(560da) Analysis of PEFC Cathode Catalyst Layer Based on Rate Dependency on Oxygen Partial Pressure | AIChE

(560da) Analysis of PEFC Cathode Catalyst Layer Based on Rate Dependency on Oxygen Partial Pressure

Authors 

Kageyama, M. - Presenter, Kyoto University
Yamaguchi, K., Kyoto University
Kawase, M., Kyoto University
Matsuda, T., JARI

Introduction

In a polymer electrolyte fuel cell (PEFC), protons generated on the anode are transported through the anode catalyst layer and the proton exchange membrane (PEM) to the cathode. Electrons are generated on the anode and transported through the external circuit to the cathode. On the cathode of the PEFC, the oxygen reduction reaction (ORR) takes place as follows:

O2 + 4 H+ + 4 e ¨ 2 H2O. (1)

Phenomena on the cathode are governed by interaction between the reaction and the transport of oxygen and protons. In our previous study, the effectiveness factor of cathode catalyst layer (CCL) was reported to be a function solely of 4 dimensionless moduli, MO(C)m, Mp(C)m, PO(C)m and yOc,1,2) which were derived from the isothermal one-dimensional model of the cathode consisting of oxygen balance, oxygen transport, proton transport, and reaction stoichiometry. MO(C)m does not depend on the oxygen partial pressure, while Mp(C)m depends on it. In this study, a method to determine the dimensionless moduli which control the effectiveness factor from experimental data measured at different oxygen partial pressures is proposed.

Theory

The ORR rate per unit volume of the CCL, rvc, is expressed as follows:

rvc = kvc* exp(–Ec/bc) pO [mol/(m3·s)], (2)

where kvc* is the reaction rate constant per unit volume of CCL [mol/(Pa·m3·s)] at Ec = 0, Ec is the cathode electromotive force (emf) [V], bc is the Tafel slope [V] and pO is the oxygen partial pressure [Pa]. The cathode emf is the difference between the electron potential, φe, and the proton potential, φp. The reaction rate constant is not constant since Ec derivates in the CCL from the boundary condition, and rvc depends on the oxygen partial pressure and cathode emf. The current density, i, is a product of the reaction rate without mass transfer resistances, kvcmpOc, and the effectiveness factor, Fe(C).

i = 4 F δ(C)kvcmpOcFe(C) [A/m2], (3)

where the subscript m means the PEM–CCL boundary and the subscript c means the CCL–gas diffusion layer (GDL) boundary. δ(C) is the CCL thickness and F is the Faraday constant. Fe(C) is a function of 4 dimensionless moduli, MO(C)m, Mp(C)m, PO(C)m, and yOc, which have the following meanings respectively:

MO(C)m: Ratio of reaction to oxygen diffusion (Thiele modulus),

Mp(C)m: Ratio of reaction to proton conductivity (Our modulus),

PO(C)m: Ratio of convection to oxygen diffusion (Peclet number), and@

yOc: Oxygen mole fraction.

MO(C)m and Mp(C)m are defined as follows:

MO(C)m = δ(C)(kvcmRT/DeO)1/2 and (4)

Mp(C)m = δ(C)[4FkvcmpOc/(σepbc)]1/2, (5)

where DeO is the effective oxygen diffusion coefficient [m2/s], and σep is the effective proton conductivity [S/m].

If the temperature, T, is identical, MO(C)m is constant regardless of the oxygen partial pressure since both the reaction rate and the oxygen diffusion rate are proportional to the oxygen partial pressure. On the other hand, Mp(C)m changes with the oxygen partial pressure since the proton transport rate is independent of the oxygen partial pressure. Under the reaction control conditions, the current density, i, is proportional to pOc since Fe(C) = 1. Under the oxygen transfer control conditions, i is proportional to pOc since Fe(C) is inversely proportional to MO(C)m and MO(C)m is not affected by pOc. Under the proton transport control conditions, i is proportional to pOc0.5 since Fe(C) is inversely proportional to Mp(C)m. Therefore, the slope of log i versus log pOc is located between 0.5 and 1 when δ(C) and Ecm are fixed. Eqs. (3) and (5) indicate the relationship between i / pOc and pOc0.5 corresponds to the relationship between Fe(C) and Mp(C)m. The former relationship is obtained from measurements and the latter is obtained from the theory. In case of PO(C)m = 0, if MO(C)m and yOc are given, the theoretical curve of Fe(C) versus Mp(C)m is determined. Then MO(C)m is obtained by fitting the experimental results with the theoretical curve.

