Regeneration methods to remove carbon monoxide from reformate fuel using an adsorption/electro-catalytic oxidation (ECO) approach

BACKGROUND OF THE INVENTION

[0002] The invention relates to carbon monoxide (CO) removal from a hydrocarbon reformate fuel. More specifically, it relates to apparatus and methods that use a catalytic material to adsorb carbon monoxide and an electrical current to initiate a chemical reaction between an oxidizing agent and carbon monoxide that has been adsorbed by the catalytic material, thereby regenerating the material.

[0003] Fuel cells convert the chemical energy in the fuel directly into electrical energy through an electrochemical reaction. Because they do not operate on the principle of gas expansion through combustion, they do not suffer the same limitations of thermodynamic efficiency commonly found in internal combustion engines. Accordingly, it is possible for fuel cells to achieve a level of conversion efficiency far greater than that seen in most traditional industrial processes. Additionally, fuel cells make it possible to use renewable forms of energy such as methanol and ethanol through a fuel processor, thereby conserving the limited fossil fuel resources of the planet. Moreover, because of the operating environment of a fuel processor and fuel cell, hydrocarbon, nitrogen oxide and carbon monoxide emissions are negligible, approaching a zero emission state.

[0004] While there are several types of fuel cells existing in practice, this invention is targeted mainly for applications in polymer electrolyte fuel cells (PEFCs), which are also known as proton exchange membrane fuel cells (PEMFCs). A very efficient PEFC uses pure hydrogen for fuel and oxygen for an oxidant. Pure hydrogen, however, has traditionally been difficult to handle and relatively expensive to store and distribute. Consequently, attempts have been made to use hydrogen rich gas mixtures obtained from reforming of various hydrocarbon fuels known as the reformate. To obtain a convenient and safe source of hydrogen for the automotive fuel cell application, on-board reforming of hydrocarbon based fuels, such as gasoline and methanol, is expected to be utilized. However, the reformate from the reforming usually contains nitrogen, carbon dioxide, and low levels of carbon monoxide in the range from 100's of ppm to a few percent. While the presence of carbon dioxide generally has little effect on the efficient operation of a fuel cell, even relatively low concentrations of carbon monoxide can degrade fuel cell performance. The degradation results from the carbon monoxide chemically adsorbing over the active sites in the anode of the fuel cell. Thus, the removal of carbon monoxide from fuel has become a major concern in the advancement of PEFC technology.

[0005] Prior attempts to remove carbon monoxide from a gas mixture include a pressure swing adsorption method disclosed in Nishida et al., U.S. Pat. No. 4,743,276. They disclose a method for selectively absorbing carbon monoxide by means of Cu(l) disposed on a zeolite support, including the step of adiabatically compressing a gas mixture in the pressure range of 0.5 kg/cm2. Golden et al., U.S. Pat. No. 5,531,809 disclose a vacuum swing method as a variation of the pressure swing method disclosed in Nishida. A solid absorbent is selected which physically absorbs carbon monoxide under pressure. When the pressure is reduced to the range of approximately 20 to 100 torr, the carbon monoxide is released from the solid absorbent. By cyclically repeating this process, carbon monoxide may be removed from a gas.

[0006] There are, however, multiple limitations to applying the pressure swing adsorption method to fuel cell applications. First, bulky and expensive pressure resistant tanks, as well as pressure and vacuum pump apparatus, are required to carry out the process. The parasitic weight and volume of these devices make it extremely difficult to apply the pressure swing adsorption method for transportation applications such as a fuel cell power plant for an automobile. A second disadvantage of this approach is the significant power expenditure necessary to cycle the pressurization and depressurization steps. This additional power consumption will result in the reduction of overall efficiency of the fuel cell system. A third disadvantage of this process is that the toxic carbon monoxide released from desorption has to be converted to carbon dioxide with additional process steps and equipment.

[0007] Another prior art process has been referred to as preferential catalytic oxidation (PROX) of carbon monoxide which was documented in U.S. Pat. No. 5,271,916 by Vanderborgh et. al. In the PROX process, a small amount of pure oxygen or air is mixed into the reformate fuel before it enters a single or multiple stage catalytic reactor. The catalyst in the reactor, which usually contains dispersed precious metals such as platinum, ruthenium, iridium, etc., preferentially reacts with carbon monoxide and oxygen to convert them to carbon dioxide. Due to the limited selectivity, however, more than a stoichiometric amount of oxygen is needed to reduce carbon monoxide to an acceptable level. The excess oxygen will oxidize the hydrogen in the reformate fuel therefore to reduce the overall fuel cell efficiency. Even with the PROX process, the concentration of CO in the reformate stream is often still significantly higher than the desirable level for sustainable PEMFC operation. Furthermore, the temperature of the PROX process has to be strictly balanced to promote the oxidation of carbon monoxide and to avoid CO being formed from the reaction of carbon dioxide and hydrogen via a process known as the reverse water-gas shift reaction.

[0008] To further eliminate residual carbon monoxide that escapes from the pretreatment or forms from the reverse water-gas-shift reaction inside of the fuel cell, a direct oxygen injection to fuel cell method was developed. For example, Gottesfeld, U.S. Pat. No. 4,910,099 discloses a method of injecting a stream of oxygen or air into the hydrogen fuel to oxidize the carbon monoxide. Pow et al., U.S. Pat. No. 5,316,747 disclose a similar means of eliminating carbon monoxide directly by introducing pure oxygen or an oxygen containing gas along the latter portion of a reaction chamber in an isothermal reactor in the presence of a catalyst that enhances the oxidation of the carbon monoxide. Wilkinson et al., U.S. Pat. No. 5,482,680 disclose the removal of carbon monoxide from a hydrogen fuel for a fuel cell by means of introducing a hydrogen rich reactant stream into a passageway having an inlet, an outlet and a catalyst that enhances the oxidation of carbon monoxide; introducing a first oxygen containing gas stream into the hydrogen rich reactant stream through a first port along the passage way, thereby oxidizing some of the carbon monoxide within the reactant stream; and introducing a second oxygen containing gas at a subsequent point, further oxidizing the remaining carbon monoxide. Wilkinson et al., U.S. Pat. No. 5,432,021 similarly oxidize carbon monoxide to carbon dioxide by means of an oxygen containing gas introduced into a hydrogen rich reactant stream in the presence of an unspecified catalyst.

[0009] There are several significant limitations for the PROX process and oxygen injection process. One of these limitations is the parasitic consumption of hydrogen. Due to the limited selectivity, the oxidant injected into a hydrogen rich fuel is always higher than the stoichiometric amount necessary for oxidizing the carbon monoxide. The unreacted oxygen will consume hydrogen in the stream to therefore reduce the overall fuel efficiency. Another significant limitation of these methods is their poor tolerance towards the variation of CO input level in the reformate. To minimize the parasitic hydrogen loss, the oxygen to CO ratio needs to be kept at a relatively low level in both approaches. Yet, the CO input level often varies as the result of change of the fuel cell power output and, thus, the reformate throughput. It is difficult to constantly match the CO input level with the oxygen level in a dynamic environment. Consequently, unreacted CO will exceed the fuel cell tolerance level, leading to poor performance. Yet another limitation of these two approaches is the concern over safety. The oxygen to hydrogen ratio in the mixture has to be strictly controlled below the explosion threshold.

