A

Tcic1TW©

K

1

ieS2c(tftciiaaitaa

0h

Available online at www.sciencedirect.com

Journal of the European Ceramic Society 33 (2013) 2943–2951

Silicon carbide–titanium diboride ceramic composites

Derek S. King ∗, William G. Fahrenholtz, Greg E. HilmasMissouri University of Science and Technology, Materials Science and Engineering Department, Rolla, MO 65409, United States

Received 21 November 2012; received in revised form 12 March 2013; accepted 28 March 2013Available online 2 May 2013

bstract

he effect of TiB2 content on mechanical properties of silicon carbide–titanium diboride ceramic composites was studied. The hardness of theeramics decreased from 27.8 GPa for nominally pure SiC to 24.4 GPa for nominally pure TiB2. In contrast, fracture toughness of the ceramicsncreased from 2.1 MPa m1/2 for SiC to ∼6 MPa m1/2 for SiC with TiB2 contents of 40 vol.% or higher. Flexure strengths were measured for threeomposites containing 15, 20, and 40 vol.% TiB2 and analyzed using a two parameter Weibull analysis. The Weibull modulus increased from 12 for

5 vol.% TiB2 to 17 for 20 and 40 vol.% TiB2. Microstructural analysis revealed microcracking in the ceramics containing 20 and 40 vol.% TiB2.he ceramic containing 40 vol.% TiB2 had the best combination of properties with a fracture toughness of 6.2 MPa m1/2, hardness of 25.3 GPa,eibull modulus of 17, and a strength of 423 MPa.

2013 Elsevier Ltd. All rights reserved.

cal pr

oa

tdpfi3hdSSs(ttiwi

eywords: Silicon carbide; Titanium diboride; Particulate composite; Mechani

. Introduction

Silicon carbide (SiC) is a strong, hard, and chemicallynert ceramic used in several applications that involve extremenvironments.2–6 In armor applications, the high hardness ofiC, which is commonly reported to be in the range of0–27 GPa, is advantageous for projectile defeat.7,8 Like mosteramics, SiC is brittle due to its low fracture toughness2–5 MPa m1/2), whereas a high fracture toughness is advan-ageous for multi hit capability in armor materials.5,6,9,10 Theracture toughness of SiC based ceramics can be improvedo 6–9 MPa m1/2 with additives that promote densification andrack propagation along the grain boundaries as well as increas-ng the grain size.2,9,11 However, an increase in toughnesss typically accompanied by a drop in hardness using thispproach.9,10,12–14 Flinders et al. reported a hardness of 20 GPand a fracture toughness of 2.6 MPa m1/2 for SiC with no sinter-ng additives, but saw the hardness drop to 14 GPa while fractureoughness increased to 6.7 MPa m1/2 for SiC containing 3 wt.%

9

luminum as a sintering aid. Titanium diboride (TiB2) exhibits high hardness (25–35 GPa) and may help combat the tradeoff

∗ Corresponding author. Tel.: +1 573 341 7205; fax: +1 573 341 6934.E-mail address: dsk2nf@mst.edu (D.S. King).

aistcns

955-2219/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.jeurceramsoc.2013.03.031

operties; Armor

f decreasing hardness with increasing fracture toughness whendded as a reinforcing phase in SiC.15–18

Studies on SiC–TiB2 composites have mainly focused onhe increase in toughness, which has been attributed to crackeflection and bridging effects due to the presence of TiB2articles.18–22 Alpha-SiC (6H) has a thermal expansion coef-cient (CTE) along the c-direction of 4.16 × 10−6/◦C and.63 × 10−6/◦C along the a-direction. TiB2, however, has aigher CTE in each direction: 8.6 × 10−6/◦C along the c-irection and 6.6 × 10−6/◦C along the a-direction.15,23,24 ForiC–TiB2 ceramics, the mismatch in CTE values puts theiC matrix in compression and the TiB2 particles in ten-ion after cooling from the typical densification temperatures1900–2200 ◦C).18,25,26 The mismatch promotes toughening dueo the thermal residual stresses and resulting crack deflectionhat can occur as cracks are drawn from the SiC matrix, whichs in compression, toward the interface with the TiB2 particleshere the highest tensile stresses exist.27 Blanc et al. showed an

ncrease in toughness from 3.5 MPa m1/2 to 3.9 MPa m1/2 with hardness decrease from 30 GPa to 23 GPa as TiB2 contentncreased from 5 vol.% to 15 vol.%.16 Bucevac et al., mea-uring toughness using the indentation method, observed aoughness boost from 4.3 MPa m1/2 to 5.3 MPa m1/2 as TiB

2ontent increased from 12 vol.% to 24 vol.%, although hard-ess data were not presented.28 Along with thermal residualtresses that develop due to thermal expansion mismatch, when

2 ean C

tcmpsap

S0mns2acmi

2

fawG1S(muSmdpeSb

e(gwGnpaathtosao2f

tat5mB5t

AwTmlfbil5BtfspmIA

CtHkpadiwecwEtir

K

tNc

944 D.S. King et al. / Journal of the Europ

he residual stresses are large enough, stress-induced micro-racking can also occur, enhancing crack deflection and theeasured fracture toughness.29–33 In areas with large enough

articles, thermal stresses may become significant and inducepontaneous microcracking, which can render toughening mech-nisms such as thermal residual toughening ineffective near sucharticles.29,34–37

The purpose of this paper is to examine the properties ofiC–TiB2 ceramic composites with TiB2 contents ranging from

vol.% to 100 vol.%. Fracture toughness and hardness wereeasured for each composition. Based on their measured hard-

ess and fracture toughness values, Young’s modulus and flexuretrength were measured for SiC–TiB2 ceramics containing 15,0, and 40 vol.% TiB2. The results show how both hardnessnd fracture toughness change over the range of SiC–TiB2eramics. The Weibull analysis also shows the importance oficrostructural changes on the performance of SiC–TiB2 ceram-

cs.

