Elsevier

Journal of Power Sources

Volume 196, Issue 9, 1 May 2011, Pages 4264-4269
Journal of Power Sources

Non-dimensional analysis of PEM fuel cell phenomena by means of AC impedance measurements

https://doi.org/10.1016/j.jpowsour.2010.11.004Get rights and content

Abstract

AC impedance or electrochemical impedance spectroscopy (EIS) is becoming a fundamental technique used by researchers and scientists in proton exchange membrane (PEM) fuel cell analysis and development. In this work, in situ impedance measurements are presented for a series of operating conditions in a 50 cm2 fuel cell. The electrode charge transfer resistance was determined from the corresponding arcs of the Nyquist diagrams. The analyses were performed for H2/O2 and H2/air operation at different stoichiometric factors and reactant gases humidification. Characteristic time scales of charge transfer processes at the different operating conditions were estimated from the corresponding Bode plots. These values were used for a non-dimensional analysis of the different fuel cell electrochemical and transport processes, namely electrochemical reaction versus GDL reactant transport. Fuel cell adapted Damkhöler numbers are thus presented, where the results indicate that the GDL diffusion transport is the limiting process for the cases under analysis, especially when air is used as oxidant. Additional analysis of channel convective mass transport versus GDL diffusive mass transport is also presented.

Introduction

AC impedance (electrochemical impedance spectroscopy) is an experimental technique widely used in the analysis of electrochemical systems, and has been established as a fundamental diagnostic and research tool for PEM fuel cells during the last years [1], [2]. One of the main advantages of AC impedance techniques is that it allows for the determination of the different resistances occurring in a PEM fuel cell, corresponding to the different electrochemical and transport processes: activation or charge transfer resistance, ohmic resistance, and transport or concentration resistance. On contrast, the polarization curve itself does not provide detailed information about the different processes as they are tightly coupled and the polarization curve describes only the integral output of all processes. AC impedance is widely applied to multiple MEA analysis such as influences of the catalyst loading, PTFE content, Nafion content, GDL structure, manufacturing methods, membrane thickness, and others [1]. In an important number of studies EIS analysis has been applied to characterize fuel cell processes. Parthasarathy et al. [3], Antoine et al. [4] and Neyerlin et al. [5] analysed the oxygen reduction reaction (ORR) kinetics. Xu et al. [6] and Neyerlin et al. [7] also analysed the effect of relative humidity conditions on the ORR kinetics of PEM fuel cells. HOR (hydrogen oxidation reaction) kinetics and the effect of CO poisoning were studied by Wagner and Gülzow [8]. Transport losses were analysed by Springer et al. [9], Eikerling and Kornyshev [10], Lefebvre et al. [11], and Saab et al. [12], [13], who also analysed the influence of the processing conditions of the electrode. Liu et al. [14] also studied the resistance to the protonic conduction at the cathode ionomer layer using AC impedance. The cell ohmic resistance is also commonly determined by means of AC impedance, as in the work of Cooper and Smith [15]. Romero-Castañón et al. [16], Song et al. [17] or Wagner [18] have applied AC impedance for the evaluation and optimization of the membrane electrode assembly (MEA). Additional applications are the analysis of electrode and GDL flooding or membrane dry-out, what is commonly referred to as MEA State-of-Health (SOH), as in the work reported by Fouquet et al. [19] or Mérida et al. [20]. The results are presented for the stack under analysis, but the analysis of single cells in a stack is also feasible as reported by Hakenjos et al. [21], who presented a measurement set-up featuring a multi-channel frequency response analyser (FRA) for the simultaneous measurement of impedance spectra of single cells in a fuel cell stack.

Non-dimensional analysis has proven to be a powerful tool for engineering design and analysis, especially in the chemical engineering field, although its application to complex systems such as fuel cells is not yet fully established [22]. However, there is a need for comprehensive and numerically less expensive description of cell performance in order to implement it into full stack models that cannot be analysed by means of Computational Fluid Dynamics (CFD) models except for small-scale units [23]. Therefore, a detailed description based on dimensional analysis of the cell phenomena would be of high relevance for further stack research and development, and important contributions are already in progress as established in the work of Gyenge [22].

This work presents AC impedance results for a 50 cm2 PEM fuel cell. The electrode activity and membrane protonic conductivity were determined elsewhere [24] based on the measurements performed for different operating conditions, aimed at the fuel cell parameter estimation needed for numerical CFD simulations [25]. In this work, the application of AC impedance measurements to the non-dimensional analysis of PEM fuel cell phenomena is presented, and results for a 50 cm2 PEM fuel cell are reported. In particular, Nyquist and Bode plots are presented for a fuel cell with serpentine flow field, operating with oxygen and air at different humidification conditions. Measurements at several current densities are presented in order to better assess the influence of the current density on the limiting processes affecting the performance of the cell.

Section snippets

PEM single cell description

The single fuel cell analysed consists of commercial graphite bipolar plates with a five-channel serpentine flow field design (Fig. 1) from ElectroChem Inc. (USA). The GDL is a Sigracet 10 CC from SGL Group (Germany), with 420 μm thickness, porosity 0.82 and 10% PTFE content, featuring a micro porous layer (MPL). The 50 cm2 membrane used is a CCM type (catalyst coated membrane) from Baltic fuel cells (Germany), with Nafion-117 and 0.3 mg Pt/cm2 and 0.6 mg Pt/cm2 catalyst load in anode and cathode

Results

The results of the measurements are presented in the Nyquist diagrams shown in Fig. 2, Fig. 3. The diagrams feature two impedance arcs. The ZRe axis intercept at high frequencies represents the overall resistance of the cell (electronic at the bulk materials, electronic contact resistances, and ionic resistances at the membrane) and it corresponds to the high frequency resistance (HFR). In general, the next semicircle diameter at medium frequencies represents the activation overpotential

Electrochemical and GDL diffusion mass transport time scales: Damkhöler number

The results plotted in the Nyquist diagram can also be interpreted as characteristic time scales of the different processes occurring within the fuel cell. The time constant τ of a particular physical or chemical phenomenon corresponds to the frequency peak of the corresponding arc in the Nyquist plot. Therefore, τ = 1/fpeak [29]. As frequency is not explicitly shown in Nyquist plots, the representation of the impedance versus frequency in a Bode diagram is more convenient.

By determining the time

Conclusions

AC impedance results for a 50 cm2 PEM fuel cell are presented. Besides analyzing the electrode charge transfer resistance from the Nyquist plots obtained for different operating conditions, this work outlines a non-dimensional analysis of PEM fuel cell transport phenomena. Characteristic time scales for electrochemical reaction and reactant transport processes in channels and GDL are derived either from experimental Bode plots or from correlations found in the literature. Damkhöler numbers were

Acknowledgements

This work has been funded by the Secretariat-General of Universities, Research and Technology, from Junta de Andalucía, under the P08-TEP-4309 project. Authors also acknowledge INTA for its close collaboration in PEMFC CFD modelling and experimental validation.

References (41)

  • X. Yuan et al.

    Int. J. Hydrogen Energy

    (2007)
  • J. Wu et al.

    Int. J. Hydrogen Energy

    (2008)
  • O. Antoine et al.

    J. Electroanal. Chem.

    (2001)
  • N. Wagner et al.

    J. Power Sources

    (2004)
  • M. Eikerling et al.

    J. Electroanal. Chem.

    (1999)
  • K.R. Cooper et al.

    J. Power Sources

    (2006)
  • T. Romero-Castañón et al.

    J. Power Sources

    (2003)
  • J.M. Song et al.

    J. Power Sources

    (2001)
  • N. Fouquet et al.

    J. Power Sources

    (2006)
  • W. Mérida et al.

    J. Power Sources

    (2006)
  • A. Hakenjos et al.

    J. Power Sources

    (2006)
  • E.L. Gyenge

    J. Power Sources

    (2005)
  • S. Shimpalee et al.

    Electrochim. Acta

    (2009)
  • A. Iranzo et al.

    Int. J. Hydrogen Energy

    (2010)
  • A. Iranzo et al.

    Int. J. Hydrogen Energy

    (2010)
  • H.A. Gasteiger et al.

    Appl. Catal. B: Environ.

    (2005)
  • J.H. Nam et al.

    Int. J. Heat Mass Transfer

    (2003)
  • P.T. Nguyen et al.

    J. Power Sources

    (2004)
  • X.D. Wang et al.

    J. Power Sources

    (2008)
  • W. Sun et al.

    J. Power Sources

    (2005)
  • Cited by (31)

    • Dimensionless numbers on columnar structured electrodes in hydrogen/oxygen polymer electrolyte fuel cells

      2023, Chemical Engineering Journal
      Citation Excerpt :

      This is probably due to the presence of attractive lateral interactions between oxygen adsorbates on columnar platinum electrodes. The difficulties of transport phenomena in PEMFCs were discussed in a critical review [43]. The problem has to be foreseen as multiphase flows to get knowledge of the interactions in electrochemical catalytic interphases.

    • 2D dimensionless numbers in isothermal fuel cells with smooth electrocatalysts

      2022, Chemical Engineering Science
      Citation Excerpt :

      We prefer direct measurements in one single channel (Fuller et al., 1966; Gu et al., 2018) by electrochemical sensors designed in our lab. Thus, the measurement of the cathodic jo at the beggining of the experiment was 0.5 Acm−2 similar to those found before (Kulikovsky, 2014; Chevalier et al., 2018; Iranzo et al., 2011). On the other hand, large values of diffusivity for oxygen in the diffusion layer were found, that is, 0.02 cm2s−1.

    View all citing articles on Scopus
    View full text