Synthesis and Morphological Analysis of TitaniumCarbide Nanopowder

Debasish Sarkar,z Mincheol Chu,w Seong-Jai Cho, and Yong Il Kim

Korea Research Institute of Standards and Science, Yusong-Gu, Daejeon 305-340, Korea

Bikramjit Basu*

Department of Materials and Metallurgical Engineering, Indian Institute of Technology, Kanpur, Uttar Pradesh208016, India

One of the major challenges in the development of nanostruc-tured/ultrafine nonoxide ceramics is the synthesis of nonagglom-erated-nanosized ceramic powders. Here, we investigatecarbothermal reduction to obtain spherical titanium carbide(TiC) nanoparticles smaller than 100 nm in diameter. Using acombination of characterization techniques, it has been con-firmed that optimum thermal reduction conditions include atemperature of 13501C with 1 h holding in an argon atmo-sphere. A three-parameter Weibull distribution function pro-vides a satisfactory fit for the size distribution of the synthesizedTiC nanoparticles. Varying the synthesis time at 13501C re-vealed the particle coarsening exponent ‘‘n’’ to be 2.9 and thatparticle shapes change from spherical at 1 h to more faceted atlonger holding times of 4 h.

I. Introduction

BULK nanoceramic materials, having grain sizes typicallysmaller than 100 nm, possess appealing mechanical, phys-

ical, and tribological properties.1 They have therefore been thesubject of significant research activity in recent years. Oneof the major challenges in the research on bulk nanoceramicsand nanoceramic composites is the synthesis of nanosizedceramic powders. While the synthesis of nanosized oxidepowders has been reported, such efforts in the case of nonox-ide ceramics are not widespread. In this context, we have carriedout a research program to develop nanosized titanium carbide(TiC).

TiC has been successfully integrated into structural compo-nents, including magnetic recording heads, transparent opticalmaterials, and composite reinforcing materials.1,2 TiC can beprepared through several methods. One such method is viacommon carbothermal reduction of titanium dioxide (TiO2)powder using carbon.3–5 Other methods use polymeric precur-sors based on either titanium alkoxides or other organic com-pounds.6–8 Another alternative method is through the directreaction of Ti and carbon.9 Gas phase reactions of TiCl4 andhydrocarbons are a relatively new approach to synthesizingTiC.10 In addition, self-propagating high-temperature synthesis,Mg-thermal reduction,11,12 gas-phase laser-induced reactions,13

mechanical milling,14 and laser chemical vapor deposition15 arealso well-established techniques for preparing TiC powders. Themost popular among these methods are the high-energy mec-

hano-chemical and carbothermal reduction (with mineralizer)techniques.10,16 Recently, Woo et al.17 successfully synthesizedpure TiC nanoparticles from TiO2 and carbon resin without anyaddition of mineralizer. A processing temperature of 15001Cwas required for the reduction of 20–30-nm-TiO2 nanoparticlesinto 80 nm (average size) TiC nanoparticles. Razavi et al.18

demonstrated that the temperature for the synthesis of TiC canalso be reduced by increasing the milling time of both the Tichips and carbon black precursors. Fenga et al.19 developed alow-temperature (5001C) direct chemical synthesis process usingTiCl4 and CaC2. The as-synthesized TiC nanoparticles were 40nm in size and small amount of free carbon as impurity. Gotohand colleagues reported the TiC nanoparticle formation mech-anism via carbothermal reduction of 5 nm TiO2 sol and methylcellulose. Interestingly, the reduction temperature was lowercompared with that of conventional reduction techniques.20

Here, we propose a new additive-free synthesis techniquefor the preparation of spherical and pure TiC nanoparticles us-ing nanorutile (prepared in-house) and an amorphouscarbon powder mixture. The sizes of the obtained TiC particleswere evaluated and quantified using Weibull distribution fit pa-rameters.

II. Experimental Procedure

(1) Powder Preparation

Before carbothermal reduction, we prepared a nanorutile pow-der in-house; details can be found elsewhere.21 Briefly, titaniumtetrachloride, TiCl4 (Junsei, Tokyo, Japan) liquid was used asthe starting precursor for the synthesis of TiO2. Initially, anethanol solvent was slowly added until the entire solution be-came a dark yellowish viscous liquid. Subsequently, ice-coldwater was slowly added to the mixture and stirred continuouslyuntil a clear solution was obtained. It was then kept at 601C for4 h at which time white rutile nanoparticles precipitated. Thefiltrate residue was then washed with hot water to remove anyexcess chloride ions, and it was then dried at 601C for 4 h. Thedried powder was mixed with carbon black (Goldstar Cable,Seoul, Korea) and acetone in a high-density polyethylene (Nal-gene, NY) jar. The powdered mixture was ball-milled for 2 hthrough 3Y-TZP grinding media. The dry black powder mixturewas then placed in an alumina crucible and finally positionedwithin the graphite chamber of a brazing furnace. Thermaltreatment was carried out in argon (Ar) gas at atmosphericpressure. Before heat treatment, pyrolysis behavior of the pow-der mixture was studied through DTA-thermogravimetric (TG)up to 14001C at a heating rate of 101C/min in an Ar gas flow. Inorder to identify the formation of intermediate phases, the pre-cursor mixture was heat treated at 12001C for 1 h. Finally, thepowdered mixture (TiO2–C) was calcined at 13501C for 1, 2, and4 h.

T. Besmann—contributing editor

*Member, The American Ceramic Society.zPresent address: Faculty of Department of Ceramic Engineering, National Institute of

Technology, Rourkela, India.wAuthor to whom correspondence should be addressed. e-mail: chumin@kriss.re.kr

Manuscript No. 25937. Received March 2, 2009; approved July 2, 2009.

Journal

J. Am. Ceram. Soc., 92 [12] 2877–2882 (2009)

DOI: 10.1111/j.1551-2916.2009.03316.x

r 2009 The American Ceramic Society

2877

(2) Powder Characterization

Phase identification of the as-obtained powders was carried outusing X-ray diffraction (XRD, Rigaku, Tokyo, Japan) with Cuaradiation. The voltage and current settings were 40 kV and 40mA, respectively. XRD patterns of the as-synthesis TiO2 andTiC nanopowders were taken in 0.021 steps from 201 to 1451. Theparticle size and residual strain of the synthesized TiC nanopar-ticles was calculated using the Williamson–Hall equation22

b cos y ¼ ldþ 2Z sin y (1)

where b is full-width (radians) of the peak at half-intensity(FWHM), y is the position of peak obtained from the XRDpattern, l is the X-ray wavelength (1.54056), Z powder micro-strain, and d is the average crystallite size. The FWHM wasmodified as b5 (b2meas�b2equip)1/2, where bmeas and bequip are themeasured FWHM and FWHM due to instrumental broadening,respectively. A commercial 2 mm TiC powder (CERAC, WI) wasused in this study to obtain standard FWHMvalues. The particlesizes of the as-synthesized TiO2 and TiC powders were charac-terized through transmission electron microscope (TEM, Tecnai-G2, FEI, OR). X-ray fluorescence (XRF, WDXRF 9800 model,Almelo, the Netherlands) was used to determine the elementalanalysis of TiC nanopowders, both at intermediate and finalstages. The particle size distributions were measured with a scan-ning mobility particle size (SMPS, Grimm, Ainring, Germany)instrument. Here, the collected particles were electrically chargedwith an AM241 neutralizer, detected with a differential mobilityanalyzer, and subsequently counted with a concentration particlecounter. Before estimation of these size distributions, the TiCparticles were dispersed using a 1 wt% polyacrylic acid in water.

