Mench M.M. Fuel Cell Engines

Подождите немного. Документ загружается.

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

470 Experimental Diagnostics and Diagnosis

∆V

∆

∆V

Figure 9.16 Current mapping technique.

opposing cathode from the main cell in order to measure the performance of the desired

location, and current mapping [16–21]. In current mapping, individual shunt resistors are

located normal to the electrode surface, between the ﬂow ﬁeld and a current-collecting

plate, as illustrated in Figure 9.16. Voltage sensors passively determine the potential drop

across each resistor and, through Ohm’s law, current distribution though the ﬂow plate is

determined. It is also possible to directly measure the current without the use of shunt

resistors using a multichannel galvanostat/potentiostat, eliminating error associated with

distributed voltage measurement [22–25].

Figure 9.17 shows an example of a distributed current measurement for a PEFC operat-

ing in an oxygen-starved, constant-ﬂow-rate condition. Obviously, the lumped polarization

curve that would be obtained without a segmented cell would hide the large local variations

in the current that exist in this condition. The current-mapping approach is actively used by

many research groups and industry to monitor steady, transient, and long-term distributed

performance in PEFCs. Fuel cells with over a hundred separate signals are now used to

examine the fundamental distributed degradation and performance over a variety of condi-

tions. Other electrochemical techniques can be combined with a distributed cell to obtain a

wealth of data, such as distributed EIS or ECSA measurement. Use of embedded wires in

the catalyst layer has also been applied to determine the local in-plane variation in current

across the channel-land interface in PEFCs [26].

Species Distribution Measurement Knowledge of the species distribution is critical for

understanding the local performance because, depending on the ﬂow ﬁeld design, there

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

9.2 Physical Probes and Visualization 471

Figure 9.17 Nonhomogeneous current distribution observed in a 50 cm

fuel cell with 100% RH

anode, 0% RH cathode inlet, at 80

◦

C. (Adapted from Ref. [22].)

may be locations were the performance is low due to reactant limitation. In some fuel cell

systems with limited crossover of reactant or products through the electrolyte, the species

proﬁle can be predicted simply from knowledge of the current distribution and Faraday’s

law. In PEFCs and some liquid electrolyte systems, where signiﬁcant ﬂux of water and other

species can travel through the electrolyte, this approach is complicated by the lack of highly

precise transport parameters. If these were perfectly known, the same approach could be

used, and direct measurement of the species distribution would not be needed. Conversely,

measurement of the species distribution can be used to precisely deﬁne transport parameters

as a function of materials and environment. Ideally, we would like diagnostics that tell us the

precise reactant mole fractions in the through-plane direction from the electrode surfaces

into the ﬂow channels, but this is a very difﬁcult task. Some laser-based diagnostics have

been used to try to understand the intermediate species involved in reaction at the catalyst

surface [27]. At a larger scale, there are techniques for online measurement of the distributed

species mole fraction in a fuel cell ﬂow channel. Of particular interest is the water vapor

distribution within the gas channels of PEFCs, since this is closely tied to the issue of

ﬂooding and electrode dryout.

One common way to measure the water balance in an operating fuel cell is through

efﬂuent collection. In this approach, the humidiﬁed efﬂuent gas from the fuel cell is chilled

through a condensing bath or desiccant, and the total water content removed over time

can be determined. While this approach can be used to measure net drag coefﬁcients in

steady state, it does not provide transient or distributed data on the water throughout the

cell, which could vary widely depending on operating conditions, current distribution, and

local nonisotropic transport parameters. Real-time measurement of species distribution

in the ﬂow channels can be accomplished with real-time gas analyzer (RTGA) systems

[28]. These devices continuously sample a small stream of ﬂow diverted from the channel

location of interest so that species mole fractions at selected locations along the reactant

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

472 Experimental Diagnostics and Diagnosis

Figure 9.18 Water vapor distribution near the middle of a single serpentine path cathode. Liquid

droplet peaks are evident at this downstream location near the cathode channel exit. (Adapted from

Ref. [28].)

gas ﬂow paths can be monitored. This provides detailed understanding and perspective

of the time scales of the various intercoupled multiphase dynamic transport phenomena.

