MEA & Stack Durability for PEM Fuel Cells

2006 DOE Hydrogen Program Review

MEA & Stack Durability for PEM Fuel Cells

3M/DOE Cooperative Agreement No. DE-FC36-03GO13098

Project ID # FC8

3 Mike Hicks

3M Company May 16, 2006

This presentation does not contain any proprietary or confidential information

Overview

Timeline 9/1/2003 6/30/2007* 70% complete * Revised end date subject to

DOE approval

Budget Total $10.1 M

DOE $8.08 M Contractor $2.02 M

Funding received in FY05: $2.43 M

Funding for FY06: $2.60 M

Barriers & Targets A. Durability: 40k hrs

Team Members Plug Power Case Western Reserve

University University of Miami

Consultant Iowa State University

MEA & Stack Durability for PEM Fuel Cells 2 3 Fuel Cell Components

Objectives Develop a pathway/technology for stationary PEM fuel cell systems for enabling

DOEs 2010 objective of 40,000 hour system lifetime to be met

Goal: Develop an MEA with enhanced durability Manufacturable in a high volume process Capable of meeting market required targets for lifetime and cost Optimized for field ready systems 2000 hour system demonstration

Focus to Date MEA characterization and diagnostics MEA component development MEA degradation mechanisms MEA nonuniformity studies Hydrogen peroxide model Defining system operating window MEA and component accelerated tests MEA lifetime analysis


Approach To develop an MEA with enhanced durability .

Optimize MEAs and Components for Durability

Optimize System Operating Conditions to Minimize

Performance Decay

Utilize proprietary 3M Ionomer Improved stability over baseline ionomer

Utilize ex-situ accelerated testing to age MEA components Relate changes in component physical properties to changes in MEA

performance Focus component development strategy

Optimize stack and/or MEA structure based upon modeling and experimentation

Utilize lifetime statistical methodology to predict MEA lifetime under normal conditions from accelerated MEA test data


Accomplishments GDL Characterization

Developed new test equipment to measure capillary pressure in GDLs Membrane

Completed investigation of reinforced membranes reinforcement may not be necessary for membrane durability

Identified membrane failure mode and implemented solution to mitigate it Ongoing monitoring of membrane properties in accelerated tests

Membrane Degradation Mechanism Analyzed experimental and literature data more than just end group degradation Utilized ionomer model compounds to identify likely points of attack and provide insight

into ionomer degradation mechanism Developed initial hydrogen peroxide model to study peroxide in operating fuel cell

MEA Nonuniformity Studies Completed 121-channel segmented cell and investigated the effects of flow rate, load

setting and GDL type; determined high gas stoichiometry yields current uniformity Utilized theoretical 3D fuel cell model to investigate effects of catalyst, membrane and

GDL nonuniformity; determined that electrode defects result in highly, nonuniform current distribution

System Test Initiated Saratoga system test with a preliminary, durable MEA design

MEA Lifetime Modeling Demonstrated that load profile affects MEA durability Developed initial lifetime prediction model to estimate MEA lifetime relative to DOEs 2010

stationary system goals Related initial fluoride ion to lifetime method to increase sample throughput


GDL Characterization Capillary Pressure Background Solution Measured GDL permeability in humid and Design your own instrument

dry air CWRU has designed, machined and Humid air yields lower gas permeability assembled the sample holders, load cell

Pores fill with water and strain sensor CWRU collaborated with Porous Materials

Problem Inc, Ithaca, NY to fabricate the instrument Need technique to characterize water PMI will integrate the syringe pump, the

transport in GDL pores press and automation There are no available instruments for

measuring capillary pressures for hydrophobic porous media

measuring Capillary Forces in hydrophobic GDLs

GDLs

Developed an instrument for

New method to characterize


Reinforced Membrane Activities Membrane Stress Model Evaluation of Various Reinforcing Members

Highest Stress Lowest Stress

Lands

Channels

Hypothesis 0 20 40 60 80

100 120 140 160 180 200

0 5 10 15 20 25 30 35 40

Tear (MPa)

I

m

p

e

d

a

n

c

e

(

m

c

m

2

)

conductivity than neat Nafion

with 3M Ionomer

- Need reinforcing member to carry stress to eliminate mechanical failure or reduce mechanical failure rate

Desired Result stronger and higher

Lines 3M Cast Nafion Membrane Symbols Various reinforced membranes

RH Cycle Test to Evaluate HypothesisTest Conditions: 80C Cycle equally between 0 and 150% RH

MEA (electrode and GDL) made Time to failure with: (hours) DuPont Nafion (NR-111)1 260 330 Ion Power Nafion (N111-IP)1 1330 + Gore Primea1 400 470 3M Cast Nafion (1000 EW) 1200 +

Neat membrane most durable

props and durability predict

mechanical durability predict

mechanical durability Less shrinking does not correlate to

more mechanical durability

No relationship between mechanical

Tensile test does not

Tear resistance does not

What is the benefit of reinforcement? 1. Gittleman et al, Fall AIChE Meeting, October 2005.


