GB2527104A - Redox probe - Google Patents

Abstract

A redox probe for measuring the redox state of a test electrolyte comprising: a first electrode in contact with the test electrolyte; a second electrode, electrically connected to the first electrode, in contact with a reference electrolyte; and means for effecting ion transfer extending between the reference electrolyte and the test electrolyte; wherein the reference electrolyte has a pH value within 4 pH units of the pH of the test electrolyte. The similarity in pH between the reference and test electrolytes lowers the junction potential and allows for increasing accuracy in the voltage readings used to obtain the redox potential. The pH of the electrolytes may be the same and they may contain similar compounds that may or may not have the same oxidation state. Ion transfer may be via a salt bridge. The invention may be applied to a fuel cell, redox battery or an electrolyser.

Description

Redox Probe Redox probes are used to measure the redox potential of liquids and have numerous applications in the prior art. To measure the redox state of the liqthd, two electrodes are required -one that gives a constant potential (the reference electrode) and one that is in the liquid to be measured (the working Sectrode). The two Sectrodes need to be connected via an ionic conducting media (for example an electrolyte and/or membrane), often referred to as a salt bhdge". A voRmeter can then be connected to both electrodes and the reading from the voltmeter indicates the redox state of the liquid.

The most common reference electrodes used in redox probes are: Calomel, Hg/Hg2CI2 Silver-silver chloride, Ag/AgCl Mercurymercury oxide, Hg/HgO Mercurymercurous sulphate, Hg/Hg2SO4 Silversilver sulphate, Ag/Ag2504 Copper-copper sulphate. Cu/CuSO4 However, the redox measurements arising from these reference electrodes vary from the theoretical value due to (liquid) junction potentials (sometimes referred to as diffusion potentials") within the system. Further, these reference electrodes are often expensive and are complicated to manufacture. They can also suffer from contamination as species from one electrolyte can pass through the salt bridge to the other electrolyte, which affects the redox state reading.

The following demonstrates how the redox state of a test electrolyte can be measured using a hydrogen reference electrode, The reaction occurring in the reference electrode half ced is: 2H(aq) + 2W H2(g) (RI) The standard redox potentia. E. for this reaction is 0 V. The potential of the hydrogen reference electrode, E(H2), can be written as: R7' ( P E(H2 = I (El) F [Ht}Y) Where: * Ih is the partial pressure of hydrogen * F is the standard pressure (1O Pa) * R is the universal gas constant * T is the temperature (K) * F is the Faraday constant * [H4] is the proton concentration For most applications with a hydrogen reference electrode, the logarithm in Equation I (El) is approximately zero, so the potential of the electrode when zero current flows is near 0 V. When the test electroMe comprises a polyoxornetate (POM), the electrochemical half cefl of the working electrode can be w'itten as (for simptity the reaction is written in terms of just one redox species undergoing a single electron transfer reaction without protons): POMOX(aq) + a-= POM(aq) (R2) In which POM°" is the oxidised POM and POMr is the reduced POM. For this simplified case, the potential of the working electrode. E(POM), when zero current flows is given by: RT (rpçfl41\ E(POM)=E' (E2) The potentl of the cathode is therefore dependent on the relative concentrations of reduced and oxidised POM. When a voltmeter is connected between the reference and the working electrodes, the reading on the voltmeter, E, is given by: E = E(POM) E(H2) + (E3) Where E3 is the sum of any junction potentials present in complete electrochemical ce [R, 0. Compton, 0. H. W. Sanders, Electrode Potentials, Oxford University Press, 1998]. The junction potentials in the complete ceH therefore affect the reading on the voltmeter. As discussed above, the reference eiectrode potential is approximately zero and constant (providing no current flows). Theoretically, if no junction potentials are present. Equation 3 (E3) can therefore be written as: E E(POM) (E4) Hence, the reading on the voltmeter theoreticay directly indicates the raflo of reduced to oxidised POM and the sum of any junction potentials introduces inaccuracies in the reading.

Redox probes are often used in PEM fuel ceDe, such as that shown in Figure 1, which use an aqueous cathode system with a catholyte consisting of soluble active species. Oxygen is reduced in solution by a homogeneous catalyst which is then reduced back to its original state at the electrode. Completion of this cycle creates a regenerative redox cathode. This system increases the efficiency of the reaction and means that expensive catalysts such as platinum are not required.

However, knowledge of the redox state of the catholyte is required. This indicates the level of reduction of the catholyte and gives information on the efficiency of the regeneration reaction (the reaction between oxygen and the reduced catholyte). Redox probes can be used to measure the potential of the individual electrolytes entering and leaving the ceU/stack. An example of this is disclosed in WO 90/03666 in relation to a vanadium redox flow battery, which uses a saturated calomel electrode and an Hg/HgSO4 electrode to measure the potentials of the &ectrolytes.

The redox state of the catholyte can also be measured using another fuel cefi, called a "dummy cell". The dummy cell consists of a porous carbon cathode, a membrane electrode assembly and a gas diffusion layer on the anode side. For systems like those in Figure 1, the membrane electrode assembly is typically bare on the cathode side and loaded with platinum on the anode side. The anode compartment of the cell is filled with hydrogen gas (at the same pressure as the working fuel ceil) and acts as the reference electrode. The catholyte flows through the porous carbon cathode, which acts as the working electrode. No current is passed through the dummy cell. Note that a working fuel cell can also be used to measure the redox state of the catholyte when zero current flows Le. at open circuit.

Although the dummy cell is a durable and accurate method for measuring the redox state of the catholyte, there are a number of disadvantages that make dummy cells unsuitable for mass production: Dummy cells are relatively expensive as they require membrane electrode assemblies (both platinum and the ionomer membrane are relatively expensive materials) * The assembly of dummy cells is time consuming Dummy ceDs are rather rbulky and require a relatively large space Dummy cefls require a connection to the hydrogen supply and corresponding exhaust Therefore, an afternative redox probe with the foflowing properties is desirable for mass production: * Constructed of ow cost materials Easy to assemble Sma and easfly miniaturized Only requires connecting to the catholyte flow * Durable Thus, according to a first aspect of the present invention, there is provided a redox probe for measuring the redox state of a test electrolyte comprising a first electrode in contact with the test electrolyte; a second electrode, electricaHy connected to the first electrode, in contact with a reference electrolyte: and means for effecting ion transfer extending between the reference electrolyte and the test electrolyte; wherein the reference electrolyte has a pH value within 4 pH units of the pH of the test electrolyte.

