progress report for propulsion materials materials is enabling the advanced integrated power module...

U. S. Department of EnergyOffice of Advanced Automotive Technologies1000 Independence Avenue S.W.Washington, D.C. 20585-0121

FY 1999

Progress Report for Propulsion Materials

Energy Efficiency and Renewable EnergyOffice of Transportation TechnologiesOffice of Advanced Automotive TechnologiesEnergy Conversion Team

Steven Chalk Energy Conversion Team Leader

October 1999

FY 1999 Progress Report Propulsion Materials

iii

CONTENTS

1. INTRODUCTION ...................................................................................................... 1

2. POWER ELECTRONICS......................................................................................... 7

A. Carbon Foam Thermal Management Materials for Electronic Packaging.......... 7

B. dc Buss Capacitors for PNGV Power Electronics .............................................. 12

C. Mechanical Reliability of Electronic Ceramicsand Electronic Ceramic Devices ......................................................................... 17

D. Low-Cost, High-Energy-Product Permanent Magnets ....................................... 20

E. Lead-Free Solders for Automotive Electronics................................................... 24

3. FUEL CELLS RESEARCH AND DEVELOPMENT ............................................ 29

F. Composites for Bipolar Plates............................................................................. 29

G. Carbon Composite for PEM Fuel Cells............................................................... 32

H. Cost-Effective Metallic Bipolar Plates Through InnovativeControl of Surface Chemistry ............................................................................. 36

I. Low-Friction Coatings for Fuel Cell Air Compressors/Bearings ....................... 39

J. Inorganic Proton Exchange Membrane Electrode/Support Development ......... 42

4. ADVANCED COMBUSTION ENGINE AND EMISSIONS R&D ...................... 44

K. Microwave-Regenerated Diesel Exhaust Particulate Filter ................................ 44

L. Rapid Surface Modification of Aluminum Engine Block Boresby a High-Density Infrared Process .................................................................... 48

M. Optimization of NFC Coatings for Light-Duty CIDI Applications .................... 53

N. Material Support for Non-Thermal Plasma Development .................................. 58

O. Nanofluids for Thermal Management Applications............................................ 60

5. CERAMICS FOR GAS TURBINES ........................................................................... 63

P. Agreement to Bring into Production and Commercialize a ManufacturingProcess for Silicon Nitride Turbomachinery Components ................................. 63

APPENDIX A: ABBREVIATIONS, ACRONYMS, AND INITIALISMS................... 66


1

Advances in propulsion material technologies willimpact all vehicle systems

Patrick B. Davis,Program Manager

1. INTRODUCTION

Advanced Propulsion Materials R&D: Enabling Technologiesto Meet Technology Program Goals

On behalf of the Department of Energy’s (DOE’s) Office of AdvancedAutomotive Technologies (OAAT), I am pleased to introduce the FY 1999Annual Progress Report for the Automotive Propulsion Materials Research andDevelopment Program. Together with DOE national laboratories and inpartnership with private industry and universities across the United States,OAAT engages in high-risk research and development (R&D) that providesenabling technology for fuel-efficient and environmentally-friendly light dutyvehicles.

The Automotive Propulsion Materials Research and Development Programsupports the Partnership for a New Generation of Vehicles (PNGV), agovernment-industry partnership striving to develop by 2004 a mid-sizedpassenger vehicle capable of achieving 80 miles per gallon while adhering tofuture emissions standards and maintaining such attributes as affordability,

performance, safety, and comfort. Automotive propulsion materials research is key to PNGV programsuccess as it focuses on improving the materials used in propulsion systems components and subsystems.New propulsion materials will facilitate higher efficiencies, lower emissions, improved alternative fuel

capabilities, and lower specific weight andvolume, without compromising cost, safety,and recyclability.

Reorganized in FY 1998, the AutomotivePropulsion Materials program is now anintegral partner with the Power Electronics,the Advanced Combustion and EmissionsControl, and the Fuel Cells for TransportationR&D programs. This change reflects theelimination of the gas turbine engine as acandidate for the 80-mpg automobile and theemphasis on direct-injection engines and fuelcell technologies as the selected candidatepowerplants for 80-mpg hybrid electric

vehicles. Projects within the Automotive Propulsion Materials program address materials concerns thatdirectly impact the critical technical barriers in each of these programs—barriers such as thermalmanagement, emissions reduction, and reduced manufacturing costs.

Enabling TechnologiesThe technologies developed in the Automotive Propulsion Materials program are what are known asenabling technologies—those necessary for the success of the power electronics, fuel cell, andcombustion engine and aftertreament research programs. One of the most important technical barriersbeing addressed by the materials program, for example, is preventing the overheating and failure ofvehicle electronics (thermal management). The components necessary for the high-fuel-economy, low-emission PNGV vehicles require high-power electronics to be smaller and lighter in weight. This R&D inelectronics materials is enabling the Advanced Integrated Power Module program to address newrequirements for vehicle electronics, such as controlling electricity generated from fuel cells and other

FY 1999 Progress Report Advanced Propulsion Materials R&D

2

Fuel

Advanced Propulsion Materials

Pow

er E

lect

roni

cs

Com

bustion Engine

and Aftertreatm

ent

Low costLightweight

Non-corrosivePhysically robust

Electrically conductive

Low costLightweight

Thermally conductive

Low costLightweightFrictionless

Physically robust

Cells

Advanced Propulsion Materials Enable Other Program Areas to Meet Technical Goals

hybrid-electric configurations. These requirements increase the temperatures realized by circuitry andchallenge the ability to keep power electronics and capacitors from overheating. R&D for new thermalmanagement materials is at the core of the activities in power electronics materials.

The successful development of fuel cells will require major materialsbreakthroughs. Not only must low-cost, durable fuel cell systems bedeveloped, but also specific fuel cell components require advancedmaterials development to turn research concepts into working systems.The bipolar plates used in the proton-exchange membrane (PEM) fuelcell stack, for example, must be resistant to corrosion, electricallyconductive, and physically robust. The propulsion materialsprogram has included projects specifically tasked withdeveloping less costly materials to provide thesecharacteristics. Two successful bipolar plate projects havebeen transferred to the Fuel Cells for Transportationprogram, and materials work is now focusing on othercritical materials-related barriers for fuel cells, such as thedevelopment of a high-temperature PEM.

Compression-ignition, direct-injection (CIDI) engine and aftertreatment development will greatly benefitfrom the propulsion materials program through research in advanced component coatings and thedevelopment of improved particulate filters for diesel engines. Wherever there are moving parts within asystem, there is concern with the amount of friction between these parts and how it places demands onoverall durability and operating life. For both fuel cell compressors and CIDI engine applications,development of advanced materials for coatings can minimize friction to ensure long component life orreduced wear of components such as fuel injectors. Wear of components can lead to increased emissions,lower efficiency, and lower durability.

In addition, current CIDI engine technology faces the difficult balance of engine efficiency versus tailpipeemissions. The propulsion materials program is working to maximize efficiency while reducing emissionsthrough the development of advanced filters to reduce particulate matter from combustion engines. Thepropulsion materials program is also investigating the development of other materials that can furtherreduce tailpipe emissions.

Advanced Propulsion Materials Program


3

Laboratory/contractor-industry collaboration

Laboratory Industrial partners

ArgonneNationalLaboratory

Ability Engineering Technology, Inc.

Atlas Cylinders, Inc.

Bronson and Bratton, Inc.

Cryomagnetics, Inc.

CRUMAX

DaimlerChrysler Corporation

Ford Motor Company

Lucas-Varity

Purdue University

UGIMAG, Inc.

Los AlamosNationalLaboratory

Bulk Molding Compounds, Inc.

Premix, Inc.

Plug Power, LLC

Oak RidgeNationalLaboratory

AlliedSignal

AVX

Bronson and Bratton, Inc.

Conoco Corporation

Cryomagnetics, Inc.

CRUMAX


Dow Chemical Company

Ford Motor Company

Florida International University

Industrial Ceramic Solutions

Kemet

Motorola

Murata

Plug Power, LLC

UGIMAG, Inc.

University of Dayton ResearchInstitute

University of Tennessee

University of Wisconsin

Sandia NationalLaboratory

AVX


Ford Motor Company

General Motors

Degussa

Ferro

Kemet

Materials Research Associates

Murata

Pennsylvania State University

TAM

Tokay

TPC Ligne Puissance

TPL, Inc.

TRS Ceramics

Collaboration and CooperationAs with other programs under PNGV, collaboration and cooperation across organizations is acritical part of the Advanced Propulsion Materials program. Across the materials projects, scientistsat the national laboratories are collaborating with manufacturers to identify and refine the necessary

characteristics for meetingperformance requirements.Component manufacturersand scientists from nationallaboratories and contractorsare also working together toidentify the technologicalbarriers to manufacturingoptimal materials to meetcomponent requirements.

There is also cooperationamong national laboratoriesto take advantage of theexpertise of each facility.Argonne and Oak RidgeNational Laboratories, forexample, are collaboratingin the development of a low-cost, high-energy-productpermanent magnet. For thisproject, Argonne providesexpertise in fabrication, whileresearchers at Oak Ridgecharacterize permanentmagnets fabricated atArgonne, as well as thosefrom other commercialmanufacturers. In anotherproject, Oak Ridge National

Laboratory (ORNL) is providing ceramic materials support to Pacific Northwest National Laboratory forthe development and fabrication of ceramic components for new non-thermal plasma after treatmentsystems to reduce diesel exhaust emissions.

In addition to national laboratory and large industry participation, the FY 1999 Advanced PropulsionMaterials program also included some breakthrough research and development conducted by a smallbusiness. Industrial Ceramic Solutions, LLC, located in Oak Ridge, Tennessee, is developing a ceramicfilter to reduce particulate emissions from diesel engines. As in the collaborative efforts of nationallaboratories with industry, researchers at Industrial Ceramic Solutions are working closely withrepresentatives from DaimlerChrysler, Ford, and General Motors to develop a filter that will help meetPNGV emissions targets.

AccomplishmentsFY 1999 featured notable accomplishments in all three materials program areas. While the remainder ofthe report provides summaries of all of the Advanced Propulsion Materials projects, this section providesa highlight of some of the major accomplishments during FY 1999. Materials development in support of


4

400 m

Carbon foam structure at hightemperature (1000oC)

power electronics, for example, featured breakthroughs in the use of high-conductivity carbon foam forheat exchangers and heat sinks. Fuel cell–related materials activities demonstrated improvements in thecapability to mass-produce low-cost composite material for fuel cell stack bipolar plates. Finally,materials research in the support of combustion engine and aftertreatment technologies led to thedevelopment of a microwave-regenerated exhaust particulate filter.

Power ElectronicsOne of the greatest challenges to the successful development of advanced power electronic devices ismanaging the heat generated by their operation. Currently, most heat sinks for cooling high-powerelectronics use a water-cooled aluminum or copper platemounted below the electrical circuitry to transfer heat fromthe electronics to the water-cooled metal. Using high-conductivity carbon foam as the core material for these heatexchangers can significantly increase the effective transfer ofheat while reducing the overall size and weight.

During FY 1999, researchers at ORNL improved processingtechniques for the production of carbon foam, increasing thethermal conductivity by more than 50% and meeting theprogram’s goal. Collaborating with potential manufacturers,ORNL researchers defined targets and manufacturingrequirements for using this highly conductive carbon foam inelectronic component heat exchangers and heat sinks.Researchers varied processing conditions and characterizedfoam structures to understand the effects on thermalproperties. As a result of these findings, ORNL researchersdeveloped a new, faster process for fabricating carbon foams,eliminating several steps in the fabrication process.

The outcome of this project has generated positive responses. First, based upon confidence in this newtechnology, ORNL has successfully licensed the carbon foam material technology to Poco Graphite ofDecatur, Texas, for large-scale commercial production. Second, extensive meetings are being conductedwith DaimlerChrysler, Ford, Lockheed Martin, Boeing, Modine, Peterbilt, and a NASCAR racing team todevelop this material for use in radiator systems and other electronics cooling applications. Third, there ispotential for using carbon foam in desktop computers and laptops to better disperse heat, allowing the sizeof the computer to be reduced. Finally, it is expected that the same technology may be applied for use infuel cell vehicles (to remove heat) and large trucks (smaller, lighter radiators make for moreaerodynamically shaped trucks).

Fuel Cells R&DThe successful development of fuel cell technology in vehicles requires breakthroughs in cost, durability,size, and performance. During FY 1999, researchers at Los Alamos National Laboratory (LANL), incollaboration with Premix, Inc., and Bulk Molding Compounds, Inc., took the final steps necessary formass production of composites for PEM fuel cell stack bipolar plates that are low-cost, corrosion-resistant, electronically conductive, and physically robust.

Although laboratory-scale plates produced at LANL in FY 1998 exhibited good properties, it wasrecognized that the processing of molding compounds would need to be improved prior to massproduction. Using the manufacturing expertise of Premix and Bulk Molding Compounds, researchersworked together to produce new molding compounds that exhibited reduced cure times, extended shelf


5

Industrial Ceramic Solutions’ microwave-regenerated ceramic fiber filter

life, and more uniform flow. Researchers used vinyl ester resins in these compounds to minimize processcycle times and enhance plate resistance to corrosion. In addition, they eliminated machiningrequirements by forming plate molds, allowing for a single molding step. All of these accomplishmentswill lead to lower-cost production of stacks for PEM fuel cell systems. As a result of these breakthroughs,LANL has submitted an application for a U.S. patent for this advanced composite molding processtechnology.

Advanced Combustion Engine and Emissions R&DA major PNGV goal is to develop a vehicle with outstanding fuel economy that meets stringent emissionsstandards. Balancing high fuel economy with low emissions is a challenge that is being addressed throughthe materials activities in support of the Advanced Combustion Engine and Emission Control R&Dprogram. Specifically, work conducted at Industrial Ceramic Solutions has led to the development of anadvanced exhaust filter system capable of capturing more than 90% of the carbon particulates from dieselengine exhaust.

Based upon characteristics identified by the PNGVrepresentatives at DaimlerChrysler, Ford, and GeneralMotors, researchers at Industrial Ceramic Solutionsdesigned and fabricated a ceramic filter system. Thissystem was demonstrated on the exhaust of the Ford 1.2liter DIATA diesel engine. The filter system features aceramic-fiber filter that can be automatically cleanedthrough the use of microwave power. Researchers testedthe capability of the system to self-clean during engineidling, varying the temperature of operation, air flowthrough the filter, and microwave power input. As a resultof the tests, Industrial Ceramic Solutions developed aprototype exhaust filter system that can capture more than90% of diesel particulates from exhaust and be regeneratedor cleaned under idle conditions.

This exhaust filter system signifies a breakthrough inemission control technology. If successful, it could beapplied in a wide variety of diesel engines that generatevery fine particulate matter. Beyond the application indiesel-powered cars, this technology could be applied todiesel engines used in pick-up trucks, sport utility vehicles,

and large trucks. In addition, with the near-elimination of particulate emissions, the focus on engineoptimization could be steered toward reducing other types of emissions, such as nitrogen oxides (NOx), amajor contributor to air pollution.

Awards and Achievements

Carbon foam thermal management (ORNL)

✟ License granted to Poco Graphite

✟ Three Patent Applications Submitted

Near-frictionless coating (ANL)

✟ Argonne National Laboratory DirectorsAward—1999

✟ Patent Application Submitted


6

Future DirectionsAs this program focuses on enabling technologies to support technology program goals, the AdvancedPropulsion Materials program will continue to work closely with PNGV partners and industry tounderstand propulsion materials-related requirements. Building upon the recent advances in materialstechnologies, many of this year’s projects will be moved out of the laboratory and over to industry fortesting. Projects such as the microwave regenerative exhaust filter will continue to optimize thetechnology design in preparation for vehicle testing. Other projects will continue to refine manufacturingrequirements and characteristics necessary to meet the challenges of the PNGV program. Severaltechnologies, such as the composite molding process technology for fuel cells, will be transferred toindustry for commercialization.

The most promising new activity in the Advanced Propulsion Materials program for FY 2000 is thecollaboration between Oak Ridge and Argonne National Laboratories in the development of smaller,lighter, and more efficient radiators. To reduce the size and weight of the radiator, heat transfer must beimproved on the airside; once that is accomplished, heat transfer characteristics of the coolant must beimproved. Carbon foam materials developed at Oak Ridge and nanofluids developed at Argonne areadvanced heat transfer materials with significantly higher thermal conductivities and improved heattransfer characteristics than are presently available. Therefore, combining carbon foam technology withnanofluid technology could lead to a breakthrough in the design of advanced vehicle thermal managementsystems that can meet the needs to improve heat transfer on both the air side and coolant side of theradiator.

As advanced automotive technology developments uncover new challenges, the Advanced PropulsionMaterials program will continue to provide breakthrough technology solutions. Collaborating withindustry, PNGV partners, national laboratories, and small businesses, the Advanced Propulsion Materialsprogram will continue to serve the needs of power electronics, fuel cells, combustion engine andaftertreatment, and all other critical technology areas.

Project AbstractsThe remainder of this report communicates the progress achieved during FY 1999 under the AutomotivePropulsion Materials program. It consists of 14 abstracts of national laboratory projects—5 that addresspower electronics, 4 that address combustion and emission technologies, 4 that address fuel cells, and 1that summarizes the closeout of the ceramic gas turbine activities. The abstracts provide an overview ofthe critical work being conducted to improve these systems, reduce overall cost, and maintain componentperformance. In addition, they provide insight into the challenges and opportunities associated withadvanced materials for high-efficiency automobiles.

Patrick B. Davis, Program ManagerEnergy Conversion TeamOffice of Advanced Automotive TechnologiesOffice of Transportation Technologies


7

2. POWER ELECTRONICS

A. Carbon Foam Thermal Management Materials for Electronic Packaging

James Klett, Lynn Klett, Tim Burchell, Claudia WallsOak Ridge National Laboratory, Oak Ridge TN, 37831-6087(865) 574-5220; fax: (865) 574-6918; e-mail: [email protected]

DOE Program Manager: Patrick Davis (202) 586-8061; fax: (202) 586-9811; e-mail:[email protected] Technical Advisor: David Stinton (865) 574-4556; fax: (865) 574-6918; e-mail:[email protected]

Contractor: Oak Ridge National Laboratory, Oak Ridge, TennesseePrime Contract No.: DE-AC05-96OR22464

Objectives

Collaborate with potential manufacturers with regard to designing, testing, and manufacturing ofsmaller, lighter, and more efficient heat exchangers and heat sinks for power electronics, utilizingthe high conductivity carbon foam.

Utilize improved understanding of effects of processing conditions on foam structure to increasethe thermal conductivities of high conductivity carbon foams by more than 50%.

OAAT R&D Plan; Task 4; Barriers B, C, D

Approach

Vary processing conditions and characterize foam structures to gain understanding of effects ofprocess conditions on structures that affect thermal properties.