. Experimental procedure

SiC–TiB2 ceramic composites with TiB2 contents rangingrom 0 vol.% to 100 vol.%, were batched using commerciallyvailable SiC (H.C. Starck; Grade UF-25; �-SiC; Newton, MA,ith an average particle size of 0.14 �m) and TiB2 (Momentive;rade HCT-F; Columbus, OH, with an average particle size of.45 �m) powders. To promote densification, 1 wt.% B4C (H.C.tarck; Grade HS) and 2 wt.% C, in the form of phenolic resinGP 2074, Georgia Pacific Chemicals, Atlanta, GA), were ballilled with the SiC and TiB2 powders in a polyethylene jar

sing TiB2 milling media and acetone. For the nominally pureiC specimen, the same procedure was used except that SiCedia were used instead of TiB2 media. Compositions were

esignated as SiC–“X”TiB2 where “X” is the nominal volumeercentage of TiB2. Resulting slurries were dried by rotaryvaporation to minimize segregation of the constituents. TheiC–TiB2 powders were then ground and sieved to −60 meshefore hot-pressing.

Initial billets of each composition, nominally 2.5 cm in diam-ter, were densified by hot pressing in a graphite element furnaceThermal Technologies, HP20-3060, Santa Rosa, CA) using araphite die. To minimize reaction between dies and billets, diesere lined with graphite paper (0.254 mm thick GTA, Leaderlobal Technologies, Deer Park, TX) that was coated with boronitride spray (SP-108, Materion, Milwaukee, WI) before theowder was loaded. Die assemblies were heated under vacuum at

rate of 50 ◦C/min. Isothermal holds were employed at 1450 ◦Cnd 1650 ◦C to facilitate removal of oxide contamination fromhe surfaces of the powder particles. At 1450 ◦C the isothermalold time was 2 h. The furnace was then held at 1650 ◦C untilhe chamber pressure returned to the nominal vacuum pressuref ∼27 Pa (200 mtorr). Above 1650 ◦C, and up to the final den-ification temperature, the atmosphere was changed to flowing

5

rgon at a pressure of ∼10 Pa (1 atm) and a uniaxial pressuref 32 MPa was applied. The final densification temperature was100 ◦C for nominally pure SiC and TiB2 billets, but 1980 ◦Cor the intermediate compositions. Pressing ceased after ram

pBad

eramic Society 33 (2013) 2943–2951

ravel had stopped for a period of 10 min, and the furnace wasllowed to cool to room temperature naturally. For compositionshat were selected for flexure testing, billets that were nominally5 mm by 55 mm by 5 mm were hot pressed in a graphite ele-ent furnace (Thermal Technologies Inc., Model HP50-7010G).ecause of the larger die size, and increased powder volume,0 ◦C was added to the isothermal hold temperatures, includinghe final densification temperature.

The bulk densities were measured for all billets using therchimedes’ method. The theoretical density for each compositeas calculated based on the nominal volumetric ratios of SiC andiB2. The 55 mm by 55 mm by 5 mm billets were machined intoechanical test bars using an automated surface grinder (Cheva-

ier Machinery Inc., Model FSG-3A818, Santa Fe Springs, CA)ollowing the guidelines of ASTM C1161-02c for B-bars (3 mmy 4 mm by 45 mm). Specimens were tested in four point bend-ng with a fully articulated fixture (20 mm upper span × 40 mmower span), using a screw-driven load frame (Instron, Model881, Norwood, MA) that was computer controlled (Instron,luehill 2, Norwood, MA). For SiC–15TiB2, 39 bars were tested

o failure while 36 bars were tested for SiC–20TiB2, and 38or SiC–40TiB2. For each of these three compositions, failuretrengths of all of the specimens were analyzed using a twoarameter Weibull distribution. Elastic constants were deter-ined using the impulse excitation method (Grindosonic Mk5

ndustrial, J.W. Lemmens Inc., Heverlee, Belgium) followingSTM C1259-08e1.Vickers’ hardness was measured (Struers Inc., Duramin 5,

leveland, OH) using a load of 9.8 N (1 kg) with a 15 s dwellime. Reported values were an average of 10 hardness indents.ardness specimens were prepared by mounting sections of bro-en flexure bars and cross-sectioned billets in an epoxy resin andolishing to a mirror finish using successively finer diamondbrasives to a 0.25 �m surface finish. Fracture toughness wasetermined using direct crack measurements. Specimens werendented (Leco Corporation, Model V-100-A2, St. Joseph, MI)ith a load of 49 N (5 kg) on a polished surface with a Vick-

rs diamond tip followed by measurements of the radial medianracks formed during indentation.38 Fracture toughness valuesere obtained from an average of ten indents using Eq. (1) where

is the Young’s modulus determined through impulse excita-ion, H is the Vickers’ hardness measured at a load of 9.8 N, Ps the load used for indention (49 N), and 2c is the length of theadial median cracks.38

IC = 0.016

(E

H

)1/2 (P

c3/2

)(1)

Polished specimens were examined using scanning elec-ron microscopy (S-570, Hitachi, Tokyo, Japan or Heliosanolab 600, FEI, Hillsboro, OR). Area fractions of the

onstituent phases and porosity were determined using com-

uter image analysis (ImageJ, National Institutes of Health,ethesda, MD). Particle sizes were determined by equatingverage particle areas to the areas of spheres with equivalentiameters.

D.S. King et al. / Journal of the European Ceramic Society 33 (2013) 2943–2951 2945

Table 1Density, hardness, and fracture toughness of SiC–TiB2 ceramics.

Composition Densificationtemperature (◦C)

Theoreticaldensity (g/cm3)

Bulk density(g/cm3)

Relativedensity (%)

Hardness(GPa)

Fracture toughness(MPa m1/2)

SiC 2100 3.22 3.20 99 27.8 ± 1.1 2.1 ± 0.1SiC–15TiB2 2030 3.42 3.37 99 26.0 ± 1.8 4.3 ± 0.4SiC–20TiB2 2030 3.48 3.37 97 25.9 ± 1.8 4.5 ± 0.4SiC–25TiB2 1980 3.55 3.50 99 27.2 ± 1.5 4.2 ± 0.2SiC–30TiB2 1980 3.61 3.56 99 26.6 ± 1.3 4.4 ± 0.3SiC–35TiB2 1980 3.68 3.59 98 25.5 ± 1.3 4.7 ± 0.5SiC–40TiB2 2030 3.74 3.64 97 25.3 ± 1.4 6.2 ± 1.0SiC–60TiB2 1980 4.00 3.90 98 25.3 ± 0.9 6.1 ± 0.6SiC–80TiB2 1980 4.26 4.15 97 24.2 ± 2.0 5.7 ± 0.7T

3

ci(wocrsdmwp

pdsBitgbnptTsoawmostetsaia

atSpcs

3

iiUfbTSTwitbtdtSdmw

3

Ti6∼4

iB2 2100 4.52 4.39

. Results and discussion

Table 1 summarizes the density values for each SiC–TiB2eramic. Theoretical densities were calculated based on the nom-nal SiC and TiB2 contents without accounting for sintering aidsC and B4C) since some of them were consumed by reactionith surface oxide impurities during hot pressing.39–41 Becausef the retention of some low density B4C and/or C in the finaleramics, the calculated values represent a lower bound for theelative density and likely underestimate the actual relative den-ity. Likewise, pores can reduce the value of the measured bulkensity of the ceramics. Even though the relative density valuesay underestimate actual density, all of the relative densitiesere more than 97%, with most being 99%. Hence, hot-pressingroved to be effective for densifying all of the compositions.

Microstructures were analyzed for all of the hot pressed com-osites (Fig. 1). The SEM images were consistent with theensity measurements and supported the conclusion that thepecimens contained a small volume fraction of porosity (<1%).oth residual sintering aids and porosity appear dark, but poros-

ty could be identified as dark inclusions with a light ring aroundhem due to beam charging. Charging can also be seen aroundrain pullout that occurred through the polishing procedure,ut can be distinguished from porosity due to the jagged cor-ers, whereas pores appear circular. While Fig. 1f (nominallyure TiB2) appears to be porous, the composition was foundo be ∼97% dense by Archimedes’ measurements (Table 1).he residual B4C or carbon sintering aids that are retained afterintering were apparent in the micrographs as dark regions with-ut charging around them. In addition to porosity and sinteringdditives, the SEM images also showed that the SiC and TiB2ere uniformly distributed. Average TiB2 particle sizes wereeasured for each ceramic and were used in the interpretation

f strength data. Based on equivalent circular diameters, theize of TiB2 particles increased from 2.7 �m for SiC–15TiB2o 6.3 �m for SiC–20TiB2 and to 26.2 �m for SiC–40TiB2. Inach ceramic, several particles formed larger particle clustershat were analyzed as a single inclusion by the image analysisoftware. Therefore, the sizes of TiB clusters in SiC–20TiB

2 2nd SiC–40TiB2, were larger than the individual particles. Thencrease in cluster size as TiB2 content increased also led ton increase in the average distance between TiB2 clusters. As

tacn

97 24.4 ± 1.7 5.9 ± 1.0

result, a crack propagating through the microstructure mustravel farther through the SiC matrix between TiB2 clusters iniC–40TiB2 than in SiC–20TiB2 or SiC–15TiB2. The longerath through SiC should increase the toughening in SiC–TiB2omposites due to the presence of compressive thermal residualtresses in the SiC matrix.17,28,42,43

.1. Hardness

The measured Vickers hardness values for SiC–TiB2 ceram-cs decreased as TiB2 content increased (Table 1). In contrast, anncrease in hardness was expected for increasing TiB2 contents.sing reported hardness values of 27 GPa for SiC and 33 GPa

or TiB2, a linear volumetric rule of mixtures calculation shoulde useful in estimating the increase in hardness with increasingiB2 content.8,15,16,44 For the measured values, nominally pureiC had the highest hardness at 28 GPa while SiC–80TiB2 andiB2 had the lowest values at 24 GPa. The drop in the hardnessith increasing TiB2 content indicates a reduction in the work of

ndentation, which may be related to the distribution of residualensile stresses in the ceramics.45,46 Microcracking, which cane caused by crystal anisotropy in nominally pure TiB2, or byhermal residual stresses within SiC–TiB2 composites, can alsoecrease measured hardness values because the work of indenta-ion is decreased through the closing of microcracks.15,44 In theiC–TiB2 ceramics, tensile stresses that develop during coolingue to the CTE mismatch between SiC and TiB2 may lead toicrocracking, which could have led to the decrease in hardnesshen an increase was expected based on a rule of mixtures.

.2. Fracture toughness

Fracture toughness measurements are also summarized inable 1. With increasing additions of TiB2, fracture toughness

ncreased from ∼2 MPa m1/2 for nominally pure SiC to over MPa m1/2 for SiC–40TiB2. The toughness values plateaued at6 MPa m1/2 for SiC–TiB2 ceramics with TiB2 contents above

0 vol.%. For compositions containing more than 40 vol.% TiB2,

he microstructures transition from isolated TiB2 particles in

SiC matrix to isolated SiC particles in a TiB2 matrix. Thehange in the volume fraction of SiC particles appeared to haveo effect on the measured toughness since both SiC–60TiB2

2946 D.S. King et al. / Journal of the European Ceramic Society 33 (2013) 2943–2951

F (d),

s positip

anScotpttatDtf

at

vnhwpinn

ig. 1. SEM images of SiC (a), SiC–15TiB2 (b), SiC–20TiB2 (c), SiC–40TiB2

till containing minimal porosity. Grain pullout in (f), however, makes the comore.

nd SiC–80TiB2 have the same nominal fracture toughness asominally pure TiB2 (∼6 MPa m1/2). Whereas TiB2 additions toiC had a major effect on fracture toughness due to the residualompressive stresses generated in the SiC matrix, the additionf SiC to TiB2 did not appear to affect the toughness comparedo nominally pure TiB2. The values measured for nominallyure SiC and TiB2 fall within the range of values reported forhose materials by a variety of methods.6,15 In addition, the frac-ure toughness values for TiB2 contents below 20 vol.% werelso similar to previously reported values for SiC–TiB2 frac-ure toughness determined by the direct crack method.1,16,25,47

espite the inherent variability of this toughness measurementechnique (as high as 25% by some estimates), the large dif-erences (∼200%) in values noted between nominally pure SiC

San

SiC–80TiB2 (e), and TiB2 (f). Compositions (a)–(e) appear to be dense, whileon appear as if it is not dense. Arrows point to examples of grain pullout and a

nd SiC containing TiB2 indicates that the increase in fractureoughness is likely real in these materials.

