DOE/OR-2003

CONTINUOUS FIBER CERAMI(2 C 9MPOSITE

PROGRAM

Technical Progress Report for Phase I

FEBRUARY 1994

INDUSTRY ENERGY EFFICIENCY DIVISIONTHE OFFICE OF INDUSTRIAL TECHNOLOGIES

DEPARTMENT OF ENERGY

OI$'_UfiON OFTHISOOCIIMENTill UNLIMITED

This report has been reproduced directly from the best available copy,

Availableto DOE and DOE contractors from the Office of Scientificand Teohnt-

; cal Information,P,O, Box 62, Oak Ridge, TN 37831; prices available from (615)676-8401, FTS 626-8401,

Available to the public from the National Technical Information Service, U.S,Department of Commerce, 5285 Port Royal Rd,, Springfield,VA 22161,

i, ,,i i ., , , i ,i , i ,,,|,,= , i , i | ,

This report was prepared as an account of work sponsored by an agency of .the United States Government, Neither the United States Government nor any

agency thereof, nor any of their employees, makes any warranty, express or *implied, or assumes any legal liability or responsibility for the accuracy, com-

pleteness,or usefulnessofany information,apparatus,product,or processdis- t

closed,or representsthatitsuse would not infringeprivatelyowned rights,

Referencehereinto any specificcommercialproduct,process,or serviceby

tradename,trademark,manufacturer,orotherwise,does notnecessarilyconsti-

tuteorimplyitsendorsement,recommendation,orfavoringby theUnitedStates

Government or any agency thereof.The views and opinionsof authorsexpressedhereindo not necessarilystateorreflectthoseoftheUnitedStates

Governmentorany agencythereof.

I, . •.... *'" °

TECHNICAL PROGRESS DURING PHASE I OF THECONTINUOUS FIBER CERAMIC COMPOSITES

. PROGRAM

• Prepared for:

U.S. Department of EnergyEnergy Efficiency and Renewable Energy

Office of Industrial Technologies

March 15, 1994

Prepared by:

David W. RichersonRicherson and Associates

• Salt Lake City, Utah

gASTE,D_RDUTION OFTIfI_ DOC'UMENTIB UNLIMITF]

9

• _.. _., ..... .q¢

EXECUTIVE SUMMARY

. United States industry has a critical need for materials that are lightweight, strong, tough, corrosionresistant and capable of performing at high temperatures. The availability of such materials willenable substantial increase in energy efficiency and reduction in emissions of pollutants.

• Continuous fiber ceramic composites (CFCCs) are an emerging class of materials which have thepotential for the desired combination of properties to meet the industrial needs. A $10 billion annualmarket has been estimated for CFCC products by the year 2010, which equates to over 100,000industrial sector jobs. 1

The Department of Energy's Office of Industrial Technologies completed in 1990 a comprehensiveprogram plan for the development of CFCC materials. 2 The resulting CFCC program began in thespring of 1992 as a three-phase 10-year effort to (1) assess potential applications of CFCCmaterials, (2) develop the necessary supporting technologies to design, analyze and test CFCCmaterials, (3) conduct materials and process development guided by the applications assessmentinput, (4) fabricate test samples and representative components to evaluate CFCC materialcapabilities under application conditions and (5) analyze scaleability and manufacturability plusdemonstrate pilot-scale production engineering. The program is a model partnership betweenI

industry, government, academia and national laboratories to provide energy savings and emissionreductions for industry and a contribution to the economic prosperity of the nation.

• DOE awarded 10 Phase I cooperative agreements to industry-lead teams plus identified genericsupporting technology projects to benefit all of these contracts. This document highlights the broad

, progress and accomplishments on these contracts and support technology projects during Phase I:

* Effective teams have been established to identify applications with high payoff, to conduct designstudies to define material property and configuration goals, to conduct the necessary CFCCmaterials development, to establish a property database, to evaluate representative components in arealistic use environment, and to later scale up to commercialization. Each team includes a CFCCfabricator and one or more end user, plus several support team members such as a fiber supplier, afiber coater, a property test laboratory and a composite design specialist.

* Each team has identified one or more industrial applications that could benefit from use of CFCCcomponents and has conducted sufficient design study to define the temperature, stress andconfigurational requirements for the CFCC material. These requirements have guided CFCCdevelopment during Phase I and have provided a basis for assessing feasibility to proceed to PhaseII. The applications and benefits are described in the document "Continuous Fiber CeramicComposites (CFCCs) for Low Cost Energy and a Cleaner Environment." 1

- * CFCC contractors have achieved substantial technical progress during Phase ! as summarized inTable 1. Key challenges that have been focused on are achieving high temperature stability in anoxidizing or corrosive environment, retaining high strain-to-failure composite fracture behavior at

" high temperature, increasing the matrix cracking yield stress and establishing low cost fabricationprocesses. The matrix cracking yield stress is the stress at which cracks form in the matrix. PriorCFCC materials have not survived for long time when stressed at high temperature above the matrixcracking yield stress.

1

TABLE 1 SUMMARY OF PHASE 1 ACCOMPLISHMENTS

Team Status of Technology Status of Technology ,Before CFCC Program After Phase I

AlliedSignal * Concept for diesel * OEM Caterpillar identified twoapplication concepts and l 1 options '

• No existing CFCC or prior * OEM conducted finite elementdevelopment analysis

• Viable design concept defined• Critical material property

requirements defined• 60 compositions fabricated and

tested

• CFCC successfully developed....... • CFCC judged feasible by OEM

Alzeta * Concept for burner application * High payoff applications defined• No existing CFCC or prior * Two radiant burner concepts

development effort designed, fabricated and tested• Radiant output doubled compared

to prior burner technology• 4500 hours, 12,000 cycles

demonstrated with no material or

performance degradation• Increase in temperature capability ,

demonstrated

• Commercialization in three yearsAmercom * Existing baseline material * High-payoff applications

• Inadequate for oxidation, identifiedcorrosion conditions of high * Improved corrosion resistanttemperature applications CFCC developed

• Pitting and >0.1 mil/h recession * No visible degradation afterat 950°C in sodium sulfate 950°C in sodium sulfate saturatedsaturated air air

• Expensive non-reusable tooling * Process demonstrated with >50%• Inadequate controlled porosity reduction in tooling cost

for hot gas filtration application * Controlled porosity CFCCdemonstrated

TABLE 1 SUMMARY OF PHASE 1 ACCOMPLISHMENTS (cont'd)......

, Babcock and * Existing baseline material * Improved properties, shapesWilcox * 24 ksi tensile strength * Doubled strength to 44 ksi

• Low strain-to-failure (0.09%) * Improved strain-to-failure (.26%), * Fabrication required over 30 * Fabrication achieved with <15

infiltration cycles infiltration cycles; major cost• Batch fabrication reduction

• No effective interface coating * Semi-automated fabricationdemonstrated; more costreduction

• Interface coating demonstrated• Property database established• Modeling capabilities established• Applications identified, assessed• Representative components

designed, fabricated andsubmitted to OEMs for testing

Dow chemical * No existing CFCC; concept * Matrix slurries successfullyCompany only formulated and coated onto fibers

• No CFCC fabrication * Flat plate CFCC samples• technology successfully fabricated with

toughness eight times higher thanmonolithic matrix material

• Major progress in tubefabrication; limit in initial trials to26" radius of curvature; achievedI" by end of Phase I

• Application identified and rig builtto assess material stability underapplication conditions

Dow Corning * Existing baseline CFCC * Improved CFCC (40 ksi strength)• Not stable in oxidizing with no strength decrease to

atmosphere above about 1200°CIO00°C * 12" by 12" plates demonstrated

• 6" by 6" plates demonstrated * Decreased cost interface layer• Expensive CVD interface layer from liquid precursor

• 75% strength retention afterthermal shock from 1000°C towater

• Equivalent strength with 50% lessinfiltration cycles

• * High payoff applicationsidentified

TABLE 1 SUMMARY OF PHASE 1 ACCOMPLISHMENTS (cont'd)

f

DuPont Lanxide * Existing baseline CFCC * Six options successfullyComposites CVI * High temperature durability fabricated

d

limited; failed in less than 100 * One improved CFCC did not failhours at 1200°C at 12 ksi after 4700 hours at 1200°C at 12

• Composition options defined, ksibut no database * No CFCC strength loss after 500

• No testing in application hours exposure to simulatedenvironments steam reformer gas stream at

1040°C• Survived 1500 hours at 1000°C

in flowing coal gas with alkalisulfate additions

DuPont Lanxide * Existing baseline CFCC * Four modified CFCCs evaluated

Compos_es * Limited database * Baseline still bestDIMOX * Unsafe interface coating process * Safe interface coating process

developed; viable for productionscale

• Improved uniformity of interfacecoating

• Strength constant to 1200°C andonly decreases to 33 ksi at1370°C

• Strength decrease of only 16% forO .

thermal shock from 1200 C intowater

• Baseline material survived 1000

hours of stress rupture exposureat 11.6 ksi and 1200°C in air....

TABLE 1 SUMMARY OF PHASE 1 ACCOMPLISHMENTS (cont'd)

GE Corporate * Existing baseline material * Tensile strength of unidirectionalR&D * Small coupons fabricated in architecture doubled to 100 ksi

laboratory by hand layup * Tensile strength ot' cross ply, * Baseline fiber/matrix interface architecture nearly doubled to 50

coating not oxidation resistant ksi• Matrix cracking yield stress

increased over 20%• Oxidation resistant interface

coating developed with cross, plystrength of 51 ksi and 1204°dstrength of 30 ksi

• No property degradation after 100hours in air at 1200°C

• Semi-automated preform processdemonstrated for substantial cost

reduction versus pre-CFCC, * Continuous fiber coating process

demonstrated for further cost

reduction and improvedreproducibility

• Application feasibility projected, by OEM assessment

Textron * No existing CFCC; concept * All steps in fabrication processSpecialty only successfully developedMaterials * CFCC with high strength and

toughness demonstrated• Tensile strength of 60-65 ksi

achieved

• Drum wrapping and filamentwinding, both production-viableprocesses, demonstrated

• Tube 24" long by 4" diametersuccessfully fabricated; suitablefor initial testing by OEM

• Joining of CFCC to monolithicmaterial demonstrated

• Surface densification processdemonstrated to seal surface

porosity and improve oxidationresistance by factor of I0

' * Team of CFCC fabricator, OEMand end user formed to

aggressively pursue testing in.............. application environment

Technical Progress Highlights

, AlliedSignal Ceramic Components and Caterpillar Inc. teamed in Phase I to explore thefeasibility of achieving a self-lubricating CFCC material for exhaust valve guides in a low heatrejection diesel engine. Caterpillar conducted 2D finite element analysis to explore three design

' options and to identify material requirements and performance targets. One design appears feasible.AlliedSignal Ceramic Components evaluated four matrix compositions of silicon nitride and 10grades of carbon fibers. The optimum fiber type, fiber loading and silicon nitride matrix wereexperimentally determined. A test sample was fabricated and shown to provide solid lubricationrunning against A.I.S.I. 4140 mild steel at 300°C, thus demonstrating technical feasibility of thematerial for the valve guide application.

The Alzeta program has demonstrated potential for near-term commercial production. Theprogram is directed towards development of a new generation of radiant burners with greaterdurability, longer life and higher temperature capability. 10% market penetration of these burnerswould save 50 billion cubic feet of natural gas and reduce nitrogen oxide emissions by 35,000 tonsper year. A burner incorporating a CFCC outer shell operated with minimal material degradationfor 3000 hours at 250 ° F higher temperature than a conventional burner. A CVD SiC infiltratedporous mat was tested for 1500 hours with no degradation. A concept to utilize a CFCC"reverberatory screen" was explored and demonstrated to increase radiant output up to 100%.

Amercom teamed with OEMs from four industry sectors and identified applications ranging fromhot gas candle filters to industrial furnace components to turbine and diesel components that would

, have high payoff from the use of CFCC materials. Amercom also conducted materials developmentfocused on improvement in high temperature stability, on cost reduction and on achieving acontrolled pore structure for hot gas filters. A graded matrix CFCC was developed with a siliconcarbide interior and a mullite oxide surface that exhibits improvement in corrosion resistance. Aprocess modification was demonstrated that eliminates costly hot tooling and is projected to reducefabrication cost by 50%. A filter concept was demonstrated that meets flow and pressure droprequirements of the candle filter application.

Babcock and Wilcox achieved broad progress during Phase I. They increased the tensilestrength oftheir baseline oxide-oxide CFCC material from 24 ksi to 44 ksi and the strain-to-failurefrom 0.09 to 0.26. They improved their matrix precursor materials to eliminate 50% of theinfiltration cycles, thus substantially reducing fabrication cost. They fabricated representativecomponents for several applications and worked with end-users to begin exploratory testing ofthese components. They substantially increased the property database and began developing designmethodology and life prediction codes.

Dow Chemical began Phase I with a concept and a few exploratory trials rather than an existingCFCC material. Their concept was to incorporate SiC fibers into their self-reinforced silicon nitride(SRSN) material. They had to start from zero in establishing key composite technologies such aspreform fabrication, interface layer composition and deposition, and composite densification. Theyhave made good progress in preform fabrication, but are still early in the tasks to demonstrate aneffective interface layer and densification process. Key preform fabrication accomplishmentsinclude development of successful matrix slurries, demonstration of techniques and apparatus to

infiltrate these slurries into candidate fiber architectures, and fabrication of an infiltrated tubular

preform two inches in diameter and I1 inches long. Dow has demonstrated some progress withinterface layers. One resulted in composite strength greater than the matrix material strength andwith a factor of 8 increase in work of fracture ( a measure of toughness).

Dow Corning demonstrated improved high temperature stability of their polymer infiltrationpyrolysis (PIP) CFCC material. Specifically, their improved material retained a tensile strength ofover 40 ksi from room temperature to at least 1200°C and fractured in a non-catastrophic mode withhigh strain-to-failure (0.7%). This material survived 500 hours stress rupture testing in air at1100°C at a stress of 20 ksi. This is particularly encouraging because the test stress was above thematrix cracking yield stress (10 ksi) for the material. Prior CFCC materials exhibited severeoxidation and embrittlement degradation when tested in air above the matrix cracking yield stress.Dow Corning also achieved improvement in fabrication capability. They scaled up to panels 12inches by 12 inches. They also demonstrated modifications in their fabrication process that have thepotential to significantly reduce cost.

DuPont Lanxide Composites (DLC) accomplished improvements to processing of theirbaseline 2D Nicalon TM SiC-based fiber/alumina matrix CFCC. The primary improvement was toestablish a production-viable process for CVD of their fiber interface coating without a reduction inproperties. DLC also fabricated four alternate CFCC materials to explore the potential of SiCparticulate filler, alumina fibers, higher temperature SiC-based fibers and chopped fibers. Usefulinformation was gained from these studies, but none of the materials exhibited propertiescomparable to the baseline material. DLC teamed with Foster Wheeler Development Corporation,GE Industrial and Power Generation Division, Solar Turbines Inc. and Westinghouse Electric toevaluate a broad range of applications. Heat exchanger tubes and gas turbine components wereidentified as viable applications.

A second program at DuPont Lanxide Composites focused on SiC-based fibers in a SiCmatrix produced by chemical vapor infiltration (CVI). Potential applications were assessed withFoster Wheeler Development Corporation and GE Industrial and Power Generation Division.Oxidation and corrosion resistance were identified as key properties for most of the applications.Six variants of CVI SiC-SiC composites were fabricated and evaluated under coal flue gas, steamreformer and stressed oxidation conditions. Variants included two fibers, a standard plus anenhanced oxidation resistant matrix and two external oxidation protection coatings. Substantialimprovements in high temperature stability were demonstrated for a couple of the variants comparedto baseline material. For example, the enhanced matrix material with standard fiber and no surfacecoating did not fail after >4500 hours exposure in air at 1200°C under a tensile stress of 12 ksi.

General Electric Corporate Research and Development achieved excellent success duringPhase I. Tensile strength was nearly doubled to 100 ksi for uniaxial and to 50 ksi for cross-plyfiber architectures. Matrix cracking yield point was increased to over 25 ksi for uniaxial and over20 ksi for cross ply. An oxidation resistant interface coating was developed which resulted inincreased material stability at high temperature. Low cost fabrication procedures were demonstratedwith an increase in reproducibility and strength. Application studies were completed, and theCFCC material properties were shown to provide a reasonable margin over design stresses.

