Multiple Manifold Fuel Cell

1. Field of the Invention

The present invention relates to electrolytic fuel cells and assemblies of such cells. More particularly, the invention relates to an electrolytic fuel cell comprising a plurality of manifolds and assemblies of such cells.

2. Background Art

Fuel cells are electrochemical devices which react a hydrogen containing fuel gas with oxygen or an oxygen-containing gas to produce an electric current in an external circuit. Since the process is electrochemical, the level of pollutants released is typically much lower than for combustion based power generation technologies. The conversion efficiency is also higher: 40% to 55% compared 30% to 40% for typical combustion based electric power generation technologies. Fuel cell power generation systems also tend to be quiet and unobtrusive, and can be sited in residential and commercial areas.

Fuel cells have been used in space and other special applications for a number of years, but cost has so far prevented their widespread use for electric utility, industrial and commercial power generation. Certain systems providing up to 11 megawatts output have been demonstrated, and the cost has been greatly reduced through improved design and manufacturing technology and larger scale production. At this point, low enough costs are projected that substantial commercial applications seem possible before the year 2000. These applications are expected to be concentrated in situations in which low pollution is critical, or other attributes such as quiet, unobtrusive operation are necessary. Additional significant cost reductions could make fuel cells attractive to a broader and more cost conscious market.

Two key factors affect the cost per kilowatt of generating capacity. The first is the power output per unit area. As this increases, the cost per kilowatt of output decreases. Unit output is a result of the cell technology and operating variables such as system ambient pressure. The second factor is the material, labor and purchased component cost per unit area. Clearly, this cost directly affects the cost per kilowatt. In typical present day fuel cell designs the cost of the active electrochemical elements is often less than half the total expense for the particular fuel cell. The majority of the cost is for support, gas channeling and similar inactive elements.

Phosphoric acid, molten carbonate and solid oxide electrolyte fuel cells are at various stages of development for utility scale power generation. Phosphoric acid and molten carbonate designs are generally stacks of planar cells, and solid oxide designs include both planar and tubular configurations.

The active portion of a molten carbonate fuel cell is a layer of ceramic powder, saturated by molten carbonate salts, sandwiched between a porous metallic nickel anode and a porous nickel oxide cathode. The cell operating temperature typically is about 650.degree. C. Fuel gas containing hydrogen is passed over the anode, and air containing oxygen is passed over the cathode. The gases react indirectly with each other through the electrolyte, causing an electrical potential to develop between the cathode and the anode. The cell can therefore supply power to an external circuit connecting the cathode and the anode. Typical electrical potential of such molten carbonate cells can be from 0.6 to 0.9 volts.

Multiple cells are combined by stacking individual cells in electrical series. The voltage generated by the cell stack is the sum of the individual cell voltages. Additional components must be added to the active components to build a stack. For example, separator plates are placed between the cells to separate the anode gases of one cell from the cathode gases of the adjacent cell. These are typically stainless steel with a protective nickel cladding layer on the anode side. Seals and flow conduits must be provided to channel fuel gas to the anodes and air to the cathodes, without mixing of the gas streams. Corrugated metal current collectors may be added to space the anode and cathode away from the separator plates, and provide passages to facilitate gas flow over the electrodes. The current collectors must provide a good electrical connection between the anodes and cathodes and the separator plates, and withstand the compressive loads which clamp the stack together. Additional components such as end plates, clamping mechanisms, gas supply and return manifolds, power takeoffs and insulation are required for a complete stack assembly.

The size of molten carbonate fuel cells has been scaled up over the years to achieve manufacturing and installation economies. Larger components require less handling labor per square foot. Dead area at the edge seals decreases as a percentage of total material area, increasing the proportion of material which is actively utilized. Further, the cost of non-repeat parts such as manifolds, end plates, axial load systems, and containment vessels tends to increase more slowly than active area and power output as the cell size increases. System plumbing and interconnection costs decrease per unit of output as the generating capacity of individual stacks increases. Increases in cell size have been paced by the ability to manufacture wide sheets of metal and active cell material with acceptable quality. Widths of up to 3 feet and areas of 4 to 10 square feet are now the state-of-the-art.

