Fuel Cell Program

2005 DOE Hydrogen Program ReviewWednesday May 25, 2005

Rod BorupLos Alamos National Laboratory

PEM Fuel Cell Durability

John Davey, David Wood, Fernando Garzon, Michael Inbody, Dennis Guidry

Project #: FC40

This presentation does not contain any proprietary or confidential information.

Fuel Cell Program

Overview

Timeline2001: Project started as Fuel Cell Stack Durability on Gasoline Reformate2004: Changed focus to concentration on PEM H2Durability

Barriers• Durability (Barrier P)• Electrode Performance (Barrier Q)• Stack Material & Manf. Cost (Barrier O)

DOE Technical Targets (2010)

• Durability 5000 hours• Precious metal loading (0.2 g/rated kW)• Survivability (includes thermal cycling and

realistic driving cycles)

Budget• FY04: $900 k• FY05: $950 k

Fuel Cell Program

Technical Objectives:Quantify and Improve PEM Fuel Cell Durability

• Define degradation mechanisms• Design materials with improved durability

• Identify and quantify factors that limit PEMFC Durability• Measure property changes in fuel cell components during life testing

• Life testing of materials• Examine testing conditions, esp. drive cycle

• Membrane-electrode durability• Electrocatalyst activity and stability• Electrocatalyst and GDL carbon corrosion• Gas diffusion media hydrophobicity• Bipolar plate materials and corrosion products

• Develop and apply methods for accelerated and off-line testing• Improve durability

Fuel Cell Program

Approach to Durability Studies• Fuel Cell MEA Durability Testing and Study

• Constant roltage/current/power and power cycling (drive cycle)• 5 cm2, 50 cm2 and full-size active area (200 cm2) stack• VIR / cell impedance• catalyst active area• effluent water analysis

• in situ and post-characterization of membranes, catalysts, GDLs• SEM / XRF / XRD / TEM / ICP-MS / neutron scattering / H2 adsorption

• Develop and test with off-line and accelerated testing techniques• Potential sweep methods• Environmental component testing and characterization (GDL)• in situ XRD• Component interfacial durability property measurements• Membrane thinning and analysis

Fuel Cell Program

DOE Review MeetingReviewer Comments

FY2004 Reviewer Comments (Scores of 4 Outstanding and 1 Good)• The team seems to have the big picture well understood, which

consequently elevated its capability to do this analytical focused project

Strengths• Important work about durability• Excellent project execution with comprehensive set of tasks

Weaknesses• Did not show any collaboration outcome even with LANL group

FY2003 Reviewer comments:• The durability objective of this project is very important and I hope it will be actively addressed.

• I especially like the proposal of operating the system in a duty cycle operating mode.

• Introduction of drive cycle dynamics and start-up for next year is a plus …

• Need more fundamental work.

Fuel Cell Program

Interactions/Collaborations• Fuel Cell Materials

– MEAs (3M, Gore, LANL)– GDLs (Spectracorp, Toray, SGL, ETEK)– Catalysts (ETEK, SMP)

• Supporting measurements/interactionsAugustine Scientific (Chris Rulison – Contact angle)Oak Ridge National Laboratory (Karren More– TEM / SEM/EDS)Celanese Ticona (Rong Chen - SEM)LANL - NMT Division (Dave Wayne ICP/MS, Laser Ablation)LANL - LANSCE (Jaroslaw Majewski, Eric Watkins - Neutron Reflectivity)UNM (Plamen Atanassov – ICP/MS, Ion chromatography, Washburn Adsorption)NREL (Tony Markel – Fuel Cell Vehicle Drive Cycle)

• Stack: Teledyne Energy Systems

Fuel Cell Program

Electrocatalyst DurabilityDurability testing shows loss of active Pt surface area

0

5

10

15

20

25

120284 120384 120484 120584 120684 120784 120884 120984 121084 121184 121284

Time / sec

Pow

er /

Wat

t

Command PowerRead Power

• Drive Cycle testing shows faster degradation than steady-state testing

• Greater Pt Surface Area Loss• Larger Pt Particle

0

5

10

15

20

25

30

0 100 200 300 400 500 600 700 800 900 1000Time (hr)