Experimental

The membrane electrode assembly (MEA) was consisted of the catalyst layers (CLs) made of TEC10E50E (Tanaka Kikinzoku Kogyo K.K.) catalyst with Nafion® ionomer and membrane (DuPont NR-211). Pt/C weight ratio was 0.50, ionomer/carbon ratio was 0.90, and the CLfs thickness was 9.0 μm. The cell temperature was 80 °C, the relative humidity (RH) was 100 % and the total pressure was 101.3 kPa. I–V measurement was carried out at the oxygen partial pressure of 3.24, 6.47, 11.3, and 53.9 kPa.

Results and discussion

When i / pOc was plotted against pOc0.5, the slope of log i versus log pOc obtained from the experiments was between 0.5 and 1. This indicates that the experiments were carried out neither under the proton transport control conditions nor the reaction control conditions.

The slope of the theoretical line of log Fe(C) versus log Mp(C)m is nearly 0 when Mp(C)m is negligibly low. The slope of the line is –1 when Mp(C)m is extremely high. When the experimental data are located on a horizontal line, it means that the proton transport resistance is negligible. When the experimental data are located on an Mp(C)m–1 line, it means the data were measured in the proton transport control region. In both the extreme cases, Mp(C)m cannot be determined from the log i versus log pOc plot. For determination of MO(C)m and Mp(C)m, data should be analyzed where proton transport resistance is not extremely low or high.

By the data fitting it was found that MO(C)m was 1.11 at Ecm = 0.74 V and MO(C)m was 2.1 at Ecm = 0.7 V, for instance. By substituting MO(C)m, Mp(C)m, and Fe(C) in eqs. (3), (4), and (5), kvcm, σep, and DeO were determined as follows at 101.3 kPa, 80 °C, RH = 100 %, and I/C = 0.9:

kvc = 1.03~108 exp ( –Ec / 0.0383 V ) mol/(Pa·m3·s),

DeO = 7.33~10–8 m2/s, and

σep = 1.26 S/m.

The obtained Tafel slope, 88.2 mV/decade, is close to the values reported by others for the same catalysts.3,4) The apparent Tafel slope appears greater than the true value if the transport resistances in the catalyst layer increase. A high value such as 118 mV/dec. at 35 °C5) was obtained from measurements without model analysis. Although the obtained effective oxygen diffusion coefficient is lower than that reported by other researchers,6,7) the effect of convective flow was ignored in this and other usual studies, which resulted in the scattered values of DeO. The effective proton conductivity as well as the effective oxygen diffusion coefficient depends remarkably on the pore and agglomerate structure. The value of the effective proton conductivity is in good agreement with the values, 0.9 S/m and 2.0 S/m, measured at similar I/C and similar RH reported in literture.8,9)

Conclusion

A method to determine the dimensionless moduli governing the PEFC cathode behavior from the dependency of the ORR rate on the oxygen partial pressure measured using a usual cell and MEA is proposed. The electrochemical reaction rate constant and transport properties were successfully determined by the proposed method. Measurements of dependency of the rate on the oxygen partial pressure should be carried out at identical total pressure, temperature, and RH by using a same MEA. In the analysis, only current and voltage data measured under the conditions where proton transport resistance is not extremely low or high can be used for determining the dimensionless moduli, MO(C)m and Mp(C)m. From the dimensionless moduli determined by the analysis, the reaction rate constant, kvc, and the transport properties, σep and DeO, can be estimated.

References

[1] M. Kawase, K. Yamaguchi, M. Kageyama, K. Sato, and G. Inoue, ECS Trans. 75(14), 147–156 (2016).
[2] M. Kawase, K. Sato, R. Mitsui, H. Asonuma, M. Kageyama, K. Yamaguchi, and G. Inoue, AIChE J. 63(1), 249–256 (2017).
[3] S. Takaichi, H. Uchida, and M. Watanabe, Electrochim. Acta. 53(14), 4699–4705 (2008).
[4] W. Zhang, S. Shironita, and M. Umeda, Int. J. Hydrogen Energy 41(15), 6526–6533 (2016).
[5] A. Parthasarathy and C. R. Martin, J. Electrochem. Soc. 138(4), 916–921 (1991).
[6] G. Inoue and M. Kawase, J. Power Sources 327, 1–10(2016).
[7] S. Litster, W. K. Epting, E. A. Wargo, S. R. Kalidindi, and E. C. Kumbur, Fuel Cells 13(5), 935–945 (2013).
[8] R. Makharia, M. F. Mathias, and D. R. Baker, J. Electrochem. Soc. 152(5), A970–977 (2005).
[9] Y. Liu, M. W. Murphy, D. R. Baker, W. Gu, C. Ji, J. Jorne, and H. A. Gasteiger, J. Electrochem. Soc. 156(8), B970–980 (2009).