-- R63 442 SILICON NITRIDE CERNMIC FIBERS FROM PEEMCPLMR /(U) SRI INTERNATIONAL MENLO PARK CAR M LRINE ET AL.JUN 97 TR-8 MNSSI4-84-C-392

UNCLASSIFIED F/0 7/3 NL

lilllllllu

lllllilli

1.5 LA .6

MICROCOPY RESOLUTION TEST CHART

NA. J&NALBUREAU OF STAt4DARDS1-194-A_

Wr IV qw im sW .. V w A.W W w

% *. -0.~

UnclaRiSEC .R-1r _A_ C S

AUMENTATION PAGE

.a RPOR SIC AD - 183 42 b. RESTRICTIVE MARKCINGS

12a SECtiITY CLASSIFICATION AUT _43 DISTRIBUTION IAVAILABILJTY OF REPORTAB 0:5 Approved for public release.

2b. DEC&ASSiF,CATION, DOWNG S EDULE Distribution unlimited.

4 PERFORMING ORGANIZATION REPORT NUMB S S MONITORING ORGANIZATION REPORT NUMBER(S)'A

Technical Report No. 8

6 NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION

SRI International ONR

6c. ADDRESS (City. State. and ZIPCode) 7b ADDRESS (City. State. and ZIP Code)Organometallic Chemistry Program Department of the Navy333 Ravenswood Avenue Arlington, VA 22217

Menlo Park, CA 94025

So. NAME OF ;UNDNG, SPONSORING Sb OFF,CE SYMBOL 9 PROCUREMENT iNSTRUMENT iDENTIFICATION NUMBERORGANIZATION (If appicabloe N00014-84-C0392Office of Naval Research ONR N00014-85-C0668

8C. ADDRESS (City, State, and ZIP Code) 10 SOURCE OF FUNDING NUMBERSDepartment of the Navy PROGRAM PROJECT '-ASK WORK .,NIT

Arlington, VA 22217 ELEMENT NO NO NO ACCESSION NO

I1 TiTLE (Include Security Classification)

Silicon Nitride Ceramic Fibers from Preceramic Polymers

:2 PERSONAL AUTHOR(S)R. Laine, Y. Blum, A. Chow, and K. Schwartz

'3& TYPE OF REPORT j3b TIME COVERED 14 DATE OF REPORT "Year, Month, Day) LPAGECOuNTPresentation I ROM TO L

'6 S PRLEME NTARY .NOTATIONTo be published in the SDIO Composite Materials Consortium Report, Woodshole, Mass.June, 1987

7 COSATI CODES 18 SUBJECT TERMS (Continue on reverse of necessary and Odentify by block numnber)::ELD GROUP SUB-GROUP Polysilazanes; preceramic polymers; Si-H bond activation

catalysis; Si N4 ; ceramic coatings, precursor fibers,titanium nitride

:g 9A PACT (Continue on reverse if necessary and identify by block number)

The program objectives are to develop: (1) Transition metal catalyzed synthetic routes

to designed, tractable silicon nitride (Si'N4 ) preceramic polymers (polysilazanes) based on

SRI developed technology; (2) Methods of spinning the resultant polysilazanes into continuous

preceramic fibers; and, (3) Pyrolysis techniques for transforming the preceramic fibers into

high strength Si3 N and silicon carbide nitride (SiCN) fibers. In the past year, we have

learned to prepare polysilazanes drived from precursors of the type-[H>SiNMe 1= , whose vis-

coelastic properties can be carefully controlled by type of catalyst and/or Raction condi-

tions. This control has permitted us to draw preceramic fibers of diameters as small as

