ration. In this case, variations in the ratio of photo/dark respiration may cause some changes in  max , especially in conditions of inorganic carbon limitation favorable to photorespiration such as in terrestrial C 3 vegetation. However, aquatic plants developed mechanisms to con- centrate CO 2 and suppress photorespiration [A. Kaplan and L. Reinhold, Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 539 (1999)]. As a result, photorespiration is not likely to affect oceanic  max . 11. H. Craig and L. I. Gordon, in Conference on Stable Isotopes in Oceanographic Studies and Paleotem- peratures, E. Tongiorgi, Ed. (Laboratory of Geology and Nuclear Science, Pisa, Italy, 1965), pp. 9 –130. 12. Air-water equilibrium was attained in less than 24 hours by bubbling outside air into seawater (25°C). The  eq value was determined in five separate ex- periments. Its average value was 16 per meg with a standard error of 2 per meg. For simplicity in calculations we assumed that the deviation of  eq from zero is the result of isotopic fractionation occurring only during O 2 invasion. In the cases discussed here, errors due to this assumption are negligible. 13. J. F. Clark et al., in Air-Water Gas Transfer, B. Jaehne and E. C. Monahan, Eds. (Aeon Verlag & Studio, Hanau, Germany, 1995), pp. 785– 800. 14. P. D. Nightingale et al., Global Biogeochem. Cycles 14, 373 (2000). 15. H. Craig and T. Hayward, Science 235, 199 (1987). 16. W. J. Jenkins and J. C. Goldman, J. Mar. Res. 43, 465 (1985); W. S. Spitzer and W. S. Jenkins, J. Mar. Res. 47, 169 (1989). 17. B. Luz and E. Barkan, data not shown. 18. A. F. Michaels and A. H. Knap, Deep-Sea Res. II 43, 157 (1996); S. C. Doney, D. M. Glover, R. G. Najjar, Deep-Sea Res. II 43, 591 (1996). 19. B. B. Benson and D. Krause Jr., Limnol. Oceanogr. 29, 620 (1984). 20. Data is found at www.bbsr.edu/Weather/climatology. html; R. Wanninkhof, J. Geophys. Res. 97, 7373 (1992). 21. We are grateful to M. Bender for numerous discussions on all aspects of this research. Comments by A. Kaplan, Y. Kolodny, and three anonymous reviewers significantly improved the manuscript. We appreciate the help of the Bermuda Biological Station in sampling at BATS and extend special thanks to S. Bell. The support of the Kinneret National Laboratory is appreciated. Y. Yacobi and Y. Sagi helped in gross-production measurements in the Sea of Galilee, and J. Erez and K. Schneider helped with the coral experiment. The support of the U.S.-Israel Binational Science Foundation, The Israel Science Foun- dation, MARS-2, and the Moshe-Shilo Minerva Center is greatly appreciated. 6 December 1999; accepted 11 May 2000 A Low-Operating-Temperature Solid Oxide Fuel Cell in Hydrocarbon-Air Mixtures Takashi Hibino, 1 * Atsuko Hashimoto, 1 Takao Inoue, 2 Jun-ichi Tokuno, 2 Shin-ichiro Yoshida, 2 Mitsuru Sano 2 The performance of a single-chamber solid oxide fuel cell was studied using a ceria-based solid electrolyte at temperatures below 773 kelvin. Electromotive forces of 900 millivolts were generated from the cell in a flowing mixture of ethane or propane and air, where the solid electrolyte functioned as a purely ionic conductor. The electrode-reaction resistance was negligibly small in the total internal resistances of the cell. The resulting peak power density reached 403 and 101 milliwatts per square centimeter at 773 and 623 kelvin, respectively. Fuel cells are widely viewed as a promising source of low-emission power generation for vehicles. There is great controversy over which fuel should be used. Polymer electro- lyte fuel cells (PEFCs) exhibit high power densities at low temperatures (353 K), but they require hydrogen as the fuel, which is impractical in terms of storage and handling. An external reformer can be used to convert alcohols and hydrocarbons into hydrogen, but their portability is inferior. There have been recent successes with solid oxide fuel cells (SOFCs), which perform well between 823 and 973 K using methane (1) and n-butane (2) directly as the fuels. A further reduction in the operating temperature of SOFCs and an enhancement in their thermal and mechanical shock resistance would make this technology a promising alternative to PEFCs. A type of fuel cell that consists of only one gas chamber, where both the anode and the cathode are exposed to the same mixture of fuel and air, has been proposed by many researchers (3–8). This design is more shock resistant than conventional fuel cells, both thermally and mechanically. We have recent- ly succeeded in applying this single-chamber cell design to a SOFC constructed from yttria-stabilized zirconia (YSZ), which is commonly used as a solid electrolyte in SOFCs, with a Ni-based anode and a perov- skite cathode (9). This SOFC, however, must operate at the high temperature of 1223 K to achieve sufficient ionic conduction in the sol- id electrolyte. Different cation-doped ceria, notably sa- maria-doped ceria (SDC), have much higher ionic conduction than YSZ in an oxidizing at- mosphere, whereas they show n-type semicon- duction in a reducing atmosphere (10, 11). Be- cause the resulting electromotive force (EMF) of the SOFC is lower than the theoretical value, the SDC electrolyte has so far been regarded as unsuitable for such applications. However, the partial pressure of oxygen at the boundary of the two atmospheres becomes gradually lower as the operating temperature decreases (12), which suggests that the SDC electrolyte can be used even under fuel cell conditions, provided it operates at extremely low temperatures. In this report, we demonstrate a low-temperature SOFC by combining the advantages of the SDC electrolyte with the single-chamber cell design. The SDC electrolyte we used here was pre- pared by pressing a commercial ceramic pow- der, Ce 0.8 Sm 0.2 O 1.9 (Anan Kasei Co. Ltd.), hy- drostatically into a pellet at 2  10 3 kg cm -2 and then sintering in air at 1773 K for 10 hours. After the pellet was cut into a disk (diameter 14 mm, thickness 1 mm), the SDC disk surface was polished to a given thickness (0.15 to 0.50 mm) with an abrasive paper. YSZ (8 mol% yttria) and La 0.9 Sr 0.1 Gd 0.8 Mg 0.2 O 3 (LSGM) were used as solid electrolytes for com- parison. Preliminary experiments revealed that 10 weight % SDC–containing Ni and Sm 0.5 Sr 0.5 CoO 3 electrodes best functioned as the anode and the cathode, respectively. These pastes were smeared on the opposite surfaces (area 0.5 cm 2 ) of the SDC disk, followed by calcining in air at 1223 K for 4 hours. The cell thus fabricated was placed in an alumina tube (inner and outer diameters 15 and 19 mm, respectively). Methane, ethane, and propane were mixed with air for each of the respective concentrations—30 volume % for methane, 18 volume % for ethane, and 14 volume % for propane—so that the oxidation would proceed safely without exploding (the explosive limits of methane, ethane, and pro- pane in air are 15.0, 12.5, and 9.5 volume %, respectively). The gas mixture was supplied to the cell at flow rates of 300 ml min -1 between 623 and 773 K (Fig. 1). When the single-chamber SOFCs using SDC, YSZ, and LSGM with a thickness of 0.50 mm were supplied with a mixture of ethane and air at 773 K, all three cells gen- erated stable EMFs of 920 mV, where the 1 Department of Structure Formation Process, Nation- al Industrial Research Institute of Nagoya, Nagoya 462-8510, Japan. 2 Graduate School of Human Infor- mation, Nagoya University, Nagoya 466-0804, Japan. *To whom correspondence should be addressed. E- mail: thibino@nirin.go.jp Fig. 1. A schematic illustration of single-cham- ber SOFC in a flowing mixture of hydrocarbon and air. R EPORTS www.sciencemag.org SCIENCE VOL 288 16 JUNE 2000 2031