j
Architectures for Individual and Stacked Micro Single Chamber Solid Oxide Fuel Cells
by
Ethan Jon Crumlin
B.S., Mechanical Engineering
Massachusetts Institute of Technology, 2005
Submitted to the Department of Mechanical Engineering
in Partial Fulfillment of the Requirements for the Degree of
Masters of Science in Mechanical Engineering
at the
Massachusetts Institute of Technology
June 2007
©2007 Massachusetts Institute of Technology
All rights reserved
Signature of Author
4e
Certified by.......
7V/
Accepted by.................... ............
MASSACHUSETTS INSTITUTE
TCH
0
etOf Mechanical Engineering
May 19, 2007
. . . . . . . . . .
....... . ........
Yang Shao-Horn
Assistant Professor of Mechanical Engineering
Thesis Supervisor
.. ......~~~~~~.......
...........................
Lallit Anand
Professor of Mechanical Engineering
Chairman, Department of Graduate Committee
MASSACHUSETTS INST!TUTE
OF T~C;H
A.
L~o
B~ARKER
JUL 13 207
Architectures for Individual and Stacked Micro Single Chamber Solid Oxide Fuel Cells
by
Ethan Jon Crumlin
Submitted to the Department of Mechanical Engineering
On May 19, 2007 in Partial Fulfillment of the
Requirements for the Degree of Masters of Science in
Mechanical Engineering
ABSTRACT
Solid oxide fuel cells (SOFCs) are electrochemical conversion devices that convert various
fuel sources directly into electrical energy at temperatures ranging from 600'C to 1000'C.
These high temperatures could potentially allow the direct use of various hydrocarbon fuel
sources and hydrogen, without the need for expensive noble metal catalysis. Conventional
SOFCs are designed in a two-chamber system, separating the fuel and oxidant flow to the
anode and cathode, respectively. However, fuel cell manufacturing cost and robustness have
proven to be the main challenges to rapid commercialization. A promising alternate method
to achieve these requirements and to open up new architecture designs for the SOFC is the
development of single-chamber solid oxide fuel cells (SC-SOFCs). SC-SOFCs avoid many
of the manufacturing challenges associated with conventional SOFCs, and have shown
optimal performance between 500*C and 800'C. This reduces the need for high temperature
sealing and a complicated manifold structure; however it also reduces the partial pressure of
the gases at the electrodes, which reduces the theoretical obtainable voltage.
Microfabrication techniques such as photolithography, sputtering, and photo-resist liftoff
were used to create various micro SC-SOFC that are 25-400microns long and 15-40microns
wide, utilizing platinum and gold for the electrodes and YSZ as the electrolyte. After
successfully fabricating these micro SC-SOFCs, the fuel cells were tested in a
microprobestation with a custom gas chamber enclosure, which was exposed to CH 4 :0 2:N 2 at
20:20:100 ccm or 40:20:100 ccm.
A switch in the OCV from a negative voltage to a positive voltage was observed around
600'C, possible indicating change in electrochemical reactions with temperature. An OCV
of -0.4V and peak power density of 27 [W/cm 2 at 900'C in a 1:1 methane:oxygen ratio was
achieved. A stack of 10 micro SC-SOFCs as fabricated showing a cumulative OCV of 3.3
V, of an average 0.33 V per cell at 600'C in a 2:1 methane:oxygen ratio.
Ongoing research will involve characterizing micro SC-SOFCs to understand the
fundamental reaction mechanisms, electrode materials, and architectures to obtain dense,
high performing stacks of micro SC-SOFCs.
Thesis Supervisor:
Title:
Yang Shao-Horn
Assistant Professor of Mechanical Engineering
4
LIST OF PUBLICATIONS
1.
Crumlin, E. J., G. J. la 0', and Y. Shao-Horn, Architectures and Performance of
High-Voltage, Microscale Single-Chamber Solid Oxide Fuel Cell Stacks, the
Electrochemical Transactions, in press (2007).
2. la 0', G.J., H.J. In, E. Crumlin, G. Barbastathis and Y. Shao-Horn, Recent
Advances in Microdevices for Electrochemical Conversion and Storage, invited
review paper, International Journal of Energy Research, 31, 548-575 (2007).
5
6
ACKNOWLEDGMENTS
I would like to give my sincere gratitude to my family. Mom, dad and bud: you have
always been there to support me through the hard times and share with me the moments
of success. You have loved me, shaped me, guided me to become the person I have
become. Thank you and I love you all deeply.
Professor Yang Shao-Horn: you are my research advisor, academic advisor, and personal
advisor. I thank you immensely for your guidance, wisdom, challenges, and most
important, your respect. You have helped me control and grow my strengths as well as
pushed me to improve my weaknesses. Your effort, patience, candidness, intellect,
wisdom, and energy are all characteristics I admire and benefit from. I look forward to
working with you in many more endeavors.
To my fellow lab mate, graduate student Gerardo Jose la 0': thank you for all of your
guidance, advice. Thank you for always being there to help even when it was
inconvenient for you. You are a great person to learn from and work with, and I hope we
can do so for a long time to come. A special thanks to Kurt Broderick, one of the key
support staff at MIT's Micro Technology Laboratory; thank you for all your help with the
fabrication involved in this thesis. Thanks to all of my lab members for their continuous
support and effort to help me move forward with my research.
In addition, I would like to thank all of my college mentors who have provided
continuous guidance and support throughout all of my undergraduate and graduate career.
Professor Kip Hodges, Professor Alex Slocum, Professor David Wallace, you have all
helped me reach the goals and gain the experiences I have hoped to achieve. Thank you
very much for everything.
I would like to extend a special thanks to Lu Chen for editing my thesis and putting a
smile on my face everyday.
This work was supported by the GEM fellowship, MIT-Lemelson Fund, Ford-MIT
Alliance, MRSEC Program of the National Science Foundation under award number
DMR 02-13282.
7
8
ABOUT THE AUTHOR
I was born in Northampton, Massachusetts, on May 6, 1983. I attended Flag Street
Elementary School in Worcester, Massachusetts where I developed my interest in science
and engineering by participating in my school's science fair and winning 3 out of 4 years.
After the 6 th grade, I attended Forest Grove Junior High School, where for three days a
week, Principle Joseph Murphy would drive several of us to the local high school for a
special math class. Junior high is when I started to do Tae Kwon Do and developed an
interest in music. I spent two years at Doherty Memorial High School where I played
football for two years. I also joined a swim team, became a lifeguard, and taught swim
lessons. After two years I attended a focused junior-senior high school, Massachusetts
Academy of Mathematics and Science. My science and engineering abilities where
nurtured and developed through academics as well as extracurricular activities, namely,
FIRST. Through participating in FIRST, I learned how to work in a machine shop, use
engineering principles and techniques to design and build. This experience was a
fundamental tool that helped me truly excel and learn on my own.
In 2001, I graduated from high school and moved to Cambridge, Massachusetts to study
Mechanical Engineering at MIT. I have felt both humbled and honored during my time at
MIT; learning discipline and control, while absorbing the tools and resources to excel in
my field. During the summers, I have held internships with Intel in Phoenix, Arizona,
Texas Instruments in North Attleboro, Massachusetts, and Apple, Inc (for three summers)
in Cupertino, California. Upon graduating in 2005 with my Bachelor of Science degree
in Mechanical Engineering, I decided to pursue further academic goals at MIT. During
the last two years while working towards my Masters of Science degree in Mechanical
Engineering, my academic focus has been on new alternative energy technologies, such
as various low temperature and high temperature fuel cell projects, one of which is the
culmination of my master's thesis research to follow. Professionally I focus on product
design; I founded a startup company, OneWorld Medical Devices, which I led to become
Runners-Up in MIT's $100k business-plan competition. I was also recruited to join
another startup company, RallyPoint, where I am the lead mechanical engineer for an
intelligent device for your hands. I have carried my passion for product design towards
teaching others as well. I have mentored or been a lab instructor for a senior product
design class (2.009), sophomore design class (2.007), and a toy design class.
Looking ahead, I am planning to obtain my PhD with the hopes of becoming a professor
at a leading institution such as MIT. Additionally, I'd like to continue my entrepreneurial
efforts and create new startups to continue working on a variety of new products and
innovations to enhance our lives.
9
10
LIST OF FIGURES
Figure 1: A traditional two chamber fuel cell assembly comprised of two electrodes on
either side of an electrolyte. For SOFCs the cathode distributes electrons to reduce
oxygen to make an oxygen anion. The anion travels through the electrolyte to the
anode where it oxidizes with the fuel and donates the electron, completing the
24
electrochem ical reaction. .....................................................................................
Figure 2: High performing SOFC with 5micron thick LSGM film electrolyte using a
400nm thick SDC buffer layer to interface electrolyte and electrode. Using H2 and
02 over anode and cathode respectively at a flow rate of 120mL/min. (7)............. 25
Figure 3: Single chamber solid oxide fuel cell (SC-SOFC)...........................................
25
Figure 4: Two chamber SOFC results (a) SC-SOFC results (b) (41). ............................
26
Figure 5: An example of proposed electrochemical reactions for a SC-SOFC exposed to
27
methane and oxygen.............................................................................................
Figure 6: Planar SC-SOFC Architecture.......................................................................
29
Figure 7: Through-Plane SC-SOFC Architecture........................................................
29
The difference between active area of Planar and Through-Plane
Figure 8:
configurations. The right most image compares the relative size of each active area
and shows that the area of the Planar configuration is more than two times that of
30
the Through-Plane configuration. ..........................................................................
Figure 9: The effect of changing the distance between electrodes in a planar SC-SOFC
configuration as shown in Figure 6. Note that the wider the gap the worse the
32
perform ance (19).................................................................................................
Figure 10: Analytical model predicting the effect of distance between two electrodes (a)
and electrolyte thickness (b) on planar SC-SOFC performance (13). .................... 33
Figure 11: Planar SC-SOFC varying electrolyte thickness. If the electrolyte is too thin
the fuel cells performance could be reduced. If it is too large, fuel cell performance
is not significantly affected. (figure is representative, not actual experimental or
. . 34
analytical values)...............................................................................................
Figure 12: The effect of electrolyte thickness on performance in a through-plane SCSOFC configuration shown in Figure 7. Note increasing thickness decreases
. . 35
perform ance (19).................................................................................................
Figure 13: Schematic for stacking SC-SOFCs ..............................................................
35
Figure 14: Schematic for individual micro SC-SOFC test cells.....................................
38
Figure 15: Schematic for planar single (a) or multiple (b) interconnect micro SC-SOFC
39
stac k .........................................................................................................................
Figure 16: Steps for micro SC-SOFC fabrication........................................................
40
Figure 17: Typical IV graph for SOFC operating at 8000C(5)......................................
41
13
Figure 18: Simplified fuel cell circuit diagram where R,, and C. represent the effects
imposed by the electrolyte and R2 and C2 represent the electrode effects (a). Ideal
ac impedance plot for the circuit above (b) (46)....................................................
43
Figure 19: An individual micro SC-SOFC with 15micron by 400micron Pt and Au
electrodes in 1:1:5 methane:oxygen:nitrogen fuel ratio, OCV at various
temperatures (a), current-voltage at various temperatures (b), current density-power
density at various temperatures (c),.......................................................................
46
Figure 20: Au and Pt electrodes on YSZ electrolyte upon a silicon substrate at 400'C
exposed to various gas compositions: Nitrogen (a), Oxygen (b), Methane (c),
Methane:Oxygen:Nitrogen in a 1:1:5 volume flow rate ratio(d)............................ 48
Figure 21: Au and Pt electrodes on YSZ electrolyte upon a silicon substrate at 850'C
exposed to various gas compositions: Nitrogen (a), Oxygen (b), Methane (c),
Methane:Oxygen:Nitrogen in a 1:1:5 volume flow rate ratio(d)............................ 49
Figure 22: Individual micro SC-SOFC with 5micron wide Pt and Au electrodes
separated by a gap distance of 15micron and exposed to methane:oxygen:nitrogen
in a 1:1:5 fuel ratio at 900C, average OCV as a function of fuel cell area holding
the electrode width constant at 5microns and varying the electrodes length (a)
current-voltage (b) current-power (c)...................................................................
52
Figure 23: Individual micro SC-SOFC with 5micron wide Pt and Au electrodes
separated by a gap distance of 5micron and exposed to methane:oxygen:nitrogen in
a 1:1:5 fuel ratio at 900*C, average OCV as a function of fuel cell area holding the
electrode width constant at 5microns and varying the electrodes length (a) currentvoltage (b) current-pow er (c)................................................................................
53
Figure 24: Two separate trials for various 'PtAu' width electrodes that are 400micron
long exposed to 1:1:5 ratio of methane:oxygen:nitrogen at 700 0 C, average electrode
resistances (a), average electrolyte resistances (b). ................................................
55
Figure 25: Various 'PtAu' width electrodes that are 400micron in length exposed to
2:1:5 ratio of methane:oxygen:nitrogen at 600'C and 700'C, average electrode
resistances (a), average electrolyte resistances (b). ................................................
56
Figure 26: Various 'Pt _Au' width electrodes that are 100micron, 200micron and
400micron in length exposed to 2:1:5 ratio of methane:oxygen:nitrogen at 600 0 C,
average electrode resistances (a), average electrolyte resistances (b). .................... 57
Figure 27: 5micron wide Pt and Au electrodes separated by a gap distance of 5micron,
10micron, and 15micron exposed to 2:1:5 ratio of methane:oxygen:nitrogen at
600 0 C, average electrode resistances (a), average electrolyte resistances (b)......... 58
Figure 28: 10 cells connected in series measuring the overall OCV as well as the OCV
between the first 8, 6, 4, 2 and 1 cells. Each individual cell is 400micron long with
5micron wide electrodes, exposed to 2:1:5 methane:oxygen:nitrogen at 600'C......... 59
Figure 29: CHEMKIN thermodynamic equilibrium simulation over various
temperatures for a 1:1:5 ratio of methane:oxygen:nitrogen.....................................
61
14
TABLE OF CONTENTS
A b stract...........................................................................................................................
List of Publications ...................................................................................................
Acknowledgm ents......................................................................................................
About the Author .....................................................................................................
L ist o f F ig ures ...............................................................................................................
L ist o f T ab le s ................................................................................................................
3
5
7
9
13
19
1.0
21
21
22
22
22
In tro d u ctio n ..........................................................................................................
1.1 Motivation .............................................................................................
1.2 Solid Oxide Fuel Cell Fundamentals ........................................................
1.2.1 Solid Oxide Fuel Cell .................................................................
1.2.2 Thermodynamics of Electrochemical Reaction ..........................
1.2.3 Traditional Two-Chamber SOFC Architecture and Performance ... 23
1.3 Single-Chamber Solid Oxide Fuel Cell ...................................................
25
1.3.1 Single-Chamber SOFC Concept ...............................................
25
1.3.2 Design of Electrode Activity and Selectivity for SC-SOFCs.....27
1.4 Single-Chamber Solid Oxide Fuel Cell Architectures................................ 29
1.4.1 Planar Design ............................................................................
29
1.4.2 Through-Plane Design...............................................................
29
1.4.3 Comparing Planar vs. Through-Plane ........................................
30
1.5 Relating Electrode and Cell Configurations to SC-SOFC performance.........31
1.5.1 Electrolyte Resistance.................................................................
31
1.6 Single-Chamber Solid Oxide Fuel Cell Stacking .......................................
35
2.0 Experimental...........................................................................................................
37
2.1 Material Selection and Material Synthesis for micro-SC-SOFCs..............37
2.2 SC-SOFC Design.....................................................................................
37
2.2.1 Individual SC-SOFC design .....................................................
37
2.2.2 Stacked M icro SC-SOFC Design...............................................
38
2.3 SC-SOFC Fabrication..............................................................................
39
2.4 Experimental Setup...................................................................................
41
2.4.1 Testing Conditions and Apparatus .............................................
41
2.4.2 Current-Voltage Data Collection ...............................................
41
2.4.3 AC Impedance Spectroscopy ......................................................
42
2.4.4 Scanning Electron Microscopy (SEM) and Transmission Electron
M icroscopy (TEM ) ...........................................................................
44
2.4.5 X-ray Diffraction........................................................................
44
3 .0 Resu lts ....................................................................................................................
45
3.1 Individual Micro SC-SOFC Operation at Various Temperatures .............. 45
3.2 Micro SC-SOFC Electrode Changes at Various Temperatures and Gas
Compositions..................................................................................................
47
3.3 Single M icro SC-SOFC Performance........................................................
51
3.4 AC Impedance Spectroscopy: Further Micro SC-SOFC Characterization and
Isolation of Electrode and Electrolyte Effects .................................................