True

Pt E

lect

roca

taly

tic

Surf

ace

Are

a (m

2 /g P

t)

Average

Linear (Average)

CV Set Points:100 mV/sCell Temp. = 80°CAnode/Cath. Humid. Temp. = 105/80°C

Average Rate of Loss of True Pt Electrochemically Active Surface Area = 71.4 cm2/g-Pt/hr

Catalyst Surface Area DuringConstant Power Operation Drive cycle ‘controls’ power

• Catalyst surface area decreases during testing• Pt Particles increase in size

Fuel cells in automotive applications cycle in power

Fuel Cell Program

Current Density (amps cm-2)0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Volta

ge (v

olts

)

0.0

0.2

0.4

0.6

0.8

1.0

prior to CV cyclingafter 300 cyclesafter 900 cyclesafter 1200 cyclesafter 1500 cycles

Electrocatalyst DurabilityPotential Cycling Measurements

Polarization Curves For Fuel CellCycled between 0.1V - 1.2V

Voltage (mV)0 200 400 600 800 1000

Cur

rent

(mA

)

-200

-150

-100

-50

0

50

100

prior to CV cyclingafter 300 cyclesafter 600 cyclesafter 900 cyclesafter 1200 cyclesafter 1500 cycles

Number of CV Cycles0 200 400 600 800 1000 1200 1400 1600%

Initi

al P

t Cat

alyt

ic S

urfa

ce A

rea

50

60

70

80

90

100

0.1V - 1.2V

Cycling:10 mV/sec 100 mV/sec

Active Pt Catalytic Surface Area

Characterization CVsAfter Cycling between 0.1V - 1.2V

Cell 80 °CH2 226% RHAir 100% RH

• Potential cycling:• Simulates drive cycle• Accelerated catalyst aging

Variables Examined:Potential RangeTemperatureCycles vs. Time (Scan Rate)Relative HumidityCatalyst Loading

Fuel Cell Program

Potential & Temperature Effect on Catalyst Growth

Number of CV Cycles0 200 400 600 800 1000 1200 1400 1600%

of I

nitia

l Pt C

atal

ytic

Sur

face

Are

a

0

20

40

60

80

100

0.1V - 1.2V0.1V - 1.2V0.1V - 1.2V0.1V - 1.0V0.1V - .96V0.1V - .96V

% of Initial Active Pt Surface AreaAfter 1500 CV Cycles

0 20 40 60 80 100

Cat

hode

Pt G

rain

Siz

e (n

m)

2

3

4

5

6

7

8

90.4V - 0.96V0.1V - 0.96V0.1V - 1.2V0.1V - 1.0V

Correlation Between Pt Surface Area and Pt Particle Size

Cell 80 °CH2 226% RHAir 100% RH

Active Pt Catalytic Surface Areavs. Cycling Voltage Range

50%

60%

70%

80%

90%

100%

0 100 200 300 400

Time Above 0.9V (min)

% In

itial

Sur

face

Are

a 10 mV/sec50 mV/sec

Comparison of Scan Rates: Time

Cycling from 0.1V - 0.96V

• Peak potential has large effect on particle growth

• Direct relation between XRD particle size and surface area

• Pt Sintering correlates with # of cycles

• Time at peak potential has lower correlation

Comparison of Scan Rates: Cycles

50%

60%

70%

80%

90%

100%

0 300 600 900 1200 1500

Number of Potential Cycles

% o

f Ini

tial S

urfa

ce A

rea 10 mV/sec

50 mV/sec

Cycling from 0.1V - 0.96V

Fuel Cell Program

Temperature Effect on Catalyst Growth

0

1

2

3

4

5

6

Initial 60 80 100 120

Cycling Temperature / oC

Pt P

artic

le S

ize

/ nm

50%

60%

70%

80%

90%

100%

60 80 100 120

Cycling Temperature / oC

% P

erfo

rman

ce Initial300 Cycles600 Cycles900 Cycles1200 Cycles1500 Cycles

Pt particle size Normalized performance decaybased on polarization curve current density at standard conditions: comparison at 0.65 V, 80 oC, 100 % RH

