The present invention relates to fuel cell systems, preferably but not exclusively incorporating a liquid electrolyte, for example an alkaline electrolyte, and to methods of operating such fuel systems.

Background to the invention

Fuel cells have been identified as a relatively clean and efficient source of electrical power. Alkaline fuel cells with a liquid electrolyte are of particular interest because they operate at relatively low temperatures, are efficient and mechanically and electrochemically durable. Acid fuel cells and fuel cells employing other liquid electrolytes are also of interest. Such fuel cells typically comprise an electrolyte chamber separated from a fuel gas chamber (containing a fuel gas, typically hydrogen) and a further gas chamber (containing an oxidant gas, usually air). The electrolyte chamber is separated from the gas chambers using electrodes. Typical electrodes for alkaline fuel cells comprise a conductive metal, typically nickel, that provides mechanical strength to the electrode, and the electrode also incorporates a catalyst coating which may comprise activated carbon and a catalyst metal, typically platinum.

In operation, it is conventional to operate the fuel cells at a temperature above 50 °C, for example at between 60° and 70 °C, because the efficiency of the cells is greater at such a temperature than at near ambient temperature. However operation at such a temperature leads to evaporation from the aqueous electrolyte, such as potassium hydroxide; and so there is a risk of forming crystals for example of potassium carbonate or potassium hydroxide. Such crystals are detrimental to the operation of the cell because they inhibit gas diffusion.

Furthermore, degradation of the electrode materials may also occur during operation.

A single fuel cell typically operates at below 1 .0 V, and it is therefore usual in practical applications to combine fuel cells into a fuel cell stack, that is to say a stack of fuel cells. In such a stack the fuel cells can be connected electrically in series, to increase the output voltage. Such a stack may for example consist of between two fuel cells and two hundred or more fuel cells, more typically between eight fuel cells and one hundred fuel cells. However it has been found that temperature non-uniformity within such a stack may have a detrimental effect on cell performance in the long term. Discussion of the invention

The fuel cell system of the present invention addresses or mitigates one or more problems of the prior art.

According to the present invention there is provided a fuel cell system comprising multiple fuel cells forming a fuel cell stack with output terminals to provide electric power to a load in a load circuit, wherein the fuel cells at each end of the stack are electrically connected to an end-cell circuit, whereas the remaining fuel cells of the stack are electrically connected to the output terminals of the stack.

In a second aspect the invention provides a method of operating a fuel cell system comprising multiple fuel cells forming a fuel cell stack with output terminals to provide electric power to a load in a load circuit, wherein the method comprises electrically connecting the fuel cells at each end of the stack to an end-cell circuit, while electrically connecting the remaining fuel cells of the stack to the output terminals of the stack.

When the load is connected to the terminals, the fuel cell system provides electric power to the load in the load circuit.

There may be a single end-cell circuit, to which both the end cells are connected, or there may be one end-cell circuit for each end cell. The or each end-cell circuit does not form part of the load circuit. Each end-cell circuit may for example carry a current up to the maximum operating current of the fuel cell, but typically the current density is no more than 0.5 times and preferably no more than 0.2 times the maximum fuel cell current density, more preferably no more than 0.1 times the maximum fuel cell current density when in steady-state operation at peak operating efficiency. Each end-cell circuit may include a resistor to dissipate electrical energy. Alternatively the end-cell circuit may be used to supply electrical power to a secondary external load.

Each fuel cell may be a liquid electrolyte fuel cell, for example an alkaline fuel cell, for example with an aqueous alkaline electrolyte. In any event each fuel cell comprises electrolyte between two electrodes, an anode and a cathode. During operation the anode and the cathode are provided with different gases: an oxidant and a fuel.