10 um as seen in the attached photographs. Furthermore, we have developed pyrolysis methodo

logy that permits us to obtain ceramic yields of 50-70% with Sf 3 N7 purities ranging from 80-99%. We have discovered that polymer molecular weight greatly in luences the ceramic yield

but only monomer design or pyrolysis under a reactive atmosphere seems to influence selecti-

vity to specific ceramic products as shown in the attached Table...--

O 0,STRIBUTiON, AVAILABILITY OF AB9WRACT 21 ABSTRACT SECtRITY CLASSIF-iCATIONLQ' NCLASSIED/UNLIMITED Q SAME A- RPT C3 OTIC USERS Unclassified

22a NqAME OF RESPONSiBLE iNDIVIDUAL ='22b TELEPMONE (Include AreaCode) 22c OF;,CE SYMBOL

Kenneth Wvnne 1 202) 606-4410 1 NC

DO FORM 1473.84 MAR 63 APR edition may be used until exhausted SECjRITY CLASSIFICA'ION OF 'HIS PAGEAll other editions are obsolete Unclassified

87 8 4 029

OFFICE OF NAVAL RESEARCH

Contract No. N00014-84-C-0392Contract No. N00014-85-C-0668

Technical Report No. 8

SILICON NITRIDE CERAMIC FIBERS FROM PRECERAMIC POLYMERS

Richard M. Lamne, Yigal D. Blum,Andrea Chow, and Kenneth S. Schwartz

Inorganic and Organometallic Chemistry,Physical Polymer Chemistry Program and

the Ceramics Program -Acceskrai F oSRI International, Menlo Park, CA 94025

-TSCR&

DTiC I A3

By.

To be Published Ain the 30'%",to

SDI0 Composite Materials Consortium ReportWoodshole, Mass

S~it I

June 1987

oric

Reproduction in whole or in part is permitted for anypurpose of the United States Government.

This document has been approved for public releaseand sale; its distribution is unlimited.

I

POLYMER PRECURSORS TO SILICON NITRIDECOATINGS, BINDERS AND FIBERS

R. M. Laine, Y. D. Blum, K. Schwartz, R.Hamlin, A. Chow, and P. L. Lundquist

Inorganic and Organometallic Chemistry,Physical Polymer Program,

and Ceramics Program

SRI INTERNATIONALMenlo Park, CA

Work Sponsored By SDIO/IST As Managed By The OfficeOf Naval Research

ONR CONTRACT NOS. N00014-84-C-0392 ANDN00014-85-C-00668

PROGRAM OBJECTIVES

* Synthesize tractable polymer precursors tosilicon nitride, using SRi's catalytic

dehydrocoupling process nitride that can be spunand that give high ceramic yields of high purityS13N4 .

* Develop an understanding of the kinetics andmechanisms of the catalytic process.

" Detail the conditions necessary to shape thepolymer precursor into a finished, Infusiblepreceramic form.

" Detail the pyrolysis conditions necessary totransfrom the Infusible shape into a finished,high density ceramic product.

" Develop analytical methods of characterizing thefinal ceramic product.

0~ 'W %EV%*~ ~ . % L]

Polymer precursor design, synthesis and pyrolytictransformation make up the three steps in thedevelopment of preceramics useful for the preparationof coatings and fibers or for binder applications.

'

Polymer Design Pyrolysis Analysis

.5,

'

BASIC CONCEPTS IN MATERIALS CHEMISTRY

MOLECULAR ANALOGS OF MATERIALS

INTHELOY, Given the Empirical Formula for a Material,It Should be Possible to Prepare a ChemicalAnalog

CERAMC CHMIALMONOMERIC UNIT MONOMERIC UNIT

SI3 N4 H6 SI3N4

H

II

K >THIS ANALOG REPRESENTS A POTENTIAL PRECURSOR TO

THAT MATERIAL

. , ,.. .' *; :. : ; t' % * ;::

MONOMERS ARE OFTEN VOLATILE and Therefore notSuitable as Precursors to Ceramics--One NeedsOligomeric or Polymeric Species

LINEAR OLIGOMERS AND POLYMERS- MUST RETAIN LATENTREACTIVITY.-So They Can Be Made Infusible ByCrosslinking:

-S

\ Si-N

SI-NNN\,

S'I

NiNIE R O L G M R A N D P O Y E S4UT E A N L T NR E A C I V I T --S T h e C a B e a d e nf u ible B y.

