Proceedings of the 7th Annual ISC Graduate Research Symposium
ISC-GRS 2013
April 24, 2013, Rolla, Missouri
Sriram Isanaka
Department of Mechanical and Aerospace Manufacturing Engineering
Missouri University of Science and Technology, Rolla, MO 65409
DESIGN AND MANUFACTURE OF LIGHT-WEIGHT, AIR BREATHING PROTON
EXCHANGE MEMBRANE FUEL CELLS
ABSTRACT
Research conducted was aimed at addressing the size and
weight limitations of Proton Exchange Membrane (PEM) fuel
cells by building cells that are fraction of the weight of
traditional fuel cells with excellent hydrogen sealing, without
adversely affecting performance. Two designs, an Axis
Symmetric Architecture (ASA) and a modified flat plate, were
proposed to overcome the weight, size and cost limitations. α –
prototypes of a single cell PEM fuel cell designed and built
were measured at 181 grams and 142 grams respectively for
ASA and modified flat plate designs. Comparatively a single
cell traditional flat plate design weighed 2875 grams. Their
performance also compared favorably with a traditional fuel
cell design. The use of polymers for plate materials and
welding and silicone combination for sealing makes these
designs built significantly lighter and easy to manufacture and
assemble. The results of mass transport finite element analysis
and manufacturing and assembly assessment will be shown to
validate the concepts proposed.
1. INTRODUCTION
Fuel cells can lead to substantial energy savings and reductions
in imported petroleum and carbon emissions. While a lot of
research has been focused on fuel cell performance [1-3],
bipolar plate design [4-6], and membrane materials [7, 8], it
was found that balance-of-plant (BOP) which refers to
supporting and/or auxiliary components based on the power
source or site-specific requirements and integrated into a
comprehensive power system package, has become very critical
to successfully commercialize a fuel cell [9-11]. Since a fuel
cell in operation will generate heat and water, its BOP often
includes an air management system, which could be composed
of pumps, fans, compressor, and blower which determine
output performance of overall system on the preferential basis.
The proper matching of a fan's speed/torque curve to
aerodynamic output is especially important in increasing
efficiency thereby reducing its demand on a fuel cell system.
Portable power systems that use fuel cells have many
applications. They can be used in the leisure sector, such as
RV's, Cabins, Marine applications, and the industrial sector,
such as power for remote locations including gas/oil well sites,
communication towers, security, weather stations etc., or in the
military sector, such as computers and communication systems.
However, a portable fuel cell should be light weight, efficient,
and cost effective. This proposed research offers innovative
concepts to greatly simplify the BOP of portable fuel cells and
yield light weight, compact, efficient, and cost-effective fuel
cell designs.
PEM fuel cells use hydrogen fuel and oxygen/air to
produce electricity. Most PEM fuel cells produce less than 1.16
volts of electricity which is insufficient for most applications.
Therefore, multiple cells must be assembled into a fuel cell
stack. The potential power generated by a fuel cell stack
depends on the number and size of the individual fuel cells that
comprise the stack and the surface area of the PEM. Current
designs are almost exclusively based on the flat plate design as
shown in Figure 1.
Figure 1: Assembly image of the components involved in the
construction of a flat plate PEM fuel cell
In a PEM fuel cell the membrane is sandwiched
between conductive plates which collect the current for the
output of the cell known as flow plates. These plates even
when made of commonly used graphite or stainless steel add
significant weight to the device. Flow plates, which account for
40-50% cost and 60-80% weight of the whole fuel cell stack,
are an important part of the Proton Exchange Membrane (PEM)
fuel cell. These stacks need to be assembled and aligned while
at the same time significant pressure must be put on the stack to
produce electrical contact, and sealing against the leak of
hydrogen. In addition to these requirements others like
managing moisture on the air side, and managing heat, must
1
Approved for public release, distribution unlimited
also be met. The final output is fuel cells that adhere to design
requirements including
 Economical (Sub 100 $).
 Light weight (under 2 lbs).
 Efficient (5 – 6 recharges per canister).
 Long working life (> 1000 hours) per fuel cell module.
 Portable. (Small form factor).
 Aesthetic. (Functionally adept while being pleasing to the
customer).
 Reliable (Work without problems in a variety of
temperature and humidity ranges).
 Air breathing. (Without the use of a fan).
 Ease of operation.