While such increases of cell sizes provide some advantages, larger cell size can also pose significant disadvantages. As cell size increases, the mass flow of gas across the electrodes increases in rough proportion to the flow path length. This in turn requires higher flow passages which add to the stack height without adding to power output per unit area.

Larger cells also increase cell complexity and material content. The separator plates and current collectors must bridge the gas flow passages while carrying compressive load and current, and must limit the contact pressure applied to the electrodes. Two or three formed heat and corrosion resistant alloy metal sheets are typically required per cell in current systems to form a structure which carries out these functions. These sheet metal assemblies are generally the largest cost element in the stack.

Furthermore, as cell complexity increases cell performance can decline. Multiple layers of sheet metal in pressure contact introduce resistive potential losses, and the electrode areas blocked by the current collectors introduce concentration gradient potential losses.

Still further, temperature differences across the cell structure can increase as cell size increases. The changing reaction and gas temperature conditions across the cells induce temperature differences and consequent stresses. The large cell size and the thinness of the components inhibit heat flow in the plane of the cell which would tend to even the temperatures out.

Small cells, in contrast, have comparatively short flow paths, and do not require current collectors. The anode and cathode porosity, augmented by dimples or grooves in the surfaces in contact with the separator plates, provides sufficient flow area. Thus in the case of small cells, bulky separator plate flow conduits are not required. This results in comparatively thin cell assemblies which have high volumetric power density and low material content. Further, the small size results in moderate thermal gradients.

The present invention is directed to improved fuel cells and assemblies of such cells. The invention is based in part upon the discovery of breaking large fuel cells into smaller areas by use of plurality of internal manifolds or shafts that provide reactant gas flow paths of a sufficiently short length so that current collectors and formed separator plates typical of prior large cells are not needed. The present invention thus provides substantial reductions in the size and material content of a cell stack of a given output relative to current large stack cell systems.

A single fuel cell of the invention is comprised of an anode layer, a cathode layer and an electrolyte layer. The three layer assembly is penetrated by a plurality of holes which serve as fuel gas and oxidant manifolds. Typically, about half the manifolds of the fuel cell are for fuel gas, and communicate with the anode layer. Suitably these fuel gas manifolds are further functionalized wherein a portion of said manifolds that carry fuel gas serve as fuel gas inlets and carry fuel gas to an anode layer, and the remaining portion of manifolds serve as fuel gas outlets, and carry fuel exhaust from an anode layer. The perimeters of these fuel gas manifolds are sealed where they pass through a cathode layer of the cell to prevent fuel gas flow into or out of the cathode layer. Typically the fuel gas manifolds are distributed across the area of the cell, whereby the flow path length between each gas inlet and exhaust opening in a cell layer is less than the planar dimensions of the cell. The balance of manifolds of the cell are for oxidant gas, and communicate with the cathode layer of the cell. These manifolds are also distributed across the area of the cell whereby the path length between each oxidant and gas inlet opening is less than the planar dimensions of the cell. Suitably these oxidant gas manifolds are further functionalized wherein a portion of said manifolds that carry oxidant gas serve as oxidant gas inlets and carry oxidant to a cathode layer, and the remaining portion of manifolds serve as oxidant outlets, and carry oxidant exhaust from a cathode layer. The perimeters of the oxidant gas holes are sealed where they pass through an anode layer of the cell to prevent oxidant gas flow into or out of the anode layer.

Fuel cell assemblies of the invention comprise a stack of a plurality of cells of the general type described above. The fuel cells of the assembly are typically substantially identical. Separator layers are interleaved between the fuel cells of the assembly to separate one cell from another. The separator layer(s) operate to separate the fuel gas of one cell from the oxidant gas of an adjacent cell. The separator layers are suitably metallic separator plates penetrated in the same pattern as the fuel cells of the assembly. The holes of all the stacked components of the assembly (i.e., cells and separator layers) line up in registration, and form manifold passages which penetrate the height of the stack. Each manifold passage of the assembly communicates exclusively with either the cathode or anode layer of each cell in the stack. The manifold passage may therefore be used to route fuel gas in and out of the anode layers, and oxidant gas in and out of the cathode layers, without mixing of the gases.