Fig. 2 shows the measured hardness and fracture toughnessalues as a function of TiB2 content. The classic hard-ess/toughness tradeoff is apparent in these materials as theardness generally decreased across the composition rangehile the toughness increased.10,12–14 Compared to nominallyure SiC, hardness decreased by nearly 10% as TiB2 contentncreased from 0 vol.% to 40 vol.% while toughness increased byearly 200% for SiC–40TiB2 compared to SiC. With both hard-ess and toughness being key properties for armor applications,

iC–40TiB2 appeared to have the best combination of hardnessnd fracture toughness of the compositions tested, with a hard-ess of 25 GPa and fracture toughness of 6.2 MPa m1/2.10,12,13,48

D.S. King et al. / Journal of the European Ceramic Society 33 (2013) 2943–2951 2947

Fd

ectemwwdwt

λ

�

ttSstatewmtCrccefcig

S2oawptSfftwtPctttwswrttt

ecatsopd

ig. 2. Hardness and fracture toughness as a function of TiB2 content. Hardnessecreases as TiB2 content increases while fracture toughness increases.

The magnitude of thermal residual stresses, and the tough-ning due to compressive stresses within the matrix, werealculated over the range from processing temperature (2030 ◦C)o room temperature (25 ◦C) using the method outlined by Tayat al. by assuming that tensile stresses were not relaxed by anyechanism during cooling.17 The average particle spacing (λ),hich is equivalent to the parameter (λ − d) used by Taya et al.,as calculated using Eq. (2) where dp is the average particleiameter and φF is the volume fraction of particles.49 Then, λ

as combined with the average matrix stress to estimate theoughening due to thermal residual stresses using Eq. (3).17

= dp

[(π

6φF

)1/3

− 1

](2)

KI = 2q

√2(λ)

π(3)

As the volume fraction of TiB2 increased, the magnitude ofhe tensile thermal residual stresses estimated for the TiB2 par-icles decreased from 1975 MPa in SiC–15TiB2 to 1389 MPa iniC–40TiB2. Over the same composition range, the compres-ive stress in SiC increased from −349 MPa to −926 MPa. Ashe TiB2 content increases, the residual compressive stressesre distributed over a smaller volume fraction of SiC. As TiB2ransitions from isolated particles to the matrix phase, thestimated compressive stresses continue to build to ∼1.5 GPaithin the SiC particulate phase of SiC–80TiB2. Based on theagnitude of these stresses, the amount of toughening due

o compressive thermal residual stresses can be calculated.onversely, a decrease in toughening due to tensile thermal

esidual stress can also be calculated using this method. Thishange in toughening, �KI, was calculated for compositesontaining 15, 20, and 40 vol.% TiB2. The amount of tough-ning was −0.58 MPa m1/2 for SiC–15TiB2, −1.16 MPa m1/2

1/2

or SiC–20TiB2, and −2.32 MPa m for SiC–40TiB2. Thesealculated �KI’s are equivalent to an increase in the toughen-ng of the ceramic composites, �KR.17 Since TiB2 pins grainrowth of the SiC matrix during densification, a fine grained

iiSa

Fig. 3. Crack deflection at the corner of a Vickers indent in SiC–15TiB2.

iC matrix is produced. As a result, a fracture toughness of MPa m1/2 was used as a baseline value for the matrix, basedn the reported fracture toughness for nominally pure SiC withn average grain size of ∼110 nm.6 This value also agrees wellith the measured fracture toughness of the nominally pure SiCroduced in this study (2.1 MPa m1/2). Based on thermal residualoughening, the composite toughness should be 2.6 MPa m1/2 foriC–15TiB2, 3.2 MPa m1/2 for SiC–20TiB2, and 4.3 MPa m1/2

or SiC–40TiB2. However, the predicted increase in toughnessor SiC, due to thermal residual stresses, value falls short ofhe measured KIC values for all of the compositions. In otherords, additional �KR from other sources is needed to explain

he toughness values measured for each ceramic composite.otential error in estimation of the residual thermal stressesould allow for an even bigger discrepancy. In a similar sys-em (ZrB2–SiC), Watts et al. used neutron diffraction to showhat residual thermal stresses started to accumulate at tempera-ures of ∼1400 ◦C, while the maximum processing temperatureas ∼1900 ◦C.50,51 Because of this, the actual thermal residual

tresses are likely to be lower than the calculated values, whichould reduce �KR even further. Based on this analysis, thermal

esidual stresses present in SiC–TiB2 ceramics are not sufficiento account for the observed increase in toughness, even whenhe magnitude of their effect is overestimated by overestimatinghe values of the thermal residual stresses.

Crack deflection around particles is another possible tough-ning mechanism in ceramic composites. In many SiC–TiB2omposites, crack deflection is thought to be the main mech-nism of toughening.18,20–22,25 Indeed, as a crack propagateshrough a SiC matrix that is in compression, the radial tensiletresses in the TiB2 particles draws the crack(s) to these regionsf the microstructure, which reduces the driving force for crackropagation in the tensile regions.27 In the present study, crackeflections were observed near TiB2 grains (Fig. 3) by examin-ng the paths of the radial/median cracks produced by Vickers

ndentations used in the determination of fracture toughness. ForiC–15TiB2, a total of 53 deflections, with an average deflectionngle of 29◦, were counted in the four radial/median cracks

2948 D.S. King et al. / Journal of the European C

Fa

ptdmndmipmt

K

3

hTsSaoasttepCiHFv(flm

W

Scu1ftoSt

maacfSc44osiritota

3

oenoc(tsomrtatctohmomparticle size may be large enough to allow spontaneous micro-

ig. 4. Two-parameter Weibull analysis of SiC–TiB2 compositions with 15, 20,nd 40 vol.% TiB2. Weibull modulus (m) increases as TiB2 content increases.

roduced by a 49 N indent in a previous study.37 Eq. (4) was usedo predict the fracture toughness of the material with no crackeflection, or other toughening mechanism, where K(θ) is theeasured fracture toughness, KIC(θ = 0) is the fracture tough-

ess with no toughening mechanisms, and theta is the crackeflection angle.52,53 This shows that if all other tougheningechanisms are neglected, the effect of crack deflection would

ncrease the toughness of the material by less than 7%. Still, theresence of crack deflection shows that in SiC–TiB2 ceramicsore than one toughening mechanism enhances fracture

oughness.

(θ) = KIC(θ = 0)sec2(

θ

2

)(4)

.3. Strength, Weibull modulus, and elastic modulus

The SiC–TiB2 ceramics with the best combinations ofardness and toughness were selected for strength testing.he strongest composition was SiC–20TiB2, with an averagetrength of 522 MPa. The next highest strength was 500 MPa foriC–15TiB2 followed by 423 MPa for SiC–40 TiB2. The char-cteristic strengths were determined from the Weibull analysisf the failure strengths. The characteristic strength is the valuet which the probability of survival is 37%. The characteristictrengths follow the same pattern as the average strengths withhe highest characteristic strength of 539 MPa for SiC–20TiB2,he middle strength of 520 MPa for SiC–15TiB2, and the low-st strength of 436 MPa for SiC–40TiB2. The increasing valuesredicted for compressive stresses in the SiC matrix due to theTE mismatch between TiB2 and SiC correlate to the observed

ncrease in strengths between SiC–15TiB2 and SiC–20TiB2.owever, the trend did not follow for higher TiB2 contents.or SiC–40TiB2, the average strength was lower despite higheralues predicted for compressive stresses in the SiC matrix−926 MPa). In this case, the lower strength indicates a largeraw size in SiC–40TiB2, which may indicate the presence of

icrocracking.27

Flexure strengths were analyzed using a two parametereibull analysis. Fig. 4 shows the Weibull plot for each

ceh

eramic Society 33 (2013) 2943–2951

iC–TiB2 composition that underwent flexure testing. As TiB2ontent increased, so did the Weibull modulus. A Weibull mod-lus of 12 was calculated for SiC–15TiB2, which increased to7 for SiC–20TiB2 and SiC–40TiB2. Unlike the trend observedor strength, increasing TiB2 content did not have a detrimen-al effect on the Weibull modulus. The higher Weibull modulusf SiC–20TiB2 and SiC–40TiB2 indicated that both of theseiC–TiB2 ceramics had smaller critical flaw size distributions

han SiC–15TiB2.Young’s modulus measurements using the dynamic impulse

ethod were made on each composition that underwent Weibullnalysis. Using Young’s modulus values of 450 GPa for SiCnd 560 GPa for TiB2, a linear volumetric rule of mixturesalculation was also used to predict modulus values for the dif-erent compositions.15,54 A measured value of 464 ± 4 GPa foriC–15TiB2 agreed well with the predicted value of 467 GPa. Inontrast, measured values of 454 ± 5 GPa for SiC–20TiB2 and82 ± 1 GPa for SiC–40TiB2 fell below the values predicted of72 GPa and 494 GPa predicted for those compositions. Basedn the strength trends, thermal residual stresses seem to boost thetrength of SiC–20TiB2 due to the higher compressive stressesn the SiC matrix compared to SiC–15TiB2; however, thermalesidual stresses may have been high enough in SiC–40TiB2 tonitiate spontaneous microcracking, which gave that composi-ion the lowest strength.26,55,56 Since porosity was minimal in allf the compositions, the decreases in Young’s moduli supporthe conclusion of spontaneous microcracking in SiC–40TiB2nd SiC–20TiB2.

.4. Analysis of microstructure

Since many of the differences between measured valuesf mechanical properties and the expected values would bexplained by the presence of microcracks, the particle sizeeeded for spontaneous microcracking was calculated for eachf the compositions examined using Weibull analysis.31 Eq. (5)an be used to estimate the critical particle size for microcrackingbc) where η is a constant ranging from 2 to 8, KIC is the fractureoughness, ν is the Poisson’s ratio for the matrix or particle (sub-cript m or p, respectively), β is the ratio of the Young’s modulif the matrix (Em) to inclusion particles (Ep), �α is the CTE mis-atch, and �T is difference between processing temperature and

oom temperature. The critical TiB2 particle size was estimatedo be ∼2.6 �m for each SiC–TiB2 ceramic composite based onn interfacial fracture toughness equivalent to the matrix frac-ure toughness (2 MPa m1/2) and an η value of 8 to maximize thealculated particle size.31 However, since the interfacial frac-ure toughness may range between that of the matrix and thatf the reinforcing phase, and the η value given by Magley et al.as a range of 2–8, the calculated particle size for spontaneousicrocracking can also range from 0.7 �m to 26 �m.57 Based

n the calculation for spontaneous microcracking and decreasedechanical properties of SiC–20TiB2 and SiC–40TiB2, the TiB2

racking. However, in SiC–15TiB2, the values of hardness andlastic modulus suggest that the interfacial fracture toughness isigher than the assumed 2 MPa m1/2 and the critical particle size

D.S. King et al. / Journal of the European C

FS

tiphvFcmd

b

c

FS

afstbeia

cmpifltacRipcnc

ttSiSmct

ig. 5. SEM image of a circumferential microcrack around a TiB2 grain iniC–20TiB2.

hat would result in spontaneous microcracking in SiC–15TiB2s larger than the measured cluster size of 2.7 �m. Further sup-ort for the conclusion that SiC–20TiB2 and SiC–40TiB2 couldave microcracks, but that SiC–15TiB2 should not was pro-ided by high resolution SEM analysis. The images shown inigs. 5 and 6 reveal what appear to be circumferential micro-racks at the interface between the TiB2 particles and the SiCatrix grains in SiC–20TiB2 and SiC–40TiB2. Similar analysis

id not find any evidence of micorcracking in SiC–15TiB2.

c = ηK2IC

{(1 + vm/2) + β(1 − 2vp)

}2

(5)

Em�α�T

Even though crack deflection was prevalent in each SiC–TiB2omposites, microcracking can also lead to a rise in toughening

ig. 6. SEM image of a circumferential microcrack around a TiB2 grain iniC–40TiB2.

wpgecaSsa

itmtsehS

4

il

eramic Society 33 (2013) 2943–2951 2949

s has been reported for SiC–TiB2 ceramics.29,30,36 Circum-erential microcracking leads to an increase in toughness byhielding the tip of a propagating crack.30,31,58 As circumferen-ial microcracking is seen in both SiC–20TiB2 and SiC–40TiB2,ut not SiC–15TiB2, microcracking appears to have a majorffect on increasing the fracture toughness of SiC–TiB2 compos-tes compared to toughening through thermal residual stressesnd crack deflection.

