Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Synthesis and Application of Calcium Doped Lanthanum
Strontium Titanate as Anode Support for
Fuel Cell Applications
Islamabad
A dissertation submitted to the Department of Chemistry,
Quaid-i-Azam University, Islamabad, in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
in
Physical Chemistry
by
Azra Yaqub
Department of Chemistry
Quaid-i-Azam University
Islamabad
2014
DECLARATION
This is to certify that this dissertation submitted by Ms. Azra Yaqub is accepted
in its present form by the Department of Chemistry, Quaid-i-Azam University,
Islamabad, Pakistan, as satisfying the dissertation requirements for the degree of Doctor
of. Philosophy in Physical Chemistry.
Supervisor: _____________________
Dr. Naveed Kausar Janjua
Assistant Professor
Department of Chemistry
Quaid-i-Azam University
Islamabad.
Head of Section:
_____________________
Prof. Dr. M. Siddique
Department of Chemistry
Quaid-i-Azam University
Islamabad.
External Examiner 1: _____________________
External Examiner 2: _____________________
Chairman: _____________________
Prof. Dr. Amin Badshah
Department of Chemistry
Quaid-i-Azam University
Islamabad.
IN THE NAME OF
ALLAH
THE COMPASSIONATE
THE MERCIFUL
Read in the name of thy Lord, Who created man from a clot of blood. Read! Thy Lord is most bounteous Who taught by the pen. Taught man what he did not know.
(96, 1-5)
DEDICATED
TO
MY LOVING PARENTS
Contents Page
Acknowledgements (i)
Abstract (iii)
List of Tables (v)
List of Figures (viii)
List of Abbreviations (xv)
Chapter-1 Introduction 1 - 11
1.1 Background 1
1.2 Fuel cell Research in Pakistan 3
1.3 Preceding Studies 3
1.4 Direction of Research 5
1.5 Research Objectives 7
1.6 Thesis Layout 7
References 8
Chapter-2 Fuel Cells 12 - 38
2.1 The Fuel Cell 12
2.1.1 Working 12
2.1.2 Historical background 14
2.1.3 Fuel cell characteristics 14
2.1.4 Fuel cell efficiency 17
2.1.5 Advantages of fuel cells 18
2.1.6 Applications of fuel cells 19
2.1.7 Types of fuel cells 20
2.2 Solid Oxide Fuel Cell 23
2.2.1 Operating principles of SOFC 23
Table of Contents
Contents Page
2.2.2 SOFC materials 24
2.2.3 Classification of SOFC 28
2.3 SOFC Anode 31
2.3.1 Anode triple-phase boundary 32
2.3.2 Criteria for selection of anode materials 32
2.3.3 Alternate anode materials 33
References 36
Chapter-3 Perovskite Oxides 39-54
3.1 Perovskite Oxides 39
3.1.1 Perovskite structure 39
3.1.2 Non stoichiometry in perovskites 40
3.2 Defect Chemistry of Perovskites 41
3.2.1 Defects 42
3.2.2 3.2.2 Rules for writing defect reactions 44
3.2.3 Electronic vs. ionic compensation 45
3.3 Electrical Conductivity in Oxides 45
3.3.1 Electrical conductivity 46
3.3.2. Effects 3.3.2 Effect of temperature on conductivity 49
3.3.3 3.3.3 Effect of oxygen partial pressure on conductivity 51
References
53
Chapter-4 Characterization Techniques 55-71
4.1 Thermal Gravimetric Analysis 55
4.2 X-Ray Diffraction (XRD) 55
Table of Contents
Contents Page
4.2.1 Generation of X-rays 56
4.2.2 Bragg’s law 57
4.2.3 Calculations for crystallite size 58
4.2.3 Calculations for theoretical density 58
4.3 Scanning Electron Microscopy (SEM) 59
4.3.1 Principle of SEM 59
4.4 Particle Size Analysis 60
4.4.1 Basic principle of laser diffraction 60
4.5 Dilatometry 61
4.6 Ac Impedance 62
4.6.1 Theory 62
4.6.2 Equivalent circuits 64
4.7 Electrical Conductivity Measurement 67
4.7.1 Four probe measurement 67
4.7.2 van der Pauw set up 68
4.8 Infrared Spectroscopy 69
References 70
Chapter-5 Synthesis and Characterization of LSCTA- 72-98
5.1 Introduction 72
5.2 Experimental 73
5.2.1 Sample preparation 73
5.2.2 Sample characterization 75
5.3 Results and Discussion 76
5.3.1 Thermal gravimetric analysis 76
5.3.2 X-Ray diffraction 77
Table of Contents
Contents Page
5.3.3 Particle size analysis and BET area 81
5.3.4 Scanning electron microscopy 82
5.3.5 Dilatometric analysis of LSCTA- samples 84
5.3.6 Ac impedance 86
5.3.7 Dc conductivity 91
5.4 Conclusions 96
References 97
Chapter-6 Aqueous Tape Casting 99-124
6.1 Introduction 99
6.2 Experimental 101
6.2.1 Aqueous tape casting of LSCTA- powder 101
6.2.2 Lamination and sintering 102
6.2.3 Impregnation procedure 102
6.2.4 Conductivity measurement of bars 103
6.3 Results and Discussion 104
6.3.1 Aqueous based slurry characteristics 104
6.3.2 Microstructure of dense and porous tapes 106
6.3.3 Conductivity of bars 107
6.3.4 Effect of impregnates on the kinetics of conductivity
evolution
116
6.3.5 Comparison of conductivity 120
6.4 Conclusions 122
References 123
Chapter-7 Microstructure Optimization with Pore Formers 125-146
7.1 Introduction 125
7.2 Microstructure Optimization with Commercial Pore Formers 127
Table of Contents
Contents Page
7.2.1 Experimental 127
7.2.2 Results and discussion 129
7.3 Microstructure Optimization with Synthesized Carbon
Microspheres as Pore Former
132
7.3.1 Experimental 133
7.3.2 Results and discussion 133
7.4 Conclusions 143
References 144
Chapter-8 Symmetrical and Button Cell Testing 147-185
8.1 Introduction 147
8.2 Electrochemical Impedance Spectroscopy for Symmetrical
and Button Cell Characterization
149
8.3 Symmetrical and Button Cell Testing 151
8.4 Symmetrical Cell Testing 151
8.4.1 Experimental 152
8.4.2 Results and discussion 155
8.5 Button Cell Testing 166
8.5.1 Experimental 166
8.5.2 Results and discussion 170
8.6 Conclusions 183
References 184
Chapter-9 Synthesis and Characterization of Doped Analogues of
LSCTA-
186-200
9.1 Introduction 186
9.2 Experimental 187
Table of Contents
Contents Page
9.3 Results and Discussion 188
9.3.1 X-Ray diffraction 188
9.3.2 Scanning electron microscopy 190
9.3.3 Dilatometry 191
9.3.4 Electrical conductivity 193
9.5 Conclusions 197
References 199
Chapter-10 Conclusions and Recommendations 201-204
10.1 Final Remarks 201
10.2 Conclusions 202
10.3 Recommendations for Future Research 203
Appendix-A8 205-207
List of Publications 208
i
All praises to Almighty Allah, the omnipotent, the omniscient and the creator of
the universe Who enabled me to complete this research project. Peace, blessings and
salutations upon His last and beloved Prophet, Hazrat Muhammad (peace be upon him)
who guided us to the perfect code of life.
I consider myself very fortunate to work under the supervision and guidance of
Dr. Mrs. Naveed Kausar Janjua whose personal interest and valuable suggestions
enabled me to complete this tedious work. She encouraged all my attempts in designing
this research work and helped me at each and every stage of my project.
I would like to express my sincere gratitude to Prof. Dr. Amin Badshah,
Chairman Department of Chemistry and Prof. Dr. M. Siddique, Head of Physical
Section, Department of Chemistry, Quaid-i-Azam University for all the facilities.
It is an honour for me to work under supervision of Prof. Dr. John TS Irvine
during my stay in University of St-Andrews, Scotland, UK where I had full access to all
the facilities in his group. I would like to thank him for his support, thought provoking
guidance and worthy discussions. I am thankful to Dr. Cristian Savaniu for his help from
the beginning of the project till its end, from synthesis to testing and for the fruitful
discussions. Thanks are due to Dr. Maarten Verbraeken especially for teaching me the
steps of tape casting and for useful discussion for symmetrical and button cell testing. I
owe my compliments to Dr. Paul Connor who was always there to answer my relevant
and non-relevant “quick” questions in detailed way. I would highly appreciate Dr. David
Miller for his help in conductivity measurements.
For financial support, I would like to thank Higher Education Commission of
Pakistan for the indigenous scholarship and IRSIP scholarship which enabled me to do
my research work at University of St-Andrews, Scotland, UK.
ii
I express my gratitude to Sylvia Williamson for demonstrating me operation of
TGA and dilatometry. We always had hours of discussion on variety of topics, from
politics to culture. I owe my thanks to Ross Blackley especially for his patience while
demonstrating me SEM. Lab technician Julie Nairn was bit strict but kind enough to
help me in different lab related problems. I would like to thank George Anthony who was
always willing to fix the problems which I had with the testing jigs.
I would like to extend my thanks to the Dr John (JTSI) research group especially
Herald, Ahmed, Elena, Dragos, Chengsheng, Lanying and Fedrica. Thanks are due to
my friends Misbah, Mazlina, Sana, Mujeeba and Khadija for their moral support and
whose company never ever made me home sick during my stay in UK.
I would like to express my thanks to my lab fellows and friends in Pakistan,
Ayesha, Sadia, Humaira, Maryam, Farhat, Fouzia and Javeria for the support, help
and co-operation.
This study would not have been possible without the prayers, love and affection
of my family members especially my parents who always supported me and boosted up
my morale. No words can express their care, support and sacrifices. Without their
constant support and encouragement, I would not have been able to accomplish this task.
I would like to thank my brother Tayyab, sister Tahira and Bhabi Najma for the support
given to me throughout my whole studies and especially during thesis write up. I always
enjoyed the company of my nieces Rafia and Tooba who were a source of joy and
happiness whenever I felt depressed.
AZRA YAQUB
iii
La0.2Sr0.25Ca0.45TiO3 is a carefully selected composition to provide optimal
processing and electrical characteristics for use as an anode support in solid oxide fuel
cells (SOFCs). In the present study, the optimization of the preparation process of A-site
deficient perovskite, La0.2Sr0.25Ca0.45TiO3 (LSCTA-) powders and their characterization
for integration into the SOFC anode supports have been focussed. LSCTA- powder was
investigated in different yet connected important aspects using high-tech methods like
tape casting, microstructure optimization and testing in symmetrical and button cell set
ups.
The major part of the present research deals with the process optimization of
LSCTA-. A modified Pechini method was successfully applied to produce single phase
perovskite at 900 oC. The effect of calcination temperature on the phase, morphology and
sintering characteristics was studied using XRD, SEM and dilatometry techniques. The
optimal calcination temperature of 1000 o
C was selected for further studies as the powder
calcined at this temperature displayed a similar sintering profile to commercial 8 mol%
yttria-stabilized zirconia (YSZ), the typical choice for electrolyte. LSCTA- showed an n-
type conduction nature where conductivity of a dense LSCTA- specimen sintered in air
increased by three orders of magnitude after in-situ reduction in 5% H2/Ar. These
encouraging characterization results supported the SOFC anode candidateship of LSCTA-.
In the second part of study, the synthesized powder was processed in aqueous tape
casting which is a quick and rapid technique to fabricate thin SOFC anodes. Slurry
formulation was optimized for both the dense and porous green tapes. The rectangular
bars fabricated from green tapes by lamination were sintered and tested for conductivity
measurements using van der Pauw set up. The effect of ceria impregnation on the
conductivity of porous LSCTA- bars was studied. The conductivity behaviour of porous
bars under redox cycling showed a two-stage process that exhibited strong reversibility.
For the reduction process, addition of impregnated ceria reduced the onset delay period
and increased the apparent rate constant, k values by 30-50% for both stages. The co-
impregnation of Ni further resulted in an increase of conductivity of porous bars.
iv
Another aspect of the study was the microstructure optimization of LSCTA- tapes.
To introduce the porosity in LSCTA- tapes, commercial pore formers like graphite,
polymethylmethacrylate (PMMA) and glassy carbon (GC) were used. It was observed
that pre-sintering the powder helps to get a good microstructure with commercial pore
formers. An interesting feature for inducing porosity in LSCTA- tapes was the synthesis
of homogeneous and well dispersed carbon micro spheres (CMS) from an optimized
hydrothermal method and their further application as pore formers.
As a part of the research, the anode performance of LSCTA- was tested in YSZ
electrolyte supported symmetrical cells. The effect of impregnates like ceria (CeO2),
gadolinium doped ceria (CGO), with and without Ni, on the performance of symmetrical
cells was investigated. It was found that co-impregnation of CeO2 and CGO with Ni have
pronounced effect in decreasing the impedance of bare LSCTA- in symmetrical cells.
Further, the anode performance was tested in button cells using a three electrode set up.
A significant improvement in cell performance could be achieved by optimizing the
anode support with various impregnates both qualitatively and quantitatively.
Finally, LSCTA- was doped at B site with Ni (LSCTN) and Fe (LSCTF). The
doped compositions offered higher conductivity values than the parent LSCTA-.
Compared to pre-reduced LSCTA- having conductivity of 38 S cm-1
, the pre reduced 5%
Ni doped LSCTA- (LSCTN-5) and 5% Fe doped LSCTA- (LSCTF-5) offered conductivity
values of 47 S cm-1
and 66 S cm-1
at 880 oC, respectively.
In conclusion, structurally stable LSCTA- could be a good alternative to state of
the art SOFC anode exhibiting good mechanical, morphological and electrical properties.
Catalyst introduction via impregnation or doping could enhance the electrical and
catalytic properties of these perovskites making them viable alternatives for
electrochemical applications.
v
Table Title Page
Chapter-2
2.1 Different types of fuel cells 22
Chapter-3
3.1 Kröger-Vink notation for point defects in binary oxide, MO 43
Chapter-4
4.1 Relations between the four basic immittance functions 64
4.2 Impedances and admittances of different circuit elements 65
4.3 Typical capacitance values and the corresponding phenomena 65
Chapter-5
5.1 Crystallite size, mean particle size and BET area of LSCTA- samples 82
5.2 Shrinkage percentages and relative density values for LSCTA- samples 85
5.3 Activation energy, Ea calculated from ac impedance 91
5.4 Conductivity value of LSCTA- pellets under different conditions at 880 oC 94
Chapter-6
6.1 Tape casting recipe for LSCTA- anode substrate 102
6.2 Codes of the bars used in present study 103
6.3 Rate constant k (cm s-1
) calculated for two fold relaxation kinetics for
oxidation cycles of La0.2Sr0.25Ca0.45TiO3 (LSCTA-) and CeO2 impregnated
LSCTA- (LSCTA-:CeO2) at 880 oC
119
List of Tables
vi
Table Title Page
6.4 Rate constant k (cm s-1
) calculated for two fold relaxation kinetics for
reduction cycles of La0.2Sr0.25Ca0.45TiO3 (LSCTA-) and CeO2 impregnated
LSCTA- (LSCTA-:CeO2) 880 oC.
119
6.4 Conductivity of bars in air and 5% H2/Ar at 880 °C 121
Chapter-7
7.1 Recipe for YSZ ink 128
7.2 Set of optimized parameters for synthesis of carbon microspheres from
hydrothermal treatment of sucrose at 180 oC
140
7.3 Porosity calculated in the back scattered images of sintered LSCTA- tapes
containing carbon microspheres as pore formers
142
Chapter-8
8.1 Slurry recipe for LSCTA- ink 152
8.2 Symmetrical cells studied 154
8.3 EIS derived polarization resistance of impregnated symmetrical cells in
air at 850 oC
157
8.4 Polarization resistance of symmetrical cells in 5% H2/Ar at 850 oC 160
8.5 Slurry recipes for LSM and LSM/YSZ inks 167
8.6 Types of button cells studied 168
8.7 OCV values for button cells A & B at different conditions at 850 oC 170
8.8 Polarization resistances extracted from EIS spectra of button cells under
different conditions 172
8.9 Values of OCV for button cells at different temperatures with humidified
H2 as fuel at anode and air at cathode 174
8.10 Energy of activation, Ea calculated from resistance-temperature plots 177
List of Tables
vii
Table Title Page
Chapter-9
9.1 Studied doped analogues of LSCTA- 187
9.2 Atomic and ionic radii of cations 189
9.3 Unit cell parameters for doped analogues of LSCTA- 189
9.4 Shrinkage percentage of doped analogues in air calculated from
dilatometric data 192
9.5 Conductivity of doped analogues upon in-situ reduction in reducing
atmosphere (5% H2/Ar) at 880 oC
194
9.6 Conductivity of pre-reduced doped analogues in reducing atmosphere
(5% H2/Ar) at 880 oC
195
9.7 Standard reduction potentials of redox couples in doped samples 196
viii
Figure Title Page
Chapter-2
2.1 Schematic of fuel cell operation 13
2.2 Ideal and actual fuel cell voltage/current characteristics 15
2.3 Schematic of an oxide ion conducting solid oxide fuel cell 23
2.4 Diagrammatic presentation of tubular SOFC 29
2.5 Schematic of a planar SOFC 29
2.6 The configurations possible for the two reactants’ flow in a fuel cell;
a) co-flow and b) cross flow 30
2.7 SOFC configurations depending on support; a) electrolyte supported
and b) anode supported 31
Chapter-3
3.1 Corner sharing octahedra in perovskites 40
3.2 Thermodynamics of defect formation 42
3.3 Approximate temperature dependence of mobility with lattice and
impurity scattering 50
3.4 Brouwer diagram showing dependence of defect concentration on
oxygen activity in case of n and p-type oxides 51
3.5 Effect of oxygen partial pressure on the partial conductivity 52
Chapter-4
4.1 Working of an X-ray powder diffractometer 57
4.2 Incident and reflected X-rays from a specified crystal plane 58
4.3 Schematic of scanning electron microscopy 60
4.4 Scattering of beam from large and small particles 61
4.5 Functional diagram of a pushrod dilatometer 62
4.6 Impedance represented in Nyquist plot 63
List of Figures
ix
Figure Title Page
4.7 Some typical equivalent circuits and the impedance in complex plane 66
4.8 Four probe set up for conductivity measurement 67
4.9 Schematic of van der Pauw set up 68
Chapter-5
5.1 a) TGA (dashed line) and DTA (solid line) and b) TGA of sample
after calcination at 1000 °C 76
5.2 XRD patterns of LSCTA- synthesized via; a) solution phase Pechini
method and b) solid state route 77
5.3 XRD patterns of LSCTA-; a) before reduction and b) after reduction 78
5.4 XRD patterns after firing at 1400 oC for; a) pure YSZ, b) LSCTA- and
c) 1:1 mixture of LSCTA-:YSZ 79
5.5 X-ray diffraction patterns of LSCTA- calcined in air at various
temperatures; a) 900 oC (S1), b) 950
oC (S2), c) 1000
oC (S3) and d)
1100 oC (S4)
80
5.6 Particle size distribution of LSCTA- calcined at various temperatures;
a) 900 oC (S1), b) 950
oC (S2), c) 1000
oC (S3) and d) 1100
oC (S4)
81
5.7 Micrographs of LSCTA- powder after calcination at various
temperatures; a) 900oC (S1), b) 950
oC (S2), c) 1000
oC (S3) and d)
1100oC (S4)
83
5.8 SEM micrographs showing effect of sintering at 1400 °C on LSCTA-
powders calcined at temperatures; a) 900 oC (S1), b) 950
oC (S2), c)
1000 oC (S3) and d) 1100
oC (S4)
84
5.9 Dilatometric sintering curves of pellets from LSCTA- powder calcined
at various temperatures in air; a) 900 oC (S1), b) 950
oC (S2), c) 1000
oC (S3), d) 8-YSZ and e) 1100
oC (S4)
85
5.10 Cole Cole plots of air sintered samples in frequency range of 1 Hz to
13 MHz at different temperatures; a) S1, b) S2, c) S3 and d) S4 87
5.11 Dependence of imaginary part of impedance on frequency for air
sintered samples (S1 to S4) in frequency range of 1 Hz to 13 MHz in
air at different temperatures; a) S1, b) S2, c) S3 and d) S4
88
5.12 Dependence of real part of impedance on frequency for air sintered
samples (S1 to S4) in frequency range of 1 Hz to 13 MHz in air at
different temperatures; a) S1, b) S2, c) S3 and d) S4
89
List of Figures
x
Figure Title Page
5.13 Arrhenius dependence of conductivity calculated from ac impedance
for LSCTA- samples calcined at various temperatures 90
5.14 Temperature dependence of conductivity of LSCTA- (S3) sintered
pellet in air 92
5.15 Conductivity profile of in-situ reduced LSCTA- (S3) pellet in 5%
H2/Ar at 880 °C 93
5.16 Conductivity profile of pre-reduced LSCTA- (S3) during
thermocycling in 5% H2/Ar 93
5.17 Conductivity profile of LSCTA- (S3) sintered in 5% H2/Ar in reducing
atmosphere upon heating 95
5.18 Micrographs of LSCTA- (S3) pellet sintered at 1400 °C in reducing
atmosphere of 5% H2/Ar under different magnifications 95
Chapter-6
6.1 Schematic of a laboratory tape-casting set-up 101
6.2 Particle size analysis of LSCTA- slurry; a) in the absence and b) in the
presence of PMMA 105
6.3 Viscosity profile of LSCTA- slurry; a) in the absence and b) in the
presence of PMMA 105
6.4 Visual effect of sintering on green samples; a) before and b) after
sintering at 1400°C in air 106
6.5 Micrographs of surface view of dense tape (~ 92% ρth ) at different
magnifications; a) 1500X and b) 3500X 106
6.6 Micrographs of porous tape (~76% ρth ).; a) surface view and b) cross
sectional view 107
6.7 Conductivity profile of bar A; a) in air and b) in 5% H2/Ar 108
6.8 Conductivity profile of bar B; a) in air and b) in 5% H2/Ar 109
6.9 Redox cycling of bar A as a function of time at 880 oC. Dashed lines
show change of partial pressure of oxygen over time 110
6.10 Conductivity profile of bar C; a) in air and b) in 5% H2/Ar 111
6.11 Redox cycling of bar C as a function of time at 880 oC. Dashed lines
show change of partial pressure of oxygen over time 112
List of Figures
xi
Figure Title Page
6.12 Micrographs of CeO2 impregnated bar 112
6.13 Conductivity profile of bar D; a) in air and b) in 5% H2/Ar 113
6.14 Micrographs of CeO2 -Ni co-impregnated bar 114
6.15 Redox cycling of bar D as a function of time at 880 oC. Dashed lines
show change of partial pressure of oxygen over time 114
6.16 Thermocycling of bar E in 5% H2/Ar 115
6.17 Redox cycling of bar E as a function of time at 880 oC. Dashed lines
show change of partial pressure of oxygen over time 115
6.18 Resistivity variation vs. time for LSCTA- and ceria impregnated
LSCTA- at 880 oC on; a) 5 oxidation and b) 5 reduction cycles
117
6.19 Resistivity/conductivity relaxation of LSCTA- and ceria impregnated
LSCTA- at 880 oC upon; a) oxidation and b) reduction. Two different
kinetic processes are indicated by dotted lines with different slopes.
118
6.20 Conductivity profile of bars; a) Bar A, b) Bar B, c) Bar C, d) Bar D
after in situ reduction in 5% H2/Ar at 880 oC
121
Chapter-7
7.1 Sintering profile for green LSCTA- samples in air 128
7.2 Micrographs of LSCTA- porous green tapes after sintering at 1400 oC.
Amount of pore formers in green tapes being; a) 20 wt% PMMA + 10
wt% Graphite and b) 30 wt% Graphite
129
7.3 a) Micrographs of LSCTA- porous green tapes containing 40 wt%
Graphite after sintering at 1400 oC and b) Corresponding sintering
profile
130
7.4 Sintering profile for green LSCTA- samples containing 30% graphite
in 5% H2/Ar 130
7.5 Micrographs of LSCTA- tape after sintering at 1400 oC. Slurry
formulation was prepared using; a) calcined LSCTA- powder and b)
thermally treated LSCTA- powder
131
7.6 Effect of co-sintering at 1400 °C; a) LSCTA- co-laminated with YSZ
and b) LSCTA- with screen printed YSZ 132
7.7 XRD pattern of carbon spheres synthesized from hydrothermal
treatment of 0.5 M sucrose solution 134
List of Figures
xii
Figure Title Page
7.8 TGA of carbon spheres synthesized from hydrothermal treatment of
0.5 M sucrose solution. 134
7.9 FTIR of carbon spheres synthesized from hydrothermal treatment of
0.5 M sucrose solution 135
7.10 Micrographs of carbon spheres synthesized from hydrothermal
treatment of sucrose solution. The concentration of sucrose solution
being; a) 0.1 M, b) 0.5 M, c) 1.0 M, d) 1.0 M (on high magnification)
137
7.11 Micrographs of carbon spheres synthesized from hydrothermal
treatment of 0.5M sucrose solution at different pH; a) pH 4, b) pH 10,
c) pH 7
138
7.12 Micrographs of carbon spheres synthesized from hydrothermal
treatment of sucrose solution in the presence of different solvent
media; a) H2O, b) H
2O:EtOH = 1:2 and c) H
2O:EtOH = 2:1
139
7.13 Micrographs of LSCTA- tape after sintering at 1400 °C using CMS as
pore former, concentration of CMS being; a) 10 wt%, b) 20 wt%, c)
30 wt%, d) 30 wt% (Graphite:HT-C=1:1)
141
7.14 Micrographs of porous LSCTA- tape (with 20 wt% CMS) co-laminated
with YSZ after sintering at 1400 °C 142
Chapter-8
8.1 Cole-Cole representation of impedance showing ohmic (Rs),
polarization (Rp) and total resistance (RT)
150
8.2 Diagrammatic representation for; a) symmetrical cell and b) button
cell
151
8.3 LSCTA- in electrolyte supported symmetrical cell with gold contacts 154
8.4 Nyquist plot of symmetrical cell A with LSCTA- electrodes in 5%
H2/Ar at 850 oC in the frequency range of 50 mHz to 1 kHz
155
8.5 a) Plot of Z// vs. Z/ in the frequency range of 50 mHz to 1 kHz and b)
Z// vs. log f for impregnated symmetrical cells in air at 850 oC
156
8.6 Plot of Z// vs. Z for impregnated symmetrical cells (B-D) in the
frequency range of 50 mHz to 1 kHz; a) before, b) after 10 min and c)
after 10 hours of in-situ reduction using 5% H2/Ar at 850 oC
158
8.7 a) Cole Cole plots of impregnated symmetrical in 5% H2/Ar at 850 oC
cells and b) under magnification in the frequency range of 50 mHz to
1 kHz
159
List of Figures
xiii
Figure Title Page
8.8 Nyquist plots for impregnated symmetrical cells at different
temperatures in reducing atmosphere (5% H2/Ar) in the frequency
range of 50 mHz to 1 kHz
162
8.9 Dependence of Z// on frequency for impregnated symmetrical cells in
reducing atmosphere ( 5% H2/Ar) at 850 oC
163
8.10 Experimental and simulated impedance spectra of symmetrical cells in
reducing atmosphere (5%H2/Ar) at 850 oC; a) cell B, b) cell C, c) cell
D and d) cell E in the frequency range of 50 mHz to 1 kHz
164
8.11 Variation of resistances extracted from fit models with temperature (in
5% H2/Ar) for; a) cell B, b) cell C, c) cell D and d) cell E
165
8.12 Diagrammatic presentation of fabricated button cells using LSCTA- as
anode support
167
8.13 Impedance spectra of button cells under different conditions at 850
oC
in the frequency range of 0.1 Hz to 1 MHz; a) cell A and b) cell B 171
8.14 Plots of cell potential and power density as a function of current
density for button cells under different conditions at 850 oC; a) cell A
and b) cell B
173
8.15 Impedance spectra of button cells under OCV at different
temperatures with humidified H2 at anode and air at cathode in the
frequency range of 0.1 Hz to 1 MHz; a) cell A and b) cell B
175
8.16 The total resistance Rt, the polarization resistance Rp, and the ohmic
resistance Rs of button cells determined from impedance plots at
different temperatures; a) cell A and b) cell B
176
8.17 Plots of cell potential and power density as a function of current
density for button cells at different temperatures with humid H2 as fuel
at anode and air at cathode; a) Cell A and b) cell B
178
8.18 Plots of OCV as a function of number of redox cycles for button cells;
a) cell A and b) cell B
179
8.19 Plots of Rs and Rp a function of number of redox cycles for button
cells; a) cell A and b) cell B
180
8.20 Experimental and simulated impedance spectra of button cells in
reducing atmosphere (humid H2) at 850 oC in the frequency range of
0.1 Hz to 1 MHz; a) cell A and b) cell B
181
8.21 Micrographs of button cells; a) cell A and b) cell B after cell tests at
850 oC
182
List of Figures
xiv
Figure Title Page
Chapter-9
9.1 XRD patterns of doped analogues of LSCTA-; a) LSCTA-, b) LSCTN1,
c) LSCTN5, d) LSCTF1 and e) LSCTF5 188
9.2 Micrographs of doped analogues of LSCTA-; a) LSCTN1, b) LSCTN5,
c) LSCTF1 and d) LSCTF5 after calcination at 1000 oC for 5 hours
190
9.3 Micrographs of doped analogues of LSCTA-; a) LSCTN1, b) LSCTN5,
c) LSCTF1 and d) LSCTF5 after sintering at 1400 oC for 6 hours
191
9.4 Dilatometric sintering curves doped analogues of LSCTA- in air; a)
LSCTN1, b) LSCTN5, c) LSCTF1 and d) LSCTF5 192
9.5 Conductivity profile of in-situ reduced LSCTN1 pellet in 5% H2/Ar at
880 °C 193
9.6 Conductivity profile during thermocycling of pre-reduced samples in
reducing atmosphere (5% H2/Ar); a) LSCTN5 and b) LSCTF5 195
9.7 Micrographs of pre reduced samples; a) LSCTN5 and b) LSCTF5 196
xv
A Area of the electrode
AC Alternate Current
AFC Alkaline Fuel Cells
BET Brunauer, Emmett and Teller
C Capacitance
CGO Gadolinium doped ceria
CHP Combined Heat and Power
CMS Carbon Micro Spheres
CPE Constant Phase Elements
CTE Coefficient of Thermal Expansion
DC Direct Current
DMFC Direct Methanol Fuel Cell
DSC Differential Scanning Calorimetry
DTA Differential Thermal Analysis
Ea Activation Energy
EC Electrochemical
EDS Energy Dispersive X-ray Spectroscopy
F Faraday constant
FTIR Fourier transformed Infra Red spectra
h Planck constant
IS Impedance Spectroscopy
IT-FC Intermediate Temperature Fuel Cells
J Imaginary unit
k Boltzmann constant
l Thickness of the pellet
LSCTA- A site deficient Calcium doped Lanthanum Strontium Titanate,
La0.2 Sr0.25 Ca0.45 TiO3
LSCTF1 La0.2 Sr0.25 Ca0.45 Ti0.99 Fe0.01 O3
LSCTF5 La0.2 Sr0.25 Ca0.45 Ti0.95 Fe0.05 O3
List of Symbols and Abbreviations
xvi
LSCTN1 La0.2 Sr0.25 Ca0.45 Ti0.99 Ni0.01 O3
LSCTN5 La0.2 Sr0.25 Ca0.45 Ti0.95 Ni0.05 O3
LSM Lanthanum Strontium Manganite
LST Lanthanum Strontium Titanate
MCFC Molten Carbonate Fuel Cells
me Mass of an electron
MIEC Mixed Ionic and Electronic Conductor
NA Avogadro number
NTCR Negative Temperature Coefficient of Resistance
OCV Open Circuit Voltage
PAFC Phosphoric Acid Fuel Cells
PF Pore Former
PMMA Poly Methyl Methacrylate
PVA Polyvinyl Alcohol
PVB Polyvinyl Butyral
R Gas constant
Rp Polarization resistance
Rs Ohmic resistance
SEM Scanning Electron Microscopy
SOFC Solid Oxide Fuel Cell
T Absolute temperature
TGA Thermogravimetry Analysis
TMA Thermomechanical Analysis
TPB Triple Phase Boundary
XRD X-Ray Diffraction
YSZ Yittria-Stabilized Zirconia
Z/ Real impedance
Z// Imaginary impedance
µi Mobility of ion
β Full width at half maximum of the peak (in radians)
η Viscosity
List of Symbols and Abbreviations
xvii
θ Incident angle
λ Wavelength of the light
ρe Experimental density
ρth Theoretical density
σ Conductivity
ω Angular frequency
Ω Resistance
Chapter 1
Introduction
1
Introduction
Abstract
Keeping in mind the importance of fuel cells, the present chapter addresses their
role to compete global energy crisis. In this chapter, a brief literature review of the
preceding studies carried out in the development of anode materials for solid oxide fuel
cells is detailed. At the end, objectives and layout of the whole thesis are presented.
1.1 Background
Fossil fuels like petroleum, oil, natural gas and coal have been used as the main
energy source in different sectors including industry, utilities, transportation and others
throughout the 20th century. Conventional methods to convert fossil fuels to useful
energy include internal combustion engines, gas turbines and steam turbines which suffer
from some serious draw backs; mainly high levels of pollution, low efficiencies due to
Carnot limitation and a fast depletion of these non-renewable energy sources [1-3].
Keeping in view the fact that global energy demand increases every year, serious efforts
are required to meet increasing energy needs [4-5]. Thus research has been focused to
search for alternate fuel and energy conversion systems.
Among alternate fuels, significant attention has been given to hydrogen and
biomass because they are not only environment friendly but they can also reduce
dependency on fossil fuels [6, 7]. In the quest for efficient energy conversion systems,
electrochemical fuel cells, and particularly solid oxide fuel cells (SOFCs), have been
given significant attention as the chemical energy of the fuel is directly converted into
electricity [8-11]. Fuel cells can reduce air pollution since they do not emit NOx and
SOx. They operate quietly thus can alleviate noise pollution as well. The efficiencies of
fuel cells are much higher than internal combustion engines with normal efficiencies of
slightly over 35%. Fuel cells particularly SOFCs can easily reach efficiencies greater than
60% owing to the fact that they are involved in conversion of chemical energy from fuel
to electrical energy without involving any intermediate steps [12, 13]. Furthermore, even
higher efficiencies could be obtained by integration of fuel cells with other technologies
CH-1 Introduction
2
[14, 15]. Thus, fuel cells have significant advantages over conventional power generation
systems. Among other benefits offered by fuel cells are minimal placing restriction,
modularity and portability.
Solid oxide fuel cells (SOFCs) require high operating temperatures, usually
within the range of 600-1000 °C, to achieve reasonable ionic conductivity in the yttria-
stabilized zirconia (YSZ), the usual choice for the electrolyte (~0.1 S cm-1
) [16-18]. The
use of high temperature has some advantages which make SOFCs more attractive from
application point of view. One of these is the fuel flexibility, thus virtually any fuel
(besides hydrogen) can be fed to the anode making external fuel reforming equipment
unnecessary. SOFCs are less sensitive to fuel impurities and catalyst poisons, such as
sulfur and carbon monoxide owing to the high temperature. Thus SOFCs can operate
with CO as a fuel which acts as poison to most cells [19, 20]. One other aspect of high
temperature is the potential for a cogeneration system or combined heat and power
system (CHP) [21, 22].
The development of the fuel cell and in particular SOFCs has included a major
investigation of component materials, their fabrication techniques and cell designs [23-
27]. The typical SOFC materials used at present are yttria-stabilized zirconia (YSZ) as
the electrolyte, strontium-doped lanthanum manganite (La1-xSrxMnO3) as cathode,
nickel/zirconia cermet as anode, and calcium-doped lanthanum chromite (La1-xCaxCrO3)
as interconnect. Searching for new SOFC materials with improved properties is still an
active area of research to overcome the limitations of these SOFC components [28-30].
Ni/YSZ cermet is regarded as the state of the art anode material for solid oxide fuel cell
(SOFC) due to low cost, good catalytic activity, high ionic and electronic conductivities
and better chemical and mechanical compatibility with other cell components [31].
However, it has some inherent drawbacks: upon redox cycling anode degradation occurs
due to large and facile Ni to NiO volume change, low tolerance to sulphur also limits the
application of this anode in SOFC conditions and its high catalytic activity causes carbon
deposition when hydrocarbons are used as fuels. Moreover, at high operating
temperatures, the catalytic active surface area decreases due to agglomeration and
sintering of Ni [32]. All of these factors affect the anode performance and long term
stability of SOFC.
CH-1 Introduction
3
Thus researchers are motivated to design alternative anode systems to overcome
the limitations of Ni/YSZ cermet without compromising the electrical conductivity and
stability of SOFC anodes. Ample literature data is available on the development of
different anode materials for SOFCs, (as discussed in chapter 2).
1.2 Fuel Cell Research in Pakistan
Like any country, energy is the lifeline of Pakistan’s economy where natural gas,
oil, hydro, nuclear, coal and liquefied petroleum gas (LPG) contribute to 48.3%, 32.1%,
11.3%, 7.6% and 0.6% of primary energy supplies respectively [33, 34]. Since Pakistan
has to spend huge foreign exchange reserves on importing oil, thus it is essential to
reduce the dependency on continuously increasing imports of oil to strengthen the
economy. At the moment, Pakistan is facing an energy crisis which is expected to
become very acute in coming years [35]. To cope with this situation, there is a need to
focus on renewable energy sources and to use the existing fossil fuel reservoirs in an
efficient way.
The fuel cell technology could be considered as a good option in this scenario
considering the high efficiency as compared to conventional power generation methods.
However, in Pakistan, the commercialization of fuel cells has not been started due to the
high cost and lack of advanced technology.
1.3 Preceding Studies
Among different Ni-cermet alternate materials, perovskite oxides appear to be the
suitable anode candidates for their remarkable properties [36]. Perovskite oxides have
general formula of ABO3 where A and B cations are 6-fold and 12-fold coordinated to
the oxygen anions, respectively. The structure consists of BO6 octahedra sharing the
corners of the cube containing an A cation at the centre. The A-site is usually occupied
by alkaline earth and/or rare earth metal ions while small transition metal ions (usually
from 3d series) reside on the B site [37]. Particular attention has been given to
perovskites containing transition metals such as Ti, Cr, Mn or Mo due to the existence of
multiple oxidation states which assist the electrocatalytic processes and facilitate
electronic conductivity.
Interesting defect chemistry is offered by perovskites by partial or full substitution
of A and/or B-sites with aliovalent cations. The properties in perovskites can be tuned
CH-1 Introduction
4
and tailored as desired defects can be introduced into the structure by careful selection of
dopants. SrTiO3 is a typical perovskite that has been extensively studied and exhibits n-
type semiconducting behaviour when donor doped or reduced. Both A and/or B-sites of
the strontium titanate have been doped to enhance the properties of the parent compound.
Special attention has been given to enhance its electrical conductivity by partial
substitution of Sr+2
on the A-site and/or Ti +4
on the B-site to yield interesting compounds
with excess or substoichiometric oxygen that largely affects perovskite properties.
The nature of B-site dopant affects structure, redox properties, conductivity and
electro-catalytic properties of the parent compound [38]. In this respect, various B-site
dopants have been investigated such as Nb [39], Mn [40], Ga [41], Sc [42], Fe [43], Al
and Cr [44]. Good conductivity values have been found for Nb doped SrTiO3, for
example, SrTi0.98Nb0.02O3-δ presents conductivity value of 339 S cm-1
at 800 oC after
being reduced in hydrogen at 1400 °C [45].
A-site substitution is effective in enhancing electrical conductivity of SrTiO3.
Donor doping SrTiO3 with trivalent cations like La+3
has been discussed in the literature
where an increase in conductivity has been observed [46, 47]. Moreover, there are reports
about the high resistance to carbon deposition or sulphur poisoning [48, 49]. Marina has
found pronounced effect of La+3
doping on the A-site of strontium titanate; among
investigated compositions, the maximum conductivity of 500 S cm-1
has been observed in
hydrogen at 800–1000 oC for La0.3Sr0.7TiO3 sintered at 1650 °C in reducing atmosphere
[50]. As reported in various papers, the charge compensation mechanism changes from
ionic (under oxidized conditions) to electronic in reduced atmosphere [51]. Under
reducing atmosphere, Ti+4
reduces to a lower oxidation state, a process accompanied by
the formation of oxygen vacancies (see equation below), freeing electronic carriers that
enhance the conductivity [52].
/ ..
