Accelerated degradation of bipolar plates in the PEMFC

(1)

Accelerated degradation of bipolar plates in the PEMFC

Accelererad nedbryning av bipolära plattor i en PEMFC

Marcus Eriksson

Master of Science Thesis in

Chemical engineering for energy and the environment KTH, Royal Institute of Technology

Department of applied electrochemistry Stockholm, Sweden, 2016

Supervisor:

Göran Lindbergh Björn Eriksson

Contact person, Sandvik:

Anna Jansson

(2)

Abbreviations

ADT Accelerated degradation test BPP Bipolar plates

CE Counter electrode

CTR Charge transfer resistance COR Carbon oxidation reaction CV Cyclic voltammetry

EIS Electrochemical impedance spectroscopy GDL Gas diffusion layers

GLC Graphite like coating

HER Hydrogen evolution reaction HFR High-frequency resistance HOR Hydrogen oxidation reaction ICR Interfacial contact resistance LFR Low-frequency resistance LSV Linear sweep voltammetry MEA Membrane electrode assembly OCV Open circuit voltage

OER Oxygen evolution reaction ORR Oxygen reduction reaction

PEMFC Polymer electrolyte membrane fuel cell PFSA Polyperfluorosulphonic acid

PTFE Polytetrafluorethylen (Teflon) RH Relative humidity

RHE Reversible hydrogen electrode SEM Scanning electron microscope SS Stainless steel

SU/SD Start-up and shutdown VDC Volts of direct current

WE Working electrode

(3)

Abstract

The aging of bipolar plates in the polymer electrolyte membrane fuel cell was evaluated using different accelerated degradation tests. From previous studies, it is well known that the startups and shutdowns of a polymer electrolyte membrane fuel cell is the primary cause of fuel cell component degradation. Therefore, the aim of these tests was to simulate the number of startups/shutdowns that normally occur during the lifetime of a polymer electrolyte membrane fuel cell, e.g. for automobile applications. The tests were carried out in situ in order to be as realistic and close to operational conditions in real applications as possible. Mechanical stress and degradation normally occurring during polymer electrolyte membrane fuel cell operation was thereby included. The accelerated degradation tests were designed for maximum fuel cell stress, including use of the no purge and the air purge strategy, short cycle duration as well as an increased number of startups/shutdowns. Since the no purge strategy avoids prevention of local H₂/O₂ fronts, this strategy was implemented first.

Other strategies e g the air purge strategy, where air is used to purge the anode, where also implemented. Parameters influencing the tests were varied and the cathode gas was changed between O₂ and air depending on the test. Electrochemical methodology was implemented for the detection of corrosion in the tests and for analysis/ studies of the test results. These techniques include cell voltage/current readings, polarization curves, electrochemical impedance spectroscopy, contact resistance and current density decrease at constant cell voltage. In addition, scanning electron microscope was used to visualize the actual corrosion of the bipolar plates. It was found that implementation of the air purge strategy resulted in localized corrosion, i.e. oxide film formation, on the surface of the bipolar plates increasing both the corrosion resistance and the interfacial contact resistance.

Sammanfattning

Åldrandet av bipolära plattor i polymerelektrolytbränslecellen undersöktes med hjälp av olika accelererade nedbrytningstest. Från tidigare studier är det känt att uppstarterna och avstängningarna av denna typ av bränslecell är den främsta orsaken till nedbrytningen av bränslecellskomponenterna. Av den anledningen var syftet med dessa test att simulera antalet uppstarter och avstängningar som normalt sker under en polymerelektrolytbränslecells livstid t ex för tillämpningar i fordon. Testen utfördes in situ för att vara så realistiska och så nära driftsförhållandena i verkliga tillämpningar som möjligt. Därmed inkluderades mekanisk påfrestning och nedbrytning som normalt sker under drift i en polymerelektrolytbränslecell.

De accelererade nedbrytningstesterna utformades för maximal bränslecellspåfrestning. Detta åstadkoms bl a genom tillämpning av den s k ”no purge”och ”air purge ” strategin, kort cykeltid samt ett utökat antal uppstarter/avstängningar. Eftersom ”no purge” strategin eliminerar förebyggandet av H₂/O₂ fronter, testades denna strategi först. Andra strategier tillämpades också som t ex ”air purge” strategin, där luft fick ”rena” anodsidan. Parametrar som påverkar testen varierades och katodgasen skiftades mellan O₂ och luft beroende på experimentet. Elektrokemisk metodik användes för att detektera nedbryning i de olika testen och för att studera testresultaten. Dessa tekniker omfattade cellspänning/ -ström mätningar, polarisationskurvor, elektrokemisk impedans spektroskopi, kontaktresistens och strömtäthetsminskning vid konstant cellspänning. Svepelektronmikroskopi användes för att visualisera den fysiska nedbryningen av de bipolära plattorna. Det visade sig att ”air purge”

strategin gav upphov till korrosion i form av lokal oxidbildning på de bipolära plattornas yta.

Detta ökade både korrosionsmotståndet och kontaktresistensen i gränsskiktet.

(4)

Introduction

Preface

This master thesis is a collaboration between Applied electrochemistry at KTH the Royal Institute of Technology in Stockholm and Sandvik Materials Technology in Sandviken. Both literature studies and laborations were performed at KTH.

Boundaries

The contents of this thesis deals only with degradation of the polymer electrolyte membrane fuel cell caused by SU/SDs. Degradation of other types of fuel cells is not included. In addition, the experiments were performed in situ and the main focus is on bipolar plate degradation.

Background

The PEMFC is today considered as a promising candidate for use as engines in vehicles due to their high efficiencies, power densities and environmentally friendly operation (Sabir & Li 2004). Transport applications of PEMFCs include the cars and buses from Ballard, DaimlerChrysler, Honda, Nissan, Toyota, Mercedes, Ford, General Motors etc (Pu, 2014). In contrast to the internal combustion engine, no carbon emissions exist assuming that the hydrogen fuel is produced in an environmentally friendly way. Other applications for PEMFCs include the use as stationary and portable power sources. However, the demanding fuel cell environment requires components made of durable materials that resist the acidic conditions without being expensive. Hence, a lot of research is done on improving the durability of materials intended for PEMFC components. Not much investigation has been done on the durability of components such as the bipolar plates of the PEMFC stack. Thus, the degradation of this particular component will be examined in this thesis.

Fig 1. One of many applications for PEMFCs, the Honda Clarity Fuel Cell 2016.

(7)

Objective of the thesis

One of the most important components of the fuel cell stack are the bipolar plates whose main functions include gas distribution, to serve as a current collector and to separate each cell in the stack. Due to the corrosive environment of a fuel cell during operation i.e. increased temperature, high humidity and acidity, the bipolar plates need to have a high resistance to corrosion. Today, metallic BPP’s are most commonly used since these are very thin, provide an increase of the volumetric power density of the fuel cell stack and are cheap to manufacture (Oyarce, 2013). However, not much research has been done on the study of degradation or corrosion of bipolar plates. Normally, most PEMFC components including the bipolar plates degrade during operation. Hence, this thesis focuses on implementing accelerated degradation test strategies or methods to preferentially degrade the bipolar plates in situ. In addition, reproducibility of tests and electrochemical methodology for the evaluation of the tests are further aims of this thesis.

(8)

Polymer electrolyte membrane fuel cell

Working principle

The basic principle of the PEMFC is the conversion of chemical energy to electrical energy using a complex reaction where the reactants H₂ and O₂ undergo several catalyzed reaction steps to eventually form the product H₂O. During the reaction, heat is formed which is then released at the cathode.

H₂ enters at the anode while pure O₂ or air is fed at the cathode. At the anode, H₂ is oxidized to H⁺ ions releasing electrons, which are transported to the cathode through an external circuit. The H⁺ ions are transported from the anode through a polymer electrolyte membrane to the cathode, where they undergo a reaction (Gao, Blunier & Miraoui 2013). At the cathode, O₂ react with H⁺ ions and electrons in a 3-phase zone to form H₂O and heat is released. Only in this 3-phase zone consisting of the reactant gas, the liquid electrolyte and the solid electrode catalyst is the reaction between H₂ and O₂ possible (Hamann, Hamnett & Vielstich 2007).

Anode

H₂ ⇔ 2H⁺ + 2e^- E⁰=0 V vs. RHE Cathode

O₂ + 4H⁺ + 4e^- ⇔ 2H₂O E⁰=1.229 V vs. RHE Total reaction

2H₂ + O₂ ⇔ 2H₂O

Fig 2. Working principle of the PEMFC.

(9)

Detailed reactions

The Hydrogen oxidation reaction (HOR) and the oxidation reduction reaction (ORR) are in fact much more complicated than the basic reactions previously described. The slow kinetics of the ORR is explained by e.g. the transfer of twice as many electrons as in the HOR and by a more complex reaction mechanism involving more reaction steps. Both the HOR and the ORR consist of several steps where adsorption and desorption on the catalyst, i.e. platinum, is involved (Zhang et al. 2013). A complete reaction mechanism of each of these reactions is presented here.

Anode

Pt + H₂ ⇔ Pt-H₂

Pt-H₂ + Pt ⇔ 2 Pt-H (rate determining step) 2 Pt-H ⇔ 2 Pt + 2H⁺+ 2e^- (Zhang et al. 2013) Cathode

Pt + O₂ ⇔ Pt-O₂

Pt-O₂ + H⁺ + e^- ⇔ Pt-O₂H (rate determining step) xPt-O₂H + xe^- + xH⁺ ⇔ xPt-O₂H₂ (Zhang et al. 2013).

xPt- O₂H₂⇔ xPt + xH₂O₂

(1-x)Pt + (1-x)Pt-O₂H ⇔ (1-x)Pt-O + (1-x)Pt-OH (1-x)Pt-O + (1-x)H⁺ +(1-x)e^- ⇔ (1-x)Pt-OH

2(1-x)Pt-OH + 2(1-x)H⁺ + 2(1-x)e^- ⇔ 2(1-x)Pt + 2(1-x)H₂0

In the ORR, x is used to represent the proportion of the respective reactions. In addition, intermediates such as H₂O₂ and OH, not mentioned previously, exist temporarily during the reaction steps of the ORR.

The second step of both reactions is slow and thereby the rate determining step.

(10)

MEA and its components

Each MEA of the PEMFC consists of components such as electrodes, catalysts and polymer electrolyte membranes. However, between each MEA, further components are required such as current collectors, GDLs, BPPs and gaskets. Among these, BPPs are needed for gas distribution, current collection and separation of one cell from another in the stack. A setup of a single cell is shown in figure 3.

Fig 3. The components of a single cell for experimental use.

Bipolar plates

In order to distribute gas, collect current and separate each cell in the stack, BPPs, are placed between cells, see figure 4. Moreover, BPPs remove reaction products, facilitates water and thermal management through the cell. The ideal characteristics of a BPPs material include low surface contact resistance, high corrosion resistance, high electrical conductivity, high mechanical strength, high surface tension, low weight and low cost (H Tawfik et al. 2012).

Today, most BPPs intended for PEMFCs are made of metals such as stainless steels, either coated or bare. Coatings for BPPs are usually made of graphite, but also of metals such as Cr, Ti, Zr or Au. Moreover, BPPs usually consist of a blend of additives of which Cr, Ni and Mo are the most common ones.

(11)

Fig 4. The construction of a PEMFC stack. (See fig. 4)

Sandvik Coated 316 Stainless steel

The material used for the bipolar plates is coated stainless steel manufactured by Sandvik in Sweden. The stainless steel is surface protected by a graphite like coating, which increases the corrosion resistance of the material in order to resist the corrosive fuel cell environment.

The GLC provides the steel with a low ICR. In addition, the 316 SS includes several additives that form protective oxide films in contact with oxygen and increase the corrosion resistance.

The approximate chemical composition of 316 SS is shown in table 1.

Element: Fe Cr Ni Mo Mn Si P C S

wt%: 68.7 17 10 2.1 1.8 0.3 0.04 0.03 0.03 Table 1. Chemical composition of 316 SS. (Sandvik)

Gas diffusion layers

GDLs are used to distribute reactant gases over the electrodes but also to conduct electricity and heat between the electrodes and the BPPs (Oyarce, 2013). The role of GDLs is i.e. to transport reactant gas and electrons (Gao, Blunier & Miraoui 2013).

Degradation and corrosion causes

Startups and shutdowns (SU/SDs)

SU/SDs, of the PEMFC are believed to be the greatest cause of degradation of the components (H Tawfik et al. 2012).

Metal corrosion

Corrosion or oxide film formation of metals is likely to occur at low voltages that exist between SU and SD of the PEMFC. This occurs since low voltages causes different metals to oxidize depending on their electrode potential. The acidic environment of the PEMFC is also likely to cause metals to passivate. Moreover, the release of fluorine ions from the polymer membrane during operation of the PEMFC creates an aggressive environment that could result in the breakdown of passive film of metals and in significant corrosion.

Carbon corrosion

(12)

The fact that both H₂ and O₂ exist at the anode during SU/SDs causes high cathodic potentials, which in turn causes cathodic corrosion or degradation. Corrosion occurs since the kinetics of carbon is much faster at high cathode interfacial potentials (Oyarce, 2013) and may result in degradation of the graphite coating of the BPP.

1) Carbon corrosion at the anode is caused by gross fuel starvation i.e. lack of sufficient H₂ to produce the current that is being drawn.

Lack of H₂ ⇒ anode potential increases ⇒ Carbon corrosion + O₂ evolution at the anode

⇒ Higher anode potential compared to the cathode.

2) Carbon corrosion at the cathode is caused by localized fuel starvation i.e. both H₂ and O₂ exist simultaneously at the anode. This occurs if:

• H₂ is fed into an air-filled anode (pic 1)

• O₂ eventually diffuse into the H₂-filled anode through the anode exhaust (pic 2), the inlet when purging with air (pic 3) or through the membrane due to O₂ crossover (pic 4).

• Local fuel blockage e.g. by water droplets; O₂ may cross over from the cathode side creating local H₂/O₂ fronts (pic 4).

Fig 5. Unwanted situations that cause H2/O2 fronts and possible degradation of PEMFCs. 1) H2 enters the anode inlet, 2) O2

enters through the anode outlet, 3) O2 enters the anode inlet at air purge and 4) O2 crossover from cathode to anode.

Simultaneous existence of H₂ and O₂ at the anode ⇒ High cathode interfacial potentials ⇒ Much faster carbon kinetics ⇒ Carbon corrosion and OER at the cathode (H Tawfik et al.

2012).

Fuel crossover

Fuel, i.e. hydrogen or O₂, may pass through the membrane either from the anode to the cathode during normal operation, or vice versa, during the SDs of the fuel cell operation, see this phenomenon is partially responsible for H₂/O₂ fronts created at the anode, shown in figure 5.

(13)

Release of fluorine ions from the polymer electrolyte

Since it is important that the polymer intended for use as electrolyte in the MEA is durable, the originally polyethylene is perfluorinated and then sulphonated. This includes a side chain which ends with sulphonic acid (Larminie, 2013), see figure 6.

Fig 6. The structure of the sulphonated fluoroethylene PFSA-PTFE copolymer (See figure 6)

The resulting PFSA polymer is durable and resistant to chemical attacks, but also strongly hydrophobic. The last property enables the polymer to drive the product water out of the electrodes and prevents in this from flooding. However, it is also important that it is partially hydrophilic in order to attract water. This is achieved by the clusters of hydrophilic sulphonate side chains, where water is collected (Larminie et al. 2013), represented by gray spheres in fig 7.

Fig 7. The structure of the Nafion membrane materials including clusters of hydrophilic sulphonate side chains. (See figure 7)

This results in that Nafion and other fluorosulphonate ionomers are chemically highly resistant, mechanically strong, acidic, able to absorb large quantities of H₂O and good proton conductors if well hydrated due to free movement of protons within the material (Larminie et al. 2013).

However, at acidic conditions during the operation of a PEMFC, aggressive F^- ions tend to be released from the polymer chains and transported to e.g. the BPPs where they cause corrosion This occurs due to the attacking of peroxy- or hydroxy radical on polymer

(14)

endgroups with residual H-containing terminal bonds (Yang et al. 2011). Suggested reaction steps of a hydroxy radical attack on the Nafion polymer is shown here (Healy et al. 2004).

R-CF2COOH + •OH → R-CF2• +CO2+H2O R-CF2• + •OH → R-CF2OH → R-COF + HF R-COF + H2O → R-COOH + HF

Corrosion reactions and mechanisms

• Carbon corrosion

of the graphite coating may occur, see figure 8.

C + 2 H₂O → CO₂ + 4H⁺ + 4e^- E⁰=0.207 V vs. RHE C + 2 H₂O → CO + 2H⁺ + 2e^- E⁰=0.518 V vs. RHE

Fig 8. Possible oxide film degradation and carbon corrosion

of the BPP. Fig 9. Oxide film breakdown and bulk corrosion.

• Pitting corrosion

of the 316 SS due to an acidic environment and low voltages that cause oxidation of Cr and Ni. Since Cr is the most protective additive of a SS, oxidation of Cr will result in a significant decrease of the corrosion resistance thus enabling BPP degradation.

At acidic conditions, F^- ions released from the polymer may be transported to the BPP and cause pitting corrosion. F^- ions may either cause aggressive environments which break down the passive films formed by the additives of the SS, or participate in the corrosion reaction themselves. Some possible reactions and mechanisms follow here.

Example, chromium corrosion mechanism

1) Cr → Cr²⁺ + 2e^- E⁰=-0.74 V vs. RHE 2) Cr²⁺ + 2H2O → Cr(OH)2 + 2H⁺

3) H⁺ + F^- → HF 4) Oxide film removal

5) SS more susceptible to corrosion

Ni → Ni²⁺ + 2e^- E⁰=-0.25 V vs. RHE Fe → Fe²⁺ + 2e^- E⁰=-0.44 V vs. RHE

If the graphite coating is degraded the corrosion resistance of the BPP is significantly reduced.

In this case, the only corrosion resistance left is that of the protective oxide film that is formed on the bulk of the SS, see figure 9.

(15)

In conclusion, degradation of the BPP may for instance occur due to:

1. Oxide film formation at acidic conditions.

2. Oxidation of metals at low voltages.

3. Release of F^- ions from the polymer membrane creating an aggressive environment and causing passive film breakdown.

4. High cathodic overpotentials that increase the rate of carbon corrosion.

5. Pitting corrosion, i.e. localized corrosion.

Theory

Irreversible losses

Degradation of the BPP occurs mostly due to electrode potentials at which the additive metals oxidize, e g -0.74 and -0.25 V for Cr and Ni respectively. Depending on the potential of the cell certain reactions can take place. However, irreversible losses, e.g. ohmic losses in form of increased ICR, may cause degradation as well. These irreversible losses occur mostly during the SUs/SDs of the fuel cell.

The cell voltage, E, is lowered by different types of irreversible losses, of which especially ohmic losses contribute to the degradation of the fuel cell components. The theoretical potential E_eq,is higher than the real potential, since it does not include irreversible losses.

Hence, when a PEM fuel cell is activated for the first time, the potential is less than the theoretical one.

E=Eeq - ηcross - ηHOR - ηORR - ηohm - ηtx,gas (Eq. 1) where:

Eeq = Eeq0 + ^{𝐑𝐑𝐑𝐑}_{𝟐𝟐𝟐𝟐} ln�𝐩𝐩(𝐇𝐇𝟐𝟐)∗𝐩𝐩(𝐎𝐎𝟐𝟐)^{𝟏𝟏/𝟐𝟐}

𝐩𝐩(𝐇𝐇𝟐𝟐𝐎𝐎) � theoretical potential or reversible OCV (Eq. 2)

Eeq0 = ^−𝐆𝐆_{𝟐𝟐𝟐𝟐} standard potential (Eq. 3)

ηcross crossover losses

η_HOR ,ηORR activation losses

η_ohm ohmic losses

η_tx,gas mass-transport losses

Activation losses occur due to slow reaction kinetics. Since the cathode reaction, ORR, is much slower than the anode reaction, HOR, the activation losses for the anode reaction can be considered as negligible.

Mass-transport losses occur when reactants are consumed faster than it takes for them to reach the catalyst surface. As previously mentioned, lack of reactants causes e.g. carbon corrosion and oxygen evolution. In addition, pure reactant gases have less mass-transport losses than reformate and air.

Ohmic losses occur due to electrical and ionic resistances in the fuel cell. The ohmic losses are proportional to the current density, meaning these are highest at steady operation of the fuel cell i.e. when a load is drawn from the cell, rather than during activation or deactivation of the fuel cell. The relation between current density and resistance is given by Ohm’s law.