Textron Specialty Materials achieved remarkable progress during Phase I. They started with aconcept and progressed all the way to establishing high strength and component fabrication

• capability. The concept consisted of incorporating silicon carbide or carbon fibers in a nitride-bonded silicon carbide matrix. They successfully developed the procedures to fabricate a CFCCmaterial and demonstrated a strength of over 60 ksi and strain-to-failure of over 0.6%. They

• identified viable radiant and immersion heater tube applications for the aluminum remelt industry,formed an industry team to design and test a specific component and fabricated a full sizecomponent suitable for testing.

Eighteen Supporting Technology projects were established in Phase I to address key issues incomposite design, material characterization, test methods and performance-related phenomena. Theobjective of these projects was to provide broad expertise from academia, government laboratoriesand technology companies to help understand and solve problems basic or generic to all theindustrial CFCC contracts. Examples include fiber-matrix-h_terface interactions, failuremechanisms and allowable operation limits, design database generation and use, time-dependentbehavior (creep, stress rupture, fatigue) and composite modeling. Excellent progress was achievedin establishing the supporting technology projects and in integrating them with the industrialprograms through informal linkages and partnerships, joint meetings and technical highlightsreports.

The remainder of this document highlights the progress and accomplishments of each contract.Phase I Material and Scope identifies the nature of the CFCC material selected fordevelopment, the state of the technology at the start of the CFCC program and the key goals ofPhase I for that specific contract. The baseline technology varied broadly for the different contracts.In some cases the CFCC already existed and the program focused on applications assessment,fabrication and extensi_n of database• In other cases the program began with a concept and focusedon material development to demonstrate feasibility for a selected application or applications.Partnerships Established identifies the teams that were formed during Phase I. PropertyImprovements highlights the improvements in strength and high temperature stability achieved inPhase I. Fabrication Improvement/Cost Reduction describes progress towards cost-effective fabrication of the types of shapes required for the selected application(s).Applications/CFCC Viability Assessment discusses the final status at the end of Phase I ofthe material capabilities versus the projected application design requirements.

The document concludes with a brief description of the Supporting Technologies scope andprojects and a review of some of the Phase I highlights.

The AlliedSignal Ceramic Components Team

• Phase I MateriM and Scope clearly viable both from material and engineperspectives.

The AlliedSignal Ceramic Components (CC)' program began with a concept rather than a Fabrication/Property Evaluation

pre-existing CFCC material. Tile conceptwas to develop a self-lubricating CFCC * Four silicon nitride matrix materials werematerial for exhaust valve guides and other evaluated: reaction bonded, sintered reactionsimilar components in a low heat rejection bonded, and hot isostatically pressed GS-44diesel engine. The proposed CFCC material and GN-10 compositions. The best matrix-was carbon fibers in a silicon nitride matrix, only friction results were obtained with GS-The silicon nitride would provide strength and 44.corrosion resistance and the carbon would

provide self-lubrication. * 16 carbon fiber grades were evaluated.These were first screened for stability in

The objectives of Phase I were (1) to conduct flowing air at 600°C. Four grades werefabrication screening trials to determine if a selected for composite development trials.carbon fiber/silicon nitride composite could befabricated with favorable properties, (2) to * Composite fabrication was explored withdefine valve guide design concepts and both continuous and chopped fibers. Testevaluate their feasibility by finite element samples were successfully fabricated withanalysis and (3) to conduct friction and wear chopped fibers but not with continuous

. measurements of CFCC development fibers.iterations to estimate whether a coefficient of

friction of 0.2 and an average wear coefficientof lxl0 -8 mm/N-m without externallubrication were feasible.

Partnerships Established

A partnership was established during Phase I

between CC and Caterpillar. CC conducted _:_:the CFCC material development and someproperty screening tests. Caterpillar _._.:conducted design studies, benefits/economicsanalysis, friction and wear tests and material ,,exposure tests in simulated diesel exhaust. . .,,_--r'---

• Design Studies Typical Microstructure of Dense SiliconNitride Reinforced with Chopped Carbon

Caterpillar identified two exhaust valve Fibers• concepts andeleven installation options.

These were narrowed to three design options * Over 60 CFCC compositions werefor 2-D finite element analysis. Analysis fabricated and tested for friction and weardetermined that one of these designs was using a pin-on-plate type of test. This test is

11

considered more severe than the application,but suitable for material screening. Twogrades of carbon resulted in coefficient offriction of 0.2 or lower. However, thecoefficient cf friction increased with time due

to wear debris from the CFCC imbedding inthe A.I.S.I. 4140 mild steel.

* CFCC variations were exposed to asimulated diesel exhaust atmosphere at 650°Cfor 250 hours. All exhibited reduction in

strength. The typical sample of the preferredcomposition decreased from 55.8 ksi to 43.5ksi. Strength decrease was not judged to be aconcern because tne ceramic is in

compression in the application. Further studyis needed to determine effects on friction andwear.

Applications/CFCC ViabilityAssessment

Caterpillar studies determined that CFCCmaterials would provide a benefit for diesel 0

exhaust valve guides and that allowablecomponent cost is in the range of feasibilityfor candidate CFCC materials. They alsoidentified a viable component design. CFCCmaterials were developed in Phase I that wereclose to the unlubricated frictional

requirements. Longer term testing isnecessary under conditions closer to theapplication requirements to demonstratematerial feasibility.

12

The Alzeta Team

. Phase I Material and Scope can be used as a reverberatory screensupported above the burner surface as

The Alzeta program differed from the other illustrated:programs. Alzeta is a manufacturer ratherthan a CFCC material developer. Alzetamanufactures a ceramic fiber radiant burner

for residential, commercial and industrial nEvERSE.ATORVSURFACE

applications.The Alzeta programfocusedonassessment of a broad range of radiant burner PoRousapplications, design of burners with CFCC BURNERRADIANT

components, and laboratory testing of CFCC _ ...:, ...materials and first generation burner : _.J ',.: ?:::

components• ,::" -• ' :.'j';,',.,', _,._

Partnerships EstablishedWALL

The Alzeta team includes Southern California

Gas Company (SCG) as a funding partnerand a number of CFCC manufacturers as

material suppliers. The role of SCG in the .ADIANTANt,program is to promote the development of SENSIBLEHEAT

low-NOx high performance radiant burnersthat will benefit end users across a broad

spectrum of applications. Material suppliers winclude DuPont Lanxide Composites as a CFCC Radiant Burner Design Conceptsupplier of porous mats and ceramiccomposite screens, and 3M Corporation, The purpose of the reverberatory screen is toAmercom, Dow Coming and Textron as increase radiant output. If strong and tough,suppliers of ceramic composite screens and the screen could also protect other burnertensile samples, parts.

Design Concepts Identified Concepts Testing

Two design concepts were identified. One * Burners have been constructed using highconsists of refractory ceramic fibers bonded porosity mats of a SiC-based fiber bondedtogether by a low volume fraction of with a CVD SiC matrix, and several iterationschemically vapor deposited silicon carbide, on this basic design have been tested in aThis is intended for use as a more durable thermal cycling test fixture. The bestradiant surface that will perform in a manner performing mat to date was supplied bythat is similar to current generation burners. DuPont Lanxide Composites and has

• The second concept consists of a strong experienced minimal performance degradationlathwork or screen of CFCC that can be used after 4500 hours of testing and 12,000as a protective sheath over a conventional thermal cycles.ceramic fiber burner. This screen alternatively

13

* Reverberatory screens and screen tensile 50samples were fabricated by 3M, Amercom, _-Dew Coming, DuPont Lanxide Composites o• 40 - ,, "., FleverberatoryScreenBurnersand Textron Specialty Materials. Tests of _- _" "

"-

burners with reverberatory screens have z 30 • "O

shown that radiant output from the burner r-.O

surface can be nearly doubled when compared '_ 20tl

to conventional radiant burners. Life testing _- " " "7, urface Combustors

of screen materials was also performed on 1- _< 10D tensile samples to aid in the material < All data taken at nominal 315 kW/sq meter surtace tiring rate

selection process. Tensile strength of 0 ..........60 80 100 120 140 160 180

samples was measured before and after 1200hours of exposure to a combustion THEORETICALAIR (percent)

environment to quantify loss of strength and Comparison o[Radiant Output from CFCCtoughness over time. The best performing Reverberatory Screen Burners Versussample group shoed a 13 percent decrease in Conventional Surface Combustors atstrength as a result of combustion Constant Surface Firing Rateenvironment exposure, and this level of

degradation has been determined to be Applications/CFCC Viabilityacceptable. A sse ss me n t

* A radiant burner was constructed with a Applications assessment indicated that CFCCdurable CFCC outer shell over a conventional materials could provide a high payoff inceramic fiber burner surface. The shell was radiant surface burners by improving3M Siconex TM material which consists of 3M efficiency and reducing NO x emissions. ANextel TM fibers and a SiC matrix. The outer 10% market penetration was predicted to saveshell increased the durability of the ceramic 50 billion cubic feet of natural gas and reducefiber surface and was shown to slow material NOx emissions by 35,000 tons per year.aging in a high temperature environment. The Successful Phase I testing suggests that flatburner was operated for 3000 hours at a plate radiant surface burners could be a near-burner surface temperature of 1120°C term commercialization success tbr CFCC

radiating to a 800°C load. materials. Field tests are projected in one totwo years and commercialization for initialapplications in three years.

14

The Amercom, Inc. Team

. Phase I Material and Scope available.

Amercom had a relatively mature family of Property ImprovementsCFCC materials prior to the CFCC program.These materials consisted of aluminum Three tasks were directed towards

borosilicate fibers or SiC fibers in a matrix improvement in properties in a highproduced in situ by chemical vapor temperature oxidizing atmosphere:infiltration/deposition (CVI). These baseline (1) evaluate the stability of alternate fibers tomaterials had been successfully demonstrated the Amercom CFCC fabrication conditionsfor thousands of hours in low stress and to oxidation, (2) develop a non-oxidizingapplications, but were not adequate for the interface to replace the pyrolytic carbon fiber-high payoff industrial applications targeted by matrix interface in the baseline material andthe CFCC program. The objectives of the (3) develop a graded matrix with SiC in theAmercom Phase I program were (1) to further interior and mullite (an oxide) at the surface.study potential high payoff applications andbetter define the material requirements, (2) to * Nine alternate fibers were studied. All weremodify their SiC-SiC material and fiber- determined to be compatible with the CVImatrix interface to increase high temperature process, but none provided strengthstability, (3) to develop a controlled porosity improvement over the baseline material.material for candle filters and (4) to exploreoptions for fabrication cost reduction, * A graded matrix was successfully

. especially by reducing the cost of tooling, demonstrated. This material had ultimatetensile strength of 27.6 ksi and strain-to-

Partnerships Established failure of 0.64% compared to 29.8 ksi and0.56% for a SiC-SiC control. Corrosion

Amercom established partnerships with ABB- testing was very promising. 20 hoursCombustion Engineering, Surface exposure to sodium sulfate saturated air atCombustion Inc., Solar Turbines and Detroit 950°C resulted in no visible attack to the

Diesel to explore options for CFCC graded matrix CFCC. Standard SiC-SiCapplications, to select viable applications for CFCC exhibited >0.1 mil per hour surfacedetailed study and to conduct benefits recession and severe pitting.assessments. MSNW was added to the team

to conduct micromechanical analysis to * Many variants were evaluated in an effortestimate the viability of composite to achieve a non-oxidizing interface betweenarchitectures under estimated application the fibers and matrix. All resulted instresses. University of Virginia participated in decreased strength and strain-to-failure.electron microscopy and auger analysis of the However, improved understanding was

' fiber-matrix interface and in fracture analysis gained of the interface and the factorsin an effort to understand (and find solutions influencing strength and toughness. Someto) high temperature degradation mechanisms, promising options were identified for study inMaterials Sciences Corp. also participated in Phase II.Phase I. They used micromechanicscomputer models to predict a full set of designproperties from the limited property data

15

Fabrication Improvement/CostReduction

* Major effort was focused on reduction intooling cost. The baseline CVI processrequires a rigid graphite tool to support eachCFCC fiber preform in the CVI furnace.These support tools are expensive, consume

valuable space in the fumace and can only be ,_used once. A technique was developed inPhase I to rigidize the fiber preform at a lowtemperature using metal tooling andsubsequently position the preform in the CVIfurnace without requiring support tooling. Acost savings of 50% is projected. Strength of Test Sample to Evaluate the Porous Candlethe resulting product is lower than for Filter Conceptbaseline material, but still above 20 ksi.

* Amercom conducted development on threeconcepts for achieving controlled porosity in acandle filter tube. Tubular test samples werefabricated and evaluated for face velocity andpressure drop at Industrial Filter and PumpMfg. Co. One concept performed close to the )

application requirements and was selected forrefinement in Phase II. This concept consistsof a filament wound structure supporting afilter membrane. The filter membrane contains

straight smooth-surfaced flow channels with atailorable diameter (4-10 microns) and flowpath length (1-3 mm).

Cross Section Showing the ControlledPorosity Flow Channels

Applications/CFCC ViabilityAssessment

Many applications were evaluated. Thefollowing were selected as having high payoffand being the most viable for near-termdevelopment: candle filter tubes, hot furnacefans, radiant burner tubes, diesel piston ringsand a heat exchanger manifold.

16

The Babcock and Wilcox Corporation Team

Phase I Material and Scope under the CFCC program.

Babcock and Wilcox (B&W) demonstrated a Property Improvements. baseline CFCC material prior to the CFCC

program. They proposed to continue Baseline B&W alumina-alumina CFCC withdevelopment of their CFCCs using aluminum 23.3% porosity had a tensile strength of 24oxide fibers in matrices of aluminum oxide, ksi and strain-to-failure of 0.09%. Efforts to

zirconium oxide and yttrium aluminum garnet increase strength and composite fracture(YAG). An all-oxide CFCC was selected to behavior in Phase I focused on developmentprovide high stability and corrosion resistance of an interface layer between the fibers andin the high temperature oxidizing matrix to minimize bonding of the matrix toenvironments of the intended applications, the fiber. A tin oxide fiber coating wasThe general fabrication process selected by developed that resulted in strength increase toB&W involves a series of liquid infiltrations 44 ksi and strain-to-failure improvement tointo a fiber preform, alternated with a series 0.26% for an alumina-alumina CFCC withof drying cycles and heat treatments to 23.2% porosity.convert the liquid precursors to the desiredoxide. The objectives during the Phase Iprogram were (1) to establish an effectiveteam, (2) to improve properties, (3) to reducecost by decreasing infiltration cycles and _oo00developing filament winding, (4) to generate a Unidirectional Tensile Test

property database suitable for design ,0000she2FiberCoating"

modeling, (5) to evaluate and select viableapplications and fabricate/test initial _ '°°°°

representative components. .,.., 30.000tO

Partnerships Established 20000

B&W established during Phase I an ,000o

interdisciplinary team consisting of end users ,, , _(B&W industrial divisions and General 0 ,ol 02 03 04

Electric Company), a fiber coater and matrix Strain(%)

precursor developer (Basic Industrial hnprovements in Strength and CompositeResearch Laboratory at Northwestern Fracture BehaviorUniversity), composite characterizationlaboratories that are also participating in Test bars of the improved CFCC materialdeveloping design methodology and life were fabricated and submitted to the testprediction codes (Cleveland State University laboratory team members to generate aand VPI and State University) and a mechanical property database. Poisson'sbraider/weaver of fiber preforms (Fiber ratio, shear strength, shear modulus, andMaterials, Inc.). B&W also established an tensile and compressive strength in both theagreement with Solar Turbines to test under a longitudinal and transverse directions wereseparate program combustor liners fabricated measured for flat plates and tubes. Thermal

17

conductivity was also measured. These data coatings and (3) stability in the candidatepermitted for the first time for the B&W team matrix materials.to analytically predict whether the CFCCmaterial would survive in candidate * The filament winding process was refinedapplications, to achieve larger tubes and to reduce

interlaminar debond defects. NDE techniquesFabrication Improvement/Cost were identified at B&W and in collaborationReduction with Argonne National Laboratory to detect

fabrication defects and were used to helpMajor cost reduction and improvement in the guide fabrication development.quality of fabricated parts were achieved inPhase I.

* New sols for alumina, zirconia and YAGwith increased solids yield were developedand refined to allow successful infiltration

into fiber preforms. This resulted in areduction of infiltration cycles from over 30for the baseline CFCC to less than 15 for the

improved CFCC, equating to a labor costreduction of over 50%.

7o ............................ -;,-._bccc,.,..............18 I1_ POqOtlly 13S% I_O_OSlly ~_

\ CFCC Heat Exchanger Tube Preform Being

.=_,o Fabricated on B& W Filament Winder40

* Representative components were fabricated

._ I . _ for a gas turbine combustor, a heat exchanger

20/,0] ____ _...o_ boiler.tUbe'a soot blower tube and tube shields for a0 l i l__ ......... _.........