2
12 2
2
X X
Ti O Ti OTi O Ti V O (1.1)
Other than lanthanum, Y+3
has also been explored as an A-site dopant [53]. Li et
al. have found that Y0.09Sr0.91TiO3 sintered at 1300 o
C in forming gas possesses an
electrical conductivity of 73.7 S cm−1
at 800 oC in the same atmosphere [54].
CH-1 Introduction
5
A-site deficient titanates have gained interest in last few years because they show
good electrical conductivity, enhanced sintering, thermal stability and good performance
as SOFC anode/anode support [55-58]. John et al. have reported A-site deficient systems
to have good conductivity, thermomechanical compatibility with yttria-stabilized zirconia
(YSZ) and performance in fuel cell conditions [59-61].
Using the research by Ahmed et al. as a springboard, calcium doped lanthanum
strontium titanate was chosen as the anode candidate for this research [62]. In that study,
A-site deficient lanthanum strontium titanate was doped with Ca+2
from x value of 0.1 to
0.7 since Ca+2
offers good solubility in SrTiO3 due to size compatibility with Sr+2
. It was
found that conductivity, σ, increases with an increase of Ca content and the maximum
value was achieved at a dopant level of x = 0.45, followed by a drop of conductivity. This
composition with maximum conductivity, La0.2Sr0.25Ca0.45TiO3 is the focus of present
research.
The solid state synthetic route adopted earlier [62] involves high temperature
sintering and many firing stages. To avoid these problems, a solution phase synthetic
approach was applied to synthesize this composition at low temperature. The present
dissertation is focused on synthesis, characterization and application of this A-site
deficient, Ca+2
doped composition, La0.2Sr0.25Ca0.45TiO3, (hereafter called as LSCTA- as
suggested in the earlier study [59]) as an anode support for SOFCs. The intent of this
study was the development of new ceramic SOFC anode materials which possess good
electrical conductivity as well as redox stability. In the present project, this composition
was studied in different yet connected aspects that are important for anode development.
1.4 Direction of Research
The first step in the present research was synthesis of LSCTA- via the solution
phase Pechini method involving sol and gel formation [63]. In the Pechini method, the
product is obtained by calcination of a dried gel. To get the optimized calcination
temperature, the LSCTA- dried gel was calcined at various temperatures from 900 °C to
1000 °C and characterized using different techniques like XRD, TGA, SEM, EIS and
dilatometry. From these results, an optimized calcination temperature was selected for
further studies. This material was then investigated in different directions that are vital for
anode development.
CH-1 Introduction
6
In the next step, the synthesized powder was processed in aqueous tape casting
which is a technique to fabricate SOFC anode/ anode support components [64, 65]. Apart
from the tape quality produced, the aqueous-based tape casting method has the obvious
advantages of being environment friendly, less health hazardous and cost-effective. For
uniform, homogeneous and crack free green tapes, the correct slurry formulation is
essential. Slurry formulation was optimized and both dense and porous green tapes were
fabricated. The sintered bars prepared from green tapes by lamination were tested for
conductivity measurements.
Porosity is a basic requirement for the electrodes of solid oxide fuel cells (SOFC)
because it controls gas permeability, electrical conductivity, mechanical strength,
electrochemical catalytic activity, surface exchange kinetics and compatibility with other
fuel-cell components [66, 67]. Although some porosity may be generated in the anode
through the control of the sintering process, however engineered porosity is produced in
electrodes of SOFC by addition of organic additives known as pore formers (PF) [68, 69]
having carbon as the main constituent, which pyrolyzes during sintering e.g., synthetic
polymers, natural biopolymers, or carbon-based materials. It is believed that the burn out
of carbon particles with desired shape and size helps in getting the anticipated porosity.
Tape casting method has been applied to fabricate SOFC anodes using suitable pore
formers like graphite, glassy carbon, PMMA, rice starch or combination of pore formers.
In the present work, porosity was introduced in LSCTA- tapes using commercial pore
formers like graphite, PMMA and glassy carbon. Both glassy carbon and PMMA are
very expensive where as with graphite, the tapes are often dry and lead to de-lamination.
A new aspect of this study employs the use of carbon spheres as a pore former in LSCTA-
tapes which were synthesized by in-expensive hydrothermal method [70, 71].
In the next stage of research, the performance of LSCTA- was studied in
electrolyte supported symmetrical cells because this configuration is easy to use. The
effect of impregnates on the performance of symmetrical cells was investigated with ac
impedance. Further, the anode performance of impregnated LSCTA- was explored in
button cell mode using three electrodes set up with lanthanum strontium manganite,
(LSM) as cathode and Pt as reference electrode.
CH-1 Introduction
7
Finally, LSCTA- was doped at B-site with Ni and Fe to increase the electronic
conductivity of parent composition. The doped compositions were synthesized by the
Pechini method and further characterized by XRD, dilatometry, SEM and dc
conductivity.
1.5 Research Objectives
The main objective of this research was in particular, to explore new anode candidates for
SOFCs. Majorly, these objectives can be described as follows:
To synthesize LSCTA- at low temperature via a solution phase method and study
effect of calcination temperature on its properties and thus to optimize its
processing.
To optimize its fabrication in aqueous tape casting (by optimizing slurry
formulation and microstructure optimization).
To check the anode performance of this compound in symmetrical and button
cells.
To improve its conductivity by doping and thus to suggest new SOFC anode
candidates.
1.6 Thesis Layout
The present thesis deals with the synthesis, characterization and application of
LSCTA- as an anode for SOFC. The thesis is divided into 10 chapters including an
introduction chapter. Chapter 2 provides a brief introduction to fuel cell technology
followed by an overview of alternative anode materials. Chapter 3 gives an account of
perovskites and their defect chemistry. In chapter 4, the techniques employed in the
current study are given. Chapter 5 is dedicated to the results regarding synthesis and
characterization of LSCTA-. Aqueous tape casting of LSCTA- is detailed in chapter 6.
Chapter 7 is devoted to the microstructure optimization of LSCTA-. Symmetrical and
button cell results are presented in chapter 8. Synthesis and characterization of doped
analogues of LSCTA- is described in chapter 9. Conclusions and recommendations of the
study are furnished in Chapter 10.
CH-1 Introduction
8
REFERENCES
1. C. Song, Catal. Today, 2002, 77, 17 – 49.
2. A.E. Atabania, A.S. Silitonga, I.A. Badruddina, T.M.I. Mahliaa, H.H. Masjukia,
S. Mekhilef, Renew Sust. Energ. Rev., 2010, 16, 2070 – 2093.
3. R. Saidur, E.A. Abdelaziz, A. Demirbas, M.S. Hossain, S. Mekhilef, Renew Sust.
Energ. Rev., 2011, 15, 2262 – 2289.
4. M.S. Dresselhaus, I.L. Thomas, Nature, 2001, 414, 332 – 337.
5. A. Szklo, R. Schaeffer, Energy, 2006, 31, 2513 – 2522.
6. N.Z. Muradov, T.N. Vezirog˘lu, Int. J. Hydrogen Energy, 2008, 33, 6804 – 6839.
7. H. Balat, E. Kırtay, Int. J. Hydrogen Energy, 2010, 35, 7416 – 7426.
8. F. Zabihian, A.S. Fung in Paths to Sustainable Energy, Ed., Dr Artie Ng, ISBN
978-953-307-401-6, Intech, 2010.
9. A.B. Stambouli, E. Traversa, Renew Sust. Energ. Rev., 2002, 6, 433 – 455.
10. J.M. Andu´ jar, F. Segura, Renew Sust. Energ. Rev., 2009, 13, 2309 – 2322.
11. A.B. Stambouli, Renew Sust. Energ. Rev., 2011, 15, 4507 – 4520.
12. W.T. Hong, T.H. Yen, T.D Chung, C.N. Huang, B.D. Chen, App Energ., 2011,
88, 3990 – 3998.
13. R.J. Kee, H. Zhu, D.G. Goodwin, P. Combust. Inst., 2005, 30, 2379 – 2404.
14. J. Palsson, A. Selimovic, L. Sjunnesson, J. Power Sources, 2000, 86, 442 – 448.
15. W. Winkler, H. Lorenz, J. Power Sources, 2002, 105, 222 – 227.
16. E. Baniasadi, I. Dincer, Int. J. Hydrogen Energy, 2011, 36, 111281 – 11136.
17. N.Q. Minh, Solid State Ionics, 2004, 174, 271 – 277.
18. H.A. Taroco, J.A.F. Santos, R.Z. Domingues, T. Matencio in Advances in
Ceramics - Synthesis and Characterization, Processing and Specific Applications,
Ed., C. Sikalidis, Intech, 2011.
19. B.C.H. Steele, Nature, 1999, 400, 619 – 621.
20. N.M. Sammes, Fuel Cell Technology: Reaching Towards Commercialization,
Springer, 2006.
21. R.M. Ormerod, Chem. Soc. Rev., 2003, 32, 17 – 28.
22. W.L. Becker, R.J. Braun, M. Penev, M. Melaina, J. Power Sources, 2012, 200, 34
– 44.
CH-1 Introduction
9
23. W.Z. Zhu, S.C. Deevi, Mater. Sci. Eng. A., 2003, 348, 227 – 243.
24. A.S. Nesaraj, J. Sci. Ind. Res., 2010, 69, 169 – 176.
25. J. Huang, F. Xie, C. Wang, Z. Mao, Int. J. Hydrogen Energy, 2012, 37, 877 – 883.
26. B.C.H. Steel, Solid State Ionics, 1996, 86-88, 1223 – 1234.
27. T.L. Wen, D. Wang, H.Y. Tu, M. Chen, Z. Lu, Z. Zhang, H. Nie, W. Huang, Solid
State Ionics, 2002, 152– 153, 399 – 404.
28. C. Zuo, M. Liu, in Sol-Gel processing for conventional and alternative energy,
Eds.; M. Aparicio, A. Jitianu, L.C. Klein, Springer, 2012.
29. S. Jiang, S.H. Chan, J. Mat. Sci., 2004, 39, 4405 – 4439.
30. J. B. Goodenough, Y.H. Huang, J. Power Sources, 2007, 173, 1 – 10.
31. N.Q. Minh, J Am. Chem. Soc. 1993, 76, 563 – 588.
32. W. Bao, H. Guan, J. Cheng, J. Power Sources, 2008, 175, 232 – 237.
33. M.A. Ghaffar, Renew Energ., 1995, 6, 941 – 976.
34. M. Amer, T.U. Daim, Energ. Sust. Develop., 2011, 15, 420 – 435.
35. M.A. Chaudhry, R. Raza, S.A. Hayat, Renew Sust. Energ. Rev., 2009, 13, 1657 –
1662.
36. S. Tao, J.T.S. Irvine, Chem. Rec., 2004, 4, 83 – 95.
37. M.A. Pen˜a, J.L.G. Fierro, Chem. Rev., 2001, 101, 1981 – 2017.
38. D.N. Miller, J.T.S. Irvine, J. Power Sources, 2011, 96, 7323 – 7327.
39. J. Karczewski, B. Riegel, M. Gazda, P. Jasinski, Kusz, J. Electroceram., 2010, 24,
326 – 330.
40. A. Ovalle, J.C. Ruiz-Morales, J. Canales-Vasquez, D. Marrero-Lopez, J.T.S.
Irvine, Solid State Ionics, 2006, 177, 1997 – 2003.
41. D. Neagu, J.T.S. Irvine, Chem. Mater., 2011, 23, 1607 – 1617.
42. X. Li, H. Zhao, F. Gao, N. Chen, N. Xu, Electrochem. Commun., 2008, 10, 1567
– 1570.
43. D.P. Fagg, V.V. Kharton, J.R. Frade, A.A.L. Ferreira, Solid State Ionics, 2003,
156, 45 – 57.
44. N.G. Eror, U. Balachandran, J. Am. Cer. Soc., 1982, 65, 426 – 431.
45. T. Kolodiazhnyi, A. Petric, J. Electroceram., 2005, 15, 5 – 11.
CH-1 Introduction
10
46. J. Canales-Vazquez, S.W. Tao, J.T.S. Irvine, Solid State Ionics, 2003, 159, 159-
165.
47. K.B. Yoo, G.M. Choi, Solid State Ionics, 2009, 180, 867 – 871.
48. R. Mukundan, E.L. Brosha, F.H. Garzon, Electrochem. Solid St., 2004, 7, A5 –
A7.
49. M. Roushanafshar, J.L Luo, A.L. Vincent, K.T. Chuang, A.R. Sanger, Int. J.
Hydrogen Energ., 2012, 37, 7762 – 7770.
50. O.A. Marina, N.L. Canfield, J.W. Stevenson, Solid State Ionics, 2002, 149, 21 –
28.
51. P. Blennow, A. Hagen, K.K. Hansen, L.R. Wallenberg, M. Mogensen, Solid State
Ionics, 2008, 179, 2047 – 2058.
52. P.R. Slater, D.P. Fagg, J.T.S. Irvine, J. Mater. Chem., 1997, 7, 2495 – 2498.
53. X. Huang, H. Zhao, W. Qiu, W. Wu, X. Li, Energ. Convers. Manage., 2007, 48,
1678 – 1682.
54. X. Li, H. Zhao, W. Shen, F. Gao, X. Huang, Y. Li, Z. Zhu, J. Power Sources,
2007, 166, 47 – 52.
55. D. Burnat, A. Heel, L. Holzer, D. Kata, J. Lis, T. Graule, J. Power Sources, 2012,
201, 26 – 36.
56. Q. Ma, F. Tietz, A. Leonide, E. Ivers-Tiffée, Electrochem. Commun., 2010, 12,
1326 – 1328.
57. H. Zhao, F. Gao, X. Li, C. Zhang, Y. Zhao, Solid State Ionics, 2009, 180, 193 –
197.
58. X. Li, H. Zhao, X. Zhou, N. Xu, Z. Xie, N. Chen, Int. J. Hydrogen Energ., 2010,
35, 7913 – 7918.
59. C.D. Savaniu, J.T.S. Irvine, J. Mater. Chem., 2009, 19, 8119 – 8128.
60. C.D. Savaniu, J.T.S. Irvine, Solid State Ionics, 2011, 192, 491 – 493.
61. D. Neagu, J.T.S. Irvine, Chem. Mater., 2010, 22, 5042 – 5053.
62. A.D. Aljaberi, J.T.S. Irvine, J. Mater. Chem. A, 2013, 1, 5868 – 5874.
63. A. Ries, A.Z. Simoes, M. Cilense, M.A. Zaghete, J.A. Varel, Mater. Charac.,
2003, 50, 217 – 221.
64. M. Cologna, V.M. Sglavo, Int. J. Appl. Ceram. Technol., 2010, 7, 803 – 813.
CH-1 Introduction
11
65. L.H. Luo, A.I.Y. Tok. Mater. Sci., Eng. A, 2006; 429, 266 – 271.
66. J.H. Lee, J.W. Heo, D.S. Lee, J. Kim, G.H. Kim, H.W. Lee, H.S. Song, J.H.
Moon, Solid State Ionics, 2003, 158, 225 – 232.
67. T. Suzuki, Z. Hasan, Y. Funahashi, T. Yamaguchi, Y. Fujishiro M. Awano,
Science, 2009, 325, 852 – 855.
68. S.F. Corbin, J. Am. Ceram. Soc., 1999, 82, 1693 – 1701.
69. J.J. Haslam, A.Q. Pham, B.W. Chung, J.F. Di Carlo, R.S. Glass, J. Am. Ceram.
Soc., 2005, 88, 513 – 518.
70. M. Sevilla, A.B. Fuertes, Carbon, 2009, 47, 2281 – 2289.
71. Y. Mi, W. Hu, Y. Dan, Y. Liu, Mater. Lett., 2008, 62, 1194 – 1196.
Key to References: e.g.,
68. S.F. Corbin, J. Am. Ceram. Soc., 1999, 82, 1693 – 1701.
S.F. Corbin Name(s) of Author(s)
J. Am. Ceram. Soc. Name of Journal
1999 Year of Publication
82 Volume Number
1693 – 1701 Page Numbers
Chapter 2
Fuel Cells
12
Fuel Cells
Abstract
The environmental concern and ever increasing dependence on oil has stimulated the
research on fuel cells which could replace conventional power generation methods due to
their potential for use in stationary, portable and transport applications. In this chapter, a
general introduction is given on fuel cells regarding their history, operational principle,
types and applications. Principles and components employed in solid oxide fuel cell
(SOFC) are given with emphasis on chemistry, research and developmental aspects of
SOFC anode.
2.1 The Fuel Cell
2.1.1 Working
The fundamental structure of a fuel cell (Fig. 2.1) consists of an electrolyte
sandwiched between the porous anode (negative electrode) and the porous cathode
(positive electrode). The fuel and oxidant (air/O2) are continuously fed to the anode and
cathode respectively. The presence of dense electrolyte prevents the direct mixing of fuel
and the oxidant. The electrochemical (EC) reaction takes place at the triple phase
boundary established at the gas-electrolyte-electrode interface. The ions which are
produced during the electrochemical reaction at one of the electrodes are conducted to the
other electrode through the electrolyte while electrons travel round an external circuit
delivering electric power.
The fuel cell works similar to that of the battery. However, in battery, the
components (electrodes and electrolyte) themselves react in the energy conversion
process whereas the working of a fuel cell requires the fuel and air/oxygen to be fed
continuously and the removal of the products of the reactions. This implies that the
batteries need to be discarded or recharged once their fuel is exhausted, but ideally, the
fuel cells can operate continuously as long as fuel is supplied and the products and by-
products are removed. Thus, a fuel cell can theoretically produce electrical energy as long
13
as fuel and oxidant are supplied to the porous electrodes, which is practically limited by
the degradation or malfunction of some of its components.
Fig. 2.1: Schematic of fuel cell operation.
In the case of O2-
conducting electrolytes like yttria-stabilized zirconia (YSZ),
oxygen is electro-reduced at the cathode to produce O2-
ions, which migrate through the
dense electrolyte to the anode where they react with fuel to produce useful energy. The
anode releases electrons that are consumed again at the cathode. The half-cell reactions
can be represented by:
Cathode: 2
2 4 2O e O (2.1)
Anode: 2
2 22 2 2 4H O H O e (2.2)
Analogous electrode reactions for proton conducting electrolytes would be;
Anode: 22 4 4H H e (2.3)
Cathode: 2 24 2 4 2H O e H O (2.4)
As the electrolyte should be a pure ion conducting material, the electron current
flowing through an external circuit creates the electrical power. If the fuel is clean, the
effluents are in principle only water, heat and CO2. Thus fuel cells are a source of
generating clean and pollution free electricity at high efficiencies.
A single cell can only generate a small amount of power. To have reasonable
power output, many single cells are combined together by a process referred as stacking.
14
Stacking involves connecting single cells in series using bipolar plates which have
channels for air and fuel to flow inside the stack [1].
2.1.2 Historical background
In 1839, Sir William Grove operated the first successful hydrogen-oxygen cell at
room temperature using a liquid electrolyte, generally stated as the start of fuel cell
history [2]. The term “fuel cell” was coined in 1889 by Ludwig Mond and Charles
Langer [3]. In 1899, Nernst discovered the solid oxide electrolyte when using stabilized
zirconia in making filaments for electric glowers which is still considered as state of the
art electrolyte material for SOFC [4]. The first ceramic fuel cell was operated at 1000 °C
by Baur and Preis in 1937 [5]. In the middle of the 20th
century the development got a
rapid boost and several types of fuel cells were developed for different applications [6].
Fuel cell technology is considered to be an expensive technology thus the primary
challenges are cost and durability which are being explored by material scientists and
engineers.
2.1.3 Fuel cell characteristics
The net reaction shown in equations, 2.1 and 2.2, can be given as
2 2 2
1
2H O H O (2.5)
For the above reaction, the Nernst equation takes the form;
2 2
2
1/2
lnO Ho
eq
H O
p pRTE E
nF p (2.6)
where Eo is the standard cell potential defined as the difference between the standard
reduction potentials of the cathode and anode reactions. Further, R is the gas constant, T
is the absolute temperature, n is the number of electrons involved (in this case, n = 2) and
F is Faraday’s constant (96,475 C/equiv). This voltage output as a function of current
drawn from the cell gives the performance of a fuel cell.
In an ideal (reversible) fuel cell, the cell voltage is independent of the current
drawn. However, in actual practice, cell potential is decreased by various irreversible
drops which are termed as polarization losses [7]. Thus the potential of the cell becomes
15
less than the potential predicted by Nernst equation by factors represented by following
equation,
conceq L act iRE E E (2.7)
In Eq. 2.7, Eeq is the theoretical equilibrium voltage as calculated from the Nernst
equation and EL corresponds to voltage loss due to leaks in the electrolyte. The slow
electrode reactions contribute to the activation overpotential, ηact. ηiR represents
overpotential due to ohmic losses in the entire cell while the slow gas diffusion processes
in the electrodes cause a concentration overpotential ηconc. A plot of cell voltage vs.
current density is known as a polarization curve. A typical polarization curve for a fuel
cell is shown in Fig. 2.2. The combined contributions of these overvoltages decrease the
cell voltage output with increasing current density. The power output of the fuel cell (in
mWcm-2
) is given by the product of voltage and current density. In fuel cell technology,
the research has always remained focused on minimizing the differences between actual
and theoretically predicted cell voltage to improve the performance.
Fig. 2.2: Ideal and actual fuel cell voltage/current characteristics.
However, the polarization losses can be minimized by selection of the right
electrode materials with optimized microstructure. A brief description of these
polarization losses [8-9] is given below:
16
2.1.3.1 Ohmic polarization ( iR )
Ohmic polarization is caused due to the resistive losses in the electrolyte and in
the electrodes. These losses obey ohm’s law, thus a linear relationship between voltage
drop and current density is observed.
IR totalIR (2.8)
where I is the current drawn from the cell and Rtotal represents the total cell resistance
which has contributions from the electrolyte, electrodes, lead wires and interfaces within
the cell. The intrinsic electrolyte resistance may be determined by;
lR
A (2.9)
In the above equation denotes electrolyte resistivity while l is the thickness and
A is the area of electrolyte. The contribution to total resistance from other components
like electrodes, contacts and lead wires is usually determined experimentally by using
electrochemical impedance spectroscopy. These losses are dominant at intermediate
current densities (~100 to 500 mA cm-2
).
2.1.3.2 Activation polarization ( act )
Activation polarization is result of sluggish kinetics of oxidation or reduction
processes at the electrodes which limits the electrochemical processes occurring at the
cell electrode(s). For examples, the slow oxygen reduction kinetics at the cathode of
polymer electrolyte membrane fuel cell and sluggish methanol oxidation kinetics at the
anode of direct methanol fuel cell result activation polarization. This overpotential is
pronounced when low currents (1-100 mA cm-2
) are drawn from an electrochemical cell.
2.1.3.3 Concentration polarization ( conc )
At high current density, the mass transfer limitation results in inadequate flow of
reactants to or removal of products from the cell electrodes, causing the concentration
gradient of the gas at the electrochemically reactive sites and the bulk of gas stream. This
results in voltage drop which is termed as concentration polarization.
17
2.1.4 Fuel cell efficiency
Fuel cells are more efficient then internal combustion engines [10, 11] because
the latter involves a couple of intermediate steps for conversion of chemical to electrical
energy namely; (a) from chemical energy into heat energy, (b) from heat energy into
mechanical energy and (c) from mechanical energy into electricity. The energy is lost in
each of these steps, making these systems less efficient.
The maximum efficiency of heat engines defined by the Carnot cycle is given by;
1
2
1T
T (2.10)
where T2 is the temperature of hot reservoir and T1 being the temperature of sink. The
maximum Carnot efficiency limit of a heat engine operating at 400 oC, with the water
exhausted through a condenser at 50 ◦C is about 52%.
In the case of fuel cells, the chemical energy of the fuel is directly converted to
electrical energy without involving any intermediate step. The chemical energy is related
to the standard enthalpy change (ΔHo) or Gibbs free energy change (ΔG
o) of the reaction.
In fact, in fuel cell operation, the direct conversion of free energy (ΔGo) to electrical
energy takes place. ΔGo of the overall reaction is related to the cell potential by the
equation;
ΔGo = –nFE
o (2.11)
Since n, F, and Eo
are positive numbers, the standard free energy change of the overall
reaction is negative which indicates the reaction spontaneity and shows the
thermodynamic feasibility of fuel cell operation.
The maximum efficiency for a fuel cell is calculated from these thermodynamic
parameters and is given by;
o
o
G
H
(2.12)
A simple calculation can show that the efficiency of fuel cells is greater than heat
engines. For example, in case of H2/O2 fuel cell, at standard conditions (298.15 K and 1
atm), the reaction given in Eq. 2.5, has enthalpy ΔH° = -285.83 kJ mol-1
and Gibbs free
energy ΔG° = -237.09 kJ mol-1
and the efficiency comes out to be 83%.
18
2.1.5 Advantages of fuel cells
Generally, the following advantages [12-13] may be described;
1. High energy conversion efficiency
The conversion of free energy into electrical energy eliminates the usual losses
encountered in conventional power generation methods which involve multi steps for the
conversion of fuel to useful energy. The efficiency is further improved when the by-
product heat is fully utilized.
2. Environment friendly
Fuel cells are a source of clean energy as they do not emit toxic gases to the
environment. The amounts of released CO2 and NOx produced per kWatt power are very
much less compared to grid electricity produced from fossil fuel-burning power plants.
This characteristic makes them an attractive alternative energy conversion system.
3. Modularity
Modularity is another big advantage of fuel cells. The size of a fuel cell can be
easily increased or decreased (from button cell to big cell stacks) and its electric
efficiency is relatively independent of size.
4. Portability
The size flexibility makes them portable so they can be transported easily to their
application sites as they have minimum siting restrictions.
5. Noise pollution reduction
The absence of moving parts makes fuel cell operation quiet. Thus they can be
used near urban residential areas.
6. High Power Density
Fuel cells offer significantly higher power densities and longer life times than
batteries.
Unfortunately, the major obstacle of this technology is its high cost. However,
efforts are being carried out to reduce the cost by finding alternate materials and
fabrication techniques.
19
2.1.6 Applications of fuel cells
Fuel cells can be used in different application areas depending on the power
requirements (ranging from a few Watts for small scale to hundreds of MW for large
scale distributed power generation. Some of the fields [14-17] in which fuel cells are used
are described below.
1. Niche Applications
Niche applications include mobile phones, camcorders, digital cameras and
laptops etc. Higher power densities of fuel cells make them attractive to be used for
above applications where only a few Watts are required.
2. Transportation
The high efficiency and significantly reduced discharge of toxic gases during the
fuel cell operation has triggered their use in the transportation sector. The low
temperature PEM fuel cells are considered for use in transport applications due to their
rapid start up, high power density, simple design and low (< 120 ºC) operating
temperature.
3. Combined Heat and Power Applications
One of the important applications of fuel cells is the generation of combined heat
and power (CHP) from a single system which could be used to power individual
households, larger residential units and business and industrial premises. The high
operational temperature makes solid oxide fuel cells attractive for combined heat and
power (CHP) applications. Such fuel cell CHP units offer significantly greater efficiency.
4. Military defense Applications
Due to the high power density, the fuel cells find their applications in accessories
used for military defense including night vision devices, global positioning systems,
target designators, climate controlled body suits and digital communication systems.
Their use as a source of remote and backup power is also valuable in defense
applications.
5. Stationary applications
Stationary power and heat generation is another application area of fuel cells.
MCFC and SOFC are the most promising fuel cell types for this kind of application due
to their high operational temperature. These fuel cells can be used alone or together with
20
other technologies such as gas turbine, steam turbine and gasification systems in
combined heat and power, i.e. cogeneration, applications.
2.1.7 Types of fuel cells
Since the fabrication of the first fuel cell in 1839, different types of fuel cells have
been developed depending on the fuel nature and the operating temperature [18-20].
Mainly, the fuel cells have been classified on the basis of the nature of ion conducting
electrolytes. The ion conduction process strongly depends on the material thus the
operational temperature also varies from one fuel cell to the other. Table 2.1 lists
different fuel cell types, along with their mobile ion, temperature of operation, fuel and
electrolyte. Large differences exist in application, design, size, cost and operating range
for the different types of fuel cells.
Among these, molten carbonate fuel cells (MCFCs) and solid oxide fuel cells
(SOFCs) are known as high-temperature fuel cells since their operating temperatures are
considerably higher than the other fuel cell types. Proton exchange membrane fuel cells
(PEMFCs), direct methanol fuel cells (DMFCs) and alkaline fuel cells (AFCs) are the
low temperature fuel cell types.
The operating temperature has consequences for design, efficiency, the choice of
materials needed and the kind of fuel that may be used in the fuel cell. For low
temperature fuel cells, usually the operating temperature is too low to enable direct
oxidation of hydrocarbon fuels like natural gas, therefore fuels like hydrogen and
methanol are used. These cells find use for small scale applications, e.g., cars, notebooks
and phones etc.
However, in high temperature fuel cells, it is possible to use natural gas which can
be reformed internally into hydrogen and carbon monoxide. The high temperature fuel
cells are used for the decentralized generation of heat and power. A brief overview of
each of these systems is presented below, with emphasis on the SOFC system.
2.1.7.1 Alkaline fuel cell (AFC)
In the alkaline fuel cell (AFC), an aqueous alkaline solution (e.g., KOH) having
hydroxyl ( OH ) as the mobile ions is used. Pure oxygen or purified air and pure
hydrogen are required to avoid poisoning of the alkaline electrolyte [21].
21
2.1.7.2 Proton exchange membrane fuel cell (PEMFC)
The polymer electrolyte membrane fuel cell (PEMFC) uses solid, proton
conducting electrolyte such as Nafion. These cells are poisoned by CO and CO2 and
require pure hydrogen and oxygen. These cells are expensive as they use carbon-
supported platinum catalysts in the electrodes.
The low operational temperatures and rapid start-up times make PEMFCs
applicable in portable applications such as vehicles and mobile devices. Another type of
PEMFCs is direct methanol fuel cells (DMFCs) which use methanol as a fuel rather than
hydrogen [22].
2.1.7.3 Phosphoric acid fuel cell (PAFC)
The phosphoric acid fuel cell (PAFC) uses phosphoric acid (H3PO4) as a proton-
conducting electrolyte. The cell is resistant to small concentrations of CO and CO2. Often
reforming of natural gas is done on site to produce hydrogen. PAFCs have found utility in
combined heat and power applications [23].
2.1.7.4 Molten carbonate fuel cell (MCFC)
In molten carbonate fuel cells (MCFCs), a carbonate-conducting mixture of
sodium and potassium carbonates is used as the electrolyte. This fuel cell operates at
relatively high operational temperature (650 °C). This type of fuel cell is tolerant to CO
and CO2 thus allows hydrocarbons to be used as a fuel. To generate the 2
3CO ions, CO2 is
required which can be supplied if air (rather than pure oxygen) is used as an oxidant or
can be recycled from the anode exhaust gas [24].
2.1.7.5 Solid oxide fuel cell (SOFC)
The solid oxide fuel cell (SOFC) uses a solid-phase oxide ion (O-2
) conducting
electrolyte usually yttria-stabilized zirconia (YSZ). The operational temperature of SOFC
is quite high (typically 600 °C - 1000 °C) which renders fuel flexibility to this system.
This fuel cell type is resistant to CO and CO2.
Table 2.1 Different types of fuel cells
PEMFC AFC PAFC MCFC SOFC
Anode Pt black Ni Pt/C Ni-10%Cr Ni-YSZ cermet
Cathode Pt Black Li-doped NiO Pt/C Li-doped NiO Sr-doped LaMnO3
Electrolyte Nafion 85% KOH 100% H3PO4 62% Li2CO3 –38% K2CO3 Yttria-stabilized ZrO2
Working T
(°C) 80 260 200 650 1000
Fuel H2, CH3OH H2 H2 H2, Hydrocarbons, CO H2, Hydrocarbons, CO
Mobile Ion (H2O)n, H+ OH
- H
+ CO3
2- O
2-
Anode
Reaction _
2 2 2H H e 2 22H OH H O _
2 2 2H H e 2
2 3 2 2 2H CO H O CO e 2
2 22 2H O H O e
Cathode
Reaction
2 24 4 2O H e H O
2 22 4 4O H O e OH
2 22 4 4O H O e OH
2
2 2 32 4 2O CO e CO 2
2 4 2O e O
PEMFC: Polymer electrolyte membrane fuel cell, AFC: Alkali fuel cell, PAFC: Phosphoric acid fuel cell, MCFC: Molten carbonate fuel cell,
SOFC: Solid oxide fuel cell [7, 9].
CH-2 Fuel Cells
23
2.2 Solid Oxide Fuel Cell
Solid oxide fuel cell (SOFC) is an energy conversion device that contains solid
oxide ion conducting electrolyte and operates at temperatures ranging from 600 °C to
1000 °C. The high temperature is required to create sufficient ionic conductivity in dense
electrolyte. Solid oxide fuel cells have gained recognition as high temperature fuel cells.
By the use of solid electrolytes, corrosion problems that are faced by using liquid
electrolytes can be avoided. One other aspect of the high temperature is the choice of fuel
flexibility because the high operation temperatures (usually > 550°C) allow internal
reforming of the fuels and promote rapid kinetics with non-noble catalysts. Meanwhile,
byproduct heat generated during operation could be utilized in other ways and give a high
total efficiency [25-29].
2.2.1 Operating principles of SOFC
The operating principle of a SOFC with an oxide ion conductor is schematically
shown in Fig. 2.3. Fuel and air are fed to the fuel and air channels at the anode and
cathode respectively. At the cathode, oxygen is electro-reduced at the porous air electrode
to produce oxide ions. These ions migrate through the solid electrolyte to the anode, and
they react with the fuel to produce effluents, H2O or CO2.
Fig. 2.3: Schematic of an oxide ion conducting solid oxide fuel cell.
CH-2 Fuel Cells
24
The open-circuit voltage, Eo, of the cell can be calculated using equation 2.13.
lnco
o o
a
o
PG RTE
nF nF P (2.13)
where ΔGo is the free energy change of the electrochemical reaction. Po
c and Po
a are the
partial pressure of the oxygen at the cathode and at the anode respectively.
For a solid oxide fuel cell working with pure hydrogen as a fuel and air as the
oxidant, the cell voltage is about 1 V at 1000 °C. However, the polarization losses result
in a cell voltage to drop from the theoretically predicted Nernst voltage. The total
polarization of a cell, η, is given as the sum of anode polarization, ηa, cathode
polarization, ηc, and resistance polarization, ηr .
a b r (2.14)
The polarization depends on the electrode materials, the electrolyte, the cell
design and the operating temperature.
2.2.2 SOFC materials
The material choice for different cell components is governed by the following
criteria [30];
a. The cell components (cathode, anode and interconnect) should have sufficient
electrical conductivity.
b. All the cell parts should have adequate chemical, mechanical and thermal stability
at the operating conditions.
c. Different components should have close thermal expansion behaviour to avoid
thermal stress.
d. Finally, the fabrication process should be relatively easy so that every component
could be fabricated without any complication.
A brief description about the cathode, anode, electrolyte and interconnect is given
below.
2.2.2.1 Cathode (air electrode)
In SOFC, the electro-reduction of oxygen takes place at the cathode by the
following reaction;
CH-2 Fuel Cells
25
2
2
12
2O e O (2.15)
The cathode must fulfill some of the requirements like high electronic and ionic
conductivity, chemical and mechanical stability in an oxidizing atmosphere, matching
thermal expansion coefficient with other SOFC components, minimal reactivity with
electrolyte and interconnect materials, high catalytic activity for dissociation of oxygen
and large triple phase boundary to have high reaction rate. Among a large number of
materials, lanthanum strontium manganite, LaSrMnO3 (LSM) is considered as leading
cathode material due to its thermal and mechanical compatibility with zirconia
electrolytes and good electronic conductivity. LSM is often mixed with YSZ to reduce
electrode polarization and extend the triple phase boundary. However, LSM reacts with
YSZ at temperatures above 1300 oC. To overcome this issue, composites of LSM with
gadolinium doped ceria (GDC) have been investigated which have displayed good
performance at low temperatures. Lanthanum strontium ferrite (LSF) has also been
shown to replace LSM between 650 and 800 oC despite of its lower electrical
conductivity.
Researchers are exploring different domains to have better cathode materials with
higher conductivity than LSM and good power density. Different cathode materials have
been studied with different electrolyte materials with every electrode-electrolyte system
having some limitations. The SOFC cathode development has been detailed in various
reviews [31-33].
2.2.2.2 Anode (hydrogen electrode)
In SOFCs, the anode or the fuel electrode is the site where the fuel is reduced
within each cell. To work efficiently, the anode should have high electronic/ionic
conductivity, high electrocatalytic activity, resistance to sulphur, chemical and
mechanical stability, a large triple phase boundary, chemical and thermal compatibility
with other cell components, stability in reducing atmosphere and optimized
microstructure.
For anode supported SOFCs, sufficient porosity is needed to promote gas
transport through the thick electrode. The low cost, good catalytic activity, high ionic and
CH-2 Fuel Cells
26
electronic conductivity and good chemical and mechanical stability make nickel-YSZ as
state of the art anode material. However, it has some limitations: it undergoes
microstructural changes upon redox cycling due to large and facile Ni to NiO volume
change. This cermet has low tolerance to sulphur and accelerates coke formation due to
high catalytic activity when hydrocarbons are used as fuels without excess steam being
present. Moreover, at high operating temperatures, the catalytic active surface area
decreases due to agglomeration and sintering of Ni. All of these factors affect the
performance and long term stability of the SOFC.
To overcome these issues, research is being directed to search for alternate anode
materials [34-36]. The alternate anode materials that have been used to replace Ni/YSZ
are discussed later in this chapter.
2.2.2.3 Electrolyte
In SOFCs, the electrolyte is a solid oxide that allows O2−
ions to migrate from the
cathode to anode. In planar designs, the electrolyte can also function as the support
during fabrication. An operational electrolyte should have good ionic conductivity but no
electronic conductivity, chemical and mechanical stability both at high temperatures and
in reducing and oxidizing environments, gas tightness (fully densified) and thermal
compatibility with other cell components.
The research on electrolyte development has its main focus to improve the oxide
ion conductivity and to decrease the operational temperature. In most solid oxide
conducting materials, the desired ionic conductivity is obtained at above ~600 °C which
puts severe restrictions on the types of materials and thus increases cost.
In SOFCs, 8 mol% yttria-stabilised zirconia (8-YSZ) is the usual choice of
electrolyte owing to its good oxide ion conductivity and mechanical and chemical
stability at the operating conditions. Upon doping of yttria (Y2O3), zirconia is transformed
from the monoclinic phase into the stable fluorite structure cubic phase. Secondly, the
acceptor doping of Y+3
at Zr+4
sites results in the creation of oxygen vacancies in the
zirconia lattice which increases oxygen ion conductivity. 8 – 10 mol% Y2O3 doping
results in the highest oxygen ion conductivity in the zirconia lattice. Further increase of
dopant concentration causes the positive oxygen vacancies and negative yttria ions to
CH-2 Fuel Cells
27
combine, which lowers the concentration of free oxygen vacancies and hence decreases
the conductivity.
To search for alternative electrolyte candidates, zirconia has been doped with
other oxides including Sm2O3, MgO, Yb2O3 etc. Scandia (Sc2O3) doped zirconia is a
good example having comparable stability but higher ionic conductivity to yttria-
stabilized zirconia.
Doped ceria has also been investigated as the electrolyte candidate due to its high
oxide ion conductivity. However, it exhibits electronic conductivity at high temperature
due to partial reduction of Ce+4
to Ce+3
in reducing atmosphere which limits its
application as an electrolyte. LaGaO3 is another material with sufficient ionic
conductivity to be used as an electrolyte. The progress in SOFC electrolyte development
has been discussed in reviews [37, 38].
2.2.2.4 Interconnect
In a planar SOFC stack, the interconnect has two important roles. First, it
separates the reducing gas at the anode of one cell from the oxidizing gas (air, oxygen) at
the cathode of the adjacent cell and secondly, it provides the electrical connection
between adjoining cells. The good interconnect should have redox stability, very high
electrical conductivity, good thermal conductivity, phase stability under temperature
range, resistance to sulfur poisoning and thermal stability with other cell components.
In SOFCs, the metallic (chromium alloys, ferritic stainless steels, austenitic
stainless steel, iron super alloys, and nickel super alloys) interconnects have been used
due to their mechanical stability, easy fabrication, high thermal and electrical
conductivities. However, their thermal expansion is higher than other cell components.
One other drawback is the facile oxidation at the cathode side which results in formation
of poorly conducting chromium oxide and is prone to cracking during long-term
operation. To cope with these problems, ceramic interconnects have been used.
Among the ceramic materials, doped lanthanum chromate (LaCrO3) is the most
common because of the desired properties. To tune the properties, lanthanum chromate
has been doped with vanadium, magnesium, copper, cobalt, iron, strontium, nickel etc.