(16)

η_ohm= iRcell where:

Rcell = Rmem + RH+ ,eff + RCR + Re (Eq. 4)

total area specific resistance of the PEMFC

Rmem ionic resistance of the membrane (not constant)

RH+ ,eff ionic resistance of the electrodes

R_CR electrical contact resistance between catalyst layer- GDL and GDL-BPP

Re electrical bulk resistances in the electrodes, GDLs and BPPs

Resistances in BPPs

As previously mentioned, an ideal BPP material would have a low interfacial contact resistance, a high corrosion resistance and a high electrical conductivity. These properties are explained in this section. Both ICR and corrosion resistance is affected by the formation of surface oxides on the BPP. Hence, the formation of surface oxides has an effect on the lifetime and performance of the BPP (Tian et al. 2011). In addition, surface oxides decrease the surface conductivity i.e. less current is able to flow on the surface of the BPP and consequently ICR is increased (Tian et al. 2011).

Contact resistance BPP-GDL

Contact resistance is an electrical property of a material and may simply be defined as the ratio between potential and current. The ICR between BPP and GDL is a part of the total ohmic losses, Increases in the ICR may occur due to the formation of oxide films that protect the material against corrosion. However, increases in the ICR may also occur due to decreases in oxide film thickness. The protective oxide film increases the corrosion resistance of the BPP.

Z = | ^𝑬𝑬_𝒋𝒋| ^{(Eq. 5)}

Fig 9. Contact resistance between BPP and GDL (See fig 6).

(17)

The surface topography of the GDL and BPP in contact and the formation of surface oxides, determines the ICR. Due to roughness features, the contact area between GDL-BPP is decreased and current is limited to flow only through the contact asperities. Consequently, there is a voltage drop across the interface (Kraytsberg et al. 2007). Current lines, where the current is able to flow, are created in the bulk of each component, see figure 9.

It follows that the ICR is also dependent on the clamp pressure since the contact area between BPP-GDL increases with increasing clamp pressure, (Tian et al. 2011). The better the contact between BPP and GDL, the lower the ICR.

Corrosion resistance

The resistance of a material or metal against reaction with unfavorable elements, either gases or liquids, that may corrode the material is known as the corrosion resistance. A material’s corrosion resistance may be increased e.g. by the addition of additives or coatings since these form protective oxide films when attacked. Sandvik’s graphite coated 316 SS BPPs are protected by a graphite coating, which increases its corrosion resistance and lowers its ICR, along with a variety of additives.

Rcorr = ^𝑬𝑬_𝒊𝒊^{𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄}

𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄 (Eq. 6)

E_corr= corrosion potential i_corr= corrosion current density

The corrosion resistance of BPP’s can be measured in several ways including, amount of released ions from the BPP to the electrolyte, potentiostatic and potentiodynamic measurements (Dundar et al, 2010). However, potentiostatic and potentiodynamic methods are the most commonly used methods for the detection of changes in the corrosion resistance of metallic BPPs.

Potentiodynamic polarization curves generated using potentiodynamic scans provide information on the different passivation regions of the SS in the BPP i.e. active, passive and transpassive regions. The formation of oxide films and corrosion is dependent on the region and in turn on the voltage. The breakdown of oxide film usually occur at high voltages, around 0.65 V, in the transpassive region or due to a highly acidic environment. If oxide film removal occurs and the BPP corrodes, the ICR is likely to decrease (H Tawfik et al. 2012).

Graphite does not seem to have any passive or transpassive regions (H Tawfik et al. 2012) and may for that reason impede the corrosion of Sandviks graphite coated 316 SS. However, if the graphite coating is degraded e.g. through the COR, the passive and transpassive regions may be valid for the SS.

Corrosion rate

May be defined as the rate at which an element or metal corrodes in a specific environment.

Since the corrosion rate is proportional to the corrosion current density, an increasing corrosion current density is equals an increasing corrosion rate. The relation between the two is given by the equation:

r = ^{𝒊𝒊∗𝒂𝒂}_{𝒏𝒏𝒏𝒏} ^{(Eq. 7)}

(18)

Measures for BPP degradation

Fig 10. Potentiostatic polarisation curve showing the different zones of corrosion i e the active, passive and transpassive zones.

(See figure 5)

The goal of the ADTs will be to access the active zones in the potentiostatic polarization curves of Cr especially, but also of other metal additives such as Ni. This zone is usually found at negative voltages where oxidation of these additives takes place. Since Cr is the most protective additive in SS, the corrosion resistance will be significantly reduced and the SS will be susceptible to corrosion once Cr is oxidized.

From the theory information yielded so far, it is possible to theoretically determine some measures or guidelines that are likely to degrade the BPPs. Hence the following measures are proposed:

• Perform ADTs in order to corrode e g GLC and then the Cr additives ⇒ Lowered corrosion resistance and increased ICR

• Maximum possible RH and temperature ⇒ facilitates corrosion

• High current density ⇒ increased corrosion rate according to eq. 7

• Intermediate voltages e g 0.2 or 0.5 V ⇒ carbon corrosion of GLC

• Low voltages e g -0.74 or -0.25 V ⇒ oxidation of Cr and Ni

• Low pH by utilizing an anodic BPP placement ⇒ facilitates corrosion

• Extend the time at which oxidation of Cr and Ni occurs

• Decrease the clamp pressure ⇒ Increased ICR

(19)

Methodology

Techniques for the detection of degradation in BPPs

Cell potential/current density measurements using multimeter and potentiostat

The current density and cell voltage between the BPP and GDL will be measured using a multimeter. From these measurements, the ICR will be obtained. Increases in ICR will reveal possible BPP degradation. In addition, current density and cell voltage measurements using a potentiostat provide information on the development of each cycle of the strategy. Low cell voltages below zero at a drawn current will thereby be monitored.

Polarization curves

For the detection of irreversible losses, polarization curves will be recorded using an Autolab PGSTAT302N potentiostat unit in combination with a BSTR10A booster. E.g. the more mass transport losses, the greater the changes in the morphology and the higher the degradation.

However, it may be difficult to determine whether the losses are caused by degradation of the intended component or not. For that reason, an additional performance test, performed by replacing the degraded MEAs with fresh ones, may be needed, according to previous studies (K. Eom et al. 2012).

Electrochemical Impedance Spectroscopy, EIS

The contributors to the impedance of a PEMFC are interfacial charge-transfer resistance, membrane resistance, contact resistance, oxidant and fuel transport resistance, water transport resistance, proton transport resistance, double-layer capacitance, and Faradaic pseudo-capacitance. EIS is able to identify the individual contributors to the fuel cell performance including ohmic resistance, interfacial charge transfer resistance, mass transport resistances in the catalyst layer and back-diffusion layer in a short period of time (Yuan et al.

2010).

Electrochemical impedance may simply be defined as resistance or the ratio between potential and current. Hence, EIS gives information on the contact resistance of a material.

Different kinds of plots may be obtained using EIS, the most common ones being Nyquist plots, see figure 10, and Bode plots. Both these plots clearly show any failures that may have occurred in coatings due to degradation.

Fig 10. Nyquist plots of the resistance in a fuel cell including charge transfer resistance, membrane resistance and CPE which represents the double layer capacitance. Nyquist plot of different charge transfer resistances (left figure) and of different values of the constant n (right figure). (See fig. 8)

The total area-specific resistance of the PEMFC, R_cell,consists of the ionic resistance of the membrane and electrodes, electrical bulk resistances in the electrodes, GDLs and BPPs plus electrical ICR between catalyst layer-GDL and GDL-BPP. See equation 4.

(20)

R_cell can be estimated using the high-frequency resistance (HFR) from EIS measurements (Oyarce, 2013). EIS will be employed for the comparison of the cell resistance before and after each test period. If the ICR has increased after the test period, the corrosion resistance may have decreased and corrosion may have occurred. This technique will be used mainly to measure the ICR between BPPs and GDLs in the PEMFC.

Impedance and Nyquist plots are obtained according to:

Z = ^𝑬𝑬_𝑰𝑰 = Z0 (cosΦ + j sinΦ) ^{(Eq. 8)}

where:

E ^potential I ^current Z0 magnitude

Φ phase shift

Scanning Electron Microscopy, SEM

SEM will be employed to identify differences in the morphology of the coated 316 SS in order to detect any degradation or corrosion of the material.

The SEM technology uses an electron gun that emits an electron beam towards a sample. The electron beam passes through several electromagnetic lenses, two condenser lenses and one objective lens, whose function is to reduce the crossover diameter of the electron beam and to focus the electron beam as a probe with the diameter of a few nanometers.

A detector collects and amplifies the signal electrons emitted from the sample and reconstructs an image on the screen of a cathode ray tube or a liquid crystal display. SEM can provide image magnification from 20 x to more than 100 000 x and effective magnifications of 20 000x (Leng 2013). The resolution of this microscopy is from 10 nm upwards, which makes it suitable for the detection of changes in the morphology of a BPP material.

Accelerated degradation tests, ADTs

Since a real test of a PEMFC would take thousands of hours to perform, and consequently be both expensive and time consuming, shorter tests simulating the total amount of SUs/SDs during the lifetime of a PEMFC can be done in-situ. These are one type of accelerated degradation tests or ADTs and usually consist of cycles of several hundreds of SUs/SDs using different shutdown strategies.

In order to degrade coated metallic BPPs in PEMFCs, different ADT procedures or strategies are evaluated. These are based on the most corrosive strategies in a previous study on the degradation of PEMFC electrodes (Oyarce, 2013). Since the simultaneous existence of H₂ and O₂ at the anodic compartment also affects the degradation of the BPPs, the previous strategies may be applicable in this thesis. Repeated hydrogen/air cycling during SUs/SDs has a tendency of being detrimental to the stability of the passive film of the BPPs. For this reason, SUs/SDs cycles are highly recommended for material qualification testing of BPPs (Hinds &

Brightman, 2015).

The proposed strategies and a list of proposed parameters values are shown in the degradation strategies section below.

(21)

Testing conditions/parameters

To simulate real automobile or portable applications, some general operating parameters, such as the operating temperature, are kept almost constant. The parameter values that will be used in the suggested ADT strategies are listed in Tables 2-3.

General

Parameter Description Value

T Operating temperature 70-80 C

RH Relative humidity 0-100%

I Current 1-3.5 A

i Current density 0.14 – 3.5 A/cm²

P Clamp pressure 1-4 bar

Constant Description Value

A Area of bipolar plates 7 cm²

Table 2. General parameter values for the proposed strategies.

A current of 3.5 A will be drawn from the fuel cell to simulate real operating conditions. This value is calculated from the area of the BPP and the proposed current density. The relative humidity at a specific temperature is calculated from equation 9.

