ADVANCED DIAGNOSIS ON
PROTON EXCHANGE MEMBRANE FUEL CELL STACKS
QUENTIN MEYER
SUPERVISORS : DANIEL J.L. BRETT
PAUL ADCOCK, TOBIAS REISCH, OLIVER CURNICK
Supergen, Birmingham, 16-18 December 2013
INTRODUCTION
PROTON EXCHANGE MEMBRANE FUEL CELLS (PEMFC)
Water,
air, heat
out
Load
unit
Air
in
Flow channel
Diffusion
medium
Gas diffusion
H2O
layer
O2
Cathode
~ 200μm
Active layer
2H+
Membrane
Nafion
Membrane
eeH2
H2
out
Anode
Cathode
1/27
Active layer
Anode
~ 200μm
Gas diffusion
layer
Flow channel
+4H++4e-
~ 50μm
Diffusion
medium
H2
in
Nafion Ionomer
Platinum
Carbon
4H2O
PEM fuel cell operation, macro and micro structure.
GDL : Gas diffusion Layer
PROJECT OUTLINE
THE AIR COOLED (AC), OPEN CATHODE 5 CELLS STACK
INTELLIGENT ENERGY AC64-5
H2 out
H2 in
SPECIFICATIONS:
• Membrane active area of 60 cm2
• No external heating for the plates or
the gases. Self heated fuel cell.
• Combined cooling
technology.
Anode
plate
Cathode Active
plate
channel
MEA and
GLDs
Cooling
channel
and
oxidant
• Oxygen is supplied using fans
blowing air through the channels.
Over stoichiometry of 25
.
• Dry hydrogen
Picture and simplified scheme of the fuel cell stack showing ‘active’ and ‘cooling’ channels.
MEA : membrane electrode assembly. GDL : gas diffusion layer
2/27
FUEL CELL EFFICIENCY
AIR COOLED, OPEN CATHODE FUEL CELL
1.0
Voltage
Temperature
0.9
70
60
0.8
50
0.7
40
0.6
30
0.5
20
0.4
0.0
0.2
0.4
0.6
0.8
1.0
-2
Current density / A cm
Polarisation and evolution of temperature of the fuel cell stack
4/27
Temperature / oC
Cell voltage / V
CURRENT AND TEMPERATURE INTERDEPENDENCE ( K TYPE THERMOCOUPLE).
FUEL CELL EFFICIENCY
AIR COOLED, OPEN CATHODE FUEL CELL
CURRENT AND TEMPERATURE INTERDEPENDENCE (THERMAL IMAGING).
0.92 A cm-2
T
/
o
C
0.69 A cm-2
0.46 A cm-2
0.22 A cm-2
0 A cm-2
5/27
PROJECT OUTLINE
FUEL CELL TEST STATION FLOW DIAGRAM
K -type
thermocouple
Heated air
and vapour
T
H2 outlet
Water exhaust
H 2 inlet
Extractor
P
Manual
valve
Electric
Valve 1
Electric
Valve 2
+
Pressure
regulator
Purge
bottle
H at 5 bars
2
H2
5 bars
+
H at 0.4 bar vs atm
2
Load unit
Air
Illustration of the system test rig
3/27
Purge
valve
Pressure
transducer
OPTIMUM CURRENT
DENSITY
FUEL CELL OPERATION
1.0
0.6
Optimum power
density
0.9
0.5
Ohmic
0.8
0.4
0.7
0.6
Mass
transport
Charge
transfer
0.3
0.2
0.5
0.1
0.4
0.0
0.0
0.2
0.4
0.6
0.8
1.0
-2
Current density/A.cm
Optimum current density region
6/27
Power density / A cm-2
Cell voltage / V
WHERE SHOULD WE OPERATE THE FUEL CELL? AT WHAT CURRENT?
CURRENT OF LOWEST RESISTANCE
THE MINIMUM IN THE POLARISATION RESISTANCE
1. Polarisation curve differentiation :
𝜕𝑈
Δ𝑈
𝐸 𝑗𝑛+1 − 𝐸(𝑗𝑛−1 )
)𝑛 =
𝑗𝑛 =
𝜕𝑗
Δ𝑗
𝑗𝑛+1 − 𝑗𝑛−1
2. Low frequency impedance :
𝜕𝑈
Δ𝑈
)𝑛 =
𝑗𝑛 = 𝑍𝑛 0
𝜕𝑗
Δ𝑗
[1] N. Wagner, Journal of Applied Electrochemistry, no. 32, pp. 859–863, 2002.
