Cvd Silicon Carbide Characterization (1994)

8/11/2019 Cvd Silicon Carbide Characterization (1994)

1/54

AD-A285 667RL TR 94 1 I IIliIIM1nIIIFinal Technical ReportAugust 1994

CVD SILICON CARBIDECHARACTERIZATION

DTICS ELECTE '':

University of Dayton Research Institute 0CT 2 11994

F

S b

Ballistic Missile Defense Organization C)

APPARVED,RPULI R E I ; D/SUTR/770N UN/IMITED

The views and conclusions contained in this document are those of the authors and shouldnot be Inlerpreted as necessarily representing the official policies, either expressed orimplied, of the Ballistic Missile Defense Organizationor the U.S. GovernmenL

Rome LaboratoryAir Force Materiel Command

GriffissAir Force Base, New York

941 2 94-32733; - ~~2, _4j. IIIllIIllIlI 1I


2/54

, This report has been reviewed by the Rome Laboratory Public Affairs Office(PA) and is releasable to the National Technical Information Service (NTIS). AtNTIS it will be releasable to the general public, including foreign nations.

RL-TR-94-122 has been reviewed and is approved for publication.

APPROVED: I --

RICHARD FEDORSProject Engineer

FOR THE COMMANDER: ; cf tz

LUKE L. LUCAS, Colonel, USAFDeputy DirectorSurveillance & Photonics

If your address has changed or if you wish to be removed from the Rome Laboratorymailing list, or if the addressee is no longer employed by your organization,please notify RL ( OCPC ) Griffiss AFB NY 13441. This will ass is t us in maintaininga current mailing list

Do not return copies of this report unless contractual obligations or notices on aspecific document require that it be returned.


3/54

CVD SILICON CARBIDE CHARACTERIZATION

G. A. GravesD. Iden

Contractor: University of Dayton Research Inst i tu teContract Number: F30602-90-D-0105, Task 6Effective Date of Contract: 25 August 1992Contract Expiration Date: 24 August 1993

Short Title of Work: CVD Silicon CarbideCharacterization

Period of Work Covered: Aug 92 - Oct 93

Principal Investigator: G. A. Graves, Jr.Phone: (513) 229-4415

RL Project Engineer: Donald LuczakPhone: (315) 330-3144

Approved for public release; distribution unlimited.

This research was supported by the Ballist ic MissileDefense Organization of the Department of Defense andwas monitored by Richard Fedors, RL (OCPC),25 Electronic Pky, Griffiss AFB NY 13441-4515 underContract F30602-90-D-0105.

Ace


4/54

p ormpprovedREPORT D UMENTATION PAGE 0MB No. 0704o1v88PAqigob' b~miIwu -m di d . a*vd towwugm o p r u p w = u r k d x i g f,, tkun-fwrum smud* am in dcaswe .9guinO' nti' r ,s. , w: r ,vr, - c n d Ir Uon Sand'w rtoClng Uor urden .d6WWi 0- c-w aspeaittscAIcdiw Of hrtfb ig s.0U tf r.a imt f to Waindlep H . o u t w SuI*.s, DfruIwg. fn b f a 0g a Rqs, 1215JteiwsonDo*i Hq.AV Sa.* 12 a*VW ,VAZM43OZ mid tatsl Offbe d MwugwTwt and AcOK P~rwiak Rakaft Pvmgt 0704-QWI. Ww*irUw DC 205W3

1. AGENCY USE ONLY (LeaveB l k 2. REPORTDATE 39REPORTTYPE AND DATESCOVERED

I ugust 1994 Final Aug 92 - Oct 934. TITLEANDSUBTITLE 5. FUNDING NUMBERS

CVD SILICON CARBIDE CHARACTERIZATION C - F30602-90-D-0105,Task 6

PE - 63215C6. UTHOR(S) PR - 1101G. A. Graves, Jr. and D. Iden TA - QN

W - 157. PERFORMINGORGANIZATIONNAME(S)AND ADDRESS(ES) 8, PERFORMINGORGANIZATION

University of Dayton Research Institute REPORTNUMBERDayton OH 45469-0172 UDR-TR-93-103

9. SPONSORINGJMONITORINGAGENCY NAME(S) AND ADDRE$S(ES) 10 SPONSORING/MONITORINGBall is t ic Missile Defense Organization AGENCYREPORTNUMBER

7100 Defense Pentagon Rome Laboratory (OLT C) RL-TR-94-122Wash DC 20301-7100 25 Electronic Pky

Griffiss AFB NY 13441-45J5

11 SUPPLEMENTARYNOTESPrepared for Science Applications International Corporation, Rome NY Ti440Rome Laboratory Project Engineeri Richard Fedors/OCPC/(315) 330-3144

12a. DISTRIBUTIONIAVALABIUTYSTATEMENT 12b. DISTRIBUTION CODE

Approved for public release; dist r ibut ion unlimited.

13. ABSTRACT MU*-Lh 2 o w a d s

Chemically vapor deposited CVD) silicon carbide is a candidate material for highquality ground and space-based mirror substrates and high quality reflectiveoptics. Stat is t ical ly valid material property data has not been available, however,to make durability and l ifetime predictions for such optics. The primary purpose ofthis study was to determine the Weibull and slow crack growth parameters for CV Dsilicon carbide. Specimens were cut from various locations in a 25 mm thick, 50 cmdiameter piece of SiC to analyze bulk material property homogeneity. Flexuralstrength was measured using a four-point bend technique. In addition to mechanicaltesting for strength, hardness, and fracture toughness, the material crystallographyand microstructure were studied. Thermal expansion, thermal diffusivi ty, specificheat, optical absorption, and infrared ref lect ivi ty measurements were also conducted.Raman spectroscopy was used to check for any residual stress. Test resul ts show thisCVD silicon carbide is a high-purity, homogeneous, fine-grained substrate material withvery good mechanical, optical, and thermal properties.

14. SUBJECT TERMS IS NUMBER OFPAGES60

silicon carbide, material properties, Weibull analysis, IaPRICECODEmirror substrates, reflective optics

17. SECURITY CLASSFICATION 18. ECURITY CLASSIFICATION 9 SECURITYCLASSIFICATION 20. UMrTATION OF ABSTRACTOF REPORT OF THIS PAGE OF ABSTRACT

UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED ULNSN 7540.01 210- Stardwd Foym 298 1Rev-89)

Prneatuld by ANSI S'd Z39.2w1-1m


5/54

ABSTRACT

Chemically vapor deposi ted (CVD) si l icon carbide (SiC) is aserious candidate for the mirror substrate required for high-qual i ty, ref lec t ive optics based on mechanical, opt ica l , andthermal proper t ies obtained from research quant i t ies of material .Unti l recently, suff ic ient quant i t ies of bulk material were notavailable for obtaining s ta t i s t i ca l ly val id quant i t ies of datarequired for brittle mater ia l l i fe t ime predict ions.

The primary purpose of t h i s material charac ter iza t ion studywas to obtain suff ic ient mechanical property data for CVD SiC todetermine the Weibull and slow crack growth parameters. Data wasobtained on flexural strength (four-point bend) specimens takenfrom a bulk piece of mater ia l or ig inal ly 0.5 m in diameter byapproximately 25 mm thick. Specimens were cut from variouslocations within the mater ia l to allow for a st tistic l analysisto determine material property homogeneity throughout the bulk.In addit ion to the mechanical strength, crystal lography,microstructure, fracture toughness, hardness, thermal expansion,thermal diffus ivi ty, speci f ic heat, opt ica l absorpt ion, andre f l ec t iv i ty were studied. Raman spectrographic analysis was alsoused to determine if residual s t ress was present in the bulkmaterial .

The conclusions from th i s study were tha t chemical vapordeposi ted s i l icon carbide can be produced which is pure,homogeneous, f ine grained material . The samples evaluatedexhibi ted very good mechanical, optical , and thermal propert ieswhen compared to other materials being considered for opticalsubs t ra tes .