RH = 𝐩𝐩(𝐬𝐬𝐬𝐬𝐬𝐬,𝐑𝐑)^{𝐩𝐩(𝐇𝐇𝟐𝟐𝐎𝐎)} * 100 ^{(Eq. 9)}

The volumetric flow rates were obtained from Faraday’s law in combination with the ideal gas law, shown in equation 10 and 11 respectively.

𝐧𝐧

𝐬𝐬 = _{𝐳𝐳𝟐𝟐}^𝐈𝐈 ^{(Eq. 10)}

𝐧𝐧 =_{𝐑𝐑𝐑𝐑}^{𝐩𝐩𝐩𝐩} ^{(Eq. 11)}

⇒ ^𝐩𝐩

𝐬𝐬 = ^{𝐈𝐈 𝐑𝐑 𝐑𝐑}_{𝐳𝐳 𝟐𝟐 𝐩𝐩}∗ λ ^{(Eq. 12)}

where:

I current R gas constant T temperature

z number of electrons F Faraday constant p gas pressure λ stoichiometry Polarization curves and EIS

Description Value

Anode gas H₂

Anode flow rate 540 ml/min or 162 ml/min

Anode stoichiometry 1 or 3.33

Cathode gas O2 or Air

Cathode flow rate 374 ml/min or 540 ml/min

Frequency 1 Hz – 100 kHz

Amplitude 10 mV

Table 3. Parameter values for the generation of polarization curves and EIS in the proposed strategies.

(22)

The cathodic flow rate is equals to 374 ml/min when pure O₂ is used as cathode gas. The corresponding anodic flow rate is set to 540 ml/min. When air is used as cathode gas the flow rate is 540 ml/min, and an anodic flow rate of 162 ml/min is set for pure H₂. In this last case, the anodic stoichiometry was varied between 1 to 3.33. This was done to stabilize and maintain a high cell voltage when current was drawn from the cell. At stoichiometry 1 the cell voltage dropped quickly to below zero when current was drawn. To void this potential drop, the stoichiometry was increased from 1 up to 3.33.

Degradation strategies

The following strategies are evaluated for the corrosion of the BPPs. In each of these strategies, all the start-ups are unprotected i.e. H2 is fed to an air filled anode at the same time as air is fed to the cathode. Since BPP degradation requires maximum corrosion, these strategies are suggested.

No purge strategy

The no purge strategy is the most harmful to the fuel cell performance since H₂/O₂ fronts are created during the SU and the long SD phase. This causes maximum carbon corrosion and MEA degradation. However, the idea of this strategy is, in particular, to cause BPP degradation by integrating long load phases and cycle times.

1. Fuel supply turned on and run cell at OCV 20 s SU

2. Run cell at 0.5 A/cm² 40 s Load

3. Shut off both fuel and air supply (open exhausts) 820 s SD

4. Purge anode with air 30 s Purge

Total 910 s / cycle

Cycle repeated 230 times.

Air purge strategy

A more traditional strategy utilized by automotive companies and other fuel cell developers. It includes an extra OCV phase and a relatively long air purge phase that removes residual H₂. Carbon corrosion and degradation of the MEA is significantly reduced. The SU/SDs are more intensified because of its decreased cycle time. In addition, more H₂/O₂ fronts will be created due to the extended final purge phase of the strategy. Furthermore, the clamp pressure is lowered somewhat.

3. OCV 10 s Idle

4. Purge anode with air 50 s SD/Purge

Total 110 s / cycle

(23)

New ADT 3

Since the air purge strategy, characterized by short cycle times and a relatively long air purge phase, resulted in a higher increase in ICR, these durations were further investigated in the new ADTs. Furthermore, the no purge strategy incorporating half cycle times also resulted in an increased ICR. In the new ADT3 strategy, even shorter cycle times were applied to achieve more SU/SDs during the same time as in previous strategies and thereby cause increased degradation of the BPPs. However, in order to shorten cycle time, the air purge and load phases also have to be shortened. Consequently, there will be less H₂/O₂ fronts and less degradation caused by drawn currents.

3. Shut off both fuel and air supply (open exhausts) 20 s SD

Total 60 s / cycle

New ADT 1

A new strategy developed for this project only. It utilizes shorter phase times in order to maximize BPP degradation. It features an extra OCV phase, to stabilize the cell voltage, and decreased phase times. The whole idea of this strategy is to shorten the cycle time and phase times in order to intensify the SU/SDs even more than in previous strategies while featuring all the phases of the air purge strategy. This is, however, done at the cost of shorter SD and load phase times meaning there will be less H₂/O₂ fronts and degradation. For this strategy, the clamp pressure was lowered further.

3. OCV 10 s Idle

4. Purge anode with air 20 s SD/Purge

Total 70 s / cycle

Equipment

Task Equipment specification

Polarization curves, Electrochemical impedance spectroscopy (EIS)

Autolab PGSTAT302N Potentiostat Autolab BSTR10A Booster

SEM Zeiss Sigma Gemini FE SEM

Gas flow rates control Brooks Instruments 5850E Mass flow controller NI USB-6008 data acquisition devices

LabVIEW software

Humidifier and purge control ASCO BS 5501 Computer controlled solenoid valve NI USB-6008 data acquisition devices

LabVIEW software

VDC Circuit control Eurotherm LR59793 Solid state relay Load for shutdown strategies Autolab PGSTAT302N Potentiostat

LabVIEW software Table 4. Equipment specification for the techniques in the proposed strategies.