[2]
[2] D. a. Harrington et al, Electrochimica Acta, vol. 56, no. 23, pp. 8005–8013, 2011.
7/27
POLARISATION CURVE
DIFFERENTIATION
POLARISATION DIFFERENTIATION
1.0
Voltage
Differentiated voltage
0.9
1.0
Δ𝐸
𝐸 𝑗𝑛+1 − 𝐸(𝑗𝑛−1 )
𝑗𝑛 =
Δ𝑗
𝑗𝑛+1 − 𝑗𝑛−1
0.8
0.7
0.5
0.6
0.5
0.4
0.0
0.0
0.2
0.4
0.6
0.8
1.0
Current density / A cm-2
Polarisation and differentiated polarisation
8/27
Differentiated polarisation /  cm2
Cell voltage / V
Δ𝐸 𝐸 𝑗𝑛+1 − 𝐸(𝑗𝑛−1 )
=
Δ𝐼
𝑗𝑛+1 − 𝑗𝑛−1
POLARISATION CURVE DIFFERENTIATION
LOW RESOLUTION OF THE VOLTAGE READINGS
9/27
Current
density /
A cm-2
Voltage /
V
ΔV /
V
1E-5
0.00839
0.01679
0.02518
0.03358
0.04197
0.05037
0.05876
0.06716
0.07555
0.08395
0.12592
0.1679
0.20987
0.25185
0.29382
0.3358
0.37777
0.41974
0.46172
0.50369
0.54567
0.58764
0.62962
0.67159
0.71357
0.75554
0.79752
0.83949
0.88146
0.92344
0.96541
1.00739
0.95
0.82575
0.79894
0.78284
0.77132
0.76415
0.75814
0.75172
0.7436
0.73945
0.73434
0.71397
0.70196
0.68625
0.67596
0.66607
0.65736
0.64731
0.63901
0.63087
0.62056
0.61319
0.60752
0.60212
0.59746
0.59037
0.58288
0.57506
0.56847
0.55614
0.53988
0.52028
0.49276
0.15106
0.04291
0.02762
0.01869
0.01318
0.01273
0.01454
0.01227
0.00927
0.02548
0.03238
0.02772
0.027
0.02018
0.0186
0.01876
0.01835
0.01644
0.01845
0.01768
0.01304
0.01107
0.0276
0.01175
0.01458
0.01531
0.01441
0.01892
0.02859
0.03586
0.04742
0.12028
Δ𝐸
/
Δ𝑗
2
 cm
9.00238
2.55569
1.64503
1.11316
0.78499
0.74032
0.86599
0.73079
0.55152
0.50586
0.38571
0.3302
0.30971
0.27038
0.22156
0.22347
0.21861
0.19583
0.21977
0.2106
0.15533
0.13186
0.11983
0.13996
0.17367
0.18237
0.17165
0.2254
0.34056
0.42716
0.56486
2.38793
11.02027
ΔV is in the same range than the
measurement noise
EIS provides a much higher sampling
rate (150kHz) and therefore should
provide a lower noise.
LOW FREQUENCY
IMPEDANCE
ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY
Current
Oscillations of voltage and current
10/27
Imaginary Impedance
Voltage
Voltage and Current
GENERAL PRINCIPLES
Relative
amplitude
Phase
shift
Decreasing
frequency
Phase
angle
Real Impedance
ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY
W
Real Impedance Re(Z)
EIS allows the various losses in an
electrochemical cell to be separated and
quantified
11/27
Voltage and Current
Ohmic
Mass transport
Imaginary Impedance
Kinetic
Voltage
Imaginary Impedance -Im(Z)
EQUIVALENT CIRCUITS
Relative
amplitude
Phase
shift
Decreasing
frequency
Phase Real Impedance
angle
POLARISATION RESISTANCE
NYQUIST PLOTS OVER THE CURRENT DENSITY RANGE.
CORRELATION BETWEEN THE LOW FREQUENCY LIMIT AND THE SLOPE [1]
• Low
frequency
resistance
corresponds to the slope of the
I - V [1] (performed on
methanol oxidation reaction).
• Possible to perform similar
experiments in PEM fuel cells,
in order to find the optimum
resistance, and highlight the
minimum in the V-I slope.
[1] F. Seland, et al,” Electrochimica Acta, vol. 51, no. 18, pp. 3827–3840, 2006.
12/27
ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY
ISSUE WITH THE POLARISATION RESISTANCE
-Im(Z) /  cm2
GRADIENT APPROXIMATION USING THE LOW FREQUENCY INTERCEPT
0.15
10 Hz
0.10
0.05
100 Hz
1 kHz
0.00
High frequency
intercept
Low frequency
intercept
Polarisation
resistance
(extrapolated)
-0.05
1 Hz
0.1 Hz
All EIS measurements
performed using LabVIEW,
a load unit, a current
clamp
and
a
data
acquisition system and a
custom made EIS software
[1].
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
Re(Z) /  cm2
•
The polarisation resistance has been modelled, and the inductive arc is attributed to PtO
formation [2], and/or adsorbed intermediate species in the oxygen oxidation reaction [3].
•
The low frequency intercept can be easily measured and used as an approximation of
the polarisation gradient, for steady state conditions of 3 s.
[1] Meyer, Q., et a;. J. of Electrochem. Sci. and Eng., 3(3), 107–114.
[2] S. K. Roy, M. E. Orazem, and B. Tribollet, “J. Electrochem. Soc., vol. 154, no. 12, p. B1378, 2007.
13/27
[3] S. Cruz-Manzo, R. Chen, J. Electrochem. Soc., vol. 160, no. 10, pp. F1109–F1115, 2013.
ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY
NYQUIST PLOTS OVER THE CURRENT DENSITY RANGE : 0 TO 1 A CM-2
FREQUENCY RANGE : 3 kHz TO 0.1 HZ
Cell voltage
0.75
Low frequency intercept
0.50
0.25
Ohmic
de
0.8
1.0
0.25
0.50
0.75
1.00
1.25
1.50 1.75 2.00
Re (Z) /  cm2
2.25
2.50
2.75
Nyquist plots between 0 and 1 A cm-2
14/27
Cu
r
re
nt
0.6
Mass
transport
0.00
ity
ns
0.4
/A
0.2
cm
Charge
transfer
-2
0.0
-Im (Z) /  cm2,
Cell voltage / V
1.00
LOW FREQUENCY INTERCEPT
1.00.6
0.9
0.8
Voltage
Low frequency intercept
Charge
transfer
Charge
transfer
Ohmic
Ohmic
2.5
MassMass 2.0
transport
transport
0.5
1.5
0.7
Lowest
resistance
0.6
1.0
0.50.4
0.5
0.4
15/27
Low frequency intercept /  cm2
2
/V
Cell voltage
/  cm
Low frequency intercept
LOW FREQUENCY RESISTANCE
0.0
0.2
0.4
0.6
0.8
1.02
Current0.7
density
/ A cm
0.2 0.3 0.4 0.5 0.6
0.8
0.9 1.0
Current density / A cm-2
Low frequency intercept for the entire current range : 0 and 1 A cm-2
PARTIAL CONCLUSION
• EIS provides a better accuracy than the differentiated polarisation
curve in the determination of the overall resistance.
• Indicates a minimum around 0.55 A cm-2.
• No need for entire Nyquist measurement, low frequency region is
sufficient in order to study the intercept (1 Hz – 0.1 Hz).
• Location of the minimum depends on the operating conditions.
16/27
APPLICATIONS TO
DIFFERENT OPERATING
CONDITIONS
APPLICATIONS
FLOW RATE AND COMPRESSION
Influence of the cathodic flow rate (Intelligent Energy Stack).
• Temperature management
• Membrane water management
Influence of the compression (Pragma System –
Collaboration work with Erik Engebretsen- EIL).