In SiC–20TiB2 and SiC–40TiB2, microcracking is the likelyause of the higher Weibull moduli compared to the non-icrocracked SiC–15TiB2. Since both ceramics had TiB2

article sizes large enough to expect spontaneous microcrack-ng, stress induced microcracking would also be expected duringexure testing.34 In ceramics, a microcracked process zone due

o stress induced microcracking can lead to crack tip shielding,llowing the material to exhibit R-curve behavior.30,58,59 R-urve behavior was not tested in the present study, but increasing-curve behavior has been reported to accompany an increase

n Weibull modulus in previous studies.30,60–62 Based on theresent study, the potential for R-curve behavior in SiC–TiB2eramics warrants a more detailed analysis of the fracture tough-ess to better understand the fracture mechanics of these ceramicomposites.

While microcracks were observed in SiC–20TiB2 (Fig. 5),hey appeared to have no significant effect on the strength ofhe ceramic as the strength of SiC–20TiB2 was higher thaniC–15TiB2. However, microcracking can lead to an increase

n the flaw size, which then reduces the strength, as seen iniC–40TiB2. Therefore, the measured strength of SiC–20TiB2ay be lower than what would be measured if a SiC–20TiB2

eramic that was free of microcracks were produced.27,63 Fur-her support for that conclusion is provided by a previous study inhich the properties of SiC–15TiB2 was compared to a ceramicroduced using commercially available SiC–TiB2 spray driedranules containing 15 vol.% TiB2. While SiC–15TiB2 did notxhibit microcracking, the ceramic produced from commer-ially available spray dried granules exhibited microcrackingnd suffered a drop in strength compared to SiC–15TiB2.57

pontaneous microcracking had a detrimental effect on thetrength of SiC–40TiB2 while possibly lowering the maximumttainable strength of SiC–20TiB2.

Titanium diboride was added to SiC with the intent ofncreasing both hardness and fracture toughness. While frac-ure toughness increased with increasing TiB2 content, the

easured hardness dropped. Increasing TiB2 content also ledo an increased Weibull modulus, allowing for narrower flawize distributions in SiC–TiB2 ceramics. Overall, SiC–40TiB2xhibited an increased fracture toughness (6.2 MPa m1/2), highardness (25.3 GPa) and a high Weibull modulus (17) makingiC–40TiB2 a good candidate for ceramic armor consideration.

. Conclusions

The mechanical properties were studied for SiC–TiB2 ceram-cs as a function of TiB2 content. Hardness decreased almostinearly from 27 GPa for nominally pure SiC to 24 GPa for

2 ean C

n2fuSe5SvStT4uiv(4ia6ptitmnwmSiSwofaa

A

Aecaeotn

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

950 D.S. King et al. / Journal of the Europ

ominally pure TiB2. Fracture toughness increased from MPa m1/2 for nominally pure SiC to a value of ∼6 MPa m1/2

or TiB2 contents of 40 vol.% and higher. Weibull analysis wassed to evaluate the strength of three compositions, SiC–15TiB2,iC–20TiB2, and SiC–40TiB2. SiC–20TiB2 had the high-st flexural strength at 522 MPa followed by SiC–15TiB2 at00 MPa and SiC–40TiB2 at 423 MPa. The lower strength ofiC–40TiB2 combined with a lower-than-expected hardnessalue of 25 GPa suggested that microcracks were present iniC–40TiB2. Young’s modulus measurements also suggested

hat microcracks existed in both SiC–20TiB2 and SiC–40TiB2.he measured Young’s modulus value for SiC–15TiB2, was64 GPa, which agreed well with a predicted value of 467 GPasing the nominal composition and modulus values for nom-nally pure SiC and TiB2. In contrast, the Young’s modulusalues measured for SiC–20TiB2 (454 GPa) and SiC–40TiB2482 GPa) were below the predicted values of 472 GPa and94 GPa, further indicating that microcracks may be presentn these compositions. Analysis by SEM revealed that TiB2gglomerate size increased from 2.7 �m in SiC–15TiB2 and.3 �m in SiC–20TiB2 to 26.2 �m in SiC–40TiB2. Thresholdarticle sizes for spontaneous microcracking were calculatedo be ∼2 �m for every SiC–TiB2 ceramic, however, mechan-cal property evaluation suggested that the interfacial fractureoughness used for calculating the particle size for spontaneous

icrocracking in SiC–15TiB2 was low and SiC–15TiB2 shouldot exhibit spontaneous microcracking. High resolution SEMas used to find evidence of the presence of circumferentialicrocracks at the interface between the TiB2 particles and theiC matrix grains in SiC–20TiB2 and SiC–40TiB2. Microcrack-

ng also helped boost the Weibull modulus of SiC–20TiB2 andiC–40TiB2 to 17 compared to a value of 12 for SiC–15TiB2,hich did not have any microcracks. While it was the weakestf the ceramics that were evaluated, SiC–40TiB2 exhibited highracture toughness (6.2 MPa m1/2), high hardness (25.3 GPa) and

high Weibull modulus making it a good candidate for ceramicrmor consideration.

cknowledgements

Research was sponsored by Saint-Gobain Ceramics and thermy Research Laboratory and was accomplished under Coop-

rative Agreement Number W911NF-08-2-0001. The views andonclusions contained in this document are those of the authorsnd should not be interpreted as representing the official policies,ither expressed or implied, of the Army Research Laboratoryr the U.S. Government. The U.S. Government is authorizedo reproduce and distribute reprints for Government purposesotwithstanding any copyright notation hereon.

eferences

1. Blanc C, Thevenot F, Treheux D. Wear resistance of �SiC–TiB2 composites

prepared by reactive sintering. J Eur Ceram Soc 1999;19(5):571–9.