0 ,0 _o _o ,0 Applications/CFCC ViabilityNumberofInfiltrations Assessment

Reduction in Infiltration Cycles Needed to

Achieve Acceptable Density A number of applications, especially for

* A technique was developed to deposit the boiler components, were identified where anoxide-oxide CFCC would decreasetin oxide fiber coating from a liquid precursor

as an in-line step in the filament winding maintenance, conserve energy and reduceprocess. This makes possible a rapid emissions. The Phase I database andautomated preform fabrication process which modeling capabilities were used to assesswill further reduce cost. feasibility of the material for specific

applications. Representative components* Four commercially available alumina fiber were fabricated and delivered to the end users

materials were evaluated for (1) ability to be for a first evaluation under the applicationpreformed by filament winding, braiding and conditions.weaving, (2) compatibility with interface

18

The Dow Chemical Company Team

. Phase I Material and Scope encouraging. The material fractured in a non-catastrophic mode as shown 3elow. The

Dow Chemical began Phase I with a concept work of fracture was eight times higher thanand a few exploratory laboratory trials rather for the monolithic matrix material. Highthan an existing CFCC material. Their temperature testing has not yet beenconcept involved incorporation of SiC fibers conducted.into their self-reinforced silicon nitride

(SRSN) material. The advantage of theSRSN material is its high strength andtoughness and corrosion resistance. The highstrength and toughness were projected to leadto a CFCC with unusually high interlaminar 3.0- ]P_shear and interlaminar tensile strengths. 2.s-

2.0-Nicnlon TM Fiber Reinforced

The objectives of Phase I included _1.5- _.._._o s,_(1) develop low cost processing routes based 3 _.0-on tape casting of SRSN, (2) develop 0.5-formulations, procedures and equipment to 0.o- I I i I

infiltrate slurries into fiber tows, (3) develop o.0 0._ 0.2 0.3 0.4lamination procedures to form infiltrated fiber Displacement(ram)

into shaped preforms, (4) demonstrate the Fracture Behavior at Room Temperature of ancapability to densify a CFCC with reasonable Early SRSN matrixSiC Fiber CFCCstrength and toughness, (5) select and screeninterface coatings to maximize toughness,(6) evaluate the potential of an application of * Testing the stability of the material underThe Dow Chemical Company and, the application conditions is critical. A testing(7) conduct material stability tests in a facility to evaluate potential CFCCs undersimulated application environment, simulated operating conditions has been

constructed under Dow Chemical funding.Partnerships Established

Fabrication Improvement/CostDow Chemical corporate research teamed Reductionwith an operating divisior_ of Dow Chemicalto select and evaluate a chemical processing * Matrix slurries were developed thatapplication. Dow obtained fibers from 3M, successfully infiltrated fiber yarns.Dow Coming and Textron Specialty Materialsand interface coatings from 3M and General * Equipment and procedures were developedAtomics. to apply the slurry to aligned fiber using a

tape caster to form a flexible tape. This hasProperty Improvements reduced layup time by a factor of three and

has increased reproducibility.Property testing is at an early stage.

• Techniques were developed to laminate cutResults with one CFCC material have been strips of the tape into flat plates and into

19

tubes. Initial tapes were nonflexible and isostatic pressing (FLIP)are required tocould not be bent to a diameter smaller than achieve densification of the present52 inches. The application required four composition. Over 97% of theoretical density 'inches. Modifications in the slurry binder and has been achieved by HIP. However, it isplasticizer increased the flexibility to ',allow encouraging to note that some cost reductionsbending to a 16 inch diameter. A damp in the densification process have already beenforming method was developed to incorporated into the process, and furthersuccessfully yield a two inch diameter tube. improvements are expected once coated fibers

become available for use. Additional work is

also ongoing to examine other densificationtechniques.

Applications/CFCC ViabilityAssessment

A proprietary Dow Chemical chemic',dprocessing application was identified that

i

would provide economic benefits andpossibly increase U.S. internationalcompetitiveness.

Two Inch Diameter 11 Inch Long TubeFabricated By Damp Drum Wrapping

* A fiber/matrix interlayer is needed toprevent fiber interaction with the matrixduring fabrication and to allow slip betweenthe fiber and matrix during mechanicalloading. This portion of the program wasdelayed awaiting receipt at Dow of acceptablecoated fibers.

* Densifying fiber/matrix preforms has beena major challenge. Initially a temperature of1825°C was required. This resulted in severedegradation of fibers. Matrix compositionmodifications have reduced the densification

temperature to 1600°C. Even at thistemperature, glass encapsulation and hot

20

The Dow Coming Corporation Team

Phase I Material and Scope MSC conducted micromechanical modelingand data verification. Argonne National

Dow Coming developed a baseline CFCC Laboratory performed NDE, mechanical• material prior to the CFCC program using a testing and microstructural evaluations. NDE

polymer infiltration and pyrolysis (PIP) techniques were demonstrated for verifyingtechnique analogous to that used for carbon- fiber architecture, locating delamination flawscarbon composites. The polymer is infiltrated and identifying the residual stress in theinto the fiber preform, cured at low CFCC.temperature and then pyrolyzed above1000°C to yield porous ceramic. The Property Improvementsprocedure is repeated until most of theporosity has been filled with ceramic matrix. * Tensile strengths over 40 ksi at roomThe Phase I objectives included (1) to identify temperature and 1200°C have beenindustrial applications with high payoff for demonstrated. The composite fractures in ause of a CFCC material, (2) to compare the non-catastrophic mode with strain-to-failureproperties of the baseline PIP material and of up to 0.7%.projected improvements to the applicationneeds, (3) to improve the properties of thePIP CFCC, especially the stability at hightemperature, (4) explore processimprovements that will decrease cost andincrease reproducibility, (5) demonstratescaleup in size of 2D CFCCs and (6) improvethe understanding of composite failure _o

SIC Filler

mechanisms. -_"_._o=5o3OO2s./. _...:::%-S:::: ...."'""""Partnerships Established _ 200

I ....... IDow Coming assembled a team that included ,oo _-'_;'_- i-i 90_*e___2o,-c IIKaiser Aerotech, Solar Turbines, Sundstrand ,_o/IFluid Handling, MSNW, Techniweave, MSC °o.0' o.1 0.2 o._ o., o.._ 0.8 °.7 o.Band Argonne National Laboratory. Dow st_°.,,cz_Coming and Kaiser Aerotech carried out Composite Fracture Behavior of PIP 2Dmaterial and process development and CFCCestablJ:;hed the Kaiser Ceramic Compositesjoint venture to manufacture and marketCFCC components. Solar Turbines and * PIP 2D CFCC retained 75% of its strengthSundstrand conducted application after thermal shock from lO00°C into water,assessments. MSNW identified a suitable compared to retention of only 20% strengthinterface coating and developed a perceramic for monolithic silicon carbide and siliconpolymer coating process. Techniweave nitride.designed 3D architectures for improvedinterlaminar properties and fabricated 3Dpreforms for processing and evaluation.

21

'_': ' ;i _ _ was demonstrated, at Kaiser Aerotech in aproduction environment, compared to the

,0. + prior fabrication of 6" by 6" panels at Dowi A + Corning in a laboratory environment.

i +- o ._,,N,] * Studies during Phase I indicate that the]., ,. ,, s,_-j number of infiltration cycles can be reduced

+ CFCC. .... ' from 12 to 8 without reduction in CFCCII

strength by control of polymer, filler and fiber20-

"_ ° o coating characteristics. This reduces0 .......... fabrication cost.

Wirer Quench AT (°C_

Applications/CFCC ViabilityThermal Shock Behavior of PIP CFCC Assessment

Fabrication Improvement/Cost The team identified a gas turbine combustorReduction liner, a containment shell for a sealless

chemical pump and bladed and unbladed• A liquid precursor technique was turbine rotor disk as high payoff applicationsdeveloped to coat fiber tows and woven fabric for the PIP CFCC material. Critical stresses

with the improved interface coating. Cost is for the combustor liner were predicted to beprojected to be substantially lower than for the 7.2 ksi hoop and 0.8 ksi interlaminar shear.CVD coating process used previously. The PIP 2D CFCC has interlaminar shear

Flexural strength of the resulting CFCC is strength of 3-3.5 ksi and matrix crackingover 50 ksi. yield stress of about 10 ksi and thus is a

viable candidate ['orthe combustor liner.