54
59
3.5 Stacked SC-SOFC Data ...........................................................................
11
4.0 Discussion...............................................................................................................61
4.1 OCV verse Temperature: Electrode Catalytic Activity Reversal................61
66
4.2 Electrode Transform ation........................................................................
76
4.3 Carbon N anowire Form ation....................................................................
4.4 Characterization of Individual m icro SC-SOFCs....................................... 84
84
4.4.1 Current-Voltage Data .................................................................
86
4.4.2 AC Impedance Data ...................................................................
91
4.5 Stacked SC-SOFC scaling........................................................................
93
5.0 Sum mary.................................................................................................................
97
References ....................................................................................................................
103
Appendix ....................................................................................................................
A. Temperature Dependence on Ion Conductivity for Selective Oxides ........... 103
12
Figure 30: Reaction mechanisms for micro SC-SOFC w/ gold and platinum electrodes
in a 1:1:5 methane:oxygen:nitrogen mixture at >700*C include water-gas shift
reaction of CO, H2 is oxidized on the platinum electrode, while 02 is reduced on
th e A u electrode. ......................................................................................................
63
Figure 31: Platinum, palladium and iridium CO to CO 2 formation rates as a function of
inverse temperature for single crystal supported catalysts (50)..............................64
Figure 32: Reaction mechanisms for micro SC-SOFC w/ Au and Pt electrodes in a 1:1:5
methane:oxygen:nitrogen mixture at <600'C include water-gas shift reaction for
the methane, carbon monoxide blocks the platinum electrode which promotes
oxygen reduction, while carbon monoxide is oxidized on the gold electrode.........65
Figure 33: SEM images documenting the transformation of Pt-YSZ-Au micro SCSOFC after testing at 900C in a 1:1:5 ratio methane:oxygen:nitrogen gas mixture.
Individual micro SC-SOFC(a), 5 400micron long micro SC-SOFCin series with
single interconnect (b), different sized single and multi-interconnect series stack
micro SC-SOFC(c), multi-interconnect series stack of micro SC-SOFC(d)...........67
Figure 34: SEM and EDS results to investigate platinum electrode. SEM image of
porous platinum electrode indicating where EDS measurement is measured on the
platinum surface(a), EDS results for measurement over platinum surface
confirming the presence of platinum(b), SEM image of porous platinum electrode
indicating where EDS measurement is taken within a platinum pore(c), EDS
results for measurement within platinum pore, showing no evidence for platinum
w ithin the pore(d).................................................................................................
68
Figure 35: SEM and EDS results to investigate platinum diffusion into YSZ. SEM
image of YSZ where platinum electrode had been; box indicates area within which
EDS is measured (a), EDS results for measurement YSZ surface, showing no
evidence for platinum (b) .....................................................................................
69
Figure 36: Micro SC-SOFC exposed to methane at 850*C for over one hour................ 71
Figure 37: XRD results for samples with gold and platinum on YSZ and exposed to
pure methane, pure oxygen, pure nitrogen and a 1:1:5 ratio of
methan:oxygen:nitrogen. XRD results only detect platinum, gold and YSZ......71
Figure 38: Optical images of Pt-Au patterned directly onto silicon with no electrolyte,
exposed to pure nitrogen at various temperatures. Individual micro SC-SOFCs at
400*C no visible changes observed (a), individual micro SC-SOFCs at 700*C
initiation of platinum transformation (b), series stacks of micro SC-SOFCs at
700C more consistent changes occur to the platinum(c), individual micro SCSOFCs at 800'C more platinum and all of the gold starts to transform(d), series
stacks of micro SC-SOFCs at 800C all platinum and gold transform (e-f). .......... 73
Figure 39: Optical images of Pt-Au patterned directly onto silicon with no electrolyte,
exposed to a 1:1:5 ratio of methane:oxygen:nitrogen at various temperatures.
Individual micro SC-SOFCs at 400C no visible changes observed (a), individual
micro SC-SOFCs at 700 C platinum appears to deform (b), series stacks of micro
SC-SOFCs at 700'C all platinum in contact with gold does not deform(c), series
15
stacks of micro SC-SOFCs at 750'C platinum in contact and within a certain
distance of gold does not transform(d), individual micro SC-SOFCs at 800'C more
platinum and all of the gold starts to transform(e), series stacks of micro SC-SOFCs
at 800'C all platinum and gold transform (f-g)...................................................... 74
Figure 40: XRD results for samples with gold and platinum deposited directly on
silicon (no YSZ electrolyte) and exposed to pure nitrogen or a 1:1:5 ratio of
methan:oxygen:nitrogen. XRD indicates the existence of polycrystalline Pt, Au,
75
and the formation of a new substance PtSi............................................................
Figure 41: SEM images of the carbon nanowire formation on the surface of the sample
with platinum being the catalysts. At a magnification of 250, it is apparent that
platinum electrodes on all individual micro SC-SOFCs have been utilized to form
the carbon nanowires (a) at higher magnification of 1000, there is no platinum
remaining on the electrolyte, and gold is unaffected (b) at a magnification of 4000,
it becomes clear that there is a tangled web of carbon nanowires(c)...................... 77
Figure 42: TEM image of a small platinum particle with a single carbon nanowire
growing from its surface. The solid layers of carbon prove the formation of a
79
carbon nanowire verses a carbon nanotube. ...........................................................
Figure 43: TEM image of a large platinum nano particle with multiple carbon
80
nanow ires grow ing from its surface. ......................................................................
Figure 44: SEM and EDS results to investigate platinum and carbon nanowires. SEM
images of gold electrode, YSZ electrolyte and the transformed platinum electrode
with carbon nanowires (a), EDS results taken from the platinum and carbon
nanowire, showing carbon and platinum (gold is shown because the energy levels
for gold and platinum are very similar and the software has trouble distinguishing
between the two) (b), depicts the atomic ratio of elements present, indicating
81
substantially more carbon than platinum(c)..........................................................
Figure 45: CHEMKIN thermodynamic equilibrium simulation over various
82
temperatures for a 2:1:5 ratio of methane:oxygen:nitrogen.....................................
Figure 46: Power density versus current density for 5micron wide electrodes separated
by a gap distance of 15microns, experimented at 900*C in a 1:1:5
84
m ethane:oxygen:nitrogen environm ent. ................................................................
Figure 47: Power density versus current density for 5micron wide electrodes separated
by a gap distance of 5microns, experimented at 900*C in a 1:1:5
84
m ethane:oxygen:nitrogen environment .................................................................
Figure 48: Power density vs. fuel cell area. For each electrode gap distance, there is a
given power density that is within the same range for the different fuel cell sizes. . . 85
Figure 49: A side by side comparison of Figure 9 and Figure 12 showing IV data for
86
planar (a) and through-plane(b) SC-SOFCs(19)....................................................
Figure 50: resistance ratios with respect to electrode gap separation............................
16
87
Figure 51: Analytical model of a micro SC-SOFC with 5micron and 15micron wide
gold and platinum electrodes. The parameters for this model are based on extracted
ASR values for the platinum and gold electrodes at 700*C. ..................................
89
Figure 52: Experimental data for a micro SC-SOFC with 15micron wide gold and
platinum electrodes at 700'C in a 1:1:5 ratio of methane:oxygen:nitrogen. ............ 89
Figure 53: OCV across different quantities of micro SC-SOFCs stacked in series........ 91
Figure 54: average OCV across different quantities of micro SC-SOFCs stacked in
serie s........................................................................................................................9
2
17
18
LIST OF TABLES
Table 1: Overall Chemical Reaction and Change in Gibbs Free Energy for Methane,
Hydrogen and Carbon Monoxide in a stochiametric mixture with oxygen at 30*C. 23
Table 2: Anode and Cathode electrochemical half cell reactions and overall
electrochemical reaction for Methane, Hydrogen and Carbon Monoxide in a
stoichiom etric mixture with oxygen...................................................................
27
Table 3: Individual micro SC-SOFC, with YSZ thickness of 200nanometers and
conductance of 0.0035Q-'cm' at 600 0 C (43), at varying electrode lengths and
separation distances ...........................................................................................
32
Table 4: AC Impedance experimental nomenclature.................................................
43
Table 5: micro SC-SOFC electrode transformation summary.....................................
50
Table 6: Thermodynamic equilibrium of gas species in the chamber with initial starting
ratio of 1:1:5 methane:oxygen:nitrogen for 500'C and 800 0 C............................62
Table 7: Mole fraction of the primary gas species at thermal equilibrium for a 2:1:5
mixture of methane and oxygen in nitrogen at 800'C.........................................78
Table 8: Computed values for the change in resistance and the ASR.........................88
Table 9: Comparison of electrode and electrolyte resistances....................................
90
19
20
1.0 INTRODUCTION
1.1 Motivation
As our world becomes increasingly integrated with ever more powerful mobile
electronics, providing enough energy to run these devices become increasingly difficult.
The theoretical capacity limit for a lithium ion battery is about 400Wh/kg (1), and the
current state-of-the-art batteries have the ability to reach the 150Wh/kg (1)
threshold.
Electronic devices will eventually require energy densities unreachable by lithium ion
battery systems. Development of small fuel cells with liquid fuels could potentially
provide 3-5 times more energy than batteries (2). This dramatic improvement in energy
storage density is the primary motivation for fuel cell research for portable electronics
products.
The motivation for solid oxide fuel cell (SOFC) research is in its potential for high power
densities and efficiencies and the intent for stationary applications or integration into a
power generation cycle. SOFCs are high temperature (800*C-1 100*C) solid state energy
conversion devices. The electrolyte is composed of an oxide ceramic material and is
used to transport oxygen anions from the cathode to the anode, thus completing an
electrochemical reaction and providing electrical energy in the process.
Despite being
able to provide high power densities at high efficiencies, SOFCs have several significant
disadvantages. At such high operating temperatures, the types of materials that can be
used for structure and sealing are very limited and costly. In addition, the sealing process
at such temperatures is very complex and at least 2 seals per cell are required in stacked
SOFC systems. These material and process barriers prevent traditional SOFCs from being
a commercially viable energy conversion technology. Traditional SOFCs are further
discussed in Section 1.2.
The aforementioned shortcomings of traditional SOFCs provide great motivation for the
research and development of single-chamber solid oxide fuel cells (SC-SOFC) which will
be discussed in detail in Section 1.3. The objective for SC-SOFCs is to leverage the
21
benefits of traditional SOFCs with a simple innovative design concept, which possesses a
dense stacking density, simplified architecture, ability to conform to countered surfaces,
and lower operating temperatures (400-700*C (3, 4)). With these advantages, SC-SOFC
technology can be implemented to run portable electronics for a longer period of time or
be embedded into the exhaust stream of conventional power generation systems to
recover and utilize un-reacted fuel.
1.2 Solid Oxide Fuel Cell Fundamentals
1.2.1 Solid Oxide Fuel Cell
A SOFC is an electrochemical conversion device which operates at high temperatures of
800-1100 0 C (5). The high operating temperatures allow the usage of many different fuels
such as hydrogen and various hydrocarbons, which react with oxygen to complete the
electrochemical reaction. Even carbon monoxide can serve as a fuel at high temperatures,
whereas at low temperatures, it would poison a PEMFC. SOFCs do not require noble
metal catalysts when operating at these elevated temperatures.
SOFCs are primarily
composed of an oxide material. At the cathode, a reduction reaction occurs that
transforms oxygen into an oxygen anion. This oxygen anion then transports through the
solid oxide electrolyte to the anode, where it oxidizes the fuel, which gives off an
electron and completes the electrochemical reaction (5).
1.2.2 Thermodynamics of Electrochemical Reaction
Obtaining energy from any system is accomplished from a change in energy states.
Table 1 provides the overall chemical reactions that occur within the fuel cell as well as
the change in Gibbs Free Energy from this reaction and the corresponding equilibrium
voltage.
22
Table 1: Overall Chemical Reaction and Change in Gibbs Free Energy for Methane,
Hydrogen and Carbon Monoxide in a stochiametric mixture with oxygen at 30'C.
Fuels
Methane
Hydrogen
Carbon Monoxide
Overall
CH4 +202 -2H20+C02
H2+
Reaction
1O2
-> H20
C+102
-1 C02
AG
-818 kJ/mol
-237 kJ/mol
2
-257 kJ/mol
E
1.41 V
1.23 V
1.33 V
2
The electrical potential can be evaluated with the Nernst equation, Eq. 1 (6):
Erxn
= E
+-In
RT
zF
E0
H
[Hrea avJ
AG
zF
Eq. 1
Eq.2
where E0 is the standard equilibrium voltage, Eq. 2 (6), at standard pressure of latm and
is based off of the change in Gibbs Free Energy of the reaction shown in Table 1, R is
the gas constant, T is the reaction temperature in Kelvin, z is the number of electrons
being transferred, F is Faraday's constant, and within the brackets are the products of the
reactants and products activities. For an ideal gas, the activity is equivalent to the partial
pressure; therefore, the lower the partial pressure, the lower the open circuit voltage
(OCV).
The OCV for an SOFC operating at around 800'C is about 0.98 V, and
decreases at higher temperatures and increases at lower temperatures. The decrease in
OCV at elevated temperatures is compensated for by an increase in kinetics, making
noble metals unnecessary for sufficient performance (5).
1.2.3 Traditional Two-Chamber SOFC Architecture and Performance
Traditional SOFCs have two chambers, one that isolates and exposes the fuel to the
anode, and another that isolates and exposes oxygen to the cathode, as shown in
Figure 1.
Traditional SOFCs have several benefits and drawbacks from their design.
Some of the benefits as previously mentioned are increased rate or reactions, no need for
noble metals, and high obtainable power densities of -2-3 W/cm 2 (7). However, these
23
results are difficult to achieve as the high operating temperatures demand very sturdy
structure and seals. All of the materials need to match in thermal expansion to reduce
thermally induced stresses and strains, which create cracks and limit the fuel cell's life.
Additionally, if an SOFC electrode area were scaled up to 100cm 2 or larger, the solid
oxide layers would be very thin and fragile and would be extremely susceptible to
defects, cracking. Thus, it would be very difficult to create a mobile system, and SOFC
technology would be limited to stationary applications (5).
02
Fuel
02
H201
Co
2
Figure 1: A traditional two chamber fuel cell assembly comprised of two electrodes on
either side of an electrolyte. For SOFCs the cathode distributes electrons to reduce
oxygen to make an oxygen anion. The anion travels through the electrolyte to the anode
where it oxidizes with the fuel and donates the electron, completing the electrochemical
reaction.
For these reasons, there is a large push to reduce the operating temperature for SOFCs to
800*C or lower. An innovative sandwiching technique used for the electrolyte layer
improves interfacial bonding and increases power density. This technique yields over
3.25W/cm 2 at an operating temperature of 700*C (Figure 2), thus providing relatively
high performance at lower temperatures (7).
24
1 .0~
3500
0000
* 0
0.8
~0-0
290D
0.8
0
* .e
0
. o0 0600rC
00
00 1*9*
OA
0.2
I A:
00
50
1W
I
0.
0.0
0
100(
20W0
300
4000
MDW
XMt dnweny (nWarn
Figure 2: High performing SOFC with 5micron thick LSGM film electrolyte using a
400nm thick SDC buffer layer to interface electrolyte and electrode. Using H2 and 02
over anode and cathode respectively at a flow rate of 120mL/min. (7)
1.3 Single-Chamber Solid Oxide Fuel Cell
1.3.1 Single-Chamber SOFC Concept
One way to further simplify SOFC architecture and thus avoid sealing issues and
complex gas manifolds is to create an SC-SOFC. Figure 3 shows the SC-SOFC places
the entire fuel cell into one chamber and simultaneously feeds fuel and oxygen. The idea
behind the SC-SOFC is that the cathode and anode will have different catalytic
selectivities, where the cathode is designed to primarily reduce oxygen and the anode
designed to primarily oxidize the fuel (3, 4, 8-42).
Fuel+0
2
CO2 + H20
Figure 3: Single chamber solid oxide fuel cell (SC-SOFC).
25
Placing the same fuel cell into a two chamber experimental setup and into a singlechamber experimental setup will help frame the effects from the different conditions.
Using methane as the fuel source, Figure 4 shows the I-V curves for the operation of the
fuel cells at various temperatures (41). In both fuel cell configurations, OCV decreases
as temperature increases, as predicted by the Nernst equation, Eq. 1. Additionally, in the
SC-SOFC, the oxygen and fuel gases are mixed; as the reaction proceeds, the product
gases for both oxidation and reduction reactions act to decrease partial pressure of each
species. Thus the extent of reaction in a SC-SOFC is less than that in a traditional SOFC,
and therfore, the OCV of a SC-SOFC is lower than that of a two-chamber SOFC. In
addition, the two-chamber SOFC does not show a plateau in performance with respect to
temperature while the SC-SOFC seems to reach its maximum performance at around
500 0C.
