1
Diagnosis of PEMFC operation using EIS
International Symposium on
DIAGNOSTIC TOOLS FOR FUEL CELL TECHNOLOGIES
Trondheim, Norway, June 2009
Hydrogen and
Fuel Cells Group
Electrical Research Institute
Félix Loyola, Ulises Cano-Castillo
2
MEA STACK SYSTEM
quality performance reliability
degradation state of health
long life
3
Characterization using d.c. techniques:
voltammetry – electrochem. active area
linear voltammetry – H2 crossover (int. short circuit)
polarization curves – performance
0 0,25 0 ,50 0,75 1,00-0,0005
0
0,0005
0,0010
0,0015
0,0020
E (Volts)
I (A
mps
/cm
2)
M6 00 s_t C ruce H 2_2 B.corM6 00 C ruce H2_A_ 2.cor
0 0,25 0 ,50 0,75 1,00-0,00 05
0
0,00 05
0,00 10
0,00 15
0,00 20
E (Volts)
I (Am
ps/c
m2)
M6 00 s_t C ruce H 2_ 2B.corM3 20 C ruce H 2_A_ 2.cor
Cruce H2 Cruce de H2
MEA i max (A/cm2) (mol/s-cm2) (mL/min-cm2)
M600 s_t 0.0013523 7.008E-09 0.01030
M600 0.00080397 4.166E-09 0.00612
M320 s_t 0.0010818 5.606E-09 0.00824
4
Interest on EIS as applied to Fuel Cells
MEA development
• platinum activity follow-up
• bulk membrane resistance
• ionic conductivity at CL (through distributed element model)
Stack
• sensitivity to operating conditions
Systems control
• Potential flooding/drying detection
-0.04
-0.035
-0.03
-0.025
-0.02
-0.015
-0.01
-0.005
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Z'
Z''B i2 DE
B i2 SM
C i2 DE
C i2 SM
0
0.2
0.4
0.6
0.8
1
0 0.05 0.1 0.15 0.2 0.25
A/cm2
V
A
B
C
D
5
- Combination of experimentally determined kinetic parameters with EIS-determined parameters for PEMFC dynamic model
Steady-state and transient models
6
Facts & warnings:
• EIS highly sensitive to changes within the fuel cell
• EIS response is strongly dependent on design
• If gradients (heterogeneities) along active area exist,
EIS will pick them up
• “Dry” and excess of water may coexist in different
regions of FC (as well as other effects)
• EIS should not be seen as a black or white result but
as a color pallette (i.e. interpreted as such)
Q = can EIS bu used as a diagnostic technique during operation near
flooding/drying conditions?
7
water dynamics near dry/wet limit
Dehydration Stage: (previously conditioned and purged)
• Tcell = 40 °C
• Cathode: P: 10psi; flow: air 0.5 L/min
• Anode: P= 10psi; gas exit closed
• t = 1hr, EIS for ohmic resistance @ Eoc
Rest Stage (no humidification):• Tcell = 40 °C
• Cathode, P = 10psi, Flow: air 0 L/min (just after dehydration process)
• Anode, P = 10psi, gas exit closed.
• t = 1hr, EIS for ohmic resistance @ Eoc
MEA: 25cm2, Gore, Pt = 0.7 mg/cm^2, carbon paper, DL = 50µµµµ,
GFF: simple serpentine
8
0
10
20
30
40
50
60
70
80
0 20 40 60 80 100 120 140
Time (min)
Drying process 3
Redistribution ofwater
cell’s ohmic resistancedetermined by high frequency intercept
redistribution of
water from bulk
membrane
*low conductivity final current collectors plates used
9
DLCLM
DLCLM
DLCLM
after conditioning
after high air flux
CL w/H2O gradient
after 1 hr rest
H2O redistributes in CL
10
Experimental approach:
- Initial conditioning
- Operating near the limit of drying/flooding
- EIS (1Hz to 1KHz), 10mV (Ecell = 0.4V)
• low T & stoich’s
• dead ended configuration
• purge stages
• dry feeds
• use of O2
Real life is cruel for EIS:
Practical operating conditions are hardly under steady-state, it
depends on specific application and design
Own 50 cm2 MEA, GFFc = 4ps, GFFa pch. 0.7 mg Pt/cm2, Nafion NRE-212.
GDL-30-BC (C paper w/MPL). T=343.15 K (70°C) & 69 kPa (10 psi)
11
2 3 4 5
-3
-2
-1
0
Z'
Z''
4SP EIS 78.z4SP EIS 79.z4SP EIS 80.z
4SP EIS 81.z4SP EIS 82.z
4SP EIS 83.z4SP EIS 84.z
gradual drying
out-of-phase shifts
in-phase
gradual drying
100 101 102
-10
0
Frequency (Hz)
thet
a
phase angle
increases and
shifts right
Drying
cathode stoichiometry = 4
not humidified
both Z” & Z’ increase
12
2.0 2.5 3.0 3.5
-1.5
-1.0
-0.5
0
Z'
Z''
4SP EIS 63.z4SP EIS 64.z4SP EIS 65.z
4SP EIS 66.z4SP EIS 67.z
4SP EIS 68.z
out-of
-pha
se
grad
ual f
lood
ing
100 101 102 103
-14
-4
Frequency (Hz)
the
ta
Flooding
gradual flooding
1Hz
3.98Hz
5.01
6.30
7.94
10.00
12.58
freq. @
max. Imflooding
1
2
3
4
5
6
EIS: every 20 mins.
13
Rs CPEa
Rct a
CPEc
Rct c
Rs = total ohmic resistance
Rct a = anode charge transfer resistance
CPEa = non-ideal double layer capacitance anode
Rct c = anode charge transfer resistance
CPEc = non-ideal double layer capacitance anode
X
CPE = distributed element, diffusional process
14
0
0.5
1
1.5
2
2.5
3
3.5
4
0 20 40 60 80 100 120
t (min)
Rs o
hm
Rs (flooding)
Rs (drying)
resistive losses increase during drying
15
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 20 40 60 80 100 120
t (min)
Rct
oh
m
Rct (flooding)
Rct (drying)
kinetic losses increase during drying and flooding
16
0
0.01
0.02
0.03
0.04
0.05
0.06
0 20 40 60 80 100 120
t (min)
CP
Ec
CPE (Flooding)
CPE (drying)
(sp/o
hm
)
during drying CPE impedance
increases (distributed nature of CL?)
during flooding CPE
impedance reduces
17
100
1 01
102
103
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0
F reque ncy (H z)
Z''
4 SP E IS 6 3 .z4 SP E IS 6 4 .z4 SP E IS 6 5 .z4 SP E IS 6 6 .z4 SP E IS 6 7 .z4 SP E IS 6 8 .z
100
1 01
102
-15
-10
-5
0
Frequency (H z)
theta
4SP EIS 63.z4SP EIS 64.z4SP EIS 65.z4SP EIS 66.z4SP EIS 67.z4SP EIS 68.z
100
1 01
102
103
- 15
- 10
-5
0
Freque ncy (H z)
theta
4 S P EI S 78 .z4 S P EI S 79 .z4 S P EI S 80 .z4 S P EI S 81 .z4 S P EI S 82 .z4 S P EI S 83 .z4 S P EI S 84 .z
100
1 01
102
103
-0 .60
-0 .35
-0 .10
Freque ncy (H z)
Z''
4S P EIS 78.z4S P EIS 79.z4S P EIS 80.z4S P EIS 81.z4S P EIS 82.z4S P EIS 83.z4S P EIS 84.z
Flooding Drying
18
Conclusions:
- During drying, both in-phase and out-of-phase
impedance content increase
- During flooding only out-of-phase content increases
- For both cases it appears that there is one single
frequency threshold (~10Hz) from which out-of-phase
content starts to shift to:
• smaller frequencies for flooding
• larger frequencies for drying
19
Phase angle vs. Imaginary content:
- θ seems to better define initial drying/flooding process
(one single frequency?)
- θ can be associated with concentration profiles (ac
voltammetry?)
20
Recommendations:
• Minimize FC design effects (ε, GFF, GDL, CL, etc.), better
base-line during testing
• specific frequencies might be design-dependent: need
further studies
• dry/flooding case: comparison of states only as a short
time forcast
• Different FC sizes might need different approaches for
diagnosis
• Isolation of true dry and true flooding conditions is only
possible if homogenous internal conditions are achieved
• Structural effects should be studied (carbon support as a
conducting grid, i.e. additional capacitance or inductance
effects?)
21
• Properties of components are needed particularly
substack layers (i.e. capillary properties, etc.)
This presentation was AH1N1 freeP!
For your safety: