-~0

r ~

:I

MTL TR 91-45[AD JAD-A243 66S,

EX,",MINING Si3N4 BASE .MA,4'.TER;ALS

WIATH VARIOUS RARE EARTH

ADDITIONS

GEORGE E. GAZZACi:,cAMICS RESEARCH BRANCH

December 1991

Approved for public' release; distribution unlimited.

O00

u,•.,000 F o o/ 0,9I0~US ARMY'LAORATORY COMMAND U.S. ARMY MATERIALS TECHNOLOGY LABORATORY

1fiMa.s TwoJ'a un"Timv Watertown, Mamchusetts 02172-0001

The tintiIgs in this report are riot to be construed as an officialOelpa'tnmnt of the Army position. uniso so desiqna.di by otherautloruted doe-jer.•n

Mention of Ma de names or 'manufacTurs in this report"*IalI not be construed as adv rtising nor "s an ,officiatindorsement or aoprovr4 of sucl products or companies byhe Liited Sates Governnment.

OISPOSITION 1ST RUCION3

this, tr,,, fugaf 1 to m~qe I* gned".00 not rFtu'it t eo u I mnavn

UNCLASSIFIEDSECURITY CLASSIFICATION O0 THIS rAGE .r.,n Data Enterd)

RED NSTRUCTONS 1REPOT DOUMETA-i10N AGEBEFORE COMPLETING FO)RM

L OT REPORT DOCUMENTA. iON PAGE •oSTV OS

1. TNUMER 2. GOVT ACCZSSION NO 3. RECIPIENT'S CATALOG NUM1ER

MT. TR 91-45

4. TITLE (and Subflttl) S. TYPE OF REPORT & PERIOO COVEREO

Final ReportEXAMINING Si3N4 BASE MATERIALS WITH VARIOUSRARE EARTH ADDITIONS 4. PERFORMING ORG. Re ORT NUMOER

7. AUTHOR(s) 6. CONTRACT OR GRANT NUMOER(s)

George E. Gazza

9. PERFORMING ORGANIZATION NAME ANO ADDRESS 10. PROGRAM ELEMENT. PROJECT, TASK

U.S. Army Materials Technology Laboratory ARCA & WORK UNIT NUMSERS

Watertown, Massachusetts 02172-0001 D/A Project: 1L162105AH84ATTM: SLCMT-EMC

II. CONTROLL1NG O•FICE NAME ANO AOORESS 12. REJPRT DATEU.S. Army Laboratory Comand December 19912800 Powder Mill Road 13. wuMEIROF PAGES

,.•Adephi, Maryland 10783-1145 914. MONITORING AGENCY NAME & ADORES5(it di•,ie.e. trai Cor-i-l,,itn Olifi) I5. SECURITY CLASS. (of ti• r 'epo.r)

Unclassified1e5. OECL ASSI FICATION/OOWNGRAOING

SCHEDULE

14. DISTRIBUTION STATEMENT (of rt:,.i !:eport)

Approved for public release; distribution unlimited.

17. OIST.IIUTION STATEMIENT (of the abstra• t mne•..d In Block 20. i t ddioevaPre Rebate)

1I. SUPPLEMENTARY NOTES/

Presented at the American Ceramic Society, Cincinnati, Ohio, May 1991.

I9. KEY WORDS (L oa,. reves iei oosr n identIty 6b' blc cilnatilSilicin nit-ride Micros truc ture CreepHot pres3ing Hardness StrengthRare earths Oxidation

20. ASSTRACT (Coet..e , Powers*.. aid, i° nocess. a" Ident.,y by bla o U0,•,

(SEE REVERSE SIDE)

FOMDD I j..', 1473 E O,1 hN, NO, , .osS 0OS T 'UNCLASSIFIED

qr l~lrveýýA-smotAf~ 40T 4 &%I %= f .71t-

UNCLASSIFIEDSECURITY CLASSIFICATION Of 'WIS PAGE (.,on Data FM.Vd,

Block No. 20

ABSTRACT

Silicon nitride base compositions were prepared by using various rareearth oxide addit.ves and -hot pressing the powders to full density. Composi-tions were primarily explored in Si 3 N4 -SiO 2 -RE2 Si 2 0 7 compatibility regions

,although other Si 3 N4 -RE 20 3 reactions were also investigated. Hot pressingdensification data and high temperature behavior of the various silicon nitride-rare earth compositional systems are examined.

UNCLASSIFIED

CONTENTS

Page

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

EXPERIMENTAL ............ ......... ............... 2

MECHANICAL BEHAVIOR .............. ...... ......... 5

CREEP MEASUREME1rT.S ........ ................... .... 10

OXIDATION ...................... ................. .. 15

SUMMARY AND CONCLUSIONS ....... ................ ... 16

ACKNOWLEDGMENT ....... ............ .......... .. 16

2

• •e ,,a. It !cat I

D isi ... a....;Av~tlab~lll , CodlJ

,.ail ,n/s

INTRODUCTION

The use of various rare earth oxides as additives' to promote sintering ofsilicon nitride is receiving greater attention in recent years.I- 5 In part, arenewed effort is directed toward determining whether a rare earth other than Y2 0 3can be used to densify Si 3 N4 and avoid some of the problems associated with Y2 0 3 -doped Si 3 N4 material. ,7 In addition, studies are focusing on the use of multiplerare earth oxide additions to produce duplex microstructures with high fracturetoughness values. 4,8

Certain material characteristics and use of preferred compositions are con-sidered in selecting specific rare earth oxides for study as additives to Si 3 N4 .Some of these characteristics are listed in Table 1. The formation of a rare earthpyrosilicate phase (RE 2 Si 2 O7 ) at the grain boundaries is often chosen to avoidmaterial instability problems at intermediate temperatures (700 0 C to 0000C) whichare caused by the oxidation of certain phases formed by Si 3 N4-rare earth reactions.Crystallization of the pyrosilicate by asuitable heat treatment after densifica-tion of the material reduces creep and oxidation at elevated temperatures. Thehigh temperature properties of the particular pyrosilicate phase formed are relatedto its melting point (refractoriness), the temperature at which it forms a eutecticwith SiO2 , its crystallization behavior (degree of crystallization and crystallitesize), and the thermal expansion coefficient of the boundary phase compared to thesilicon nitride grains. The degree of anisotropy associated with this phase will

