xVyQi2-y2C o wherein Z is selected from B, P, S, As, Si, Ge, Ni, Rh, Sn, Al, Cu, l ₍ Br, F, Fe, Co,

Cr, Zn, H ₂ , Te, n and Se and combinations of two or more thereof;

Q is selected Mo or W;

M is selected from the alkali metal ions;

x is between 0 and 4;

y is between 1 and 8, more preferably between 2 and 4, even more preferably between 3 and 4; and

y is greater than or equal to x.

Z is preferably phosphorus. Further, x is preferably between 0 and y and y is preferably between 2 and 4, or mixtures thereof. In a further embodiment, such a polyoxometallate is present at a concentration of up to around 0.6 by concentration of the central atom (Z), which is commonly phosphorus.

In another embodiment, the high potential polyoxometallate for use in the present invention may be represented by the formula:

HaP _c MOyV _x Ob Wherein a is a number of H necessary to charge balance the [P _c o _y V _x O _b ] ^a' anion;

c is between 1 and 3;

y is between 8 and 20;

x is between 1 and 12; and

b is between 40 and 89.

Preferably, such a po!yoxometallate has the formula HiaPsMQisVjOas], Hi ₂ [P2Moi2V ₆ 082]. H ₁₄ [P2Mo ₀ V8O ₆₂ ] or Hi5[P3 o ₁₈ V ₆ 0 ₈ 4]-

In one preferred embodiment of the invention, the ion selective PE is a cation selective membrane which is selective in favour of protons versus other cations.

The cation selective polymer electrolyte membrane may be formed from any suitable material, but preferably comprises a polymeric substrate having cation exchange capability. Suitable examples include fluororesin-type ion exchange resins and non~fluororesin-type ion exchange resins, Fluororesin-type ion exchange resins include perfluorocarboxylic acid resins, perfluorosulfonic acid resins, and the like, Perfluorocarboxyiic acid resins are preferred, for example " afion" (Du Pont Inc.), "Flemion" (Asahi Gas Lid),"Acipiex" (Asahi Kasei Inc) and the like. Non-fluororesin-type ion exchange resins include polyvinyl alcohols, polyalkylene oxides, styrene-divinylbenzene ion exchange resins and the like, and metal salts thereof. Preferred non-fluororesin-type ion exchange resins include polyalkylene oxide-alkali metal salt complexes. These are obtainable by polymerizing an ethylene oxide oligomer in the presence of lithium chlorate or another alkali metal salt, for example. Other examples include phenolsulphonic acid, polystyrene sulphonic, polytriflurostyrene sulphonic, sulphonaied trifluorostyrene, sulphonated copolymers based on α,β,β triflurostyrene monomer and radiation-grafted membranes. Non-fluorinated membranes include sulphonated polyCphenylquinoxaiines), poly (2,8 diphenyl-4~phenylene oxide), poiyCary!ether sulphone), poly(2,8-dipheny!enol), acid-doped polybenzimidazo!e, sulphonated polyimides, styrene/ethylene-butadiene/styrene triblock copolymers, partially sulphonated polyaryiene ether sulphone, partially sulphonated polyether ether ketone (PEEK) and polybenzyl suphonic acid siloxane (PBSS).

In some cases it may be desirable for the ion selective polymer electrolyte membrane to comprise a bimembrane. The bimembrane if present will generally comprise a first cation selective membrane and a second anion selective membrane. In this case the bimembrane may comprise an adjacent pairing of oppositely charge selective membranes. For example the bimembrane may comprise at least two discreet membranes which may be placed side-by-side with an optional gap therebetween. Preferably the size of the gap, if any, is kept to a minimum in the redox cell of the invention. The use of a bimembrane may be used in the redox battery of the invention to maximise the potential of the cell, by maintaining the potential due to a pH drop between the anode and catholyte solution. Without being limited by theory, in order for this potential to be maintained in the membrane system, at some point in the system, protons must be the dominant charge transfer vehicle. A single cation-selective membrane may not achieve this to the same extent due to the free movement of other cations from the catho!yte soiution in the membrane,