III. Results and Discussion

A first step toward the production of TiC nanoparticles was thesynthesis of TiO2 precursor materials. As described in the pre-vious section, titania precursor was produced and TEM wascarried out. The bright field TEM image of the as-synthesizedTiO2 particles is provided in Fig. 1. The selected area diffractionpattern analysis indicated the as-synthesized TiO2 to be in therutile phase. In Fig. 1(a), a needle-like irregular structure of thenanosized rutile powder is evident. The rutile agglomerates con-sisted of 5 nm needle-like TiO2 clusters (Fig. 1(b)). This specificshape can be attributed to the high TiCl4 concentration in water,for the most part. From the synthesis point of view, a high con-centration of TiCl4 initiates rapid hydrolysis and TiO2 phase

formation proceeds by the rearrangement of a TiO6 octahedralnetwork through vertex sharing. As a result, elongated rutilecrystal was produced preferentially. The hydrolysis of TiCl4 re-duces the number of OH ligands. This suppresses edge-sharedbonding and enhances corner-shared bonding—nonsphericaloval-shaped TiO2 nanoparticle is promoted.23 From the firstprinciples analysis for the total energy calculation, the vertex(111) plane of the rutile phase has a higher surface energy thanthat of the edge (110) plane.24 Such anisotropies encourage theformation of head-to-tail agglomeration, resulting in needle-likesecondary particles.

In order to understand the heating induced reactions/disso-ciation, DTA-TG analysis on TiO2–C powder mixture was per-formed in Ar atmosphere at a heating rate of 101C/min (Fig. 2).The pyrolysis process illustrates a significant endothermic peakat 11141C that ranges from 8001 to 13801C. In addition, the TGcurve of the powder mixture shows a gradual weight loss ofabout 8% in the temperature range 6001–9151C and a sharpweight loss of about 34% in the temperature range 9151–11891C. Further weight loss of about 6% occurs up to a tem-perature of 13781C. The first-step weight loss and endothermicpeak are presumed to be due to the rapid formation of an un-stable oxygen-deficient TnO2n�1 (one of the Magneli phases,n5 4–10) phase with a lower oxidation state of Ti. The second-step weight loss is mainly attributed to the conversion ofTnO2n�1 to TiCxOy.

25 Moreover, the final weight loss is gov-erned by the continuous formation of TiC from a nonstoichio-metric TiCxOy phase. The overall weight loss during thispyrolysis process was observed to be 47.89%, which agreeswith the theoretical reduction of TiO2 to TiC. Hence, the entirepyrolysis procedure can be described by the above-mentionedthree-step process, and the carbothermal reduction of TiO2

completes at approximately 13781C. This is possible becauseof the fact that the formation of CO is almost complete in theaforementioned dynamical conditions. In a recent article, it wasobserved that the reduction of TiO2 to TiC is promoted via theformation of various suboxides at different temperatures.20

However, in that work, the intermediate suboxide peak at10501C was not observed. This indicates a rapid and direct con-version of TiO2 to TiCxOy. Such variation in transformationtemperature is due to the starting precursor particle size and re-action atmosphere. In our study, the powder processing tem-perature was lower compared with that of the dynamic pyrolysisbehavior. This difference in temperature appears to be beneficialfor the synthesis of ceramic nanoparticles.26

XRD pattern indicates the precursor TiO2 nanoparticles to beentirely composed of the low crystalline rutile phase and the as-synthesis pure TiC crystalline phase without the presence of any

Fig. 1. Transmission electron microscope image of synthesized needle-like rutile (TiO2) particles (a) and 5 nm oval-shaped primary particles adheringthrough head-to-tail arrangement and form agglomeration (b).

2878 Journal of the American Ceramic Society—Sarkar et al. Vol. 92, No. 12

TiO2 or intermediate phases (Fig. 3). The Pawley refinement re-sult shows that the synthesized nano-TiC particles are of cubicsymmetry with Fm-3m space group having Z5 4 and lattice pa-rameters a5 b5 c5 4.3049(6) . The inset XRD pattern of thepowders calcined at 12001C exhibits low angle shift, which in-dicates an increment in lattice spacing due to the presence ofoxygen ions. In other words, the TiO2 and carbon powder mix-ture forms a TiCxOy phase at low temperature. XRF confirmedthat the Ti:C:O ratio at the intermediate stage was 1:0.68:0.32,

whereas the as-synthesis TiC powder calcination at 13501C for1 h is free from any element or impurity.