6/54

TABLE OF CONTENTS

Section PageINTRODUCTION 1

OBJECTIVE 1APPROACH 1BACKGROUND 2TECHNICALAPPROACH 2

Weibull Analysis 3Slow Crack Growth and MinimumTime to Failure 4

Dynamic Fatigue Analysisof Slow Crack GrowthParameters 5

EXPERIMENTALPROCEDURE 7

TEST MATRIX 7MATERIAL 9X-RAY DIFFRACTION 9MECHANICAL PROPERTIES 9

Flexural Strength 9Dynamic Fatigue 10Fracture Toughness 10

PHYSICAL PROPERTIES 11Young's Modulus and Poissons Ratio 11Vickers Microhardness 11

THERMAL PROPERTIES 11Coefficientof Thermal Expansion(CTE) 11Thermal Diffusivity 11

Specific Heat 11OPTICAL PROPERTIES 11

CalorimetryMeasurement Apparatus 11Spectral Reflectance Measurements 12Total Integrated Scatter (TIS) 12Raman Data Analysis 12

ii


7/54

TABLE OF CONTENTS (concluded)

Section PaeDISCUSSIONOF RESULTS 14

CRYSTALLOGRAPHY 14MICROSTRUCTURALANALYSIS 15MECHANICALPROPERTIES 17

Flexural Strength 17Inert-Strength Weibull Parameters 22Slow Crack Growth 23

PHYSICAL PROPERTIES 24Young's Modulus and Poisson's Ratio 24Fracture Toughness 25Microhardness 25

THERMAL PROPERTIES 26Coefficientof Thermal Expansion (CTE) 26Thermal Diffusivity 26Specific Heat 28

OPTICALPROPERTIES 28Absorption Measurement Data Reduction 28

Reflectance Measurement Results 29Total Integrated Scatter 29Raman Microprobe Measurements 31

CONCLUSIONS 36

ACKNOWLEDGMENTS 36

REFERENCES 37

iii


8/54

LIST OF ILLUSTRATIONS

Figure Page

1 Bulk CVD SiC specimen cutting diagram. 8

2 TIS measurement apparatus. 13

3 Schematic of 03SiC crystallographicstructure. 15

4 Microstructure of CVD SiC near deposition surface. 15

5 Microstructure of CVD SiC near substrate surface. 16

6 Microstructure of CVD SiC in the plane of the growth direction. 16

7 Linear regressionfit of Weibull flexural test data. 228 Weibull failure probabilityversus strength plot of flexural

specimens tested in LN 2 . 22

9 Dynamic fatigue resultsfor CVD SiC. 23

10 Least squares fitof CTE data. 27

11 Plot of the specific heat results for CVDSiC. 28

12 Typical reflectivitycurve for CVD SiC. 30

13 Diagramof Raman microprobe specimen. 3114 Raman spectral line broadening due to residual strain on the

side of the specimen that hadbeen cut, ground, and beveled(i.e., chamfered). 32

15 Raman line broadening due to residual strainon end and sideof specimen. 32

16 Raman scan on thebevel of the cut and ground side 150 g.mfrom the growth surface. 33

17 Raman scan on the diamond cutend at the growth surface (0 gim)

of the specimen. 3318 Raman scan on thediamondcut end 3500 pm from the growth

surface of the specimen. 34

19 Raman spectralline broadening due to residual strainonpolished surfaceof the mirror specimen (surface roughness-3A rms) and theground and beveled side (or edge). 35

iv


9/54


10/54

INTRODUCTION

OBJECTIVE

For several years Morton International(MI) was supported by the Air Force/SDIOto perform researchand development of fabrication methods for preparing cubic J3) silicon

carbide (SiC) by chemical vapordeposition (CVD) for optics applications. Earlier

characterization of the material taken from variousresearch fabrication processingruns

indicated thatthe materialhad extremely high-purity,was very homogeneous,with highaverage strength and Young'smodulus values for a polycrystalline ceramicmaterial. Duringthe development process, however,the resultsof various characterization studies were

somewhat inconsistent, as MI worked to improve control over its process, while scaling

equipment to fabricate 1.5 m diameter pieces up to 25 mm thick. There wasan insufficientamount of material and inadequate fundingavailable to prepare a statistically meaningful

compilation of material propertydata. It became apparent duringthe final phase of theprogram that an extensive material characterizationstudy, producing consistent,reliablevalues, would be necessary if designers were to have sufficientconfidence in the lifetime

predictions required for incorporatingthe material into their optical systems.

The objectiveof this program was to provide the required thermal, mechanical, andoptical property datafor SiC synthesizedby the CVD process. Thiswas necessary to establisha validated design guide for use of this material in critical optical/ structuralapplications.Major emphasis was placed on obtaining sufficient mechanicalproperty data to determine

Weibull and slow crack growthparameters. An analysis of variance (ANOVA)was alsoperformed to determine if significant differencesin mechanical strengthexisted in specimens

obtained from different locations in the bulk material. The m easurementof thermal properties,

conductivity, diffusivity, specificheat, and coefficientof thermal expansion(CTE) as well assome optical propertieswere also performed.

The test matrix design for thisstudy andthe test procedures used to acquire the datawere based on UDRI's past experience in dealing with lifetimepredictionsfor brittle materialsbeing used in structural applications.

APPROACH

The Government provided bulk CVD silicon carbide preparedby Morton Internationalin sufficient amounts to conductthis study. The bulk material was representativeof the best

material prepared in an Air Force/SDIO-sponsoredmaterial development program at MI.That program demonstrated that careful controlof CVD processing parametersproduced

material with reproduciblemechanical and opticalproperties. UDRI verified this consistency


11/54

by choosing random samplesof new material fromthe scaled-up process and comparing themwith data taken from material made duringthe developmentphase. Both materials were made

using identical techniques,the original material was made in small experimental furnaceswhere aore precise control was possible.

BACKGROUND

Design engineers have long recognized the need forestablishing statisticallysignificant propertyvalues for structural material. Because of their ductilityand associatedplastic deformation,the prediction of failure of metals can be quite precise. On the other hand.

ceramics typicallydo not exhibit ductile behavior and their mechanical failure is much lesscertain. Structuresmade of ceramics most often fail in tension at a flaw on theobject's surface.Typically, the flaw is small and cannot be easily detected. The tlaw tip is usually quite sharp(of atomic dimensions) and, unlike metals, doesnot blunt when put in tension due to the lack

of plastic flow. Catastrophic failure occurs at stresses dependenton the statistical distributionof surface flaws; thus, the relative unpredictdbilityof the failure conditions.

Regardless of their brittle behavior, ceramicshave come to the foreground of materialsresearch becauseof their unique properties. These includehigh specificstiffness, chemicalstability in severe environments,strength retentionat elevated temperatures,dimensionalstability overlarge temperature excursions,a wide range of thermal properties, and in somematerials, excellentoptical properties overa wide spectral range.

Despite the lack of precision in predicting a ceramic object's mechanicalstrength,systems can be appropriately designedusing thesematerials in structural applications. Witha reliable property values established,designers can have a high degree of confidencethata ceramic componentwill survive in use. It is important that data collectionbe sufficientlycomplete to ensure a design based on the data is neither overly conservative or dangerouslyunderdesigned.

TECHNICAL APPROACH

The primary goal of the program was to obtain sufficient mechanical property data forCVD SiC, to perform a meaningful analysis,and provide designengineers with the tools forlifetime prediction. Due to the variation of flaw sizes that lead to failure in a brittle material,statistical methodsmust be used to analyze mechanicalproperty strength data. Acceptablemethods have been found thatuse the Weibull distributionfunction as the basis for determiningthe statistical parametersrequired to predict the probabilityof failure for a brittle ceramicmaterial. Over the past 20 years, severalmodels usingthe Weibull parameters have beendeveloped for use in designing structural opticalcomponents.

2


12/54

Weibull Analysis

The Weibull two-parameter statistical strength modelis most oftenused to characterize

the strength of brittle materials I11. The probability, F, that a component or specimen will fail

at a surface flaw is given by.

F = I - exp -{LiJdS ( )

F = 0fora


13/54

where 1gis the flexural specimen inner gage length, w is the specimen width, and t is thespecimen thickness. The factor k accounts for the possibilityof failures between the innerand outer gage points, and is given by,

k=I[irn+(L/lg)]/[m I+ (5)

where L is the outer gage length.

When strength distributionsbased on experimentalevaluation of the strength of brittle

materials deviate from a Weibull distribution,such deviation may be due to the existence ofmore than one distributionof failure-initiatingdefects in the as-fabricatedspecimens.

The Weibull cumulative probabilityof failure, F = I - R, for a group of identicalspecimens tested underidentical conditions can be evaluated through the use of a rankingestimator as a function of specimen strength. Forthis analysiseffort the Median Rank Estimator

was used [4]. For each data set the specimen strengthswere ordered from weakest to strongestand the failure probabilitywas determined for each specimen. The median rank estimator is

F = ( ni- 0 5 )/N (6)

where ni is the rank of the ith specimen and N is the total number of specimens in the group.

The determination of the Weibull parameters was accomplished through a linearregression analysis based on the two-parameterWeibull distribution (Equation l):

log{ -ln(R) I = log{ 1/),n I + m logi a 1 (7)

Equation (7) correspondsto the linear equation

Y = A + b X (8)

where Y = log( -ln(R) 1, A = log I 1/f'n ), b = m, and X = log{ a }.

Slow Crack Growth and Minimum Time to Failure

Ceramic materials are also susceptibleto delayed failure due to a phenomenonknownas slow crack growth. Failure results from stress-assistedgrowth of subcritical flaws to acritical size in the presence of moisture or other corrosiveenvironments. The factors governing

the rate of subcriticalcrack growth in a given environmentmust be determinedbefore a relation-ship between applied stress and failure probability can be determined.

Tests on a number of brittle materialshave demonstratedthat the rate of crack growthincreases as the water contentof the environmentincreases [13]. The rate of crack growthhasbeen demonstrated to exhibit a power-law relationshipwith the stress intensitylevel, K, [5]:

4


14/54

' = c(Kc) A K (9)

where K, = Y ya, s, n is the slow crack growth exponent, and vo and A are alternative second

crack growth parameters.

Integration of Equation (9) yields a relationship that can be used to calculate the

minimum time, tmin, that a specimen or component will survive when exposed to a constant

maximum applied tensile stress, Y, [2 ].

2mi 1( 2/1 - 1n )2 (10)= (n- 2) A Yna" c}{I(n2 121 c

[his can be simplified to

2"m , 2,(n_2) ( 11)train (n-2) A Y2 aKin

where the initial value of the stress intensity is given by0.5

Kii = Y Ga ci (12)

The variable ci is the size of the maximum flaw in the region of maximum stress, Ga, when

the component enters operation. The parameter cf is the size of the largest flaw when t = tmin-

Assuming that the part fails at t = tri,

( 92cf = ,Y'J (13)

Using the above relationships, the time to failure can be estimated for any applied stress and

failure probability if the Weibull parameter m, 7i, Kic, crack growth rate, n and B are known.

Past experience at UDRI has shown that n and B are best determined by using "dynamic

fatigue" techniques (most economical relative to test time and material required). The material

of interest is tested to obtain its inert strength at a very high loading rate in liquid nitrogen.

Three or more loading rates, each separated by at least an order of magnitude, are then performed

in a corrosive environment. Using a relationship between the applied stressing rates and the

fracture stress, n and B can then be calculated.

Dynamic Fatigue Analysis of Slow Crack Growth Parameters

The minimum time to failure Equation 11) may be written in a simpler form using an

alternative second slow crack growth parameter, B, which may be determined through a dynamic

fatigue approach [6,71:

5


15/54

(n-2)a-

tmin = B - (14)Oa

The intrinsic, fast-fracture(zero crackgrowth), strengthof the componen: is oi and the

alternative second slow crack growth parameter, B, is given by2B 2 15 )y2(n-2)AK )n-2) A yKIC-2

The analytical basisfor the dynamic fatigueapproach is the crack growthrate power

law relationshipdescribed in Equation (9). The integrationof the crack growth rate equation

is carried out with the assumptionthat the stress rate, d, is constant. With such an

assumption, n+l) (n-2) (16)af "-6B (n +1)ao (6

where of is the failure stress of a specimen or parttested at a constant stress rate,d. Taking

the natural logarithm of Equation (16) yields,

Ind In[ B ( n+l )]+ ( n-2 ) In(a1 ) (17)11 k(f) (n+l (n+l)

This maybe linearized to,

In(of) = aIn + a 2 (18)

where

a1 - (n+1) (19)

and

ln[B(n+l )]+(n-2)In(a.) (20)( n+l )

Flexure testsat two or more constant stress ratesin a water environment, combinedwith

knowledgeof the intrinsic strengthof the material, provides the basis for evaluation of the

linear equation variablesa, and a2 , and subsequently, for the slowcrack growth parametersn and B. Because of the scatter in the strength of brittle materials it is necessary to employ

strength valuesat equivalentreliability levels. Themedian strength(R = 0.5) for eachconstant rate group of specimens, as well as the median strength of the intrinsic strengthgroup, providesa reliable basisfor this calculation.

6


16/54

EXPERIMENTALPROCEDURE

TEST MATRIX

The test matrixfor this studywas developedbased on Weibull analysisof four-point

flexure tests on 30 specimensand dynamic fatigue testing on 30 specimens at each of fourloadingrates (Table 1). Past experience has shown thata sample of 30 specimens is usuallyadequate to perform a Weibull analysis on data obtained from dense, wellprepared,homogeneous,brittle materials. In this study extra specimens were prepared for eachsampleset to ensure that at least 30 specimenswould be available forflexural strength testswhenWeibull parameter determinationswere required. If more than 30 specimens survivedthecutting and finishingprocess, all the specimens were tested. As indicated in Figure 1, theodd numberedspecimenswere obtained from near the deposition surface (top)and the evennumbered specimenswere obtained from the substrate side of the bulk material. All of the

specimens tested in this study were obtainedfrom the locations shownin Figure 1.

Table 1. Number and Dimensionsof Specimens Tested

Specimen Numberand Dimension

CTE jCpTE) 6 specimens (50 x 6.5 x 6.5 mm)

Reflectivity and Zygo 6 specimens (38 mm dia. x 6 mm thick)

Optical Absorptionand TIS 6 specimens (38 mm dia. x 6 mm thick)

Thermal Diffusivity(aTD) 6 specimens 13 mm x 13 mm x 2.5 mm)

Specific Heat (Cp) 3 specimens (scrap material)

Density (p) 3 specimens (flexure specimenswere used)

Flexural Strength (OMOR) total of 150 four-point bend specimens (50 mm x 4 mm x 3 mm)30 specimens testedat -I 800C30 specimens testedat room temperature, 4stressingrates (120)4-point bend bars (50mm x 4 mm x 3 mm)

FractureToughness (Klc) 10 specimens (50 mm x 5 mm x 4 mm)Young's Modulus and

Poisson'sRatio 8 specimens (75 mm x 13 mm x 6.5 mm)

Microhardness(VH) used Kic specimensfor this study

X-Ray Diffraction 5 specimens (broken MORspecimens were used)

Metallography 10 specimens (broken MOR specimenswere used)

7


17/54

6 Ev

/ 4-..

ox

.;i: Z) n3 C0 M >

L(Ui

-4r-

19 L

w -- , . .. ......... ... .. . . .O--s_ E'E Ee

-OL 4-6. 6 ILn

0_ . . iS E~n

W r

U,' E *

cjOxx

*0 C.

8 L


18/54

MATERIAL

The bulk material usedto prepare specimens was a pie-shaped (1200 included angle)

piece of chemically vapor deposited (CVD) silicon carbide(SiC) cut froma 0.5 mm diameter x

25 mm thick plate originally depositedby Morton International Inc. (MI), Woburn, MA. The

final thickness of the piece was 16 mm; 8 mm was removed from the as-deposited side and 1 mmfrom the substrate side by diamond grinding. All of the rectangular specimensused in the study

were cut from the bulk material as shown in the cutting diagram(Figure 1). A diamond cutoff

saw,I using diamond embedded saw blades,2 was used to cut the samples. The sample sideswere ground flat andparallel witha grinding wheel3 on the same machine. When available,four

specimens from each group wereselected for each set. Disk-shapedoptical specimenswereultrasonically machinedfrom the bulk material before being groundflat to within 1/10 X and the

edges beveledby Chan Kare Corp, Worchester,MA. Disk specimenswere finalpolished to a

surface roughnessof approximately3A rms by Janos Corp, Townshend,VT.

X-RAY DIFFRACTION

Crystallographicstudies were conducted witha Norelco x-ray diffractometer using Ka

Cu radiation.

MECHANICAL PROPERTIES

Flexural Strength

Table 2 contains the actual number of test specimens usedin this program and the testconditions for each groupof specimenstested.

Table 2. Program Test Matrix for MechanicalProperties

Group Number Rate (m/s) Environment

1 32 8.4 x 10-6 liquidN2

2 32 8.4 x 10-5 distilled water




1Harig618 slicer.2Norton ASD120 R 100B99,Norton Company, Worchester,MA.3 Radiac ASD 120R75.

9


19/54

The inert or fast fracturestrength of CVD SiC was determinedby four-pointbendflexural tests conducted in a liquid nitrogen environmentat a load displacementrate of8.4 x 10-6m/s (0.02 in/min), group 1 in Table 2. The conditions for the tests were chosen tominimize effects of slow crack growthand provide a sound basis for the determination of fast-

fracture or inert-strengthWeibull parametervalues.The inert strength flexural tests employedSize B specimens as identified in Mil-Std-

1942A, 4 mm wide, 3 mm thick, with an outer span of 40 mm and an inner span of 20 mm.Uniform tensile stress loads were applied over the 20 mm span of the tensile surface of thespecimens. A group of 32 specimensobtained from each MOR group (Figure 1) was tested.

Dynamic Fatigu

The dynamicfatigue measurements,groups 2 to 5 in Table 2, were taken using the sameequipment, specimen preparation, and specimensize as group 1. Dynamic fatigue tests wereconducted to obtain surface-flawbased, slow crack growthparametervalues forCVD SiC.When possible, four specimens from eachMOR group, two fromthe top and two from thebottom, were selected. The waterenvironmentprovideda worst-case condition to assure thatparameter valueswere conservative. The specimens were immersed in water contained in aplexiglass box resting on the testing machine platen.

Fracture Toughness

The fracture toughness(K,,) determinationswere performedusing a controlledflawtcchnique. A diamond Vickers indent

was placed on the tensile surface of a flexure bar usinga microhardnesstest apparatus.4 A load sufficient to obtain cracks at the corners of the longdiagonal of the indent was used. The lengthof the cracks w'ere measured using low poweroptical microscopy. The specimens were then loadedto failure in a Universal testing machine5.The KIc value was determined using the followingformula:

Kic2 = 2.06 af(c/lh)/2 (24)

where of = flexure strength (MPa)

c = radial crack extension(m)

4 Zwicke Hardness 3212 Testing Machine, E. Windsor, CT51nstron Model 1123, Instron Corp, Canton,MA

10


20/54

PHYSICAL PROPERTIES

Young's Modulus and Poisson's Ratio

A resonantfrequency apparatus6 was used to analyze the flexural and torsional vibrations

on a 75 mm x 13 mm x 6 mm specimenafter an impulse excitation. From these resultsthe

Young's modulusand Poisson's ratio are calculated.