Table 1. RARE EARTH ADDITIVESISILICATES.SOME CHARACTERISTICS CONSIDERED

1. Ionic radius of the rare earth (RE),

2. Melting point of RE2Si20 7,

3. Eutectid melting poimt of RE2 Si2O7 -SiO2,

4. Lowest eutectic in the Si3 N4 -RE2 0 3 -SiO2 system,

5. Polymorphism of RE2 Si2 O7,

6. Crystallization behavior,

7. Thermal expansion coefficient; anisotropy and

8. Valency; oxidation state

1. HIROSAKI, N., OKADA, A., and MATOBA. K. Sintering of Si_'N'4 with the AJdition of Rare Earth Oxides. Communications of the Amer.Ceram. Soc., C-144-147, March 1988.

2. SANDERS, W. A., and MEESKOWSKI, D. M.. Strength and Microst'ructure of Sij 4 with Rare Eart Oxide Additions. kmer. Ceram. Soc.BulL, v. 64, 1985, p. 304.

3. TANI. E., NISHLIIMA, M., ICHLNOSE. H.. KISII, K., and UMEBAYASH!, S. Gas Presurw Sintering of Si.rV4 with an Oxide Addition.Yogyo-Kyokri-Shl, v. 94, no. 2, 1986, p. 300-305.

4: TANI, E., UMEBAYASHI. S., KISHI, K., KOBAYASHI, K., and NISHIJIMA, M. 0 as.Presunre Sinterng of SiyV4 with Concurrent Additionof A12 0 3 and 5 wr% Rare Earth Oxide: High Fracture Toughness S•yV 4 with Fiber.Like Struciue. Amer. Ceram. Soc. BulL., v. 65, no. 9,1986, p. 1311-1315.

5. HAMPSHIRE, S., LEIGH, M., MORRISSEY, V. J., POMEROY, M. J., and SARUHAN, B. Crystallization Heat Treatments of Silicon ,VitrdeCeramics and Glast.Ceramics Containing Neodvmia. Proc. 3rd Internat. Symp. on Ceramic Materials and Components for Engines, Las Veras.NV, November 1988, p. 4324-42.

6. PATEL,'J. K., and THOMPSON. D. P. The Low Temperarare Oxidation Problem iw Ytrrfa.Densijed Silico- .Vitrde Ceramics. Br. Ceram.Trans. 3., v. 87, 1988, p. 70-73.'

7. LANGE. P. F., SINGHAL, S. C., and KUZNICKI, R. C. Phase Rela.,ions and Stability Studies in the S1iJi4.SiO•Y'O3 PreudoternarySystemL ]: Amer. Ceram. Soc., v. 60, 1977, p. 249-252.

8. 11, C. W.,YAMANIS. J., WHALEN, P. J., GASDASIA, C. ]., and BALLARD. C' P. In Situ Reinforced'Sificon ,VWiide Ceramics. The'37th Sagamore Asrmy Materials Reseauch Conference Proceedings: Structural Ceramics. Plymouth. MA. 1990. p. 284.-293.

II

also influence properti-s. Other criteria used for the selection of a particularrare earth are its characteristic valency (single or variable valency) or the ionicradius of the cation. In the rare earth series, the ionic radius of the cationdecreases (also the size of the unit cell) with increasing atomic number. This isknown as the lanthanide contraction.

In a study conducted by Andersson and Bratton, 9 the influence of this effecton the high temperature strength (modulus of rupture) of hot-pressed silicon nitridecompositions doped with various rare earth additives was examined. Standard compo-sitions contained approximately 6.5 m/o rare earth oxide and 13.0 m/o SiO2 . Rareearths used in the study were Yb2 03 , Er 20 3 , Sc2 03, Y2 03 , La 2 03 , CeO 2 , Pr 2 03, Nd2 03 ,Sm203, Gd203, and DY20 3 . They found that the high temperature strength increasedwith decreasing ionic radius of the rare earth additive and suggested that corres-ponding decreases in the bond length of the silicates formed may be responsible forimproving the high temperature properties.

Since it was of interest to determine whether this effect would influence thehigh temperature time-dependent properties (creep),or other properties, such asstrength and hardness, the present study was conducted using rare earths withrelatively small ionic radii' as additives to determine both the room temperatureand high temperature (1200 0 C to 14000C) properties of the resultant materials.

EXPERIMENTAL

The silicon nitride starting powder selected for use in the study was TOSOHTS-lO. The powder is approximately 95% to 96% alpha phase. Rare earths obtainedfor use as additives were Sc 2 03 , Lu 203, Yb2 03 , Er 20 3 , and DY20 3. Chemical analysesof the starting materials and their 'supply source are shown in Table 2. The-puritylevel of the additives is essentially 99.9%.

The additives were milled with the silicon nitride powder and a small amountof silica in a nylon jar with hot isostatically pressed Si 3 N4 milling balls inisopropanol for 18 hours. Compositions with each rare earth additive were preparedto produce a standard composition of 90.0 m/o Si 3 N4 -3.0 m/o rare earth oxide-7.0m/o SiO2 ., This composition was located on'the SiO2 rich side of the Si 3 N4 -Y 2 Si 2 07tie line in the Si 3 N4 -Y 2 O3-SiO2 .phase diagram. The milled' powder mixtures weredried and hot'pressed in graphite dies, at 25 MPa uniaxial pressure, under 0.1 MPanitrogen, for 120 minutes at final hot pressing temperatures within the 1790 0 C to1805 0 C range. Bars 3 = x 4 rA x 46 mm were machined from hot-pressed discsapproximately 63.5 mm diameter x 8.0 to 9.0 mm thick. Immersion densities weredetermined with the machined bars for each Si3N4 -rare earth additive composition.The mean values are listed in Table 3.

Pieces from machined bars were powdered 'and analyzed by XRD.. The phase compo-sitions formed with'the Er 2 03, Lu 20 3 , and Yb2 03 additives were comprised of beta-Si 3 N4 , Si 2 N20, and the pyrosilicate of the respective additive used, i.e.,Er 2 Si 20'7, Lu2 Si 2 07 , or Yb2 Si 2 07 . A small amount of alpha-Si 3 N4 also appeared to hepresent in the Er 20 3 and Lu 2Q3-doped materials. The DY2 03 -doped Si 3 N4 may have hada glassy grain boundary phase in thatthe XRD results did not indicate the presenceof crystalline phases other than beta-Si 3 N4 and Si 2 N2 0.

9. ANDERSSON, C. A.. and BRATTON, R. Cermic .faterwiz for High Tempemturf Turbinet U.S. En•rgy' Res Dev. Xam. ContrcEy-76-C-05,S210, Final Report, 1977.

2

Table 2. CHEMICAL ANALYSFS OF STARTING MATERIALS (Wt%)(SUPPLIERS ANALYSIS)

Rare Earth Oxides (99.9%)

Si3N4TOSOM TS.10 Dy2O3 Yb2O�" Er2O3t bj2Q3t Sc203�

N-38.0 Gd-0.04 Gd-0.03 Ho-0.Q1 Yb-0.05 Yb-0.0'I

0.1.0 Ho-0.01 Tb-0.01 Si-0.01 Si-0.01 Y-0.01

C-O.06 V.0.03 Dy-0.0i Fe-0.01 F.-0.01 Ca-0.005

CI-0.03 V.0.03 Mg-0.01 Mg-0.01

Fe-0.01 Ca-0.01 Ca-0.01

AI-0.01 AJ-0.01

Malycorp lnr�., Louviers, COtRes�rch Chemicals, Phoenix, AZ.Johnson Matthey Alpha, Ward Hill, MA

Table 3. MATERIAL COMPOSITION, HOT-PHESSED DENSITY,AND DENSIFICATION PARAMETERS

DensityComposition (g(cc)

90 rn/a Si3N4 + 3.0 m/o Dy203 + 7.0 rn/o 5i02 3.320

90 mlo Si3N4 + 3.0 rn/a Er2O3 + 7.0 rn/a 5i02 3*34$

90 rn/a Si3N4 * 3.0 rn/a Yb2O3 + 7.0 rn/a 5i02

90 m/o Si3N4 �- 3.0 rn/a Lu2O3 . 7.0 m/o 5i02 3.357

90 ri/a Si3N4 + 3.0 in/c 5c203 + 7.0 m/o 5i02 . 3.195

Densification Parameters

Applied Pressure - 25 MPa

Final Hot Pressing Temperature Range- 1790C to 1SOSOC

Hot Pressing Time at Final Temperature- 120 Minutes

"The densification behavior of che powders was studied by monitoring theshrinkage during the hot pressing run. In particular, for each composition, thetemperatures at which the onset of rapid shrinkage occurs were noted as thesetemperatures should approximate when liquid first forms in the reacting powdermixture, that is, when the temperature approaches that of the lowest eutectic inthe compositional system. These temperatures are listed in Table 4 'along withthe incipient melting points (lowest melting eutectics) found by Andersson et al. 9

by heating compacted powder mixtures of Si 3 N4 , SiO2 , and varidus rare earth oxides.Where melting was not observed up to the temperature shown in the table, theincipient melting points were indicated to be greater than this. t.mperature.Also shown are temperatures reported for the melting point of the various pyro-silicate phases for rare earths used in the study and melting points of thepyrosilicate-SiO2 eutectic. Reasonable agreement appears to exist between theincipient melting points found by Andersson et al. 9 for their Si 3 N4 -SiO 2 rareearth oxide mixtures and the temperatures ac which the onset of rapid shrinkageoccurs for the corresponding Si 3 N4 -SiO2 rare earth oxide powders hot pressed inthis study.

Tab.e 4. COMPARISON OF MELTING POINT TEMPERATURES WITH ONSET OFSHRINKAGE TEMPERATURES DURING HOT PRESSING OF Si3N4 -RE20 3 -SiO2

COMPOSITIONAL SYSTEMS

Melting Point of the Est. Lowest Eut. Temp. OnsetMelting Point of RE2Si2 0 in the Si3 N4 of Shrinkage

Rare Earth RE2 Si2 O7* Eutecticl RE2 Si2 03 -SiO 2 System During Hot PressingtAdditive (0C) (a) (*Q (OC)

DY203 1720 1640 155011570 1565-1575(I)

Er20 3 1800 1680 >1625 1630-1640(C)

SYb2 03 1850 1650 >1565 1595-1605(C)

Lu20 3 1850 -- 1675-1685

Sc20 3 1850 1660 >1675 1725-1735(C)

Y20 3 1775 1660 1580 1570-1580(I)

*Data from Handbook of Phase Diagrams of Silicate Systems, v. 1, N. A. Toropov, V.' P. Barzakovskii,V. V. Lapin, N. N. Kurtseva. U.S. Department of Commerce, Nati. Tech. Information Service,Springfield, VA 22151

tOata from Ceramic Materials for High Temperature Twbines, C. A. Andersson and R. J. Bratton,Final Report. ERDA E-76-C-05-5210, 1977

tHeating RatebDuring Hot Pressing - 10C!Minute (Present Study)

I i incongruent Melting

C Congruent Melting

4

MECHANICAL BEHAVIOR

Bars machined from the hor-pressed discs we e used to determine room tem-perature modulus of rupture (MOR) and high temperature creep and oxidationresistance. MOR testing was conducted in four-point bending on fixtures with a20 rmn loading span on the compressive surface and a 40 nmm load span on the ten-sile surface. Platen rate was 0.5 mm/minute. Results obtained irom the MORtests are shown in Table 5. The data was generated with specimens as-machinedfrom the hot-pressed discs and with specimens that had been used in high temperaturecreep tests. In the latter case, it was of interest ý.o determine the retainedstrength of specimens after they had been exposed to high temperature creep condi-tions (to be described later in the report).

