This invention relates to fuel processing and more particularly to processing fuel for use with fuel cells. More particularly still, the invention relates to a fuel processing system for use with low operating temperature fuel cell power plants, such as a PEM fuel cell.
Fuel cell power plants are used increasingly to provide electrical power for a variety of end uses. The fuel cell power plant typically includes one or more fuel cell stack assemblies (CSA), each consisting of an anode, a cathode, and an electrolyte that separates the anode and cathode. Fuel reactant (for a PEM fuel cell), which is typically a hydrogen-rich stream, enters the anode of the CSA, and an oxidant reactant, typically air, enters the cathode. A catalyst in the anode causes the hydrogen to oxidize, resulting in the creation of hydrogen ions, which pass through the electrolyte to the cathode to create the electrical current of the power plant. In addition to the CSA, the power plant typically also includes a fuel processing system (FPS) for converting a hydrocarbon feedstock, such as natural gas, LPG, gasoline, and/or numerous others, to the hydrogen-rich fuel stream, which fuel stream may be referred to as “reformate” because of an included process of reformation.
Although the types of fuel cells vary according to their electrolytes, they may also be viewed as varying according to their operating temperatures. One type of fuel cell that is receiving considerable attention for application to automotive and other uses is the PEM fuel cell that employs a solid polymer electrolyte referred to as a proton exchange membrane. These fuel cells typically operate at temperatures of 60° C. to 83° C., though temperatures to less than 38° C. and as great as 120° C. are possible, and within a pressure range of about one to five atmospheres. Moreover, the catalyst associated with the anode of a PEM fuel cell is particularly susceptible to being “poisoned” by carbon monoxide (CO) that may be contained in the reformate.
Accordingly, the fuel processing system (FPS) of the power plant is not only required to reform a hydrocarbon feedstock to a hydrogen-rich stream to fuel the anode, but it is also required to convert significant levels of CO in the reformate to carbon dioxide (CO2) to thereby reduce the concentration of CO to a level acceptable at the anode. The FPS may similarly be required to remove or convert objectionable sulfur species in the hydrocarbon feedstock in order to avoid damage to the CSA.
In a typical FPS 10 of a fuel cell power plant 15 of the prior art, a portion of which is depicted in
CO+H2O<=>CO2+H2ΔH1=−41 kJ/mole H2 (1)
This reaction is exothermic (in the forward direction) and equilibrium-limited, with lower temperatures favoring higher CO conversions. However, the reaction rate of the HT WGS catalyst increases exponentially with temperature. Thus, the existing practice that optimizes thermodynamics and kinetics of prior existing HT WGS catalysts is to use a second, or low-temperature, water gas shift reactor (LT WGS) 20. The LT WGS 20 typically includes a water vaporizer or cooler (heat exchanger) 22 preceding a catalytic reactor 24. The vaporizer 22 serves as a cooling device and also provides additional steam for the reactor 24. In some architectures the vaporizer may be replaced or assisted by a cooler which will serve to cool the reformate and at the same time use the heat to pre-heat the feed. The catalyst in the LT WGS reactor 24 has typically been Cu/ZnO or the like, or more recently may have been noble metal-based.
Referring further to
CO+0.5 O2→CO2 ΔH2=−283 kJ/mole CO (2)
H2+0.5 O2→H2O ΔH3=−242 kJ/mole H2 (3)
Uniform mixing of the reformate with inlet air assures homogeneous mixing and therefore, effective operation by avoidance of hot spots. To reduce the amount of H2 consumed, this preferential oxidation process has been performed in two, usually adiabatic, stages, as depicted.
The reformate from LT WGS 20 is fed to a high temperature CO preferential oxidizer subsystem (H PROX) 30 (sometimes also referred to as PROX 1), which includes an air mixer 32 followed by a cooler 34 in turn followed by a catalytic reactor 36. An O2/CO ratio slightly above the stoichiometry of reaction (2) is used (in the 0.5-2 regime) and the CO is reduced from the range of 10,000-5,000 ppmv to the range of about 2,000 to 500 ppmv. The reformate from H PROX reactor 36 is then fed to a low temperature CO preferential oxidizer subsystem (L PROX) 40 (sometimes also referred to as PROX 2), which includes an air mixer 42 followed by a cooler 44 in turn followed by a catalytic reactor 46. Here, a significantly higher O2/CO ratio (1.0-4.0, or more) than the stoichiometric ratio for the reaction (2) is used to ensure elimination of CO to concentrations less than 50 ppmv, and typically less than 10 ppmv. Finally, reformate from the L PROX 40 is flowed through an anode precooler 50 and thence to the anode 54 of fuel cell stack assembly (CSA) 56.
Further consideration is given here to various limitations or complexities that arise with the use of the 2-stage PROX section, and particularly the H PROX section, or subsystem, 30. The combined CO selectivity of the two-stage PROX process is about 35%, i.e., for each CO molecule consumed, two H2 molecules are consumed. Thus, there is an efficiency penalty during this process. If it is assumed, for example, that 8,000 ppmv of CO is oxidized in the two-stage PROX system, the overall H2 consumed is roughly 2×8,000 ppmv=16,000 ppmv, or 1.6%, of which up to 75%, i.e. up to 12,000 ppmv, is consumed in the H PROX reactor 36. Moreover, since the CO reaction is highly exothermic and the levels of CO that are consumed are relatively high, the temperature in this adiabatic reactor typically increases more than 100° C. Due to the high reaction exothermicity, it may be prone to overheating, with a concomitant reverse shift reaction that converts CO2 back to CO, or to overcooling, which fails to adequately oxidize the CO. To avoid overheating, it is necessary to carefully regulate the temperature of reformate entering the H-PROX reactor 36, this typically being done via the heat exchanger/cooler 34 in the H PROX section 30. However, as noted, it is equally important to avoid overcooling since the catalyst of the H PROX reactor 36 has a high sensitivity to the inlet temperature of the reformate and to the CO concentration. While the latter is governed by the HT WGS 12 and the LT WGS 20, the former is governed principally by careful regulation of the heat exchanger/cooler 34 in the H PROX section 30. Indeed, if overcooling occurs and the reformate gas stream temperature at the inlet to the H PROX reactor 36 is lower than the catalyst “light-off” temperature, the catalyst will remain inactive, thereby passing unacceptably high levels of CO to the L-PROX section 40 and also creating a “cold” reformate gas stream. Since the L-PROX reactor is designed to operate for an inlet CO level of less than about 2,000 ppmv (wet basis), then the high CO levels may be passed to the anode 54 of the CSA 56 and poison the catalyst there.
A further illustration of the sensitivity of the WGS sections 12, 20 and PROX sections 30, 40 to inlet flow rates, with respect to CO levels and to gas temperatures, is depicted in
A similar behavior is depicted in
Further, this high exothermicity of the H PROX reactor 36 can lead to unstable operation states referred to as “bifurcations”, in which the system may be seen to operate in multiple “steady states”. This is due to the coupling effects in the changes of the O2/CO ratio and the coolant flow rates during operation of the system in a transient mode, and leads to either serious overheating, i. e., temperature runaway, or to overcooling, i. e., process extinguishing. As in the examples described above, these modes may be mitigated only with the use of sophisticated, and therefore expensive, coolant and air control systems.
In view of the foregoing discussion of the operating dynamics of the various reaction sections presently used, the further requirement for sophisticated and costly coolant flow control and air control will be understood and appreciated. Such need, or burden, is particularly manifested in the operation of the 2-stage PROX section in the FPS 15, which imposes certain burdens on the power plant 10, to wit, the cost, weight and volume of the PROX hardware itself as well as the extra cost of the sophisticated coolant flow and air controls required to avoid the limitations of a simpler proportional control system.
Accordingly, it is an object of the invention to provide an improved fuel processing system for a fuel cell power plant.
It is a further object of the invention to provide a fuel processing system that requires relatively less equipment.
It is a still further object of the invention to provide a fuel processing system that does not require relatively sophisticated/costly associated controls for thermal and/or air management.
It is an even further object of the invention to provide an efficient, smaller size and weight (compact) and cost effective fuel processing system for providing reformate with an acceptably low CO concentration, to a fuel cell stack assembly.
In accordance with the invention, it has been determined that use of a high performance, low-temperature water gas shift (LT WGS) catalyst, possessing relatively lower exothermicity and being active at temperatures below 250° C., can have a significant positive impact on the fuel processing system (FPS), principally through the elimination of the high temperature, first stage, H PROX reactor 36, and also the air mixer 32 and cooler 34 associated with that reactor, and secondarily by allowing the utilization of a simpler coolant control system. Additionally, or alternatively, the use of such catalyst in the LT WGS may allow the use of a relatively smaller catalyst bed and reactor for a given requirement of CO reduction in the reformate flow. The use of such a catalyst in the LT WGS reactor enables that reactor to process the reformate to exit temperatures as low as about 250° C. and below, as governed by the thermodynamic equilibrium equation, which concomitantly permits the CO concentrations to be reduced to as low as about 2,000 ppmv, or less. With the reformate at that temperature and particularly at that CO concentration, it is then possible to eliminate the H PROX and still attain the requisite CO level of less than about 50 ppmv at the output of the L PROX.
Through use of such high-performance, low-temperature, WGS catalyst in the low-temperature water gas shift reactor, the hardware and coolant and air processing in the FPS may be simplified. Specifically, it becomes possible to eliminate the high-temperature PROX subsystem previously required. Moreover, this elimination of the H PROX subsystem allows simplification of the associated coolant and air systems, both in terms of hardware requirements and control complexity. Finally, an increase in the overall fuel cell power plant efficiency of at least 1% is achieved, since the catalyst in the H PROX would also otherwise burn hydrogen, with the amount of the efficiency increase depending upon the CO/H2 ratio and the catalyst selectivity at H PROX operating conditions.
The foregoing improvements and advantages in the design and operation of the fuel processing system include reduced system cost, volumetric reduction or minimization, and enhanced performance due to higher efficiency and improved controllability during transients. These improvements and advantages arise from the recognition that by increasing the activity of a catalyst that consists of noble metals over doped ceria to at least 0.18 moles of CO per moles of Pt per sec. (moles CO/moles equiv. Pt-sec.) at 250° C. for a gas composition of 1.5%-2.5% CO, 25%-35% H2O, 30-50% H2, 10-14% CO2, balance N2, and simultaneously decreasing the CO reaction order in the LT WGS reaction from that previously assumed with the use of commercial Cu/ZnO catalysts (˜1), the activity of a catalyst would become substantially independent of the CO partial pressure and the shift reaction could effect reduction of CO to lower concentrations and at lower temperatures. This allows reduction of CO concentration along the LT WGS reactor to levels predicted by the thermodynamic equilibrium equation with relatively smaller reactor volume than required by another catalyst of similar activity but a high CO reaction order, thus affording savings in reactor hardware costs and a reduction in system parasitic power.
In regard to the foregoing and in accordance with the invention, the catalyst employed in the LT WGS is high performance, and is capable of activity at relatively low temperatures below about 250° C. Certain noble metal-on-ceria-based, nanophase-structured catalysts are seen to provide such properties, with a preferred example being a supported platinum (Pt) and rhenium (Re) catalyst, with at least 1 wt %, preferably 1.5 to 8.0 wt %, and most preferably 2.0 to 5.0 wt %, of Pt and Re having a dispersion greater than 90% on a porous mixed metal oxide support.
The foregoing features and advantages of the present invention will become more apparent in light of the following detailed description of exemplary embodiments thereof as illustrated in the accompanying drawings.
The description of the prior art, with reference to
The relationship between CO, H2O, CO2, and H2 (and N2) concentrations in the reformate stream are governed by thermodynamic equilibrium considerations. The equilibrium concentration and conversion values for CO in the WGS reaction are plotted in
In regards to the foregoing, it was determined to assess the impact of increasing the catalyst activity per mole of Pt by at least 1.5 times and decreasing the CO reaction order to the range of 0.0-0.10 from that associated with the recently-developed, high activity, noble metal-based, catalyst associated with the prior art systems. For these high performance, low-temperature WGS catalysts, it was found that the reaction rate can be expressed as a function of concentrations (or partial pressures) and temperature with the following empirical equation, based on the Arrhenius equation:
where [A] is proportional to the number of active sites in the catalyst, Ea is the activation energy, beq=[H2][CO2]/{[CO][H2O]Keq(T)} and Keq(T)≈10(2073/T−2.029). In the above rate expression, the concentrations are in kmoles/m3. The temperature is in degrees Kelvin (K). The factor A has units such that the rate of reaction is finally expressed in gmoles of CO/(Kg catalyst-sec). The Activation Energy is expressed in Joules/mole, and R, the universal gas constant is 8.314 Joules/(mole-K). Kinetic experiments were run to estimate the reaction rate parameters, for a broad range of reformate compositions and temperatures. Depending on the catalyst, temperature (200° C.-260° C.) and reformate composition, the values of these parameters were found in the following regimes, using state-of-the-art mathematical software to analyze and fit the experimental data:
Referring to
In general, the CO2 in the reformate has a negative impact on the activity of these catalysts that is expressed with a negative reaction order dependence on the CO2 concentration. However, the activity of the high activity catalysts in accordance with the invention is practically independent of the CO2 concentration, with a CO2 reaction order between −0.02 and −0.08, while similar recently-developed, noble metal-based catalysts have relatively more negative CO2 reaction orders.
Cu/ZnO is a commercially available, low cost, LT WGS catalyst that operates at low temperatures in the 200-240° C. regime. However, in addition to its pyrophoricity when exposed to air which is a significant limitation in using this catalyst in a fuel cell power plant, especially for automotive applications that employ on board fuel reforming, the dependence of its reaction rate on CO concentration results in the required reactor being at least 10 times larger relative to that required for the high activity catalysts in accordance with the invention, as comparatively depicted in FIG. 7. This Figure specifically corresponds to a LT WGS reactor in a 150 KWe fuel cell power plant that uses natural gas as a fuel in the fuel processor. State-of-the-art mathematical models and software that integrate reaction engineering, thermodynamics, transport phenomena, energy, mass and momentum conservation differential equations with the appropriate boundary conditions were applied to estimate the reactor size. ⅛″ extrudates were assumed for calculating the Cu/ZnO reactor volume while a 200 cpsi honeycomb monolith, wash-coated with 250 grams of catalyst per liter of reactor was assumed for the Pt-based high activity LT WGS catalyst with the previously described characteristics.
The importance of having a high catalyst activity per weight of Pt is illustrated in
Again, estimation of the Pt amount in
In view of the foregoing discussion, the selection of catalyst for use in at least the LT WGS reactor is important. That catalyst should be capable of activity at relatively low temperatures below about 250° C. Catalysts possessing these characteristics are to be found amongst the noble metal-on-ceria-based nanophase structure catalysts. A particular example of a suitable high activity catalyst is a Pt—Re mixed metal cluster on a nano-crystalline, large pore, controlled density, cerium oxide-based catalyst, as in
Additional description of the catalyst discussed in the preceding paragraph, and particularly the Pt—Re mixed metal cluster embodiment, as well as the manufacture thereof, may be found in U.S. application Ser. No. 10/402,808, entitled Ceria-Based Mixed-Metal Oxide Structure, Including Method of Making and Use (C-2727A) and filed Mar. 28, 2003, and assigned to the same assignee as the present application, which application is incorporated herein by reference to the extent required herein.
Reference is now made to
The omission of the H PROX section represents a significant savings in terms of hardware and associated weight, volume, cost, and maintenance. Moreover, the elimination of the H PROX section obviates some of the need for a sophisticated coolant control system, and the coolant and air supply scheme for the sections of FPS 110 may now be generally proportional, and thus less complex and less expensive. Finally, the efficiency of the fuel cell power plant system 115 is increased by at least 1% because an additional 0.5-1 mole % H2 is now produced in the LT WGS subsystem 120 and an additional 1-2 mole % H2 is saved that would otherwise have been consumed in the omitted H PROX reactor (36 of FIG. 1).
Although the invention has been described and illustrated with respect to the exemplary embodiments thereof, it should be understood by those skilled in the art that the foregoing and various other changes, omissions and additions may be made without departing from the spirit and scope of the invention.
Number | Name | Date | Kind |
---|---|---|---|
3565830 | Keith et al. | Feb 1971 | A |
3825501 | Muenger | Jul 1974 | A |
3870455 | Hindin | Mar 1975 | A |
4021366 | Robin et al. | May 1977 | A |
4157316 | Thompson et al. | Jun 1979 | A |
4170573 | Ernest et al. | Oct 1979 | A |
4171288 | Keith et al. | Oct 1979 | A |
4297246 | Cairns et al. | Oct 1981 | A |
4308176 | Kristiansen | Dec 1981 | A |
4331565 | Schaefer et al. | May 1982 | A |
4476246 | Kim et al. | Oct 1984 | A |
4585752 | Ernest | Apr 1986 | A |
4587231 | Sawamura et al. | May 1986 | A |
4629612 | van der Wal et al. | Dec 1986 | A |
4835132 | Sambrook | May 1989 | A |
4868148 | Henk et al. | Sep 1989 | A |
5039503 | Sauvion et al. | Aug 1991 | A |
5057483 | Wan | Oct 1991 | A |
5073532 | Domesle et al. | Dec 1991 | A |
5139992 | Tauster et al. | Aug 1992 | A |
5254519 | Wan et al. | Oct 1993 | A |
5275997 | Ganguli et al. | Jan 1994 | A |
5464606 | Buswell et al. | Nov 1995 | A |
5480854 | Rajaram et al. | Jan 1996 | A |
5490977 | Wan et al. | Feb 1996 | A |
5491120 | Voss et al. | Feb 1996 | A |
5492878 | Fujii et al. | Feb 1996 | A |
5500198 | Liu et al. | Mar 1996 | A |
5516597 | Singh et al. | May 1996 | A |
5532198 | Chopin et al. | Jul 1996 | A |
5744118 | Imamura et al. | Apr 1998 | A |
5788950 | Imamura et al. | Aug 1998 | A |
5830425 | Schneider et al. | Nov 1998 | A |
5843195 | Aoyama | Dec 1998 | A |
5895772 | Grigorova et al. | Apr 1999 | A |
5945369 | Kimura et al. | Aug 1999 | A |
5948683 | Koermer et al. | Sep 1999 | A |
5990040 | Hu et al. | Nov 1999 | A |
6033634 | Koga | Mar 2000 | A |
6040265 | Nunan | Mar 2000 | A |
6107240 | Wu et al. | Aug 2000 | A |
6120923 | Van Dine et al. | Sep 2000 | A |
6120925 | Kawatsu et al. | Sep 2000 | A |
6133194 | Cuif et al. | Oct 2000 | A |
6159256 | Bonville, Jr. et al. | Dec 2000 | A |
6204219 | Brezny et al. | Mar 2001 | B1 |
6299994 | Towler et al. | Oct 2001 | B1 |
6322917 | Acker | Nov 2001 | B1 |
6326329 | Nunan | Dec 2001 | B1 |
6409939 | Abdo et al. | Jun 2002 | B1 |
6455182 | Silver | Sep 2002 | B1 |
6562088 | Ukai et al. | May 2003 | B2 |
6692545 | Gittleman et al. | Feb 2004 | B2 |
6733552 | Taguchi et al. | May 2004 | B1 |
6787118 | Roark et al. | Sep 2004 | B2 |
20010002248 | Ukai et al. | May 2001 | A1 |
20020071806 | Sabacky et al. | Jun 2002 | A1 |
20020073895 | Barnes et al. | Jun 2002 | A1 |
20020110519 | Ying et al. | Aug 2002 | A1 |
20020131915 | Shore et al. | Sep 2002 | A1 |
20030129100 | Ukai et al. | Jul 2003 | A1 |
20030235526 | Vanderspurt et al. | Dec 2003 | A1 |
20040048114 | Shore | Mar 2004 | A1 |
Number | Date | Country |
---|---|---|
0298351 | Jan 1989 | EP |
1 046 612 | Oct 2000 | EP |
1 256 545 | Jul 2001 | EP |
1 161 991 | Dec 2001 | EP |
1 256 545 | Nov 2002 | EP |
10202101 | Aug 1998 | JP |
1133294 | Dec 1999 | JP |
2000-178007 | Jun 2000 | JP |
WO 9623573 | Aug 1996 | WO |
WO 9744123 | Nov 1997 | WO |
WO 0103828 | Jan 2001 | WO |
Number | Date | Country | |
---|---|---|---|
20040187384 A1 | Sep 2004 | US |