Vickers Microhardness

Vickers hardness measurementswere madeusing the microhardnesstester7 with a

Vickers indenter. Measurementswere obtained from the polished surface of fractured flexural

strength test specimens usinga 100 gm load.

THERMAL PROPERTIES

Coefficient of Thermal Expansion (GTE)

The coefficientof thermal expansion (CTE)measurementswere obtained on a double

rod quartz tube dilatometer8 using a NIST 739-LI fused silicastandard.

Thermal Diffusivity

Thermal diffusivity measurementswere carried out on a Xenon flash lampapparatus.

The frontface of a 13 mm x 2.5 mm specimen is irradiatedwith an energy pulsefrom the Xenonflashlamp. Theresulting temperature profileis acquired on the back side and storedon a digital

storage oscilloscope9 for furtheranalysis.

Specific Heat

Specific heat measurementswere madeon a differential scanning calorimeter'0 using

sapphire as the reference material.

OPTICAL PROPERTIES

Calorimetry Measurement Apparatus

Absorptanceof SiC was measured using laserrate calorimetry at X =10.6 mm. This

consists of a source laser'1 operating single line at 10.6 g~m, beam steeringand alignment optics,

a vacuum calorimeter chamber, and data collectionapparatus. A 10.6 p.m beam enters the

6Grindo Sonic,MK3, W.Lemmens Co., St. Louis, MO.7ZwickeModel 3212 HardnessTesting Machine.8Dilatronic Dilatometer,Theta Industries,Inc., Port Washington, NY.9Nicolet Model 310Digital StorageScope.IDuPontInstruments DSC2910,New Castle, PA.11CoherentModel 41 CO2

11


21/54

chamber through an AR coated window, strikesthe sample, andis reflected backthrough the

window and steered into a power meter.12

Temperature rise within the sampledue to bulk absorptionis monitoredby twothermocouples separated diametricallyby 1800 which contact the sample side. Sensitivity

is 80 gic/C of temperature change. A nano-voltmeter13 amplifies the signal and the data arerecorded on an x-y recorder. Powerreflected is also recorded.

Absorptance was measured at one location, the approximatecenter of the sample disc.

The beam size was approximately 5-6 mm.

Spectral Reflectance Measurements

A Fourier transferinfrared spectrophotometer (FTIR)14 was used to collect spectralreflectance survey scans from 2,500to 25,000 nm on all of the samples. This instrument

has a demonstrated precision of four decimal places;sufficient to clearly separate spectral

differences between specimens.

Total Integrated Scatter (TIS)

TIS was m easuredat X= 0.6328 gm. The apparatusused to determine the TIS is

illustrated in Figure 2. The specimen is located nearthe midplane of a hemispherical collecting

mirror comm only referredto as a Coblentz sphere. The radiation scattered fromthe specimenis imaged on a detector, and the reflected radiationis directed back through the center of the

entrance hole in the sphere to an absorbing target. The circular entrancehole spans an angle of

approximately 2.50. The angle of incidence is maintained constantat 10 for all measurements.The radiation sourceis a HeNe laser operatingat 633 nm. Rotating polarizersmay be used for

attenuating the laser beam,and apertures may beused to eliminate diffractionand extraneousscattered radiation. The Si detector outputsignal is measured witha suitable amplifier.

Raman Data Analysis

It has been wellestablished thatthe width orthe position of a Raman line canchangedue to strains introduced by either residualor applied stresses. The transverseoptical mode

(TO) Raman line at 795 cm-1 was recorded usinga double spectrometer modifiedto use a

liquid-nitrogen-cooled CCD array detector. The 795 cm-1

line is two-fold degenerate andhasbeen observed to split with some specimens of cubic CVDSiC, presumablydue to layering.

12CoherentModel 20113Kiethley Model 14814Nicolet Model 740

12


22/54

SSHUTTER,ATTENUATORCHOPPER A UER..

.ARGETON APERTUREPNATE

ONLY ONA DETECTOR

tT IS POSITIONED

VRSV3 . V~s

SCATTERED.INCIDENT.R REFLECTED

RADIATIONLOCK-INI

AS RERED.SYNC

R A R P

Vi

i

Figure 2. TIS measurementapparatus.

The specimens studied here showed no splitting.Therefore, the behavior of the full width at halfmaximum (FWHM) of the Raman line, which is generally sensitive to residual strain introduced

by the fabrication process, was measured.he laser spot size of the Raman microprobedimensions for these studies was less than2 microns wide by ~10icrons long. The narrow dimension was aligned parallel to the top-to-bottom direction to give high spatial resolution in the direction of translation. The laser wasoperated at 532 nm with less than 1 mW average power incident on the specimen. Spectrometerresolution was about two pixels or 0.7 cm- 1 with the measured widths ranging from 2.8 to4.7 cm" . All measurements were taken with

the specimens at room temperature.

13

IT IS POSIIONE


23/54

DISCUSSION OF RESULTS

The numberof specimens prepared for testing and their dimensionsare shown inTable 1. The specimen locations were selected froma variety of locations withinthe plate toprovide informationon the consistency of the material's mechanical properties. An analysis ofvariance was performed on the data to determine if a statistically significant differenceexistedin the mechanicalproperties of specimens taken from the various locations.

In this report any reference tothe substrate side (bottom) of the bulk material is referringto the material depositednearest the graphite substrate (initial materialdeposition). Reference tothe deposition side (top) refersto the material nearestthe final deposition surface.

In order to ascertain if CVD SiC from the large arc section usedin this study wascomparable to previouslycharacterizedmaterial obtained from smaller furnaces,x-ray

diffraction, and metallographic analyses were performed. Thesetests were usedto determinecrystal structure,possiblepreferred crystallographic orientation,grain size,void and/or inclusioncontent, and general surface conditionsof polished specimens priorto further testing. Specimens

for these studieswere selected from various positions within the bulk material(Figure 1).Specimens also were prepared andtested to obtain thermal, optical, and physical propertiesfrom various positions in the bulk to further verify the homogeneityof the material.

The specific heat (Cp), coefficient of thermal expansion(CTE), thermaldiffusivity gi),density (p), and Vickersmicrohardness (VMH) weredetermined. Opticalproperties studied

includedtotal integratedscatter, absorption,and reflectance.

CRYSTALLOGRAPHY

X-ray diffraction results (Table3) obtained from a specimen 16 mm thick indicated thematerial was 13SiC (Figure3) throughoutthe thickness. Minimal differences of intensity in thethree orthogonaldirections showed the amount of preferred orientation to be minimal even

though thegrowth habit of the grains is needle like. There was no indication of any hexagonalphases.

Table 3. X-ray Diffraction Results

SiC Side I SiC Bottom SiC Top

20 I d (A) 20 I d 20 1 d

35.75 100 2.512 35.80 100 2.508 35.90 100 2.501

41.50 9 2.176 41.55 4 2.173 41.65 4 2.168

60.05 38 1 541 60.10 21 1.539 60.25 41 1.536

14


24/54

Si

Figure 3. Schematic of 13SiC crystallographicstructure.

MICROSTRUCTURALANALYSIS

Micrographs of the as-received materialare represented in Figures 4through 6. Theseshow typicalmicrostructures observedon specimens taken from variouslocations in the bulk

material.

Figure 4 shows the small grainsize found on material deposited latein CVD depositionrun. Much larger grains, shownin Figure 5, are a result of the bulk material being held at thedeposition temperature fora long periodof time (-72 hours). The typicalelongated needle likeCVD SiC grain structurein the plane of the growth directionis shown in Figure 6.

Ai

Figure4. Microstructureof CVD SiC near depositionsurface.

15


25/54

4z~

~ r Pr

ki I

A'f100 p

Figure 5.icrostructureof CVD SiCt hp lnea r f btrae suwtdrfac.on

16


26/54

MECHANICAL PROPERTIES

Flexural Strength

Inert flexural strengthvalues for specimens tested in liquid nitrogenat a loading rate of

8.4 x 10-6 m/s (0.02 in/min) are shown in Table 4. The results of the dynamic fatigue flexural

strength testing in water, at four loading rates over four orders of magnitude,are shown in Tables

5 through 8.

