**1. Introduction**

20 Will-be-set-by-IN-TECH

304 Mass Transfer - Advanced Aspects

Yasuda, I., Matsuzaki, Y., Yamakawa, T. & Koyama, T. (2000). Electrical conductivity and

*Ionics* 135(1-4): 381–388.

mechanical properties of alumina-dispersed doped lanthanum gallates, *Solid State*

The performance of Proton Exchange Membrane Fuel Cells (PEMFC) and Electrolyzers (PEME) is subject to mass transport limitations. Within this chapter we will discuss the origination of those limitations and the current research efforts for mitigation. Hydrogen powered fuel cells operate based on the reaction of hydrogen and oxygen, (Figure 1) where the anode reaction is found in Eq. 1, the cathode reaction in Eq. 2 and the overall reaction in Eq. 3. The reverse of this reaction (Eq. 4) is electrolysis. Where, in the electrolyzer the anode reaction is Eq. 5 and the cathode reaction is Eq. 6.

$$\text{H}\_2 \xrightarrow{} 2\text{H} \text{\textquotedblleft} + 2\text{e} \qquad \text{E} \equiv 0\text{V} \tag{1}$$

$$\text{H}\_2\text{O}\_2 + 2\text{H} + 2\text{e}\_2 \xrightarrow{} \text{H}\_2\text{O} \qquad \text{E} = 1.229\text{V} \tag{2}$$

$$\text{H}\_2 + \text{V} \text{O}\_2 \xrightarrow{} \text{H}\_2\text{O} \qquad \text{E} = 1.229 \text{V} \tag{3}$$

$$\text{H}\_2\text{O} \xrightarrow{\Delta} \text{H}\_2 + \text{H}\_2\text{O}\_2 \qquad \text{E} = \text{-1.229V} \tag{4}$$

$$\text{H}\_2\text{O} \xrightarrow{} \text{V}\_2\text{O}\_2 + 2\text{H}^\* + 2\text{e}^\cdot \qquad \text{E} = \text{-1.229V} \tag{5}$$

$$2\text{H}^\* + 2\text{e}^\cdot \rightleftharpoons \text{H}\_2 \qquad \text{E} = 0\text{V} \tag{6}$$

Basic cell construction is very similar for both PEMFC and PEME. During electrolysis a voltage is applied to the cell while an ion conductor with electrocatalyst layers, such as Pt black on Nafion®, is used to split water into hydrogen and oxygen, as in Figure 2. As water is split into hydrogen and oxygen ions at the anode, the hydrogen ions travel across the PEM and oxygen is collected and exhausted at the bipolar plate. At the cathode, hydrogen ions recombine to create diatomic hydrogen, which can be then be stored for later use. The cell components are similar to those used in a PEM fuel cell, but different bipolar plates must be used due to the corrosive environment. PEMFCs typically use graphite bipolar plates that will degrade under the conditions used in a PEME. Corrosion resistant bipolar plates are substituted for graphite. Titanium plates are typically used, but are very expensive. Stainless steel bipolar plates have also been used, but there is a risk of leaching iron into the water, which would affect the performance of the catalysts and the membrane.

Mass Transport Limitations in Proton Exchange Membrane Fuel Cells and Electrolyzers 307

Figure 3 (Yeager and Steck 1981). In this model, the structure of Nafion® is represented in three separate regions. Zone A is the fluorocarbon based backbone of the polymer. Zone C represents the ionic clusters, where ion transport occurs via either a vehicular motion or through the Grothaus mechanism. Zone B is representative of the interfacial region between A and C, consisting mostly of sulfonated ether side chains of the fluorocarbon backbone.

Fig. 3. Structure of Nafion®. Reproduced from (Yeager and Steck 1981)

transport, or the availability of fuel supply for the reaction.

base catalyst.

