Abstract
A process for forming a metal supported solid oxide fuel cell is provided. The process can include the steps of: a) applying a green anode layer including nickel oxide and a rare earth-doped ceria to a metal substrate; b) prefiring the anode layer under non-reducing conditions to form a composite; c) firing the composite in a reducing atmosphere to form a sintered cermet; d) providing an electrolyte; and e) providing a cathode; wherein the reducing atmosphere comprises an oxygen source, a metal supported solid oxide fuel cell formed during this process, fuel cell stacks and the use of these fuel cells.
Claims
1. A process for forming a metal supported solid oxide fuel cell, the process comprising the steps of: a) applying a green anode layer including nickel oxide and a rare earth-doped ceria to a metal substrate; b) thereafter, prefiring the green anode layer under non-reducing conditions to form a composite and a passivation layer interposed between the composite and the metal substrate; c) thereafter, firing the composite in a reducing atmosphere to form a sintered cermet, form nickel metal, maintain the rare earth-doped ceria in a partially-reduced state, and retain the passivation layer, wherein the reducing atmosphere comprises a reducing agent and an oxygen source, and wherein an oxygen partial pressure in the reducing atmosphere is in the range 10.sup.?14 to 10.sup.?22 bar; d) thereafter, providing an electrolyte; and e) thereafter, providing a cathode.
2. The process according to claim 1, wherein the reducing agent is a gaseous reducing agent, and wherein the oxygen source is a gaseous oxygen source, and further wherein the reducing atmosphere of firing step c) comprises an inert gas.
3. The process according to claim 2, wherein the reducing agent is selected from hydrogen, carbon monoxide and combinations thereof.
4. The process according to claim 2, wherein the gaseous oxygen source is selected from carbon dioxide, water vapour and combinations thereof.
5. The process according to claim 2, wherein the reducing atmosphere of firing step c) comprises in the range 0.01 to 50 volume % of the oxygen source and/or 0.5 to 50 volume % reducing agent.
6. The process according to claim 1, wherein in firing step c) the nickel oxide is reduced to nickel metal prior to sintering.
7. The process according to claim 1, wherein in firing step c) the nickel oxide is at least partially sintered prior to reduction to nickel metal.
8. The process according to claim 1, wherein at least one of the prefiring of the green anode layer and the firing of the composite occurs at a temperature in the range 950? C. to 1100? C.
9. The process according to claim 1, comprising bracing the metal substrate during at least one of a heating step selected from: prefiring the green anode layer, firing the composite, or combinations thereof.
10. The process according to claim 1, wherein the nickel oxide and rare earth-doped ceria are powdered, the powders being of particle size distribution d90 in the range 0.1 to 4 ?m.
11. The process according to claim 1, wherein the nickel oxide and rare earth-doped ceria are applied as an ink.
12. The process according to claim 11, wherein the application of the green anode layer includes an initial application of the ink to the metal substrate, and drying the ink to provide a printed layer of thickness in the range 5 to 40 ?m.
13. The process according to claim 1, further comprising a step of compressing the green anode layer at pressures in the range 100 to 300 MPa.
14. The process according to claim 1, further comprising the step of reoxidising the sintered nickel prior to the provision of the electrolyte.
15. A metal supported solid oxide fuel cell formed by the process according to claim 1.
16. A fuel cell stack comprising two or more fuel cells according to claim 15.
17. The process according to claim 1, wherein the rare earth-doped ceria has the formula Ce.sub.1?xRE.sub.xO.sub.2?x/2, wherein RE is a rare earth and 0.3?x?0.5.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) In order that the present invention may be more readily understood, it will be described further with reference to the figures and to the specific examples hereinafter.
(2) FIG. 1 is a schematic representation of a SOFC as described in GB 2 368 450;
(3) FIG. 2 is a scanning electron micrograph (SEM) showing a cross section through a SOFC of FIG. 1 (15.0 kV, 7.9 mm?1.50 k);
(4) FIG. 3 is a thermodynamic phase diagram for a nickel/nickel oxide system covering the temperature range 500 to 1100? C. and oxygen partial pressures in the range log pO.sub.2 0 to ?40;
(5) FIG. 4 is a thermodynamic phase diagram for a chromium/chromium oxide system covering the temperature range 500 to 1100? C. and oxygen partial pressures in the range of log pO.sub.2 0 to ?40;
(6) FIG. 5 is a thermodynamic phase diagram for a nickel/nickel oxide system at 1030? C. and 1 bar total pressure as a function of hydrogen and steam partial pressures;
(7) FIG. 6 is a thermodynamic phase diagram for a chromium/chromium oxide system at 1030? C. and 1 bar total pressure as a function of hydrogen and steam partial pressures;
(8) FIG. 7 is a SEM showing a cross-section through a metallic support and anode of an SOFC of the invention after pre-firing in air (15.0 kV, 7.0 mm?4.0 k);
(9) FIG. 8 is a SEM also showing a cross-section through the metallic support and anode of firing in the reducing hydrogen atmosphere and reoxidation as described below (20.0 kV, 4000?);
(10) FIG. 9 is a SEM showing the cross-section of FIG. 8 at higher magnification (20.0 kV, 13000?);
(11) FIG. 10 is a SEM showing a cross section through a SOFC made using the process of the invention;
(12) FIG. 11 is a current-voltage curve for the SOFC of FIG. 10 as a function of cell operating temperature (56% hydrogen-44% nitrogen fuel, excess air fed to cathode);
(13) FIG. 12 is a power-cycle graph of the SOFC of FIG. 10; and
(14) FIG. 13 is a table of the results of mechanical strength tests undertaken on SOFC cells both after initial manufacture and after cells have operated in an initial performance characterisation test, for both standard nickel-CGO anodes as illustrated in FIG. 2, and reduced fired nickel-CGO anodes as illustrated in FIG. 8
DETAILED DESCRIPTION
(15) A SOFC 10 as described in GB 2 368 450 is shown schematically in FIG. 1, and in SEM cross-section in FIG. 2. Both figures show a ferritic stainless steel substrate 1, made partially porous by laser-drilling thousands of holes though the central region of the substrate 2. The porous substrate includes a chromium oxide passivation layer 11, a nickel oxide and CGO anode layer 3 covering the porous region 2 of the substrate 1. Over the anode layer 3 is deposited a CGO electrolyte layer 4 (10 to 20 ?m, CGO), which overlaps the anode 3 onto the undrilled area 9 of the substrate 1, thus forming a seal around the edge of the anode 3. The cathode 5,6 has a thin active layer 5 (CGO composite) where the reduction of oxygen takes place, and a thicker current collector layer 6 (lanthanum strontium cobaltite) to allow current to be collected from the cell 10 in a stack. FIG. 2 additionally shows a very thin stabilised zirconia layer 7 and an even thinner doped ceria layer 8, which block electronic conductivity (preventing short circuiting from undesirable chemical reactions between the cathode 5,6 and zirconia layer 7) and form the interface between the anode and electrolyte respectively.
(16) SOFC 10 of FIGS. 1 and 2 was prepared by applying a screen-printing ink containing suspended particles of nickel oxide powder and CGO powder (d90=0.7 to 1.2 ?m, ratio of nickel oxide to CGO in the ink being 1:0.55 by weight). The ink was screen printed onto ferritic stainless steel substrate 1 using conventional methods, and dried in an oven to evaporate the solvents and set the binders thereby forming a dried, printed layer of thickness 9 to 15 ?m. The dried, printed layer was compressed using cold isostatic pressing at pressure of 300 MPa. The green anode layer was placed in a furnace and heated to a temperature of 960? C. in air atmosphere for 40 minutes, to produce a sintered anode layer 3. A CGO electrolyte layer 4 was sprayed onto the anode layer 3 and fired in a furnace at 1020? C. for 40 minutes. Finally, zirconia layer 7 was applied to the fired electrolyte layer by means of the method disclosed in GB 2 456 445 followed by application of the doped ceria layer 8 and the two cathodic layers 5,6 also using the methods of GB 2 456 445, before firing at a temperature of 825? C. to produce the SOFC 1 structure.
(17) In contrast the SOFC 10 of the invention, whilst appearing to have a similar structure to the SOFC 10 of FIGS. 1 and 2, is prepared in a different way and (as shown in FIGS. 7 to 10) exhibits a good sintering of the nickel oxide phase, a porous anode structure and a contiguous chromium oxide passivation layer 11, between the support 1 and the anode 3. In FIG. 10 the electrolyte layer 4, cathodic layers 5,6, zirconia layer 7 and doped ceria layer 8 are also shown.
(18) The SOFC of FIGS. 7 to 10 is prepared by applying screen printed ink containing suspended particles of nickel oxide powder and CGO powder (d90=0.7 to 1.2 ?m, ratio of nickel oxide to CGO being 1:0.78). The ink was screen printed onto a ferritic stainless steel substrate using conventional methods and dried to evaporate the solvents and set the binders thereby forming a dried, printed layer of thickness 9 to 15 ?m. The dried printed layer was fired in air at a temperature of 1020? C. for 40 minutes to produce a sintered anode layer 3. The furnace was then allowed to cool to room temperature and the air purged from the system using a 5% hydrogen/argon mix.
(19) An atmosphere comprising 4.85 volume % hydrogen, 2.91 volume % water vapour, the remainder being argon was introduced and the furnace heated to 1045? C. The water vapour was introduced into the dry mixture of hydrogen and argon by bubbling the hydrogen and argon mixture through deionised water resulting in an oxygen partial pressure in the reducing atmosphere in the range 10.sup.?17 to 10.sup.?19 bar. The composite was fired in this atmosphere and at this temperature for a time period of 40 minutes allowing reduction of nickel oxide to metallic nickel and sintering of the nickel and rare earth-doped ceria to form a cermet.
(20) After 40 minutes the furnace was allowed to cool and the atmosphere switched to nitrogen bubbled through deionised water. This allowed the partial pressure of oxygen to rise to above 10.sup.?13 bar, leading to oxidation of nickel metal to nickel oxide.
(21) After cooling completely, the anode was re-oxidised by heating it in a furnace in air to 700? C. for 60 min.
(22) The sintered anode 3 was then treated as described above for FIGS. 1 and 2 in order to form a complete solid oxide fuel cell comprising CGO electrolyte layer 4, zirconia layer 7, doped ceria layer 8, and two cathodic layers 5,6.
EXAMPLES
Nickel Oxide and Chromium Oxide Stabilities
(23) The stability of nickel, nickel oxide, chromium, and chromium oxide are of interest in the systems of the invention, as the reduction of nickel oxide to nickel is a key to the functioning of the anode. The formation and preservation of the passivation layer on the SOFC support, which will typically be chromium oxide as ferritic stainless steel substrates are the substrates most commonly used, is important to the prevention of diffusion between the support and the anode, which can potentially contaminate both the anode, reducing it's efficiency, and the support, forming austenitic phases and reducing the supports structural integrity. In addition, the passivation layer prevents degradation of the support during the firing steps used in formation of the fuel cell, and then in use.
(24) FIG. 3 shows a thermodynamic phase diagram for a nickel/nickel oxide system showing the limits of thermodynamic stability of metallic nickel as a function of temperature and oxygen partial pressure. It can be seen that at 1000 to 1100? C., the metallic nickel is stable at an oxygen partial pressures as high as 10.sup.?13 to 10.sup.?14 bar. Therefore, at these and lower partial pressures of oxygen, nickel oxide will reduce to metallic nickel.
(25) FIG. 4 shows the equivalent phase diagram for a chromium/chromium oxide system showing that at 1000 to 1100? C., metallic chromium is only stable at oxygen partial pressures of 10.sup.?22 to 10.sup.?24 bar or below. Therefore, at oxygen partial pressures above around 10.sup.?22 bar a chromium oxide passivation layer will be retained.
(26) FIG. 5 shows a phase diagram for the nickel/nickel oxide system at 1030? C. and 1 bar total pressure as a function of hydrogen and steam partial pressures, showing that any gas mixture containing 0.5-10% water vapour and 1-20% hydrogen is sufficiently reducing that the only stable phase is metallic nickel.
(27) FIG. 6 shows the equivalent phase diagram for the chromium/chromium oxide system showing that for the same range of gas mixtures the only thermodynamically stable phase is chromium oxide.
(28) SOFC Structure
(29) FIG. 7 shows a SEM cross-section of an anode 3 produced by the method described herein, after the initial firing in air. This image shows the ferritic stainless steel substrate 1, a thermally grown chromium oxide scale 11 on the substrate 1, and a weakly sintered porous anode structure 3 consisting of nickel oxide (dark phase45 volume %) and CGO (light phase55 volume %). FIG. 8 is a cross-section of this anode 3 after firing in the reducing atmosphere subsequent reoxidation, and FIG. 9 a higher magnification image of the same anode 3 microstructure. These figures show that the chromium oxide passivation layer 11 remains intact after firing, and that a good sintering of both the nickel oxide phase 12 and the lighter CGO phase 13 is present. Good sintering is evidenced by a clear distinction between ceramic and metallic regions. The ceramic regions appearing as light regions and the metallic regions as dark patches.
(30) FIG. 10 shows a complete SOFC cell 10 with an anode 3 produced by the method described herein after operation of the fuel cell 10. The anode structure 3 can be seen after reduction of the nickel oxide in the anode 3 back to metallic nickel during SOFC operation, along with the other parts of the SOFC 10 as described above.
(31) The resulting anode structure has been demonstrated to be highly REDOX-stable at operating temperatures of <650? C., being capable of withstanding hundreds of high-temperature fuel interruptions without significant cell performance degradation.
(32) SOFC Performance
(33) FIG. 11 is a current-voltage polarisation curve for the fuel cell of FIG. 10, at different operating temperatures. Fuelling rate was calculated to give approximately 60% fuel utilisation at 0.75V/cell at each of the measured temperatures, showing that the system can be operated across a range of temperatures at least as broad as 492 to 608? C., allowing the operational temperature to be optimised for application, number of cells in the stack, output required etc.
(34) FIG. 12 shows the very good REDOX stability possible with this anode structure. A series of cycles are run at 600? C. on a seven-layer short stack, where a current-voltage curve is run to establish the stack performance. The stack is then returned to open circuit, and the hydrogen supply to the stack is cut whilst maintaining the stack at 580-600? C. Air and nitrogen are maintained to the stack during this period. The fuel interruption is sustained for 20 minutes, allowing time for the anode to partially reoxidise. The hydrogen feed is then restored, and after giving the stack a few minutes to recover, another current-voltage curve is run to determine if stack performance has been lost as a result of the REDOX cycle of the anode. This sequence continues until stack performance starts to fall, indicating damage to one or more cells as a result of REDOX cycling.
(35) It can be seen from FIG. 12 that with the SOFC cell of FIG. 10, the seven cells within the stack will tolerate more than 200 REDOX cycles without any measurable loss of performance after a small initial burn-in, with 291 cycles being run in total. A loss of performance observed after 200 cycles was in this instance was due to the failure of one cell at the bottom of the stack; it is believed that mechanical optimisation of the stack design can avoid failure of that layer leading to even greater REDOX stability.
(36) FIG. 13 is a table of the results of mechanical strength tests undertaken on SOFC cells both after initial manufacture and after cells have operated in an initial performance characterisation test, for both standard nickel-CGO anodes as illustrated in FIG. 2, and reduced fired nickel-CGO anodes as illustrated in FIG. 8. The after operating test for the reduced fired nickel CGO anodes included over 250 REDOX cycles.
(37) In the as-manufactured cells, the anodes are in the oxidised state and prior to the mechanical test they are reduced in order to mimic the anode structure in the cell at the start of operating, whereas the anodes in the after operating cells are in the final cermet state of the working anodes.
(38) In order to perform the mechanical strength measurement on the cells, the metal substrates of the cells are first glued to a flat steel plate to prevent the cells flexing when a pulling force is applied. The cathodes of the cells are removed mechanically, exposing the electrolyte.
(39) To assess the mechanical strength of the anode and/or the anode-electrolyte bond, circular metal test pieces are glued to the electrolyte surface in the four corners of the electrolyte and the middle of the cell. A diamond scribe is used to cut through the ceramic layers of the cell around the metal test piece. A calibrated hydraulic puller is then attached to the test piece and used to measure the stress required to pull the test piece off the cell substrate. A maximum pulling stress of 17 MPa may be applied using this technique, after which the glue holding the test piece to the electrolyte tends to fail rather than the fuel cell layers on test. Should the test piece be pulled off at less than 17 MPa this indicates the failure stress of the weakest cell layer (usually the internal structure of the anode).
(40) It can be seen that whilst the standard nickel-CGO anodes are strong in the as-manufactured state, they fail at much lower stresses after reduction of the nickel oxide to metallic nickel in the after operating cell. Without being bound by theory, it is believed this is largely because of the lack of a contiguous ceramic structure within the anode, meaning the mechanical strength of the anode is provided entirely by relatively weak necks between nickel particles. By contrast it can be seen that the reduced fired nickel-CGO anodes retain their strength after reduction to the cermet structure, indicating much greater sintering of both metallic and ceramic phases.
(41) It should be appreciated that the processes and fuel cells of the invention are capable of being incorporated in the form of a variety of embodiments, only a few of which have been illustrated and described above.