Although generally discussed in the context of molten carbonate fuel cells, the present invention is has application to phosphoric acid, solid oxide and other planar fuel cells.

A more complete understanding of the invention may be provided by reference to the accompanying drawings wherein like reference numerals refer to like elements throughout the various figures.

FIG. 1 is an above view of a portion of a fuel cell of the invention;

FIG. 2 is a cross-sectional view taken along the line 2--2 of FIG. 1;

FIG. 3 is a cross-sectional view taken along the line 3--3 of FIG. 1;

FIG. 4 is a top view of a cell layer of the present invention;

FIG. 5 is a cross-sectional view taken along the line 5--5 of FIG. 4; and

FIG. 6 is a front view of a cell stack assembly of the present invention.

The present invention reduces flow paths of fuel and oxidant reactive materials of a fuel cell by a novel multiple manifold or shaft design. The manifolds axially penetrate a cell, i.e., the manifolds longitudinally extend through a cell perpendicular to the planar surface of the cell layers. The manifolds provide distribution of the reactive materials to areas surrounding manifolds outlets. By employing a plurality of manifolds, the fuel and oxidant gases can be distributed throughout a large cell. That is, a large cell is effectively broken into smaller areas with flow paths short enough that current collectors and formed separator plates are not needed. The small flow area required by the short flow paths is provided by the porosity of the cathode and anode structures, augmented by dimples or shallow grooves formed in their surfaces. The small flow areas are served by the noted array of internal fuel and oxidant manifolds 3 as shown in the figures.

Preferably, the manifolds are further functionalized to provide four different "species" of manifold passages. More specifically, as shown in FIG. 1, preferably manifolds 3 are segregated manifolds to provide either gas inlet and outlet functions. Thus, manifold 3A supplies oxidant gas to a cathode layer; manifold 3B collects oxidant exhaust from a cathode layer; manifold 3C supplies fuel gas to an anode layer; and manifold 3D collects fuel exhaust from an anode layer. Directions of flow of fuel and oxidant gases and exhaust to and from the respective manifolds are also depicted in FIG. 1. Specifically, arrows proximate to and pointing away from the manifold positioned in the upper left hand side of FIG. 1 indicate the direction of gas flow into the cathode layer; arrows proximate to and pointing away from the manifold positioned in the lower left hand side of FIG. 1 indicate the direction of gas flow into the anode layer; arrows proximate to and pointing toward the manifold positioned in the upper right hand side of FIG. 1 indicate the direction of gas flow to the anode exhaust manifold; arrows proximate to and pointing toward the manifold positioned in the lower right hand side of FIG. 1 indicate the direction of gas flow to the cathode exhaust manifold.

As used herein, the term "flow path length" refers to the distance a reactive gas (oxidant gas or fuel gas) travels over an electrode layer between a gas inlet manifold and a exhaust or gas outlet manifold. Thus, in FIG. 1, the flow path length between manifolds 3C and 3D is the distance Y. The flow path length between manifolds 3A and 3B is the distance X.

As the reactive gas travels across an electrode layer, the gas reacts with the electrode and its content of reactive gas is depleted. The greater the flow path length, the larger the mass flow rate must be to assure that enough of the reactant remains to carry out a useful reaction near the end of the flow path across the electrode near the adjacent exhaust manifold.

Manifolds 3 are suitably formed by a plurality of holes 2 in each cell layer (e.g., cell layers of anode, electrolyte matrix, cathode, separator plate) where each of holes 2 in each cell layer is in at least partial registration with a hole in adjacent cell layer(s). Preferably holes 2 in each cell layer are of the same diameter and are in full registration with mating holes 2 in each adjacent cell layer. While the shape of manifolds 3 is generally discussed herein as being tubular, it should be appreciated that the present invention is not limited to tubular-shaped manifolds and that manifolds 3 suitably can be of other geometries provided the manifolds can provide communication of reactive gases to a fuel cell. FIG. 1 shows anode seal 7, cathode seal 10 and gas flow passages 11 formed by the dimpled electrode surface, all of which are more fully discussed infra.