High temperature synthesis induces strain in nanoparticles;therefore, an estimation of strain is important. The Williamson–Hall method was used to calculate the volume-weighted crys-tallite size as well as the residual strain of the synthesized TiCnanoparticles.22 The process of determining particle size andstrain from the obtained slope is illustrated in Fig. 4, where Yrepresents b cos y and X represents 2 sin y. Hence, the sloperepresents the strain (Z) and the volume weighted crystallite sizewas calculated from the intercept (l/d). The average crystallitesize and strain of nano-TiC particles was calculated as 91 nmand 0.2%, respectively. The strain of the particle reduced to0.07% and its morphology tended to near-equilibrium shapewith increasing time up to 4 h. In a recent article, the strain of107-nm-sized TiC nanoparticles was reported to be 0.61%.18

In the present case, the volume weighted crystallite size, ob-tained from XRD analysis, was further confirmed by TEManalysis. The changes in particle morphology after synthesizingfor different lengths of time were also investigated using TEM.Typical TEM images of TiC particles synthesized at 13501C for1, 2, and 4 h are shown in Figs. 5(a)–(g), respectively. The latticeparameter obtained for 1 h calcination matches well with thatobtained from the XRD analysis. The particles synthesized at13501C for 1 h are nearly spherical; however, particle size in-creases with time. The development of faceted grain morphol-ogy with sizes up to 200 nm was noticed for longer thermaltreatments. The anisotropic grain growth possibly reduces thestrain in TiC particles, as reported in Table I.

A detailed investigation of particle size distribution is helpfulin predicting the consolidation behavior of nanopowders. Weemphasize that agglomeration during compaction, which leads

Fig. 2. Nonisothermal DTA-thermogravimetric (TG) analysis of tita-nium dioxide and carbon black mixture in argon atmosphere at a heat-ing rate of 101C/min. An endothermic peak is observed at 11141C. TheTG curve is linear beyond 13781C and resultant weight loss is 47.89%.

20 40 60 80 100 120 140

Inte

nsity

(a.

u)

2θ

1200°C/1h

1350°C/1h

40.0 40.5 41.0 41.5 42.0 42.5 43.020 40 60 80 100 120 1402θ

Inte

nsity

(a.

u)

TiO2 (Rutile)

2θ

Inte

nsity

(a.

u)

TiC

Fig. 3. X-ray diffraction patterns of the synthesized titanium carbide (TiC) phase at 13501C for 1 h. Inset indicates the presence of rutile phase inprecursor titanium dioxide (TiO2) and peak shifting (at 12001C) due to nonstoichiometric TiCxOy phase.

December 2009 Synthesis and Morphological Analysis of TiC Nanopowder 2879

to interparticle and interagglomerate pores, is a critical problemand leads to inhomogeneous densification of the compact. Fornanopowders, interagglomerate pores are typically larger than

interparticle pores. In general, commercially available ceramicpowders are designated by their mean particle size, for example100 and 500 nm, 1 mm, etc, without mentioning particle sizedistributions. Here, we address this deficiency using a mathe-matical approach. Usually, the normal distribution representsthe average particle size (D50); however, we used Weibull anal-ysis, which demonstrates the maximum appearance of the par-ticles in sizes, when compared with normal distribution(D634D50). Recently, Ramakrishnan27 studied the particle sizedistribution of different ceramic particles, such as TiO2 andZrO2, through the two-parameter Weibull distribution. How-ever, the degree of distribution (a) or maximum appearance ofparticles in size (b) has not been studied. Thus, it is difficult tounderstand the real meaning of the size distribution withoutrepresentative data for the distribution parameters. We selecteda three-parameter Weibull distribution function that includes acorrection parameter in order to model particle size distributionsobtained from SMPS measurements28 (Fig. 6(a))

f ðxÞ ¼ 1� exp � x� gb

� �a� �(2)

Here, x, a, b, and g are all positive, and f(x) is the cumulativeundersize percent of particle size x present in the distribution.

0.5 0.6 0.7 0.8 0.9 1.00.004

0.005

0.006

0.007

0.008

0.009

0.010

Y = 0.00169 + 0.00824 * X

Βco

sθ

sinθ

Dv= 91.15 nm

� = 0.00206

Fig. 4. Titanium carbide nanoparticle size and strain measurements af-ter synthesis at 13501C for 1 h. The particle size (b5 91.15 nm) andstrain (slope (Z)5 0.00206) was calculated using the Williamson–Hallequation.

Fig. 5. Transmission electron microscope analysis of the titanium carbide phase synthesized at 13501C for 1 h (a–c), 2 h (d–e), and 4 h (f–g), respec-tively. The particles are more faceted and that average particle size increases with processing time.

2880 Journal of the American Ceramic Society—Sarkar et al. Vol. 92, No. 12

Three parameters a, b, and g represent shape, scale, and loca-tion, respectively. The values of these parameters can be ob-tained using linear fits between ln ln[1/(1�f(x))] and Ln(x�g)(Fig. 6(b)). In the present case, we are able to achieve good fits tothe SMPS data using the above three-parameter distributionfunction using a, b, and g equal to 0.96, 84, and 11, respectively.The shape parameter for the synthesized TiC nanoparticlesrepresents smaller value, indicating wider size distribution.The scale parameter indicates that most particles are 84 nm insize, and the location parameter indicates the minimum particlesize to be 11 nm. The as-realized wide distribution is presumablydue to the presence of nanoparticles in water and partial for-mation of agglomerates during SMPS analysis. The as-obtainedb value from the Weibull analysis gradually increases with in-creasing reaction time, which was also confirmed through TEManalysis. Table I represents the physical properties of as-synthe-

sized TiC nanoparticles at 13501C. The shape parameter (a) in-dicates a wider distribution, which is presumably due to thecoarsening effect with increasing time. Based on the Weibullanalysis, the size parameters were fitted with a power-law(dn�dno5At, where, do particle diameter at initial time ‘‘t’’ inmin, A5 constant) of particle coarsening. The calculated coars-ening exponent (n) for the presently described synthesis processis found to be 2.9 (34n42), which agrees with previous graingrowth study.29,30 Volume diffusion is the probable growthmechanism for this process.

IV. Conclusions

Based on the present work, the following conclusions can bedrawn:

(1) The present work demonstrates that carbothermal re-duction of nanosized rutile with amorphous carbon at a (rela-tively) low temperature of 13501C for 1 h leads to the nanosizepure TiC particles. Longer holding times leads to the well-faceted TiC particle.

(2) The synthesis mechanism has been described by a three-stage process of reduction of TiO2 to TiC phase: first, formationof an unstable oxygen-deficient TnO2n�1 phase; second, conver-sion of TnO2n�1 to TiCxOy; and third, formation of TiC fromnonstoichiometric TiCxOy phase.

(3) The as-synthesized TiC nanoparticles are nearly spher-ical with sizes below 100 nm and compositionally pure. Suchparticles can be synthesized at relatively low temperatures.

(4) The three-parameter Weibull function is useful for de-scribing the particle size distribution and their magnitudes. Ex-periments carried out at varying synthesis time of at 13501Crevealed the coarsening behavior and the particle coarseningexponent (n) to be 2.9. Additionally, the particle shape changedfrom spherical (1 h) to more faceted for longer holding timesof 4 h.

