PEM Fuel Cell Metallic Bipolar Plates: Technical Status and Nitridation Surface Modification for Improved Performance M. P. Brady a (bradymp@ornl.gov), T. J. Toops a, Karren L. More a (morekl@ornl.gov), H. M. Meyer III a, P.F. Tortorelli a, M. Abd Elhamid b, G. Dadheech b, J. Bradley b, H.Wang c, and J. A. Turner c a Oak Ridge National Laboratory, Oak Ridge TN USA 37831-6115 b General Motors Technical Center, Warren, MI USA 48090 c National Renewable Energy Laboratory, Golden, CO USA 80401 Technoport 2012 Trondheim Norway April 16, 2012
Outline Considerations for Metallic Bipolar Plates -short commentary on state-of-the-art for protection of metallic bipolar plates addressed Metallic Bipolar Plates and the Nitridation Concept Exploratory Single-Cell Fuel Cell Evaluation of Nitrided 15 cm 2 FeCrV Stainless Steel Stampings -1000 h test results -rapid cycle quartz/plasma lamp nitriding Single-Cell Evaluation of 50 cm 2 Stamped/Laser Welded/Nitrided FeCrV Bipolar Plate Assembles -collaboration with General Motors
Considerations for Protection of Metallic Bipolar Plates ORNL Viewpoint
Electrically Conductive, Corrosion-Resistant Coating Candidate Materials Long Established After Borup and Vanderborgh, MRS Proceedings 1995 Carbon-Based Metallic Graphite Inert metals (e.g. Nb) Conductive Polymer Metal Nitrides Diamond, Diamond Like Metal Carbides Carbon Organic Self-Assembled Noble Metals (e.g. gold) Monolayers *conductive oxides have also been examined in past decade Manufacturing considerations of coating approach cost, flow-field coverage, and defect incidence are key to success Not clear if can be completely prime reliant on coating for protection -alloy substrates likely need good degree of inherent corrosion resistance (challenge for steel and Al substrates, favors stainless steel)
Evolution of Coating Architectures to Better Protect Metallic Bipolar Plates Surfaces Deposited Coating Reacted/Graded/Nano Surfaces Thin Carbon Films Alloy Thin Gold layers Gas Nitridation Doped Alloys Electrochem. Treatment Multi-Layer Surface pin-holes, flaws Alloy Reduces Defects Alloy Nitride/Metal (Cr x N/Cr) Carbon/Metal Nitride/Nitride (TiN/Cr x N) Composite Surface Alloy conductive phase Boride/Passive Oxide Gold/Polymer or Oxide Carbon/Polymer or Oxide V x N/Cr 2 O 3 Nitride/Rubber
State-of-the-Art Performance Benchmark is Electroplated Gold on Stainless Steel What are current literature trends for lower cost coated metal alternatives? Nano Gold variations (layers and composites) -may be only option if encounter frequent excursions >1V Carbon-based surfaces Cr-nitrides Electrochemical treatments Very dependent on end use application: some coatings that fail automotive performance or cost targets may be acceptable for stationary or portable applications
How Do We Usually Assess Candidate Bipolar Plate Alloys and Coatings? Desired lifetimes of ~5000 h (auto) to 40,000 h + (stationary) Accelerated single-cell and stack testing protocols available -manufacturing bipolar plates of exploratory alloys or coatings can be difficult, costly, and complicated Early screening typically accomplished with ex-situ flat sheet coupon immersion polarization corrosion tests less than 8 h -metallurgical surface conditions of flat plates may be significantly different than cold work stamped foils -flat coupons far easier to coat than stamped plates
Corrosion and Interfacial Contact Resistance (ICR) are Key for Metallic Bipolar Plates Electrochemical screening by immersion in hot sulfuric acid with high levels of F - may now be overly aggressive -better membrane-electrode assemblies (MEAs) and improved water management reduce corrosivity of bipolar plate operating environment Oxide layer formation and increased ICR may be bigger issue than metal corrosion/dissolution and MEA contamination Current density target numbers from hot sulfuric acid polarization corrosion tests are of limited use -how current density reflects ICR increase or metal ion dissolution material dependent (e.g. different coating materials may need different targets) Strong need for more open literature studies that relate ex-situ corrosion and ICR assessment to in-situ fuel cell test results
Nitridation Approach
Gas Nitridation: Thermally Grown Cr-Nitride to Protect Metallic Bipolar Plates Cr-Nitride Nitrogen-containing gas Cr Cr-Containing Bipolar Plate Alloy Surface conversion not a deposited coating Model Ni-50Cr Alloy (SEM Cross-Section) Protective Cr-Nitride Layer 10 μm Metal High temperature favors reaction of all exposed metal surfaces -few, if any, pin-hole defects -not line of sight limited, amenable to complex geometries e.g. flow fields -drawback: very substrate alloy dependent, most compositions won t form continuous Cr x N surface layer Stamp then nitride: Industrially established and low cost
Cr x N Formed on Nitrided Model Ni-50Cr Alloy Exhibits Good Corrosion Resistance Potential (mv, SHE) Anodic Polarization Data at ph3 H 2 SO 4, 80 o C 1100 900 700 500 300 100 Ni 0-100 1x10-8 1x10-7 1x10-6 1x10-5 1x10-4 1x10-3 Ti Ni-50Cr Metal Current Density (A/cm 2 ) Nitrided Ni-50Cr (1100 C/2h)
ICR mωcm 2 800 600 400 200 Cr x N Surface on Model Nitrided Ni-50Cr Exhibits and Maintains Low ICR Low Initial Contact Resistance DOE Goal Twice ICR mωcm 2 No Contact Resistance Increase after 4100 h Corrosion Exposure As-Nitrided 300 4100 h corrosion test (ph 3 H 2 SO 4 +F -, 80 C) 200 100 0 50 100 150 200 Compaction force, N/cm 2 0 50 100 150 200 Compaction Force N/cm 2 4100h in LANL bipolar corrosion test cell (Air/H 2 ph 3 H 2 SO 4 + F- 80 C)- No attack of Cr x N and minimal metal ion dissolution (K. Weisbrod test cell) Good single-cell fuel cell test behavior also observed
Need Fe-Base Stainless Steel to Meet $3-5/kW Cost Goals for Auto Applications Commercial Ni-Cr Alloys in Range of ~$15-25/lb -far too expensive for automotive use Focus on Fe-Base Stainless Steels: ~$2-8/lb High N 2 Permeability Makes Surface Cr-Nitride Formation Difficult on Fe-Cr: poor corrosion resistance N 2 Floods in Little Surface Cr-Nitride Internal Cr-nitride 10 μm mm Model Fe-27Cr Alloy
Pre-Oxidation Followed by Nitridation to Form Cr x N Surface on Stainless Steels Nitrogen Nitrogen Cr x N Internal Cr 2 N Nitrided Bare metal Cr 2 O 3 Pre-oxidized N-modified Cr 2 O 3 Nitrided Pre-oxidized Stainless steels internally nitride: corrosion Form Cr 2 O 3 by preoxidation to keep N 2 at surface -convert surface Cr 2 O 3 to surface Cr x N by nitridation V added to stainless steel assists conversion to nitride -good ICR and corrosion results with model Fe-27Cr-6V alloy
ORNL Developed Fe-(20-23)Cr-4V Wt.% Alloy Foils Protective Nitride Cr V Ni Overlap Yields Successful Alloy Stampability Cr Ni Cost V Ni Co-optimize ductility (for stamping) and low alloy cost with protective nitride surface formation Pre-oxidation: 900 C, 5 min to 1 h, N 2-4H 2-0.5O 2 ; Nitridation: 1000 C, 1-3 h, N 2-4H 2 (separate steps here, can be combined in single step)
Composition (at.%) 80 60 40 20 0 Surface Chemistry Changes for Fe-20Cr-4V On Pre-Oxidation and Nitriding O Metal Pre-Oxidized Nitrided Cr-Ox Fe 0 1 2 3 4 5 Depth (nm) Cr-Met V Composition (at.%) 80 60 40 20 0 O V Cr-Ox Fe Cr-Met Composition (at.%) 0 2000 4000 Depth (nm) 80 60 40 20 0 N O V Cr-Ox Fe 0 2000 4000 Depth (nm) Cr-Met Under ideal circumstances Fe completely eliminated from surface by pre-oxidation and nitridation -yields (V,Cr) x N in (Cr,V) 2 O 3 Above for prepped laboratory sheet coupon, foil tends to retain a few % Fe in surface (900-1000 C: 1-2 h peak temperature hold)
Continuous Cr x N to Cr x N+Oxide to Through-Thickness V x N in Cr 2 O 3 Decreased Alloy Cost Better Corrosion Resistance Lower alloy Cr, V content and shorter nitriding cycles (1-2 h at 1000 C) to reduce cost result in V x Nin Cr 2 O 3 surface rather than continuous Cr x N
Low ICR and Good Corrosion Resistance Achieved with Nitrided FeCrV Foils 2x ICR mω-cm 2 200 180 160 140 120 100 80 60 40 20 0 Polarized Nitrided Fe-20Cr-4V Foil Polarized Nitrided Model NiCr/FeCr Target ICR 0 50 100 150 200 250 Compaction Force (N/cm 2 ) Polarization in 1M H 2 SO 4 + 2 wppm F - for 7 h at 0.14V (H 2, anode simulation) or 0.84 V (aerated, cathode simulation) vs SHE Corrosion resistance of nitrided FeCrV foils (3 to 5 μm A/cm 2 ) not quite as good as FeCrV sheet coupons or model NiCr/FeCr sheet alloy coupons Moderate ICR increase for nitrided FeCrV foils after polarization
Single-Cell Fuel Cell Testing of 15 cm 2 Active Area Stampings
Voltage Single-Cell Fuel Cell Testing of Nitrided Foils Benchmarked to Stainless Steels and Graphite 0.9 0.6 0.3 Fuel Cell Test Cycle (1000-1200 total h) 1 min OCV Stamped and Nitrided Fe-20Cr-4V Foil 5 cm 0 0 50 100 Time (minutes) 7 cm Operating conditions: 80 C, 25 psig performance curves (V-I):0.9-0.4V, 0.05V steps, 20 min./step, repeat 3x Simple serpentine ~15 cm 2 active area stamped foils for metals, machined graphite block of similar flow-field design
1000-h Aged Nitrided Fe-20Cr-4V, 904L, and Graphite Plates All Showed Good Durability V-I Curves of Aged Plates Using Fresh Gore PRIMEA MEA Voltage (mv) 1000 900 800 700 600 500 400 300 200 100 0 0 500 1000 1500 Current (ma/cm 2 ) Initial Nitrided Fe20Cr4V Aged (1114h) Initial Graphite Aged (1175h) Initial Untreated 904L-SS Aged (1122h) Slight decline in nitrided Fe-20Cr-4V data within fuel cell build-to-build variation (< 5-10% variation of peak power output) Lower performance of graphite attributed to flow-field differences w/stampings
Nitrided Surface on Fe-20Cr-4V Protected MEA from Metal Contamination Nitrided Fe-20Cr-4V Plates Metal content (μg/cm 2 ) 1.2 1 0.8 0.6 0.4 0.2 X-ray Fluorescence (XRF) of MEAs Foil/plate used N/A (fresh MEA) Graphite Nitrided Fe20Cr4V Untreated 904L-SS 0 Fe Cr V Ni No visible attack of nitrided Fe-20Cr-4V plates (slight staining-gdl contact) XRF found MEAs from graphite and nitrided Fe-20Cr-4V plates clean Small ( 1 μg/cm 2 ) level of metal ion contamination with 904 L 321 tested for 120 h showed regions of 1-70 μg/cm 2 metal contamination
Robustness/Repeatability Concern for Fe-20Cr4V and Short 1-2 h Nitridation Cycles (Multiple Tests) Oxidation seen in 1.5h Nitrided Stamping -Performance decline in test -Significantly higher V-O signal in flow-field (no Fe oxidation) -Oxidation of V x N -Likely inadequately nitrided (needed longer cycle this run) Composition (at.%) Composition (at.%) 50 40 30 20 10 0 50 40 30 20 10 O Fe-O Mn-O(x10) Cr-O Cr -metal/n V-O V-N 0 400 800 1200 1600 V-O O Mn-O(x10) N Depth (nm) N Cr-O Depth (nm) V-N Fe Cr -metal/n Fe 0 0 400 800 1200 1600
Rapid/More Robust Nitriding with Quartz Lamps? Quartz Lamp Furnace Plasma Arc Lamp Plasma arc and quartz lamps for rapid nitriding heating/cooling -potential nitriding in minutes instead of hours (faster cycle, lower cost!) Rapid heating may favor more nitride-rich surface -less transient oxidation, less Fe at surface -potential for more robust surfaces? Minimizes brittle σ phase formation permitting higher-cr stainless steel alloys (if can be stamped)
Desired Surface Formed by Rapid Nitriding Using Quartz Lamp Heating Technology XPS Surface Chemistry of 1000 C/10 min Treated Fe-20Cr-4V 70 Composition (at.%) 60 50 40 30 20 10 0 V N O Cr Fe Driver is Lower-Cost Processing, Better Surface 0 400 800 1200 1600 2000 Depth (nm) Surface consistent with V(Cr) x N dispersed in (Cr,V) 2 O 3 Virtually no Fe in treated region (usually correlates w/ good corrosion resist.) 900-1000 C; 5 to 15 min peak temperature holds
Quartz-Lamp-Treated Stampings Show Promising Single-Cell Fuel Cell Behavior Initial V-I Curves Using N212 MEA Voltage (mv) 1000 900 800 700 600 500 400 300 200 100 0 904L-SS 321-SS 1000 C: 5,10 Min Quartz Lamp Treated N-Fe-20Cr-4V_Aged-1000+h N-Fe-20Cr-4V_10m (quartz-1) N-Fe-20Cr-4V_5m (quartz-3) 0 200 400 600 800 1000 1200 1400 Current (ma/cm 2 ) Better V-I behavior than standard nitriding procedure -reasons why not understood (ex-situ ICR coupons showed similar values) Durability testing yielded generally positive results
General Motors (GM) Interaction: Manufacturability and Single-Cell Assesments Details to be published soon. Presented slides in this section not yet available for release
Summary Promising manufacturability and single-cell fuel cell results obtained for stamped, laser welded, gas nitrided Fe-(20-23)Cr-4V alloy foils Quartz lamp nitriding shows promise for low cost, robust surfaces May be able to extend approach to existing commercial SS foils -favor 20 wt.% Cr ferritic (nitride forming additions Ti, V, etc) -best chance with quartz lamp (grow passive oxide layer with nitride particles) Concern for nitrided surfaces if frequent operation excursions > 1V
Acknowledgements Funding from USA DOE Fuel Cell Technologies Program Extensive Collaborations, Input, and Advice from: -Arizona State University -ATI Allegheny Ludlum -AGNI-GenCell -General Motors -Los Alamos National Lab -National Renewable Energy Lab