50%

60%

70%

80%

90%

100%

0 500 1000 1500 2000# of Potential Cycles

% In

itial

Act

ive

Surf

ace

Are

a

60 C80 C100 C120 C

After 1500 cycles

Increasing temperature leads to increased catalyst sintering

Fuel Cell Program

Humidity Effect on Catalyst Growth

Number of CV Cyles0 200 400 600 800 1000 1200 1400 1600

% o

f Ini

tial S

urfa

ce A

rea

50

60

70

80

90

100

10% RH50% RH100% RH

0

0.5

1

1.5

2

2.5

3

3.5

10 50 100

Cycling Relative Humidity / %

Pt P

artic

le S

ize

/ nm

0

0.2

0.4

0.6

0.8

1

1.2

10 50 100

Cycling Humidity / RH%

% P

erfo

rman

ce Initial300 Cycles600 Cycles900 Cycles1200 Cycles1500 Cycles

Normalized performance decaybased on polarization curve current density at standard conditions: comparison at 0.65 V, 80 oC, 100 % RH

Pt particle size After 1500 cycles to 1.0 V

• Pt particle growth rate increases with humidity

• Growth mechanism enhanced by H2O (Pt mobility?)

Active Pt Catalytic Surface Area

(Cell Temp 80 °C, Cycled from 0.1V - 0.96V)

Fuel Cell Program

X-ray scattering provides Pt particle size and size distribution

• X-ray peak widths• Larger particles show narrower diffraction peaks• Least square minimization to Pt peaks

Pt Cathode Catalyst Aging Simulated particle size distribution:- continuous line log-normal type profile (coalescence growth mechanism)-- dashed line for an Oswald Ripening profile Ascarelli, Contini, and Giorgi J. Appl. Phys., Vol. 91, No. 7, 1 April 2002

OswaldRipening

LogNormal

• Catalysts have non-monotonic size distributions

– Log normal distribution typical for synthesized particles (J. App. Physics V47 5 1976)– Oswald Ripening if catalyst grows by atom migration from small crystals to large ones– Coalescence mechanism for particle-particle growth

Crystal size • 38 Å average (distribution)• 51 Å average (volume)Particle size distribution- redCumulative distribution-blueAqua –log normal Brown –sample log normal

Pt-Cathode Cycled 0.1-1.0 V All Samples show Log Normal Distribution(up to 1.5V cycling)

Suggests catalysts grow by particle coalescence for all testing and conditions

Fuel Cell Program

Electrode Carbon Corrosion

0

1

2

3

4

5

6

7

8

0.96 V 1.0 V 1.2 V 1.5 V

Cycling Potential

XRD

Pt /

Car

bon

Rat

io

0

1

2

3

4

5

6

7

8

10% RH 50% RH 100% RH

Cycling Relative Humidity

Pt P

artic

le S

ize

/ nm

Carbon corrosion measured by monitoring XRD Pt/C ratios

0

1

2

3

4

5

6

7

8

60 C 80 C 100 C 120 C

Cycling Temperature

XRD

Pt/C

Rat

io

Potential Effect Relative Humidity Effect

Temperature Effect

80 oC 80 oC100 % RH

100 % RH

0.1 – 0.96 V

0.1 – 0.96 V• Carbon corrosion occurs at:

• High potentials (> 1.0 V)• Low Humidity

• Temperature not observed to have an effect at 100 % RH and 1.0 V

Fuel Cell Program

Electrocatalyst Size GrowthXRD analysis of electrocatalysts

0123456789

Condit

ioned

60 C

0.4-0

.9680

C 0.

4-.96

100 C

0.4-0

.96

120 C

0.4-0

.9680

C - 0

.11 - .