CH-2 Fuel Cells
28
However, the ceramic interconnects are costly, have sintering difficulties, and often
suffer from deformation [39, 40].
2.2.2.5 Fuels and fuel processing in SOFCs
The choice of fuel in a fuel cell is partly governed by the operating temperature.
The high temperature helps to internally reform practical hydrocarbon fuels. Thus,
SOFCs can virtually operate with different fuels which are reformed to hydrogen before
entering the anode chamber. Thus, the SOFC system is simple because the need of
external reformer and associated heating arrangements is abandoned. Also, due to high
operational temperature, SOFCs are less sensitive to fuel impurities and catalyst poisons,
such as sulfur and carbon monoxide. This renders significant advantages to SOFCs in
comparison with other fuel cells [41, 42].
2.2.3 Classification of SOFC
The SOFCs have been majorly classified according to their temperature level,
nature of electrolyte, stack design, type of support, flow configuration and fuel reforming
type [43-46].
2.2.3.1 Classification according to the temperature level
SOFCs may be classified as low-temperature (LT-SOFC), intermediate-
temperature (IT-SOFC), or high-temperature (HT-SOFC) depending on operational
temperature. The high working temperature in HT-SOFC decreases ohmic polarization of
cell components while electrode kinetics is increased which reduces activation
polarization. However, they require larger startup and shut down time and are also
limited by material costs.
2.2.3.2 Classification according to the cell and stack design
SOFCs may be classified as tubular, planar, and monolithic on the basis of cell
and stack design. Among these cell designs, tubular is the most commonly developed one
which has the thin layered cell components deposited on a cylindrical tube also known as
a SOFC roll.
CH-2 Fuel Cells
29
Fig. 2.4: Diagrammatic presentation of tubular SOFC [47].
However, research is being directed to overcome the problems of sealing in this
design. In the monolithic design, adjacent fuel and oxidant channels are in the form of
honeycomb like array. In SOFCs, the highest power density is achieved in the monolithic
design, but this design fabrication still has many challenges.
Fig. 2.5: Schematic of a planar SOFC [48].
In the planar design, the cells can be stacked without creating voids which is a
problem in tubular design. Also, the ohmic losses and fabrication costs are lower. The
disadvantage of planar design over tubular design is the need for gas-tight sealing.
CH-2 Fuel Cells
30
However, in tubular design, the expansion and contraction of cells is without any
constraints which gives significance to this design.
2.2.3.3 Classification according to the flow configuration
The direction of fuel and oxidant in SOFCs can be cross-flow, co-flow or counter-
flow. The nature of flow affects the temperature distribution within the stack. It has been
found that most uniform temperature distribution is achieved with co-flow configuration
under similar fuel utilization and operating conditions.
a b
Fig. 2.6: The configurations possible for the two reactants’ flow in a fuel cell; a) co-flow
and b) cross flow [49].
2.2.3.4 Classification according to the fuel reforming type
As has been stated earlier, SOFCs accept all types of fuel, which is reformed to
H2 and/or CO before fuel cell operation. This reforming process can be inside the stack,
called internal reforming, or outside stack known as external reforming. Internal
reforming is further divided into indirect internal reforming (IIR-SOFC) and direct
internal reforming (DIR-SOFC). In the IIR-SOFC, the reformer section is separate from
the other components inside the cell but in close thermal contact with the anode section.
In the DIR-SOFC, the reforming takes place directly on the anode catalyst.
CH-2 Fuel Cells
31
2.2.3.5 Classification according to the type of support
SOFCs may be fabricated as anode-supported, cathode-supported or electrolyte-
supported. As the operating temperature of a SOFC is increased, the ionic conductivity of
its electrolyte is also increased. Thus, the electrolyte-supported configuration is generally
selected for high temperature SOFC. For intermediate and low temperature SOFCs, the
electrolyte is usually in a very thin form and electrode supported (cathode or anode)
configuration is chosen. These three types of manufacturing may be called self-
supporting configuration.
a b
Fig. 2.7: SOFC configurations depending on support; a) electrolyte supported and
b) anode supported.
2.3 SOFC Anode
The anode plays an important role in the performance of a fuel cell. It not only
offers reaction sites for electrochemical oxidation of fuel, but also provides path for
transportation of electrons from reaction site to external circuit generated by the
following general reaction;
2
2 2 2, 2 , 2CO H O CO H O e (2.16)
Principally, the anode must be a good electronic conductor with high surface area
and catalytic activity towards oxidation reaction in addition to the general properties
discussed earlier. Porosity is also required for efficient diffusion of fuel and removal of
by products from the reaction sites.
CH-2 Fuel Cells
32
2.3.1 Anode triple-phase boundary
The charge transfer reaction occurs at the triple phase boundaries (TPB)
established at the junction between electrolyte, electrode and gas phase. In other words, it
can be said that TPB is the effective area at which the desired reaction takes place. The
breakdown in connectivity in any of the three phases lowers the performance of the cell
as the desired reaction cannot occur. For example, if ions from the electrolyte are unable
to reach the site or if the removal of electrons could not be accomplished from the site,
then that site would not contribute to the performance of the cell. It has been established
by various theoretical and experimental methods that under normal conditions, TPB
exists approximately 10 µm from the electrolyte into the electrode [50, 51].
SOFCs with enhanced anode TPB have demonstrated improved fuel cell
performance such as higher power density and low resistance loss. The active area of the
anode increases from the contact point between electrolyte and metal for metal anodes, to
the whole surface area for mixed ionic and electronic conductors. Therefore, ionic
conduction is also required for anode materials to achieve advanced fuel cell
performance. Thus a better anode material would be the one having large triple phase
boundary with sufficient ionic conductivity along with the desired electronic
conductivity.
2.3.2 Criteria for selection of anode candidates
The candidate materials should have high electrical conductivity. It has been
shown that materials with electrical conductivity as low as 1 S cm-1
can also be used as
anode for SOFC. Materials should be chemically and mechanically stable under operating
conditions because the anode materials are exposed to reducing atmosphere with fuels at
high temperature (600-1000 °C). They should be sufficiently resistant to oxidation
because of possible presence of H2O and other oxidizing gases (CO2, CO) in the fuel. It
should also be noted that candidate materials should have sulfur tolerance for H2S
containing fuels. Candidate materials should not have any undesirable reactions with
other cell materials during the operation or fabrication process.
The candidate materials should have sufficient catalytic activity for
electrochemical oxidation of various fuels so that polarization losses could be avoided.
CH-2 Fuel Cells
33
Normally, the interfacial resistance of Ni-YSZ is about 0.2 Ω cm2
in the range of 950 -
1000 °C. Thus the interfacial polarization resistance of candidate materials should be
comparable to that of Ni-YSZ.
Thermal compatibility with other cell components is an important criterion which
the candidate materials should meet. Thermal incompatibility may cause de-lamination
and fracture of cell components during fuel cell operation. Thus, new materials should
have a similar coefficient of thermal expansion (CTE) with that of YSZ. Because the
CTE of YSZ is about 10.8×10-6
K-1
the marginal CTE of candidate materials should be
close to that of YSZ e.g., 10 – 12×10-6
K-1
. The properties described above are essential
requirements that must be considered in the development of new anode materials.
2.3.3 Alternate anode materials
Ni-YSZ is regarded as the state of the art anode material because of its excellent
properties. The high catalytic activity of Ni results in a facile anode reaction while the
good electronic conductivity helps in fast transport of electrons from the anode
electrolyte interface to the external circuit. The presence of YSZ provides adequate ionic
conductivity, thus the cermet has a large triple phase boundary. Due to the excellent
catalytic steam reforming activity of Ni, hydrocarbon fuels can be directly fed to the
anode without the need for an external fuel reforming set up.
However it has some inherent drawbacks: importantly, the redox instability,
sulphur poisoning, carbon deposition with hydrocarbons as fuel and sintering of Ni
particles. All these factors affect the anode performance and long term stability of
SOFCs. The performance Ni/YSZ mainly depends on fabrication process as well as on
the characteristics of the initial NiO and YSZ powders. Jiang et al has given good
discussion of important issues in the fabrication and optimization of Ni/YSZ cermet
anodes [52].
Good reviews regarding anode development are available for interested readers
[53-57]. Briefly speaking, research is being carried out in different directions from
modifying the state of the art Ni/YSZ cermet to mixed conductive ceramic oxides as
alternate anode materials.
CH-2 Fuel Cells
34
Ni/YSZ has been modified by replacing Ni with other metals like Ru or Cu and
replacing YSZ with other alternate oxide ion conductors such as ceria stabilized zirconia,
calcium doped ceria, yttria doped ceria or samarium doped ceria [52, 55].
The modification and improvement of metal cermets has resulted in good
performance using either hydrogen or methane as the fuel however none of them is as
efficient as Ni/YSZ. The metal cermets also suffer from general problems like sintering
and volume instability. To reduce the structural mismatch between anode and electrolyte,
single phase oxide anodes have been developed. The major attraction has been given to
mixed ionic and electronic oxides because they result in enhancement of reaction zone
over the three phase boundary.
Among single phase oxides, the fluorite structures having general formula of AX,
with coordination of cation and anion being 4 and 8 respectively have been studied. Well
known examples of oxides with the fluorite structure include yttria-stabilized zirconia
(YSZ) and ceria (CeO2). The former is used as an electrolyte for high temperature SOFCs
where as the latter is used as an electrolyte for low temperature SOFCs due to its
electronic conductivity at high temperature. Both of these are oxide ion conductors. To
have reasonable electronic conductivity, these oxides have been doped to serve as
potential anode materials. Thus, the anodes having the fluorite structure are further
classified into zirconia or ceria based depending on the oxide ion conductor chosen [52 –
54].
Anode materials having rutile, spinel, tungsten bronze, pyrochlores and
perovskites have also been focused on [53]. However, many of the promising oxides used
as SOFC anodes form perovskite-related structures. The perovskite oxide formula can be
written as ABO3 where A is a large cation with a coordination number of 12 and B is a
small cation with a coordination number 6. The small B-site in the perovskite allows first
row transition elements to be introduced into the lattice. These elements exhibit multi-
valence under different conditions, which may be the source of high electronic
conductivity.
In the family of perovskites, titanate-based oxides have good stability and
reasonable electronic conductivity in reducing conditions [54, 56]. They show n-type
CH-2 Fuel Cells
35
behavior at low oxygen partial pressures where the Ti+4
is reduced to Ti3+
, thus
contributing to electronic conductivity. This behavior makes them attractive as potential
anode materials for SOFCs. One of the important members of titanate family is SrTiO3
which is a good electronic conductor in fuel cell conditions and it also exhibits resistance
to sulphur which is one of the limitations of Ni-YSZ cermet anodes. Both A and/or B
sites of the strontium titanate have been doped to tune and tailor the properties of SrTiO3.
Special attention has been given to enhance its electrical conductivity by partial
substitution of Sr2+
on A-site and/or Ti 4+
on B-site. Various lanthanides especially La+3
have been doped at the A-site to increase the electronic conductivity of SrTiO3. Among
B-site dopants, good conductivity values have been found for niobium and yttrium doped
SrTiO3. To further improve the properties, (La,Sr)TiO3 has been doped with several
transition metals (Ni, Co, Cu, Cr and Fe) and Ce. The most effective among these
dopants is cerium, which significantly decreases the polarization resistance, although iron
also produces modest improvements.
Other perovskites investigated as anode materials include vanadates, ferrites,
gallates, niobates and cerates having V+3
, Fe+3
, Ga+3
, Nb+3
and Ce+4
as B-site cations
respectively [53, 57]. The general chemistry of perovskites is detailed in chapter 3.
CH-2 Fuel Cells
36
REFERENCES
1. N.Q. Minh, J Am. Chem. Soc., 1993, 76, 563 – 588.
2. W.R. Grove, Philos. Mug., 1839, 14, 127 – 130.
3. E.A. Merewether in Alternative sources of energy - An introduction to fuel cells,
U.S. Geological Survey Bulletin 2179, 2003.
4. H.H. MoÈ bius, J. Solid State Electrochem., 1997, 1, 2 – 16.
5. N.Q. Minh, T. Takahashi in Science and technology of ceramic fuel cells, Science,
1995.
6. J.M. Andu´ jar, F. Segura, Renew Sust. Energ. Rev., 2009, 13, 2309 – 2322.
7. V. Ramani, Electrochem. Soc. Interface, Spring, 2006, 41 – 44.
8. S.M. Haile, Acta Mater., 2003, 51, 5981 – 6000.
9. J.O.M. Bockris, S. Srinivasan in Fuel cells: Their electrochemistry, McGraw–
Hill, 1969.
10. D. Shekhawat, J.J. Spivey, D.A. Berry in Fuel cells: Technologies for fuel
processing, Elsevier, 2011.
11. F. Zabihian, A.S. Fung in Paths to Sustainable Energy, Ed., Dr Artie Ng, ISBN
978-953-307-401-6, Intech, 2010.
12. A.B. Stambouli, E. Traversa, Renew Sust. Energ. Rev., 2002, 6, 433 – 455.
13. S.K. Pratihar in Applied physics in 21st century, Ed., X. Chen, Research
Singapost, India, 2008.
14. K. Huang, J.B. Goodenough in Solid oxide fuel cell technology: Principles,
performance and operations, Woodhead Publishing Limited, 2009.
15. N. Laosiripojana, W. Wiyaratn, W. Kiatkittipong, A. Arpornwichanop, A.
Soottitantawat, S. Assabumrungrat, Eng. J., 2009, 13, 65 – 83.
16. S. Singhal, K. Kendal, High temperature solid oxide fuel cells: Fundamentals,
design and applications, Elsevier, 2003.
17. R.M. Ormerod, Chem. Soc. Rev., 2003, 32, 17 – 28.
18. B.C.H. Steele, A. Heinzel, Nature, 2001, 414, 345 – 352.
19. A.B. Stambouli, Renew Sust. Energ. Rev., 2011, 15, 4507 – 4520.
CH-2 Fuel Cells
37
20. A.J. Appleby, F.R. Foulkes, Fuel Cell Handbook, Van Nostrand Reinhold, New
York, 1989.
21. G.F. McLean, T. Niet, S.P. Richard, N. Djilali, Int. J. Hydrogen Energy, 2002, 27,
507 – 526.
22. J.H. Wee, Renew Sust. Energ. Rev., 2007, 11, 1720 – 1738.
23. N. Sammes, R. Bove, K. Stahl, Curr. Opin. Solid State Mater. Sci., 2004. 8, 372 –
378.
24. W.M. Vogel, L.J. Bregoli, H.R. Kunz, S.W. Smith, in Proceedings of the
Symposium on Molten Carbonate Fuel Cell Technology, Eds.; R.J. Selman, T.D.
Claar, The Electrochemical Society, Inc., Pennington, 1984.
25. N.Q. Minh, T. Takahashi, Science and Technology of ceramic Fuel Cells,
Elsevier, Amsterdam, 1995.
26. A.S. Nesaraj, J. Sci. Ind. Res., 2010, 69, 169 – 176.
27. S.C. Singhal, Solid State Ionics, 2000, 135, 305 – 313.
28. S.C. Singhal, Electrochem. Soc. Interface, Winter, 2007, 41 – 44.
29. N.Q. Minh, Solid State Ionics, 2004, 174, 271 – 277.
30. J. Fergus, JOM - J. Min. Met. Mat. S. 2009, 59, 56 – 62.
31. G. Stochniol, E. Syskakis, A. Naoumidis, J. Am. Ceram. Soc., 1995, 78, 929 –
932.
32. J. Richter, P. Holtappels, T. Graule, T. Nakamura, L.J. Gauckler, Monatsh Chem.,
2009, 140, 985 – 999.
33. C. Sun, R. Hui, J. Roller, J Solid State Electrochem., 2010, 114, 1125 – 1144.
34. I.M. Ge, S.H. Chan, Q.L. Liu, Q. Sun, Adv. Energy Mater., 2012, 2, 1156 – 1181.
35. M. Gong, X. Liu, J. Trembly, C. Johnson, J. Power Sources, 2007, 168, 289 –
298.
36. P.I. Cowin, C.T.G. Petit, R. Lan, J.T.S. Irvine, S. Tao, Adv. Energy Mater., 2011,
1, 314 – 332.
37. F.M.L. Figueiredo, F.M.B. Marques, WIREs Energy Environ., 2013, 2, 52 – 72.
38. J.W. Fergus, J. Power Sources, 2006, 162, 30 – 40.
39. J.W. Fergus, Mater. Sci. Eng. A. 2005, 397, 271 – 283.
CH-2 Fuel Cells
38
40. W.Z. Zhu, S.C. Deevi, Mater. Sci. Eng. A, 2003, 348, 227 – 243.
41. B.C.H. Steele, Nature, 1999, 400, 619 – 621.
42. M. Krumpelt, T.R. Krause, J.D. Carter, J.P. Kopasz, S. Ahmed, Catal. Today,
2002, 77, 3 – 16.
43. K. Kendall, S.C. Singhal in Solid Oxide Fuel Cells, Elsevier, 2004.
44. J. P. Ackerman, J. E. Young, "Solid oxide fuel cell having monolithic core," US
Patent 4476198, 1984.
45. P. Aguiar, N. Lapena-Rey, D. Chadwick, L. Kershenbaum, Chem. Eng. Science,
2001, 56, 651 – 658.
46. A.L. Lee, R.F. Zabransky, W. J. Huber, Ind. Eng. Chem. Res., 1990, 29, 766 –
773.
47. S.C. Singhal, Mater. Res. Soc. Bull., 2000, 25, 16 – 25.
48. http://people.bath.ac.uk/cf233/sofc.html.
49. http://www.doitpoms.ac.uk/tlplib/fuel-cells/printall.php.
50. V.M. Janardhanan, V. Heuveline, O. Deutschmann, J. Power Sources, 2008, 178,
368 – 372.
51. W. Zhu, D. Ding, C. Xiaz, Electrochem. Solid-State Lett., 2008, 11, B83 – B86.
52. S.P. Jiang, S.H. Chan, J. Mat. Sci., 2004, 39, 4405 – 4439.
53. S. Tao, J.T.S. Irvine, Chem. Rec., 2004, 4, 83 – 95.
54. J.W. Fergus, Solid State Ionics, 2006, 177, 1529 – 1541.
55. A. Atkinson, S. Barnett, R.J. Gorte, J.T.S. Irvine, A.J. Mcevoy, M. Mogensen,
S.C. Singhal, J. Vohs, Nature, 2004, 3, 17 – 27.
56. J.B. Goodenough, Y.H. Huang, J. Power Sources, 2007, 173, 1 – 10.
57. W.Z. Zhu, S.C. Deevi, Mater. Science Eng. A, 2003, 362, 228 – 239.
Chapter 3
Perovskite Oxides
39
Perovskite Oxides
Abstract
The intrinsic limitation of Ni-YSZ cermet has triggered the development of
alternate anode systems for SOFCs. In this context, perovskite oxides appear to be
suitable anode candidates due to their remarkable properties. This chapter gives a brief
overview of perovskites and their underlying defect chemistry which governs their
electrical conductivity.
3.1 Perovskite Oxides
Perovskite oxides are a versatile class of single phase oxides which have been
extensively investigated because of their important physical characteristics such as
ferroelectricity, piezoelectricity, pyroelectricity, magnetism, high temperature
superconductivity, catalytic activity and electro-optic effects [1-4].
Perovskite oxides have general formula of ABO3 where A is the larger cation and
is 12- fold coordinated with oxygen atoms. Usually it belongs to alkaline earth and/or
rare earth metal ions, while, small transition metal ion (usually from 3d series like Ti, Cr,
Mn) resides on the B-site and is 6-fold coordinated with oxygen atoms. A general
skeleton of perovskite depicting corner sharing octahedra is shown in Fig. 3.1 where
centre position is occupied by the A cation. The structure can also be viewed with the A
cation in centre of cube while the B cation in the centre of oxygen octahedra [5]. This
frame work offers numerous cations to be incorporated, thus a variety of perovskites
structures are possible.
3.1.1 Perovskite structure
The ideal perovskite structure displays cubic symmetry with Pm3m-Oh space
group [6]. Ideal perovskite symmetry can be characterized by a tolerance factor (t), as
introduced by Goldschmidt [7]. It is used to measure the deviation from ideal situation
and is defined by:
CH-3 Perovskite Oxides
40
( )
2
A O
B O
r rt
r r
(3.1)
t is unity for an ideal perovskite, however even for lower t-values i.e., 0.75 < t < 1.0, the
cubic structure is observed [8, 9]. The symmetry is lowered to orthorhombic,
rhombohedral, tetragonal, monoclinic and triclinic for small t values [10, 11]. A simple
distortion and/or enlargement of the cubic unit cell results in a distorted structure which
transforms to cubic symmetry through intermediary distorted phases in a number of steps
at high temperature.
Fig. 3.1: Corner sharing octahedra in perovskites.
Interesting defect chemistry is offered by perovskites by full or partial substitution
of A and/or B cations. Substitution of the B-site with cations having different size and
charge can modify the simple perovskite structure. For a perovskite having equal atomic
substitution of two ions at the B-site, the general formula could be written as A2BB/O6 (or
AB0.5 B/0.5O3). In the case of differently charged cations, the octahedral symmetry of B
and B/ is preserved although oxygen atoms are slightly shifted to more charged cation.
3.1.2 Nonstoichiometry in perovskites
One of the conditions for substitution in perovskites is the preservation of
electroneutrality besides ionic requirements. Thus, a variety of compounds can exist
CH-3 Perovskite Oxides
41
depending on appropriate charge distribution of A and B cations leading to compounds
such as A+1
B+5
O3, A+2
B+4
O3, or A+3
B+3
O3. In addition to this, defected perovskites also
exist depending on cation or anion nonstoichiometry. A brief discussion about
nonstoichiometry in perovskites is given below;
3.1.2.1 Oxygen nonstoichiometry
In perovskites, oxygen deficient stoichiometry is mostly found and the general
formula is written as AnBnO3n-δ where δ denotes the oxygen deficiency. The browmillerite
structure with ordered anion vacancies exhibited by Ca2Fe2O5 and La2Ni2O5 can be
quoted as an example of oxygen deficient perovskites [12, 13].
However, oxygen excess stoichiometry is less common as the incorporation of
extra oxygen in the interstices is thermodynamically unfavorable. Ba1-λLaλTiO3+λ/2,
LaMnO3+λ and EuTiO3+λ are the examples of oxygen excess stoichiometry. Among these,
λ is found to be 0.12 for LaMnO3+λ. It was shown from neutron diffraction studies that
the excess oxygen is accommodated with partial removal of La as La2O3 and formation of
vacancies at the A and B sites [14, 15].
3.1.2.2 Cation nonstoichiometry
In perovskites, cationic vacancies are comparatively less common than anion
(oxygen) vacancies. B-site vacancies in perovskite oxides are rarely found. In fact, due to
large charge to size ratio of B cations, vacancies at the B-site are thermodynamically
unfavourable. Ba5Ta4O15 shows vacancies at the B-site where the vacancy is present at
the octahedral site [16].
A-site vacancy is commonly found in perovskites as the stable network of BO3
array favors vacancy at the A site. Cu0.5TaO3 is an interesting example of A-site deficient
perovskite which has orthorhombic unit cell and displays a pseudocubic perovskite
structure [7].
3.2 Defect Chemistry of Perovskites
Interesting defect chemistry is offered by the perovskites by full or partial
substitution of cations at A and/or B-sites. In this section, a general discussion about
defect chemistry is given with an aim to understand defect chemistry of perovskites.
CH-3 Perovskite Oxides
42
3.2.1 Defects
A perfect crystal is characterized by all its atoms having a defined lattice position
in the structure. Above 0 K there are always defects in the structure of a perfect crystal,
which may be extended defects such as dislocations or they can occur at isolated atomic
positions, known as point defects [17, 18].
Fig. 3.2: Thermodynamics of defect formation.
Point defects are caused due to deviations from the perfect atomic arrangement or
stoichiometry. They significantly affect the chemical and physical properties of the
crystalline solid, such as the diffusion, electric conductivity and the reactivity. The
creation of point defects in an elemental, crystalline solid is entropy driven as the
enthalpy of the defect formation is positive [19] as shown in Fig. 3.2. The increase of
configurational entropy of the system is enough to provoke an increase in the Gibbs free
energy, which makes the incorporation of more defects easier. Thus, point defects will
always be present in a crystal above 0 K thermodynamically. These are further divided
into ionic and electronic defects [20-22].
3.2.1.1 Ionic defects
Atoms and ions occupy regular positions to define the respective crystalline
system. If some of these ions are missing from their position, the ionic defect is termed as
a vacancy. Similarly, the occupancy of atoms or ions at interstitial positions of a perfect
CH-3 Perovskite Oxides
43
crystal also results in an ionically defective crystal. Also, the substitution of atoms by
small amounts of impurities in a crystal causes ionic defects.
3.2.1.2 Electronic defects
Formation of electronic defects (formation of electrons, holes) can be analyzed by
Fermi statistics. At 0 K, only the lower energy levels are occupied up to the level called
the Fermi energy level (valence band). Electrons in higher energy levels are located in the
conduction band and they contribute in the conduction process. Above 0 K, thermal
excitation causes the jump of some of the electrons from the valence band to the
conduction band enabling them to participate in the conduction process. When a
transition of an electron from the valence band to the conduction band occurs, it creates
an electron plus an electron hole pair (e′+h˙) according to the following equation:
' .Null e h
Kröger-Vink notation is used to express the electronic and ionic defects. Table 3.1
provides the most common notations for ionic and electronic defects for incorporation of
defects in a binary oxide, MO where M corresponds to a divalent cation.
Table 3.1 Kröger-Vink notation for point defects in binary oxide, MO
Defect Symbol Effective charge
Free electron 'e -1
Free electron (hole) .h +1
Vacancy at M site ''
MV -2
Substitution of M at M site MM +2
A+ (acceptor) dopant at M site
'
MA -1
D+3
(donor) dopant at M site .
MD +1
M at interstitial site ..
iM +2
O at interstitial site ''
iO -2
CH-3 Perovskite Oxides
44
3.2.2 Rules for writing defect reactions
The following rules should be considered while writing defect reactions;
1. No mass can be created or lost in a defect chemical reaction.
2. The ratio of cation and anion sites of the crystal must be preserved, although the
total number of sites can be increased or decreased.
3. The total effective charges should be balanced in the Kröger-Vink notation
system.
3.2.2.1 Examples of defect chemical reactions
The defect formation reactions typically include;
1. Predominant intrinsic and ionic defects (Frenkel or Schottky)
2. Intrinsic electronic defects
3. Oxidation and reduction reactions
4. Solute incorporation
These defects are expressed in Kröger Vink notation as follows;
3.2.2.1.1 Frenkel defect
For a Frenkel defect in AgCl, the dominant mechanismis written in Kröger-Vink
notation is as;
. 'x x
Ag i i AgAg V Ag V (3.2)
3.2.2.1.2 Solute incorporation
The substitution of an isovalent cation like Ni+2
on the Mg+2
site in MgO is
expressed as
x x
Mg oNiO Ni O (3.3)
In the case of aliovalent cation substitution like Al2O3 addition to MgO, Al+3
will
substitute Mg+2
and oxygen ions are likely to occupy additional oxygen lattice, the
respective defect reaction is written as;
. ''
2 3 2 3 x
Mg o MgAl O Al O V (3.4)
These examples display ionic compensation upon solute incorporation into oxide
structure.
CH-3 Perovskite Oxides
45
3.2.2.1.3 Oxidation and reduction reactions
The reduction of an oxide in which oxygen is removed is accompanied by the
creation of oxygen vacancies and is written as
.. '
2
12
2
x
O oO O V e (3.5)
This mechanism also explains the increase in conductivity of an n-type oxide in
reducing conditions where electrons act as charge carriers.
The oxidation can be written as the consumption of oxygen vacancies
.. .
2
12
2
x
o OO V O h (3.6)
These reactions reflect electronic compensation as electronic defects (electrons or
holes) are created.
3.2.3 Electronic vs. ionic compensation
In oxide semiconductors, the effectiveness of a donor or an acceptor is governed
by their ionization energies as well as the extent of oxidation and reduction. In fact, an
aliovalent impurity in an ionic compound can be charge-compensated by ionic defects
(ionically compensated) or by electrons or holes (electronically compensated) or by a
combination of the two. The variables which govern the mode of compensation are the
oxygen partial pressure; pO2, the dopant concentration and temperature.
Upon Nb2O5 doping in TiO2, the incorporation of Nb+5
onto Ti+4
sites can be
compensated by any of these defect reactions;
Ionic compensation . ''''
2 52 4 10 x
Ti O TiNb O Nb O V (3.7)
Electronic compensation . '
2 5 22 4 8 4x
Ti ONb O Nb O O e (3.8)
Thus Nb doping of TiO2 tends to be ionically compensated (formation of titanium
vacancies) if Nb2O5 concentration is large, pO2 is high and the temperature is low,
whereas the reverse conditions favor the electronic compensation. Similar effects are
observed in case of titanates such as BaTiO3.
3.3 Electrical Conductivity in Oxides
Oxides have conductivity ranging from insulators, through semiconductors and
metallic conductors, to superconductors [23]. However, certain features characterize the
conductivity of oxides. For example, their conductivity profile shows a metal insulator
CH-3 Perovskite Oxides
46
transition in which an insulator shows metallic behaviour under certain conditions
(composition, temperature of pressure) [24]. Another characteristic feature is the
dependence of their conductivity on oxygen activity where the change of oxygen activity
causes either an increase or decrease in carrier concentration thus affecting the
conductivity [25].
3.3.1 Electrical conductivity
According to Wagner [26], an ionic crystal can be considered as a mixed
conductor and its conductivity can be written as the sum of the partial conductivity
associated with each type of charge carrier as given by
j e i
j
(3.9)
where j is partial conductivity associated with each type of charge carrier. e and i signify
the conductivity contribution from electronic and ionic charge carriers respectively. As
discussed, the electronic conductivity of oxides has oxygen partial pressure dependence;
thus e can be further expressed as
2 2
1/ 1/n n
e n O p OP P (3.10)
In above equation, n and p implies the generation of electrons and holes
(electronic compensation) due to change of partial pressure of oxygen and their role in
resulting conductivity.
The partial conductivity associated with each type of charge carriers can be
obtained by multiplying the concentration with the respective charges and mobility. Thus,
j can be written as
( )j j j jn z e (3.11)
where jn is the concentration of charge carrier j, jz e is its charge, and j is the
mobility.
The fraction of the total conductivity contributed by each charge carrier is given
by transference number jt ;
j
jt
(3.12)
CH-3 Perovskite Oxides
47
This equation can be broken down into contribution from both ionic (ti) and electronic
contribution (te) where the electronic transference number te contains the contribution from
electron and hole transference numbers tn and tp respectively. Since neither ti nor te is
truly zero, all solids are in principle mixed conductors.
Both of these conduction modes (ionic and electronic) are independent of each
other. Usually the ionic conduction in a crystal is determined by its structure where as the
band gap decides the electronic conductivity. Thus a good ionic conductor may or may
not be a good electronic conductor. Mostly, et ≠ and thus only one type of charge carrier
predominates. The nature of electrical conduction in oxides is determined by relative
magnitude of these transference numbers.
3.3.1.1 Electronic conductors: te >> ti
In electronic conductors, the electrical conductivity is due to the electrons in the
conduction band or missing electrons (electron holes) in the valence band. The
conduction occurs either by movement of delocalized electrons (free electrons) in the
conduction band or by hopping of localized electrons from one potential well to the other
in valence band. The width of the band governs the mode of conduction where the free
movement of electrons occurs in the case of a wide band while a hopping mechanism
occurs in the case of a narrow band. Mostly hopping mechanism occurs in the case of
compounds containing transition metals which have more than one oxidation state.
A good electronic conductor is characterized by a small band gap [23]. By
selecting the optimal temperature, oxygen partial pressure and dopant (nature and
concentration), the electronic compensation can be achieved. A variable valent cation
leads to high conductivity due to following defect reaction
.. '
2
12
2
x
O oO O V e (3.13)
CeO2 can be considered as an example which shows higher conductivity than
ThO2 or ZrO2 which is attributed to existence of Ce in two oxidation states of +3 and +4.
In reducing conditions, Ce+4
goes to Ce+3
accompanied by defect reaction shown in Eq.
3.13.
it
CH-3 Perovskite Oxides
48
3.3.1.2 Ionic conductors: te << ti
Ionic conductors have ionic charge carriers as the dominant mode of charge
transport. The ionic conductivity involves mass transport resulting from the diffusion of
the ionic defects like vacancies and/or interstitials. The ionic mobility is a thermally
activated process which increases strongly with temperature. The ionic conductivity
becomes appreciable only at high temperatures, where defect concentrations become
quite large and ions have high thermal energy.
The ionic defects may be present intrinsically in the form of Schottky, Frenkel or
antisite disorder. In oxides, the Frenkel disorder on the oxygen sublattice is common and
given by
'' ..x
O i OO O V (3.14)
The reduction reaction shown by Eq. 3.13 also shows ionic defect formation.
Also, the doping of aliovalent impurities cause the ionic defects in which charge is
compensated by oxygen vacancies. One of the well known examples of oxide ion
conductors is yttria-stabilized zirconia (YSZ) which has large defect concentration [27]
and is a good oxide ion conductor at high temperature, thus is used as an electrolyte for
solid oxide fuel cell.
3.3.1.3 Mixed ionic and electronic conductors: te ≈ ti
Mixed ionic–electronic conductors (MIECs) conduct both ions and electronic
charge carriers [28]. These conductors are characterized by the close magnitude of ionic
and electronic transference numbers. Thereby, the condition, te ≈ ti, requires that
i i e en n
Due to the negligible mass of electrons, the electronic mobilities are more than 100
times more than ionic mobilities, thus for a mixed ionic and electronic conductor to have
the same electronic and ionic transference number, it must possess 100 times greater
concentrations of ionic carriers than electronic carriers. The oxides having high densities
of mobile ions induced by doping or by crystallographic disorder as well as relatively
small energy band gap satisfy this condition. These mixed ionic and electronic
conductors have been used as electrodes of fuel cells where the triple phase boundary is
increased due to their mixed mode of conduction.
CH-3 Perovskite Oxides
49
3.3.2 Effect of temperature on conductivity
Equation 3.11 shows that conductivity depends on carrier concentration as well as
on their mobility. In the case of metals, the large density of charge carriers remains
unaffected by temperature and conductivity majorly depends on the mobility, which
decreases with temperature increase, resulting in a decrease in conductivity. In the case of
semi-conductors, temperature affects both the concentration and mobility of charge
carriers. Mathematically, we can write
q T n T T p Tn p (3.15)
where n and p signify the concentration of electrons and holes while n and p are the
respective mobilities.
3.3.2.1 Temperature dependence of mobility of semiconductor
The mobility of charge carriers in a semiconductor is affected by two scattering
mechanisms; lattice scattering and impurity scattering. In lattice scattering, the scattering
of charge carriers is the outcome of thermal vibration of the lattice atoms. As the
temperature increases, the thermal vibration becomes greater and the frequency of such
collisions increases. Thus, lattice scattering dominates at high temperature resulting in
decreased mobility [29].
The mobility due to lattice scattering is related to temperature by following
equation
32
l T
(3.16)
The other scattering mechanism operational at low temperatures is impurity
scattering caused by ionized impurities. When a charge carrier moves across such
impurities, the coulombic forces result in deflection in its path and hence mobility is
decreased. The probability of this scattering depends on the total concentration of the
ionized impurities. The impurity scattering becomes insignificant at high temperature
because the carriers move faster and thus, are less influenced by the coulombic forces.
Actually, the slow movement of carriers at low temperatures causes their interaction with
charged impurities. The carrier mobility due to impurity scattering is related to
temperature and doping concentration, N by
CH-3 Perovskite Oxides
50
32
i
T
N (3.17)
Fig. 3.3 shows the dependence of mobility on temperature where impurity
scattering is seen only at low temperatures while lattice scattering plays a dominant role
at high temperatures [29].
Fig. 3.3: Approximate temperature dependence of mobility with
lattice and impurity scattering [29].
3.3.2.2 Temperature dependence of carrier concentration
The promotion of electrons from the valance band into the conduction band is a
thermally activated process. On increasing the temperature, the electrons in the valance
band gain energy and go in the conduction band where they contribute to conductivity. In
fact, the number density of free carrier electrons, n, in the conduction band is an
exponential function of temperature [23] as given by Eq. 3.18.
ni T 22kT
h2
3
2
mn*m p
* 3
4 expEg
2kT
(3.18)
Changing the temperature in a semiconductor has a much greater effect on the
carrier concentration than on the mobility, and the conductivity normally increases with
temperature.
CH-3 Perovskite Oxides
51
3.3.3 Effects of oxygen partial pressure on conductivity
One of the most important features of conducting oxides is the dependence of
their conductivity on oxygen partial pressure. It is assumed that equilibrium between
oxygen partial pressure, oxygen vacancies, oxygen interstitials, and electronic defects
(electrons and holes) determine the conductive behaviour of a compound [25, 30].
A Brouwer diagram is used to describe the defect concentration as a function of
temperature, oxygen activity and dopant concentration on a double log plot. One of such
plots is shown in Fig. 3.4, which explains the dependence of defects (electrons, electron
holes or ions) concentration on oxygen partial pressure [31].
Fig. 3.4: Brouwer diagram showing dependence of defect concentration on oxygen
activity in case of n and p-type oxides [31].
Three regions can be distinctly seen in Fig. 3.4;
1. Region I (low oxygen partial pressure, reducing conditions)
In this region, the material acts as an n-type conductor and shows pO2-1/6
dependence.
CH-3 Perovskite Oxides
52
2. Region II (intermediate oxygen partial pressure)
The concentration of charge carriers remains constant with changing oxygen
partial pressure in this region.
3. Region III (high oxygen partial pressure, oxidation conditions)
In this region, the material is a p-type conductor and exhibits pO21/6
dependence.
By analogy, the conductivity of a material can also be expressed as function of oxygen
partial pressure as shown in Fig. 3.5.
Fig. 3.5: Effect of oxygen partial pressure on the partial conductivity [31].
Such figures are very useful in determing the conducting nature of material by
observing the conductivity profile as a function of oxygen activity.
CH-3 Perovskite Oxides
53
REFERENCES
1. L.J. Tejuca, J.L.G. Fierro in Properties and Applications of Perovskite Type
Oxides, Marcel Dekker, 1993.
2. A.S. Bhalla, R. Guo, R. Roy, Mat. Res. Innovat., 2000, 4, 3 – 26.
3. G. Centi, F. Trifiro, Catal. Rev. Sci. Eng., 1986, 28, 165 – 184.
4. N. Mizuno, M. Misono, Chem. Rev., 1998, 98, 199 – 218.
5. R.J. Brook in Concise Encyclopedia of Advanced Ceramic Materials, Pergamon
Press.1991.
6. W.R. Moser in Catalytic Chemistry of Solid State Inorganics, New York
Academy of Sciences, 1976.
7. M.A. Pen˜a, J.L.G. Fierro, Chem. Rev., 2001, 101, 1981 – 2017.
8. A.F. Wells in Structural Inorganic Chemistry, Oxford Science publications.1995.
9. U. Müller in Inorganic Structural Chemistry, Wiley & Sons Ltd, 1993.
10. C.P. Khattak, F.F.Y. Wang in Handbook of the Physics and Chemistry of Rare
Earths, Eds.; K.A. Gschneider, Jr., L. Eyring, North-Holland Publisher,
Amsterdam, 1979.
11. J.B. Goodenough in Solid State Chemistry, Ed.; C.N.R. Rao, Marcel Dekker,
1974.
12. D.M. Smyth in Properties and Applications of Perovskite-Type Oxides, Eds.; L.G.
Tejuca, J.L.G. Fierro, Marcel Dekker, 1993.
13. M.J. Sayagues, M.V. Reg, A. Caneiro, J.M. Gonzalez-Calbet, J. Solid State
Chem., 1994, 110, 295 – 304.
14. B.C. Tofield, W.R. Scott, J. Solid State Chem., 1974, 10, 183 – 194.
15. R.J.H. Voorhoeve, J.P. Remeika, L.E. Trimble, S.A. Cooper, F.J. Disalvo, K.P.
Gallagher, J. Solid State Chem., 1975, 14, 395 – 406.
16. C.N.R. Rao, J. Gopalakrishnan, K. Vidyasagar, Indian J. Chem. Sect. A, 1984,
23A, 265 – 284.
17. F. Agullo-Lopez, C.R.A. Catlow, P.D. Townsend in Point defects in materials,
Academic Press Limited, 1988.
18. L.E. Smart, E.A. Moore in Solid State Chemistry: An introduction, CRC Press,
2005.