It has surprisingly been found that using a reference electrolyte with a pH similar to that of the test electrolyte significantly reduces the junction potentials of the electrochemical cell as a whole (ic. the system comprising the two half cefi reactions). As shown in the equations discussed above, this will improve the accuracy of the reading on the voltmeter, thereby allowing a more accurate measurement of the redox state of the test electrolyte.

By "electricaily connected' we mean preferably that means are in place to enable measurement of the voltage and/or potential difference between the reference electrode and the test electrode. This may be achieved not necessarUy through a direct connection but by other means such as a voltmeter, potentiometer or potentiostat, for example.

The reference electrolyte may have a pH value within 3.5, 3.0, 25. 2.0, t5, tO or 0.5 pH units of the pH of the test electrolyte. The reference electrolyte may have the same pH value as that of the test electrolyte.

The reference electrolyte may comprise one or more compounds having the same chemical composiflon as one or more compounds that are present in the test electrolyte. When this is the case, the one or more compounds having the same chemical composition may or may not have the same oxidation state in the reference electrolyte and in the test electrolyte at any given time.

The reference electrolyte may comprise compounds that are present in the test electrolyte.

The reference electrolyte may consist of compounds that are present in the test electrolyte.

This reduces the effects of contamination, as the species in each electrolyte are the same.

The movement of a species from one &ectrolyte into the other electrolyte wifi therefore have htte or no effect on the voftmeter reading. However, the compounds need not be in the same redox state in each electrolyte, afthough the oxidation state of the reference electrolyte must be known.

In one embodiment, the reference electrolyte comprises a polyoxometalate. The polyoxometalate may have a high oxidation state or a ow oxidation state. By this, it is meant that the polyoxometalate solution, or any ndividuai redox active coordinated metal within the polyoxometalate, may be fully oxidised or close to fuhy oxidised, or partiafly reduced, or fully reduced or close to fufly reduced. Intermediate oxidation states may also be used. ft has been found that consistent results are obtained when the polyoxometalate solution is close to fuHy oxidised as the oxidation state does not change greatly on contact with air, It has also been found that the voltmeter reading settles much faster when the polyoxometalate is partially reduced or. conversely, partiafly oxidised (i.e. there are significant amounts of reduced and oxidised species present).

The first andior second electrodes may be solid electrodes, This eliminates a junction potential that can exist within the reference &ectrodes of the prior art. The reference electrodes of the prior art discussed above could be used as the second (reference) &ectrode in present invention, as long as they extend into a reference electrolyte with the claimed pH. However, these electrodes a include a solid electrode in a further electrolyte solution. The further electrolyte solution is separated from the reference electrolyte via a salt bridge. This arrangement does decrease the junction potential of the system as a whole as the reference electrolyte and the test electrolyte have the claimed pH and so the junction potential between the two is decreased. However, this arrangement creates an additional junction potential between the reference electrolyte and the further electrolyte within the reference electrode. It is therefore advantageous to use a solid electrode in direct contact with the reference electrolyte to avoid creating additional junction potentials.

The first and/or second electrodes may comprise a disc, rod, plate or foam. The first and/or second electrodes may comprise glassy carbon, graphite, other graphite materials (such as a graphite/polymer composite), platinum, boron doped diamond or metals such as titanium or stainless steeL It is desirable that the electrode material does not react with the test or reference electrolytes.

The salt bridge must adequat&y separate the two electrochemical halt cells and allow ions to travel between both half cells. It must also be electrically insulating. PossibHe salt bridges include: Membrane materials such as sulfonated fluoropolyrner (e.g. Nation) or hydrocarbon membranes Microporous separators, such as those used in lithium batteries * Porous hits, such as those used in traditional" reference electrodes (eg. the R.E2B calomel reference electrode manufactured by ALS Co. Ltd, Tokyo, Japan), which may be made from a number of materials, including vycor glass or ceramic * Fibrous materials, such as those used in specialised reference electrodes (e.g. the XRI5O Calomel Reference Electrode produced by Radiometer Analytical SAS, 72 rue dAisace, 69527 ViUeurbanne CEDEX, Lyon, France) Ideally, the salt bridge should not aHow the separated electrolytes to diffuse across the boundary and mix. This contamination would cause the potential of the reference electrode to vary, thus limiting the accuracy of the redox probe. Microporous separators, porous hits and fibrous materials aV allow mixing of the separated electrolytes to some degree.

However, membrane materias such as Nation have been found to effectively separate the electrolytes.

The ion conducting means of the redox probe may comprise a Nation rod or tube. The ion conducting means may comprise a Nation tube fifled with a curable elastomer. As discussed above, this reduces the exchange of species from one electrolyte to another. Further, the length of such an ion conducting means can be varied if contamination is an issue, as the longer the ion conducting means, the less contamination will occur, A long ion conducting means is therefore beneficial. A Nation tube filled with a curable elastomer does also not easily become contaminated, unlike some porous hits known in the art, which are hard to clean.

S

The first &ectrode may compdse a hollow body through which the test electrolyte can flow.

This increases the contact between the electrolyte arid the electrode and h&ps to provide a reUable reading. The on conducting means can extend through a wall of the hollow body into the test electroMe.

According to a second aspect of the present invention, there is provided a fuel cell, a redox battery or an electrolyser including the redox probe of the present invention, wherein the test electroiyte comprises an electrolyte of the fuel cell, redox battery or electrolyser.

Instead of the dummy cell generay used to measure the redox state of an electrolyte within a fuel cell, the present invention uses electrochernical half cells that have the desirable properties of a redox probe according to the present invention.

This arrangement reduces the contamination resulting from species in one electrolyte passing through the means for effecting ion transfer to the other electrolyte over time, which can affect the redox state measurement. Dummy cells also have robustness issues, as they can burst due to the differential pressure, as we as requiring regular cleaning. Additionally, the test electrolyte is normally in direct contact with the membrane separating the reference and the test electrolytes in a dummy cell arrangement. Since the membrane is relatively thin (50 microns or less), this can result in species in the test electrolyte passing through the membrane, which is detrimental to the accuracy of the redox state measurement. Leaks may also occur in the fuel cell, which can have the same effect.

