Andrew J. Steinbach3M Company, St. Paul, MN

U.S. DOE 2017 Annual Merit Review and Peer Evaluation MeetingWashington, DCJune 7, 2017

Highly Active, Durable, and Ultra-low PGM NSTF Thin Film ORR Catalysts and Supports

This presentation does not contain any proprietary, confidential, or otherwise restricted information

Project FC143

Project OverviewTimeline

Budget

Partners

DOE 2020 Technical Targets

Project Start: 1/1/2016Project End: 3/30/2019

Barriers

Total DOE Project Value: $4.360MM*

Total Funding Spent: $1.073MM*

Cost Share Percentage: 23.72%*Includes DOE, contractor cost share and FFRDC funds as of 3/31/17

Johns Hopkins University (J. Erlebacher)Purdue University (J. Greeley)Oak Ridge National Laboratory (D. Cullen)Argonne National Laboratory (D. Myers, J. Kropf)

A. DurabilityB. CostC. Performance

PGM total content (both elec.): 0.125 g/kWPGM total loading: 0.125 mg/cm2

Loss in initial catalytic activity: < 40%Loss in performance at 0.8A/cm2: < 30mVLoss in performance at 1.5A/cm2: < 30mVMass activity (0.90VIR-FREE): 0.44A/mg

2

Project Objective and RelevanceOverall Project Objective

Develop thin film ORR electrocatalysts on 3M Nanostructured Thin Film (NSTF) supports which exceed all DOE 2020 electrocatalyst cost, performance, and durability targets.

Project RelevanceORR catalyst activity, cost, and durability are key commercialization barriers for PEMFCs.3M NSTF ORR catalysts have intrinsically high specific activity and support durability, and approach many DOE 2020 targets in state-of-the-art MEAs.Project electrocatalysts will be:• compatible with scalable, low-cost fabrication processes.• compatible into advanced electrodes and MEAs which address traditional NSTF challenges:

operational robustness, contaminant sensitivity, and break-in conditioning.

Overall ApproachEstablish relationships between electrocatalyst functional response (activity, durability), physical properties (bulk and surface structure and composition), and fabrication processes (deposition, annealing, dealloying) via systematic investigation.Utilize high throughput material fabrication and characterization, atomic-scale electrocatalyst modeling, and advanced physical characterization to guide and accelerate development.

3

BP1 Milestones and Go/No-GoTask Number, Title Type

(M/G), Number

Milestone Description/ Go/No-Go Decision Criteria Status Date (Q)

4.1 Proj. Management M4.1 Intellectual Property Management Plan Completed, Signed 100% 0

3.1 HT FabricationM3.1.1 HT Catalyst Deposition Process Reproducible 100% 1M3.1.2 HT Catalyst Treatment Process Reproducible 60% 2

3.2 HT CharacterizationM3.2.1 HT EC Characterization Reproducible 100% 2M3.2.2 HT XRD Characterization Reproducible 100% 3

3. HT Development M3.1 HT Activity, Area Agrees w/ Homogenous MEA 65% 33.2. HT Characterization M3.2.3 HT EXAFS/XANES Characterization Reproducible 100% 4

2.1 KMC Refinement G2.1.1 KMC predicts specific area and composition trends of PtxNi1-x(≥3 mole fractions) during EC dealloying. 100% 4

2.2 DFT Refinement G2.2.1 DFT predicts specific activity trends of PtxNi1-x (≥3 mole fractions). 100% 5

1.2 Catalyst EC Characterization

G1.2.1 (PROJ)

Electrocatalyst achieves (≥0.44A/mg and ≤50% mass activity loss) or(≥0.39A/mg and ≤ 40% mass activity loss).

100% 4

• Validation of high throughput catalyst development largely complete; delay in (key) M3.1 resolved; now progressing.• Refinement of KMC and DFT models complete; models validated.• Project BP1 GNG (activity and durability) achieved.

4

Task Number, Title Type (M/G),

Number

Milestone Description/ Go/No-Go Decision Criteria Status Date (Q)

3.2 HT Characterization M3.2 Combinatorial/HT Physical Char. Agrees w/ Homogenous Catalyst 60% 62.3 New NPTF Simulation M2.3.1 Model(s) predict ≥2 new NPTF alloys w/ ≥30m2/g area 40% 62.4 New UTF Simulation M2.4.1 Model(s) predict ≥2 new UTF alloys w/ ≥4mA/cm2 activity 50% 72.5 Dur. Cat. Simulation M2.5.1 Model(s) predicts ≥4 alloys w/ ≤ 30% loss under ASTs. 0% 82. Catalyst Simulation M2.1 Model(s) predict ≥4 alloys w/ ≥0.8A/mg and ≤ 20% loss w/ ASTs. 0% 8

1.5 Catalyst Integration G1.5.1 (PROJ)

Electrocatalyst achieves ≥ 0.6A/mg, ≤ 30% loss, and MEA PGM content ≤ 0.13 g/kW @ 0.70V 85% 8

• BP2 milestones mostly reflect KMC and DFT simulations of new alloy candidates with improved activity, area, and durability.