CH-3 Perovskite Oxides
54
19. F.A. Kroger in The chemistry of imperfect crystals, North-Holland Publishing
company, 1974.
20. S.M. Naido in Applied Physics, Dorling Kindersley, India, 2009.
21. R.W. Siegel in Point Defects and Defect Interactions in Metals, Eds.; J.I.
Takamura, M. Dōyama, M. Kiritani, North Holland, 1982.
22. J.H. Crawford, L.M. Slifkin in Point Defects in Solids, Plenum Press, 1975.
23. P.A. Cox in Transition Metal Oxides, Clarendon Press, 1992.
24. M.A. Imada, A. Fujimori, Y. Tokura, Rev. Mod. Phys., 1998, 70, 1039 – 1263.
25. P. Kofstad in Nonstiochiometry, Diffusion and electrical conductivity in binary
metal oxides, R.E. Krieger Publishing, 1983.
26. C. Wagner, Proc. Intern.Comm. Electrochem. Thermodyn. Kinet., 1957, 7, 361.
27. S. Hull, Rev. Prog. Phys., 2004, 67, 1233 – 1314.
28. I. Riess, Solid State Ionics, 2003,157, 1 – 17.
29. N. Dasgupta, A. Dasgupta in Semiconductor Devices: Modeling and technology,
PHI Learning Pvt. Ltd., 2004.
30. A.J. Moulson, J.M. Herbert, in Electroceramics, Chapman and Hall, 1990.
31. W. Gao, N.M. Sammes in An introduction to electronic and ionic materials,
World Scientific, 1999.
Chapter 4
Characterization Techniques
55
Characterization Techniques
Abstract
After synthesis, different techniques are used to characterize the samples. This
chapter gives a brief detail of techniques including thermal gravimetric analysis, X-Ray
diffraction, scanning electron microscopy, particle size analysis, dilatometry, ac
impedance and electrical conductivity measurements employed in the present research
project for characterization and investigation of the samples.
4.1 Thermal Gravimetric Analysis
Among thermal analysis techniques, the thermogravimetry analysis (TGA) and
differential thermal analysis (DTA) are the most common. TGA measures the change in
the sample mass as a function of temperature in a controlled atmosphere and provides
both qualitative and quantitative analysis. A TGA curve provides information related to
thermal stability and thermal decomposition profile of initial powders as well as of any
intermediate compounds formed during the process. In addition to this, this technique can
also be used to measure the oxygen stoichiometry. A sensitive balance is used to
accurately weigh the variation of the sample mass. While in DTA, the difference in
temperature ΔT between a sample and an inert reference material as a function of
temperature is measured [1].
4.2 X-Ray Diffraction (XRD)
X-Ray diffraction is a non-destructive, versatile, rapid and sensitive analytical
technique used in the characterization of solid crystalline materials. It gives information
about the crystallographic structure like (texture, crystallinity, phases, grain size and
crystal defects) of materials and thin films at the atomic level as the wavelength of X-rays
(0.5 and 2.5 Å) lies in the same order as spacing (d) between the crystal planes [2, 3].
CH-4 Characterization Techniques
56
Every crystalline material has its unique diffraction pattern which serves as finger print
for their characterization and phase identification.
4.2.1 Generation of X-rays
The X-rays are generated in vacuum by applying a potential difference, of tens to
hundreds kV, between cathode and a metallic target working as an anode. Usually,
copper and molybdenum are used as the targets. When the cathode filament of tungsten is
heated, highly energetic electrons released by thermo ionic effect are accelerated through
vacuum to the target due to potential difference. The inelastic collision of electrons with
the target results in knocking out the electrons in the internal layers of target creating a
high-energy excitation state. On de-excitation, electrons of the external layers jump to the
internal layer, causing the emission of X-ray radiation characteristic of the target [4].
Monochromatic incident beam is obtained by directing the X-ray through a filter. On
striking the sample, the X-rays are scattered from each set of lattice planes at specific
angles depending on its crystal structure. The result is displayed in the form of a spectrum
of the scattering intensity as a function of the incident or scattering angle (diffractogram).
An X-ray diffraction experiment can be done in two ways; by using reflection or
transmission mode. In the reflection method, the incident beam penetrates the top layers
of the sample and then reflected towards the collector. The incident angle of the beam
varies in order to sweep all possible angles for constructive interference to obtain
diffraction pattern. For the transmission method, the beam is transmitted through the
sample. Samples of different thickness have to be used in the two different methods. In
the reflection case, a thick sample is used for the incident beam to penetrate the top
sample layer only without reaching the sample support while in the transmission method,
the sample should be thin enough so that the incident beam can pass through the sample
The main differences between these methods are the beam collector position, in relation
to the sample position and the thickness of the sample [5].
CH-4 Characterization Techniques
57
Fig. 4.1: Working of an X-ray powder diffractometer.
4.2.2 Bragg’s law
It is a simple mathematical formulation derived by the English physicists Sir
W.H. Bragg and his son Sir W.L. Bragg in 1913 which explains the phenomenon of
reflection of X-rays in crystals.
The law can be stated as,
n λ = 2dsin θ (4.1)
In this equation, d is the distance between atomic layers in a crystal. λ is the
wavelength of the incident X- ray beam, θ is the angle of incidence and n is an integer
[6].
Equation 4.1 dictates the conditions of diffraction and explains why X-ray
reflection occurs only at a certain angle, θ in a crystalline substance. In case of
constructive interference, the waves reflected by the planes are in phase with one another,
producing diffraction peaks of different intensities. To satisfy this condition, the distance,
d between the planes should be equal to an integral of wavelength, λ (Fig. 4.2). If the
waves reflected by the planes are out of phase with one another, destructive interference
would result and no diffraction peak would be observed in the diffraction pattern.
CH-4 Characterization Techniques
58
Fig. 4.2: Incident and reflected X-rays from a specified crystal plane.
4.2.3 Calculations for crystallite size
From XRD spectrum, mean crystallite size can be calculated by using Scherrer
equation [7, 8] as given;
57.3
cos
kD
(4.2)
where, D is the crystallite size, is the broadening of diffraction line measured at
half of its maximum intensity known as full width at half maximum (FWHM). k is the
shape factor and λ is the wavelength of the X–ray beam. It can be seen that crystallite size
is inversely related to FWHM of an individual peak; the narrower the peak, the larger
would be the crystallite size.
4.2.4 Calculations for theoretical density
From the XRD data, the theoretical density of crystalline solids can also be
calculated using following equation;
1.66th
m M Z
V V
(4.3)
In above equation, m is the mass and V is the volume of a unit cell given as
a b c where a, b and c are unit cell parameters. M is the formula weight in g mol-1
while Z stands for numbers of formula units in each cell.
CH-4 Characterization Techniques
59
4.3 Scanning Electron Microscopy (SEM)
Electron microscopy is an extremely versatile and powerful technique capable of
providing a direct and detailed description concerning the morphologies (microstructural
information) or compositions of a specimen over a wide range of magnification. It is
preferred over the light microscopy as the resolution of a light microscope cannot go
beyond 400 nm due to the wavelength limitation of visible light. However, the much
smaller wavelength of electrons as given by de Broglie relationship (Eq. 4.4) reveals very
fine details due to high resolution in an SEM image [9, 10].
h
mv (4.4)
where h is the Planck constant 6.63×10-34
Js and m is the mass and v is the velocity of the
electrons. This technique gives information about particle size, shape of the powder and
also can be used to study porosity and microstructure of fully or partially dense bodies.
4.3.1 Principle of SEM
In SEM, a beam of accelerated electrons is applied to the sample to give
information about the surface topography, composition and other properties of the
sample. The working of scanning electron microscope is shown in Fig. 4.3.
A potential difference (5 to 30 kV) is applied to the tungsten filament which
results in thermo-ionic emission of an electron beam. Several condenser lenses are used
to focus the electrons to a spot around 50-100 Å in diameter. The electron beam interacts
with the material generating secondary electrons (electrons from the surface atoms of the
material which get ejected due to inelastic collision of the electron beam with atoms of
the sample), backscattered electrons (electrons from the beam that have undergone
interaction with the nuclei of the atoms in the sample), characteristic X-rays and so on
which are collected selectively to form the image [11].
For a poorly conductive sample, gold sputtering becomes necessary to obtain a
good micrograph. Gold is used due to its inert nature and good conductivity.
CH-4 Characterization Techniques
60
Fig. 4.3: Schematic of scanning electron microscopy [12].
4.4 Particle Size Analysis
Particle size analysis is a simple, fast and flexible technique which measures the
particle size distribution in the powder. Ideally, it is desired to have uniform shaped and
sized particles in a sample. However, particles have a range of sizes and may have
different shapes in practice. For such samples, particle size distribution (PSD) is a useful
parameter that describes distribution of particles in different size ranges quantitatively
[13]. Laser diffraction based on scattering of laser light by particles of different sizes is
used to measure particle size distribution.
4.4.1 Basic principle of laser diffraction
The basic principle of laser diffraction is based on scattering of laser beam by
particles in suspension. The scattering angle depends on size of particles and the
scattering angle is inversely related to the sizes of particles in a sample (Fig. 4.4). A
photo detector array is used to detect the scattered light. A mathematical algorithm is
used to convert the intensity of light on each detector to particle size distribution plots.
CH-4 Characterization Techniques
61
This technique allows one to measure particle sizes from sub-micron to a few millimeters
[14]. A well-dispersed and homogenous suspension is needed for reliable and precise
particle size measurement.
Fig. 4.4: Scattering of beam from large and small particles.
4.5 Dilatometry
Dilatometry is a useful technique in which the change in linear dimension of a
sample as a function of temperature is recorded under negligible mechanical load. It
provides information about thermal expansion, coefficient of thermal expansion, density
changes, decomposition, sintering, phase transition and glass transition temperature of a
sample [15]. This technique is also called as thermomechanical analysis (TMA). In the
case of fuel cell research, this technique has got quite importance as it enables to
investigate the shrinkage behavior of different components of the cells.
Among different types of dilatometer, a connecting rod (push rod) dilatometer has
simple design as shown in Fig. 4.5. The sample is fixed in the dilatometer tube with the
help of push rod which lies in the axis of the tube. A thermocouple is placed close to the
sample to measure the temperature. When the sample is thermally treated, the
corresponding dimensional change affects the movement of pushrod which is sensed by a
transducer [16]. However, the push-rod movement is the outcome of the expansion of the
sample as well as of the dilatometer tube. To eliminate this error, the calibration of
CH-4 Characterization Techniques
62
dilatometer is done with a well characterized sample of known expansion called a
standard or reference. For accurate calibration, the length of both sample and reference
should be close to each other and both should undergo the same temperature profile.
Fig. 4.5: Functional diagram of a pushrod dilatometer.
4.6 Ac Impedance
Impedance spectroscopy is a powerful technique to characterize electrochemical
properties of materials by measuring the characteristic response of a process toward
applied voltage occurring within a material.
In this technique, an ac excitation voltage of different frequencies is applied to the
sample under consideration and the current response is determined. The beauty of this
technique lies in the fact that it can separate different processes occurring in a system on
the basis of their characteristic relaxation times [17].
4.6.1 Theory
In the impedance spectroscopy, a small amplitude ac voltage ( )V of different
frequencies is applied between the working and the reference electrode and the resulting
current response ( )I is measured [18, 19]. The impedance ( )Z is obtained by dividing
the applied voltage (V) by the current response (I) as given by Eq. 4.5.
( )
( )( ) (cos sin )
( )
o
o
j t
o
j t
o
V e VVZ j
I I e I
(4.5)
CH-4 Characterization Techniques
63
In this equation, ω is the angular frequency, j the imaginary unit and φ is the
phase shift. Since impedance Z(w) is a complex vector, it can be written as sum of real
and imaginary components. Thus, we can write
real imaginaryZ Z Z (4.6)
where
' cosrealZ Z Z (4.7)
and
'' sinimaginaryZ Z j Z (4.8)
The modulus of impedance is given by
' 2 '' 2( ) ( )Z Z Z (4.9)
A simple mathematics shows that the phase angle is related to these components by
relation;
''1
'tan
Z
Z
(4.10)
It is customary to represent the impedance in a complex plane called as Nyquist
plot where imaginary part ( ''Z ) is plotted versus the real part ( 'Z ) as shown in Fig. 4.6.
Fig. 4.6: Impedance represented in Nyquist plot.
CH-4 Characterization Techniques
64
A Nyquist plot does not incorporate frequencies, thus, sometimes impedance is
also presented in Bode mode, where imaginary and real components are plotted as a
function of frequency. A Bode plot is capable to separate processes that occur at different
frequencies having different relaxation times.
The impedance data can be represented by some other quantities which are
derived from the impedance. These include the admittance (Y), modulus (M) and
permittivity (ε) and are listed in Table 4.1.
Table 4.1 Relations between the four basic immittance functions [17]
M Z Y ε
M (modulus) M µZ µ Y-1
ε-1
Z (impedance) µ-1
M Z Y-1
µ-1
ε -1
Y (admisttance) µM-1
Z-1
Y
µε
Ε (permittivity) M-1
µ-1
Z-1
µ-1
Y ε
µ = jωCo where Co is the capacitance of the empty cell.
4.6.2 Equivalent circuits
EIS data is commonly analyzed by fitting it to an equivalent electric circuit which
is proposed according to the model that gives a reasonable explanation of observed
impedance of a system. The model fitting helps to identify different underlying processes
occurring in the system. For a good simulation, both the experimental and simulated
impedance spectra should be close to one another. At the same time, good fitting does not
ensure the model to be the best, the correctness of model is a must. The model should
provide the physical interpretation of the system.
Elements that are commonly used in equivalent circuits are listed in Table 4.2
along with their impedances and admittances. In the table, R is a resistor, C is a capacitor,
L is an inductor and W is a Warburg element which is used to describe semi-infinite
diffusion processes. Q is a constant phase element (CPE) whose value is adjusted by its
exponent α which can lie between -1 and 1 (-1≤ α ≤1). If α is zero, Q acts as a resistor
and for α = -1, it becomes an inductor. Q denotes a capacitor for α = 1. Finally, G is
CH-4 Characterization Techniques
65
Gerischer impedance describing impedance originating from a coupled electrochemical-
chemical process.
Table 4.2 Impedances and admittances of different circuit elements
Element Description Impedance Admittance
L Inductor j L 1 j L
R Resistor R 1 R
C Capacitor 1 j C j C
W Warburg oZ j oY Z
Q Constant phase element ( )oZ j ( )oY j
G Gerischer ( )oZ j k ( )oY j k
To understand nature of processes, knowledge of capacitance values is quite
helpful which can be used to identify different processes. Values of capacitances
associated with some processes are given in Table 4.3.
Table 4.3 Typical capacitance values and the corresponding phenomena [20]
Capacitance, F Interpretation of phenomenon
10-12
Bulk
10-11
Minor, secondary phase
10-11
-10-8
Grain boundary
10-10
-10-9
Bulk ferroelectric
10-9
-10-7
Surface layer
10-7
-10-5
Sample-electrode interface
10-4
Electrochemical reactions
CH-4 Characterization Techniques
66
Some typical equivalent circuits and their impedance spectra are given in Fig. 4.7.
The impedance data is simulated using the equivalent circuits and the values of
resistances and capacitances are extracted to have an insight into the system.
Fig. 4.7: Some typical equivalent circuits and the impedance in complex plane [21].
The impedance data is simulated using the equivalent circuits by different
softwares.
CH-4 Characterization Techniques
67
4.7 Electrical Conductivity Measurement
For the conductivity measurement, two methods are usually used which are
briefly discussed below.
4.7.1 Four probe measurement
In a four probe measurement, four parallel contacts are applied to the sample. The
basic system used is shown in Fig. 4.8. The current source is connected to outer probes
while a voltmeter is attached to the inner probes of the sample. The current source applies
a fixed known current to the sample. The current then travels through the sample. The
voltmeter measures the voltage drop (V) between the internal terminals. Resistance is
measured by well known ohm’s law;
VR
I (4.11)
Fig 4.8: Four probe set up for conductivity measurement [22].
The advantage of the four terminal method is that there is no current flowing
through the voltage sensor wires, so there is no IR drop and this resistance will not
influence the conductivity measurement.
CH-4 Characterization Techniques
68
From Eq. 4.11, resistivity can be computed as;
AR
l (4.12)
where A is the cross-sectional area of the specimen and l is the separation of the two inner
probes.
4.7.2 van der Pauw set up
The van der Pauw technique [23] measures the resistivity of thin samples having
arbitrary shape. This set up is applicable if the thickness of sample is known, the sample
is uniform and contacts are small and at the periphery of the sample.
In this setup, two resistance measurements are made on four contacts as shown in
Fig. 4.9. In the first measurement, current is passed to contacts 1 and 2 and voltage is
measured between contacts 3 and 4. The resistance (R1) is calculated using equation
4.11. In the next measurement, current is applied to contacts 1 and 4 and voltage is
measured between contacts 2 and 3 followed by another resistance calculation (R2).
From the average value of resistance, resistivity is calculated using Eq. 4.13.
ln 2
dR
(4.13)
where R is the average resistance calculated from these two measurements and d is the
thickness of the sample.
Fig. 4.9: Schematic of van der Pauw set up [24].
CH-4 Characterization Techniques
69
4.8 Infrared Spectroscopy
Infrared spectroscopy is one of the most common and applied techniques to
identify the chemical bonds in a molecule. It also verifies the structure of a molecule as
each functional group has a characteristic peak in the IR region, which ranges from 400-
4000 cm-1
[25].
IR radiations do not have enough energy to induce electronic transitions as in the
case of UV radiations. The frequencies that are absorbed are closely related to the
structure of the molecules, i.e., atom species, bonding types and ways of possible
vibration (stretching, scissoring, rocking and twisting).
For a typical IR spectrum, studied samples are exposed to a beam of infrared light
and transmitted light is collected which reveals the absorption of the samples. From the
characteristic absorption frequencies, different functional groups can be quickly
identified. In an FTIR spectrum, usually broad peaks are observed as the absorption of IR
radiations depends upon conjugation and proximity effects.
CH-4 Characterization Techniques
70
REFERENCES
1. P. Slade, J. Lloyds in Thermal characterization techniques, M. Dekker, New
York, 1970.
2. D.W.L. Hukins in X-ray Diffraction by Disordered and Ordered System,
Pergamon Press, Oxford, 1981.
3. C. Hammond in The Basic of Crystallography and diffraction, Oxford Science
Publications, 2001.
4. H.H. Willard, J.L. Merritt, J. Dean in Instrumental Methods of Analysis, D. Van
Nostrand Company, New Jersey, 1965.
5. W.F. Smith, Principles of Materials Science and Engineering, McGraw-Hill.
1998.
6. R.E Dinnebier, S.J.L. Billinge in Powder diffraction: Theory and Practice, The
Royal Society of Chemistry, 2008.
7. L.E. Smart, E.A. Moore in Solid State Chemistry, An introduction, Taylor &
Francis Group, 2005.
8. A.R. West in Basic Solid State Chemistry, John Wiley & Sons, 1999.
9. I.M. Watt in The Principles and practice of electron microscopy, Cambridge
University Press, 1997.
10. P.W. Hawkes, J.C.H. Spence in Science of Microscopy, Springer, 2007.
11. P.J. Goodhew, F.J. Humphreys in Electron Microscopy and Analysis. Taylor &
Francis, 1998.
12. http://www.purdue.edu/rem/rs/sem.htm
13. J.P.M. Syvitsk in Principles, Methods and Application of Particle Size Analysis,
Cambridge University Press, 1997.
14. T. Allen in Particle Size Measurement; 4th Ed., Chapman & Hall, 1992.
15. B. Wunderlich in Thermal Analysis of Polymeric Materials, Springer, 2005.
16. Dilatometry, Methods, Instruments, Applications, NETZHCH.
17. E. Barsoukov, J.R. Macdonald in Impedance Spectroscopy Theory, Experiment,
and Applications, 2nd
Ed, John Wiley & Sons, 2005.
18. S.M. Park, J.S. Yoo, Anal Chem., 2003, 1, 455A – 461A.
CH-4 Characterization Techniques
71
19. B.Y Chang, S.M Park, Annu. Rev. Anal. Chem., 2010, 3, 207 – 229.
20. J.T.S. Irvine, D.C. Sinclair, A.R. West, Adv. Mater., 1990, 2, 132 – 138.
21. V.F. Lvovich in Imprdance Spectroscopy, Applications to Electrochemical and
Dielectric Phenomenon, John Wiley & Sons, 2012.
22. http://www.imagesco.com/articles/superconductors/four-pt-schematic.gif
23. L.J. van der Pauw, Philips Res. Repts., 1958, 13, 1 – 9.
24. H. Czichos, T. Saito, L. Smith in Materials measurement methods, Springer,
2006.
25. A.D. Cross in An introduction to practical infra-red spectroscopy, Butterworth &
Co Ltd. 1964.
Chapter 5
Synthesis and Characterization of LSCTA-
72
Synthesis and Characterization of LSCTA-
Abstract
In the present chapter, the characterization results of Pechini method derived A-
site deficient calcium doped lanthanum strontium titanate (LSCTA-) powder are
presented. LSCTA- was synthesized by the Pechini Method. The effect of calcination
temperature on phase, microstructure and shrinkage characteristics was investigated and a
calcination temperature of 1000 oC was optimized in terms of close match of sintering
behaviour to that of yttria-stabilized zirconia (the electrolyte) for further studies. The
initial results have demonstrated LSCTA- to be a suitable anode candidate.
5.1 Introduction
In quest of alternate anode materials, A-site deficient titanates have gained major
attention because they show good electrical conductivity, enhanced sintering, thermal
stability and good performance as SOFC anodes [1-6]. For the present research project, a
major direction was taken from work of Ahmed where he focused on A-site deficient,
lanthanum strontium titanate, La0.2Sr0.7TiO3 [7]. In that study, A-site deficient lanthanum
strontium titanate was doped with Ca2+
from an x value of 0.1 to 0.7 (La0.2Sr0.7-xCaxTiO3)
as Ca+2
has good solubility in SrTiO3 besides its size compatibility with A site. The
maximum value of conductivity was achieved at a dopant level of x = 0.45 followed by a
drop of conductivity upon further doping. The composition with maximum conductivity
is the focus of the present research. The synthetic route adopted earlier was a solid state
which is a high temperature method with many firing stages. Moreover, the solid state
route (SSR) results in the formation of bigger particles [8]. In recent years, solution phase
methods like sol-gel, combustion, Pechini and co-precipitation have replaced solid state
routes [9-11]. These solution based methods are relatively simple, cost effective and
facilitate the reaction at lower temperatures and result in homogeneous and fine particles.
Among wet chemical methods, polymeric precursor-based Pechini method is an
alternative to the conventional sol gel method [12] to produce homogeneous powders
Chp-5 Synthesis and characterization of LSCT
73
using an aqueous medium and commonly available metal salts at much lower synthesis
and treatment temperatures [13, 14].
In Pechini’s method, an alpha-hydroxycarboxylic acid such as citric acid is used
to chelate various cation precursors forming a polybasic acid. In the presence of a
polyhydroxy alcohol, such as ethylene glycol, these chelates polymerize with the alcohol.
Better polymerization leads to homogeneous distribution of the metallic ions in gels.
Constant heating results in polyesterification yielding a homogeneous sol having metal
ions uniformly distributed throughout the organic matrix. Further heat treatment results in
formation of a solid resin. Finally, the polyester is decomposed to eliminate the excess of
organic material and the dried resin is calcined to form the desired stoichiometric phase
with high chemical and structural homogeneity [15].
In the present work, a solution phase Pechini method was applied to synthesize A-
site deficient, Ca2+
doped composition, La0.2Sr0.25Ca0.45TiO3, hereafter called as LSCTA-
[16]. The synthesized LSCTA- powder was characterized by XRD, TGA, SEM,
dilatometry, ac impedance and dc conductivity and effect of calcination temperature was
studied.
5.2 Experimental
5.2.1 Sample preparation
The sample preparation involved following steps.
5.2.1.1 Pechini synthesis
A modified Pechini method was adopted to synthesize A-site deficient calcium
doped lanthanum strontium titanate, LSCTA-. Metal nitrate salts were used as precursors
because metal nitrates have more favorable decomposition kinetics compared to the
carbonates, acetates and chlorides bases [17, 18]. An aqueous solution containing
stoichiometric amounts of lanthanum nitrate (Aldrich, 99.9%), strontium nitrate (Aldrich,
>99%), calcium nitrate (Aldrich, 99%) and titanium(IV)-bis-(ammoniumlactato)
dihydroxide, 50% w/w in water (Aldrich, 99% ) was mixed with a solution of ethylene
glycol and citric acid to have final molar ratio of metal ions to citric acid to ethylene
glycol, 1:4:16. The beaker containing this mixture was then placed on a hot plate and the
temperature of the solution was raised to 80 °C. Heat treatment increased the viscosity of
Chp-5 Synthesis and characterization of LSCT
74
the solution without any visible phase separation. The resulted gel was heat treated at 300
°C, leading to the expansion to three times of its original volume. The gel was dried and
the residue was calcined in air (for 5 hours) at various temperatures. (LSCTA- calcined at
900 oC, 950
oC, 1000
oC and 1100
oC are abbreviated as S1, S2, S3 and S4 respectively in
subsequent sections).
5.2.1.2 Firing
Raw powders are usually fired at certain temperatures with controlled procedures
to produce desired phase and microstructures. The thermal changes occurring during
firing depend on the temperature programs and atmospheres. Two terms are used to refer
the thermal treatment, i.e., calcination and sintering.
5.2.1.2.1 Calcination
Calcination is a thermal treatment applied to raw materials resulting in thermal
decomposition, phase transition or removal of a volatile fraction to give physical and
chemical stability. Calcination is normally done at temperatures below the melting point
of the product materials.
For calcination, the powders were introduced in the muffle furnace at room
temperature. Then the furnace was programmed for temperature increase by 1 °C min-1
to
allow slow decomposition of organics present in the dried resin till 500 °C with a dwell
of 30 minutes. Further, the temperature was increased to 1000 °C with dwell of 5 hours
followed by cooling down to room temperature at the rate of 5 °C min-1
. An X-ray
diffraction was performed to confirm the phase purity.
5.2.1.2.2 Sintering
Sintering is a high temperature thermal treatment in which a compact powder is
heated usually to 0.5~0.9 times of its melting temperature in Kelvin to densify it.
Densification is achieved by atomic diffusion at the sintering temperature which leads to
particle necking and thus reduction in porosity. The process leads to an increase in
strength but reduces the surface energy of the system due to particle grain growth.
In the present project, sintering of LSCTA- pellets was done at 1400 °C for 6
hours. Powders were uniaxially pressed into pellets using cylindrical steel dies of
typically 13 mm diameter. These sintered pellets were used in dc conductivity
measurements and ac impedance studies.
Chp-5 Synthesis and characterization of LSCT
75
5.2.1.3 Pre-reduction
The n-type conductors show good conductivity when pre-reduced at high
temperatures in reducing environment [19]. For the present case, the high temperature
pre-reduction step was performed using a Carbolite STF 15/180 tube furnace to heat the
dense pellets to 1050 °C for 72 hours in dry 5% H2/Ar at a ramp speed of 3 oC min
-1.
Following the reduction process, the pellets were cleaned and tested for XRD and
conductivity.
5.2.2 Sample characterization
Thermogravimetric TGA was performed on a Netzch STA 449c equipped with
ProteusTM
thermal analysis software in air at a heating rate of 3 oC min
-1. The phase
formation was studied using Philips XRD diffractometer using Cu-Kα1 radiation in the
range of 20o to 80
o. Lattice parameters were fitted with STOE WinXPOW software.
Particle size analysis was carried out on a Malvern Instruments Mastersizer 2000. For
particle size analysis, the LSCTA- powder was dispersed in 2 wt% of triton in isopropyl
alcohol. BET (Brunauer, Emmett and Teller) measurements were taken on a
Micromeritics TriStar II 3020 instrument. The morphology of the calcined powders was
studied using JEOL 6700F field emission microscope. Sinterability of LSCTA- powder
was investigated using Netzch DIL 402C instrument. For dilatometry, powder was
pressed into pellets of 13 mm diameter under pressure of 1 ton and sintering behavior
was investigated in air upto 1400 K using ramp rate of 2 oC min
-1. For ac impedance,
LSCTA- pellets were sintered in air at 1400 oC for 6 hours and the surface of sintered
pellets was polished and coated with Pt paste which was then consolidated at 900 oC for
one hour. Impedance data were taken using a Solartron 1260 impedance/gain phase
analyzer in the frequency range of 1 Hz to 13 MHz. Dc conductivity was measured by
van der Pauw method [20] on LSCTA- pellets. For van der Pauw set up, four gold mesh
contacts were attached at the edges of the sample using gold paste. The contacts were
consolidated by firing at 900 oC for one hour. The density of sintered pellets was
calculated by measuring their mass and dimensions and compared to the theoretical
density computed using the unit cell parameters.
Chp-5 Synthesis and characterization of LSCT
76
5.3 Results and Discussion
5.3.1 Thermal gravimetric analysis
Fig. 5.1a shows thermo-gravimetric analysis of LSCTA- resin in air.
0 200 400 600 800 100020
40
60
80
100
0
500
1000
1500
2000
DT
A (
mW
/mg
)
% M
ass l
oss
Temperature (oC)
a
0 200 400 600 800 1000
96
99
102
105
%
Ma
ss
lo
ss
Temperature (oC)
b
Fig. 5.1: a) TGA (solid line) and DTA (dotted line) curves of LSCTA- resin in air and b)
TGA of sample after calcination at 1000 °C.
It can be noted that major mass loss (~65%) occurs in the range of 280 oC to 410
oC and is attributed to burnout of organic components, consistent with the formation of
the oxide. Accordingly, a strong exothermic peak was observed in the DTA indicative of
combustion of organic components. After ~410 oC, the removal of all organic matter is
complete and further heat treatment does not cause any mass loss. No mass modification
is observed on cooling down to room temperature. However, the calcined sample showed
stability in air upon heating (Fig. 5.1b). The stability of the sample was tested in air after
Chp-5 Synthesis and characterization of LSCT
77
calcination. It can be seen that no weight loss is observed after calcining the sample.
Similar thermogravimetric behavior was observed for all samples calcined at different
temperatures.
5.3.2 X-ray diffraction
X-ray diffraction is a useful technique to study the phase purity of samples. In
case of material characterization, it can be considered as the first step towards phase
identification and structure determination.
5.3.2.1 Comparison of solid and solution method
Fig. 5.2 displays comparison of XRD patterns of LSCTA- synthesized by solution
phase Pechini Method (a) and the reported solid state method (b).
20 40 60 80
0
20
40
60
80
100
120
Diffraction Angle (2)
Re
lati
ve
In
ten
sit
y
(332)
(420)(022)
(121)
(242)(400)
(042)(040)
(200)
a
b
Fig. 5.2: XRD patterns of LSCTA- synthesized via; a) solution phase Pechini
method and b) solid state route [7].
Both the XRD patterns match closely to one another showing the same phase
evolution using the solid state and solution phase synthetic routes.
5.3.2.2 XRD pattern of reduced LSCTA-
LSCTA- is a supposed anode candidate, thus XRD pattern was taken after its pre-
reduction at 1050 oC for 72 hours in 5% H2/Ar to check the structural integrity (section
Chp-5 Synthesis and characterization of LSCT
78
5.2.1.4). Comparison of XRD of calcined and pre-reduced sample is shown in Fig. 5.3. It
was observed that on reduction, LSCTA- retained its perovskite structure and no extra
peaks were detected showing no phase separation although an expansion in unit cell was
noted. The slight increase in lattice parameters for reduced sample is due to reduction of
Ti+4
to Ti+3
[21]. The small change in lattice parameters upon reduction suggests that the
structural integrity of this suggested anode candidate is tolerant to redox cycles or
variation of oxygen partial pressure during fuel cell operation.
20 40 60 80
0
20
40
60
80
100
120
a
b
Re
lati
ve
In
ten
sit
y
Diffraction Angle (2)
(332)
(420)
(022)
(121)
(242)(400)
(042)
(040)
(200)
Fig. 5.3: XRD patterns of LSCTA-; a) before reduction and b) after reduction.
5.2.2.3 Compatibility with yttria-stabilized zirconia (YSZ)
One of the basic requirements for SOFC components is chemical inertness and
stability to avoid undesired reactions. For anode candidates, there should be no reaction
between the electrolyte and electrode because in the final fabrication, the anode comes in
contact with the electrolyte. The reaction between electrolyte and anode is detrimental for
anode performance. To check the chemical stability of LSCTA- towards YSZ, LSCTA-
and 8 mol% YSZ (YSZ) were mixed in 1:1 ratio and fired at 1400 oC for 2 hours. After
this treatment, the XRD pattern was taken which is shown in Fig. 5.4.
Chp-5 Synthesis and characterization of LSCT
79
No extra peak was observed showing chemical inertness and stability of LSCTA-
towards YSZ which implies that a physical mixture of LSCTA- and YSZ could retain its
integrity after firing at 1400 oC.
0 20 40 60 80 100
0
20
40
60
80
100
120
*
*
**
*
*
**
* LSCTA-
b
cRel
ativ
e In
ten
sity
Diffraction Angle (2)
a
YSZ*
Fig. 5.4: XRD patterns after firing at 1400 oC for; a) LSCTA- , b) pure YSZ and c) 1:1
mixture of LSCTA-:YSZ.
5.2.2.4 XRD’s of LSCT calcined at different temperatures
XRD patterns of LSCTA- samples that had been calcined at various temperatures
in air, S1-S4 are shown in Fig. 5.5. All the samples show single perovskite structure and
no impurity peak was detected in any of the XRD patterns. The XRD pattern was indexed
and cell parameters were refined using WinXPOW software. All the peaks were indexed
in orthorhombic symmetry with space group of Pbnm. The values of lattice parameters; a,
b and c were found to be 5.4661(7) Å, 5.4638(6) Å and 7.7343(6) Å, respectively for S3.
This follows a relation close to √2ap x √2ap x 2ap where ap is the unit cell parameter of the
ideal cubic symmetry.
It has been reported that A-site deficient La0.2Sr0.7TiO3 (LSTA-) has cubic
symmetry [22] while CaTiO3 exhibits orthorhombic symmetry at room temperature [23].
Detailed discussion about origin of different symmetries with calcium doping in LSTA-
Chp-5 Synthesis and characterization of LSCT
80
system could be found elsewhere [7]. Ca doping to lanthanum strontium titanate has been
shown to decrease the symmetry from cubic through tetragonal to orthorhombic. Thus the
symmetry of A-site deficient lanthanum strontium titanate changes from cubic to
orthorhombic with calcium doping, as the replacement of large Sr2+
(1.44 Å) with smaller
size Ca2+
(1.35Å) is likely to increase the distortion of the perovskite structure by
decreasing the tolerance factor from 0.907 for La0.2Sr0.7TiO3 to 0.891 for
La0.2Sr0.25Ca0.45TiO3 (LSCTA-).
.
20 40 60 80 100
0
25
50
75
100
125
150
c
b
a
d
Rela
tive I
nte
nsit
y
Diffraction angle (2)
Fig. 5.5: X-ray diffraction patterns of LSCTA- calcined in air at various temperatures; a)
900 oC (S1), b) 950
oC (S2), c) 1000
oC (S3) and d) 1100
oC (S4).
The crystallite size was calculated by the Scherrer equation using the peak at 2θ =
32.6o.
cos
kD
(5.1)
where λ is the incident X-ray wave length in angstroms, β is the full width half maximum
of the peak (in radians) at diffraction angle θ and k is shape factor. Average crystallite
size increased with increasing calcination temperature where D value almost doubled for
S4 sample than S1 i.e., from 0.033 µm to 0.06 µm, (Table 5.1).
Chp-5 Synthesis and characterization of LSCT
81
5.3.3 Particle size analysis and BET area
It is likely that the calcination process produces agglomerates of primary
particles, thus the particle size was determined after ultrasonicating the powders in a
mixture of isopropanol and dispersant (triton) for 15 minutes. The results of particle size
analysis are presented in Fig. 5.6 and confirmed the observed increase of mean crystallite
size with the calcination temperature.
The figure shows that a narrower distribution is observed in the case of lower
calcination temperature. The distribution of primary particles broadens with
corresponding decrease of volume fraction as calcination temperature is increased. The
shoulder observed for S4 is attributed to large agglomerates of LSCTA- or bubbles due to
in-situ ultrasonication in the apparatus.
0.01 0.1 1 10 100 1000
0
2
4
6
8
10
Vo
lum
e%
Particle diameter/m
a
b
c
d
Fig. 5.6: Particle size distribution of LSCTA- calcined at various temperatures; a) 900 oC
(S1), b) 950 oC (S2), c) 1000
oC (S3) and d) 1100
oC (S4).
Table 5.1 shows variation of mean particle size, primary crystallite size and
average BET area with calcination temperature.
Chp-5 Synthesis and characterization of LSCT
82
Table 5.1 Crystallite size, mean particle size and BET area of LSCTA- samples
Samples *d (0.5)
µm
Crystallite Size
µm
BET area (± 0.15)
m2
g-1
S1 3.16 0.033 11.55
S2 3.56 0.047 8.34
S3 3.79 0.053 3.90
S4 6.94 0.060 1.53
*The d(0.5) is the average particle diameter where 50% particles of the distribution have
size below this value.
All of these parameters follow the general trend; the higher the calcination
temperature, the larger the mean diameter and crystallite size and lower the BET area.
BET area decreases with increasing calcination temperature due to inverse relationship
between BET area and particle size.
5.3.4 Scanning electron microscopy
The morphology of the samples was studied using scanning electron microscope
to look into the microstructure and grain sizes.
5.3.4.1 Calcined samples
The effect of calcination temperature on particle growth and nucleation was
investigated by looking at the microstructure for these S1-S4 LSCTA- samples (Fig. 5.7)
where particle size enlarged with increase in the calcination temperature.
From SEM micrographs, it can be noted that the sample S1 consists of particles
having an average grain size between 100-150 nm whereas S4 shows considerably larger
particles. One can observe that while S1 comprises rather isolated particles, the powder
calcined above 1000 oC exhibits particle necking and formation of clusters of larger
submicron sizes. It is expected that the fine particles produced from the solution method
to have high sinterability which may be exploited favorably in processing the SOFC
electrodes.
Chp-5 Synthesis and characterization of LSCT
83
Fig. 5.7: Micrographs of LSCTA- powder after calcination at various temperatures; a)
900oC (S1), b) 950
oC (S2), c) 1000
oC (S3) and d) 1100
oC (S4).
5.3.4.2 Sintered samples
The difference in microstructure affects the sintering process where smaller size is
beneficial to densification and results in higher density after sintering. This can be
manifested by looking at the micrographs of sintered samples that were calcined at
different temperatures.
In Fig. 5.8, we see well defined grains with limited porosity after sintering at
1400°C. It was observed that maximum densification occurred in the case of pellets
derived from S1. From the determination of mass and dimensions of LSCTA- pellets
sintered in air at 1400 oC, relative density was calculated which is given in Table 5.3.
Relative density was found to decrease with initial calcination temperature. Smaller
particle size helps in achieving more densification, thus S1 acquired higher density as
compared to the others.
Chp-5 Synthesis and characterization of LSCT
84
Fig. 5.8: SEM micrographs showing effect of sintering at 1400 °C on LSCTA- powders
calcined at temperatures; a) 900 oC (S1), b) 950
oC (S2), c) 1000
oC (S3) and d) 1100
oC
(S4).
5.3.5 Dilatometric analysis of LSCTA- samples
One of the basic requirements of an effective anode material is the thermal
compatibility with the other cell components, especially the electrolyte. In terms of
sinterability, there should be a match between these two SOFC components not only in
shrinkage extent, but also in the onset sintering temperature. Fig. 5.9 shows sintering
behaviour of pellets made of S1-S4 powders, in air up to 1400 °C as compared to the
usual choice of electrolyte, 8 mol% YSZ.
b
Chp-5 Synthesis and characterization of LSCT
85
0 300 600 900 1200 1500
-24
-18
-12
-6
0
300
600
900
1200
1500
1800
Te
mp
era
ture
(oC
)
e
dc,
b
dL
/Lo%
Time/min
a
Fig. 5.9: Dilatometric sintering curves of pellets from LSCTA- powder calcined at various
temperatures in air; a) 900 oC (S1), b) 950
oC (S2), c) 1000
oC (S3), d) 8-YSZ and e)
1100 oC (S4).
The shrinkage is directly related to particle size of the powder e.g., smaller size
leads to more sinterability. It is obvious that S1 sinters much more and the sintering starts
earlier in comparison to YSZ. The sinterability decreases as calcination temperature for
the initial powder is increased. Shrinkage percentages are given in Table 5.2.