FIGS. 2 and 3 show separator plates 1 which are suitably employed in a multiple cell assembly. Plates 1 are typically flat sheets suitably constructed of stainless steel, nickel clad on the anode side. Separator plates 1 have a plurality of holes 2 which communicate with openings 2 of other cell layers to form manifold passages 3.

Three layer packages of active elements, each consisting of an electrolyte-matrix layer 4 sandwiched between cathode layer 5 and anode layer 6, are clamped between separator plates 1. These layers also have manifold holes 2 which are in at least partial registration with the holes 2 in the separator plates. When the components are stacked, the aligned holes 2 of each cell layer form internal manifold passages 3 which communicate with each cell layer. As discussed above, during use of the cell, a portion of the total number of internal manifold passages 3 carry fuel gas to and exhaust from the cell anodes (said manifold passages sometimes referred to as the anode passages), and another portion of the total number of manifold passages 3 carry oxidant gas to and exhaust from the cathodes (said manifold passages sometimes referred to as the cathode passages). Typically half of the total number of manifold passages carry fuel gas and half the manifold passages carry oxidant.

The anode gas passages 3 are isolated from the cathode electrodes as they pass through the cathode cell layers, Similarly, the anode cell layers are isolated from the axially extending cathode passages 3. This is accomplished in the anode layers as follows. Anode seal 7 is bonded to the top of the separator plate 1 so that they surround each of the cathode passages 3. Seals 7 are suitably constructed of aluminized stainless steel. Holes 2 in anode sheet 6 are made of a size sufficient to clear the rings so that the anode bears uniformly on the separator plate. An anode seal 7 is suitably about the same thickness as the anode so that in combination seal 7 forms a level surface to support the electrolyte matrix sheet 4 in the case of a molten carbonate cell, the molten carbonate saturating the matrix layer 4 can wet the seal top surface 8 and form a gas seal 8' which prevents air in the cathode manifold passage from mixing with the fuel gas in the porous anode sheet. The gases can, however, react with each other through the electrolyte matrix by carbonate ion transport and a return flow of electrons through the seal ring. To prevent or inhibit corrosion of seal 7 by such action, seal 7 can be of an aluminized stainless steel construction. The aluminum oxide layer is generally insoluble in the electrolyte, and acts as a barrier to the electron flow.

An important aspect of the invention is seal ring joints 9 positioned between seal 7 and separator plates 1. Joints 9 preferably both prevent gas leakage and avoid any capillary paths between the air in a manifold passage 3 and the fuel in the anode. Such a path could fill up with electrolyte and form a corrosion cell which would attack the separator plate. A preferred design of joints 9 is to fusion weld or nickel braze the seal ring to the separator plate. Other potential designs of joints 9 include use of fused glass or ceramic cement to block gas leakage and exclude electrolyte from the joint.

Cathode seals 10 operate in similar fashion as anode seals 7 to separate the cathodes from the anode passages. Cathode seals are suitably constructed of aluminized stainless steel. The cathode seal rings are bonded to the bottom surface of the separator plate by cathode seal joints 9'. A preferred design of joints 9' is to fusion weld or nickel braze the seal ring to the separator plate, although other bonding means will be suitable such as those discussed above for the anode joints 9. The wet seals and corrosion concerns are essentially the same as with the anode seals.