• BP2 Go/No-Go reflects increased activity, durability, and MEA performance. • Current BP2 GNG status (UTF PtNi) is 85%: 0.37A/mgPGM, 43% loss, and 0.13g/kW.

BP2 Milestones and Go/No-Go

5

Approach – Two Distinct Thin Film Electrocatalyst MorphologiesNanoporous Thin Film (NPTF)

NPTF Approach:1. Maximize area and minimize leachable TM

by structure, composition, and process optimization.2. Stabilize against nanopore coarsening and TM

dissolution via additives.

UTF Approach:1. Develop active, stable and thin surface facets

by structure, composition, and process optimization.2. Increase catalyst absolute area by integration with

higher area NSTF supports to enable increased absolute power density.

Ultrathin Film (UTF)

NPTF PtNiIr “#3”Status Project Target

Mass Activity (A/mgPGM) 0.38 0.80Specific Area (m2/gPGM) 19.3 30

Spec. Activity (mA/cm2Pt) 2.1 2.6

Mass Act. Change (%) -45 -20

UTF PtNiIr “#2”Status Project Target

Mass Activity (A/mgPGM) 0.44 0.80Specific Area (m2/gPGM) 16.1 20

Specific Activity (mA/cm2Pt) 2.7 4.0

Mass Act. Loss (%) -45 -20

6

Status versus DOE and Project Targets

2020 Target and Units

Project Target UTF PtNi, PtNiIr NPTF PtNiIr2016 2017 2016 2017

Platinum group metal (PGM) total content (both electrodes)

0.125 g/kW(Q/∆T ≤ 1.45)

0.1(@ 0.70V) NA 0.132 0.18 0.151

PGM total loading (both electrodes) 0.125 mg/cm2 0.10 NA 0.0772 0.127 0.1221

Loss in catalytic (mass) activity 40 % 20 NA 454 40 453

Loss in performance at 0.8 A/cm2 30 mV 20 NA 234 28 213

Loss in performance at 1.5 A/cm2 30 mV 20 NA NA NA NAMass activity @ 900 mViR-free 0.44 A/mg

(MEA) 0.80 0.31 0.444 0.24 0.383

12017(Jan.) NPTF BOC MEA. 0.016mgPt/cm2 NSTF anode, 0.090mgPGM/cm2 NPTF PtNiIr cathode, 0.016 mgPt/cm2 cathode interlayer. 22017 (Jan.) UTF BOC MEA. 0.015mgPt/cm2 NSTF anode, 0.046mgPGM/cm2 UTF PtNi cathode, 0.016 mgPt/cm2 cathode interlayer.3NPTF PtNiIr/NSTF cathode from 2017 (Jan.) NPTF BOC MEA, 0.09mgPGM/cm2 after 30k Electrocatalyst AST cycles. 4UTF PtNiIr. BOL and after 30k Electrocatalyst AST CyclesPGM content values at 90°C cell, 1.5atmA H2/Air, 0.70V (Q/∆T = 1.41kW/°C)GREEN: Meets or exceeds DOE 2020 target. YELLOW: Within ca. 15% of DOE 2020 target.• 2017 catalysts approaching or meeting many DOE 2020 catalyst and MEA targets.• PGM content and loading values include cathode interlayers (16µg/cm2) for operational robustness.

7

Accomplishments and Progress – New NPTF Alloy Development

8

NPTF Status Values. 6 Alloys(Ni, B-F) x 3 Mole Fractions. 0.08-0.10mgPGM/cm2. 50cm2 MEA Format.

Pt/V Pt PtNi B C D E F0

5

10

15

2040

Pt/V Pt PtNi B C D E F0.0

0.1

0.2

0.3

0.4

0.5Mass Activity (A/mg

PGM)

Pt/Vulcan

Pt/Vulcan

Specific Area (m2/gPGM) • Status values reflect current

optimal composition and processing for each alloy.

• Highest mass activity (0.47A/mgPGM) to date with annealed, dealloyed PtNi.

• Mass activity variation mostly due to ca. 4x variation in specific area.

0.0 0.5 1.0 1.5 2.00.40.50.60.70.80.91.0

0.00 0.05 0.100.800.820.840.86

0.880.90

CPtNi

PtB

F

DE

Cel

l V (V

olts

)

J (A/cm2)

Pt/V

80/68/68C, 1.5/1.5atmA inlet, CS2/2.5 H2/Air, 120s/pt. 80/68/68C, 1.5/1.5atmA inlet, CS2/2.5 H2/Air, 120s/pt.

E

DCPtPt/V

BF

PtNi

Cell

V (V

olts

)

J (A/cm2)

• At lower loading, several NPTF candidates yield improved performance vs. Pt/V.

• Up to 90mV @ 1A/cm2.

• Up to 35mV @ 0.02A/cm2

• New alloys “B”, “F” have improved high current H2/Air performance over PtNi, but lower kinetic performance.

H2/Air Performance - NPTF (0.08-0.09 mgPGM/cm2) vs. Pt/V (0.10mgPGM/cm2). 80oC, 1.5atmA.

0.0 0.4 0.8 1.2 1.60.4

0.5

0.6

0.7

0.8

0.9

PtNiIr #3PtNiIr #4

PtNiIr #1

PtNiIr #2

PtAlloy/C

Cell

Volta

ge (V

olts

)

J (A/cm2)

Accomplishments and Progress – NPTF PtNi Stabilization via IrIr Impact on BOL Activity and DOE Cat. AST Durability. 0.08-0.11mgPGM/cm2. 50cm2 MEA Format

Ir Content (Arb.) Ir Content (Arb.)00

0.1

0.2

0.3

0.4

0.5

PtNiIr #4

Pt Alloy/CPtNiIr #2

PtNiIr #3

PtNiIr #1

Mass Activity (A/mgPGM)

-80

-60

-40

-20

0

PtNiIr #4

PtNiIr #3PtNiIr #2

PtNiIr #1

Pt Alloy/C

Mass Activity % Change

• Ir integration with NPTF PtNi reduces mass activity loss after Electrocatalyst AST.

• 70% loss (Ir-free) to as little as 45% loss (w/ Ir).

• Ir primarily improves specific activity retention.

• Beginning of life PGM mass activity varies from 0.28-0.42A/mgPGM; depends on Ir integration method.

Ir Impact on H2/Air Performance Retention After Electrocatalyst AST

• Pre-AST performance depends on Ir integration method (impacts dealloying).

Before AST

Ir Content (Arb.)0

-60

-40

-20

0

20

80/68/68C, 1.5/1.5atmA H2/Air, CS2/2.5PtAlloy/C

PtNiIr #4

PtNiIr #1 PtNiIr #3

PtNiIr #2

H2/Air∆ V @ 0.32A/cm2 (mV)

Ir Content (Arb.)0 -60

-40

-20

0

20

40

80/68/68C, 1.5/1.5atmA H2/Air, CS2/2.5PtAlloy/C

PtNiIr #4

PtNiIr #1PtNiIr #3PtNiIr #2

H2/Air J @ 0.50V % Change

• Monotonic improvement w/ Ir content for #1, #3 integration; all < 30mV loss.

• PtAlloy/C: 50mV loss.

• PtNi, PtNiIr limiting current density change: -10% to +15% (improves).

• PtAlloy/C: -45% (local transport)

30k (0.6-1.0V, 50mV/s)

9

Accomplishments and Progress – New UTF Alloy DevelopmentUTF Status Values. 6 Alloys (Ni, B-F) x 4 Mole Fractions. 25-30µgPGM/cm2. 50cm2 MEA Format.

• Status values reflect optimal composition and processing for each alloy.

• Three alloys identified with >0.38A/mgPGM and > 2.5mA/cm2

Pt.• ~7x higher specific activity than Pt/V.

• 50% specific activity gain needed to achieve 0.8A/mg project target.• Approach: Pt skin optimization.Pt/V Pt PtNi B C D E F0.0

0.1

0.2

0.3

0.4

0.5Mass Activity (A/mgPGM)

Pt/Vulcan

Pt/V Pt PtNi B C D E F0

1

2

3

Pt/Vulcan

Specific Act. (mA/cm2Pt)

0.0 0.5 1.0 1.5 2.00.30.40.50.60.70.80.9

0.00 0.05 0.100.780.800.820.840.860.88

C

PtNi

Pt

B

F

D

E

Cell

V (V

olts

)

J (A/cm2)

0.05Pt/V

80/68/68C, 1.5/1.5atmA inlet, CS2/2.5 H2/Air, 120s/pt. 80/68/68C, 1.5/1.5atmA inlet, CS2/2.5 H2/Air, 120s/pt.

E

D

CPt

Pt/V

BF

PtNi

Cell

V (V

olts

)

J (A/cm2)

H2/Air Performance - UTF (0.025-0.030 mgPGM/cm2 vs. Pt/V (0.05mgPGM/cm2). 80oC, 1.5atmA • Four UTF alloys yield improved

performance vs. Pt/V, with 40-50% less PGM loading.

• Up to 69mV @ 0.5A/cm2

• Up to 43mV @ 0.02A/cm2.

• Absolute performance increase needed to achieve targets.

• UTF development phases:1. Optimize electrocatalyst

(activity, durability, ECSA).2. Integrate with higher area supports to

allow increased loading (0.08mg/cm2) and absolute cathode surface area.

10

Accomplishments and Progress –UTF PtNi Property Correlations

11

Structure-Activity Correlations (Composition) - ANL Surface Strain Mapping-ORNL

• (One example): Pt skin on PtNi is thick and relaxed relative to bulk.

• Challenging measurement/analysis.• Not representative of all UTF PtNi.

UTF PtNi

Pt

Pt-P

t Str

ain

(%)

-5

-4

-3

-2

-1

0

AnnealedUntested

UnannealedUntested

AnnealedTested

UnannealedTested

Pt

As. Dep. Mole Fraction

Strain vs. Catalyst State

• Bulk strain increases w/ alloy fraction.• Strain relaxes after FC testing.• Annealed catalysts retain more strain.

-6 -5 -4 -3 -2 -1 00.00.51.01.52.02.53.0

Spec

ific

Activ

ity (m

A/cm

2 Pt) Spec. Activity vs. Bulk Strain

Calc. Pt-Pt Strain (%)• Activity of annealed and unannealed

PtxNi1-x correlates with bulk strain.Structure-Composition-Activity Correlations (Process) - ORNL

Ni R

eten

tion

Frac

tion

Grain size (Arb.)• UTF PtNi compositional stability varies

with grain size (anneal process).

Spec. Activity vs. Bulk Strain

-6.1 -5.1 -4.1 -3.1 -2.0 -1.0 00.00.51.01.52.02.53.0

Spec

ific

Act.

(mA/

cm2 P

t)

Calc. Pt-Pt Strain (%)

Ni Retention Fraction, Post- FC Test

• Activity correlates with post-test composition (bulk strain).

Pt/NSTF0.00

0.25

0.50

0.75

1.00

As Dep. Anneal1 Anneal2

Accomplishments and Progress – Ir Integration into UTF Pt, PtNi

• Ir improves post-AST H2/Air limiting current densities of UTF PtNi.

• PtNiIr has ~75% higher limiting current than PtAlloy/C, with ~40% less PGM.

• PtNiIr > PtNi >> Pt.

• Ir impact on mass activity retention after AST is unclear.

• Specific area retention increases monotonically with Ir content.

Electrocatalyst AST Durability of UTF PtIr, PtNiIr

Ir Content (Arb.)0

H2/Air Performance After ASTActivity Impact of Ir

• At optimal content, Ir increases UTF Pt and PtNi PGM mass activity due to increased specific activity.

Ir Content (Arb.)0 0 Ir Content (Arb.)-100-75-50-25

02550

0.0 0.4 0.8 1.2 1.60.30.40.50.60.70.80.9

PtNiIr#2-Lvl10.05mgPt/cm2

PtNiIr#2-Lvl20.05mgPt/cm2

PtNi0.05mg/cm2

Pt/NSTF0.05mg/cm2

PtAlloy/C0.09mg/cm2

H2/Air J @ 0.50V % Change

80oC Cell, 1.5/1.5atmA H2/Air, CS2/2.5

PtAlloy/C0.09mg/cm2

PtNiIr#2-Lvl20.05mgPt/cm2

PtNiIr#2-Lvl10.05mgPt/cm2

PtNi0.05mg/cm2Pt/NSTF

0.05mg/cm2

Cell

Volta

ge (V

olts

)

J (A/cm2)

-50

-25

0

-80

-60

-40

-20

0

PtIrPtNiIr#1

Mass Activity Change (%)

Pt

PtNiIr#2

PtNiIr#1

PtNiIr#2 PtIr

Specific Area Change (%)

30k (0.6-1.0V, 50mV/s)

0

1

2

3

0.0

0.1

0.2

0.3

0.4

0.5PtNiIr#2

PtNiIr#1

Mass Activity (A/mgPGM)

Pt

PtNiIr#2

PtNiIr#1

Pt

Specific Act. (mA/cm2Pt)

12

Accomplishments and Progress – Best of Class MEAs

GREEN: Meets or exceeds DOE 2020 target. YELLOW: Within ca. 10% of DOE 2020 target

Total PGM Loading (mg/cm2)

Spec. Power @ Q/∆T=1.45(kW/gPGM)

Rated Power @ Q/∆T=1.45

(W/cm2)1/4 Power (A/cm2

@ 0.80V)

ORR MassActivity

(A/mgPGM)

Electrocatalyst AST Durability (NSTF Cathode Only)

Mass Act.Loss (%)

∆V @ 0.8A/cm2

(mV)DOE 2020 Target 0.125 8.0 1.000 0.300 0.44 40 30

2015 (Sept.) NPTF PtNi 0.131 6.8 0.891 0.310 0.39 ~65 NA2017 (Jan.) NPTF PtNiIr 0.122 7.3 0.897 0.308 0.40 45 21

2017 (Jan.) UTF PtNi 0.077 8.1 0.626 NA 0.37 43 502017 (Mar.) UTF PtNiIr < 0.089 >6.6 0.584 < 0.200 0.44 45 23

• 2017 NPTF PtNiIr MEA (0.122mgPGM/cm2 total) • Improved specific power, durability over 2015 status.• MEA achieves total loading and ¼ power targets.• NPTF cathode achieves H2/Air performance loss target.