[0010] Another prior art process for removing carbon monoxide involves membrane separation, whereby the hydrogen in the reformate can be separated by a metallic membrane. For example, R. E. Buxbaum, U.S. Pat. No. 5,215,729 discloses a palladium based metallic membrane, which provides the selectivity for hydrogen separation up to 100%, therefore removing carbon monoxide and other components from hydrogen. Although highly selective, the process has several disadvantages. Since it uses precious metal as membrane material, it is expensive. Furthermore, the reformate has to be pressurized to facilitate the separation process which results in parasitic power loss and equipment complexity.

[0011] Methanation is another prior art process to remove carbon monoxide through the catalytic reaction of carbon monoxide with hydrogen to form methane. An example of this method is given by Fleming et. al., U.S. Pat. No. 3,884,838. Methane does not have a detrimental impact and is regarded as non-reactive in the fuel cell. The methanation reaction, however, requires hydrogen as a reactant and, therefore, increases the parasitic consumption of the fuel for the fuel cell. Furthermore, under the methanation condition, not only carbon monoxide but also carbon dioxide participates in the reactions. The reaction of carbon dioxide with hydrogen generates carbon monoxide through chemical equilibrium. It is therefore difficult to reduce the carbon monoxide level to the desirable limit for PEFC operation.

[0012] Because of the sensitive nature of the polymer electrolyte fuel cells, it is vital that carbon monoxide removal approaches 100% efficiency. In addition to the process limitations, such as cost, excess volume and weight, system complexity and high parasitic hydrogen consumption for the above mentioned methods, there is another common shortcoming, i.e., slow response during the cold start of the fuel cell power plant. Most of these approaches need the system to reach a certain temperature before they are operable, which often represents an undesirable delay between start-up and normal operation.

[0013] In W/O 00/16880, an electrolytic oxidation (ECO) processor for removing carbon monoxide from a hydrocarbon reformate comprises a cell containing an electrode catalytic material that preferentially adsorbs and reacts with carbon monoxide. An oxidizing agent, activated by the electric current, reacts with the carbon monoxide absorbed by the catalytic material, converting the carbon monoxide to carbon dioxide, and consequently regenerates an adsorption capacity of that catalytic material. Electrical leads disposed on opposite sides of a proton permeable membrane electrode assembly form a circuit capable of discharging a current through the catalytic material, thereby triggering the regeneration process. The circuit may comprise a separate external D.C. power supply.

[0014] However, an external D.C. power supply leads to higher cost, additional power consumption and more system complexity for the carbon monoxide removal system. Consequently, alternative methods need to be developed which can effectively remove the carbon monoxide adsorbed in ECO processor without using an external power supply.

SUMMARY OF THE INVENTION

[0015] In the present invention, a method for removing carbon monoxide from a hydrocarbon reformate in an electro-catalytic oxidation processor having a first electro-catalytic oxidation cell comprises producing an electrical current from a first and a second electro-catalytic oxidation cell and through the first electro-catalytic oxidation cell that has adsorbed the carbon monoxide and through the second electro-catalytic oxidation cell that may or may not have adsorbed the carbon monoixde; and converting the carbon monoxide that has adsorbed in the first and the second electro-catalytic oxidation cells to carbon dioxide to thereby regenerate an adsorption capacity of the electro-catalytic oxidation processor.

[0016] In another aspect of the present invention, a method of removing carbon monoxide adsorbed in an electro-catalytic oxidation processor having a first electro-catalytic oxidation cell comprises producing an electrical current from a first charge storage device such as a capacitor, which is previously charged by the first electro-catalytic oxidation cell, and through the first electro-catalytic oxidation cell that has adsorbed the carbon monoxide; and converting the carbon monoxide that has adsorbed in the first electro-catalytic oxidation cell to carbon dioxide to thereby regenerate an adsorption capacity of the electro-catalytic oxidation processor.

[0017] In a further aspect of the present invention, an electro-catalytic oxidation processor for removing carbon monoxide from a hydrocarbon reformate comprises a first electro-catalytic oxidation cell; a second electro-catalytic oxidation cell; and a switch disposed between the first and second catalytic oxidation cells.

[0018] In a still further aspect of the present invention, an electro-catalytic oxidation processor for removing carbon monoxide from a hydrocarbon reformate comprises a first electro-catalytic oxidation cell; a first charge storage device; and a switch disposed between the first electro-catalytic oxidation cell and first charge storage device.

[0019] These features, aspects and advantages of the present invention will become better understood with reference to the following drawings, description and claim.

BRIEF DESCRIPTION OF THE DRAWINGS

[0020]
FIG. 1 is a block diagram of a method and apparatus for a system of removing carbon monoxide from a reformate fuel according to one embodiment of the present invention;

[0021]

FIG. 2

a
schematically depicts the cross-section of a membrane electrode assembly (MEA) in an electrode catalytic oxidation (ECO) cell according to one embodiment of the present invention;

[0022]

FIG. 2

b
schematically depicts a bipolar plate that can be attached on anode and cathode sides of the MEA shown in FIG. 2a;

[0023]
FIG. 3 is a block diagram of an ECO processor comprising two ECO cells and a switch electrically connected in series according to an embodiment of the present invention;

[0024]

FIG. 4

a
is a block diagram of an ECO processor comprising multiple ECO cells and switches electrically connected in parallel with all of the switches set at the same direction according to another embodiment of the present invention;

[0025]

FIG. 4

b
is a block diagram of the ECO processor shown in FIG. 4a with all of the switches set at the same direction except one;

[0026]

FIG. 4

c
is a block diagram of the ECO processor shown in FIG. 4a with reformate flow direction running through all the ECO cells.

[0027]
FIG. 5 is a block diagram of an ECO processor comprising an ECO cell, a capacitor as the charge storage device, and a switch according to another embodiment of the present invention;

[0028]

FIG. 6

a
is a block diagram of an ECO processor comprising multiple ECO cells and a single capacitor as the charge storage device with a first switch closed according to yet another embodiment of the present invention.

[0029]

FIG. 6

b
is a block diagram of an ECO processor shown in FIG. 6a with a second switch closed.

[0030]

FIG. 7

a
is a block diagram of an ECO processor comprising multiple ECO cells and a single capacitor as the charge storage device with a fourth ECO cell reverse connected according to yet another embodiment of the present invention.

[0031]

FIG. 7

b
is a block diagram of an ECO processor shown in FIG. 7a with a first switch closed.

[0032]
FIG. 8 is a block diagram of an ECO processor comprising two ECO cells with double pole double throw switches and a charge storage according to yet another embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

[0033]
FIG. 1 is a block diagram of a method and apparatus for the overall system of removing carbon monoxide (CO) from a fuel and producing D.C. power according to the present invention. In general, a hydrocarbon fuel source 56, such as gasoline, natural gas or methanol, is introduced into a fuel processor 57. In the fuel processor 57, the hydrocarbons can react with air or water through partial oxidation or steam reforming to form a hydrocarbon reformate mixture containing hydrogen, carbon monoxide, carbon dioxide, steam, nitrogen and other minor components. The reformate mixture usually undergoes additional steps of catalytic reactions, such as a water-gas-shift reaction, to further promote the reaction between steam and CO to form hydrogen and CO2.