• Contact resistance and mass transport limitations
17/27
APPLICATIONS : CHANGE OF FLOW RATE
Cell voltage / V
INFLUENCE ON THE VOLTAGE
1.0
1.65 m3/min
1.35 m3/min
1.05 m3/min
0.9
0.8
0.7
0.6
0.5
0.4
0.0
18/27
0.2
0.4
0.6
0.8
1.0
Current density / A cm-2
Influence of the flow rate (1.05 m3min-1, 1.35 m3/in, 1.65 m3min-1) on the
voltage
APPLICATIONS :
Temperature/ oC
CHANGE OF FLOW RATE
INFLUENCE ON TEMPERATURE
70
1.65 m3/min
1.35 m3/min
1.05 m3/min
60
50
40
30
20
0.0
0.2
0.4
0.6
0.8
Current density / A cm-2
1.0
Influence of the cathode flow rate (1.05 m3min-1, 1.35 m3min-1, 1.65
m3min-1) on the temperature
19/27
APPLICATIONS : CHANGE OF FLOW RATE
4 DIMENSIONAL REPRESENTATION
I,V, FLOW RATE AND TEMPERATURE
Cell voltage / V
T/ oC
Influence of the flow rate (1.05, 1.35, 1.65 m3 min-1 on the voltage and temperature)
20/27
APPLICATIONS :
CHANGE OF FLOW RATE
Low frequency intercept / cm2
INFLUENCE ON THE CURRENT OF LOWEST RESISTANCE
1.65 m3/min
1.2
1.35 m3/min
1.05 m3/min
1.0
0.8
0.6
0.4
1.0
21/27
0.8
0.6
0.4
0.2
0.0
Current density / A cm-2
Influence of the cathode flow rate (1.05 m3 min-1, 1.35 m3 min-1, 1.65 m3
min-1) on the low frequency intercept
APPLICATIONS :
Low frequency resistance /  cm2
CHANGE OF FLOW RATE
I,V, LOW FREQUENCY RESISTANCE AND FLOW RATE
22/27
T/ oC
Influence of the cathode flow rate (1.05 m3 min-1, 1.35 m3 min-1, 1.65 m3 min-1 on
the low frequency intercept and temperature
APPLICATIONS :
CHANGE OF FLOW RATE
High frequency resistance / cm2
INFLUENCE ON THE OHMIC RESISTANCE
1.65 m3/min
1.35 m3/min
1.05 m3/min
0.20
0.15
0.10
0.05
1.0
0.8
0.6
0.4
0.2
0.0
Current density / A cm-2
Influence of the cathode flow rate (1.05 m3 min-1, 1.35 m3 min-1, 1.65 m3
min-1) on high frequency resistance
23/27
APPLICATIONS :
High frequency resistance /  cm2
CHANGE OF FLOW RATE
I, OHMIC RESISTANCE, FLOW RATE AND TEMPERATURE
T/ oC
Similar results over 50
degrees
also
been
reported on self breathing
(no convection) systems
[1].
Increase
and
decrease
significantly
more important here.
[1] T. Fabian, et all, J. Power Sources, vol.
161, no. 1, pp. 168–182, 2006.
Influence of the cathode flow rate (1.05 m3 min-1, 1.35 m3 min-1,
1.65 m3 min-1) on the high frequency resistance and temperature
24/27
APPLICATIONS :
CHANGE OF COMPRESSION LEVEL
• The applied compression
can be varied (0 - 2.5 Mpa).
Bronkhorst EL-FLOW
controllers (Bronkhorst, UK)
Cell Compression Unit (Pragma
Industries SAS, France)
• Enables to study the effect
of the compression on the
membrane and GDL [1,2],
and its influence on the
hydration level [3].
In-house humidification
bottles
[1] T. J. Mason, et al, Journal of Power Sources, vol. 219, pp. 52–59, 2012.
[2] T. J. Mason et al, International Journal of Hydrogen Energy, vol. 38, no. 18, pp. 7414–7422, 2013.
[3] T. J. Mason et al,, J. Power Sources, vol. 272, pp. 70–77, 2013.
25/27
APPLICATIONS :
1.0
0.5 MPa
1.5 MPa
2.5 MPa
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Current density / A cm-2
Low frequency resistance/ cm2
Voltage / V
CHANGE OF COMPRESSION LEVEL
2.7
0.5 MPa
1.5 MPa
2.5 MPa
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.0
0.1
0.2
0.3 0.4 0.5 0.6 0.7
Current density/ A cm-2
Influence of the compression level (0.5 MPa, 1.5 MPa, 2.5 MPa) on the voltage and low frequency resistance
26/27
CONCLUSION
• Low frequency intercept enables to find the optimum
current density and the optimum operating conditions.
• Operating near the CLR should improve the durability of
the fuel cell, as the effects of the charge transfer
resistance and mass transport resistance are minimised.
• The optimum operating region is strongly depending on
the operating conditions, and needs to be determined prior
to long term endurance tests.
27/27
ACKNOWLEDGEMENT
Dan Brett, Paul Shearing, Simon Barrass,
James Robinson, Krisztian Ronaszegi
Paul Adcock, Toby Reisch, Oliver Curnick, Sean
Ashton
Questions ?