Table 4. Flexural Strength Results for CVD SiC Tested in Liquid Nitrogen at a Loading Rate of8.4 x 10-6 m/s (0.02 in/min)

Flexural Strength

Specimen Number (ksi) (MPa)C-4 34 236G-4 36 253J-4 36 254H-2 38 265B-4 40 279C-2 43 297H-3 44 303F-2 44 308D- 1 44 308D-4 45 317E- 1 46 322C-1 47 325G-2 48 334

E-4 50 350E-2 52 361F-3 53 368H-1 53 370E-3 54 372H-4 54 374G-1 54 374G-3 54 376F-4 54 377F- 1 55 380B-2 57 398D-2 59 409J-1 59 41 0J-2 60 417C-3 61 427B-3 62 432J-3 63 438B-1 66 460D-3 68 475

Average 51.5 355Std. Dev. 9.1 62

17


27/54

Table 5. Flexural StrengthResults for CVD SiC Tested in H20 at a Loading Rate of

8.4 x 10-6 m/s (0.02 in/min)

Flexural Strength

Specimen Number (ksi) (MPa)

H- 12 22.6 156H-10 29.3 202C-9 30.8 212C- 11 36.8 254E- 12 37.6 259G-10 38.7 267E-1 1 39.4 272G-9 43.4 300G 1 1 45.3 312J- 1 47.7 329F-9 48.2 332F-12 48.3 333D-11 48.4 334J-9 48.4 334H- 1 49.2 339B- 1 49.4 341B-10 49.4 341J-12 50.8 350D-9 51.2 353B-12 52.0 359J-10 52.0 359B-9 52.2 360C- 12 52.7 364

F- 1 53.837 1

G-12 54.2 374D-10 54.3 374D-12 55.8 385F-10 55.9 385E-10 56.4 389H-9 57.6 397E-9 62.0 427Average 47.5 328Std. Dev. 8.9 62

18


28/54


29/54

Table 7. Flexural Strength Resultsfor CVD SiC Tested in H20 at a Loading Rate of8.4 x l0-7 m/s (0.002 in/min)

Flexural Strength

Specimen Number (ksi) (MPa)

H-16 23 160H-14 32 224C-15 33 229J-15 35 247E-15 38 263F-13 39 272E-14 40 276B-16 40 278J-13 42 290F-16 43 297F-14 43 302H-13 44 305

E-13 45 315G-16 47 324G-13 47 325C-13 48 334B-14 49 339C-16 50 347J-14 51 355J-16 52 361B-15 53 367D-15 53 368C-14 53 369F-15 54 375B-13 54 375G-14 54 377D-13 55 384D-14 56 386D-16 57 398G-15 57 399E-16 59 410H-15 59 413


20


30/54

Table 8. Flexural Strength Results for CVD SiC Tested in H20 at a LoadingRate of8.4 x l0-8 mrs (0.0002 in/min)

Flexural Strength

Specimen Number (ksi) (MPa)H-18

20 14 1H-20 23 158J-17 32 222F-19 33 233G-19 36 251E-19 37 261F- 17 42 291B-17 42 296C-20 43 298E-17 44 305B-18 44 307E-20 44 309

J-19 44 310F-20 45 315B-20 45 315F-18 47 326D-18 47 327G-18 47 328G-17 47 328D-19 49 339E-18 49 340C-19 49 340G-20 50 347J-18 50 349J-20 50 351C-17 51 354D-20 52 365H-19 54 375D-17 55 384H-17 57 395C-18 61 421


Although the average strength results are lower than previously tested CVD SiCdepositing runs, the strengths obtained are still quite high for monolithicceramics. The lowerstrengths may be due to the larger than usualgrains resulting from the longer CVD depositiontimes for this deposit which is 4-10 timesthicker than previous MI CVD material testedatUDRI.

21


31/54

Inert-Strength Weibull Parameters

The linear regression relationshipemployedto evaluate the inert strength Weibullparameters is illustrated in Figure 7, and the corresponding Weibullfailure probability versus

failure stress is plotted in Figure 8.

1.000

Specimens tes ted in LN2Loading ra te = 8.4 x 106 m/s

n 6.77; Tt

-0.5000 ........... 0 o fd n eBn

00

-2.0000 i I I I i

2.3,50D 2.40M0 2.4510D 2.5000 2.5500 2.6]00D 2.6500 2.70W0

log (Strength. MPa)

Figure 7. Linear regression fitof Weibull flexural test data.

0.9 -

Specimens tes ted in LN 20.8 Loading ra te - 8.4 x 10 6 m/s

Sm~r = 6.77 /-O.( ,7.465 MPo m^2/m2 0.6S952 m^

0.9

L-L0.4

.2

a

0 0.

S0.2 0 - LnarRgeso

0.4

0.00 100.00 200.3W.00 4W.00 5 60.0

Strength (MPa)

Figure 8. Weibull failure probabilityversus strength plotof flexural specimenstested in liquid nitroge

22


32/54

The results of the Weibull inert surfacestrengthevaluationare:

Weibull Modulus, m: 6.77Weibull Scale Parameter, (o: 97.46 MPa m( 2Jm)Average Strength: 355.2 MPaWeibull-Based Median Strength: 360.4 MPa

Slow Crack Growth

The linear relationship between In(stress rate) and ln(medianstrength) employed by thedynamic fatigue approach to evaluate the slow crack growth parametersis illustrated in Figure 9.

The results of the Weibull analysis of all flexure tests listedin Table 2 are summarized in Table 9.

The jesults of the dynamic fatigueevaluation of slow crack growthparameters for CVD

SiC SizeB Flexure specimens are:

Slow Crack Growth Exponent, n: 216

Slow Crack Growth Parameter, B: 7.88E-6 sec MPa 2 .

CVD SICD ~ m FatgwueWarW Enrvror fvtat 21C

U M~dm 24WUN. .

5.8m0

:: _ __.----- ----- ---- -.750W-2.0O0 -I OOM 001D0 I D 2.O3O0 & 0 0 5.0 0 6,00W

Wo tr rate) WIVIPOISec)

OYkAAMCAT1GUE NALYSIS OR CVD SiC SIZE W FLEXURE KPECIMENS

LINEAREL.ATIONMP

Rde Rate Stwo Medhll Ratel Sinu I al. o.00461

stumgdmw __ n:____Mw- il - mu Mpg NIP I y ] hhrIO (1W) -XOETn=- 216

5.06E,04 I99E.01 317 320 .16.147 $.757 5.717 B = w (e .2141(11 - 3) in(somd I I1 815.08603 I9690.0 332 324 06S 5.6049 5.7793 8 7.8E96.00 uMIPW5.09E42 1 99E.01 319 327 110 75. 5.7M ALTENATE P5.0801 1 0.9gE02 333 330 52931 59079 5.M06 v0.2 KIc-2/IB Y11 (nl-)}

Y=t ; : .. 0. UIPS W.5vim YO. 4650.03 tWuc

506E.02 5 13E.00 240 325 16349 5.67" 57937 A vO t KICq Ioq(49) lqA+ rlog(KICI

IN A = -Mt4 tg ID, m.5)

Figure 9. Dynamic fatigue resultsfor CVD SiC.

23


33/54

Table 9. CVD SiC FlexureTest and Weibull Analysis Results

CVD SIC FOUR-POINTBENDFLEXURESPECIMENS.SIZE , 3rmm x 4mm x 50mm

DisplacementRate cm/min 5.08E-02 5.08E-01 5.08E-02 5.08E-03 5.08E-04

Test Envrionment Liq.N,. WaterWater Water Water

Test Purpose Inert Dyn.Fat Dyn.Fat Dyn.Fat Dyn.Fat

Weibull Modulus 6.77 5.53 5.26 5.90 5.08Welbull Scale Parameter MPa'm^(2/m) 97.46 67.52 59.78 74.31 56.05

Flexure Bar Median Strength MPo 360.3 332.9 319.4 331.9 317.2Average Strength MPa 355.3 327.8 315.2 327.2 312.3

Effective Surface Area. S' mA2 9.85E-05 1.02E-04 1.03E-04 1.01E-04 1 4E 4Average Stress Rate MPa/sec 5.13E+00 1.99E+02 1.99E+01 1.99E+00 1.99E-01

Numberof SpecimensTested 32 31 32 32 31Inspectionof Figure 9 indicates that the slope of the plotted line is very close to zero

which indicates that CVD SiCis not subject to slow crackgrowth . The resultsof this evaluation

indicate thatCVD SiC exhibits a negligible propensity forslow crack growth. Resultsofdynamic fatigueflexure testsshow that the stress ratehas very little effecton the median strength

of the four test groups.

PHYSICALPROPERTIES

Young's Modulus and Poisson's RatioThe resultsof the resonantfrequency room-temperatureYoung's modulus determin-

ations, obtained from eight specimens, taken from various locations (Figure1) within the bulk

material, are shown in Table 10. These results are typical of homogeneous, pore-freeCVDSiC. They are among thehighest values ever recorded for polycrystallineSiC.

Table 10. CVD SiC Young's Modulusas Determinedby Resonant FrequencyTechnique

Nearest Surface Locations Young's Modulus(see Figure 1) GPa (x10 6 psi)

Deposition Surface Location N 460.9 (66.9)Substrate Surface LocationN 461.5 (67.0)Deposition Surface Location 0 464.1 (67.4)Substrate Surface Location0 460.5 (66.8)

Deposition Surface LocationP 461.7 (67.0)Substrate Surface LocationP 462.0 (67.1)

Deposition Surface LocationQ 463.7 (67.3)Substrate Surface Location Q 461.6 (67.0)

24


34/54

Fracture Toughness

The fracture toughness(Kic) values for 10 specimens is given in Table 11. These values

are typical forCVD SiC material obtained from MI during the developmentprogram.

Table 11. Fracture Toughness (Kic) Results Obtainedby the Controlled Flaw Technique

Specimen Average Crack Fracture ToughnessNumber Extension gmi) (MPa m,/2)

A-1 87.5 2.5A-2 71.5 2.5A-3 83.0 2.2A-4 97.5 2.5A-5 67.5 2.1A-6 97.5 2.6A-7 77.0 2.4

A-8 63.0 2.3A-9 55.0 2.1

A-10 98.5 2.5I-I 81.8 2.81-2 97.0 2.61-4 90.0 2.5I-5 102.0 2.51-6 101.5 2.51-7 94.0 2.71-9 98.8 2.2

1-10 108.5 2.6

Microhardness

Results of the Vickers Microhardnesstests are found in Table 12.

Table 12. Microhardnessof CVD SiC

Test Temperature: 70F (21 C) Average 2.4Crosshead Speed: 8.4x10- 6m s Standard Dev. 0.2

(0.02 in/min)Relative Humidity:55%

IndenterLoad Used: 100 g S.Dc. No. Vickers Hardness(ka/mm)Al 2264A3 2083A7 224112 223118 2174

Avg. 2199

25


35/54

THERMALPROPERTIES

Coefficient of Thermal Expansion(CTEM

The resultsof the CTE analysis, overvarious temperatureranges, obtained from sixspecimens are given in Table 13. A least squares fit of the data for the six specimens is shownin Figure 10.

Table 13. CVD SiC Thermal Expansion(CTE)

Temperature Coefficient of Thermal Expansion(mrnmmm./C)Range Top - deposition surface; Bottom - substrate surface

(0C) R-Top R-Bottom S-Top S-Bottom T-Top T-Bottom

-150 to -100 0.53 0.55 0.44 0.64 0.42 0.43

-100 to -50 1.08 1.09 0.94 1.15 1.01 0.98

-50 to0 1.66 1.67 1.50 1.69 1.55 1.560 to 50 2.09 2.17 2.02 2.20 1.98 1.90

5o to 100 2.59 2.62 2.53 2.67 2.53 2.48

100 to 150 3.15 3.06 3.05 3.03 2.99 2.95

150 to 200 3.28 3.45 3.48 3.43 3.43 3.45

200 to 250 3.67 3.67 3.81 3.66 3.75 3.78

250 to 300 3.97 4.02 4.12 4.01 4.07 4.08

Heating Rate: 300C/hour

Atmosphere: AirReference: FusedSilica

Thermal Diffusivity

The thermal diffusivity valuesover the temperaturerange 21C to -175C is given inTable 14 for the six specimens measured. The reason for thelow values obtainedfrom the

U-Top specimenare unknown. The specimen was retested with the same results (withinexperimentalerror).

26


36/54

AVERAGE COEFFICIENT OF THERMAL EXPANSION RESULTSOF MORTON INTERNATIONAL CVD SIC

-. 4

E

E 3

2

c IU1

-100 0 100 200 30 0

TEMPERATURE ( C)

Figure 10. Least squares fit of CTE data.

Table 14. Thermal Diffusivity Obtained from Bulk Material (Top and Bottom), at Locations

U, V, and W.

Test Thermal Diffusivity(cm 2/sec)Temperature Top - deposition surface; Bottom - substrate surface

C U-Top U-Bottom V-Top V-Bottom W-Top W-Bottom

21 0.635 1.082 1.004 1.076 0.963 1.148

-25 0.787 1.529 1.321 1.404 1.339 1.455

-75 1.019 2.275 1.836 1.891 1.824 2.032

-125 1.384 3.175 2.449 2.664 2.386 2.893

-175 1.847 No Data 3.550 No Data 3.118 3.986Note: An Armco Iron standard was run before each set of tests. The ArmcoIron standard test resultswere adjusted,using a correction factor, to a thermal diffusivityof 0.202 cm 2/s. This factor was then used tocorrect the test data.

27


37/54

Specific eat

A plot of the specific heat from - 140'C to 600'C is shown in Figure 11. The plotteddata are the average values of three differentialscanning calorimetry (DSC) measurementsobtained from three specimens.

1.0

0.9

0.8

2. 0.7

w, 0.6

0.5.s

U 0.4

0.3

0.2

0.1-200 -100 0 100 200 300 400 500 600

TEMPERATURE (0 C)

Figure 11. Plot of the specific heat results for CVD SiC.

OPTICAL PROPERTIES

AbsorptionMeasurement Data Reduction

Measurement of optical absorptionwas accomplishedusing laser rate calorimetrywhich is the most sensitive and most accurate (15%) conventional technique. Precision ofthe measurement as related from sample to sample variations and measurementrepeatabilityis better than 0.0005 for the specimens tested.

Table 15 summarizesthe values of absorptance (Pa/Pr)measured. Each columnrepresents a remounting and re-measurementof the absorption.

28


38/54

Table 15. CVD SiC Optical Absorption Results for X = 10.6 gim

SpecimenLocation Absorptance (Pa/PR)

No. 10-1 10-1 10-' 10-1 Mean x 10-1

L-top SN 1 1.26 1.30 Aleonox Cleaned1.20 1.25 .05

L-bottom SN 2 1.28 1.29 1.28 1.28 .006

K-top SN 3 1.15 1.15 1.15 .0003

K-bottomSN 4 1.16 1.16 1.16 .0007

M-top SN 5 1.30 1.32 1.32 1.30 1.31 .01

M-bottom SN 6 1.29 1.28 1.25 1.27 .02

The absorption datafall into two statistically significant groups.Four specimens groupabout the 0.1278 0.025 absorptancevalue. The other two specimensaverage absorptance is0.1155 0.007. Repeated measurements confirmedthe accuracy of these values.

Reflectance Measurement Results

Reflectance of all samples was measured between 2,500to 25,000 nanometersusing a bare goldmirror for reference. This technique did not yield absolute reflectancevalues but did yield highlyaccurate relativereflectancevalues.

One representative sample fromthe absorptance groupis shown in Figure 12.No attemptwas made to analyze the data, however,there was indication of a slight broadeningof the reflectance peak in the 10.6 gim area for the samples with lower absorptance.

Total Integrated Scatter

Measurement of the surface finish of optical components is of considerableinterest in assessing the quality of the componentsand predicting their performance. The

measurement of total integrated scatter(TIS) provides a method of quantifyingsurface micro-roughness. TIS is related to the rms surface roughness for whosetopographysurfaces satisfy

certain statisticalcriteria with respect to the randomnessof the amplitudesand the spatialfrequency distribution of defects. Table 16 summarizes the results of the measurementsmade

on the six silicon carbidespecimens identified forthe TIS measurementsand calculated values

of hrms (Figure 1).

Janos Corp. reportedsurface roughness, measuredby contact profilometry,to be better than 3A rms onall six specimens.

29


39/54

Table 16. CVD SiC Total Integrated Scatter (TIS) Results for X = 0.6328 gtm

Specimen Number TIS (x10 4 ) hrms (A)

SN-1 6.10 12.2

SN-2 8.17 14.4

SN-3 4.92 11.1

SN-4 4.14 10.2

SN-5 4.01 10.1

SN-6 9.91 15.7

SIJI.CON CARBIUE SN t VS. BORF ,OID

- X 10601.761 NM Y 83.1573 % R

In

LO

_lu

z

-iuLfl

LLLl

0F)

0

500 5000 7500 10000 12500 15000 17900 20000 22500 25000NANOMETERS

Figure 12. Typical reflectivity curve for CVD SiC.

30


40/54

Raman MicroprobeMeasurements

Three series of spectral measurementsof the fuil width at half maximum(FWHM) of the 795 cm-1 Raman line were made to determine residual stress in the specimen.Figure 13 is a schematic of the Raman specimen indicating its orientation in the initial CVD

material and the processing of the surfaces. The first series shownin Figure 14 (i.e., thechamfer) was across the side (planeof growth direction) from the deposition (top) surfacein 10 jim steps, down the bevel and onto the side of the specimen towardsthe ,ubstrate surface(bottom). In the plane of the side, the bevel was determined to be 180 pm wide. The secondseries given in Figure 15 was again across the side of the specimen from the top to the bottom,except that the first four steps wereat 50 gtm intervals, the fifth step 400 gim from the top andthe remaining foursteps at 1000 ptm (i.e., 1mm) intervals. At the first six positions fromthetop, two measurementswere made at different lateral locationsto characterize the variations at

the same transverse position. The third series across the end of the specimen, also in Figure15, consisted of the first step at the top surface, two steps at 50 Pm intervals followedby astep at 500 microns, with the remaining four steps at 1000 micron intervals towards the bottomsurface.

TOP (Deposition surface)Ground & Polished

GrwthIIDlrctlon

END

(Plane of Deposition)CutIDE Cu

Cut & GroundBOTTOM(Substrate surface)

Ground

Figure 13. Diagram of Raman microprobe specimen.

31


41/54

350 . .

0 DATA- FIT

300

w0 250D

< 200

15 0 . . . ... . . . . . . . . . . I . . . . . . . .780 790 800 810

WAVENUMBER v (cm-1)

LORENTZIAN FIT: Amp-=188*1 . 0,0 -795.15*0.002 . FWHM-3.34* 1.9%. Bock- 158.7-+0.4%

Figure 14. Raman spectralline broadeningdue to residual strain on the sideof the specimenthat had been cut,ground, and beveled (i.e., chamfered).

350 0 DATA- FIT

:3o 300

wh

-j 250

200

780 790 800 810

WAVENUMBER v (cm- 1)

LORENTZIAN FIT: Amp- 149*0.9%. &' -795.83*0.003 . FWHM-4.68*2.1%, Bock- 197.7*0.4%

Figure 15. Raman line broadening due to residual strainon end and side of specimen.

32


42/54

Data reduction was accomplished by a least squares fit of the Lorentzian lineshape function. For each data set, the peak height above background, the line center position,the FWHM, and the background level were fit as free parameters. Three examples of the spectraare included (Figures 16 through 18), the solid curve is the fitted function and the data are given

400 , ,

0 DATA350 - FIT

300

0w

O 250

: 200

150

780 790 800 810WAVENUMBER v (cm-1)

LORENTZIANFIT: Amp=,227*O.9%, vo-794.95*0.002-%, FWHM-3.23*1l.5X, Bock-135.3*0.4%

Figure 16. Raman scan on the bevel of the cut and ground side 150 pRm from the growth surface.

........... i......... i......... i......... ....

o CUT EN D

,, e CUT & GROUND SIDE

E4.0

W 3.5z

z


43/54

0 DIAMOND CUT &S4.5 GROUND SIDETT

E

4.0

UJ 3.5z-i

2 3.0BEVEL

0 50 100 150 200 250

DISTANCE FROM GROWTH SURFACE pu,m)

Figure 18. Raman scan on the diamond cut end 3500 pgm from the growth surface of the specimen.

by the open circles. Fitting uncertainties in the widths (see FWMH on the graphs) were typically

1.5 to 2.0 percent. These uncertainties are plotted as error bars on the graphs.

In materials where the optical penetration depth is small, typically _ 1 mm, the

width of a Raman line is a sensitive indicator of the relative amount of residual stress in the near

surface region. This width depends on the long range order of the crystal or crystallites being

probed. In general, cleaved, single crystal surfaces will show the smallest width.

Surface processing, such as grinding or polishing, introduce localized strains

which reduce the long range ordering of the local crystalline environment. This loss of orderbroadens the Raman line. For example, polishing with a coarse grit will broaden the line more

than polishing with a fine grit. Thus, the broader the line, the greater the relative residual stress.

Since the specimens studied here are polycrystalline, the residual stress can vary

from grain-to-grain, particularly in ground surfaces. This is clearly seen in Figures 14 and 15.

In contrast, the data in Figure 15 for the cut end (no grinding or polishing) shows a constant

width over more than 4 mm suggesting a uniform but small or zero residual stress as might be

34


44/54

expected. The data confirmthe high precisionof our results. Moreove r,the cut end data

strongly supportsthe conclusion that the bulk materialhas very small or zero strain gradientswhich suggests no bulkstrains or nearly so.

A CVD SiC polished specimen surface (-3A rms roughness) and its ground

beveled (i.e., chamfered)edge were also scanned. Figure 19 shows a comparisonof the Ramanline broadening along the mirror surface and down along the side (or edge) of the mirror. Notethe high residual strain in the polishedsurface versus the ground bevel. The figure shows clear

evidence of a strain gradientover a distance of 0.3-0.4 mm in the beveled edge region. From

there thewidth on the polished face are nearly uniform and significantly smallerthan for the cut

end data in Figure 15. This is consistent with the fineness of the final polishing stepnecessaryto fabricate the mirrorsurface. The small deviationsof the widthare outside the fittinguncertainties and therefore may indicate some smallnonuniform surface strainsin the polished

face. The data on the side of the mirror specimen showsthe typical variations expected foraground surface, but with somewhat smaller peak-to-peak variations than shownin Figures 14

and 15.4.0............................ .............

0 POUSHED FACE0 GROUND BEVEL SIDE

E 3.5

I-

z

- 2.5

0 500 1000 1500 2000 2500 3000

POSITION FROM EDGE ,um)BEVELIS ABOUT310 pm WIDE. MEASUREMENTSON THE MIRROR AREALONG A RADIUS FROM OUTER EDGE OF BEVELON THE SWE THEY ARE ALONG A UNE PERPENDICULARTO THE POUSHED FACE STARING BELOW THE BEVEL

Figure 19. Raman spectralline broadeningdue to residual strain on polished surfaceof the mirror specimen(surface roughness -3A rms) and theground and beveledside (oredge).

35


45/54

CONCLUSIONS

From the test resultsobtained in this characterizationstudy of CVD SiC,it can be

concluded that:

"* the material is chemicallyand crystallographicallyhomogeneous;"* due to the longdeposition run times required to depositthe material usedin this studyand

esulting larger grainsize, the averagemechanical strengthwas slightly lower thanthe:ngth of material (-6 mm thick) fromshorter depositionruns measured in previous

studies;"* there is no statisticallysignificantdifferences in mechanical strength when comparing

material obtainedfrom various locationswithin the bulkmaterial;"* the thermal andoptical properties were consistent for material taken from various

locationswithin the bulk, further confirmingthe homogeneityof the material; and

" Raman microprobe spectroscopy indicated that there is no detectable residual strain

gradients in the bulkSiC as deposited,however, detectablestrain is present after grindingand polishing of the material.

ACKNOWLEDGMENTS

We thank Mr. William Kaveney and Mr. Donald Luczakof Rome Laboratoryfor

their support in arranging forthe funding toallow this program and their encouragementthroughout the program. We thank Ms. MaryDursi for aiding us in obtaining the subcontract toaccomplish the work through her organization, ScientificApplications InternationalCorporation.

We thank Mr. D. Wolf for the XRD work; Dr. P. Yaney for the Raman microprobe studies;Mr. S. Goodrich and Mr. D. McCullum forthe mechanical propertytesting; Mr. D. Grant forthe microstructural work; andespeciallyMr. S. Hilton for the preparation and arduous taskof

identifying and trackingthe nearly two hundred specimens usedin the study.

36


46/54

REFERENCES

1. Weibull, W., A Statistical DistributionFunction of Broad Applicability, J. Appl. Mech.

1B(3), 1951, pp. 293-297.

2. Iden, D. J., Detrio, J. A., and Fox, J. A., Optical Window Mechanical Strength and

Reliability, AFWAL-TR-83-4059,June 1983, Contract F33615-80-C-5137.

3. Graves, G. A., McCullum, D. E., and Wimmer, J. M., Investigation of the Thermal,

Electrical,Mechanical, and Physical Propertiesof Infrared Laser Window and IR

Transmitting Materials, AFML-TR-77-23,April 1977, Contract F33615-75-C-5011.

4. Lipson, C. and Sheth, N. J., Statistical Design and Analysis of EngineeringExperiments,

(McGraw Hill Book Co., NY, 1973).

5 Evans, A. G., Fracture Mechanics Determinations, in Fracture Mechanics of Ceramics,

Vol.2, ed. Bradt,R.C. (Plenum Press, NY, 1974).

6. Ritter, J. E., EngineeringDesign and Fatigue Failure of Brittle Materials, in Fracture

Mechanics of Ceramics, Vol. 4, ed. Bradt, R.C. (Plenum Press, NY, 1978).

7. Smyth, K. K. and Magida, M. B., Dynamic Fatigue of a Machinable Glass Ceramic,

J. Amer. Cer. Soc., 66(7), July 1983, pp. 500-505.

37


47/54

DISTRIBUTION LIST

addresses numberof cop ies

ROME LABORATORY/OCPC SATTN: RICHARD FEDORS

Z5 ELECTRONIC PKYGRIFFISS AFa NY 13441-4515

UNIVERSITY OF DAYTON 5RESEARCH INSTITUTEDAYTON, OH 45669-0172

RLISUL ITECHNICAL LIBRARY26 ELECTRONIC PLYGRIFFISS AFB My 13441-4514

ADNXNISTRATOK 2DEFENSE TECHNICAL INFO CENTERDTIC-FDACCAMERON STATION BUILDING 5ALEXANDRIA VA 22304-6145

BALLISTIC MISSILE DEFENSE 2

ORGANIZATION1100 DEFENSE PENTAGONWASH DC 20301-7100

NAVAL WARFARE ASSESSMENT CENTER 1GIDEP OPERATIONS CENTER/CODE QA-50ATTN: E RICHARDSCORONA CA 91718-5000

ASC/ENENS 1WRIGHT-PATTERSON AFB OH 45433-6503.