[0034] Upon exiting the fuel processor 57, a hydrogen rich reformate 58 containing a small amount of carbon monoxide (usually less than a few percent) enters an electrode catalytic oxidation (ECO) processor 59 where carbon monoxide is removed from the reformate 58. The reformate 62 exiting from the ECO processor 59 contains very low to near zero concentration of CO. Subsequently, it enters a fuel cell assembly or stack 63 where the hydrogen in reformate 62 is electrochemically oxidized at an anode, and air or oxygen is reduced at a cathode to produce a D.C. power output 64. The operation of the fuel processor 57, the ECO processor 59 and the fuel cell stack 63 can be controlled by a central subsystem 41 which manages the necessary air, water and heat, as well as the operation commands for each stage or step in FIG. 1.

[0035] More specifically, the fuel processor 57 converts the hydrocarbon fuel 56 to the reformate 58 through multiple steps. These steps consist of fuel reforming which includes steam reforming or partial oxidation, high temperature water-gas-shift reaction, low temperature water-gas-shift reaction, as well as reformate conditioning such as humidification and temperature control through a heat transfer process.

[0036] In the steam reforming stage, the hydrocarbon fuel 56 reacts with a water steam 42 over a reforming catalyst at an elevated temperature to form a mixture containing mainly hydrogen, carbon monoxide, carbon dioxide and others. This process is endothermic but energy efficient. In place of the steam reforming, a partial oxidation process can be used in which the hydrocarbon fuel 56 reacts with a small amount of oxygen or air 43 to form a mixture of hydrogen, carbon monoxide, carbon dioxide and others. This process is exothermic and self-sustaining but nonetheless less energy efficient.

[0037] Following the steam reforming or partial oxidation stage, the gas mixture undergoes high temperature (i.e. about 350° to 550° C.) and low temperature (i.e. about 200 to 300° C.) water-gas shift reactions in which the carbon monoxide reacts further with additional steam 42 to form hydrogen and carbon dioxide over the water-gas-shift catalysts. In the present invention, the water-gas-shift reactions not only improve the overall yield of hydrogen in the fuel processor 57, it also reduces the carbon monoxide concentration to typically less than a few percent in the reformate output 58. The above fuel reforming and water-gas shift reaction are well known in the art and described, for example, in “Synthesis Gas and Associated Processes” by Charles N. Satterfield, Chapter 10, page 419-465, McGraw-Hill, New York, 1991, which is incorporated herein by reference.

[0038] Following the water-gas shift reaction, the reformate undergoes a conditioning process during which the humidity and temperature of the reformate output 58 is adjusted to be suitable for PEMFC application. The humidity adjustment is accomplished by mixing water steam and the temperature adjustment is accomplished by thermal transfer through a heat exchanger. In a preferred operating condition, the temperature of the reformate output 58 should be in the range of about 70 to 100° C. and the humidity should be close to 100% relative humidity (RH) at the corresponding temperature.

[0039] The fuel processor 57 is preferably controlled by the central management subsystem 41 as a result of operational data 44 being transmitted there between. The central management subsystem 41 may control any number of operational parameters, such as a water vapor flow 42, an airflow 43, and a coolant flow 45 to the fuel processor 57. In a preferred embodiment, a single integrated electronic management subsystem 41 controls not only the fuel processor 57, but also the ECO processor 59 and the fuel cell (or fuel cell stack) 63, both of which are further described below. It is, however, envisioned to control these components with separate management subsystems.

[0040] Sensing devices (not pictured) well known in the art can be installed in the fuel processor 57, the ECO processor 59, as well as the fuel cell stack 63. These sensors monitor the overall system 40 performance by measuring the parameters which include, but are not limited to, the pressure, temperature, carbon monoxide concentration, output voltage/current, etc. This data will be part of the operational data 44, 46, 49 which will be directed to the management subsystem 41 and receive feedback from the management subsystem 41 for the control of each operation of the unit. Particularly relevant to this invention is the carbon monoxide data collected as part of each operation data 44, 46, and 49. To generate the operational data with respect to carbon monoxide levels, the sensing device according to one embodiment of the present invention is a broadband infrared absorption-detector, although other similar devices can also be used.

[0041] The carbon monoxide concentration in the humidified reformate 58 is too high to be used directly as the fuel for PEM fuel cell stack 63. Therefore, ECO processor 59 is used to remove CO in reformate 58. The reformate 58 enters an anode side 66 of the ECO cell 29 (FIG. 2a) through a flow field 76 of a bipolar plate 74 (described below). It passes across an electrode catalytic material 73a of the anode side 66, which includes a catalyst metal component 68a, and a catalyst support further described below. Thereby, the catalytic metal component 68a chemisorbs the carbon monoxide in the reformate 58. The reference to “chemisorbs” herein is intended to refer to chemical adsorption where the electronic orbital interaction between CO and the active site in the catalytic metal 68a occurs to form a quasi-chemical bond. Subsequent references herein to “adsorb” and “chemisorbs” are used interchangeably unless otherwise specified, such as “physi-adsorption”. The chemisorption of carbon monoxide occurs preferentially over hydrogen. This preferential adsorption is due to a significant difference in the Gibbs free energies of adsorption between carbon monoxide and hydrogen with the catalytic sites. Consequently, the catalytic metal component 68a preferentially absorbs carbon monoxide despite the composition of the hydrocarbon reformate typically comprising far greater percentages of hydrogen.

[0042] Over time and in the process of adsorbing carbon monoxide, the catalytic metal component 68a eventually approaches a point of carbon monoxide saturation, thereby reducing or altogether eliminating its adsorption capacity to adsorb more carbon monoxide from the reformate 58. To maintain an efficient removal of CO from the reformate 58, the catalytic metal component 68a, and specifically its adsorption capacity, should be regenerated. The regeneration preferably occurs before there is any substantial increase of CO concentration at the exit of ECO processor 59. The preferred CO level in the processed reformate 62 is less than 100 PPM.

[0043] Regeneration can occur by removing the CO from the catalytic metal component 68a via an oxidizing agent such as condensed water 42v from the steam in the reformate 58. Specifically, the transient species such as a hydroxyl radical, a hydrogen peroxide radical, etc. are formed from the condensed water adsorbed over the surface of the catalytic metal 68a during an electrochemical process (described below). The transient oxidizing agents from the activated water chemically reacts with the CO that had been adsorbed by the catalytic metal component 68a, CO is subsequently oxidized to carbon dioxide that is generally not harmful to the performance of the fuel cell 63. The carbon dioxide has only weak physi-adsorption (i.e. adsorption due to van der Waals interaction). Therefore, it is easily released by the catalytic metal component 68a and swept away by the continuing flow of reformate 58. With the adsorbed CO now removed, the catalytic metal component 68a is again able to adsorb additional CO. Accordingly, the adsorption capacity of the catalytic metal component 68a has been regenerated.