Work with industrial partners to define targets for foam-based heat exchangers.

Develop joining/bonding techniques to laminate the foam to heat exchangers without sacrificingthermal conductivity.

Accomplishments

Met program goal for increasing thermal conductivity by over 50%.

Licensed the technology to Poco Graphite, Inc. to commercialize the material and scale up forproduction for large volume markets.

Developed bonding techniques with epoxies and brazing for laminating foam to differentmaterials.

Developed electroplating technique for improving the ruggedness of the foam.


8

Future Direction

Develop logic and system design for integration of the foam into application-specific devices.

Improve understanding of pressure drop/density/thermal conductivity/heat transfer coefficientrelationships to ensure optimum weight-efficiency tradeoffs.

Increase heat flux/weight by up to 10-fold

Introduction

Two devices are currently used for thermalmanagement: heat exchangers, which transferheat energy from one area of a device to another;and heat sinks, which dissipate heat into the air.Currently, most cooling heat exchangers for high-power electronics use layers of water-cooledaluminum or copper plate mounted below theelectrical circuitry to transfer heat from hotterareas to cooler areas. Using carbon foam as thecore material for these heat exchangers, theeffective transfer of heat can be significantlyincreased while the size and weight of the heatexchanger is reduced.

A new, less time-consuming process forfabricating mesophase pitch-based graphiticfoams without the traditional blowing andstabilization steps has been developed at OakRidge National Laboratory (ORNL) and is thefocus of this research. Initially these foamspossess a thermal conductivity of 106 W/m∙K at arelatively low density of 0.54 g/cm3. Potentially,the process will lead to a significant reduction inthe cost of carbon-based thermal managementmaterials (i.e., foam-reinforced composites andfoam core sandwich structures).

Experimental

In this research, two 100% mesophase pitcheswere used to produce graphitic foam: MitsubishiARA24 naphthalene-based synthetic pitch(melting point of 237°C) and a proprietarymesophase pitch from Conoco Corporationlabeled “Conoco Dry Mesophase” (melting pointof 355°C). All foam samples were carbonized at0.2°C/min to 1000°C and then graphitized at10°C/min in argon to 2800°C with a 2-hour soak,at temperature.

In order to develop a fundamentalunderstanding of the foam structure and graphitic

morphology, and therefore develop an ability totailor the thermal properties, samples wereexamined using a JOEL scanning electronmicroscope. Also, the thermal conductivity, , ofthe foam was determined with a xenon flashdiffusivity technique. The thermal diffusivity, ,was first measured on samples of 12-mmdiameter by 12-mm thickness on a custom-builtmachine in the High Temperature MaterialsLaboratory at ORNL. The sample density, , andspecific heat capacity, Cp, (assumed to be 713J/Kg∙°C) were then used to calculate the thermalconductivity with the following relation:

= ∙ ∙ Cp.

Finally, the same foam samples used forthermal diffusivity were machined to cylinders of12-mm diameter by 6-mm thickness for X-raydiffraction studies, giving an understanding of therelationship between processing conditions andthe graphitic structure of the foam.

Several 38.1-mm-thick foam blocks weremade from AR mesophase pitch with the standardORNL process. Sandwich panels wereconstructed from 12.7-mm-thick, 152.4-mm-diameter foam core sections machined from thethick blocks. Both aluminum 3003-H14 andcopper 110, 0.635-mm-thick, were used as facesheets. A thermally conductive film adhesive,T-gon 1/KA-08-128 (0.203 mm, 8 W/m∙K), wasused to bond the face sheets to the foam core witha cure at 0.241 MPa, 150°C for 30 minutes.Although a slightly higher pressure wasrecommended for curing the film adhesive, 0.241MPa was found to be sufficient for bonding to thefoam.

Flexural tests were conducted on 107-mm by27.9-mm samples according to ASTM C393-94for 4-point bending with two-point loading andone-quarter span. This specimen geometry was


9

chosen to produce core shear or bond failure.Compression testing was conducted at 5.08mm/min for 19-mm-square samples.

The thermal conductivity of the laminatedsamples was determined by the same method asthat used for the basic foam.

Results and Discussion

Figures 1 (a) and (b) are scanning electronmicrographs of the pore structure of theMitsubishi ARA24- and Conoco-derived foams,respectively, heat treated at 1000°C. The Conocopitch yielded foams with marginally higherdensities than foams produced with the ARA24mesophase pitch. The ARA24 pitch-derivedfoams exhibited a larger mean pore size than theConoco pitch-derived foams (275 m vs. 60 m).The higher melting temperature of the Conocopitch yields higher viscosities during processing,and therefore smaller bubble sizes.

Both foams exhibit a spherical cell structurewith open, interconnected pores (P in Figure 1)between most of the cells. It is evident from theimages in Figure 1 that the graphitic structure inboth foams is oriented parallel to the cell wallsand highly aligned along the axis of the ligaments(L in Figure 1).

It can be seen in the ARA24-derived foamsthat the graphitic structure is less aligned in thejunctions between ligaments (J in Figure 1) and

possesses more folded, mosaic texture. It ispostulated that this arises from the lack of stressesat this location during forming.Figure 2 is the X-ray diffraction spectra for bothfoams. The d002 spacing was calculated to be0.3355 nm for the ARA24 foam and 0.3360 nmfor the Conoco-derived foam. This is significantlybetter than in existing high-performance carbonfibers such as K1100 (0.3366) and vapor-growncarbon fibers (0.3366) and better than mostsynthetic carbons. The crystallite sizes (La and Lc)were similar to typical high-thermal-conductivitycarbon fibers.

The thermal conductivity of the ARA24 foamgraphitized at 10°C/min ranged from 50 to 150W/m∙K, and the Conoco derived foams exhibitedthermal conductivities ranging from 40 to 135W/m∙K (Figure 3). These conductivities areremarkable for materials with such low densities,0.27 to 0.57 g/cm3 (density was varied by varyingprocessing conditions).

Under close examination of the scanningelectron microscope images in Figure 1,microcracks and delaminations of the graphiteplanes can be observed. These are most likely dueto the thermal stresses induced duringcarbonization and graphitization as a result of thelow thermal conductivity of the foam prior tographitization. It was postulated that reducing theheating rates during this process would minimizethermal gradients, and thus minimize thermal

Figure 1. Structure of (a) Mitsubishi ARA and (b) Conoco Mesophase pitch-derived carbonfoams carbonized at 1000°C.

J

J

P

P

L L

400 m

(a) (b)


10

stresses, resulting in fewer microcracks anddelaminations. In fact, when the graphitizationrate was slowed to 4°C/min, the thermalconductivity of the ARA24-derived foamsincreased significantly to nearly 190 W/m∙K(Figure 3). This is significantly better than thetargeted 50% increase in thermal conductivity.

Four sandwich panels were fabricated fortesting: two with aluminum face sheets and twowith copper face sheets. The flexure specimensexhibited classic shear failure with only a slightdelamination of the foam from the face sheet. Thecore shear stress ranged from 1.49 to 2.35 MPa,while the shear modulus ranged from 47.9 to 111MPa. The values for panels with copper facesheets had a significantly higher core shear stressand core shear modulus that has not beenexplained.

A typical load-displacement curve for thecompression tests shows that the foam corecrushes with a fairly uniform load over a largedisplacement. The compression strength andmodulus ranged from 1.2 to 2.5 MPa and 44 to176 MPa, respectively.

The results of the thermal conductivity testingof the sandwich panels indicated that thesandwich specimens had a through-the-thicknessthermal conductivity of between 50 and65 W/m∙K with little difference between thealuminum and the copper sandwich panels.Although the thermal conductivity was lower thanthat of the basic foam because of the relativelylow-conductivity adhesive, the specificconductivity of the sandwich panels iscomparable to that of aluminum.

The average adhesive thickness in thesandwich panels was between 0.127 and0.203 mm. With a thermal conductivity of only8 W/m∙K, the interface was the limiting factor forthe through-thickness conductivity. Methods toimprove the through-thickness thermalconductivity include using a higher-conductivityadhesive and decreasing the adhesive thickness.Several additional sandwich panels have beensuccessfully bonded with thinner bond lines offilled epoxies (approximately 0.0254 mm). Also,a brazing technique has been developed forbonding aluminum face sheets (thermalconductivity of the brazing material isapproximately 45 W/m∙K).

Applications

Since the foam is open cellular, it is a primecandidate for use as a porous media heatexchanger for a power electronics substrate.Currently, most substrates for high-powerelectronics include a water-cooled aluminum orcopper plate mounted below the circuitry. It canbe shown that the effective heat transfercoefficient can be raised from ~ 250 W/m2∙K forcurrent designs to over 10,000 W/m2∙K for flowthrough a porous graphite foam. It was proposedthat a once-through foam core/aluminum-platedsubstrate be fabricated to replace the currentsubstrates (see Figure 4). The foam corethickness, geometry, channel patterns, foam cellsize, and heat treatment temperature will be

2

20 40 60 80

Inte

nsity

0

5000

10000

Conoco Derived Foam

ARA24 Derived Foam

Figure 2. X-ray diffraction patterns of ARA24 andConoco derived foams graphitized at2800°C.

0

20

40

60

80

100

120

140

160

180

200

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Specific Gravity

Th

erm

al

Co

nd

uc

tiv

ity

[W

/m∙K

]

AR Derived Foam

Conoco Derived Foam

Figure 3. Thermal conductivity of ARA24 andConoco derived foams as a function ofdensity.

Graphitization at 4°C/min

Graphitization at 10°C/min

ARA24 Conocod002 0.3355 nm 0.3360 nmLa 20 nm 18 nmLc 51 nm 84 nm


11

IGBT

Coo l Air

InH ot Air

O ut

A ir Flow

A ir Flow

A luminumH ou sing

Blind ho les to m in imi ze pr essure d rop

A ir F low ThroughF oam P orous M ed ia

Figure 4. Schematic and photograph of design forfoam-based heat exchanger substrate forpower electronics cooling.

evaluated to give optimum heat removal at thelowest flow rate of cooling fluid.

In a separate test, heat transfer coefficients fora shell-and-tube heat exchanger with carbon foamas the core were measured as high as 11,000W/m2∙K. This test validated the capability of thefoams to remove large amounts of heat withcooling air instead of water.

Furthermore, independent tests of the foammaterial at Florida International University haveconfirmed the dramatic improvement in heattransfer coefficients when the foam is used as aporous heat transfer medium. The majorconclusion of this research is that naturalconvection is not enough to improve the heattransfer: the air must be forced through the foamto realize the full potential of the material.

A closed-form mathematical model wasdeveloped at the University of Tennessee topredict the thermal conductivity of the foambased on the structure and density. This model isbeing expanded to study the heat transfercharacteristics of the foam under forced flow.

Last, extensive meetings are being conductedwith DaimlerChrysler, Lockheed Martin, Boeing,Modine, Peterbilt, and a NASCAR racing team todevelop this material for radiator systems andother electronic cooling applications.

Conclusions

The manufacture and properties of high-thermal-conductivity carbon foams have beenreported. It was shown that pitch precursorcharacteristics will affect foam structure andproperties such as bubble size, ligament structure,and thermal conductivity. The highly alignedligaments have similar structures to high-thermal-conductivity carbon fibers, such as K1100 andVGCF. In fact, the d-spacing was less than that ofVGCF, which has exhibited thermalconductivities as high as 1950 W/m∙K. Theseproperties, combined with the continuousgraphitic network, result in a specific thermalconductivity of up to 6 times greater than that ofcopper. Through this essential materialscharacterization, it was determined that slowerheating rates during carbonization andgraphitization would result in a dramaticimprovement in thermal conductivity, nearly 75%better than the initial values. Hence, the mainprogram objective was met.

Standard laminating techniques were shownto be viable for producing foam core sandwichpanels. However, either thin bond lines or brazedinterfaces were found necessary to preserve thehigh thermal conductivity. With furtherdevelopment, carbon foam can replacehoneycomb in applications that require highthermal conductivity and low weight.

High heat-transfer coefficients have beenmeasured, and heat exchangers have beendesigned and built with the knowledge learned inthis program. In future research, extensive testsand redesigns will be conducted to build a properheat exchanger substrate for power electronicscooling and other cooling applications for thePartnership for a New Generation of Vehicles.

References/Publications

1. J. W. Klett, “Mesophase Pitch-based CarbonFoam: Effects of Precursor on ThermalConductivity,” The First Annual Carbon


12

Foam Workshop, Kettering, Ohio, August16–18, 1999.

2. J. W. Klett, “Mesophase Pitch-based CarbonFoam: Effects of Precursor on ThermalConductivity,” The 23rd Annual Conferenceon Ceramic, Metal, and Carbon Composites,Materials, and Structures, Cocoa Beach,Florida, January 1999.

3. J. W. Klett, “High Thermal Conductivity,Pitch-based Carbon Foam for ThermalManagement Applications,” The 22nd AnnualConference on Ceramic, Metal, and CarbonComposites, Materials, and Structures, CocoaBeach, Florida, January 1998

B. dc Buss Capacitors for PNGV Power Electronics

B.A. Tuttle, D. Dimos, G. Jamison, P.G. Clem, P. Yang, and D. WheelerMS 1405, Sandia National LaboratoriesP.O. Box 5800, Albuquerque, NM 87185-1405(505) 845-8026, fax: (505) 844-2974; e-mail: [email protected]


Contractor: Sandia National Laboratory, Albuquerque, New MexicoPrime Contract No.: 04-94AL85000

Objectives

Develop a replacement technology for the presently used aluminum electrolytic dc busscapacitors for year 2004 new-generation vehicles.

Develop a high-temperature polymer dielectric film technology that has dielectric propertiestechnically superior to those of Al electrolytic dc buss capacitors and is of a comparable orsmaller size.

Develop a low-cost, multilayer ceramic technology that results in capacitors that are technicallysuperior to presently used Al electrolytic capacitors.

OAAT R&D Plan; Task 4; Barriers A, B, C, D

Approach

Contact automobile design and component engineers, dielectric powder and polymer filmsuppliers, and capacitor manufacturers to determine state-of-the-art capabilities and to definemarket enabling, technical goals.

Synthesize conjugated polyaromatic chemical solution precursors that result in dielectric filmswith low dielectric loss and excellent high-temperature dielectric properties.


13

Use a molecular engineering approach to create higher polarizabilities in polymer films, leadingto higher dielectric constants, by substitution of side groups and bridging molecules.

Fabricate, microstructurally analyze, and electrically characterize barium titanate and lead-basedceramics dielectrics in layers of suitable thickness for use in 2004 automobiles.

Accomplishments

Determined critical issues for dc buss capacitors: (1) cost and hard breakdown for ceramicmultilayers and (2) higher dielectric constants for polyfilm capacitors.

Developed technologies for single-layer deposition thickness of 3 m for polyfilm capacitors andlayer thickness ranging from 20 m to 50 m for ceramic capacitors using computer-controlledmicro-pen and tape cast technologies (these would be suitable for 600-V dc buss capacitoroperation).

Designed and fabricated polyfilm dielectrics with dielectric constants of 4 and dielectric loss lessthan 0.003, which exceeded the properties of state-of-the-art polyphenylene sulfide.

Fabricated ceramic dielectrics with high permittivity that met or exceeded manufacturer’sspecifications and that were compatible with low-cost electrode approaches (ultralow fire andbase metal electrode). Performed a trade-off study of performance versus electrode cost for low-fire barium titanate dielectrics.

Future Directions

Enhance dielectric constant of polymer film dielectrics to greater than 6 and keep dissipationfactor below 0.01. Commercial manufacturers have stated that these properties would spurcommercial development.

Develop low-cost electrode, ceramic dielectric capacitor technology that can withstand highelectrical breakdown fields.

Fabricate multilayer polyfilm and ceramic dielectrics that meet PNGV electric field operationcriteria to reduce size of units.

Introduction

Sandia National Laboratories (SNL) hasactively interacted with a number ofrepresentatives from the automobile industry toobtain their perspective on what is needed for2004 automobiles. These representatives includedGary Crosbie and Paul Crosby (Ford), BalaramaMurty and Jim Nagashima (General Motors), andAlvin White (Daimler Chrysler). Sandia has alsoactively interacted with a variety of capacitormanufacturers and dielectric powder suppliers.These suppliers included AVX, Murata, Kemet,Degussa, TAM, Ferro, TPL Inc., TRS Ceramics,Materials Research Associates, Tokay, and TPCLigne Puissance.

These interactions led us to the conclusionthat the two most viable replacement technologiesfor the electrolytic dc buss capacitors by 2004were multilayer ceramic and multilayer polymerfilm capacitors. Although the greatest concernregarding the multilayer ceramic capacitors(shown in Figure 1) was the cost, reducing thesize of the polymer capacitors was most oftencited by automobile design engineers as a neededtechnology-enabling breakthrough. Thus, Sandiais targeting technical solutions that will reducethe cost of ceramic multilayer capacitors and thatwill increase the dielectric constant of thepolyfilm dielectrics, thereby leading to smallercapacitors. The industry would like to havehigher field operation for low-cost multilayerceramic technologies, such as base metal (BME)


14

Figure 1. Multilayer ceramic capacitor schematic and micrograph of cross-sectional view.

and ultra-low-fire (ULF) dielectrics. A dielectricconstant of 1000 at a dc field of 200 kV/cm hasbeen stated as a goal for power electronics byAVX and the Center for Dielectric Studies atPenn State. Both the BME and ULF technologiessubstantially reduce the cost of multilayerceramic capacitors by permitting the use oflower-cost Ni or Ag electrodes instead of high-Pd-content electrodes. The electrodes compriseroughly 95% of the total cost of a ceramicmultilayer capacitor. Other considerationsexpressed by the auto manufacturers were cost,hard breakdown behavior in ceramics, and thehigh-temperature performance of polymer filmcapacitors. Inverter designs and operatingconditions for dc buss capacitors vary frommanufacturer to manufacturer. We presented ourassumption of 450-volt dc operation with 125-volt spike voltages for a nominal 500 F dc busscapacitor, which was well received at the DOEMerit Review Meeting.

Based on these criteria, an individualdielectric layer thickness of approximately 3 mfor polyfilm and 40 m for ceramic capacitors isprojected. Operating field strengths of 2 MV/cmand 150 kV/cm are projected for polyfilm andceramic dielectric capacitors, respectively. Basedon these assumptions and on measurement of

presently available commercial capacitors, sizecomparisons of 500 F dc Buss capacitors fordifferent technologies were obtained (Figure 2).

While size and outstanding temperatureperformance are advantages for the ceramictechnologies, soft breakdown behavior and lowercost are assets for polymer film capacitors.