Table 5. MODULUS OF RUPTURE DETERMINED AT ROOM TEMPERATURE WITH SPECIMENS AS-MACHINED FROMHOT-PRESSED MATERIALS AND WITH SPECIMENS USED FOR HIGH TEMPERATURE CREEP TESTS

Specimens As-Machined Retained MOR on Specimens UsedFrom Hot-Pressed Matena, For Creep Tests

Rare Earth MOR MOR Test Temp. ("C) PermanentAdditive tMPa) (MPa) (For Creep) Strain (%)

DY20 3 High-810 828 1250 0.12Mean-751 668 1300 0.30Low-698 888 1350. 095

Er20 3 High-832 770 1250 0.13Meav,-760 757 1300 0.29Low-712 510 1350 0.78

867 1300 0.21'653 1350 0.51t

Yb20 3 High-762 852 1200 0.14Mean-706 79P 1250 0.26Low-648 645 1300 0.60

Lu2 03 High-860 790 1300 0.10Mean-773' 718 T300 0.12Low-683. 870 1350 0.30

712 1400 065472 '400. 0.490

Sc20 3 843 1250 0.21803 1300 0.30613 1350 0.42

*Creep tested for 166 hourstCreep tested for 195 hours$Creep tested for 1813 houit

• %' •5

For specimens tested in the as-machin,:d condition, mean, high and low MORvalues are shown in Table 5 for each composition. The mean values ranged from706 to 773 MPa for the Si 3N4 -Yb 2 03-SiO2 material, and the Si 3 N4 -Lu 20 3 -SiO 2 mat.-rial, respectively. Some MOR values higher thsn 800 MPa were observEd forDY20 3 , Er 2 03 , and Lu 2 03-doped specimens.

Retained MOP. (at room temperature) was determined for specimens that hadbeen creep tested The temperature at which these specimens had been testedalong with the amount of permanent strain (as a result of creep) measured on eachspecimen are shown in the table with each MOR value. Some high values, 888, 867,852, and 843 MPa were noted for the Dy20 3 , Sr 2 03, and Sc 203 -doped compositions,respectively. Lower values were obtained, in general, when increasing tempera-tures we-e used for the creep tests or with increasing permanent strain in thebar.

Using geometrical analysis, a l.U% strain value would be equivalent to acurvature/specimen thickness ratio of 50 for the specimen and load span dimen-sions used in the MOR tests. Since the bars tested had permanent strain valuesof <1.0%, the curvature/specimen thickness ratio would be >50. Errors associatedwith testing curved specimens'have been shown to~be small (<1%) fros. curved beamanalysis 1 0 when the curvature specimen thickness ratio is large (>40).

Fracture surfaces of Si 3 N4 specimens dcped with Lu 2 03, Lt203, Yb2 03 , andSc20 3 are shown in Figures la, Ib, 2a, and 2b, respectively. The microstructuresare typical of hot-pressed Si 3 N4 material where the conversion of alpha to betaphase forms high aspect grains by solution reprecipitation through a liquid phaseproduced from Si 3 N4 -additive reactions. Grain diameters are in the 1 to 2 umrange. Fracture strengths shown in Table 5 were limited by the presence of ironimpurity found at the fracture origin of most specimens. Typically, the affectedarea appears as a dark.spot in the material and is observed in Figures 3a and 3b,the fracture origin of a Yb2 03 -doped Si 3 N4 specimen. In Figure 3a, the affectedarea is approximately 35 to 40 lm in size resulting in a specimen MOR value of701 MPa. At higher magnification, Figure 3b, the coarsening of the microstructureproduced by the Fe impurity is observed surrounded by a finer grain structure.

The hardness, iderntation fracture toughness, and elastic maodulus of thematerials were measured and listed in Table 6. The hardness was' determined usinga Vickers indenter under a 2 kg load. The Sc 20 3 -doped composition had the highesthardness while 'Yb20 3 and DY2 0 3 -doned material had the lowest values. It isuncertain whether the higher hardness values found for the Sc20 3 , Lu 2 03 , andEr 2 03 -doped Si3N4 materials were-affected'by the presence of residual alpha phaseSi 3 N4 . indentation fracture toughness was calculated using a technique reportedby Niihara et al."ll Indentation loads applied were II kg. Fracture toughnessvalues were calculated using both median and Palmqvisr cracks. The crack length/indentation size ratio favored the use of the ?almqvist expression and thesevalues are listed in Table 6. The Er 203 and Lu 2 0 3 -doped compositions gave thehighest toughness values. Although not reported here, when the equation formedian cracks was used, higher fracture toughness values were calculated for allcompositions. Elastic modulii were derived from ultrasonic measurements. Again,except for the Yb202-doped material, higher values appeared to be favored bydoping with rare earths having smaller ionic radii.

10. BARATTA. F. L.. MATTHEWS, W. T., and QUINN, G. D. ErnoAnoiartd wit Flexure Tesdng of ariltle .Warefwas. U.S ArmyMaterials Technology Laboratcry, MTL TR 87.35, July 1937.

11. NIIHARA. K., MORENA. R., and HASSELMAN. D. P. H. Furher Reply to Comment on Els/el1•tstc indentation Damag n Cmi:The .Wfedn/Rdia Crack System. Communicatias of the Amer. Ceramic S c.. C-1 16. July 1982.

6

a. Lu2O3

b. Er2 O3

Figure 1. Fracture surfaces of doped Si3 N4 specimens. Mag. ý0QOQX

7

a. Yb2O3

Lob

; ur,? 2. F-3C-ure ;-r'jc-s ioco>d :z. ý -ŽC1-i. ¶2 C

VOW I

7r~5 ki

a. ~ V A 5t OmLnurv edfetd ein a.80

~. oudar Btu ev oree einadFfrGandMti.Mg 70

F~gure3. Frctureorigi n Y~O3 ~oud S13 4 se6re

S,9

Table 6. HARDNESS, INDENTATION FRACTURE TOUGHNESS, ANDELASTIC MODULUS OF HOT-PRESSED Si3 N4 -RE20 3-SiO 2 COMPOSITIONS

Rard Earth Hardress*(VHN) Fracture Toughnesst Elastic Modulus$Additive (GPa) (MPam) (GPa)