Mitigation of Membrane Edge Failure in Modules Problem In module testing, observe infant

Active Area

Site of mortality of MEAs due to edge failure at edge the membrane catalyst interface failure

Solution Developed edge protection component

for MEA

R

e

l

a

t

i

v

e

M

E

A

F

a

i

l

u

r

e

R

a

t

e

120

100

80

60

40 No Failures

20

0

w/o Edge Protection w/ Edge Protection

failure mode

solution to significantly reduce infant mortality failure rate

Identified MEA

Implemented a


3M Ionomer Membrane Properties vs Decay Membrane Aging Procedure

Pre-condition w/

Received H+ Form Membrane H+ Form

H2SO4 (0.1M) Ion exchange w/ FeSO4 (0.1M) 70C, 1 hour Fe(II) Form

70C, 1 hour

Degraded MembraneFe(II) Form

H2O2 (0.1M) 70C, ~ 35 hours H2SO4 (0.1M) Ion exchange w/

70C, 2 hours

Measure degraded membrane properties over time

As Received H+ Form Degraded Sample @ 125 hrs

132C

125C

131C

Mechanical

20 40 60 80 100 120 140 160 180 0.0

0.1

0.2

0.3

0.4

0.5

0.6

T

a

n

d

e

l

t

a

Temperature [C]

experiments in progress

after 125 hrs

0 100 200 300 400 500 600 0

20

40

60

80

100

120

W

e

i

g

h

t

[

%

]

Temperature [C]

Thermal Gravimetric

Dynamic

Analysis

Aging

No change Analysis


Membrane Decay Mechanism Via Model Compounds 208th ECS Meeting, Abstract 1195,

Non-zero intercept

mechanism(s)

F

-

g

e

n

e

r

a

t

e

dConventional Wisdom: Los Angeles, CA, October 2005 H2O2 generated during fuel cell

operation HO or other radicals are

attacking species -COOH end group unzipping

primary route 0 0

Demands other degradation

[ -COOH] Investigate alternative degradation mechanism(s) via

model compounds reactive sites

Utilize analytical capabilities Better isolation of effect from different Age MCs via Fentons test or UV light (200 - 2400 nm @ 100W)

MC1 MC2 MC3 O O O

F F2 F2 F F2 F2 F2 F2 F2 F2 F2HO C C O C C CF3 HO C C O C C C C SO3H HO C C C C SO3H

CF3 CF3

MC4 MC7 MC8

F3C F2 C C6 OH F3C

F2 C O

F2 C

F2 C

F2 C SO3H F3C

F2 C O

F2 C

F C O

F2 C

F2 C SO3H

O CF3


MC3 > MC1 MC2 > MC4 > MC7 & MC8 MC3 MC1 MC2

Model Compounds Relative Degradation Rates

HO C

O

F2 F2 F2 > C F C O O

CF3

HO F2 F2 HO C F C O

CF3

O

F2 F2 F2 F2 C C C SO3H C C CF3 C C C C SO3H

MC4 MC7 MC8

O F2

O F2 F

C O

CF3

F2 F2F2 F2C OH

O

F2 F2 F2> C C C SO3HCF3C F3C C C C C SO3H F3C C>6

effect

products?

& MC2

hydrolysis

COOH containing MCs exhibit low stability Comparison of MC3 & MC4

Is it really a reactivity effect or solubility

Is there a change in reactivity hydrolysis

Hydrolysis observed (by NMR) for MC1

Need to evaluate MC7 & MC8 for

Identified MC1 & MC2 Reaction Products O O O

C C F3C

COH F3C CF2

MC3 Isomer Degradation6 7 11

O CF3O F F 1 3

HO CF2 CF2

CF2 4 CF 10 SO3H

2 SO3H HO CF SO3H HO

O CF3 F F 8 95

Same degradation rate Decarboxylation is rate determining step


Membrane Decay Mechanism Hydrogen Peroxide Model Objective To define simple model to study peroxide behavior in an MEA Equations:

(CH O ) = Rate of production electrochemical +Chemical recombination )d (dt 2 2+ Rate of consumption

+ electrochemical reductionIonomer degradation + catalytic disproportionation

+ Transport through the electrode Diffusion +Convection )(

Peroxide to membrane

Peroxide Concentration Profile as f(L) O2 inlet No peroxide 0.75 V =

Z= Z= 0 1

Experiments to Determine Input Parameters 1. Rate of Peroxide Production2. Rate of Peroxide Disproportionation

Model provides insight into hydrogen peroxide distribution in an operating fuel cell and the degradation of ionomer by hydrogen peroxide

Geometry Model Output


MEA Nonuniformity Studies Motivation - MEA Durability Is MEA durability a function of current

distribution/uniformity?

C

u

r

r

e

n

t

D

e

n

s

i

t

y

(

A

/

c

m

2

)

4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Increasi urrentng avg. cV. Gurau, H. Liu and S. Kakac, A Two Dimensional Non-Isothermal Mathematical Model for Proton Exchange Membrane Fuel Cells, AIChE Journal, Vol. 44 (11), pp. 2410 2422, 1998

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless Channel Length

Approach Measure experimentally segmented cell Theoretical modeling


Segmented Cell

Inlet

Outlet A B C D E F G H I J K

1 2 3 4 5 6 7 8 9 10 11

0 5 10 15 20 25

V

o

l

t

a

g

e

(

V

)

Current (A)

100 sccm Air 200 sccm Air 500 sccm Air

1000 sccm Air

210 sccm O2

50 cm2, 121 segments

Validation of Cell Design

0.60

0.65

0.70

0.75

0.80

0.85

0.90

0.95 Filled Symbols Sum of Individual Segments Hollow Symbols Fuel Cell (Segments shorted together)

F

r

a

c

t

i

o

n

o

f

T

o

t

a

l

C

u

r

r

e

n

t

a

t

0

.

6

6

V

(

S

e

g

m

e

n

t

C

u

r

r

e

n

t

/

T

o

t

a

l

C

u

r

r

e

n

t

)

Effect of Air Flow Rate on Current Distribution

O2 Utilization = 0.99 0.96 0.56 0.31

200 sccm100 sccm 500 sccm 1000 sccm0.00

0.04

0.08

0.12

0.16

0.20

0.24

Inlet Outlet MEA & Stack Durability for PEM Fuel Cells 14 3 Fuel Cell Components

at high stoichiometry for uniformity

load

Cell design validated Design fuel cell systems to operate

Recently completed 121 channel


0

0.1

0.2

0.3

0.4

0.5

0.6

c

u

r

r

e

n

t

d

e

n

s

i

t

y

0

0.2

0.4

0.6

0.8

x

1

1.02

1.04

1.06

y

0.78590.68110.57630.47160.36680.26200.15720.0524

CL +1

0

0.1

0.2

0.3

0.4

0.5

0.6

c

u

r

r

e

n

t

d

e

n

s

i

t

y

0

0.2

0.4

0.6

0.8

x

1

1.02

1.04

1.06

y

0.58900.51050.43200.35340.27490.19630.11780.0393

CL -1

0

0.2

0.4

0.6

0.8

1

1.2

c

u

r

r

e

n

t

d

e

n

s

i

t

y

0

0.2

0.4

0.6

0.8

x

1

1.02

1.04

1.06

y

1.13380.98260.83140.68030.52910.37790.22680.0756

CL +3

0

0.1

0.2

0.3

0.4

0.5

0.6

c

u

r

r

e

n

t

d

e

n

s

i

t

y

0

0.2

0.4

0.6

0.8

x

1

1.02

1.04

1.06

y

0.58900.51050.43200.35340.27490.19630.11780.0393

CL -1/2

Hydrogen

Air

Collector Plate

Membrane

x

y

z

Cathode catalyst layerGas diffusion layer

Anode catalyst layer Gas diffusion layer

Collector Plate

Gas channel

Gas channel

MEA Nonuniformity StudiesVariables Investigated Ionic Conductivity Catalyst Loading GDL Porosity Electrode Thickness Membrane Thickness GDL Thickness

Electrode Thickness

Surface defects resulted in highly non-uniform current distribution


Objective Investigate possible interaction between system design and durable MEA design

0

10

20

30

40

50

60

70

80

0 100 200 300 400 5000

0.2

0.4

0.6

0.8

1

Stack DC voltage System Efficiency

Cell Ratio

Saratoga System Test First Durable MEA TestingS

t

a

c

k

V

o

l

t

a

g

e

(

V

)

S

y

s

t

e

m

E

f

f

i

c

i

e

n

c

y

(

%

)

S

t

a

c

k

C

e

l

l

R

a

t

i

o

Run Hours No negative MEA System interaction Program approach validated

System Restarts

Statistical MEA Lifetime Predictions from Accelerated Test Data

00

0.50.5

Solid Lines Dashed Lines

I

(

A

/

c

m

2

)

0

0.5

Dotted Lines

I

(

A

/

c

m

2

) Modified Load CycleConstant Load Cycle

I

(

A

/

c

m

2

)Near-OCV Load Cycle

Time Time Time


Model Assumes Class model load profiles

Accelerated Lifetime (Hrs)

F

r

a

c

t

i

o

n

F

a

i

l

i

n

g

Decreasing Stress

Predicted Lifetime 70C 100% RH

ibution

.001

.003

.005

.01

.02

.03

.05

.1

.2

.3

.5

.7

.9 .98

10^01 10^02 10^03 10^04 10^05 10^06

Censored data No censored data

Baseline Components

.001

.003

.005

.01

.02

.03

.05

.1

.2

.3

.5

.7

.9 .98

10 20 50 100 500

F

r

a

c

t

i

o

n

F

a

i

l

i

n

g

Comparison of MEA Designs

~ 4x New 3M PEM MEAs

Baseline MEAs

Weibull distribution Arrhenius for temp Humidity model for RH

Lifetime probability distrReasonable predictive values No OCV load cycle offers ~ 13X lifetime improvement New MEAs with 3M ionomer ~ 4x more durable

200 1000

Accelerated Lifetime (Hrs)

Fluoride Ion Mapping of Accelerated Test Data 1.0E+05

A

c

c

e

l

e

r

a

t

e

d

L

i

f

e

t

i

m

e

(

H

r

s

)

Predicted Lifetime New 3M PEM MEAs 70C 100% RH Hollow symbols: In-Progress

R2 = 0.77

R2 = 0.89

R2 = 0.83

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

0.00 0.01 0.10 1.00 10.00

Initial Fluoride Release (g/min)

I

(

A

/

c

m

2

)

3M PEM MEAs under accelerated, near-OCV load cycle test conditions

Time

Pathway towards ~ 20,000 hour MEA lifetime with

Means to increase sample throughput

Near-OCV Load Cycle


0

0.5

Future Work To the End of the Project MEA & Stack Development & Testing

MEA Component optimization & integration 3M Saratoga stack tests Plug Power Complete MEA evaluation in modules/single cells Plug Power Select Final stack and MEA design and test Plug Power/3M

MEA Degradation Studies Peroxide model CASE

Incorporate realistic kinetic and transport parameters Model compounds CASE

Determine degradation kinetic constants MEA nonuniformity studies 3M/Plug/University of Miami

Determine operating conditions/MEA designs that yield current distribution uniformity

Post mortem analysis CASE/Plug Power Mechanical properties-morphology relationship CASE

MEA Statistical Lifetime Predictions MEA lifetime modeling 3M/Plug Power


Project Summary Relevance:

Approach:

Progress:

Developing MEA and system technologies to meet DOEs year 2010 stationary durability objective of 40,000 hour system lifetime. Providing insight to MEA degradation mechanisms.

Two phase approach (1) optimize MEAs and components for durability and (2) optimize system operating conditions to minimize performance decay.

Demonstrated pathway towards 20,000 hour MEA lifetime with 3M PEM MEAs under accelerated near-OCV load cycle test conditions. Initiated durable MEA-stack system tests.

DOE 2010 FY 05 FY 06 Goal (hrs)

Accelerated Lifetime Predictions (hrs) 16,000 > 20,000 40,000

Technology Transfer/Collaborations: Active partner with CWRU, Plug Power and the University of Miami. Presented 9 presentations and 2 papers on work related to this project in last 12 months.

Future Work: Complete studies on MEA degradation mechanism. Select final MEA and stack design and test system for 2,000 hours.


Publications and Presentations M. Yandrasits, Mechanical property measurements of PFSA membranes at elevated temperatures and

humidities, 2nd International Conference on Polymer Batteries and Fuel Cells, Las Vegas, NV, June 2005. D. Stevens, M. Hicks, G. Haugen, J. Dahn, Ex situ and in situ stability studies of PEMFC catalysts: Effect of

carbon type and humidification on degradation of the carbon, J. Electrochem. Soc., 152 (12), A2309 (2005). D. Schiraldi and C. Zhou, Chemical durability studies of PFSA polymers and model compounds under mimic

fuel cell membrane conditions, 230th ACS Meeting, Washington, D.C., August 2005. M. Hicks, D. Pierpont, P. Turner, T. Watschke, M. Yandrasits, Component Accelerated Testing and MEA

Lifetime Modeling, 2005 Fuel Cell Testing Workshop, Vancouver, BC, September 2005. J. Dahn, D. Stevens, A. Bonakdarpour, E. Easton, M. Hicks, G. Haugen, R. Atanasoski, M. Debe, Development

of Durable and High-Performance Electrocatalysts and Electrocatalyst Support Material, 208th Meeting of The Electrochemical Society, Los Angeles, CA, October 2005.