Additionally, the dummy cell arrangement of the prior art runs with the dummy cell itself at two different temperatures. The test electrolyte will be at the temperature at which the fuel cell is run, which is generally 50°C or higher. In contrast, the reference electrode in the dummy cell is at room temperature. The temperature of the half ceV reactions affects the redox state measurement and so calibration is required, which can be difficult when different temperatures are involved.

The fuel cell, redox flow battery or electrolyser may be used in an electronic, automotive, combined heat and power or any other suitable equipment.

According to a third aspect of the present invention, there is provided a use of a redox probe comprising: a first electrode that can contact a test electrolyte; a second electrode, electrically connected to the first electrode, in contact with a reference electrolyte; and a means for effecting ion transfer able to extend between the reference electrolyte and the test electrolyte; to measure the redox potenfial of the test electrolyte; wherein the reference electrolyte has a pH vakie withfr 4 pH units of the pH of the test eectrolyte.

The use of a probe comprising a reference &ectroiyte with a similar pH to the test electrolyte has surprisingly been found to reduce the junction potentials in the cell as a whole, As discussed above, this results in a more accurate measurement of the redox state of the test electrolyte.

The optional features discussed above in relation to the first aspect of the invention all apply to the third aspect of the invention.

The probe of the present invention is flexible as it allows the reference electrolyte to be selected with regard to the test electrolyte. In other words, a reference electrolyte with the required pH can be selected for each test electrolyte to be measured. If desired, a reference electrolyte consisting of components that are present in the test electrolyte can be s&ected for each test electrolyte.

The test electrolyte may be an electrolyte in a fuel cell, redox flow battery or electrolyser.

The test electrolyte may be a cathotyte or an anolyte in a fuel ceO, redox battery or electrolyser. The specific benefits of the use of such a redox probe in a fuel cell are discussed above.

Figures The present invention will now be described in more detafi, by way of example only, with reference to the figures, in which: Figure 1 iflustrates a conventional fuel cell of the prior art; Figure 2 illustrates a schematic diagram of the experimental set up for the results in

Table I below;

Figure 3a illustrates the two half cells separated by the Nafion tube assembly salt bridge discussed below; Figure Sb iflustrates the experimental set up where 0.5 M KCI solution and 0.3 M fresh NaV4 polyoxometalate (POM) are on efther side of the Nation tube assembly salt bridge; Figure 4a Dustrates a schematic diagram for the POM reference electrode used for stationary appcations; Figure 4b illustrates the POM reference &ectrode in Figure 4(a); Figure 5 illustrates cydllc voltammograms for 0.3 M Na'.!4 POM at a 3 mm diameter edge plane pyrolytic graphite working electrode where the reference &ectrode is a POM reference electrode (PRE) and a saturated calomel reference electrode (SCE); Figure 6 illustrates a schematic diagram of the POM reference electrode used for flow apphcations; Figure 7 iilustrates a schematic showing the fuel cell rig including the reference electrodes; Figures 8 to 11 illustrate the catholyte redox potential at different redox states and temperatures, as measured by three reference electrode systems (the potential is corrected so that all reference electrode potentials are relative to the Hydroflex reference electrode). Approximate catholyte potentials are as foUows: 1.04 V in FigureS 1 V in Figure 9, 0.8 V in Figure 10 and 0.7 V in Figure 11; Figures 12 to 15 illustrate the catholyte redox potential at different redox states and temperatures measured with a Hydroflex reference electrode. Also shown are the corresponding ceU potentials at open circuit and the potential of the anode measured using the same Hydroflex reference electrode (the latter is plotted on the right hand y axis). Approximate catholyte potentials are as follows: 1.04 V in Figure 8, 1 V in Figure 9, 0.8 V in Figure 10 and 0.7 V in Figure 11; Figure 16 illustrates a scale drawing of a redox probe housing made from graphitic carbon, where the housing acts as the working electrode, eliminating the need for an extra working electrode; Figure 17 is a more detailed drawing of Figure 16; and Figure 18 illustrates a plot of catholyte potential agarnst moles of electrons passed dudng Sectrochemical reduction, where the potential was measured using the redox probe in Figure 16 and a Hydroflex reference &ectrode attached via a Nation tube bridge.

Examples

In the examples below a number of electrodes/electrode materials were used: Platinum thee electrodes were 2 mm diameter supplied by ALS Co. LW, Tokyo? Japan Platinum wire (>99,9%) was either 0.127 mm or 0.5 mm diameter purchased from Ma Aesar, Heysham, UK Glassy carbon disc electrodes were 3 mm diameter supphed by ALS Co. Ltd, Tokyo, Japan Glassy carbon rods (Type 1) were 3 mm diameter purchased from Ma Aesar, Heysham, UK Graphite rods (99.9995%) were 3.05 mm diameter purchased from Al Aesar, Heysham, UK Edge plane pyrolytic graphite (EPPO) disc Sectrodes were 3 mm diameter supphed by ALS Co. Ltd, Tokyo, Japan * Saturated calomel &ectrodes supplied by ALS Co. Ltd, Tokyo, Japan * Stainless steel sheet was 0.1 mm thickness 316 stainless steel purchased from Knight Strip Metals Ltd, Potters Bar, Hertlordahire, UK * ETOID Hydroflex reference electrodes were supplied by eDAG, Warszawa, Poland * JP045 rods were 3 mm diameter purchased from Mersen, Portslade, UK Plastic formed carbon electrodes (PFCE) were 3 mm diameter discs eiectrodes supplied by ALS Co. Ltd, Tokyo, Japan The Nafion tube was purchased from Vaisala (Moisture removing Nation tubing for GM7O hand-held meter), Birmingham. UK. iii

Example I

Hgure 2 iUustrates the experimental set up for Example I with two electrodes (a reference electrode and a test electrode) in contact with a single electrolyte solution. An electrical connection includlilg a digital voltmeter (DVM) extends between the two electrodes. A saturated calomel electrode, SCE (RE2G. ALS Co. Ltd, Tokyo, Japan), was placed in a beaker of fufly oxidised, aqueous 0.3 M Na4HFPV2MoO4 (NaV4 POM) alongside another electrode materiaL in the same eclution. In this case, the salt bridge is the porous tilt on the calomel electrode that separates the saturated KCI solution from the fresh POM. The two electrodes were then connected to a DVM with the saturated calomel electrode connected to the negative terminal and the other electrode material connected to the positive terminaL Table I sts the different electrode materials used, the reading obtained from the digital voltmeter, E (against the saturated calomel electrode), and the time taken for the reading to stabize, Also listed is the potential vs. the standard hydrogen electrode. This is obtained by adding 0244 V to the digital voltmeter reading since the potential of the saturated calomel electrode at 2000 is +0.244 V vs. the standard hydrogen electrode [http:i/en.wikipedia.org/wiki/SaturatedpalomelelectrodeJ.