DY20 3 16.1 5.53 303

Yb 2 0 3 16.5 5.76 307

Er20 3 17.5 6.85 309

Lu2 0 3 17.5 6.86 315

Sc 2 0 3 1'.0 5.70 318

°2 kg loadt1 1 kg load?Ultrasonic Method

CREEP MEASUREMENTS

The objective of the high temperature, time-dependent study was limited todetermining the extent of permanent strain that would develop in the bars under aconstant stress (300 MPa)'at temperatures between 1200 0 C and 1400 0 C for times upto approximately 350 hours. Machined bars were loaded in four-point bending onSiC fixtures with the 'same loading span dimensions as the room temperature fix-tures. The bars were placed under static loading calculated to produce 300 MPaflexur3l' stress. Testing temperatures used were between 1200 0 C and 1400 0 C withtime under stress approximately 350 hours for most specimens although, in somecases, shorter times, 185 to 1.95 hours, were chosen to produce other time-dependent values of permanent strain in the bars. All testing was conducted inan air environment. After the required number of hours under stress and tempera-ture had elapsed, the specimens were unloaded and cooled to room temperature forpermanent strain measurements and weight gain determinations. Strain measurementswere performed using a method reported by Hollenberg et al.1 2 The formula usedis based on me.lsuring specimen curvature, shown in Figure 4. Permanent strain,e, was derived from measurements on the bars where L is the distance between toploading points, t is the bar thickness, and Y is the rise generated by the arc

.formed by the specimen curvature and the chord subtended by that arc. Dimensionalmeasurements used for the calculations were taken on the .ceep'.tested specimenswith a stereoscope. Limitations of the method are that the strain is restrictedto I to 1.5% maximum, the curvature is reasonably uniform, and the creep propertiesin tension and compression are similar, that is, the neutral axis is not displaced'.

Specimens tested in this study did not produce strains greater than 1% andmeasurements along the curved surfaces indicated the curvature to be reasonablyuniform between the loading points'. The similarity of creep behavior in tensionand compression was determined by measuring curvature on both tensile and compres-sive surfaces and calculating the permanent strain values. These are compared in

11 HOLLENSERG, G. W.. TERWILLIGER, G. R.. and CORDON, R. S. Calcukaion of Stresses and Strains in Fou-Point Bending Creep

Tests. J. Amer. Ceram. Soc., 1971. p. 196-19*.

10

L -- - - - -tt

YI

4 (t) (Y)Permanent Strain (e) = (L 2 ,

Figure 4. Schematic diagram for permanent strain measurement in flexure specimen.

Table 7 for several different compositions exposed to test temperatures of 13000Cto 1400 0 C for approximately 350 hours. The strain values are within 0.03% for agiven test specimen. A complete listing of permanent, strain values determined inthis investigation is shown in Table 8. For each rare earth-doped Si 3 N4 composi-tion, the permanent strain value shown (%) was determined for a specimen of thatcomposition tested at a temperature indicated at the top of the column. Theactual duration of the test, in hours, is indicated by the number enclosed inparentheses under the strain value. Also shown are permanent strain determinationsfor Er 2 0 3 -doped Si 3 N4 and Lu 2 0 3 -doped Si 3 N4 held for shorter times at temperatureand for specimens of these compositions which were prefired in nitrogen at 1300 0 Cfor eight hours attempting to promote further crystallization of the grain boundaryprior to creep testing. A commercially available material, Kyocera grade 252, anin situ toughened Si 3 N4 doped with Yb 2 0 3 (and small amounts of Y2 0 3 and A1 2 0 3 )was included in the study for comparative purposes. As expected, for each composi-tion, the amount of permanent strain increased with increasing test temperature.The specimens of Lu 2 0 3 -doped Si3N4 that were prefired in nitrogeni had a similaramount of permanent strain as the as-hot-pressed specimens after testing at temper-atures of 1300 0 C and 1350 0 C. This may be due to a significant amount of crystal-lization already occurring during the cooldown from the hot pressing temperaturefor this composition. However, the specimens of Er 2 03 -doped Si 3 N4 prefired innitrogen showed a reduction in permanent strain as compared with' as-hot-pressedmaterial. 'At 1300 0 C, the permanent strain was reduced from 0.29% to 0.23% and at1350 0 C from 0.78% to 0.59%.

Table 7. COMPARISON OF'PERMANENT STRAIN MEASUREMENTS ON THECOMPRESSION AND TENSILE SURFACES

Rare Earth Test Temp. Perm. Strain (%) Perm Strain (%)Additive (C0 (Compressive) (Tensile)

OY203 1350 0.95 0.92

Er203 1350 0.59 0.57

Lu20 3 1350 0.27 0.28

Lu2 0 3 1400 0.65 0.63

Kyocera 252" 1300 0.43 0.46

*A high toughness Yb2 03 -doped Si 3 N4 containing small amounts-of Y203 andAX2 0 3. (Kyocera Corp., Japan)

•'11

Table 8. PERMANENT STRAIN IN Si3 N4-RE2 0 3 -SiO 2 COMPOSITIONS AFTER APPROXIMATELY350 HOURS UNDER 300 MPa FLEXURE STRESS AT TEMPERAT JRES OF 12000C to 14000C

Temperature (0C)Rare EarthAdditive 1200 1250 1300 1350 .1400

Yb 2 0 3 014% 026% 0.60% -

(355) (351) (355)

DY20 3 - 0.12% 0.30% 0.95%(354) (356) (355)

Sc 2 0 3 0.21% 0.30% 0.42%(355) (353) (354)

Er 2 0 3 0.13% 0.23% 0.78%(353) (351) (352)

Er 20 3 0.21% 0.51%(166) (192)

Er 203* 0.23% 0.59%

(355) (355)

Lu 2 03, 0.10% 0.30% 0.65%

(352) (352) (345)

Lu 2 0 3 - 0.49%

(1,88)

Lu 203* 0.12% 0.27%

(355) (355)

Kyocera ,252 0- .19% 0.44% 0.98%(355) (355) (353)

'Specimen prefired in N2 at 13000C for 8 hours before creep testing.

NOTE: Numbers in parenthesis, beneath strain values, represent the actual test durationtimes (hours).