• 3D fiber architectures were explored to Stresses for the turbine disk are higher, butincrease interlaminar properties. 3D fiber the use temperature is lower. The usepreforms are more rigid than 2D preforms and temperature for chemical pumps is alsothus more difficult to infiltrate, especially with relatively low, but the fluids can be verypolymers containing high level of fillers, corrosive. Further assessment is underway toDow Coming succeeded in infiltrating 3D determine viability of the PIP CFCC for thepreforms with resins containing up to 45% turbine disk and the chemical pumpSiC filler particles, which is comparable to the containment shell.prior 2D capabilities.

• PIP technology was transferred to KaiserAerotech. Fabrication of 12" by 12" panels

22

The DuPont Lanxide Composites CVI Team

• Phase I Material and Scope Property Improvements

DuPont Lanxide Composites (DLC) Six material variants were evaluated: twodeveloped commercial scale capabilities prior fibers, baseline versus enhanced oxidationto the CFCC program for fabrication of resistance matrix and two external oxidationcomposites by chemical vapor infiltration protection coatings. Pre-test measurements(CVI) of SiC matrix into SiC-based fiber included room temperature and highpreforms. They determined that their baseline temperature tensile strength, thermalmaterial had limitations in oxidation and diffusivity, specific heat, coefficient ofcorrosion resistance and thus developed thermal expansion and room temperatureseveral modified materials projected to have compressive strength, shear strength andimproved stability. The objectives of the toughness.Phase I CFCC program were (1) to studypotential applications and identify test * Enhanced oxidation resistance matrixconditions to validate CFCC materials for CFCC did not fail during 4700 hoursselected applications, (2) to fabricate test exposure to 12 ksi tensile stress at 1200°C.coupons of the baseline material and modified Baseline material fractured in less than 100materials, (3) to measure key mechanical and hours at the same stress and temperature.thermal properties needed for componentdesign analysis and (4) to compare the * Enhanced matrix CFCC prestressed to 20modified materials with baseline material after ksi (above the matrix cracking yield stress)exposure to simulated coal gas, steam did not fail after > 1000 hours at 12 ksi andreformer and stressed oxidation conditions. 1200°C.

Partnerships Established * All six CVI CFCC materials survived 500hours under simulated steam reformer

DLC teamed with Foster Wheeler conditions with no visual degradation orDevelopment Corporation (FWDC). Foster change in tensile strength. ExposureWheeler is a leading producer of technology consisted of 1040°C in a flowing CH 4, H20,for coal-fired utilities and for steam CO gas stream. 1000 hour and 1500 hourgeneration. The team compared the CVI tests are completed and propertyCFCC fabrication capability and properties measurements are in progress.with industry needs and selected viableapplications. FWDC defined and conducted * Testing in a simulated coal flue gaskey tests to experimentally determine the environment was conducted for 500, 1000stability of candidate materials in simulated and 1500 hours at temperatures up toapplication environments. DLC also provided 1260°C. The sample test surfaces wereproperty database to the GE CFCC program intermittently coated with ash and exposed toto include in the GE Industrial and Power flowing flue gas. Samples tested at 1260°CGeneration Division turbine evaluation exhibited 0-30% strength reduction,studies, depending on the material. The CFCC with

protective surface coatings performed best.Some coal flue gas tests were conducted at1000°C with extra alkali sulfate added to

23

assess resistance to hot corrosion. One Applications/CFCC Viabilitycontained 7% alkali sulfate, the other 44%. AssessmentAll materials survived the low alkali

exposures with no degradation. Some FWDC identified property requirement targetsdegradation was observed for the high alkali for selected applications and compared themtesting, to the material properties. Tubes for a high

temperature air heater for a coal-fired powerFabrication hnprovement/Cost generation application were identified asReduction viable. The steam reformer was identified as

another application, but requires furtherPhase I included no fabrication improvement material development to achieve hydrogen-tasks, impermeable tubes. Hot gas transfer pipes are

also a candidate, but also requirenonpermeable material. Westinghouserecommended hot gas filtration as animportant application. Trials at DLC andWestinghouse determined that the desiredporosity and corrosion resistance could beachieved with the CVI material.

i lThe CFCC material tests in simulated

_ application environments were very

promising. These tests, plus comparisons ofmaterial strength versus application designstress, indicate that the DLC CVI materials aretechnically viable for the candidateapplications.

DLC Commercial Scale CVI Fabrication

Facility

Examples of CVI CFCC Parts Fabricated byDuPont Lanxide Composites

24

DuPont Lanxide Composites DIMoxTMTeam

• Phase I Material and Scope stress or the strength. Microcrack yield stressremained at 6-9 ksi. Strength decreased from

DuPont Lanxide Composites established a 36 ksi to 26 ksi,, CFCC material prior to the CFCC program.

This baseline material consisted of 2D fabric * CFCC samples were successfullyof a SiC-based fiber in an aluminum oxide fabricated with alumina fibers. The strength

matrix. The matrix was grown i_o a fabric was low (19 ksi) and degraded further as thepreform by the Lanxide DIMOX directed temperature was increased above 700°C.metal oxidation process. This processinvolves growth of a ceramic-metal composite * CFCC samples were fabricated with anmaterial at the surface of a molten metal alternate fiber. All had low strength andthrough controlled oxidation. If a fiber fractured in a brittle mode.preform is placed adjacent to the moltenmetal, the ceramic-metal material will grow * CFCC samples were fabricated withthrough the fiber preform to form a non- chopped fibers. These fractured in a non-porous matrix. The residual metal is then catastrophic mode, but had decreased strengthextracted to leave a CFCC. compared to baseline material due to the

decreased fiber volume fraction.

The objectives of Phase I were (1) to assessapplications for potential use of the DLC * Tested under a parallel program theCFCC, (2) to fabricate and evaluate four baseline CFCC: Retained room temperaturealternate alumina matrix CFCC materials, strength to 1200°C and only decreased to 33(3) to establish an interface coating process ksi at 1370°C. Water quench thermal shockthat is safe and can be scaled to production testing from 1200°C resulted in a strengthlevels and (4) to increase the matrix cracking decrease of only 16%.yield stress.

• Survived over 1000 hours of stress rupturePartnerships Established exposure at 11.6 ksi and 1200°C in air.

DLC formed applications evaluation * Survived over two million cycles withoutpartnerships with Foster Wheeler failure at 1000°C with an applied tensileDevelopment Corporation, GE Industrial and stress between 8 and 83 MPa at a frequencyPower Generation Division, Solar Turbines of 5 Hz.Inc., and Westinghouse Electric.

• A property database was provided to allProperty Improvement OEM partners for the baseline CFCC

material.The four alternate materials were fabricated

and evaluated. Fabrication Improvement/CostReduction

• SiC particles were added to the matrix.This increased thermal conductivity by 20% Effort focused on establishing a production-and decreased matrix microcracks. However, viable process for coating the fiber fabric withit did not increase the microcracking yield a SiC layer to protect the fibers during the

25

fabrication process. A process was Applications/CFCC Viabilitysuccessfully developed which was safe and Assessmentalso improved the uniformity of coatingdeposition within the fiber preform. FWDC identified heat exchanger tubes forMechanical properties of the CFCC fabricated coal-fired combined-cycle utility generation aswith the modified coating process were a prime candidate for the DLC CFCC •identical to mechanical properties of the material. They also identified steamCFCC with the original process, reformers as a potential application. This

application requires high reliabilitycontainment of high pressure hydrogen gas.The current CFCC material contains openporosity, so further material development isneeded before the reformer application can be

,.. a0 30 considered. Westinghouse identified hot gasE

25 --D-- Original CVI Conditions 25 filtration as an area with high payoff. Trialsvl

_= ---O-- ModifiedCVlConditions determined that the required levels of porosity= 20 p 20._ _ / could not be achieved readily in the alumina.t,_ , matrix CFCC, but could in a CVI SiC/SiC"= 15 _\ _A lu

= _ material. Phase II effort was recommended to

.ff._ 10 10 be transferred to the DLC CVI program. GEO

o 5 5 and Solar identified gas turbine combustor9_ and shroud components as viable for the DLC¢n 0 ,,,_ 0

0 1 2 :3 4 5 (J 7 8 9 CFCC alumina matrix material. Further high

Ply Number temperature durability testing in air undercyclic stress loading is required to verify

Improvement in Coating Thickness material feasibility for these long lifeUniformity with the New Process applications.

26

GE Corporate R & D Team

• Phase I Material and Scope selected turbine components. This linked theGE program to CFCC programs at Babcock

GE demonstrated a baseline CFCC material and Wilcox, Amercom and DuPont Lanxide• prior to the CFCC program. This baseline Composites. These programs included tasks

material consisted of SCS-6 TM monofilament to gather or generate property data forSiC fibers in a Silcomp TM matrix. The compilation into the GE IPG database. Thebaseline material was prepared by casting a Textron role was to provide fiber, continuoussh, rry of SiC plus carbon particles over hand fiber coatings and to establish andaligned fibers to form a thin tape, laminating demonstrate inhouse capability to fabricate thetapes to form a flat panel, and infiltrating with GE CFCC material.molten silicon to react the carbon to SiC and

fill the remaining pores with silicon. The Property Improvementsadvantages of this process are high strengthand toughness, near-zero porosity, short High strengths (approaching the theoreticalfabrication time and no dimensional change potential based on fiber strength) andduring densification, toughness were demonstrated. Of special

importance is the high matrix cracking yieldThe objectives of Phase I were (1) to apply stress in the 26-29 ksi range. This is wellthe baseline interface coating to fibers in a above the predicted application stress.continuous rather than batch process, (2) to Studies have shown that most CFCCdevelop an interface coating with improved materials are susceptible to fatigue andhigh temperature stability, (3) to increase oxidation damage if stressed above the matrixstrength and matrix cracking yield point by cracking yield stress.use of a smaller diameter fiber, (4) to achieveautomated fiber alignment and tape fabrication * Use of 3 mil versus 5.5 mil diameter fibersby drum winding, (5) to fabricate tubes, increased the tensile strength from 55 ksi to(6) to determine the properties needed for 100 ksi and the matrix cracking yield stresscomponent design and assemble a database of from about 22-24 ksi to 26-29 ksi. Strain-to-these properties for the GE material and other failure was increased slightly from aboutCFCC materials and (7) to transfer the GE 0.8% to 0.9%. These results were for

CFCC fabrication technology to a CFCC samples with 20 volume% uniaxially orientedmanufacturer (Textron Specialty Materials). fibers. Increase to 30 volume % should

further increase the properties.Partnerships Established

A team was established headed by GECorporate Research and Develcpment(GE

' CR&D) and including GE lndtistrial andPower Generation (GE IPG) and TextronSpecialty Materials. The role of GE IPG wasto identify candidate gas turbine components,define and gather a design database forcandidate CFCC program materials andassess the viability of the materials for

27

• • t • • ! • • i , • I • • I •

loo 8/1°s lo3 Fabrication Improvement/Cost

/_'_,/_74_ /' Reduction8o * Continuous coating of the SiC fibers was

s/J_ demonstrated for the baseline interface

60

coating. The coating was more uniform than '40 batch coatings and resulted in strength

29.3 improvements up to 50%, as well as in

20 , substantial cost reduction.0 ,i i l , A I , i

oo 0.3 06 0.9 4.2 4.5 * A computer-controlled drum windingstrain(%) apparatus was constructed which dramatically

reduced fabrication time and improvedStress Strain Curves for the Improved CFCCwith 3 mil Diameter Fibers reproducibility.

* A new oxidation resistant fiber coating wasdeveloped and demonstrated to improve roomtemperature and high temperature properties.For cross-ply fiber architecture using the _baseline 5.5 mil fibers, the room temperaturestrength was increased from 35 ksi to 51 ksi.Matrix cracking yield stress was increasedfrom about 15 ksi to over 20 ksi. The new

material retained a strength of 30 ksi at1204°C with the matrix cracking yield stressstill near 20 ksi. The strength reductionversus temperature matches the strength Near-N et Shape Tube 4" Diameter and 4"reduction of the fiber versus temperature, so Long with O.1 inch Wall Thicknessis not due to composite degradation. 100 Reinforced with 5.5 mil Diameter Fibers

o , O ,

hours exposure in air at 1200 C resulted m noproperty degradation.. * Fabrication technology was transferred to

Textron and fabrication trials initiated.' ' I ' ' I ' ' I ' ' I

so- Applications/CFCC ViabilityAssessment

40

"" GE IPG completed the property database_ ao compilation for the various CFCC materials

and conducted analysis to predict the viability20 - .

of the materials for gas turbine components.10 A combustor liner and a turbine shroud were

identified as viable applications. The GE0 CR&D CFCC had the highest probability of0.0 0.3 0.6 0.9 1.2

survival due to its high values of thermalStrain(%) conductivity, matrix cracking yield stress and

Stress Strain Curves at Various Temperatures interlaminar shear strength compared to otherfor CFCC with Improved Interface Layer CFCC materials.

28

Textron Specialty Materials Team

. Phase I Material and Scope composites with monofilament SiC fiber,graphite yam and SiC-based yarn. Over 80

Textron Specialty Materials (TSM) proposed uniaxial tensile strength tests were conducted. a new CFCC concept that would combine the with bars cut from these composites.

fiber and composite fabrication capabilities atTSM with the nitride-bonded silicon carbide * NB SiC matrix with 18 volume % SCS-6

(NB SiC) capabilities of Nova Industrial monofilament SiC fibers exhibited tensileCeramics. The NB SiC is a low cost material strength of 60-65 ksi and strain-to-failure ofin large scale commercial production. The 0.8-1.0%. CFCCs with only 6% fiber hadobjectives of the Phase I program were (1) to tensile strength of 23-30 ksi and strain-to-develop slurry formulations for the NB SiC failure of 0.5-0.6%. CFCCs with 8 and 16%precursors, (2) to use drum wrapping to fiber had intermediate strength. These areapply these slurries to candidate fibers to form excellent properties, especially for such lowtape containing a single layer of fiber, (3) to volume fractions of fibers. Since fibersstack the tape layers to form a multilayer currently dominate cost, low fiber content islaminate, (4) to establish the time/temperature beneficial.parameters to densify the matrix in a nitrogenatmosphere, (5) to measure the strength foreach iteration to guide subsequent iterations, s,,o,,(6) to demonstrate the TSM Rapid _"" ''_''

• • TM

Denslficatlon process to seal surface- I ,,,,o.,,,-,,,-,,dconnected porosity with silicon carbide, (7) to '_ - _ /"_'°"_" [ _"'"'""'"'°'"/

demonstrate fabrication by lowcost filament 5° /_'_/

winding, (8) to establish a team of end users _ ,o _,o,,,.,

and OEMs to study applications and prepare apreliminary commercialization plan.

Partnerships Established ,0o ,o

o I I I , , ,

TSM established during Phase I a vertically o_ o, o_ o_ ,o ,_integrated team consisting of the matrix s,,,,,_,,_material manufacturer (Nova Industrial Typical Stress-Strain Curves for NB SiC withCeramics), original equipment manufacturers Monofilament SiC Fibers(Hauck Burners Corp. and WilliamsInternational), a plant design and construction * NB SiC matrix with 8-10% carbon yamcompany (Stone and Webster), a specialty reinforcement exhibited tensile strength of 15-industrial furnace manufacturer (Schaefer), an 25 ksi and strain-to-failure of 0.2-0.6%.

, end user (Doehler-Jarvis) and a compositesproperty and architecture analyzer (Materials * NB SiC matrix with Tyranno SiC-basedSciences Corp.). yarn had low strength (5-7 ksi) and fractured

' in a brittle mode. The fibers were not

Property Improvements thermochemically stable under the fabricationconditions.

During Phase I, TSM fabricated NB SiC

29

Fabrication Improvement/CostReduction

Extensive process development wasconducted in Phase I, and componentfabrication feasibility was clearlydemonstrated.

* 15 slurry formulations were explored, andthree were selected that adequately coatedfibers and produced homogeneous matrices.

* Monolayer tapes were prepared using drumwrapping. These tapes were then laminatedto form flat plates for property testing andsmall tubes. Tube fabrication time was one

day, which was later reduced to two hoursusing filament winding. Filament windingcan be automated and be very low cost.

