Kein Folientitel

1
The Polarization of LaMnO3 Cathodes in Solid
Oxide Fuel Cells Microelectrode Measurements and
Numerical Calculations
J. Fleig, H.-R. Kim, V. Brichzin, J.
Maier Max-Planck-Institute for Solid State
Research Stuttgart
Electroceramics VIII
Rome
August 27, 2002
2
Introduction
Solid oxide fuel cell (SOFC)
Internal resistance (polarization)

cathodic
resistance


ohmic resistance


anodic resistance
3

Topic of this talk
Elucidation of the cathodic polarization
mechanisms in solid oxide fuel cells
O
1/2 O
2e

-
2-
Mechanism of the oxygen reduction reaction at
LaMnO
2
3
Possible pathways (literature)
surface path
bulk path
O
2
O
O
d
-
2
ad
three phase
2e
-
O
cathode
2-
boundary
)
d
(2-
e
d

cathode
-
-
O
ad
electrolyte
O
electrolyte
2-
O
2-
O
-incorporation into electrolyte
2-
O
-incorporation via cathode bulk
2-
close to three phase boundary
and two phase boundary
4
- surface path often assumed to be dominant at
LaMnO
cathodes

3
- a generally accepted reaction model does not
exist yet
O
2
O
d
-
ad
)
(2-
d
e
-
d
-
O
ad
Experiments which could yield information on the
path
O
2-
Geometry-dependent electrical measurements
O
- varying three phase boundary length
2
2e
-
O
2-
- varying contact area
- varying surface area
O
2-
LaMnO
3
Porous cathodes
difficult to
characterize and
to
control
geometrical parameters
ZrO
2
K. Sasaki, et. al.
J. Electrochem. Soc. 143 (1996) 530
5
To overcome the problem with respect to defined
electrode geometry
well-defined thin, dense microelectrodes
Depending on reaction path
or rate determining step

typical geometry dependencies
of the polarization resistance
varying
- three phase boundary length
- surface area
- thickness
6
Experimental
LSM-microelectrodes on yttria-doped ZrO
(YSZ) single crystals
2
SEM-image
(La
Sr
)
MnO
LSM (700 nm)
0.8
0.2
0.92
3
YSZ
A-site deficient to avoid formation
of La
Zr
O
-layer at interface
2
2
7

Pulsed laser deposition
dense film
Lithographic techniques, ion beam etching

arrays of microelectrodes (20..200
m
m diameter)
100 nm
impedance measurements
microelectrode thickness 100-800 nm
(0.05 Hz - 1MHz)
microscope
ac
YSZ
sapphire
Pt- counter electrode
heating table
7
LSM-microelectrode
contacted by tungsten carbide tip
set-up (without heating table)
heating table at 800 C
8
Results
9
Thickness dependence
30 microelectrodes of 100 nm thickness
30 microelectrodes of 250 nm thickness
30
resistance
25
20
W
/ M
15
W
R
10
5
0
0
50
100
150
200
250
300
350
400
thickness / nm

factor 2.5 in thickness
factor 2.2 in R
Conclusion
polarization resistance mainly due to transport
through the cathode

bulk path dominates the rate of oxygen reduction
reaction
at LSM-microelectrodes

s


ionic conductivity of LSM from R
(
800 C)
8
10
S/cm
-8
10
(No Transcript)
11
Reason
Defect chemical model
(F.W. Poulsen,
Solid State Ionics 129 (2000) 145)

V
O
partial pressure dependence of
vacancy concentration

Nernst equation
anodic () bias high p(O
) at YSZ/LSM interface
2

and ionic conductivity
anodic bias
decrease of C

V

bulk path blocked

reaction via three phase boundary
12
Transition from bulk to surface path depends on
electrode geometry


geometry as "weighting factor"
I
I
Bulk
Surface
thin broad microelectrodes

bulk path higher weighted
I
Bulk
I
Surface
13
Advantage of microelectrode experiments

quantitative parameters for bulk and surface path
Simulation of the
polarization resistance of porous cathodes
1. Step Only bulk path considered
numerical finite element calculations
equipotential lines in model cathode
cathode
0.95
O
O
O
2
2
0.8
O
2
2
O
O
0.5
2
2
electrolyte
Interpolation formula
d
R
area-related polarization resistance
d particle size of the cathode
s
15
ion
14

Using measured ionic conductivity of doped LaMnO

3

m
polarization resistance of
a porous cathode (d 0.6
m) at ca. 800 C

W

50
cm
2

O
True polarization is probably somewhat larger
2
2e
-
since oxygen incorporation reaction also
contributes to the resistance
O
2-
cathode

(first estimate
120
W
cm
2
)
electrolyte
O
2-
15
Summary

For cathodic voltage and U0
bulk path dominates oxygen
reduction
reaction at LSM-microelectrodes

Anodic bias causes mechanism change

predominant surface path

Numerical calculations
using electrochemical parameters
from microelectrode experiments


voltage at which the mechanism changes depends
on electrode geometry (voltage is cathodic for
porous electrodes)


Reaction rates due to bulk path and surface path
in doped LaMnO
cathodes
are
3
more similar than generally assumed
16
Consideration of the entire impedance spectrum
C

Equivalent circuit predicted
fit circuit II
if bulk path dominates and





40.0
transport
is rate determining





R
W





1 Hz
/ M
W






20.0
imag
-Z





10 Hz

reasonable fit





(J. Jamnik, J. Maier, and S. Pejovnik,
0.1 Hz
Electrochim. Acta
44 (1999) 4139 )
0.0
0.0
20.0
40.0
60.0
W
Z
/ M
real
W finite Warburg element
0.25
T
T
h
/D
hthickness
2
W
W
0.20
0.15
/ s
W
T
0.10
strong thickness dependence
0.05
supports bulk path interpretation
0.00
0
50
100
150
200
250
300
350
400
thickness / nm
17


estimate of chemical diffusion coefficient of O
in LSM from T
W


(
800 C) D
3 10
cm
/s
-9
2
s



estimate of oxygen vacancy concentration from D
and


C


8 10

-6
V
However
Deviations from theoretical thickness dependence
certain effect of additional step
(possibly oxygen incorporation into LSM)
18
(No Transcript)