12

The data may be graphically represented by a semilogarithmic plot of permanentstrain versus temperature, illustrated in Figure 5. Plots of data for Si 3 N4 dopedwitr Dy203, Yb203, Er 2 0 3 , or Lu 2 0 3 appear to have reasonably similar slopes sug-gesting a common mechanism for producing time-dependent strain at hich temperaturesin these materials. Only the Sc 2 0 3 -doped Si 3 N4 exhibits a significantly differ-ent slope. However, further testing and characterization of this material isrequired. The Lu 2 0 3 -doped material clearly shows the highest resistance to strainof the materials tested.

,/

1.0Kyocera 252 /-/7' /

0.5 /

/ SC

Yb

Lu

S0.1 -L.CU

0.05

F Stress-300 MPa

I I - 1 I I1200 1250 1300 1350 1400

Tefanperature (OC)

Figu. anert strain (%) versus temperature for Si3 N4 -RE20 3 -SiO 2 compositions tested under300 MPa flexure stress for approximately 350 hours.

The bars were analyzed by XRD to determine the phases existing after thelong exposure to high temperature during creep testing. All specimens showedbeta-Si 3 N4 and Si 2 N20 to be present. The compounds Lu 2 Si 2O7 , Yb2 Si 2O7 , Er 2 Si 2 O7 ,and Sc 2 Si 2 O7 were also found in specimens duped with their respective oxides.Peak intensities for these pyrosilicate phases were significantly greater thanfound in the as-hot-pressed materials indicating that appreciable crystallizationof these phases had occurred during high temperature testing. In the DY20 3 -dopedspecimens, the presence of a phase which appeared analagous to the'Y 5 (Si0 4 ) 3 N

13

(nitrogen-apatite) phase formed during high temperature testing rather than theexpected pyrosilicate phase. Apparently, a compositional shift occurred duringhot pressing which was not detected until crystallization of the phase occurredduring the high temperature tests. The development of this phase, which appearsto be dysprosium N-apatite, at the grain boundaries may increase 'resistance tchigh temperature creep over DY2 Si 2O7 containing material. Oxidation of theDY20 3 -doped material at 900 0 C and 1000 0 C did not indicate the existence of anyintermediate temperature problems. However, the oxidation studies were conductedwith as-hot-pressed specimens, before crystallization of this phase was observed.

Creep behavior in silicon nitride is primarily governed by vis.ous flow ofthe grain boundary phase, grain boundary sliding/cavitation, or diffusional pro-cesses. 1 3-15 In flexure, the creep behavior of Si 3 N4 has been described astransient 1 6 rather than attaining steady state conditions. Determination ofcreep mechanisms usually involves calculations of activation energy and stressexponents based on the temperature and stress dependence of the strain (creep)rate accompanied by TEM to observe microstructural changes in the specimens. Itwas not an objective of this study to s,.ggest creep mechanisms responsible forthe development of permanent strain observed in the test specimens. However,from the permanent strain data available, activation energies were calculatedfrom log strain rate versus I/T. Strain rate values were derived from permanentstrain versus time data. From the calculations, based on the data used to cqn-struct Figure 5, it was determined that the activation energies fell between 70and 100 kcal/mol for all compositions except for the Sc2 0 3-doped Si 3 N4 . Thesevalues are lower than would be obtained under steady-state conditions because thetotal strain values used represent cumulative strain from both primary and second-ary stages of creep. The activation energy value rises to approximately 140 kcal./mol if only the strain data obtained after testing Er 20 3 -doped Si3N4 specimensfor 166 and 351 hours at 1300 0 C and 192 and 352 hours at 1350 0 C (data from Table 8)is used for the calculation. This value would essentially be associated withcreep behavior in the secondary stage (steady-state region) of the creep test.Activation energies of this magnitude have been reported for a suggested mechanismof creep deformation by viscous flow with strain accommodation by grain boundarysliding.14, 15 This may be related to the existence of an amorphous layer betweenthe Si 3 N4 grains and the crystalline phaze formed in the boundaries, as observedby Kleebe and Ruhle1 7 and Clarke.1 8 As the boundary phase crystallizes, thezomposition of the amorphous phase formed should be driven toward a eutecticcomposition ,in the system. 19 Mechanical behavior, particularly at high tempera-ture-, would be significantly influenced by the thickness and composition of theamorphous phase layer as well as the structure and composition of the crystalline

13. 'KOSSOWSKY, q., MILLER. D. G., and DIAZ. E. S. Tensie and Creep Strengths of IHot.Prewd Si.V 4. I..Mals. S.i..'v. l0. 1975.o. 983-997.

14. SELTZER. H. S. High remperitue Ceep of silicon.Sase Cor •pmds. Amer. Ceram. Soc. BBuI v. 56. no. 4. 1977, p. 418-423.15. MOSHER, D. f., RAJ, R.. and KOSSOWSKY, R. •easuremett of Viscosity of the Gram BounderY Phase in Hot-Pressed Silicon

,Viride. J. Matls. Sc., v. 11. 1976. p. 49-53.16. WIEDERHORN. S. M., and TIGHE. N. J. Structural Reliabiliiy of Ytrria.Doped Hot.A'ested Silicon .Vitrde at Elevated Temperatures.

J. Amer. Ceram. Soc..v. 66, no. 12.1983. p. 884-889.!7. KLEEBE. H-J.. anJ RUHLE. M. HiTh Resolution Electron .icroscopy Studies on Cystalline Secondary Phase$ ir Silicon .Vitride Based

. Maerials. , Psented at the 93rd Annual Amer. Ceramic Society Meeting. Cincinnati. OH, 30 April 1991.I1. CLARKE, D. R. The .Vcrostucrure of.Vitrogen Ceramics in Progress in Nitrogen Ceramics. Proc. of the NATO Adv. Study Inst. on

Nitrogen Ceramics. Univ. of Sussex. U.K.. July 27 - August 7,1981, F. L Riley, ec1.. M. Nijhoff. Pub.. 1983. p. 341-357.19. LANGE. F. F. Importance of Phase Equilibria on P'ocess Control of SiyV4 Fabrication in Ceramics for High Performance Applications

ll[-Reliability, Proc. of the 6th Army Materials Technology Cont., Orcas Island. WA. July 10-13. 979. E. M. Lenoe, R. N. Katz,aidJ.J. Burke. edL., lenum Pres. NY. 1983. p. 275-29E.

14

phase(s) formed in the grain boundaries and at triple points. The solubility ofnitrogen in these grain boundary phases is unknown and its effect on eutectictemperatures as well as resultant mechanical behavior, such as creep, needs clari-fication. It was noted that the Si 3 N4 -rare earth compositions with i.ncreasingcreep resistance were similar in order to the compositions having higher onset ofshrinkage temperatures during hot pressing.

OXIDATION

Oxidation measurements were made with unstressed test bars fired in air at900 0 C and 1000°C and from bars used for the creep testing at 1200 0 C to 14000 C.All bars were weighed to the nearest 0.1 mg before exposure to high temperatures.The unstressed Si 3 N4 specimens doped with Dy 2 0 3 , Yb2 0 3 , Er 20 3 , and Lu 2 0 3 hadnegligible weight gains (0.1 mg max.) after 41 hours at 9000 C and 192 hours at10000C. The specimens used for creep testing were reweighed after the tests andweight gains observed (mg/cm2 ) are shown in Table 9. Time at temperature wasapproximately 350 hours. Excellent resistance to oxidation at all temperaturesis indicated 'for each composition. The Sc 2 0 3 -doped material shows somewhat higherweight gains than the other compositions btut the values are reasonably consistentwith those reported by Cheong and Sanders 2 0 for material of similar composition.-

Table 9. OXIDATION OF Si3 N4 -RE 2 0 3 -SiO 2 COMPOSITIONS(WEIGHT GAN AFTER CREEP TESTS - mg/cm2)

Temperature (C)Rare EarthAdditive 1200 1250 1300 1350 1400

DY20 3 0.08 0.12 0.13

Er2 0 3 - 0.06 0.09, 0.13

Yb 2 0 3 0.06 0.09 0.11 -

Lu2 0 3 - - 0.10 0.12 0.20

Sc2 0 3 - 0.09 '0.16 '0.23

NOTES: 1. Time at temperature approximately 350 hours.2. Specimens of each inomposition oxidized at 900C for 41 hours and at 10000C for

192 hours did not exhibit any significant weight change. (Weighing balance sens-itivity was t 0.1 mg.)

20. CHIEONG, D-S.. and SANDERS. W. A. High Temperturc Deformawlon and Microsrructural A ndysft for Si3 cV4-oJj. NASA Tech.Meem. 103239. NASA Lewis Research Center. Cleveland. OH. August 1990.

' 15

SUM•?ARY AND CONCLUSIONS

Si 3 N4 can be fully densified with additions of Lu 2 03 , Er 2 0 3 , Yb2 03 , or DY2 03 .The pyrosilicate phase, (RE2 Si 2 O7 ), was formed with the Lu 2 03 , Er 2 03, and Yb2 03dopants while a phase which appears to be an N-apatite [Dy 5 (SiO4 )3N], crystallizedduring high temperature testing of the DY20 3 -doped specimens. However, DY2 Si 2O7can be formed with appropriate processing conditions.

Hardness and creep resistance of the materials examined generally increasedwith decreasing ionic radius of the rare earth used for doping, the exceptionbeing the Yb2 03 -doped Si 3 N4 which exhibited the lowest hardness and creep resistanceof the materials tested in the study. However, another explanation for the levelsof hardness and creep resistance observed may involve the solubility of nitrogenin the grain boundary phases and its effect on eutectic temperatures and mechanicalproperties of phases formed in the Si 3 N4 -RE20 3 -SiO 2 systems.

Fermanent strain values (<1%) were determined for each Si 3 N4 -rare earthcomposition tested under 300 MPa flexural stress for up Lc approximately 350 hoursat temperatures of 1200 0 C to 14000 C. The Lu 2 03 -doped Si 3 N4 exhibited the mostresistance to high temperature creep up to the highest test temperature, 1400 0 C.similarities in the slopes of semilogarithmic plots of permanent strain (e) versusT, for each composition, suggest.that creep was controlled by a common mechanism.

ACKNOWLEDGMENT

The author thanks G. Gilde for assisting with the hardness/indentation tough-ness measurements, R. Hinxman for XRD data, A. Zani for polished specimen prepara-tion, P. Wong for SEM microscopy, and L. Tardiff for ultrasonic modulus measurements.

16

DISTRIBUTION LIST

No. ofCopies To

I Office of the Under Secretary of Defense for Research and Engineering,The Pentagon, Washington, DC 20301

Commander, U.S. Army Laboratory Command, 2800 Powder Mill Road, Adelphi,MD 2.0783-1145

1 ATTN: AMSLC-IM-TL1 AMSLC-CT

Commander, Defense Technical Information Center, Cameron Station, Building s,5010 Du'e Street, Alexandria, VA 22304-6145

2 ATTN: DTIC-FOAC

1 MIAC/CINDAS, Purdue University, 2595 Yeager Road, West Lafayette,IN 47905

Commander, Army Research Office, P.O. Box 12211, Research Triangle Park,NC 27709-2211

1 ATTN: Information Processing Office

Commander, U.S. Army Materiel Command, 5001 Eisenhower Avenue,Alexandria, VA 22333

1 ATTN: AMCSCI

Commander, U.S. Army Materiel Systems Analysis Activity,Aberdeen Proving Grouna, MD 21005

1 ATTN: AMXSY-MP, i. Cohen

Commander, U.S. Army Missile Command, Redstone Scientific Information Center,Redstone Arsenal, AL 35898-5241

1 ATTN: AMSMI-RD-CS-R/Doc1 AMSMI-RLM

Commander, U.S. Army Armament, Munitions and Chemical Command, Dover, NJ' 078011 ATTN: Technical Library

Commander, U.S.' Army Natick Research, Development and Engineering Center,Natick, MA 01760-5010

1 ATTN: Technical Library

Comnander, U.S. Army Satellite Communications Agency, Fort MonmOuth, NJ 07703I ATTN: Technical Document Center

Commander, U.S.,Army Tank-Automotive Command, Warren, M.I 48397-50001 ATTN: AMSTA-7SK1 AMSTA-TSL, Technical Library