Table I. Potentials recorded between a saturated calomel electrode and a range of eectrode materials in a solution of freab POM. The time taken to obtain a stable reading is also listed.

EEètrode Material E I V vs. E I V vs. Stabilization SCE SHE time! mine Platinum disc 0.766 1010 C I Platinum wire 0.785 1.029 c1 Gssycarbon disc 0178 1.022 <1 Glassy carbon rod 0,783 1.027 Graphite rod 0.764 1.008 > 5 Edge plane pyrolytic graphite disc 0.779 1.023 <1 >5 The voltmeter reading with the stainless steel sheet electrode took a long time to stabilize and gave a reading that was too low (compared to the other electrode materials). This suggests stainless steel is not suitable for this application. Also, the stabilization time when using the graphite rod was long. The graphite rod was found to be microporous and slowly absorbed the electrolyte. The other electrodes in Table I performed well and suggested the redox potential of the fresh POM was 1.02 +/ 0.01 V vs. SHE. A major contribution to the 10 Ii rrV accuracy range most Ukely arises from the potential of the SCE drifting whilst it was immersed in POM. For example, a 6 my difference was observed between a freshly prepared SCE and an SCE that had been immersed in POM for several hours (a voltmeter was connected across the two SCEs whilst immersed in 0.5 M KCI(aq) solution).

With the potential of the fresh POM determined, the electrodes were used in the experimental set up illustrated in Figure 3. A Nation tube was filled with silasfic elastomer then inserted into a FEP tube of 0125 inches internal diameter and 025 inches external diameter, This tubeassembiy bridged two glass vessels, thus forming a salt bridge. The glass vessel on the left was fills with 0.5 M KCI solution and the vessel on the right was tilled with fresh 0.3 M NW4 POM (the same s&ution from above). The same SCE used for the experiments in Table I was placed in the 0.5 M KCl solution and a range of different electrodes were placed in the fresh POM. The two electrodes were then connected to a digital voltmeter with the saturated calomel electrode connected to the negative terminal and the other electrode material connected to the positive terminal. Table 2 lists the different electrode materials used, the reading obtained from the digital voRmeter, E (against the saturated calom& electrode), and the time taken for the reading to stabite. Also listed is the potential vs. the standard hydrogen electrode.

Table 2. Potentials recorded between a saturated calomel electrode immersed in 0.5 M KCI(aq) and a range of electrode materials in a solution of fresh POM, where the two ha]f cells were connected via the Nation tube salt bridge thscussed in the text. The time taken to obtain a stable reading is also listed.

[iiictrode Material E I V vs. 2 / V vs. Stabilization SCE SHE tinielmins Platinum disc 0.668 0912 c I Platinum wire 0.668 0.912 <3 Glassy carbon disc 0.680 0.924 <1 Glassy carbon rod Graphite rod Edge plane pyrolytic graphite disc 0.668 0.912 C 2 Stainless steel sheet 0.818 0.862 > 3 Ideally, the results in Table 2 should be the same as those in Table I. The relatively large difference in potentials suggests there may be a large junction potential across the Nation tube salt bridge. The solutions were left in the vessels for 3 hours and the experiment was repeated. The results are listed in Table 3. As observed, the potentials have drifted closer to the expected values, but are still over 50 my too low, again suggesting a large junction potential across the Nafion tube salt bridge.

Table 3. Potentials recorded between a saturated calomel electrode immersed in 0.5 M KCi(aq) and a range of electrode materials in a solution of fresh POM, where the two half cells were connected via the Nation tube salt bridge discussed in the text. The salt bridge was in contact with the two solutions for 3 hours before taking potential measurements. The time taken to obtain a stable reading is also Usted.

Electrode Material if / V vs. if I V vs. Stabilization BCE SHE time I mins 0107 0,9511<1 Platinum we 12 Glassy carbon disc 0721 0.965 <1 Glassy carbon rod 0.724 0968 c I Graphite rod 0.715 0.959 > 5 Edge plane pyrolytic graphite disc 0.721 0.965 C I Stainless steel sheet 0.655 0.890 > 5 These results suggests a KCI(aq) Nafion POM cell is not suitable for accurately measuring the redox potential of the POM.

Example 2

The same experimental set up in Figure 2 (Example 1) was used on a different day (i.e. an BCE and working electrode immersed in fresh POM in the same beaker). The results are listed in Table 4. As observed, the results are similar to those in Table I with all the electrodes (apart from the graphite rod) gMng a POM redox potential of 1.024 +/0.008 V vs. SHE.

Table 4. Potentials recorded between a saturated calomel electrode and a ranqe of electrode materials in a solution of fresh POM. The time taken to obtafri a stable reading is also listed, ectrodeMateri&iE vs E"TVs Sta BCE SHE timelmins Platinum disc 0.772 1.O16<1 Platinum wire 0373 1017 <1 Glassy carbon disc 0779 1023 c Glassy carbon rod 0786 1.032 <1 Graphite rod 0759 It 003,> 2 1 Edge pne pyrolytic graphite disc 0.776 1102 C I Next, the experimental set up from Figure 3 was used, except this time fresh POM was placed in both glass vessels, thus eflminating the junction potential across the Nation tube salt bridge. The same SCE used in Table 2 was placed in the left hand vessel and a range of different eiectrodes were placed in the right hand vessel. The two electrodes were then connected to a digftal voltmeter, with the saturated calomel electrode connected to the negative terminal and the other electrode material connected to the positive terminal. Table lists the different electrode materials used, the reading obtained from the digital voltmeter, E (against the saturated calomel electrode), and the time taken for the reading to stabilize.

Also listed is the potential vs. the standard hydrogen electrode.

Table 5, Potentials recorded between a saturated c&omel electrode immersed in fresh POM and a range of electrode materials immersed in a solution of the same POM, where the electrodes were separated via the Nafion tube sat bridge discussed in the text. The time taken to obtain a stable reading a also hated, iIe E v vt E / V vs. Stabilization SCE SHE time / mine Platinum disc 0.750 0.994 C 3 Platinum wire 0.753 0.997 C I Glassy carbon disc 0.757 1.001 C 3 1036011.004 C I Graohrte rod 0740 10964 >4 Edge plane pyrolytic graphite disc 0.752 0.996 <3 As observed, the potentials measured are much closer to the results in Table 4 compared to the analogous case in Example 1. This suggests using POM on both sides of the Nation salt bridge has removed a significant junction potential. However, the potentials in Table 5 are still over 20 my too low, suggesting the presence of another significant junction potentiaL The most logical location for this junction potential is the saturated KCl porous frit POM junction at the boundary between the SCE and the fresh POM. Hence, another experiment was conducted using a GC rod in place of the SCE, with all the other experimentai details the same as above. Table 6 lists the different electrode materiais used, the reading obtained from the digital voltmeter. E. and the time taken for the reading to stabilize, Table 6. Potentials recorded between a glassy carbon rod immersed in fresh POM and a range of electrode materials immersed in a solution of the same POM, where the electrodes were separated via the Nafion tube salt bridge thscussed in the text. The time taken to obtain a stable reading is also listed.

Mteria Elm Stabilization time! mins Platinum disc 3.0 <3 Platinum wire -0.7 <2 Glassy carbon disc 0.9 <1 Glassy carbon rod *-0.6 c 3 Graphite rod -40 > 3 Edge plane pyrolyUc graphite disc -4.4 <3 In the absence of any junction potentials, the voltmeter reading should be 0 V1 as the same redox couple is on both sides of the salt bridge (ignoring any electrode surface effects).

Apart from the graphite rod, all the electrode materials gave a potential of 0 +I 5 my, in agreement with that expected. This suggests a similar experimental set up could be used to measure the redox potential of a different POM (i.e. a reduced POM), as long as the potential of the fresh POM is known. As before, the graphite rod is thought to perform badly because of its microporous nature.

The experiment was repeated using a Pt wire as the negative electrode (in place of the glassy carbon rod) with all the other experimental details the same as above. Table 7 lists the different electrode materials used, the reading obtained from the digital voltmeter, E, and the time taken for the reading to stabilize. As before, all the electrode materials (apart from the graphite rod) result in a potential of 0 +1-5 mV, in agreement with the expected result.

Table 7. Potentials recorded between a platinum wire immersed in fresh POM and a range of electrode materials immersed in a solution of the same POM, where the electrodes were separated via the Nation tube salt bridge discussed in the text. The time taken to obtain a stable reading is also sted, Electrode Material El nW Stab llzation tUne! mins Platinum disc -2.8 <3 Platinum wire -0.5 <1 Gssy carbon disc -0.5 <1 Glassy carbon rod -1.6 c 1 Edge plane pyrolyhc grapftte dsc -4 8 <2

Example 3

The fIndings from Example 2 were used to investigate if the redox potential of a reduced POM can be measured using the Nation bridge approach. First the redox potential of a reduced POM was measured using the experimental set-up in Figure 2, where an SCE and another electrode were immersed in the same sokition of reduced POM, The SCE was connected to the negative terminal of the digital voltmeter and the other electrode was connected to the positive terminaL Table 6 US the different electrode materials used, the reading obtained from the digital voltmeter, E (against the saturated calomel electrode), and the time taken for the reading to stabilize. The voltmeter reading was found to settle much faster with the reduced POM than with the fresh POM.

Table 5. Potentials recorded between a saturated calomel electrode and a range of electrode materbals in a solution of reduced POM. The time taken to obtain a stable reading is also listed.

Electrode Material El V vs. E I V vs. Stabilization SCE SHE time! mine P!atinumdiscO.610 Platinum wire 0.612 0.856 <1 Glassy carbon disc 0.610 0.854 <1 Glassy carbon rod 0.611 0.855 < I Edge plane pyrolytic graphite disc 0.811 0855 < I Next. the experimental set up in Figure 3 was used with fresh POM on the left side of the Nation bridge and reduced POM on the right side. A glassy carbon rod was placed in the fresh POM and connected to the negative terminal of a digita voltmeter. From Table 4, we know this etectrochemical half cell has a potential of 0.788 V vs. SCE or 1.032 V vs. SHE. A range of other electrode materials were placed in the reduced POM and were connected to the positive terminal of the digital voltmeter. Table 9 lists the different electrode materials used, the reading obtained from the digital voltmeter, E, and the lime taken for the reading to stabifize. Also listed is the potential of the reduced POM (EoM) vs. the SCE and the standard hydrogen electrode calculated using the potential of the glassy carbon fresh POM half ce in Table 4 (0186 V vs. SCE or 1.032 V vs. SHE).

Table 9. Potentials recorded between a glassy carbon rod immersed in fresh POM and a range of eiectrode materials immersed in reduced POM, where the electrodes were separated via the Naflon tube salt bridge as discussed in the text. The ootential readings have also been corrected to SCE and SHE readings. The Ume taken to obtain a stable reading is also hsted.

Electrode Material E I my vs. Fresh POM time I mins vs. SCE vs. SHE Platinum disc 175.6 ci 0.6124 0.8564 Platinum wire 175,6 < 1 0.5124 0.6564 Glassy carbon disc i75,6 c 1 0.6124 08564 Edge plane pyrolytic 175.4 <1 06126 0.8566 graphite disc As observed, there is excellent agreement between the second column in Table 8 and the fourth column in Table 9, and between the third column in Table B and the final column in Table 9. ThIs suggests that a suitable Sectrode material immersed in POM can be used (in counction with a suitabie salt bridge) to measure the redox potential of a different POM.

This is a much simpler and cheaper redox probe than the currently used dummy cell.

To summarize, the above examples clearly demonstrate that more accurate redox state measurements can be made when both sides of the salt bridge (for example the Nation tube assemby) are chemicay similar. This similarity may refer to the actual chemical species or the pH.

Example 4

Figure 4(a) illustrates a schematic diagram of a reference electrode made using the findings above. The electrode consists of a Nation tube filled with silastic elastomer. essentially preventing any fluids from entering the interior of the tube. The filled dry Nation tube is then inserted into a FEP tube of 0.125 inches internal diameter and 0.25 inches external diameter tube, forming a snug fit. Soaking the tube assembly in water causes the Nation to expand thus forming a Ught seal with the FEP tube. At the top of the Naflon tube assembly there is a small reservoft of &ectroiyte (in this case fresh HV4 POM, 0,3 M) that wets one end of the Nation tube. A suitable electrode material is also immersed in the electrolyte, forming the reference electrode half cell (in this case the material is a glassy carbon rod). The reservoir is sealed, The photograph in Figure 4(b) shows how the reference &ectrode can be made using standard Swagelok materials (PFA fithngs). The bottom of the reference electrode is then immersed into the solution of interest and used in conjunction with other electrodes.

Figure 5 illustrates a cyclic voltammogram of 0.3 M Nay4 POM using an edge plane pyrolytic graphite working electrode, a platinum wire counter electrode and the reference electrode in Figure 4(b). Also shown is a cyclic voltammogram obtained using an SCE as the reference electrode, with all other experimental detalls the same. As observed, the two voitammograms are identical but shifted apart by 789 my. In a separate experiment, the same SCE and POM reference electrodes were both placed in the same 0.3 M NW4 POM solution and the potential difference between the two electrodes was measured using a digital voltmeter The reading was 782 my, which agrees with the potential shift in Figure 5.

The results clearly demonstrate that the POM based reference electrode can be used for stationary electrochemistry applications, such as cyclic voltammetry. Furthermore, the traditional' reference electrodes with porous frits as salt bridges (e.g. SCE, Ag/AgGI) often contaminate solutions with unwanted ions (such as chloride anions from SCE electrodes, which often contaminate the test electrolytes). Due to the nature of the Nation salt bridge, the POM based reference electrodes cause very little contamination of the test solutions.

Furthermore, the POM based reference electrodes can be designed such that they only contain ions that are present in the test solutions, so even if ion crossover occurs, there

Example 5

Figure 6 illustrates a schematic diagram of a POM based redox probe that can be used to measure the potential of a working POM solution in a flowing environment. The glassy carbon rod on the right acts as the connection for the reference electrode (in stationary solution) and the glassy carbon rod on the left is in contact with the flowing electroiyte.

Connecting a digital voltmeter to both rods allows the redox potential of the flowing catholyte to be determined. The same Nation tube assembly in Example 4 is used as the salt bridge.

This redox probe was used on a FlowCath fuel cell rig for over 24 hours and gave consistent results.

Example 6

Three redox probes were constructed using a Hydroflex hydrogen reference electrode in 0,5 M H2S04, almost fuHy oxidised aqueous 0,3 M H7PVMo5O40] (HV4)/glassy carbon and regenerated HV4/glassy carbon as the reference systems. Details of the redox probes are

given in Table 10.

Table 10. Details of the redox probes used in ExampleS.

Reference Reference system Working electrode Reference electrode electrode number potential vs. _________I_______________ _____________ Hydroflex I V I Hydroflex Glassy Carbon Rod 0.0 (Hydrogen/Pt) (GLR) protruding into immersed ri 0.5 M catholyte stream H2504 2 Fresh H')4 / Glassy GLR protruding into 1.041 Carbon Rod (GLR) catholyte stream 3 Regenerated H')4 / GLR protruding into 0853 GLR cathclyte stream AD reference electrodes were constructed using an FEP sheathed, sUastic silicone filled Nation tube and buift as shown Figure 6. The Nation tube was of sufficient length so as to keep the reference electrodes thermafly isolated from the catholyte steam.

An ACAL Energy 21 cm2 active area single cefi fuel ccli was assembled and attached to a test rig. The cathode consisted of a 50 x 62 mm piece of SOL Carbon GFD2.5 graphite felt sat in a weli of 1.1 mm, which was engraved into the carbon cathode block, The cathode block also housed the catholyte inlet and outlet channels with a plug4low flow field. The cathode was sealed against a Johnson Matthey JMV3+ membrane electrode assembly of 21 cm2 active area using a 1.1 mm silastic seal placed in a seal groove around the cathode material. The cathode side of the membrane was naked whereas a standard platinum loading was used on the anode side, The anode consisted of a 50 x 62 mm piece of SGL Carbon GDL 346C surrounded by a 0.3 mm thick hard silicon gasket on top of a serpentine flow field engraved into the anode carbon bbck. Current cohectars were placed at each end of the fuel cell and the whole assembly was compressed between two steel end plates. The test rig was modified so that the three reference electrodes were plumbed into the catholyte loop between the pump and the fuel cell inlet. The reference electrode voltages were monitored using a Squirrel data logger. Squirrel Chann& 4 was also connected wfth the positive terminal connected to the Hydroflex reference &ectrode and the negative terminal connected to the fuel cell anode. The system was charged with 250 ml of MV4 polyoxometalate catholyte and sealed. The system temperature was maintained via the reservoir jacket heater and the cell heaters. All quoted temperatures are those taken from the fuel cell inlet thermocouple.

For all open circuit voltage (OCV) measurements, the fuel cell anode exhaust valve was open and the anode was suppUed with a flow of 25 ml miri1 hydrogen, The catholyte flow rate was approximately 150 ml rnin at all times. Measurements were taken when a stable cell CCV and stable voltages at the reference electrodes were observed. When required, the catholyte was reduced using the fuel cell, the hydrogen flow was then increased and the exhaust valve was closed.

All data was logged using the test rig andlor the squfrrel data logger. The values of the voltage recorded by the squirrel data logger were calibrated post experiment to match those taken sporadically using a cahbrated DVM. The arrangement of the fuel cell rig including the reference electrodes is shown in Figure 7.