Typically, each cell comprises an electrolyte chamber between two electrodes that are spaced apart, one electrode being an anode and the other electrode a cathode, the anode separating the electrolyte chamber from a gas chamber through which a fuel gas is passed (which may be referred to as an anodic gas chamber), and the cathode separating the electrolyte chamber from a gas chamber through which an oxidising gas is passed (which may be referred to as a cathodic gas chamber). The system includes a fuel gas supply duct to supply a fuel gas to all the anodic gas chambers in the stack, and an oxidising gas supply duct to supply an oxidising gas to all the cathodic gas chambers in the stack. So, in the method, a fuel gas may be supplied through the fuel gas supply duct to all the anodic gas chambers in the stack, and an oxidising gas may be supplied through the oxidising gas supply duct to all the cathodic gas chambers in the stack.

The gas supply system is thus straightforward, as all the cells of the stack are supplied with fuel gas through the same fuel gas supply duct, and all the cells are supplied with oxidising gas through the same supply duct. The electrical connections are somewhat different, as the end cells are connected to an electrical circuit which is different to that that to which the remaining cells are connected. Since each end cell is connected to an end-cell circuit, the current through the end cells can be controlled independently of the current through the remaining cells. Typically the fuel cell stack operates at an elevated temperature which may be between 50 °C and 90 °C, more preferably between 60 °C and 70°C. The current in the or each end-cell circuit may be adjusted, during operation, in accordance with the temperature difference between the end cells of the stack and the remaining fuel cells of the stack. It has been found that if all the cells carry the same current, the end cells have a lower voltage at the start of operation; and that there is a tendency for the end cells in a stack to deteriorate more rapidly than the other cells, if all the cells carry the same current. These differences are assumed to be due to the difference in heat transfer from the cells, as the end cells are not blanketed by adjacent cells, and may therefore be cooler than the remaining cells in the stack. If the end cells carry a lower current, this deterioration is suppressed, while the electric current ensures that some heat is generated in the end cells. In one example the current density provided by the fuel cells that are connected to the load may be more than 100 mA/cm ² , for example 125 mA/cm ² , whereas the current density in the end cells may be less than 20 mA/cm ² , for example 10 mA/cm ² . More generally, each end-cell circuit is typically arranged such that the current density in each end cell is no more than 0.5 times that of the remaining fuel cells, preferably no more than 0.2 times that of the remaining fuel cells, for example no more than 0.1 times that of the remaining fuel cells.

The invention will now be further and more particularly described, by way of example only, and with reference to the accompanying drawings in which:

Figure 1 shows a cross-sectional view of a fuel cell;

Figure 2 shows a cross-sectional diagram of a fuel cell stack;

Figure 3 shows a schematic diagram of an alternative arrangement of fuel cells to form a stack;

Figure 4 shows a fuel cell system incorporating the fuel cell stack of figure 2;

Figure 5 shows a modification to the fuel cell system of figure 4; and

Figure 6 shows a fuel cell system incorporating a fuel cell stack arranged as in figure 3.

Cell Structure

Referring to figure 1 , a fuel cell 25 consists of two electrodes 10a, 10c which are spaced apart to define an electrolyte chamber 26 between them. By way of example each electrode 10a, 10c may comprise a sheet 1 1 of a metal such as nickel or ferritic stainless- steel, which may be of thickness 0.3 mm. In this embodiment most of the sheet - the central region 12 - is perforated for example by drilling, punching or by chemical etching to produce a very large number of through holes 14, while a margin 15 around the periphery of the sheet 1 1 , of width 5 mm, is not perforated.

After forming the through holes 14, one surface of the perforated central region 12 is then covered in a gas-permeable layer 16; the exposed surface of the gas-permeable layer 16 is then covered with a layer 18 which contains catalytically active material. The metal sheet 1 1 and the layers 16 and 18 are bonded together. In this example the gas-permeable layer 16 is of thickness 350 μιη, while the catalyst-containing layer 18 is of thickness 250 μιη. One electrode 10a is an anode, and the other electrode 10c is a cathode, so the catalytically active materials in the layers 18 may be different.

The gas-permeable layer 16 comprises carbon with a hydrophobic binder. The catalyst-containing layer 18 is similar, combining carbon with an appropriate active catalytic material in particulate form, and with a hydrophobic binder. In each case one suitable binder would be polytetrafluoroethylene (PTFE). By way of example the electrode for the anode may incorporate 10% palladium or 10% palladium/platinum on activated carbon, optionally combined with carbon to provide electrical conductivity, the catalyst-loaded activated carbon constituting between 50% and 100% of the total mass of carbon; and the electrode for the cathode may incorporate carbon for electrical conductivity combined with a spinel Mn _x Co _y 0 ₄ , which is a catalytic material, with the spinel forming up to 80% of the total mass of the layer. A wide range of other catalytic materials may be used instead. The catalyst-containing layer 18 incorporates a smaller proportion of PTFE binder than does the gas-permeable layer 16. The catalyst-containing layer is therefore less hydrophobic; and may indeed incorporate hydrophilic components. The electrode 10a that is an anode separates the electrolyte chamber 26 from a gas chamber 27 through which a fuel gas such as hydrogen is passed, while the other electrode 10c, which is a cathode, separates the electrolyte chamber 26 from a gas chamber 28 through which an oxidising gas such as air is passed. The anode 10a and the cathode 10c have the structure described above, but as mentioned above, may differ in the nature of the catalytic material.

The electrolyte chamber 26 is defined by a circumferential frame 30 which is sealed by gaskets 32 onto the non-perforated margin 15 of each electrode 10 (the frame 30 and the gasket 32 are shown by broken lines). The gaskets 32 are stepped so they enclose the periphery of the layers 16 and 18, so holding those layers 16 and 18 securely onto the metal sheet 1 1 .

In operation of the fuel cell 25 a fuel gas such as hydrogen is supplied to the gas chamber 27 and air is provided to the gas chamber 28 respectively, and an electrolyte solution such as aqueous potassium hydroxide solution is supplied to or passed through the electrolyte chamber 26. The fuel cell 25 provides an electromotive force of slightly less than 1 V, for example between 0.6 V and 0.8 V. Within the cell 25 the electrolyte permeates through the outer portion of the catalyst-containing layer 18 (which is at least partly hydrophilic) and so contacts the catalytic material in the catalyst-containing layer 18 of each electrode 10.

Cell Stack Structure

Since the electromotive force of a single working fuel cell 25 is less than 1 V, fuel cells 25 are typically formed into fuel cell stacks, to provide a higher output voltage. Referring now to figure 2, there is shown a cross-sectional view through the structural components of a cell stack 200 with the components separated for clarity. The stack 200 consists of a stack of moulded plastic plates 202 and 206 arranged alternately. Each plate 202 is equivalent to the frame 30 shown in figure 1 , as it defines a generally rectangular through-aperture 26 surrounded by a frame 204; the apertures 26 constitute the electrolyte chambers 26, and immediately surrounding the aperture 26 is a 5 mm wide portion 205 of the frame which projects 0.5 mm above the surface of the remaining part of the frame 204. The plates 206 are bipolar plates; each defines rectangular blind recesses 27 and 28 on opposite faces, each recess being about 3 mm deep, surrounded by a frame 210 generally similar to the frame 204, but in which there is a 5 mm wide shallow recess 21 1 of depth 1 .0 mm surrounding each recess. The blind recesses 27 and 28 constitute the gas chambers.

It will thus be appreciated that between one bipolar plate 206 and the next in the stack 200 (or between the last bipolar plate 206 and an end plate 230), there is an electrolyte chamber 26, with an anode 10a on one side and a cathode 10c on the opposite side; and there are gas chambers 27 and 28 at the opposite faces of the anode 10a and the cathode 10c respectively. These components constitute a single fuel cell, equivalent to the fuel cell 25 shown in figure 1 .