Polymer Film Dielectric Development

SNL polymer film dielectric development hasbeen based on the request from manufacturersthat the new polyfilm dielectrics have voltage andtemperature stability that is equivalent to presentpolyphenylene sulfide (PPS) technology. Thus, astructural family of polymer dielectrics has beendesigned and synthesized to meet two of the moststringent PNGV requirements: (1) low dielectricloss and (2) extremely good temperature stability.Figure 3 shows a schematic diagram of Sandia’sconjugated, polyaromatic-based structure andindicates the large number of molecularmodifications to this structure that are possible.Our present effort emphasizes molecularengineering of higher-polarizability structuresthat will enhance dielectric constants, yet retainacceptable dielectric loss characteristics. A patentdisclosure has been issued covering the designand synthesis techniques for this polymeric

Electrolytic PPS polyfilm Projected Ceramic Projected 400 cm3 933 cm3 SNL poly 167 cm3 SNL ceramic K = 3 352 cm3 84 cm3

K = 8 2*Eb

Figure 2. Size diagram of 500 F dc buss capacitors of different technologies.


15

Figure 3. Schematic diagram of SNL conjugated polyaromatic film base structure.

family. Three initial molecular modifications tothe base structure were made: (1) propyl bridgesubstitution, (2) sulfur bridge substitution and(3) replacement of R side groups with high-electronegativity fluorine ions to enhancepolarizability.

The initial dielectric properties of a seriesof films of approximately 0.4 m thickness areshown in Table 1. The industry standard high-temperature performance polyfilm dielectric,PPS, has a dielectric constant of 3, a dissipationfactor of 0.003 and a breakdown field of2.2 MV/cm at 25 C. We have increased thedielectric constant to approximately 4, whilemaintaining similar loss and breakdown fieldcharacteristics. Breakdown field is very much afunction of the processing environment and, inproduction-type environments, it is anticipatedthat the breakdown field will be substantiallyenhanced. An example is shown in the bottomrow of the table for polyvinylidene fluoride(PVF2) films that were processed using achemical vapor deposition technique at Sandia.Breakdown fields of 5.5 MV/cm were obtainedfor these films and indicate the potential 6-foldincrease in energy density possible for the nextgeneration of chemically deposited polyfilms.Although the fluorine substitutions did notyield the hoped-for enhancement in dielectricconstant, numerous other promising

molecularly modified polymer dielectrics willbe synthesized in FY 2000.

Ceramic Dielectric Layer Fabrication andCharacterization

Ceramic dielectric layers were fabricatedusing two different techniques: (1) conventionaltape casting and (2) computer-controlledmicropen technology. For the micropendeposition, a ceramic slurry in the form of a30-micron-diameter tube is deposited onto thedesired substrate (Figure 4). We havedemonstrated high-quality dielectric layers ofthickness as low as 20 microns using thistechnique. Both chemically prepared bariumtitanate–based dielectric powders from Degussaand conventional state-of-the-art mixed oxidepowders from TAM have been fabricated intodielectric layers using this technique. Further,an initial study to determine dielectric propertiesas a function of firing temperature has beencompleted. Dielectric layers processed at 850 Chad lower dielectric constants, but they could beprocessed using inexpensive electrodes withhigh Ag content. Layers processed at highertemperatures had higher dielectric constants butrequired more expensive electrodes with high Pd

Table 1. SNL polyfilm dielectric properties for initial molecular modifications

Sample description Dielectric constant Dissipation factor Breakdown field(MV/cm)

Propyl bridge 4.0 .003 1.9Sulfur bridge 3.8 .002 1.7Fluorine side groupsubstitution

3.9 .001 1.8

PVF2 vapordeposited

4.1 .005 5.5

Y

Ar

ArAr

Ar

Ar Ar

R1 R2

R3 R4

X

n


16

Figure 4. Micropen CAD/CAM Direct Write(30 m to 500 m lines).

content for compatibility. The layers processed athigh temperature (1320 C) had dielectricconstants of 3000, while films processed at lowertemperature (900 C) had dielectric constants of1000. Studies of breakdown strength versustemperature are in progress for the layersproduced by the different processes.

In addition to micropen deposition,conventional tape casting processes weredeveloped. Presently, a 40- m-per-layer processhas been optimized for high-altitude depositionsand has resulted in dielectric layers withproperties comparable to those specified bycommercial manufacturers in state-of-the-artmultilayer ceramic fabrication facilities. Table 2shows the dielectric constants obtained fromchem-prep (Degussa) and mixed oxide (TAM)

barium titanate–based powders using differentdeposition methods.

SummaryCritical economic and technical issues for

improvements of dc buss capacitors for new-generation vehicles were determined throughdiscussions and visitations with automobiledesign engineers and capacitor manufacturers.We have been able to fabricate new polymer filmdielectrics with a 33% increase in dielectricconstant compared with industry standards, whilemaintaining voltage and dielectric loss stability.The goal in FY 1999 is to increase the dielectricconstant of the Sandia-designed polyfilms by100% while keeping losses below 1%. Thus, thepolymer film task is at the stage where theappropriate foundation or backbone chemistrydevelopment is completed. In the future, mole-cular modifications to this backbone chemistryshould lead to substantial enhancements indielectric constant.

Two different techniques were developed tofabricate ceramic layers of a thickness suitable forhigh-voltage dc buss capacitor applications.Dielectric properties of both tape-cast andmicropen-deposited layers were equivalent tostate-of-the-art commercial multilayer dielectrics.Thus, the foundation has been formed for theevaluation and enhancement of both ULF andBME ceramic dielectrics via proper selection ofdopants and low-level additives in FY 2000.


None.

Table 2. Dielectric properties of micropen and tape cast layers

Sample K, commercialmultilayer

K, SNLmicropen

K, SNLtape cast

K, SNLbulk ceramic

Degussa, X76CP, 1310 C

3200 3200 3400 2595

TAM X7R262L, 1140 C

2242 2240 2056 2062


17

C. Mechanical Reliability of Electronic Ceramics and Electronic Ceramic Devices

A. A. Wereszczak and K. BrederMechanical Characterization and Analysis Group (MCAG)Oak Ridge National Laboratory, P.O. Box 2008, MS 6069, Bldg. 4515Oak Ridge, TN 37831-6069(865) 574-7601; fax: (865) 574-6098; e-mail: [email protected]



Objectives

Develop testing algorithms that can be used to assess electronic ceramic (EC) and electronicceramic device (ECD) mechanical reliability.

Mechanically characterize EC and ECD alternatives that are less expensive and that can be usedto promote device miniaturization.

OAAT R&D Plan: Task 4; Barriers A, D

Approach

Utilize micromechanical testing and ceramic-specific-characterization testing facilities to measure insitu mechanical properties of ECs in the ECDs.

Characterize presently used and developmental EC and ECDs supplied from their manufacturersor their end-users.

Use already-developed ceramic component prediction codes (whose development was funded bythe Office of Transportation Technologies/Ceramic Technology Project for structural ceramics)and their statistical analysis capabilities in analyzing the mechanical strength and fatigue of ECsand ECDs.

Provide results and insights back to manufacturers that will result in the improved reliability ofECs and ECDs.

Accomplishments

A mechanical properties microprobe (MPM) was used to characterize BaTiO3 dielectrics insnubber multilayer capacitors (MLCs). It was found that mechanical performance (e.g., fracturetoughness, strength) of ceramic dielectrics can be significantly different in equivalent snubberMLCs.

Strength and fatigue of two alumina substrates and an aluminum nitride substrate werecharacterized and mechanical design data were generated for use with ECDs.


18

The fatigue and strength testing of candidate tape-cast aluminas for automotive gas exhaustsensors was initiated in a collaboration with Motorola.

Future Direction

Acquire commercially available dc buss MLCs and measure the in situ mechanical performanceof their dielectric ceramics. Collaborate with Sandia National Laboratories in the characterizationof the dielectric ceramic in their developmental dc buss MLCs for automotive power electronicsbuilding blocks.

Characterize less expensive Ni-electrode snubber MLCs.

Study the utility of non-destructively measuring residual stresses in MLCs withpiezospectroscopy and non-destructively identifying damaged MLCs with resonant ultrasoundspectroscopy.

Introduction

A lack of mechanical reliability of ECs inECDs can often limit the reliability of theirelectronic function. Three classes of ECs or ECDsthat this project examined in FY 1999 weremultilayer capacitors (MLCs), EC substrates, andoxide ceramics for automotive exhaust gassensors; the service life of all three can in fact belimited by their mechanical reliability. Theapplication of ceramic life prediction codes(developed for structural ceramic componentdesign in high-temperature gas turbine engines) isused in concert with the mechanical testinganalyses of the ECs because they portray theprobabilistic strength and fatigue properties ofECs in an appropriate (but underutilized) manner.

One MLC manufacturer states that 40–50%of its capacitor failures are mechanically induced.Although its MLC failure rate is only on the orderof 1 part per million, since the companymanufactures 30–40 million capacitors a day, thisfailure rate is actually significant. Therefore, thecompany (and its customers) have a vestedinterest in improving MLC mechanical reliability.A goal of this project is to help MLCmanufacturers improve this reliability.

Measuring the mechanical performance ofdielectric ceramics in MLCs is not trivial becauseof their very small size. An MPM in theMechanical Characterization and Analysis Groupwas used to measure in situ fracture toughness ofthe dielectric ceramic in equivalent, commerciallyavailable, snubber MLCs. Although the MLCshad equivalent capacitances, the fracture

toughnesses of their dielectrics were statisticallydifferent, as shown in Figure 1. Fracturetoughness is an indicator of mechanicalrobustness, and an end-user of these three MLCs(an automobile manufacturer) stated that it hadexperienced increasing reliability with the MLCsin this examination with higher fracturetoughnesses. These results suggest that MLCsthat have a dielectric ceramic with maximumtoughness will offer better mechanical reliability.

Additional techniques to assess mechanicalrobustness of MLCs were also examined anddeveloped. Image analysis was coupled withfracture mechanics and Weibull theory to“calculate” the strength distribution of dielectricceramics in MLCs. This technique, along withfracture toughness measurement, helps portraythe whole mechanical performance of dielectricceramics.

��

��

��

0.0

0.5

1.0

1.5

2.0

MLC "A" MLC "B" MLC "C"

Ave

rage

Fra

ctur

e T

ough

ness

, KIc

, (M

Pa¦

m)

Supplier's Capacitor

1.53

1.36

1.11

Average of49 values

Average of65 values

Average of65 values

Figure 1. The fracture toughness of the dielectricceramic in equivalent snubber multilayercapacitors was different. The differencewas statistically significant.


19

The strength and fatigue of threecommercially available ceramic substrates werecharacterized. Two of the substrates werealuminas (a tape-cast and a roll-compactedmaterial), and the third material was an aluminumnitride (AlN). The tape-cast alumina (Al2O3) wasstronger and more fatigue-resistant than the roll-compacted Al2O3 (commonly used in ECDs), butit was more costly. The AlN was not as strong asthe tape cast Al2O3, but it was stronger than theroll-compacted Al2O3 and much more fatigue-resistant, as compared in Figure 2 (and was themost expensive of the three). The results fromthis study will help end-users of ceramicsubstrates appropriately design with and choosethem.

-6

-5

-4

-3

-2

-1

0

1

2

100 1000

ln ln

( 1

/ (

1 -

Pf )

)

Failure Stress (MPa)

Initial

Probability of F

ailure, Pf , (%

)

99.999.0

90.0

50.0

20.0

10.0

2.0

5.0

1.0

0.5

200 300 400 600 800

1 day

1 week

1 month

1 year

96% Al2O

3

Figure 2a

-6

-5

-4

-3

-2

-1

0

1

2

100 1000

ln ln

( 1

/ (

1 -

Pf )

)


Initial = 1 day =

1 week =1 month =

1 year

Probability of F

ailure, Pf , (%

)

99.999.0

90.0

50.0

20.0

10.0

2.0

5.0

1.0

0.5

200 300 400 600 800

AlN

Figure 2b

Figures 2 a and b. A 96% alumina material (2A) wasfound to be susceptible to fatigueat 20°C, while AlN (2B) was not.The alumina tested is a commonlyused ceramic substrate; the AlNsubstrate is more costly but hasbetter thermal conductivity and athermal expansion similar to thatof silicon.

Motorola is developing an automotiveexhaust gas sensor; this project is assisting itsdevelopment. The sensor is a multilayer designand will be using Al2O3 as a matrix; however, analumina with optimum strength and fatigue(which is also inexpensive) has not yet beenidentified. The mechanical testing of three Al2O3swas initiated during FY 1999 to measure theirstrength as a function of temperature, as well astheir fatigue performance. The highest purity ofAl2O3 examined had the lowest strength at roomtemperature, while the least pure Al2O3 had thehighest strength at room temperature, as shown inFigure 3. However, recent results show that thelowest-purity Al2O3 has the worst fatigueresistance of the three and was no longer thestrongest Al2O3 of the three at 1000°C (anexpected service temperature of the sensor).

Summary

The present project’s FY 1999 mechanicaltesting results, characterization, and analyses ofall three classes of these ECs and ECDs will helpmanufacturers (1) identify ECs that will prolongthe service life of ECDs and (2) design theirECDs so that deleterious stresses are not imposedwithin them during their manufacture or service.Fracture toughness and strength results andanalyses of dielectric ceramics in MLCs are beingused by MLC manufacturers to improve theoverall mechanical robustness of their MLCs.Strength and fatigue results and analyses of Al2O3

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

2.0

40 60 80 100 300

ln ln

( 1

/ (

1 -

Pf )

)


Probability of F

ailure, Pf , (%

)

99.999.0

90.0

50.0

20.0

10.0

2.0

5.0

1.0

0.5

200 50030

92% Alumina

96% Alumina

100% Alumina

Figure 3. The 20°C strength of 92% alumina was thehighest, followed in turn by the 96% and100% aluminas. All aluminas arecandidates for Motorola’s automotiveexhaust gas sensor.


20

and AlN ceramics will help ECD manufacturerschoose an appropriate ceramic substrate (anddesign) for their ECD that promotes longerservice life. Finally, high-temperature strengthand fatigue results and analyses of candidateAl2O3s will help this original equipmentmanufacturer identify the optimum and best-performing Al2O3 for their developmentalautomotive exhaust sensor.


1. A. A. Wereszczak, T. P. Kirkland, and K.Breder, “Biaxial Strength, Strength-Size-Scaling, and Fatigue Resistance of Aluminaand Aluminum Nitride Substrates,” in review,International Journal of Microcircuits andElectronic Packaging, 1999.

2. K. Breder, A. A. Wereszczak, L. Riester, andT. P. Kirkland, “Determination of Strengthfor Reliability Analysis of MultilayerCeramic Capacitors,” in press, Ceramic

Engineering and Science Proceedings, Vol.20, 1999.

3. A. A. Wereszczak, A. S. Barnes, and K.Breder, “Probabilistic Strength of{111} n-Type Silicon,” in review,International Journal of Microcircuits andElectronic Packaging, 1999.

4. A. A. Wereszczak, R. S. Scheidt, and K.Breder, “Probabilistic Thermal ShockStrength Testing Using Infrared Imaging,” inreview, Journal of the American CeramicSociety, 1999.

5. A. A. Wereszczak, K. Breder, M. K. Ferber,R. J. Bridge, L. Riester, and T. P. Kirkland,“Failure Probability Prediction of DielectricCeramics in Multilayer Capacitors,” pp. 73–83 in Multilayer Electronic Ceramic Devices,Ceramic Transactions, V97, AmericanCeramic Society, 1999.

D. Low-Cost, High-Energy-Product Permanent Magnets

Tom M. Mulcahy and John R. HullArgonne National Laboratory, Energy Technology-335Argonne, IL 60439(630) 252-6145; fax: (630) 252-5568; e-mail: [email protected]

Andrew E. PayzantOak Ridge National Laboratory, MS-6064, P. O. Box 2008Oak Ridge, TN 37831-6064(865) 574-6538, fax: (865) 574-4913; e-mail: [email protected]


Contractor: Argonne National Laboratory, Argonne, IllinoisPrime Contract No: W-31-109-Eng-38

Objectives

Develop a low-cost process to fabricate NdFeB permanent magnets with up to 25% higherstrengths. The higher-strength magnets will replace ones made by traditional powder metallurgyand enable significant size and weight reductions of traction motors for hybrid vehicles.


21

Utilize high-strength superconducting magnets to improve the magnetic alignment of grains priorto pressing and sintering, therefore producing higher-strength magnets.

Collaborate with magnet manufacturers, who will provide powder, sinter the green compacts, andperform characterizations of the engineering magnetic properties.

OAAT R&D Plan: Task 3; Barriers B, C, D

Approach

Develop facilities to align magnetic domains of NeFeB powders within a high-strength magneticfield, created by a superconducting magnet, during forming operations.

Characterize, compare, and correlate engineering and microscopic magnetic properties ofmagnets processed under varying conditions, including some in current production.

Utilize a reciprocating feed to automate insertion of loose and compacted magnet powder into andout of the steady field of a superconducting solenoid.

Accomplishments

Completed design of an axial-die press for making 1- to 2-cm-diameter permanent magnets in abatch mode.

Fabricated and assembled an Axial-Die Press facility at Argonne National Laboratory thatincludes a 9-Tesla superconducting magnet.

Characterized magnetic properties of permanent magnets provided by manufacturers.

Future Directions

Optimize fabrication processing and powders from different industrial partners using axial-dieand isostatic pressing in batch mode operations.

Design, fabricate, and operate a reciprocating press in a continuous mode to demonstrate thefeasibility of competitive factory operation.

Optimize fabrication processing and powders from different industrial partners, using transverse-die pressing and a reciprocating press.

Provide design rules for the fabrication of permanent magnets, including knowledge for scale-upto larger size magnets at commercial rates of production.

A Cryomagnetics, Inc., superconductingsolenoid has been specified and purchased, afterconsultation with permanent magnetmanufacturers. The magnitude of the steady fieldin the bore of the solenoid can be continuouslyvaried up to 9 Tesla. Magnet powder can bealigned in a field that is uniform within 5%, overa volume that is large enough to axial-die press1- to 2-cm-diameter cylindrical magnets that havesimilar lengths. Alternatively, a 2.5-cm-diameter

by 12.5-cm-long volume of powder, contained ina rubber mold, can be aligned for subsequentisostatic pressing.