The catholyte was circulated around the system at room temperature until a stable cell OCV and stable voltages at the reference electrodes were observed. The system was then heated to 40°C, 60°C and 80°C, and allowed to stabilise at each temperature. The system was then allowed to cool to assess the extent of any slow reduction of the catholyte due to a short circuit. The catholyte was reduced to the point at which the Hydroflex E 1 00 V. The system was heated stepwise to 80°C and measurements were taken at the previously defined temperatures. The catholyte was then reduced to E 080 V and allowed to cool stepwise to 20°C, Finally the catholyte was reduced to E 070 V and heated stepwise to 80°C.

As shown in Figures 8 to 11, the redox potentials measured using the reference electrodes all show similar responses to catholyte temperature and redox state (the potential of the catholyte varies from Figure 8 to 11, with the following approximate values: 1.04 V in Figure 8, 1 V in Figure 9, 0.8 V in Figure 10 and 0.7 V in Figure 11), This indicates that the reference electrode configuration of the present invention is robust. The measurements made with the Hydrofiex and fresh HV4 electrodes are especially coherent, with a maximum difference of 4 mV. The potentials recorded with the regenerated HV4 &ectrode seem to have drifted negatively from those taken with the Hydrofiex and fresh HV4 as the experiment proceeded. This could be due to oxidation of the reference catholyte, which would cause a negative measurement error of this sort.

The open circuit potential measured using the fuel ceO varies substantiaUy from the redox potentials of the catholytes measured using the reference electrodes. This is illustrated in Figures 12 to 15, where the ce potential and the Hydroflex results from Figures 6 to 11 are plotted on the left hand yaxis. As before, the potential of the catholyte varies from Figure 12 to 15, with the following approximate values: 104 V in Figure 6, 1 V in Figure 9, 0.8 V in Figure 10 and 0.7 V in Figure 11. Apart from Agure 15, the decrease in potential with temperature is noticeably sharper for the cell than for the catholyte, In Figure 15 the cell potential is almost independent of temperature whereas the catholyte potential increases with temperature The reference probe assembly also allows the potential of the anode to be measured, independently of the cathode (at open circuit). These results, (i.e. the potential of the anode vs. the Hydroflex reference electrode) are plotted using the right hand axis in Figures 12 to 15 and explain the difference between the cell potential and the catholyte potential. In all of Figures 12 to 15. the anode potential increases with temperature. Given that the cell potential is the cathode (or catholyte) potentia minus the anode potential, this explains the differences between cell potential and catholyte potential. Thus, the reference probe housing can be used to decouple anode and cathode temperature effects, allowing greater insights into the operation of the fuel cell.

xample I Figure 16 illustrates an alternative redox probe housing to that in Figure 6. In Figure 16, the redox probe housing is made of graphitic carbon. In this case the material is an impermeable form of graphite used in bipolar plates (JP945 manufactured by Mersen, which is an electrically conductive carbon composite material. A generic example of such a material is carbon particles held in a resin matrix). The carbon block has 3 ports one for the electrolyte inlet, one for the electrolyte outlet and one for the reference electrode. These are seen in more detail in the cutaway" drawing. The electrolyte flows through the carbon block via the inlet and outlet and the reference electrode protrudes into the catholyte stream.

In this case, the carbon block is the working electrode. The potential is then measured between the carbon block and the reference electrode. A conductive pin can be placed in a hole in the carbon block for the potential measurement, Figure Il shows more detailed drawings of the alternative probe housing.

Example 8

The alternative redox probe housing was used in a similar fuel cell system to that described in Example 6. The housing was connected to the catholyte flow just after the outlet of the fuel cefi. A Nafion tube bridge similar to that described in Example 6 was secured into the reference electrode port, such that the Nation tube protruded into the catholyte flow. The other end other of the Nafion tube was connected to a 0.5 M H2504 reservoir. into which a Hydroflex reference electrode was immersed. The fuel cell was used to reduce the catholyte and the voltage between the carbon block and the Hydroflex reference electrode Was recorded over the course of the experiment. Figure 18 shows the potential of the catholyte (measured via the redox probe) plotted against the number of moles of electrons passed.

Example 9

The redox potentials of solutions of 1.6 M vanadium in 4 M H2804 were measured using the reference electrodes of the present invention. The solutions were made by dissolving vanadium oxides in H2802 and the oxidation state of the vanadium was changed using electrochemical reduction or electrochemical oxidation. The solutions were placed ri a beaker. A working electrode and reference electrode were then immersed in the beaker and a digital voltmeter was used to measure the potential between them. The working electrode was selected from a 3 mm diameter glassy carbon disc electrode, 2 mm diameter Pt disc electrode, 3 mm diameter EPPG disc electrode, 3 mm diameter PFCE disc electrode and a 3 mm diameter rod of JP945 (manufactured by Mersen). The reference electrode was either a Hydroflex reference electrode immersed in a reservoir of 0.5 M H2804 and connected to the vanadium solution via a Nafion tube bridge, as used in the previous examples. or the reference electrode shown in Figure 4b where the reference electrolyte reservoir was filled with fully oxidised 0.3 M Na4Ha[PMouVO4o], The latter is referred to as a POM reference electrode.

Table 11 lists the measured potentials for the different combinations of reference electrodes and working electrodes immersed in a solution of 1.6 M V5 in 4 M 112804. The potential for all the combinations settled within 60 seconds of immersion into the solution. As shown, alt the measured potentials are in good agreement with each other, suggesting all the working &ectrodes are suitable materials to use for solutions of V4. In addition, the difference between the Hydroflex and POM reference electrode results, E(POM+!ydroflex), is close to 1.019 V, which was the voltage recorded between the 2 reference electrodes.

Table 11. Results of potential measurements between different combinations of reference electrodes and working electrodes immersed in a solution of 1.6 M V' in 4 M H2804.

Working Electrode Glassy Platinum EPPG PFCE [JP945 Carbon E vs. POM 255,6 255 255 254.9 254.4 reference I my InN E(POM 1011 1016 1018 1018 1022 Hydroflex) I my Settled within Yes Yes Yes Yes Yes 60s? . I Table 12 Vets the measured potenflals for the different combiriafions of reference &ectrodes and working &ectrodes immersed in a solution of 0.8 M \P and 0.8 M V in 4 M H2S04.