Figure 2: Pin type flow field design
2. DESIGN CONCEPTS
Current research at Missouri S & T aims at reducing the
number of components and the weight of the PEM fuel cell
assembly. Two design ideas were proposed to accomplish this
task. The first was a variation of the existing flat plate design,
and the other was the ASA design. Mass transport Finite
Element Analysis (FEA) models were used to analyze the
design concepts in detail and their results will be discussed in
the forthcoming sections. Manufacturing and assembly
assessment of the designs will also be elaborated.
2.1. Design 1: Modified flat plate design
The major components that induce weight to
traditional design were the clamp plates and flow plates as
shown in Figure 1. These coupled with the fastener assembly
not only introduced weight but also significant manufacturing
and assembly complexity. To reduce weight, elimination of the
end plates and the fastener assembly was mandated. The flow
plates were redesigned using polycarbonate to make them light
and welding was used in conjunction with silicone adhesion to
ensure leak proof sealing. A 100 micron hole size sheet of
porous stainless steel was used as both Gas Diffusion Layer
(GDL) and current collector. A number of flow field designs
were identified to strike the right balance between fuel
distribution and manufacturing ease. The designs investigated
include pin, straight, serpentine and hybrid designs. Examples
of the designs investigated are shown in Figure 2 and Figure 3.
Figure 3: Hybrid flow field design with a combination of straight
and serpentine elements
After analyzing the mas transport simulations and identifying
the ideal flow field pattern to suit our application the αprototype of the modified flat plate design was constructed as
shown in the Figure 4.
Figure 4: Prototype of modified flat plate design
2
Approved for public release, distribution unlimited
2.2. Design 2: ASA design
In the ASA an inner cylinder serves as a mandrel for
the membrane assembly as well as a hydrogen reservoir. Any
hydrogen leaks go towards the membrane. An electrode is
wrapped around the mandrel and the membrane is wrapped
around the electrode. Another electrode is wrapped next,
finished with an enclosure which serves as a permeable air
source, so that the fuel cell breathes from all sides. A model of
the ASA design is shown in Figure 5.
3. MASS TRANSPORT FINITE ELEMENT ANALYSIS
Before manufacturing the designs, the flow parameters of the
reactants through the fuel cell designs were analyzed using
mass transport finite element analysis models. The models were
created using FLUENT software.
3.1. Design 1: Modified flat plate design
The model of the flow plate was created along with
Gas diffusion layer (GDL) and the Membrane Electrode
Assembly (MEA). The following were assumed for both the
GDL and MEA: viscous permeability of 0.44 * 10 -12 (m2) and
inertial permeability of 34 * 10-8 (m). The simulations were
performed to identify inlet and outlet positions, reactant dead
zones, velocity, purge time, channel cross-sections, land width,
and fuel distribution across the membrane surface area. One of
the aims of the analyses was to find the effect of the width of
the channel on flow and reactant distribution.
Figure 5: Assembly drawing of the ASA design
The shape of the ASA by its nature ensures fewer
sections where hydrogen can leak as compared to a
conventional flat plate fuel cell design. The ASA design has
been made with extensive use of polycarbonate and by virtue of
its cylindrical shape exhibits a very small form factor. The
current collecting electrodes in this case are windings of
stainless steel wire with a diameter of 0.5mm. The choice of
stainless steel and polycarbonate provides better conductivity
and strength than graphite based designs while ensuring the
design remains light. Also the innovative use of welding and
silicone based adhesion ensures leak proof sealing and ease of
assembly. A comparison of the ASA design and a traditional flat
plate design are shown in Figure 6.
Figure 7: Mass fraction of air in a straight channel fuel cell design
Figure 7 and Figure 8 show the mass fraction of the
reactant i.e. in this case hydrogen which is purging air from the
system in the case of straight channel and pin type channel
designs respectively. From the analysis, it was observed that at
higher channel widths, the reactant gas underwent a turbulent
flow through the channels which in turn would affect the rate of
diffusion on the gas into the GDL and MEA. To ensure streamlined flow it was found that channel widths in excess of 5mm
were to be avoided. The choice of the channel and land widths
would have to be a balancing act as having too thin or small
land areas would lead to sharp edge geometries that cause tears
in the MEA and GDL during assembly and the possibility of
creating a short in the fuel cell. The other significant effect of
the widths exceeding this value is that the reactant gas doesn’t
flow into the area in the GDL adjoining the land areas i.e. the
straight/parallel obstructions or the pins in the corresponding
plates. This effect was observed even after assuming three
dimensional porosity in the FEA models.
Figure 6: Size comparison of conventional flat plate vs. a prototype
ASA fuel cell design
3
Approved for public release, distribution unlimited
Table 1: Convective velocities at various ambient temperatures
generated using FLUENT
S.
No
1
2
3
4
5
Figure 8: Mass fraction of air in a pin type channel fuel cell design
The depth of the channels also has an effect on the
flow of the reactants. Deeper the channel, the better it is for
flow of the reactant gas but a large amount of reactant is not
utilized for the chemical reaction. For portable applications, it’s
found to be ideal to keep the channels depth at around 1mm.
The above conclusions can be supported by research conducted
at the University of Alabama [12].
3.2. Design 2: ASA design
One of the major advantages of the ASA design is
theorized as the creation of a natural convective effect across its
surface known as the chimney effect. Chimneys have only been
recently been used to improve the natural air convection
cooling of electronic components [13, 14]. The draft of the
chimney is caused because of the tendency of hot air to rise.
This could be a boon for the ASA enabling a natural flow of air
over the fuel cell membrane bringing fresh oxygen rich air into
the immediate environment on a regular basis. Preliminary
Finite Element analysis simulations performed using FLUENT
flow modeling software, support the hypothesis as shown in
Figure 9 and Table 1.
Figure 9: FEA generated velocity plot in an ASA design indicating
the occurrence of the chimney effect.
Fuel cell
temperature (C)
75
75
75
75
75
Ambient air
temperature (C)
-50
-25
0
25
50
Maximum draft
velocity (mm/s)
270
190
160
110
50
Table 1 shows values of draft velocity generated using
flow simulations in FLUENT. The fuel cell temperature is
retained at 75 o C, owing to the fact that Nafion membrane used
commonly in PEM fuel cells will dehydrate beyond this
temperature. Also the ASA was always designed as an air –
breathing, portable fuel cell that could be used in wide ranging
atmospheric conditions with differing temperatures and
humidity. Hence, multiple simulations were performed showing
their use in common and also demanding environments. As it
can be seen a draft is produced in nearly all environments while
it is more predominant at lower ambient air temperatures. This
was expected and validates the basic concept of the chimney.
This leads us to believe that if the packaging around the ASA is
designed and built without hindrance to flow it is theoretically
possible to create the chimney effect as part of the ASA design.
4. MANUFACTURING
AND
ASSEMBLY
ASSESSMENT
In the conventional flat plate design metal and graphite are used
extensively in general to ensure solid construction and a design
that has good contact and conductivity without the tendency to
bow. The choice of material also has to be stainless steel 316
which has a much higher corrosion resistance while having
good electrical properties as compared to other materials like
aluminum and copper. To better understand the manufacturing
and assembly complexity that is involved in making
conventional flat plate design it is to be noted that since the flat
plate design uses graphite or stainless steel extensively there
will be requirements for a large number of high cost tooling
such as end mills and drill bits. The large numbers of fasteners
in this design are mandatory to ensure proper surface contact
and also to reduce the contact electrical resistance inherent in
this design. Also insulating washers and sleeves are necessary
to ensure that the hydrogen side and air side plates are isolated
from one another and will never come in contact to produce a
short. It also to be noted that owing to the large number of
components there will be significant number of alignment
issues and hence the assembly and labor requirements to make
this design will be considerable.
4.1. Design 1: Modified Flat plate design
To ensure that the design is lighter and has a smaller
form factor, the conventional flat plate has been redesigned
with different choice of materials. To improve the sealing
welding is employed. To ensure strength, conductivity and low
4
Approved for public release, distribution unlimited
weight a combination of polycarbonate for the flow plates and
stainless steel for the current collector are utilized. The
polycarbonate plates are assembled with sintered porous
stainless steel sheets which now act as the current collectors
and gas diffusion layers. These leads are directly drawn out of
the fuel cell thereby reducing contact losses. In the traditional
flat plate design the electrons produced at the membrane are
transferred to the carbon cloth and then to the bipolar plate and
finally to the end plate before being drawn into external circuits
to generate current. These triple layer contact losses will not be
an issue in the new design since leads are directly drawn
outside from the carbon cloth. An assessment of its
manufacturing ease can be performed by analyzing the cost to
manufacture the design. Table 2 shows the comparison of
manufacturing costs between conventional and modified flat
plate designs.
Table 2: Comparison of manufacturing cost estimates between
conventional and modified flat plate designs.
S.
No
1
2
3
Cost
Raw material
Machining (100 $/hr)
Labor (60 $/hr)
4
Miscellaneous
Total cost per fuel cell($)
Percentage reduction in cost
Conventional
Modified
124
260 mins
131 mins
22
16 mins
29 mins
-------
21
515
98
81 %
The latest prototype while being a fraction of the
weight of the traditional flat plate fuel cell design produces
similar power output of 0.25 W. The use of welding and
polycarbonate completely negates the use of nut and bolt
assemblies thereby significantly reducing manufacturing and
assembly cost. Also in the new design the need for end plates
has been negated. The welding bonds the plates together
completely forming a single piece and this ensures that bulky
ends plates are no longer a necessity. The savings achieved by
this redesign are significant as can be seen in Table 3.
Table 3: Comparison of assembly ease and portability between
conventional and modified flat plate design
Type of
Material of
Number of
design
construction
components
Conventional SS - 316
47
Modified
Polycarbonate
6
Percentage reduction in number of components
Weight
(grams)
2875
142
Percentage reduction in weight
95.1 %
87.2 %
With the use of polycarbonate machining costs,
tooling costs are significantly reduced and with the reduced
number of components assembly costs are also reduced. This
prototype with a weight of 142 grams is now more suitable for
mobile applications.
4.2. Design 2: ASA design
In comparison, in the ASA design the frame and end
caps of the cylindrical fuel cell are made of polycarbonate,
which being a much softer material as compared to graphite or
stainless steel has lesser cutting requirements. Also the use of
wire winding which by nature is compressive eliminates
machining requirements and nut and bolt assembly based
fastening requirements. Owing to choice of materials (stainless
steel for conventional flat plate and polycarbonate for ASA)
and its small form factor the ASA saves significantly on the raw
material and machining costs. The ASA also does not require
the high cost carbide tooling that conventional flat plate design
requires. Owing to use of significantly less fasteners the labor
cost of the ASA is also significantly lower. Table 4 shows the
comparison of manufacturing costs between conventional and
ASA designs.
Table 4: Comparison of manufacturing cost estimates between
conventional and ASA designs.
S.
Cost
No
1
Raw material
2
Machining (100 $/hr)
3
Labor (60 $/hr)
4
Miscellaneous
Total cost per fuel cell($)
Percentage reduction in cost
Conventional
ASA
124
260 mins
131 mins
-------
20
28 mins
52 mins
21
515
120
77 %
A cylindrical fuel cell design has vastly reduced component list
of just 10. This will produce a drastic reduction in the assembly
time and labor requirements and will ultimately drive down
costs as can be seen in Table 4. The number of fasteners in this
design is significantly lesser than the flat plate design because
the wire winding by virtue of its nature is a compressive design
and will reduce contact resistance as shown in Table 5. Two sets
of these windings will ensure sufficient surface contact without
the need for excessive fasteners and their insulating elements.
Table 5: Comparison of assembly ease and portability between
conventional and ASA design
Type of
Material of
Number of
Weight
design
construction
components (grams)
Conventional SS - 316
47
2875
ASA
Polycarbonate
10
181
Percentage reduction in number of components
78.7 %
Percentage reduction in weight
93.8 %
Currently the ASA requires assembly fixturing which adds an
extra component to its total cost, but as this design matures and
more prototypes are made this cost can be negated. Since the
5
Approved for public release, distribution unlimited
membrane in both fuel cells is to be the same its cost is not
included in this analysis.
5.
5. CONCLUSIONS
6.
A variety of flow field configurations have been assessed
during the course of this research including pin-type,
straight/parallel channels, serpentine and multiple serpentine
channels, and also a hybrid inter-digitated design. Each of these
designs has their own advantages and disadvantages and they
can be used for different applications. Also, the design alone is
not a guarantee for the successful operation of the fuel cell.
Sealing between the plates and the MEA is also a vital factor
which can be ensured with the use of welding. If the cell is
prone to reactant leakage it will reduce the efficiency of the cell
and also lead to a hazardous environment for the user. Also
unconventional designs proposed and built show that lighter,
smaller fuel cells can be made that could become the ideal
portable power choice. Two working single cell prototypes
have been developed, namely the modified flat plate and the
ASA each weighing less than 200 grams. Mass transport finite
element analysis has been employed in both the designs to
assess preliminary feasibility of concept. Either could be scaled
up to a multi cell design that will be suitable for portable power
applications.
6. ACKNOWLEDGMENTS
This research was conducted in collaboration with Air Force
Research Labs. The assistance provided by Megamet Solid
Solutions in the manufacture of initial prototypes is
appreciated. The contribution of the Intelligent Systems Center
(ISC) at Missouri University of Science and Technology
towards the successful completion of the research is also
greatly appreciated. The timely assistance provided by the
machine shop technicians of the Mechanical Engineering
department at Missouri University of Science and Technology
is appreciated.
7.
8.
9.
10.
11.
12.
13.
14.
7. REFERENCES
1. Viral Mehta, Joyce Smith Cooper, “Review and Analysis of
PEM fuel cell design and Manufacturing”, Journal of
Power Sources, Volume 114, Issue 1, 25 February 2003,
pages 32- 53.
2. T. Berning, D. M. Lu, N. Djilali, “Three Dimensional
Computational Analysis of transport phenomena in a PEM
fuel cell”, Journal of Power Sources, Volume 106, Issue 12, 1 April 2002, pages 284-294.
3. Lin Wang, Attila Husar, Tianhong Zhou, Hongtan Liu, “A
parametric study of PEM fuel cell performances”,
International Journal of Hydrogen Energy, Volume 8, Issue
11, November 2003, pages 1263-1272.
4. J. Wind, R. Spah, W. Kaiser, G. Bohm, “Metallic bipolar
plates for PEM fuel cells”, Journal of Power Sources,
Volume 105, Issue 2, 20 March 2002, pages 256-260.
H. Tawfik, Y. Hung, D. Mahajan, “Metallic bipolar plates
for PEM fuel cells – A review”, Journal of Power Sources,
Volume 163, Issue 2, 1 January 2007, pages 755 - 767.
Xianguo Li, Imran Sabir, “Review of bipolar plates in PEM
fuel cells : Flow – field designs”, International Journal of
Hydrogen Energy, Volume 30, Issue 4, March 2005, pages
359 - 371.
K. D. Kreuer, “On the development of proton conducting
polymer membranes for hydrogen and methanol fuel cells”,
Journal of Membrane Science, Volume 185, Issue 1, 15
April 2001, pages 29 - 39.
Haolin Tang, Shen Peikang, San Pin Jiang, Fang Wang, Mu
Pan, “A degradation study of Nafion proton exchange
membrane of PEM fuel cells”, Journal of Power Sources,
Volume 170, Issue 1, 30 June 2007, pages 85 - 92.
Isa Bar-On, Randy Kirchain, Richard Roth, “Technical cost
analysis of PEM fuel cells”, Journal of Power Sources,
Volume 105, Issue 1, 15 June 2002, pages 71 - 75.
G. Gigliucci, L. Petruzzi, E. Cerelli, A. Garzisi, A. La
Mendola, “Demonstation of a residential CHP system
based of PEM fuel cells”, Journal of Power Sources,
Volume 131, Issue 1-2, 14 May 2004, pages 62 - 68.
Attila Ersoz, Hayati Olgun, Sibel Ozdogan, “Reforming
options for hydrogen production from fossil fuels for PEM
fuel cells”, Journal of Power Sources, Volume 154, Issue 1,
9 March 2006, pages 67 - 73.
Kumar, A and Reddy, R G. , “Effect of channel dimensions
and shape in the flow-field distributor on the performance
of polymer electrolyte membrane fuel cells.” 2003, Journal
of Power Sources, Vol. 113, pp. 11-18.
Ishizuka, M., Hatakeyama, T.; Nakagawa, S.; Kitamura, Y.;
Funawatashi, Y. , “Chimney Effect on Natural air cooling
of Electronic Equipment Under Inclination”, Thermal
Issues in Emerging Technology and Applications
Conference, 2010.
May, G De, M. Wojcik, “Chimney Effect on Natural
Convection Cooling of a transistor mounted on a Cooling
Fin”, Journal of Electronic Packaging, Volume 131, Issue
1, 2009.
6
Approved for public release, distribution unlimited