Table 5.2 Shrinkage percentages and relative density values for LSCTA- samples
Samples Shrinkage % *Relative Density %
S1 27.65 92.9
S2 24.84 91.7
S3 21.46 86.0
S4 19.69 84.6
YSZ 21.03 ~100
* using theoretical density of 4.70 g cm-3
Chp-5 Synthesis and characterization of LSCT
86
S3 shows an interesting behaviour. Its shrinkage matches very well with
electrolyte, YSZ, in terms of extent and onset. This feature makes it suitable for a co-
sintering process together with YSZ, a typical electrolyte choice in SOFC manufacturing.
Based on this promising result, the calcination temperature of 1000 oC was selected for
further studies.
5.3.6 Ac Impedance
The electrical properties of sintered LSCTA- pellets (S1-S4) with Pt electrodes
were investigated by ac impedance in air at various temperatures in the frequency range
of 1 Hz to 13 MHz. The measured impedance data were analyzed by Z view TM
program.
The results are discussed below;
5.3.6.1 Impedance in air
The Cole-Cole plots of S1 to S4 samples are given in Fig. 5.10. One well defined
arc starting from the origin could be seen in all of the plots for the temperatures
mentioned. The semicircular arc of the impedance spectrum can be expressed as an
equivalent circuit consisting of a parallel RC circuit for each sample.
From the impedance plots, the values of resistances and capacitances were
extracted by fitting and modeling the experimental data. It is observed that increasing the
temperature results in decreased resistance which is indicative of a negative temperature
coefficient of resistance (NTCR) behavior, as expected for an electronic semiconductor.
The corresponding capacitance values fall roughly in the range of ~10-12
F cm-1
which are
typically attributed to the contribution of intragranular or bulk phases [24]. This indicates
a relatively homogeneous semiconducting material; especially since the resistance
decreased significantly and consistently with increasing temperature.
The same electrochemical behavior was observed in all of the investigated air
sintered pellets where samples became less resistive with temperature.
Chp-5 Synthesis and characterization of LSCT
87
0 3 6 9 12 15 180
-2
-4
-6
-8
-10
a 700oC
750oC
800oC
Z''(k
cm
2)
Z'(kcm2)
0 4 8 12 16 200
-2
-4
-6
-8
-10b 700
oC
750oC
800oC
Z''(k
cm
2)
Z'(kcm2)
0 1 2 3 4 50
-1
-2
-3
c 700
oC
750oC
800oC
Z''(k
cm
2)
Z'(kcm2)
0 2 4 6 8 100
-1
-2
-3
-4
-5
-6
-7
700oC
750oC
800oC
Z'(
k
cm
2)
Z'(kcm2)
d
Fig. 5.10: Cole Cole plots of air sintered samples in frequency range of 1 Hz to 13 MHz
at different temperatures; a) S1, b) S2, c) S3 and d) S4.
Dependence of the imaginary part of the impedance on frequency at different
temperatures is shown in Fig. 5.11 for the samples studied. Distinct peaks appear in the
impedance spectrum where increase in temperature resulted in asymmetric broadening
and a decrease in Z″ magnitude due to a loss in resistive property of the sample with rise
in temperature. The graphs also show the shifting of Z″max with increase in temperature.
At high frequencies, all the graphs show similar behavior irrespective of temperature [25,
26].
Such results have been attributed to existence of temperature dependent relaxation
with a spread of relaxation times in the material. The relaxation species may possibly be
Chp-5 Synthesis and characterization of LSCT
88
immobile species/electrons at low temperature and defects/vacancies at higher
temperatures [27, 28].
This effect is obvious for all the samples where the imaginary part of impedance
decreases prominently with rise in temperature (Fig. 5.11).
3 4 5 6 70
-2
-4
-6
-8a
Z''(k
cm
2)
log f
700oC
750oC
800oC
850oC
3 4 5 6 70
-2
-4
-6
-8 700
oC
750oC
800oC
850oC
Z''(
k
cm
2)
log f
b
3 4 5 6 70.0
-0.5
-1.0
-1.5
-2.0
-2.5
c
Z''(
k
cm
2)
log f
700oC
750oC
800oC
850oC
3 4 5 6 70
-1
-2
-3
-4
-5
Z''(
k
cm
2)
log f
700oC
750oC
800oC
850oC
d
Fig. 5.11: Dependence of imaginary part of impedance on frequency for air sintered
samples (S1 to S4) in frequency range of 1 Hz to 13 MHz in air at different
temperatures; a) S1, b) S2, c) S3 and d) S4.
The dependence of real part of Z as a function of frequency at different
temperatures is shown in Fig. 5.12. The impedance-frequency trends merge in the high
frequency region irrespective of temperature. This may be due to the release of space
Chp-5 Synthesis and characterization of LSCT
89
charges as a result of reduction in the barrier properties of the material with rise in
temperature and may be the responsible factor for the enhancement of ac conductivity of
the material with temperature at higher frequencies [29, 30].
0 2 4 6 8
0
4
8
12
16 700
oC
750oC
800oC
850oC
Z'(
k
cm
2)
log f
a
0 2 4 6 80
4
8
12
16
20
Z'(
k
cm
2)
log f
700oC
750oC
800oC
850oC
b
3 4 5 6 70
5
10
15
20c
Z'(
k
cm
2)
log f
700oC
750oC
800oC
850oC
0 2 4 6 8
0
2
4
6
8
10
Z'(
k
cm
2)
log f
700oC
750oC
800oC
850oC
d
Fig. 5.12: Dependence of real part of impedance on frequency for air sintered samples
(S1 to S4) in frequency range of 1 Hz to 13 MHz in air at different
temperatures; a) S1, b) S2, c) S3 and d) S4.
5.3.6.2 Comparison of ac conductivity
Ac conductivity of the samples was calculated from the values of bulk (total)
resistance using the relation,
Chp-5 Synthesis and characterization of LSCT
90
1 l
R A
(5.2)
where l is the thickness and A is the area in cm2 of the pellet. The slope of the plot of ln
σT vs. 1/T shown in Fig. 5.13 gives the activation energy using Arrhenius equation:
ln ' aET A
RT (5.3)
where 'A is the temperature dependent frequency factor and Ea is the activation energy.
0.9 1.0 1.1 1.2 1.3 1.4
-6.0
-4.5
-3.0
-1.5
0.0
1.5
3.0 S1
S2
S3
S4
ln
T(
Sc
m-1
K)
1/T x 103
(K-1
)
Fig. 5.13: Arrhenius dependence of conductivity calculated from ac impedance for
LSCTA- samples calcined at various temperatures.
Variation in ac conductivity with temperature in air points that all samples show
Arrhenius type behavior i.e., the conductivity increases linearly with temperature. The
differences are not very significant and would clearly relate to differences in
microstructure and thermal history. The activation energy values estimated from slope of
Arrhenius conductivity plots are presented in Table 5.3.
Analysis of the table shows that S3 offered least energy of activation so it is
expected to have a better conductivity profile. The same sample also showed close
Chp-5 Synthesis and characterization of LSCT
91
sintering match to yttria-stabilized zirconia (YSZ), therefore this sample was further
tested for dc conductivity.
Table 5.3 Activation energy, Ea calculated from ac impedance
Samples Ea (eV)
S1 1.478 ± 0.013
S2 1.423 ± 0.014
S3 1.280 ± 0.011
S4 1.404 ± 0.020
5.3.7 Dc conductivity
The van der Pauw set up was used to measure dc conductivity of S3. The
conductivity was monitored under different conditions;
a) In air
b) After insitu reduction at 880 oC in reducing atmosphere
c) For pre-reduced sample in 5% H2/Ar
d) For sample sintered in 5% H2/Ar.
where each case is briefly discussed below;
5.3.7.1 Dc conductivity measurement in air
The dc conductivity of a dense LSCTA- pellet (88% of theoretical value) sintered
in air at 1400 oC was found to increase with temperature as measured in air indicating
semiconducting behavior (Fig. 5.14). The sample attained conductivity value of 1.24 mS
cm-1
in air at 880 oC.
Chp-5 Synthesis and characterization of LSCT
92
200 400 600 800 1000 1200
0.0
0.4
0.8
1.2
1.6
mS
cm
-1
Temperature (K)
Fig. 5.14: Temperature dependence of conductivity of LSCTA- (S3) sintered pellet in air.
5.3.7.2 Dc conductivity measurement in reducing atmosphere
After taking conductivity in air (oxidizing environment) at 880 oC, 5% H2/Ar was
purged in the system to create reducing conditions. The initial conductivity value in air
(1.24 mS cm-1
) increased by three orders of magnitude to 1.30 S cm-1
upon 24 hours in-
situ reduction at 880 oC due to reduction of Ti
4+ to Ti
3+ [19].
The time dependence of conductivity graph is shown in Fig. 5.15. After an initial
delay, the reduction proceeds in two stages, rapidly in the first two hours, followed by a
much slower subsequent increase. It takes more than 18 hours for less than 10% increase
in conductivity after initial span of two hours. These two stages might be related to the
fast removal of oxygen from the surface of the perovskite that is followed by a slow
diffusion into the bulk of the micron size grains [21].
Chp-5 Synthesis and characterization of LSCT
93
0 5 10 15 200.0
0.3
0.6
0.9
1.2
1.5
S
cm
Time (hrs)
Fig. 5.15: Conductivity profile of in-situ reduced LSCTA- (S3) pellet in 5% H2/Ar
at 880 °C.
5.3.7.3 Dc conductivity measurement for pre-reduced sample
It has been reported that n-type SrTiO3-based materials show good conductivity
when pre-reduced or sintered in a reducing atmosphere [31-32]. For pre-reduction dense
LSCTA- sample from powder calcined at 1000 o
C (S3) and sintered in air at 1400 oC was
reduced in 5% H2/Ar at 1050 oC for 72 hours. Fig. 5.16 shows the temperature
dependence of the electrical conductivity of pre-reduced LSCTA- in 5% H2/Ar on heating
and subsequent cooling.
200 400 600 800 1000 1200
40
60
80
100
120
(
Sc
m-1)
Temperature(K)
cooling
heating
Fig. 5.16: Conductivity profile of pre-reduced LSCTA- (S3) during thermocycling in 5%
H2/Ar.
Chp-5 Synthesis and characterization of LSCT
94
It can be seen that on cooling in reducing atmosphere, the conductivity increases
showing positive temperature coefficient of resistance indicative of metallic behavior. It
could be attributed to electronic conduction as predominant conduction mechanism in the
pre-reduced sample. A metal insulator transition could also be observed [22] at ~ 350 K.
Below this temperature, conductivity decreases with decrease in temperature. At 880 oC,
conductivity of 38 S cm-1
was obtained at log pO2 = -16.65 atm. The value is comparable
to the one reported for the sample prepared by solid state synthesis (27.53 S cm-1
at 900
oC at pO2 = 10
-19 atm) [7] and would be sufficient for using this material as an anode
current collector backbone in anode supported SOFC configuration.
The conductivity values of LSCTA- samples processed under different conditions
at 880 oC is tabulated in Table 5.4.
Table 5.4 Conductivity value of LSCTA- pellets under different conditions
at 880 oC
Conditions σ (S cm-1
)
In air 1.24x10-3
After in-situ reduction 1.30
Pre reduction 38.0
5.3.7.4 Dc conductivity measurement for sintered sample in 5% H2/Ar
From the conductivity results obtained so far, it is expected to have good
conductivity for LSCTA- sample sintered in reducing atmosphere. Thus, LSCTA- pellet
was sintered in 5% H2/Ar at 1400 oC for 6 hours and conductivity was monitored by
thermocycling in same environment shown in Fig. 5.17. For this sample, a conductivity
value of 144 S cm-1
was measured at room temperature.
Chp-5 Synthesis and characterization of LSCT
95
200 400 600 800 1000 12000
50
100
150
200
250
S
cm
-1)
Temperature (K)
Fig. 5.17: Conductivity profile of LSCTA- (S3) sintered in 5% H2/Ar in reducing
atmosphere upon heating.
The micrographs of LSCTA- pellet sintered in reducing atmosphere are also shown
in Fig. 5.18.
Fig. 5.18: Micrographs of LSCTA- (S3) pellet sintered at 1400 °C in reducing atmosphere
of 5% H2/Ar under different magnifications.
Chp-5 Synthesis and characterization of LSCT
96
From Fig. 5.18, it can be inferred that the microstructure is a strong function of
the sintering atmosphere. The sample sintered in 5% H2/Ar has smaller grain size with
different morphology as compared to sample sintered in air (compare with Fig. 5.8). The
former sample shows more grain boundaries than the latter. Therefore, S3 pellet sintered
in 5% H2/Ar, could ultimately offer better conductivity in cell testing as compared to that
sintered in air.
5.4 Conclusions
The powder characterization results presented here confirmed that single phase
LSCTA- can be produced via a solution combustion method that can be easily scaled up
for larger quantities required for large anode supported SOFC production. Compared to
traditional techniques such as solid state synthesis, this preparation method offers
improved homogeneity, lower preparation temperatures and less preparation steps, saving
time, energy and minimizing powder contamination. A precursor powder calcination
temperature of 1000 oC seemed to be very promising in terms of further processing for
anode fabrication via tape casting and anode-electrolyte co-sintering or screen printing.
The selected powder also showed good conductivity under various reducing conditions
that could be exploited for their application as SOFC anodes/anode support. In
conclusion, structurally stable LSCTA- could be a good alternative to state of the art
SOFC anodes exhibiting good mechanical, morphological and electrical properties.
Chp-5 Synthesis and characterization of LSCT
97
REFERENCES
1. D. Burnat, A. Heel, L. Holzer, D. Kata, J. Lis, T. Graule, J. Power Sources, 2012,
201, 26 – 36.
2. Q. Ma, F. Tietz, A. Leonide, E. Ivers-Tiffée, Electrochem. Commun., 2010, 12,
1326 – 1328.
3. H. Zhao, F. Gao, X. Li, C. Zhang, Y. Zhao, Solid State Ionics, 2009, 180, 193 –
197.
4. X. Li, H. Zhao, X. Zhou, N. Xu, Z. Xie, N. Chen, Int. J. Hydrogen Energ., 2010,
35, 7913 – 7918.
5. C.D. Savaniu, J.T.S. Irvine, Solid State Ionics, 2011, 192, 491 – 493.
6. C.D. Savaniu, J.T.S. Irvine, ECS Trans., 2009, 25, 2213 – 2222.
7. A.D. Aljaberi, J.T.S. Irvine, J. Mater. Chem., A, 2013, 1, 5868 – 5874.
8. V. Fruth, M. Popa, J.C. Moreno, E. Tenea, M. Anastasescu, Processing Appl.
Ceram., 2010, 4, 167 – 182.
9. S.M. Selbach, M.A. Einarsrud, T. Tybell, T. Grande, J. Am. Ceram. Soc., 2007,
90, 3430 – 3434.
10. C.J .Brinker, G.W. Scherer in Sol-Gel Science: The Physics and Chemistry of Sol-
Gel Processing, Academic Press, San Diego, 1990.
11. E. Traversa, P. Nunziante, M. Sakamoto, Y. Sadaoka, R. Montanari, Mater. Res.
Bull., 1998, 33, 673 – 681.
12. M. Pechini, U.S. Patent no. 333069, 1967.
13. J.D.G. Fernandes, D.M.A. Melo, L.B. Zinner, C.M. Salustiano, Z.R. Silva, A.E.
Martinelli, M. Cerqueira, C. Alves Ju´nior, E. Longo, M.I.B. Bernardi, Mater.
Lett., 2002, 53, 122 – 125.
14. L.W. Tai, P.A. Lessing, J. Mater. Res., 1992, 7, 502 – 510.
15. A. Ries, A.Z. Simoes, M. Cilense, M.A. Zaghete, J.A. Varel, Mater. Charac.,
2003, 50, 217 – 221.
16. C.D. Savaniu, J.T.S. Irvine, J. Mater. Chem., 2009, 19, 8119 – 8128.
17. N. Osman, A.M. Jani, I.B. Talib, Ionics, 2006, 12, 379 – 384.
18. A.M. Azad, S. Subramaniam, Mater. Res. Bull., 2002, 37, 85 – 97.
Chp-5 Synthesis and characterization of LSCT
98
19. O.A. Marina, N.L. Canfield, J.W. Stevenson, Solid State Ionics, 2002, 149, 21–
28.
20. L.J. van der Pauw, Philips Res. Repts., 1958, 13, 1 – 9.
21. D. Neagu, J.T.S. Irvine, Chem. Mater., 2010, 22, 5042 – 5053.
22. P.R. Slater, D.P. Fagg, J.T.S. Irvine, J. Mater. Chem., 1997, 7, 2495 – 2498.
23. N. Lamrani, B. Itaalit, S. Marinel, M. Aliouat, Mater. Lett., 2011, 65, 346 – 349.
24. J.T.S. Irvine, D.C. Sinclair, A. R. West, Adv. Mater., 1990, 2, 132 – 138.
25. Z.G. Yi, Y.X. Li, Y. Wang, Q.R. Yin, J. Electrochem. Soc., 2006, 153, F100 –
F105.
26. K.Verma, S. Sharma, Phys. Status Solidi B., 2012, 249, 209 – 216.
27. R. Rizwana, T.R. Krishna, A.R. James, P. Sarah, Cryst. Res. Technol., 2007, 42,
699 – 706.
28. A.M.M. Farea, S. Kumar, K.M. Batoo, A. Yousef, Alimuddin, Physica B., 2008,
403, 684 – 701.
29. S. Sumi, P.P. Rao, M. Deepa, P. Koshy, J. App. Phys., 2010, 108, 063718 (1 – 9).
30. S. Brahma, R.N.P. Choudhary, A.K. Thakur, Physica B., 2005, 355, 188 – 201.
31. N.G. Eror, U. Balachandran, J. Am. Cer. Soc., 1982, 65, 426 – 431.
32. T.R.N. Kutty, S. Philip, Mater. Sci. Eng. B, 1995, 33, 58 – 66.
Chapter 6
Aqueous Tape Casting
99
Aqueous Tape Casting
Abstract
Tape casting is a low cost and well known technique used in the fabrication of
SOFC components. For uniform, homogeneous and crack-free green tapes, optimized
slurry formulation is essential. The present chapter gives an overview of LSCTA-
powder processing using aqueous tape casting for dense and porous tapes. Further, the
green tapes were laminated in bars and the conductivity of sintered bars was
determined. It was observed that impregnation resulted in significant improvement in
conductivity of porous bodies.
6.1 Introduction
Tape casting is a cost effective well known colloidal shaping technique for
thin ceramic components which are often employed in electronic applications [1-3]. It
has advantages over other methods such as pressing and extruding in terms of
production of large-area, thin and flat ceramic tapes [4] having a wide variety of
controlled morphologies, from highly porous to fully dense microstructures. It is an
attractive technique to produce solid oxide fuel cell components [5-9].
Both aqueous and non aqueous solvents can be used for processing of tapes.
As far as process control is concerned, a solvent-based tape casting method leads to
higher quality green tapes [10]. Usually solvent-based tape casting has been used to
prepare SOFC stacks [11-13]. However, the solvent-based tape casting technology has
some draw backs; the most important is the use of noxious solvents and perilous
additives which raises the production cost and also poses much harm to human health
and the environment. Therefore, this concern has stimulated the interest in water-
based tape casting processes [14-17]. Apart from the tape quality produced, an
aqueous-based tape casting method is not only environment friendly and less health
hazardous but is also inexpensive [18].
The tape casting process involves consecutive steps: the first one and probably
the most important is the fabrication of the slurry with the powder under
consideration. It is to be noted that final microstructure of the sintered tape depends
CH-6 Aqueous Tape Casting
100
on the initial powder, so the powder is sometimes pre-treated to have the appropriate
phase content, grain size distribution and morphology [19].
The slurry formulation is a crucial step for tape-casting as the arrangement,
dispersion and homogeneity of the starting ceramic particles in the slurry affects the
sintering behaviour of green tapes, and hence their final microstructure. A good slurry
is characterized by a stable dispersion i.e., it should have well dispersed particles with
no agglomeration and sedimentation. Usually, the dispersion of particles and thus the
stability of the slurry is promoted by thermal agitation and electrostatic and steric
repulsive forces. The second important slurry characteristic is the good slip rheology,
which controls the casting of tape. In terms of viscosity, it should be low to allow an
easy casting and high enough for the green tape to have a sufficient creep resistance to
maintain its geometry [20-23].
For slurry formulation, organic and/or inorganic additives (anti-foaming,
dispersing agent), the binder and one or more plasticizers are added to the powder to
form the slurry having proper slip rheology which increases the mechanical strength
and flexibility of green tape. The slurry is then mixed and ground mostly by ball-
milling to ensure homogenization and destruction of agglomerates. After formulation,
the tape is cast on the mylar sheet with the help of a doctor blade where the thickness
is governed by the height of the doctor blade above the substrate. In further steps the
tape is removed from the carrier film, followed by drying [24, 25]. A general tape
casting setup is shown in Fig. 6.1.
Green tapes of same and/or different components can be co-laminated
together. Lamination results in formation of multilayered ceramic with good
mechanical strength [26, 27]. The laminated tapes are then cut in to desired shapes
followed by sintering by placing them on a proper support [28, 29]. There should be
no reaction between the support and the sample. To avoid the reaction, the support
should be made with the same material as the layer in contact with it. Mostly, the
flatness of the sample is attained by placing a light weight on top of the samples. Non
appropriate support might result in vertical/longitudinal shrinkage rather than radial
shrinkage, which affects the planarity and flatness of the sintered samples.
After sintering, the thickness of the sample is influenced by different factors
like;
the powder nature
size, composition and formulation of the slurry
CH-6 Aqueous Tape Casting
101
temperature profile of sintering
the nature of the support for drying or sintering
Fig. 6.1: Schematic of a laboratory tape-casting set-up [30].
It was established from characterization results (chapter 5) that LSCTA- could
be considered as an anode/anode support for SOFC. In the next stage, LSCTA- was
processed in aqueous tape casting. This chapter gives an overview of results regarding
aqueous tape (dense and porous) casting of LSCTA-. The green tapes were laminated
into rectangular bars and conductivity of the sintered bars was determined using four
probe dc conductivity. Effect of impregnation on the conductivity of bars was also
investigated.
6.2 Experimental
6.2.1 Aqueous tape casting of LSCTA- powder
The slurries for the tape casting process were prepared by a ball milling
process that included two steps. In the first step, the ceramic powders were milled in
distilled water for 24 h with dispersant D3005 (The Dow Chemical Company) to
break down agglomerates in the powder. While in the slurry for porous tapes, a
required amount of PMMA (20 wt% to the weight of ceramic powder) was added as
pore former at this stage. In the second stage, other organic additives, such
plasticizers, binder and defoamer were added, followed by additional milling for 9 -
12 h. The recipe adopted to formulate both dense and porous LSCTA- slurries is given
in Table 6.1. The slurries were cast manually onto a Mylar sheet and the height of the
doctor’s blade was adjusted to give a 100 µm thick tape.
CH-6 Aqueous Tape Casting
102
Table 6.1 Tape casting recipe for LSCTA- anode substrate
Chemicals Function Amount (g)
Dense Tape Porous Tape
LSCTA- Ceramic Powder 15.0 12.0
Poly methyl methacrylate
(PMMA)
Pore Former ------- 3.0
De-ionized Water Solvent 12.0 12.0
Ammonium poly electrolyte
35 wt% (D 3005)
Dispersant 0.25 0.25
Polyethylene glycol Plasticizer Type 1 0.9 0.9
Glycerol Plasticizer Type 2 1.8 1.8
Poly vinyl alcohol 15wt% Binder 12.0 12.0
2,4,7,9 Tetramethyle (5-decyne)
4,7 diol
Defoamer
0.20 0.20
The mean particle sizes and particle size distributions of slurries were
determined by a particle analyzer (Model Horiba LA920, Delta Analytical, Inc.). The
viscosity was determined by using Brookfield DV-E Viscometer.
6.2.2 Lamination and sintering
After drying, the green tapes were laminated by placing different layers one on
another and passing through the laminator. The laminated green tapes were cut into
discs and bars. The sintering was done at 1400 °C for 2 hours in air. The dimensions
and weight of the green and sintered tapes were measured to determine their density
and degree of sintering. Micrographs of sintered tapes were taken with JEOL 5600
SEM.
6.2.3 Impregnation procedure
After sintering, the bars were impregnated with CeO2 and CeO2–Ni to improve
the electrocatalytic activity of LSCTA-. For CeO2 impregnation, the aqueous solutions
of 0.1 M Ce(NO3)3.6 H2O (99.99%, Alfa Aesar) was infiltrated drop-wise into the bar.
CH-6 Aqueous Tape Casting
103
Since the amount of catalyst that can be added in a single infiltration step is limited by
the pore volume of the LSCTA- scaffold and the concentration of the nitrate solution,
multiple impregnations with heat treatments at 400 °C between infiltrations were done
in order to achieve the desired weight loading. Finally, the sample was heated at 700
°C for 1 hour. After this step, the bar was weighed to calculate the amount of
infiltrated material. The process was repeated until loading level of CeO2 reached ~7
wt.%.
For CeO2–Ni impregnation, the first impregnation was done with CeO2. Then,
Ni was wet-impregnated from aqueous 0.1 M Ni(NO3)2. 6H2O. (99.9%, 5% max Pd,
Alfa Aesar) solution to a final loadings of ~ 0.9 wt%. The following bars were studied
for conductivity (see Table 6.2).
Table 6.2 Codes of the bars used in present study
Bars Bars Codes
Dense Bar A
Porous Bar B
7.7 wt% CeO2 Impregnated Porous Bar C
7.7 wt% CeO2 & 0.86 wt% Ni Impregnated Porous Bar D
*Pre-reduced Dense Bar E
* Reduced in a flow of 5% H2/ Ar at 1050 ºC for 24 hours to see the effect of pre
reduction.
6.2.4 Conductivity measurement of bars
The bars were subjected to four probe dc conductivity. Four Pt strips were
attached parallel to each other with the help of Pt paste to improve the contact
resistance and fired at 900 oC for 1 hour for consolidation. An R-type (Pt–Pt/Rh 10%)
thermocouple was used to measure the temperature of the sample. The current was
supplied from the current source (Keithley 224, USA) while the digital multimeter
(Keithley 2000, USA) was used to measure voltage drop across the probes. After
preparation, the bars were tested for conductivity under different conditions as listed
below;
CH-6 Aqueous Tape Casting
104
6.2.4.1 Conductivity measurement in air
The bars were heated in air and the conductivity was monitored as a function
of temperature.
6.2.4.2 Conductivity measurement during in-situ reduction
After heating the sample in air at about 880 ºC, the system was flushed with
flow of 5% H2/Ar, thus changing environment to reducing atmosphere for in-situ
reduction. The conductivity was monitored with time and partial pressure of oxygen
in 5% H2/Ar at 880 ºC.
6.2.4.3 Redox cycling at 880 oC
For redox cycling experiments, the atmosphere was changed from reducing to
oxidizing once the system attained a stable conductivity value at a constant
temperature. In the case of redox cycling of bars, after achieving a stable value at 880
oC in a reducing environment, the flow of 5% H2 was cut off and air was allowed to
leak into the furnace. Eventually, an increase of oxygen pressure resulted and the
conductivity dropped. After achieving a constant conductivity value in an oxidizing
atmosphere, 5% H2/Ar was again flushed into the system and above steps were
repeated.
6.3 Results and Discussions
6.3.1 Aqueous based slurry characteristics
After the first stage of ball milling, particle size analysis was carried out to
check the dispersion of ceramic powder. The uniform particle distribution as
illustrated by Fig. 6.2 shows that slurries are well dispersed in first step of stirring.
It could be attributed to optimal stirring times which ensured a smooth
homogeneous mixture for tape casting. The small shoulder in case of porous slurry is
due to PMMA present in the slurry. Additional evidence that both the LSCTA- and
pore former powders were well dispersed was obtained from the uniform appearance
of the final porous ceramics that were produced from the tapes.
CH-6 Aqueous Tape Casting
105
0.01 0.1 1 10 100 1000 100000
2
4
6
8
Vo
lum
e%
Particle diameter (m)
a
b
PMMA
Fig. 6.2: Particle size analysis of LSCTA- slurry; a) in the absence and b) in the
presence of PMMA.
After ensuring the homogeneous mixing, binder, plasticizers and defoamers
were added in a second step followed by another milling. The viscosity of the slurry at
the final stage plays an important role in casting. Thus, proper slip rheology is always
required for easy casting as well as for the mechanical stability.
The viscosity profile (Fig. 6.3) shows the decrease in viscosity of both the
dense and porous slurries with an increase in shear rate thus showing pseudo-plastic
profile.
0 5 10 15 20 25
500
1000
1500
2000
2500
3000
Shear Rate (rpm)
(c
P)
a
b
Fig. 6.3: Viscosity profile of LSCTA- slurry; a) in the absence and b) in the presence
of PMMA.
Pseudoplastic slips are characterized by their shear-thinning nature. This
behavior is helpful in the tape casting process because the slip displays a lower
CH-6 Aqueous Tape Casting
106
viscosity under the shear of the doctor blade and a higher viscosity downstream from
the blade, thereby resisting motion within the casted film [19].
6.3.2 Microstructure of dense and porous tapes
The green samples were sandwiched between zirconia coated porous alumina
plates for sintering. The zirconia bed prevented the reaction between porous alumina
plates and the samples. After sintering, both dense and porous samples came out to be
slightly different in colour as seen from Fig. 6.4. The dense samples appeared darker
than the porous ones.
Fig. 6.4: Visual effect of sintering on green samples; a) before and b) after sintering at
1400 °C in air.
Also, the samples came out to be flat and it can be gauged that the plates
helped in maintaining the flatness of the samples. The effect of PMMA addition can
be seen by looking at the microstructure of tapes (Fig. 6.5 & Fig. 6.6).
Fig. 6.5: Micrographs of surface view of dense tape (~ 92% ρth ) at different
magnifications; a) 1500X and b) 3500X.
In the case of a dense tape, a very compact microstructure with fully sintered
grains having 5-10 micrometers size is observed. However, porosity could be
CH-6 Aqueous Tape Casting
107
observed in porous tape where burning of fine particles of PMMA resulted in small
pores. It can be seen that PMMA addition not only induced porosity in the
microstructure but also limited the grain growth. Thus porosity can be tuned and
tailored by careful selection of pore former in terms of its nature and quantity. The
relative density of the sintered samples was calculated by measuring the dimensions
of the sintered bodies.
Fig. 6.6: Micrographs of porous tape (~76% ρth ).; a) surface view and b) cross
sectional view.
6.3.3 Conductivity of bars
The conductivity of the bars was investigated using four point dc conductivity
set up. Each bar is discussed below.
6.3.3.1 Dense bar (Bar A)
The sintered bar laminated from dense tape was heated in air and the
conductivity monitored is shown in Fig. 6.7a. It is seen that conductivity increases
with the temperature which implies that the sample becomes more conductive as the
temperature increases, showing semi conducting behaviour.
After heating the sample in air at 880 oC, 5% H2/Ar was flushed into the
system for in-situ reduction of the sample. A significant increase in conductivity is
observed as the sample is in-situ reduced. The conductivity increases with the extent
of reduction or with decrease of pO2. It can be seen that on changing the atmosphere
from air to 5% H2/Ar at 880 ºC, a three orders of magnitude increase was observed
(Fig. 6.7b). The increase in conductivity is attributed to reduction of Ti+4
to Ti+3
in
reducing conditions freeing electrons which result the increase in conductivity [31,
32].
CH-6 Aqueous Tape Casting
108
600 800 1000 1200
0.0
0.4
0.8
1.2
1.6
m
S c
m-1
Temperature (K)
a
0 10 20 30 40 50
-8
-6
-4
-2
0
2
-18
-15
-12
-9
-6
-3
0b
log
pO
2
ln m
S c
m-1
Time (hours)
Fig. 6.7: Conductivity profile of bar A; a) in air and b) in 5% H2/Ar.
6.3.3.2 Porous bar (Bar B)
The conductivity profile of sintered bar laminated from porous tape is shown
in Fig. 6.8a. It can be observed that conductivity increases with increase in
temperature in air, however, the conductivity value is comparatively less as compared
to bar A under similar experimental conditions.
CH-6 Aqueous Tape Casting
109
400 600 800 1000 1200
0.0
0.2
0.4
0.6
0.8
1.0
1.2
m
Sc
m-1)
Temperature (K)
a
0 3 6 9 12 15 18
-6
-4
-2
0
2
-18
-15
-12
-9
-6
-3
0b
log
p
O2
lnS
cm
-1
Time (hrs)
Fig. 6.8: Conductivity profile of bar B; a) in air and b) in 5% H2/Ar.
It might be attributed to the porous network in the case of bar B where the
grains are not as much connected due to porosity. However, it is seen that the same
bar offers a considerably higher conductivity after in-situ reduction at 880 oC (Fig.
6.8b) which might be attributed to facile reduction due to the porosity [33].
The porous bar was also subjected to redox cycling. For redox cycling, a
stable conductivity value (3.35 S cm-1
) was achieved at ~880 oC in a flow of 5%
H2/Ar. Then, the flow of 5% H2/Ar was cut off and air was allowed to leak into the
furnace. It resulted in an increase of oxygen pressure and consequent decrease in
conductivity to 1.65 mS cm-1
. After achieving a constant conductivity value in
oxidized atmosphere, 5% H2/Ar was again flushed into the system and the above steps
were repeated. The conductivity is plotted against time for five subsequent redox
CH-6 Aqueous Tape Casting
110
cycles in Fig. 6.9. The LSCTA- system was found redox stable as it recovered the
same value of conductivity after each cycle.
0 5 10 15 20 25-3
-2
-1
0
1
-16
-12
-8
-4
0
log
pO
2
log
S
cm
-1
Time (hrs)
Fig. 6.9: Redox cycling of bar A as a function of time at 880 oC. Dashed lines
show change of partial pressure of oxygen over time.
It can be noted that the oxidation occurs much faster than the (re)reduction
indicating a fast oxygen diffusion into the material. The removal of oxygen in a
subsequent reduction takes hours and most likely occurs very fast in a very thin
surface layer and then slows down as the oxygen removal progresses inside the
micron size grains of material. Reduction occurs fast, initially, followed by a much
slower step associated with the oxygen removal from the bulk of the LSCTA- grains
[34]. The redox stability makes this material an attractive candidate for a conductive
anode backbone for further impregnation in SOFC applications.
6.3.3.3 CeO2 (7.7 wt%) impregnated porous bar (Bar C)
It is well known that the electrocatalytic activity of the strontium titanate-
based materials is very poor compared to nickel, there is a need for an (or a couple of)
electrocatalyst(s) such as ceria that has to be introduced into the anode porous
backbone via impregnation to obtain a robust anode component. Thus the effect of
impregnates like ceria (CeO2) and ceria-nickel (CeO2-Ni) was studied on the
conductivity of the porous bars to investigate the role and stability of impregnated
catalysts into a porous backbone of LSCTA- prepared from solution phase method.
CH-6 Aqueous Tape Casting
111
The bar impregnated with CeO2 depicted a better conductivity profile in
oxidizing and reducing atmospheres. Fig. 6.10a shows the dependence of conductivity
on temperature in air where a marked increase in conductivity could be observed.
In case of in-situ reduction (Fig. 6.10b), an improvement in conductivity was
observed. The role of the ceria catalyst on the improvement of conductivity from 3.35
S cm-1
for the bare backbone to 5.87 S cm-1
for the ceria impregnated sample can be
noted.
200 400 600 800 1000 1200
0.0
0.5
1.0
1.5
2.0
2.5(
mS
cm
-1)
Temperature (K)
a
0 5 10 15 20 25
-6
-4
-2
0
2
-18
-15
-12
-9
-6
-3
0b
log
pO
2
lnS
cm
-1
Time (hrs)
Fig. 6.10: Conductivity profile of bar C; a) in air and b) in 5% H2/Ar.
Figure 6.11 presents the conductivity evolution upon 6 redox cycles
(following the same procedure as above) for ceria impregnated porous LSCTA-. It also
shows remarkable tolerance to pO2 changes (less than 10% decrease in conductivity
CH-6 Aqueous Tape Casting
112
compared to the initial value) and, therefore, has promising stability upon redox
cycling.
0 7 14 21 28 35-3
-2
-1
0
1
2
-20
-16
-12
-8
-4
0
log
Scm
-1
Time (hrs)
log
pO
2(a
tm)
Fig. 6.11: Redox cycling of bar C as a function of time at 880 oC. Dashed lines show
change of partial pressure of oxygen over time.
This increase in conductivity might be due to presence of catalytically
active CeO2 in the pores of the bar as seen from micrographs (Fig. 6.12).
Fig. 6.12: Micrographs of CeO2 impregnated bar.
6.3.3.4 CeO2 and Ni co-impregnated porous bar (Bar D)
The effect of a small amount of Ni was also studied along with CeO2
impregnation in bar D. It is anticipated that higher values of conductivity would result
from the synergic effect of CeO2 which is an oxidation catalyst and Ni which has
good catalytic activity.
CH-6 Aqueous Tape Casting
113
As expected, the bar co-impregnated with CeO2 and a catalytic amount of
Ni exhibited higher conductivity in air and 5% H2/Ar as shown in Fig. 6.13.
200 400 600 800 1000 1200
0.00
0.75
1.50
2.25
3.00
3.75
m
Sc
m-1
Temperature (K)
a
0 5 10 15 20 25 30 35-6
-4
-2
0
2
-16
-12
-8
-4
0b
log
pO
2
ln S
cm
-1
Time (hrs)
Fig. 6.13: Conductivity profile of bar D; a) in air and b) in 5% H2/Ar.
It is clear that a catalytic amount of Ni resulted in enhancing the
conductivity as compared to bare CeO2 impregnated bar. The results also suggest
that co-impregnation of CeO2 and Ni is good option to increase the conductivity. The
microstructure (Fig. 6.14) shows CeO2 and Ni particles in pores which are
responsible for enhanced value of conductivity. This fact has been evidenced in ref
[35] where improved cell performance was observed in Ceria-Nickel co-impregnated
cell configuration.
CH-6 Aqueous Tape Casting
114
Fig. 6.14: Micrographs of CeO2 -Ni co-impregnated bar.
Bar D was also tested for redox stability. Upon redox cycling, this bar also recovered
the same value of conductivity, thus is redox stable as depicted by Fig. 6.15.
0 9 18 27 36 45-3
-2
-1
0
1
2
-18
-15
-12
-9
-6
-3
0
log
pO
2
log
S
cm
-1
Time (hrs)
Fig. 6.15: Redox cycling of bar D as a function of time at 880 oC. Dashed lines show
change of partial pressure of oxygen over time.
6.3.3.5 Pre-reduced dense bar (Bar E)
One of the sintered bars from the dense tape was reduced at 1050 ºC for
24 hours before conductivity measurements. Upon reduction, the bar turned
completely black, giving an indication of complete reduction of Ti4+
to Ti3+
.
Thermocycling of pre-reduced bar in 5% H2/Ar is shown in Fig. 6.16.
CH-6 Aqueous Tape Casting
115
200 400 600 800 1000 120015
30
45
60
75
S
cm
-1
Temperature (K)
Heating Up
Cooling Down
Fig. 6.16: Thermocycling of bar E- in 5% H2/Ar.
It can be seen that upon cooling, the conductivity increases with
decreasing temperature showing positive temperature coefficient indicative of
metallic behavior. The behavior continues till ~350 K, marking the metal insulator
transition. Below this temperature, the conductivity decreases with decreasing
temperature. This bar offered the maximum value of conductivity at 880 oC.
However, subjecting this bar to redox cycling does not yield encouraging
results, to be expected from its compact and dense microstructure. The bar did not
recover same value of conductivity although was given much longer times for re-
reduction as shown in Fig. 6.17.
0 15 30 45 60 75 901.2
1.3
1.4
1.5
1.6
-16
-12
-8
-4
0
log
pO
2(a
tm)
log
S
cm
-1
Time (hrs)
Fig. 6.17: Redox cycling of bar E as a function of time at 880 oC. Dashed lines
show change of partial pressure of oxygen over time.
CH-6 Aqueous Tape Casting
116
The dense bar did not possess a redox stable conductivity at 880 °C and the
conductivity value was lost upon oxidation of the sample. The loss in conductivity
upon oxidation might be due to the incorporation of oxygen back into the perovskite,
causing conversion of Ti3+
to Ti4+
. The increased positive charge might have resulted
in a decrease in the number of electrons available for conduction, and the net result
would be decrease in the conductivity of the material.
6.3.4 Effect of impregnates on the kinetics of conductivity evolution
To understand the effect of impregnates on kinetics of conductivity, porous
(bar B) and ceria impregnated (bar C) bars were focused on. In Figs. 6.18a (oxidation)
and 6.18b (reduction) sections of the resistivity evolutions are plotted on the same
time scale for the native porous LSCTA- bar together with the ceria impregnated
sample at 880 °C, for five subsequent redox cycles.
There is a remarkable overlapping within cycles. Ceria addition clearly
accelerates the onset of changes in observed conductivity via redox. When the
atmosphere is changed back to a reducing atmosphere (see Fig. 6.18b) there is a
sudden decrease in resistivity after the first 0.5 hours that indicates that once a certain
concentration of oxygen vacancies is achieved within the material during the initiation
step, the reduction process tends to proceed faster. This delay was much smaller for
oxidation, probably because of higher vacancy content at the process onset. The
addition of ceria also improves the conductivity in both the reduced and oxidized
samples, possibly due to improved conductivity at the grain boundary.
It can be noted that the oxidation occurs much faster than the (re)reduction. On
closer analysis of the curves, we can note that upon the change in atmosphere and
after the time delay, the redox processes occur via two sequential rate determining
steps, a fast initial one followed by a slower evolution in time and in addition there is
also a time delay before the conductivity is seen to respond to changes in
atmosphere.
CH-6 Aqueous Tape Casting
117
Fig. 6.18: Resistivity variation vs. time for LSCTA- and ceria impregnated LSCTA- at
880 oC on; a) 5 oxidation and b) 5 reduction cycles.
A more detailed analysis is obtained using the diffusion equation which
assumes the surface reaction to be first order with the rate constant k proposed by
Song and Yoo [36] and plotting ln(1-((σt- σ0)/( σ∞ – σ0)) against time for the reduction
process. In this equation, σt is the mean conductivity at time t where as σ0 and σ∞
denote the initial conductivity at t=0 and the final conductivity at the new equilibrium
i.e., at t→∞. If we consider the surface diffusion as the rate determining, the slope of
the curve is -2k/a is given by:
0
0
2ln(1 )t kt
a
(6.1)
Here a/2 is the half grain size of the material, as this is the best estimate for the
minimum diffusion distance in a porous material. For a dense sample or crystal this
CH-6 Aqueous Tape Casting
118
value would come from the smallest dimension of the body. By analogy the diffusion
process on oxidation can be considered in terms of resistivity rather than conductivity.
Figure 6.19 shows the relative resistivity/conductivity change in semi-
logarithmic scale with time for oxidation and reduction cycles of LSCTA- and CeO2-
impregnated LSCTA- at 880 oC. Clearly two distinguishable slopes can be seen
indicating two-fold relaxation kinetics. Such twofold relaxation kinetics has been
attributed to fast relaxation in the oxygen sublattice followed by slow relaxation in a
cation sublattice for TiO2 [37].
Fig. 6.19: Resistivity/conductivity relaxation of LSCTA- and ceria impregnated
LSCTA- at 880 oC upon; a) oxidation and b) reduction. Two different kinetic processes
are indicated by dotted lines with different slopes.
CH-6 Aqueous Tape Casting
119
From the slope of these curves, rate constants for both oxidation and reduction
kinetic processes were calculated. Tables 6.3 gives values of onset delay, rate
constants and relevant times for each relaxation type extracted from observed
oxidation cycles.
Table 6.3 Rate constant k (cm s-1
) calculated for two fold relaxation kinetics for
oxidation cycles of La0.2Sr0.25Ca0.45TiO3 (LSCTA-) and CeO2 impregnated LSCTA-
(LSCTA-:CeO2) at 880 oC
LSCTA- LSCTA-:CeO2
Onset delay (hrs) 0.07 0.03
Region I
koxIx10
7 (cm s
-1) 3.00 2.69
Time range (hrs) 0.07 - 0.11 0.05 - 0.10
Region II
koxIIx10
7 (cm s
-1) 0.40 0.20 - 1.20
Time range (hrs) 0.30 0.20 - 1.40
Similarly, the values were extracted from reduction cycles and are tabulated in
Table 6.4.
Table 6.4 Rate constant k (cm s-1
) calculated for two fold relaxation kinetics for
reduction cycles of La0.2Sr0.25Ca0.45TiO3 (LSCTA-) and CeO2 impregnated LSCTA-
(LSCTA-:CeO2) 880 oC
LSCTA- LSCTA-:CeO2
Onset delay (hrs) 0.40 0.20
Region I
kredIx10
7 (cm s
-1) 1.40 2.18
Time range (hrs) 0.40 - 0.50 0.20 - 0.30
Region II
kredIIx10
7 (cm s
-1) 0.18 0.24
Time range (hrs) 0.90 – 2.5 0.40 – 1.40
CH-6 Aqueous Tape Casting
120
Considering reduction, ceria impregnation accelerates the process decreasing
the onset delay and increasing the rate constant for both relaxation stages by 33-50%.
The oxidation processes are more facile than reduction, see Fig. 6.18, with shorter
onset times and higher rate constant for stage I but not for stage II. Ceria impregnation
results in a decrease in onset time but also slightly lower rate constants. Thus,
different factors determine the influence of ceria impregnation on oxidation and
reduction rates in these experiments.
It is important to note that the reduction experiments start from a highly
oxidized sample with few oxygen vacancies and so catalysis of surface exchange by
ceria has a significant influence on both the initial delay whilst percolation is achieved
and also on the reduction process. In the oxidation stage, the sample has a high
number of vacancies at the start of oxidation; hence surface exchange at ceria has
little beneficial influence, apart from decreasing the onset delay.
6.3.5 Comparison of conductivity
6.3.5.1 Conductivity of bars in air
It is quite obvious that the conductivity of all the bars increases with
temperature in air. However, the impregnated bars offered more conductivity.
Generalizing the conductivity profile of all the bars in air, it can be said that
impregnated bars are more conductive than their non impregnated counter parts
because impregnation results in improved conductivity. Comparison of the value of
conductivities observed in air at 880 °C is given in Table 6.5.
6.3.5.2 In-situ reduction at 880 oC
Upon in-situ reduction at 880 oC, the conductivity of all the bars increased
with extent of reduction. This is a characteristic of n-type semiconductors. The
impregnation resulted in significant increase of conductivity as can be seen from Fig.
6.20 where a comparative graph is shown for all the bars at 880 oC.
CH-6 Aqueous Tape Casting
121
0 4 8 12 16
0.0
0.4
0.8
1.2
1.6
2.0
ln S
cm
-1
Time (hrs)
a
b
c
d
Fig. 6.20: Conductivity profile of bars; a) Bar A, b) Bar B, c) Bar C, d) Bar D after
in-situ reduction in 5% H2/Ar at 880 oC.
It can be seen that impregnated samples offer higher values of conductivities than
bare skeletons under the same experimental conditions (section 6.2.4.2). The catalytic
amount of Ni plays a role in contributing to higher conductivity in bar D. The
conductivity values in reducing atmosphere at 880 oC are tabulated in Table 6.5.
Table 6.5 Conductivity of bars in air and 5% H2/Ar at 880 °C
*Bar codes Air
(mS cm-1
)
5% H2/Ar
(S cm-1
)
A 1.49 2.99
B 1.03 3.40
C 2.09 5.90
D 3.24 7.57
E ---- 38.0
*Table 6.2
From the table 6.5, it is evident that the conductivity increases by about
three orders of magnitude upon reduction in 5% H2/Ar for all the bars tested.
CH-6 Aqueous Tape Casting
122
6.3.5.3 Redox cycling of bars
From the redox cycling graphs, it becomes clear that redox stability is also a
function of microstructure. For redox stability, the microstructure should facilitate the
diffusion of gases. Thus, the bars laminated from porous tape recovered the same
value of conductivity, hence are redox stable while the dense bars are not redox stable
under same experimental conditions. Encouraging results were obtained from the
impregnated bars where high conductivity as well as redox stability was observed.
6.4 Conclusions
Aqueous tape casting is a quick and rapid technique to fabricate thin SOFC
anodes. For uniform, homogeneous and crack free green tapes, the correct slurry
formulation is essential. Slurry formulation was optimized for both the dense and
porous green tapes of LSCTA-. The rectangular bars fabricated from green tapes by
lamination were sintered and tested for conductivity measurements using van der
Pauw set up. All the bars show semi-conducting behavior in air where the
conductivity increases with temperature. Upon reduction, the conductivity increases
by about three orders of magnitude thus giving a clue to n-type conduction. The
impregnated bars offered high values of conductivity. The kinetic studies revealed
that CeO2 impregnation increased conductivity by enhancing reduction kinetics, but
had limited effect on the oxidation processes, which were a little faster in absence of a
catalyst. Whilst the obtained rate constants were derived using some approximations,
all samples were treated similarly, hence the increase of rate constant kred, by about
50% due to ceria impregnation is significant. Redox cycling experiments showed
encouraging redox stability of the ceramic system and the impregnated counterparts
thus imparting a suitable anode support candidateship. The results show that the
conductivity can be modified and tuned by impregnating with suitable agents.
Impregnation along with pre-reduction is a cost effective and easy way to enhance the
conductivity.
CH-6 Aqueous Tape Casting
123
REFERENCES
1. G.N. Howatt, R.G. Breckenridge, M. Brownlow, Am. Ceram. Soc., 1947, 30,
237 – 242.
2. G.N. Howatt, U.S. Patent 2,582,993, 1952.
3. A.I.Y. Tok, F.Y.C. Boey, K.A. Khor, J. Mat. Process. Tech., 1999, 89-90, 508
–512.
4. H. Moon, S.D. Kim, S.H. Hyun, H.S. Kim. Int. J. Hydrogen Energy, 2008, 33,
1758 – 1768.
5. K.C. Wincewicz, J.S. Cooper, J. Power Sources, 2005, 140, 280 – 296.
6. W.J. Quadakkers, H. Greiner, M. Hansel, A. Pattanaik, A. S. Khanna, W.
Mallener, Solid State Ionics, 1996, 91, 55 – 67.
7. S.B. Savignat, M. Chiron, C. Barthet, J. Eu. Ceram. Soc., 2007, 27, 673 –
678.
8. M.P. Albano, L.B. Garrido, Mat. Sci. Eng. A, 2006, 420, 171 – 178.
9. D. Montinaro, V.M. Sglavo, M. Bertoldi, T. Zandonella, A. Arico`, M.L. Faro,
V. Antonucci, Solid State Ionics, 2006, 177, 2093 – 2097.
10. K.J. Yoon, P. Zink, S. Gopalan, U.B. Pal. J. Power Sources, 2007, 172, 39
– 49.
11. L.C. Thomas, W.S. Stephen, J Power Sources, 2007, 174, 221 – 227.
12. S. Linderoth, J. Electroceram., 2009, 22, 61 – 66.
13. S.H. Nien, C.S. Hsu, C.L. Chang, B.H. Hwang, Fuel Cells, 2011, 11, 178 –
183.
14. T. Chariter, A. Bruneau, J. Eur. Ceram Soc., 1993, 12, 243 – 247.
15. J. E. Smay, J.A. Lewis, J. Am. Ceram. Soc., 2001, 84, 2495 – 2500.
16. M, Cologna, V.M. Sglavo, Int. J. Appl. Ceram. Technol., 2010, 7, 803 –
813.
17. L.H. Luo, A.I.Y. Tok. Mater. Sci., Eng. A, 2006; 429, 266 – 271.
18. F. Snijkers, A. de Wilde, S. Mullens, J. Luyten, J. Eu. Ceram. Soc., 2004, 24,
1107 – 1110.
19. R.E. Mistler, E. R. Twiname in Tape Casting: Theory and Practice, 2000,
The
American Ceramic Society, Westerville, 2000.
20. R. Moreno, Am. Ceram. Soc. Bull., 1992, 71, 1521 – 1531.
CH-6 Aqueous Tape Casting
124
21. R. Moreno, Am. Ceram. Soc. Bull., 1992, 71, 1647 – 1657.
22. J.H. Feng, F. Dogan, J. Am. Ceram. Soc., 2000, 83, 1681 – 1686.
23. S.B. Reddy, P.S. Singh, N. Raghu, V. Kumar, J. Mater. Sci., 2002, 37, 929
– 934.
24. D. Hotza, P. Greil, Mat. Sci. Eng. A, 1995, 202, 206 – 217.
25. W. Wunderlich in Ceramic Materials, Intech, 2010.
26. H. Park, H. Moon, S.C. Park, J. Power Sources, 2010, 195, 2463 – 2469.
27. X. Li, J. Lu, H. Wang, Sci. Sinter., 2011, 43, 305 – 312.
28. A. Grosjean, O. Sanséau, V. Radmilovic, A. Thorel, Solid State Ionics, 2006,
177, 1977 – 1980.
29. R. Costa, J. Hafsaoui, A.P. Almeida de Oliveira, A. Grosjean, M. Caruel,
A. Chesnaud, A. Thorel, J. App. Electrochem., 2009, 39, 485 – 495.
30. http://multilayer.4m-association.org/node/37
31. P.R. Slater, D.P. Fagg, J.T.S. Irvine, J. Mater. Chem., 1997, 7, 2495 –
2498.
32. O.A. Marina, N.L. Canfield, J.W. Stevenson, Solid State Ionics, 2002, 149, 21
– 28.
33. D. Neagu, J.T.S. Irvine, Chem. Mater., 2011, 23, 1607 – 1617.
34. D. Neagu, J.T.S. Irvine, Chem. Mater., 2010, 22, 5042 – 5053.
35. M.C. Verbraeken, B. Iwanschitz, A. Mai, J.T.S. Irvine, J. Electrochem.
Soc., 2012, 159, F757 – F762.
36. C.R. Son, H.I Yoo, Solid State Ionics, 1999, 120, 141 – 153.
37. D.K. Lee, H.I. Yoo, Solid State Ionics, 2006, 177, 1 – 9.
Chapter 7
Microstructure Optimization with Pore
Formers
125
Microstructure Optimization with Pore
Formers
Abstract
The microstructure of an anode is crucial for the performances of the entire cell.
The present chapter incorporates results obtained for getting optimized microstructure of
LSCTA- sintered bodies using commercial pore formers (PFs). To introduce the porosity
in LSCTA- tapes, commercial pore formers like graphite, polymethylmethacrylate
(PMMA) and glassy carbon (GC) were used. It was observed that pre-treated powder
leads to good microstructure with commercially available pore formers. In parallel,
porosity in LSCTA- tapes was achieved with carbon microspheres (CMS) which were
synthesized by an inexpensive hydrothermal method. From the results, it was inferred
that these spheres could be used as effective PF.
7.1 Introduction
It is well-known that solid oxide fuel cell (SOFC) performance is dependent
strongly on the composition, fabrication process and the resulting microstructure of the
electrodes because the gas permeability and thus electrical conductivity strongly depends
on microstructural parameters such as porosity, particle size, pore shape and distribution
[1-4]. The porous ceramic network provides the mechanical strength to the fuel cell and
also allows an easy flow of the gases to and from the electrolytic membrane. Many
researchers have investigated electrode performance with respect to the microstructure of
the electrodes, especially for the long-term stability of the electrodes [5-8].
The proper microstructural conditions also reduce the concentration polarization,
which is related to the diffusion of the reactant or product of the electrode reaction. The
correlation between the microstructure of anode and the electrochemical performance of
SOFC has been reported [9-13].
For an optimized microstructure, the minimum degree of porosity is required to
create sufficient electrochemical reaction sites and gas permeability in the conduction
CH-7 Microstructure Optimization
126
layer. Although controlling the sintering process may yield some porosity, however it is
not effective at producing large pore pathways within the microstructure particularly
when only fine ceramic particles (<1 µm) are used to fabricate the anode, which makes
the final porosity generally low to fulfill the anode requirements. It can particularly be an
issue with anode supported designs because the anode is relatively thick.
To produce engineered porosity in ceramic laminates, usually organic additives
(pore formers) are mixed into the ceramic powder slurry before tape casting [14]. The
burning of these organic additives leads to fine pores and thus help to have desired
porosity in the sintered tapes. The idea of using pore-forming agents (PFA) to control the
size and distribution of porosity in tape cast ceramics has been employed in various
studies [15-17].
Tape casting with different pore formers like graphite, glassy carbon, polymethyl
methacrylate (PMMA), rice starch or a combination of these has been used to fabricate
the anodes for SOFC [18-22]. Both glassy carbon and PMMA are very expensive
whereas with graphite, horizontal planes of pores are formed, the tapes are often dry
which could lead to de-lamination.
Basically, a good microstructure is characterized by interconnected porosity along
with desired mechanical strength. The fabricated anode should have good adherence to
the electrolyte. The present chapter gives an account of different strategies to get the
functional microstructure for LSCTA- anode substrate. The synthesized LSCTA- powder
was processed in aqueous tape casting to yield porous tapes using pore formers. The
results are given into two parts.
The first part of the chapter is dedicated to the results obtained by using
commercial pore formers where as the second part gives an account of synthesis and
application of carbon microspheres (CMS) as pore former and resulting morphology.
CH-7 Microstructure Optimization
127
7.2 Microstructure optimization with commercial pore formers
The following section gives a brief overview of results obtained to optimize
microstructure with commercially available pore formers.
7.2.1 Experimental
7.2.1.1 Tape casting with different pore formers
The method suggested by Corbin and Apt´e [14] was chosen while working with
pore formers. In this method, the pore formers are considered as additional ceramic
powder and the tape formulation is adjusted keeping the weight ratio, (total amount of
organics in the green tape):(ceramic powder + pore formers) constant. Different pore
formers along with their weight percent used in the study are listed below;
20 wt% PMMA
20 wt% PMMA+ 5 wt% Graphite
20 wt% PMMA+ 10 wt% Graphite
25 wt% Graphite+5 wt% Glassy carbon
30 wt% Graphite
40 wt% Graphite
Slurries were prepared in the same way as described in the previous chapter in section
6.2.1. The micrographs were taken with JEOL 5600 SEM.
7.2.1.2 Co-lamination and co-sintering
Green tapes were laminated by putting them one on the other and then passing
them through the laminator. After laminating the tapes, they were cut into discs form and
sintered using the following temperature program (Fig. 7.1).
CH-7 Microstructure Optimization
128
Fig. 7.1: Sintering profile for green LSCTA- samples in air.
7.2.1.3 Ink preparation of YSZ
The recipe used for YSZ ink formulation is given in Table 7.1.
Table 7.1 Recipe for YSZ ink
Chemicals Function Mass (g)
YSZ (8 mol% Pi Kem) Ceramic 10.0
KD1 Dispersant 0.3
Acetone To assist milling 20.0
5 wt% PVB in terpineol Binder 4.3
For ink preparation, first the YSZ was ball milled with dispersant in acetone for
24 hours at 160 rpm. Then the contents were emptied into a beaker followed by the
addition of the binder. The beaker was covered with perforated Nafion film and acetone
was let to evaporate by constant stirring, to have homogeneous and consistent ink. When
the acetone evaporated, the ink was collected in a sample vial.
CH-7 Microstructure Optimization
129
7.2.2 Results and Discussion
7.2.2.1 Playing with pore formers
To induce porosity, pore formers of different types and in different quantities
were used. It was found that pore formers did not reveal their effect in the final
microstructure and dense and compact microstructures were obtained after sintering.
Examples can be seen by looking at the microstructure obtained by using graphite and
mixture of graphite and PMMA (Fig.7.2).
Fig. 7.2: Micrographs of LSCTA- porous green tapes after sintering at 1400 oC. Amount
of pore formers in green tapes being; a) 20 wt% PMMA + 10 wt% Graphite and
b) 30 wt% Graphite.
It is clear that the microstructure is not porous enough to be used as an anode for
solid oxide fuel cells although pore formers were used in large quantities. The behavior
might be explained that LSCTA- again densifies after the burning of the pore formers.
Similar micrographs were obtained with other pore formers.
7.2.2.2 Changing sintering program
The sintering program was changed in the following respects;
7.2.2.2.1 Increasing sintering rate
To avoid sintering of LSCTA- powder, the sintering program was changed. The
idea was that increasing the sintering rate might not allow the sample to sinter very
much. The sintering program and resultant microstructure while using 40 wt% graphite
are shown in Fig. 7.3.
CH-7 Microstructure Optimization
130
Fig. 7.3: a) Micrographs of LSCTA- porous green tapes containing 40 wt% Graphite after
sintering at 1400 oC and b) corresponding sintering profile.
7.2.2.2.2 Changing sintering atmosphere
The above results suggest a very sinteractive nature of LSCTA- powder
which results in collapsing of the pores after the burning of pore formers, leading to
dense and compact microstructures even in the presence of pore formers. Therefore, it is
essential to have a structure in which particles don’t combine and fill the pores after the
organic content is burnt. Thus the sintering atmosphere was changed. The tape containing
30 wt% graphite was first heated up in Ar using the following sintering profile:
Fig. 7.4: Sintering profile for green LSCTA- samples containing 30% graphite in 5%
H2/Ar.
CH-7 Microstructure Optimization
131
After which, the sample was sintered in air. But this strategy did not work well
and again a dense microstructure was observed.
7.2.2.3 Pretreatment of powder
Sintering is closely related to the particle size; small sized particles result in
more sintering. Thus, sintering could be controlled by changing the particle size. In the
present case, the sinteractivity of the powder was quenched by use of coarse particles. To
achieve, the powder was heat treated at 1100 ºC for 6 hours. The temperature of 1100 oC
was chosen by looking at the dilatometric profile of LSCTA- powder. It was expected that
thermal treatment of calcined powder would result in particle coarsening and hence
limiting the sintering. The desired porosity (functional microstructure) could be achieved
by adopting this strategy using pre-treated powder as depicted in Fig. 7.5.
Fig. 7.5: Micrographs of LSCTA- tape after sintering at 1400 oC. Slurry formulation was
prepared using; a) calcined LSCTA- powder and b) thermally treated LSCTA- powder.
7.2.2.4 Adherence to YSZ
Since the anode has to be in close contact with YSZ electrolyte in a cell
configuration, there should be no delamination between them during or after cell testing
as any defect with the electrolyte would deteriorate the functioning of solid oxide fuel
cell. To check the adherence to YSZ, the green tapes of both LSCTA- (tape obtained by
pre-treated powder) and YSZ were co-laminated and co-sintered using sintering program
shown in Fig. 7.1. After co-sintering, the sample came out to be flat and without any
flaw.
CH-7 Microstructure Optimization
132
Fig. 7.6: Effect of co-sintering at 1400 °C; a) LSCTA- co-laminated with YSZ and
b) LSCTA- with screen printed YSZ.
Good adherence with YSZ was found as seen from Fig. 7.6a. The effect of screen
printed YSZ was also shown in Fig. 7.6b. YSZ ink was prepared (section 7.2.1.3) and
screen printed on LSCTA- green tape and then co-sintered. The SEM image shows a well
adhered thin layer of YSZ (about 10 µm), sandwiched between LSCTA- tapes with no de-
lamination.
7.3 Microstructure Optimization with Synthesized Carbon
Microspheres as Pore Former
Among different forms of carbon, carbon spheres have found a range of
applications e.g., in catalyst supports [23], lithium-ion secondary batteries [24], drug
delivery [25] and energy storage medium [26] because of their interesting properties.
Another application of these carbon spheres could be their use as a pore former
because they could be synthesized by in-expensive methods like hydrothermal treatment
[27-29]. The variation of different experimental parameters during the synthetic route can
tune and tailor the morphology of carbon spheres [30].
In current research work, carbon spheres were synthesized by hydrothermal
treatment and characterized by XRD, FTIR and TGA and electron microscopy. The effect
of different parameters on hydrothermal synthesis was studied. Further they were used as
pore former in LSCTA- tapes.
CH-7 Microstructure Optimization
133
7.3.1 Experimental
Carbon micro-spheres (CMS) were prepared with hydrothermal treatment of
sugar via a very simple synthesis. 15 mL of aqueous sugar solution was sealed in 40 mL
autoclave which was heated at 180 °C in a furnace for 12 hours. The final black product
was collected in Falcon tubes, washed repeatedly with distilled water and ethanol using
ultracentrifugation. The solid was vacuum dried at 80 °C.
Thermogravimetric TGA was performed on a Netzch STA 449c equipped with
ProteusTM
thermal analysis software in air at heating rate of 3oC min
-1. The phase
formation was studied using a PANalytical Empyrean diffractometer using Cu-Kα1
radiation in the range of 20o to 90
o. Particle size analysis was carried out on a Malvern
Instruments Mastersizer 2000. Density measurement was done with Micromeritics
AccuPyc 1340. BET (Brunauer, Emmett and Teller) measurements were taken on a
Micromeritics TriStar II 3020 instrument. The morphology of CMS was studied using
JEOL 6700F field emission microscope. FTIR was done with Perkin Elmer Spectrum GX
FT-IT System.
The slurries for tape casting process were prepared in the same way as given in
section 7.2.1.1 by a two step ball milling process. However in the first step, the ceramic
powder, pre-treated LSCTA- (section 7.2.2.3) was milled in distilled water for 24 h with
dispersant KD6 and carbon micro spheres as pore formers in different weight ratio. In the
second stage, other organic additives, such as plasticizers, binder and defoamer were
added, followed by additional milling for 9–12 h. The slurries were then cast manually
onto a Mylar sheet followed by drying. After drying, the green tapes were co-laminated
to increase the mechanical strength and cut into 3 cm diameter discs shape and then
sintered at 1400 oC.
7.3.2 Results and discussion
7.3.2.1 X-ray diffraction
The XRD pattern for the pure carbon spheres is shown in Fig. 7.7. The peak could
be assigned to graphitic 002 plane where the broadening suggests that the carbon spheres
synthesized by means of hydrothermal process have a low degree of crystallinity and
graphitization and the possible presence of amorphous carbon [31, 32].
CH-7 Microstructure Optimization
134
0 20 40 60 80 100
400
800
1200
1600
2000
*
Sample Holder
*
*
Inrt
en
sit
y
Diffraction angle (2)
*
Fig. 7.7: XRD pattern of carbon spheres synthesized from hydrothermal treatment of
0.5 M sucrose solution.
7.3.2.2 Thermal gravimetric analysis
Oxidation of carbon spheres was done in air at heating rate of 3° min-1
as shown
in Fig. 7.8.
0 200 400 600 800 1000
0
20
40
60
80
100
% M
as
s lo
ss
Temperature (oC)
Fig. 7.8: TGA of carbon spheres synthesized from hydrothermal treatment of 0.5 M
sucrose solution.
CH-7 Microstructure Optimization
135
It can be seen with the oxidation starts at 300 °C and below this temperature, a
smooth mass loss of carbon could be observed whereas complete mass loss (seemingly as
CO and/or CO2) was observed before reaching 600 °C.
7.3.2.3 Fourier transform infrared spectroscopy
To identify the functional groups after hydrothermal treatment, the FTIR
spectrum was taken. The presence of hydroxyls groups is evidenced in Fig. 7.9 by the
strong wide peak at 3420 cm-1
attributed to the O–H stretching vibrations implying the
existence of a large number of residual hydroxyl groups and intermolecular H-bonds
[33]. The bands in the 2880–2980 cm−1
originated from the stretching vibration of the C-
H groups of the saturated alkyl hydrocarbon. There are several other characteristic
absorption bands positioned at 1704 (C=O stretching), 1616 (C=C stretching of aromatic
and furanic rings), 1290 and 1211 (C–O–C stretching), 1023 (characteristic furan 1030 to
1015 cm-1
band), and 798 and 756 cm-1
(furanic out-of-plane C–H deformation) [19, 21].
4000 3600 3200 2800 2400 2000 1600 1200 800
s(C=O)
s(C-H)
s(O-H)Tra
nsm
itta
nce
(a.
u.)
Wavenumber (cm-1
)
s(C=C)
b(C-Haromatic)oop
Fig. 7.9: FTIR of carbon spheres synthesized from hydrothermal treatment of 0.5 M
sucrose solution.
These results indicate that the surface of colloidal carbon spheres is hydrophilic
and has a distribution of hydroxyl and carboxyl groups [34]. Usually the oxygen
CH-7 Microstructure Optimization
136
functionalities of these surface functionalized carbon spheres makes them attractive
candidates for a number of applications like template for synthesis of hollow spheres of
different inorganic materials like (Ga2O3, GaN, WO3, SnO2, etc.) [35-37]. The presence
of functional groups also helps to encapsulate nanoparticles in their cores [38].
7.3.2.4 Scanning electron microscopy
Effects of sucrose concentration, pH and solvent polarity were investigated for
achieving desired carbon spheres.
7.3.2.4.1 Effect of sucrose concentration
The SEM images indicate that carbon spheres with perfect spherical
morphology were obtained after hydrothermal treatment of sucrose solution in de-ionized
water (Fig. 7.10). It was found that with 0.1 M sucrose, 2-3 µm sized carbon spheres
were isolated. The product yield of carbon spheres increased with the concentration of
sucrose as the increased concentration of sugar leads to an increase in the amount of
product as the processs of polymerization and aromatization are favoured. However, with
1.0 M, particle necking was found under experimental conditions as could be seen from
Fig. 7.10d. The micrograph also shows that the spheres formed under these conditions are
solid and not hollow.
With 0.5 M sucrose solution, according to carbon weight percentage, the yield of
the product is more than 90%. The density of the carbon spheres was found to be ca. 1.46
g cm-3
, a value close to glassy carbon, ca. 1.50 g cm-3
, and lower than graphite, ca. 2.26 g
cm-3
. A concentration of 0.5 M was selected for further experiments based on the yield of
carbon spheres formed and the desired morphology of carbon spheres (no interconnected
particle or particle necking and good product yield).
CH-7 Microstructure Optimization
137
Fig. 7.10: Micrographs of carbon spheres synthesized from hydrothermal treatment of
sucrose solution. The concentration of sucrose solution being; a) 0.1 M, b) 0.5 M, c) 1.0
M, d) 1.0 M (on high magnification).
7.3.2.4.2 Effect of pH
To check the effect of pH, on the morphology of carbon spheres, the formation of
spheres was carried out in acidic, basic and neutral media (Fig. 7.11) where pH was
adjusted using CH3COOH in case of acidic, Na
2CO
3 in case of basic and combination of
these for neutral while all other experimental conditions were kept constant.
CH-7 Microstructure Optimization
138
Fig. 7.11: Micrographs of carbon spheres synthesized from hydrothermal treatment of 0.5
M sucrose solution at different pH; a) pH 4, b) pH 10, c) pH 7.
It could be seen from the micrographs that in acidic media, the spheres were
merged into one another. In case of basic media, the spheres could be seen but it seems
that if the reaction did not complete well. No spherical shaped particles were observed in
case of neutral pH. Thus none of the acidic, basic and neutral pH helped to get the
spheres with desired morphology.
7.3.2.4.3 Effect of solvent polarity
When pure water was replaced by a binary mixture of water and ethanol (2:1),
very nicely dispersed spheres were obtained (Fig. 7.12). Increasing the ethanol to water
ratio (2:1) did not give encouraging results.
CH-7 Microstructure Optimization
139
Fig. 7.12: Micrographs of carbon spheres synthesized from hydrothermal treatment of
sucrose solution in the presence of different solvent media; a) H2O, b) H
2O:EtOH
= 1:2 and c) H2O:EtOH = 2:1.
Some of the experiments were also carried out to optimize the ultrasonication
time for preparing sucrose solutions prior to hydrothermal treatment and dwelling time.
Following these above described experiments, a set of optimized parameters were
established and are summarized in Table 7.2.
Summarizing the results, it can be said that homogeneous and well dispersed
microspheres could be produced by hydrothermal treatment of 0.5M sucrose at 180 oC for
CH-7 Microstructure Optimization
140
12 hours. Ultrasonication the sucrose solution for about 20-25 minutes prior to
hydrothermal treatment also helps to get the microspheres with desired morphology.
Table 7.2 Set of optimized parameters for synthesis of carbon microspheres from
hydrothermal treatment of sucrose at 180 oC
Parameters Optimized Value
Sucrose concentration 0.5 M
Solvent H2O:EtOH = 1:2
Dwell time 12 h
Ultrasonication time for sucrose solution
before hydrothermal treatment ~25 min
7.3.2.5 Application of CMS as pore former
The synthesized carbon monospheres (CMS) with optimal morphology were then
used as pore formers. The experimental detail is given in section 7.3.2.
The CMS were applied to LSCTA- slurry in 10, 20, 30 weight percent ratio. The
back scattered images obtained by FEG -6700F (Fig. 7.13) show that after sintering, the
carbon spheres burn leaving voids that induce interconnected porosity in the ceramics. It
can be seen the porosity is proportional to amount of microspheres used, higher
concentration of microspheres leads to larger porosity. These microspheres were also
used with graphite where the combination of both resulted in a good distributed porosity.
The results indicate that CMS prepared from simple hydrothermal methods could be used
as effective pore formers.
CH-7 Microstructure Optimization
141
Fig. 7.13: Micrographs of LSCTA- tape after sintering at 1400 °C using CMS as pore
former, concentration of CMS being; a) 10 wt%, b) 20 wt%, c) 30 wt%, d) 30 wt%
(Graphite:HT-C=1:1).
Image J software was used to estimate the porosity in the samples. The calculated
porosity is tabulated in Table 7.3.
CH-7 Microstructure Optimization
142
Table 7.3 Porosity calculated in the back scattered images of sintered LSCTA- tapes
containing carbon microspheres as pore former
Weight % pore former Porosity %
10% CMS 25 ± 2
20% CMS 33 ± 3
30% CMS 45 ± 2
CMS+ Graphite (15+15) 50 ± 3
It can be noted that porosity increases proportionally with the addition of CMS
pore former. The % porosity increases with the added pore formers while the extra
porosity may arise due to burnout of the organics added during slurry formulation for
tape casting.
7.3.2.6 Adherence to YSZ
As discussed earlier, for a functional anode, there should be no delamination
between anode and electrolyte during or after cell operation. To check the adherence to
YSZ, the green tapes of both porous LSCTA- (having CMS as microspheres) and YSZ
were co-laminated and co-sintered using sintering program shown in Fig. 7.14.
Fig. 7.14: Micrographs of porous LSCTA- tape (with 20 wt% CMS) co-laminated with
YSZ after sintering at 1400 °C.
After co-sintering, the sample came out to be smooth and planar. Good adherence
with YSZ could be seen as seen from Fig. 7.14 which is useful for its SOFC application
as a stable anode support.
CH-7 Microstructure Optimization
143
7.3.2.7 Mechanism of formation
A number of studies have focused the mechanism of the formation of carbon
spheres from hydrothermal treatment of mono-saccharides [28, 34, 35]. A recent study
has demonstrated stepwise formation of carbon spheres where the sugar is dehydrated
first. Then dehydrated products self-decompose into organic acids (acetic, levulinic, and
formic acids) where the hydronium ions formed from these acids act as a catalyst in
subsequent reaction stages [36]. In the next step, the dehydrated and fragmented products
are polymerized and condensed followed by aromatization of polymers via keto-enol
tautomerization or intramolecular dehydration to form C=C bonds. Finally, nucleation
and subsequent growth by diffusion and linkage of species from the solution with the
reactive oxygen surface functionalities like hydroxyl, carbonyl, carboxylic and ester to
the nuclei surface takes place leading to carbon sphere formation [28].
7.4 Conclusions
LSCTA- produced by Pechini’s method is very sinter active. The pre-treated
LSCTA- powder at 1100 oC gave good microstructure with commercial pore formers
which also showed good compatibility with YSZ. The pretreatment causes coarsening of
particles which helps to quench the sinterability of the powder. Carbon microspheres
were successfully prepared by an optimized hydrothermal treatment of sucrose.
Furthermore, these spheres were used as pore former in the green tape of LSCTA-; an
emerging anode for SOFC. It is anticipated that microspheres produced by the
economical and in-expensive hydrothermal method could be used as pore former.
CH-7 Microstructure Optimization
144
REFERENCES
1. W.Z. Zhu, S.C. Deevi, Mat. Sci. Eng. A, 2003, 362, 228 – 239.
2. J.B. Goodenough, Y.H. Huang, J. Power Sources, 2007, 173, 1– 10.
3. O.A. Marina, N.L. Canfield, J.W. Stevenson, Solid State Ionics, 2002, 149, 21–
28.
4. R.M.C. Clemmer, S.F. Corbin, Solid State Ionics, 2004, 166, 251– 259.
5. A.V. Virkar, J. Chen, C.W. Tanner, J.W. Kim, Solid State Ionics, 2000, 131, 189–
198.
6. J.H. Lee, H. Moon, H.W. Lee, J. Kim, J.D. Kim, K.H. Yoon, Solid State Ionics,
2002, 148, 15 – 26.
7. J.H. Lee, J.W. Heo, D.S. Lee, J. Kim, G.H. Kim, H.W. Lee, H.S. Songa, J.H.
Moon, Solid State Ionics, 2003, 158, 225 – 232.
8. T. Suzuki, Z. Hasan, Y. Funahashi, T. Yamaguchi, Y. Fujishiro M. Awano,
Science, 2009, 325, 852 – 855.
9. H. Koide, Y. Someya, T. Yoshida, T. Maruyama, Solid State Ionics, 2000, 132,
253 – 260.
10. A. Atkinson, S. Barnett, R.J. Gorte, J.T.S. Irvine, A.J. Mcevoy, M. Mogensen,
S.C. Singhal, J. Hohs, Nat. Mater., 2004, 3, 17 – 27.
11. H. Abe, K. Murata, T. Fukui, W.J. Moon, K. Kaneko, M. Naito, Thin Solid Films,
2006, 496, 49 – 52.
12. T. Fukui, K. Murata, S. Ohara, H. Abe, M. Naito, K. Nogi, J. Power Sources,
2004, 125, 17 – 21.
13. H. Itoh, T. Yamamoto, M. Mori, T. Horita, N. Sakai, H. Yokokawa, M. Dokiya, J.
Electrochem. Soc., 1997, 144, 641 – 646.
14. S.F. Corbin, J. Am. Ceram. Soc., 1999, 82, 1693 – 701.
15. J. Hu, Z. Lua, K. Chena, X. Huanga, N. Ai, X. Du, C. Fu, J. Wang, W. Su, J.
Membrane Sci., 2008, 318, 445 – 451.
16. J.J. Haslam, A.Q. Pham, B.W. Chung, J.F. Di Carlo, R.S. Glass, J. Am. Ceram.
Soc., 2005, 88, 513 – 518.
CH-7 Microstructure Optimization
145
17. L. Mingyi, Y. Bo, X. Jingming, C. Jing, Int. J. Hydrogen Energy, 2010, 35, 2670
– 2674.
18. D. Neagu, J.T.S. Irvine, Chem. Mater., 2010, 22, 5042 – 5053.
19. A. Sanson, P. Pinasco, E. Roncari, J. Eur. Ceram. Soc., 2008, 28, 1221 – 1226.
20. S.F. Corbin, J. Lee, X. Qiao, J. Am. Ceram. Soc., 2001, 84, 41 – 47.
21. W. P. Pan, Z. Lü, K. F. Chen, X. B. Zhu, X. Q. Huang, Y. H. Zhang, B. Wei, W.
H. Su, Fuel Cells, 2011, 11 172 – 177.
22. C.D. Savaniu, J.T.S. Irvine, J. Mater. Chem., 2009, 19, 8119 – 8128.
23. Z. Wen, Q. Wang, Q. Zhang, J. Li, Electrochem. Commun., 2007, 9, 1867 – 1872.
24. Y.Z. Jin, Y.J. Kim, C. Gao, Y.Q. Zhu, A. Huczko, M. Endo, H.W. Kroto, Carbon,
2006, 44, 724 – 729.
25. S. Zhao, C.Y. Wang, M.M. Chen, Z.Q. Shi, N. Liu, New Carbon Mater., 2010,
25, 438 – 443.
26. K. Wilgosz, X. Chen, K. Kierzek, J. Machnikowski, R. J. Kalenczuk, E.
Mijowska, Nanoscale Res. Lett., 2012, 7, 269 – 273.
27. M. Sevilla, A.B. Fuertes, Chem. Eur. J., 2009, 15, 4195 – 4203
28. M. Sevilla, A.B. Fuertes, Carbon, 2009, 47, 2281 – 2289.
29. Y. Mi, W. Hu, Y. Dan, Y. Liu, Mater. Lett., 2008, 62, 1194 – 1196.
30. G. Patrinoiu, M. Tudose, J.M.C. Moreno, R. Birjega, P. Budrugeac, R. Ene, O.
Carp, J. Solid State Chem., 2012, 186, 17 – 22.
31. Y.Z. Jin, C. Gao, W.K. Hsu, Y.Q. Zhu, A. Huczko, M. Bystrzzejewski, M. Roe,
Carbon, 2005, 43, 1944 – 1953.
32. L.H. Kao, Y.C. Chang, P.W. Hung, H.T. Lee, P.Hs Chi, Colloids and Surfaces A:
Physicochem. Eng. Aspects, 2012, 410, 170 – 177
33. J. Chen, N. Xia, T. Zhou, S. Tan, F. Jiang, Di. Yuan, Int. J. Electrochem. Sci.,
2009, 4, 1063 – 1073.
