Characterization of the Evolution and Properties of Silicon Carbide DerivedFrom a Preceramic Polymer Precursor

Suraj C. Zunjarrao, Arif Rahman, and Raman P. Singh†

School of Mechanical and Aerospace Engineering, Oklahoma State University, Stillwater, Oklahoma 74075

This study reports on the fabrication and characterization

of polymer-derived amorphous and nano-grained SiC, bycontrolled pyrolysis of allylhydridopolycarbosilane (AHPCS)

under inert atmosphere. Processing temperatures and hold

times at final temperatures are varied to study the influence of

processing parameters on the structure and resulting properties.Chemical changes, phase transformations, and microstructural

changes occurring during the pyrolysis process are studied.

Polymer cross-linking and polymer to ceramic conversion is

studied using infrared spectroscopy (FTIR). Thermogravimetricanalysis (TGA) and differential thermal analysis (DTA) are

performed to monitor the mass loss and phase change as a func-

tion of temperature. X-ray diffraction studies are performed to

study the intermediate phases and microstructural changes.Hardness and modulus measurements are carried out using

instrumented nanoindentation to establish processing-property-

structure relationship for SiC derived from the polymer precur-sor. It is seen that the presence of nanocrystalline domains in

amorphous SiC significantly influences the modulus and hard-

ness. A nonlinear relationship is observed in these properties

with optimal mechanical properties observed for SiC processedto 1150°C for 1 h hold duration, having average grain size of

3 nm. In addition, bulk mechanical characterization, in terms

of biaxial flexure strength, is done for SiC–SiC particulate

composites purely derived from the polymer precursor.

I. Introduction

ALLYLHYDRIDOPOLYCARBOSILANE (AHPCS) is an ultrahighpurity polymer precursor that is gaining great interest

for potential use in wide-ranging, technologically advanced,and in some instances, critical applications.1 Its high ceramicyield, relatively low shrinkage, ability to be handled, and tobe processed in ambient conditions have attracted wideattention, especially as a precursor to SiC fibers and morerecently as a matrix material.2–5 AHPCS has been researchedas a binder for ceramic powders and matrix source for poly-mer-derived ceramic matrix composites.1,2,4,6–9 It has alsobeen used to join monolithic10 and composite ceramicparts.11 Researchers have also used AHPCS-derived SiC as apossible matrix for ceramic nuclear fuels.12,13 Previously,Singh and coworkers have researched ceramic foams basedon AHPCS-derived SiC and hollow alumino-silicatespheres,14 and also have pursued AHPCS-derived SiC as apotential matrix material for the fabrication of ceramic com-posite fuels for gas cooled fast reactors (GCFR).15

Nonetheless, there are limited measurements of mechanicalproperties of SiC derived from polymer precursors. Most ofthese studies have focused on processing and analytical char-

acterization. Investigation of mechanical properties has gen-erally been focused on the resulting composites rather thanspecifically on the AHPCS-derived SiC matrix. Bulk charac-terization of AHPCS-derived SiC sample, prepared by PIPprocess, has been reported by Mores et al.5 The materialswere characterized in terms of density (by immersionmethod), fracture toughness (by bulk V-notched beammethod), and hardness (by bulk Vickers indentation test).The highest values for fracture toughness and hardness werefound to be ~1.67 MPa m1/2 and 13 GPa, respectively.Porosity was found to adversely affect the properties. Evenin this study, however, the micro/mesoscale characterizationwas not done directly on the SiC matrix in an attempt todeconvolute the effect of porosity.

Mitchell et al.16 examined the nucleation and crystalliza-tion process in Si–C and Si–N–C systems produced by pyroly-sis of granulated polymethylsilane and found that processingtemperature and time both influence the crystal sizes. Grainsizes were calculated to be 5–20 nm. Nechanicky et al.17

reported TGA, FTIR, TEM, and XRD studies on a-SiC–b-SiC particulate reinforced composites prepared by PIP usinga hyper-branched polymethylsilane (mPMS). Grain growthof b-SiC was found to be seeded by the presence of a-SiCpolycrystals indicating processability of dense ceramic at atemperature range of 1000°C–1200°C.

Despite these studies, there is still a need to investigatethe fundamental processing–microstructure–property–appli-cation relationships for this organic polycarbosilane whichwill aid in the characterization of composites derived fromAHPCS.

II. Experimental Procedure

AHPCS has a nominal structure of [Si(CH2CH =CH2)2CH2]0.1[SiH2CH2]0.9.

5,18 Thus, it has a Si:C ratio that isclose to 1:1 and yields a near stoichiometric SiC uponpyrolysis. The polymer, designated as SMP-10, is acquiredfrom Starfire Systems Co., Malta, New York, USA. At roomtemperature, it is in the form of clear amber colored viscousliquid.

(1) Material FabricationOne of the main advantages of precursor polymer route toceramics is the ease of fabrication. Preparing SiC fromAHPCS simply requires heating the polymer to temperatureof above 900°C in an oxygen free atmosphere. When theprecursor is heated from room temperature, cross-linkingstarts at ~100°C and a cured green body is formed which fur-ther yields amorphous SiC at 900°C. To understand theeffect of processing parameters two different approaches wereundertaken as discussed in the following section.