[0044] To initiate the catalytic oxidation reaction between the oxidizing agent and the carbon monoxide, a current is discharged through the area containing the electrode catalytic material 73a and, specifically, the catalytic metal component 68a. The current will initiate an electrochemical process, which transforms the water vapor 42v adsorbed on the surface of the catalytic metal 68a to highly reactive oxidizing species. Such current discharge can occur in various preferred manners that are referred to herein as electrolytic. Irrespective of the manner of current discharge, during the regeneration period, the catalyzed oxidation reaction produces the carbon dioxide described above.

[0045] The present invention also includes an adsorption cycle, which is distinguished from the regeneration cycle by an absence of electrical current flow and thus an absence of catalytic oxidation reactions. Preferably, a regeneration period or cycle alternates with an adsorption period or cycle as the level of carbon monoxide adsorbed to the catalytic metal component 68a rises and falls. In other words, and for example, during the adsorption cycle, the amount of adsorbed carbon monoxide rises towards the maximum adsorption capacity of the catalytic metal component 68a. Before or upon the catalytic metal component 68a reaching saturation, the regeneration cycle commences, during which time the amount of adsorbed CO falls sharply in a short duration due to the electro-catalytic oxidation of CO to CO2. As can be appreciated, the alternation of regeneration and adsorption can theoretically continue indefinitely.

[0046] Thus, for example, the adsorption cycle is carried out by precluding an electrical current from being developed across the area of catalytic metal component 68a. But upon partial or full carbon monoxide saturation of the catalytic metal component 68a, an electric current can be discharged through the area of the catalytic metal component 68a to initiate the regeneration cycle. Consequently, during the adsorption cycle, essentially no protons flow across a proton permeable membrane 65 of the ECO processor 59, as further described below. But such proton flow does occur during the regeneration cycle.

[0047] Accordingly, no hydrogen consumption takes place during the adsorption period. During the regeneration process, however, a residual amount of hydrogen chemisorbed on the surface of the catalyst metal 68a (further described below) of the ECO cell 29 will participate in electrochemical oxidation reactions on the anode 66. The electrochemical oxidation of hydrogen competes with the electrochemical oxidation of carbon monoxide and water, which are adsorbed on the catalytic metal 68a. Generally, the electro-catalytic oxidation of hydrogen is more facile than that of carbon monoxide and water. However, because the surface of the catalyst metal 68a has already saturated with chemisorbed CO, the rate of CO oxidation is comparable with or even more favorable than that of H2 under the circumstance. The electro-oxidation of hydrogen and water results in the formation of protons. The protons migrate across the proton permeable membrane 65 to the cathode 67 of the ECO cell 29 and react with reduced oxygen at the surface of cathode metal component 68c of the electrode catalyst material 73c to form water. Since the electrochemical process occurring during the regeneration period is usually much faster than the cumulative adsorption process, the adsorption period generally comprises a major portion (i.e., about 50 to 99%) of the overall ECO operation cycle.

[0048] During both the regeneration and adsorption cycles, a substantial amount (i.e., about 60 to 99.9%) of carbon monoxide is removed from the reformate 62 exiting the ECO processor 59. However, it can be appreciated that during the adsorption cycle, the amount of CO in the exit reformate 62 will be rising as the adsorption capacity of the catalytic metal component 68a decreases. To prevent the leakage of CO in the exit reformate 62, the ECO processor 59 is preferably regenerated so that the adsorption capacity of catalytic metal component 68a can be restored in a timely fashion. In any event, the reformate 62 with CO concentration less than about 100 ppm can then enter the fuel cell stack 63, the reformate 62 can react with an oxidant, such as air 47, through an electrochemical process which produces a D.C. electrical power 64. The fuel cell byproducts that include an oxygen-depleted air 51 and a hydrogen-depleted reformate 52 can then be exhausted by the management subsystem 41 in the form of an exhaust 53.

[0049] As mentioned above, FIG. 2a depicts the internal structure of a membrane electrode assembly (MEA) of an ECO cell 29. FIG. 2b depicts a bipolar plate 74 that is used on both sides of the MEA of FIG. 2a. Thereby, the ECO cell 29 is generally constructed in a fashion similar to well known proton exchange membrane (PEM) fuel cells. Such PEM cells, including the construction of bipolar plates and membrane electrode assemblies, are described in the article “Polymer Electrolyte Fuel Cells” by S. Gottesfeld and T. A. Zawodzinski in ADVANCES IN ELECTROCHEMICAL SCIENCE AND ENGINEERING, R. C. Alkire, H Gerischer, D. M Kolb and C. W. Tobias eds., Volume 5, page 195-302, Wiley-VCH, Weinheim, Germany, 1997 and incorporated herein by reference. The ECO cell 29 will be typically operated between ambient temperature to about 100° C. and at about 1 to 5 atmospheres of pressure. The ECO cell 29 includes a first portion and a second portion—namely, the anode 66 and the cathode 67—together with the proton exchange membrane 65 therebetween. Various proton permeable membrane materials that are well known in the art can be used as the proton exchange membrane 65, such as perflourinated polymers like NAFION®.

[0050] The CO adsorption and electrochemical oxidation occurs on the anode side of the ECO cell 29. The anode side consists of the anode 66 and the bipolar plate 74 with the gas flow field 76. As shown in FIG. 2a, the electrode catalytic material 73a includes the catalyst metal component 68a dispersed over a conductive high surface area support 69. On one side of the electrode catalytic material 73a, and in close contact therewith, is the proton exchange membrane 65. On the other side of the electrode catalytic material 73a is a porous conductive gas diffusion backing material 71. The backing material 71 distributes reformate 58 uniformly to the anode 66 and can be made of conductive materials with a gas diffusion property such as carbon cloths or porous carbon papers. An example of a commercial backing material 71 is ELATM made by E-TEK, Inc. The side of the gas diffusion backing material 71 opposite the catalytic material 73a is in close contact with the bipolar plate 74, which is connected to a first conductive lead (not shown). Through the bipolar plate 74 and the first conductive lead, electrons are transferred between the anode 66 and cathode 67 through an external circuit. The backing material 71 can be coated with a hydrophobic coating 72 to prevent local flooding by water from the electrochemical process and from the humidified reformate 58.

[0051] In operation, the reformate 58 containing CO enters the ECO processor 59 through an inlet 75 of the bipolar plate 74. The reformate 58 follows the flow path or feed channel 76 across a conductive surface 77 and to an outlet or exhaust 78. During the process, the reformate 58 will constantly pass through the gas diffusion backing material 71 and interact with the catalyst metal component 68a. As it was mentioned earlier, the CO in the reformate 58 will be selectively chemisorbed over the catalyst metal component 68a. The majority (i.e., about 50 to 99.9%) of the CO at the outlet 78 of the bipolar plate 74 is therefore being removed from the reformate 58.

[0052] To facilitate the proton transfer process during the regeneration cycle, the electrode catalyst material 73a and support 69 are bound to the proton exchange membrane 65 in a matrix of proton conductive ionomer composite 70. Alternatively, the electrode catalyst material 73a and support 69 can be bound to the backing material 71 through the matrix of proton conductive ionomer composite 70 and collectively pressed against the proton exchange membrane 65 upon assembling of the ECO cell 29.

[0053] The cathode 67 is preferably of design similar to the anode 66 to insure that an oxidant such as oxygen is channeled to interact with the protons traversing the membrane 65.

[0054] The catalytic metal component 68a comprises noble and/or transition metals in a highly dispersed form upon the support 69. The support 69 is generally characterized as being electrically conductive, chemically inert, and having a high surface area. The conductivity of the support 69 may vary, but is generally comparable to that of carbon. The need for the support 69 to be chemically inert is to avoid reactions between the reformate 58, the oxidizing species and the support 69 during both adsorption and regeneration cycles and to maintain the structural stability of the anode 66 during long term ECO process operation. In this embodiment, the surface area of support 69 may range from about 5 to 1500 m2/g and, more preferably, range from about 150 to 300 m2/g. Some examples of suitable materials for the support 69 include carbon black, graphitic carbon powder, metal nitride and metal carbide such as titanium nitride, tungsten carbide, etc. The support material for the cathode catalyst metal 68c can be of the same types as that of anode.

[0055] In another embodiment of this invention, the catalytic metal component 68a can be finely divided metal crystallite powder without a supporting material 69. These metal crystallites are generally highly dispersed with particle dimensions ranging from 10 nm to >1000 nm. The benefit of using an unsupported metal crystallite is eliminating the requirement and limitation of the support 69. The unsupported metal crystallite, however, generally provides less available surface area than that of supported catalytic metal component 68a. The cathode metal component can also be metal crystallites with similar particle size.

[0056] The noble metals that are suitable for use as the catalytic metal component 68a include, but are not limited to, ruthenium, rhodium, iridium platinum, palladium, gold, silver, etc. The useful transition metals include, but are not limited to, molybdenum, copper, nickel, manganese, cobalt, chromium, tin, tungsten, etc. The present invention contemplates that two and three noble or transition metals can be used in any combination as the catalytic metal component 68a in the form of a multiple component alloy. However, it is preferred that one or two noble metals and/or one or two transition metals be utilized in any form of combinations as a bimetallic or a trimetallic alloy, which are demonstrated by the examples below.

[0057] Although the catalyst metal component 68a in the anode 66 and the catalyst metal component 68c in the cathode 67 can be the same, catalyst metal component 68c at the cathode 67 is preferably different from that at the anode 66. The preferred catalyst metal component 68c at the cathode 67 includes platinum and platinum-transition metal alloys such as Pt—Co, Pt—Cr. The preferred catalyst metal component 68a at the anode 66 is ruthenium, rhodium, iridium, palladium, platinum and their corresponding transition metal alloys.

[0058] The performance of ECO cell 29 depends on the amount of catalyst metal component 68a and 68c used in the membrane electrode assembly that is normally represented by the weight of the catalyst metal per unit MEA surface area. In this invention, the preferred amount of catalyst metal component 68a for the anode 66 ranges from about 0.1 to 5 mg/cm2. The preferred amount of catalyst metal component 68c for the cathode 67 ranges from about 0.1 to 3 mg/cm2.

[0059] For the electrode catalyst material 73a at the anode 66, the amount of catalyst metal component 68a loading in the support 69 can also affect the performance of the ECO cell 29. For a noble metal based catalytic metal component 68a, the metal loading over the support 69 preferably ranges from about 2 to 70 wt. %. More preferably, the loading is from about 20 to 50 wt. %. Below about 2 wt. %, the net amount of catalyst needed for constructing the anode 66 maybe too high to fully utilize the metal in an electrochemical process where the proton transfer needs to be connected throughout the anode 66. Above 70 wt. %, it is difficult to achieve high metal dispersion which results in lower metal utilization because of the relatively lower surface metal atom to overall metal atom ratio. It is generally believed that the surface metal atoms of the catalyst metal component 68a are the active sites during a catalytic or an electro-catalytic reaction. For a transition metal based catalyst metal component 68a, the metal loading preferably ranges from about 0 to 40 wt. % and, more preferably, from about 3 to 30 wt. %. Loading outside such range tends to result in similar types of performance degradation described above for noble metals. The same loading range is also applicable to the catalyst metal 68c of the electrode catalyst material 73c at the cathode 67.

[0060] As noted above, the catalyst metal component 68a at the anode 66 is dispersed on the substrate 69 with a high dispersion coefficient. The dispersion coefficient is defined as the ratio of the number of surface atoms of an active catalyst metal to the total number of atoms of the metal particles in the catalyst. In this embodiment, it is preferred that the catalyst metal component 68a be characterized by a dispersion coefficient between about 5 to 100% and, more preferably between about 30 to 100%. If below about 5%, the catalyst surface area provided by the catalyst metal component 68a can be too low to utilize the catalyst metal efficiently. The low utilization of the catalyst metal can result in a higher amount of the catalyst metal needed for the anode 66, hence leading to a higher cost of the ECO cell 29. The catalyst metal component 68c at the cathode 67 also has the same preferred range for metal dispersion.

[0061] As mentioned above, the regeneration cycle is initiated by a discharge of electrical current through the ECO cell 29. Without intending to be limited by any electro-catalysis theory, it is believed that the following chemical and electro-chemical processes occur during the adsorption and the regeneration cycles at the surface of catalyst 73a in anode 66. During the adsorption stage, the carbon monoxide in the gas phase will chemisorb over the active site of catalyst metal component 68a, designated as M, to form a chemisorbed CO species, CO/M, through the reaction:

CO+M→CO/M (1)

[0062] Meanwhile, the hydrogen in the gas phase will also participate in a dissociative adsorption over the active site M through the reaction:

H2+2M→2H/M (2)

[0063] Due to the significant difference in the heat of adsorption, the surface concentration of CO/M is higher than that of H/M through the cumulative adsorption of CO. The water vapor in the humidified reformate 58 will also be adsorbed over the surface of the anode 66 to form H2Oads. The surface on which the water adsorbs includes, but is not limited to, the surface of active site M through the following equation:

H2O(gas)→H2Oads (3)

[0064] During the regeneration stage, the following electro-oxidation reactions occur on the surface of the anode catalyst material 73a:

H/M→M+H++e (4)

and

H2Oads→OHads+H++e (5)

[0065] OHads is the hydroxyl group formed over the surface of the anode 66, which is highly reactive and can oxidize the chemisorbed CO/M through the following electro-catalytic reaction:

CO/M+OHads→M+CO2+H++e (6)

[0066] or through the direct catalytic reaction:

CO/M+2OHads→M+CO2+H2O (7)

[0067] Another way to express the electro-catalytic oxidation of chemisorbed carbon monoxide by the activated water is by the following equation:

CO/M+H2Oads→M+CO2+2H++2e (8)

[0068] The carbon dioxide formed through the equations (6) to (8) has a weak interaction with the anode 66 surface and, therefore, will be swept out of the anode 66 after the regeneration.

[0069] As mentioned previously, various approaches can be used to generate current during the regeneration cycle. FIG. 3 depicts a preferred embodiment of an electro-catalytic oxidation (ECO) processor 59 for removing CO from a reformate fuel 58. Instead of comprising an external D.C. power supply, the ECO processor 59 comprises a first electrode catalytic oxidation cell 79, a second electrode catalytic oxidation cell 80, and a first polarity switch 81, all of which are electrically connected in series. Thereby, the need for an external power supply is eliminated. Instead, ECO cells 79, 80 become mutual internal power supplies in lieu of the external power source. The first and second ECO cells 79, 80 may be the same or dissimilar in design.

[0070] As described above, the CO removal process is accomplished in two alternate cycles. In the adsorption cycle, humidified reformate 58 enters anode side 66 of either or both of the first 79 and the second 80 ECO cells (depending upon whether one or both of the cells is in the adsorption cycle), and a majority of the CO in the humidified reformate 58 is then selectively adsorbed by the ECO material 73a on the anode surface 66. During adsorption, the first polarity switch 81 is left open electrically. Therefore, no electric current is passing through either the first ECO cell 79 or the second ECO cell 80 in this time period.

[0071] The CO removal adsorption cycle can be terminated upon or before or after the saturation of the ECO material 73a on the anode surface 66. At this point, the CO removal process can enter the regeneration cycle of regenerating the adsorption capacity of the ECO material 73a on the anode surface 66 through electrochemical oxidation. Upon or before saturation by the CO, the first polarity switch 81 can be closed for a time period, such as about 0.01 to 10 seconds. During this closed time period, the first ECO cell 79 or the second ECO cell 80 can be exposed to the reversed potential that is provided by the opposite ECO cell. An electric current boost will be generated from the “reversed” potential provided by both ECO cells. The oxidizing species will be produced at the surface of the anode 66 of both ECO cells, as a result of reactions similar to that given in equation (5) listed above. The oxidizing species include but are not limited to hydroxyl groups, peroxide groups, etc. These oxidizing species can be formed over the surface of catalyst metal component 68a or other parts of the anode 66 and migrate to the catalyst metal component 68a. The oxidizing species can then react with the adsorbed carbon monoxide to form carbon dioxide. The weakly adsorbed carbon dioxide will then be stripped by the flow of reformate 62, leaving the anode 66 surface “clean.” As the result, the ECO material 73a on the anode 66 surface is cleaned and the ECO processor 59 is ready for another cycle of CO removal.

[0072] The duration of the regeneration is controlled by the central subsystem 41. The regeneration duration can be greater than about 0.001 to about 100 seconds. The preferred embodiment of the present invention is from about 0.01 second to 10 seconds.

[0073] While the above description deals with the ECO processor 59 comprising two ECO cells 79, 80, more than two ECO cells can be used in the ECO processor 59 to enhance the overall CO removal capacity. These multiple ECO cells can be electrically connected in a series or in a parallel pattern, depending on the need of regeneration voltage. It should be noted that the open circuit potential (OCP) for ECO cells connected in parallel equals the OCP of an individual ECO cell. The open circuit potential of several ECO cells connected in series equals a sum of the OCPs of an individual ECO cell.

[0074]

FIG. 4

a
represents multiple ECO cells electrically connected in a parallel configuration. Alternatively, and like with fuel cells, the ECO cells can be stacked in which the individual cells are electrically connected in series. Additionally, ECO processor 59 may contain several modules. In each module, there may be one or more ECO cells that are connected electrically either in series or in parallel. The ECO modules can also be connected electrically either in series or in parallel among themselves in the ECO processor 59 in order to provide necessary voltage and current for regeneration. The regeneration process of multiple ECO cells can occur simultaneously or individually in a sequential manner. Further, the manner in which the reformate 58 gas flows into ECO processor 59 with multiple ECO cells can also be in parallel or in series, depending on the balance between the pressure drop and the mass transfer of CO to the surface of anode 66. Additionally, the reformate 58 gas flow pattern within an ECO cell module with two or more cells can also be either in parallel or in series. The overall flow pattern inside of ECO processor 59 can be the combination of both. For example, the reformate 58 may flow in parallel through the ECO cells within the module while the modules are electrically connected in series. The series flow pattern is preferred for more complete carbon monoxide removal.

[0075]

FIG. 4

a
depicts a variation of the embodiment of the ECO processor 59 shown in FIG. 3 and comprises multiple ECO cells and multiple switches. In FIG. 4a, a first 82, a second 83, a third 84, a fourth 85, and a fifth 86 ECO cell are electrically connected in parallel through a first 87, a second 88, a third 89, a fourth 90, and a fifth 91 double pole double throw (DPDT) switch and wiring 92. Thus, during the adsorption cycle, humidified reformate 58 can pass through ECO cells 82 to 86, either in series or in parallel, while the ECO material 73a on the anode surface 66 of each ECO cell is adsorbing CO from the reformate stream 58. At this point, all of the DPDT switches are set in the same direction as shown in FIG. 4a. Thus, no electric current is generated and no electrode chemical reaction is activated.

[0076] Upon or before or after the saturation of CO over the anode surface 66, the DPDT switch of one or more ECO cells can then be switched to the other poles for a time period that is sufficient to create a reversed potential for regeneration. The switching action may be initiated through an electric signal pulse 93 from the central management system 41. The time period of the regeneration is labeled as Δt in FIG. 4b. At usually ranges from about 0.001 to 10 seconds. For illustration, and as shown in FIG. 4b, the first DPDT switch 87 is switched to the other pole upon receiving the electric signal pulse 93 while the reversed potential is experienced equally by all of the ECO cells 82 to 86 and activates the regeneration cycles. Since only the first DPDT switch 87 is switched to the other pole, while the other DPDT switches remain the same position as shown in FIG. 4b, the first ECO cell 82 receives the sum of the electric current from all of the ECO cells 83 to 86, while the ECO cells 83 to 86 sharing the total electric current because they remain to be electrically connected in parallel. The electric current is created due to the electrochemical activation process of the oxidizing species on the anode 66 surfaces throughout all the ECO cells 82 to 86. However, because the first ECO cell 82 receives a higher current density than the others, it generally results in more complete regeneration. The switch 87 may be returned to its original position shown in FIG. 4a after a brief regeneration time period, Δt, followed by a time period of adsorption, ta labeled in FIG. 4b, before the ECO cells are triggered again for regeneration. The following regeneration, however, may be accomplished by switching another DPDT, for example the switch 88, to provide more complete regeneration for ECO cell 83. Accordingly, all of the DPDT switches should preferably be switched alternatively to the other pole individually or by a group to achieve regeneration completion for each of the ECO cells. The adsorption period ta can be the same or different for each ECO cell during alternative regeneration. The switches used in the current invention can be either mechanical switches or solid-state semiconductor switches for faster and more accurate timing control.

[0077]

FIG. 4

c
is an example that shows the reformate flows through ECO cells in sequence for processor 59 with multiple ECO cells such as that demonstrated in FIGS. 4a and 4b. The ECO cells 82 to 86 are connected in series to process the incoming reformate flow 58. According to the above description, each cell is sequentially regenerated through switching of each polarity switch connected to each cell upon receiving the trigger signal. The advantage of such flow configuration is to maximize the utilization of the CO adsorption capacity on anode 66 surface before regeneration with minimum leakage of carbon monoxide level in the treated reformate 62.

[0078] In another embodiment wherein an external power source is eliminated, one or more charge storage devices such as capacitors may be included in a single or multiple ECO cell circuit to form an ECO processor 59 (FIG. 5). The charge storages may have the same or different capacities compared to each other. Typically, the charge storage used in this application has very high unit weight charge storage capacitance such as the ultracapacitor. In FIG. 5, and for purposes of illustrating a single charge storage and single cell circuit, a first ECO cell 96 is electrically connected with a first charge storage 94 through a first DPDT switch 95 in series. During an adsorption cycle, the first charge storage 94 is being charged by the electric current generated by the first ECO cell 96 until the potential of the first charge storage 94 is equal to that of the open circuit potential of the first ECO cell 96. Once the ECO material 73a on the anode surface 66 is partially or completely saturated with CO, the CO removal process can then enter the regeneration cycle.

[0079] Upon initiating the regeneration cycle, the first DPDT switch 95 is switched to the opposite direction for a time period sufficient to create a reversed potential, such as about 0.01 to 10 seconds, depending on the capacitance of charge storage device 94 and the rate of discharging. At this time, the positively charged end of the first charge storage device 94 is electrically connected with the first ECO cell 96 on the anode side 66, and the negatively charged end of the first charge storage device 94 is connected to the cathode side 67 of the first ECO cell 96. An electric current flow can thereby be produced through the discharge of the first charge storage device 94 as well as the first ECO cell 96. As a result, absorbed CO on the anode surface 66 can be converted to carbon dioxide which exits the ECO processor 59 with the rest of the reformate 62. The discharge process can continue until the potential across the first charge storage device 94 equals about zero. After this point, the current in the circuit can continue, and the first charge storage device 94 can be charged by the first ECO cell 96 except the opposite potential will be applied to the first charge storage device 94. The electric current will not stop until the voltage of charge storage device 94 is equalized by the potential of ECO cell 96. Meanwhile, the regeneration process stops and the adsorption cycle restarts. And again, the first DPDT switch 95 can be switched to the opposite side upon CO saturation for regeneration purpose.

[0080] According to another embodiment of the present invention, FIGS. 6a and 6b depict an ECO processor 59 comprising a first charge storage device 97, such as a capacitor, electrically connected in parallel with a first ECO cell 98, a second ECO cell 99, a third ECO cell 100, and a fourth ECO cell 101 through a first single pole single throw (SPST) switch 102, a second SPST switch 103, a third SPST switch 104, and a fourth SPST switch 105. This embodiment can use the first charge storage device 97 as an intermediate charge storage device to regenerate each of the ECO cells when required. The charging and regeneration cycle is initiated by the pulse signal 107 sent from the central management system 41. For example, the switch 102 is closed when triggered by the signal 107 and the charge storage device 97 is charged by the electric current from the ECO cell 98 until the voltage of charge storage device 97 is equalized by the potential of ECO cell 98, as is shown in FIG. 6a. The switch 102 will be in an open position again after the charge storage device 97 is fully charged. Meanwhile, the adsorption of carbon monoxide continues on the anode 66 surface of remaining ECO cells until the second trigger signal closes the switch 103 upon or before or after saturation of CO over anode catalyst 73a of ECO cell 99, as is shown in FIG. 6b. A reversed potential from charge storage device 97 from FIG. 6a will be applied to the ECO cell 99 hence to activate the regeneration for ECO cell 99 by the discharging current. Again, the electric current will continue to flow until the charge storage device 97 is fully charged, except that the polarity of charge storage device 97 is reversed in comparison with the situation in FIG. 6a. This process can be continued sequentially to the other ECO cells, shown in FIGS. 6a and 6b, until all the cells are regenerated. Subsequently, the process can be repeated for continuous operation.

[0081] Under certain operation conditions, it is preferred that the polarity of the charge storage device such as an ultracapacitor is not frequently reversed for extended usage. Thus, according to yet another embodiment of the present invention, no reversal of charge storage polarity is needed when using a different switching mechanism. One approach is to designate one or multiple ECO cells of the processor 59 for charging purposes only. For example, in FIGS. 7a and 7b, an ECO processor 59 comprises a first charge storage device 108, a first ECO cell, 109, a second ECO cell 110, a third ECO cell 111, and a fourth ECO cell 112 ECO with a respective first SPST 113, a second SPST 114, a third SPST 115, and a fourth SPST 116, which are all connected with a circuit 117. ECO cell 112 is designated to be used only to charge the charge storage device 108 by tentatively closing switch 116 while the other switches 109 to 111 remain open, and re-open switch 116 when the charge storage device 108 is fully charged. The remaining ECO cells 109 through 111 will be used for CO removal through adsorption and regeneration. When one of the cells 109 to 111, for example cell 109, receives a regeneration signal 118, the switch 113 will be briefly connected and leads the discharge current from the charge storage device 108 to the ECO cell through the circuit 117 for quick regeneration. The connection period should be short enough so that the charge storage device 108 is not completed discharged or receiving the charging current flow with opposite polarity from ECO cell 109. After the regeneration, the charge storage device 108 can be charged again by the ECO cell 112 without changing the charge polarity. Consequently, the same regeneration process can be repeated through other ECO cells with the electric connection similar to that of cell 109. Such approach, however, reduces the CO removal capability of the charging cell 112 because it will not receive the reversed potential for completed regeneration.

[0082] An alternative approach is shown in FIG. 8. For the purpose of illustration, first and second ECO cells 120 and 121 are connected with a first charge storage device 119, such as a capacitor, through the circuit 124 and first and second double pole double throw (DPDT) switches 122 and 123. Each DPDT switch has three positions, a first or left position 122-1, 123-1; a second or center-off position 122-2, 123-2; and a third or right position 122-3, 123-3. In this design, each ECO cell can serve as a charging device for itself or for other ECO cells depending on the position of the switch. It can also be regenerated by the discharging current from the charge storage device 119 without reversing the charge storage device's polarity. For example, the DPDT switch 122 can first be in left position 122-1 so that the charge storage device 119 can be charged by ECO cell 120. The DPDT switch can then be moved to center-off position 122-2 (neutral) upon the charge storage device 119 being charged to a desirable level and ready to be used for regeneration. The regeneration can then be applied via switching DPDT 122 to the right position 122-3 briefly before returning to center-off position 122-2. Similarly, this charging and regeneration cycle can be applied to other ECO cells in the circuit with the same electric connection.

