Elsevier

Corrosion Science

Volume 48, Issue 5, May 2006, Pages 1036-1058
Corrosion Science

Stainless steels in bipolar plates—Surface resistive properties of corrosion resistant steel grades during current loads

https://doi.org/10.1016/j.corsci.2005.05.012Get rights and content

Abstract

Instances of an increase in the contact resistance of stainless steel surfaces, resulting a decrease in the efficiency of fuel cells, were reported by automotive developers of PEM fuel cells, who had used the AISI 316L grade (EN 1.4404) in bipolar plates. To identify the reason for this, surface resistive properties were measured for a number of stainless steel grades, using cyclic current density loads. Electrochemical impedance spectroscopy (EIS) was used as a tool for these measurements of the resistive properties before and after the current density loads. In addition, a high continuous galvanostatic current density was imposed on one of the most stable and corrosion resistant stainless steel grades, 904L, to evaluate any conductive changes of the passive film due to temperature, current or a combination of the two.

Introduction

Attempts to bring polymer electrolyte membrane (PEM) fuel cells into mass production have emphasised the importance of the availability and quality of materials used in the cell. Stainless steel has been recognised as a candidate material for use in the bipolar plates [1]. Instances of a loss of efficiency due to increases in the contact resistance of stainless steel surfaces were reported by automotive developers of PEM fuel cells who had used the AISI 316L steel grade (EN 1.4404) in bipolar plates. However, for metal screen and foil hardware in PEM fuel cells, Zawodzinski et al. tested the stainless steel grade AISI 316L and concluded that this steel grade should be sufficient with regard to its corrosion properties [1]. On the other hand, a recent study showed that unsuitable corrosion rates have occurred when using a 316L type steel as the end plates in a single cell configuration [2], thus contaminating the membrane. In a previous study, Makkus et al. demonstrated that there were variations in the contact resistance properties of stainless steel grades used as bipolar plates. The authors also referred to another stainless steel grade, not specified in the publication, which performs better than AISI 316L, when given the proper pre-treatment and possibly used with a gasket. This unnamed grade also satisfied the requirements specified for using stainless steel for the bipolar plates in PEM fuel cells [3]. In addition, Davies et al. identified large variations in electrical resistance depending on the elemental composition of the steel grade. As an example, the authors showed that for a number of chosen steel grades tested under fuel cell conditions, the relative interfacial resistance decreased when comparing the steel grades of type 304 > 347 > 316 > 904L [4]. A limit of the value of the contact resistance of bipolar plates in real fuel cells was reported as <20  cm2, which can be regarded as a low electrical resistivity [3]. Other requirements of the material are that it should be low weight, high strength, corrosion resistant, cost efficient, easy to machine and it should be possible to produce it in thin dimensions.

Holm defined the contact resistance, Rc, for a circular contact spot, Eq. (1) in which the resistivity of the bulk material is of major importance [5], [6].Rc=ρ/Dwhere ρ is the resistivity of the bulk material and D is the conducting contact diameter.

Furthermore, Eq. (1) indicates that the surface roughness of the sample will influence the contact diameter and indirectly the contact resistance. The contact resistance was also reported to be dependent not only on the materials properties but on the internal pressure in the fuel cell stack [7].

The passive film properties of a stainless steel surface, such as corrosion resistance and other oxide properties, e.g. surface conductivity, composition of the oxide and thickness, change in relation to the surrounding environment [9]. The thickness of a passive film is commonly considered to be in the range of 1–3 nm depending on the environment and the steel grade, but it also depends on pH values and the potential obtained during polarisation [10], [11]. In addition, a passive film on a stainless steel surface is in electrical terms regarded as a semiconductor consisting of an inner layer of mixed iron/chromium oxides and an outer layer of chromium hydroxide [3], [12], [13].

Imposing a current density on any stainless steel will result in an increased potential in the positive, anodic direction. This increased potential alters the passive film depending on the chemical equilibrium reactions at the received voltage of the stainless steel and the alloy elements in the steel grade. For an iron–chromium alloyed stainless steel, in an acid environment, at anodic potentials, iron will be selectively dissolved and there will be an oxidation of Fe(II) to Fe(III) [14]. Several authors have also showed that for iron–chromium alloys, the composition of the passive film varies depending the chromium content of the steel [2], [15], [16].

The increase of the potential in the anodic direction may further increase the risk of pitting corrosion if the potential rises above the breakdown potential of the steel grade’s passive layer. The type of solution in the service environment is of major importance and it is well established that chlorides have an adverse effect on stainless steels in terms of corrosion. On the other hand, sulphates are effective inhibitors due to their properties of competitive adsorption on a stainless steel surface.