The solid polymer electrolyte is in contact with the catalyst layer. Typically, the catalyst layer consists of a carbon supported Pt based catalyst mixed with ionomer (typically similar materials as the polymer electrolyte). The catalyst electrode must provide channels for the transport of reactants and products, and electrically conductive path for the transport of electrodes from the electrochemical reaction and an ion conductive path for the transport of protons from the electrode to the membrane. As it is shown in Figure 4, the electrode must have a balance in order to avoid performance losses and maximize the utilization of the Pt

The theoretical open circuit voltage for a PEMFC with a pure hydrogen feed is 1.23V. However, actual performance of the fuel cell is considerably lower due to cell resistances, slow reaction kinetics and gas transport limitations. At potentials above 0.9V, losses are attributed slow reaction kinetics at the cathode. Between 0.9 and 0.5V, internal cell resistances govern the incurred losses, while below 0.5V losses can be attributed to gas

PEME are operated at higher potentials in order to drive the electrolysis reaction of water. During operation of a PEME in the voltage range below approximately 1.4 V the cell is

Fig. 2. Schematic of a PEM electrolyzer

Typically, the electrolyte is a solid polymer electrolyte, such as Nafion®, a sulfonated polytetrafluoroethylene based ionomer. One of the most widely sited structures in found in

e-

Anode Electrolyte Cathode

O2

H2O

H2

H+

e-

Anode Electrolyte Cathode

H+

e-

Typically, the electrolyte is a solid polymer electrolyte, such as Nafion®, a sulfonated polytetrafluoroethylene based ionomer. One of the most widely sited structures in found in

H2

Fig. 1. Schematic of a PEM fuel cell

H2O

O2

Fig. 2. Schematic of a PEM electrolyzer

Figure 3 (Yeager and Steck 1981). In this model, the structure of Nafion® is represented in three separate regions. Zone A is the fluorocarbon based backbone of the polymer. Zone C represents the ionic clusters, where ion transport occurs via either a vehicular motion or through the Grothaus mechanism. Zone B is representative of the interfacial region between A and C, consisting mostly of sulfonated ether side chains of the fluorocarbon backbone.

Fig. 3. Structure of Nafion®. Reproduced from (Yeager and Steck 1981)

The solid polymer electrolyte is in contact with the catalyst layer. Typically, the catalyst layer consists of a carbon supported Pt based catalyst mixed with ionomer (typically similar materials as the polymer electrolyte). The catalyst electrode must provide channels for the transport of reactants and products, and electrically conductive path for the transport of electrodes from the electrochemical reaction and an ion conductive path for the transport of protons from the electrode to the membrane. As it is shown in Figure 4, the electrode must have a balance in order to avoid performance losses and maximize the utilization of the Pt base catalyst.

The theoretical open circuit voltage for a PEMFC with a pure hydrogen feed is 1.23V. However, actual performance of the fuel cell is considerably lower due to cell resistances, slow reaction kinetics and gas transport limitations. At potentials above 0.9V, losses are attributed slow reaction kinetics at the cathode. Between 0.9 and 0.5V, internal cell resistances govern the incurred losses, while below 0.5V losses can be attributed to gas transport, or the availability of fuel supply for the reaction.

PEME are operated at higher potentials in order to drive the electrolysis reaction of water. During operation of a PEME in the voltage range below approximately 1.4 V the cell is

Mass Transport Limitations in Proton Exchange Membrane Fuel Cells and Electrolyzers 309

kinetically limited and the current increases exponentially with the cell potential. Between 1.4 V and 1.7 V the cell is transitioning to a mass transfer limited mode of operation. Above 1.7 V, the cell current is completely limited by the diffusion rate of water across the membrane and further increases in the cell voltage do not result in higher cell current. The steady state current that is reached above 1.7 V is known as the mass transfer limited current density. At the mass transfer limiting current density, the rate of water diffusion across the membrane minus the rate of electroosmotic drag is equal to the reaction rate of water at the

In general, the net water flux occurs from the anode to the cathode and higher water content is related to higher performance (Falcao, Rangel et al. 2009). Thus, the influence of water content at the cathode has a higher impact than the water content at the anode. At lower humidification levels, the hydrophilic fraction of the membrane, where the water travels, decreases and overall membrane permeation becomes limited by water diffusion (Majsztrik,

Figure 4 and 5 shows the representative fuel cell performance outlining the different losses arising from the different components. At low current densities the losses are dominated by the the activation polarization, which occur at the cathode under operation with clean hydrogen. The losses are followed by the ohmic resistance, which is mostly attributed to the solid electrolyte. Finally at high current densities, the performance is limited by the mass transport of reactancts and products. Semi-empirical approaches have been used to predict and analyze the fuel cell performance. Such an approach is the one by (Pisani, Murgia et al.

( )

*N N E EI K b <sup>a</sup>*

where *Vcell* is the cell potential, *E*0 is the standard cell potential, *Rcell* cell resistance, *I* is the current density, *b* is the Tafel slope, *<sup>l</sup> I* cell current density at the limiting current

proportionality constant, <sup>0</sup> *Nd* is the diffusion mechanism parameter at the zero current

exponent of the species in the Butler-Volmer equation, *<sup>l</sup> Nd* is the diffusion mechanism

In real life operation, the use of pure fuel and oxidant gases results in an impractical system. A more realistic and cost efficient approach is the use of air as an oxidant gas and hydrogen from hydrogen carrier molecules (i.e., ammonia, hydrocarbons, hydrides). The short and long term effect of impurities in these gases may have an overriding effect on the fuel cell performance. Common atmospheric impurities in the cathode gas stream that have an effect on the performance of the fuel cell include SO2, NO2, H2S, O3 (Veldhuis, deBrujin et al. 1998). Even though the hydrogen oxidation reaction occurs at higher rates than the oxygen reduction reaction at the cathode (Sukkee, Wang et al. 2000), the effect of hydrogen

β

= =+ = =

*<sup>I</sup> V E RIb I a S*

<sup>1</sup> <sup>0</sup>

<sup>0</sup> ln ln 1 *<sup>l</sup>*

α β

⎛ ⎞ ⎜ ⎟ ≈− − + − ⎜ ⎟

1

⎛ ⎞ − − ⎜ ⎟ ⎝ ⎠

μ

⎝ ⎠

*d d c F c F*

γ

 αβ

is an empirical constant, *E* is the potential, *K* is the

*<sup>F</sup>* is the Faraday constant,

*l*

*I*

+ +

0

*I I*

*l*

γ

is the kinetic

2002). where the performance curve can be represented by:

*cell cell*

( )

μ

*c* is the cathode transfer coefficient,

0

density, *S* flooding parameter,

parameter at the limiting current density.

anode.

with

density,

α

Bocarsly et al. 2008).

Fig. 4. Schematic representation of the catalytic layer. (A) where at low Nafion content not all the catalyst particles are connected to the membrane for ionic conduction (B) the optimal Nafion contentwhere there is good ionic and electronic conduction for all the catalyst particles. (C) When there is too much Nafion and not all of the catalyst particles are electronically connected to the diffusion layer. Reproduced from (Passalacqua, Lufrano et al. 2001)

kinetically limited and the current increases exponentially with the cell potential. Between 1.4 V and 1.7 V the cell is transitioning to a mass transfer limited mode of operation. Above 1.7 V, the cell current is completely limited by the diffusion rate of water across the membrane and further increases in the cell voltage do not result in higher cell current. The steady state current that is reached above 1.7 V is known as the mass transfer limited current density. At the mass transfer limiting current density, the rate of water diffusion across the membrane minus the rate of electroosmotic drag is equal to the reaction rate of water at the anode.

In general, the net water flux occurs from the anode to the cathode and higher water content is related to higher performance (Falcao, Rangel et al. 2009). Thus, the influence of water content at the cathode has a higher impact than the water content at the anode. At lower humidification levels, the hydrophilic fraction of the membrane, where the water travels, decreases and overall membrane permeation becomes limited by water diffusion (Majsztrik, Bocarsly et al. 2008).

Figure 4 and 5 shows the representative fuel cell performance outlining the different losses arising from the different components. At low current densities the losses are dominated by the the activation polarization, which occur at the cathode under operation with clean hydrogen. The losses are followed by the ohmic resistance, which is mostly attributed to the solid electrolyte. Finally at high current densities, the performance is limited by the mass transport of reactancts and products. Semi-empirical approaches have been used to predict and analyze the fuel cell performance. Such an approach is the one by (Pisani, Murgia et al. 2002). where the performance curve can be represented by:

$$V\_{cell} \approx E\_0 - R\_{cell}I - b \ln\left(I\right) + a \ln\left(1 - \frac{I}{I\_l}S^{-\mu\left(1 - \frac{I}{I\_l}\right)}\right)$$

with

308 Mass Transfer - Advanced Aspects

Catalist particles

Fig. 4. Schematic representation of the catalytic layer. (A) where at low Nafion content not all the catalyst particles are connected to the membrane for ionic conduction (B) the optimal Nafion contentwhere there is good ionic and electronic conduction for all the catalyst particles. (C) When there is too much Nafion and not all of the catalyst particles are electronically connected to the diffusion layer. Reproduced from (Passalacqua, Lufrano et al. 2001)

Membrane Diffusion Layer Catalist Layer

Intermixed Nafion

A)

B)

C)

$$E\_0 = E\left(I = 0\right) + K \qquad b = \frac{1 + N\_d^0}{\alpha\_c \beta\_F} \qquad a = \frac{\gamma + N\_d^I}{\alpha\_c \beta\_F}$$

where *Vcell* is the cell potential, *E*0 is the standard cell potential, *Rcell* cell resistance, *I* is the current density, *b* is the Tafel slope, *<sup>l</sup> I* cell current density at the limiting current density, *S* flooding parameter, μ is an empirical constant, *E* is the potential, *K* is the proportionality constant, <sup>0</sup> *Nd* is the diffusion mechanism parameter at the zero current density, α*c* is the cathode transfer coefficient, β *<sup>F</sup>* is the Faraday constant, γ is the kinetic exponent of the species in the Butler-Volmer equation, *<sup>l</sup> Nd* is the diffusion mechanism parameter at the limiting current density.

In real life operation, the use of pure fuel and oxidant gases results in an impractical system. A more realistic and cost efficient approach is the use of air as an oxidant gas and hydrogen from hydrogen carrier molecules (i.e., ammonia, hydrocarbons, hydrides). The short and long term effect of impurities in these gases may have an overriding effect on the fuel cell performance. Common atmospheric impurities in the cathode gas stream that have an effect on the performance of the fuel cell include SO2, NO2, H2S, O3 (Veldhuis, deBrujin et al. 1998). Even though the hydrogen oxidation reaction occurs at higher rates than the oxygen reduction reaction at the cathode (Sukkee, Wang et al. 2000), the effect of hydrogen

Mass Transport Limitations in Proton Exchange Membrane Fuel Cells and Electrolyzers 311

**No impurities** 

Fig. 7. Modified cell performance curve to include losses from impurities in cell feed

**With catalyst and membrane impurities** 

> **Anode**

> > **IR**

**Electrolyte**

Fig. 8. Modified cell loss curves to include feed impurities

**Loss**

> **Loss**

**(with**

> **(with**

**impurities)**

**impurities)**

Fig. 5. Performance curve of a PEMFC. Reproduced from the DOE Fuel Cell Handbook (2004)

Fig. 6. Cell loses due to feed. Reproduced from the DOE Fuel Cell Handbook (2004)

Fig. 5. Performance curve of a PEMFC. Reproduced from the DOE Fuel Cell Handbook (2004)

Fig. 6. Cell loses due to feed. Reproduced from the DOE Fuel Cell Handbook (2004)

Fig. 7. Modified cell performance curve to include losses from impurities in cell feed

Fig. 8. Modified cell loss curves to include feed impurities

Mass Transport Limitations in Proton Exchange Membrane Fuel Cells and Electrolyzers 313

electrode is significantly higher than on the membrane. On the other hand during fuel cell

An investigation on the effects of slightly higher concentrations of ammonia on PEMFC performance (Uribe, Gottesfeld et al. 2002), found the damage to the fuel cell to be irreversible, unlike previous results (Halseid, Vie et al. 2006). Even at 30ppm levels it was found the cell performance to drop considerably after several hours of exposure. The authors were able to successfully trap the ammonia using an ion exchange resin and

Fuel cell systems are even more sensitive to sulfur containing compounds, yet few systematic studies have been completed on the phenomenon. Mohtadi et al. found that exposure to 5ppm of H2S would cause a 96% performance loss in a Pt catalyst based PEMFC (Mohtadi, Lee et al. 2005). This rate of poisoning was approximately 69% lower at 50oC than at 90°C. There was also evidence that sulfur crossed over at the cathode and affected the

Recent research by Ballard Power Systems on a commercial stack suggests that not only is the source of a hydrogen impurity important, but it's point of induction also (Knights, Jia et al.). The following impurities were found to effect cell performance in decreasing order: H2S in fuel >SO2 in air > NO2 in air > NH3 in air > CO in fuel > NH3 in fuel. This suggests that the control of environmental air pollutants is as important for PEMFC operation as a high purity hydrogen standard. The changes in air quality could result in up to 30mV performance loss, which was most noticeable on cold, clear days. In order to address problems such as performance loss due to impurity effects, new catalysts or membranes are

Recent studies have been investigating the effect of trace halide contaminants on performance (Martínez-Rodríguez, Fox et al. 2011). The study of tetrachloroethylene, a common cleaning and degreasing agent, found that even at levels equal to the current ISO standards for hydrogen purity (ISO under development) detrimental impacts on fuel cell performance occur. At overpotentials above 0.2V, cell performance was fully recoverable. Poisoning that occurred at lower potentials was recoverable either by purging the cell or by changing the

PEM electrolyzers have a thermoneutral voltage of 1.48V, below which H2 or O2 cannot be generated. Testing of single cell PEM electrolyzers, operated at 75°C, have produced cell efficiencies of 82% at 1 A/cm2 and 69% at 2 A/cm2 (Badwal, Giddey et al. 2006). Results indicate that the voltage losses experienced are ohmic in nature, or the voltage drop is the resistance of electron flow across the electrodes and interconnects of the cell. The cell was found to have better performance with thinner membranes, but these membranes have a shorter lifetime and are more fragile. The optimal operating current density of a water electrolyzer is between 0.5-1 A/cm2 , where resistances are minimized (Wendt and Imarisio 1988). Minimizing the ohmic resistance of the cell is important due to the high internal resistance and overvoltages experienced during operation. Cell efficiency will increase with decreasing resistance. Cell voltages will decrease with increasing cell temperature due to the decrease in overpotential and resistive losses (Onda, Murakami et al. 2002). If the individual cell is upgraded to small stacks of approximately fourteen cells, enough heat is generated due to internal resistive losses to make the cell thermally self-sustaining (Badwal, Giddey et

testing, at 0.1 ppm the performance is unaffected by the ammonia.

continue use of the fuel cell without further damage.

oxygen reduction reaction.

being developed.

operating voltage.

**2.1.2 Electrolyzers** 

al. 2006).

impurities on fuel cell performance can be devastating. Trace impurities arising from different hydrogen production processes include carbon monoxide, carbon dioxide, ammonia, water, sulfur, hydrocarbons, oxygen, helium, nitrogen, argon, formaldehyde, formic acid and halogenates. The effect of the impurities can alter the catalytic activity of the catalyst, the ohmic resistance due to poisoning on the solid electrolyte and changes in the hydrophobicity of the pores affecting the water management in the system, which in turn affects the mass transport. Figures 7 and 8 shows a simplified schematic of the losses on the performance.