(A) Effect of Processing Parameters: To study theeffect of heat treatment on physical properties, AHPCS-derived amorphous SiC (900°C) was heated to different tem-peratures of 1150°C, 1400°C, and 1650°C. The materialswere heated starting from room temperature at a rate of

R. Reidel—contributing editor

Manuscript No. 32369. Received November 26, 2012; approved February 7, 2013.†Author to whom correspondence should be addressed. e-mail: raman.singh@

okstate.edu

1869

J. Am. Ceram. Soc., 96 [6] 1869–1876 (2013)

DOI: 10.1111/jace.12273

© 2013 The American Ceramic Society

Journal

5°C/min and held at final temperature for 30 min. The pro-cess was done in a high temperature oven under a constantflow of argon.

To study the effect of holding the material at the finaltemperature for different durations on the structure of SiCformed and its mechanical properties, SiC samples weremade by pyrolysis of precursor, starting with the liquid form,and heating to 900°C, 1150°C, 1400°C, and 1650°C in singleruns at 4°C/min. The materials being processed were held atthe final temperatures (i.e. 900°C, 1150°C, 1400°C, and1650°C) for time durations of 2 min, 60 min, and 4 h indifferent experiments.

The resulting material was in the form of chunks of SiC.Samples for nanoindentation were prepared by embeddingpart of the chunk in epoxy and polishing. Samples for X-raydiffraction and TEM were prepared by crushing part of thechunk by hand using mortar and pestle for XRD, and usinga ball mill for TEM.

(B) Fabrication for Bulk Testing: Other than meso-scale characterization using nanoindentation, bulk character-ization was done using ring-on-ring (ROR) biaxial flexuretest. Polymer infiltration and pyrolysis (PIP) process wasused for fabrication of polymer-derived ceramic compositesfor bulk scale testing.

First, SiC powder was prepared by pyrolysis of AHPCS to900°C, 1150°C, 1400°C, and 1650°C in single runs at 4°C/min. The samples were held at the final temperature for60 min to ensure thermal equilibrium. These were crushedinto powder by planetary ball mill (PM-100; Retsch GmbH,Haan, Germany) in a tungsten carbide bowl (WC) with10 mm WC balls for 12 min at 300 rpm. These powders inturn form the reinforcing fillers in the next step of fabricatingcomposites using PIP process. Thus, by using SiC powderderived from AHPCS, we could make bulk samples made ofSiC purely derived from AHPCS alone. The powdersobtained from the initial pyrolysis of the precursor were thenmixed with small amounts of polymer precursor (3% byweight of the milled powder). These mixtures were com-pacted into short cylinders, 25.4 mm 9 15 mm, using a sim-ple steel ram-cylinder setup. Hand compaction with a load of~445 N (~100 lb) was sufficient to form plugs that could beeasily handled. The resulting compacts were then heated tothe final temperature (900°C, 1150°C, 1400°C, or 1650°C).To increase the material density, the samples were subjectedto multiple polymer infiltration cycles. Figure 1 shows theschematics of the complete PIP processing used in this study.The infiltration of the cylindrical plugs was carried out undervacuum, with repeated 1 h cycles for 4 h with intermediate1 min purges. After third infiltration and pyrolysis cycles, thecylindrical specimens were cut to get disks of 1 mm thicknessusing a precision sectioning saw (Isomet 1000; Buehler, 202Lake Bluff, Illinois). Further infiltration and pyrolysis, in the

aforementioned way, was carried out on the disks up to atotal of eight cycles. Work by Ozcivici et al.14 withpolymer-derived ceramic composites using this polymersystem has showed that eight infiltration cycles are sufficientto obtain the maximum achievable density. These diskswere tested for biaxial flexure modulus and strength usingring-on-ring test.

(2) Physical CharacterizationBulk density and porosity of the ceramic composite disksfabricated by PIP process using AHPCS-derived SiC powderand AHPCS-SiC matrix were determined using the buoyancymethod19 using a density determination kit in conjunctionwith a high-resolution analytical balance. Density measure-ments were also performed on finely crushed powders of SiCderived from AHPCS heated to 300°C, 500°C, 700°C, 900°C,1150°C, 1400°C, and 1650°C using helium pycnometry (Ultr-apycnometer 1000; Quantachrome, Boynton Beach, FL).This method yields mesoscale density of inherently porousmaterials, which cannot be accessed by bulk density measure-ments.

Fourier transform infrared spectroscopy (FTIR) analysisperformed on polymer precursor material heated from roomtemperature to final temperatures of 300°C, 500°C, 700°C,900°C, and 1150°C using a Nicolet Model Magna 760 FTIRspectrometer (Nicolet Instrument Corp., Madison, WI) withZnSe ATR crystal with 4 cm�1 resolution and average over256 scans.