34. M. Zheng, J. Cao, X. Chang, J. Wang, J. Liu, X. Ma, Mater. Lett., 2006, 60, 2991
– 2993.
35. X. Sun, Y. Li, Angew. Chem. Int. Ed., 2004, 43, 3827 – 3831.
CH-7 Microstructure Optimization
146
36. X. Sun, J. Liu, Y. Li., Chem. Eur. J., 2006, 12, 2039 – 2047.
37. C.Y. Lee, S.J. Kim, I.S. Hwang, J.H. Lee, Sensors and Actuators B, 2009, 142,
236 – 242.
38. X. Sun, Y. Li, Angew. Chem. Int. Ed., 2004, 43, 597 – 601.
Chapter 8
Symmetrical and Button Cell Testing
147
Symmetrical and Button Cell Testing
Abstract
The present chapter deals with the results obtained using LSCTA- as anode in
electrolyte supported symmetrical and button cells because of simplicity of the design.
The effect of impregnates like CeO2 and CGO along with Ni on the performance of
symmetrical cells was investigated. It was found that co-impregnation of CeO2 and CGO
with Ni has pronounced effect in reducing the impedance of bare LSCTA- in symmetrical
cells. Further, the anode performance was tested in button cells using a three electrode set
up. From the results, it was inferred that a significant improvement in performance could
be achieved by optimizing the anode support with various impregnates both qualitatively
and quantitatively.
8.1 Introduction
Ni/YSZ cermet has been considered as the state of the art anode material due to
low cost, high electronic and ionic conductivity, excellent catalytic properties and
stability for the H2 oxidation under SOFC operation conditions [1]. However, it has some
serious drawbacks: upon redox cycling, anode degradation occurs due to large and facile
Ni to NiO volume change due to a decrease of triple phase boundary. Low tolerance to
sulphur limits the application of this anode in SOFC conditions and its high catalytic
activity causes coke formation when hydrocarbons are used as fuels, without excess
steam being present which results in a loss of cell performance. Moreover, at high
operating temperatures, the catalytic active surface area decreases due to agglomeration
and sintering of Ni [2-4]. These factors affect the working and long term stability of
SOFCs.
In quest of Ni free anodes, Cu-ceria cermets were suggested where Cu is believed
to provide electronic conductivity and ceria is the mixed conductor as well as oxidation
catalyst responsible for the electrochemical oxidation of hydrocarbons. These cermets
also offer good sulphur resistance [5, 6]. Thus, the hydrocarbons can be fed directly to the
CH-8 Symmetrical and Button Cell Testing
148
cermet. Also, Cu does not catalyze the formation of carbon fibers. But this cermet has
poor thermal stability above 700 oC [7] due to low melting point and surface energy of
Cu [8].
As alternative functional anode materials, ceramic oxides like ceria, chromite and
lanthanum titanate-based oxides have attracted a great attention. Among ceramic oxides,
donor doped strontium titanates based on the perovskite structure, for instance, La doped
strontium titanates, have been considered as potential anode candidates due to good
chemical and dimensional stability, n-type conductivity in reducing conditions and
resistance to sulphur [9, 10].
However, when compared to Ni, lanthanum doped strontium titanates have low
electronic conductivity which could result in an increase in contact resistance between
electrodes and current collector [11, 12]. Also, they have poor electro-catalytic activity
for oxidation of fuel [13, 14]. To overcome this issue, ion impregnation has been
considered as an effective approach for incorporation of nano size catalytic oxide
particles in electrode scaffolds. Infiltration of various materials to porous electrodes has
been introduced in SOFC electrodes to enhance the electro catalytic activity for better
performance [15-17]. The impregnated anodes have shown marked improvement in
performance implying that electrode reactions are catalyzed by impregnation [18-20].
Among the impregnates, ceria and doped ceria have been given special attention because
CeO2 is not only an oxidation catalyst but it also suppresses sulphur poisoning [21-24].
The performance of electrode materials can be tested in single chamber and two
chambers testing [25]. In single chamber testing, both electrodes on each side of the
electrolyte are the same and the tests are carried out in the same environment. In two
chamber testing, the anode and cathode are separated by exposing them to reducing gas
and air respectively.
For the present studies, single chamber testing was done via symmetrical cell
testing. It puts some constraints on the selection of the electrode material, for example,
the material should have acceptable conductivity in both reducing and oxidizing
environments. The material should be catalytically active towards oxygen reduction and
fuel oxidation. It should be chemically stable and compatible with other cell components.
CH-8 Symmetrical and Button Cell Testing
149
In addition, it should have sufficient porosity to allow gas diffusion to the active sites
[26].
Single testing setup is simple as the number of different components is reduced
and anode-electrolyte and cathode electrolyte interfaces are similar, helping to overcome
problems associated between cell components and their thermal mismatch. It is also
assumed that the reversal of gas flows would be beneficial to remove the carbon deposits
or sulphur contamination.
Two chamber testing was done through button cell testing. The performance of
symmetrical and button cells is usually characterized by electrical impedance
spectroscopy which is a convenient, versatile and informative technique which gives
valuable insight into the systems under investigation.
8.2 Electrochemical Impedance Spectroscopy for Symmetrical
and Button Cell Characterization
In impedance spectroscopy, a sinusoidal current perturbation, i(t) = iocos(ωt) is
imposed, onto a working cell (symmetrical or fuel) and the frequency-dependent
sinusoidal output V(t) = Vo cos(ωt+φ) is measured. The ratio of voltage V(t) and I(t) gives
impedance, Z=V(t)/i(t), across a range of frequencies. Both frequency dependent and
independent processes manifest themselves in the impedance response.
In the case of frequency-independent processes, the ratio of potential and current
is constant and both the perturbation and the output have the same phase. The frequency
independent processes of a fuel cell are associated with the ohmic losses, i.e., the
conductivity of the electrodes and electrolyte. The impedance offered is called as ohmic
resistance.
However, the frequency-dependent processes are characterized by a phase shift in
the output. These processes are associated with the non-ohmic losses, such as losses due
to the transport of gases through the porous electrodes and the reactions at the electrodes.
The impedance due to these processes is termed as polarization resistance.
The impedance data is usually represented by Cole-Cole plot (Nyquist plot) where
imaginary impedance is plotted against real impedance. Each data point reflects
measurement at a specific frequency, ω.
CH-8 Symmetrical and Button Cell Testing
150
The high frequency real-axis intercept gives the time independent resistance of
the cell known as the ohmic resistance (Rs) where as the low frequency intercept gives
the total resistance that is the sum of ohmic resistance and time dependent impedance of
the electrodes called the polarization resistance (Rp) of the cell. Thus the polarization
resistance can be calculated from the difference of high and low frequency intercepts.
One of such Cole Cole plots is shown in Fig. 8.1.
Ideally each frequency-dependent process would manifest itself in a semi-circle in
the impedance spectrum.
Z/
Rs RT
Z//
Rp
Fig. 8.1: Cole-Cole representation of impedance showing ohmic (Rs), polarization
(Rp) and total resistance (RT).
In actual practice, various processes occur in a fuel cell, thus the observed
impedance spectrum is the outcome of multiple overlapping arcs representing each of the
frequency dependent processes, making it difficult to completely separate them.
To understand the underlying processes, the impedance data is analyzed by fitting
using equivalent circuits modeling. Most of the circuit elements in the model are common
electrical elements such as resistor, capacitor and inductor. To be useful, the elements in
the model should have basis in physical electrochemistry. Each semicircle in the
impedance spectrum can be fitted to single parallel RC circuit with the diameter of
semicircle being R and peak frequency equal to 1/2πRC.
CH-8 Symmetrical and Button Cell Testing
151
8.3 Symmetrical and Button Cell Testing
Another aspect of present research was to investigate the performance of LSCTA-
in symmetrical and button cell configuration via electrical impedance spectroscopy (EIS).
The chapter discusses the results of symmetrical and button cell testing of LSCTA- (see
Fig. 8.2). Since LSCTA- lacks electrocatalytic activity for fuel oxidation, thus effect of
ceria (CeO2) and gadolinium doped ceria, Ce0.80Gd0.20O3 (CGO) was studied as both of
these are good ionic conductors and thus are expected to extend the triple phase boundary
increasing the effective electrode area [27, 28]. Enhanced catalytic and better current
collection properties were also investigated in ceria and CGO infiltrated cells co-
impregnated with NiO.
a b
Fig. 8.2: Diagrammatic representation for; a) symmetrical cell and b) button cell
configurations.
8.4 Symmetrical Cell Testing
For symmetrical cell testing, LSCTA- ink was painted on each side of the thick
electrolyte, YSZ of ~2 mm thickness. The fabrication of symmetrical cells is given in the
experimental section followed by results.
CH-8 Symmetrical and Button Cell Testing
152
8.4.1 Experimental
8.4.1.1 Fabrication of symmetrical cell
The symmetrical cell fabrication involves following steps.
8.4.1.1.1 Preparation of YSZ pellets
Dense YSZ pellets of approximately thickness of ~2 mm were obtained by
uniaxially pressing about 3.60 g of YSZ powder with pressure of 2.5 T followed by
sintering at 1500 oC for 12 hours.
8.4.1.1.2 Preparation of LSCTA- ink
For LSCTA- ink, required amount of pre-treated LSCTA- (section 7.2.2.3), graphite
and glassy carbon (as pore former) were milled down using dispersant D3005 in milling
media (acetone + zirconia balls) at 160 rpm overnight to deagglomerate the particles.
Then, the contents were emptied into beaker and a set amount of 5% PVB was added
with constant stirring. The acetone was left to evaporate to have a final ink of uniform
consistency. The recipe is tabulated in Table 8.1.
Table 8.1 Slurry recipe for LSCTA- ink
Ingredients Function Amount (g)
LSCTA- Ceramic material 7.0
Graphite Pore former 1.50
Glassy Carbon Pore former 1.50
KD1 Dispersant 0.20
Acetone To assist milling 20.0
5 wt% PVB in α-terpineol Binder 4.30
PVB = Polyvinyl butyral
8.4.1.1.3 Screen printing of LSCTA- ink on sintered YSZ pellets
In the next step, LSCTA- ink was screen printed on both sides of sintered YSZ
pellets using a screen printing machine DEK-248. To have final anodes of 20 micron
thickness, two layers of LSCT ink were screen printed on both sides of YSZ. Finally the
sintering of anodes was done at 1200 °C for 2 hours.
CH-8 Symmetrical and Button Cell Testing
153
8.4.1.1.4 Catalyst infiltration
LSCTA- lacks significant electrocatalytic activity for the anode reaction, thus
catalytically active components CeO2 and CGO catalyst were infiltrated to increase the
conductivity of LSCTA- anode backbone. The catalyst was infiltrated drop-wise into the
porous anode of the sintered symmetrical cells from aqueous solutions of Ce(NO3)3.6H2O
(99.99%, Alfa Aesar) and CGO solution. The amount of catalyst that can be added in a
single infiltration step is limited by the pore volume of the LSCTA- scaffold and the
concentration of the nitrate solution, thus multiple impregnations with heat treatments at
400 °C between infiltrations were done in order to achieve the desired weight loading.
Finally, the sample was heated at 700 °C for 1 hour. The process was repeated until
loading levels of CGO and CeO2 reached 10 wt%.
To prepare Ni - CeO2 (Ni@CeO2) and Ni – CGO (Ni@CGO) samples, first wet-
impregnation was done with CeO2 using aqueous solutions of Ce(NO3)2.
6H2O (Aldrich,
99%) and CGO (Ce0.80Gd0.20O3) solution into the LSCTA- substrates with loading levels
of 10 wt %. Then, Ni was wet-impregnated from aqueous Ni(NO3)2.6H2O. (99.9%, 5%
max Pd, Alfa Aesar) to final loadings of 3% in each case. Ni impregnation was done in
the same way as described above. The amount of infiltrated material in the different
samples was calculated by weighing the small symmetrical cells and button cells before
and after the infiltration steps.
The deposition of impregnates (caused during infiltration steps) on the edges of
all cells was removed with gentle polishing to avoid current leakage between the
electrodes.
8.4.1.1.5 Application of Au as current collector
In the last step of fabrication, gold was painted on each side in spider web form as
the current collector which was then consolidated at 900 oC for 1 hour prior to testing. A
fabricated symmetrical cell is diagrammatically shown in Fig. 8.3.
CH-8 Symmetrical and Button Cell Testing
154
Fig. 8.3: LSCTA- in electrolyte supported symmetrical cell with gold contacts.
8.4.1.2 Types of symmetrical cells studied
The following symmetrical cells were prepared and tested for impedance studies
using Solartron 1260 Hz.
Table 8.2 Symmetrical cells studied
S.No Codes Electrode Description Representation
1 A LSCTA- LSCTA-/YSZ/ LSCTA-
2 B 10 wt% CeO2 impregnated
LSCTA-
CeO2@LSCTA-/YSZ/ CeO2@LSCTA-
3 C 10 wt% CGO impregnated
LSCTA-
CGO@LSCTA-/YSZ/ CGO@LSCTA-
4 D 10 wt% CeO2 +3 wt% Ni
impregnated LSCTA-
CeO2-Ni@LSCTA-/YSZ/ CeO2-
Ni@LSCTA-
5 E 10 wt% CGO + 3wt% Ni
impregnated LSCTA-
CGO -Ni@LSCTA-/YSZ/ CGO-
Ni@LSCTA-
8.4.1.3 Symmetrical cell set up and data acquisition
The impedance of symmetrical cells with LSCTA- as backbone was characterized
by electrochemical impedance spectroscopy (EIS) in a one-atmosphere (single chamber)
CH-8 Symmetrical and Button Cell Testing
155
setup. The impedance spectra of the symmetrical cells were obtained at open circuit
voltage (OCV) with a Solartron 1260 FRA (frequency range 50 mHz to 1 kHz with 20
mV amplitude). Measurements were conducted between 650 – 850 ºC in a reducing
atmosphere (5% H2/Ar). The impedance data were analyzed with a complex non-linear
least squares fitting routine (CNLS) using equivalent circuit modeling employing Zview
software. For symmetrical cells, both identical electrodes contribute to the measured
polarization resistance so the average value of resistance offered by each electrode is
calculated by dividing the polarization resistance by two.
8.4.2 Results and Discussion
The ac impedance spectra of symmetrical cells (cell A to E) were recorded first in
air at 850 oC and then the cells were in-situ reduced at the same temperature using 5%
H2/Ar and the impedance was monitored again. The data collected is discussed below.
8.4.2.1 LSCTA- as symmetrical cell (Cell A)
Fig. 8.4 shows typical Nyquist plot for electrochemical response of symmetrical
cell A having LSCTA- as electrode backbone at 850 oC in 5% H2/Ar. Under these
conditions, a large polarization resistance could be seen which might be attributed to low
ionic conductivity and poor catalytic activity of LSCTA-. However, the low series
resistance, Rs corresponds to reasonable electronic conductivity in the back bone.
0 75 150 225 3000
-50
-100
-150
-200
Z''(
cm
2)
Z'(cm2)
Fig. 8.4: Nyquist plot of symmetrical cell A with LSCTA- electrodes in 5% H2/Ar
at 850 oC in the frequency range of 50 mHz to 1 kHz.
CH-8 Symmetrical and Button Cell Testing
156
8.4.2.2 Impregnated Cells
To improve the electro-catalytic activity of LSCTA-, impregnation was done with
catalytically active compounds. The impregnated cells were studied in air and in reducing
atmosphere, maintained by using 5% H2/Ar. The electrode polarization resistance of
various infiltrated electrodes with LSCTA- as backbone on symmetrical cells has been
characterized by electrochemical impedance spectroscopy in a one atmosphere setup.
8.4.2.2.1 Impedance characterization of all cells in air
Fig. 8.5 depicts the impedance spectra of impregnated symmetrical cells in air at
850 oC. The difference between high frequency and low frequency intercepts gives the
polarization resistance. Both Figs. 8.5 (a & b) show that impedance decreases
significantly with impregnation. Such pronounced effect could be seen by comparing the
polarization values.
The calculated polarization resistance is tabulated in Table 8.3. It can be inferred
from the table that total resistance decreases remarkably via impregnation. Both ceria and
gadolinium doped ceria (CGO) result in significant drop of polarization resistance.
0 30 60 90 120 1500
-15
-30
-45
-60
Z''(
cm
2)
Z'(cm2)
Cell B
Cell C
Cell D
Cell E
a
0 1 2 3 4 5 60
30
60
90
120
150b
Cell B
Cell C
Cell D
Cell E
Z''(
cm
2)
log f
Fig. 8.5: a) Plot of Z// vs. Z
/ in the frequency range of 50 mHz to 1 kHz and b) Z
// vs. log
f for impregnated symmetrical cells in air at 850 oC.
CH-8 Symmetrical and Button Cell Testing
157
It can be seen that the impregnation with CGO has more pronounced decrease in
the impedance value than ceria impregnation which might be due to better ionic
conductivity of CGO than ceria. The co-impregnation with Ni further improves the
performance by drastic drop in impedance in NiO impregnated cells as compared to ceria
or CGO only.
Table 8.3 EIS derived polarization resistance of impregnated
symmetrical cells in air at 850 oC
Codes Polarization resistance
(Ω cm2)
B 59.90
C 29.12
D 30.41
E 16.72
8.4.2.2.2 Impedance in 5% H2/Ar
After taking impedance at 850 °C in air, the symmetrical cells were in situ
reduced by purging 5% H2/Ar in the set up. In a reducing atmosphere, the impedance
decreased as expected for n-type semi conductors. The very first reading was taken just
after 10 min of in-situ reduction where a drastic decrease in resistance is observed as
shown in Fig. 8.6 (Cells B-D). The reaction was monitored for further 10 hours where the
decrease in the resistance was observed to be continuous.
CH-8 Symmetrical and Button Cell Testing
158
0 30 60 90 120 1500
-10
-20
-30
-40
-50
Z''(
cm
2)
Z'(cm2)
a
b
c
Cell B
0 10 20 30 40 50 600
-4
-8
-12
-16
-20
a
b
c
Cell C
Z''(
cm
2)
Z'(cm2)
0 15 30 45 60 750
-4
-8
-12
-16
-20
Cell D
Z''(
cm
2)
Z'(cm2)
a
b
c
0 8 16 24 32 400
-3
-6
-9
-12
-15
Cell E
Z''(
cm
2)
Z'(cm2)
a
b
c
Fig. 8.6: Plot of Z// vs. Z for impregnated symmetrical cells (B-D) in the frequency range
of 50 mHz to 1 kHz; a) before, b) after 10 min and c) after 10 hours of in-situ reduction
using 5% H2/Ar at 850 oC.
Furthermore, it seems that the reduction process proceeds in two stages as
discussed in chapter 6, section 6.3.4. It can be seen that the first process is pretty fast
which results in maximum drop of impedance ~ 10 min. The second process is slow
which does not cause appreciable decrease in impedance and even after 10 hours of
reduction, a slight decrease in impedance was observed (compare curves b & c in Fig. 8.6
CH-8 Symmetrical and Button Cell Testing
159
(B-D). These two stages might be related to the fast removal of oxygen from the surface
of the perovskite that is followed by a slow diffusion into the bulk of the micron size
grains. The thin electrode (~20 microns) might facilitate the fast removal of oxygen
causing a drastic decrease in the impedance [29].
To compare the performance of the symmetrical cells, the impedance of all of the
cells (Cell B – Cell E) at OCV value measured at 850 oC in a reducing atmosphere
established by purging 5% H2/Ar in the setup is shown in Fig. 8.7.
5 10 15 200
-2
-4
-6
-8
-10
Z''(
cm
2)
Z'(cm2)
Cell B
Cell C
Cell D
Cell Ea
4 6 8 100
-1
-2
-3
-4
-5
Cell B
Cell C
Cell D
Cell E
Z''(
cm
2)
Z'(cm2)
b
Fig. 8.7: a) Cole Cole plots of impregnated symmetrical in 5% H2/Ar at 850 oC cells and
b) under magnification in the frequency range of 50 mHz to 1 kHz.
CH-8 Symmetrical and Button Cell Testing
160
The values of polarization resistance calculated as the difference of high and low
frequency intercepts from these plots is tabulated in Table 8.4. The polarization resistance
of bare LSCTA- is also given for comparison. It can be observed that impregnation results
in a pronounced drop of impedance in reducing conditions, where the impedance dropped
by approximately ~ 12 times in the case of cells B & C with ceria and CGO impregnated
LSCTA- scaffolds, respectively. For co-impregnated cells (D & E), the drop in impedance
is ~ 75 times.
Table 8.4 Polarization resistance of symmetrical cells in 5% H2/Ar at 850 oC
Codes Rp
(Ω cm2)
A 162 10
B 12.30
C 13.82
D 2.28
E 1.92
Recently, Zhangbu [30] has demonstrated that impregnation with ceria and doped
ceria significantly improves the anode performance mainly due to the catalytic activity.
Both ceria and CGO are good oxidation catalysts due to the presence of CeIV/CeIII couple
under high temperature reducing conditions. In fact, the catalytic activity of ceria is
dependent on the activity of redox couple CeIV/CeIII which is formed from partial
reduction of CeO2 to Ce2O3 under reducing conditions.
Furthermore, the cells with Ni co-impregnation exhibit better performance
compared to those with CeO2 or CGO only under same operating conditions. It might be
attributed to high catalytic activity of Ni towards oxidation of fuel [15, 31]. The synergic
effect makes co-impregnation a better strategy to lower the electrode polarization
resistance of symmetrical cells with LSCTA- as backbone.
Thus it can be said that these impregnates improve the anode performance by
extending the triple phase boundary of the anode which actually increases the electrode
CH-8 Symmetrical and Button Cell Testing
161
effective area. Also, the improvement in performance could be attributed to an increase
in electro-catalytic activity for fuel oxidation with these impregnates [32].
Generalizing the results, it can be concluded that pronounced improvement in
performance could be achieved by careful selection of impregnates.
8.4.2.2.3 Effect of temperature on impedance
The effect of temperature was also studied in the impedance behavior of the
symmetrical cells. Fig. 8.8 shows that the impedance increases with a decrease in
temperature, indicating negative temperature coefficient of resistance behavior (NTCR).
On decreasing the temperature from 850oC to 650
oC, the impedance increases as shown
in Fig. 8.8 (a-d).
CH-8 Symmetrical and Button Cell Testing
162
0 100 200 300 400 500 6000
-50
-100
-150
-200
-250
-300
-350
Z''
cm
2)
Z'( cm2)
650oC
700oC
750oC
800oC
850oC
Cell Ba
0 35 70 105 140 1750
-40
-80
-120
-160
-200b
650oC
700oC
750oC
800oC
850oC
Z''(
cm
2)
Z'( cm2)
Cell C
10 20 30 40 50
0
-5
-10
-15
-20
-25
650oC
700oC
750oC
800oC
850oC
c
Z''(
cm
2)
Z'( cm2)
Cell D
0 10 20 30 40 500
-4
-8
-12
-16
-20
650oC
700oC
750oC
800oC
850oC
bZ
''(
cm
2)
Z'( cm2)
Cell E
Fig. 8.8: Nyquist plots for impregnated symmetrical cells at different temperatures in
reducing atmosphere (5% H2/Ar) in the frequency range of 50 mHz to 1 kHz.
8.4.2.2.4 Analysis of impedance
Among different presentation formats of impedance, the Bode plot is quite
informative as it gives information of different frequency dependent processes. To
illustrate the number of underlying processes, the impedance of impregnated cells in
represented in Bode plot form as shown in Fig. 8.9.
CH-8 Symmetrical and Button Cell Testing
163
-2 -1 0 1 2 3 4 5 60
-2
-4
-6
-8
-10
-12
Z''
(cm
2)
log f
Cell B
Cell C
Cell D
Cell E
Fig. 8.9: Dependence of Z// on frequency for impregnated symmetrical cells in reducing
atmosphere ( 5% H2/Ar) at 850 oC.
Fig. 8.9 shows a broad peak in Z// vs. log f, reflecting a distribution of frequency
dependent processes in the symmetrical cells [33]. In the case of co-impregnated cells (D
& E), two peaks appear which shows that at least two processes are responsible for the
impedance observed. To get a better understanding, the impedance responses measured at
open circuit voltage in 5% H2/Ar were deconvoluted by Z view software using equivalent
circuit fitting having series combination of parallel RQ elements. Each RQ element
defines the resistance (R) parallel with constant phase element CPE of an electrochemical
process. The constant phase element is defined by;
n
CPE oY Y j (8.1)
Where Yo and n are frequency independent parameters and ω is the angular
frequency. The overall polarization resistance is obtained by addition of the individual
resistances of different processes.
Symmetrical cells impregnated with ceria or CGO have been represented by
equivalent circuit, LRs (RQ)1 (RQ)2 [34]. In this circuit, L is the inductance of the wires,
CH-8 Symmetrical and Button Cell Testing
164
Rs is the ohmic resistance having a contribution from the electrolyte and electrodes and Q
depicts constant phase element. Two RQ circuit elements show that two electrochemical
processes govern the observed behavior. In the case of doubly impregnated cells, (RQ)1
(RQ)2 (RQ)3 shows three electrochemical processes at high, medium and low frequency
that are responsible for the impedance [26, 35]. The experimental and simulated graphs
are shown in Fig 8.10, with the corresponding equivalent circuits used for simulating the
data.
4 8 12 16 200
-2
-4
-6
-8
-10
Z''(
cm
2)
Z'(cm2)
Experimental
Simulated
a
4 8 12 16 20 240
-4
-8
-12
-16
Z''(
cm
2)
Z'(cm2)
Experimental
Simulatedb
5 6 7 80.0
-0.5
-1.0
-1.5
-2.0
Z''(
cm
2)
Z'(cm2)
Experimental
Simulatedc
5 6 7 8 90.0
-0.5
-1.0
-1.5
-2.0
Z''(
cm
2)
Z'(cm2)
Experimental
Simulatedd
Fig. 8.10: Experimental and simulated impedance spectra of symmetrical cells in
reducing atmosphere (5%H2/Ar) at 850 oC; a) cell B, b) cell C, c) cell D and d) cell E in
the frequency range of 50 mHz to 1 kHz.
CH-8 Symmetrical and Button Cell Testing
165
Good fitting as seen in Fig. 8.10 shows that the impedance of symmetrical cells B
& C can be analyzed with an equivalent circuit consisting of two processes at low and
high frequencies, while in the case of co-impregnated cells (D & E), the impedance is
analyzed by three processes at low, medium and high frequencies. The good fit also
shows the applicability of these equivalent circuit models. These models remain valid for
all the investigated temperatures. The extracted values of resistances for each cell are
plotted against temperature in Fig. 8.11 which shows that all these processes are
thermally activated as interfacial resistance decreases with increase in temperature.
650 700 750 800 850
0
100
200
300
400
500
Re
sis
tan
ce
(
cm
2)
Temp (oC)
Rs
R1
R2
a
650 700 750 800 850
0
150
300
450
600
750b
Re
sis
tan
ce
(
cm
2)
Temp (oC)
Rs
R1
R2
650 700 750 800 850
0
10
20
30
40
50
Re
sis
tan
ce
(
cm
2)
Temp (oC)
Rs
R1
R2
R3
c
650 700 750 800 8500
7
14
21
28
35
Re
sis
tan
ce
(
cm
2)
Temp (oC)
Rs
R1
R2
R3
d
Fig. 8.11: Variation of resistances extracted from fit models with temperature (in 5%
H2/Ar) for; a) cell B, b) cell C, c) cell D and d) cell E.
CH-8 Symmetrical and Button Cell Testing
166
The values of resistivity at different temperatures for each cell are given in
appendix A8.
8.5 Button Cell Testing
From the results of symmetrical cell testing, it was observed that the co-
impregnation of ceria and CGO with Ni resulted in significant decrease in polarization
resistance. With this inference, NiO co-impregnated cell configuration was further used
in button cell using three electrodes setup. The fabrication of button cells is given in
experimental section followed by results and discussion.
8.5.1 Experimental
8.5.1.1 Fabrication of button cell
The fabrication of button cells involves the following steps.
8.5.1.1.1 Preparation of YSZ pellets and LSCTA- ink
The synthetic procedure has already been discussed in the experimental section of
symmetrical cell testing (8.4.1.1).
8.5.1.1.2 Preparation of LSM and LSM-YSZ inks
LSM (La0.8Sr0.2MnO3) is mostly used as a cathode material for SOFCs as it has
good catalytic activity for the dissociation of oxygen to O2 anions and is also a good
electronic conductor. It is often mixed with YSZ to form the composite, LSM-YSZ, for
good thermo-mechanical stability. Both LSM and LSM-YSZ inks were prepared by
following the procedure discussed earlier (section 8.4.1.1.2). The slurry recipes are given
in Table 8.5.
8.5.1.1.3 Screen printing of LSCTA- and LSM and LSM-YSZ inks on sintered
YSZ pellets
The button cells were prepared by screen printing both cathode and anode. First
LSCTA- ink was screen printed on one face of sintered YSZ pellet followed by sintering
at 1200 oC. After screen printing and sintering the anode side, the cathode side was first
screen printed with a layer of LSM-YSZ ink while LSM which itself is a good electronic
conductor was applied as a current collector to form bilayer ink (LSM-YSZ׀LSM).
Finally the cell was sintered at 1100 oC for 2 hours to form the required perovskite phase.
CH-8 Symmetrical and Button Cell Testing
167
Table 8.5 Slurry recipes for LSM and LSM/YSZ inks
Ingredients Function LSM ink
(g)
LSM-YSZ ink
(g)
LSM Ceramic 7.544 3.772
YSZ (8 mol% Pi-
Kem)
Composite ceramic
with LSM ---------- 3.772
Graphite Pore former 1.406 1.406
Glassy Carbon Pore former 1.647 1.647
KD1 Dispersant 0.202 0.202
Acetone To assist milling ~20 ~20
5 wt% PVB in
terpineol
Binder 4.310 4.310
8.5.1.1.4 Preparation of impregnated button cells
Wet impregnation was done in the same way as has been explained before in
symmetrical cell fabrication. Both the cells were impregnated with 10 wt% CeO2 or CGO
along with 3 wt% of Ni. Au was used as the current collector on LSCTA- face while Pt
was used as the reference electrode (as shown in Fig. 8.12).
Fig. 8.12: Diagrammatic presentation of fabricated button cells using LSCTA- as
anode support.
CH-8 Symmetrical and Button Cell Testing
168
8.5.1.1.4 Types of button cells studied
The following table lists the tested cells:
Table 8.6 Types of button cells studied
Codes Anode Cathode
A Ni-CeO2@LSCTA- LSM-YSZ׀LSM
B Ni-CGO@LSCTA- LSM-YSZ׀LSM
8.5.1.2 Button cell set up
After the fabrication, the button cells were sealed on the anode side to an alumina
tube using ceramic adhesive (Ceramabond 552-VFG, Aremco). Then the testing
assembly was positioned inside a furnace with a programmable temperature controller. In
order to cure the adhesive, the testing assembly was heated to 93 °C and 260 °C and held
at each temperature for 2 hours. After the drying of adhesive, the assembly was
completed by supplying oxygen to the cathode by exposing it to ambient air and reducing
atmosphere to anode. A water bubbler at room temperature was incorporated into the fuel
line to raise the humidity level of the fuel to approximately 3% H2O. The fuel flow rate
was regulated by a flow meter and maintained well above 20 mL/min at all times to
ensure low conversion in the anode. Similarly, the air flow was regulated by flow meter.
Measurements were taken between 750 and 850 °C at 50 °C intervals.
8.5.1.3 Electrical characterization and data acquisition of button cells
The electrochemical characterization of button cells is usually done by measuring
the polarization curve (current-voltage I-V characteristics) and impedance spectrum.
The I-V curves show the voltage output of the fuel cell as a function of the current
density drawn. Fuel cells ideally operate at the Nernst potential, but irreversible processes
within the electrodes and electrolyte result in drop of cell potential. For IV
measurements, a voltage range (from the open circuit voltage value to 0.1 V) with a set
ramp rate (10-50 mV s-1
) is applied between the working and the reference electrode. The
deviation of the I-V curve from the OCV gives the total loss in the system, consisting of
ohmic and non-ohmic losses such as the activation of the electrochemical reactions, the
CH-8 Symmetrical and Button Cell Testing
169
ohmic resistance resulting from the ion conductivity through the electrolyte or mass
transfer losses due to gas diffusion. The slope of the I-V curve gives the cumulative
impedance due to these processes.
Although I-V curves provide insight into fuel cell performance, they cannot be
used to differentiate between the various sources of loss within the cell. Electrochemical
impedance spectroscopy (EIS) is a powerful technique that helps in identifying such
different processes.
For button cell measurements, the working electrode is connected to the anode
while the reference electrode is connected to the cathode of the fuel cell. For IV
measurements, the voltage was applied to the anode using a Solartron Electrochemical
Interface (model 1287). CorrWare software was used for input of experimental details
where as CorrView was used for viewing the output results. The ramp rate for the I-V
measurements was 10 mV s-1
. Measurements were taken between 750 and 850 °C at 50
°C intervals. The data obtained from these measurements was used to generate a I-V plot.
For impedance measurements a Solartron Frequency Response Analyser 1255
was used in combination with a Solartron Electrochemical Interface 1287. Impedance
was recorded in the range of 1 MHz to 1 mHz with amplitude of 10 mV. Measurements
were performed using Zplot®, a commercially available computer program for
impedance measurements. These EIS measurements were done under open circuit
conditions and at biased conditions, as well.
Redox cycling was also done to investigate the redox stability of the button cells
at 850 oC. Redox cycling was done by changing the reducing atmosphere (humidified H2)
to completely oxidizing by cutting off the flow of H2 at the anode. The anode was
exposed to the oxidizing atmosphere for ~1 hour and then H2 was again purged in. To
investigate the effect of redox cycling, the impedance was monitored after changing the
gases at OCV conditions.
CH-8 Symmetrical and Button Cell Testing
170
8.5.2 Results and Discussion
To investigate the anode performance in the button cells A and B, they were
tested under different conditions which are briefly discussed below:
8.5.2.1 OCV values in reducing conditions at 850 oC
The prepared button cells were exposed to gases at anode and cathode and OCV
was monitored. To the cathode, air was supplied while anode was exposed to reducing
atmosphere from 5% H2/Ar to pure H2 sources. To check the sealing of the button cell,
OCV values were determined and are tabulated in Table 8.7.
Table 8.7 OCV values for button cells A & B at different conditions at 850 oC
Gas at Anode Gas at Cathode -OCV (V)
Button Cell A Button Cell B
Dry 5% H2/Ar Air 0.97 0.950
Dry H2(g) Air 1.17 1.147
Humidified H2(g) Air 1.07 1.084
The OCV values show well sealed samples as the OCV is close to Nernst predicted
voltage.
8.5.2.2 Button cell performance in reducing conditions at 850 oC
The impedance spectra of button cells A and B with impregnated LSCTA- as
anode and LSM as cathode are shown in Fig 8.13. All the data were collected at 850 oC
under OCV conditions. Mainly two arcs could be seen in impedance plots of both button
cells. The high frequency intercept gives the ohmic resistance which arises from the ion
conducting resistance of the electrolyte and interface resistance between the electrolyte
and the electrode. Interfacial (polarization) resistance is obtained from the difference of
high and low frequency intercepts on x-axis. In the present case, this is the anode
resistance due to the three electrode set up used [30].
CH-8 Symmetrical and Button Cell Testing
171
2 3 4 50.0
-0.4
-0.8
-1.2
-1.6
-2.0
Z''(
cm
2)
Z'(cm2)
5% H2/Ar
Dry H2
Wet H2
Galvanostatic
a
2.0 2.4 2.8 3.20.0
-0.2
-0.4
-0.6
-0.8
5% H2/Ar
Dry H2
Wet H2
Galvanostatic
Z''(
cm
2)
Z'(cm2)
b
Fig. 8.13: Impedance spectra of button cells under different conditions
at 850 oC in the frequency range of 0.1 Hz to 1 MHz; a) cell A and b) cell B.
CH-8 Symmetrical and Button Cell Testing
172
Values of polarization resistance extracted from above EIS spectra are tabulated
in Table 8.8.
Table 8.8 Polarization resistances extracted from EIS spectra of button cells under
different conditions
Fuel at anode Rp (Ω cm
2)
Button Cell A Button Cell B
5% H2/Ar 3.486 1.381
Dry H2 2.218 1.110
Wet H2 1.905 1.054
*After Galvanostatic treatment 1.758 1.036
* after passing constant current of 150 mA for 20 min at OCV
The cell impedance decreases as the fuel at the anode is changed from 5% H2/Ar
to pure H2. The impedance is less in humidified hydrogen then in dry hydrogen. Better
reducing conditions help in decreasing the impedance values. The galvanostatic treatment
further decreases the polarization resistance. Both short and long time scale polarization
of electrodes via constant current load affect the impedance. Usually, short scale current
load causes the process activation decreasing the impedance at OCV than before [36].
However, the electrode activation also depends on measurement conditions. In the
present case, a constant current of 150 mA was applied at OCV for 20 min and then the
impedance was monitored. The electrode process seems to be activated where decrease in
impedance was noted. It could be attributed to the availability of new reaction sites under
measurement condition [37].
By comparing button cells A and B, it is inferred that cells co-impregnated with
Ni-CGO offered less resistance then Ni-CeO2 under same experimental conditions. The
performance of button cells A and B under different conditions at 850 oC is shown in the
form of IV curves in Fig. 8.14.
It was observed that maximum power is delivered after short time scale
galvanostatic polarization of the electrode. These results are at par with impedance
graphs where least impedance was observed upon galvanostatic treatment.
CH-8 Symmetrical and Button Cell Testing
173
0.00 0.04 0.08 0.12 0.16 0.20
-0.60
-0.75
-0.90
-1.05
-1.20
0
30
60
90
120
Ce
ll V
olt
ag
e (
V)
Current density (Acm-2)
Po
we
r d
en
sit
y (
mW
cm
-2)
Wet H2
Dry H2
Galvanostatic
a
0.00 0.04 0.08 0.12 0.16 0.20
-0.60
-0.75
-0.90
-1.05
-1.20
0
20
40
60
80
100
120
Po
we
r d
en
sit
y (
mW
cm
-2)
Ce
ll V
olt
ag
e (
V)
Current density (Acm-2)
Dry H2
Wet H2
Galvano static
b
Fig. 8.14: Plots of cell potential and power density as a function of current density for
button cells under different conditions at 850 oC; a) cell A and b) cell B.
CH-8 Symmetrical and Button Cell Testing
174
8.5.2.3 Button cells performance in wet H2 at 850 oC at different
temperatures
The performance of button cells was evaluated while cooling down from 850 oC.
OCV values found at these temperatures are tabulated in Table 8.9. The OCV values
were found to closely match to that predicted by the Nernst equation, indicating well
sealed button cells at all temperatures. The near ideal OCV values also suggest the
mechanical and chemical stability of button cells at different temperatures.
Table 8.9 Values of OCV (V) for button cells at different temperatures with humidified
H2 as fuel at anode and air at cathode
Temp. OCV (V)
Button Cell A Button Cell B
750 oC -1.064 -1.152
800 oC -1.108 -1.157
850 oC -1.154 -1.162
Once the working temperature, 850 oC was achieved, the cell system was
equilibrated for half an hour and the impedance was recorded while cooling down the
button cells in humidified H2. It is seen that impedance decreases while decreasing the
temperature as shown in Fig. 8.15. The cell becomes more resistive as temperature is
lowered. However, it can be noted that the cell B seemed to be less resistive than cell A at
all temperatures in accordance with the results found in symmetrical cell testing where
better performance was achieved in cell co-impregnated with CGO and NiO.
CH-8 Symmetrical and Button Cell Testing
175
2 4 6 8 10 120
-1
-2
-3
-4
-5
-6
Z''(
cm
2)
Z'(cm2)
800oC
750oC
700oC
a
2 3 4 5 6 70.0
-0.4
-0.8
-1.2
-1.6
-2.0b
800oC
750oC
700oC
Z''(
cm
2)
Z'(cm2)
Fig. 8.15: Impedance spectra of button cells under OCV at different temperatures with
humidified H2 at anode and air at cathode in the frequency range of 0.1 Hz to 1 MHz; a)
cell A and b) cell B.
EIS results also point to enhanced activation in CGO-based cells than ceria and
show less impedances under similar cell testing conditions. From the impedance data, the
values of ohmic, polarization and total resistance were determined and plotted as a
function of temperature in Fig. 8.16. The Arrhenius type dependence allows to calculate
the activation energy for all of these processes in both the cells.
CH-8 Symmetrical and Button Cell Testing
176
0.88 0.90 0.92 0.94 0.96 0.98
0.5
1.0
1.5
2.0
2.5
3.0
ln R
(
cm
2)
1000/T (K-1
)
Rs
Rp
Rt
a
0.88 0.90 0.92 0.94 0.96 0.98
0.4
0.8
1.2
1.6
2.0
b
ln R
(
cm
2)
1000/T (K-1
)
Rs
Rp
Rt
Fig. 8.16: The total resistance Rt, the polarization resistance Rp, and the ohmic resistance
Rs of button cells determined from impedance plots at different temperatures; a) cell A
and b) cell B.
The resistance vs. temperature plots clearly differentiate the performance of cells,
A and B, whereby the low intercept in graph b imparts better conductivity to cell B. The
calculated energy of activation for all types of resistances in both cells, A and B is
tabulated in Table 8.10.
CH-8 Symmetrical and Button Cell Testing
177
Table 8.10 Energy of activation, Ea calculated from resistance-temperature plots
Type of
resistance
Ea (eV)
Button Cell A Button Cell B
Rs 0.590 0.613
Rp 1.925 0.931
Rt 1.432 0.742
It can be seen that the activation energy in the case of cell B is less than that of
cell A which explains the observed better performance of cell B as compared to cell A.
The significant difference of activation energy also manifests itself in power density
curves. Fig. 8.17 gives polarization curves typical of button cells at different temperatures
with humidified H2 as fuel at anode and air at cathode.
The power density increases with temperature while both of the cells deliver
maximum power at 850 oC. The button cell B offered higher power density than A and by
all studies, it is proved a better configuration than cell A.
CH-8 Symmetrical and Button Cell Testing
178
0.000 0.045 0.090 0.135 0.180
-0.60
-0.75
-0.90
-1.05
-1.20
0
25
50
75
100
Po
we
r D
en
sit
y (
mW
cm
-2)
Ce
ll V
olt
ag
e (
V)
Current Density (Acm-2
)
750oC
800oC
850oC
a
0.00 0.05 0.10 0.15 0.20 0.25
-0.4
-0.6
-0.8
-1.0
-1.2
0
20
40
60
80
100
Ce
ll V
olt
ag
e (
V)
Po
we
r d
en
sit
y (
mW
cm
-2)
Current Density (Acm-2
)
750oC
800oC
850oC
b
Fig. 8.17: Plots of cell potential and power density as a function of current density for
button cells at different temperatures with humid H2 as fuel at anode and air at cathode; a)
Cell A and b) cell B.
8.5.2.4 Redox cycling at 850 oC
Upon redox cycling at 850 oC, it was found that OCV remained stable over a
couple of cycles showing good stability of the tested systems. The stable OCV values
CH-8 Symmetrical and Button Cell Testing
179
confirmed that the cells remained intact after redox cycling at high temperature (Fig.
8.18) owing to mechanical and chemical stability, as well.
1 2 3 4 5-1.20
-1.16
-1.12
-1.08
-1.04
-1.00
OC
V (
V)
No. of cycles
a
1 2 3 4 5-1.20
-1.18
-1.16
-1.14
-1.12
-1.10
b
OC
V (
V)
No. of cycles
Fig. 8.18: Plots of OCV as a function of number of redox cycles for button cells; a) cell
A and b) cell B.
CH-8 Symmetrical and Button Cell Testing
180
Corresponding graphs of evolution of Rs and Rp with redox cycling is given in
Fig. 8.19 where it was found that both ohmic and polarization resistances were decreased
upon redox cycling [38].
0 1 2 3 4 5 6
2.5
2.6
2.7
2.8
2.9
3.0
Rs
Rp
No. of Cycles
2.5
2.6
2.7
2.8
2.9
3.0
Rp
cm
2)
Rs (
cm
2)
a
0 1 2 3 4 5 62.5
2.6
2.7
2.8
2.9
3.0
Rp
(
cm
2)
Rs
Rp
No. of Cycles
Rs (
cm
2)
b
1.5
1.6
1.7
1.8
1.9
2.0
Fig. 8.19: Plots of Rs and Rp a function of number of redox cycles for button cells; a)
cell A and b) cell B.
CH-8 Symmetrical and Button Cell Testing
181
8.4.2.5 Impedance analysis by model fitting
In reducing conditions with humid H2 as the fuel at the anode and air at the
cathode, at least two arcs are observed which were fitted with an equivalent circuit
represented as L Rs (RQ)1 (RQ)2 (RQ)3 showing that the observed spectra could be
attributed to three electrochemical processes as shown in Fig. 8.20.
2.0 2.5 3.0 3.5 4.0 4.50.0
-0.5
-1.0
-1.5
-2.0 a
Z'' (
cm
2)
Z''(cm2)
Experimental
Simulated
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.20.0
-0.3
-0.6
-0.9
-1.2
Z'' (
cm
2)
Z''(cm2)
Experimental
Simulated
b
Fig. 8.20: Experimental and simulated impedance spectra of button cells in reducing
atmosphere (humid H2) at 850 oC in the frequency range of 0.1 Hz to 1 MHz; a) cell A
and b) cell B.
CH-8 Symmetrical and Button Cell Testing
182
The non-ohmic part from the curve fitting of EIS data reveals three frequency
dependent processes at low, medium and high frequency, (RQ)1 (RQ)2 (RQ)3. The fit
obtained using the model is in good agreement with the experimental curves which shows
the applicability of the model. The same model remained valid for all the temperatures
where the button cells showed NTCR behavior with temperature. The extracted values of
resistances from the fit data at different temperatures are given in appendix 8A-II.
8.5.2.6 SEM imaging after cell test
The SEM technique is quite useful in characterizing the materials as well as
visualizing the after effects of cell testing in fuel cell technology. Both the tested cells
were fractured and imaged by SEM-5600.
Fig. 8.21: Micrographs of button cells; a) cell A and b) cell B after cell tests at 850 oC.
CH-8 Symmetrical and Button Cell Testing
183
Fig. 8.21 shows the anode electrolyte interface after cell testing of both the button
cells. The examination of interface reveals good adhesion between anode and the
electrolyte, YSZ after test. The anode microstructure shows limited porosity. It is
expected that better performance could be achieved by microstructure optimization.
8.6 Conclusions
The poor electrocatalytic activity of LSCTA- was modified by catalytically active
components like ceria and CGO. The addition of these impregnates lowers the
polarization resistance significantly. The co-impregnation with Ni is an effective
approach to drastically decrease the impedance. The symmetrical cells co-impregnated
with CGO-Ni offered less resistance and better performance than that of CeO2-Ni cells
and the same configuration was found to have improved performance in button cells, as
well. It was also concluded that our prepared impregnated anode support is more
workable at higher temperatures. The performance could be enhanced further by
optimizing the microstructure of anode as well as the quantity/quality of impregnates.
CH-8 Symmetrical and Button Cell Testing
184
REFERENCES
1. N.Q. Minh, J Am. Chem. Soc. 1993, 76, 563 – 588.
2. H. Kim, S. Park, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc., 2001, 148, A693 –
A695.
3. Y. Matsuzaki, I. Yasuda, Solid State Ionics, 2000, 132, 261 – 269.
4. D. Waldbillig, A. Wood, D.G. Ivey, Solid State Ionics, 2005, 176, 847 – 859.
5. S. McIntosh, R.J. Gorte, Chem. Rev., 2004, 104, 4845 – 4865.
6. S. Park, J.M. Vohs, R.J. Gorte, Nature, 2004, 404, 265 – 267.
7. M.D. Gross, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc., 2006, 153, A1386 –
1390.
8. K.B. Yoo, B.H. Park, G. M. Choi, Solid State Ionics, 2012, 225, 104 – 107.
9. K. Ahn, S. Jung, J.M. Vohs, R.J. Gorte, Ceram. Int., 2007, 33, 1065 – 1070.
10. P.R. Slater, D.P. Fagg, J.T.S. Irvine, J. Mater. Chem., 1997, 7, 2495 – 2498.
11. S.P. Jiang, X.J. Chen, S.H. Chan, J.T. Kwok, K.A. Khor, Solid State Ionics, 2006,
177, 149-157.
12. K.P. Ong, P. Wu, L. Liu, S.P. Jiang. Appl. Phys. Lett., 2007, 90. 044109 (1-3).
13. J.C. Vazquez, S.W. Tao, J.T.S. Irvine, Solid State Ionics, 2003, 159, 159 – 165.
14. O.A. Marina, N.L. Canfield, J.W. Stevenson, Solid State Ionics, 2002, 149, 21 –
28.
15. A. Babaei, S.P. Jiang, J. Li, J. Electrochem. Soc., 2009, 156, B1022 – B1029.
16. X.C. Lu, J.H. Zhu, Solid State Ionics, 2001, 178, 1467 – 1475.
17. J.W. Yun, S.P. Yoon, Han J, Park S, Kim HS, Nam SW. J. Electrochem., Soc.,
2010, 157, B1825 – B1830.
18. S.P. Jiang, X.J. Chen, S.H. Chan, J.T. Kwok, J. Electrochem. Soc., 2006, 153,
A850 – A856.
19. X.J. Chen, Q.L. Liu, S.H. Chan, N.P. Brandon, K.A. Khor, Electrochem.
Commun., 2007, 9, 767 – 762.
20. H. Kurokawa, T.Z. Sholklapper, C.P. Jacobson, L.C. De Jonghe, S.J. Visco,
Electrochem. Solid-State Lett., 2007, 10, B135 – B138.
21. H.P. He, R.J Gorte, J.M. Vohs, Electrochem. Solid State Lett., 2005, 8, A279 –
A280.
CH-8 Symmetrical and Button Cell Testing
185
22. R.J. Gorte, S. Park, J.M. Vohs, C.H. Wang, Adv. Mater., 2000, 12, 1465 – 1469.
23. X. Wang, R.J. Gorte, Catal. Lett., 2001, 73, 15 – 19.
24. T. Suzuki, H.I. Iwanami, T. Yoshinari, Int. J. Hydrogen Energy, 2000, 25, 119 –
126.
25. K.B. Yoo, G.M. Choi, Solid State Ionics, 2009, 180, 867 – 871.
26. J.C.R. Morales, J.C. V´azquez , D.M. L´opez, D.P. Coll, J.P. Mart´ınez, P.
N´u˜nez, J. Power Sources, 2008, 177, 154 – 160.
27. W. Zhu, D. Ding, C.R. Xia, Electrochem. Solid State Lett., 2008, 11, B83 – B86.
28. D. Ding, Z.B. Liu, L. Li, C.R. Xia, Electrochem. Commun., 2008, 10, 1295 –
1298.
29. D. Neagu, J.T.S. Irvine, Chem. Mater., 2010, 22, 5042 – 5053.
30. Z. Liu, D. Ding, B. Liu, W. Guo, W. Wang, C. Xia, J. Power Sources, 2011, 196,
8561 – 8567.
31. B.B. He, L. Zhao, W.D. Wang, F.L. Chen, C.R. Xia, Electrochem. Commun.,
2011, 13, 194 – 196.
32. S.P. Jiang, Mater. Sci. Eng. A., 2006, 418, 199 – 210.
33. K. Verma, S. Sharma, Phys. Status Solidi B., 2012, 249, 209 – 216.
34. L.L. Zheng, X. Wang, L. Zhang, J.Y. Wang, S.P. Jiang, Int. J. Hydrogen Energy,
2012, 37, 10299 – 10310.
35. P. Blennow, K.K. Hansen, L.R. Wallenberg, M. Mogensen, ECS Trans., 2008, 13,
181 – 194.
36. M.K. Bruce, M. van den Bossche, S. McIntosh J. Electrochem. Soc., 2008, 155,
B1202 – B1209.
37. M.J. Jorgensen, M. Mogensen, J. Electrochem. Soc., 2001, 148, A433 – A442.
38. M.C. Verbraeken, B. Iwanschitz, A. Mai, J.T.S. Irvine, J. Electrochem. Soc.,
2012, 159, F757 – F762.
Chapter 9
Synthesis and Characterization of Doped
Analogues of LSCTA-
After Reduction
Ex-solved Particles
186
Synthesis and Characterization of Doped
Analogues of LSCTA-
Abstract
Doping at both (A and/or B sites) in the perovskites has always remained a
strategy to tailor the properties. The common B-site dopants used to improve the
properties of strontium titanate include Fe, Ni, Cr etc. In the present research project,
Fe and Ni were doped at 1% and 5% at B-site of A-site deficient lanthanum doped
strontium titanate, LSCTA-. The doped compositions were synthesized by the Pechini
method and characterized by XRD, SEM, dilatometry and conductivity. The doped
analogues have the same orthorhombic symmetry as the parent; however an
expansion in the unit cell volume was observed, which is in accordance with the ionic
sizes of the dopants. The doped compositions offered higher conductivity values than
LSCTA-.
9.1 Introduction
SrTiO3 is a typical perovskite that has been extensively studied as anode
template for solid oxide fuel cell (SOFC) because of its n-type nature in reducing
conditions. Both A and/or B sites of the strontium titanate have been doped to tune the
properties of parent compound as interesting defect chemistry is achieved by partial or
full substitution with alio-valent cations. The careful selection of dopants is essential
to avoid mismatch in the crystal system in order to keep the stoichiometry intact.
Special attention has been given to enhance the electrical conductivity of SrTiO3 by
partial substitution of Sr+2
on A-site and/or Ti+4
on B-site to yield interesting
compounds with excess or substoichiometric oxygen that largely affects the properties
[1-5].
The nature of B-site dopants affects structure, redox properties, conductivity
and electro-catalytic properties of the parent compound [6]. In this respect, various B-
site dopants have been investigated such as Nb [7], Mn [8], Ga [9], Sc [10], Fe [11],
Al and Cr [12]. Good conductivity value was found for Nb doped SrTiO3, for
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
187
example, SrTi0.98Nb0.02O3-δ presents conductivity value of 339 S cm-1
at 800 oC after
being reduced in hydrogen at 1400 °C [13].
As a last part of the study, LSCTA- was doped at the B-site to enhance its
conductivity. Fe and Ni have good catalytic activity so they were chosen as B-site
dopants. Additionally, the doping level (1% and 5%) was kept low to have good
solubility and preservation of dense network of Ti on B-site of perovskite. The doped
compositions were synthesized, characterized and investigated for dc conductivity.
9.2 Experimental
The Pechini method was used to synthesize doped analogues of LSCTA-.
Briefly, stoichiometric amount of iron nitrate (Aldrich, 99%) or nickel nitrate
(Aldrich, 99%) was dissolved in an aqueous solution containing stoichiometric
amounts of lanthanum nitrate (Aldrich, 99%), strontium nitrate (Aldrich, > 99%),
calcium nitrate (Aldrich, 99%) and titanium (IV)-bis-(ammoniumlactato) dihydroxide,
50% w/w in water (Aldrich, 99%). A solution of ethylene glycol and citric acid (both
Sigma) was added to the above solution to have a final molar ratio of metal ions to
citric acid to ethylene glycol as 1:4:16. The resulting solution was heated on the hot
plate at 80-100 oC. The resulting gel was dried and was calcined in air for 5 hours.
Room temperature powder X-ray diffraction (XRD) was performed on a Philips
XRD diffractometer using Cu-Kα1 radiation in the 2θ range of 20o to 80
o in the
reflection mode. The morphology of the calcined powders was studied using a JEOL
6700F field emission microscope. Sinterability of doped analogues was investigated
using a Netzch DIL 402C instrument. van der Pauw method was used to measure dc
conductivity of the synthesized samples.
Table 9.1 lists the investigated doped analogues of LSCTA-.
Table 9.1 Studied doped analogues of LSCTA-
Doped analogues Codes
La0.2 Sr0.25 Ca0.45 Ti0.99 Ni0.01 O3 LSCTN1
La0.2 Sr0.25 Ca0.45 Ti0.95 Ni0.05 O3 LSCTN5
La0.2 Sr0.25 Ca0.45 Ti0.99 Fe0.01 O3 LSCTF1
La0.2 Sr0.25 Ca0.45 Ti0.95 Fe0.05 O3 LSCTF5
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
188
9.3 Results and Discussion
9.3.1 X-ray diffraction (XRD)
Room temperature XRD of as prepared doped analogues show characteristic
reflections of perovskite crystal structure shown in Fig. 9.1. Similar XRD pattern was
observed in all compositions and no impurity peak was detected in any of X-ray
diffraction patterns showing full solubility of these dopants up to the doping level
added. XRD of parent LSCTA- is also given for comparison.
15 30 45 60 75
0
25
50
75
100
125
Rel
ativ
e In
ten
sity
Diffraction angle (2)
a
b
c
d
e
(332)
(420)(022)
(121)
(242)(400)
(042)
(040)
(200)
Fig. 9.1: XRD patterns of doped analogues of LSCTA-; a) LSCTA-, b)
LSCTN1, c) LSCTN5, d) LSCTF1 and e) LSCTF5.
The atomic and ionic radii of Ni and Fe are given in Table 9.2 and compared
with Ti. The ionic radii of Fe+3
and Ni+2
are greater than the Ti+4
, thus the unit cell
volume is expected to increase with these dopants. A similar trend was observed in
the XRD pattern where slight shift to low 2θ is observed upon doping. The shifting
suggests that Ti+4
(0.605 Å) was successfully substituted by larger Ni+2
(0.690 Å)
and Fe +3
(0.645 Å). The ionic radii of Ni+2
is greater than Fe+3
so it is anticipated that
Ni+2
doping would result in more expansion in the unit cell volume.
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
189
Table 9.2 Atomic and ionic radii of cations [14]
The peaks were indexed using WinXPOW software. All the doped analogues
were found to be iso-structural with the pristine having orthorhombic symmetry with
space group Pbnm. The values of cell parameters extracted are given in Table 9.3.
Table 9.3 Unit cell parameters for doped analogues of LSCTA-
Samples a (Å) b(Å) c (Å) V(Å)
3
LSCTA- 5.4661(7) 5.4638(6) 7.7343(6) 230.99
LSCTN1 5.4709(17) 5.4753(14) 7.7350(5) 231.70
LSCTN5 5.4760(7) 5.4830(5) 7.7370(4) 232.21
LSCTF1 5.4683(8) 5.4680(3) 7.7315(15) 231.17
LSCTF5 5.4721(3) 5.4681(17) 7.7348(9) 231.44
Analysis of the table shows that the doping results in expansion of unit cell.
The expansion in volume with Ni+2
as dopant is more than Fe+3
in accordance with
bigger ionic size of Ni as compared to Fe.
Atomic No Coordination Crystal (Å) Ionic (Å)
Ti 22 VI 0.745 0.605
Fe 26 VI 0.785 0.645
Ni 28 VI 0.830 0.690
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
190
9.3.2 Scanning electron microscopy
The morphology plays an important role in final properties of a ceramic. The
morphology of the doped analogues was studied with scanning electron microscopy.
9.3.2.1 Calcined samples
The calcined samples show small particles having size in sub microns (Fig.
9.2).
Fig. 9.2: Micrographs of doped analogues of LSCTA-; a) LSCTN1, b) LSCTN5, c)
LSCTF1 and d) LSCTF5 after calcination at 1000 oC for 5 hours.
However, the size appeared to increase with doping if we compare with the
parent LSCTA-. It can also be seen that size depends on doping level as well. The
average size of 5% doping at B-site is larger than 1% doping with LSCTN5 size
greater than all. It might be explained considering the size difference of these dopants.
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
191
9.3.2.2 Sintered samples
The calcined powders were pressed in pellets which were sintered in air at
1400 oC for 6 hours. The SEM micrographs of surfaces of dense pellets of synthesized
compositions after sintering are presented in Fig. 9.3. The micrographs show well
sintered grains with averages size of ~10 micron in diameter with no surface pores.
The obtained compact microstructure shows good densification after sintering.
Fig. 9.3: Micrographs of doped analogues of LSCTA-; a) LSCTN1, b) LSCTN5, c)
LSCTF1 and d) LSCTF5 after sintering at 1400 oC for 6 hours.
Sintering also depends on the particle size of initial powder. Usually, small
size results in more shrinkage and hence more sintering.
9.3.3 Dilatometry
Thermal shrinkage of doped analogues is studied by dilatometry. Fig. 9.4
shows sintering behaviour of pellets prepared from calcined powder of doped
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
192
analogues, on heating up to 1400 °C in air. The shrinkage is directly related to the
particle size of the powder e.g., smaller size leads to more sinterability.
0 300 600 900 1200 1500 1800-30
-24
-18
-12
-6
0
0
350
700
1050
1400
Tem
pera
ture
(oC
)
dL/L
o%
Time (min)
a
b
c
d
Fig. 9.4: Dilatometric sintering curves doped analogues of LSCTA- in air;
a) LSCTN1, b) LSCTN5, c) LSCTF1 and d) LSCTF5.
From these sintering profiles of doped analogues, the shrinkage percentage
was calculated which is tabulated in Table 9.4.
It can be seen from the table that the shrinkage of 5% doped analogues is less
than that of 1% doped, as expected from micrographs of respective calcined powders.
An increase of doping level caused an increase in particle size which resulted in less
shrinkage of these doped analogues. The results indicate that the thermal shrinkage
can be controlled via a doping strategy.
Table 9.4 Shrinkage percentage of doped analogues in air calculated from
dilatometric data
Codes % Shrinkage
LSCTN1 26.59
LSCTN5 21.12
LSCTF1 25.24
LSCTF5 24.82
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
193
9.3.4 Electrical conductivity
Doping is usually done to tailor the properties of the parent compound, one of
which is usually to increase the conductivity. In the present study, doping was aimed
to increase the conductivity of LSCTA- to suggest new anode candidates based on the
parent composition.
Dc conductivity measurements were conducted on pellets of doped LSCTA-
samples sintered in air at 1400 oC. Fig. 9.5 represents the conductivity-time profile of
LSCTN1 in reduced atmosphere maintained by flushing 5% H2/Ar at 880 oC. The dc
conductivity increases with extent of reduction showing n-type nature.
0 10 20 30 40 50-8
-6
-4
-2
0
2
ln (S
cm
-1)
Time/hours
Fig.9.5: Conductivity profile of in-situ reduced LSCTN1 pellet in 5% H2/Ar
at 880 °C.
After an initial delay, the reduction proceeds in two stages, rapidly in the first
few hours, followed by a much slower subsequent increase. The equilibrium value
could not be reached even after 50 hours of reduction but the increase in conductivity
value was gradual as time progressed. These two stages might be related to the fast
removal of oxygen from the surface of the perovskite that is followed by a slow
diffusion into the bulk of the micron size grains.
Similar trends were observed in all other samples. The values of conductivity
recorded after 24 hours of in-situ reduction at 880 oC in each case is tabulated in
Table 9.5.
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
194
Table 9.5 Conductivity of doped analogues upon in-situ reduction in reducing
atmosphere (5% H2/Ar) at 880 oC
For comparison, the conductivity of LSCTA- is also given. All the doped
compositions showed higher conductivity values than the parent. It can also be seen
that the conductivity also depends on the dopant level. 5% doping resulted in higher
conductivity than 1% doping in case of both, iron doped and nickel doped LSCTA-.
Thus these new compositions can be further tested for anode applications. It is
expected that these doped analogues would perform better as an anode owing to their
better conductivity.
The 5% doped samples offering high conductivity (LSCTN5 and LSCTF5)
were pre-reduced in a reducing atmosphere with 5% H2/Ar at 1050 oC for 24 hours.
Then the conductivity-temperature profile was monitored in same atmosphere as
shown in Fig. 9.6.
Samples Conductivity (S cm
-1
)
LSCTA- 1.30
LSCTN1 2.12
LSCTN5 3.41
LSCTF1 2.99
LSCTF5 4.48
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
195
200 400 600 800 1000 1200
60
80
100
120
140
160a
(S
cm
)
Temperature (K)
Heating
Cooling
200 400 600 800 1000 1200
50
75
100
125
150
175b
(S
cm
)
Temperature (K)
Cooling
Heating
Fig. 9.6: Conductivity profile during thermocycling of pre-reduced samples in
reducing atmosphere (5% H2/Ar); a) LSCTN5 and b) LSCTF5.
On cooling, the conductivity increases with the decrease in temperature until
~320 K indicative of positive temperature coefficient of resistance and metallic type
behaviour. This suggests electronic conduction to be the pre-dominant mechanism in
these materials. Further decrease of temperature from ~320 K to room temperature
causes drop of conductivity showing metal insulator transition.
From above graphs, the value of conductivity was determined at 880 oC and is
tabulated in Table 9.6.
Table 9.6 Conductivity of pre-reduced doped analogues in reducing atmosphere
(5% H2/Ar) at 880 oC
Samples Conductivity (S cm-1
)
LSCTA- 38.0
LSCTN5 66.1
LSCTF5 46.8
A significant improvement in conductivity of the parent composition LSCTA-
is observed upon 5% doping with Fe and Ni. The pronounced increase in conductivity
values in pre-reduced samples may point to their applications in fuel cells where the
reducing atmosphere is an essential environment at the anode site.
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
196
The same samples were then tested for morphology. The micrographs show that
surface of reduced samples is decorated with small particles which were not present in
the sintered samples. The micrographs of pre-reduced LSCTF5 and LSCTN5 are
shown in Fig. 9.7.
Fig. 9.7: Micrographs of pre-reduced samples; a) LSCTN5 and b) LSCTF5.
The micrographs show that surface of reduced samples is decorated with small
particles which were not present in the sintered samples. It shows that reduction has
resulted in ex-solution of some particles. The nature of these particles can be
understood by comparing the reduction potentials of the couples present in the
systems. The reduction potentials of the couples are given in Table 9.7.
Table 9.7 Standard reduction potentials of redox couples in doped samples
It can be seen that in all of these couples, titanium is the most difficult to
reduce. However, both Fe and Ni couples are easily reducible. Thus, it might be
anticipated that upon reducing these samples, these dopant reduce to metallic particles
because of their lower reduction potentials than Ti as given in Table 9.7. The
Couples Electrode Reaction E
o (V)
Ti+4
/Ti 4 4 ( )Ti e Ti s -0.88
Fe+3
/Fe 3 3 ( )Fe e Fe s -0.04
Ni+2
/Ni 2 2 ( )Ni e Ni s -0.23
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
197
exsolution has been attributed to the inability of the host lattice to accommodate
vacancies (A-site vacancies and inherent and introduced oxygen vacancies) beyond a
certain limit upon reduction. Thus the defect chemistry provides the driving force for
the exsolution of B-site dopants.
The proposed mechanism for exsolution involves creation of oxygen
vacancies in titanate upon reduction in first step. In the second step, the limit of
oxygen vacancies is reached and the perovskite lattice cannot hold any more
vacancies. Further reduction results in exsolution of small proportion of B-site
dopants to the surface with simultaneous reduction to the respective metals [15].
The exsolution of B-site dopants has been discussed in the literature [16-19].
Among various applications, the metal nano-particle precipitation has been shown to
improve catalytic properties of the SOFC anodes where the degradation of SOFC
anode was successfully eliminated by repeated redox cycling. Upon oxidation, these
metal nano-particles redissolve in the oxide lattice and subsequent reduction causes
precipitation of fresh metal nano-particles again available for enhanced performance.
The regenerative behaviour of nano-particles reduces the anode degradation [20, 21].
Thus it is anticipated that these doped analogues would serve to be better anode
candidate then parent LSCTA-.
The key requirement to prepare such SOFC anodes is to have a catalyst
element having good solubility in the lattice in air (at high oxygen partial pressure)
and a relatively low free energy of oxide formation so that precipitation of separate
metallic phase could take place upon reduction [22].
Thus it is anticipated that these doped analogues would serve to be better
anode candidate then parent LSCTA-. Nevertheless, further investigations, especially
symmetrical and button cell testing are required to evaluate the effect of the Ni and Fe
doping on the electrochemical performance of LSCTA-.
9.4 Conclusions
Fe and Ni doped analogues were successfully synthesized via the Pechini
method with an aim to further improve the conductivity of parent LSCTA-. All the
doped analogues have the same orthorhombic symmetry as of the parent however; an
expansion in unit cell volume was observed which is in accordance with the ionic
sizes of the dopants. The doped analogues offered better conductivity than the parent,
imparting these new compositions as suitable anode support candidates. The B-site
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
198
dopants were ex-solved upon reduction due to manipulation of the defect chemistry of
doped compositions. It is anticipated that these doped analogues would be better
anode candidate then parent LSCTA-. It could be concluded that B-site doping is an
effective approach to improve the conductivity of the parent composition.
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
199
REFERENCES
1. J. Canales-Vazquez, S.W. Tao and J.T.S. Irvine, Solid State Ionics, 2003, 159,
159 – 165.
2. K.B. Yoo, G.M. Choi, Solid State Ionics, 2009, 180, 867 – 871.
3. R. Mukundan, E.L. Brosha F.H. Garzon, Electrochem. Solid St., 2004, 7, A5 –
A7.
4. M. Roushanafshar, J.L Luo, A.L. Vincent, K.T. Chuang, A.R. Sanger, Int. J.
Hydrogen Energ., 2012, 37, 7762 – 7770.
5. O.A. Marina, N.L. Canfield, J.W. Stevenson, Solid State Ionics, 2002, 149, 21
– 28.
6. D.N. Miller, J.T.S. Irvine, J. Power Sources, 2011, 96, 7323 – 7327.
7. J. Karczewski, B. Riegel, M. Gazda, P. Jasinski and B. Kusz, J. Electroceram.,
2010, 24, 326 – 330.
8. A. Ovalle, J.C. Ruiz-Morales, J. Canales-Vasquez, D. Marrero-Lopez, J.T.S.
Irvine, Solid State Ionics, 2006, 177, 1997 – 2003.
9. D. Neagu, J.T.S. Irvine, Chem. Mater., 2011, 23, 1607 – 1617.
10. X. Li, H. Zhao, F. Gao, N. Chen, N. Xu, Electrochem. Commun., 2008, 10,
1567 – 1570.
11. D.P. Fagg, V.V. Kharton, J.R. Frade A.A.L. Ferreira, Solid State Ionics, 2003,
156, 45 – 57.
12. N.G. Eror, U. Balachandran, J. Am. Cer. Soc., 1982, 65, 426 – 431.
13. T. Kolodiazhnyi, A. Petric, J. Electroceram., 2005, 15, 5 – 11.
14. R. Shannon, Acta Crystallogr., Sect. A: Found. Crystallogr., 1976, 32, 751 –
767.
15. G. Tsekouras, D. Neagu, J.T.S. Irvine, Energy Environ. Sci., 2013, 6, 256 –
266.
16. Y. Nishihata, J. Mizuki, T. Akao, H. Tanaka, M. Uenishi, M. Kimura, T.
Okamoto, N. Hamada, Nature, 2002, 418, 164 – 167.
17. Y. Wang, B. D. Madsen, W. Kobsiriphat, S. A. Barnett and L. D. Marks,
Microsc. Microanal., 2007, 13, 100 – 101.
18. D. M. Bierschenk, E. Potter-Nelson, C. Hoel, Y.G. Liao, L. Marks, K.R.
Poeppelmeier, S.A. Barnett, J. Power Sources, 2011, 196, 3089 – 3094.
CH-9 Synthesis and Characterization of Doped Analogues of LSCT
200
19. B. D. Madsen, W. Kobsiriphat, Y. Wang, L. D. Marks, S. A. Barnett, ECS
Trans., 2007, 7, 1339 – 1348.
20. L. Adijanto, V.B. Padmanabhan, R. Kungas, R.J. Gorte, J.M. Vohs, J. Mater.
Chem., 2012, 22, 11396 – 11402.
21. W. Kobsiriphat, B.D. Madsen, Y. Wang, L.D. Marks, S.A. Barnett, Solid State
Ionics, 2009, 180, 257 – 264.
22. B.D. Madsen, W. Kobsiriphat, Y. Wang, L. D. Marks, S. A. Barnett, J. Power
Sources, 2007, 166, 64 – 67.
201
Conclusions and Recommendations
Abstract
The current chapter concludes the dissertation by giving an overview of
conclusions drawn from different research aspects carried out in present project. At the
end, recommendations for future work are furnished.
10.1 Final Remarks
The primary goal of present work was to identify, characterize and develop
calcium doped lanthanum strontium titanate (LSCTA-) as an anode/anode support for
solid oxide fuel cells.
In the present project, processing conditions have been optimized for LSCTA- as
an anode support for SOFC. The suggested anode candidate was also found to have good
electronic conductivity as well as redox stability and chemical stability with yttria-
stabilized zirconia YSZ. Also, a close matching of thermal shrinkage with yttria-
stabilized zirconia was observed which could avoid de-lamination before and during fuel
cell operation.
Further the LSCTA- was processed in aqueous tape casting which is often used for
fabrication of SOFC anodes. The conductivity profile of sintered bars prepared from the
green tapes provided useful insight into nature of LSCTA-. The effect of impregnates on
conductivity was also studied. It is known from literature that ceria is an oxidation
catalyst and thus helps to lower the polarization resistance in fuel cell operations.
However, the present study explores the effect of ceria impregnation on the kinetics of
reduction quantitatively. The conductivity behavior of porous bodies showed a two stage
process on both oxidation and reduction cycling that exhibits strong reversibility. For the
reduction process, addition of impregnated ceria reduced the onset delay period and
increased the apparent rate constant, k values by 30-50% for both stages.
CH-10 Conclusions and Recommendations
202
An interesting aspect of the present research project was the use of carbon
microspheres (CMS) as pore formers in LSCTA- green tapes. The optimal porosity in
LSCTA- tapes was achieved by successful incorporation of carbon microspheres (CMS)
which were synthesized by an optimized hydrothermal method to yield CMS of desired
morphology and well dispersed nature.
The present study also directs the use of co-impregnated catalysts for better
performance in symmetrical and button cell testing where the performance seemed to
improve due to synergic effect of oxidation catalysts CeO2 and CGO with Ni.
Finally the LSCTA- was successfully doped at B-site with catalytically active
dopants, Fe+3
and Ni+2
without phase separation via the Pechini method. It was also found
that the doped analogues offered better conductivity then the parent LSCTA- imparting
these new compositions a suitable anode candidateship.
10.2 Conclusions
The first part of the study dealt with the synthesis of calcium doped lanthanum
strontium titanate. A solution phase Pechini method emerged to be quite effective in
producing fine homogeneous powders. XRD spectra show that a single orthorhombic
phase could be obtained at relatively low calcination temperatures. A calcination
temperature of 1000 °C was considered as optimum for a further promising processing.
LSCTA- showed n-type conduction nature where conductivity of a dense LSCTA-
specimen sintered in air increased by three orders of magnitude after in-situ reduction in
5% H2/Ar. Pre-reduction resulted in enhancement of conductivity to a value of 38 S cm-1
at 880 oC. Redox cycling showed encouraging redox stability of the ceramic system thus,
imparting it a suitable anode support candidateship.
The conductivity measurements of ceria impregnated bars fabricated from green
LSCTA- tapes showed that CeO2 impregnation resulted in further improvement in
conductivity by enhancing reduction kinetics, but had limited effect on the oxidation
processes, which were a little faster in the absence of a catalyst. Whilst the obtained rate
constants were derived using some approximations, all samples were treated similarly,
CH-10 Conclusions and Recommendations
203
hence the increase of rate constant kred, by about 50% due to ceria impregnation is
significant.
In an effort to optimize microstructure, it was found that thermal pre-treatment of
LSCTA- powder resulted in good microstructure with commercial pore formers which
also showed good compatibility with YSZ. As an extension of this part of study, it was
established that carbon microspheres spheres successfully prepared by hydrothermal
treatment of sucrose could be used as in-expensive pore former in the green tape of
LSCTA-.
From the results of symmetrical and button cells, it was observed that the poor
electrocatalytic activity of neat LSCTA- could be modified by catalytically active
components like ceria and CGO. The addition of these impregnates improved the
polarization resistance significantly. The co-impregnation with Ni is an effective
approach to drastically reduce the impedance. The performance can be improved by
optimizing the microstructure of anode. Optimization of quantity/quality of impregnates
could also help to improve the performance.
From the last part of the study, it could be concluded that B-site doping is an
effective approach to improve the conductivity of the parent composition. The B-site
dopants were ex-solvated upon reduction due to manipulation of the defect chemistry of
doped compositions. It is anticipated that these doped analogues would serve as better
anode candidate then parent LSCTA- due to their enhanced conductivity and regenerative
behaviour of nano particles ex-solved upon reduction.
10.3 Recommendations for Future Research
Although the results presented here successfully show that LSCTA- could be
considered as a suitable anode support, further research prospects still remain covered.
Great efforts have been made to optimize the synthetic procedure and basic
developmental aspects of LSCTA- as anode candidate. However, still additional work is
needed to improve the performance of this anode in fuel cell testing. It was found that co-
impregnation with ceria or CGO with Ni seems to be an effective catalyst for significant
CH-10 Conclusions and Recommendations
204
reduction in polarization resistance of the cells but it is anticipated that optimizing the
quantity of these impregnates can further improve the performance.
Another aspect could be replacement of Ni with Pd, Rh in co-impregnated cells to
gauge the activity of LSCTA-. After optimization, a good idea is to use anode supported
cell configuration having thin electrolyte for testing in order to achieve good
performance.
In the continuing search for Ni-YSZ alternate materials, the present study has also
suggested Ni and Fe doped LSCTA- compositions as anode/anode supports for SOFCs.
This has opened a door of research to explore these materials for anode support
candidateship.
205
Appendix A8-I
Table 1 Extracted resistivity values from fitting the impedance data of symmetrical cell
B at different temperatures
Resistivity
(Ω cm2)
650 °C 700 °C 750 °C 800 °C 850 °C
R1 18.52 11.59 8.10 6.14 4.56
R2 652.70 263.20 111.50 55.23 6.54
R3 187.70 141.00 92.40 48.58 12.70
Table 2 Extracted resistivity values from fitting the impedance data of symmetrical cell
C at different temperatures
Resistivity
(Ω cm2)
650 °C 700 °C 750 °C 800 °C 850 °C
R1 17.61 11.01 7.67 5.90 4.80
R2 25.91 14.73 11.99 10.04 5.11
R3 435.20 243.30 131.80 67.70 20.89
206
Table 3 Extracted resistivity values from fitting the impedance data of symmetrical cell
D at different temperatures
Resistivity
(Ω cm2)
650 °C 700 °C 750 °C 800 °C 850 °C
R1 19.65 12.26 8.47 6.54 4.59
R2 48.59 28.19 14.64 8.00 0.04
R3 4.857 2.20 1.66 0.90 0.31
R4 10.47 1.90 1.70 1.62 1.27
Table 4 Extracted resistivity values from fitting the impedance data of symmetrical cell E
at different temperatures
Resistivity
(Ω cm2)
650 °C 700 °C 750 °C 800 °C 850 °C
R1 17.60 11.07 7.75 6.00 4.84
R2 3.04 0.82 0.44 0.40 0.19
R3 32.75 23.46 11.39 6.07 2.01
R4 16.27 2.45 2.43 1.96 1.66
207
Appendix A8-II
Table 1 Extracted resistivity from fitting the impedance data of button cell A at different
temperatures
Resistivity
(Ω cm2)
750 °C 800 °C 850 °C
R1 3.907 2.905 2.303
R2 3.229 2.276 0.006
R3 2.455 1.269 0.077
R4 6.974 4.381 0.683
Table 2 Extracted resistivity from fitting the impedance data of button cell B at different
temperatures
Resistivity
(Ω cm2)
750 °C 800 °C 850 °C
R1 3.67 2.55 1.96
R2 0.86 0.42 0.22
R3 1.62 1.59 0.45
R4 0.44 0.43 0.39
208
List of Publications
1. Azra Yaqub, Cristian Savaniu, Naveed K. Janjua, John T.S. Irvine, J. Mat.
Chem. A, 2013, 1, 14189-14197.
2. Azra Yaqub, Cristian Savaniu,
Naveed K. Janjua, John T.S. Irvine, Synthesis and
Characterization of B-site Doped La0.20Sr0.25Ca0.45TiO3 as SOFC Anode
Materials, submitted to Int. J. Hydrogen Energy.
3. Azra Yaqub, Cristian Savaniu,
Naveed K. Janjua, John T.S. Irvine, Application of
carbon microspheres as pore former for SOFC electrodes, to be submitted in
Carbon.
Posters
1. Azra Yaqub, Cristian Savaniu,
Naveed K. Janjua,
John T.S. Irvine, “Synthesis
and Characterization of LSCTA- (La0.20Sr0.25Ca0.45TiO3) as SOFC Anode Material,
Poster presented in RSC Solid State Group Christmas Meeting, University of
Liverpool, UK, 2011.
2. Azra Yaqub, Cristian Savaniu,
Naveed K. Janjua, John T.S. Irvine, “Application
of carbon microspheres as pore former”, Poster accepted in International
Conference on Diamond and Carbon Materials, 2-5 September 2013, Italy.