in situ
PEM Fuel Cell Electrocatalyst Durability Measurements 2006 Fuel Cell Seminar
Rod L. Borup, John R. Davey, Fernando H. Garzon and Paul M. Welch Los Alamos National Laboratory and Karren More Oak Ridge National Laboratory This presentation does not contain any proprietary or confidential information.
Fuel Cell Barriers
Electrocatalyst has major impact on: • Durability • Cost • Electrode performance
*Nancy Garland, 2006 DOE H2 Program Review Meeting
Electrocatalyst Durability Studies • Operating condition effects • Drive cycle and steady-state • Accelerated testing • Effect of temperature, relative humidity • Examine degradation mechanisms • Performance loss • Surface area loss • in situ and post-characterization • Changes in particle size distribution • SEM / XRF / XRD (ex situ and in situ) / TEM
Catalyst/Ionomer Structure Is Not Ideal TEM of prepared MEA Structure • Pt not strongly anchored to Carbon • The ionomer tends to “clump” around and between the C-support • Ionomer “picks up” Pt during ink preparation • Picture not typical of entire MEA structure * Karren More from ORNL
Electrocatalyst Particle Growth Stationary vs. Automotive Catalyst Particle Size / nm
4.0 3.5
Untested Anode Cathode
3.0 2.5 2.0 1.5 1.0 0.5 0.0 Fresh Catalyst
Prepared MEA
900 hr 3500 hr 1200 hr Steady State Steady State Drive Cycle
• Cathode electrocatalyst particle growth during operation • Particle growth increased with power/voltage cycling
Fuel Cell Drive Cycle Testing Transportation Vehicle Power Rapidly Cycles Power transients will > 100,000’s US06 Drive Cycle
0.6 50 0.4 25
0.2
0
0.09
0.18
0.27
0.0 0.36
Voltage / Current
0.8
75
0
DOE Drive Cycle
1.0
Voltage / V
Current / Amps
100
1.0 0.8 0.6 0.4 0.2 0.0
Cell Voltage Current Density
0
0.03
0.06
0.09
Time / hr
Time / hr
1 cycle over 20 minutes
1 cycle over 6 minutes
• Fuel Cell provides power transients • ~ 400,000 potential transients
0.12
• Battery/hybrid provides power transients • Fewer potential transients
Bimodal Particle Size Distribution After Drive Cycle Testing
• Cathode electrocatalyst layer post 2000 hrs US06 drive cycle • High concentration of large particles are circled.
Potential Cycling Measurements Simulates drive cycle and accelerated catalyst aging technique
Voltammograms during Cycling 100
Current (mA)
50 0 -50 prior to CV cycling after 300 cycles after 600 cycles after 900 cycles after 1200 cycles after 1500 cycles
-100 -150 -200 0
200
400
600
800
Voltage (mV)
1000
% Initial Pt Catalytic Surface Area
Repeatedly cycle cathode potential Pt Catalytic Surface Area 100 0.1V - 1.2V
90 80 70
Cell 80 °C H2 226% RH Air 100% RH
60 50 0
200
400 600
800 1000 1200 1400 1600
Number of CV Cycles
Variables Examined: Potential Range, Temperature, Cycles vs. Time (Scan Rate), Relative Humidity, Catalyst Loading, Catalysts
Electrocatalyst Surface Area and Particle Size Cathode Pt Grain Size (nm)
Particle size vs. electrochemical active surface area 9 0.4V 0.1V 0.1V 0.1V
8 7 6
-
Pt 20%/Carbon Fresh Crystal size • 19 Å average (distribution) • 21 Å average (volume)
0.96V 0.96V 1.2V 1.0V
5 4
0
25
50
75
100
125
150
Size Angstroms
3
Pt Post Cycled 0.1-1.0 V
2 0
20 40 60 80 % of Initial Active Pt Surface Area After 1500 CV Cycles
100
Crystal size • 38 Å average (distribution) • 51 Å average (volume)
• Surface area loss due to particle size growth • Particle distribution remains log-normal 0
25
50
75
100
Size Angstroms
125
150
Pt Coarsens within Ionomer Pockets* Fresh cathode Pt catalyst particles
Pt particles after 80°C cycling
• Increased Pt coarsening associated with Pt concentrated within ionomer “pockets” • Indicates particle coalescence Æ particle agglomerates are not single crystals * Karren More from ORNL
Pt Particle Size / nm
Temperature Effect on Particle Size Growth 6 5 4 3 2 1 0 Initial
60
80
100
120
Cycling Temperature / oC
• Pt particle growth rate increases with temperature Pt particle size after 1500 cycles (Cycled from 0.1V - 0.96V, 100 % RH)
Humidity Effect on Particle Size Growth 8
Pt Particle Size / nm
Cycling 0.1 - 1.0 V 6
Cycling 0.1 - 1.2 V
4
2
0 10
50
100
Relative Humidity / % RH
• Pt particle growth rate increases with humidity • Growth mechanism enhanced by H2O (Pt mobility/solubilization) (Pt particle size after 1500 cycles, 80 oC)
Comparison of Scan Rates: Time
Comparison of Scan Rates: Cycles
100%
10 mV/sec 50 mV/sec
90% 80% 70% 60% 50% 0
300
600
900
1200
Number of Potential Cycles
1500
100%
% In itia l S u rfa c e A re a
% of Initial Surface Area
Cycle vs. Time Effect on Catalyst Growth 10 mV/sec 50 mV/sec
90% 80% 70% 60% 50% 0
100 200 300 Time Above 0.9V (min)
Cycling from 0.1V - 0.96V
• Pt sintering correlates strongly with # of cycles • Time at peak potential has lower correlation
400
Pt-Co Alloy Catalyst
% Initial Surface Area
120%
• Pt-Co Alloys show slower reduction in surface area compared with Pt
110% 100% 90%
• Higher humidification enhances particle growth
80% 70%
• Humidification shows
50% RH 100% RH
60%
similar effects as with Pt
50% 0
2000
4000
6000
# of Cycles
8000
10000
Kinetic Modeling of Pt Particle Growth Particle Population
1st Order Particle Growth Rate 500
Time = 2 Time = 3 Time = 4 Time = 5 Time = 6 Time = 7
400 300 200 100 0
0
20
40 60 Particle Size
80
100
• Start with initial distribution and 1st order particle growth rate • Rgrowth = k*(PtSA) dPtSA/dt = k * PtSA • Pt Particles show change in diameter • Growth in particle size occurs, distribution form does not change
Kinetic Modeling of Pt Particle Growth Particle Population
Particle Growth with Particle Coalescence 8000
Time = 1 Time = 2 Time = 3 Time = 4 Time = 0
6000
4000
2000
0
0
10
20
30
40
50
Particle Size
• Initial distribution with 1st order particle growth and particle coalescence – Particles show change in diameter and distribution form – Particle size distribution form changes – bimodal distribution • Similar to post drive cycle particle distribution
Equilibrium concentration of dissolved Pt depends on potential* •
o 1.E-05Pt/C-80 C, GM-MIT
Dissolved Pt conc. (M)
ANL-Pt wire ANL-Pt/C-Nafion-23oC From Pourbaix
~ 2 orders magnitude higher solubility at 1.0 V compared with 0.6 V
1.E-06
– Potential cycling during transportation operation likely between OCP (1.0 V) and ~ 0.6 V
1.E-07
•
1.E-08
• 1.E-09
1.E-10 0.6 0.7 0.8
0.9 1.0 1.1 1.2
Both polycrystalline Pt and Pt/C showed a maximum Pt equilibrium concentration, at 1.1 V and 1.0 V, respectively At >1.2 V, dissolved Pt concentration decreases for polycrystalline Pt, but continues to increase for Pt/C
1.3 1.4 1.5
Potential (Volts vs. SHE)
*Deborah Myers, ANL
Conclusions • Loss of Pt surface area correlates with loss of MEA performance • Loss of surface area caused by Pt particle growth • Pt particle growth on cathode occurs during steady-state operation • Pt particle growth enhanced with cycling operation • Operating conditions effect rate of particle sintering • Higher temperatures increase rate of particle growth • Higher RH increase rate of particle growth • Rate of growth electrochemical potential dependent • # cycles more important than time at high potentials • Particle growth mechanism • Catalyst particles not strongly anchored to carbon support • Particle growth correlates with solubilization and reprecipitation • Pt solubility changes 3 orders of magnitude with potential • Particle coalescence occurs during cycling
Thanks to • DOE -EERE Hydrogen, Fuel Cells and Infrastructure Program for financial support of this work – Program manager: Nancy Garland
Rod L. Borup, John R. Davey, Fernando H. Garzon and Paul M. Welch Los Alamos National Laboratory and Karren More Oak Ridge National Laboratory This presentation does not contain any proprietary or confidential information.
Fuel Cell Barriers
Electrocatalyst has major impact on: • Durability • Cost • Electrode performance
*Nancy Garland, 2006 DOE H2 Program Review Meeting
Electrocatalyst Durability Studies • Operating condition effects • Drive cycle and steady-state • Accelerated testing • Effect of temperature, relative humidity • Examine degradation mechanisms • Performance loss • Surface area loss • in situ and post-characterization • Changes in particle size distribution • SEM / XRF / XRD (ex situ and in situ) / TEM
Catalyst/Ionomer Structure Is Not Ideal TEM of prepared MEA Structure • Pt not strongly anchored to Carbon • The ionomer tends to “clump” around and between the C-support • Ionomer “picks up” Pt during ink preparation • Picture not typical of entire MEA structure * Karren More from ORNL
Electrocatalyst Particle Growth Stationary vs. Automotive Catalyst Particle Size / nm
4.0 3.5
Untested Anode Cathode
3.0 2.5 2.0 1.5 1.0 0.5 0.0 Fresh Catalyst
Prepared MEA
900 hr 3500 hr 1200 hr Steady State Steady State Drive Cycle
• Cathode electrocatalyst particle growth during operation • Particle growth increased with power/voltage cycling
Fuel Cell Drive Cycle Testing Transportation Vehicle Power Rapidly Cycles Power transients will > 100,000’s US06 Drive Cycle
0.6 50 0.4 25
0.2
0
0.09
0.18
0.27
0.0 0.36
Voltage / Current
0.8
75
0
DOE Drive Cycle
1.0
Voltage / V
Current / Amps
100
1.0 0.8 0.6 0.4 0.2 0.0
Cell Voltage Current Density
0
0.03
0.06
0.09
Time / hr
Time / hr
1 cycle over 20 minutes
1 cycle over 6 minutes
• Fuel Cell provides power transients • ~ 400,000 potential transients
0.12
• Battery/hybrid provides power transients • Fewer potential transients
Bimodal Particle Size Distribution After Drive Cycle Testing
• Cathode electrocatalyst layer post 2000 hrs US06 drive cycle • High concentration of large particles are circled.
Potential Cycling Measurements Simulates drive cycle and accelerated catalyst aging technique
Voltammograms during Cycling 100
Current (mA)
50 0 -50 prior to CV cycling after 300 cycles after 600 cycles after 900 cycles after 1200 cycles after 1500 cycles
-100 -150 -200 0
200
400
600
800
Voltage (mV)
1000
% Initial Pt Catalytic Surface Area
Repeatedly cycle cathode potential Pt Catalytic Surface Area 100 0.1V - 1.2V
90 80 70
Cell 80 °C H2 226% RH Air 100% RH
60 50 0
200
400 600
800 1000 1200 1400 1600
Number of CV Cycles
Variables Examined: Potential Range, Temperature, Cycles vs. Time (Scan Rate), Relative Humidity, Catalyst Loading, Catalysts
Electrocatalyst Surface Area and Particle Size Cathode Pt Grain Size (nm)
Particle size vs. electrochemical active surface area 9 0.4V 0.1V 0.1V 0.1V
8 7 6
-
Pt 20%/Carbon Fresh Crystal size • 19 Å average (distribution) • 21 Å average (volume)
0.96V 0.96V 1.2V 1.0V
5 4
0
25
50
75
100
125
150
Size Angstroms
3
Pt Post Cycled 0.1-1.0 V
2 0
20 40 60 80 % of Initial Active Pt Surface Area After 1500 CV Cycles
100
Crystal size • 38 Å average (distribution) • 51 Å average (volume)
• Surface area loss due to particle size growth • Particle distribution remains log-normal 0
25
50
75
100
Size Angstroms
125
150
Pt Coarsens within Ionomer Pockets* Fresh cathode Pt catalyst particles
Pt particles after 80°C cycling
• Increased Pt coarsening associated with Pt concentrated within ionomer “pockets” • Indicates particle coalescence Æ particle agglomerates are not single crystals * Karren More from ORNL
Pt Particle Size / nm
Temperature Effect on Particle Size Growth 6 5 4 3 2 1 0 Initial
60
80
100
120
Cycling Temperature / oC
• Pt particle growth rate increases with temperature Pt particle size after 1500 cycles (Cycled from 0.1V - 0.96V, 100 % RH)
Humidity Effect on Particle Size Growth 8
Pt Particle Size / nm
Cycling 0.1 - 1.0 V 6
Cycling 0.1 - 1.2 V
4
2
0 10
50
100
Relative Humidity / % RH
• Pt particle growth rate increases with humidity • Growth mechanism enhanced by H2O (Pt mobility/solubilization) (Pt particle size after 1500 cycles, 80 oC)
Comparison of Scan Rates: Time
Comparison of Scan Rates: Cycles
100%
10 mV/sec 50 mV/sec
90% 80% 70% 60% 50% 0
300
600
900
1200
Number of Potential Cycles
1500
100%
% In itia l S u rfa c e A re a
% of Initial Surface Area
Cycle vs. Time Effect on Catalyst Growth 10 mV/sec 50 mV/sec
90% 80% 70% 60% 50% 0
100 200 300 Time Above 0.9V (min)
Cycling from 0.1V - 0.96V
• Pt sintering correlates strongly with # of cycles • Time at peak potential has lower correlation
400
Pt-Co Alloy Catalyst
% Initial Surface Area
120%
• Pt-Co Alloys show slower reduction in surface area compared with Pt
110% 100% 90%
• Higher humidification enhances particle growth
80% 70%
• Humidification shows
50% RH 100% RH
60%
similar effects as with Pt
50% 0
2000
4000
6000
# of Cycles
8000
10000
Kinetic Modeling of Pt Particle Growth Particle Population
1st Order Particle Growth Rate 500
Time = 2 Time = 3 Time = 4 Time = 5 Time = 6 Time = 7
400 300 200 100 0
0
20
40 60 Particle Size
80
100
• Start with initial distribution and 1st order particle growth rate • Rgrowth = k*(PtSA) dPtSA/dt = k * PtSA • Pt Particles show change in diameter • Growth in particle size occurs, distribution form does not change
Kinetic Modeling of Pt Particle Growth Particle Population
Particle Growth with Particle Coalescence 8000
Time = 1 Time = 2 Time = 3 Time = 4 Time = 0
6000
4000
2000
0
0
10
20
30
40
50
Particle Size
• Initial distribution with 1st order particle growth and particle coalescence – Particles show change in diameter and distribution form – Particle size distribution form changes – bimodal distribution • Similar to post drive cycle particle distribution
Equilibrium concentration of dissolved Pt depends on potential* •
o 1.E-05Pt/C-80 C, GM-MIT
Dissolved Pt conc. (M)
ANL-Pt wire ANL-Pt/C-Nafion-23oC From Pourbaix
~ 2 orders magnitude higher solubility at 1.0 V compared with 0.6 V
1.E-06
– Potential cycling during transportation operation likely between OCP (1.0 V) and ~ 0.6 V
1.E-07
•
1.E-08
• 1.E-09
1.E-10 0.6 0.7 0.8
0.9 1.0 1.1 1.2
Both polycrystalline Pt and Pt/C showed a maximum Pt equilibrium concentration, at 1.1 V and 1.0 V, respectively At >1.2 V, dissolved Pt concentration decreases for polycrystalline Pt, but continues to increase for Pt/C
1.3 1.4 1.5
Potential (Volts vs. SHE)
*Deborah Myers, ANL
Conclusions • Loss of Pt surface area correlates with loss of MEA performance • Loss of surface area caused by Pt particle growth • Pt particle growth on cathode occurs during steady-state operation • Pt particle growth enhanced with cycling operation • Operating conditions effect rate of particle sintering • Higher temperatures increase rate of particle growth • Higher RH increase rate of particle growth • Rate of growth electrochemical potential dependent • # cycles more important than time at high potentials • Particle growth mechanism • Catalyst particles not strongly anchored to carbon support • Particle growth correlates with solubilization and reprecipitation • Pt solubility changes 3 orders of magnitude with potential • Particle coalescence occurs during cycling
Thanks to • DOE -EERE Hydrogen, Fuel Cells and Infrastructure Program for financial support of this work – Program manager: Nancy Garland
