Development of Glassy Alloy Separators for a Proton Exchange Membrane Fuel Cell (PEMFC)

Transcription

1 Materials Transactions, Vol. 46, No. 7 (2005) pp to 1710 #2005 The Japan Institute of Metals Development of Glassy Alloy Separators for a Proton Exchange Membrane Fuel Cell (PEMFC) Akihisa Inoue 1, Takayoshi Shimizu 2, Shin-ichi Yamaura 1, Yuichiro Fujita 2, Shinobu Takagi 2 and Hisamichi Kimura 1 1 Institute for Materials Research, Tohoku University, Sendai , Japan 2 Daido Steel Co. Ltd., Nagoya , Japan The Nb 15 Ti 15 Zr 10 glassy alloy was developed as a separator material for a proton exchange membrane fuel cell (PEMFC). The corrosion rate of the glassy alloy in sulfuric acid was approximately 3 digits lower than SUS316L. A precise groove forming of the glassy alloy can be achieved by hot-pressing in the supercooled liquid state. The result of the power generation test of the single cell assembled with the groove-formed glassy alloy sheets revealed that sufficient power was generated. Moreover, the deterioration of the morphology of the glassy separator could hardly be recognized even after the durability test for 350 h. (Received March 9, 2005; Accepted May 17, 2005; Published July 15, 2005) Keywords: metallic glass, separator, bipolar plate, solid polymer fuel cell (SPFC), polymer electrolyte fuel cell (PEFC) 1. Introduction A fuel cell exhibits higher efficiency as compared to internal combustion engines by directly converting the chemical energy of fuels to electric energy. 1) It is thus capable of considerably reducing the amount of discharged carbon dioxide. Furthermore, fuel cells can contribute to the suppression of fossil fuel consumption due to their capability of using various types of fuels such as natural gas and methanol, and can contribute to the environmental preservation in that they discharge small amounts of nitrogen- and sulfur-oxides that cause air pollution. Among various types of fuel cells, the proton exchange membrane fuel cell (PEMFC) is currently under development but is soon expected to be prevalent as an electric source mainly in homes and electric cars, owing to its advantageous characteristics such as high output power density and low temperature operation. 2) Some problems concerning the main constituents of fuel cells such as electrolyte films, catalysts, diffusion layers and separators have yet to be resolved. 3) Many researchers are engaged in investigating the separators 4,5) in the viewpoint of the material design 6) and the flow route design. 7) The functions of the separator are to transfer the fuel and the oxidizer to the reaction site, contain the reaction products, accumulate the electricity generated, and mechanically support the cell. The separator having such significant roles, is the constituent that accounts for more than 60% of the weight and more than 30% of the cost of the entire fuel cell. Therefore, the mass, volume and cost of the fuel cell can be expected to drastically decrease by the substitution of the present separator substrate, fine graphite, with another suitable material. As one of the simplest substitutes, metallic alloy, especially stainless steel, may be considered as a candidate. 6) However, stainless steel is not practical as a substitute yet because it results in a voltage drop due to the generation of a passive film. Therefore, we are now discussing and investigating the use of the metallic glass as a new separator material. As a noncrystalline material, metallic glass is assumed to be one of the most promising substrates owing to its high corrosion resistance, high strength, and precise processing capability. The results obtained so far are reported. 2. Experimental Alloy ingots were prepared by arc-melting the mixture of pure metals in an Ar atmosphere. Thin sheet-shaped specimens of 20 and 50 mm in width were produced by using a single-roller melt-spinning equipment. In order to verify the amorphicity of the alloys, these sheet specimens were examined by X-ray diffraction (XRD). Thermal stability and the supercooled liquid region of the specimens were examined by a differential scanning calorimeter (DSC) at the heating rate of 40 K/min. The PEMFC environment comprises weak acidic media. In order to simulate the PEMFC atmosphere, dilute sulfuric acid solution was used as a test solution. The corrosion resistance test was performed as follows: 400 ml of 1 mass% sulfuric acid water solution was poured into a beaker attached to a return flow-cooling pipe, and the thin as-spun specimens of mm was soaked in the solution; finally, the boiling test was performed for 168 h at around 368 K to measure the weight reduction and amount of dissolved ions. The corrosion test apparatus is shown in Fig. 1. For the groove-forming test, hot-pressing with precisely grooved dies was performed under the controlled atmosphere. For the electricity generation test, the standard cell from Japan Automobile Research Institute was used with a grooved thin separator set in the graphite frame. The humidifying temperature of both the anode and the cathode was 343 K. The electrode area was 103 mm 2. The cell operating temperature was also fixed to be 343 K with the humidity of 100%RH. Hydrogen and oxygen were used as the fuel gas and the oxidizing gas, respectively. Gas flow rate of both hydrogen and oxygen was 0.5 L/min.

2 Development of Glassy Alloy Separators for a Proton Exchange M 1707 Return flow-cooling pipe Boiling 1.0mass% H 2 SO 4 solution Specimen Heater Fig. 2 Alloy compositions prepared in this study. Ni content is fixed to 60 at% in the Ni Nb Ti Zr system. Fig. 1 Boiling corrosion test apparatus. Table 1 Summary of the metallic glass formability, thermal stability and brittleness of the Fe-based melt-spun alloys. Chemical composition DSC Bending XRD of the alloy (at%) T g (K) T x (K) T x (K) test at R.T. Fe 43 Cr 16 Mo 16 C 15 B Fe 64 Cr 10 Mo 5 C 8 P Fe 69 Cr 10 Mo 5 C 8 P Fe 65 Cr 10 Mo 5 B 4 P (?) Fe 67 Cr 10 Mo 5 B Fe 72 Si 9:6 Nb 4 B 14: Fe 67 Si 9:2 Nb 4 Cr 6 B 13:8 895 Note 1) XRD; Amorphous, Bending test; broken Note 2) T g : Glass transition temperature, T x : Crystallization temperature T x ¼ T x T g : Supercooled liquid region 3. Results and Discussion 3.1 Fe-based glassy alloys Since Fe-based alloys are suitable for this application in terms of cost, the alloy design was made taking corrosion resistance into consideration. Previous references report on some Fe-based glassy alloys in Fe Cr Mo C B 8) and Fe Cr C B P 9) systems having good corrosion resistance. As shown in Table 1, although a glassy phase can be obtained in all the alloys prepared in this study, the as-spun sheet specimens of the glassy alloys easily break during the bending tests at room temperature because of their severe brittleness. Consequently, further investigation on the Febased glassy alloys was discontinued because of the difficulty to apply these alloys to the separator. 3.2 Ni-based glassy alloys Since the problem of brittleness could not be resolved in Intensity (a.u.) Fig. 3 Nb 15 Ti 10 Zr 15 Nb 20 Ti 10 Zr 10 Nb 5 Ti 25 Zr 10 Nb 5 Ti 30 Zr 5 Nb 20 Ti 20 Nb 10 Ti θ Results of the XRD investigation on the Ni-based alloy. Fe-based alloys, probably due to a harmful influence of nonmetallic elements, an alloy design of Ni-based alloys without non-metallic elements was performed. The Ni-based glassy alloys in the Ni Nb Ti Zr system, developed by Inoue et al., have a comparatively large supercooled liquid region T x (¼ T x T g ) of 76 K at maximum. 10) The range of the alloy chemical compositions examined in the present study is shown in Fig. 2. As shown by the XRD results in Fig. 3 and the DSC measurement in Fig. 4, all the alloys transform into their respective glassy states with the apparent glass transition temperatures. Judging from the corrosion rate of these glassy alloys in the sulfuric acid boiling test, as shown in Fig. 5, all the alloys have much better corrosion resistance in comparison with SUS316L. It was proved that an increase in the amount of Nb results in a decrease in the corrosion rate, and in the case of more than 15% Nb, the corrosion resistance may be saturated.

3 1708 A. Inoue et al. Exothermic (a.u.) 40 K/min X=20 X=15 X=10 X=5 Tg Tx Temperature, T / K 1073 Stress, σ / MPa K 863 K 843 K 823 K 803 K RT Elongation (%) Fig. 6 Results of the tensile test on the as-spun thin specimens (strain rate: s 1 ). Fig. 4 Results of the DSC measurement of the Nb x Ti ð30 xþ Zr 10 glassy alloys. Nb 5 Ti 30 Zr 5 Nb 5 Ti 25 Zr 10 Nb 10 Ti 30 Nb 10 Ti 25 Zr 5 Nb 10 Ti 20 Zr 10 Nb 15 Ti 25 Nb 15 Ti 20 Zr 5 Nb 15 Ti 15 Zr 10 Nb 15 Ti 10 Zr 15 Nb 20 Ti 20 Nb 20 Ti 15 Zr 5 Nb 20 Ti 10 Zr 10 ( SUS316L ) Fig E E E E E+00 Corrosion rate, c / mm y 1 Corrosion rates of various Ni-based alloys in sulfuric acid. 3.3 Groove-forming test In the case of the separator, it is essential for the gas to distribute uniformly over the reaction site. Therefore, it is very important that the separator is designed with a precise gas flow route. It is known that the metallic glasses exhibit Newtonian viscous flow in the supercooled liquid region, thereby allowing precise grooves to be formed in that temperature range. In this study, a sheet specimen of the Nb 15 Ti 15 Zr 10 glassy alloy with a width of 50 mm and a thickness of approximately 60 mm was used to make tensile test specimens of the JIS-6 type (width of the parallel part: 10 mm, distance between the standard points: 50 mm, the length of the entire specimen: 150 mm) for conducting the tensile test at a desired high temperature. The result is shown in Fig. 6. The glass transition temperature T g of this alloy is 828 K, while the crystallization temperature T x is 890 K. At room temperature, the alloy exhibits elastic deformation but not plastic deformation. The tensile strength at room temperature is measured to be 1874 MPa at the strain rate of s 1, which is three times higher than that of SUS316. However, as the temperature increases, plastic deformation can be observed. In particular, in the supercooled liquid range, it is observed that the strength decreases drastically while the elongation caused by the deformation increases. By the tensile test, it is confirmed that the specimens at 863 K and 873 K (near the T x temperature) are brittle, indicating that they are in the partially crystallized state. On the basis of this result, 843 K is concluded to be the most suitable temperature for the groove-forming. The TTT diagram of this alloy is shown in Fig. 7. It is clarified that the alloy remains in a glassy phase state for 600 s at 843 K and changes to the crystalline phase after 180 s at 863 and at 873 K. The crystallization occurs at 863 and 873 K during the heating for the tensile test, since the tensile tests are conducted after the test piece is maintained at the desired temperatures for 60 s. The groove-forming test was carried out in the range of 823 K to 843 K. The groove has a depth of 0.6 mm, a bottom Temperature, T / K Fig. 7 Tx Tg Supercooled liquid Crystallization Time, t / s TTT diagram of the Nb 15 Ti 15 Zr 10 glassy alloy.

4 Development of Glassy Alloy Separators for a Proton Exchange M Fig. 8 Schematic illustration of the cross section of the desired groove (mm). Table 2 Summary of groove formability. width of 0.9 mm, and a pitch of 2.7 mm as shown in Fig. 8. The results listed in Table 2 reveal that the deformation was insufficient only at 823 K. Figure 9 shows the photograph as well as the cross section profile after the groove was formed. As seen in the figure, the cross section profile exhibits an extremely good shape. In case of plural serpentine flow-routes with an increase in the number of grooves, the perfect deformation was observed at 843 K. The photograph of the resulting grooves is shown in Fig. 10. As mentioned above, the actually forming result is in agreement with the tensile test results, proving the existence of the temperature range for good formability. Forming temperature (K) Insufficient in the forming Intensity (a. u.) After forming Before forming θ Fig. 11 XRD patterns before and after the groove-forming. Electric voltage, V / V I V curves Nb 15 Ti 15 Zr 10 BMG SUS316L Electric current density, I / 100 A m 2 Fig. 12 I V characteristics of the cell with the metallic glass separator. 1mm Fig. 9 Appearance and the cross sectional view of the groove-formed specimen. Fig. 10 Plural serpentine grooves formed on the thin specimens. Fig. 13 Appearance of the metallic glass separators after power generation for a long duration.

5 1710 A. Inoue et al. It is confirmed that the glassy phase is maintained and the corrosion speed does not change even after the grooveforming. According to the XRD result shown in Fig. 11, the groove-forming at 843 K does not cause any change in XRD patterns. The corrosion rate was also measured to be 5: mm/y for the as-spun sheet specimen and 4: mm/y after the forming, thereby demonstrating that there is no change in corrosion rate before and after the forming test. 3.4 Electricity-generating test Using the Ni-based glassy alloy with the formed groove, the electricity-generating performance was examined. A single cell was assembled by inserting the groove-cut specimen produced in this study into the standard cell carbon separator and the early stage of the test was carried out to estimate the electricity-generating performance. The humidifying temperature of both the anode and the cathode was 343 K. The electrode area was 103 mm 2. The cell operating temperature was also fixed to be 343 K with the humidity of 100%RH. Hydrogen and oxygen were used as the fuel gas and the oxidizing gas, respectively. Gas flow rate of both hydrogen and oxygen was 0.5 L/min. The result is shown in Fig. 12 in comparison with SUS316L. As shown in the figure, the passive film causes an extreme voltage drop in SUS316L, while the newlydesigned glassy alloy enables the generation of even higher electric current, implying a practical current level. Subsequently, a durability test was conducted by assembling the cell in the same manner and then conducting the electricity-generating test for a long duration to investigate the voltage variation. The durability test was performed at the current density of 0.5 A/cm 2. The test was continued for 350 h. The morphology of the glassy alloy separator after the durability test was shown in Fig. 13. As shown in the figure, no degradation in the separator was observed. 4. Summary We investigated the potential of a metallic glass for the separator of PEMFCs. The results obtained in this study are summarized as follows: (1) The Ni-based glassy alloy sheets were produced by the single-roller melt-spinning technique. The corrosion rate of the Nb 15 Ti 15 Zr 10 glassy alloy in sulfuric acid was approximately 3 digits lower than that of SUS316L. (2) A precise groove forming can be achieved by hotpressing under the supercooled liquid condition. (3) The result of the power generation test of the single cell assembled with the groove-formed sheets Nb 15 Ti 15 Zr 10 glassy alloy revealed that sufficient power was generated. Moreover, the deterioration in performance could hardly be recognized even after the durability test performed for 350 h. It was confirmed that the Ni-based metallic glass has the significant potential as a separator material. Acknowledgement This work was supported by the New Energy and Industrial Technology Development Organization (NEDO) of Japan under the research program, Research and Development of Polymer Electrode Fuel Cell. REFERENCES 1) F. T. Bacon: Electrochim. Acta 14 (1969) ) M. Warshay, P.R. Prokopius, M. Le and G. Voecks: Proceedings of the Intersociety Energy Conversion Engineering Conference (1997), Vol. 1, pp ) F. N. Buechi and S. Srinivasan: J. Electrochem. Soc. 144 (1997) ) N. P. Brandon, S. Skinner and B. C. H. Steele: Ann. Rev. Mater. Res. 33 (2003) ) V. Mehta and J. S. Cooper: J. Power Sources 114 (2003) ) H. L. Wang and J. A. Turner: J. Power Sources 128 (2004) ) T. Bewer, T. Beckmann, H. Dohle, J. Mergel and D. Stolten: Proceedings of the First European PEFC forum (EFCF) (2001), pp ) S. Pang, T. Zhang, K. Asami and A. Inoue: Mater. Trans. 43 (2002) ) S. Pang, T. Zhang, K. Asami and A. Inoue: Acta Mater. 50 (2002) ) A. Inoue, W. Zhang and T. Zhang: Mater. Trans. 43 (2002)