800
1200
350
1000
1000
80300
250
600
600
200
400
600
5
40010
0
0
400
800
1200
Current density / mA cm
1600
2000
2
a
0
0
200
MC00 600
400
0
000
Current density / mA cmi
b
Figure 4: Two chamber SOFC results (a) SC-SOFC results (b) (41).
This shows that for a SC-SOFC, there may be an optimal operating temperature that is
lower than 800'C. Although the decreased operating temperature may eliminate sealing
complications and allow better integration into mobile devices, the draw is in the
decreased actual power density. This SC-SOFC has been shown to produce a max power
of just over 300mW/cm 2 , while the two-chamber SOFC easily surpasses 1 W/cm2 (41).
However, comparing the two fuel cell set ups at the low operating temperature of 500'C,
the SC-SOFC has about 20% higher power output than the two-chamber SOFC (41).
This suggests that if SOFCs are to operate at low temperatures, the more efficient way to
do so would be by using a SC-SOFC.
26
1.3.2 Design of Electrode Activity and Selectivity for SC-SOFCs.
With SC-SOFCs, the fuel and oxidant are simultaneously introduced into the chamber.
This results in a decrease in activities of the reacting chemical species, which according
to the Nernst equation, Eq. 1, would decrease fuel cell performance. However, this effect
can be influenced by the catalytic selectivity of the electrodes. In an ideal SC-SOFC, the
electrodes have complete selectivity, where each electrode participates only in its
designated reaction. Figure 5 shows that the SC-SOFC cathode selectively reduces
oxygen and does not react with the methane, while the anode facilitates the oxidation of
the fuel, as well as potential gas shift reactions, but does not react with oxygen (23).
Table 2 details the ideal half reactions at each electrode.
jMi6
PF
2CH4
4#20
402
0,
CH t
>~
.,'~
Cathode
2 CH +402=
4H2 +2C,6--,
. .....Anode
-Electrolyte
Figure 5: An example of proposed electrochemical reactions for a SC-SOFC exposed to
methane and oxygen.
Table 2: Anode and Cathode electrochemical half cell reactions and overall
electrochemical reaction for Methane, Hydrogen and Carbon Monoxide in a
stoichiometric mixture with oxygen.
Fuels
Anode Half
Cell Reaction
Cathode Half
Cell Reaction
Overall
Reaction
-> 2H 20+
CH4 +402
CO2 +8e-_
202+ 8e- -->402CH4 +20
2
Carbon Monoxide
Hydrogen
Methane
-> 2H 2 0+
CO 2
2H 2 + 202- -+ 2H 20+ 4e- 2CO + 202- - 2CO 2 + 4e02+
4e
-
2H2 +0 2 -
202
02+
2H20
4e
2CO+0
-
2
2022CO2
I
27
Incomplete electrode selectivity negatively impacts the OCV. The two main speculations
for poor electrode selectivity are related to a geometrical effect and a material effect.
Geometry is believed to influence the OCV when the electrodes are oriented such that the
products of each electrode dilutes and reduces the reaction potential of the other
electrode(8, 28, 42). Ahns, et.al. showed that an OCV ~0.35 V in a SC-SOFC can be
obtained using interdigitated electrodes that have a gap spacing of 50microns (8). This
effect is further supported by Jacques-Bedard, et.al., who shows that decreasing the
separation distances between cathode and anode from 300microns to 40microns reduces
the OCV from 0.98 V to an unstable OCV ranging from 0.4-0.7V (28). Both studies use
the same anode (nickel-oxide), cathode (LSM) and electrolyte (YSZ).
This work
supports the idea that the SC-SOFC's geometry can influence the OCV and that mixing
electrode products negatively influences the reactivity of the electrodes.
Choice of
electrode material can impact OCV in one of two ways: the electrodes may be extremely
reactive with both the fuel and the oxidant, or not very reactive with either (17, 23).
Hibino, et.al. shows that when using the same electrolyte, Pt/YSZ/Au obtains an OCV of
less than 0.6 V while Ni/YSZ/LSM obtains an OCV of 0.8 V. This reduction in OCV can
be caused by the material's insufficient ability to react with the gas species; Hibino et.al.
showed that the conversion of methane using platinum was only ~4%, while that using
nickel was more than 50% (17).
The OCV is also affected if the electrodes are over
reactive. A difference in reactivity at each electrode is what generates a potential, to
achieve a high OCV, the cathode needs to have a higher rate of oxygen reduction than the
anode, and anode needs to have a higher rate of fuel oxidation than the cathode;
therefore, rapid reactions at both electrodes are non-conducive to generating a high
potential.
28
1.4 Single-Chamber Solid Oxide Fuel Cell Architectures
1.4.1 Planar Design
In planar SC-SOFC architecture, Figure 6, the anode and cathode are aligned on the
same face of the electrolyte (20, 21). The advantage of a planar SC-SOFC is that it is the
simplest architecture to manufacture since the electrodes can be visually aligned to each
other.
Figure 6: Planar SC-SOFC Architecture.
1.4.2 Through-Plane Design
In a through-plane SC-SOFC architecture, Figure 7, higher performances can be
achieved than in a planar configuration by minimizing the electrolyte thickness, which
reduces the electrolyte resistance (19).
However, this is more challenging to fabricate
due to the difficulties in aligning the anode and cathode when they are not necessarily
visible. To obtain a thin electrolyte, this design would require support by a thick anode
or cathode.
Figure 7: Through-Plane SC-SOFC Architecture.
29
1.4.3 Comparing Planar vs. Through-Plane
Assuming the same cell size, materials, and process steps, the through-plane architecture
provides a larger power density than a planar configuration. There are two contributing
factors for this, the more trivial factor is how power density is calculated. Power density
is based off of a two-dimensional area, which is the active area of the fuel cell. As shown
in Figure 8, the planar configuration has its electrodes side by side, and its active area is
effectively double the electrode area plus the spacing area between the two electrodes. In
contrast, in the through-plane SC-SOFC configuration, the electrodes are on top of each
other, consuming less two-dimensional space.
Planar Configuration
Active Area Comparison
Through-Plane Configuration
I
I~
I
I
I
I
I
I
I
I
I'
I
I
I
I
I
I
]
I
I
I
I
I
1
I
I
I
I
I
___
I
Planar Area> 2x(Opposite Area)
Top View
Top View
I
E
Cross Section View
Cross Section View
Anode
Electrolyte
Cathode
Active Area
The difference between active area of Planar and Through-Plane
Figure 8:
configurations. The right most image compares the relative size of each active area and
shows that the area of the Planar configuration is more than two times that of the
Through-Plane configuration.
30
1.5 Relating Electrode and Cell Configurations to SC-SOFC performance
Changing the geometric characteristics for the SC-SOFC, such as the distance between
electrodes and the thickness of electrolyte, results in different effects discussed in the
following sections. The two main parameters of the planar configuration, Figure 6, are
the distance between the electrodes and the electrolyte thickness.
1.5.1 Electrolyte Resistance
Eq. 3 can be used to calculate the electrolyte resistance, R:
AR
=R
Eq. 3
The symbols 1 and A are both geometric properties, where / is the distance oxygen anions
have to travel, and A is the cross-sectional area available for anion transport.
The
conductance, o, for various high temperature electrolytes such as YSZ changes as a
function of temperature, a plot of which is provided in Appendix A (43). For example,
an individual micro SC-SOFC that is 400microns long, with YSZ thickness of
200nanometers and conductance of 0.0035Q-'cm-1 at 600'C (43), and whose electrodes
are separated by 5microns, yields a resistance of 1.79x10
5
Q. Table 3 provides the
resistance values for different electrode and separation distances. Two main trends are
apparent: the larger the electrode area, the lower the resistance, and the smaller the gap,
the smaller the resistance. Therefore in a planar SC-SOFC configuration, it is expected
that the best performing fuel cell would consist of a long electrolyte layer with the
electrodes placed as closely as possible without touching.
31
..........
Table 3: Individual micro SC-SOFC, with YSZ thickness of 200nanometers and
conductance of 0.0035Q- cm' at 600 0 C (43), at varying electrode lengths and separation
distances
Electrode-Electrode Separation Distance
9 at 5micron
Q at 10micron
9 at 15micron
400microns
1.79 e5 Q
3.58 e5 Q
5.37 e5 Q
100microns
7.15 e5 Q
1.42 e6 Q
2.15 e6 Q
25 microns
2.86 e6 Q
5.72 e6 Q
8.58 e6 Q
Length
1.5.2 Previous Findings
As shown in Figure 9, increase in gap distance between the electrolytes decreases
performance (19). This result is in accordance with Eq. 3, since increased gap distance, 1,
results in increased resistance, R, and thus decreased voltage. Figure 9 clearly shows that
the changes in current and power are linearly affected by the change in the distance
between electrodes. This effect is also demonstrated with an analytical model shown in
Figure 10a (13).
I (W
140
-0.5nim
pp
80
400
60
40
200 -Wider
20
0
k)
0
200
410
600
- 3nim
00
Current density / mA cmi~
Figure 9: The effect of changing the distance between electrodes in a planar SC-SOFC
configuration as shown in Figure 6. Note that the wider the gap the worse the
performance (19).
32
An analytical model was created to determine if there was an optimum electrolyte
thickness (13). A micro planar SC-SOFC system with electrodes of width of 20micron
and a gap distance of 20micron has an optimum electrolyte thickness of about 80microns,
as shown in Figure 10b (13). For this analytical model, it is assumed that Ni-GDC is
used as the anode, and SSC-SDC is used for the cathode, each with 100nm diameter pore
size. Additionally, hydrogen and oxygen are the fuel and oxidant, and the system operates
under the temperature 500'C. However, this model should only be considered as a
foundation for the idea that a thicker electrolyte for planar SC-SOFCs can provide
improved performance, because as will be discussed later, micro SC-SOFCs have
difficulty experimentally obtaining a conventional OCV as shown in this model. Micro
SC-SOFCs have experimentally demonstrated an OCV closer to 350-400mV (44), while
many larger SC-SOFCs have an OCV of 0.95V or lower (3, 4, 9, 16-23, 26, 45).
Experimental results to be discussed later show the OCV to be much lower than that
predicted by the Nernst Equation as shown in this model.
*100
U 9S
&b
(
l
) '~tJ0
40)X
KX)
1 S3A) 4000
1 V0
zoo
.1$'0
0
W0
tl10
b
a
Figure 10: Analytical model predicting the effect of distance between two electrodes (a)
and electrolyte thickness (b) on planar SC-SOFC performance (13).
Electrolyte thickness can play a role in planar SC-SOFC performance because as
thickness decreases, the oxygen anion flow can be hindered, preventing its ability to flow
freely and quickly from the cathode to the anode. This idea is visually expressed in
Figure 11, where the electrodes have the ability to handle a fixed exchange current.
When the electrolyte, the carrier for the exchange current, is too thin, it reduces the
amount of current that can travel from one electrode to the other.
Increasing the
33
=_ _
-
- _..
---
-
-1
-
-
MM
thickness of the electrolyte will increase the current flow; however, further increase in
thickness will have diminishing improvements to fuel cell performance. The trend
explained above and conveyed in Figure 10b can also be extracted from Eq. 3. In this
case, the electrode length and electrolyte conductivity are fixed, but the cross-sectional
area is changed.
With these constraints, the electrolyte resistance is now inversely
proportional to area: increasing area decreases resistance and vice versa. Additionally,
when the cross-sectional area of the electrolyte gets too large, its ability to improve the
fuel cell performance is negligible. Alternatively, when the cross-sectional area becomes
too small, it quickly increases the resistance towards the limit of infinity.
Too thin of an electrolyte
Anode
Optimal electrolyte thickness
Cathode
Electrolyte
Very thick electrolyte
Exchange Current Pathway
Figure 11: Planar SC-SOFC varying electrolyte thickness. If the electrolyte is too thin
the fuel cells performance could be reduced. If it is too large, fuel cell performance is not
significantly affected. (figure is representative, not actual experimental or analytical values)
In Through-Plane SC-SOFC configuration shown in Figure 7, the gap between the
electrodes is controlled by the thickness of the electrolyte; therefore electrolyte thickness
is the major geometric factor that contributes to the fuel cell performance. In this case,
the thinner the electrolyte, the higher the performance (19).
Figure 12 provides
experimental data demonstrating the relationship between electrolyte thickness and fuel
cell performance of through-plane configured SC-SOFCs.
In this configuration, the
resistance is directly proportional to the electrolyte thickness, which is the length anions
have to travel, Eq. 3. This trend is reflected in Figure 12. Despite slight deviations,
which can be associated with experimental error, the resistance is strictly linear with
respect to the thickness (13, 19).
34
>
250-0_
000
wo
'nitii thickhicker
-0.6mmn
T0
600
40
0
mmn Lhicktiess!
200 400 600
SWf
I(
)200
1400
Cunetii density / mA m -
Figure 12: The effect of electrolyte thickness on performance in a through-plane SCNote increasing thickness decreases
SOFC configuration shown in Figure 7.
performance (19).
1.6 Single-Chamber Solid Oxide Fuel Cell Stacking
Thermodynamically, a single fuel cell can obtain a maximum voltage of ~0.98 volts at
800'C regardless of its size (see Section 1.2.2 for OCV thermodynamic dependencies).
Note that Eq. 1 contains no area or volume term to scale the voltage. Changing the
physical size of the fuel cell contributes to the amount of current that can flow through
the system, but to increase the system's voltage, individual cells need to be connected in
series, which will contribute linearly to an increase in voltage (21, 26, 37). Figure 13 is a
schematic depicting the SC-SOFC stacking connecting the anode of one cell to the
cathode of another, resulting in additive voltages.
Figure 13: Schematic for stacking SC-SOFCs
35
36
2.0 Experimental
2.1 Material Selection and Material Synthesis for micro-SC-SOFCs
The primary materials utilized for the micro SC-SOFC are platinum and gold to serve as
the anode and cathode respectively, and YSZ as the electrolyte. YSZ is selected for its
well-known electrochemical properties, and while platinum and gold are not conventional
high temperature fuel cell catalyst materials, they do possess different catalytic reaction
properties which are desirable for high electrode selectivity.
Platinum has a higher
affinity for oxidizing the fuel, while the gold has the ability to reduce oxygen. In addition,
since the electrodes are pure noble metals, they simplify the potential reaction
mechanisms, and are physically easier to handle and process. To further improve micro
SC-SOFCs performance, future studies will investigate more complex materials such as
solarium doped ceria (SDC) as the electrolyte, nickel combined with SDC as the anode,
and barium strontium cobalt iron (BSFC) as the cathode (4).
2.2 SC-SOFC Design
2.2.1 Individual SC-SOFC design
The experimental platform for studying individual micro SC-SOFCs is shown in Figure
14. These individual cells are created to test a wide range of parameters in an attempt to
isolate the fundamental electrochemical processes. The control cell, which every test fuel
cell in this study is compared to, consists of two equal sized electrodes (5 by 100
microns) with a separation distance of 5microns. Each pair of electrodes is placed upon
its own individual electrolyte layer that borders the pair of electrodes by 5microns. Sets
of experiments were performed on the control fuel cell where only one of the following
parameters were changed each time: anode width, cathode width, gap separating the
electrodes, lengths of both electrodes (varied simultaneously), and widths of both
electrodes (varied simultaneously).
37
. ...
For every variable that is changed, a change in performance is expected to occur. By
changing the gap separating the electrodes, the relationship between gap distance and
electrolyte performance can be studied, which can help determine the sensitivity of
performance due to electrode alignment and positioning. Individually changing the width
of the anode or cathode can provide insight into several design conditions for micro SCSOFCs, such as what electrode impacts the rate limiting reaction, and if increasing the
size of the rate limited electrode could improve fuel cell performance. Changing the
electrodes length and width can help determine if the reactions are dependent on the
electrodes area or triple phase boundary (electrode perimeter, i.e. the boundary between
the electrode, electrolyte, and chemical species).
Anode Length
Anode Width
4
hoarder
Cathode Width
K-Cathode
Length
b
Figure 14: Schematic for individual micro SC-SOFC test cells.
2.2.2 Stacked Micro SC-SOFC Design
The stacked micro SC-SOFC design shown in Figure 15 uses the individual control
micro SC-SOFC discussed in Section 2.2.1 as the base fuel cell. Experiments were done
with the stacks varying three variables: number of cells in a stack, the length of each cell
in the stack, and the number of interconnections between fuel cells (either one
interconnect between fuel cells, or one interconnect every 25 microns).
Experiments with changing the number of fuel cells in a stack are used to validate the
ability of micro SC-SOFCs stack in series to increase its voltage linearly. To understand
how a micro SC-SOFCs stacks total current and current density change with size, several
different stacks are fabricated with fuel cells of different lengths. Experiments with
changing the amount of interconnects used for stacking micro-SC-SOFCs are used to
38
determine if a single interconnect is sufficient (i.e., does not limit current), or if several
interconnects are needed to fully utilize a series stack of fuel cells.
0
0
0
S
Stack
*
Length
Slogi Interconnect mkro SC-SOFC
.
_
_
~c
colic
Stack Longth
tultiple lcttecolntt micro SC-sOtC
Stok Dwft
b
a
Figure 15: Schematic for planar single (a) or multiple (b) interconnect micro SC-SOFC
stack.
2.3 SC-SOFC Fabrication
In these micro SC-SOFCs, 8mol% yttria-stabilized zirconia (YSZ) (Praxair Specialty
Ceramics, USA) was used for the electrolyte; platinum (Pt) and gold (Au) were used for
the electrodes. Figure 16 depicts the fabrication steps used to produce micro-sized SCSOFCs: (1) a 4 inch silicon wafer is used as the SC-SOFC support; (2) spin-coating of
negative photoresist (Clariant, Switzerland) is applied; (3) wafer is covered by photomask with electrolyte features, then exposed; (4) YSZ is deposited by sputtering for 2
hours at 100W power and 60:8 ratio of Ar:0 2; (5) excess YSZ is lifted off by removing
the resist with a photoresist stripper, leaving the desired electrolyte features; (6) sample is
re-coated with negative photoresist; (7) wafer is covered by the photo-mask with the first
electrode features, then exposed; (8) Au is deposited by sputtering for 25 minutes at
100W power and pure Ar environment; (9) excess Au is lifted off by removing the resist
39
with a photoresist stripper, leaving the desired Au electrode features; (10) sample is recoated with negative photoresist; (11) wafer is covered by the photo-mask with the
second electrode features, then exposed; (12) Pt is deposited by sputtering for 25 minutes
at 100W power and pure Ar environment; (13) excess Pt is lifted off by removing the
resist with a photoresist stripper, leaving the desired Pt electrode features, and completing
the micro SC-SOFC. The YSZ, Au, and Pt thicknesses were all found to be
approximately 200nanometers using AFM measurements.
Deposit 1st
Electrode
Si Wafer
-0
Lift off
NCoat w/
2resist
UV expose
~O
& Develop
Coat w/
resist
Deposit
P0
Electrolyte
D
UV expose
& Develop
Lift off
Deposit 2nd
Electrode
Coat w/
resist
UVexpose
-
&Develop
Si Wafer
Resist
Electrolyte
1st Electrode
Figure 16: Steps for micro SC-SOFC fabrication.
40
Lift off
2nd Electrode
2.4 Experimental Setup
2.4.1 Testing Conditions and Apparatus
Samples were placed inside a microprobe station (Karl Suss, Germany) equipped with an
optical microscope (Mitutoyo, Japan) and a temperature controlled stage (Linkam
TS1500, UK), with a custom-sealed gas chamber enclosure. A Princeton Applied
Research Potentiostat 263A was used to measure OCV and collect current verse voltage
data. OCV measurements were taken between 400 and 900*C in increments of 100*C
and the data were averaged over a 120-second span. A mixture of methane, oxygen, and
nitrogen, in a volume ratio of 1:1:5 or 2:1:5 and at a flow rate of 140 mL/min or
160mL/min respectively, was supplied to the chamber.
2.4.2 Current-Voltage Data Collection
A standard method for monitoring the operation of a fuel cell is to create current-voltage
(IV) plots, as shown in Figure 17. A preferred energy conversion device would be able
to provide a constant voltage for any current drawn from the cell. However activation
losses, ohmic resistances and mass transport limitations all contribute to an overall
reduction in performance. This voltage loss relationship is shown in Eq. 4.
1.2
l
1.0 -- - - - - - - - - - - - - -
--.-Vvotg
-f
Graph is fairly linear
0.8 0.6
V
small initial tall in voltage, and open circuit
voltage is very close to theoretical value
=
0.4
-
Voltage begins to fall faster
at higher currents
0 0
200
400
600
800
1000
2
Current density (mA cm- )
Figure 17: Typical IV graph for SOFC operating at 800*C(5).
V
=
Vocy
-
A Vactivaionoss -
(Ranode + Rcathode + Reecroyte )I - AV.sstransport
Eq 4
41
SOFCs have very little activation losses due to the high operating temperatures. In wellfueled systems that do not demand high current densities, mass transport losses can also
be neglected.
Therefore voltage will be proportionally dependent upon the combined
resistances of the anode, cathode, and electrolyte. To help determine which resistance
has the largest contribution, AC impedance spectroscopy can be performed to determine
the extract electrolyte and electrode resistances. Once these resistances are determined,
they can be used in Eq. 4 to provide insight to the fuel cell performance.
2.4.3 AC Impedance Spectroscopy
AC Impedance Spectroscopy is a common analytical tool for evaluating solid state
devices(46). By modeling the fuel cell as a simple network of resistors and capacitors, it
becomes a customary electrical circuit. An alternating voltage at a certain frequency is
applied to the system, the current output is measured, and an impedance is calculated. As
the voltage frequency is changed, a spectrum of impedances can be created which can be
correlated to the device's operation. Figure 18a shows a simplified circuit diagram for a
fuel cell and Figure 18b shows an ideal ac impedance plot(46).
From the AC impedance plot, Figure 18b, R, represents the resistance of the fuel cell's
electrolyte layer. Resistor and capacitor, R2 and C2, correlate to the electrode's resistance,
which, when plotted with ac impedance, creates a semi-circle graph where the diameter
of the semi-circle is the resistance associated with the electrodes.
All R" (electrolyte
resistance) and R 2 (electrodes/fuel cell resistance) are extracted from the collected data in
order to compare test results.
AC impedance spectroscopy measurements were conducted on a Solartran 1260
frequency response analyzer connected to a Solartron 1296 dielectric interface in the
frequency range of -OMHz to -100
Hz using an ac voltage amplitude of 30-5OmV with
zero dc bias. The integration time varied between 10-30seconds depending on the clarity
of the data obtained. Each data experiment was run 2-3 times and averaged together to
account for measuring variability.
42
a
lectrodes/Fuel Cell
Resistance
Electrolyte Resistance
-Im(Z)
(RIO-R 2) Re(Z)
b
where R. and C. represent the effects
diagram
circuit
cell
Figure 18: Simplified fuel
imposed by the electrolyte and R2 and C2 represent the electrode effects (a). Ideal ac
impedance plot for the circuit above (b) (46).
The main experiments conducted evaluate the resistance characteristic for different
electrode configurations. The four main electrode configurations evaluated are all the
combinations of gold and platinum electrodes with widths of 5micron and 15micron.
Table 4 lays out the main types of cells tested and the nomenclature for the collected
data.
Table 4: AC Impedance experimental nomenclature
Nomenclature
Width
Electrode
'Pt Au'
(micron)
Material
Cell 1
Platinum
Gold
5
5
'5 5'
Cell 2
Platinum
Gold
5
15
'5_15'
Cell 3
Platinum
Gold
15
5
'15_5'
-
Cell 4
Platinum
Gold
15
15
'15 15'
43
2.4.4 Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy
(TEM)
Images of the tested samples were obtained by scanning electron microscopy (SEM) on a
Jeol 6320 operating at an accelerating voltage of 15kV for imaging and x-ray energy
dispersive spectroscopy (EDS). Images revealing the molecular structure for certain
samples were achieved by transmission electron microscopy (TEM) on a Jeol 201 Of
operating at an accelerating voltage of 200kV.
2.4.5 X-ray Diffraction
Thin film X-ray diffraction was performed to help confirm the initial molecular
composition of the samples before and after experimentation to determine if any changes
occurred in the materials structure or composition after testing.
Thin film X-ray
diffraction patterns were collected by Bruker D8 DISCOVER equipped with a twodimensional general area-detector diffraction system (GADDS). The K alpha radiation of
1.5418 angstroms was generated in a Cu tube at 40 kV and 40 mA. The X-ray beam was
focused using a collimator with a 0.5 mm beam diameter. The angle between a thin film
sample and the incident X-ray beam (omega angle) was set to 10 degrees.
44
3.0 Results
3.1 Individual Micro SC-SOFC Operation at Various Temperatures
Operation characteristics for individual micro SC-SOFCs at various temperatures are
shown in Figures 19a-c.
The micro SC-SOFC gold and platinum electrodes in this
experiment are 15microns by 400microns separated by a 5micron gap. The fuel cell was
placed into a custom sealed gas chamber. The chamber was purged for 60 minutes with
the operating gas composition of 1:1:5 of methane:oxygen:nitrogen with a total flow rate
of 140ccm. Once the purge was complete, the temperature of the fuel cell was raised to
400*C at a rate of 10*C per minute. After holding the temperature at 400*C for 15
minutes, microprobes were put in contact with each electrode to monitor the OCV for
three minutes and to collect the current-voltage data. After repeating the data collection
process at least once, the probe tips were lifted off the sample, and the chamber
temperature was increased by an additional 1 00*C at a rate of 10*C per minute. Once the
new temperature was reached, the data collection process was repeated until data had
been collected from the fuel cell up to a chamber temperature of 900*C.
For the individual micro SC-SOFCs, as the fuel cell temperature increases, the OCV goes
from a negative to a positive value between 500*C and 600'C. Once the voltage becomes
positive, a stable voltage of ~0.4V is maintained from 700*C to 900*C, Figure 19a.
Current is pulled from the fuel cell only when the temperature is above 600*C. Below
600*C either no current can be measured from the fuel cell or the value results from the
instrumentation. Above 600'C, as temperature increases, the amount of current drawn
from the cell increases, reaching current density of up to 500 A/cm 2 when the electrodes
are shorted, Figure 19b.
Conversion of the current-voltage data into power density
values shows a maximum power density of 27 tW/cm 2 when operating at 900'C, Figure
19c.
45
Open Circuit Voltage at various SC-SOFC Temperatures
0.5
0.4
0.3
>
0.2
0.1
f
-0.10
-0.2
-0.2
500
400
900
800
700
600
Temperature (C)
a
Voltage - Current Density
at Temperature 400-900C
0.5
0.40.3
400C
*500C
600C
70C
9 02
0.1
~00
800C
900C
.0.2-0.3
-0.4
100
0
200
300
400
500
600
Current (uA/cmA2)
b
Power Density- Current Density
at Temperature 400-900C
30
25
400C
a 500C
600C
700C
800C
< 20
..
is
10
900C
5
0
-
0
100
200
300
Current (uA/cmA2)
400
500
600
C
Figure 19: An individual micro SC-SOFC with 15micron by 400micron Pt and Au
electrodes in 1:1:5 methane:oxygen:nitrogen fuel ratio, OCV at various temperatures (a),
current-voltage at various temperatures (b), current density-power density at various
temperatures (c),
46
3.2 Micro SC-SOFC Electrode Changes at Various Temperatures and Gas
Compositions
During the initial micro SC-SOFC experiments, some of the gold and platinum electrodes
had transformed from solid uniform layers to porous or rough edged layers. In an attempt
to understand this observation, various samples were placed into the heated gas chamber
and exposed to different gas environments: pure methane, pure oxygen, pure nitrogen,
and a 1:1:5 mixture of all methane:oxygen:nitrogen. The samples were purged with the
selected gas mixture for one hour, and then heated to 400*C at a rate of 10*C per minute.
Keeping the temperature at 400C, digital photos of the sample were taken through the
gas chamber in order to capture the effects temperature and the gas composition had on
individual and stacked micro SC-SOFCs. Once the data were collected and documented
at 400*C, the temperature was increased 50C at a rate of 15C per minute.
Upon
reaching the new temperature, several pictures were quickly taken. The fuel cell was
kept at the new temperature for 15 minutes, and additional pictures were taken to
document any quasi steady state changes that may have occurred.
This process was
repeated up to the final resting temperature of 850*C, and held at that temperature for an
additional 60 minutes to determine if any additional transformation of the electrodes
would occur.
Figures 20a-d shows the initial condition for each sample at 400*C.
Figures 21a-d shows the final condition for each sample after being held at 850C for
over an hour. Summaries of the key changes that occurred are highlighted in Table 5.
This experiment was conducted in an attempt to isolate the effects that are caused by
chemical reactions between the gas species and the electrodes, phase transformation of
the electrodes, temperature, or possible electrochemical reactions occurring in the fuel
cell.
47
ah
d
c
Figure 20: Au and Pt electrodes on YSZ electrolyte upon a silicon substrate at 400*C
exposed to various gas compositions: Nitrogen (a), Oxygen (b), Methane (c),
Methane:Oxygen:Nitrogen in a 1:1:5 volume flow rate ratio(d).
48
9
iib
C
d
Figure 21: Au and Pt electrodes on YSZ electrolyte upon a silicon substrate at 850C
exposed to various gas compositions: Nitrogen (a), Oxygen (b), Methane (c),
Methane:Oxygen:Nitrogen in a 1:1:5 volume flow rate ratio(d).
49
400
700
Table 5: micro SC-SOFC electrode transformation summary
All
CH4
N2
02
Onset of Pt
Au defects over Pt defects on all Pt defects on all
cells
defects in
Si and on Series cells*
Series cells
cells
750
Slight increase
in Pt defects
Slight increase
in Au defects
Onset of Pt
800
Pt defects
increase in
Series cells,
Onset of Pt
defects in
Single cells
Pt defects
increase on all
cells, but
minimally
Increase in Pt
defects, onset
of Au defects*
Increase in Pt
defects
Slight increase
of defects in
both Au an Pt
Increase in Au
defects*
Pt uniformly
deformed in all
cells
Slight increase
of Au defects,
a few large Pt
defects
Au
substantially
degraded,
increased Pt
defects, but not
as severe as
No change, Pt
uniformly
deformed
defect
850
Au*
Summary
Slight Pt
defects, no
affect on Au
Overall
minimal
effects, smaller
uniform
deformation in
Au, and a few
Pt degrades,
then Au, a
switch occurs*
Pt uniformly
became porous,
Au unaffected
I large Pt defects
this is the most severe results observed at this test condition, but the results are inconsistent from test to
test, it is unclear at this time what effects influence reproducibility
*
When the micro SC-SOFCs are exposed to nitrogen or oxygen, Figure 21a and Figure
21b respectively, there is very little change to the gold and platinum electrodes. Some
small features develop primarily over the silicon, but several small features also develop
on the electrodes. When exposed to methane, Figure 21c, gold starts to transform at
lower temperatures followed by the platinum transformation at higher temperatures (note:
these are the most severe results observed at this test condition, but the results are
inconsistent. It is unclear at this time what effects influence the experiment's
reproducibility). When exposed to a mixture of gases, Figure 21d, gold is unaffected
50
while the platinum electrodes of the single cells deform and develop pores at the surface,
and these for the stacked cells completely degrade off of the surface of the YSZ.
3.3 Single Micro SC-SOFC Performance
To further understand how these micro SC-SOFC perform, additional OCV and currentvoltage plots were obtained for cells at 900*C exposed to methane:oxygen:nitrogen in a
1:1:5 ratio and a cumulative flow rate of 140ccm. The fuel cells tested all had platinum
and gold electrodes that were 5microns in width and lengths varying from 1 00microns to
400microns. The other variable evaluated is the gap between electrodes. The data shown
in Figure 22 has a gap distance of 15microns, while Figure 23 shows the data for a gap
distance of 5microns.
For individual micro SC-SOFCs with 5micron wide platinum and gold electrodes
separated by a 5micron or 15micron gap and exposed to methane:oxygen:nitrogen in a
1:1:5 fuel ratio at 900*C, a clear trend is evident between the fuel cell area and
performance; holding the electrode width constant at 5microns and changing the
electrode length results in lager area and higher power and current output. The OCVs for
both types of cells, Figure 22a and Figure 23a, are both slightly lower but comparable to
earlier results shown in Figure 19. This slight reduction in OCV is possibly attributed to
the thin electrodes being more affected by gas species mixing than the larger 15micron
wide electrodes. Mixing of the gas species may lower the partial pressure and hence
reduce the OCV as shown in Eq. 1.
At the electrode gap distance of 15microns, the
maximum current and power are approximately 9.5nA and 0.95nW when the electrodes
are at the largest size of 400microns in length.
Similarly, at the gap distance of 5
micron, the peak current and power are 15nA and 1.1 nW.
51
ocV @ 900C
Electrolyte: YSZ
Electrodes: Au & Pt Smicron, XXXmicron length with 15micron gap
Gas: CH4:02:N2 [20:20:100 ccm]
0.4
0.3
0.2
01
0A
0
-
-- - -
-------
-~
400micron
200micron
100micron
fuel Cell Length
a
IV © 900C
Electrolyte: YSZ
Electrodes: Au & Pt
5micron,
XXXmicron length with 15micron gap
Gas: CH4:02:N2 [20:20:100 ccm]
0.4
O100micron
0.3
+
0 0.2
+
0.1
trial 1
a
100micron trial 2
x
100micron trial 3
200micron trial 1
200micon trial 2
200micron trial 3
400micron trial 1
400micron trial 2
i,
U
2
0
4
12
10
8
6
Currant (nA)
b
Power 0 900C
Electrolyte: YSZ
Electrodes: Au & Pt Smicron, XXXmicron length with
Gas: CH4:02:N2 (20:20:100 ccm]
15micron gap
100micron trial 1
*
0.8
0.6
trial 3
trial 1
troal 2
trial 3
400micror trial 1
400micron trial 2
0A6
0.2
0
100micror trial 2
100micron
200micron
200micron
200micron
-
2
8
6
10
12
Currant (nA)
C
Figure 22: Individual micro SC-SOFC with 5micron wide Pt and Au electrodes
separated by a gap distance of 15micron and exposed to methane:oxygen:nitrogen in a
1:1:5 fuel ratio at 900 0 C, average OCV as a function of fuel cell area holding the
electrode width constant at 5microns and varying the electrodes length (a) currentvoltage (b) current-power (c).
52
Electrodes: Au & Pt
ocV a 900C
Electrolyte: YSZ
5mlcron, XXXmicron length with
Gas:
E
>
Smicron gap
C1H4:02:N2 (20:20:100 ccm]
2
fuel
......
..............
..
.....
..
...
Coli length
a
IV 0 900C
Electrolyte: YSZ
Electrodes: Au & Pt 5micron, XXXmicron length with 5micron gap
Gas: CH4:02:N2 [20:20:100 ccm]
0.4
0.3
100micron
7
0.2 V
9
trial 1
a 1Omicron trial 2
100micron trial 3
9 400micron trial 1
9
>
400micron
trial 2
0
9
6
3
0
15
12
Current (nA)
b
Power 0 900C
Electrolyte: YSZ
Electrodes: Au & Pt Smicron, XXXmicron length with Smicron gap
Gas: CH4:02:N2 [20:20:100 ccm]
1.2
,08
Il00micron trial
100micron trial
K400micron trial
400micron trial
0.6
0.4
0.2
2
3
I
2
--
0
0
2
10
4
12
14
Current (nA)
C
Figure 23: Individual micro SC-SOFC with 5micron wide Pt and Au electrodes
separated by a gap distance of 5micron and exposed to methane:oxygen:nitrogen in a
1:1:5 fuel ratio at 900 0 C, average OCV as a function of fuel cell area holding the
electrode width constant at 5microns and varying the electrodes length (a) currentvoltage (b) current-power (c).
53
3.4 AC Impedance Spectroscopy: Further Micro SC-SOFC Characterization
and Isolation of Electrode and Electrolyte Effects
AC impedance spectroscopy is conducted to help analyze the characteristics for micro
SC-SOFCs.
In an attempt to obtain results uninfluenced by corrosion, lower
temperatures of 600'C to 700'C were used. However, due to the long time it takes to run
AC impedance, the platinum electrode in a mixed gas mixture would still transform even
at the low temperatures. Thus, the fuel cells were held at the desired temperature to allow
for the electrode transformation to reach steady state in order to avoid measuring
dynamic changes to the electrode.
The four main micro SC-SOFCs studied are outlined in Table 4.
The electrode and
electrolyte resistances are shown in Figure 24a and Figure 24b respectively. These
figures show the results of two trials conducted on two different fuel cells. Each trial was
conducted
with
a
400micron
long
micro
SC-SOFC
at
700C
with
a
methane:oxygen:nitrogen ratio of 1:1:5 and a cumulative flow rate of 140ccm. Repeating
this experiment on two different samples helps to establish a baseline for the data's
repeatability among different tests.
While the two different trials do not completely
overlap, they are within experimental error. The electrolyte resistance is within 1Mohm
for each trial as expected since the electrolyte resistance should be the same for each cell.
The electrolyte resistance shown in Figure 24b is relatively unchanged and within the
same tolerance of 1 Mohm, indicating that the electrolyte resistance is fairly constant from
one cell configuration to another. The average resistance, however, does change with
respect to the electrode configuration, Figure 24a.
The overall resistance decreases as
the width of one electrode is increased. However when the widths of both electrodes
increase by the same amount, the resistance increases, contrary to what's expected.
These experiments seem to indicate that creating electrodes that are asymmetrical in size
benefit the fuel cells operation. This observation has been reproduced with the two trials
54
above, however the physical origin for this observation is not understood and additional
tests will be conducted to try and understand these results.
Electrode Resistance for
400micron Long SC-SOFC
CH4:02:N2=> 1:1:5
5
.10E
Electrolyte Resistance for
400micron Long SC-SOFC
CH4:02:N2=> 1;1:5
1.00E+08
08
S1E00++07
1.00E+06 i
3.10E+08
2,10E+08
1.00E+05
E
1 105-08
100E+07
-
Pt 5_A 5
-*-
--
~
1: 1 700C trial 1
1: @C70CC trial 2
E+04
~,
Pt 5_ Au 15
Pt 15 Au 5
Pt 15_Au
Electrode Type and Width (microns)
15
-
Pt 5 A 5
--
-1-
5
Pt 5 Au 15
Pt 15 Au
Electrode Type and Width (microns)
Pt
15_Au 15
1:1 a 700C trial 1
1@ 700C trial2
a
b
Figure 24: Two separate trials for various 'PtAu' width electrodes that are 400micron
long exposed to 1:1:5 ratio of methane:oxygen:nitrogen at 700'C, average electrode
resistances (a), average electrolyte resistances (b).
In an attempt to improve the data collection using AC impedance, the methane to oxygen
ratio was increased to 2:1, resulting in a cumulative flow rate of 160ccm when combined
with 5 parts nitrogen. Using this fuel mixture, the four main fuel cells as mentioned in
Table 4 were tested at 600 0 C and 700 0 C, the electrode and electrolyte resistances are
shown in Figure 25a and Figure 25b respectively.
attempted.
Data collection at 800 0 C was
However, at this temperature, a new form of platinum transformation
occurred, tainting the AC impedance measurements. This new transformation effect will
be discussed in Section 4.3. In this mixture at 600*C and 700 0 C, the average resistance
has the same trend as shown in Figure 25a. With these parameters, the overall resistance
follows the expected trend, and the lowest resistance is achieved when both electrodes are
the largest. The relationship between electrolyte resistance and temperature is also as
expected for each temperature tested, Figure 24b: the higher the temperature, the lower
the electrolyte resistance.
55
Electrolyte Resistance for
400micron Long SC-SOFC
Electrode Resistance for
400micron Long SC-SOFC
SOE+08
7
-
-----
1.OE+08
4.OE+08
E
.C
E
C
3.OE+08
2,OE+08
:1.
-
1.OE+06
0E+08
1.0E+06
0 1.OE+07
-
1.E+05
Pt 15_Au 5 Pt 15_Au 15
Pt 5 Au 15
Pt S-A 5
Electrode Type and Width (microns)
- -...... - -.....
Pt 5_ Au 15 Pt 15Au 5 Pt 15-Au 15
Electrode Type and Width (microns)
--
Pt 5_A 5
2:1 of CH402 @ 600C
-0 2:1 of CH4:02 ( 700C
of CH4:02 @ 600C
-2:1
4- 2:1 CH4:02 @ 700C
a
b
Figure 25: Various 'PtAu' width electrodes that are 400micron in length exposed to
2:1:5 ratio of methane:oxygen:nitrogen at 600'C and 700*C, average electrode
resistances (a), average electrolyte resistances (b).
In the 2:1 methane to oxygen environment at 600*C, the electrode and electrolyte
resistances are respectively shown in Figure 26a and Figure 26b for three different
length of micro SC-SOFCs (100micron, 200micron and 400micron). As length increase,
the average resistance and electrolyte resistance decreases, as expected. Additionally, the
electrolyte resistance is virtually constant for all of the fuel cell configurations that have
the same length.
The final AC impedance test is used to analyze the effect of electrode-electrode gap
distance on performance. 5micron by 400micron gold and platinum electrodes separated
by 5, 10, and 15micron gaps were exposed to 2:1 methane to oxygen, at 600'C. The
effect of electrode gap separation on electrode and electrolyte resistances are shown in
Figure 27a and Figure 27b, respectively. An increase in gap distance increases the
electrolyte resistance and the overall resistance.
56
Micro SC-SOFC Overall Resistance
CH4:02:N2= >40:20:100
@600C
-
5.OE+09
0
--
4.OE+09
3.OE+09
2.O E+09
1.QE+09
-
- -
-
-- 1
-
1.OE+07
Pt5_A 5
-4-
400micron
-U-
100micron
Ptl15_Au 5
PtS5_ Aul15
and
Width
(microns)
Electrode Type
Ptl15_Au 15
-k- 25micron
a
Electrolyte Resistance
CH4:02:N2=>40:20:100
@600C
1.OE+08
E
o
1.0E+07
1.OE+06
1.OE+05
Pt 5_A 5
Pt 5 Au 15
Pt
(5Au5
Pt 15_Au 15
Electrode Type and Width (microns)
-*-
400micron
-U-
100micron
-1r-
25micron
b
Figure 26: Various 'Pt Au' width electrodes that are 100micron, 200micron and
400micron in length exposed to 2:1:5 ratio of methane:oxygen:nitrogen at 600 0 C,
average electrode resistances (a), average electrolyte resistances (b).
57
Electrode Resistance for
400micron Long, 'x' Electrode Gap Distance
CH4:02:N2=>40:20:100
@600C
3.50E+08
~~~~
-
~~..
_-......7 ~7~.~
.............
----..
.....
-_ - _ .- - --. - ..........
...
-..
---..
- -...
-------
3.00E+08
IA
0
.
...
..-..
2.50E+08
.....
....
- ...
--...
--
.....
........
......
....
.-..
......
2.OOE+08
C
1.50E+08
1.OOE+08
5.00E+07
5 micron
10 micron
Electrode-Electrode Gap Distance
15 micron
a
Electrolyte Resistance for
400micron Long, x' Electrode Gap Distance
CH4:02:N2=>40:20:100
@600C
9.OOE+06
8.00E+06
'
E
7.OOE+06
6.00E+06
C
5.00E+06
4.00E+06
3.00E+06
15 micro
2.00E+06
1.00E+06
10 micron
5 micron
15 micron
Electrode-Electrode Gap Distance
b
Figure 27: 5micron wide Pt and Au electrodes separated by a gap distance of 5micron,
0
10micron, and 15micron exposed to 2:1:5 ratio of methane:oxygen:nitrogen at 600 C,
average electrode resistances (a), average electrolyte resistances (b).
58
3.5 Stacked SC-SOFC Data
One of the ultimate objectives of this research is to develop high voltage micro SC-SOFC
stacks. Several design concepts were proposed in Section 2.2.2 and later fabricated and
tested using 1:1 and 2:1 methane to oxygen. However, these stacked fuel cells were the
first to show platinum transformation. This transformation happened so quickly that it
rapidly severed the series connections between the stacked fuel cells. As a result, the
multi-connection stacked fuel cells, as depicted in Figure 15b, degraded quickly, making
it impossible to obtain a stack OCV. Therefore, to study the effects of serial stacking,
stacked micro SC-SOFCs with a single interconnect were exposed to a 2:1:5 ratio of
methane:oxygen:nitrogen at 600*C.
Figure 28 shows the data of 10 cells stacked in
series where each cell is composed of 400micron by 5micron wide electrodes separated
by a 5micron gap. After obtaining the overall stack OCV, the voltage between the first 8,
6, 4, 2, and 1 cells was obtained to determine the voltage change along the fuel cell stack.
The peak OCV obtained with all 10 cells stacked in series was 3.3 volts.
OCV of micro SC-SOFC Stack
3.5
S
3
2.5
* 10 cells in series
* 8 cells in series
6 cells in series
* 4 cells in series
* 2 cells in series
1 single cell
22
5
1.5
1
0.5
0
100
150
200
Time (s)
250
300
Figure 28: 10 cells connected in series measuring the overall OCV as well as the OCV
between the first 8, 6, 4, 2 and 1 cells. Each individual cell is 400micron long with
5micron wide electrodes, exposed to 2:1:5 methane:oxygen:nitrogen at 600*C.
59
60
4.0 Discussion
4.1 OCV verse Temperature: Electrode Catalytic Activity Reversal
The influence of temperature on the characteristics of noble metal micro SC-SOFC
operation shows rather interesting behavior. As temperature increases, the OCV reverses
direction, going from a negative voltage at 400*C to a positive voltage at 600'C and
above, as shown in Figure 19a. Knowing the composition of the gas exposed to the fuel
cell may help determine which electrochemical reactions may be occurring at the
different temperature regimes. Figure 29 shows the thermodynamic equilibrium states
for a 1:1:5 ratio of methane:oxygen:nitrogen over the temperature range of 400*C1 000*C, and is similar to those found in literature (47). The main chemical species in the
chamber exposed to the fuel cell are hydrogen, carbon dioxide, carbon monoxide, water,
methane, and nitrogen. Table 6 gives the molar fractions of these chemical species at
500'C and 800*C. Oxygen is not listed since it is present at a very low concentration, but
its presence facilitates the electrochemical reaction at the cathode.
0 12
C 0.10
S0 08
2
OC
H2 0
C--
E-
0,04
02
S004
000
400
Equilibrium Mole traction H
Equilibrium Mole traction CO
Equilibrium Mole traction CO
Equilibrium Mole fraction HH4
Equilibrium Mole traction H2
1000
800
900
600
700
500
Temperature C1 Equilibrium Clusteri (Cl) (C)
Figure 29: CHEMKIN thermodynamic equilibrium simulation over various temperatures
for a 1:1:5 ratio of methane:oxygen:nitrogen.
61
Table 6: Thermodynamic equilibrium of gas species in the chamber with initial starting
ratio of 1:1:5 methane:oxygen:nitrogen for 500'C and 800*C.
Species molar fraction at
'X' temperature
800 0 C
500 0 C
0.12
0.06
H2
0.04
0.05
CO 2
0.055
0.01
CO
0.055
0.05
H2 0
<0.001
0.025
CH 4
0.73
0.805
N2
The potentiostat is connected to the fuel cell via the platinum anode and the gold cathode.
When the fuel cell temperature is below 600'C, the OCV is measured and, surprisingly, a
negative voltage is observed. However, when the fuel cell is operating above 700*C, an
OCV of approximately 0.35V is observed. Although 0.35V is not particularly close to
the theoretical OCV value of 0.98V, it closely correlates with reported results for microand standard-sized single chamber fuel cells (8, 23, 28). Previous studies have attributed
the discrepancy between the theoretical OCV value and the lesser observed OCV value
to one of two possibilities: product gas species mixing with reactant gas species, thereby
reducing electrode reactivity, or poor catalytic electrode materials (8, 17, 23, 42). At
higher temperatures the fuel cell obtains an OCV similar to reported values, confirming
that the platinum anode is oxidizing hydrogen, while the gold cathode reduces oxygen.
In addition to hydrogen and hydrocarbon fuels, SOFCs are known to be able to use
carbon monoxide as a fuel. However, it is known that in the presence of water, carbon
monoxide will undergo a water-gas shift reaction to form hydrogen and carbon dioxide.
This hydrogen can then be reutilized in the electrochemical reaction (48). The results
shown above indicate the possibility that at 800 C, when there is an almost 1:1 ratio of
water to carbon monoxide, the carbon monoxide will undergo a water-gas shift reaction
to form hydrogen and carbon dioxide.
This water-gas shift reaction is beneficial in
several ways: it creates more hydrogen for an electrochemical reaction, and it transforms
the carbon monoxide into species that do not reduce the activity of platinum towards
62
27t
hydrogen oxidation (48). A schematic for the anode and cathode reaction at 700*C and
higher the can be seen in Figure 30.
ee
Co
117
Ii()
C(
O7
CO)
H'
H2
C- --
H2
H2 + C0 2
H2 O *co
4GH,2
Cathode
02+
4e- -> 202-
Anode
+700 0 C
2H 2 +02 -+ 2H 20
+ 202- -> 2H20+ 4e-
Figure 30: Reaction mechanisms for micro SC-SOFC w/ gold and platinum electrodes in
a 1:1:5 methane:oxygen:nitrogen mixture at >700'C include water-gas shift reaction of
CO, H2 is oxidized on the platinum electrode, while 02 is reduced on the Au electrode.
In the low temperature regime, the proposed electrochemical reaction is more
complicated. It would appear that the same process that occurs in the high temperature
regime could happen in the low temperature regime, but there are a few differences. It is
known that nitrogen and carbon dioxide should not be reactive and their only influence
on the performance of the fuel cell is reducing the partial pressure of the reacting fuels.
The methane that has not already decomposed will react with the water to form more
hydrogen and carbon dioxide. Once methane is reformed there is significantly less water
available at 500'C than at 800*C remaining to drive the water-gas shift reaction to push
carbon monoxide towards creating carbon dioxide. The presence of the remaining carbon
monoxide is the key property in this process at temperatures below 600'C. Carbon
monoxide is known to coat platinum easily, which then renders it inactive as a hydrogen
catalyst. This event is often referred to as platinum carbon monoxide poisoning, which
can occur for low temperature polymer exchange membrane fuel cells (49). The reason
63
for platinum carbon monoxide poisoning is due to the similarity between the binding
energy of carbon monoxide and its activation energy with platinum. Figure 31 shows
how carbon monoxide is converted into carbon dioxide on a platinum surface over a
range of temperatures. Platinum can only react with another species if carbon monoxide
is desorbed from its surface or if the carbon monoxide undergoes a water-gas shift
reaction to form carbon dioxide.
If there is not enough water for a water-gas shift
reaction to occur, the binding and activation energy levels are too similar for carbon
monoxide to be desorbed from the platinum surface, and the rate of carbon dioxide
formation is very slow (50). Thus, at lower temperatures, the platinum electrode is very
inactive with regards to reacting with hydrogen or another species, due to carbon
monoxide acting as a catalytic barrier.
1000
Pco. I$
S
10
1
\~.--5% IriSiO2
W
o
e
to(?
0.1t100K0/ s
$%Pd
0) 1
es
(K
)
0)
.
0.01
0.001
1.4
_____
1.8
1.8
2.0
2.2
2.4
2.8
1
1000/T (K- )
Figure 31: Platinum, palladium and iridium CO to CO 2 formation rates as a function of
inverse temperature for single crystal supported catalysts (50).
Interestingly, gold can oxidize carbon monoxide, even in a mixed gas environment. Gold
is very inactive towards methane, hydrogen, and nitrogen, but studies have shown that
gold nanoparticles can oxidize carbon monoxide (51, 52).
Therefore, while carbon
monoxide reduces the performance of the platinum, gold has the ability to oxidize carbon
monoxide to form carbon dioxide, thereby making gold act as the anode while platinum
reduces oxygen to complete the reaction. This switch in activity would explain why a
64
-7.
M-TT
-
.
, _
1.--
__
2
__-
____
__
,
I"
____
A schematic for the proposed
negative voltage is observed at lower temperatures.
reactions at lower temperatures is shown in Figure 32.
e--
(I'I11, 0I
CWH0
+ 2H 2 0
C114
I
II
4H2 + C
C1111
co
2CHA
2CO2
402
4H
Anode
2CO + 202
-> 2CO,
+ 4e
<600C
2CO+0
2
2+
Cathode
4e~ -> 202
-> 2CO 2
Figure 32: Reaction mechanisms for micro SC-SOFC w/ Au and Pt electrodes in a 1:1:5
methane:oxygen:nitrogen mixture at <600*C include water-gas shift reaction for the
methane, carbon monoxide blocks the platinum electrode which promotes oxygen
reduction, while carbon monoxide is oxidized on the gold electrode.
The role of carbon monoxide at lower temperatures also explains why no current can be
drawn from the fuel cell below 600*C. It is difficult for carbon monoxide to be removed
from the surface of the platinum electrode, thereby making it the rate limiting electrode.
There are several theories to explain why the operating OCV above 700*C is ~0.4 V
instead of the theoretical value of 0.98V.
One proposed theory is that when the
electrodes are too close to each other, the products of each electrode reaction mix and
dilute the reactants for the opposite electrode. This mixing reduces the partial pressures
of the fuel cell's reactants, which in turn reduces the fuel cells performance (44). The
effect on OCV of increasing the distance between the anode and cathode has been
studied, and shows that the OCV becomes smaller and less stable as the distance between
the electrodes gets smaller (28). Both of these studies observed OCVs around O.4V with
65
the distance between electrodes being less than half a millimeter. The other theory is that
the low OCV is directly correlated with the material properties of the electrodes.
Platinum is susceptible to coking, which can reduce the voltage over time. Additionally,
platinum and gold have high overpotentials associated with their ability to adsorb and/or
ionize oxygen (23).
There is reasonable evidence to show that the low OCV observed in these noble metal
micro SC-SOFCs is due to the effects of mixing and high overpotentials.
The effect
caused by having the electrodes close together makes these cells very susceptible to
product and reactant mixing.
Without using more selective materials, this geometric
constraint is certainly a factor that affects the observed OCV. The large overpotentials of
platinum and gold further support the observation that these materials are poor selective
catalysts and can also explain why the observed OCV is smaller than the theoretical
value.
Given these constraints, the micro SC-SOFCs created have met or surpassed
reported OCV values for platinum and gold electrodes as well as for other micro SCSOFCs (23, 28, 44). Future studies will utilize alternative materials to help improve
selectivity and reduce overpotentials.
4.2 Electrode Transformation
When the micro SC-SOFCs in a mixed gas environment of methane, oxygen and nitrogen
are operating at temperatures above 600'C, it is apparent that the electrodes undergo
physical transformation as previously shown in Figure 21.
SEM pictures of this
transformation demonstrate the changes to the platinum electrode on an individual micro
SC-SOFC, Figure 33a, a stack of 5 cells in series with a single interconnect Figure 33b,
several smaller 5 cell stacks with a single interconnect, Figure 33c, and a 5 cell stack
with multi-interconnects, Figure 33d.
Looking at the individual cells, the platinum appears to become porous, but the electrodes
do not appear to be changing in size. This observation eliminates the possibility for the
66
deposition of an introduced substance such as carbon or the formation of new compounds
such as PtN. In single cells, the platinum transforms, but can still be seen on the
electrolyte; in series stacks, however, the transformation occurs so rapidly that in most
cases, the electrodes disappear. This may be a result of platinum breakdown or its
diffusion into the electrolyte.
Ium
1plm
b
a
4
Pt
A
Al L
100pm
-4P_
1O0ijm
c
d
Figure 33: SEM images documenting the transformation of Pt-YSZ-Au micro SC-SOFC
after testing at 900*C in a 1:1:5 ratio methane:oxygen:nitrogen gas mixture. Individual
micro SC-SOFC(a), 5 400micron long micro SC-SOFCin series with single interconnect
(b), different sized single and multi-interconnect series stack micro SC-SOFC(c), multiinterconnect series stack of micro SC-SOFC(d).
67
To determine if the platinum could have diffused into the YSZ or if it formed some new
compound with the YSZ, the samples were placed into an SEM, and EDS was performed
to look for the electron scattering signature for platinum within the YSZ. Since the EDS
has a penetration depth of several micrometers and the sample's total height is less than
500nm, this technique is able to determine if any platinum is embedded within the YSZ.
Figures 34a-b show a close up of a platinum electrode that has become porous. The label
in Figure 34a indicates the location of the focused electron energy over the platinum, and
Figure 34b shows the matching EDS results. The same process is repeated in one of the
pores and is shown in Figures 34c-d. From this data, there appears to be platinum on top
of YSZ, however there does not appear to be any within the pore or within the YSZ.
Pt
YSZ
BOe
cAe 104dM Cursor
Elecron Imse I
V,64keV (37 ctmH
b
a
Pt
YSZ
SIm
FIl-c0on
c
image I
)
lS
4
e 1062
cis
Crsor 3.104
keV
(12
5
cts
kIt
d
Figure 34: SEM and EDS results to investigate platinum electrode. SEM image of
porous platinum electrode indicating where EDS measurement is measured on the
platinum surface(a), EDS results for measurement over platinum surface confirming the
presence of platinum(b), SEM image of porous platinum electrode indicating where EDS
measurement is taken within a platinum pore(c), EDS results for measurement within
platinum pore, showing no evidence for platinum within the pore(d).
68
Performing EDS over an area where only a small amount of a platinum electrode remains
from a multi-interconnect series cell further proves that platinum did not diffuse into the
YSZ.
The scanned area shows no signs of platinum, Figures 35a-b.
These results
eliminate the possibility that platinum pores were created by the platinum diffusing into
the YSZ.
I
5pm
Electron Image 1
a
3
2
1
0
u11 Scale 7351 cts Cursor: 2.967 keV (41 cts)
4
5
6
7
8
9
keV
b
platinum diffusion into YSZ. SEM
investigate
to
results
EDS
and
Figure 35: SEM
image of YSZ where platinum electrode had been; box indicates area within which EDS
is measured (a), EDS results for measurement YSZ surface, showing no evidence for
platinum(b),
69
If the transformation is not formed from platinum diffusion into the YSZ, it is possible
that transformation is caused by either a chemical reaction between the electrodes and the
gas species, or by an electrochemical reaction facilitated by an electric potential created
between the gold and platinum electrodes.
Figures 21a-b show the cells after being exposed to nitrogen and oxygen. From these
exposures, the gold is essentially unaffected and the platinum has a few small
inconsistent defects, which could be influenced by potential dust or other particle
contaminants during fabrication. These effects are minimal and do not seem correlated to
the formation of the pores in the platinum. Thus, nitrogen or oxygen must not be the sole
contributors to the transformation.
When the fuel cells are exposed to pure methane, the electrodes transform differently
than when exposed to all three gases. Figure 36 is a close up of the micro SC-SOFC
exposed to methane after being held at 850*C for over an hour. Under these conditions,
the gold electrode transforms, which is not seen during standard operation.
The
transformation is similar to wetting or melting of gold. When exposed to pure methane,
the platinum does not become porous; instead, it acquires small black features. These
features could be the result of carbon deposition onto the platinum surface from methane
decomposition. XRD results show that gold, platinum, and YSZ seem unaffected but
does not indicate the formation of new compounds or carbon deposition, Figure 37.
However, this does not discount the possibility for carbon deposition; the carbon could be
amorphous, making XRD detection difficult.
70
Figure 36: Micro SC-SOFC exposed to methane at 850C for over one hour.
with YSZ
N
CO
N
N
.D
All gas
oxygen gas
nitrogen gas
methane gas
20
30
40
50
60
70
20 (degree)
Figure 37: XRD results for samples with gold and platinum on YSZ and exposed to pure
methane, pure oxygen, pure nitrogen and a 1:1:5 ratio of methan:oxygen:nitrogen. XRD
results only detect platinum, gold and YSZ.
71
The platinum does not appear to diffuse into the YSZ, it does not visibly transform when
exposed to nitrogen or oxygen, and methane seems to primarily aid the carbon deposition
onto the platinum surface. The last parameter to evaluate is if the transformation occurs
due to an electrochemical reaction induced by the gas mixture. To evaluate this, the gold
and platinum electrodes were patterned directly onto the silicon substrate. The electrodes
were no longer electrochemically coupled by an electrolyte, and thus any transformation
cannot be caused by an electrochemical reaction. This experiment was performed twice
once with pure nitrogen and again with all three gases, shown in Figures 38a-f and
Figures 39a-g, respectively.
In both environments, the electrodes are susceptible to
transformation; however, the transformation that occurs is different for each environment.
In both gas conditions, when the gold and platinum are directly on the silicon, the gold's
transformation is very similar to its transformation in the micro SC-SOFCs exposed just
to methane as shown in Figure 36. When the platinum on silicon is exposed to the
nitrogen at 850*C, only some of the electrodes transformed, whereas when it is exposed
to a mixture of gases, all the platinum transforms. However, these transformations are
different from those seen with electrodes on an electrolyte. When examining the XRD
data for these experiments, Figure 40, it becomes clear that the compound PtSi has
formed and can be attributed to the platinum defects observed on silicon, but the data
does not indicate that any of the gases influence the deformation, and therefore this does
not eliminate the possibility that the pores are caused by an electrochemical reaction.
72
a
b
d
C
e
f
Figure 38: Optical images of Pt-Au patterned directly onto silicon with no electrolyte,
exposed to pure nitrogen at various temperatures. Individual micro SC-SOFCs at 400*C
no visible changes observed (a), individual micro SC-SOFCs at 700 0 C initiation of
platinum transformation (b), series stacks of micro SC-SOFCs at 700C more consistent
changes occur to the platinum(c), individual micro SC-SOFCs at 800*C more platinum
and all of the gold starts to transform(d), series stacks of micro SC-SOFCs at 800 0 C all
platinum and gold transform (e-f).
73
a
b
c
d
e
f
g
Figure 39: Optical images of Pt-Au patterned directly onto silicon with no electrolyte,
exposed to a 1:1:5 ratio of methane:oxygen:nitrogen at various temperatures. Individual
micro SC-SOFCs at 400*C no visible changes observed (a), individual micro SC-SOFCs
at 700*C platinum appears to deform (b), series stacks of micro SC-SOFCs at 700*C all
platinum in contact with gold does not deform(c), series stacks of micro SC-SOFCs at
750*C platinum in contact and within a certain distance of gold does not transform(d),
individual micro SC-SOFCs at 800C more platinum and all of the gold starts to
transform(e), series stacks of micro SC-SOFCs at 800*C all platinum and gold transform
(f-g).
74
without YSZ
n
NT
0
en CD
C%4
N
0-C)<
<
CL
Nitrogen gas
20
30
40
50
60
70
20 (degree)
Figure 40: XRD results for samples with gold and platinum deposited directly on silicon
(no YSZ electrolyte) and exposed to pure nitrogen or a 1:1:5 ratio of
methan:oxygen:nitrogen. XRD indicates the existence of polycrystalline Pt, Au, and the
formation of a new substance PtSi.
Regardless of the compounds formed, the changes do not correlate to the transformations
documented when the micro SC-SOFCs are exposed to methane, oxygen and nitrogen.
This confirms that the platinum transformation is caused by the electrochemical reactions
occurring within the fuel cell. Since platinum does not diffuse into the YSZ, and chemical
reactions have proven to at most deposit carbon onto the platinum versus removing the
platinum from the fuel cell, something is pushing or bubbling through the platinum to
create the pores.
Revisiting the electrochemical reactions occurring in the fuel cell
(Section 4.1), the result of the reaction of methane and oxygen is the formation of water
at the anode. Hydrogen has the ability to be absorbed by platinum (53). Therefore when
methane decomposes to form carbon dioxide and hydrogen, the hydrogen can be
75
adsorbed by the platinum and travel to the platinum-YSZ interface to meet up with an
oxygen anion to form water. At the high operating temperatures, the water is created in
vapor form, causing significant expansion and breaks through the platinum electrode.
With this understanding, it is important to consider two options. First, if it is desirable to
have a solid anode electrode, it would be important to select a stiff material that does not
soften at elevated temperatures like platinum. To avoid changing materials to improve
the stiffness, having a thicker platinum electrode might alleviate the effects from the
water vapor formation. Secondly, if a porous electrode is acceptable or desired, this
could resolve the problem, since a porous electrode would be able to accommodate the
water vapor formation and allow it to vent without changing the electrodes structure.
4.3 Carbon Nanowire Formation
In the data collection process, an interesting phenomenon occurred. Under certain
conditions, carbon nanowires were created at the platinum electrode, shown in Figures
41a-c. These were created when exposed to a ratio of 2:1:5 methane:oxygen:nitrogen at
800*C for 15-30mins. Figure 41a shows that every platinum electrode facilitated this
carbon nanowire growth. The platinum electrodes on the series stacks degraded away
before reaching 800'C, so it is unclear if these electrodes would facilitate the reaction to
create carbon nanowires. The entire platinum electrode is involved in the creation of the
nanowires. Figure 41b shows that there is no platinum remaining on the surface of the
YSZ, while the gold electrode remains unaffected.
76
a
b
C
Figure 41: SEM images of the carbon nanowire formation on the surface of the sample
with platinum being the catalysts. At a magnification of 250, it is apparent that platinum
electrodes on all individual micro SC-SOFCs have been utilized to form the carbon
nanowires (a) at higher magnification of 1000, there is no platinum remaining on the
electrolyte, and gold is unaffected (b) at a magnification of 4000, it becomes clear that
there is a tangled web of carbon nanowires(c).
77
TEM images show that the platinum is acting as the catalyst and confirms the growth of
carbon nanowires versus carbon nanotubes. Figure 42 shows a small platinum particle at
the end of the carbon nanowire structure, creating a layering of carbon in a solid
configuration.
nanotubes.
Since there are no center voids, these structures must not be carbon
Figure 43 shows that many carbon nanowires can be formed on a large
platinum particle. Performing EDS on the nanowires confirms the solid deposition of
carbon onto the platinum electrode, Figures 44a-c.
Additional investigations will be needed to completely understand the reaction
mechanism and determine what can be done to improve the formation of carbon
nanowires. Initial speculation is that the new 2:1:5 methane:oxygen:nitrogen
gas
composition is responsible for the carbon deposition. Figure 45 shows the thermal
equilibrium of the chemical species when the chamber has a 2:1 ratio of methane to
oxygen. Table 7 provides the approximate mole fractions of the main gas species under
this conditions.
Table 7: Mole fraction of the primary gas species at thermal equilibrium for a 2:1:5
mixture of methane and oxygen in nitrogen at 800C.
H2
CO
CH 4
IH 2 0
CO 2
N2
0.61
0.02
0.02
0.05
0.1
Mole Fraction 0.2
78
Figure 42: TEM image of a small platinum particle with a single carbon nanowire
growing from its surface. The solid layers of carbon prove the formation of a carbon
nanowire verses a carbon nanotube.
79
Figure 43: TEM image of a large platinum nano particle with multiple carbon nanowires
growing from its surface.
80
a
F
1~4
1C
7 C
is
is
kej
b
Element
Weight%
Atomic%
CK
OK
Si K
Pt M
Au M
515.21
30.37
24.53
74.13
10.09
93.05
4.12
1.89
0.82
0.11
Totals
654.33
C
Figure 44: SEM and EDS results to investigate platinum and carbon nanowires. SEM
images of gold electrode, YSZ electrolyte and the transformed platinum electrode with
carbon nanowires (a), EDS results taken from the platinum and carbon nanowire,
showing carbon and platinum (gold is shown because the energy levels for gold and
platinum are very similar and the software has trouble distinguishing between the two)
(b), depicts the atomic ratio of elements present, indicating substantially more carbon
than platinum(c).
81
0.25
020
2
H20
-
Equilibrium Mole traction CO2
Equilibrium Mole fraction CO
CO-
015-
Equilbrium Mole traction H
Equilibrium Mole fraction VH4
oiE 0,10
-
005
-
--
Equilibrium Mole fraction H2
-9
400
1000
900
800
700
600
500
(C)
(C1)
Temperature C1 Equilibrium Cluster1
Figure 45: CHEMKIN thermodynamic equilibrium simulation over various temperatures
for a 2:1:5 ratio of methane:oxygen:nitrogen
The formation of nanowires may be primarily facilitated by a chemical reaction, or by an
electrochemical reaction. If a chemical reaction is the basis for this phenomenon, it is
most likely the hydrogen reacting with carbon monoxide on the surface of platinum, as
shown in Eq. 5. This chemical reaction would deposit solid carbon onto the surface of
the platinum.
H2 +CO -H
2
0+CS
Eq. 5
One study suggests that a high ratio of carbon monoxide to carbon dioxide is responsible
for the carbon deposition(54). At equilibrium, methane is present at 800*C when feed
gas is 2:1 methane:oxygen, whereas no methane is present when feed gas is in a 1:1 ratio,
shown in Figure 29. This suggests that methane could be used to facilitate a chemical or
electrochemical reaction to create carbon nanowires.
Chemically, Eq. 6 shows how
methane could react to form carbon nanowires.
CH4 ->C()
82
+2H2
Eq.6
However, methane is probably not the sole cuase of carbon nanowire formation since the
micro SC-SOFC operates at 850*C in a pure methane environment and only shows some
deposition, Figure 21c, and does not react with every electrode to form carbon
nanowires, Figure 42.
If formation of carbon nanowires is due to an electrochemical reaction, the anode would
have the following half cell reaction, Eq. 7.
CH4 +2-
-> C() + 2H 20 + 2e-
Eq 7
To determine if this reaction controls the formation of the carbon nanowires, a platinum
and gold electrode sample patterned directly onto silicon was used to decouple the
electrochemical effects.
When platinum and gold on silicon were tested with 2:1
methan:oxygen gas mixture at 800*C, very few to no nanowires were formed on the
platinum.
However, when a SC-SOFC was operating, the fuel cell temperature was
higher than the furnace by about 100'C (3, 15). Thus to mimic the increase in operating
temperature, the electrodes on bare silicon were tested again at 900*C with the same gas
mixture. At 900*C, the platinum on silicon started to produce structures that resembled
nanowires.
However, the reaction was not uniform.
Increasing the temperature to
1 000*C destroyed the platinum and no visible carbon deposits were observed. It appears
in some areas that the platinum may have combined with silicon substrate to form PtSi.
This investigation is inconclusive regarding the role of electrochemical reactions in the
formation of carbon nanowires. The formation of nanowires can be attributed to either the
ratio of carbon monoxide to carbon dioxide or electrochemical reactions. Future studies
will further investigate this process in greater detail.
83
4.4 Characterization of Individual micro SC-SOFCs
4.4.1 Current-Voltage Data
The power density-current density plots with varying electrode lengths shown in Figures
22 and Figure 23 indicate, as expected, that the larger the fuel cell, the higher the power
output. In order to compare how the cells perform with respect to each other, the power
and current density are normalized by area. Area is computed using the planar
configuration area shown in Figure 8. The normalized power and current density data
for the 15micron and 5micron gaps are shown in Figure 46 and Figure 47, respectively.
A comparison of the max power densities is provided in Figure 48.
Power density (area) @ 900C
Electrolyte: YSZ
Electrodes: Au & Pt Smicron, XXXmicron length with l5micron gap
(20:20:100 ccm]
CH4:02:N2
Gas:
.. ...*
12
l0omicron trial
100micron trial
100micron trial
200rmicron trial
trial
0 200micron trial
400micron trial
- 400micron trial
+
10
+
+
W
*
j
a200micron
4
a
2
-
..
- -.
0
0
a
^
100
75
50
Currant Density (uA/cm2)
25
I
2
3
1
2
3
1
2
125
Figure 46: Power density versus current density for 5micron wide electrodes separated
by a gap distance of 15microns, experimented at 900C in a 1:1:5
methane:oxygen:nitrogen environment.
Power Density (area)@ 900C
Electrolyte: YSZ
Electrodes: Au & Pt 5micron, XXXmicron length with Smicron gap
Gas: CH4:02:N2 [20:20:100 ccm)
20
18a
-
1
6
14
P
4
g2r
100emcaon trialed
d
aa
gam
a
b
r
mtaellntcrrgn
Curntdntt
(Acm2
0
0
25
s0
7S
craptrial
m2
100
120
IS0
175
200
225
250
Carrat donsity taA/cma2)
Figure 47: Power density versus current density for 5micron wide electrodes separated
by a gap distance of 5microns, experimented at 900'C in a 1:1:5
methane: oxygen: nitrogen environment.
84
SC-SOFCs with electrodes separated by 5 microns show higher power density than those
with electrodes separated by 15 microns. This is expected since gap area increases overall
fuel cell area but is catalytically inactive and does not contribute to power. Additionally,
increasing the gap area increases the electrolyte resistance. The normalization also shows
that for each gap distance, the max power density falls within the same value range when
accounting for experimental uncertainty, Figure 48.
Maximum Power Density versus Fuel Cell Area
20
16
1E14
12
10-
1.E 05
2,E-C5
3.5-CS
4.E-05
55
E05
6 E-05
7.E-05
8.505
9.E 35
:.E-34
Fuel Cell area (cm^2)
S13micron
Gup
* 5micron~
Gup
Figure 48: Power density vs. fuel cell area. For each electrode gap distance, there is a
given power density that is within the same range for the different fuel cell sizes.
There are some clear benefits for moving to a through-plane geometry. As discussed in
Section 1.4.3, the "active area" for planar SC-SOFCs is defined to be the area of both
electrodes and the area that separates the electrode. A micro SC-SOFC in the planar
configuration with 5 micron wide electrodes separated by a 5 micron gap exhibits a final
power density at least 3 times lower than that of the analogous through-plane
configuration. In addition, electrolyte resistance decreases by several orders of magnitude
in a through-plane SC-SOFC, as the length decreases and the cross-sectional area
increases.
Resistance reduction improves fuel cell performance.
Although the above
analysis is overly simplified, it is clear that there are potential benefits for moving to a
85
through-plane geometry. However, such geometry requires more complex fabrication
methods and may have other complications such as higher mass flow limitations.
Assuming Hibino, et. al.(19) correctly calculated the area for both the planar and throughplane SC-SOFC configurations, Figure 49 (Figure 9 and Figure 12) shows that going
from a planar to through-plane SC-SOFC design may not increase the power output.
Assuming that the electrolytes conductivity, cross-sectional area, and path lengths are the
same for the planar and through-plane configuration, the power output does not increase
by a factor of three. This suggests that a through-plane configuration has additional
limitations. However the performance of a through-plane SC-SOFC can be improved by
reducing the thickness of the electrolyte layer.
800
0(
E
MY)
0
40
-f
fE
f
I)
b
a
12 showing IV data for
Figure
Figure 49: A side by side comparison of Figure 9 and
planar (a) and through-plane(b) SC-SOFCs( 19)
4.4.2 AC Impedance Data
As seen in Section 3.4, the data show a consistent electrolyte resistance for each fuel cell
under each testing condition. The previous observation made in Section 4.4.1 about the
reduction in power density with respect to an increase in the gap between electrodes is
validated using AC impedance by measuring the electrolyte resistances. Using the data
from Figures 27a-b, a resistance ratio between the fuel cells with a 5, 10, and 15Smicron
electrode gap distance is created to indicate how these resistances scale with gap size.
The ratios for the average resistance and electrolyte resistance are shown in Figure 50.
Both the average resistance and electrolyte resistance increase almost linearly with the
electrode gap distance.
86
This expected trend validates the current-voltage data results
predicted in Section 1.4.4 and shown in Table 3.
The trend between electrolyte
resistance and fuel cell size is also shown experimentally, Figure 26b. Even though the
predicted trends are validated experimentally, the measured value for the electrolyte
resistance is an order of magnitude larger than that calculated.
This suggests that either
the selected YSZ conductivity does not match that of the experimental YSZ,
experimental tolerance for measuring within the high frequency domain (the domain that
provides information about the electrolyte) is a factor, or the conductivity is suffering
from the electrolyte being too thin to allow the anions to travel freely through the
electrolyte as demonstrated in Figures 10-11 and as proposed by Chung (13).
If the
electrolyte is indeed too thin to function as predicted in Eq. 3, a new thin film
conductance model would be needed to adequately predict the electrolyte resistance for
planar micro-SC-SOFC devices.
Normalized Change in Resistance as Separation of
Electrodes Increases
3.50
3.00
2.0
2.50
M
2.00
U.2
2
L
VS
1.50
1.00
0.50
0
0.00
5 micron
10 micron
15 micron
Electrode Separation Distance (microns)
-
--*-
Electrolyte Resistance Ratio
Electrode Resistance Ratio
ideal Resistance Ratio
Figure 50: resistance ratios with respect to electrode gap separation.
To determine the rate limiting electrode, an 'Area Specific Resistance'(ASR) is extracted
from the AC impedance data. A control experiment was performed using a micro SCSOFC with 15 micron wide electrodes. When one electrode is changed to 5 microns, the
change in resistance can be attributed to the change in area of that specific electrode.
Multiplying the change in resistance by the change in area can provide an ASR for that
electrode.
Applying this technique to the data shown in Figure 25a provides the
following ASR for the platinum and gold electrodes, Table 8.
87
Table 8: Computed values for the change in resistance and the ASR
attributed to specific electrodes.
Platinum (5_15)
Gold (15_5)
Change in resistance (AQ)
[(15-15) - (xx-yy)]
Electrode specific ASR
1.03E+07 Q
3.87E+08 Q
412 Qcm 2
15463 Qcm 2
(AQ*AArea)
The platinum and gold ASR values are both several orders of magnitude larger than what
has been reported in literature(55). One potential reason for the discrepancy is the size
difference between electrodes. Herle, et. al. used gold electrodes that were 20 microns
thick with a cross-sectional area of 2.27cm 2 . In comparision, the micro SC-SOFCs have
thin film electrodes that are 200 nanometers thick, occupying an area several orders of
magnitude smaller. Since the gold is a dense thin film, the reduction in surface area from
a larger electrode can reduce its reactivity. Additionally this can be indicating the poor
selectivity of gold and platinum catalysts. Since the platinum electrode is porous, it can
reduce the resistance; however, platinum is susceptible to coking which can reduce the
reaction sites, resulting in higher resistance. If the gold were porous, it could be more
reactive and take on properties similar to gold nanoparticles, which might reduce its
resistance(52).
With the ASRs from Table 8, an average electrolyte resistance obtained from Figure
25b, and assuming an OCV of 0.4V, a current-voltage plot can be generated to show the
effect of the resistances on fuel cell performance, shown in Figure 51. In the case where
there are no losses, the voltage would remain constant over all current ranges. However,
with the introduction of the electrolyte resistance, performance does decrease slightly, but
the effect is often negligible. The same can be said with respect to the losses attributed to
the platinum electrode.
The gold cathode, however, creates the largest resistance and
limits the current output. This makes sense for several reasons: firstly, the gold electrode
is not porous so it has less surface area available for reaction. Secondly, oxygen reduction
is known to be a slow and rate limiting process. Lastly, gold can reduce oxygen, but is
ineffective. Materials such as BSCF and LSF are much better at reducing oxygen than
gold and will be investigated in future studies (34). When comparing the theoretical
current-voltage data shown in Figure 51 to collected data in Figure 52, a close
88
correlation can be seen between the analytical model and experimental data.
This
validates the approach to extracting an ASR value as well as the methods used to
characterize fuel cell operation and relate the information directly to fuel cell
performance. Finally, to emphasize how much improvement is needed at the cathode,
Table 9 provides the ratios of resistances used to create Figure 51. Table 9 shows that
the cathode resistance is tens to hundreds of times larger than the electrolyte or platinum
resistances.
Current-Voltage Predictions from Experimentally Determined Electrode
ASR:
Platinum/Gold Electrodes 5micron by 400micron
&
Platinum/Gold Electrodes 15micron by 400micron
0.45
0.4;,-
0.35
y
0
IM
no v loss
A
0.3
.
25
PtAu
Aa
.
A
S0.2
Electrolyte
Pt&Au 15micron, Cathode Losses
AA
> 0.15
A
0.1
I5micron,
Losses
Pt&Au iSmicron, Anode Losses
A
A
0
Pt&Aul Smicron, electrolyte
Losses
Pt&Au Smicron, Anode Losses
A
Pt&Au Smicron. Cathode Losses
A
0.05
0
0.OOE+00 2,OOE-10
4,00E-10
6.OOE-10 8.OOE-10
Current (A)
1.OOE-09
1.20E-09
1.40E-09
Figure 51: Analytical model of a micro SC-SOFC with 5micron and 15micron wide gold
and platinum electrodes. The parameters for this model are based on extracted ASR
values for the platinum and gold electrodes at 700*C.
Voltage verse Current for
15micron
by 400micron micro SC-SOFC
at 700C in 1:1:5 CH4:02:N2
0.410.3
i0.2
0
0.00E+00
1.OOE-10
2.00E-10
3.00E-10
4.00E-10
5.00E-10
6.00E-10
7.OOE-10
8.00E-10
9.00E-10
1.00E-09
Current (A)
Figure 52: Experimental data for a micro SC-SOFC with 15micron wide gold and
platinum electrodes at 700 0 C in a 1:1:5 ratio of methane:oxygen:nitrogen.
89
Table 9: Comparison of electrode and electrolyte resistances.
Pt Q/ YSZ Q
Au Q /Pt Q Au Q /YSZ i
Pt Au
15micron_1
5micron
37.51
193.12
5.15
5micron_5
micron
37.51
579.37
15.45
90
4.5 Stacked SC-SOFC scaling
One of the ultimate goals of this project is to be able to obtain high voltage stacks in a
very small area. Figure 28 shows the initial data evaluating a stack of 10 micro SCSOFCs attached in series. The OCV across different quantities of micro SC-SOFCs
stacked in series is shown in Figure 53, and can be fit to a trend line with slope of 0.33.
This agrees with conventional theory and it behaves similarly to reported trends in
literature(21, 26, 37). The average voltage per number of cells in the stack is
approximately 0.325V for stack sizes of 4 or more, Figure 54. This average voltage is
similar to the single cell tests previously mentioned as well as the voltages reported in
literature. To my knowledge, this is one of the highest numbers of fuel cells in a SCSOFC stack to date that maintains a consistent voltage across all of its cells. The slight
reduction in voltage for the individual and 2 cells stacked in series may possibly be
attributed to cell degradation since they were the last cells to be tested. The degradation
of stacked cells is a problem that will need to be addressed in futures studies.
Stack Voltage for 'n' Cells
3.50
-
3.00
y = 0.3304x
R
2.50
2
0.985
> 2.00
U
0
.
150
1.00
0.50
0.00
0
2
4
6
8
10
Number of Cells in a Stack
Figure 53: OCV across different quantities of micro SC-SOFCs stacked in series.
91
Unit Cell OCV for "n" Stacked Cells
0.4
U 0.35
r
0.3
U=0.25
0ow0U
n4-0.2
M 0
-
0.1
U
(U
4'0.05
0
0
2
4
6
8
10
12
# of Cells in a stack
Figure 54: average OCV across different quantities of micro SC-SOFCs stacked in
series.
92
5.0 Summary
This work demonstrates several novel and insightful results of individual and stacked
micro SC-SOFCs using noble metal electrodes: First, a change in direction of the OCV at
600'C has indicated a change in the chemical species exposed to the fuel cell electrodes.
This suggests that different electrochemical reactions are occurring on the gold and
platinum depending on temperature and gas composition. At low temperatures, the gold
and platinum electrodes are respectively performing the anodic (CO+0
2
->
CO 2 + 2e~)
and cathodic (/202 + 2e- -> 02.) reactions, while at high temperatures the gold and
platinum are hypothesized to be respectively performing the cathodic (V20
and anodic (H2+02- ->
2+
2e~ ->
02-)
H20 + 2e-) reactions. To my knowledge, this is the first reported
electrochemical potential reversal as a function of temperature. These findings suggest
the importance of electrode materials and gas composition.
Second, higher temperature results in better performance. For an individual micro SCSOFC at 900'C, an OCV of ~0.4V and a peak power density of 27p,W/cm2 was achieved.
There is a lot of room for improvement. Experiments indicate that having the smallest
electrode-electrode separation is desirable. The cathodic reaction is shown to be rate
limiting; therefore, increasing the size of the cathode by several factors may be able to
make up for its inadequate performance.
Third, at 800'C, using a 2:1 ratio of methane to oxygen, the platinum electrode has
shown coking to form carbon nanowires. It is unclear at this time if this finding is novel,
or has any new applications. However, this is a distinct byproduct of this study and is
worth investigating in the future.
Lastly, a stack of 10 micro SC-SOFCs in series produces an OCV of 3.3V, where each
fuel cell has an average OCV of approximately 0.33V.
This validates the micro
fabrication methods utilized to create micro SC-SOFC stacks, and has provided a method
to easily create the largest SC-SOFC stack reported.
This proof of concept will be
extended in the future to further increase the number of cells in a stack.
93
Among these experimental results, it is important to consider the potential of the micro
fabrication process developed to create individual and stacks of micro SC-SOFCs. By
utilizing the micro-fabrication techniques of sputtering and photolithography outlined in
Section 2.3, it will be possible to make hundreds or thousands of stacked fuel cells in an
area of less than one square centimeter. This work validates the micro fabrication stack
concept. Comparing the collected data to literature, it becomes apparent that our lack of
performance can be largely attributed to the poor selectivity of the gold and platinum
electrodes.
Therefore, with the use of more selective materials and an optimized
geometry design, this technology will gain potential to become commercially viable.
Future studies will go in two directions: materials and design. The micro SC-SOFCs will
be fabricated using materials with enhanced selectivity such as BSFC and SDC. This
should increase fuel cell performance by several orders of magnitude, from 1x0~5 watts
per square centimeter to lx10.1 watts per square centimeter. Using these new materials
will also reduce the operating temperature of micro SC-SOFCs. In addition, future work
will further develop planar design and different stacking structures, such as the throughplane configuration shown in Figure 7, as well as other geometric layouts including
three-dimensional structures and contoured surfaces.
These innovations have the
potential to create a system that is not only simple, inexpensive, and robust, but also one
that has adequate power density.
If micro SC-SOFCs can be improved to have a
comparable power density to a traditional SOFC, then micro SC-SOFCs may be able to
break down the current barriers to entry of fuel cell technology into commercial
applications.
94
95
96
References
1.
Linden, D.R., Thomas B., Handbook of Batteries. Third Edition ed,
ed. D.R. Linden, Thomas B. 2002: McGraw-Hill.
2.
Burke, K.A., Fuel Cells for Space Science Applications. American
Institute of Aeronautics and Astronautics, 2003. 2003-5938.
3.
Hibino, T., A. Hashimoto, T. Inoue, J. Tokuno, S. Yoshida, and M.
Sano, A solid oxide fuel cell using an exothermic reaction as the heat
source. Journal of the Electrochemical Society, 2001. 148(6): p.
A544-A549.
4.
Shao, Z.P. and S.M. Haile, A high-performance cathode for the next
generation of solid-oxidefuel cells. Nature, 2004. 431(7005): p. 170173.
5.
Larminie, J. and A. Dicks, Fuel Cell Systems Explained 2nd Edition.
book, 2003.
6.
Allen J. Bard, L.R.F., Electrochemical Methods Fundamentals and
Applications. Vol. 2nd edition. 2001, Hoboken, NJ: John Wiley &
Sons, Inc.
7.
Yan, J., H. Matsumoto, M. Enoki, and T. Ishihara, High-Power SOFC
Using La[sub 0.9]Sr[sub 0.1]Ga[sub 0.8]Mg[sub 0.2]0[sub 3 delta]/Ce[sub 0.8]Sm[sub 0.2]0[sub 2 - delta] Composite Film.
Electrochemical and Solid-State Letters, 2005. 8(8): p. A389-A391.
8.
Ahn, S.-J., J. Moon, J.-H. Lee, and J. Kim, Single-chambersolid oxide
micro-patterned interdigitated electrodes.
cell with
fuel
Electrochemical and Solid State Letters, 2006. 9(5): p. A228-A23 1.
9.
Asahara, S., D. Michiba, M. Hibino, and T. Yao, Single Chamber
SOFC Using BaLaIn[sub 2]O[sub 5.5] Solid Electrolyte.
Electrochemical and Solid-State Letters, 2005. 8(9): p. A449-A45 1.
10.
Buergler, B.E., M.E. Siegrist, and L.J. Gauckler. Single chamber solid
oxide fuel cells with mixed ionic electronic conducting electrolyte.
2004. Honolulu, HI, United States: Electrochemical Society Inc.,
Pennington, NJ 08534-2896, United States.
11.
Buergler, B.E., M.E. Siegrist, and L.J. Gauckler, Single chamber solid
oxide fuel cells with integrated current-collectors. Solid State Ionics,
2005. 176(19-22): p. 1717-1722.
97
12.
Choi, S.H., W.S. Kim, H.Y. Jung, S.-J. Ahn, J.-H. Lee, H.-W. Lee,
and J. Kim. Structure and conductingproperties of porous NiO films
for single chamber micro-solid oxide fuel cell. 2005. Quebec, Canada:
Electrochemical Society Inc., Pennington, NJ 08534-2896, United
States.
13.
Chung, C.Y. and Y.C. Chung, Performance characteristicsof micro
single-chamber solid oxide fuel cell. Computationalanalysis. Journal
of Power Sources, 2006. 154(1): p. 35-41.
14.
Hao, Y., C. Pantano, and D.G. Goodwin. A two-dimensional model of
a single-chamber SOFC with hydrocarbon fuels. 2005. Quebec,
Canada: Electrochemical Society Inc., Pennington, NJ 08534-2896,
United States.
15.
Hao, Y., Z. Shao, J. Mederos, W. Lai, D.G. Goodwin, and S.M. Haile,
Recent advances in single-chamber fuel-cells: Experiment and
modeling. Solid State Ionics, 2006. 177(19-25): p. 2013-2021.
16.
Hibino, T., Hashimoto, A., Yano, M., Suzuki, M., Yoshida, S., Sano,
M., High Performance Anodes for SOFCs Operating in Methan-Air
Mixture at Reduced Temperatures. Journal of Electrochemical
Society, 2002. 149(2): p. A133-!136.
17.
Hibino, T., A. Hashimoto, T. Inoue, J. Tokuno, S. Yoshida, and M.
Sano, Single-chamber solid oxide fuel cells at intermediate
temperatures with various hydrocarbon-airmixtures. Journal of the
Electrochemical Society, 2000. 147(8): p. 2888-2892.
18.
Hibino, T., A. Hashimoto, T. Inoue, J. Tokuno, S. Yoshida, and M.
Sano, A low-operating-temperature solid oxide fuel cell in
hydrocarbon-airmixtures. Science, 2000. 288(5473): p. 2031-2033.
19.
Hibino, T., H. Tsunekawa, S. Tanimoto, and M. Sano, Improvement of
a single-chamber solid-oxide fuel cell and evaluation of new cell
designs. Journal of the Electrochemical Society, 2000. 147(4): p.
1338-1343.
20.
Hibino, T., K. Ushiki, and Y. Kuwahara, New conceptfor simplifying
SOFC system. Solid State Ionics, 1996. 91(1-2): p. 69-74.
21.
Hibino, T., K. Ushiki, T. Sato, and Y. Kuwahara, A Novel Cell Design
for Simplifying Sofc System. Solid State Ionics, 1995. 81(1-2): p. 1-3.
22.
Hibino, T., S.Q. Wang, S. Kakimoto, and M. Sano, Single chamber
solid oxide fuel cell constructed from an yttria-stabilized zirconia
98
electrolyte. Electrochemical and Solid State Letters, 1999. 2(7): p.
317-319.
23.
Hibino, T., S.Q. Wang, S. Kakimoto, and M. Sano, One-chamber
solid oxide fuel cell constructed from a YSZ electrolyte with a Ni
anode and LSM cathode. Solid State Ionics, 2000. 127(1-2): p. 89-98.
24.
Hori, M., K. Nagasaka, M. Miyayama, G. Trunfio, and E. Traversa,
Evaluation of electrode performances of single-chamber solid oxide
fuel cells, in Electroceramics in Japan Viii. 2005. p. 155-158.
25.
Hori, M., K. Nagasaka, M. Miyayama, G. Trunflo, and E. Traversa,
Evaluation of electrode performances of single-chamber solid oxide
fuel cells. Key Engineering Materials, 2006. 301: p. 155-158.
26.
Horiuchi, M., S. Suganuma, and M. Watanabe, Electrochemical
power generation directly from combustion flame of gases, liquids,
and solids. Journal of the Electrochemical Society, 2004. 151(9): p.
A1402-A1405.
27.
Hyoungchul, K., S.H. Choi, W.S. Kim, J.-H. Lee, H.-W. Lee, and J.
Kim. Development of single chamber solid oxide fuel cells with
coplanar micro-electrode array. in 207th Meeting of the
Electrochemical Society. 2005. Quebec, Canada: Electrochemical
Society.
28.
Jacques-Bedard, X., T. W. Napporn, R. Roberge, M. Meunier,
Coplanar Electrodes Designfor a Single-Chamber SOFC. Journal of
Electrochemical Society, 2007. 154(3): p. B305-B309.
29.
Jacques-Bedard, X., T.W. Napporn, R. Roberge, and M. Meunier,
Performance and ageing of an anode-supportedSOFC operated in
single-chamber conditions. Journal of Power Sources, 2006. 153(1):
p. 108-113.
30.
Jasinski, P., T. Suzuki, F. Dogan, and H.U. Anderson, Impedance
spectroscopy of single chamber SOFC. Solid State Ionics, 2004.
175(1-4): p. 35-38.
31.
Napporn, T.W., X. Jacques-Bedard, F. Morin, and M. Meunier,
Operating conditions of a single-chamber SOFC. Journal of the
Electrochemical Society, 2004. 151(12): p. A2088-A2094.
32.
Napporn, T.W., F. Morin, and M. Meunier, Evaluation of the actual
working temperature of a single-chamber SOFC. Electrochemical and
Solid State Letters, 2004. 7(3): p. A60-A62.
99
33.
Napporn, T.W., S. Savoie, R. Roberge, X. Jacques-Bedard, and M.
Meunier. Single-chamber SOFC. Comparing dry and humidified
conditions. 2005. Quebec, Canada: Electrochemical Society Inc.,
Pennington, NJ 08534-2896, United States.
34.
Shao, Z., J. Mederos, W.C. Chueh, and S.M. Haile, High powerdensity single-chamber fuel cells operated on methane. Journal of
Power Sources, 2006. 162(1): p. 589-596.
35.
Shiratori, Y. and Y. Yamazaki, Study of high voltage single chamber
SOFC with series connected cells I. IV-IF characteristics of two
segments cell. Electrochemistry, 2001. 69(2): p. 92-97.
36.
Stefan, I.C., C.P. Jacobson, S.J. Visco, and L.C. De Jonghe, Single
chamber fuel cells: Flow geometry, rate, and composition
considerations.Electrochemical and Solid State Letters, 2004. 7(7): p.
A198-A200.
37.
Suzuki, T., P. Jasinski, H.U. Anderson, and F. Dogan, Single chamber
electrolyte supported SOFC module. Electrochemical and Solid State
Letters, 2004. 7(11): p. A391-A393.
38.
Suzuki, T., P. Jasinski, H.U. Anderson, and F. Dogan, Role of
composite cathodes in single chamber SOFC. Journal of the
Electrochemical Society, 2004. 151(10): p. A1678-A1682.
39.
Suzuki, T., P. Jasinski, V. Petrovsky, H.U. Anderson, and F. Dogan,
Anode supported single chamber solid oxide fuel cell in CH4-air
mixture. Journal of the Electrochemical Society, 2004. 151(9): p.
A1473-A1476.
40.
Suzuki, T., P. Jasinski, V. Petrovsky, H.U. Anderson, and F. Dogan,
Performance of a porous electrolyte in single-chamber SOFCs.
Journal of the Electrochemical Society, 2005. 152(3): p. A527-A531.
41.
Tomita, A., S. Teranishi, M. Nagao, T. Hibino, and M. Sano,
Comparative performance of anode-supported SOFCs using a thin
CeO. 9GdO. 101.95 electrolyte with an incorporatedBaCeO.8Y0. 203alpha layer in hydrogen and methane. Journal of the Electrochemical
Society, 2006. 153(6): p. A956-A960.
42.
Ahn, S.-J., Kim Yong-Bum., Moon, Jooho., Lee, Jong-Ho., Kim,
Joosun, Co-planar type single chamber solid oxide fuel cell with
micro-patterneed electrodes. Journal of Electroceram, 2006. 17: p.
689-693.
100
43.
Yamamoto,
0.,
Solid
State
Electrochemistry,
Chapter
II
"Applications",ed. P.G. Bruce. 1995: Cambridge University Press.
44.
Ahn, S.-J., J. Moon, J.-H. Lee, and J. Kim. Single-chamber solid oxide
fuel cell with micro-patternedinterdigitatedelectrodes. 2005. Quebec,
Canada: Electrochemical Society Inc., Pennington, NJ 08534-2896,
United States.
45.
Hibino, T., K. Asano, and H. Iwahara, Solid Oxide Fuel-Cell Which
Can Work in Uniform-Gas Phase Using Partial Oxidation of
Methane. Nippon Kagaku Kaishi, 1994(7): p. 600-604.
46.
Macdonald, J.R., Impedance Spectroscopy Emphasizing Solid
Materials and Systems, ed. J.R. Macdonald. 1987, New York: John
Wiley & Sons.
47.
Buergler, B.E., A.N. Grundy, and L.J. Gauckler, Thermodynamic
equilibrium of single-chamber SOFC relevant methane-air mixtures.
Journal of the Electrochemical Society, 2006. 153(7): p. A1378A1385.
48.
Virkar, A.V., J. Chen, C.W. Tanner, and J.W. Kim, The role of
electrode microstructure on activation and concentration
polarizations in solid oxide fuel cells. Solid State Ionics, 2000. 131(12): p. 189-198.
49.
Cheng, X., Z. Shi, N. Glass, L. Zhang, J. Zhang, D. Song, Z.-S. Liu,
H. Wang, and J. Shen, A review of PEM hydrogen fuel cell
contamination: Impacts, mechanisms, and mitigation. Journal of
Power Sources, 2007. 165(2): p. 739-756.
50.
Berlowitz, P.J., C.H.F. Peden, and D.W. Goodman, Kinetics of carbon
monoxide oxidation on single-crystal palladium, platinum, and
iridium. J. Phys. Chem., 1988. 92(18): p. 5213-522 1.
51.
Grisel Rudd, K.-J.W., Andrea Gluhoi, Bernard E. Nieuwenhuys,
Catalysis by Gold Nanoparticles. Gold Bulletin, 2002. 35(2): p. 3945.
52.
Haruta, M., Size- and support-dependency in the catalysis of gold.
Catalysis Today, 1997. 36(1): p. 153-166.
53.
Markovic, N.M., Ross, P.N. Jr., , Surface science studies of modelfuel
cell electrocatalysis.Surface Science Reports, 2002(45): p. 117-229.
Murray, E.P., T. Tsai, and S.A. Barnett, A direct-methanefuel cell
with a ceria-basedanode. Nature, 1999. 400(6745): p. 649-651.
54.
101
55.
102
Huang Keqin, P.Y.H., John B. Goodenough, Characterizationof ironbased alloy interconnects for reduced temperature solid oxid fuel
cells. Solid State Ionics, 2000. 129: p. 237-250.
Appendix
A. Temperature Dependence on Ion Conductivity for Selective Oxides
Temperature ()
800
1000 900
1200
s7-4T
100
700
j
600
*ZI,
E
102
S J/L7~
C
C
10
60
8.0
10.0
12.0
10IT(K ')
Temperature Dependence on Ion Conductivity for Selective Oxides (43)
103