EXAMPLES

Example 1

[0083] An ECO membrane electrode assembly (MEA) was prepared as follows. For the anode, an electrode catalyst ink was prepared by mixing 0.38 grams of Ru supported on XC-72R [E-TEK] with 5.40 grams of 1100 EW ionomer solution [Solution Technologies] and 0.38 grams tetrabutyl ammonium hydroxide [Aldrich]. The electrode catalyst ink was mixed in excess of 8 hours. A teflon decal support was coated until a 0.6 mg/cm2 loading of electrode catalyst was placed on the support. For the cathode, an electrode catalyst ink was prepared by mixing 0.19 grams of Pt supported on XC-72R [E-TEK] with 1.80 grams of 1100 EW ionomer solution [Solution Technologies] and 0.09 grams tetrabutyl ammonium hydroxide [Aldrich]. The electrode catalyst ink was mixed in excess of 8 hours. A teflon decal support was coated until a 0.3 mg/cm2 loading of electrode catalyst was placed on the support. The decals were then transferred to the surface of a Nafion 112 membrane [DuPont] under pressure of 1290 psi with heated platens (411° F.) to form the MEA. The MEA was then washed in sulfuric acid, sandwiched by two layers of gas diffusion back material ELAT™ (E-TEK) and loaded into a fuel cell bipolar plate fixture to form an ECO cell.

Example 2

[0084] An ECO membrane electrode assembly (MEA) was prepared as follows. Anode and cathode electrode catalyst inks were prepared as described in Example 1. A 0.6 mg/cm2 anode decal was prepared using the anode electrode catalyst ink and a 0.3 mg/cm2 cathode decal was prepared using the cathode electrode catalyst ink. A second electrode catalyst ink was prepared by mixing 0.19 grams of Ru supported on XC-72R [E-TEK] with 1.35 grams of 1100 EW ionomer solution [Solution Technologies] and 0.19 grams tetrabutyl ammonium hydroxide [Aldrich]. The electrode catalyst ink was mixed in excess of 8 hours. The decals were then transferred to the surface of a Nafion 112 membrane [DuPont] under 1290 psi with heated platens (411° F.). After pressing, an additional layer was applied to the surface of the anode with the second electrode catalyst ink, which created a low density around a loading of 0.6 mg/cm2. The total loading of the ruthenium metal on the anode side would be 1.2 mg/ cm2. The MEA is subsequently loaded into an ECO cell according to Example 1.

Example 3.

[0085] The MEA's of two 5 cm2 ECO cells, Cell A and Cell B, were prepared according to procedure described in Example 1. Cell A had a ruthenium loading of 0.6 mg/cm2 and Cell B had a ruthenium loading of 0.67 mg/ cm2. Fully humidified CO/H2 mixture was passed through the anode side of the ECO cells while fully humidified air was passed through the cathode side. For reference, the regenerative CO removal capacity for each ECO cell was measured through either a galvanic regeneration or an electrolytic regeneration method. In the galvanic regeneration mode, the electric leads from anode and cathode were briefly connected for about one second after the anode was fully saturated with carbon monoxide. In the electrolytic regeneration mode, the positive lead of a D. C. power supply (Hewlett Packard 6552A) with an output voltage of 0.4 volt was connected to the anode of the ECO cell and the negative lead was connected to the cathode. Again, the connection was in a very brief period of around one second after CO saturated the ECO cell capacity. The regenerative CO adsorption per unit ECO surface area was calculated based on the amount of carbon monoxide removed. For the galvanic method, the value ranged from 0.3 to 0.35 μmol/cm2 per regeneration for Cell A and B. For the electrolytic method, the value ranged from 0.7 to 0.8 μmol/cm2 per regeneration for Cell A and B. Subsequently, the ECO Cell A and Cell B were connected together according to the electric diagram shown in FIG. 3. Fully humidified CO/H2 mixture was passed through the anode sides of the Cell A and B simultaneously until both ECO cells were fully saturated by carbon monoxide. The switch in FIG. 3 was then quickly closed for about one second and re-opened again. The amount of CO removed from both Cell A and Cell B was measured. It was found that the CO removal capacity during the regeneration for Cell A was essentially the same as that obtained through electrolytic regeneration method while the CO removal capacity for Cell B was the same as that obtained through galvanic regeneration.

Example 4

[0086] Two ECO cells, one with MEA surface area of 5 cm2 and the other one with MEA surface area of 25 cm2 were prepared according to Example 1. The regenerative CO removal capacity of the 5 cm2 ECO cell was measured under the electrolytic mode as a reference point. Two ECO cells were then connected according to the electric diagram shown in FIG. 4a with 25 cm2 ECO cell representing multiple, parallel connected 5 cm2 ECO cells. Fully humidified CO/H2 mixture was passed through the anode sides while fully humidified air was passed through the cathode sides of both ECO cells. The process was continued until anodes were fully saturated by carbon monoxide. The DPDT switch for 5 cm2 ECO cell was then switched to the opposite direction for about one second, similar to what was shown in FIG. 4b, and was switched back to the original position. The regenerative CO removal capacity for the 5 cm2 ECO cell was then measured and compared with that obtained through the electrolytic regeneration via an external DC power supply. It was found that regenerative CO removal capacity obtained in this experiment was about twice as much as that obtained by electrolytic regeneration with the external DC power supply.

Example 5

[0087] A 5 cm2 ECO cell was prepared according to Example 2 with a Ru loading of 1.2 mg/cm2. The regenerative CO removal capacity of this cell was previously measured under both galvanic mode and electrolytic mode by a DC power supply, according to the method described in Example 3. The ECO cell was then connected with an ultracapacitor (PF-10 from Maxwell PowerCache®) according to the electric diagram shown in FIG. 5. A fully humidified synthetic reformate with 8000 ppm of CO and 37% of hydrogen was directed through the anode of ECO cell and fully humidified air was passed through the cathode side. The electric current produced by the ECO cell gradually charged the ultracapacitor until the potential of the capacitor was equal to that of the ECO cell. The DPDT switch, as is shown in FIG. 5, was then switched to the opposite end for a brief period of around one second to discharge the ultracapacitor. The discharge current passed through ECO cell oxidized CO in the cell through an electrolytic oxidation process. The regenerative CO removal capacity measured by this approach was compared with that obtained by galvanic method and the electrolytic method with a 0.4 volt DC power supply. The results are shown in the following table.

1ElectrolyticMethod withElectrolyticGalvanic0.4 V externalMethod with anRegeneration MethodMethodPower SupplyUltracapacitorUnit Area Regenerative0.31.41.6CO Removal Capacity(μmol/cm2)

[0088] As can be appreciated by those skilled in the art, the present invention provides an approach to improve fuel cell operation efficiency by removing carbon monoxide from the hydrogen fuel externally. The present invention provides advantages of a high degree of carbon monoxide removal, simple system configuration, low parasitic hydrogen consumption, increased tolerance to the dynamics of carbon monoxide output from the reformer, and ease of operation with robust electronic control. Although a primary application of the invention is to reduce the concentration of carbon monoxide with the hydrogen fuel for fuel cell operation, the present invention can have other applications where carbon monoxide removal is necessary.

[0089] It should be understood, of course, that the foregoing relates to preferred embodiments of the invention and that modifications may be made without departing from the spirit and scope of the invention as set forth in the following claims.

	Number	Date	Country
Parent	09393103	Sep 1999	US
Child	09849473	May 2001	US

Regeneration methods to remove carbon monoxide from reformate fuel using an adsorption/electro-catalytic oxidation (ECO) approach

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

US Classifications

International Classifications

Abstract

Description

Claims

CROSS-REFERENCE TO RELATED APPLICATIONS

Provisional Applications (1)

Continuation in Parts (1)