The axial-die and punch set shown inFigure 1 was designed in consultation with thepermanent-magnet manufacturer UGIMAG, Inc.and the tooling fabricator Bronson and Bratton,Inc. The set was made using very low-magnetic-permeability material (<1.002), which will notaffect significantly the uniformity of the


22

��

��

�� 0.63

0.68

0.63

1.25

3.50

Right Punch (Inconel 718)

Left Punch (Inconel 718)

Die Case-2.5 OD (Inconel 718)

Die Insert-1.5 OD (6% Ni Carbide)

Powder

Figure 1. Axial-die and punch set for 5/8 in. magnet;powder shown is compacted.

superconducting solenoid's magnetic field or thealignment of the magnet powder. However, self-demagnetization fields are inherent in shortcylinders of magnet powder and can affect theapplied field uniformity and grain alignment ofpermanent magnets. This happens in currentproduction, where, typically, electromagnetsprovide alignment fields of less than 2 Tesla,which approaches the saturation limit of the steelpresent in commercial presses. Vector fieldelectromagnetic code calculations were made thatshowed demagnetization effects were stillpronounced up to applied alignment fields of7 Tesla, which can be exceeded with thesuperconducting solenoid purchased.

An unorthodox press-in-tube method foraxial-die pressing in a batch mode was devisedand built by Ability Engineering Technology,Inc., which met the magnetic, geometric andprogram cost constraints associated with usingsuperconducting solenoids (see Figure 2). Thepress tube and ram are made from very low-magnetic-permeability materials (< 1.002). Thehydraulic cylinder was custom made, by AtlasCylinders, Inc., from stainless steel with a lowpermeability (< 1.010). The length of the press

tube was made sufficient to locate the moremagnetic hydraulic cylinder in the far field of thesolenoid, where the magnetic-field gradient isweak, and to access the solenoid, which isthermally shielded deep within its helium Dewar.Electromagnetic code calculations were made todetermine the far magnetic field and maintainminimal magnetic forces between the solenoidand the axial-die press.

The press insertion mechanism enablescomplete and controlled removal of a greencompact from the Dewar’s warm-bore tube, whilethe superconducting solenoid is operating and thehydraulic cylinder is activated. This featureallows the experimental determination ofextraction and insertion forces, which will beused to calibrate electromagnetic code modelsalready developed. Understanding the forces andtheir minimization is key to the design of areciprocating press. Also, the mechanism allowsrotation of the end of the press tube into thesolenoid’s weak fields (< 250 Gauss) for removalof the green compacts.

The capability to move the axial-die pressinto the far field of the solenoid allowsexploration of another processing method. Inisostatic pressing, the powder is aligned separatefrom the pressing operation. Argonne NationalLaboratory maintains an isostatic press thatexceeds the capabilities of magnet manufacturers.Discussions with manufacturers and a review ofthe journal and patent literature indicates that thegreater alignment fields of superconductingsolenoids will be less effective in energy-productimprovement, because self-demagnetizationeffects are much less for the long, powder-filledcylindrical rubber molds used in isostaticpressing. Also, because the commercialelectromagnetic alignment solenoids are separate

Figure 2 (a). Axial-die press in the solenoid. 2 (b). Press-tube interior; no compaction.

He Dewar

Superconducting Solenoid

Warm-Bore Tube

Access: He & Power

��

��

��

��

��

Press Tube

Hydraulic Cylinder

��

Insertion Mechanism

��

��

��

��

��

��

��

��

RamBase Plug

Press Tube

2.111


23

from the press and can be isolated fromferromagnetic material, they can providemultiple, higher-field alignment pulses (< 6 Tesla) to the powder. In addition,enhancements can be achieved by mechanicalmanipulation of the rubber molds. Currently,isostatic pressing is claimed to producepermanent magnets that are within 5–10% ofhaving optimal alignment and magneticproperties. However, the alignment of powder hasnever been attempted in the steady, high fieldsavailable with the superconducting solenoid.Other processing improvements may be possible,such as an ability to align a more economicalpowder that has a wider particle-size distribution.The magnet producer CRUMAX will participatein identifying any benefits.

The magnet manufacturers providedcharacterized permanent magnets: transverse-diepressed 26-mm cubes and two sets of isostaticallypressed cylindrical magnets with diameters of14–15 mm. The energy products of the cubeswere 41 MGOe. One set of cylinders had a45-MGOe energy product, which wassignificantly different from the other’s 35 MGOe.The two sets were made using different processes.At Argonne, the remnant induction of eachmagnet was measured by rotating it in aHelmholtz coil. The manufacturer’s values, whichranged from 1.2 to 1.35 Tesla, were confirmed.

At Oak Ridge, an examination ofmicrostructure and preferred orientation revealedclear differences between samples. Electronmicroprobe analysis of each manufacturer’sproducts showed the composition of their grainboundary phases to be slightly different; however,the primary phase was chemically identical. X-ray diffraction pole figures were used tocharacterize the texture on the face normal to themagnetization axis.

Data were collected at 5-degree tilt androtation increments using Cu k-alpha X-rayradiation with 1-mm diameter incident beamcollimation. The strongest preferred orientationwas observed for one of the isostatically pressedcylindrical magnets, as shown in Figure 3,whereas the other cylindrical magnet had theweakest texture, as shown in Figure 4. Both of

Figure 3. Oblique view of 006 pole figure of thecylindrical sample showing axial symmetrycharacteristic of strong “fiber” texture.Peak height = 105 cps.

Figure 4. Oblique view of 006 pole figure for thecylindrical sample showing axial symmetrycharacteristic of moderate “fiber” texture.Peak height = 49 cps.

these showed a fiber texture with cylindricalsymmetry.

The cube samples were strongly textured,only slightly less so than the strongest cylindricalsample, and both were nearly identical. Theirtexture was not perfectly cylindrical, as shown inFigure 5. This was not unexpected for transverse-die pressing, where compaction is unidirectionaland normal to the direction of magnetization.

The correlation of these texturemeasurements with the energy product andmethods of pressing is a clear demonstration ofthe importance of improved grain alignment andthe utility of microstructure analysis, in particular,X-ray diffraction analysis, in quantifying subtlevariations in texture produced by differentmethods of processing.


24

Figure 5. Top view of 006 pole figure for a cubesample, showing slightly distortedcylindrical symmetry. Peak height = 100cps.

E. Lead-Free Solders for Automotive Electronics

M. L. Santella, R. J. Barkman, and M. J. GardnerOak Ridge National Laboratory, Metals & Ceramics DivisionOak Ridge, TN 37830-6096(865) 574-4805, fax: (865) 574-4928; email: [email protected]



Objective

Support development of lead-free, high-temperature “engineered” solder alloys for automotiveelectronics applications such as power electronics devices by systematically determining theeffects of alloying on the properties, soldering characteristics, and microstructure of Sn alloys.

OAAT R&D Plan; Task 4; Barrier D

Approach

A series of binary alloys are formulated to determine the effects of individual alloying elementson mechanical properties, the wetting behavior on copper, and melting characteristics.

A micromechanical testing technique is being used to determine the mechanical properties ofsolder alloys in situ in solder joints.

Accomplishment

The feasibility of using automated ball indentation testing to measure the tensile properties ofsolder bumps was demonstrated. This indicates that measurement of solder joint properties is alsofeasible.


25

Future Direction

The effects of major alloying elements such as Bi, Cu, In, Sb, and Zn on tensile properties,wetting behavior, microstructure, and melting characteristics will be determined.

Alloying elements will be identified that improve the ductility of Sn, as this is expected tofavorably impact solder joint reliability and the resistance to fillet lifting in through-holeconnections.

Some effort will be directed toward perfecting micromechanical testing of solder joints.

Introduction

Lead is recognized as a significant threat toboth the environment and public health, andpending legislation could result in the banning oflead from a wide variety of products, includingelectronics. The overall objective of this project isto develop lead-free, high-temperature“engineered” solder alloys that will increase theoperating temperature of automotive electronicpackages from 120 C to 180 C. The initialemphasis is on studying the properties andcharacteristics of tin alloys because they are mostlikely to form the basis of high-temperaturesoldering alloys such as eutectic Sn-Ag.

During FY 1999, significant progress wasmade toward determining the feasibility of usingautomated ball indentation (ABI) testing tomeasure the tensile properties and strain ratesensitivity of solder joints. Because of itsimportance to reliability, the mechanical behaviorof existing and newly developed solder alloys is akey concern in the design and engineering ofadvanced electronic components. Standard tensiletesting is commonly used to determine alloystrength and ductility properties, but there isconcern that tensile testing may not provide anaccurate indication of actual solder jointperformance. Part of the basis for this concern isthat the thermal processing conditionsexperienced during joint manufacturing may notbe accurately reproduced in bulk alloy tensilespecimens. This means that the mass transportconditions and chemical reactions that occur injoints and that will influence their properties areunlikely to be reflected in standard test data. TheABI testing was viewed as a novel way tomeasure the mechanical properties of actualjoints.

Subsized tensile specimens were made from1.5-mm-thick rolled sheets of pure Sn, Sn–3.5Ag,Sn-37Pb, and Pb–5Sn (nominal compositionswt %). The Pb alloys were included as standardreference alloys. These specimens were tested atroom temperature to determine baseline tensileproperties.

Subsequently, ABI tests were performed onthe shoulder areas of the same specimens asindicated schematically in Figure 1.

1.000

.350

.070

.140 R.100 DIA

.300

Figure 1. Schematic of solder alloy tensile specimenshowing relative size of ABI indent onspecimen shoulder (dot on upper leftcorner).

Images of the ABI indentations taken in ascanning electron microscope (SEM) are shownin Figure 2. The indentations are circular, asrequired for data analysis, and the plastic zonessurrounding them are evident. During the tests,both the load and indentation depth are veryaccurately measured and used to determine aplastic flow curve where the true strain, p, andthe true stress, p, are given by

D

d pp 2.0

2

4

p

pd

P .


26

Figure 2. ABI indentations on tensile specimens of pure Sn (right) and Sn–3.5 Ag solder alloy (left).

The quantity dp is geometrically related toindentation depth, D is the indenter diameter, P isthe applied load, and is a constant determinedfrom the test data. The yield strength is alsodetermined from the ABI test using a separatecalculation.

Overall, however, there was good agreementbetween the two data sets. The ABI test wasoriginally developed for use on steels, for whichit provides very accurate measurements of tensileproperties. This experiment established that theABI test was also capable of determining tensileproperties of solder alloys with reasonableaccuracy (see Figure 3).

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

True strain

0

1000

2000

3000

4000

5000

6000

7000

Tru

e s

tre

ss

, p

si Tensile test

ABI test, 0.030" ballPower fit to ABI data

Sn-3.5Ag wt%, rolled sheet

Room temperature tests

Loading rate = 0.0001 in/s for both tests

Figure 3. Comparison of tensile and ABI test datafrom bulk specimen of Sn–3.5Ag solderalloy.

Next, small FR4 printed-circuit test boardswere obtained that contained arrays of solderbumps made with Sn–3.5Ag solder alloy. Thesolder bumps had nominal dimensions of 2.5 mm

length by 0.8 mm width by 0.30–0.50 mmthickness. The ABI test requires a flat surface foraccurate measurements, so flat spots were madeon the solder bumps using a polishing technique.The sizes and shapes of the solder bumps beforeand after preparation of flat spots are shown inFigure 4.

The appearance of several solder bumps afterABI testing is presented in Figure 5. Theindentations on the bumps at the top left positionand bottom right position of Figure 5 showuniform circular impressions that are consistentwith obtaining acceptable data. The indentationon the lower left has an irregularity on itsdiameter that was produced by a defect in thesolder bump. The impression on the upper rightis actually a porosity defect in the solder bumprather than an indentation from testing. Theexistence of defects was unexpected, and theyonly became apparent in the solder bumps as theABI tests were being conducted. Figure 5 furtherillustrates the accuracy with which the ABIindentations can be located on small specimens.

ABI test results from the solder bumps agreedwell with those from the tensile specimenshoulders, as shown in Figure 6. Compared withthe ABI data from the tensile specimens, thesolder bump data showed slightly higher flowstresses and work hardening rate. Also, the yieldstrength determined for the solder bump wasstrength determined for the solder bump washigher than that of the bulk alloy.


27

Figure 4. Sn-3.5Ag solder bumps on FR4 test board before (left) and after (right) polishing.

Figure 5. ABI indentations on Sn-3.5Ag solder bumps.

Figure 6. Comparison of ABI results from the bulkalloy and a solder bump of Sn-3.5Ag solderalloy.

There are a number of possible reasons forthe discrepancies between the tensile data and theABI data sets. Calculation of the yield strengthand the flow curve from the ABI data usesempirical constants that are determined from thematerials being tested. Confidence in the valuesof these constants increases proportionately withthe amount of available tensile test data and ABItest data. Also, specific tests, which were notperformed in this experiment, may be done todetermine the values of some of the constants.Another source of error relates to the strain ratesensitivity of the solder alloys, which is relativelyhigh. The strain rate during tensile testing isnearly constant until the onset of necking.However, during ABI testing, the strain ratedecreases by over an order of magnitude because

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

True strain

1000

2000

3000

4000

5000

6000

7000

8000

Tru

e s

tre

ss

, p

si

ABI test comparison for Sn-3.5Ag wt%

Solder bump B7Bulk alloy


28

of the geometry of the deformation under theindenter. Traditionally, controlling strain rateduring ABI testing was not a concern because it isnot important for determining the properties ofsteels.

Another source of error in the ABI tests onthe solder bumps is defects in the material. Aradiograph of the FR4 test board used for thisexperiment (Figure 7) indicates that most of thebumps contain porosity defects, as indicated bythe dark spots. Arrows are used in Figure 7 toindicate the bumps that were ABI tested; whilesome of the spots are indentations, most of themare porosity defects. Finally, one importantdifference between the data from the bulk alloyand that from a solder bump (Figure 6) is that thesolder bump has chemically reacted with thecopper on the test board. It is possible thecombined effects of reaction with the Cu and thedifferent thermal processing of the bumpproduced a measurable difference in properties.

These results show that valid ABI tests can beconducted on relatively small volumes of solderalloys, and that the data from such tests agree

Figure 7. Radiograph of FR4 test PC board used forABI testing.

well with tensile test data from the bulk material.The results indicate the reasonable possibility ofobtaining valid tensile property measurements onactual solder joints.


29

3. FUEL CELLS RESEARCH AND DEVELOPMENT

F. Composites for Bipolar Plates

M. S. Wilson and D. N. BusickMaterials Science and Technology Division, MST-11, MS D429, Los Alamos National LaboratoryLos Alamos, NM 87545(505) 667-6832; fax: (505) 665-4292; e-mail: [email protected]


Contractor: Los Alamos National Laboratory, Los Alamos, New MexicoPrime Contract No.: W-7405-Eng-36

Objective

Develop low-cost, mass-producible composites for fuel cell stack bipolar plates that arecorrosion-resistant, electronically conductive, and physically robust.

OAAT R&D Plan; Task 13; Barrier B

Approach

Use low-cost, commercially available raw materials (graphite powder and polymeric resinbinder).

Use thermosetting vinyl ester resins for corrosion resistance and short process cycle times.

Eliminate machining requirements by forming net shape plates (including fluid flow channels) ina single compression, injection, or injection-compression molding step.

Accomplishments

Collaborated with industry to provide materials suitable for mass production with enhancedproperties and processability.

Submitted U.S. Patent application.

Future Direction

Transfer technology to industry and continue to provide feedback, especially in the areas of fiberreinforcement and corrosion resistance.


30

A major focus area of FY 1999 research wasimprovement of the processability andmechanical properties of our baseline compositematerial. Although laboratory-scale platesproduced at Los Alamos National Laboratory(LANL) exhibited good properties, it wasrecognized that mass production would requiresignificant improvements in the processability ofthe molding compound. To achieve theseimprovements, we have been collaborating withPremix, Inc., and Bulk Molding Compounds, Inc.(BMC), both of which have expertise inmanufacturing vinyl ester molding compounds.Both companies produced molding compoundswith reduced cure times (2–5 minutes comparedwith 10–15 minutes for LANL formulations),extended shelf lives (several weeks compared toseveral hours for LANL formulations), and moreuniform flow. In addition, both companies haveutilized mold release agents within theircompounds. This prevents parts from sticking tothe mold, while avoiding the time- and labor-intensive application of a release agent to themold surfaces before each part is molded. Suchinternal mold release agents will be especiallycritical in molding fluid flow fields directly intobipolar plates. A mold designed by Plug Power,LLC, was used to mold plates with fluid flowchannel dimensions on the order of 1 mm fromLANL compounds in FY 1998 and from Premixcompounds in FY 1999. To date, even though thefeasibility of direct flow field molding has beendemonstrated, the bipolar plates used for fuel celltesting have had machined flow fields because offrequent design changes and the high cost oftooling.

Although compression molding has been theintended processing route for these graphite/vinylester compounds because of their high solidscontent, other manufacturing alternatives nowappear feasible. Premix has developed moldingcompounds that, with slight modifications, itexpects to be suitable for injection or injection-compression molding. BMC has already producedplates successfully from some of its compoundsusing injection-compression molding. Althoughthe injection-compression–molded plates had anelectrical conductivity somewhat lower than ourhistorical target value of 100 S/cm, goodperformance in fuel cells can be achieved with

plates of lower conductivity. Figure 1 comparesthe fuel cell performance of graphite plates andPremix composite plates with a conductivity of85 S/cm. Recent feedback from the fuel cellcommunity indicates that conductivities of60 S/cm, or possibly as low as 15 S/cm, may beacceptable for bipolar plates. With these lowerconductivity targets, the solids content of themolding compounds can be reduced so thatinjection molding is more easily achieved.

Figure 1. Polarization curves obtained from singlefuel cell tests using POCO graphite platesand Premix composite plates.

In LANL formulations developed inFY 1998, the incorporation of certain fibers intothe baseline bipolar plate material resulted insignificant strength improvements. However,similar improvements were not observed withcompounds from BMC and Premix incorporatingthe same fibers; in fact, the incorporation ofthese fibers actually appeared to affect compositestrength negatively in the new formulations. Twopossible reasons have been postulated. First, theLANL compounds were based on a fairly dilute,low-viscosity vinyl ester resin. In the case ofcotton fibers, strengthening presumably resultedfrom resin absorption by the fibers and thesubsequent formation of an essentially continuousfiber-matrix interphase. The cotton fibers may notbe able to imbibe the thicker resins used by thecompounders. Second, the LANL formulationsdid not incorporate mold release agents. Thelubricating action of the internal mold releasesused by Premix and BMC may hinder fiber-resinadhesion. Thus, the utility of fiber additivesremains a topic of investigation.

However useful fiber reinforcementseventually prove to be, the strength of composites

��

��

��

��

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.2 0.4 0.6 0.8 1

Graphite platesGraphite plates (corrected)Composite platesComposite plates (corrected)

Cel

l Vol

tage

(V

)

Current Density (A/cm 2)


31

with high particulate loading is determinedprimarily by the matrix material. Significantimprovements in strength can be achieved withthe use of different resin binders. However,changing the resin system often requiresadjustments to the graphite loading level tomaintain the target electrical conductivity. Acomposition of 68% graphite in a fairly dilutevinyl ester resin provided good conductivity intests at LANL, and it served as a “baseline” forsubsequent formulations developed by Premix.But when other binder systems were used, thedifferent rheologies and reduced extrusion duringmolding resulted in significantly lowerconductivities. When graphite loading wasincreased, good conductivities were regained, andhigher strengths were still obtained bycapitalizing on the improved resin systems. Ifconductivity requirements can be relaxed asexpected, graphite content can be reduced to giveeven stronger plates, since the resin matrix is thedetermining factor in composite strength. Typicalproperties of composites using different resinsystems with and without cotton fiberreinforcement are shown in Figure 2.

Figure 2. Property comparison for composites usingthree different resin binders, with andwithout cotton fibers.

Corrosion in the fuel cell environment is amajor issue in bipolar plate material development.Vinyl esters are widely known and used in thechemical process industry for their superiorcorrosion resistance, and therefore they areexcellent matrix material candidates forcomposite bipolar plates. Indeed, none of thegraphite/vinyl ester samples subjected tocorrosion testing sustained measurable weightloss or visible degradation. However, degradation

of mechanical properties during exposure is apotential problem, especially in the case of directmethanol fuel cells. While vinyl esters in generalare quite corrosion resistant, some vinyl esters areadversely affected by methanol. In preliminarytests, the tensile and flexural strengths of onematerial were reduced to about 60% of theiroriginal values after exposure to 6M methanol at80°C for 500 hours. Although the methanolconcentration in an actual fuel cell would bemuch lower, this strength reduction is still ofsignificant concern, and more extensive tests areplanned.

Besides physical corrosion or weakening ofthe bipolar plate, the leaching out of ions fromresin additives may also be problematic if theseions bind to the proton exchange membrane.Resin additives such as mold release agents oftencontain heavy atoms such as calcium,magnesium, or zinc. Based on elemental analysisof the liquid solutions and Nafion membranesused in corrosion testing, no ions were leachedfrom the resin additives. However, significantamounts of calcium and iron were leached out ofash impurities in the graphite powder, in somecases tying up as many as 20% of the active sitesin membranes immersed with composite samples.Purifying the graphite powder prior tocompounding would eliminate these impurities,but such an extra processing step would be costly,time consuming, and probably unnecessary. Theobserved effect of leachants is exaggerated, sincethe samples and membranes were immersed for aperiod of 3–6 weeks in liquids that were notcirculated or changed. Long-term fuel cell testinghas not indicated any adverse effects fromleachable ions binding to the proton exchangemembrane, probably because the continuous flowof liquid in an operating fuel cell prevents theions from reaching and/or binding to themembrane.


1. Deanna Busick and Mahlon Wilson,“Development of Composite Materials forPEFC Bipolar Plates,” 1999 SpringMeeting of the Materials Research Society,San Francisco, April 1999 (proceedingsvolume in press).

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

0

1000

2000

3000

4000

5000

6000

0

20

40

60

80

100

120

LANLunreinforced

LANLreinforced

Premixunreinforced

Premixreinforced

BMCreinforced

��Flexural Strength

�� Tensile Strength

�� Bulk Conductivity

Str

engt

h (p

si)

Conductivity (S

/cm)


32

2. Deanna N. Busick and Mahlon S. Wilson,“Low-Cost Composite Bipolar Plates forPEFC Stacks,” in Proton ConductingMembrane Fuel Cells II, Vol. 98–27, pp.435–445, The Electrochemical Society,1998.

3. Deanna N. Busick and Mahlon S. Wilson,“Low-Cost Composite Materials for PEFCBipolar Plates,” Fuel Cells Bulletin No. 5,pp. 6–8, February 1999.

G. Carbon Composite for PEM Fuel Cells

Ted Besmann, James Klett, John Henry, Jr., and Edgar Lara-curzioOak Ridge National Laboratory, P.O. Box 2008, Bldg. 4515, MS-6063Oak Ridge, TN 37830-6063(865) 574-6852; fax: (865) 574-6918; e-mail: [email protected]



Objectives

Develop a low-cost slurry molded carbon fiber material with a carbon chemical-vapor-infiltratedsealed surface for proton exchange membrane fuel cell stack bipolar plates

Collaborate with potential manufacturers with regard to testing and manufacturing of suchcomponents.


Approach

Fabricate fibrous component preforms for the bipolar plate by slurry molding techniques usingcarbon fibers of appropriate lengths and filler.

Fabricate hermetic plates using filler and a final seal with chemical-vapor-infiltrated carbon.

Develop commercial-scale components for evaluation.

Accomplishments

Fabrication of prototypical 100-cm2 active area plate (single- and two-sided)

Measured high strength (175 MPa in biaxial flexure) and observed good flexibility

Demonstrated ability to impress/emboss features

Material of very low density: 0.96 g/cm3


33

Initial production cost estimates within specification

Single-sided specimens provided to industry for evaluation

Program goal exceeded for material electronic conductivity of greater than 200 S/cm.

Program goal met of less than 1 cm3/cm2/h leakage under 2 atm. of hydrogen.

Single-cell test indicating good kinetics and very low cell resistance

Future Direction

Scale to 15 x 15-cm plates

Transfer technology to partners such as Plug Power, AlliedSignal, and others.

In FY 1999, the ORNL carbon compositebipolar plate effort has achieved severalprogrammatic goals in scaling, electricalproperties, cell performance, strength, and ofparticular importance, weight (0.96 g/cm3).Single-sided components 1.5 mm in thickness andtwo-sided plates 2.5 mm in thickness with a00-cm2 active area have been produced. Samplecomponents have been tested at Los AlamosNational Laboratory (LANL) and have beenprovided to Plug Power and AlliedSignal for theirevaluation. Previously it was demonstrated thatprojected costs would meet program goals.

Fibrous component preforms for the bipolarplate are prepared by slurry molding techniquesusing 400- m carbon fibers (Amoco DKD-xmesophase pitch fiber) in water containingphenolic resin (Figure 1). The approach is suchthat a vacuum molding process produces a low-density preform material. A phenolic binder isused to provide green strength, and also assists inproviding geometric stability after an initial cure.

Figure 1. Schematic of slurry-molding system.

The surface of the preform is sealed using achemical vapor infiltration (CVI) technique in

which carbon is deposited on the near-surfacefibers sufficiently to make the surface hermetic.This is accomplished by placing the preformsin a furnace (Figure 2) which is heated to1400–1500 C and allowing methane underreduced pressure to flow over the component.The hydrocarbon reacts and deposits carbon onthe exposed fibers of the preform, and whensufficient deposition has occurred, the surfacebecomes sealed. Thus the infiltrated carbonprovides both an impermeable surface and thenecessary electrical conductivity so that powercan be obtained from the cell. Processing timesare in the range of 5 h.

Figure 2. CVI system for carbon infiltration.

Initial specimens were provided to LANL fordetermining bulk conductivity. The values weremeasured by four-point probe to be 200–300S/cm. Surface resistivity measured at ORNL was

4 0 0 mM ille d

F ib e r

P h e n o licD u re z®

R e s in

S lu rry

V a cu u m M o ld in g S ta tio n

4 0 0 m e sh S ie ve S c re e n

V a cu u mT a n k

M o ld e d P re fo rm

REACTANTGASES IN

EXHAUSTGASES TOVACUUMSYSTEM

COMPONENTSBEINGINFILTRATED


34

12.2 4.2 /cm, compared with POCO graphitehaving 7.8 2.62 /cm.

During this period, 100-cm2 active area plateswere prepared to the specifications provided byLANL (Figure 3). The flow field was machinedfor these initial developmental components.

Figure 4 is a polished cross-section of asingle-sided plate sample showing the flow fieldchannels and the sealed surface. Also apparent isthe low-density volume (> 50 vol % of voidvolume) of the component, which consists offibers bonded by the CVI carbon.

Figure 3. A fabricated, 100 cm2 active area plateshowing the flow field.

Figure 4. Optical image of a cross-section of a single-sided plate.

A single-sided, 100-cm2 active area plate wasprepared and forwarded to LANL for evaluationin a proton exchange membrane fuel cell. Theplate tested very well, exhibiting good kineticsand exceptional low cell resistance (Figure 5).An observed drop-off in cell voltage at highcurrents was likely due to leakage from sealsaround the edge of the plate in the cell.

Figure 5. Fuel cell resistance and voltage test resultsfrom LANL using the 100-cm2 active areabipolar plate sample.

The mechanical properties of the bipolar platematerial were tested in biaxial flexure. Samples ofmaterial 3.8 cm in diameter (disks) and 1.5 mm inthickness, with one side sealed, were prepared. Abiaxial load fixture was fabricated that applies aload to a ring centered on the disk, with the edgeof the opposite side of the disk supported by asecond ring. The stress is applied so that thesealed surface is in tension. The results indicatethe material has a strength of 175 26 MPa(25.3 26 ksi), and Figure 6 is an example of thestress-displacement curve. To indicate the onsetof cracking, the fixture was fitted with an acousticdetector; and acoustic emissions did indicatecracking at relatively low loads (Figure 6).However, samples that had been subjected to 100-MPa stresses and that emitted acoustic signalsduring testing were found not to suffer through-thickness gas leakage. Thus the material can bestressed to close to failure strength without loss ofintegrity. A wetted plate was also tested underfreeze-thaw conditions and remained undamaged.


35

Figure 6. Stress-displacement curve for thecomposite bipolar plate material, alsoindicating measured acoustic emissions.

The bipolar plates tested to date have beenfabricated with machined flow fields. Inproduction, these plates would need to useembossed or otherwise impressed features, asmachining would be too costly. A preliminaryevaluation of the capability to emboss features inthe composite material prior to infiltration withcarbon was performed. A small aluminum moldwas fabricated with channels 0.78 mm (31 mil)deep and wide. The mold was used to impresschannels into the preform material, with the resultseen in Figure 7. As is apparent, the preformmaterial can take impressed features with thenecessary tolerances and dimensions.

Figure 7. Example flow field impressed on thecarbon composite bipolar plate preformmaterial.

In summary, as of the end of FY 1999, it hasbeen established that the carbon compositebipolar plate approach results in a componentwith

significantly higher strength thancompeting materials

component weight about half that of othermaterials

very high electronic conductivity very low cell resistance good cell kinetics gas impermeability low cost scalability to prototype dimensions


1. T. M. Besmann, J. W. Klett, and T. D.Burchell, “Carbon Composite for a PEM FuelCell,” in Materials for ElectrochemicalEnergy Storage, eds. D. S. Ginley, D. H.Doughty, T. Takamura, Z. Zhang, and B.Scrosati, Vol. 496, Materials ResearchSociety, Warrendale, PA.

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

60

70

80

90

100

0

0.25

0.5

0.75

1

0 50 100 150 200

am

plit

ude (

db)

disp

lace

me

nt (m

m)

Stress (MPa)

Sample: 990516B-4Biaxial Flexure

ORNL-4515-200-810.06.04.99

��

184

MP

a


36

H. Cost-Effective Metallic Bipolar Plates Through Innovative Control of SurfaceChemistry

M. P. Brady, L .D. Chitwood, J. H. SchneibelOak Ridge National Laboratory, P.O. Box 2008, MS-6115Oak Ridge, TN 37831-6115(865) 574-5153; fax: (865) 574-7659; e-mail: [email protected]

DOE Program Manager: Patrick Davis (202) 586-8061; fax: (202) 586-9811; e-mail:[email protected]

ORNL Technical Advisor: David Stinton (865) 574-4556; fax: (865) 574-6918; e-mail:[email protected]


Objective

Demonstrate “proof of principle” for producing proton exchange membrane fuel cell stack bipolarplates by forming corrosion-resistant and electrically conductive nitride scales via elevatedtemperature exposure of model alloys in a nitrogen-containing gas.


Approach

Form nitride scales on two model metallic alloys by elevated temperature exposure in a nitrogen-containing gas.

Determine if through-thickness defects are present in the nitride scales by immersion in an acidenvironment aggressive to the substrate metallic alloys but not the nitride scales.

Supply nitrided coupons to AlliedSignal (H. Dai) for potentiodynamic evaluation of corrosionresistance in a pH 5 sulfuric acid solution; corrosion current of less than 1 H 10!6 A/cm2 at 900mV vs SCE is desired.

Accomplishments

Acid exposures showed no evidence of through-thickness “pin-hole” defects typical ofconventional coating processes.

The current AlliedSignal goal of a corrosion current of less than 1 H 10!6 A/cm2 at 900 mV vsSCE was met.

Proof of principle for protection of metallic bipolar plates alloys via thermally grown scales wasestablished.


37

Future Directions

Demonstrate the formation of a corrosion-resistant and electrically conductive thermally grownscale on either a commercially available alloy or a “custom” designed alloy that can meet the costgoals set by OAAT for metallic bipolar plates in proton exchange membrane fuel cellapplications.

Demonstrate that metal bipolar plates treated in this manner will exhibit better cell performancethan graphite bipolar plates or composite bipolar plates currently under development.

For proton exchange membrane (PEM) fuelcells, thin metallic bipolar plates offer thepotential for lower weight and significantly lowercost than graphite bipolar plates. However,inadequate corrosion resistance can lead to highelectrical resistance and/or contaminate the PEM.Metal nitrides, carbides, and borides (e.g. TiN,NbC) offer electrical conductivities that are up toan order of magnitude greater than those providedby graphite and are highly corrosion-resistant.However, conventional coating methods leave“pin-hole” defects that result in accelerated localcorrosion.

Pin-hole-type defects are generally notobserved in thermally grown oxide scalesbecause both kinetic and thermodynamicconsiderations strongly favor completeconversion of surface metal to oxide. Suchdefects are also not expected for thermally grownnitride, carbide, or boride scales. Rather, the keyissues are scale cracking, adherence andmorphology (discrete internal precipitates vs.continuous external scales). These factors can becontrolled through proper selection of alloycompositions, alloy microstructure, and elevatedtemperature gas reaction conditions. The goal ofthis effort was to demonstrate proof of principlefor forming defect-free, corrosion-resistant nitridescales via gas nitridation (Figure 1).

Two alloys were selected for study: alloy A(commercially available refractory alloy) andalloy B (model alloy designed to be nitrided). Todetermine if through-thickness pin-hole defectswere present, nitrided coupons were immersedfor 24–72 hours in an acid environmentaggressive to the substrate metallic alloys but notto the nitride scales. Nitrided coupons were also

supplied to AlliedSignal (H. Dai) forpotentiodynamic evaluation of corrosionresistance in a pH 5 sulfuric acid solution. Theinitial goal set by AlliedSignal was a corrosioncurrent of less than 1 H 10-6 A/cm2 at 900 mV vsSCE.

N-containing gas

Figure 1. Schematic of gas nitridation approach.

No acid attack on the substrate was detectedin nitrided alloy A, which suggests that the nitridelayer was defect-free (Figure 2). Pin-hole-typedefects were also not detected in nitrided alloy B.However, acid attack was evident at sharp edgeregions where surface preparation was poor(Figure 3). Such edge susceptibility likely can beeliminated by rounding sharp edges prior tonitridation and/or modifying nitriding conditions.Preliminary potentiodynamic data in pH 5sulfuric acid solution indicated that nitrided alloyA and nitrided alloy B both meet the target goalof less than 1 H 10!6 A/cm2 at 900 mV vs SCE(Figures 4 a and b).

M

M-Nitride

M-Containing Bipolar Plate Alloy


38

Figure 2. SEM cross-section of nitrided Alloy Aafter 46 h exposure in a HF solution.

Figure 3. SEM cross-section micrograph of Alloy Bafter 64 h in a HNO3 solution.

The results of this preliminary study suggestthat thermally grown nitride scales caneffectively protect the underlying metallicsubstrate from corrosion (i.e., pin-hole defectstypical of coating processes are not formed).Future work will seek to demonstrate theformation of a corrosion-resistant and electricallyconductive thermally grown scale on either acommercially available alloy or a “custom”-

4 a. Nitrided Alloy A

4 b. Nitrided Alloy B

Current/Area (A/cm2)

Figures 4 a and b. Preliminary potentiodynamicdata in pH 5 sulfuric acidsolution (data of H. Dai ofAlliedSignal).

designed alloy that can meet the cost goals set bythe Office of Advanced AutomotiveTechnologies for metallic bipolar plates in PEMfuel cell applications.

40 m

Nitride

E(m

V)

Current/Area (A/cm2)

9 X 10-7 A/cm2 at900mV vs SCE

1 X 10-7 A/cm2 at900mV vs SCE

E (

mV

)

40 m

Nitride

Edge GapAttack


39

I. Low-Friction Coatings for Fuel Cell Air Compressors/Bearings

G. R. Fenske, O. Ajayi, and A. ErdemirArgonne National Laboratory, Energy Technology DivisionArgonne, IL 60439(630) 252-5190; fax: (630) 252-4798; e-mail: [email protected]


Contractor: Argonne National Laboratory, Argonne, IllinoisPrime Contract No.: W-31-109-Eng-38

Objectives

Develop and evaluate the friction and wear performance of low-friction coatings and materialsfor fuel cell air compressor/expander systems.

—Specific goals—50 to 75% reduction in friction coefficient

—One order of magnitude reduction in wear

Transfer developed technology to DOE industrial partners.

OAAT R&D Plan; Task 13; Barrier D

Approach

Identify appropriate compressor components requiring low friction and wear.

Optimize near-frictionless carbon (NFC) coatings for the identified components.

Conduct laboratory benchtop evaluation of friction and wear performance of the NFC coating.

Conduct component testing of NFC-coated compressor/expander parts.

Develop and evaluate friction and wear properties of boric acid coating and composite materials.

Accomplishment

The radial air bearings and thrust bearings of Meruit’s turbo compressor were identified ascomponents that require both low friction and low wear rate for satisfactory performance.

Thrust washer wear tests showed that NFC coating reduced the friction by about four times andwear rate by two orders of magnitude. Both met and exceeded the project goals.

A procedure was developed for laboratory fabrication of polyphenylene sulfide and B2O3 polymermatrix composite. A preliminary wear test showed a significant reduction in friction coefficientwith addition of B2O3.


40

Future Directions

Continue benchtop testing of NFC coatings and boric acid–based solid lubricant. The effect ofvarious contact parameters, such as speed and load, will be assessed.

Optimize NFC coating process for Meruit’s turbo compressor air bearings.

Conduct component rig testing of NFC-coated turbo compressor air bearings for reliability anddurability improvement.

Optimize boric acid lubricant–based polymer composites. This will include friction and weartesting of various compositions.

Conduct component rig testing of boric acid–based coating and composites.

Fuel cell technology air managementsubsystems consisting of compressors andexpanders have many critical components withtribological challenges. The seals and bearings inthese devices will require good lubrication forfriction and wear control. Recent activities atArgonne National Laboratory have lead to thediscovery of a new class of amorphous carboncoatings (near frictionless carbon or NFC) thatexhibit extremely low friction coefficients(<0.001 in dry nitrogen or dry argonenvironments). Wear rates of test samples coatedwith NFC films are also extremely low(approximately 5 to 6 orders of magnitude lowerthan those of uncoated test samples). Theseresults have been achieved under dry slidingconditions. Earlier research at Argonne has alsodemonstrated that certain boric acid–basedcompounds can be made to be extremelylubricious under the proper conditions. Underhumid conditions, such as those found in fuel cellair compressors, friction coefficients of as low as0.02 can be achieved. The main objective of thisproject is to apply these lubrication technologiesto address the tribological challenges in the fuelcell compressors and expanders.

Different components requiring lubricationhave been identified in compressor and expandersbeing developed by Vairex, A.D. Little, andMeruit. Results presented in this report are fromthe work done for the lubrication of Meruitturbocompressor radial journal and thrust airbearings. Under normal operation, a thin film ofhigh-pressure air separates the load-bearingsurfaces of an air bearing. However, fuel cell aircompressors are expected to undergo numerous

start/stop cycles during which the load-bearingsurfaces will contact each other and be subjectedto wear. To minimize wear during start/stopcycles (and during off-normal transients andbumps), the surfaces of the air bearings are oftentreated with a low-friction wear-resistant coating.Argonne’s NFC coating is being evaluated forthis application.

Benchtop thrust washer tests were conductedfor NFC-coated and uncoated 440 C stainlesssteel material. The tests duplicate the tribologicalconditions expected in a thrust bearing contactinterface. Figure 1a shows the schematic contactconfiguration of the test. Two discs were loadedagainst each other. One of the discs had a recessin the middle, as shown in Figure 1b. During thetest, one disc was stationary while the other wasrotated at a constant speed. Dimples of knowndepth were created on the disc with a recess(Figure 1b). Linear wear was measured bymonitoring the changes in the dimple height.Tests were conducted at load of 1.25 N, speed of1000 rpm, room temperature and relativehumidity of 34–45%.

NFC coating significantly improved thefriction and wear behavior of the 440 C steelmaterial. The frictional force was reduced by fourtimes, and the linear wear rate was reduced bytwo orders of magnitude, as shown in Figure 2.Oxidative wear mechanism was predominant inthe uncoated steel surfaces and was accompaniedby the generation of abrasive oxide wear debris.Polishing was the primary mode of wear in theNFC-coated surfaces. The resulting improvementin the surface roughness is expected to bebeneficial as time passes.


41

1a. 1b.

Figures 1 a and b. Thrust washer test configuration.

Figure 2. Friction and wear of uncoated and NFC-coated 440 C steel surfaces.

With the very good friction and wear resultsfrom the bench test, the plans are under way torun component tests with NFC-coated turbocompressor journal bearings. This is a highpriority because of some recent results of thecompressor bearing tests that indicated theoccurrence of localized wear. High friction atstartup of the compressor has also beenproblematic for the bearing test.


1. “Tribology of Hard Carbon Films underExtreme Sliding Conditions,” Ali Erdemir,Invited Keynote Lecture at the Cost-516Tribology Symposium, Helsinki, Finland,May 14–15, 1998, pp. 38–57.

2. “Friction and Wear Performance of Diamond-like Carbon Films Grown in Various SourceGas Plasmas,” A. Erdemir, I. B. Nilufer, O.L. Eryilmaz, M. Beschliesser, and G. R.Fenske, Presented at International Conferenceon Metallurgical Coatings and Thin Films,April 12–16, 1999, San Diego, CA.

3. “Effect of Source Gas Chemistry onTribological Performance of Diamond-likeCarbon Films”, A. Erdemir, O. L. Eryilmaz,I. B. Nilufer, and G. R. Fenske, Presented atthe 10th European Diamond Conference,Prague, Czech Republic, September 12–17,1999.

Uncoated NFC Coated0

2

4

6

8

10

12

0

50

100

150

200

250

Friction F

orce

Average Linear Wear Rate Friction Force

Line

ar W

ear

Rat

e (1

0-5 m

m/m

ina

1

2

3

4

bc

d


42

J. Inorganic Proton Exchange Membrane Electrode/Support Development

M. A. Janney and O. O. OmateteOak Ridge National Laboratory, MS 6087, P.O. Box 2008Oak Ridge, TN 37830-6087(865) 574-4281; fax: (865) 574-8271; e-mail: [email protected]



Objectives

Develop electrically conducting metallic electrodes/supports and catalytically active ceramicsandwich layers for use in ceramic electrolyte proton exchange membranes for fuel cells. Theseinorganic membranes should be less costly because of reduced Pt loading and should minimizeCO poisoning because of the higher operating temperatures.

Collaborate with Professor Marc Anderson’s project on microporous inorganic membranes at theUniversity of Wisconsin. The membrane electrodes/supports and sandwich layers developed inthis project will be evaluated in the Wisconsin project, which will provide valuable feedback forthe design and development of advanced membrane electrodes/supports and sandwich layers.


Approach

Fabricate initial electrodes/supports using sintered tape-cast nickel foils coated with a titaniasandwich layer having a pore size of about 0.1 m. A blend of submicron and nanosized titaniaparticles has been tested and determined to be appropriate for the sandwich layer.

Deposit platinum in the sandwich layer from an aqueous solution of chloroplatinic acid oraminoplatinum.

Accomplishment

Developed techniques for tape casting and sintering porous nickel foils.

Developed appropriate titania powder compositions for a sandwich layer that sinters at <400 C.

Developed a technique for coating nickel substrate with a titania sandwich layer.

Future Direction

Develop similar sandwich layer for carbon fiber paper electrodes.

Develop complete bipolar configuration for fuel cell based on the Wisconsin inorganic protonexchange membrane.


43

Microporous inorganic membranes are beingdeveloped by Professor Marc Anderson at theUniversity of Wisconsin as proton exchangemembranes (PEMs) for fuel cells. These newmembranes will operate at temperatures in excessof 100°C, will retain water at these elevatedtemperatures, and will provide protonconductivities of the same order of magnitude asthe presently employed Nafion® membranes.More important, these membranes should reducethe cost of the membrane by substantiallyreducing the amount of Pt catalyst required tooperate a fuel cell, and they should minimize COpoisoning of the Pt by operating at these elevatedtemperatures. The goal of this project is todevelop electrically conducting metallicelectrodes/supports and catalytically activeceramic sandwich layers for use in ceramicelectrolyte PEM membranes based onnanoparticles of TiO2 and Al2O3. This project wasinitiated in the last quarter of FY 1999 and isbeing conducted in coordination with Anderson’sproject on microporous inorganic membranes atthe University of Wisconsin. The membraneelectrodes/supports and sandwich layersdeveloped in this project will be evaluated in theWisconsin project, which will provide valuablefeedback for the design and development ofadvanced membrane electrodes/supports andsandwich layers.

The initial electrode/support is based on atape-cast, sintered porous nickel foil. The sinterednickel support is not a long-term solution. It will,however, allow us to provide a workableelectrode/support system to the University ofWisconsin in a timely fashion, which will allowthem to conduct electrochemical tests with theirmembranes. The sintered nickel foil has aporosity of about 55 vol % and an average poresize of about 1–2 m. A sandwich layer having asimilar porosity and a pore size of about 0.1 m isdeposited on top of the electrode foil. The small

pore size of the sandwich layer provides anappropriate surface upon which to deposit thenanoparticle membrane. A blend of submicronand nanosized titania particles has been tested anddetermined to be appropriate for the sandwichlayer. This blend produces a desirable pore size(about 0.1 m) and can be sintered to reasonablestrength at temperatures of as low as 300°C. Inaddition, the sandwich layer will be platinized.The platinum is deposited on the sandwich layerfrom an aqueous solution of chloroplatinic acid oraminoplatinum. The sandwich layer will becharacterized with respect to microstructure,including pore size and distribution, particlebonding, and Pt particle size (SEM and TEM),surface area (BET gas adsorption), surface finish(atomic force microscopy or profilometry), and Ptcontent (chemical analysis).

Once the sintered nickel foil–based substratescan be made repeatably and reliably, we will startdeveloping the sandwich layer on a better porousconducting support to replace the sintered nickelfoil support. One promising candidate is carbonfiber paper, which is currently used in theNafion-based PEM fuel cells. These papers havebeen shown to have the physical and chemicalcharacteristics required of the PEM fuel cellelectrode/support structure. We have tested thecarbon papers and shown them to be thermallystable at temperatures of up to at least 500°C inboth oxidizing and neutral atmospheres. The lowfiring temperatures (<400°C) we havedemonstrated for the blended submicron/nanoparticle titania sandwich layer compositionsallow us to contemplate using the carbon fiberpaper as the electrode/support for the inorganicPEM membranes. We can thermally process themembrane and the sandwich layer to bond andstrengthen them without fear of damaging thecarbon fiber paper. Platinizing will again be doneeither in the slurry state or after the substratelayer is bonded to the metal support.


44

4. ADVANCED COMBUSTION ENGINE AND EMISSIONS R&D

K. Microwave-Regenerated Diesel Exhaust Particulate Filter

Richard NixdorfIndustrial Ceramic Solutions, LLC ,1010 Commerce Park Drive, Suite IOak Ridge, TN 37830(865) 482-7552; fax (865) 212-6103; e-mail: [email protected]


Contractor: Industrial Ceramic Solutions, Oak Ridge, TennesseePrime Contract No: 80X-SZ896V

Objectives

Develop a ceramic filter capable of removing particulate matter (PM) from diesel engine exhauststreams at greater than 90% efficiency.

Develop a microwave cleaning system capable of regenerating the filter cartridge during dieselengine idle condition operation.

Use technology and materials that are cost-effective and scalable to large-volume commercialproduction.


Approach

Test a bench-scale microwave ceramic filter system in the laboratory during 15 cfm air flowconditions (idle condition for a 1.9-liter Volkswagen Passat diesel engine) to measure the heatingcapability of the filter during engine idle.

Test a more sophisticated microwave ceramic filter system on the exhaust of a 1.2-liter FordDIATA diesel engine to measure the PM removal filtration efficiency in the exhaust stream andthe microwave cleaning efficiency of the ceramic filter cartridge during engine operation.

Accomplishments

Designed and fabricated the bench-scale microwave ceramic filter system for measuring heatingcapability under forced air flow conditions.

Conducted a parametric study with air flow through the filter and microwave power input as theindependent variables, measuring temperature increase in the ceramic filter cartridge.

Proved in bench-scale testing that the filter cartridge could reach 600oC carbon combustiontemperatures in less than a minute under engine idle conditions.


45

Designed and fabricated a microwave ceramic filter capable of operating on the exhaust of theFord 1.2-liter DIATA diesel engine.

Conducted a debugging test of the engine filter system on a Ford Motor Company test celldynamometer.

Achieved greater than 90% elimination of diesel particulates from the exhaust stream and betterthan 90% microwave regeneration recovery (cleaning) of the filter cartridge.

Refined and improved Ford test cell system for August testing at the Ford Motor Company.

Future Direction

Optimize the materials properties of the ceramic filter media, filter insulation package,microwave power source input, and filter cartridge design for mechanical strength, filterparticulate control efficiency, and microwave energy minimization.

Produce microwave-regenerated diesel exhaust particulate filters for vehicle testing.

Enlist engineering assistance from automobile and aftertreatment systems suppliers to refine asystem, based on vehicle test results, to a viable commercial device for controlling particulateexhaust emissions.

Introduction

This was the first year of funding for thisdiesel exhaust particulate control project by theDOE OAAT Program. The primary concern of theautomotive partners in the Partnership for a NewGeneration of Vehicles (PNGV) programregarding the microwave-regenerated dieselexhaust particulate filter was the ability of theinvention to reach carbon combustiontemperatures (600oC) during diesel engineoperation. Several thermodynamic models hadshown that over 30 kW of microwave powerwould be required to heat the ceramic filter to theproper temperature for cleaning during exhaustflow. Therefore, a first step toward proving theconcept, before moving ahead with the project,required heating a ceramic filter cartridge to600oC under engine idle exhaust flow conditionsat a reasonable microwave power input. A surveyof existing small diesel engines revealed that the1.9-liter Volkswagen Passat idles at 15 cubic feetper minute (cfm) of exhaust flow. The 1.2-literFord DIATA diesel engine idles at 10 cfm exhaustflow. Industrial Ceramic Solutions (ICS) chosethe 15-cfm flow rate for the laboratory bench-scale testing.

Heating Efficiency of the Microwave-Regenerated Diesel Exhaust ParticulateFilter

Filter cartridge specifications weredetermined in meetings with PNGVrepresentatives from DaimlerChysler, Ford, andGM. A filter cartridge size of 2 in. in diameter by3 in. in length was selected according to thepartners’ specifications. The filter was fabricated,and a microwave system to accommodate theceramic filter was designed, constructed andtested. A schematic of the microwave-regeneratedfilter test system (Figure 1) and a picture of theactual experimental system (Figure 2) are shown.

The assumption by most scientists whoevaluated this technology was that all microwaveenergy converted to heat energy by the uniquesilicon carbide fibers would be immediatelytransferred to the exhaust gases. Fortunately, thesilicon carbide fibers convert the microwaveenergy to thermal energy much faster than theheat transfer to the exhaust stream occurs. Forexample, a prototype ceramic filter was installedin the system (Figure 2) with 15 cfm of airflowing through the filter. After the microwavewas turned on at a very low 0.8 kW of power, thetemperature of the filter increased rapidly tonearly 800 C in only 40 seconds. Obviously, thepower required to heat the filter was much lessthan the calculated 30 kW. A graph of theexperimental data is exhibited in Figure 3.


46

Figure 1. Demonstrated Industrial Ceramic Solutions exhaust filter can achieve regeneration temperatures atengine idle conditions.

Figure 2. Heating efficiency device test.

Figure 3. Graph of experimental data.

Site Port View of Hot Filter800 watts, 15 c40 seconds

Heat Exchanger

Ceramic Filter andMicrowave Applicator

Microwave Source

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90

18 wt. % SiC FiberTemperature Readings at 800 Watts and 15 cfm Air Flow

Temperature

²T

Heat Exchanger

Pressure

Pressure

Microwave Source

Heating Chamber

Thermocouple

OpticalPyrometer

Air Inlet

Water Inlet

HeatedFilter

FlowMeter

Waveguide

Time


47

Diesel Engine Testing on the Ford MotorCompany’s 1.2-Liter DIATA Diesel

A microwave-regenerated diesel exhaustfilter system was designed and fabricated to betested in the Ford test cell #5 DIATA enginedynamometer. This system is shown in Figure 4installed on the DIATA engine. The analyticalequipment in the Ford test cell measures thepercentage of smoke removal in Bosch smokeunits. Figure 5 illustrates that the prototype

particulate filter removed nearly 100% of theparticulates at 1800 rpm or less. The efficiencyof the filter decreased slightly at higher enginerpm.

The ceramic filter was regenerated bymicrowaves over three exhaust loading cycles.In each regeneration cycle, the filter returned toits “clean” condition for the next round offiltration. A microwave regeneration cycle isshown in Figure 6.

Figure 4. Industrial Ceramic Solutions microwave-regenerated ceramic filter installed on Ford DIATAengine test cell.

Figure 4. Industrial Ceramic Solutions microwave reg

Figure 5. Percentages of particulates removed at different speeds.

ICS Wall-Flow Filter: June 9,1999 Ford DIATA Engine Tests

0

10

20

30

40

50

60

70

80

90

100

1000 1800 2300

Engine RPM

% S

mo

ke

Rem

ova

l (B

osc

h S

mo

ke U

nit

s)


48

Figure 6. Microwave regeneration cycle in process.

Conclusions

The work has proved that concerns regardingthe ability of the ceramic filter to reach carboncombustion temperatures under diesel engine idleconditions are not a critical problem. Themicrowave ceramic filter achieved regenerationtemperatures at a fraction of the microwavepower calculated to be necessary.

The initial testing at Ford’s test cell #5 on theDIATA diesel engine was a debugging test. Anumber of issues arose related to the parties atFord and Industrial Ceramic Solutions learningeach other’s procedures and equipment. Thoseissues have been resolved. ICS recognized severalareas for potential improvements to the producttechnology during these debugging tests. Themost important of these are filter cartridge shapedesign and microwave distribution within thefilter during regeneration. These improvementsare currently being addressed in the ICSlaboratory to ensure that the next tests at Ford

will include the engineering refinements. ICSexpects significant data from future testing atFord to positively influence the future of themicrowave-regenerated diesel exhaust filter. Thecompany is currently focusing its efforts to ensurethat the improvements are ready and tested beforeproceeding with the next opportunity at Ford.


1. R. Nixdorf, “Microwave Regenerated DieselExhaust Particulate Filter,” DOE CIDI R&DReview, October 6–7, 1998, Oak Ridge,Tennessee.

2. R. Nixdorf, “Microwave Regenerated DieselExhaust Particulate Filter Status,” Partnershipfor the New Generation of Vehicles Meetingat USCAR, November 18, 1998, Detroit.

3. R. Nixdorf, “Microwave Regenerated DieselExhaust Particulate Filter Status,” DOEDiesel Cross-Cut Team Meeting at USCAR,April 13, 1999, Detroit.

4. R. Nixdorf, “Microwave Regenerated DieselExhaust Particulate Filter Status,” CIDI Mid-Year Review, March 31, 1999, Washington,D.C.

5. R. Nixdorf, “Microwave Regenerated DieselExhaust Particulate Filter Status,” MeritReview and Peer Evaluation of FY 1999DOE National Laboratory Automotive CIDIEngine Combustion and Emission ControlR&D, June 22, 1999, Argonne, Ill.

L. Rapid Surface Modification of Aluminum Engine Block Bores by a High-DensityInfrared Process

Craig A. Blue, Peter J. Blau, David C. Harper, and P. Gregory EnglemanInfrared Processing Center and Ceramic Surface Systems Group (Tribology)Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6083(865) 574-4351; fax: (865) 574-4351; e-mail: [email protected]



49


Objectives

Aluminum engine blocks for compression-ignition, direct-injection engines offer an advantageover conventional iron materials because of the significant weight savings. Unfortunately,wearing of the relatively soft aluminum cylinder walls results in increased emissions and reducedfuel economy as the engine wears. The objective of this work is to develop a new, durability-enhancing coating for aluminum engine block cylinder bores using an innovative yet inexpensive,rapid infrared surface modification process.

Treat the cylinder internal bores to enhance wear resistance and reduce friction in order toeliminate the need for heavy cast iron cylinder liners.

Collaborate with potential end users of the technology to enhance the effectiveness of the projectwhile identifying other potential applications.

OAAT R&D Plan; Task 5; Barrier C

Approach

Identify a candidate coating system capable of matching or exceeding the performance of heavycast iron cylinder liners.

Apply the candidate coating on both 4340 steel and aluminum substrate materials and optimizethe high-density infrared (HDI) fusing process using metallurgical analysis and hardness testing.

Perform appropriate wear testing to verify the performance of the coating system.

Assist the original equipment manufacturer in the development of the necessary down-holeinfrared fusing hardware.

Collaborate with the automotive industry to ensure that approaches are in line with industrialpractices.

Accomplishments

The HDI processing equipment, manipulators, and facility have been received at Oak RidgeNational Laboratory and installation has been completed.

Coating systems have been identified and applied.

Optimization of HDI fusing for 4340 steel specimens has been completed and wear testing initiated.

Initial fusing of coating systems on aluminum has been accomplished.

Discussion and collaboration with Ford Motor Company and DaimlerChrysler Corporation havebeen enhanced and other potential projects identified.

Down-hole hardware development has been initiated.


50

Future Direction

Improve coating adherence to aluminum through elemental additions to coating metal matrixsystem.

Improve coating adherence to aluminum through minor fluxing addition, similar to that used byFord with thermal spraying.

Continue wear testing of coating systems on both steel and aluminum.

Continue down-hole hardware development.

Further develop related applications with automotive and other industries.

Introduction

Cast aluminum alloys have been used toreduce the weight of internal combustion engineblocks conventionally made of cast iron. Tomaintain adequate wear and frictionalcharacteristics, cast-iron cylinder liners are ofteninstalled within aluminum blocks. Eliminating theneed for these heavy liners requires developing acost-effective surface treatment for the aluminumalloy surfaces where they mate with the slidingpiston rings. The cast iron liners currently usedcost approximately $5 per liner installed,resulting in a cost of $40 for a V-8 engine. Theirwall thickness is approximately 3 to 4 mm forcast-in and 2 to 3 mm for press fit liners.Eliminating these liners is expected to reduceengine weight by at least 8 to 12 lb., and severalmore pounds could be cut by reducing the enginesize because less space would be needed.

A high-density infrared (HDI) transient-liquidcoating (TLC) process to produce wear-resistantcoatings has been developed at the Oak RidgeNational Laboratory (ORNL). The application ofthis process on aluminum cylinder bore surfacescould result in weight reductions through theelimination of cast iron liners and a potentialreduction in overall engine size. Initialcalculations indicate that an electrical cost of 2 to3 cents per bore will be necessary to fuse thechosen coating. The coating process, performed atroom temperature, should cost in the range of $1per bore. This is an 80% reduction in cost inaddition to a weight reduction of approximately10 lb per engine.

Experiment

This project will develop and demonstratethe use of a unique, ORNL-developed HDI TLC

process to produce wear-resistant coatings onaluminum cylinder bore surfaces. The TLCprocess, which uses a high-intensity infrared heatsource to fuse wear-enhancing additives into asurface, will offer attractive manufacturing speedadvantages over spot-type methods that mustraster back and forth to get full coverage, or othermethods that require long line-of-sight paths orcomplex gas plasma-generating systems. Also,this process will treat the full bore at one time,with a processing time of 1 to 5 seconds,eliminating the reannealing that occurs with othermethods. Therefore, an entire engine block couldbe treated in less than 30 seconds with a singleapparatus. A conceptual design of the final HDIfusing apparatus is shown in Figure 1.

Figure 1. Conceptual design of the final high-densityinfrared fusing apparatus.

The experimental approach for this projectwill involve the following three stages.


51

Select and test a surface alloying system.Material compositions that are most likely toprovide wear resistance, desired frictionalbehavior, compatibility with engine lubricants,and manufacturability will be selected. Theselection will be based on the knowledge ofORNL infrared processing and tribology staff inlight of current wear-resistant coatingstechnology. Test coupons of candidate HDI-synthesized materials on 4340 steel andaluminum will be analyzed in terms ofmicrostructure and hardness. The bestcompositions and processing parameters will beselected for the next phase.

Figure 2 compares the reduced wear ratesachievable for a variety of coatings with the wearrates of aluminum and cast iron.

Figure 2. Comparison of wear rates.

Optimize HDI TLC for cylindricalsurfaces. Test coupons of candidate infrared-synthesized materials will be subjected toreciprocating wear and friction tests under hot,lubrication-starved conditions to simulate thepiston/cylinder interface. Also, ring materialsprovided by Ford Motor Company containingplasma-sprayed coatings will be incorporated inthe testing. HDI fusing of candidate coatingmaterials on aluminum substrates will continuebecause of inherent interfacial issues, which areanticipated as a result of the presence of alumina.Microstructure and hardness of the treatedsurfaces will be characterized, and the bestcandidate compositions and processingparameters will be selected for additionaloptimization on aluminum.

Demonstration using aluminum alloyengine block. ORNL will work cooperatively

with an auto company to demonstrate the coatingtechnique on curved surfaces followed by analuminum engine block. Wear tests on the treatedcylindrical surfaces will be conducted by ORNL’sTribology Group. Actual piston rings will be usedas the sliders. Tests will be run at a variety ofcontact loads and reciprocating speeds to assessthe sensitivity of coating performance metrics toimposed operating conditions. Plans are toprovide the HDI-treated engine block to anengine testing facility to verify its performanceunder conditions comparable to actual service.

Discussions will continue between materialsengineers from the automotive industry andORNL infrared processing and senior tribologystaff throughout the research to enhance theproject through industrial insight. Hardwaredevelopment for down-hole HDI TLC will beinitiated with the original equipmentmanufacturer.

Results and Discussion

Twelve tungsten-carbide particulates, 20 and60 to 70 vol %, and Cr2C3, 70 vol %, combinedwith a nickel-boron-silicon-iron-chromiumpowder, were sprayed on 4340 steel substrates.These coatings were fused using the HDI TLCprocess to initiate the wear testing. Theinstallation at ORNL used for the process isshown in Figure 3. The nickel base binder systemwith the WC reinforcement was used becausethese types of systems are being explored byother fusing techniques such as laser fusing andthermal spray. Both of these techniques haveextreme draw backs: fusing overlap and timeconsumption for laser fusing, and lack ofinterfacial bonding and porosity for thermalspray.

Through optimization, it was found thatprocessing parameters of 1000 W/cm2 at atranslation speed of 0.5 cm/s were necessary toobtain a liquid metal matrix and allow for wettingof the 4340 base material. The microstructure ofthe room-temperature-sprayed, tungsten-carbide/nickel-based, HDI-processed sample is shown inFigure 4. The coating is metallurgically bonded tothe base material and has been shown to beextremely adherent even under repeatedtemperature cycling to 660°C.

10-4

10-3

10-2

10-1

0 2 4 6 8 10 12

Abrasive Wear Data

Cast IronAl319Titanium/SiliconNickel/SiliconCopper/SiliconMolybdenum/Silicon

Vo

lum

etr

ic W

ea

r (c

m3 )

Time (minutes)


52

Figure 3. High-density infrared processing facility at the Oak Ridge National Laboratory.

Figure 4. Tungsten-carbide-reinforced coatingproduced by fusing nickel-based matrixonto 4340 steel.

Hardness testing of the HDI-fused coatingrevealed that a coating hardness of 1100 HV istypical, and thicknesses from 10 to 100 m areeasily achievable.

The same fusing process on the aluminumbase material resulted in a fused coating, butadherence with the base material was an issue.This result was expected because of the inherent

Al2O3 that forms on the aluminum. This wettingissue has been further studied. Two approacheshave been explored: (1) the addition of activeelements into the metal power matrix, which isbeing fabricated by an outside vender; and (2) theaddition of minor amounts of fluxing material.The fluxing material to be used is similar to thatpresently used by an automotive manufactureduring coating of aluminum engine bores.

Applications

The automotive industry, coating industry,and ORNL identified several other applicationsfor this process in addition to coating of cylinderbores. They include the following:

Auto companies— Ford Motor Company (proposal stage)

Robert C. McCune– Surface alloying—corrosion resistance– Wear-resistant surfaces on aluminum

alloys– Advanced joining of lightweight

materials


53

[Collaboration with Ford MotorCompany also includes a potentialcoating system for engine bores. Fordhas supplied ORNL with compressionrings with a plasma-sprayedmolybdenum (nickel-chromium)composite for incorporation into weartesting.]

—DaimlerChrysler Corporation Rosa Paulo–Ceramic coatings (boron carbide)

Industrial Companies—Dow Chemical Company

• Ceramic coatings

—Electric Boat Corporation• Cladding work• Postheat treatments

—DOE Idaho Operations Office• Yevgeny Macheret

– Powder metallurgy

—Cincinnati Milicron• Wear-resistant coatings• Down-hole hardware being developed

Conclusion

An HDI TLC process to produce wear-resistant coatings has been developed at ORNL.The application of this process on aluminumcylinder bore surfaces could result in engineweight reduction through the elimination of castiron liners and reduction of overall engine size.

Initial calculations indicate that an electrical costof 2 to 3 cents per bore will be necessary to fusethe chosen coating. Performed at roomtemperature, the coating process should cost inthe range of $1 per bore. This estimate representsan 80% reduction in cost, in addition to a weightreduction of approximately 10 lb. per engine Thisprocess will eliminate the interfacial and porosityissues involved in many metal spray techniques,while eliminating the overlap and surfaceroughness encountered with laser techniques.

The new HDI TLC facility has been installedand producing coatings for several months.Candidate coating systems have been selected forthe engine bore application through a literaturesurvey and collaboration with the automotiveindustry. Successful coating deposition has beenaccomplished on 4340 steel. Coating fusion hasbeen accomplished on aluminum but hasexperienced some interfacial bonding issues as aresult of the inherent alumina present on thealuminum substrates. Two different approacheshave been taken to enhancing the interfacialproperties between the coating and substrate.Wear testing on coatings deposited on 4340 hasbeen initiated to produce a base line for thealuminum-based material work. Also, because ofthe salient advantages of the HDI TLC process,several other applications with the automotiveindustry are at the proposal stage, and potentialapplications in other industries have beenidentified. One industrial partner is driving theneed for the down-hole hardware that is presentlybeing studied and developed with VortekIndustries, the original manufacturer of theplasma infrared equipment.

M. Optimization of NFC Coatings for Light-Duty CIDI Applications

G. R. Fenske, A. Erdemir, L. AjayiArgonne National Laboratory, Energy Technology DivisionArgonne, IL 60439(630) 252-5190; fax: (630) 2152-4798; e-mail: [email protected]

DOE Program Manager: Patrick Davis (202) 586-8061; fax: (202) 586-9811; [email protected] Technical Advisor: David Stinton (865) 574-4556; fax: (865) 574-6918; e-mail:[email protected]


54


Objectives

Develop low-friction coatings to improve performance and durability of critical enginecomponents that will run on low-sulphur diesel and dimethyl ether fuels as well as lubricating oilsand greases.

Evaluate friction and wear performance of Argonne’s near frictionless carbon (NFC) coatingsunder conditions prototypical of compression-ignition direct-injection (CIDI) engine components.

Optimize coating microstructure and chemistry to obtain maximum friction and wearperformance on light-duty CIDI applications.

OAAT R&D Plan; Task 5; Barriers A, B

Approach

Determine deposition conditions needed to achieve optimum friction and wear performance.Vary deposition parameters and determine the optimum coating conditions for components usedin fuel injection systems (such as those being developed by Lucas-Varity and Bosch). The majordeposition parameters will include gas composition, substrate bias, substrate material, and coatingthickness.

Use bench-top test machines to determine the effects of load, speed, and temperature (as wellas type of motion - e.g. unidirectional sliding, reciprocating motion, and rolling) on friction andwear performance. Characterize worn surfaces and determine failure mechanisms. Useexperimental and analytical findings to maximize the friction and wear performance of materialsoperating on low-sulfur diesel fuels.

Test coated components under conditions prototypical of advanced CIDI engines beingdesigned by Ford and Chrysler for their PNGV contracts. The major parameters will include:engine lubricant (parafinic and synthetic base stocks, and commercial products with additives,and new vs. used oil), fuels (conventional diesel, low-sulphur diesel, and DME).

Evaluate NFC coatings in actual fuel injection rigs: Fuel injection components coated withNFC coatings will be tested in component test rigs to quantify the effect of NFC coatings onmechanical energy losses and component durability. These tests will be performed oninstrumented test rigs in conjunction with Lucas-Varity and AVL.

Accomplishments

NFC coatings were successfully produced on test materials and on some actual fuel injectioncomponents. Their chemical and microstructural characterizations were completed. Uniformcoating thickness was achieved over the entire surface of the test materials and actualcomponents.

Deposition conditions were further optimized to improve coating adhesion. Different bond layersand layer thickness were tried.

Demonstrated dramatic improvements in friction and wear of NFC coated test pieces in lowsulfur diesel fuels and in a variety of oils (basestock and formulated).


55

Future Directions

Continue benchtop testing of NFC coated test pieces in clean diesel fuel and in a variety of dirtyoils recovered from engines.

Optimize deposition conditions to achieve maximum improvements in friction and wear.

Perform rig testing of NFC in actual engines and/or instrumented testers that simulate actualengine applications (such as Lucas-Varity and AVL).

Scale up and transfer optimized products to industrial companies.

Compression ignited direct-injected enginesare one of several energy conversion conceptsbeing pursued for hybrid passenger vehicles dueto their high thermal efficiency. However,emission of NOx and particualtes from CIDIengines may prohibit their use. Combustion andaftertreatment approaches are being pursued toovercome these emission barriers and the effortsdescribed below focus on the development oflow-friction, wear-resistant coatings to improvethe durability of critical engine componentsrequired for these new approaches.

In this project, we are exploring theproperties and light-duty CIDI applications of anear-frictionless carbon (NFC) film developed atArgonne National Laboratory. The project aimsto optimize the growth conditions andtribological performance of these films and totransfer the optimized process to industry for usein critical CIDI applications. The overall goal isto increase energy efficiency, improve componentreliability, and reduce toxic emissions toenvironments. Research to date has shown thatthe NFC films have the potential to overcomemany of the friction and wear problemsexperienced by various engine components (suchas fuel injectors operating in sulfur-free dieseland gasoline fuels) under sever runningconditions. The film has exceptional wearresistance and durability with a coefficient offriction 0.001 when measured in a dry nitrogenatmosphere. Deposition process is very versatileand the film can be deposited on CIDIcomponents at fairly low temperatures (roomtemperature to 200oC) without risking damage tothe base materials. Fuel lubricity tests in aspecially designed friction and wear test machineindicated that NFC coatings were compatible withethanol and sulfur-free diesel fuels and provided

dramatically improved friction and wearperformance. These coatings seem to workexceptionally well in ethanol, methanol, and low-sulfur diesel fuels and under dry or marginally-lubricated conditions. These findings have led tothe conclusion that the NFC films hold highpromise for light-duty CIDI engine applications.A series of recent tests evaluated the effects ofvarious lubricating oils on friction and wear ofNFC-coated test pieces. In these tests, we havebeen using several base stock and formulated oils.We have been also evaluating the effects ofvarious oil additives on friction and wear. Testsunder lubricated reciprocating contact conditions.at room temperature have indicated that the NFCfilms are very compatible with base mineral oilsand have the capacity to lower friction by 40%.Figure 1 shows the lubricated friction coefficientsof uncoated and NFC coated 52100 steel samples.Steel against steel under lubricated condition canprovide a friction coefficient of about 0.13 atroom temperature, but with the NFC coating, thefriction coefficient is reduced to 0.07 level. Wearon the uncoated surfaces are quite severe, but onthe coated surfaces, it is minimal. These initialresults indicated that the NFC films are quiteeffective in reducing both friction and wear underlubricated test conditions.

The results presented in Figure 1 are quiteremarkable, in that the NFC coating seems tohave a synergistic effect on the lubricity of baseoil. As is known, on lubricated surfaces friction isvery much determined by the viscosity and thenature of the boundary films formed on slidingsurfaces. It looks that the NFC coating is quitecompatible with base stock oil. Results of othertests run with different oils and oil additives aresummarized in Table 1. Tests with synthetic anddirty oils are planned and will be performed over


56

Figure 1. Effect of mineral oil on friction coefficients of uncoated and NFC-coated steel surfaces.

a wide temperature range (i.e., room temperatureto 160oC).

Experimental studies with diesel fuels havealso continued and concentrated on theeffectiveness of the NFC films on reducing wear.For this purpose, we utilized two test machines;one reciprocating ball on flat (very popular inEurope) and a ball on three flats (favored in theUnited States). Initially, we ran tests with

conventional (high-sulfur diesel containing about500 ppm sulfur) and low sulfur (containing 140ppm sulfur) diesel fuels. The results obtained sofar are very encouraging. As shown in Figure 2,the NFC film effectively reduced the wear of the52100 steel flats in low- and high-sulfur dieselfuels. Figure 3 shows the size-of-wear scarsformed on uncoated and NFC-coated flats duringtests in a low-sulfur diesel fuel.

Figure 2. Effect of NFC coating on the wear of 52100 steel disks tested in low- and high-sulfur diesel fuels.

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0 40000 80000 120000 160000 200000 240000 280000

Time [s]

Fri

cti

on

Co

eff

icie

nt

uncoated

NFC coated

0

1

2

3

4

5

uncoated NFC-coated

High Sulfur (500 ppm)

Low Sulfur (140 ppm)


57

Figure 3. Effects of sulfur content of diesel fuel onwear scar diameters of uncoated and NRC-coated 52100 flats.

As is clear from Figures 2 and 3, the NFCcoating is very effective in reducing wear,especially in a low-sulfur diesel fuel environment.A synthetic diesel fuel containing essentially nosulfur is on order and will also be used in our testprogram. Lubricated tests will also continue andconcentrate on the effects of temperature anddirtiness of the oils (i.e., heavily sooted oilsdrained from actual engines). As can be seenfrom Table 1, presence of certain additives makessignificant difference in the friction and wearperformance of uncoated steel samples in poly-alpha-olefin (PAO) oil. However, when used onNFC coated samples, the effect of additives isdiminished. In fact, the base stock PAO oil seemsto work rather well, thus raising the prospect forthe elimination of some of the unwanted sulfur,chlorine, or phosphorous bearing additives fromformulate oils.

In the near future, we will approach end-usercompanies to jointly determine the optimum filmproperties that will required for their specificapplications. We will deposit, characterize, andqualify coatings on their specific test pieces forfurther tests in their labs. If the results lookpromising, we will help scale-up the process forpilot size production and use in actual engines.

Table 1. Friction and wear performance of uncoated and NFC coated steel samples in various oils.

Test conditions: 1 km, 20 cm/s, uncoated parts, room temp and humidity unless otherwise noted

Test Fluid Average Wear (mm) Average Friction Coefficient

Test 1 Test 2 Test 3 Test 1 Test 2 Test 3

Dry 0.8351 0.543

PAO 0.715 0.793 0.121 0.114

1% TCP 0.355 0.335 0.293 0.099 0.105 0.113

1% ZDDP 0.778 0.745 0.122 0.121

1% Moly 1.048 1.005 0.132 0.131

1% Moly & 1% TCP 0.793 0.665 0.126 0.121

1% Moly & 1% ZDDP 0.273 0.285 0.107 0.070

1% TCP & 1% ZDDP 0.648 0.4782 0.713 0.115 0.130 0.129

1% TMS 0.880 0.855 0.117 0.1121 Test run for only 0.1 km2 More oil applied

Uncoated Disk(Mag: 76X)

NFC-Coated Disk(Mag: 76X)


58

Test conditions: 1 km, 20 cm/s, NFC-coated parts, room temp and humidity unless otherwise noted

Test Fluid Average Wear (mm) Average Friction Coefficient

Test 1 Test 2 Test 3 Test 1 Test 2 Test 3

Dry 0.2753 0.218 0.097 0.105

PAO 0.158 0.158 0.084 0.070

1% TCP 0.155 0.160 0.067 0.076

1% ZDDP 0.2933 0.2833 0.085 0.100

1% Moly 0.2333 0.2253 0.079 0.077

1% Moly & 1% TCP 0.185 0.175 0.063 0.071

1% Moly & 1% ZDDP 0.168 0.178 0.070 0.071

1% TCP & 1% ZDDP 0.2733 0.3083 0.077 0.067

1% TMS 0.1904 0.145 0.087 0.0793 Film appears to have worn out4 Film thickness slightly less than for test 2

Publications

1. “Tribology of Hard Carbon Films underExtreme Sliding Conditions,” Ali Erdemir,Invited Keynote Lecture at the Cost-516Tribology Symposium, Helsinki, Finland,May 14–15, 1998, pp. 38–57.

2. “Friction and Wear Performance of Diamond-like Carbon Films Grown in Various SourceGas Plasmas,” A. Erdemir, I. B. Nilufer, O.L. Eryilmaz, M. Beschliesser, and G. R.

Fenske, Presented at International Conferenceon Metallurgical Coatings and Thin Films,April 12–16, 1999, San Diego, CA.

3. “Effect of Source Gas Chemistry onTribological Performance of Diamond-likeCarbon Films,”A. Erdemir, O. L. Eryilmaz,I. B. Nilufer, and G. R. Fenske, Presented atthe 10th European Diamond Conference,Prague, Czech Republic, September 12–17,1999.

N. Material Support for Non-Thermal Plasma Development

S. D. NunnOak Ridge National Laboratory, MS-6087, P.O. Box 2008Oak Ridge, TN 37831(865) 576-1668; fax: (865) 574-8271; e-mail: [email protected]


Contractor: Oak Ridge National Laboratory, Oak Ridge, TennesseePrime Contract No. DE-AC05-96OR22464


59

Objective

Provide ceramic material support to Pacific Northwest National Laboratory (PNNL) fordevelopment and fabrication of new proprietary ceramic component designs for use in non-thermal plasma reactors for the treatment of diesel exhaust gases.

Fabricate and ship components to PNNL for testing and evaluation in prototype non-thermalplasma reactors.

OAAT R&D Plan; Task 1B; Barrier A, B

Approach

Utilize the gelcasting forming method to fabricate complex-shaped ceramic components that meetPNNL design specifications.

Modify processing as needed to accommodate material and design changes.

Evaluate metallization materials and processes to apply electrodes to the ceramic components.

Accomplishments

Fabricated and shipped ceramic components having two design variations to PNNL for testingand evaluation.

Identified a metallizing material for use in forming electrodes on the ceramic components.

Future Direction

Fabricate metallized ceramic components for testing at PNNL.

Modify processing as necessary to meet material and design modifications.

Evaluate commercially viable processes for component fabrication.

Complex-shaped ceramic components for anew design of non-thermal plasma (NTP) reactorswere successfully fabricated and sent to PacificNorthwest National Laboratory (PNNL) fortesting and evaluation. NTP reactors have showngreat potential as an effective means ofeliminating unwanted exhaust gas emissions fromdiesel engines. Researchers at PNNL aredeveloping new design configurations for NTPreactors that build on past experimental work. Toimprove effectiveness, these designs includeceramic components having complexconfigurations. Oak Ridge National Laboratory(ORNL) has extensive experience in thefabrication of complex ceramic shapes, primarilybased on prior work related to developingceramic components for gas turbine engines.ORNL’s gelcasting process for forming ceramic

shapes was used in the fabrication of componentsfor the new NTP reactors.

A well-defined processing procedure wasdeveloped for gelcasting the ceramic shapes.Careful control of the ceramic slurry preparation,the casting mold preparation and assembly, andthe casting and gelling procedure were required toproduce satisfactory green ceramic parts. Thesubsequent drying and sintering steps were alsocritical in producing parts that were free ofdefects such as cracking and distortion. Thecasting molds were designed to produce anoversize part to account for the shrinkage of theceramic component during densification. Theoptimum solids content of the ceramic slurry, theorganic monomers for gel formation, and thepolymerization initiator were determinedexperimentally to produce high-quality gelcastcomponents. High-purity ceramic raw materials


60

were used to maintain high dielectric strength anduniform dielectric properties in these components.

Three metal alloy systems were evaluated forforming the electrodes on the ceramiccomponents. These included a magnesium-basedalloy, a powdered titanium-based paste, and amoly-manganese silk screening ink. Themagnesium and titanium materials were selectedfor their lower processing temperatures.Experiments on sample parts, however, indicatedthat the fired-on moly-manganese ink providedthe best metallization on the ceramic componentmaterial. Tests were conducted under varyingheat treatment conditions. The effects of bothfiring temperature and furnace atmosphere wereevaluated.

Components will continue to be produced andprovided to PNNL for testing. As componenttesting at PNNL proceeds, any necessary materialand design changes will be incorporated in futureceramic component fabrication at ORNL. Nowthat an appropriate material has been identified, awell-defined process for applying the metalelectrodes to the ceramic components will bedeveloped and electroded components will beprovided to PNNL. Alternative processingapproaches, which may be more amenable tocommercial production of the ceramiccomponents, will be assessed.

O. Nanofluids for Thermal Management Applications

S. ChoiArgonne National Laboratory, ET 335, 9700 South Cass AvenueArgonne, IL 60439(630) 252-6439; fax (630) 252-5568; e-mail: [email protected]



Objective

To develop nanofluid technology for increasing the thermal transport of engine coolants andlubricants. The research in FY 1999 focuses on demonstration of the heat transfer potential ofnanofluids, for application in thermal management systems for advanced vehicles.

OAAT R&D Plan; Task 5; Barrier C

Approach

Modify an existing small-channel heat transfer test apparatus. Perform bench-scale testing usingvarious nanofluids and several different particle loadings. Measure and report heat transfercoefficients and pressure drop.


61

Accomplishments

An existing hot-wire cell requires about 500 ml of nanofluids to measure the thermal conductivityof nanofluids. In efforts to reduce the volume of testing samples, a small hot-wire cell thatrequires only 60 ml of nanofluids is being designed and fabricated. In addition, a new heattransfer test section is being designed and fabricated.

Future Direction

Perform heat transfer experiments using various nanofluids and several different particleloadings.

Develop smaller and lighter nanofluid/carbon foam radiators.

Conventional automotive heat transfer fluids,such as lubricants and engine coolants, areinherently poor heat transfer fluids. There is astrong need to develop advanced heat transferfluids with significantly higher thermalconductivities and improved heat transfer.Combining nanophase technology with heattransfer technology provides a new class of heattransfer fluids, called nanofluids, that areengineered by dispersing nanometer-size solidparticles in traditional heat transfer fluids toincrease thermal conductivity and heat transferperformance. The potential benefits of theapplication of nanofluid technology includeautomotive heat exchangers that are smaller,lighter in weight, and more efficient thanconventional units and have a reduced inventoryof fluid.

The objective of this activity is to developnanofluid technology for increasing the thermaltransport of engine coolants and lubricants. Thisis a new activity started in April 1999. Previousresearch with nanofluids has demonstrated thatnanofluids are stable and that thermalconductivity can be substantially enhanced overthe base fluid. For example, improvements inthermal conductivity of up to 40% of the basefluid have been demonstrated for a relatively lowmetallic nanoparticle loading (<0.3 vol %). Theincrease in thermal conductivity will result in anincrease in heat transfer over that of the base fluidwithout dispersed nanoparticles. In preliminaryresearch, Argonne National Laboratory (ANL)has shown that the heat transfer capability ofwater increased by 20% with the dispersion ofless than 1 vol % copper oxide nanoparticles.Even greater improvements in heat transfer are

expected for nanofluids that contain metals (suchas Cu and Ag) rather than oxides. In FY 1999, theresearch focuses on demonstration of the heattransfer potential of nanofluids for application inthermal management systems for advancedvehicles.

In May, Xinwei Wang, a Ph.D. candidatefrom Purdue University, started working at ANL.During the reporting period, Wang focused hisefforts on further development and evaluation ofthe transient hot wire technique for measuringthermal conductivity of nanofluids. An existinghot-wire cell requires about 500 mL of nanofluidsto measure the thermal conductivity ofnanofluids. In efforts to reduce the volume oftesting samples, a small hot-wire cell that requiresonly 60 mL of nanofluids is being designed andfabricated. In addition, a new heat transfer testsection is being designed and fabricated.

On May 5, Marty Wambsganss and SteveChoi visited DaimlerChrysler in MadisonHeights, Michigan. Choi made a presentationtitled “Nanofluids for Thermal Management.”Based on our discussions with experts in fuel celltechnology at ANL and at the auto makers, itappears that the near-term application ofnanofluid technology for cooling fuel cell systemswould be in the secondary coolant loop. That loopuses a conventional coolant and includes theradiator and a liquid-to-liquid heat exchanger.The application of nanofluid technology toimprove the heat transfer characteristics of thecoolant will lead to smaller heat exchangers and areduced fluid inventory. In the long term, it maybe possible to identify nanoparticle materials andnanofluid production methods that would allow


62

nanofluids to directly cool the fuel cell stackwithout adversely affecting operation of the stack.

The heat rejection requirements of a fuel cell–powered vehicle can be significantly greater thanthose of a conventional vehicle because nearly100% of the waste heat goes to the coolingsystem, compared with only 30% forconventional vehicles. In addition, the fuel cellsmust operate at temperatures of 80 C or less,resulting in a smaller driving temperaturedifference for heat transfer in the radiator. Theseconditions require much larger radiators and leadto the conclusion that a conventional coolingsystem will not work. As a result, thermalmanagement can be considered an enablingtechnology for fuel cell–powered vehicles.

The heat transfer coefficient on the air side isinherently low; therefore, a large air side surfacearea is required. The heat transfer characteristicsof the coolant are also relatively poor; forexample, the thermal conductivity of awater/ethylene-glycol mixture is about one-halfthat of water. To reduce the size and weight of theradiator, there is a need to improve heat transferon the air side. Once that is accomplished, theheat transfer characteristics of the coolant need tobe improved. Carbon foam materials andnanofluids are advanced heat transfer materialswith significantly higher thermal conductivitiesand better heat transfer characteristics than arepresently available. Therefore, combining carbon

foam technology with nanofluid technology couldlead to a breakthrough in advanced vehiclethermal management system design that can meetthe strong need to improve heat transfer on boththe air side and coolant side of the radiator.

Oak Ridge National Laboratory and ANLplan to develop carbon foam radiators that usenanofluid coolants. The purposes of the plannedjoint research are to develop combined nanofluidand carbon-foam technology for application inadvanced vehicle thermal management systems,and to demonstrate that the combined carbonfoam and nanofluid technology can significantlyreduce the size and weight of radiator systems forfuel cell–powered vehicles, heavy vehicles,hybrid-electric vehicles and other performance-driven systems.


1. Lee, S., U. S. Choi., S. Li, S., and J. A.Eastman, J. A., “Measuring ThermalConductivity of Fluids Containing OxideNanoparticles,” J. Heat Transfer, 121, 280–289, 1999.

2. Wang, X., X. Xu, and U. S. Choi, “ThermalConductivity of Nanofluids,” accepted forpublication in the Journal of Thermophysicsand Heat Transfer, 1999


63

5. CERAMICS FOR GAS TURBINES

P. Agreement to Bring into Production and Commercialize a Manufacturing Process forSilicon Nitride Turbomachinery Components

Danielle NewsonAlliedSignal Ceramic Components2525 West 190th StreetTorrance, CA 90504-6099(310) 512-5705; fax: (310) 512-5901; e-mail: [email protected]


Contractor: AlliedSignal Ceramic Components, Torrance, CaliforniaPrime Contract No: N0014-95-2-0006

Objective

Develop, demonstrate, and verify the ceramic technology necessary to produce large quantities ofhigh-quality silicon nitride components for automotive gas turbine engines at costs acceptable tothe automobile industry.

OAAT R&D Plan; Task 1; Barrier A

Approach

Provide improvements in yield, quality, and cost of Si3N4 components.

Develop and implement automated forming, statistical process control, and intelligent processingcontrol.

Demonstrate a gelcasting manufacturing process capable of supplying engine-quality hardware intypical production volumes.

Accomplishments

Designed, fabricated, and installed automated equipment capable of gelcasting 500 to 1000complex-shaped silicon nitride turbine wheels per month.

Developed an improved AS800 silicon nitride gelcasting system that demonstrates moreconsistent gelation that results in higher yields, reduced cycle times, improved part quality, andreduced costs.

The gas turbine engine was eliminated as acandidate for the 80-mpg Partnership for a NewGeneration of Vehicles automobile in FY 1998.

Work on this contract continued through FY1999, using funds from previous fiscal years.

Structural ceramics, which traditionally havebeen based upon non-oxides such as silicon


64

carbide or silicon nitride, have long beenconsidered primary candidates for hot-sectioncomponents in advanced gas turbines. Initialproperty limitations such as low strength, lowWeibull modulus, and poor creep resistancewere successfully addressed in a number ofmaterials development programs. However, thecommercialization of silicon nitride has beenlimited because of the high cost of thesecomponents.

Technical Progress

A local supplier that specializes in systemsdesign, integration, and fabrication of automatedequipment was identified to design and buildautomated equipment for the large-scalemanufacturing of complex-shaped Si3N4

components. A fully automated system wasdesigned with only minimal operator interactionrequired. The system currently requires twooperators; the goal is to reduce the number ofoperators to one. A closed-loop conveyor systemwas designed for the movement of gelcast moldsthrough the process stations. The GelFast TM

system was installed at AlliedSignal CeramicComponents in FY 1998 and debugged andintegrated into a full operating system during FY1999. Figure 1 shows the system prior toinstallation of heat station #2.

The production demonstration using theautomated Gelcasting equipment was delayedfor 6 months because of problems associatedwith the completion of the automated system bythe equipment manufacturer. The demonstrationwas completed during FY 1999, and a steadystate cycle time of 15 minutes per wheel wasachieved. The automated gelcasting of aTeledyne M304 turbine wheel is shown inFigure 2.

The cost study for a dedicated facility tofabricate large quantities of turbine wheels hasbeen completed. The study was based on anannual production of 50,000 turbine wheels foran industrial turbogenerator. The predicted dropin cost per wheel was exponential up to 10,000wheels, whereupon the cost began to dropslowly with each additional 5,000 wheelsproduced per year. The predicted unit cost forthe ceramic wheel was 2–3 times the cost of the

Figure 1. Automated gelcasting system, GelFast TM.

Mold Fill

Heater #1

MoldWash/Dry

Mold ReleaseUp Elevator

System MainConveyor

MoldAssembly(Hidden)

MoldDisassembly

DownElevator

SecondaryCooling

Heater #2(not shown)

ReturnConveyor


65

Figure 2. Automated gelcasting of a Teledyne M304 turbine wheel.

metallic wheel. However, the life of the ceramicwheel is predicted to exceed the life of the engine,while the metallic wheel would have no usefullife under these operating conditions.

A number of process refinements to theoriginal gelcasting process were investigated inFY 1999. An optimized AS800 silicon nitridegelcasting system was developed anddemonstrated. The system demonstrated muchmore consistent gelation than the previoussystem; and a wide variety of parts including testplates, Teledyne M304 turbine wheels, andturbine nozzles were successfully fabricated.Optimization activities have focused on partquality and yield, cost reduction, and cycle timereduction.

A binder burnout cycle for gelcast M304turbine wheels was also developed for the newsystem. Wheels were reproducibly densifiedusing this cycle (no M304 wheels made using theprevious gel system achieved density using theold pyrolysis cycle), and mechanical propertieswere verified. The pre-sintering cycle was

eliminated, contributing to reduction of cost andcycle time. The use of optical pyrometers foroptimized temperature monitoring in the furnaceis being employed as a cost reduction measure.

The high cost of gelcasting molds has been adetriment to making production of complex-shapegelcast parts feasible. Advanced mold technologywas pursued this year. Alternate mold materialswere identified and shown to be compatible withthe optimized gelcast system. The alternativetechniques have reduced the lead time requiredfor mold fabrication and have therefore reducedcosts.

The industrial turbogenerator wheel wassuccessfully redesigned during the year with asimplified single-pull vector used for the airfoilslides. The original design had a complex three-dimensional twist to the airfoil, which requiredmultiple pull vectors to be engaged duringfabrication of the part. The original design wascost-prohibitive and high-risk for productionoperations. The revised design will saveapproximately 50% in tooling costs.


66

APPENDIX A:ABBREVIATIONS, ACRONYMS, AND INITIALISMS

ABI automated ball indentationAlN aluminum nitrideAl2O3 alumina or aluminum oxideASTM American Society for Testing and MaterialsBMC Bulk Molding Compounds, Inc.BME base metalCIDI compression-ignition, direct-injectionCTP Ceramic Technology ProjectCVI chemical vapor infiltrationdc direct currentDME dimethyl etherDOE U.S. Department of EnergyEC electronic ceramicECD electronic ceramic deviceGM General MotorsHDI high-density infraredICS Industrial Ceramic SolutionsLANL Los Alamos National LaboratoryMLC multilayer capacitorMPM mechanical properties microprobeNFC near-frictionless carbonNOx oxides of nitrogenNTP non-thermal plasmaOAAT Office of Advanced Automotive TechnologiesORNL Oak Ridge National LaboratoryOTT Office of Transportation TechnologyPAO poly-alpha-olefinPEBB power electronics building blockPEM proton exchange membranePM particulate matterPNGV Partnership for a New Generation of VehiclesPNNL Pacific Northwest National Laboratorypps polyphenylene sulfideR&D research and developmentSEM scanning electron microscopeSNL Sandia National LaboratoriesTLC transient-liquid coatingulf ultra low fire

progress report for propulsion materials materials is enabling the advanced integrated power module...

Documents