POLYSILAZANE PRECURSORS TO Si3 Nj

IN PRACTICE:

It Is Difficult to Synthesize Even Simple, HighMolecular Weight Preceramic Polysilazanes That Are

Tractable; Yet Retain Latent Reactivity.

The Polysilazane, H-[Me2 SiNHJx-H , a Nitrogen Analog of

Polysiloxane Exhibits no Latent Reactivity andTherefore Depolymerizes when Pyrolyzed -Giving No

Ceramic Product

Polysilazane Syntheses byCatalytic Dehydrocoupling

SRI has recently developed a catalytic method offorming SI-N bonds from Si-H and N-H bonds that canbe used to form polysilazanes:

Et2 SIH 2 + NH 3 8M3 (-C 12L60QC> H2 + -[Et 2SI-NH] x -

Mn - 500

The products obtained by this reaction are mostly cyclomeric.However, by performing modeling studies on this type ofreaction, we been able to obtain sufficient kinetic informationto establish a preliminary picture of the reaction mechanismand use this to develop better approaches to preceramicpolymers as shown in the next slides:

I ~ * ~ q d* .~II 1I~* * ~ *

MODELING THE DEHYDROCOUPLING REACTION

Et3SIH + RNH2 %F 1 H2 + Et3SINHR

Rate = k[Et3SIH][RNH2 JI- *X for R =n-Pr, n-Bu

Rate = k[Et3SIH][RNH2Jo~X for R =s-Bu

Rate = k[Et3SIH]O.Y[RNH2] OX for R = t-Bu

Rate = 0 for piperidine

Rate = k[Ru3(CO) 1 2]-.

PROPOSED DEHYDROCOUPLING MECHANISM

RNH 2 + M -- (RNH 2)M + RNH2 <-> (RNH 2)2M

(RNH2)M + Et3SJH <- Et3SIM(H)(RNH2)

Et3SiM(H)(RNH 2) + RNH 2 ->Et 3SINHR + (RNH 2)MH 2

(RNH2)MH 2 (RNH 2)M + H2

* ' , .'6P A

The above results Indicate that transition metalcatalyzed dehydrocoupling Is extremely susceptible tosteric Inhibition. This Is supported by the phenylsilane coupling reactions wherein, the 600C reactionleads exclusively to linear oligomers and only at 900Cdoes crosslinking occur by formation of imino bridges.This latter observation suggests that Imino bridgeformation could be the mechanism whereby linearpreceramic polysilazanes can be made Infusible.

PhSiH-3 + Ni- 3 &s3LMQ12/600C/THF,, H2 +

H-[PhSIHNH] x-H

viscous oil, Mn = 1000

H-[PhSiHNH]X-H + NH-3 Ru(O1 H2 +

NH 0

-[PhSiHNH~x[PhSiNH] Ysolid, Mn =1400

Preceramic Polysilazanes

The ammonolysis of H2 SiCI2 gives oligomers,-[H2SINMe]x-, where x = 10:

H2 SICI2 + 3MeNH 2 - - -[H2 SiNMe] x - + 2xMeNH 3 CI

Pyrolysis of these oligomers gives a 38-39 wt % yieldof ceramic product that is reported to be mostlysilicon nitride. Considerable precursor volatilizationoccurs during pyrolysis.

Seyferth and Wiseman, 1984

Because the Polysilazane HNMe-[H2SiNMe]x-H has N-H

caps that can react, when heated at 60-900C, with theInternal H2 Si groups. The Dehydrocoupling Reaction can

be used to form tractable higher molecular weight(less volatile) polymers and then to crosslink(thermoset) these polymers to render them Infusible:

HNMe-[H2SiNMeJx-H Ru3(O120 polymers

polymers - gels- rubbers ;-plastics

Figure 1, following, illustrates the changes in molecular weightand dispersion that occur as polymerization proceeds.

55K6

500K OUR50

4030

50K 101 0

106 105 0 103 102

MOLECULAR WEIGHT (daltons)

JA-B-997-1 3

FIGURE 1 GPC RESULTS OF 1H2SiNMe2x POLYMERLZATK)N CATALYZED BY RU3(0O) 12

%2

The tractable polymers shown in the last slide have beenused to make 2000 A coatings of silicon nitride on stainlesssteel, aluminum, silica and graphite/graphite composites. Thecoatings on stainless steel were featureless in the SEM at thehighest magnification. Electrochemical corrosion studies wereconducted on coated aluminum 6061 coupons to establish themicroporosity of these coatings and their ability to protect thesurface from corrosive environments over extended periods. Theresults of these electrochemical corrosion studies (shown onthe following slides) reveal that the coatings contained someflaws but were unchanged upon exposure to 3.5% NaCI solutionfor as long as 21 days. These results suggest that siliconnitride coatings prepared by simple dipcoating techniques maybe useful in a variety of composite applications [e.g. asprotective coatings on graphite fibers during the fabrication ofaluminum/graphite fiber composites.

6061 Aluminum Alloy in 3.5% NaCI for 3 hours5

80

460

U 3 - f

2 40

1 20

0 0-4 -2 0 2 4 6

LOG w

AC Impedance Spectrum

Observation: The Polarization Resistance Rp is at least 103

which is relatively low.

~' xZ.''1e~e-VXZ C.11 -1'Zz

Coated 6061 Alloy in 3.5% NaCI for 5 hours5-

-80

4 3E 60

M3-

0

-2 -1 0 1 2 3 4 5Log w

AC Impedance SpectrumObservation: R pis at least 10. = 31600 ohms which isrelatively high.

Coated 6061 Alloy in 3.5% NaCI for 21 days5 90

80

4 - 70

60 1~_ 50

ti40

2 -22

-2 -1 0 1 2 3 4 5LOG w

AC Impedance SpectrumObservaion: R is still at least 104 .5 or 31600 ohms after 21days in solution.

Icr s522.6jgikcm 2 Uncoated 6061 T6'corr!52.6.tAlcm2 Coated 6061 T6

Significance: The Corrosion Protectit n Afforded by Si3N4Coatings Does not Degrade with Time over 21 Days

ANODIC POLARIZATION DIAGRAM

1.5

w 0.5U)

-0.5

-1 .5

0 2 4 6 8 10Log I/nA

Observation: Si 3N4 Coated Al 6061 T6 Alloy has a PittingPotential about 1.75 Volts Higher than Uncoated alloy.

Significance: The coated material is significantly moreresistant to pitting corrosion.

w7

CATHODIC POLARIZATION DIAGRAM

-0.9

-1.1

wE( -1.3 • E

-1.5

-1.7 ,

2 3 4 5 6 7 8Log I/nA

Ober.vatin: icorr = 7 jA/cm2 Uncoated 6061 T6 Alloy1corr = 1.4 gA/cm 2 Coated 6061 T6 Alloy

by Tafel Extrapolation

Obserati: bc = 175 mv/decade Uncoated 6061 T6 Alloy

= 340 mv/decade Coated 6061 T6 Alloy

Interpretation: Based on Geometric Surface Area, the SubstrateCorrosion Rates and Hydrogen Evolution Rates areSignificantly Lower on the Coated Specimen.

'

is

is

.5

These precursor polysilazane polymers have provedextremely useful as binders in the fabrication of fully densesilicon nitride bodies by pressureless sintering of compressionmolded silicon nitride powder. We find that pyrolysis of shapescompression molded with polysilazane at 8000C leads todensification and observable intrinsic strength in the greenbody. Densities of up to 75% have been obtained. Furtherheating at 17250C for 20h under N2 leads to full densification if

sintering aids are present.By comparsion, 800°C pyrolysis of shapes molded with a

standard organic binder does not result in densification and theresulting product is similar to chalk.

The more viscous polymer can be extruded to give fibers of100-3001m and hand drawn, as shown in the following Figure, togive smooth precursor fibers of approximately 10 pm.

HAND DRAWN 10 urn FIBERS USING POLYMER DERIVED FROM -H 2SiNMelx:

0 50 o 50Pm 450X pm 450X

JA-899' 15

Pyrolysis of -[HSNMe1x- Oligomers andPolymers

S Mn ViscoIty Ceramic Yield SI3N4

(GPC) (poise) (9000C) %

[H2 SINMex 650 1 40 80-85

x=10

[H2SINMe]x 1150 5 45-50 80-85

x=19

[H2 SINMe]x 2100 18 60-65 80-85

Ru3(CO) 12/90°C30h

[H2 SINMe]x 2300 100 65-70 80-85

Ru3(CO) 12/900C

65h

The salient features seen In the above table are that thereis a direct correlation between the molecular weight Mn of theprecursor and the ceramic yield. This is to be expected Ifprecursor volatilization results in physical loss of precursor.Also of importance is the fact that the viscosity of the polymerchanges from 1 to 100 poises while the Mn changes from only

. ',,,...,. . ,... ,, .. , .:e.,,.'j%.., ... , , %., ,* .. p:.,'. I

650 to 2300 D. These observations Indicate that thepolymerization process Is most likely a gelation process. Themostly linear macromolecular structure of the 650 D material istherefor quite different from the highly branched speciespresent in the 2300 D polymer. Of considerable Importance isthat the selectivity to ceramic products (83% S13N4 and 17%

amorphous carbon) remains unchanged despite the considerablechange in polymer molecular weight and macroscopic structure.These results suggest that it Is the monomer unit -H2 SiNMe-, at

the molecular level, that determines the selectivity to ceramicproducts.

This last conclusion, if valid, supports the concept that itis indeed feasible to design materials at the molecular level.Finally, transition metal catalyzed dehydrocoupling has nowbeen domonstrated for the formation of oligo-borazines fromBH3 and MeNH 2 . In addition, we have found a simple

condensation process that leads to polyiminotitanides which canbe used as perceramic precursors to titanium nitride.

BN OLIGOMER PRECURSORS BY CATALYTICDEHYDROCOUPLING

NMeMe I

I NeN

M Me Me MeM NMe NMMe

Pyrolysis at 8000Q: Ceramic Yield 60 Wt %Pyrolysis at 1600OC: Ceramic Yield 49 Wt %

TITANIUM NITRIDE PRECURSORS

TI(NMo2)4 + PrNH 2 -x Mo2 NH +

[TI(NMe 2 )3 (1-NPr)]x-

or

-[Ti(NMe 2)2(I-NPr)21x-

Pyrolysis at 8000C under NH 3 gives TIN

o

PROGRESS

* Developed an understanding of the mechanism(s)of the dehydrocoupling reaction.

" Learned how to polymerize oligomers of-[H2 SiNMeJx- and to control product rheological

properties.

* Demonstrated the feasibility of preparing thin,corrosion resistant coatings on metals and silicausing a preceramic polymer.

* Demonstrated the utility of using preceramicpolymers as binders for compression moldedS13 N4 .

* Prepared 10-100 gm preceramic fibers frompreceramic polymers.

" 6 publications in press or in preparation. 2patent applications. One major spin-off projectto Improve the strength of glass bottles.

* • , r", , .",*,. *, " *; .. '.* -.,' *.'. .. *. . * .,. , . V' ." .'" . 'V" .*" V'" ." *." "'

0P:.% fPs e%.