Electrodes 1 0a and 10c locate in the shallow recesses 21 1 on opposite sides of each bipolar plate 206, with the catalyst-carrying face of the electrode 10a or 10c facing the adjacent electrolyte chamber 26. Before assembly of the stack components, the opposed surfaces of each frame 204 (including that of the raised portion 205) may be covered with gasket sealant 215; this adheres to the frame 204 and dries to give a non-tacky outer surface, while remaining resilient. The components are then assembled as described, so that the raised portions 205 locate in the shallow recesses 21 1 , securing the electrodes 10a and 10c in place. The sealant 215 ensures that electrolyte in the chambers 26 cannot leak out, and that gases cannot leak in, around the edges of the electrodes 10a and 10c, and also ensures that gases cannot leak out between adjacent frames 204 and 210. The perforated central section 12 of each electrode 10 corresponds to the area of the electrolyte chamber 26 and of the gas chamber 27 or 28; the non-perforated peripheral margin 15 is sealed into the peripheral shallow recess 21 1 ; and the gas-permeable layer 16 with the catalytic coating 18 is on the face of the electrode 10 closest to the adjacent electrolyte chamber 26.

It will be appreciated that this cell stack 200 is shown by way of example only, and is a schematic view. For example the sealant 215 may be arranged in a somewhat different fashion from that shown, and the seal may instead be provided by solid gaskets or O-rings. Whatever the detailed arrangements of the cell stack 200 may be, in each case a single fuel cell consists of an electrolyte chamber 26 with electrodes 10a and 10c on either side which separate it from adjacent gas chambers 27 and 28. Within the stack 200 several fuel cells are arranged so as to be electrically in series, to provide a greater voltage than is available from a single cell.

The flows of fluids to the fuel cells follow respective fluid flow ducts, at least some of which are defined by aligned apertures through the plates 202 and 206. Only one such set of apertures 216 and 218 is shown, which would be suitable for carrying electrolyte to or from the electrolyte chambers 26 via narrow transverse ducts 220. The flows of the gases to and from the gas chambers (recesses 27 and 28) may similarly be along ducts defined by aligned apertures through the plates 202 and 206. Although the transverse ducts 220 are shown as being within the plates 202 they may instead be defined by grooves at the surface of the plates 202.

At one end of the stack 200 is a polar plate 230 which defines a blind recess 28 on one face but is blank on the outer face. Outside this is an end plate 240, which is also made of polymeric material, and which defines apertures 242 which align with the apertures 216 and 218 in the plates 202 and 206; at the outside face the end plate 240 also defines ports 244 communicating with the apertures and so with the fluid flow ducts through which fluids can flow to or from the stack 200, each port 244 comprising a cylindrical recess on the outer face. At the other end of the stack 200 is another polar plate (not shown) which defines a blind recess 27.

After assembly of the stack 200 the components may be secured together for example using a strap 235 (shown partly broken away) around the entire stack 200. Other means may also be used for securing the components, such as bolts. For example bolts (not shown) may extend through aligned apertures through all the sheets 202 and 206.

In operation of the fuel cell stack 200 gases such as hydrogen and air are provided to the gas chambers 27 and 28 respectively, as mentioned above in relation to figure 1 , and an electrolyte solution such as 6 mole/litre aqueous potassium hydroxide solution is passed through the electrolyte chambers 26. Recirculating the electrolyte solution through the fuel cell stack 200 and an electrolyte circulation duct has the benefit that the electrolyte carries away much of the heat generated during operation, and this excess heat can be removed (for example using a heat exchanger) from the circulating electrolyte. Each cell produces an electromotive force as discussed above, and the electrical output of the stack 200 depends on the number of fuel cells in the stack.

Referring now to figure 3, an alternative fuel cell stack 250 differs from that of figure 2 in not having bipolar plates 206 between successive cells. In several respects the features are the same, and are referred to by the same reference numerals. The stack 250 consists of a stack of moulded plastic plates 202, 255 and 256. The plates 255 and 256 are arranged alternately along the stack 250, and each is between successive plates 202. Each plate 202 defines a generally rectangular through-aperture 26, and the apertures 26 constitute the electrolyte chambers 26. The plates 255 define through apertures 258 which constitute oxidant gas chambers 28, whereas the plates 256 define through apertures 257 which constitute fuel gas chambers 27.

Each electrolyte chamber 26 has an anode 10a on one side and a cathode 10c on the opposite side; and there are gas chambers 27 and 28 at the opposite faces of the anode 10a and the cathode 10c respectively. These components constitute a single fuel cell, equivalent to the fuel cell 25 shown in figure 1 . At one end of the stack 250 is a polar plate 230 which defines a blind recess 28 on one face, and outside this is an end plate 240. At the other end of the stack 250 is another polar plate (not shown) which defines a blind recess 27 (equivalent to those in figure 2). In the stack 250 each oxidant gas chamber 28 has cathodes 10c on both sides, while each fuel gas chamber 27 has anodes 10a on both sides.

The fuel cell stack 250 operates in substantially the same way as that described above, but the electrical connections within the stack 250 between successive fuel cells are slightly different, because alternate cells are facing opposite directions. Electrical connections is made to each anode 10a and cathode 10c by a projecting tab 260, and these are connected in pairs by links 262, so that the cells are in series. The end link 262 is connected to an output terminal 264. One such arrangement of the links 262 is shown in figure 3, and is explained below in more detail in relation to figure 6. Electrical Connections

Referring now to Figure 4 there is shown a fuel cell system 300 which in this example incorporates the fuel cell stack 200. Each electrode 10 within the fuel cell stack 200

incorporates a projecting tab 301 (not shown in figure 2) to which electrical connection can be made outside the fuel cell stack 200. In the arrangement shown in figure 3 the third tab 301 in from each end of the stack 200 is connected to an output terminal 302; while the intervening tabs 301 are joined together in pairs by connectors 303, each such pair of tabs 301 linking the electrodes 1 0a and 10c on opposite sides of a bipolar plate 206 (as shown in figure 2). The output terminals 302 are connected electrically through a switch 304 to a load 305, which for example may be an electric motor. The circuit including the output terminals 302, the switch 304, and the load 305 constitutes a load circuit 306.

The last two projecting tabs 301 at each end of the fuel cell stack 200 connect to the anodes 10a and the cathodes 10c of the fuel cells 25 at the ends of the stack 200. In this example these two end cells are connected electrically to a pair of secondary output terminals 310: each end tab 301 is connected directly to a respective one of the secondary output terminals 31 0. The secondary output terminals 310 are connected to a respective end-cell circuit 312 which is shown as incorporating a switch 313 and an adjustable resistor 314.

In operation of the fuel cell system 300, with the switch 313 closed, the adjustable resistor 314 is used to ensure a desired current flows through the end-cell circuit 312, for example corresponding to a current density at each electrode 10 of 10 mA/cm ² . Similarly, with the switch 304 closed, the fuel cell system 300 provides electrical power from the output terminals 302 to the load 305, and the current density at each electrode 10 may be

considerably larger, for example 1 10 mA/cm ² or 125 mA/cm ² , depending on the load 305.

Operating in this manner, with a much smaller electrical current flowing through the cells 25 at the ends of the stack 200, has been found to enhance long-term performance. In particular this has been found to suppress the deterioration of individual cells 25 within the stack 200, so that the stack 200 can operate for longer.

It will be appreciated that the fuel cell system 300 may be modified in various ways. For example, as shown in figure 5, a fuel cell system 400 differs from the fuel cell system 300 in that rather than connecting each end cell to a respective end-cell circuit 312, the two end cells are instead connected in series to a single end-cell circuit 312. This may be achieved by connecting each end tab 301 directly to a respective one of the secondary output terminals 310, while the second tabs 301 in from each end are connected together by a connector 308. The secondary output terminals 310 are connected to the end-cell circuit 312, which as described above incorporates a switch 313 and an adjustable resistor 314. In another alternative, as shown in figure 6, a fuel cell system 500 may incorporate a fuel cell stack 250 as shown in figure 3. In this case the end projecting tabs 260 may be connected to respective secondary outlet terminals 310. The end link 262 is connected to an output terminal 264 at each end of the stack 250, and each output terminal 264 is connected not only to an output terminal 302 but also to a secondary output terminal 310. As described in relation to figure 4, the output terminals 302 are connected electrically through a switch 304 to a load 305, which for example may be an electric motor, so that those terminals 302 provide the output power from the fuel cell system 400.

The secondary output terminals 310 at each end of the stack 250 are connected to a respective end-cell circuit 312 which is shown as incorporating a switch 313 and an adjustable resistor 314. The fuel cell system 500 operates in substantially the same way as described above.

In one experiment, in a conventional arrangement with all the fuel cells 25 in the stack 250 connected in series to the output terminals 302, initially, with a current density at each electrode 10 of 10 mA/cm ² , each fuel cell was producing an electromotive force of about 680 mV, but as the current was gradually increased, the voltages of the end cells were found to decrease. At a current density of 125 mA/cm ² , after a period of operation, the fuel cells 25 at the ends of the stack were about 590 mV while all the intervening cells were about 650 mV. During prolonged operation all the cell voltages gradually decrease, as the electrodes 10 deteriorate, and when at least one cell voltage decreases below a threshold, operation of the stack 250 is terminated. In this particular experiment, the stack 250 operated for about 1 month before operation was terminated. In contrast, with the two end cells 25 of the stack 250 connected to the two end-cell circuits 312, as described in relation to figure 6, with a current density of 10 mA/cm ² , the stack 250 was able to operate for twice as long before the operation was terminated.

In a further modification, when using the fuel cell stack 250, the end links 262 (which connect the second and fourth projecting tabs 260 from each end) may be omitted, and the output terminals 302 connected directly to the fourth projecting tab 260 from each end. In this case the second projecting tab 260 from each end of the stack 250 can be connected directly to the secondary output terminal 310, in substantially the same way as in the system 300 of figure 4. Since, in this case, the end two tabs 260 are connected only to the end-cell circuit 312, this can be further modified as in the system 400 of figure 5 to connect both the end cells in series to a single end-cell circuit 312.

It will also be appreciated that in each of the fuel cell systems 300, 400 and 500 the current through the or each end-cell circuit 312 may be adjusted using the adjustable resistor 314 during operation. Typically, although the maximum fuel cell current density in steady-state operation at peak operating efficiency may be more than 100 mA/cm ² , the current in the or each end-cell circuit 312 would normally correspond to less than 20 mA/cm ² , for example 15 mA/cm ² or 10 mA/cm ² . This has the consequence that the end cells degrade less during operation; and also that the remaining cells in the stack are more uniform in their performance. In practice operation of the cell stack would normally be terminated when one of the cells has dropped below a threshold value; and the effect of the current through the or each end-cell circuit 312 is therefore that a larger voltage differential may be permitted between the starting value and the terminating value for cell voltages.

The use of comparatively low current through the end cells avoids or minimises damage when there is no load connected to the terminals 302, that is to say in open circuit conditions. Furthermore, during operation, the end cells provide a thermal blanket to the adjacent operating cells of the cell stack. Since the power takeoff from the cell stack, that is to say the output terminals 302, are connected to cells that are not at an end of the stack, the cells that provide the power are those that are operating in a regime of thermal and flow symmetry. Although the fuel cells have been described as using a liquid electrolyte, it will be appreciated that a fuel cell system of the invention may utilise a different type of fuel cell. It will also be appreciated that although with liquid electrolyte cells as described above the typical operating voltage is less than 1 V, the electromotive force from a single fuel cell can in some cases be significantly higher, for example more than 1 .2 V, and in any event it will depend on operating conditions such as electrolyte concentration, operating temperature, reactant gas flows, and the current density. With a potassium hydroxide aqueous electrolyte fuel cell operating at about 60° or 70 °C, and an operating current density of 125 mA/cm ² , the typical voltage is between 0.65 and 0.75 V for each fuel cell. Each end-cell circuit 312 has been described as including an adjustable resistor 314, but it will be appreciated that the end-cell circuits may include different circuit elements for control of the current, such as transistors or integrated circuits. It will also be appreciated that the end-cell circuit might be used to provide useful power output, rather than merely being perceived as dissipating energy.