The surfaces of the electrodes in contact with the separator plates are dimpled as shown, e.g., by electrode dimple ED in FIG. 3. This forms flow passages 11 which carry part of the gas flow across the electrodes. Cathode flow passage 11A and anode flow passage 11B are depicted in FIG. 3. Additional reactive gas flows through the pores of the electrodes. The design trade-offs are between the flow path length, pressure drop, dimple geometry, and the number of manifold holes required per unit area of the cell. The path lengths across the anodes and cathodes and the cathode and anode manifold passage areas do not necessarily have to be the same. Since the cathodes typically require higher flow than the anodes, it may be advantageous to adjust the geometry to make the cathode flow paths shorter and the cathode manifold areas comparatively larger to minimize pressure drops.

The arrows depicted in FIG. 2 show the direction of gas flow into the cathode layer of the cell, and the arrows depicted in FIG. 3 show the direction of exhaust gas flow from the anode layer of the cell.

Referring now to FIG. 4, each manifold opening 3, except those at the edges, serves the surrounding plane surface of the cell layer. A particularly preferred design for the edges of the cell layer is shown in FIG. 4. The active area of the manifolds 3 along the edge of a cell layer ends at the centerline of the manifolds. That is, partial manifolds are employed along the cell layer edges to reflect the fact the edge manifolds serve less area. As shown in FIG. 4, preferably half round manifolds 12 are employed on the cell sides and quarter round manifolds 13 are employed on the cells corners. Alternative designs for the "edge" manifold openings can leave "dead" active areas near the cell layer perimeter which have reduced flow on either the anode or the cathode side. Directions of flow of fuel and oxidant gases and exhaust of a preferred manifold arrangement are also shown in FIG. 4.

A "picture frame" edge seal 14 made of strips of aluminized stainless steel 15 is suitably bonded to both the cathode and anode sides of the separator plates 1 to contain the gas within the stack. The half-round and quarter-round seal rings 16 needed at the edges and corners are suitably incorporated into the edge strips. In FIG. 5, a typical edge cross section is shown where a cross section of a manifold providing oxidant gas to a cathode layer is depicted. The corner joints are suitably welded (to form a welded joint 17) or brazed. The arrows shown in FIG. 5 indicate the direction of gas flow into the cathode layers.

The process for manufacturing and assembling the fuel cells of the invention in many respects is in accordance with known methods. See, for example, U.S. Pat. No. 4,770,955, incorporated herein by reference for its disclosure of the manufacture of fuel cells and the use thereof. The cathodes 5 and anodes 6 are suitably manufactured by a porous sheet forming process such as tape casting, followed by sintering. The dimples on the electrodes may be added as part of the forming process or as a secondary operation, either before or after sintering. Holes 2 in the cathode and anode layers are suitably punched and the edges are trimmed to produce finished parts. A tooled punch process will produce highly repeatable part dimensions.

Separator plates 1 are suitably punched from flat sheets of nickel clad 300 series stainless steel sheet. Punch tooling will provide close dimensional control so that manifold openings 2 will match the cathode and anode openings. Punching has the further advantages of relatively inexpensive tooling and low press tonnage requirements.

Anode and cathode seal rings 7 and 10 are suitably punched from stainless steel and aluminized. One side of the seal ring must be protected from the aluminization process, or ground clean afterwards, to assure a good fusion weld or braze joint to separator plates 1.

Edge seal strips 15 are also suitably punched from stainless steel sheet. They are welded into "picture frames" 14 and then aluminized. As with the seal rings, one side should be clean. Alternatively, the brazing operation may be used to seal the corners and eliminate the need for welding. The separator plate assembly, consisting of the plate, seal rings and edge seals, is then made. Fixturing can be required to assure accurate alignment of the components so that they fit correctly with the electrodes. Nickel brazing in a vacuum or reducing atmosphere furnace is a good method to join the seals and the separator plate. A nickel brazing alloy applied as paste or metal preforms supplies the filler material. Alloys of nickel and phosphorus with melting points in the range of 1100.degree. C. are available and will provide suitable performance. The filler flows into the joints 9 and 9' by capillary action, and fills the space completely provided the surfaces are clean and properly positioned. The brazing process should not harm the aluminized surfaces.

An automated fusion welding process, using resistance welding, laser welding or electron beam welding may also be used. The principal requirement is a continuous weld line to prevent gas leakage or electrolyte bridging. Care should be taken to avoid damage to the aluminized coating.