In this case, the cation selective membrane may be positioned on the cathode side of the bimembrane and the anion selective membrane may be positioned on the anode side of the bimembrane. In this case, the cation selective membrane is adapted to allow protons to pass through the membrane from the anode side to the cathode side thereof in operation of the cell. The anion selective membrane is adapted substantially to prevent cationic materials from passing therethrough from the cathode side to the anode side thereof, although in this case anionic materials may pass from the cathode side of the anionic-selective membrane to the anode side thereof, whereupon they may combine with protons passing through the membrane in the opposite direction. Preferably the anion selective membrane is selective for hydroxyl ions and combination with protons therefore yields water as a product.

In a second embodiment of the invention the cation selective membrane is positioned on the anode side of the bimembrane and the anion selective membrane is positioned on the cathode side of the bimembrane. In this case, the cation selective membrane is adapted to allow protons to pass through the membrane from the anode side to the cathode side thereof in operation of the cell. In this case, anions can pass from the cathode side into the interstitial space of the bimembrane and protons will pass from the anode side. It may be desirable in this case to provide means for flushing such protons and anionic materials from the interstitial space of the bimembrane. Such means may comprise one or more perforations in the cation selective membrane, allowing such flushing directly through the membrane. Alternatively, means may be provided for channelling flushed materials around the cation selective membrane from the interstitial space to the cathode side of the said membrane.

The redox battery of the present invention also operates at different current densities to those redox batteries of the prior art, which generally operate at between 50 and 100 mA/cm ² . In contrast, the redox battery of the present invention operates at a current density of at least 250 mA/cm ² , preferably at least 400 mA/cm ² and more preferably 500 mA/cm ² .

A further aspect of the present invention provides the use of a polyoxometallate as described above as a redox couple in a redox battery. Also provided is a polyoxometallate as described above for use in a redox battery and a redox battery stack comprising the redox batteries as described above.

Various aspects of the present invention will now be more particularly described with reference to the following figures, which illustrate embodiments of the present invention:

Figure 1 demonstrates the voltage profiles for charge and discharge cycles using all-vanadium and PO 1/PO 2 polyoxometallate flow batteries; Figure 2 demonstrates the voltage profiles for charge and discharge cycles using all-vanadium and POM1/PO 3 po!yoxometallate flow batteries;

Figure 3 demonstrates cell potential during charge and discharge vs. open circuit potential for all-vanadium and PO 1/POM2 polyoxometaliate flow batteries; and

Figure 4 demonstrates cell potential during charge and discharge vs. open circuit potential for all-vanadium and POM1/PO 3 polyoxometaliate flow batteries.

The data was obtained using a test stand with separate heated anolyte and catholyte reservoirs. A redox cell with 31cm ² active area was constructed using gold plated copper current collectors, stainless steel end plates, carbon anode and cathode blocks with plug-flow flow fields and silicone gaskets. Both anode and cathode electrodes were made from 2.5mm GFD 2.5 carbon felt (SGL) that was compressed to 1.15mm thickness. An NRE-212 membrane (Ion Power) was used to separate the anode and cathode. The anolyte and catholyte reservoirs were purged with nitrogen for 3 minutes before filling. 50ml of anolyte and 100ml of catholyte were measured out and poured into the respective reservoirs. The reservoirs were then purged with nitrogen every 10 minutes and were sealed whilst not being purged.

The battery was charged and discharged using a Bio-Logic potentiostat. Open circuit potential was also measured using this instrument. Redox batteries were constructed and tested using: a) vanadium/sulphate + suiphuric acid solution as both anolyte and cathoiyte, b) silicotungstic acid as anolyte and Na ₄ H ₃ Mo _s V4P0 ₄ o] polyoxometallate as cathoiyte and c) silicotungstic acid as anolyte and Hio[P2 4 o ₈ 0 ₄ ] polyoxometallate as cathoiyte. The solutions used were:

1 .8 vanadium and 4.0M S0 ₄ in ultra pure water (proton counter ions);

0.3 silicotungstic acid (2.4 tungsten 100% oxidised) in ultra pure water (POM1);

0.3 Na4H ₃ [ oaV ₄ P0 o] (1.2 vanadium 50% oxidised) in ultra pure water {POM2); and

0.45 H-sofPa ^MoeC^] (1 .8M vanadium 50% oxidised) in ultra pure water CPG1V13).

A new cell was constructed for each different anolyte/catholyte combination. The eel! cathoiyte and anolyte inlet pressure was maintained at 1000 ±150 mbar gauge. Cell, cathoiyte and anolyte temperatures were controlled at the same system temperature ±2°C. The current for a!l experiments was 2.50A or 80.6mA/cm ² ,

The performance of the vanadium/SC flow battery was only assessed at 35°C due to the limited solubility of V ^(V) at temperatures above 40°C. The use of the polyoxometallate solution circumvents this solubility issue and therefore allows for the operation of the system at higher temperatures. Measurements were taken at 35 and 70°C for at least one charge - discharge cycle at a high current density.

Figure 1 demonstrates the voltage profiles for charge (solid Sines) and discharge (dashed lines) cycles at 2.5A (80,8mA/cm ² ) using an all-vanadium flow battery at 35°C and POM1/POM2 polyoxometallate flow batteries at 35 and 70°C. Figure 2 demonstrates the voltage profiles for charge (solid lines) and discharge (dashed lines) cycles at 2.5A (8Q.6mA/cm ² ) using an all-vanadium flow battery at 35 and PO 1/PO 3 polyoxometallate flow batteries at 35 and 70°C.

Figures 1 and 2 demonstrate that the polyoxometallate flow batteries display E under load vs Q curves that are not unlike those expected for standard vanadium redox battery systems, although it is clear that the measured potential of the polyoxometailate batteries is io er than the vanadium systems. However, it is expected that the potential of the polyoxometailate battery can be raised by:

* further reducing PO 1 ; and/or

* optimising the formulation of both catholyte and anolyte.

Figure 3 demonstrates cell potential during charge (circles with dashed lines) and discharge (squares with solid lines) at 80.8mA/cm ² vs. open circuit potential for all-vanadium, PO 1/PO 2 polyoxometailate at 35°C and PO 1/POM2 polyoxometailate at 70°C. Figure 4 demonstrates cell potential during charge (circles with dashed lines) and discharge (squares with solid lines) at 80.8mA/cm ² vs. open circuit potential for all-vanadium, POM1/POIV13 polyoxometaHate at 35°C and POM1/PO 3 polyoxometailate at 70°C.

Figures 3 and 4 demonstrate that the POM1/PO 3 polyoxometailate battery has lower charging and discharging over-potential than the PQ 1/POIVI2 polyoxometailate battery. It is also apparent that for a portion of the open circuit range, the over-potential for the PO 1/POM3 polyoxometailate battery is Iower than that for the all-vanadium battery. The coulombic efficiency has been calculated for ail systems as:

Coulombic efficiency =

Where:

AGd - charge passed during discharge at 2.5A from OCP ^™ x to OCP ~ y AQc™ charge passed during charge at 2.5A from OCP -y to OCP ~ x

This table demonstrates the calculated values for the coulombic efficiency of all redox battery systems tested in this series of experiments, it is apparent that the coulombic efficiencies of the polyoxometailate redox batteries were similar to that of the vanadium system at this high current density (the PO 1/POM3 polyoxometaliate at 35°C result may be anomalous). It is expected that the columbic and overall electrical efficiency advantage of the poiyoxometailate systems will be clearer at higher operating current densities, as the electrode kinetics of the all-vanadium system are poorer than the poiyoxometailate system, The voltage loss at higher current densities will therefore be higher for the all-vanadium system than for the poiyoxometailate systems.

It is anticipated that the charge and discharge current density of the poiyoxometailate flow batteries can be further increased (without loss of efficiency) by:

* increasing the operating temperature above 70°C;

* optimising the formulation of catholyte and anolyte; and/or

» improving the cell construction with for example thinner membranes or activated electrodes.