D. Pierpont, M. Hicks, P. Turner, T. Watschke, Accelerated Testing and Lifetime Modeling for the Development of Durable Fuel Cell MEAs, 208th Meeting of The Electrochemical Society, Los Angeles, CA, October 2005 (presentation and paper).

M. Hicks, K. Kropp, A. Schmoeckel, R. Atanasoski, Current Distribution Along a Quad-Serpentine Flow Field: GDL Evaluation, 208th Meeting of The Electrochemical Society, Los Angeles, CA, October 2005 (presentation and paper).

G. Haugen, D. Stevens, M. Hicks, J. Dahn, Ex-situ and In-situ Stability Studies of PEM Fuel Cell Catalysts: the effect of carbon type and humidification on the degradation of carbon supported catalysts, 2005 Fuel Cell Seminar, Palm Springs, CA, November 2005.

D. Pierpont, M. Hicks, P. Turner, T. Watschke, New Accelerated Testing and Lifetime Modeling Methods Promise Development of more Durable MEAs, 2005 Fuel Cell Seminar, Palm Springs, CA, November 2005.

M. Hicks, R. Atanasoski, 3M MEA Durability under Accelerated Testing, 2005 Fuel Cell Durability, Washington, DC, December 2005.

Z. Qi, Q. Guo, B. Du, H. Tang, M. Ramani, C. Smith, Z. Zhou, E. Jerabek, B. Pomeroy, J. Elter, "Fuel Cell Durability for Stationary Applications - From Single Cells to Systems, 2005 Fuel Cell Durability, Washington, DC, December 2005.


Response to 2005 Reviewers Comments Need to evaluate catalyst degradation; how does catalyst degradation affect

overall MEA durability? Reported results of commercial Pt/C catalyst durability and degradation at 2004

HFCIT Review Project not focused on development of Pt/C catalyst; separate 3M/DOE project

focused on catalyst durability (3M NSTF catalyst) Need additional characterization of membrane physical properties and effect of

aging on these properties Initiated task on measuring membrane mechanical properties & morphology as a

function of aging Need to relate effect of component improvements to overall MEA improvements.

What component improvement added most value to MEA lifetime? Integration of components is critical in terms of obtaining good MEA durability Considering possible patent applications

Need to work on reinforced membranes. Have evaluated reinforced membranes; results to be presented in the future Development out of scope of project some work done at expense to 3M

Better description of lifetime model Using std lifetime statistical analysis techniques; see W.Q. Meeker and L.A.

Escobar, Statistical Methods for Reliability Data, John Wiley and Sons, Inc. (1998) Need to address other targets (cost/performance) in concert with durability

Reported performance at the 2005 DOE Hydrogen Program Review Cost not a primary objective; it is used as a metric when deciding options

Too much emphasis on fluoride ion release. Disagree Very strong relationship between fluoride release and MEA lifetime


Critical Assumptions and Issues Validation of lifetime model analysis method

Testing baseline samples at normal test conditions Comparison to field test data

Increasing sample throughput of improved durability MEAs New, durable MEAs last too long Use initial fluoride ion release as metric (reduces test time) Plug Power test equipment online (adds more test equipment)

Understanding role of peroxide Initial peroxide lifetime model established

Demonstrate benefit of new, more durable MEAs Start lifetime accelerated tests of new MEAs Apply lifetime model to new MEAs


MEA & Stack Durability for PEM Fuel CellsOverviewObjectivesApproachAccomplishmentsGDL Characterization Capillary PressureReinforced Membrane ActivitiesMitigation of Membrane Edge Failure in Modules3M Ionomer Membrane Properties vs DecayModel Compounds Relative Degradation RatesMEA Nonuniformity StudiesSegmented CellMEA Nonuniformity StudiesVariablesSystem Test First Durable MEAFluoride Ion Mapping of Accelerated Test DataFuture Work To the End of the ProjectProject SummaryPublications and PresentationsResponse to 2005 Reviewers CommentsCritical Assumptions and Issues

MEA & Stack Durability for PEM Fuel Cells

Documents

membrane durability

stack durability

operating fuel cell

pem fuel cells

fuel cell components

d fuel cell model

hour system lifetime

lifetime modeling