AFIT/LOEE 1Z950 P STREETWRIGHT-PATTERSON AFB OH 45433-6577

DL- I


48/54

WRIGHT LABORATORY/MLPOATTN: U.L. DENISONWRIGHT-PAtTERSON AFB OH 4 5 4 3 3 - 6 S 3 3

WRIGHT LABORATORY/MTELWRIGHT-PATTERSON AFS OH 45433

AUL/LSE IBLOG 1405MAXWELL AFB AL 3 6 11 2 - 5 5 6 4

COMMANDING OFFICER INAVAL AVIONICS CENTERLIBRARY 0D765INDIANAPOLIS IN 46219-2189

COMMANDING OFFICER INCCOSC ROTE DIVISION

CODE 02748, TECH LIBRARY53560 HULL STREETSAN DIEGO CA 92152-5001

CDOR 1NAVAL WEAPONS CENTERTECHNICAL LIBRARY/C3431CHINA LAKE CA 9355S-6001

CDR, U.S. ARMY MISSILE COMMAND 2REDSTONE SCIENTIFIC INFO CENTERAMSMI-RD-CS-R/ILL DOCUMENTSREDSTONE ARSENAL AL 35898-5241

ADVISORY GROUP ON ELECTRON DEVICES 2ATTN: DOCUMENTS2011 CRYSTAL ORIVESUITE 307ARLINGTON VA 22202

LOS ALAMOS NATIONAL LABORATORY IREPORT LIBRARY"MS 5000LOS ALAMOS NM 87544

OL-2


49/54

COMMANDER/USAISCATTN: ASOP-OO-TLSLOG 61801FT HUACHUCA AZ 85613-5000

SOFTWARE ENGINEERING INST SEI)TECHNICAL LIBRARY5000 FORBES AVEPITTSBURGH PA 15213

NSAE323/HCSAB2 DOOR 22FORT MEADE NO 21055-6000

NSAATTN: D. ALLEYDiV X9119800 SAVAGE ROADFT MEAOE NO 20755-6000

DOD

R319800 SAVAGE ROADFT. MEADE MD 20755-6000

DIRNSAR5099800 SAVAGE ROADFT MEADE MD 20775

DEFENSE TECHNOLOGY SEC ADMIN (DTSA)ATTN: STTO/PATRICK SULLIVAN400 ARMY NAVY DRIVESUITE 300ARLINGTON VA 22202

OC TECHNICAL LIBRARY

ROME LABORATORY/OC26 ELECTRONIC PKIGRIFFISS AFB MY 1 3 4 4 1 - 4 5 1 4

ROME LABORATORY/OC26 ELECTRONIC PKYGRIFFISS AFB MY 1 3 4 4 1 - 4 5 1 4

OL-3


50/54

LITTON STS INC/ITEK OPTICAL STS DIV 2ATTN: HART R. LATHAN/SR. LIBRARIANFOR: MR TOM PzrTS10 MAGUIRE ROADLEXINGTON, MA 02173-3199

LOCKHEED CORP/TECHNICAL INFO CT RATTN: SUZANNA STANEK, TECH LBFOR: MU -FA CHEN, ( 0 /6 6 -60 , B586W)0/90 , OCS-L1O2, BLDG 106/BOX 3504SUNNYVALE, CA 94088-3504

MARTIN MARIETTA CORP/AEROSPACEATTN: JACKIE 0. BUNTINGDIR: TECHNOLOGY DEVELOPMENTP.O Box 17 9DENVER, CO 0201

SMDO/TNSATTN: ERWIN NYRICK7100 DEFENSE PENTAGONWASHINGTON OoC. 20301-7100

TRW SPACE E DEFENSEATTN: JERRY F. SAO/MARIE MORANSENIOR TECHNICAL LIBRARIANONE SPACE PARKREDONDO BEACH, CA 90278

GRUMMAN AEROSPACE CORP.ATTN: EDWARD LEEN/S 0 - 0 1 - 2 31111 STEWART AVENUEBETHPAGE NY 11714-3541

TINSLEY LABORATORIES INC.ATTN: ROBERT JOHNSONV.P. MARKETING3900 LAKESIDE DRIVERICHMOND, CA 94806

W. J. SCHAFER ASSOCIATES 3

ATTN: GATLE FEOLE/SECURITTFOR: RICHARD DYERZOO LIBERTY PLAZAROME, NT 13440

THE AEROSPACE CORP./LIB ACO GROUPATTN: P V GREENFOR: DR VIRENDRA NAHAJAN M5643P.O. BoX 92957LOS ANGELES, CA 90009-Z957

DL-4


51/54

DR MICHAEL PICKERINGMORTON INTERNATIONAL/CVD INC185 NEW BOSTON STWOBURN MA 01801-6203

DR GEORGE GRAVESUNIVERSITY OF DAYTON RESEARCH300 COLLFGE PARK KL-165DAYTON OH 45469-0172

HUGHES AIRCRAFT COATTN: 0. WEBaFOR: OR ROGER WITHERINGTONP.O. BOX 902 El/E110EL SEGUNDO* CA 90245

EASTMAN KODAK CO/GOVERNMENT SY S

ATTN: AUDA ROBERTFOR: MR DON GILONERPu BOX 24939

ROCHESTER MY 14624

SALL AEROSPACE SISTEMS DIVISION

ATTN: MR PETER 4AL.c/SECURlTYFOR: MR E. A. ROYBA-.P.O. BOX 1062BOULDER, CO 80306-1062

NATIONAL AERONAUTICS C SPACE ADMINOPTICAL SYSTEMS BRANCHATTN: MR ROBERT ROOD M S : 8

MARSHALL SPACE FLIGHT CENTER, AL35812

ROCKWELL INTERNATIONAL CORP/ATTN: MR DONALD J. CARLSONP.O. BOX 31053370 MIRALOMA AVEANAHEIM CA 92803-3105

DONALD LUCZAK

ROME LABORATORY/OCPC25 ELECTRONIC PKYGRIFFISS AF8 NY 13441-4515

DR GEORGE RANSEYER525 BROOKS RDGRIFFISS AFB NY 1 3 4 4 1 - 4 5 0 5

OL-5


52/54

OR YOLANDA KINGPL/XPPABERDEEN AVE SEKIRTLANO AF 1 NM 8 7 11 7 - 5 7 7 6

COL LANNY LARSONPL/LI3533 ARERDEEN AVE SEKIRTLAND AFa NM 8 7 11 7 - 5 7 7 6

LW BRANTLEYNASA MSFCPS 05MARSHALL SPACE FLIGHT CTR AL 35812

GORDON AUGASONNASA AMES RESEARCH CENTERN-245-6

MOFFETT FIELD CA 9 4 0 3 5 - 1 0 0 0

JOEL SHADYUS ARMY SOC/CSSD-KA-RH106 WYNN DRIVEHUNTSVILL AL 35807

GEORGE SLOANUS ARMY SDC/CSSO-SD-O

106 WYNN DRIVtHUNTSVILLE AL 35807

CAPT WILLIAM CAMERONWL/MLBC2)41 P ST STE 1WRIGHT-PATTERSON AF8 OH 4 5 4 3 3 - 7 7 5 0

OR RON KERANSWL/MLLMZ230 TENTH ST STE 1WRIGHT-PATTERSON AFS OH 4 5 4 3 3 - 7 7 5 0

JIM WELLS MS-160TELEDYNE BROWN ENGINEERING300 SPARKMAN OR N4HUNTSVILLE AL 35807

DL-6


53/54

OR HAROLD SENNETTNAWCCODE C023103

CHINA LAKE CA 93555

OR DONALD DECKERNAWCCODE C02316CHINA LAKE CA 93555

KIM VUSC/MGSS185 DISCOVERER BLVD STE 1315LOS ANGELES AF9 CA 90245-4695

LTC AYOELOTTESMC/MGSSl aS DISCJVERER BLVD STF 1315LOS ANGELES AF8 CA 90245-4695

OR CHIA SHIH LEESMC/MGSS

185 OISCOVER-R SLVO STE 1315LOS ANGELES AF9 CA 9 0 2 4 5 - 4 6 9 5

LTC SOLTSMC/MGSE

185 DISCqVERER BLVDLOS ANGELES AFB CA 90245-4695

ERIC HANSENGEMINI 8-M TELESCOPES PROJECT

950 NORTH CHERRY AVETUCSON AZ 85726-6732

*UA. GOVERNMENTPRINTING OFFICE: 1 9 9 4- 5 10-117-50032

DL-7


54/54

MISSION

OF

ROME LABORA TOR Y

Mission. The missionof Rome Laboratory is to advance the science andtechnologies of command, control, communicationsand intelligenceand totransition them into systems to meet customer needs. To achieve this,Rome Lab:

a. Conducts vigorous research, development and test programs in allapplicable technologies;

b. Transitions technologyto current and future systems to improveoperational capability, readiness, and supportability;

c. Provides a full range of technical support to Air Force MaterielCommand product centers and other Air Force organizations;

d. Promotes transfer of technology to the private sector;

e.Maintains leading

edgetechnologicalexpertise

in theareas ofsurveillance,communications,command and control, intelligence, reliability

science, electro-magnetic technology, photonics, signal processing, andcomputational science.

The thrust areas of technical competence include: Surveillance,Communications,Command and Control, Intelligence, Signal Processing,Computer Science and Technology, ElectromagneticTechnology,Photonics and Reliability Sciences.

Cvd Silicon Carbide Characterization (1994)

Documents

Transcript of Cvd Silicon Carbide Characterization (1994)