• 2017 UTF PtNi MEA (0.077mgPGM/cm2 total)• MEA achieves specific power and total loading targets.• UTF cathode loading is ultra-low 46µg/cm2.

• 2017 UTF PtNiIr MEA (< 0.09mgPGM/cm2)• Improved durability and activity vs. PtNi, but reduced

BOL H2/Air performance.• UTF cathode achieves DOE mass activity target.

“Best of Class” MEA format• Low PGM anode; 3M-S 725EW PFSA 14µm w/ additive.• Optimized anode GDL, cathode interlayer for operational

robustness.

0.0 0.5 1.0 1.5 2.00.6

0.7

0.8

0.9

2017 UTF PtNiIr2017 UTF PtNi

2017 NPTF PtNiIr

90oC Cell T., 7.35/7.35psig H2/Air (OUT), CS(2,100)/CS(2.5, 167), GDS(120s/pt, High->low J Only). 2012 BOC: 84/84C A/C Dewpoint

2015 BOC: 84/84C Dewpt for J > 0.4. 68/68C Dewpt for J < 0.4

Cell

Volta

ge (V

olts

)

J (A/cm2)

2015 NPTF PtNi

13

-5 -4 -3 -2 -1 0-2-101234 1ML 2ML 3ML 4ML

ln(j/

j Pt)

Surface strain vs Pt(111) (%)

Accomplishments and Progress – UTF PtNi DFT ModelingSurface Strain Stability Relaxed Pt Skin Activity

• Highly-strained Pt skins are not thermodynamically stable.

• High-strain skins will reconstruct (Moiré), reducing surface strain.

• After relaxation, Pt skin surface strain (and activity) depends on resultant skin thickness and less on bulk composition.

% Bulk Strain (EDS)

Spec

. Act

. Gai

n ov

erPt

/NST

F

Model Validation – Experimental UTF PtNi• Experimental activity trends with bulk strain

consistent with model predictions, but not magnitude.• Exp.: 2x vs. Pt. Model: ~20x vs. Pt(111).

• Discrepancy may be due to non-optimalPt skin thickness, near-surface defects, and actual surface strain of experimental materials.• Difficult to quantify.

-6.1 -5.1 -4.1 -3.1 -2.0 -1.0 00.00.51.01.52.02.53.0

Thickness Comp AnnealFixed Var FixedVar Fixed Fixed

Fixed Fixed Var.

• Activity peak between 2-3% strain for 2-4ML Pt skins on PtxNi1-x.

• Activity at high strain depends on skin thickness.

Pseudomorphic Pt Skin Activity

xML Pt skin on PtxNi1-x

1 2 3 4 5 6 7 8 9 10-8

-6

-4

-2

0Peak ActivitySkin Strain

Relaxed Skin Strain

Stra

in (%

vs

Pt)

Skin thickness (ML)

Psuedomorphic Skin Compressive Strain

Stability Limit

-5 -4 -3 -2 -1 0-101234

Pt/Ni Pt/PtNi3 Pt/PtNi

ln(j/

j Pt)

Surface strain vs Pt(111) (%)

14

Accomplishments and Progress – NPTF ModelingKinetic Monte Carlo Model Refinement, Validation (JHU)

Density Functional Theory and Kinetic Monte Carlo Modeling of NPTF PtNiIr (JHU, Purdue)

• KMC model refined to include dealloying via oxidation/reduction cycles (akin to dealloying in FC)

• Refined model accurately captures onset of nanoporosity formation with composition

• Model surface areas 2x experiment values.• Some ECSA loss during break-in.

• Using DFT-estimated interaction parameters, kMC-predicted structure strikingly similar to experiment.

• Ir is relatively immobile in simulation and experiment.

• Suggested stabilization mechanism -capillary wetting of Pt, Ni onto Ir.

0.2 0.3 0.4 0.50123456

Expe

rimen

tEC

SA E

nhan

cem

ent

Over

Pt

x in PtxNi1-x

Mod

elTo

tal s

urfa

ce a

rea/

Sphe

re s

urfa

ce a

rea

0.00.51.01.52.02.53.0

10 nm10 nm

15

Accomplishments and Progress – High Throughput Catalyst DevelopmentXAFS of Gradient Composition PtxNi1-x (ANL) Segmented Fuel Cell ORR Activity (3M)

• XAFS conducted on gradient PtNi NSTF catalyst on fabrication substrate

• Bond distances and coordination numbers vary monotonically with Pt mole fraction, as expected.

• All HT physical characterization methods validated: XRF, XRD, WAXS, XAFS.

• Good sample-sample and segment-segment reproducibility with homogenous electrodes.

• Absolute activity variation of gradient electrodes correct in trend, but not in magnitude.

• PtNi: 3-4x activity variation expected vs. 1.5x observed• Pt: 2.5x activity variation expected vs. 1.5x observed

• Lateral electronic conduction reduces sensitivity; optimization continues.

0 11 22 33 44 55 66 77 88 99 110 1210.50

0.75

1.00

1.25

1.50 FC038556 Homogenous PtCoMn 0.15mgPt/cm2 As dep. FC038914 Homogenous PtCoMn 0.15mgPt/cm2 As dep. FC038922 Homogenous PtCoMn 0.15mgPt/cm2 As Dep. FC038934 Gradient Comp. PtxNi1-x 0.10mgPt/cm2 Annealed FC038987 Gradient Load Pt 0.05-0.02mgPt/cm2

J NORM

Segment #

JAVG=0.04A/cm2 (A/cm2)

Gradient 0.10PtxNi1-x

Gradient xPt

2.50

2.55

2.60

2.65

2.70

2.75

Ni-PtPt-Ni

Ni-Ni

Bond

Dis

tanc

es (A

ng.)

Pt-Pt

0.0 0.2 0.4 0.6 0.8 1.002468

1012

Total

Pt Mole Fraction

Pt-Ni

Coor

dina

tion

Num

ber

Pt-Pt

3 ReplicateHomogenous

16

Collaborations

• 3M - Electrocatalyst Fabrication and Characterization, Electrode and MEA Integration,HT Development• A.Steinbach (PI), C. Duru, A. Hester, S. Luopa, A. Haug, J. Abulu, G. Thoma, K. Lewinski,

M. Kuznia, I. Davy, J. Bender, M. Stephens, M. Brostrom, J. Phipps, and G. Wheeler.• Johns Hopkins University – Dealloying Optimization, kMC Modeling, HT Development

• J. Erlebacher (PI), L. Siddique, E. Benn• Purdue University – DFT Modeling of Electrocatalyst Activity, Durability

• J. Greeley (PI), Z. Zeng, J. Kubal• Oak Ridge National Laboratory – Structure/Composition Analysis

• D. Cullen (PI)• Argonne National Laboratory – XAFS and HT Development

• D. Myers (PI), A. J. Kropf, D. Yang

• University of Hawaii, NREL – NSTF Transport Studies• J. St. Pierre, T. Reshetenko, K. C. Neyerlin

• FC-PAD Consortium• MEAs to be provided annually.

17

Response to Reviewers’ CommentsRated power stability of NSTF MEAs“… if membrane improvements lessen NSTF degradation by membrane fragments), they could enable the full cyclic durability promise of NSTF to finally be realized in fuel cell applications. There is a significant possibility that the changes generated by this project will provide only incremental improvements that are insufficient to get NSTF into significant fuel cell applications.”• Project is addressing several catalyst-specific factors of rated power loss. Factors directly addressed:

1. PFSA decomposition product accumulation on the NSTF cathode catalyst surface (PEM decomp. rate, ECSA).2. ECSA loss where the NSTF cathode electrode roughness factor drops below ~10cm2

Pt/cm2planar

3. Transition metal loss from the cathode electrocatalyst to the PEM• Impact of new alloys on PFSA decomp. rate and rated power loss will be assessed.• Non-catalyst modifications are under development (outside this project).

Operational robustness of NSTF MEAs“While quite solid, it is unclear whether performance and durability should be the primary focus (as laid out in the project) or whether “operational robustness,” the historic Achilles’ heel of NSTF, should have more emphasis.”• Improved operational robustness is important, but out of scope for a catalyst project.• While not formal project criterion, operational robustness is being assessed with downselected catalysts.• Electrode-extrinsic approaches developed in previous MEA integration project are effective with new catalysts.• Electrode-intrinsic approaches are promising and under development in 3M Electrode project (FC155).• See backup slide for summary figure.

18

Remaining Challenges and Barriers

1. Further improved electrocatalysts will require optimization of large composition/process space. HT electrochemical characterization necessary for acceleration not yet validated.

2. Experimental specific activities approximately 10x below entitlement model prediction of catalysts with well-defined and optimally-strained Pt skins.

3. Break-in conditioning of NSTF MEA cathodes is longer and more complex than many carbon supported Pt nanoparticle MEA electrodes.

4. Rated power loss is generally key lifetime-limiting factor for NSTF cathode MEAs.

19

Key Future Work – 2Q17-1Q18• Validate HT electrochemical characterization; implement HT electrocatalyst development.• UTF Development

• Composition, process optimization of downselected alloys towards increased specific activity.• Optimize Pt skin on stable UTF base electrocatalysts towards increased specific activity.• Initiate integration onto higher area supports for improved rated power.

• NPTF Development• Composition, process optimization of downselected alloys towards increased specific area and

rated power capability.• Additive integration optimization for further improved durability.

• Electrocatalyst Modeling• Complete modeling studies of durability additives.• Complete modeling of new electrocatalyst concepts for improved activity and durability.

• Rated Power Durability• Evaluate impact of new catalysts on MEA-level rated power durability (load cycle, F- emission).

Any proposed future work is subject to change based on funding levels20

Summary• New Electrocatalyst Development

• 6 UTF and 6 NPTF Pt alloy series have been fabricated and characterized. Several improved candidates identified for further optimization.

• Relationships between electrocatalyst functional response, physical properties, and fabrication processes have been established for UTF PtNi catalyst.

• Ir integration improves many durability characteristics of NPTF and UTF PtNi catalysts. • Durable UTF and NPTF PtNiIr catalysts can yield high specific power in MEA.

• Electrocatalyst Modeling• Refined Kinetic Monte Carlo model predicts composition and structure during dealloying

consistent with experimental NPTF PtNi.• Refined Density Functional Theory model predicts activity trends vs. strain consistent with UTF

PtNi. Significant gap in magnitude between model and experiment (experimental char. gap).• Models are providing insight into mechanism of Ir stabilization.

• High Throughput Development• Approaches for HT electrocatalyst fabrication and physical characterization (composition, bulk

structure, atomic structure) have been validated.• HT electrochemical characterization (seg. cell) is reproducible and correct in trend, but

magnitude of response is muted.

21

Technical Backup Slides

22

NPTF, UTF PtNiIrUTF PtNiIr. 50cm2 MEA Format.

NPTF PtNiIr. 50cm2 MEA Format.

Ir Content (Arb.)0 Ir Content (Arb.)0

Ir Content (Arb.)0 Ir Content (Arb.)0 Ir Content (Arb.)00.1

0.2

0.3

0.4

0.5

PtNiIr #4

Pt Alloy/CPtNiIr #2

PtNiIr #3

PtNiIr #1

Mass Activity (A/mgPGM)

0

10

20

30

40

PtNiIr #4

Pt Alloy/C

PtNiIr #2PtNiIr #3

PtNiIr #1

Specific Area (m2/gPGM)

0

1

2

3

PtNiIr #4

Pt Alloy/C

PtNiIr #2PtNiIr #3

PtNiIr #1

Specific Activity (mA/cm2Pt)

0.0 0.5 1.0 1.50.40.50.60.70.80.91.0

UTF PtNiIr(various Ir contents)

J (A/cm2)

Cell

V (V

olts

)

80/68/68C, 7.35/7.35psig H2/Air, CS(2,100)/CS(2.5, 167)GDS(0.02->2->0.02, 10steps/decade, 120s/pt, 0.4V limit, 0.1maxJstep)Upscan (high->low J) only.

UTF PtNi

• With NPTF PtNi, Ir does not strongly impact PGM-specific area or specific activity.

• With UTF, Ir does not strongly impact PGM-specific area. • H2/Air: UTF PtNiIr > UTF PtNi

1.5atmA H2/Air, 27ugPt/cm2

0

5

10

15

20

0.0

0.1

0.2

0.3

0.4

0.5PtNiIr#2PtNiIr#1

Mass Activity (A/mgPGM)

PtIr

Specific Area (m2/gPGM)

PtNiIr#1PtNiIr#2

PtIr

23

Break-in Conditioning, Operational Robustness, Area DeterminationBOC MEA Conditioning Operational Robustness ECSA Determination at Low Load

0.0

0.5

1.0

1.5

2.0

0 5 10 15 20 25 300.0

0.1

0.2

0.3

0.4

J @

0.8

0V (A

/cm

2 )

Time (hours)

2015 NPTF PtNi 2017 UTF PtNi 2015 NPTF PtNiIr 2017 UTF PtNiIr

75/70/70C0/0psig H2/Air800/1800SCCMPDS(0.85V->0.25V->0.85V, 0.05V/step, 10s/step)

J @

0.3

0V (A

/cm

2 )

• UTF modestly slower than NPTF.• Ir has little effect.• Processing variations are influential.

3M Thermal Cycle Protocol

• CVs from ultra-low load MEAs have apparent “other” reductive processes.• HUPD integration error-prone.

• “Forced” symmetry approach removes questionable HUPD integration.

0.0 0.2 0.4 0.6 0.8-2

-1

0

1

2-2

-1

0

1

2

70C, 100% RH, H2/N2, 100mV/s

Symmetric Analysis

J (m

A/cm

2 )

E v. H2 (Volts)

70C, 100% RH, H2/N2, 100mV/s

J (m

A/cm

2 )

"Standard" Analysis50µg/cm2

16.6 ± 3.4 m2/g

9µg/cm2

14.9 ± 2.4 m2/g

50µg/cm2

17.1 ± 1.4 m2/g

9µg/cm2

17.1 ± 2.2 m2/g• Two approaches are effective towards

improved NSTF operational robustness (performance sensitivity to temperature)

1. Electrode extrinsic – anode GDL and cathode interlayer (2015 BOC, prev. 3M MEA integration project)

2. Electrode intrinsic – advanced NSTF electrode (FC155).

10 20 30 40 500.0

0.4

0.8

1.2

1.6

2015 BOC MEA0.1PtNi

An. GDL, Ca. IL2013 NSTFBaseline

0.2Pt/V

100/100% RH, 100/150kPaA H2/Air

J @

0.4

V (A

/cm

2 )

Cell T (oC)

0.2Pt/NSTFAdv. Elec

24

Best of Class MEAs – Operational Robustness

Anode GDL, Cathode Interlayer Approaches Effective w/ 2017 BOC MEAs• 2016, 2017 BOC MEAs achieve stable 1A/cm2 operation down to 42C cell temperature or

lower, similar to or improved vs. 2015 NPTF PtNi Best of Class MEA.• Target is stable 30C operation.

30 40 50 60 70 80-0.2

0.0

0.2

0.4

0.6

0.8

30 40 50 60 70 80Cell T (oC)

xoC Cell T., 7.35/7.35psig H2/Air, 696/1657SCCMJ: Step from 0.02 to 1.0A/cm2

60-80C: 100% RH. 30-50C: 0% RH

2015 (Sept.) NPTF PtNi BOC MEA. 0.131mgPGM/cm2 total. 2016 (Jan.) NPTF PtNiIr BOC MEA. 0.127mgPGM/cm2 total. 2017 (Jan.) NPTF PtNiIr BOC MEA. 0.122mgPGM/cm2 total. 2017 (March) UTF Pt"A"Ir BOC MEA. ~0.08mgPGM/cm2 total.

Cell

Volta

ge (V

olts

)

Cell T (oC)

t=0s after transient t=30s after transient

25

Kinetic Monte Carlo (kMC) SimulationsAtom-Scale Simulations of Dealloying and Coarsening

Novel kMC Simulation inputs:• DFT-based activation barriers for Ni-

Pt-Ir bond energies• Oxidation and reduction transitions

to simulate cyclic voltammetry

Primary Conclusions to date:• Dealloying/short time scales

• Porosity evolution is controlledby creating mobile Pt atomsduring surface redox cycling

• Composition vs redox cyclefollow experimental trends inbase alloy composition

• Coarsening/long time scales• Ir has low mobility• Ni and Pt wetting of relatively

immobile Ir clusters slowscoarsening

nickel

platinum

Pt20Ni80 Pt25Ni75 Pt30Ni70 Pt40Ni60

dealloying simulations show correct composition and surface area trends vs. # of redox cycles

26

DFT - Pt/PtNi3 Moiré Reconstructions ; PtIr Interactions

(1ML skin, -4%) (2ML skin, -2.5%) (3ML skin, -2.0%) (4ML skin, -1.5%)

RT19 on RT21 RT25 on RT28 RT43 on RT49 RT27 on RT31

0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

dEfo

rm(Ir

) (eV

/Ir)

Ir coverage (ML)0.2 0.4 0.6 0.8 1.0

0.0

0.2

0.4

0.6

0.8

dEfo

rm(Ir

) (eV

/Ir)

Ir coverage (ML)

0.4 eV/Ir

Ir/Pt(332)

PtIr

0.2 0.4 0.6 0.8 1.0 1.2-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

dEfo

rm (e

V/P

t)

Pt coverage (ML)

IrPt

Pt/Ir(332)Pt on Ir is more energetically favorable than Ir on Pt.

Pt skin strain (and activity) on relaxed surfaces depends on skin thickness

27

Electron Microscopy Reveals Highly Durable Ir Coatings on PtNi NPTF

PtNiIrConditioned

PtNiIrAST

PtNiConditioned

PtNiAST

• Minimal Irdissolution observed following ASTs.

• Ir remains as a surface layer on the underlying PtNi alloy

28