Synthesis of TiB2 TiC Ti3SiC2 Composites by Reactive Hot Pressing of B4C SiC Ti Powder Mixtures

(1)

Synthesis of TiB

₂

-TiC-Ti

₃

SiC

₂

Composites by Reactive Hot Pressing

of B

₄

C-SiC-Ti Powder Mixtures

Hitoshi Taimatsu

1

, Shigeaki Sugiyama

2

and Shuhou Koseki

1;*

1

Department of Materials Science and Engineering, Faculty of Engineering and Resource Science, Akita University, Akita 010-8502, Japan

2_{Akita Prefectural Research and Development Center, Akita 010-1623, Japan}

TiB2-TiC-Ti3SiC2composites were reaction-sintered from B4C-xSiC-(3þ2x)Ti (x¼0to 1) powder mixtures using reactive hot pressing,

and the phase relation of reaction products and sintering behavior were investigated. The reaction B4CþxSiCþ(3þ2x)Ti!

2TiB2þ(1x)TiCþxTi3SiC2 proceeded fundamentally during hot pressing. Although the reaction demands no production of TiC phase

atx¼1, TiC phase was formed due to the nonstoichiometry of TiC and Ti3SiC2phases produced. The hot-pressed composites containing the

layered compound Ti3SiC2had no preferential texture. Both TiB2and TiC were ﬁrst produced near 960C during heating, and then Ti3SiC2near

1200_{C. The product Ti}

3SiC2contributed eﬀectively to the densiﬁcation of the composites above 1450C.

[doi:10.2320/matertrans.MRA2008174]

(Received May 30, 2008; Accepted July 10, 2008; Published August 20, 2008)

Keywords: titanium boride, titanium carbide, titanium silicon carbide, reaction sintering, solid state reactions

1. Introduction

Titanium boride and carbide, TiB2and TiC, are very hard

and chemically stable compounds with good electrical conductivity like WC. If they can be consolidated easily, their use will increase as hard engineering ceramics possible to be readily formed using electro-discharge machining. Unfortunately their sinterability is so bad that they can not be easily sintered even by hot pressing. For TiB2-TiC

compo-sites, however, a reactive hot pressing technique can be applied. Barsoum et al.1) _{and Brodkin} _{et al.}2,3) _obtained dense TiB2-TiC composites at 1600C using the reactive

hot-pressing of B4C and Ti powder mixtures, in which

the displacement reaction B4Cþ3Ti!2TiB2þTiC

pro-ceeded. This sintering temperature is much lower than that in the ordinary hot pressing of TiB2 and TiC powder

mixtures.4)_{We also successfully applied basically the same} technique to prepare dense composites of TiB2-TiC,5)B4

C-TiB2-TiC6) and TiB2-Ti(C, N).7) Thus the reactive hot

pressing accompanied by the solid state reaction is very eﬀective for the consolidation of TiB2-TiC ceramics.

Titanium silicon carbide, Ti3SiC2, have been drawn

attention as a machinable ceramic.8,9) This ceramic, which is a layered compound with a hexagonal crystal structure, is deformable at high temperatures and resistant to thermal shock.10–12)_{Because of these properties, Ti}

3SiC2addition to

improve fracture toughness of ceramics has been

investigat-ed.13–15) _{However, these properties also propose a use of}

Ti3SiC2 as a ceramic sintering aid for the hot pressing of

diﬃcultly sinterable ceramics. When B4C-SiC-Ti powder

mixtures are used as the starting powder and the reaction B4CþxSiCþ(3þ2x)Ti!2TiB2þ(1x)TiCþxTi3SiC2

is allowed to proceed during hot pressing, TiB2-TiC-base

ceramics may be consolidated at lower temperatures because of the displacement reaction and the deformable product Ti3SiC2. Furthermore, these ceramics probably promise to

have better properties in oxidation resistance as well as in fracture toughness, because Ti3SiC2is much more

oxidation-resistant than TiC.16)_{In this study, therefore, we intended to} prepare dense TiB2-TiC-Ti3SiC2ceramics by the reactive hot

pressing of the B4C-SiC-Ti powder mixtures. Reaction

products and microstructure were surveyed for sintered bodies, and the phase relation of the reaction products and sintering behavior were investigated.

2. Experimental

The starting materials for reaction sintering were B4C

powder (Rare Metallic, average particle diameter 1.05mm, purity 99 mass%, B/C molar ratio 4.0), SiC powder (Ibiden, average particle diameter 0.31mm, impurities (mass%): free SiO2 0.39, free C 1.08, total Al 0.0153, total Fe

0.0173, H2O 0.20) and Ti powder (Rare Metallic, average

particle diameter 24.5mm, purity 99.5 mass%). These pow-ders were weighed and mixed in an Al2O3mortar at ratios to

produce TiB2-TiC-Ti3SiC2 composites in accordance with

the reaction

B4CþxSiCþ(3þ2x)Ti

!2TiB2þ(1x)TiCþxTi3SiC2 (x¼0to 1)

ð1Þ

during reactive hot pressing. Since both the TiC and Ti3SiC2

practically have nonstoichiometry as described later, precise quantity of each product does not obey eq. (1). The mole fractions of the reactants are1=ð4þ3xÞfor B4C,x=ð4þ3xÞ

for SiC andð3þ2xÞ=ð4þ3xÞfor Ti.

The powder mixture was pressed at 50 MPa in a graphite die with an inner diameter of 20 mm and outer diameter of 50 mm. The powder compact was sintered in the die using a resistance-heated hot-pressing system (the Spark-Plasma Sintering Machine, Sumitomo Coal Mining, SPS-2080) in a vacuum under the following conditions: sintering temper-ature 2000C for x¼0, 1600C for x¼0:1 to 0.5 and 1500C for x¼0:5 to 1; heating rate 50C min1; applied pressure 50 MPa; holding time at each sintering temperature *_{Graduate Student, Akita University}

(2)

10 min. The temperature was measured 10 mm inside the outer surface of the die (5 mm outside the sample) through a small hole using an optical pyrometer. The displacement of a pressing ram was monitored during heating. The decrease in length per unit length of a sample for a temperature rise of 1_C, _i.e._{, the shrinkage coeﬃcient of the sample, was} calculated from net displacement obtained by correcting measured displacement for the expansion of load train spacers and work punches during heating. Sintered bodies were examined for reaction products and microstructure using a high-power X-ray diