Patent Application
20080279710
The present invention provides a composite composition composed of expanded graphite, a non-expandable component, and a matrix or binder material. The composition can be used to make fuel cell bipolar plates or flow field plates. In particular, the present invention provides a highly conducting, less anisotropic composite flow field plate composition that has an exceptionally high electrical conductivity in the plate thickness direction.
A fuel cell converts chemical energy into electrical energy and some thermal energy by means of a chemical reaction between a fuel (e.g., hydrogen gas or a hydrogen-containing fluid) and an oxidant (e.g., oxygen). A proton exchange membrane (PEM) fuel cell uses hydrogen or hydrogen-rich reformed gases as the fuel, a direct-methanol fuel cell (DMFC) uses methanol-water solution as the fuel, and a direct ethanol fuel cell (DEFC) uses ethanol-water solution as the fuel, etc. These types of fuel cells that require utilization of a PEM layer as a proton transport electrolyte are collectively referred to as PEM-type fuel cells.
A PEM-type fuel cell is typically composed of a seven-layered structure, including (a) a central PEM electrolyte layer for proton transport; (b) two electro-catalyst layers on the two opposite primary surfaces of the electrolyte membrane; (c) two fuel or gas diffusion electrodes (GDEs, hereinafter also referred to as diffusers) or backing layers stacked on the corresponding electro-catalyst layers (each GDE comprising porous carbon paper or cloth through which reactants and reaction products diffuse in and out of the cell); and (d) two flow field plates (or a bi-polar plate) stacked on the GDEs. The flow field plates are typically made of graphite, metal, or conducting composite materials, which also serve as current collectors. Gas-guiding channels are defined on a GDE facing a flow field plate or, more typically, on a flow field plate surface facing a GDE. Reactants (e.g., H2 or methanol solution) and reaction products (e.g., CO2 at the anode of a DMFC, and water at the cathode side) are guided to flow into or out of the cell through the flow field plates. The configuration mentioned above forms a basic fuel cell unit. Conventionally, a fuel cell stack comprises a number of basic fuel cell units that are electrically connected in series to provide a desired output voltage. If desired, cooling channels and humidifying plates may be added to assist in the operation of a fuel cell stack.
In one common practice, a fuel flow field plate and an oxidant gas flow field plate are separately made and then assembled together to form a bipolar plate (one side of a bipolar plate serving as a negative terminal and the other side as a positive terminal, hence the name). In some cases, an additional separator is sandwiched between the two flow field plates to form a bipolar plate. It would be highly advantageous if the flow filed plates and the separator can be mass-produced into an integrated bipolar plate assembly. This could significantly reduce the overall fuel cell production costs and reduce contact ohmic losses across constituent plate interfaces. The bipolar plate is known to significantly impact the performance, durability, and cost of a fuel cell system. The bipolar plate, which is typically machined from graphite, is one of the most costly components in a PEM fuel cell.
Fluid flow field plates have open-faced channels formed in one or both opposing major surfaces for distributing reactants to the gas diffuser plates, which are the anode and cathode backing layers, typically made of carbon paper or fabric. The open-faced channels also provide passages for the removal of reaction products and depleted reactant streams. Optionally, a bipolar plate may have coolant channels to manage the fuel cell temperature. According to the US Department of Energy (DOE), a bipolar plate should have the following desirable characteristics: high electrical conductivity (e.g., preferably having a conductivity no less than 100 S/cm and specific areal conductivity no less than 200 S/cm2), low permeability to fuel or oxidant fluids, good corrosion resistance, and good structural integrity. The specific areal conductivity is essentially the bipolar plate thickness-direction conductivity divided by the plate thickness. Hence, it is highly desirable to have a thinner plate. Current graphite bipolar plates, typically 3-5 mm thick, should preferably be reduced to below 1 mm and most preferably below 0.5 mm.
Conventional methods of fabricating fluid flow field plates require the engraving or milling of flow channels into the surface of rigid plates formed of a metal, graphite, or carbon-resin composite. Such plates are expensive due to high machining costs. The machining of channels into the graphite plate surfaces causes significant tool wear and requires significant processing times. Metals can be readily shaped into very thin plates, but long-term corrosion is a major concern. A corrosion-resistant coating may be used, but it has to be applied perfectly. The coating may also increase contact resistance.
Alternatively, fluid flow field plates can be made by a lamination process (e.g., U.S. Pat. No. 5,300,370, issued Apr. 5, 1994), wherein an electrically conductive, fluid impermeable separator layer and an electrically conductive stencil layer are consolidated to form one open-faced channel. Presumably, two conductive stencil layers and one separator layer may be laminated to form a bipolar plate. It is often difficult and time-consuming to properly position and align the separator and stencil layers. Die-cutting of stencil layers require a minimum layer thickness, which limits the extent to which fuel cell stack thickness can be reduced. Such laminated fluid flow field assemblies tend to have higher manufacturing costs than integrated plates, due to the number of manufacturing steps associated with forming and consolidating the separate layers. They are also prone to delamination due to poor interfacial adhesion and vastly different coefficients of thermal expansion between a stencil layer (typically a metal) and a separator layer. Corrosion also presents a challenging issue for metal-based bipolar plates in a PEM fuel cell since they are used in an acidic environment.
A variety of composite bipolar plates have been developed, which are mostly made by compression molding of polymer matrices (thermoplastic or thermoset resins) filled with conductive particles such as graphite powders or fibers. Because most polymers have extremely low electronic conductivity, excessive conductive fillers have to be incorporated, resulting in an extremely high viscosity of the filled polymer melt or liquid resin and, hence, making it very difficult to process. Bi-polar plates for use in PEM fuel cells constructed of graphite powder/fiber filled resin composite materials and having gas flow channels are reviewed by Wilson, et al (U.S. Pat. No. 6,248,467, Jun. 19, 2001). Injection-molded composite-based bipolar plates are disclosed by Saito, et al. (U.S. Pat. No. 6,881,512, Apr. 19, 2005 and U.S. Pat. No. 6,939,638, Sep. 6, 2005). These thermoplastic or thermoset composites exhibit a bulk conductivity significantly lower than 100 S/cm (the US Department of Energy target value), typically not much higher than 10 S/cm.
Besmann, et al. disclosed a carbon/carbon composite-based bipolar plate (U.S. Pat. No. 6,171,720 (Jan. 9, 2001) and U.S. Pat. No. 6,037,073 (Mar. 14, 2000)). The manufacture process consists of multiple steps, including production of a carbon fiber/phenolic resin preform via slurry molding, followed by a compression-molding step. The molded part is then pyrolyzed at a high temperature (1,500° C.-2,500° C.) to obtain a highly porous carbon/carbon composite. This is followed by chemical vapor infiltration (CVI) of a carbon matrix into this porous structure. It is well-known that CVI is a very time-consuming and energy-intensive process and the resulting carbon/carbon composite, although exhibiting a high electrical conductivity, is very expensive.
Instead of using pyrolyzation and CVI to produce carbon/carbon composites, Huang, et al. (US Patent Application Pub. No. 2004/0229993, Nov. 18, 2004) discloses a process to produce a thermoplastic composite with a high graphite loading. First, polymer fibers, such as thermotropic liquid crystalline polymers or polyester, reinforcing fibers such as glass fibers, and graphite particles are combined with water to form a slurry. The slurry is pumped and deposited onto a sieve screen. The sieve screen serves the function of separating the water from the mixture of polymer fibers, glass fibers and graphite. The mixture forms a wet-lay sheet which is placed in an oven. Upon heating to a temperature sufficient to melt the polymer fibers, the wet-lay sheet is allowed to cool and have the polymer material solidify. Upon solidification, the wet-lay sheet takes the form of a sheet material with reinforcement glass fibers held together by globules of thermoplastic material, and graphite particles adhered to the sheet material by the thermoplastic material. Several of these sheets are then stacked, preferably with additional graphite powder interspersed between sheets, and compression-molded in a hot press. After application of heat and pressure in the press, one or more formed bipolar plates are obtained, where the bipolar plates are a composite of glass fibers, thermoplastic matrix and graphite particles. Clearly, this is also a tedious process which is not amenable to mass production.
Alternatively, fluid flow field plates can be made from an electrically conductive, substantially fluid impermeable material that is sufficiently compressible or moldable so as to permit embossing. Flexible graphite sheet is generally suitable for this purpose because it is relatively impervious to typical fuel cell reactants and coolants and thus is capable of isolating the fuel, oxidant, and coolant fluid streams from each other. It is also compressible and embossing processes may be used to form channels in one or both major surfaces. The “flexible graphite” is the expanded or exfoliated reaction product of rapidly heated natural graphite particles which have been treated with an agent that intercalates into the crystal structure of the graphite to expand the intercalated particles typically 80-300 times in the direction perpendicular to the carbon layers in the crystal structure. The exfoliated graphite particles are vermiform in appearance, and are therefore commonly referred to as worms. Hereinafter, the term “exfoliated graphite” will be used interchangeably with the term “expanded graphite.” The worms may be compressed together into flexible sheets which, unlike the original graphite flakes, can be formed and cut into various shapes. These thin sheets (foils or films) are hereinafter referred to as flexible graphite. Flexible graphite can be wound up on a drum to form a roll of thin film, just like a roll of thin plastic film or paper.
Although flexible graphite sheets are highly conductive, they by themselves do not have sufficient stiffness and must be supported by a core layer or impregnated with a resin. For example, U.S. Pat. No. 5,527,363 (Jun. 18, 1996) discloses a fluid flow field plate comprising a metal sheet interposed between two flexible graphite (FG) sheets having flow channels embossed on a major surface thereof. These FG-metal-FG laminates are also subject to the delamination or blistering problem, which could weaken the plate and may make it more fluid permeable. Delamination or blistering can also cause surface defects that may affect the flow channels on the plate. These problems may be difficult to detect during fabrication and may only emerge at a later date. The vastly different coefficients of thermal expansion (CTE) and elastic constants between a metal and a flexible graphite layer result in many challenging problems. In particular, thermal cycling between frozen and thawed states, as are likely to be encountered in an automobile application of the fuel cell, could result in delamination between a flexible graphite layer and the metal layer.
Alternatively, Mercuri, et al. (e.g., U.S. Pat. No. 5,885,728, Mar. 23, 1999 and U.S. Pat. No. 6,037,074, Mar. 14, 2000) disclosed a flexible graphite sheet having embedded ceramic or glass fibers extending from its surface into the sheet to increase the resin permeability of the sheet for the preparation of a resin-impregnated flexible graphite bipolar plate. By allowing ceramic or glass fibers to puncture through layers of exfoliated graphite also leave these layers vulnerable to gas permeation, thereby significantly reducing the hydrogen and oxygen permeation resistance of a bipolar plate and increasing the chance of dangerous mixing of hydrogen and oxygen inside a fuel cell stack.
What follows is a summary of the state of the art of the resin-impregnated expanded graphite composite, resin-impregnated flexible graphite sheet composite, and methods of producing these composites:
Olstowski, et al. (“Novel Compressed Cohered Graphite Structures and Method of Preparing Same,” U.S. Pat. No. 3,492,197, Jan. 27, 1970) provided compressed and resin-bonded forms of expanded vermicular graphite. The resin-bonded composite is obtained by (a) providing a supply of an expanded vermicular graphite having an apparent bulk density of 0.2-2.0 pounds per cubic foot; (b) providing a supply of a bonding agent; (c) blending the expanded vermicular graphite and bonding agent in an amount of 2-35 weight percent bonding agent based on the total weight of the expanded graphite-bonding agent mixture; (d) compressing the mixture at a pressure of 5-50,000 psi in predetermined directions into predetermined forms of cohered graphite; and (e) treating the so-formed composite to activate the bonding agent thereby promoting adhesion within the compact. This invention taught about compressing vermicular-bonding agent mixture in a uniaxial direction to produce a highly anisotropic composite and in bi-axial, tri-axial, cylinder-radial, and isostatic directions to produce less anisotropic or more isotropic composites. However, it failed to teach, implicitly or explicitly, how a desired degree of isotropy could be maintained when the bi-axially, tri-axially, cylinder-radially, and isostatically compressed composite compacts (prior to curing or fusing to consolidate) were re-compressed or molded as a final operation to become a thin composite plate. This thin plate (thinner than 5 mm, preferably thinner than 3 mm, further preferably thinner than 1 mm, and most preferably thinner than 0.5 mm) is for a bipolar plate application. Further, this patent was limited to using a solid bonding agent to begin with the blending process, excluding liquid polymers from the invention due to the perceived notion that these liquid polymers “can prevent formation of highly densified composites.” This patent did not teach how bi-axial, tri-axial, cylinder-radial, and isostatic compressions could be accomplished in a real manufacturing environment for the mass production of less anisotropic composites. Furthermore, the method disclosed in this patent entailed first exfoliating graphite to obtain graphite worms and then mixing graphite worms with a bonding agent in a fine solid powder form. Once the graphite worms are formed, it would be very difficult to mix the worms with fine solid particles in a homogeneous manner without breaking up or significantly disturbing the continuous network of electron-transport paths (interconnected graphite flakes).
Caines (“Vermicular Expanded Graphite Composite Materials,” U.S. Pat. No. 4,265,952, May 5, 1981) disclosed an expanded graphite composite containing a corrosion resistant resin (e.g., polytetrafluoroethylene, PTFE). The composite was prepared by blending vermicular graphite with a suspension of fine solid resin particles in a carrier liquid medium, vaporizing the carrier, and heating the composite material to sinter the resin. No electrical property of the resulting composite was reported.
Atkinson, et al. (“Housing for Electrical or Electronic Equipment,” U.S. Pat. No. 4,530,949, Jul. 23, 1985) provided a low-density composite composition consisting of exfoliated graphite and a thermosetting resin binder. The density (<0.1 gm/cm3) and the electrical conductivity (0.1 S/cm) values are relatively low.
Fukuda, et al. (“Reinforced Flexible Graphite Sheet,” U.S. Pat. No. 4,729,910, Mar. 8, 1988) disclosed a process of producing thermosetting resin reinforced flexible graphite sheets. The process involved subjecting both the flexible graphite sheet and a phenolic resin solution to a preliminary de-aeration treatment prior to immersing the flexible graphite sheet in the resin solution. No electrical conductivity data was offered.
Chung provided a low-density (0.7 gm/cm3) exfoliated flexible graphite flake-reinforced composite with a conductivity of 2 S/cm (Chung, “Low-Density Graphite-Polymer Electrical Conductor,” U.S. Pat. No. 4,704,231, Nov. 3, 1987). Chung also provided an in-situ exfoliation method of producing graphite flake-reinforced epoxy composites (“Composites of In-Situ Exfoliated Graphite,” U.S. Pat. No. 4,946,892, Aug. 7, 1990).
Fong, et al. (“Methacrylate Impregnated Carbonaceous Parts,” U.S. patent application Ser. No. 09/896,178, filed on Jun. 29, 2001 (Pub. No. US 2001/0046560, Pub date Nov. 29, 2001)) disclosed a method of impregnating a highly porous carbon material with a methacrylate polymer. No electrical conductivity data was provided.
Öttinger, et al. (“Impregnated Bodies Made of Expanded Graphite, Process for Producing Such Bodies and Sealing Elements, Fuel Cell Components and Heat-Conducting Elements Formed of the Bodies,” U.S. Pat. No. 6,746,771, Jun. 8, 2004) provided composites of expanded graphite impregnated with isocyanate or epoxy resins. The method involved soaking expanded graphite with a low-viscosity, polymerizing resin. The achievable electrical conductivity of the resulting composites appears to be in the range of 2-10 S/cm.
Da Silva, et al. (“Method for Producing Composite Objects Using Expanded Graphite and Vermiculite,” U.S. patent application Ser. No. 10/574,803 filed on Oct. 8, 2004 (Pub. No. US 2007/0015267, Pub date Jan. 18, 2007)) disclosed a method of producing s composite object consisting of at least two distinct parts.
Mercuri, et al. (“Flexible Graphite Article and Method of Manufacture,” U.S. Pat. No. 6,432,336, Aug. 13, 2002 and U.S. Pat. No. 6,706,400, Mar. 16, 2004) disclosed a resin-impregnated flexible graphite sheet exhibiting enhanced isotropy and a method of producing resin-impregnated flexible graphite sheet. The method includes the steps of (i) reacting raw natural graphite flake-like particles with a liquid intercalant solution to form intercalated graphite particles; (ii) exposing the intercalated graphite particles to a temperature of at least about 700° C. to expand the intercalated graphite particles to form a stream of exfoliated graphite particles; (iii) continuously compressing the stream of exfoliated graphite particles into a continuous coherent self-supporting mat of flexible graphite; (iv) continuously contacting the flexible graphite mat with liquid resin and impregnating the mat with liquid resin; and (v) continuously calendering the flexible graphite mat to increase the density thereof to form a continuous flexible graphite sheet having a thickness of no more than about 1.0 inch.
It is of interest to note that this process disclosed by Mercuri, et al. involves compressing the exfoliated graphite into a flat mat prior to impregnating the mat with a resin. This sequence is disadvantageous in that the re-compressed flexible graphite, being much denser, is less permeable to resin impregnation. Furthermore, uniaxial re-compression of the exfoliated graphite prior to resin impregnation tends to align or orientate the graphite flakes along the graphite sheet plane direction (perpendicular to the re-compression vector), resulting in a more anisotropic flexible graphite sheet composite. Once these flakes were well-aligned in a sheet to form a highly cohered mat, their orientations could no longer be changed during subsequent resin impregnation and molding operations. Furthermore, no attempt was made to re-compress the mat in different directions. Thin graphite flakes are essentially single crystals with the flake plane parallel to the basal plane and, hence, exhibit a high electrical conductivity along thin flake plane directions and much lower conductivity along the thickness direction, or c-axis direction. Consequently, the bipolar plates prepared by using the Mercuri process are not expected to have a high thickness-direction conductivity.
The resin-impregnated flexible graphite sheet exhibiting enhanced isotropy as disclosed by Mercuri, et al. (U.S. Pat. No. 6,706,400) was said to contain interlocked particles of expanded graphite. A portion of these interlocked particles of expanded graphite was substantially unaligned with the opposed planar surfaces. However, Mercuri, et al. did not fairly specify how these unaligned graphite flakes were obtained. Presumably, this could be achieved by mixing large particles of exfoliated graphite with smaller particles of exfoliated graphite, as implied in a Mercuri's earlier patent (U.S. Pat. No. 5,846,459, Dec. 8, 1998). The trade literature published by Grafrech (assignee of Mercuri's patents) indicates the electrical resistivity of bipolar plates in the X-Y plane as 7 μOhm-m (in-plane conductivity=1428 S/cm) and in the Z-direction as 300 μOhm-m (thickness-direction conductivity=33 S/cm). The thickness-direction conductivity is unsatisfactory.
In addition to exhibiting high electrical conductivity, the flow field plate or bipolar plate should be constructed from inexpensive starting materials, materials that are easily formed into any plate configuration, preferably using a continuous molding process, and materials that are corrosion resistant in low temperature fuel cells and that do not require further processing such as high temperature pyrolyzation treatments. The above review clearly indicates that prior art bipolar plate material compositions and processes have not provided a satisfactory solution for the fuel cell industry.
Accordingly, an object of the present invention is to provide an exfoliated graphite composite composition that exhibits a relatively high thickness-direction conductivity.
Another object of the present invention is to provide a composite composition that can be easily molded or embossed into a flow field plate or bipolar plate.
Still another object of the present invention is to provide an exfoliated graphite composite that is intrinsically less anisotropic.
Yet another object of the present invention is to provide a process for producing exfoliated graphite composites with enhanced isotropy. Such a process can be continuous, automated, and adaptable for mass production of bipolar plates.
This invention provides a method of producing an electrically conductive, less anisotropic, and structurally sound composite composition for fuel cell flow field plate or bipolar plate applications. The composition comprises: (a) expanded or exfoliated graphite; (b) particles of non-expandable graphite or carbon, wherein these particles are between 3% and 60% by weight based on the total weight of the particles and the expanded graphite; and (c) a binder or matrix material to bond the expanded graphite and the particles of non-expanded graphite or carbon for forming a highly conductive composite, wherein the binder or matrix material is between 3% and 60% by weight based on the total composite composition weight. The composite plate exhibits a thickness-direction conductivity typically greater than 35 S/cm, more typically greater than 50 S/cm, most typically greater than 100 S/cm, and a thickness-direction specific areal conductivity greater than 200 S/cm2, more typically greater than 500-1,500 S/cm2.
In one preferred embodiment, a method of producing an electrically conductive composite composition includes the following steps: (a) providing a supply of expandable graphite powder; (b) providing a supply of a non-expandable powder component comprising a binder or matrix material (preferably also comprising an isotropy-promoting agent such as non-expandable natural graphite particles); (c) blending the expandable graphite with the non-expandable powder component to form a powder mixture wherein the non-expandable powder component is in the amount of between 3% and 60% by weight based on the total weight of the powder mixture; (d) exposing the powder mixture to a temperature sufficient for exfoliating the expandable graphite to obtain a compressible mixture comprising expanded graphite worms and the non-expandable component; (e) compressing the compressible mixture at a pressure within the range of from about 5 psi to about 50,000 psi in predetermined directions into predetermined forms of cohered graphite composite compact; and (f) treating the so-formed cohered graphite composite to activate the binder or matrix material thereby promoting adhesion within the compact to produce the desired composite composition. Step (e) may comprise an uniaxial compression, a biaxial compression, a triaxial compression, and/or an isostatic compression. Preferably, the composite composition is subjected to a biaxial, triaxial, and/or isostatic compression, prior to a final shaping operation to obtain a bipolar plate. This final shaping operation can involve an uniaxial compression, shearing, impression, embossing, compression molding, or a combination thereof to form a flow field plate or bipolar plate. The plate is preferably smaller than 1 mm and more preferably thinner than 0.5 mm. This final operation typically involves a combination of uniaxial compression and some shearing, which could bring the final composite plate back to a less isotropic state (as compared to the composition prior to this final shaping operation). The presence of a non-expandable powder component (e.g., fine particles of natural graphite) serves to eliminate or reduce this further anisotropy induced by the final shaping operation.
Other preferred embodiments involve adding the non-expandable powder (isotropy-promoting agent) and/or binder material to the exfoliated graphite worms (rather than prior to exfoliation). In the case of adding a liquid resin into the worms, it is preferred that one component (e.g., curing agent) of a two-component resin system is impregnated into the graphite worms first, followed by impregnation of the second component. Further preferably, the curing agent is diluted by a volatile liquid (e.g., acetone) to reduce the surface tension and viscosity of the curing agent to facilitate surface wetting and impregnation of the worms.
It may be noted that the US Department of Energy (DOE) target for composite bipolar plates includes a bulk electrical conductivity of 100 S/cm or an areal conductivity of 200 S/cm2, where the areal conductivity is essentially the ratio of the thickness-direction conductivity to the plate thickness. This implies that a thinner plate has a higher areal conductivity, given the same thickness-direction conductivity. One of the advantages of the presently invented composite composition is the notion that this composition can be prepared in such a manner that the resulting composite plate can be as thin as 0.3 mm or thinner, in sharp contrast to the conventional graphite bipolar plates which typically have a thickness of 3-5 mm. This, when coupled with the fact that bipolar plates typically occupy nearly 90% of the total fuel cell stack thickness, implies that our technology enables the fuel cell stack size to be reduced dramatically. The resulting plates have electrical conductivities far exceeding the DOE target values, which was an original objective of the DOE-sponsored research and development work that resulted in the present invention.
A prior art fuel cell, as shown in
In practice, the proton exchange membrane in a PEM-based fuel cell is typically coated on both sides with a catalyst (e.g., Pt/Ru or Pt) to form a catalyst-coated membrane 9 (CCM). The CCM layer 9 is then sandwiched between an anode backing layer 10 (diffuser) and a cathode backing layer 12 (diffuser). The resulting five-layer assembly is called a membrane electrode assembly 8 (MEA). Although some fuel cell workers sometimes refer to CCM as a MEA, we prefer to take the MEA to mean a five-layer configuration: anode backing layer, anode catalyst layer, PEM, cathode catalyst layer, and cathode backing layer.
The fuel cell also comprises a pair of fluid distribution plates (also referred to as fluid flow field plates) 21 and 23, which are positioned on opposite sides of membrane electrode assembly 8. Plate 21, which serves as a fuel distribution plate, is shaped to define fuel flow channels 22 facing towards anode diffuser 10. Channels 22 are designed to uniformly deliver the fuel to the diffuser, which transports the fuel to the anode catalyst layer 16. An input port and an output port (not shown), being in fluid communication with channels 22, may also be provided in flow field plate 21 so that carbon dioxide (in a DMFC) can be withdrawn from channels 22.
Flow field plate 23 is shaped to include fluid channels 24 for passage of a quantity of gaseous oxygen (or air). An input port and an output port (not shown) are provided in plate 23, which are in fluid communication with channels 24 so that oxygen (or air) can be transported through the input port to the cathode diffuser 12 and cathode catalyst layer 18, and water and excess oxygen (or air) can be withdrawn from channels 24 through the output port. Plate 23 is electrically conductive and in electrical contact with cathode diffuser 12. It can be used as a uni-polar plate (the positive terminal of the electrical current generated by the fuel cell unit) or as a part of a bi-polar plate (if integrated with fuel flow field plate 21). Shown in
As indicated earlier, bipolar plates can be made from an electrically conductive flexible graphite sheet, which is then impregnated with a resin (e.g., Mercuri, et al., U.S. Pat. No. 6,432,336, Aug. 13, 2002 and U.S. Pat. No. 6,706,400, Mar. 16, 2004). Flexible graphite sheets are compressible and embossing processes may be used to form channels in one or both major surfaces of a sheet. Conventionally, flexible graphite is obtained first by intercalating graphite with an intercalating agent (also referred to as an intercalate or intercalant) to form a graphite intercalation compound (GIC). Then, the GIC is exposed to a thermal shock at a temperature of 700-1,050° C. for a short duration of time (20-60 seconds) to expand or exfoliate graphite. The exfoliation is characterized by an expansion of graphite particles up to a ratio of typically 80-300 times in the c-axis direction perpendicular to the graphene or basal plane of the graphite crystal structure. The exfoliated graphite particles are vermiform in appearance, and are therefore commonly referred to as worms. The worms may be re-compressed together into flexible sheets which are characterized by having most of the graphite flakes oriented parallel to the two opposed exterior surfaces, which are perpendicular to the c-axis. These thin sheets (foils or films) are referred to as flexible graphite. Flexible graphite can be wound up on a drum to form a roll of thin film, just like a roll of thin plastic film or paper. Although a flexible graphite sheet is typically highly conductive along the sheet plane directions, their thickness-direction conductivity is rather poor (reported to be approximately 15 S/cm).
The present invention provides a highly conductive, less anisotropic composite composition that is based on exfoliated graphite (but generally not based on flexible graphite sheets) and can be easily molded into a flow field plate or bipolar plate. The resulting composite plates exhibit a thickness-direction conductivity typically greater than 35 S/cm, more typically greater than 50 S/cm, often greater than 100 S/cm, and in many cases, greater than 200 S/cm. These impressive conductivity values hitherto have not been known to be achievable with prior art resin-impregnated flexible graphite composites. The invented composition comprises: (a) expanded or exfoliated graphite (including, for instance, expanded graphite, expanded graphite oxide, and expanded graphite fluoride containing less than 20% of non-carbon elements); (b) particles of non-expandable graphite or carbon (e.g., natural graphite particles and carbon black serving as an isotropy-promoting agent), wherein the amount of the non-expandable graphite or carbon is between 3% and 60% by weight based on the total weight of the particles and the expanded graphite together; and (c) a binder or matrix material to bond the expanded graphite and the particles of non-expanded graphite or carbon for forming a highly conductive composite, wherein the binder or matrix material amount is between 3% and 60% by weight based on the total composite composition weight. The composite composition, after molding to form a plate, has a flexural strength typically greater than 50 MPa and can be higher than 100 MPa if other reinforcement components such as graphite fibers are incorporated. It may be noted that the exfoliated graphite in the instant invention comprises exfoliated natural graphite, synthetical graphite, highly oriented pyrolytic graphite, graphite fiber, graphitic nano-fiber, graphite oxide, graphite fluoride, chemically modified graphite, or a combination thereof. These species form a laminar or layered structure and can be intercalated and exfoliated.
We have surprisingly found that the presence of non-expandable graphite particles (whether larger or smaller than the exfoliated flake sizes) effectively promotes or facilitates more isotropic orientations of exfoliated flakes, resulting in a much higher thickness-direction conductivity, typically much greater than 50 S/cm. This is a highly desirable feature of a bipolar plate since electrons produced by a fuel cell stack flow along this direction. This increase in thickness-direction conductivity is achieved with a slightly reduced in-plane conductivity, which is still very high (500-1,000 S/cm). The slight reduction in the in-plane conductivity is not a concern for fuel cell bipolar plate applications.