(24)

Fuel cell components

Component Manufacturer/Designation

BPP Sandvik graphite coated 316 SS

MEA Gore Primea MESGA 3L

GDL Sigracet 25BC

Gasket Fiberflon PTFE coated glass fabric 628.25 0.22 mm

Table 5. Fuel cell components specification.

Software

LabVIEW software was used for the programming of the accelerated degradation test for the PEMFC. In addition, NOVA software was used to control the potentiostat, for activation procedures and to generate polarization curves. Matlab was used to obtain plots of contact resistance, potential vs time and potential vs cycle number together with additional degradation plots.

Experimental

Tests were carried out employing a single cell consisting of the fuel cell assembly, GDLs, BPP, MEA and gaskets. The components are specified in table 5.

Components

BPP

The BPPs used in the experiments are circular, have an area of 7 cm², serpentine flow field design, figure 11, and are made of Sandvik’s graphite coated 316 SS, as previously mentioned.

Fig 11. Serpentine flow field design of the BPPs from Sandvik. Fig 12. The fuel cell setup used for the experiments.

(25)

GDL

The GDLs are also circular in form and have an area of 7 cm² in order to match the BPPs. GDLs used in these experiments are made of carbon.

Fuel cell setup and assembly

The fuel cell setup used for the experiments, shown in figure 12, is equipped with one BPP and one GDL, either at the anode or at the cathode. A current collector is inserted at the center of each endplate, perpendicular to the plane. The GDL and the BPP are in contact and placed between the current collectors, figure 13.

Fig 13. Experimental setup for measurements of the in-situ ICR. (See fig.11)

Experimental setup

Specifications

An experimental setup was installed and a computer program was programmed in order to control and automate the ADTs for the fuel cell. The experimental setup had to fulfill the following specifications.

Able to:

• Set the RH of the humidifiers through humidifier controllers

• Set and measure the temperature of the entering gases and of the cell using a temperature regulator

• Apply loads and measure current/cell voltage using a potentiostat

• Measure current/cell voltage between GDL and BPP of the cell using a multimeter.

• Supply the cell with fuel and purge gas using tubes

• Adjust the clamp pressure

• Adjust the pressure of all entering gases

• Shift between different fuel gases

The software program that was developed in LabVIEW had to fulfill the following specifications.

(26)

Able to:

• Repeat a strategy sequence, a cycle, the no of times needed for each test.

• Adjust the phase times of the cycle in each strategy

• Adjust the mass flow controllers to control the amount of entering gas

• Turn on/off the solid state relays to turn on/off fuel and purge

• Control the potentiostat to apply loads on the cell at a certain time in the cycle

• Read and display the potentiostat measurements

• Read and display the multimeter measurements

• Write files, for potentiostat and multimeter measurements including date/time

Fig 14. Schematic of the experimental setup for the accelerated degradation strategies.

A probe is inserted in the cathodic current collector and another one is inserted through the fuel cell assembly making good contact with the BPP. A multimeter is then connected to the probes to enable measurement of the ICR between the GDL and the BPP. The probes consist of platina wires with a length of a couple of centimeters and a diameter of approximately 1 mm. Measurements of the ICR between the BPP and GDL are enabled according to figure 15.

Fig 15. Schematic of the setup for ICR measurements.

A thermocouple is inserted in the cathodic current collector to enable measurement of the temperature in the fuel cell. The thermocouple consists of a tin wire with similar length as the probes but with a diameter of approximately 2 mm.

(27)

In addition, cables are connected from a potentiostat to the anodic and cathodic current collectors to serve as WE and CE respectively, see figure 16. This enables both current loads to be drawn from the fuel cell as well as cell voltage/current measurements.

The SU/SDs were automated including potentiostat and multimeter measurements of each cycle. This was programmed in LabVIEW, the program is shown in figure 15. The program was used for all the strategies i.e. the no purge strategy, the air purge strategy, the new ADT3 strategy and the new ADT1 strategy.

Fig 16. Program developed in LabVIEW for accelerated degradation testing of the PEMFC.

List of performed experimental tests using open exhausts

These were the eight different experiments that were performed including the test conditions.

Exp Fuel

No Flow

rates (ml/min)

# of cycles, BPP location

j (A/

cm²) Gas pressure (bar)

Clamp pressure (bar)

RH (%) T

(C) Time 3^rd phase (s)

Purge 4^th phase No

purge

1 H2/O2 540/374 230 c 0.5 1.5/1.5 4 100 80 820 Air 2 H₂/O₂ 540/374 230 c 0.5 1.5/1.5 4 100 80 820 Air 3 H₂/O₂ 540/374 460 c 0.5 1.5/1.5 3 100 80 410 -

Air purge

4 H2/Air 162/540 2000 c 0.5 1.5/1.5 3 80 70 0 Air 5 H₂/Air 162/540 2000 a 0.5 1.5/1.5 3 80 70 0 Air

ADT 3

6 H₂/Air 162/540 4000 a 0.5 1.5/1.5 3 80 70 20 - 7 H₂/Air 162/540 3260 a 0.14 1.5/1.5 3 88 70 20 -

ADT 1

8 H₂/Air 162/540 3000 a 0.5 1.5/1.5 1 80 70 0 Air Table 6. Test conditions used in the implementation of the no purge strategy experimental tests.

Accelerated degradation of bipolar plates in the PEMFC