References

1A. Mukhopadhyay and B. Basu, ‘‘Consolidation-Microstructure-Property Re-lationships in Bulk Nanoceramics and Ceramic Nanocomposites: A Review,’’ Int.Mater. Rev., 52 [5] 257–88 (2007).

2E. K. Stroms, The Refractory Carbides, Refractory Materials Series, Vol. 2.Academic Press, New York, 1967.

3L. E. Toth, Transition Metal Carbides and Nitrides. Academic Press, NewYork, 1971.

4R. Koc and J. S. Folmer, ‘‘Carbothermal Synthesis of Titanium Carbide UsingUltrafine Titania Powders,’’ J. Mater. Sci., 32 [12] 3101–11 (1997).

5N. A. Hassine, J. G. P. Binner, and T. E. Cross, ‘‘Synthesis of Refractory MetalCarbide Microwave Carbothermal Reduction,’’ Int. J. Refract. Met. Hard Mater.,13 [6] 353–8 (1995).

6H. Preiss, L. M. Berger, and D. Schultze, ‘‘Studies on the Carbothermal Prep-aration of Titanium Carbide from Different Gel Precursors,’’ J. Eur. Ceram. Soc.,19 [2] 195–206 (1999).

7Z. Jiang and W. E. Rhine, ‘‘Preparation of Titanium Nitride (TiN) andTitanium Carbide (TiC) from a Polymeric Precursor,’’ Chem. Mater., 3, 1132–7(1991).

8S. Dutremez, P. Gerbier, C. Guerin, B. Henner, and P. Merle, ‘‘Metal Alkox-ide/Hexa-2,4-diyne-1,6-diol Hybrid Polymers: Synthesis and Use as Precursors toMetal Carbides and Nitrides,’’ Adv. Mater., 10, 465–70 (1998).

9D. B. Miracle and H. A. Lipsitt, ‘‘Mechanical Properties of Fine-GrainedSubstoichiomebic Titanium Carbide,’’ J. Am. Ceram. Soc., 66 [8] 592–7 (1983).

10H. O. Pierson, Handbook of Refractory Carbides and Nitrides. Noyes Publi-cations, Westwood, 1996.

11H. H. Nersisyan, J. H. Lee, and C. W. Won, ‘‘Self-Propagating High-Tem-perature Synthesis of Nano-Sized Titanium Carbide Powder,’’ J. Mater. Res., 17[11] 2859–64 (2002).

12D. W. Lee and B. K. Kim, ‘‘Synthesis of Nano-Structured Titanium Carbideby Mg-Thermal Reduction,’’ Scr. Mater., 48 [11] 1513–8 (2003).

13R. Alexandrescu, E. Borsella, S. Botti, M. C. Cesile, S. Martelli, R. Giorgi,S. Turt, and G. Zappa, ‘‘Synthesis of TiC and SiC/TiC Nanocrystalline Powdersby Gas-Phase Laser-Induced Reaction,’’ J. Mater. Sci., 32 [11] 5629–35 (1997).

14X. L. Cui, L. S. Cui, L. Wang, and M. Qi, ‘‘Synthesis of Titanium CarbidePowder from TiO2 and Petroleum Coke by Reactive Milling,’’ Petrol. Sci. Tech-nol., 20, 999–1007 (2002).

15M. S. Noel and D. Kovar, ‘‘Laser Chemical Vapor Deposition of TiC onTantalum,’’ J. Mater. Sci., 37 [4] 689–97 (2002).

16N. J. Welham and D. J. Llewellyn, ‘‘Formation of Nanometric HardMaterialsby Cold Milling,’’ J. Eur. Ceram. Soc., 19 [16] 2833–41 (1999).

Table I. Physical Properties of Titanium Carbide (TiC)Nanopowders Synthesized at 13501C for Varying Timescale

Processing

time (h)

Strain

(%) Morphology

Weibull parameters (size distribution)

Shape (a) Scale (b)

1 0.20 Spherical 0.9582 842 0.12 Near spherical 0.9372 1174 0.07 Faceted 0.9254 162

Fig. 6. Particle size distribution of titanium carbide nanopowders syn-thesized at 13501C for 1 h (a) and three-parameter Weibull distributionplot, where a and b represents the shape and the scale parameters,respectively; calculated from linear fit of SMPS data (b).

December 2009 Synthesis and Morphological Analysis of TiC Nanopowder 2881

17Y.-C. Woo, H.-J. Kang, and D. J. Kim, ‘‘Formation of TiC Particle duringCarbothermal Reduction of TiO,’’ J. Eur. Ceram. Soc., 27 [2–3] 719–22 (2007).

18M. Razavi, M. R. Rahimipour, and R. Mansoori, ‘‘Synthesis of TiC–Al2O3

Nanocomposite Powder from Impure Ti Chips, Al and Carbon Black by Me-chanical Alloying,’’ J. Alloys Comp., 450 [1–2] 463–7 (2008).

19X. Fenga, Y. Baia, B. Lub, C.-G.Wanga, Y. X. Liua, G. L. Genga, and L. Lic,‘‘Easy Synthesis of TiC Nanocrystallite,’’ J. Cryst. Growth, 264 [1–3] 316–9 (2004).

20Y. Gotoh, K. Fujimura, M. Koike, Y. Ohkoshi, M. Nagura, K. Akamatsu,and S. Deki, ‘‘Synthesis of Titanium Carbide from a Composite of TiO2 Nano-particles/Methyl Cellulose by Carbothermal Reduction,’’ Mater. Res. Bull., 36[13–14] 2263–75 (2001).

21D. Sarkar, M. C. Chu, S. J. Cho, and K. Y. Yoon, ‘‘Barium Titanate Nano-particles and Method of Preparing the Same, Korean-Patent, 10-2007-0119498,2007.

22G. K. Williamson and W. H. Hall, ‘‘X-Ray Line Broadening from FiledAluminium and Wolfram,’’ Acta Metall., 1 [1] 22–31 (1953).

23Y. Chen, A. Lin, and F. Gan, ‘‘Preparation of Nano-TiO2 from TiCl4 byDialysis Hydrolysis,’’ Powder Technol., 167 [3] 106–9 (2006).

24M. Ramanoorthy, D. Vanderbilt, and R. D. King-Smith, ‘‘First-PrinciplesCalculations of the Energetics of Stoichiometric TiO2 Surfaces,’’ Phys. Rev. B, 49[23] 16721–7 (1994).

25G. V. White, K. J. D. Mackenzie, I. W. M. Brown, M. E. Bowden, and J. H.Johnston, ‘‘Carbothermal Synthesis of Titanium Nitride,’’ J. Mater. Sci., 27 [16]4294–9 (1992).

26D. Sarkar, D. Mohapatra, S. Ray, S. Adak, and N. K. Mitra, ‘‘Synthesis andCharacterization of Zirconia Doped Alumina Nanopowder,’’ Ceram. Int., 33 [7]1275–82 (2007).

27K. N. Ramakrishnan, ‘‘Powder Particle Size Relationship in Microwave Syn-thesized Ceramic Powders,’’ Mater. Sci. Engg., A259 [1] 120–5 (1999).

28G. J. Hahn and S. S. Shapiro, Statistical Models in Engineering. Wiley, NewYork, 1967, 287 pp.

29L. Huimin, H. E. Jianping, Y. Bin, and Z. Jishan, ‘‘Grain Growth Law ofSemi-Solid State TiCp/7075 Al Alloy Prepared by Spray Deposition,’’ Acta Met.Sin., 42 [2] 158–62 (2006).

30G. C. Nicholson, ‘‘Grain Growth in Magnesium Oxide Containing a LiquidPhase,’’ J. Am. Ceram. Soc., 48 [10] 525–8 (1965). &

2882 Journal of the American Ceramic Society—Sarkar et al. Vol. 92, No. 12