The composite composition can further comprise a reinforcement or filler selected from the group consisting of graphite/carbon fiber, graphite/carbon nano-fiber, nano-tube, glass fiber, ceramic fiber, polymer fiber, metal fiber, metal particle, polymer particle, organic particle, inorganic particle, or a combination thereof, wherein the reinforcement or filler is between 0.5% and 30% by weight based on the total weight of expanded graphite, particles of non-expanded graphite or carbon, and reinforcement or filler. In addition to serving as an isotropy-promoting agent, this reinforcement or filler can impart other desired properties to the resulting exfoliated graphite composite. The reinforcement or filler is preferably electrically conductive (e.g., graphite fiber). Although ceramic or glass fibers were incorporated in a prior art resin-impregnated flexible graphite sheet composite, these stiff fibers were used solely or primarily for the purpose of puncturing the exfoliated graphite flakes to enhance resin impregnation (Mercuri, et al., U.S. Pat. No. 5,885,728, Mar. 23, 1999 and U.S. Pat. No. 6,037,074, Mar. 14, 2000). Furthermore, these fibers were not electrically conductive and, hence, could reduce the electrical conductivity of the flexible graphite sheet and its resin-impregnated version. It may be further noted that Mercuri (U.S. Pat. No. 5,846,459, Dec. 8, 1998) did suggest that an amount of exfoliated graphite flakes of smaller dimensions could be mixed with exfoliated graphite flakes of larger sizes to enhance the isotropy of flexible graphite sheets. However, it was not clear if this approach could be adapted to effectively improve the isotropy in the resin-impregnated flexible graphite sheet composite. The best available data published by Graffech (assignee of Mercuri's patents) indicates a thickness-direction conductivity of 33 S/cm, which is not very impressive. This data seems to suggest that thin flakes of exfoliated graphite are not very effective in enhancing electrical conductivity isotropy of the resulting flexible graphite sheet or resin-impregnated flexible graphite sheet.
In the presently invented composite composition, the binder or matrix material may be selected from a polymer, ceramic, glass, metal, carbon, polymeric carbon, asphalt, tar, coal tar pitch, petroleum pitch, mesophase pitch, or a combination thereof. The polymer binder may be preferably selected from the group consisting of polyethylene, polypropylene, nylon, polyesters, polytetrafluoroethylene, polyvinylidene fluoride, fluoro polymers, polyacrylonitrile, acrylic resins, epoxides, polyimide, bismale imide, phenol formaldehydes, vinyl ester, isocyanate resins, and combinations thereof. Many polymers (e.g., phenolic resin and polyacrylonitrile), upon exposure to high temperature (300-1,000° C.), can be converted to polymeric carbons, which are much more conductive than the un-pyrolyzed polymers.
The binder or matrix material may be an inorganic vitreous glass-forming material which contains at least one of the compounds selected from the group consisting of boric oxide, silica, phosphorous pentaoxide, germanium oxides, vanadium pentoxides, and beryllium fluoride. The binder or matrix material may be a glass-forming composition containing at least two oxides selected from the group consisting of silica, aluminum oxide, sodium oxide, potassium oxide, magnesium oxide, cuprous oxide, barium oxide, lead oxide, and boric oxide. The binder may be chosen from metals or metal alloys, which are normally very conductive.
The presently invented composite composition and the final bipolar plate can be produced by several unique and effective methods. As one example (Approach 1), schematically shown on the right-hand side of
In this method, step (e) may comprise an uniaxial compression, a biaxial compression, a triaxial compression, and/or an isostatic compression. An uniaxial compression alone tends to produce a more anisotropic composite. A biaxial, triaxial, or isostatic compression produces a composite with reduced anisotropy. As a preferred embodiment of the present invention, the composite composition preferably is prescribed to go through a biaxial, triaxial, and/or isostatic compression, prior to a final shaping operation to obtain a bipolar plate. This final shaping operation can involve an-uniaxial compression, shearing, impression, embossing, compression molding, or a combination thereof. This operation results in the formation of a flow field plate or bipolar plate typically with flow field channels built onto at least one surface of the plate. The plate is preferably thin, smaller than 1 mm and more preferably thinner than 0.5 mm. This final operation typically involves a combination of uniaxial compression and some shearing, which could bring the final composite plate back to a less isotropic state (as compared to the composition prior to this final shaping operation). We have surprisingly found that the presence of a non-expandable powder component (e.g., fine particles of natural graphite) serves to eliminate or reduce this further anisotropy induced by the final shaping operation. This is a non-trial and non-obvious discovery, achieved only after extensive, in-depth research and development efforts.
By contrast, a prior art method of producing exfoliated graphite composites (Olstowski, et al. U.S. Pat. No. 3,492,197), schematically shown on the left-hand side of
The binder or matrix material may be a char-yielding material and the method further comprises a step of baking or pyrolizing the composite at a temperature for a period of time sufficient to convert the char-yielding material into carbon or graphite. The char-yielding material may be selected from the group consisting of asphalt, tar, sugars, phenolic resins, coal tar pitches, petroleum pitches, mesophase pitches, saccharides, organic polymers, and combinations thereof.
A second method (Approach 2) of producing an electrically conductive composite composition is schematically shown in
Again, in this method, step (f) may comprise an uniaxial compression, a biaxial compression, a triaxial compression, and/or an isostatic compression. As a preferred embodiment of the present invention, the composite composition is subjected to a biaxial, triaxial, and/or isostatic compression, prior to a final shaping operation to obtain a bipolar plate. This shaping operation can involve an uniaxial compression, shearing, impression, embossing, compression molding, or a combination thereof. This final shaping operation results in the formation of a flow field plate or bipolar plate typically with flow field channels built onto at least one surface of the plate. The plate is preferably thin, smaller than 1 mm and more preferably thinner than 0.5 mm. Again, the presence of a non-expandable powder component (e.g., fine particles of natural graphite) serves to eliminate or reduce the further anisotropy induced by the final shaping operation.
Again, the binder or matrix material may be a char-yielding material and the method further comprises a step of baking or pyrolizing the composite at a temperature for a period of time sufficient to convert the char-yielding material into carbon or graphite. The char-yielding material may be selected from the group consisting of asphalt, tar, sugars, phenolic resins, coal tar pitches, petroleum pitches, mesophase pitches, saccharides, organic polymers, and combinations thereof.
In a preferred embodiment of Approach 2, step (e) comprises impregnating the compressible mixture with a first component of a two-component or multiple-component thermosetting or polymerizing resin and then impregnating the compressible mixture with a second component of the resin. In particular, step (e) may comprise impregnating the compressible mixture with a mixture of a volatile diluent and a first component of a two-component or multiple-component thermosetting or polymerizing resin, removing the volatile diluent, and then impregnating the compressible mixture with a second component of the resin. A diluent is used to reduce the viscosity and surface energy of the curing agent, promoting surface wetting and impregnation of exfoliated graphite with this curing agent. Once the interior and exterior surfaces of the pores in exfoliated graphite are wetted with the curing agent, subsequent impregnation or infiltration of the resin is essentially spontaneous. This is due to the notion that typically a curing agent is chemically compatible with its matting base resin. Preferably, the resin comprises epoxy resin and the first component of a two-component epoxy system comprises a curing agent or hardener.
Another method of producing a highly conductive bipolar plate composite composition is shown in
The process for producing bipolar plates from exfoliated graphite may be further illustrated by referring to
A series of composite compacts were prepared as follows:
Sample 1-A: Ultrafine polyethylene (PE) powder, having an average particle size of about 10μm, was dry-blended with 30% by weight of non-expandable natural graphite particles and 70% by weight of acid-intercalated, expandable graphite (based on the total weight of expandable and non-expandable graphite). The PE amounts were 5, 15, 25, and 50% by weight based on the total weight of the resulting composite composition. The non-expandable graphite was intended as an isotropy-promoting agent. The three-component mixture was enclosed in a quartz tube, which was purged with nitrogen gas and then loosely sealed from both ends of the tube with ceramic cloth. The tube was rapidly transferred to the center of a tube furnace pre-heated to a temperature of 1,100° C. and maintained at that position for 20 seconds. Rapid expansion or exfoliation of the expandable graphite occurred and, surprisingly, the PE did not suffer any significant thermal degradation as would have been expected by polymer scientists. This might have been due to the notion that PE was exposed to high heat for only a very short period of time. It may be noted that the exfoliated graphite herein used comprises graphite oxide since strong acid intercalation tends to partially oxidize natural graphite.
Sample 1-B: The compositions and process conditions were the same as Sample 1-A with the exception that there was no non-expandable graphite in 1-B.
Sample 1-C: The compositions and process conditions were the same as Sample 1-A with the exception that PE powder was added after (rather than before) exfoliation of the expandable graphite.
Sample 1-D: The compositions and process conditions were the same as Sample 1-A with the exceptions that there was no non-expandable graphite in the composite and PE powder was added after (rather than before) exfoliation of the expandable graphite. This was based on a prior art approach.
Sample 1-E: The compositions and process conditions were the same as Sample 1-A with the exception that the amount of non-expandable graphite was 20% with the remaining 10% being replaced with short graphite fibers for the purpose of enhancing mechanical strength of the resulting composite plate.
A desired amount of the various PE-graphite blends was poured into a mold and uniaxially compressed to a pressure of about 5,000 psi (34.5 MPa), heated to 160° C., and then cooled to produce thin, flat plates (approximately 1 mm thick).
A comparison of the conductivity data between Sample 1-A and Sample 1-B indicates that Sample 1-A is more isotropic, providing a much higher thickness-direction conductivity. This is due to the presence of non-expandable graphite particles that serves as an isotropy-promoting agent. Sample 1-A appears to be a much better candidate material for bipolar plate applications. A comparison of the flexural strength data between Sample 1-A and Sample 1-C indicates that Sample 1-A provides a higher strength, likely due to the notion that mixing the binder material powder with the expandable graphite (prior to exfoliation) leads to a more homogeneous composite. A comparison of the conductivity data between Sample 1-A and Sample 1-D indicates that Sample 1-A is more isotropic, providing a much higher thickness-direction conductivity, and exhibiting a higher strength. Clearly, the presently invented composition is far superior to the prior art exfoliated graphite composite composition that contains no isotropy-promoting agent in the composite and the binder material was added after (rather than before) exfoliation of the expandable graphite. A comparison of the conductivity data between Sample 1-A and Sample 1-E indicates that Sample 1-E (containing high-strength graphite fibers) exhibits a much higher flexural strength. Furthermore, the addition of graphite fibers did not seem to compromise the electrical conductivity of the resulting composite. This implies that different properties can be tailored independently and this class of composite materials and related processes are versatile.
Sample 2-A is identical to sample 1-A (15% PE) and Sample 2-D is identical to sample 1-D. However, Samples 2-A and 2-D were subjected to bi-axial compression (the first compression vector is defined as the X-axis direction and the second compression vector is the Y-axis direction) at a final pressure of 500 psi and, separately, tri-axial compression (samples compressed in X- and Y-axis directions first, followed by a final Z-axis compression (12,500 psi) to form a thin plate). The samples were consolidated (heated to above 160° C.) and then cooled under a final pressure of 500 psi (sample of biaxial compressions only) and 12,500 psi (triaxial compression sample), respectively. The electrical conductivity values in the three directions of all samples are given in Table 2:
A comparison of the conductivity data between Sample 2-A and Sample 2-D indicates that Sample 2-A (containing an isotropy-promoting agent) is more isotropic, providing a much higher thickness-direction (Z-direction) conductivity. Both samples show very impressive thickness-direction conductivity values (335 S/cm and 120 S/cm), which are much greater than that of prior art flexible graphite composites (33 S/cm at best).
Sample 3-A: Ultrafine zinc powder (approximately 220 nm in average diameter) was prepared by using a twin-arc atomization and gas phase condensation process. This powder was dry-blended with 30% by weight of non-expandable natural graphite particles and 70% by weight of acid-intercalated, expandable graphite (based on the total weight of expandable and non-expandable graphite). The Zn amount was approximately 30% by weight based on the total weight of the resulting composite composition. The non-expandable graphite was intended as an isotropy-promoting agent. The three-component mixture was enclosed in a quartz tube, which was purged with nitrogen gas and then loosely sealed from both ends of the tube with ceramic cloth. The tube was rapidly transferred to the center of a tube furnace pre-heated to a temperature of 1,100° C. and maintained at that position for 20 seconds. Rapid expansion or exfoliation of the expandable graphite occurred. For the preparation of Sample 3-B, the compositions and process conditions were the same as with Sample 3-A with the exception that there was no non-expandable graphite in 3-B. Both mixtures, on a separate basis, were uniaxially compressed in a mold to about 5,000 psi to yield a cohered composite. Then this composite was heated in an inert atmosphere to 425° C., while still under a pressure of approximately 5,000 psi, for 5 minutes and subsequently cooled to room temperature to form a consolidated metal-infiltrated expanded graphite composite plate. The in-plane and thickness-direction conductivity of Sample 3-A are 14,500 S/cm and 350 S/cm, respectively while those of Sample 3-B are 14,670 S/cm and 121 S/cm, respectively.