Figures 9.18 and 9.19 are examples of the instantaneous species mole fractions measured

with the RTGA system. In Figure 9.18, the cathode mole fraction changes are observed for

changes in the system operating voltage. Figure 9.19 shows the transient response of the

anode of a PEFC for a step change in current conditions. The transient responses in PEFCs

are strong functions of heat transfer membrane humidity and operating conditions, among

other variables. This approach has also been used to efﬁciently measure the mass transport

coefﬁcients when used in conjunction with current density techniques [30].

100

360 365 370 375 380 385 390

Time (min)

Mole fraction (%)

Water mole fraction

0.4 V

0.7 V

Hydrogen

Water

Figure 9.19 Transient anode species response at x/L = 83% for voltage perturbation (0.7–0.4 V)

for a relatively dry anode, A/C RH = 0/75. (Adapted from Ref. [29].)

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

9.2 Physical Probes and Visualization 473

Temperature Distribution Measurement The temperature distribution in fuel cells can

be of critical importance to the kinetics, electrolyte conductivity, material compatibility

issues (high-temperature fuel cells), internal reformation process (high-temperature fuel

cells), and other kinetic and transport phenomena known to be functionally dependent

on temperature. Since the SOFC is dominated by the electrolyte resistance, which is a

strong function of temperature, the current distribution in these systems closely follows the

temperature proﬁle. Several techniques can be used to measure the temperature distribution

in a fuel cell. An embedded thermistor or thermocouple can be used when carefully placed.

Additionally, infrared temperature measurement is a fascinating way to observe real-time

temperature variation in a fuel cell. Infrared scanners can be used to look at temperature

distribution in a specially modiﬁed single fuel cell and have been useful to “see” the phase

change processes from ice to liquid in a low-temperature fuel cell [31].

For the PEFC, the water balance is highly coupled to the temperature distribution, as

discussed. However, direct measurement of localized temperature is difﬁcult, due to the

two-phase nature of ﬂow in the gas channels and the small through-plane dimensions of a

typical electrolyte. Besides infrared measurement, the most commonly applied technique

is the direct embedding of a thermocouple or thermistor within the bipolar plate. This

approach is acceptable for most fuel cell varieties. If all thermal transport parameters, such

as speciﬁc heat, thermal conductivity, and contact resistance, are known, calculation of the

temperature proﬁles within the fuel cell can be accomplished using embedded thermocouple

data and analytical or computational heat transfer models.

For fundamental research where the thermal parameters are not precisely known and

to deﬁne unknown thermal transport parameters, we would like to directly measure the

temperature proﬁle within the electrolyte. To approach this problem, a thermocouple can

be embedded directly in the diffusion media of a PEFC [32, 33]. However, the contact

resistance between the diffusion media and the thermocouple becomes another unknown

parameter. To circumvent these difﬁculties, Burford et al. invented a method to embed an

array of microthermocouples directly between two 25-µm-thick Naﬁon electrolyte sheets

of a membrane electrode assembly [34, 35]. Local temperature variation in PEFCs was

determined to reach >10

◦

C at high current density for a thick diffusion media (>400 µm

for woven cloth media). This proved that an isothermal assumption is typically not justiﬁed

over a full range of performance and indicates phase change plays a role in water transport

in PEFCs. An even smaller MEMs-based thermosensor array has been developed using

vapor deposition [36] and has been embedded within a PEFC electrolyte, providing precise

locational control of the sensor position.

Impedance Distribution Measurement The localized impedance proﬁle can also be a

great tool to characterize local design, material, and performance. For a well-built and

designed fuel cell, the dominating ohmic loss is typically from the electrolyte. With a

segmented fuel cell for distributed measurements, a full electrical impedance spectra can

be used to examine the local performance losses. This information can be enormously

valuable to the design engineer, because the spatial variations in the EIS are often obscured

when the full cell is examined using EIS. Additionally, the HFR measurement can be

used in a distributed cell to yield detailed information about local membrane conductance.

For a PEFC, this is directly related to water content and also directly impacts durability.

For a higher temperature fuel cell such as an SOFC, the HFR proﬁle can be an indirect

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

474 Experimental Diagnostics and Diagnosis

Figure 9.20 Example of comparison of high-frequency resistance proﬁles at 0.7 V for three con-

ditions of very different inlet humidity; all other conditions are the same along a concurrent ﬂow

path. The current density, HFR, and species proﬁle is highly nonuniform for a underhumidiﬁed inlet

conditions. (Adapted from Ref. [29].)

measurement of temperature. Figure 9.20 is shown as an example of the detail that can come

from such measurement in a PEFC and shows a very nonhomogeneous conductance distri-

bution, indicating some localized dryout of the electrolyte is occurring near the inlet section.

Since accelerated electrolyte degradation has been associated with high-temperature, low-

moisture conditions, this ﬁgure indicates the anode inlet region should be most susceptible

to decreased performance over time. The bulk full-cell HFR cannot show this detail and can

be very misleading, since the HFR measurement will be indicative of the bulk resistance

and will not indicate the true local conditions. Thus, bulk HFR measurements are unlikely

to correlate well to localized membrane degradation behavior.

9.2.2 In Situ Measurement and Visualization

Direct Visualization Techniques Fuel cell visualization tools are also of great value,

especially for low-temperature PEFCs fueled by hydrogen or alcohol solutions where

blockage of reactant ﬂow by electrode, gas diffusion layer, and channel-level ﬂooding

hinders performance. The most common technique to directly image the water droplet

formation and motion in the ﬂow channel and on the catalyst surface of a PEFC is direct

optical imaging [37–40]. In this approach, a small window is embedded above the channel

allowing visual access. A video camera can focus on the ﬂow channel or on the catalyst

surface if the diffusion media is removed or replaced by a metal mesh at that location. Figure

9.21 shows an image of a liquid droplet emerging on a diffusion media surface taken with

an optical model fuel cell. This technique has revealed a variety of interesting phenomena

and has high spatial resolution on the order of micrometers, although some intrusion and

change in heat transfer conditions is needed to image the catalyst surface. Since one side of

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

9.2 Physical Probes and Visualization 475

Figure 9.21 Direct photographic image of water on surface of DM in visualization fuel cell.

the fuel cell is imaged in a small location, this approach is ideal for fundamental study of

liquid motion but is not appropriate for full-sized cell study. Careful attention to the thermal

boundary condition is also needed to ensure physically relevant results are achieved.

Another tool that is of great help in imaging two-phase water distribution in a fully

operational PEFC is neutron radiography (NR) [41–46]. Neutron radiography is analogous

to X-ray imaging, except that neutron attenuation on hydrogen is extremely high. As a

result, liquid water is easily distinguished from ﬂow channels, gas diffusion layers, and so

on, as shown in Figure 9.22.

On the left side of Figure 9.22, the neutron image of the 50-cm

PEFC has a water

wedge in it to provide a calibration standard from which the total liquid water content in

the fuel cell can be quantiﬁed. The image on the right is an expansion of the active area

Figure 9.22 (a) Neutron radiograph image of 50-cm

fuel cell and calibration water wedge; (b)

magniﬁed image of active area showing locations of liquid water accumulation. (Image location at

Penn State Breazeale Nuclear Reactor Fuel Cell Imaging Laboratory.)

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

476 Experimental Diagnostics and Diagnosis

of the same fuel cell, digitally processed to eliminate background noise and remove all

nonliquid water content, leaving an image of only the liquid water in the fuel cell. In Figure

9.22 the lighter areas indicate liquid accumulations. Advanced data analysis techniques

can quantify liquid water amounts and even measure ﬂow velocities of droplets or relative

liquid thickness distributions between the DM and channels. Another advantage of the

neutron imaging technique is that the entire fuel cell, even full-sized stack plates, can be

imaged in real time, so that the design of full-sized plates can be rapidly improved. Using

neutron computed tomography, a three-dimensional imaging can be accomplished, although

the spatial and temporal resolution is signiﬁcantly reduced [47]. Optimal geometric and

temporal resolution can range from 10 µm and 30 frames per second, respectively and

further development in this technology is expected.

Table 9.1 shows a comparison of some of the various imaging approaches for PEFCs.

Note that there is always a trade-off between resolution and ﬁeld of view. That is, resolution

is increased as the ﬁeld of view is decreased, which is a classic limitation of any imaging

technology. Another visualization approach is particle image velocimetry (PIV), which has

been used to study gas-phase ﬂow contours in the channels [48] but is only suitable for gas

channel manifold ﬂow and cannot look into the porous media surface. Magnetic resonance

imaging (MRI) [49] is a high-resolution technique that has been used in small operating fuel

cells, but because it relies on a strong magnetic force, magnetic metals cannot be used. As a

result, highly altered and small fuel cells must be used. Recently, X-ray microtomography

[50], used in the past for soil saturation analysis, has been adopted in ex situ studies of water

saturation in diffusion media of a PEFC in a nonoperating environment. This technique may

someday be used in an operating fuel cell with further development. Other technologies are

being developed for water imaging purposes and may become useful in the future.

Other Techniques As indicated, the ﬁeld of fuel cell diagnostics is far too broad to discuss

every technique here. Since ﬂooding in a PEFC is such an important technical challenge,

there have been many systems developed to try to measure liquid water content, distribution,

and storage in the PEFC. Besides neutron imaging, MRI, and direct visualization, several

more conventional technologies can be exploited to gain some additional information. Since

liquid water storage in the PEFC will change with time, the weight of the PEFC will also

vary with time according to the stored water content:



= g



(9.17)

The main limitation with this approach is that the fuel cell generally weighs many orders of

magnitude more than the water content change, so that ﬁnding a precision scale to handle

the weight of the fuel cell and still give adequate resolution for the water content is difﬁcult.

To circumvent this, the fuel cell can be built for a quick opening, and the soft goods (MEA)

can be extracted from the fuel cell, weighed, and resealed in the fuel cell. This approach

is obviously somewhat intrusive but can be used with relative ease on a properly designed

laboratory fuel cell.

Another technique to measure liquid water content in diffusion media is based on the

pressure drop through the media. Since the pressure drop increases with water saturation,

the pressure drop through an interdigitated ﬂow ﬁeld can be used as a qualitative measure

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

Table 9.1 Summary of Various Direct Visualization Techniques Which Have Been Applied to PEFCs

Visualization

Technique Spatial Resolution

Temporal

Resolution Field of View

Potential for use in an

operating cell? Limitations

Neutron Imaging 25–100 µm 30 fps or greater Up to full-size

stack single

cell

Yes—with an

unmodiﬁed cell

Only about a dozen institutions in the

world can perform this.

Magnetic

Resonance

Imaging (MRI)

∼25–100 µm Steady-State

measurement

∼5cm

cell Yes—with a highly

modiﬁed cell

Imaging cell must be non-metal, and

alters the water distribution

unreasonably [49, 52].

X-Ray micro

tomography

∼10 µm Steady-State

measurement

∼ 5mm

component

Yes Water must be artiﬁcially imbibed or

purged through a DM, not an operating

fuel cell [50].

Particle Image

Velocimetry

∼50 µm 5 fps 3–4 mm

Yes—with a modiﬁed

cell to allow optical

access

Accurate tool for ﬂow in the channel only

[48]. Limited to low speed laminar

ﬂow.

Optical Imaging Lens resolution-

can be

sub-micron

Limited to camera

resolution, can

be >1000 fps

Lens ﬁeld-

of-view

Yes—with a modiﬁed

cell to allow optical

access

Accurate tool for ﬂow for channel-level

ﬂow only. Catalyst layer images

require alteration of the diffusion

media and heat transfer boundary

conditions.

477

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

478 Experimental Diagnostics and Diagnosis

of the liquid content. One research group [51] has designed a unique fuel cell that has

a conventional ﬂow ﬁeld that can be quickly switched to an interdigitated design during

operation. From pressure-versus-time traces, the cell design can be used to delineate liquid

ﬂooding in the diffusion media from the catalyst layer in a PEFC and estimate the total

liquid blockage in the cell.

Dynamic response-based methods can also be used to determine ﬂooding status, and are

especially useful in stacks, where sensor options are much more limited. The characteristic

voltage time behavior of individual cells can be used to determine general cell status.

9.3 DEGRADATION MEASUREMENTS

Although beginning-of-lifetime performance and modeling are critical, the behavior of the

fuel cell as a function of load cycles, environment, or service time is also of great importance

to meet consumer needs. In order to determine the various factors governing degradation

of the components, many experimental techniques have been devised. In terms of general

performance, the techniques discussed in Section 9.1 to examine the various polarization

losses can be applied as a function of cycle number or service time to delineate the losses

observed. For example, catalyst sintering will result in a loss of ECSA that can be measured

by CV techniques.

Other online techniques can be used to examine the change in electrode or electrolyte

materials. If parts of the electrode or electrolyte are lost due to reaction, ﬁnite vapor pressure,

or other reasons, analysis of the efﬂuent product can often be correlated to the particular loss

mechanism. For example, in PEFCs, one mode of physicochemical electrolyte degradation

is accompanied by loss of the ﬂuorine ion, which can be detected by measurement of the

efﬂuent condensed water ﬂuorine content. Also, if carbon corrosion is occurring, carbon

monoxide or carbon dioxide gas is produced, and this can be measured with a sufﬁciently

sensitive device. In general, if there is a chemical reaction causing the degradation, the

product species from this can be detected in the efﬂuent and correlated with the measured

loss.

In many cases, however, the degradation is physical and affects the properties of the

materials. For example, the wetting properties of the PEFC and AFC diffusion media surface

can change with time due to contamination from gaskets and hoses in the fuel cell system.

These types of degradation are purely physical, and the only way of measuring them is

through indirect correlation (e.g., a noted performance loss) or by ex situ examination after

removal from the fuel cell. There are too many material analysis techniques to list here,

but microscopic imaging, surface area and morphological measurement, and mechanical

integrity testing are all readily available and ubiquitously used in fuel cell research.

9.4 SUMMARY

New diagnostic approaches and methods are being constantly developed and are part of

the ever-evolving and exciting ﬁeld of fuel cell research and development. In this chapter,

some of the most common techniques for analyzing fuel cell behavior and deriving various

electrochemical parameters were discussed. The ﬁeld of electrochemical analysis is too

broad for a single chapter, so only the most commonly used approaches were presented. In

c09 JWPR067-Mench December 16, 2007 15:15 Char Count=

Application Study: Developing an Experimental Test Program 479

particular, EIS and CV methods were discussed. The EIS measurement is based on applying

an AC with varying frequency and measuring the electrochemical response. When used

ex situ with a reference electrode or in situ with a dynamic hydrogen electrode, the EIS

approach can be used to comparatively examine ohmic, charge transfer, and mass transfer

resistances for different electrode and fuel cell systems. At the extreme high-frequency range

of the impedance spectra, the HFR represents the summation of purely ohmic resistances.

The current-interrupt technique is also used to determine the ohmic contribution to losses

but is difﬁcult to precisely apply. The CV approach can be used to determine the ECSA and

compare the electrode loading efﬁciency and electrode kinetic performance as a function

of operation history and environment. The same conﬁguration can also be used to measure

the crossover current through the electrode with a limiting current measurement. Crossover

can also be quantiﬁed using a ﬂowmeter or the OCV decay approach.

In order to examine the various transport and physicochemical processes within an

operating fuel cell and provide detailed model validation, techniques for distributed cur-

rent, temperature, and species measurement were discussed. Data from these distributed

diagnostic cells can be used to close energy, electrical, and mass conservation equations on

a local level, providing a much greater level of detail than a single-cell lumped polarization

curve. Various sensors and visualization techniques were discussed, particularly as applied

to PEFCs and observation of liquid water and ﬂooding phenomena. Neutron imaging, direct

visualization cells, magnetic resonance imaging, and other approaches have been used, but

all have particular limitations that must be considered. There is always a trade-off between

ﬁeld of view and spatial resolution for imaging analysis.

APPLICATION STUDY: DEVELOPING AN

EXPERIMENTAL TEST PROGRAM

In this text, we have learned of many of the various phenomena which control fuel cell

performance. Many readers will not have any laboratory or work experience on which to put

the understanding gained into a practical perspective, however. In this chapter, a few of the

basic methods to quantify fuel cell performance have been discussed. For this application

study, you will develop an experimental program. Consider that you are asked to develop

a workable fuel cell model to describe a new fuel cell you have no data for but want to

understand the following:

1. What is (are) the limiting polarization(s)?

2. What are the main factors controlling these polarizations?

3. What are the parameters needed to develop an analytical model similar to that

described in Chapter 4 to predict the fuel cell behavior as a function of operating

conditions?

Prepare a list of experimental tests that must be completed and the parameters you will

vary during testing (e.g., you will need to determine kinetic parameters as a function of

temperature and species mole fraction). List the methods you will use and the tests that

must be performed to obtain all of the values and functional dependencies you will need to

completely deﬁne a zero-dimensional model of this new fuel cell.