Simultaneous differential thermal analysis (DTA) and ther-mogravimetric analysis (TGA) was performed to study theconversion of amorphous SiC to nanocrystalline SiC. Forthis, a small amount of SiC derived from polymer precursorpyrolyzed at 900°C was heated to 1300°C at a rate of 5°C/min under nitrogen atmosphere (STA 449 C Jupiter; NET-ZSCH, Selb/Bavaria, Germany).

X-ray diffraction studies were performed on SiC powderspyrolyzed at 900°C, 1150°C, 1400°C, and 1650°C. Powdersamples were prepared by wet milling in a planetary ball mill(PM-100; Retsch GmbH, Haan, Germany) for 4 h in ethanoland then mounted on a glass slide. Powder diffractionpatterns were collected using Scintag PAD-X automated dif-fractometer with a CuKa radiation (k = 0.1540 nm) using ascanning rate of 0.5° per min and operating at 45 kV and25 mA.

To characterize the microstructure of SiC formed at differ-ent processing temperatures, TEM studies were performedusing JEOL JEM-2100 (JOEL Ltd., Tokyo, Japan) ScanningTransmission Electron Microscope System with an EDAXGenesis 2000 EDS system, (Edax Inc., Mahwah, NJ) withthe help of Dr. Susheng Tan at the Oklahoma State Univer-sity Microscopy Laboratory.

Fig. 1. Schematic of PIP process for fabrication of bulk samples for bulk characterization.

1870 Journal of the American Ceramic Society—Zunjarrao et al. Vol. 96, No. 6

(3) Mechanical CharacterizationFor mesoscale characterization by nano-indentation, chunksof SiC material from the pyrolysis products, at different pro-cessing temperatures and different hold times, were mountedand samples were then indented using a sharp Berkovichdiamond indenter (NanoTest System; Micro Materials,Wrexham, UK) to a peak load of 10, 25, 50 and 100 mN foreach material. A total of 10 indentations were performed forevery sample.

Bulk mechanical characterization was done using ring-on-ring (ROR) biaxial tests. The ROR is an axisymmetrictest, where the disk is supported by a ring and loaded fromthe opposite side by another smaller concentric ring, asshown in Fig. 2. The concentric rings were made up of stain-less steel and had bullnose edges with a radius of 0.3125 mmtoward the loading side. This configuration led to a supportring of / 19.05 mm and a loading ring of / 6.35 mm.

The specimens were loaded in the ROR fixture using astandard mechanical testing frame. In the area underneaththe smaller ring there exists an equibiaxial tensile stress statewhere the initialization of fracture is expected. The flexurestrength can then be determined from the peak load at fail-ure Eq. 1 as,20

rROR ¼ 3P

2pt2ð1� tÞða2 � r2Þ

2R2þ ð1þ tÞln a

r

� �(1)

where υ is the Poisson ratio of the specimen and assumed tobe 0.20 for SiC, a is the radius of the support ring, r is theradius of the load ring, and R and t are the radius and thick-ness of the disk specimen, respectively.

The ROR configuration is preferred over the other equibi-axial tests as it subjects a greater portion of the specimen toan equibiaxial stress state and distributes the applied contactload over a larger area. This reduces the stress concentrationat the contact locations between the fixture and specimenand lowers the likelihood of specimen failure by crushing(invalid test data).

III. Results and Discussion

(1) Polymer–to–Ceramic ConversionFigure 3 shows the IR spectra obtained for products at dif-ferent temperatures; data are offset to aid comparison. Peaksattributed to C–H (stretching), Si–H (stretching), and Si–C(rocking) bonds were clearly observed in the ranges of 2800–3000 cm�1, 2000–2140 cm�1, and 870–1070 cm�1,1,21,22

respectively. It can be seen that relative intensity of all thepeaks initially increases with increasing temperature, which isa result of increase in cross-link density as the polymer cures.A gradual shift in Si–H peak toward lower wave numberswith increasing temperature suggests conversion from Si–H3

to Si–H2 to Si–H, as hydrogen is expelled in the form of gas.The broad peak attributed to several C–H bonds reduces,and eventually disappears, along with the Si–H peak at1150°C as hydrogen is completely removed. Small peaksresulting from the presence of mono-substituted alkenes in

the polymer appear ~900 cm�1 (d) at 300°C and disappear athigher temperatures. Peaks attributed to CH3 (bending)(a and b) and SiCH2Si bonding (c) are also identified in theIR spectra. A small amount of hydrogen appears to be pres-ent in the system even at 900°C, whereas at 1150°C only aSiC peak shows complete conversion of polymer into SiC.The changes in chemical composition of AHPCS used in thisstudy, upon heating to different temperatures, have beenreported in literature.23,24 Moraes has reported presence ofcarbon clusters between 1100°C and 1600°C. In addition,they have also reported peaks corresponding to species ofSi–O and Si–O–C compounds. Additional mass loss abovethe crystallization temperature can be attributed to the lossof oxygen. Elemental composition at various temperatureswas reported by Moraes et al.24 Carbon to silicon mole ratiowas found to be close to 1.13 at various pyrolysis tempera-tures indicating slightly carbon-rich compounds formed atdifferent temperatures.

Figure 4 shows the ceramic yield obtained as a function ofdecomposition temperature. The loss in weight is attributedto the loss of low-molecular weight oligomers and hydrogengas.1 Marginal loss in mass was observed beyond 700°C and~72%–74% ceramic yield was obtained in the range 900°C–1650°C. In a separate study on reaction kinetics during thepyrolysis of AHPCS,25 the polymer pyrolysis was character-ized as a three-step process consisting of volatilization, cross-linking and crystallization; and activation energies for thevolatilization and cross-linking were determined as 83.1 and

Fig. 2. Schematic of fixture for testing biaxial flexure propertiesusing ROR.

80012001600200024002800320036004000Wave number (cm-1)

I R A

bsor

banc

e (a

.u.)

C - H Si - CSi - H

a

a b

b

c

c

d

d

(v)

(iv)

(iii)

(ii)

(i)

Fig. 3. IR Spectra for AHPCS heated to 300°C (i), 500°C (ii),700°C (iii), 900°C (iv), and 1150°C (v).

40

50

60

70

80

90

100

0

0.5

1

1.5

2

2.5

3

3.5

4

Mass LossDensity

0 300 600 900 1200 1500 1800

Sam

ple

Mas

s Va

riatio

n (%

)

Density (g/cc)

Processing Temperature (°C)

Fig. 4. Mass loss and density variation as a function oftemperature for AHPCS pyrolyzed to different temperatures.

June 2013 SiC from Preceramic Polymer 1871

149.7 kJ/mol, respectively, using the mass loss data discussedearlier.

Conversion of polymer precursor into ceramic materialwas also tracked in terms of density (shown in Fig. 4 alongwith sample mass variation) of pyrolysis products at differentstages of heating. Density measurements were performedusing helium pycnometry on finely crushed powders. Startingwith a liquid AHPCS having a density of 0.997 g/cc (as men-tioned by Starfire Systems Inc., USA), a dry and partiallycross-linked solid with density ~1.07 g/cc is obtained at300°C. Further heating results in more cross-linking accom-panied by the loss of low-molecular weight oligomers andhydrogen gas. As the processing temperatures increase, den-sity is observed to increase steadily until it reaches valuesthat are close to theoretical density for SiC at 1150°C.

(2) Amorphous to Nano-crystalline ConversionEvidence of crystallization in the SiC samples derived fromAHPCS was first seen in differential thermal analysis (DTA)experiments. Figure 5 shows the DTA and TG curvesobtained. A distinct peak is seen in the DTA curve around1100°C which is attributed to crystallization in the material.The occurrence of this peak coincides with the change inslope in the TG curve. Further evidence of onset of crystalli-zation at this temperature was seen in the X-ray diffraction(XRD) and electron-diffraction patterns obtained for thepyrolyzed products.

Figure 6 shows the XRD patterns obtained for varioussamples; data are offset to aid comparison. Amorphous SiCformed at 900°C shows a greatly diffused peak whereas thepeak intensity increases as the processing temperaturesincrease. Gradual growth of SiC peaks at 2h values of 35.7°,60.2° and 72.0° suggests increasing ordering as nano-crystal-line domains form and grow in amorphous SiC. It is notedthat small peaks for residual tungsten carbide (WC), fromthe grinding media, are seen in the patterns. Also, eventhough a peak for WC lies very close to the SiC peak at35.7°, the prominent peaks at this 2h are attributed to SiCbecause the WC peak at 35.6° and 48.3° are expected to beof same intensity according to JCPDS (ICCD 29–1131). Anestimate of the crystallite size was obtained from the peakbroadening using the Debye-Scherrer equation.26 The crystal-lite sizes were determined to be ~4, 5, and 11 nm at 1150°C,1400°C, and 1650°C, respectively.

Further evidence on the presence of amorphous SiC at900°C and its polycrystalline nature at higher temperaturewas seen from transmission electron microscopy. Figures 7–10 show the TEM micrographs obtained for SiC processed at900°C, 1150°C, 1400°C, and 1650°C and held at the final

98.6

98.8

99

99.2

99.4

99.6

99.8

100

-0.24

-0.16

-0.08

0

0.08

0.16

0.24

0.32

0 200 400 600 800 1000 1200 1400

Thermogravimetric data (TG)Differential Thermal Analysis (DTA)Pe

rcen

tage

of i

nitia

l wei

ght (

%)

DTA

(µµ µµV/mg)

Temperature (°C)

1095°C

Fig. 5. DTA and TG curves for AHPCS heated to 1300°C at a rateof 5°C/min.

20 30 40 50 60 70 80

WC (ICCD: 25-1047)SiC (ICCD: 29-1131)

2 Theta (Degree)

Rel

ativ

e In

tens

ity (a

.u.)

900°C

1150°C

1400°C

1650°C

Fig. 6. Powder diffraction patterns of SiC derived from AHPCSheated to 900°C, 1150°C, 1400°C, and 1650°C.

Fig. 7. TEM micrograph for SiC derived from AHPCS heated to900°C and hold duration of 4 h. While the microstructure is mostlyamorphous, some areas of crystalline regions are seen.

Fig. 8. TEM micrograph for SiC derived from AHPCS heated to1150°C and hold duration of 4 h. Large number of crystalline areawith average size of 2–3 nm can be seen.

1872 Journal of the American Ceramic Society—Zunjarrao et al. Vol. 96, No. 6

temperature for 4 h. While at 900°C, SiC is mostly seen inamorphous form, small domains of ordered regions are seenat some places. This is due to the long hold duration at thistemperature. A lot of small crystalline regions are seen inSiC processed to 1150°C. Similar, but larger domains areseen at 1400°C. These domains appear to be surroundedby amorphous phase as the one seen at 900°C. The TEMmicrograph for 1650°C clearly shows large domains ofwell-ordered material. All these domains are showing thenanocrystalline SiC.

Furthermore, selected area electron-diffraction (SAED)patterns were obtained during transmission electron micros-copy (TEM) studies on these SiC samples and are shown inFig. 11. As seen in Fig. 11(a), greatly diffused concentricrings for SiC processed at 900°C suggest a largely amorphousstructure. These rings are seen to gradually become distinctand sharp for SiC processed at higher temperatures[Fig. 11(b)–(d)], which is suggestive of growing crystallitesize. SAED patterns obtained for SiC processed at 1650°C[Fig. 11(d)] shows tiny bright specks intermittently along the

rings. These are typically seen for nano-sized polycrystallinematerials.27

(3) Mechanical Property CharacterizationMechanical properties of SiC derived from AHPCS pyro-lyzed to 900°C, 1150°C, 1400°C, and 1650°C and held at thefinal temperature of 2 min, 60 min, and 4 h were character-ized in terms of hardness and elastic modulus using instru-mented nanoindentation.

Nanoindentation was performed using a Berkovich tip andloaded up to peak loads of 10, 25, 50, 75 and 100 mN.Indentations with multiple loads were performed to deter-mine the loads required to obtain sufficient displacements, soas to obtain reliable results. At 25 mN, data were consistentover multiple tests and a good nanoindentation depth of 250–300 nm was observed. Hence, indentation data being dis-cussed further are from experiments performed with peakindentation load of 25 mN.

Figures 12 and 13 show the hardness and modulus as afunction of processing temperature. Error bars show stan-dard deviation. Nominal values of hardness and moduluswere obtained for SiC that was pyrolyzed at 900°C with ahold time of 1 h and 4 h. For 25 mN peak load, these valueswere ~160 GPa for modulus and ~23 GPa for hardness,which are typical for a-SiC. Material processed at 900°C witha hold time of 4 h showed higher values of ~197 GPa formodulus and ~25 GPa for hardness. There is considerablevariation seen in hardness and modulus as a function of tem-perature and these variations are also dependent on the holdtime at final temperatures. Figure 12 shows three differenttrends seen in the hardness values as a function of processingtemperature, for materials held for three different time dura-tions at the final temperature. Materials held for 2 min and1 h hold durations show a similar trend of an initial increasein the hardness values for materials processed at 1150°C andthen a drop in hardness progressively as material wasprocessed to higher temperatures of 1400°C and 1650°C.Whereas material held for 4 h shows a progressive drop inhardness with increasing processing temperatures. It is alsointeresting to note that the hardness values are very close formaterials held for 2 min and 1 h hold durations at the lowestand highest processing temperatures being considered that isat 900°C and 1650°C. The highest values for hardness of

Fig. 9. TEM micrograph for SiC derived from AHPCS heated to1400°C and hold duration of 4 h. Crystalline regions of average sizeof 5–6 nm are seen.

Fig. 10. TEM micrograph for SiC derived from AHPCS heated to1650°C and hold duration of 4 h. Large and distinct crystallineregions of 10–15 nm are seen.

(a) (b)

(c) (d)

Fig. 11. SAED patterns for SiC derived from AHPCS heated to (a)900°C, (b) 1150°C, (c) 1400°C, and (d) 1650°C.

June 2013 SiC from Preceramic Polymer 1873

around 30 GPa were seen for the material processed at1150°C for a hold time of 1 h. This is ~52% higher than thelowest value, which is seen for material processed at 1650°Cand a hold time of 4 h. Moraes et. al5 have reported Vickershardness of 9–12 GPa, measured at a load of 10 N forAHPCS-derived material heat treated in the range of 1000°C–1600°C. It should be noted that Vickers hardness dependson porosity rather than microstructure whereas nanoindenta-tion provides better correlation between microstructure andmechanical properties. Thus, hardness measured in the exper-iments here is considerably higher than that reported byMoraes et. al.

Similarly, Fig. 13 shows three different trends seen in themodulus values, as a function of processing temperature, formaterials held for three different time durations at the finaltemperature. While values for modulus were very close formaterial processed for 2 min and 1 h hold durations at tem-peratures of 900°C and 1650°C, the values peaked at 1150°Cfor 1 h hold samples and at 1400°C for those held for 2 min.In both these cases, higher modulus values were seen at inter-mediate processing temperatures of 1150°C and 1450°C ascompared to 900°C and 1650°C. Samples held for 4 h at thefinal temperature show lowest values of about 197 GPa,

which increases with processing temperature to about205 GPa at 1650°C. Similarly, highest values of moduluswere observed for materials processed at 1150°C for a holdtime of 1 h. Highest modulus values were ~218 GPa, whichis 37% higher than the lowest value which is seen for mate-rial processed with 2 min hold time at 900°C.