2. Cao JJ, MoberlyChan WJ, De Jonghe LC, Gilbert CJ, Ritchie RO. In situtoughened silicon carbide with Al–B–C additions. J Am Ceram Soc1996;79(2):461–9.

2

2

eramic Society 33 (2013) 2943–2951

3. Endo H, Ueki M, Kubo H. Hot pressing of SiC–TiC composites. J MaterSci 1990;25(5):2503–6.

4. Zhang XF, Sixta ME, Jonghe LCD. Grain boundary evolution in hot-pressedABC–SiC. J Am Ceram Soc 2000;83(11):2813–20.

5. Miura M, Yogo T, Hirano SI. Phase separation and toughening of SiC–AINsolid–solution ceramics. J Mater Sci 1993;28(14):3859–65.

6. Kim Y-W, Mitomo M, Hirotsuru H. Grain growth and fracture toughnessof fine-grained silicon carbide ceramics. J Am Ceram Soc 1995;78(11):3145–8.

7. Matchen B. Applications of ceramics in armor products. Key Eng Mater1996;122–124:333–44.

8. Engineered materials handbook: ceramics and glasses, vol. 4. MaterialsPark, OH: ASM International; 1991.

9. Flinders M, Ray D, Anderson A, Cutler RA. High-toughness silicon carbideas armor. J Am Ceram Soc 2005;88(8):2217–26.

0. Karandikar PG, Evans G, Wong S, Aghajanian MK, Sennett M. A review ofceramics for armor applications. In: Advances in ceramic armor IV: ceramicengineering and science proceedings, vol. 29(6). Hoboken, NJ: John Wiley& Sons Inc.; 2009. p. 163–75.

1. Hilmas GE, Tien T-Y. Effect of AlN and Al2O3 additions on the phaserelationships and morphology of SiC. Part I: compositions and properties.J Mater Sci 1999;34(22):5605–12.

2. Ray D, Flinders M, Anderson A, Cutler RA. Hardness/toughness relation-ship for SiC armor. In: 27th annual cocoa beach conference on advancedceramics and composites. A: ceramic engineering and science proceedings.Hoboken, NJ: John Wiley & Sons, Inc.; 2008. p. 401–10.

3. Ray D, Flinders M, Anderson A, Cutler RA, Rafaniello W. Effect of room-temperature hardness and toughness on the ballistic performance of SiC-based ceramics. In: Swab JJ, editor. International conference on advancedceramics and composites, vol. 26, Advances in ceramic armor, 2005.

4. Park SJ, Cowan K, Johnson JL, German RM. Grain size measurement meth-ods and models for nanograined WC–Co. Int J Refract Met Hard Mater2008;26(3):152–63.

5. Basu B. Processing and properties of monolithic TiB2 based materials. IntMater Rev 2006;51(6):352–74.

6. Blanc C, Thevenot F, Goeuriot D. Microstructural and mechanical char-acterization of SiC–submicron TiB2 composites. J Eur Ceram Soc1999;19(5):561–9.

7. Taya M, Hayashi S, Kobayashi AS, Yoon HS. Toughening of a particulate-reinforced ceramic-matrix composite by thermal residual stress. J Am CeramSoc 1990;73(5):1382–91.

8. Kuo D-H, Kriven WM. Mechanical behavior and microstructure of SiC andSiC/TiB2 ceramics. J Eur Ceram Soc 1998;18(1):51–7.

9. Ohya Y, Hoffmann MJ, Petzow G. Sintering of in-situ synthesizedSiC–TiB2 composites with improved fracture toughness. J Am Ceram Soc1992;75(9):2479–85.

0. McMurtry CH, Boecker WDG, Seshadri SG, Zanghi JS, Garnier JE.Microstrutcture and material properties of SiC–TiB2 composites. Bull AmCeram Soc 1987;66(21):325–9.

1. Janney MA. Mechanical properties and oxidation behavior of hot-pressedSiC–15 vol%-TiB2 composite. Bull Am Ceram Soc 1987;66(2):322–4.

2. Yoon JD, Kang SG. Strengthing and toughening behaviour of SiC withadditions of TiB2. J Mater Sci Lett 1995;14:1065–7.

3. Li Z, Bradt RC. Thermal expansion and thermal expansion anisotropy ofSiC polytypes. J Am Ceram Soc 1987;70(7):445–8.

4. Ferber MK, Becher PF, Finch CB. Effect of microstructure on the propertiesof TiB2 ceramics. J Am Ceram Soc 1983;66(1):C-2–3.

5. Cho K-S, Choi H-J, Lee J-G, Kim Y-W. In situ enhancement of toughnessof SiC–TiB2 composites. J Mater Sci 1998;33(1):211–4.

6. Pan M-J, Hoffman PA, Green DJ, Hellmann JR. Elastic properties andmicrocracking behavior of particulate titanium diboride–silicon carbidecomposites. J Am Ceram Soc 1997;80(3):692–8.

7. Wachtman JB, Cannon WR, Matthewson MJ. Mechanical properties ofceramics. 2nd ed. Hoboken, NJ: John Wiley & Sons, Inc; 2009.

8. Bucevac D, Boskovic S, Matovic B, Krstic V. Toughening of SiC matrixwith in-situ created TiB2 particles. Ceram Int 2010;36(7):2181–8.

9. Gu WH, Faber KT. Tensile behavior of microcracking SiC–TiB2 composites.J Am Ceram Soc 1995;78(6):1507–12.

ean C

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4

4

5

5

5

5

5

5

5

5

5

5

6

6

6

D.S. King et al. / Journal of the Europ

0. Gu WH, Faber KT, Steinbrech RW. Microcracking and R-curve behavior inSiC–TiB2 composites. Acta Metall Mater 1992;40(11):3121–8.

1. Magley DJ, Winholtz RA, Faber KT. Residual stress in a two-phase micro-cracking ceramic. J Am Ceram Soc 1990;73(6):1641–4.