The potential for ail the combinations settled within 60 seconds of immersion into the sok4tion. As shown, all the measured potentials are in good agreement with each other, suggesting a the working electrodes are suitable materials to use for solutions of mixtures of V and V5 in similar proportions.

Table 12. Results of potential measurements between different combinations of reference electrodes and working electrodes immersed in a solution of 0.8 M \P and 0.8 MV5 in 4 M H2S04.

Working Electrode Gassy Platinum EPPG PFCE JP945 Carbon E vs. POM 83.3 83,3 83.4 83.4 83,4 reference! mV EvtHydroflex! 1104 1104 1104 1104 1104 mY E(POM-1021 1021 1021 1021 1021 Hydroflex) / mY [Sethed within GO Yes Yes Yes Yes Yes Table 13 hsts the measured potentials for the different combinations of reference electrodes and working electrodes immersed in a solution of t6 M V4 in 4 M H2S04. Apart from the JP945 rod, the potential had not settled wIthin 2 minutes of immersion, which was the time when the voltmeter reading was recorded. The voltmeter reading when the glassy carbon electrode was used was very unstable, whereas the other materials appeared to have almost sethed by 2 minutes. As observed, the difference between the POM and Hydroflex readings are all noticeably lower than the anticipated 1019 niV, apart from the JP945 result.

This suggests that for certain soutions (i.e. high percentage of V) some electrode materials are more suitable than others i.e. JP945 is more suitable than glassy carbon for measuring the potential of V solutions, Table 13, Results of potential measurements between different combinations of reference electrodes and working electrodes immersed in a solution of 1.6 M V4 in 4 M H2SO.

Working Electrode Glassy I, Platinum EPPG PFCE JP945 Carbon E vs. POM -265 -382 -380 -370 -419 reference! mY E vs. Hydroflex 382 615 595 570 599 /nN E(POM-647 997 975 940 1018 Hydroflex)! my SeWed within No No No No Yes 120s?

Claims (10)

1. Claims 1. A redox probe for measuring the redox state of a test electrolyte compri&ng: a first electrode in contact with the test electrolyte; a second electrode, electricafly connected to the first electrode, in contact with a reference &ectrolyte: and means for effecting ion transfer extending between the reference electrolyte and the test electrolyte; wherein the reference electrolyte has a pH value within 4 pH units of the pH of the test electrolyte.
2. 2. The redox probe of Claim I wherein the reference &ectrolyte has a pH value within 3.5. 30, 25, 2.0, 1,5, 1.0 or 0.5 pH units of the pH of the test electrolyte.
3. 3. The redox probe of Claim 1 or 2 wherein the reference electrolyte has the same pH value as that of the test electrolyte.
4. 4. The redox probe of any preceding claim wherein the reference electrolyte comprises one or more compounds that are present in the test electrolyte.
5. 5. The redox probe of any preceding claim wherein the reference electrolyte comprises one or more compounds having the same chemic& composition as one or more compounds that are present in the test electrolyte.
6. 6. The redox probe of Claim 5 wherein the one or more compounds having the same chemical composition do or do not have the same oxidation state in the reference electrolyte and in the test electrolyte at any given time,
7. 7. The redox probe of any one of Claim 4 to 6 wherein the reference electrolyte consists of one or more compounds that are present in the test electrolyte.
8. 8. The redox probe of any preceding claim wherein the reference electrolyte comprises a polyoxometalate.
9. 9. The redox probe of Claim B wherein the reference electrolyte comprises a highly oxidised polyoxometalate.
10. 10. The redox probe of Claim 8 wherein the reference electrolyte comprises a highly reduced polyoxometalate.11 The redox probe of any preceding daim wherein the first andlor the second electrodes are solid electrodes.12. The redox probe of Claim 11 wherein the first and/or the second electrodes comprise a disc, rod, plate, foam or other porous material.13. The redox probe of Claim 11 or 12 wherein the first and/or the second electrodes comprise glassy carbon, graphite, impermeable forms of graphite, other graphite materials (such as a graphite/polymer composite such as resin-impregnated graphite), platinum, boron doped diamond or metais such as titanium or stainless steel.14. The redox probe of any preceding claim wherein the means for effecting ion transfer comprises a salt bridge.15. The redox probe of Claim 14 wherein the salt bridge comprises a proton exchange material, ion exchange material polymer electrolyte, pertluorinated ionomers, pertluorosulphonic acid polymer such as a perfluorosulfonic acid (PFSA) polymer membrane material or a hydrocarbon ion exchange membrane material.16. The redox probe of Claim 14 or claim 15 wherein the salt bridge comprises a rod, a tube or a sheet.17. The redox probe of Claim 16 wherein the salt bridge comprises a tube filled with a curable elastomer.18. The redox probe of any preceding claim wherein the first electrode comprises a hollow body through which the test electrolyte can flow.19. A fuel cell, a redox battery, redox flow battery or an electrolyser including the redox probe of any one of Claims 1 to 18, wherein the test electrolyte comprises an electrolyte of the fuel cell, redox battery or electrolyser.20, The fuel caD, redox battery or electrolyser of Claim 19 wherein the test electrolyte comprises the catholyte or anolyte of the fuel cell, redox battery or electrolyser.21. The fu& cell, redox battery or electrolyser of Claim 20 wherein the catholyte or anolyte comprises a polyoxometalate 22. An &ectronic, automotive and/or combined heat and power equipment comprising the fuel cell. redox battery or electro!yser of any one of Claims 19 to 21.23. A use of a redox probe comprising: a first electrode that can contact a test electrolyte: a second electrode, electrically connected to the first Sectrode. in contact with a reference electroWte; and a means for effecting ion transfer able to extend between the reference electrolyte and the test electrolyte: to measure the redox potential of the test electrolyte; wherein the reference electrolyte has a pH value within 2 pH units of the pH of the test electrolyte.24. The use of Claim 23 wherein the test electrolyte is an electrolyte in a fuel cell, redox battery, redox flow battery or electrolyser.25. The use of Claim 24 wherein the test electrolyte is a catholyte or anolyte in a fuel cell, redox battery, redox flow battery or electrolyser.