Figures 14 and 15 show the hardness and modulus as afunction of the hold durations at final temperature for SiCprocessed to different temperatures. In general, for bothhardness and modulus, progressive increase is seen for SiCprocessed to final temperature of 900°C as the hold time at900°C is increased. On the other hand, SiC processed to1650°C shows continuous drop in hardness as the hold timeis increased from 2 min to 4 h. The values of modulus donot change significantly as a function of hold time for SiCprocessed to 1650°C. For both, hardness and modulus, mate-rial processed to intermediate temperatures of 1150°C and1400°C shows slightly higher values for hold time of 1 h ascompared to other hold times.

These results, in conjunction with the microstructuralinformation at different temperatures, make for interestingobservations. Amorphous SiC is formed at 900°C with2 min hold time and increasing the hold time further densi-

0

5

10

15

20

25

30

35

800 1000 1200 1400 1600 1800

H-2minH-1hH-4h

Har

dn

ess

(GP

a)

Processing Temperature (°C)

Hold times

Fig. 12. Hardness determined by nanoindentation for SiC derivedfrom AHPCS heated to 900°C, 1150°C, 1400°C, and 1650°C, as afunction of processing temperature.

100

150

200

250

800 1000 1200 1400 1600 1800

H-2minH-1hH-4h

Mo

du

lus

(GP

a)

Processing Temperature (°C)

Hold times

Fig. 13. Modulus determined by nanoindentation for SiC derivedfrom AHPCS heated to 900°C, 1150°C, 1400°C, and 1650°C, as afunction of processing temperature.

0

5

10

15

20

25

30

35

0 50 100 150 200 250 300

900°C1150°C1400°C1650°C

Har

dn

ess

(GP

a)

Hold time (min)

Fig. 14. Hardness determined by nanoindentation for SiC derivedfrom AHPCS heated to 900°C, 1150°C, 1400°C, and 1650°C, as afunction of hold duration at final temperature.

0

50

100

150

200

250

300

0 50 100 150 200 250 300

900°C1150°C1400°C1650°C

Mo

du

lus

(GP

a)

Hold time (min)

Fig. 15. Modulus determined by nanoindentation for SiC derivedfrom AHPCS heated to 900°C, 1150°C, 1400°C, and 1650°C, as afunction of hold duration at final temperature.

1874 Journal of the American Ceramic Society—Zunjarrao et al. Vol. 96, No. 6

fies the material resulting in improvements in its mechanicalproperties. With a hold time of 4 h, the microstructureremains mostly amorphous at 900°C, but there are somecrystalline regions formed which explain the small increasein mechanical properties observed for SiC processed at900°C for 4 h. Nanocrystalline SiC with average crystal sizeof about 3 nm is formed at 1150°C and this greatlyinfluences its mechanical properties. SiC formed at higherprocessing temperatures of 1400°C and 1650°C had largergrain sizes, but lower mechanical properties. This could bedue to the classical Hall-Petch effect, which ascribes increas-ing mechanical properties for smaller grains sizes to disloca-tion pileup on grain boundaries. Varying the hold durationat final temperature has a limited effect on the mechanicalproperties. While hardness and modulus increases for amor-phous SiC processed to 900°C with increasing hold duration,for SiC processed to higher temperatures, these propertiesgenerally increased only marginally for 1 h hold and gener-ally dropped for 4 h hold durations. Thus, there appears tobe an optimum processing temperature (~1150°C) thatresults in just the right grain size to achieve higher mechani-cal properties. Similarly, an optimum hold time of 1 hresults in better properties as compared to 2 min and 4 hhold durations.

Table I lists the biaxial flexure strength obtained for thebulk samples composed purely of SiC derived from AHPCSand fabricated using the PIP process. An increase in over50% is observed in the biaxial strength for SiC processed to1150°C and 1400°C. Unfortunately, specimen processed to1650°C suffered oxidation and could not be tested.

IV. Conclusion

The use of high-pressure and/or high-temperature processingis a basic requirement for traditional powder-based fabrica-tion of ceramic components. These processing conditions notonly increase the energetic requirements but also place limitson the types of materials that can be produced. Despiteseveral different techniques developed to produce ceramics,the search for a technique that offers a combination ofadvantages such as low temperature processing, low costs,ability to fabricate complex shapes, high-density yields andmicrostructure control, still continues. Polymer-derivedceramics is an attractive alternative because of pressure-lessand lower temperature processing and near-net shape fabri-cation. Control over the microstructure of SiC formed bypyrolysis of preceramic polymer precursor is relatively easyand is presented here by fabrication of amorphous and nano-crystalline SiC. Based on the work reported in this investiga-tion, the conclusions drawn are as follows:

1. The chemical and structural changes during the pyro-lysis of allylhydridopolycarbosilane were studied bytracking mass loss, density changes, chemical bonding,and evolution of microstructure up to 1650°C. Ahydrogen-free ceramic yield of 72%–74% was obtainedin the range 900°C–1650°C using very slow heatingrate of 4°C/min. IR spectroscopy observations revealedpresence of hydrogen in SiC pyrolyzed at 900°C. Theyalso provide insight into the decomposition of polymer

precursor by dissociation of Si–H and C–H bondsalong with the evolution of a Si–C network.