96

80 C

- 0.11

- 1.0

80 C

- 0.11

- 1.2

80 C

- 0.1-

1.2

80 C

- 0.1-

1.5

0.1 - 0

.96 10

mV/se

c

0.1 - 0

.96 50

mV/se

c10

% RH

50% R

H0.2

mg P

t 0.1-

0.96

0.2 m

g Pt 0

.4-0.9

6

Pt P

artic

le S

ize

• Pt particle growth on cathode occurs for steady-state, enhanced with cycling• No growth in anodes Pt crystallites• No Oswald ripening mechanism during cycling (potential cycling up to 1.5 V)• Particle growth increases with temperature• Particle growth increases with humidification• Pt loadings (0.2 mg/cm2 to 0.4 mg/cm2) did not effect the Pt sintering• Carbon corrosion is present at high potentials, and low humidification

•Cycling increases Pt particle growth rate over steady-state operation

•# cycles has larger effect on catalyst sintering than duration at high potential

•Catalyst grows by particle coalescence regardless of cycling conditions

600 cycles

1500 cycles

in situ XRD Measurements

Fuel Cell Program

non-destructive particle size analysisFramed MEA for non-destructive testing in XRDUtilizes 0.022” Nafion to separate electrocatalyst layers

Design for potential controlled in situXRD testing, possible ‘real-time’synchrotron particle size analysis

Nafion .022”Used to separate catalyst layer signals during XRD MEA

Incident x-raysDiffracted x-rays

Fuel Cell Program

Loss of Hydrophobicity in Gas Diffusion Layer (GDL)

3 µm10k×

60 µm500×

Graphitized Fiber (~10 µm)Sintered PTFE Nanoparticles

(200-500 nm dia.)

1 µm30k×

MEA (3-Layer)

Aggressive Operating Environment (O2, H2, e- Passage, Voltage, H2O(l), low Nafion® pH, Electro-oxidation, Electro-reduction, Time)

From Membrane → SO4-2, SO3

-2, F-

Catalyst → Pt (Co, Cr, etc.)From C-Support → Na+, Ca+2, Zn+2, Fe+3 (10-100 ppm), CO2 (CO3

-2)Others (Electrocatalytic Byproducts, etc.)

Attacking Species on PTFE and Graphite

Composite Microstructure

Typical Bilayer GDL Courtesy SGL Carbon and Celanese Ticona

Fuel Cell Program

Comparison of Hydrophobicity MeasurementsBulk Material vs. Single Fiber

Fiber (10.74 ±0.10 µm dia.)

Wilhelmy Plate Contact Angle for Bulk Plain TGP-H 120 was 121.8 ± 0.3°, which

is ~41° higher than for a single fiber

Why?

108

112

116

120

124

128

132

0 2 4 6 8 10Time (min)

Th

eta

(d

eg

ree

s)

Toray TGP-H 060 (10 wt% PTFE), Slope = -0.385 deg/min

Toray TGP-H 060 (Plain), Slope = -0.957 deg/min

Dynamic Sessile-Drop θ for Toray TGP-H 060

Wilhelmy θwas 80.3 ±

0.2° for single Toray fiber

Fuel Cell Program

GDL Hydrophobicity Aging

UntreatedToray Fiber

80.3o

Toray TGP-H090 17.2 wt% FEP

94.6o

Toray TGP-H-06017.0 wt% FEP

83.2o

Toray TGP-H-09016.7 wt% FEP

88.4o

80oC 80oC

460 hr 680 hrN2 Air

116.0

118.0

120.0

122.0

124.0

126.0

128.0

130.0

132.0

0.0 200.0 400.0 600.0 Time (seconds)

Con

tact

Ang

le (d

egre

es)

B1 Unaged #1 Toray TGP-H 090B2 Unaged #2 Toray TGP-H 090B3 460 hr aged in 60C DI water N2 SpargeB4 460 hr aged in 80C DI water N2 SpargeB5 680 hr aged in 60C DI water Air SpargeB6 680 hr aged in 80C DI water with Air Sparge

SessileDrop Spreading Contact Angle

Single Fiber

o Unaged #1 Toray TGP-H 090o Unaged #2 Toray TGP-H 090o 460 hr aged in 60oC DI Water with N2 Toray TGP-H 060o 460 hr aged in 80oC DI Water with N2 Toray TGP-H 090o 680 hr aged in 60oC DI Water with Air Toray TGP-H 060o 680 hr aged in 80oC DI Water with Air Toray TGP-H 060

Loss of hydrophobicity increases with temperature and oxidation

Fuel Cell Program

Decrease in Hydrophobicity of FEP-treated Toray GDL

N2 (60°C) Loss = −0.020 deg/hrN2 (80°C) Loss = −0.024 deg/hr

90

100

110

120

130

140

150

0 100 200 300 400 500 600 700Time (hr)

Con

tact

Ang

le (d

eg)

N2 at 60°C N2 at 80°C Air at 60°C Air at 80°C

Nitrogen Atmosphere

Air AtmosphereAir (60°C) Loss = −0.032 deg/hrAir (80°C) Loss = −0.029 deg/hr

Toray TGP-H (17 wt% FEP)

020406080

100120140160180

InitialMS1

FinalMS1

InitialMS2

FinalMS2

InitialNMS1

FinalNMS1

InitialNMS2

FinalNMS2

Con

tact

Ang

le • Hydrophobicity loss on microporouslayer

• No loss in hydrophobicity of non-microporous layer

0.30

0.40

0.50

0.60

0.70

0.80

0.90

40 50 60 70 80 90 100 110Anode/Cathode Humidifier T (°C)

Cur

rent

Den

sity

(A/c

m2)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Hig

h-Fr

eque

ncy

Res

ista

nce

(mΩ

*cm

2 )

Fresh GDLs

1006-hr-Aged GDLs

Fresh GDLs

1006-hr-Aged GDLs

Cell Temp. = 80°CCell Voltage = 0.702Anode/Cathode Gas Pressure = 15/15

HRF

CurrentDensity

MS = Microporous layerNMS = non-microporous layer

Fuel Cell Program

SPEAR (Surface Profile Analysis Reflectometer )

Substrate

50Å layerθi θf

ki koutQz

Qz =4π sin(θ)/λ10 -7

10 -5

10 -3

10 -1

0.00 0.10 0.20 0.30

Ref

lect

ivit

y

Neutron Reflectometry

SPEAR:Film thicknessRoughnessCoverage

1.E-10

1.E-09

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

0.005 0.015 0.025 0.035 0.045 0.055 0.065 0.075 0.085 0.095 0.105

Qz (Angstrom-1)R

efle

ctiv

ity

1100 EW Raw Data1100 EW Error950 EW Raw Data950 EW Error5 wt% Solution Spin Coated

33 vol% Nafion (1:2) with Ethanol.

δfilm,NS3 = 2π/∆Qz = 61.5 nm

δfilm,NF4 = 2π/∆Qz = 69.8 nm

Thin-Film Nafion/Si(Comparison between 1100 and 950 EW)

Nafion 1100EW / 950EW Scattering Length Densities (SLDs) = 4.23x10-6 / 3.95x10-6

1.E-09

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

0.005 0.015 0.025 0.035 0.045 0.055 0.065

Qz (Angstrom-1)

Ref

lect

ivity

Bare Glassy C Raw DataBare Glassy C ErrorNafion 1100 on Glassy C Raw DataNafion 1100 on Glassy C Error

5 wt% Solution Spin Coated 33 vol% Nafion (1:2) with

Ethanol

δfilm,NGC1 = 2π/∆Qz = 62.1 nm

Thin-Film Nafion1100EW/Glassy-C

Calculated Scattering Length Density (SLD) for Glassy C = 5.10x10-6, Surface Roughness = 79.7 ÅGlassy Carbon (~ only) acceptable surface to study PEM Fuel cell polymer/carbon interactions

Fuel Cell Program

Membrane Thinning

Groove

Mag: 10x15

Post-characterization of membrane• Membrane degradation exemplified by cross-over, and hole formation

• Post analysis shows membrane ‘thins’• Peroxide formation a key to

membrane degradation

01020304050607080

0 200 400 600 800 1000

Time / hr

Cro

ss-o

ver C

urre

nt /

mA

• Examining membrane properties:• H form of membrane compressed in flowfield exposed to inert gas and temperature.

• No backing layers, catalyst layer, proton conduction, etc.

H2 Cross-over vs. Testing Time

0

0.2

0.4

0.6

0.8

1

1.2

80 C dry 120 C dry 80 C Humid 120 C Humid

Rel

ativ

e M

embr

ane

Thic

knes

s CompressedUncompressed

Regardless of compression, membrane showed thinning with exposure to dry gases

Fuel Cell Program

Future Plans• Catalyst Durability / Characterization

– in situ XRD analysis of Pt particle growth period– Modeling of particle growth to correlate growth conditions– Pt equilibrium diagram with PEM fuel cell conditions with F-, sulfates, etc.– Carbon bonding interaction with Pt - develop stable Pt/C catalysts– Pt alloys with higher stability– Examine non-carbon electrocatalyst supports for durability

• Carbon Corrosion– Further examine carbon corrosion in electrocatalyst layers and GDL materials

• Component Interfacial Durability Property Measurements– GDL material interfacial contact with the MEA catalyst layer– examine Nafion / PTFE degradation and carbon bonding via neutron scattering

• Membrane Degradation– examine conditions leading to membrane thinning– examine conditions leading to membrane failure

Fuel Cell Program

Publications and PresentationsMicrostructural Changes of Membrane Electrode Assemblies during PEFC Durability Testing at High Humidity Conditions, Xie et al., Journal of The Electrochemical Society, 152 5 A1011-A1020 2005

Durability Study of Polymer Electrolyte Fuel Cells at High Humidity Conditions, Xie et al., Journal of The Electrochemical Society, 152 A104-A113 2005

Effects of Long-Term PEMFC Operation on Gas Diffusion Layer and Membrane Electrode Assembly Physical Properties, Wood et al., 206th Meeting of The Electrochemical Society, Honolulu, Hawaii, October 5th, 2004

Long-Term Performance Characterization of Proton Exchange Membrane Fuel Cells, Wood et al., 206th Meeting of The Electrochemical Society, Honolulu, Hawaii, October 5th, 2004

PEM FUEL CELL DURABILITY, Borup et al., FY 2004 DOE EERE Hydrogen Program Annual Report

DURABILITY ISSUES OF THE PEMFC GDL and MEA UNDER STEADY-STATE AND DRIVE-CYCLE OPERATING CONDITIONS, Wood et al., 2004 Fuel Cell Seminar, San Antonio Texas, Nov. 1-5

PEM Electrocatalyst Durability Measurements, Borup et al., To be presented at the Electrochemical Society, June 12 – 17 2005, Las Vegas NV

PEM Electrocatalyst Durability Measurements, Davey et al., To be presented at the Fuel Cell Seminar, 2005, Palm Springs, CA, Nov. 14 -18, 2005

MASS-TRANSPORT PHENOMENA AND LONG-TERM PERFORMANCE LIMITATIONS IN H2-AIR PEMFC DURABILITY TESTING, Wood et al., To be presented at the Fuel Cell Seminar, 2005, Palm Springs, CA, Nov. 14 - 18, 2005

Fuel Cell Program

Hydrogen Safety

The most significant hydrogen hazard associated with this project is:

Hydrogen leak in the hydrogen supply coupled with ignition leading to a significant hydrogen fire.

Fuel Cell Program

Hydrogen Safety

Our approach to deal with this hazard is:

Hydrogen and carbon monoxide room sensors are electrically and computer interlocked with the test stand power and the gas supplies.

H2 sets off the H2 sensors (set at 10% of LFL)H2 also sets off the CO sensors, (set at 30 ppm)

Limits H2 far from the flammable or explosive limit

Work has been reviewed through Los Alamos National Lab’s safety programs:Hazard Control Plan (HCP) - Hazard based safety reviewIntegrated Work Document (IWD) - Task based safety reviewIntegrated Safety Management (ISM)