Electrochemical impedance spectroscopy (EIS) has been widely used for predicting the performance of coatings used for corrosion protection [17], [18]. EIS has also been widely used for investigating the electrical properties of various oxide films on metals [19], [20], [21]. Jüttner reported that when metal samples with protective oxide films, i.e. a passive film, are exposed to a non-aggressive electrolyte, the response in impedance is determined essentially by the resistive and capacitive properties of the passive film [22]. The capacitive and resistive properties of the passive films can be evaluated and related to the thickness of the film and the resistivity of the oxide [21]. The capacitive response measured is often inadequate and this is normally rectified by utilising a constant phase element (CPE) for spectra fitting instead of an ideal capacitance element. The impedance representation of CPE and resistance is produced by:CPE=Yo-1jω-nR2=ρdA-1where CPE is the constant phase element, R2 is the resistance, ρ is the resistivity of the oxide, A is the surface area, d is the passive film thickness, R1 is the electrolyte resistivity.

The elements CPE and R2 are not always dependent on corrosion resistance but they also reflect the overall electrical resistance and dielectric properties. In an oxidative environment, the passive film growth depends on the transport of cations and anions or their vacancies across the oxide film. If defects such as pores, channels or cracks are present in the passive film, the electrolyte will penetrate the film and impair its resistance. Further, if the passive film breaks down, i.e. pitting corrosion occurs, this will lead to a marked reduction in the overall resistance, R2 [21]. However, an increase in R2 can relate to the resistivity property, ρ, and the passive film thickness, d according to Eq. (3) and provide a further relation to the contact resistance according to the definition by Holm in Eq. (1). Consequently, in this work it was desirable to maintain a low resistance R2 and to avoid any decrease due to corrosion, since any corrosion products would contaminate the membrane of a PEM fuel cell.

A surface oxide film exhibits a capacitive behaviour due to the dielectric nature of the oxide. In an ideal scenario, the exponential factor of n is equal to 1 and the CPE acts like a capacitor. Heterogeneous surfaces that contain examples of surface roughness or porosity contribute to a deviation from the ideal capacitive behaviour. The constant phase element, CPE, was thus identified as a capacitance, C, and the variation of the resistance of the passive film was evaluated based on Eqs. (3), (4), taking the thickness increase into account [23], [24]. By registering the capacitive measurements, the contribution attributable to the thickness of the passive film could be identified and studied in relation to and in comparison with the overall resistance R2.C=ε0εAd-1where ε is the dielectric constant, ε0 is the permittivity of vacuum (8.85 × 10−14 F/cm).

When evaluating the increase in thickness of the passive film after imposing current loads, the dielectric constant was considered to be a constant and to be independent of the imposed current loads.

The conductive properties of the passive film can also be monitored using the Mott–Schottky analysis. The passive film capacitance can be analysed from the following equations found in the literature [25]:Csc=-(ωZ)-1where Csc is the space charge capacitance.

The space charge capacitance of a p-type semiconductor is given by:C-2=-2(ε0εeNAA2)-1(V-Vfb-kBTe-1)where NA is the acceptor concentration in the passive film, V is the applied voltage, Vfb is the flat band voltage, e is the charge of the electron (1.60219 × 10−19 C), kB is the Boltzmann’s constant.

In an ideal scenario, the junction between n and p is abrupt where an n-type region containing a constant net donor concentration is next to a region with a constant net acceptor concentration. However, in practice the transition between both these regions will be gradual. The p–n junction may be considered to be a capacitor, and as an approximation we consider the p–n junction to be a parallel plate condenser. Seeger calculates the relation between the capacitive properties as a relation to the resistivity on the p-side as an example, but since the junction is graded rather than abrupt, it is rather difficult to verify this experimentally [26]. However, by examining the linear relationship between an exponent of C and the variation in the set potential, it is possible to plot the experimental data. The slope is then a function of the resistivity according to Eq. (7).(Aε0ε/C)2=(VD+VB)ρwhere VB is the applied voltage, VD is the diffusion voltage, ρ is the resistivity on the p-side, A is the cross-section area.

Localised corrosion such as pitting is an undesired effect of the breaking down of the passive film [8]. When pitting corrosion occurs, the function of the fuel cell is destroyed by the corrosion. For this reason it is vital that the stainless steel remains in the passive state during fuel cell operation. The pitting resistance equivalent (PRE) of a specific steel grade can be estimated by formulas in which the relative influence of a few elements, i.e. chromium, molybdenum and nitrogen, are considered. In stainless steel, chromium, molybdenum and nitrogen are the most essential alloying elements for enhancing the resistance to local corrosion. The PRE is calculated on the basis of the composition of the alloy and provides information on the corrosion resistance of the stainless steel grade according to:PRE=%Cr+3.3%Mo+16%NAlfonsson and Qvarfort investigated several formulas to calculate PRE values and compared the PRE values with critical pitting temperatures (CPT) measured according to ASTM G150 using the Avesta cell [27]. They found an acceptable linearity between the CPT and PRE values for any of the PRE formulas studied. The higher the PRE value, the better the resistance to corrosion.

In this work, the aim was to study any correlation between variations in the alloy composition of stainless steels and an increase in resistive properties and passive film thickness. Furthermore, the aim was to find a stainless steel grade which would be suitable for use in bipolar plates, i.e. which demonstrated a combination of corrosion resistance and good surface conductivity properties, thereby minimising the effects of performance loss caused by increases in the resistive properties of the stainless surface during exposure to current.

Section snippets

Material

All steel grades used were produced with a bright annealed (BA) surface and all test coupons underwent a predefined laboratory surface treatment prior to the tests. The surface treatment involved a controlled pickling process in the laboratory prior to repassivation of the stainless steels in order to obtain reproducible passive films on the surfaces. The pickling solution was a mixture of 22% nitric acid and 6% hydrofluoric acid for all steel grades. Pickling times appropriate to the steel

Electrochemical impedance spectroscopy (EIS) measurements

Fig. 3 shows measured and simulated EIS data before and after the short duration current loads for the austenitic grade 904L in Nyquist representation showing the resistive and imaginary parts of the impedance measurements and Fig. 4 shows the data in Bode representation showing any phase shift and impedance data measured for 904L. Fig. 5 shows measured and simulated EIS data before and after short duration current loads for the steel grade S30815, commonly used for high temperature

Discussion

As a consequence of Eq. (1), one of the parameters influencing contact resistance is surface roughness. This is commonly independent of the steel grade, but depends more on the process and pickling methods used in the production of stainless steels [30]. The deviations from linearity in the error calculations also indicate that it was not possible to obtain completely reproducible surfaces from production.

It was not possible in this work to reproduce a real fuel cell environment. This was

Conclusions

The following conclusions can be drawn after subjecting the stainless steel grades tested to cyclic current density loads:

  • The 201 stainless steel grade (7%Mn) produced a lower increase in resistance compared with the other steel grades tested, but showed an unstable voltage response during longer periods of current load.

  • The alloy element manganese would appear to contribute to the formation of desired conductive properties in a passive film.

  • After high current density load, 904L showed

Acknowledgments

I would like to thank Anette Wallin for her assistance with the electrochemical measurements in the laboratory. Thanks are also due to Thobias Gustafsson, Dalarna University for the assistance with AES measurements and to Claes Olsson at Avesta Research Centre, for fruitful discussions. This project is a part of the Swedish National Programme MISTRA part II, fuel cells for a sustainable society and I would like to express my thanks to the participants in this research programme for their

References (37)

  • C. Zawodzinski et al.

    Metal screen and foil hardware for polymer electrolyte fuel cells

  • D.P. Davies et al.

    Bipolar plate materials for solid polymer fuel cells

    J. Appl. Electrochem.

    (2000)
  • R. Holm

    Electric Contacts

    (1967)
  • T. Hammam, Friction, wear and electric properties of tin-coated contact terminals, Licentiate thesis, Uppsala...
  • F.A. de Bruijn, R.C. Makkus, ECN Netherlands Energy Research Foundation, Petten, Personal communication,...
  • Outokumpu Stainless Corrosion Handbook,...
  • C.R. Clayton et al.

    Passivity of austenitic stainless steels

  • P. Schmuki et al.

    Metastable pitting and semiconductive properties of passive films

    J. Electrochem. Soc.

    (1992)
  • Cited by (97)

    • Corrosion resistance and conductivity of NbN-coated 316L stainless steel bipolar plates for proton exchange membrane fuel cells

      2022, Corrosion Science
      Citation Excerpt :

      The SS corrosion products pollute the catalysts and MEA, resulting in performance degradation of the PEMFC. The formation of a semiconductive passive layer on the SS surface in an acidic environment can protect the SS, but it increases the interfacial contact resistance (ICR) between the gas diffusion layer (GDL) and BP [8], which further increases the energy loss and decreases the energy density of the cell stacks. Hence, multiple modified layers with excellent corrosion resistance and electrical conductivity have been applied to overcome this problem.

    View all citing articles on Scopus
    View full text