2. The onset of crystallization was seen close to 1100°Cleading to the formation of amorphous/nanocrystallineSiC. Crystallite size, as estimated from powder diffrac-tion patterns, was found to be in tens of nanometersat 1650°C. Clear evidence of nanocrystalline SiCformed with different crystal sizes at different process-ing temperatures was seen through transmission elec-tron microscopy along with a study of selected areaelectron diffraction.

3. Nanocrystalline SiC, similar to other nanocrystallineceramics, is expected to have superior properties. Signifi-cant improvements in mechanical properties were seenafter the onset of crystallization in SiC. Mechanicalproperties of amorphous and amorphous/nano-crystal-line SiC derived from AHPCS, pyrolyzed to differenttemperatures and different hold times at final tempera-ture, were characterized in terms of elastic modulusand hardness by instrumented nanoindentation. Themechanical properties were influenced by both process-ing temperature and hold durations. The highest valuesfor hardness, of around 30 GPa, were seen for the mate-rial processed at 1150°C for a hold time of 1 h. Thesewere ~52% higher than the lowest value, which wasseen for materials processed at 1650°C with a hold timeof 4 h. Similarly, highest modulus values, observed forthe same samples, were ~218 GPa, which were 37%higher than the lowest value which were seen for materi-als processed with 2 min hold time at 900°C. Thus, opti-mum processing temperature of 1150°C and holdduration of 1 h gave the best values for hardness andmodulus.

Acknowledgment

The authors thank DOE for financial support on this project (Award No.DE-FC07-05ID14673).

References

1L. V. Interrante, C. W. Whitmarsh, W. Sherwood, H. J. Wu, R. Lewis,and G. Maciel, “High Yield Polycarbosilane Precursors to Stoichiometric SiC.Synthesis, Pyrolysis and Application,” in: Proceedings of the 1994 MRSSpring Meeting, April 4, 1994 - April 8, 1994, Mater. Res. Soc., San Fran-cisco, CA, USA, 593–603 (1994).

2M. Z. Berbon, D. R. Dietrich, D. B. Marshall, and D. P. H. Hasselman,“Transverse Thermal Conductivity of Thin C/SiC Composites Fabricated bySlurry Infiltration and Pyrolysis,” J. Amer. Ceram. Soc., 84, 2229–34 (2001).

3S. M. Dong, Y. Katoh, A. Kohyama, S. T. Schwab, and L. L. Snead,“Microstructural Evolution and Mechanical Performances of SiC/SiC Com-posites by Polymer Impregnation/Microwave Pyrolysis (PIMP) Process,”Ceram. Int., 28, 899–905 (2002).

4L. V. Interrante, J. M. Jacobs, W. Sherwood, and C. W. Whitmarsh,“Fabrication and Properties of Fiber- and Particulate-Reinforced SiC MatrixComposites Obtained with (A)HPCS as the Matrix Source,” Key Eng. Mater.,127–131, 271–8 (1997).

5K. V. Moraes and L. V. Interrante, “Processing, Fracture Toughness, andVickers Hardness of Allylhydridopolycarbosilane-Derived Silicon Carbide,” J.Amer. Ceram. Soc., 86, 342–6 (2003).

6M. Kotani, T. Inoue, A. Kohyama, Y. Katoh, and K. Okamura, “Effect ofSiC Particle Dispersion on Microstructure and Mechanical Properties of Poly-mer-Derived SiC/SiC Composite,” Mater. Sci. Eng., A, 357, 376–85 (2003).

7R. Sreeja, B. Swaminathan, A. Painuly, T. V. Sebastian, and S. Packirisam-y, “Allylhydridopolycarbosilane (AHPCS) as Matrix Resin for C/SiC CeramicMatrix Composites,” Mater. Sci. Eng. B: Solid-State Mater. Adv. Technol.,168, 204–7 (2010).

8H. Z. Wang, X. D. Li, J. Ma, G. Y. Li, and T. J. Hu, “Multi-Walled CarbonNanotube-Reinforced Silicon Carbide Fibers Prepared by Polymer-DerivedCeramic Route,” Compos. Part A: Appl. Sci. Manuf., 43, 317–24 (2012).

9H. Z. Wang, X. D. Li, J. Ma, G. Y. Li, and T. J. Hu, “Fabrication ofMulti-Walled Carbon Nanotube-Reinforced Carbon Fiber/Silicon CarbideComposites by Polymer Infiltration and Pyrolysis Process,” Compos. Sci.Technol., 72, 461–6 (2012).

10J. Zheng and M. Akinc, “Green State Joining of SiC Without AppliedPressure,” J. Amer. Ceram. Soc., 84, 2479–83 (2001).

11C. A. Lewinsohn, R. H. Jones, P. Colombo, and B. Riccardi, “SiliconCarbide-Based Materials for Joining Silicon Carbide Composites for FusionEnergy Applications,” J. Nuc. Mater., 307–311, 1232–6 (2002).

Table I. Biaxial Strength of the SiC–SiC CompositesFabricated by PIP Route and Tested Using

ROR Biaxial Flexure Test

Material Biaxial Strength (MPa) Pyrolysis Temperature (°C)

SiC-900 55 � 2 900SiC-1150 80 � 8 1150SiC-1400 85 � 5 1400SiC-1650 Oxidized 1650

June 2013 SiC from Preceramic Polymer 1875

12A. A. Solomon, J. Fourcade, S. G. Lee, S. Kuchibhotla, S. Revankar,R. Latta, P. L. Holman, and J. K. McCoy, “The Polymer Impregnation andPyrolysis Method for Producing Enhanced Conductivity LWR Fuels,” in: 2004International Meeting on LWR Fuel Performance, September 19, 2004 - Sep-tember 22, 2004, Amer. Nucl. Soc., Orlando, FL, United states, 146–55 (2004).

13C. Shih, J. S. Tulenko, and R. H. Baney, “Low-Temperature Synthesis ofSilicon Carbide Inert Matrix Fuel Through a Polymer Precursor Route,”J. Nuc. Mater., 409, 199–206 (2011).

14E. Ozcivici and R. P. Singh, “Fabrication and Characterization of CeramicFoams Based on Silicon Carbide Matrix and Hollow Alumino-SilicateSpheres,” J. Amer. Ceram. Soc., 88, 3338–45 (2005).

15A. K. Singh, S. C. Zunjarrao, and R. P. Singh, “Processing of UraniumOxide and Silicon Carbide Based Fuel Using Polymer Infiltration and Pyroly-sis,” J. Nuc. Mater., 378, 238–43 (2008).

16B. S. Mitchell, H. Zhang, N. Maljkovic, M. Ade, D. Kurtenbach, andE. Muller, “Formation of Nanocrystalline Silicon Carbide Powder FromChlorine-Containing Polycarbosilane Precursors,” J. Amer. Ceram. Soc., 82,2249–51 (1999).

17M. A. Nechanicky, K. W. Chew, A. Sellinger, and R. M. Laine, “Α-SiliconCarbide/b-Silicon Carbide Particulate Composites via Polymer Infiltration andPyrolysis (PIP) Processing Using Polymethylsilane,” J. Eur. Ceram. Soc., 20,441–51 (2000).

18C. K. Whitmarsh and L. V. Interrante, “Synthesis and Structure of aHighly Branched Polycarbosilane Derived From (Chloromethyl)Trichlorosi-lane,” Organometallics, 10, 1336–44 (1991).

19ASTM Standard C830-00, Standard Test Methods for ApparentPorosity, Liquid Absorption, Apparent Specific Gravity, and Bulk Density of

Refractory Shapes by Vaccum Pressure. ASTM int., West Conshohocken,PA, 2000.

20J. E. O. Ovri, “Parametric Study of the Biaxial Strength Test for BrittleMaterials,” Mater. Chem. Phys., 66, 1–5 (2000).

21M. S. Lee and S. F. Bent, “Bonding and Thermal Reactivity in Thina-SiC:H Films Grown by Methylsilane CVD,” J. Phys. Chem. B, 101, 9195–205 (1997).

22F. I. Hurwitz, T. A. Kacik, X. Y. Bu, J. Masnovi, P. J. Heimann, and K.Beyene, “Pyrolytic Conversion of Methyl- and Vinylsilane Polymers to Si-CCeramics,” J. Mater. Sci., 30, 3130–6 (1995).

23L. V. Interrante, C. K. Whitmarsh, T. K. Trout, and W. R. Schmidt,“Synthesis and Pyrolysis Chemistry of Polymeric Precursors to SiC andSi3N4”; pp. 243–54 in Inorganic and Organometallic Polymers with SpecialProperties, R. Laine. Springer, Netherlands, 1992.

24K. V. Moraes, “The Densification, Crystallization and Mechanical Porper-ties of Allylhydridoploycarbosilane-Derived Silicon Carbide”; PhD thesis in:Materials Science and Engineering, Rensselaer Polytechnic Institute, USA,2000.

25X. Wang, S. C. Zunjarrao, R. P. Singh, and H. Zhang, “Advanced Modelof Silicon Carbide Based Uranium Ceramic Nuclear Fuel Production,”J. Thermophy. Heat Trans., 23, 286–93 (2009).

26B. D. Cullity, Elements of X-ray Diffraction. Addison-Wesley PublishingCo. Unc., London, 1978.

27D. B. Williams and C. B. Carter, Transmission Electron Microscopy:A Textbook for Materials Science. Kluwer Academic/Plenum Publishers,New York, 1996. h

1876 Journal of the American Ceramic Society—Zunjarrao et al. Vol. 96, No. 6