Analysis of Kinetic Parameters and Effect of Pt Loading on Cell Performance of PEFC Electrodes Prepared by Inkjet Printing

Several fabrication methods are available for production of the fuel cell electrodes [1]. We have been developing inkjet printing to fabricate thin, low loading electrodes and conventional loading electrodes. Due to the reduced loading and thickness of these electrodes, the peak Pt utilization improved significantly and was one of the highest reported values in the literature [2]. A parametric study of the effect of Nafion loading in the catalyst layer (CL) revealed that these electrodes were not sensitive to Nafion loading changes between 20 – 40 wt%. The electrodes also showed a higher dependance on oxygen partial pressure than conventional spray coated electrodes [2]. An issue identified with these electrodes was the low kinetic performance when operating with H₂/air at ambient pressures due to their low Pt loadings which limited the cell efficiency.

In order to understand the poor kinetic performance at low Pt loading, electrodes with different Pt loadings, Pt/C particle type and ink solvents were fabricated and tested at varying oxygen partial pressures. First, the effect of Pt loading in the range of 0.014 to 0.2 mg/cm² is studied for 20% Pt/C and 30% Nafion. The loading was controlled by the number of printed passes. The effect of Pt loading on cell performance has been known for a quite long time [3,4]. There exists a threshold value of Pt loading above which kinetic gains are substantially lower [4]. Using 20% Pt/C catalyst, a steep increase in performance was observed as the Pt loading increased from 0.014 to 0.075 mg/cm², corresponding to a CL thickness change from around 1 µm to 4.5 µm. However, in the range of 0.075 to 0.2 mg/cm² the change in performance was substantially reduced in the kinetic region. Similar trends were observed at higher oxygen partial pressures and at different RH conditions. Kinetic parameters such as active area, reaction order, Tafel slope and exchange current density were also studied. As expected, the mass specific active area (m²/g) primarily decreased with Pt loading [5], while the overall active area for the electrodes (cm²_Pt) increased. The Tafel slope measured in the range of 20 – 100 mA/cm² at 90% RH showed a change from 0.136 V/dec to 0.09 V/dec for increasing Pt loadings from 0.014 to 0.1 mg/cm². The exchange current densities were found to be in the range of 2.7x10^-6– 5.4x10^-6 A/cm² with the lower loadings showing higher values.

Due to the increased sensitivity of low loading electrodes to oxygen partial pressure, the effect of partial pressure is studied. Recent literature has shown a change in both apparent Tafel slope and reaction order (γ) with potential [6]. Therefore, the oxygen reaction order was analyzed at different potentials. Preliminary analysis of γ for a 15 layered CCM (0.075 mg/cm²) gave a value of 0.55 when measured at a constant overpotential of 0.15 V (w.r.t OCV).

In the previous studies, an increase in loading resulted in an increase in electrode thickness. In order to decouple the two effects, electrodes were fabricated using 20% and 40% Pt/C particles. The performance of the electrodes with same Pt loading using different Pt/C catalyst particles was different. For the case of 46.7% catalyst, the optimal Pt loading was found to be 0.17 mg/cm². Effect of Pt loading on the variation in reaction order is currently being conducted.

Finally, inkjet printing necessitates the use of a high viscosity solvent. Electrodes manufactured using glycerol, ethylene glycol and propylene glycol were tested. It is found that electrode performance in the kinetic region is notably affected by the type of solvent used even though no solvent could be detected in the layer using cyclic voltammetry [7].

In summary, low and high loading cathode fuel cell electrodes are accurately manufactured using an inkjet printer. The effect of increased loading either by increasing the number of layers or changing the catalyst type is studied. Testing is performed under several partial pressures to be able to estimate reaction order. The effect of solvent is also studied.

References

[1] Litster, S. et. al. J Power Sources 130.1 (2004): 61-76.

[2] Shukla, S. et al. Electrochimica Acta 156 (2015): 289-300.

[3] Qi, Z. et al.. J Power Sources 113.1 (2003): 37-43.

[4] Gasteiger H. et. al., J Power Sources 127.1 (2004): 162-171.

[5] Saha, M. et al.JES 158.5 (2011): B562-B567.

[6] Subramanian N. et al., JES., 2012, 159(5), B531-B540.

[7] Dailey, A. et al. Electrochimica acta 44.6 (1998): 1147-1152.