* Nitridation time/temperature/gas flow '": '" ....._Radiant Heater Tube 24 Inches Long by 4cycles were successfully developed to achieveInches Diameter Fabricated from NB SiCmaximum conversion of silicon to siliconwith SCS-6 rM

nitride at minimum time and temperature toavoid degradation of fibers. This majormilestone was necessary to demonstrate Applications/CFCC Viabilityfeasibility of the NB SiC concept. Assessment

• . TM The team identified radiant and immersion* The TSM Rapid Denslficatlon process

heaters for aluminum melting and awas successfully demonstrated to fill near-combustor liner for a small industrial turbine

surface porosity of the NB SiC. This processas the first applications to evaluate, to bedeposits SiC at approximately 100 times thefollowed with pyrolysis tubes forrate achieved by CVD processes, so has good

potential for low cost manufacturing. The petrochemical processing. The materialresulting CFCC exhibited a factor of 10 strength appears more than adequate for thesereduction in oxidation at 1100°C compared to applications. Exposure of the matrix materialbaseline material, to molten aluminum for 80 hours resulted in

no sign of wetting or reaction.

* A tubular configuration for a radiant heaterwas successfully fabricated. The fabricationincluded the demonstration of capability tojoin the CFCC material to a monolithic NBSiC flange.

30

Supporting Technology Projects

Scope

The Supporting Technologies Task was established to provide basic or generic support to the, CFCC program in four categories: (1) composite design, (2) materials characterization, (3) test

methods and (4) performance-related phenomena. 18 projects were initiated in Phase I involvingsix universities, four government laboratories and two companies.

Composite Design Subtask Progress

Widespread success with polymer matrix composites has been partially due to a capability toanalytically design the architecture of the composite to meet the needs of the application. Thisrequires a basic understanding of the behavior of constituent materials plus their interactions whenfabricated into a composite, as well as the behavior of the overall composite when a stress field isapplied. The objective of the Phase I composite design subtask was to begin to establishmicromechanics and macromechanics models for CFCC materials, with the long range objective ofachieving experimentally-verified capabilities comparable to those for polymer matrix composites.

Phase I efforts focused on micromechanical modeling, assembly of measured properties to supportthe modeling, and experimental study of crack propagation.

* The micromechanics modeling was conducted at UCLA and included three aspects:(1) fiber/matrix interphase modeling for optimizing toughness and strength, (2) fabric architecturemodeling and (3) laminate modeling. Finite element and closed-form models were developed anddemonstrated to correlate well with existing data for cross-ply architecture CFCCs. Specific datarequirements were defined to provide better verification and model refinement and will be generatedin Phase II. User-friendly menus and user manuals will be prepared for these models and will bemade available to the CFCC industrial teams to support their material development and componentdesign efforts.

* Macromechanics modeling was conducted at Materials Sciences Corporation. The objective is todemonstrate a structural analysis capability that integrates micromechanics materials effects modelsinto macro models for CFCC thermomechanical response. Initial models were defined andintegrated into a finite element code. These account for fiber, matrix, interphase, porosity, fiberarchitecture, fiber/matrix debonding and matrix microcracks. Data requirements for verification andrefinement were submitted to the Test Methods Subtask.

* Bulk thermal conductivity measurements for several CFCC materials were performed at Oak. Ridge National Laboratory (ORNL) and added to the database for modeling. A scanning thermal

conductivity microprobe apparatus was developed and demonstrated to allow study of thermalconductivity effects in CFCC materials on a micromechanics level.

* Techniques were developed at NIST to investigate the initiation and propagation of matrix cracksin CFCC materials. These techniques were applied at room temperature to initial CFCC samplesand were demonstrated to provide very useful information relative to composite design, crack-fiber

31

interactions, and basic understanding of toughening mechanisms and failure mechanisms. The testtechniques will be extended to elevated temperature and simulated environments in Phase II and willbe applied to CFCC samples supplied from the industrial contracts.

Materials Characterization Subtask Progress4

Phase I plans were to address characterization of the individual composite constituents (fibers,matrix, interface) and the complete composites. However, after discussions with the industry teamsand review of other programs (such as at DOD and NASA), the plans were modified. The clearneed was for increased focus on interface characteristics.

* An interface test capability was established and demonstrated at ORNL. It allows directobservation of movement of a fiber in the matrix under an applied stress, plus measurement ofdebond stress, coefficient of friction, sliding shear stress and residual stress. These data helpprovide improved understanding of the material and design factors which control compositestoughness and strength. Samples of SiC/SiC CFCCs were fabricated and tested for mechanicalproperties and fiber-matrix interface characteristics. Preliminary models were prepared to relate themechanical behavior to the micromechanics acting at the fiber-matrix interface.

* A program was initiated at Pennsylvania State University in July 1993 to focus on determinationof optimum interface or interphase characteristics for oxide-oxide CFCC systems. This project willalso interact with the UCLA micromechanics modeling effort by providing samples and data forverification of the evolving models.

Test Methods Subtask Progress

CFCC materials have highly directional structures and properties. They require a different testmethodology than monolithic ceramics. Test methodologies are well established for polymer matrixcomposites, but not for the high temperature regimes of interest for CFCCs. The test methodssubtask is defining the test requirements, establishing thermomechanical test facilities anddeveloping reproducible test procedures. The subtask also is addressing NDE for characterization aswell as for final quality assurance.

* University of Washington and ORNL collaborated in Phase I to assess the test methods neededfor CFCC materials. They defined tests for tensile strength, compressive strength, in-plane shear,interlaminar shear, flexure strength, strain, tension-tension cyclic fatigue and creep/stress rupture.They published a technical highlights report April 1993 "Thermochemical Test Methods forCFCCs". Equipment are being established for these tests, and a sample test matrix has been draftedto obtain an initial database with SiC/SiC CFCC material.

* Non-destructive characterization (NDC) studies were initiated at Argonne National Laboratory(ANL) and at ORNL. ANL focused on infrared imaging and ORNL on ultrasonic techniques.Both showed good potential for detecting delaminations and density variation. Other NDCtechniques were also screened including x-ray radiography, computed tomography, eddy currentand electrical resistivity. NDC techniques appear capable of providing useful data on densityuniformity, voids, delaminations, matrix material distribution and fiber geometry.

32

* ORNL conducted an effort to determine if fibrous debris generated during fabrication posedhealth problems for workers. Sampling was conducted at a CFCC company that used smalldiameter SiC yarn. Airborn fibers and particles were detected, but not in dangerous levels or fiber,L

size.

. Performance-Related Phenomena Progress

Most of the CFCC applications require long life with the components exposed to mechanical and/orthermal stress at high temperature, often in a chemically-aggressive atmosphere. The life will likelybe determined by time-dependent degradation of the CFCCs or accumulation of damage.

* Performance simulation modeling was initiated at VPI. The objective is to combine pertinentdata, basic understanding and models together to predict the remaining strength and life of CFCCcomponents. Effort began in October 1993 and focused on data and supporting experiment needsto prepare simulation models.

* An environmental effects project began at ORNL July 1991. This project seeks to understand theinteractions of the CFCC materials with representative application environments. Key issues arecomposite stability, failure mode and damage tolerance. Efforts so far have focused on carefulthermogravimetric characterization of the oxidation kinetics of SiC/SiC composites fabricated byCVI. Emphasis is shifting to evaluation of the combined effects of stress and oxidation.

* A separate project was spun off the environmental project in October 1993 to focus on time-dependent behavior. The objective is to evaluate the creep and fatigue behavior. This is a#

particularly important task because most of the industrial teams do not have equipment or experiencein creep, creep rupture and fatigue testing. Creep tests were initiated at 1100-1400°C for a SiC/SiCCFCC fabricated by CVI. Creep was low, but the exposure resulted in embrittlement.

* A project has been in progress at Idaho National Engineering Laboratory since September 1991to understand failure mechanisms of CFCC materials and to apply this to prediction of structuralintegrity for applications. A draft report describing failure conditions for structures was distributedfor comment. Fracture toughness testing has begun.

* A project to evaluate the thermal shock behavior of CFCC materials was begun in December1992 at the University of Cincinnati. Effort so far has focused on evaluation of various tests forthermal shock behavior and on selection of one or more tests appropriate for CFCC materials. Theobjective is to select tests that will identify the critical composite parameters that control thermalshock resistance/damage.

33

References

1. Research Management Consultants, Inc., "Continuous Fiber Ceramic Composites (CFCCs) forLow Cost Energy and a Cleaner Environment," report to the U.S. Department of Energy undercontract DE-ACO 1-91CE40956, December 15, 1993.

i

2. Michael A. Karnitz, Douglas F. Craig and Scott L. Richlen, "Continuous Fiber CeramicComposite Program," Cerctm. Bull. 70, 430 ( 1991).

34