Y and Nb-doped SrTiO3 as a mixed conducting anode for solid oxide fuel cells

This application claims the benefit of U.S. application Ser. No. 60/482,714, filed Jun. 27, 2003.

This invention relates to SrTiO 3 co-doped with a first trivalent metal dopant atom (e.g. yttrium) replacing some of the Sr atoms on a strontium sub-lattice in the A site of the molecule and a second pentavalent dopant metal atom(e.g. niobium) replacing some of the Ti atoms on a titanium sub-lattice in the B site of the molecule, and to solid oxide fuel cells including an anode made of the co-doped SrTiO 3 .

The solid oxide fuel cell (SOFC) is one of the most advanced systems for generating electricity in an efficient and environmentally friendly way. It can operate on a variety of fuels in addition to hydrogen, and may find applications from transportation to stationary systems (1, 2).

There has been a long history of research on the application of cermets (ceramic-metal composites) of nickel and yttria stabilized zirconia (YSZ) as anodes in solid oxide fuel cells. However, Ni—YSZ has a number of drawbacks including sintering of the Ni particles at high operating temperatures, sulfur poisoning and carbon deposition when the SOFC is fuelled by natural gas, reliance on a triple phase junction for the electrochemical reaction and cost. Ni—YSZ cermet anodes cannot be used in pure methane without pre-reforming. Therefore, an alternative material with adequate conductivity (>50-100 S/cm at 800-1000° C.) and catalytic activity toward total or partial oxidation of methane is necessary. Mixed ionic-electronic conductor (MIEC) oxides, either perovskite type (3-5) or non-perovskite (6-9) type, exhibiting high ionic and electronic conductivities at elevated temperatures are attractive candidates for SOFC electrodes. Mixed conductors are believed to relax the limitation attributed to electronic conducting electrodes by expanding the active reaction zone to include the whole of the electrode-gas interface instead of so-called triple phase boundary area (10).

According to one embodiment of the invention, a novel compound of general formula I is provided

(Sr 1-1.5x M 1 x ) 1-y/2 Ti 1-y M 2 y O 3   I

wherein M 1 is first trivalent dopant metal atom replacing some strontium atoms on a strontium sublattice, x is a mole percent of said dopant atoms M 1 on the strontium sublattice and 0<x≦0.04, and M 2 is second pentavalent dopant metal atom replacing some titanium atoms on a titanium sublattice, y is a mole percent of said dopant atoms M 2 on the titanium sublattice and 0≦y≦0.2.

According to an aspect of this embodiment of the invention, M 1 is yttrium (Y +3 ) and wherein x is in a range from 0 to 4 mole % and M 2 is niobium (Nb +5 ) and wherein y is in a range from 0 to 20 mole %.

According to another aspect of this embodiment of the invention, 0<x≦0.04 and 0≦y≦0.2.

According to yet another aspect of this embodiment of the invention x=0.04 and 0≦y≦0.2.

According to yet another aspect of this embodiment of the invention, 0.03≦x≦0.04 and 0.05≦y≦0.20.

According to another embodiment of the invention, a novel compound of general formula II is provided

(Sr 1-1.5x M 1 x ) 1-y/2 Ti 1-y M 2 y O 3-δ   II,

wherein M 1 is first trivalent dopant metal atom replacing some strontium atoms on a strontium sublattice, x is a mole percent of said dopant atoms M 1 on the strontium sublattice and 0<x≦0.04, and M 2 is second pentavalent dopant metal atom replacing some titanium atoms on a titanium sublattice, y is a mole percent of said dopant atoms M 2 on the titanium sublattice and 0≦y≦0.2 and δ is in a range from 0 to about 0.07.

According to an aspect of this embodiment of the invention, M 1 is yttrium (Y +3 ) and wherein x is in a range from 0 to 4 mole % and M 2 is niobium (Nb +5 ) and wherein y is in a range from 0 to 20 mole %.

According to another aspect of this embodiment of the invention, 0<x≦0.04and 0≦y≦0.2.

According to yet another aspect of this embodiment of the invention x=0.04 and 0≦y≦0.2.

According to yet another aspect of this embodiment of the invention, 0.03≦x≦0.04 and 0.05≦y≦0.20.

According to another embodiment of the invention, a process is provided for making a compound of formula I.

That is, the compounds of formula I were prepared in air in which the A-site deficiency is introduced to balance the charge difference of partially substituting Y 3+ for Sr 2+ and Nb 5+ for Ti 4+ . The stoichiometric ratios of A site elements (Sr 2+ , Y 3+ ) and B site elements (Ti 4+ , Nb 5+ ) were adjusted to match the electrically neutral condition under oxidation atmosphere.

The compounds of formula I with all species in their fully oxidized states are then partially reduced at elevated temperature in a forming gas e.g. Ar-8% H 2 , to obtain the novel oxygen deficient compounds of formula 11.

Therefore, “δ”(delta) will be “0” under air. However, the coupling of Ti 4+ /Nb 5+ can be reduced to Ti 3+ /Nb 4+ under reducing atmosphere at high temperature. Empirically we calculated the δ (delta) of the formula II based on thermo-gravimetric data as shown in FIG. 6 . The range of δ (delta) depends on the range of y in the formula and can be defined as 0≦δ≦0.07 in case of 0≦y≦0.2 for x=0.04. In this case, both electronic and ionic conduction is achieved which is a favorable criterion for SOFC anode performance. The thermal and electrical behavior of these novel ceramics was measured, and is described below.

According to another embodiment of the invention, a solid oxide fuel cell is provided, comprising a laminate of a cathode, an anode and an electrolyte sandwiched between said cathode and anode, said anode comprising a compound of general formula (Sr 1-1.5x M 1 x ) 1-y/2 Ti 1-y M 2 y O 3 I, wherein M 1 is first trivalent dopant metal atom replacing some strontium atoms on a strontium sub-lattice, x is a mole percent of said dopant atoms M 1 on the strontium sub-lattice and 0<x≦0.04, and M 2 is second pentavalent dopant metal atom replacing some titanium atoms on a titanium sub-lattice, y is a mole percent of said dopant atoms M 2 on the titanium sub-lattice and 0≦y≦0.2.

According to an aspect of this embodiment of the invention, M 1 is yttrium (Y 3+ ) and wherein x is in a range from 0 to 4 mole % and M 2 is niobium (Nb 5+ ) and wherein y is in a range from 0 to 20 mole %.

According to another aspect of this embodiment of the invention, 0.03≦x≦0.04 and 0.05≦y≦0.20.

According to another aspect of this embodiment of the invention, x=0.04 and y=0.20.

According to another aspect of this embodiment of the invention, the electrolyte is yttrium stabilized zirconia (YSZ) and the anode has chemical stability therewith at operating temperature and also at the SOFC sintering temperature of up to 1350° C. It will be appreciated by those skilled in the art that the electrolyte and cathode may be made of any materials conventionally used for such components in commercial SOFC fuel cells. For examples of such materials see refs. 13 and 14, the disclosures of which are incorporated herein by reference.

According to another aspect of this embodiment of the invention, the anode includes a metallic catalyst and/or a ceramic catalyst mixed with said compound of formula I, for increasing conductivity and/or catalytic activity of said anode.

According to another aspect of this embodiment of the invention, the metallic catalyst is selected from the group consisting of iron, cobalt, nickel, copper and alloys thereof.

According to another aspect of this embodiment of the invention, the ceramic catalyst is selected from the group consisting of undoped cerium oxide, doped cerium oxide, lanthanum and chromium based perovskites.

According to another aspect of this embodiment of the invention, the solid oxide fuel cell is operated at a temperature ranging from 400° C. to 1000° C., and wherein an anode compartment is operated at an oxygen pressure ranging from about 10 −14 to about 10 −22 atm.

In another embodiment of the invention, an electrolyte-supported single cell SOFC with (Sr 0.94 Y 0.04 ) 0.9 Ti 0.8 Nb 0.2 O 2.93 as the anode was made and tested at 900° C. using forming gas (Ar-8% H 2 ) as a fuel.

FIG. 1 is lattice parameters of yttrium doped strontium titanate as a function of the concentration of yttrium.

FIG. 2 is a graph illustrating XRD patterns of (Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3 sintered in air at 1400° C. for 4 h.

FIG. 3 is a graph illustrating variation of the lattice parameters of (Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ sintered in air and forming gas.

FIG. 4 is a graph illustrating the electrical conductivity of (Sr 0.94 Y 0.4 ) 1-y/2 Ti 1-y Nb y O 3-δ at 800° C. as a function of oxygen partial pressure.

FIG. 5 is a graph illustrating the conductivity of (Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ versus temperature in forming gas.

FIG. 6 is a graph of oxygen deficiency versus doping level of niobium in (Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ .

Compounds of formula (Sr 1-1.5x Y x ) 1-y/2 Ti 1-y Nb y O 3 wherein 0<x≦0.04 and 0≦y≦0.2 were prepared by conventional solid-state reaction. High purity SrCO 3 (Aldrich, 99.9%), Y 2 O 3 (Aldrich, 99.99%), TiO 2 (Alfa Aesar, 99.8%, rutile), and Nb 2 O 5 (Aldrich, 99.9%) were intimately mixed in the appropriate ratios, and heated, in air, at 1000° C. for 5 h, then at 1200° C. for 10 h, and finally, reground and reheated at 1400° C. for 10 h. Dense pellets (17 mm diameter) and bars (ca. 5×2.5×20 mm) of each sample were obtained by pressing at 1000 kg/cm 2 and then sintering at 1400° C. for 4 h in forming gas (Ar-8% H 2 ). Polyvinyl alcohol (PVA) and polyethylene glycol (PEG) were used as the binder and the plasticizer, respectively. The compounds were sintered on yttria stabilized zirconia (YSZ) substrates. No reaction with the substrate was detected after sintering.

Diffraction patterns were recorded on powders and ceramics on a Bruker D8 Advanced X-ray diffractometer (Bruker AXS Inc.) with Cu K α radiation. For accurate lattice parameter determination, silicon was chosen as a standard. Complete patterns were recorded over the range of 2θ from 20 to 70°. The unit cell parameters were derived from a computerized least-squares refinement technique.

The conductivity measurements were made using a conventional four-probe technique. Pt mesh was attached to the surface with Pt paste 4082 (Ferro Inc.). The contacts were cured in situ under forming gas for 1 hour at 950° C. The temperature was then reduced to 800° C. to perform the conductivity measurements. Initially, argon with trace amounts of oxygen was allowed to leak into the measurement cell over a period of 5-6 hours and the conductivity monitored with changing oxygen partial pressure measured by a POAS micro-sensor (Setnag). The conductivity as a function of temperature was also recorded.

To determine the oxygen content in the reduced compounds, thermogravimetric analysis was performed using a Hi-Res TGA thermogravimetric analyzer (TA Instruments). The samples were oxidized by heating at 5° C./min up to 1000° C. in flowing air.

Thermal expansion coefficients were measured in the temperature range of 100-800° C. using TMA 2940 thermomechanical analyzer (TA Instruments).

The lattice parameters of perovskite yttrium doped strontium titanate as a function of yttrium contents are shown in FIG. 1 . It can be seen that single-phase cubic perovskite-type solid solutions of yttrium doped strontium titanate were observed for compositions containing up to 4 mol % of yttrium. The effect of yttrium doping into perovskite strontium titanate on enhancing mixed conductivity under reducing atmosphere was well described in literature (3). The conductivity was found to increase with increasing concentration of yttrium. Therefore compounds of Y and Nb co-doped strontium titanate with 4 mol % doping of yttrium were investigated to determine the effect of the solid solubility of niobium on the titanium sublattice. Single-phase compounds were observed by powder X-ray diffraction for y<0.2. For y=0.2 nearly single-phase samples were observed with a very small extra peak at 2θ=30.5° (see FIG. 2 ). This peak can be assigned to Y 2 Ti 2 O 7 . Therefore, niobium probably decreases the solubility level of yttrium in SrTiO 3 . Any further increase of Nb in the system results in an increase of second phase content. Cell parameters for the (Sr 0.94 Y 0.004 ) 1-y/2 Ti 1-y Nb y O 3-δ compounds sintered in air and forming gas at 1400° C. are given in Table 1. As the Nb content increases, the unit cell expands (see FIG. 3 ), due to the larger radius of Nb 5+ (0.64 Å) versus Ti 4+ (0.605 Å) (11). The diffraction patterns resemble very closely that of cubic SrTiO 3 . Lattice parameters were determined by indexing the diffraction peaks by analogy with those of SrTiO 3 . On reduction all samples exhibited an approximate 0.03-0.15% expansion in crystallographic unit cell size due to the larger sizes of Ti 3+ (0.67 Å) and Nb 4+ (0.68 Å).

Four probe dc conductivity measurements of the compounds measured at 800 ° C. over a wide range of oxygen partial pressures are shown in FIG. 4 . The conductivity of all Nb-doped titanates increases slightly with decreasing oxygen partial pressures, indicating n-type conductivity predominated under reducing conditions. The electron concentration, or Ti 3+ /Nb 4+ concentration is saturated by the Nb doping level, and the conductivities were only a weak function of oxygen partial pressure in the range 10 −22 -10 −15 atm. One may see that the Y and Nb co-doped compound with x=0.04 and y=0.2 has much higher conductivity than Sr 0.94 Y 0.04 TiO 3 at 800° C. in the whole range of oxygen partial pressures measured. It would be desirable to use this material as a porous anode on which a thin layer of electrolyte is attached. Electrical conductivity of (Sr 1-1.5x Y x ) 1-y/2 Ti 1-y Nb y O 3-δ may be influenced by factors such as excess strontium/yttrium vacancies and the existence of Ti 3+ (or Nb 4+ ) which are produced under reducing conditions. Bulk densities and conductivities of the various (Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ compositions are given in Table 2.

For the proposed application as an anode in a SOFC, it is important to investigate the variation of conductivity of (Sr0.94Y0.04) 1-y/2 Ti y Nb y O 3-δ in forming gas as a function of temperature. FIG. 5 shows the conductivity measured from low to high temperature. It is well known that the oxygen partial pressure in forming gas is determined by the equilibrium:

H 2 ( g )+½O 2 ( g )←→H 2 O ( g )

and significantly decreases with decreasing temperature. Therefore, the increase of conductivity with decreasing temperature is very likely owing to the decrease of oxygen partial pressure. However, when the temperature approaches a kinetic limit (in this case 250-350° C.) the system may not be in thermodynamic equilibrium and the temperature mainly controls the generation of electrons rather than the reduction of oxygen partial pressure.

Thermogravimetric measurements were performed on (Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ compounds as prepared under reducing conditions. The compounds were heated under flowing air up to 1000° C. to complete oxidation. Measurements showed that oxidation of the compounds begins at 840-870° C. The weight gain associated with the oxidation process was used to determine the amount of oxygen per formula unit. The calculated oxygen deficiency as a function of Nb content is plotted in FIG. 6 . The δ monotonically increases with increasing Nb concentration.

The thermal expansion coefficients of (Sr0.94Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ compounds were compared with that of (Y 2 O 3 ) 0.08 (ZrO 2 ) 0.92 . All materials showed near linear expansion, with similar thermal expansion coefficients. The observed value of α=11.6×10 −6 K −1 (from 100° C. to 800° C.) for y=0.2 was close to the 10.8×10 −6 K −1 reported for YSZ (12). The measured thermal expansion coefficients of the other Nb-doped (y<0.2) ceramics were a little higher, in the range 11.6-12.7×10 −6 K −1 .

The anode material must have chemical compatibility with the electrolyte, not only at operating temperature but also at the higher temperatures that the anode is exposed to during the fabrication of the SOFC. Chemical compatibility was tested by grinding a 1:1 (by weight) mixture of TZ-8Y (Tosoh Zirconia) and (Sr 0.94 Y 0.04 ) 0.9 Ti 0.8 Nb 0.2 O 3 and firing at 1350° C. for 10 h in forming gas. No changes in the X-ray patterns were observed after this heat treatment.

(Sr 0.94 Y 0.04 ) 1-y/2 Ti 1-y Nb y O 3-δ compounds have suitable features for application as an anode material for a SOFC in the sense that they are stable in both air and reducing conditions, exhibit quite high (62.8 S/cm at y=0.2) conductivity in low Po 2 , and do not show any evidence of chemical reaction with YSZ electrolyte even at 1350° C.

A single cell SOFC was made by screen-printing (Sr 0.94 Y 0.04 ) 0.9 Ti 0.8 Nb 0.2 O 3 -glycol slurry on one side of a YSZ disc (18 mm diameter, 0.35 mm thick) and firing at 1200° C. in forming gas. Pt paste (Ferro Inc.) was applied on other side of the disk as the cathode. Forming gas saturated with water vapor at 20° C. was delivered to the anode and air was provided to the cathode. Unfortunately the cell showed poor power density (max 15 mW/cm 2 at 900° C.). Specific electrode interfacial resistance was roughly estimated at 700 Ω·cm 2 . This indicates that the main problem may lie with the interface contacts, rather than with the bulk properties of the anode.

1. M. C. Williams, in Solid Oxide Fuel Cells VI, S. C. Singhal, M. Dokiya, Editors, PV 99-19 p. 3, The Electrochemical Society Proceedings Series, Pennington, N.J. (1999).

2. H. Yokokawa, in Solid Oxide Fuel Cells VI, S. C. Singhal, M. Dokiya, Editors, PV 99-19 p. 10, The Electrochemical Society Proceedings Series, Pennington, N.J. (1999).

3. S. Hui, A. Petric, J. Electrochem. Soc., 149 (1) J1-J10 (2002).

4. T. Kawada, N. Iizawa, L-Q. Han, K. Yashiro, A. Kaimai, Y. Nigara, and J. Mizusaki in: in Proc. 2001 Joint Int. Meeting, Electrochem. Soc., Abstract No 1558.

5. P. R. Slater, D. P. Fagg, J. T. S. Irvine, J. Mater. Chem., 7 (12), 2495 (1997).

6. D. P. Fagg, S. M. Fray, J. T. S. Irvine, Solid State Ionics, 72, 235 (1994)

7. O. Porat, M. A. Spears, C. Heremans, I. Kosacki, H. L. Tuller, Solid State Ionics, 86-88, 285 (1996).

8. P. R. Slater, J. T. S. Irvine, Solid State Ionics, 124, 61 (1999).

9. H. L. Tuller, Solid State Ionics, 94, 63 (1997).

10. H. L. Tuller, in High Temperature Electrochemistry: Ceramics and Metals, Editors F. W. Poulsen et. al. Riso National Laboratory, Roskilde, Denmark, p. 139 (1996).

11. R. D. Shannon, Acta Crystallogr. A32, 751 (1976)

12. R. Maenner, E. Ivers-Tiffee, W. Wersing, and W. Kleinlein in: Proc. 2 nd Int. Symp. on Solid Oxide Fuel Cells, eds. F. Gross, P. Zegers, S. C. Singhal and D. Yamamoto (CEC, Luxemburg, 1991) p. 715

13. B. C. H. Steele, Solid State Ionics, 134, 3-20 (2000)

14. F. Tietz et al., Solid State Ionics, 152-153, 373-381, (2002)