2. Todd RI, Boccaccini AR, Sinclair R, Yallee RB, Young RJ. Thermal residualstresses and their toughening effect in Al2O3 platelet reinforced glass. ActaMater 1999;47(11):3233–40.

3. Sigl LS. Microcrack toughening in brittle materials containing weak andstrong interfaces. Acta Mater 1996;44(9):3599.

4. Green DJ. Stress-induced microcracking at second-phase inclusions. J AmCeram Soc 1981;64(3):138–41.

5. Evans AG, Faber KT. Crack-growth resistance of microcracking brittlematerials. J Am Ceram Soc 1984;67(4):255–60.

6. Cai H, Gu WH, Faber KT. Microcrack toughening in a SiC–TiB2 com-posite. Proceedings of the American society for composites fifth technicalconference, composite materials in transition 1990.

7. King DS, Fahrenholtz WG, Hilmas GE. Microstructural effects on themechanical properties of SiC–15 vol% TiB2 particulate-reinforced ceramiccomposites. J Am Ceram Soc 2013;96(2):577–83.

8. Anstis GR, Chantikul P, Lawn BR, Marshall DB. A critical evaluation ofindentation techniques for measuring fracture toughness. I: direct crackmeasurements. J Am Ceram Soc 1981;64(9):533–8.

9. Zhang SC, Hilmas GE, Fahrenholtz WG. Pressureless sintering of ZrB2–SiCceramics. J Am Ceram Soc 2008;91(1):26–32.

0. Zhu S, Fahrenholtz WG, Hilmas GE, Zhang SC. Pressureless sintering ofzirconium diboride using boron carbide and carbon additions. J Am CeramSoc 2007;90(11):3660–3.

1. Prochazka S, Scanlan RM. Effect of boron and carbon on sintering of SiC.J Am Ceram Soc 1975;58(1–2):72.

2. Cutler RA, Virkar AV. The effect of binder thickness and residualstresses on the fracture toughness of cemented carbides. J Mater Sci1985;20(10):3557–73.

3. Chermant JL, Osterstock F. Fracture toughness and fracture of WC–Cocomposites. J Mater Sci 1976;11(10):1939–51.

4. Srivatsan TS, Guruprasad G, Black D, Radhakrishnan R, Sudarshan TS.Influence of TiB2 content on microstructure and hardness of TiB2–B4Ccomposite. Powder Technol 2005;159(3):161–7.

5. Wang L, Bei H, Gao YF, Lu ZP, Nieh TG. Effect of residual stresses on thehardness of bulk metallic glasses. Acta Mater 2011;59(7):2858–64.

6. Simes TR, Mellor SG, Hills DA. A note on the influence of residual stresson measured hardness. J Strain Anal Eng Des 1984;19(2):135–7.

7. Mestral F, Thevenot F. Ceramic composites: TiB2–TiC–SiC. J Mater Sci1991;26(20):5547–60.

6

eramic Society 33 (2013) 2943–2951 2951

8. Lundberg P, Lundberg B. Transition between interface defeat and penetra-tion for tungsten projectiles and four silicon carbide materials. Int J ImpactEng 2005;31(7):781–92.

9. Matteucci S, Kusuma VA, Sanders D, Swinnea S, Freeman BD. Gas transportin TiO2 nanoparticle-filled poly(1-trimethylsilyl-1-propyne). J Membr Sci2008;307(2):196–217.

0. Watts J, Hilmas G, Fahrenholtz WG, Brown D, Clausen B. Measurementof thermal residual stresses in ZrB2–SiC composites. J Eur Ceram Soc2011;31(9):1811–20.

1. Watts J, Hilmas G, Fahrenholtz WG, Brown D, Clausen B. Stress mea-surements in ZrB2–SiC composites using Raman spectroscopy and neutrondiffraction. J Eur Ceram Soc 2010;30(11):2165–71.

2. Weiderhorn SM. Brittle fracture and toughening mechanisms in ceramics.Annu Rev Mater Sci 1984;14:373–403.

3. Rezaie A, Fahrenholtz WG, Hilmas GE. Effect of hot pressing time andtemperature on the microstructure and mechanical properties of ZrB2–SiC.J Mater Sci 2007;42(8):2735–44.

4. Shaffer PTB, Jun CK. The elastic modulus of dense polycrystalline siliconcarbide. Mater Res Bull 1972;7(1):63–9.

5. Pan M-J, Green DJ, Hellmann JR. Influence of interfacial microc-racks on the elastic properties of composites. J Mater Sci 1996;31(12):3179–84.

6. Green DJ. Microcracking mechanisms in ceramics. In: Bradt RC, EvansAG, Hasselman DPH, Lange FF, editors. Fracture mechanics of ceramics,vol. 5. New York: Plenum Press; 1983. p. 457–78.

7. King DS, Fahrenholtz WG, Hilmas GE. Microstructural effects on themechanical properties of SiC–15 vol.% TiB2 particulate reinforced ceramiccomposites. J Am Ceram Soc 2013;96(2):577–83.

8. Ritchie RO. Mechanisms of fatigue crack propagation in metals, ceram-ics and composites: role of crack tip shielding. Mater Sci Eng A1988;103(1):15–28.

9. Bowling GD, Faber KT, Hoagland RG. Computer simulations of R-curvebehavior in microcracking materials. J Am Ceram Soc 1987;70(11):849–54.

0. Gong J, Si W, Guan Z. Weibull modulus of fracture strength of tough-ened ceramics subjected to small-scale contacts. J Mater Sci 2001;36(10):2391–6.

1. Kaur S, Shetty DK, Cutler RA. R curves and crack-stability map: applicationto Ce-TZP/Al2O3. J Am Ceram Soc 2007;90(11):3554–8.

2. Shetty DK, Wang J-S. Crack stability and strength distribution of ceramics

that exhibit rising crack-growth-resistance (R-curve) behavior. J Am CeramSoc 1989;72(7):1158–62.

3. Green DJ. Critical microstructures for microcracking in Al2O3–ZrO2 com-posites. J Am Ceram Soc 1982;65(12):610–4.