The electrolyte matrix 4 material is tape cast in sheet form and punched to final dimensions. As is typical practice, the electrolyte matrix material is put in the assembly in the "green" condition, and the binders are burned out during stack start-up.

The unit cells are assembled by nesting the electrodes into the recessed areas of the separator plate assemblies and aligning the matrix sheet on top of this subassembly. Electrolyte may be added at this stage or it may have been previously added to one or both of the electrodes. At start-up of operation of a cell, the electrolyte will melt and flow into the matrix and adjust to the proper levels in the electrodes.

The thus assembled unit cells then can be stacked to provide a stacked multiple cell assembly or fuel cell module. As is known in the art, the terms "fuel cell module" or "fuel cell power generation module" refer to, among other things, an apparatus that contains a single stack of fuel cells as well as an apparatus that comprises multiple stacks of fuel cells.

The operation of stacking assembled unit cells should be straightforward as the edge seals will make the unit cells of a sufficient rigidity for good handling, and the lack of external manifolds makes the alignment somewhat forgiving. An overall stack assembly is shown in FIG. 6. The center part 18 of the assembly as shown in FIG. 6 is the active cell stack. This assembly depicted in FIG. 6 is a simple stack of unit cells held under axial compression.

As shown in FIG. 6, gas distribution plates 19 are placed above and below the stack. Plates 19 are suitably welded or brazed assemblies of sheets and tubes with internal distribution and collection plenums 20, 21, 22, 23 for fuel and oxidant gas flow. As shown in FIG. 6, each plenum suitably serves a different species of manifold. More specifically, FIG. 6 illustrates one embodiment of the invention where plenum 20 communicates with a fuel exhaust manifold 3D through cathode exhaust tube 20A; plenum 21 supplies oxidant to an anode inlet manifold 3B; plenum 22 supplies fuel to a cathode inlet manifold 3C through cathode inlet tube 22A; and plenum 23 communicates with an oxidant exhaust manifold 3A through anode exhaust tube 23A. The faces which face the cell stack have openings 24 which align with the manifold passages and connect to the appropriate plenums. The stack is suitably fed with fuel and oxidant gases through the plenums and manifold passages from each end of the cell stack to allow greater height for a given manifold passage size. The electric power pickups (not shown) are connected to these plates. Other constructions such as cored stainless steel castings are also possible, These gas distribution plates can be salvaged at the end of the stack's life and reused.

Thermal and electrical insulation blocks 25 are placed above and below the gas distribution plates. Blocks 25 isolate the active cell stack from the support and clamping structures. A suitable material of construction for blocks 25 is a low density ceramic, for example the alumina and silica containing material sold under the trade name of KAOWOOL MB-823 Fiberboard.

A base plate 26 is placed below the stack, and pressure plates 27 are placed on each end. Plates 27 are connected by axial load tie bars 28 with axial load springs 29 or similar follow-up system to accommodate stack height changes. A thermal insulation blanket (not shown) is suitably wrapped around the cell stack and the gas distribution plates. The base plate, pressure plate and axial load hardware are typically outside the blanket in an area of relatively lower temperature.

Four pipes (not shown) each connect to the bottom and top gas distribution plates. These pipes connect to the plenums through flow transition sections 30 which are welded or bolted to the distribution plates. Insulating couplings (not shown) are used to maintain electrical isolation.

As discussed above, the present invention permits notable reductions in the size and material content of a cell stack relative to prior systems. For example, the height of a unit cell in some current molten carbonate fuel cell stack systems is between 0.200 to 0.400 inches. In comparison, the height of a distributed manifold unit cell of the present invention suitably can be about 0.100 inches or less. As a first approximation, this increases the volumetric power density of a stack by a factor of 2 to 4 at a given output per square foot of active area. This leads to a number of benefits. For example, assuming a power output of about 120 watts per square foot of active area, and that active area is about 85% of gross area, the power density of the distributed manifold unit cell of the invention is over 10 kilowatts per cubic foot of cell stack. This should render the cell suitable for use in applications such as transportation. Additionally, non-repeat part costs such as insulation, enclosures, axial load systems, and plumbing should drop by a factor of 2 to 4 on a per kilowatt basis, since these costs tend to be a factor of the number and physical size of the stacks, not their output.

The plurality of manifolds distributed across the planar area of a fuel cell of the invention adds a degree of freedom to the design process which does not exist with conventional manifold configurations. Most importantly, it uncouples the flow path length and cross section area from the overall cell size. This uncoupling allows elimination of current collectors; as a consequence significant increases in power density and reductions in material requirements can be realized.

Another advantage of a cell stack assembly of the present invention is its structural integrity. The cell stack of the present invention suitably is essentially an assembly of flat sheets, without formed current collectors which can yield or cause stress concentrations.

The fuel cell of the invention will readily scale up or down in cell area, since the flow conditions in the cells are a function of the repeating manifold pattern, and are essentially independent of the total cell size. A 1 foot by 1 foot cell, for example, should provide the same operation and performance characteristics as a 4 foot by 8 foot cell. Cell sizes as large as 4 foot by 8 feet will be suitable. At a 10 foot active height, such a cell stack would have over 3 megawatts output in a truck shippable package.

Temperature and gas composition variations will be relatively independent of cell size, and primarily a function of the repeating manifold pattern. It is therefore possible to design for specific temperature and composition values independent of the overall cell size. Further, the distributed manifold configuration of the present invention may allow faster heat-up and cool-down cycles than conventional cells of the same size, since transient temperature gradients are reduced.

To control the temperature of a stack cell assembly, a gas flow can be superimposed along the length of the stack, in addition to the flow of the reactant gases which passes between the cells. This flow should be effective in removing heat from the stack, since multiple manifolds have a large surface area, and the heat transfer distance from any part of the cell to a manifold passage is short. Such a gas flow provides means to control the reactant gas flow through the cells independently of the cooling flow. Such a gas flow may also be useful in heating the stack quickly and uniformly to operating temperature at startup.

The distributed manifold cell design of the present invention can be implemented in a number of configurations and with a range of design parameters. For example, a suitable cathode layer thickness is between about 0.015 and 0.050 inches, more preferably about 0.030 inches thick. A suitable anode layer thickness is between about 0.010 and 0.040 inches, more preferably about 0.030 inches thick. A suitable thickness of the electrolyte matrix is between about 0.010 and 0.030 inches, more preferably about 0.020 inches thick.

A separator plate used in the cell assemblies of the invention is suitably between about 0.008 and 0.025 inches thick, more typically about 0.020 inches thick.

The diameters of manifolds of the fuel cells of the invention suitably may vary within a somewhat wide range. For example, a manifold of about 0.50 inches or less will be suitable as well as manifold diameters of about 2 inches or greater. More typically, the manifold diameters will be in a range of from about 0.50 to 2.0 inches. Suitable flow path lengths also may vary somewhat widely. For example, flow path lengths of about one inch or greater will be suitable.

Preferably a plurality of manifolds of a fuel cell are separated fully from each adjacent manifold of the cell by a reactive cell area. More preferably all or essentially all the manifolds of a fuel cell are laterally separated from each adjacent manifold of the cell by a reactive cell area, i.e., each cell layer is essentially a sheet perforated with manifolds wherein the manifolds are not contiguous or otherwise abut one another. Such a configuration provides a number of advantages including imparting greater structural integrity to the cell relative to a system having openings that are in some manner contiguous or that otherwise abut adjacent openings. Cell assembly also should be facilitated by such fully separated manifolds. It is further believed that cell efficiency can be improved by use of such manifolds that are fully surrounded by reactive cell areas relative to a system that has fuel openings that are in some manner contiguous or that otherwise abut adjacent openings.

Suitably the combined cross-sectional area of the manifolds constitute only a minor portion (i.e., less than 50 percent) of the total active planar area of a fuel cell of the invention. More preferably the combined cross-sectional area of a cell's manifolds total between about 5 and 40 percent of the total active planar area of a fuel cell; still more preferably the cross-sectional area of the manifolds totals between about 5 and 30 percent of the active planar area of a fuel cell; and even more preferably the combined cross-sectional area of the manifolds total between about 10 and 25 percent of the total active planar area of a fuel cell. As is clear to those in the art, the term active planar area of the cell includes both the gas manifolds and reactive cell areas positioned between the manifolds. In accord with the above, preferred is a fuel cell of the invention wherein the ratio of the total reactive area of a cell to the combined cross-sectional area of the cell's manifolds is between about 20 and 1.5, more preferably the ratio is between about 20 and 3, and still more preferably the ratio of total reactive planar area of a cell to the combined cross-sectional area of the cell's manifolds is between about 10 and 4.

In one fuel cell of the invention, the manifolds are about 1 inch in diameter and on 2.83 inch centers. The flow path length is 4 inches center-to-center (that is, the distance X or Y shown in FIG. 1 is 4 inches as measured from the center of one manifold to the center of the corresponding adjacent manifold). The surrounding wet seal disk is 1.25 inches in diameter, leaving a 0.125 inch seal width. The active area is, therefore, about 85% of the gross area, and the manifold openings total about 12% of the active area. The electrode dimple height is about 0.010 for the 4 inch flow path length. This should result in the same order-of-magnitude of flow velocity for a given volumetric rate per square inch of electrode area as a more typical 0.050 to 0.100 inch thick flow path serving a 30 to 40 inch path length. Pressure drops should be on the order of 0.30 to 0.40 pounds per square inch through the flow fields between the inlet and outlet manifolds. The pressure drop in the manifold passages should be on the order of 0.10 pounds per square inch per foot of stack height. One inch diameter manifold openings should be able to serve on the order of 500 cells from one end, or 1,000 cells with gas supplied from both ends.

The dimple approach to creating flow fields allows flows in all directions, and can be formed on the electrodes without regard to the manifold hole locations. It is also possible to form flow channels in the faces of the electrodes which register with the manifold openings to tailor the flow patterns.

The seal ring wet seal width of 0.125 inches is narrower than typical, but is over 6 times the matrix thickness. If seal integrity is a problem, use of either wider seals or thinner matrix sheets will reduce the gasket stress. Plate flatness and clamping pressure uniformity of the fuel cells of the invention will be good due to fact that the mating metal parts are well supported, simple flat punchings.

An edge seal width of 0.50 inches is suitable and in the range of typical practice. This edge seal width should be sufficient to provide the required compressive support for the stack and to seal the gases.

A preferred distributed manifold fuel cell that comprises fuel and oxidant inlet and exhaust manifolds has the following configuration of manifolds 3. Manifolds 3 are positioned across the planar surface of the fuel cell and are positioned at the intersections of an essentially square grid according to the following rules: fuel inlet and exhaust manifolds are positioned so that fuel gas flows essentially diagonally across each square of said grid through the anode layers; oxidant inlet and exhaust manifolds are positioned so that oxidant gas flows essentially diagonally across each square of said grid through the cathode layers at essentially right angles to above-mentioned direction of fuel flow; and gas flows through each manifold opening into or out of all the squares surrounding that manifold opening.

More preferably the manifold configuration is of the above described preferred type and the outer boundary of the active cell area passes approximately through the centers of the outermost manifold openings 3, and the edge manifold openings are about half the area of the central manifold openings (i.e., manifold openings that do not intersect a planar edge of a cell layer), and the corner manifold openings are about one quarter the area of the central manifold openings. These reduced manifold areas reflect the fact that they are serving smaller cell areas at the edge of the cells. A corner manifold opening is one that intersects two planer edges of a cell layer. Such a preferred configuration is generally shown in FIG. 4.

The foregoing description of the invention is merely illustrative thereof, and is understood that variations and modifications can be effected without departing from the scope or spirit of the invention as set forth in the following claims.