Sample 4-A: First, 30% by weight of non-expandable natural graphite particles and 70% by weight of bromine-intercalated, expandable graphite (based on the total weight of expandable and non-expandable graphite) were dried blended. The non-expandable graphite was intended as an isotropy-promoting agent. The mixture was enclosed in a quartz tube, which was purged with nitrogen gas and then sealed from both ends of the tube with ceramic cloth. The tube was rapidly transferred to the center of a tube furnace pre-heated to a temperature of 600° C. and maintained at that position for 30 seconds. Rapid expansion or exfoliation of the expandable graphite occurred. The resulting graphite worms were then mixed with 25% by weight of fine phenol-formaldehyde powder, based on the total weight of the resulting composite composition. The resulting mixture was charged into a mold, heated to 200° C., and uniaxially compressed to a pressure of 15,000 psi for one hour and further cured at 270° C. for another hour to form a thin composite plate.
Sample 4-B: The compositions and process conditions were the same as Sample 4-A with the exception that there was no non-expandable graphite in 4-B. The in-plane and thickness-direction conductivity of Sample 4-A are 1,450 S/cm and 81 S/cm, respectively while those of Sample 4-B are 1,467 S/cm and 31 S/cm, respectively. Again, the presence of non-expandable graphite particles did significantly promote isotropy, leading to a better thickness-direction conductivity. With a plate thickness of 0.5 mm, the specific areal conductivity of Sample 4-A is 1620 S/cm2, a very impressive value.
Sample 4-A, retained in a steel mold, was slowly heated to 500° C. for 4 hours and then raised to and maintained at 920° C. for 24 hours in an oxygen-free environment. Phenolic resin, a char-yielding polymer, was carbonized to become a polymeric carbon. The in-plane and thickness-direction conductivity of the carbon-bonded composite are improved to become 1,750 S/cm and 320 S/cm, respectively.
In another embodiment of the instant invention, a glass binder-based vermicular glass with good electrical conductivity, dimensional stability, and corrosion resistance was prepared as follows: About 18 grams of expandable graphite and 7 grams of non-expandable graphite were mixed and then heat-exfoliated at 1,000° C. to obtain an exfoliated graphite-unexpanded graphite mixture. About 22 grams of a commercially available lime glass powder was blended with this mixture by gentle tumbling. The resulting mixture, after heating to 920° C. in a steel mold, was uniaxially compressed to a pressure of about 10,000 psi. Upon cooling back to room temperature, the composite plate exhibits an in-plane conductivity of 950 S/cm and thickness-direction conductivity of 23 S/cm. With a thickness of 1 mm, this plate still provides a specific areal conductivity of 230 S/cm2, exceeding the DOE conductivity requirement for a composite bipolar plate.
Sample 7-A: First, 30% by weight of non-expandable natural graphite particles and 70% by weight of bromine-intercalated, expandable graphite (based on the total weight of expandable and non-expandable graphite) were dried blended. The non-expandable graphite was intended as an isotropy-promoting agent. The mixture was enclosed in a quartz tube, which was purged with nitrogen gas and then sealed from both ends of the tube with ceramic cloth. The tube was rapidly transferred to the center of a tube furnace pre-heated to a temperature of 600° C. and maintained at that position for 30 seconds. Rapid expansion of the expandable graphite occurred, resulting in a compressible mixture of exfoliated graphite worms and non-expanded graphite.
This compressible mixture was impregnated with a mixture of a volatile diluent (acetone with a quantity 3 times the weight of a curing agent) and the curing agent of a two-component epoxy resin. The diluent was used to reduce the viscosity and surface energy of the curing agent, thereby promoting impregnation and wetting of exfoliated graphite with this curing agent. Upon completion of the impregnation procedure, the volatile diluent was removed under a chemical fume hood. The curing agent-impregnated compressible mixture was then impregnated with the epoxide, the second component of the epoxy resin system. Once the interior and exterior surfaces of the pores in exfoliated graphite were wetted with the curing agent, subsequent impregnation or infiltration of the resin was essentially spontaneous. This is a very effective way of impregnating graphite worms.
Sample 7-B: The compositions and process conditions were the same as Sample 7-A with the exception that there was no non-expandable graphite in 7-B.
Samples 7-A and 7-B were subjected to bi-axial compression (first compression vector is defined as the X-axis direction and second compression vector is the Y-axis direction) at a final pressure of 5,000 psi and, separately, tri-axial compression (samples compressed in X- and Y-axis directions first, followed by a final Z-axis compression (12,500 psi) to form a thin plate). Additionally, a desired amount of both samples was directly compressed uniaxially to form a thin plate. The samples were cured at 160° C. for two hours. The final composite plates contain approximately 17% epoxy resin. The electrical conductivity values in the three directions of all samples are given in Table 3:
The data again demonstrates that non-expandable graphite particles are an effective isotropy-promoting agent, resulting in exceptional thickness-direction conductivity. Pre-compressions in one or two directions (X- and Y-direction), prior to the final shaping operation (Z-direction), provides an effective way of producing exfoliated graphite-based bipolar plates with excellent electrical conductivity properties.
In summary, the present invention provides the fuel cell industry with a highly conductive flow field plate or bipolar plate component. The resulting fuel cell system is of lower costs (due to their amenability to mass production) and better performance (due to lower contact resistance and internal resistance and, hence, higher voltage and power output). The presently invented composite composition has the following additional features and advantages:
This invention is based on the research results of a project supported by the US Department of Energy (DOE) SBIR-STTR Program. The US government has certain rights on this invention.
