Patent Grant
8557485
This invention relates to the selection of anode materials for mitigating hydrogen starvation effects in a hydrogen-consuming fuel cell using a proton exchange membrane (PEM). More specifically, the present invention relates to the use of tungsten compound-containing hydrogen spillover materials mixed with an anode catalyst material to delay the onset of performance degradation of the cathode during temporary hydrogen starvation.
Proton exchange membrane (PEM) fuel cells are largely considered the main fuel cell technology candidate for, among others, portable and automotive transportation applications. A typical PEM fuel cell generally includes a membrane electrode assembly (MEA), a pair of gas diffusive backing layers positioned adjacent to each side of the MEA, and a pair of current-collecting flow plates positioned adjacent to each backing layer. The MEA, which is further composed of a solid polymer electrolyte membrane having an anode on one of its faces and a cathode on the opposite face, is the PEM fuel cell's main source of electrochemical activity where a usable electrical current that flows from the anode to the cathode is generated and intermittently utilized to power an external device. Such electrochemical activity is the result of catalyst-driven oxidation and reduction reactions that occur when a fuel is supplied to the anode side of the MEA and an oxidant is supplied to the cathode side. In many instances a large number of PEM fuel cells may be assembled in series to form what is known as a fuel cell stack to generate greater electrical current outputs. For example, stack arrangements of PEM fuel cells can generate a cumulative electrical current output capable of powering automotive devices such as electrical motors and certain auxiliary fuel cell stack equipment. Typical arrangements of multiple cells in a stack are described in U.S. Pat. No. 5,763,113, assigned to General Motors Corporation.
The solid polymer electrolyte membrane of the MEA may be formed from a proton conductive and electrically insulative polymer material. The membrane can thus provide a conductive pathway for ion transport between the anode and the cathode, can direct the flow of an electrical current through an external device in order for that current to travel from the anode to the cathode and thus complete the electrochemical cell, and separate the reactant gases supplied to the anode and the cathode. A popular polymer material used to make the solid polymer electrolyte membrane of a PEM fuel cell is a perfluorosulfonic acid ionomer such as, for example, those manufactured by DuPont and sold under the trade name Nafion®. The other two main components of the MEA—namely, the anode and cathode—are each typically formed from an ionically and electrically conductive binder material mixed with finely divided catalyst particles that are often supported on larger carbon particles. These catalyst particles, which are oftentimes precious metal particles such as those of platinum and ruthenium, catalyze the oxidation and reduction half-reactions that occur at the anode and cathode, respectively.
The operation of many kinds of PEM fuel cells involves supplying hydrogen gas to the fuel cell anode and oxygen gas in the form of either air or pure oxygen to the fuel cell cathode. The anode functions to dissociate the incoming hydrogen gas molecules into protons and electrons. The presence of an electrochemical gradient within the MEA for each of these newly-formed charged particles causes them to move towards the cathode, albeit by different pathways. The protons, as alluded to before, migrate from the anode to the cathode through the solid polymer electrolyte membrane. The electrically insulative nature of the membrane, however, is not amenable to electron transport and thus forces the electrons to flow through an external circuit in order to reach the cathode. It is the purpose of the cathode to then facilitate the reaction of the arriving protons and electrons with supplied oxygen gas to form water. But to keep the MEA operating efficiently in such a manner over an extended period of time is somewhat dependent on the ability to maintain a certain level of hydration in the membrane for optimized proton conductivity while at the same time not flooding the anode and cathode. Indeed, many PEM fuel cell water management controls seek to achieve optimum hydration of the MEA by balancing the generation of water at the cathode, the removal of water from both the anode and cathode, and the relative humidity of the supplied reactant gas streams.
One particular water management issue that can affect the performance of the MEA is hydrogen starvation. Hydrogen starvation generally occurs when the flow of hydrogen gas to the anode is blocked by water accumulation at the anode or at the anode-side flow plate. This accumulation of excess water can be caused by the external humidification of the hydrogen reactant stream and/or by the back diffusion of water from the cathode to the anode. The occurrence of hydrogen starvation is problematic because it can initiate the formation of localized cathode cells near the anode. These localized cells form, in the absence of hydrogen, as a result of oxygen cross-over from the anode to the cathode and the lack of lateral proton transport due to the anisotropic electrical properties of the solid polymer electrolyte membrane. The most available source of protons for these oxygen-rich localized cathode cells now happens to be through the oxidation of carbon materials and/or oxidation of water on catalyst (oxygen evolution reaction) in the cathode. The phenomenon of hydrogen starvation thus promotes performance degradation of the cathode as a result of carbon corrosion.
Possible solutions for mitigating the problem of hydrogen starvation include (i) frequent flushing of the anode compartment to remove any accumulated nitrogen and/or water, (ii) use of an oxygen evolution catalyst that decreases the partial current for carbon oxidation in the cathode, and (iii) use of non-carbon catalyst supports in the cathode. But unfortunately all of these options have significant operational or practical drawbacks. It is therefore desirable to devise an alternative mechanism for addressing the hydrogen starvation problem.
Each membrane electrode assembly of a PEM fuel cell may comprise a proton exchange membrane with an anode and cathode in intimate contact with opposite sides of the membrane. As mentioned before, the anode may comprise carbon support particles loaded with very small catalyst particles such as those of platinum, palladium, ruthenium, and alloys thereof for facilitating the dissociation of hydrogen gas. The anode may also comprise tungsten-containing hydrogen spillover materials to help mitigate the effects of hydrogen starvation by helping form a hydrogen reservoir near the fuel cell anode. Such a hydrogen reservoir can, during temporary periods of hydrogen starvation, compensate for any localized proton deficits and thus delay the onset of oxygen-driven cathode cells near the anode.
The tungsten-containing hydrogen spillover materials that are provided in the anode to combat the effects of hydrogen starvation may include particles of tungsten oxides and/or tungsten silicides. These tungsten-containing materials have shown the ability to interact with atomic hydrogen and protons in the presence of catalyst-loaded carbon particles to capture hydrogen/protons during periods where hydrogen gas flow to the anode is sufficient. But when hydrogen flow to localized areas of the anode is disrupted, the chemical environment of the starved localized area provokes the release of any stored hydrogen for a period of several seconds to be oxidized and thus sustain normal anode operation. The consumption of such hydrogen reserves postpones the formation of oxygen-driven cathode cells near the anode as well as the detrimental corrosion of cathode carbon materials to feed protons to those cathode cells for at least the time period in which the previously-stored hydrogen is being depleted. This time lag between the commencement of hydrogen starvation and the need for a new source of protons at the anode can provide a much needed timeframe in which the flow of hydrogen can be restored to appropriate levels by either normal fuel cell operating dynamics or affirmative remedial acts.
These tungsten-containing particles, when physically mixed with platinized carbon particles, have been observed in cyclic voltommetry behavior as providing an increase in the hydrogen adsorption/desorption (HAD) area when compared to standard anode constructions that do not include similar tungsten-containing particles. It is believed that such hydrogen storage-promoting functionalities are due to their so-called hydrogen spillover mechanism. This spillover mechanism can best be described as the net effect of platinum-hydrogen, platinum-oxide, and hydrogen-oxide interactions within the tungsten compound's crystal lattice. What this means practically is that hydrogen evolved at low potentials on platinum sites is “spilled over” to nearby tungsten compounds and then entrapped within those compounds' crystal lattice. In preferred embodiments, the tungsten-containing spillover materials may be tungsten trioxide (WO3) particles and/or tungsten disilicide (WSi2) particles. And a preferred mixing ratio of platinized carbon particles to WO3 and WSi2 particles that can achieve enhanced hydrogen storage capabilities in an anode while not otherwise impairing the anode's other functions has been found to be between 3:1 and 1:3, respectively, on a mass ratio basis when the those compounds are utilized individually.
Referring now to
To address the problem of hydrogen starvation, the use of tungsten-containing hydrogen spillover materials may be provided in an anode of a PEM fuel cell's MEA. These tungsten-containing spillover materials can facilitate the formation of hydrogen reservoirs near the anode that are capable of storing useful amounts of hydrogen gas and/or protons during periods when hydrogen supply to the anode is operationally sufficient. The hydrogen/proton reserves stored in such hydrogen reservoirs may then be released in response to a changing localized electrochemical environment that accompanies periods of insufficient hydrogen supply to the anode. This infusion of hydrogen/protons can help compensate for temporary proton deficiencies in the MEA and thus ultimately delay the corrosion of carbon support materials in the cathode. For instance, and referring momentarily back to
Suitable tungsten-containing hydrogen spillover materials that may be provided in the anode of an MEA are particles of tungsten oxides and/or particles tungsten silicides. Tungsten is an element that displays a range of valences and may form several different oxides and silicides. Tungsten trioxide, WO3, and tungsten disilicide, WSi2, are examples of such suitable tungsten oxygen-containing and silicon-containing compounds. These kinds of particles, when physically mixed with catalyst loaded carbon particles such as platinized carbon to form a composite anode, have shown an ability to promote the formation of hydrogen reservoirs through their spillover mechanism when compared to other anode constructions that do not contain such spillover materials. This has been observed in the cyclic voltammetry behavior as an increase in the hydrogen absorption/desorption area (HAD) of the composite anode material structure. And such increases in the anode's HAD are directly proportional to the transition time it takes for a hydrogen starved anode under a constant current density to reach potentials positive to 0.5 V/rhe—an indication that hydrogen is not being oxidized. The observed hydrogen spillover effects of tungsten oxides and silicides when mixed with platinized carbon is likely attributable to the net effect of platinum-hydrogen, platinum-oxide, and hydrogen-oxide interactions as well as strong platinum-support interactions. It is believed that the spillover of hydrogen starts with the surface diffusion from platinum sites followed by the chemical interaction of hydrogen within the tungsten-oxide crystal lattice. Proton intercalation into the tungsten-oxide lattice can also occur as a result of tungsten redox state variations.
In exemplary embodiments, the tungsten-containing spillover materials that may be used to form composite anodes are particles of tungsten trioxide (WO3) and/or tungsten disilicide (WSi2). Both of these compounds have shown the ability to improve the composite anode's HAD. They have also demonstrated an ability to significantly delay a voltage excursion of the composite anode to potentials of 0.5 V/rhe and greater (sometimes referred to as the oxygen evolution region) upon cessation of adequate supplied hydrogen flow. These tungsten-containing spillover materials were tested in combination with a representative anode catalyst material, platinum particles supported on electronic conductive carbon particles.
To demonstrate these beneficial effects, two experimental procedures were performed on composite anodes structures that contained particles of either WO3 or WSi2. The performance results and general observations gleaned from these procedures were ultimately compared against a benchmark anode structure for context. The first procedure conducted involved talking galvanostatic measurements of gas-diffusion electrodes while switching the gas flow from hydrogen to argon to simulate a hydrogen starvation effect. The second procedure conducted involved subjecting a flooded gas-diffusion electrode to potential cycling under simulated fuel cell conditions to investigate the electrodes hydrogen storage capacity, durability, and optimal electrode composition.
Galvanostatic Measurements
Galvanostatic experiments were performed with Teflon-bonded gas-diffusion electrodes having a 2.8 cm2 geometric area. The electrodes were placed in 0.1 M perchloric acid and subjected to a constant anodic current density of 20 mA/cm2 at room temperature (25° C.) and atmospheric pressure. Galvanostatic measurements of potential versus time for the electrodes were obtained with a Radiometer Analytical VoltaLab potentiostat. To simulate hydrogen starvation, the flow fields of each of the electrodes was initially supplied with a source of hydrogen gas that was later abruptly switched over to a source of argon gas.
Three electrode compositions were used in this galvanostatic analysis. The benchmark electrode was composed of 47.7 weight % platinum on Vulcan carbon commercial catalyst mixed with Vulcan XC-72 carbon in a 1:1 mass ratio. The WO3 electrode was composed of 47.7 wt. % platinum on Vulcan carbon commercial catalyst mixed with tungsten trioxide (Aldrich Cat #550086, BET area 17 m2/g) in a mass ratio of 1:1. And lastly, the WSi2 electrode was composed of 47.7 wt. % platinum on Vulcan carbon commercial catalyst mixed with tungsten disilicide (Noah Technologies #14963, 99.5%, −325 mesh, BET area of 0.5 m2/g) in a 1:1 mass ratio. Additionally, each of these three electrodes contained 30 wt. % polytetrafluoroethylene (PTFE), which resulted in a 0.37 (±0.01) mg/cm2 platinum loading for the electrodes. The three electrodes were also sintered for 45 minutes at 260° C. in an argon atmosphere.
In order to interpret the transition time data in Table 1 and to investigate correlations with the hydrogen desorption charge, voltammograms (current-voltage curves) at 20 mV/s scan rates in argon saturated 0.5 M perchloric acid solutions were performed for the various electrodes and plotted in
The data contained in
Potential Cycling
Potential cycling was carried out in three-electrode cell systems at a scan rate of 10 mV/s over the potential range of 0<E<1.2 V/rhe in a 0.5 M H2SO4 aqueous solution at 80° C. with an argon gas purge. The three electrodes for each cell system were a working electrode, a counter electrode, and a reference electrode. The working electrodes of each cell system were formed as circular coupons (about 9 mm in diameter) and pressed onto a gold current collector under one metric ton pressure. The coupons were cut from sheets prepared by a roll and paste technique—to deposit working electrode material—with a 10% Teflon binder. The gold current collectors were made of a 1 mm diameter gold wire that was welded to the rim of a gold disc that measured 12 mm in diameter. The counter electrodes of each cell system were composed of ultra-high purity (99.999%) carbon rods coated with a mixture of graphitized carbon and Nafion®. The reference electrodes of each cell system were composed of reversible hydrogen electrodes. Those hydrogen electrodes comprised a captive hydrogen gas bubble at room temperature sealed inside a glass tube and in contact with a platinum wire.
Seven different three-electrode cell systems were subjected to the potential cycling procedure just mentioned. The working electrodes of each cell system, however, were compositionally different. In the benchmark cell system, the working electrode was composed of 46.7 wt. % platinum on Vulcan XC-72 carbon commercial catalyst. The other six cell systems utilized composite working electrodes—three with WO3 and three with WSi2. The three WO3 cell systems included working electrodes comprised of 46.7 wt. % platinum on Vulcan XC-72 carbon commercial catalyst and mixed with WO3 nanopowder (Aldrich Cat #550086). The mass ratios of Pt/C to WO3 for the three working electrodes in the three WO3 cell systems were 1:3, 1:1, and 3:1. The three WSi2 cell systems included working electrodes comprised of 46.7 wt. % platinum on Vulcan XC-72 carbon commercial catalyst and mixed with WSi2 (Noah Technologies #14963). The mass ratios of Pt/C to WSi2 for the three working electrodes in the three WSi2 cell systems were also 1:3, 1:1, and 3:1.
For each of the seven cell-systems, two voltammograms each were recorded at 2 mV/s during potentiodynamic cycling at 10 mV/s and starting one cycle prior to, and including, cycles 10, 20, 50, 100, 200, 500, 1000, 2000, 3000, and 4000, in order to determine the hydrogen adsorption/desorption area (HAD) for their working electrodes. The HAD areas were calculated using an Arbin BT-2000 multi-channel instrument. X-ray diffraction measurements were also taken for each working electrode at 0 and 1000 cycles using Cu kα radiation on a Bruker General Area Detector Diffraction System (GADDS). Platinum crystallite sizes were calculated from the full width half maximum (FWHM) of the Pt (220) reflection using the Scherrer equation. The FWHM was measured using peak profile fitting software.
X-ray diffraction (XRD) data was collected for the working electrodes in their as-prepared state and after 1000 cycles for some of the seven potential cycling experiments—namely, the benchmark working electrode, the 1:3 mass ratio WO3 working electrode, and all three of the WSi2 working electrodes.
The above galvanostatic and potential cycling experiments reveal that composite anodes containing catalyst-loaded carbon particles admixed with tungsten oxide particles and/or tungsten silicide particles may be useful to help combat the possible adverse effects associated with hydrogen starvation in PEM fuel cells. The ability of these tungsten-containing compounds to temporarily delay an anodic potential excursion during disruptions of hydrogen gas flow can postpone the corrosion of the carbon support materials located in the cathode. This provides a much needed and ample time frame in which the hydrogen gas flow problem can be rectified before significant performance degradations occur in the PEM fuel cell's MEA.
The above description of embodiments of the invention is merely exemplary in nature and, thus, variations thereof are not to be regarded as a departure from the spirit and scope of the invention.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5763113 | Meltser et al. | Jun 1998 | A |
| 6828056 | Molter et al. | Dec 2004 | B2 |
| 20060188768 | Kanaoka et al. | Aug 2006 | A1 |
| 20060240312 | Xie et al. | Oct 2006 | A1 |
| 20060251954 | Merzougui et al. | Nov 2006 | A1 |
| 20070202386 | Matsuoka et al. | Aug 2007 | A1 |
| 20080118818 | Merzougui et al. | May 2008 | A1 |
| 20080166599 | Swathirajan et al. | Jul 2008 | A1 |
| Number | Date | Country | |
|---|---|---|---|
| 20100151295 A1 | Jun 2010 | US |