Commander, White Sands Missile Ra.ge, NM 880021 ATTN: STEWS-WS-VT

President, Airborne, Electronics and Special Warfare Board, Fort Bragg,NC 25307

I ATTN: Library

Director, U.S. Army BSllistic Research Laboratory, Aberdeen Proving Ground,MO 21005

1 ATTN: ;LCSR-TSB-S (STINFO)

Commander, Dugway Proving Ground, Ougway, UT 840221 ATTN: Technical Library, Technical Information Division

Commander, Harry Diamond Laboratories, 2800 Powder Mill Road, Adelphi, MID 20783ATTN: Technical Information Office

Director, Benet Weapons Laboratory, LCWSL, USA AJ4CCOM, Watervliet, NY 121891 ATTN: AMSMC-LCB-TLI AMSMC-LCB-R1 A4SMC-LCB-RM1 AMSMC-LCB-RP

Commander, U.S. Army Foreign Science and Technology Center, 220 7th Street, N.E.,Charlottesville, VA ?2901-5396

3..ATTN: AIFRTC, Applied Technologies Branch, Gerald Schlesiiger

No. ofCopies To

I Plastics Technical Evaluation Center, (PLASTEC), ARDEC, Bldg. 355N,Picatinny Arsenal, NJ 07806-5000

Commander, U.S. Army Aeromedical Research Unit, P.O. Box 577, Fort Rucker,AL 36360

I ATTN: Technical Library

Commander, U.S. Army Aviation Systems Command, Aviation Research and TechnologyActivity, Aviation Applied Technology Directorate, Fort Eustis, VA 23604-5577

I ATTN: SAVDL-E-MOS

U.S. Army Aviation Training Library, Fort Rucker, AL 36360I ATTN: Building 5906-5907

Commander, U.S. Army Agency for Aviation Safety, Fort Rucker, AL 363621 ATTN: Technical Library

Comnmander, USACDC Air Defense Agency, Fort Bliss. TX 79916I ATTN: Technical Library

I Clarke Engineer Scool Library, 3202 Nebraska Ave. North, Ft. Leonard Wood.MO 65473-5000

Commander, U.S. Army Engineer Waterways Experiment Station, P. 0. Box 631,Vicksburg, MS 39130

1 ATTN: Research Center Lit "ay

Commandant, U.S. Army Quartermaster School. Fort Lee, VA 2?3011 ATTN: Quartermaster School Library

Naval Research Labor~tnry, Washington, OC 20375I ATTN: Code 5'230

I Dr. G. R. Yoder - Code 6384

Chief of ýaval Reseirch, Arlington, VA 2221/I ATTN: Code 4/1

I Edward J. Morriýsey, 4ROC/NILTE, .'right-P.itterson Air Force, Base. Ol 145,133-6523

CommAndo-, U.S. Air Force Wriqht Research & D0'velopicnt Centcr.Wright.Patterson Air Force Base, OH 45433-5523

1 ATTN: WRDC/MLLP, M. 7orney, Jr.1 wRDC/MLBC, Mr. Stanley Schulman

%ASA - Marshall Space Flight Center. MSFC, AL 3581?1 ATTN: Mr. Paul Srhuerer/EHO1

U.S. Department of Commu..rce, 'Iatiunal Institute of Standards and technotogy,Gaithersburg, MO 20899

1 ATTN: Stephen M. Hs;,, Chipf, Ceramics Division, Institute for MaterialsScience and Engineprini

I Coemittee on Marine ýtructrcs, Marine Board, Nitional Research Council,2101 Constitution AvP., i.W., Wja.injton, UC ?0410

I Librarian, Materials Sciences Corporation, 930 Harvest Drive, Suite 300,

BlueBell, PA 194?2

I The Charles Stark Draper l.aboraLory, 68 Albany Street. Cambrldge, MA 02139

Wyman-Gordon Company,.Worcester, MA 01601I ATTN: Technical Library

General Dynamics, Convair Aprospace Division, P.O. Box 748, Fort Worth.TX 75101

I ATTN: Mfq. Engineering Technical Library

1 D*partment of the Army,.Aerostructures DirtCtot Ate, MS.266, U.S. Army Aviation

R&T Activity AVSCP., Langley Research Center, 1NaPton. VA 23665-5225

I NASA Langley Research CUnter, Mampton, VA 23665-5225

I U.S. Army Prooulsion Directorate, NASA Lewis Research Center, 2100 SrookpartRoad, Cleveland, OH 44135-3191

1 NASA - Lewis Research Center, 2100 Srookpark Road, Cleve114nd. ON 44135-3191

Director, U.S. Army Materials Technol~ogy Laboratory, Watertown, MA 02172-00012 ATTN: SIMT-TrI,1 " Author

0~ ~ 4JL0 Jt..,Ica a~ S-a, WW

112 43. 0;= CV)- CZ r. W3tW-0 w3 3L

tj/ D. 0 v. SL.Q 0 >, V) S. a3~ 0 0

L40 0 Cr'r CL" /0 0 C3 ~ 4 3 3

00.cm 0 0

I In c" m I

.04 -W. ml ' -2 C3

*.

& ~ ML 3U 0'CL43

0- V) I IVa -.0A01J0 -.30 U-l j ; n0%, In C L..~-

C% 4o Q 0 000Sho- L - o In a,. 3

0'o W uc I 433 ui .1 4. 4L3j -4 a 0 S. t

- o CL0ms. 4 wi

cc0- a 0 m cc z

It a

I6 L 0 . W.2WO3- -.c 40 i I 3 CL6ML

43 V,~ L 43 .

-6 jj4 Io - t-Cs

Li uI 01 410 V 0' 1''0 43 a CA -

4) 0~L 6. 5433L. 3.4

V 10S i

3.33. 0% 034 .0 0.., 44 ~~~..J 0 34 1. I3JU 0 3 4 1 . 0

3.3

4. 3.8-c 4 3,

cK w.- 433- v

If; 3.U Li43IN- -w 43 C.~'~"'QM.Ui 1,o

1020

c. v~J .2. 11, S

9.1cc 3.i *CCI U ,

DATC: