Mench M.M. Fuel Cell Engines

Подождите немного. Документ загружается.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

330 Polymer Electrolyte Fuel Cells

Channel Channel Channel

Capillaries Water reservior

Figure 6.39 Schematic of United Technologies porous plate fuel cell design concept, based on

[35].

Mesh designs are another attempt to reduce parasitic pressure losses, which can con-

sume up to 30% of the gross system power. Since ﬂow tends to follow the path of least

resistance, the mesh design can leave corner areas as deadzones. The spiral design is an

attempt to have exit channels, typically high in humidity, adjacent to inlet channels, so

that moisture can be exchanged in-plane between channels, reducing the humidity load

required for optimal performance. Unfortunately, aligning channels from near the inlet

with channels near the output is rarely wise because the pressure drop in the channels will

induce unintentional reactant bypass between the lower pressure exit and higher pressure

inlet, leaving portions of the active area without sufﬁcient reactant.

The metal foam is a highly porous structure designed to completely eliminate the

channel/land structure of the ﬂow ﬁeld and landing combination. Instead, a highly porous,

conductive material is used as the ﬂow ﬁeld. This system has similar draw backs to the

mesh design. Radial designs have been proposed for use in portable applications. Fuel is

pumped through an interior annulus of the radial stack, and air diffuses along the cathode

surface. The radial design minimizes the diffusion distance of the air, which is helpful when

no active pump is used.

An emerging concept in ﬂow ﬁeld design is that of asymmetric channel properties.

In this type of design, a basic pattern is adopted, such as the parallel–serpentine combi-

nation. However, varying channel–land ratios or channel depths are applied to maximize

performance and stability. This advanced design approach requires a deep understanding

of the engineering trade-offs and nonuniformities along the ﬂow ﬁeld so that the unique

features can be properly implemented. For example, the channel width can be varied along

the channel or between various channels to control pressure drop, ﬂow distribution, or

liquid water accumulation inside the ﬂow ﬁeld. The disadvantage of this approach is that

manufacture of the ﬂow ﬁelds is generally more complex.

Stack Orientation and Flow Direction In PEFCs, the anode and cathode ﬂow ﬁelds

should have some similar landing locations, so that most of the lands of the anode press

against the lands of the cathode, forming a good overall compression and reducing sag

of the membrane electrode assembly (MEA) into channels. Although ﬂow in the fuel cell

porous media is dominated by capillary forces, the channel-level and manifold liquid ﬂow

is strongly inﬂuenced by gravitational forces. If the fuel cell stack is oriented in a nonneutral

position with gravity (e.g., the fuel cell plates are aligned vertically or at an angle), the

inlets should be higher than the exits to allow water slugs to ﬂow from the cell into the exit

manifold. Additionally, the ﬂow ﬁeld should avoid local low points, where the water can

pool in the channels due to gravitational effects. The ﬂow should either be gravity neutral

or follow a path of continually decreasing height with respect to gravity from inlet to exit

for the anode and cathode.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

6.3 PEFC Flow Field Conﬁgurations and Stack Design 331

(a)

Anode in

Cathode in

(b)

Anode in

Cathode in

Figure 6.40 Schematic of single channel (a)coﬂowand(b) counterﬂow design. In stack plates,

the ﬂow is often not exactly one or the other, but a mixture, since the anode and cathode ﬂow designs

can be quite different from one another.

The direction of the fuel and oxidizer ﬂow can be coﬂow, counterﬂow, or a mixture of

both. In the coﬂow case, the fuel and oxidizer inlets follow the same path, with reactants

depleting from inlet to exit, as shown in Figure 6.40a.

For underhumidied inlets, the qualitative humidity, performance, and mole fraction dis-

tributions are illustrated in Figure 6.41. The ﬂooding shown will generally only occur after

the saturation limit is reached, which may not actually occur depending on the operational

parameters, such as ﬂow rate, temperature, and inlet humidities on the anode and cathode.

In the counterﬂow case, the inlets of the oxidizer and fuel are at opposite ends and ﬂow

against each other, as shown in Figure 6.40b. Depending on the stack and plate design, the

ﬂow pattern may not be completely coﬂow or counterﬂow and is some combination of both.

The reasons for choosing a particular ﬂow geometry are varied. The traditional view has

been to design the fuel cell ﬂow channel to minimize variations in the properties; that is,

the coolant ﬂow channel would be designed to minimize temperature gradients in the fuel

cell. However, if a cell is properly designed to take advantage of pressure, temperature, or

humidity gradients, these gradients can be used to effectively manage the water balance or

cooling requirements. For underhumidied inlets, the qualitative humidity, performance, and

mole fraction distributions are illustrated in Figure 6.42. For nonbalanced inlet humidity

combinations, the local distributions can be more complex, as discussed in Section 6.6. In

i (flooding loss)

RH saturated limit

, y

iRH

Figure 6.41 Schematic of typical distributions of species, current, humidity, and temperature in

PEFC with purely coﬂow design and underhumidiﬁed anode and cathode inlets. The ﬂooding loss

portion does not necessarily have to occur, depending on inlet conditions, temperature, etc.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

332 Polymer Electrolyte Fuel Cells

RH anode/cathode

, y

Figure 6.42 Schematic of typical distributions of species, current, humidity, and temperature in

PEFC with counterﬂow design with underhumidiﬁed anode and cathode inlets.

some cases, the mole fraction of the hydrogen can actually increase along the ﬂow path if

the inlet ﬂow is dehumidiﬁed along the ﬂow path.

Unintentional Reactant Flow Bypass Reactant bypass, or in-plane gas shorting, is de-

scribed as a phenomenon of unintentional reactant penetration through the backing layer,

underneath a current collecting land, and into another ﬂow channel. Optimal performance

without bypass is shown in Figure 6.43, where the reactant gases follow the intended ﬂow

path and utilize the fuel cell’s entire active area.

In the case where bypass occurs, portions of the cell’s active area do not continuously

receive a sufﬁcient amount of reactants and a local dead zone is created, as illustrated in

Figure 6.44. The bypass phenomenon is commonly observed in fuel cells with serpentine-

style ﬂow ﬁelds and may be caused by a variety of circumstances in a fuel cell, including

poorly designed ﬂow ﬁeld plates, water droplet accumulation, poor assembly, and large

Figure 6.43 Schematic of normal reactant ﬂow without bypass around 180

◦

switchback channel

turn. (Adapted from Ref. [36].)

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

6.3 PEFC Flow Field Conﬁgurations and Stack Design 333

Figure 6.44 Schematic of unintentional bypass that leaves a dead zone. The bypass can be a result

of poor compression, excessive channel pressure drop, or poor ﬂow ﬁeld design. (Adapted from

Ref. [36].)

pressure differentials along a given ﬂow channel [36–39]. In terms of the ﬂow plate design,

if some ﬂow channels with a high local pressure (e.g., inlet locations) are adjacent to

channels with lower pressure (e.g., exit locations), excessive shorting can occur. The spiral

design in Figure 6.36 is an example of this. As the ﬂow comes into the cell and goes into

the second turn, it is adjacent to the exit portion of the channel that is at the lowest overall

pressure in the ﬂow ﬁeld. In this case, resultant pressure mismatch can lead to signiﬁcant

reactant bypass and low performance.

The amount of bypass ﬂow is controlled by the ratio of the pressure drop through

the channel to the pressure drop through the DM. If the pressure drop through the DM to

the bypass emergence point is relatively low compared to the pressure drop through the

channel, signiﬁcant bypass will occur. Assuming laminar ﬂow, the pressure drop though

the channel a distance L from the bypass location to the bypass emergence point can be

derived from laminar ﬂow theory [40]:

P

ﬂow path

32LVµ

(6.36)

where V is the channel bulk ﬂow velocity and we have ignored turbulent, entrance, or

secondary ﬂow losses. An estimate of the pressure drop through the porous backing layer

can be estimated with Darcy’s law [41]:

P

Backing

ave

S (6.37)

where µ is the viscosity of the ﬂuid, U

ave

is the bulk average velocity through the porous

medium, S is the porous medium thickness in the direction of ﬂow (equivalent to the land

width), and k is the permeability of the backing layer. The DM permeability is a highly

nonlinear function of porosity, as discussed in Chapter 5.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

334 Polymer Electrolyte Fuel Cells

A dimensionless parameter termed the bypass ratio (BPR) can be deﬁned as the ratio

of pressure drop through the channel to the backing layer [42]:

BPR =

32LVk

ave

S

(6.38)

Obviously, the channel bulk ﬂow velocity and the average bypass velocity U

ave

will be

related, but the basic physics can be understood with this simple relationship. If BPR

is much greater than unity, signiﬁcant bypass should occur; if BPR is much less than

unity, signiﬁcant bypass can be avoided. Channel pressure, ﬂow viscosity, and temperature

should have little or no effect, while the hydraulic diameter, channel length, land width,

DM permeability, and liquid saturation are the key controlling physical parameters.

Stack Design Considerations for Portable and Automotive Applications The ideal fuel

cell stack design has the following qualities:

1. Compact design for maximum power density.

2. Proper ﬂow manifolding to deliver ﬂow evenly to all fuel cells in the stack without

reversal of ﬂow.

3. Proper sealing of reactants and coolant channels.

4. Low-cost, robust components.

5. High durability under a variety of operating environments.

6. Simpliﬁed purge and drainage of liquid water on shutdown.

For stationary and automotive PEFCs, the most common stack design is simply a vertically

stacked fuel cell plate conﬁguration. Figure 6.45 shows an individual fuel cell ﬂow plate

Figure 6.45 Serpentine stack plate ﬁeld design showing manifolding of coolant, hydrogen, and air

through middle of cell arrangement. (Reproduced with permission from Ref. [37].)

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

6.3 PEFC Flow Field Conﬁgurations and Stack Design 335

Figure 6.46 Fuel cell corrugation stacking for compact design. (Reproduced with permission from

Ref. [37].)

used in a hydrogen PEFC ﬂow ﬁeld design patented by Ballard. Typical power densities of

the fuel cell stack are now on the order of 1.3 kW/L and higher not including the balance

of the plant.

Typically, coolant ﬂow channels are placed every or every second fuel cell to pro-

vide uniform heat removal and temperature distribution. Cells are stacked as efﬁciently

as possible to provide a compact design. One way to increase stacking density is with

corrugated designs, as illustrated in Figure 6.46. The corrugated design is very efﬁcient for

the stack, but sealing and proper ﬂow manifolding are difﬁcult to achieve. Coolant ﬂow

channel designs are typically simpler than the hydrogen or air ﬂow ﬁeld designs but should

be tailored in concert with the anode and cathode ﬂow ﬁelds to achieve proper heat and

water management.

For portable fuel cells, as used in small consumer electronics, the shape of a common

vertical stack is usually not desirable to replace many of the thin batteries currently used.

Therefore, a planar stack design is often used, as shown in Figure 6.47. In this design,

the passive air ﬂow cathode faces the open air, and the perforated metal cover serves as

the current collector. The main issue with this conﬁguration has been limitations in the

effectiveness of the electrical connection between adjacent cells, since the path length for

electron travel is longer than in a traditional vertical stack design.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

336 Polymer Electrolyte Fuel Cells

Anode

flow

Anode 1 Anode 2 Anode 3

Cathode 1 Cathode 2 Cathode 3

Bipolar plate/cell

interconnect

Cathode

air flow

Figure 6.47 Planar stack design with open cathode design for natural breathing. Note the use of

cell-to-cell interconnects instead of structural bipolar plates common to PEFC designs.

Several different radial fuel cell stack designs have been developed. The primary reason

to develop a radial system is to have the same form factor as a common battery. The radial

design is typically fed fuel from an inner hollow core and air from the ambient around the

unsealed cathode edges of the cell, as shown in Figure 6.48. An advantage of the radial

design in this instance is that the diffusion path length is minimized.

Many other portable designs exist, although almost all follow the basic planar, radial,

or conventional stack design with some modiﬁcations.

Stack Manifold Flow One of the most difﬁcult engineering challenges in fuel cell stack

design is the proper manifold design for fuel, oxidizer, and coolant ﬂow. The manifold

design challenge centers around three main constraints:

1. The manifold must be compact to minimize stack volume yet have minimal pressure

loss.

2. The manifold must deliver ﬂow evenly to all fuel cell plates, or ﬂow maldistribution

can occur, leading to signiﬁcant performance degradation.

3. The manifolds must be properly sealed to prevent mixing of air, fuel, and coolant.

Anode

Cathode

Air in

out

Figure 6.48 Schematic of portable design with radial open channels. The hydrogen ﬂows upward

through the inner distribution tube and outward into the closed anode chambers. The cathode ﬂow

ﬁeld is an open structure, drawing air in from the ambient. Interconnects are not shown for clarity,

and different variations of this concept exist.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

6.3 PEFC Flow Field Conﬁgurations and Stack Design 337

Figure 6.49 Schematic of chimney design utilized in most stack conﬁgurations for ﬂow manifold-

ing. Generally, an air, hydrogen, and coolant inlet and outlet chimney is used.

The manifold design challenge is further exacerbated by the fact that most fuel cell

stacks contain over 100 fuel cells in series. In PEFCs, the presence of liquid water presents

a multiphase ﬂow and instability problem that further complicates design. Despite the

myriad ﬂow ﬁeld patterns, the vast majority of vertical stacks employ some sort of chimney

manifold, as illustrated in Figure 6.49, with ﬂow for coolant, fuel, and air running through

the periphery or the middle of the stack. The manifold exchanges between the individual

cells must be sealed with O-rings, gaskets, or permanent glues to prevent crossover or

coolant leakage.

In the chimney manifold design, the pressure drop in the manifold must be much less

than the pressure drops through the ﬂow ﬁeld plates, so that ﬂow is evenly distributed

among the plates. In an ideal situation, each parallel ﬂow ﬁeld plate would have an identical

pressure drop, providing uniform ﬂow distribution. However, multiphase ﬂow will exisit

in PEFC operation during startup, shutdown, and even regular operation, and assembly

conditions or DM sag can result in slight difference in plate-to-plate or channel-to-channel

pressure drop. The ﬂow inside an exit manifold is highly complex and nearly impossible

to analytically describe. The presence of liquid water droplets in the ﬂow ﬁelds increases

the pressure drop for droplet removal and causes transient pressure variations between the

ﬂow ﬁeld plates in the stack.

The ﬂow from the manifold and through the fuel cells is analogous to current ﬂow

through a parallel resistance network, as illustrated in Figure 6.50. If one of the individual

cells has a slightly different pressure drop due to liquid accumulation or DM sag, the ﬂow

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

338 Polymer Electrolyte Fuel Cells

Inlet manifold flow exit manifold flow

etc.

Figure 6.50 Flow from manifold into fuel cell stack plates is analogous to resistor in parallel net-

work. Minute variations in the pressure drop through any of the plates can result in ﬂow redistribution

and cause severe damage.

into that cell will be reduced, possibly below the minimum required ﬂow to produce the

stack current. In this case, that cell voltage will go to zero or even a negative value (voltage

reversal), reducing the efﬁciency of the stack and increasing the heat release.

Voltage reversal can lead to rapid cell degradation from excessive heat release and unde-

sired side reactions that take place to produce the necessary stack current. While impossible

galvanically for a single cell, an individual fuel cell in a stack can undergo voltage reversal

by using available power from the properly functioning cells to drive undesired electrolytic

reactions. Voltage reversal can occur when the conditions are such that it is not possible to

generate the stack current in a galvanic mode. Since the individual cells are generally in

series electrically, each cell must carry the same current, but the voltage in each cell can

be different. As long as the total stack voltage is greater than zero, individual cells in

the stack can be powered electrolytically by the rest of the stack to produce the required

current through alternative electrochemical reactions. Essentially, if the current cannot be

produced by the available galvanic cell voltage, the cell will be powered by the other cells

in electrolytic mode to a negative voltage where the current is produced by some other

electrochemical reaction. In some cases, the alternative reaction can be benign, such as

water electrolysis. In other cases, the reaction is irreversible and permanently degrades the

electrode. For example, if the anode ﬂow is blocked, the anode potential will rise, reducing

the fuel cell voltage. The ﬁrst signiﬁcant electrolytic reaction that will occur at the anode

to produce current is water electrolysis:

O → 2H

+ 2e

−

(6.39)

This reaction is benign and will not cause irreversible damage in small quantities. If this

reaction cannot generate enough current, the anode potential will continue to rise, eventually

driving the anode to oxidize the carbon in the electrode support at sufﬁciently negative cell

voltages:

C + 2H

O → CO

2(g)

+ 4H

+ 4e

−

(6.40)

Normally, this reaction occurs so slowly at PEFC temperatures and potentials that it is not a

concern. However, under a reversal situation the cell voltage can become negative enough

to accelerate this reaction to appreciable levels.

c06 JWPR067-Mench January 26, 2008 20:1 Char Count=

6.4 Direct Alcohol Polymer Electrolyte Cells 339

If there is oxidant starvation, the anode oxidation continues, but due to the absence of

oxygen, the cathode potential is decreased to around 0–100 mV, where protons are reduced.

This is called a hydrogen pump and is benign:

+ 2e

−

→ H

(6.41)

In general, voltage reversal must be avoided during operation to prevent accelerated degra-

dation.

6.4 DIRECT ALCOHOL POLYMER ELECTROLYTE

CELLS

Hydrogen is a high-efﬁciency fuel for fuel cells but has several technical drawbacks which

engender an alternative fuel in certain applications.

Ĺ Storage and transportation of hydrogen are not yet established or simple.

Ĺ Storage of hydrogen is bulky in the gas phase, heavy stored in solid-phase hydrides,

and cryogeninc and expensive in the liquid phase.

Ĺ Hydrogen is highly ﬂammable and has an invisible ﬂame. Broad mixes of hydrogen

in air between 4 and 74% hydrogen are ﬂammable.

Ĺ The H

PEFC requires ancillary components such as humidiﬁers and coolant recir-

culators that can be larger than the fuel cell stack itself.

Chapter 8 discusses the hydrogen storage, generation, and delivery in greater detail. For

now, we can identify two major applications where the use of stored hydrogen fuel may

not be the best solution:

1. Portable applications, where system compactness and not efﬁciency is the most

valuable design parameter.

2. Stationary and distributed power applications, where delivery of hydrogen is not

readily available, but natural gas is.

For stationary applications, the most common solution is to use a stream of reformed

hydrogen as fuel. In fuel reformation, a hydrogen carrying liquid fuel such as methane

or ethanol is partially oxidized into a hydrogen-rich mixture including carbon dioxide,

water vapor, and other species. If operating on reformed gas, the PEFC is still technically

a hydrogen fuel cell.

For portable applications, a liquid alcohol solution can be directly used as fuel, and the

reformation process occurs internally at the anode. The ultimate charge transfer at the anode

and the cathode for the DAFC is the same as the hydrogen fuel cell. That is, protons and

electrons are produced at the anode to generate current, and oxygen is reduced at the cathode.

Unlike hydrogen oxidation, there are many intermediate reactions which occur in the DAFC

anode to break the bonds of the original fuel until oxidation and charge transfer occur. This

logically means the electrochemical efﬁciency of the DAFC is lower than a hydrogen PEFC,

since the electrochemical oxidation reaction is more complex than hydrogen oxidation, and

will therefore have comparatively higher activation losses. However, since the density of

the liquid is much greater than gas-phase hydrogen, fuel storage is much more compact

that compressed hydrogen. Additionally, with the use of liquid fuel, other ancillary systems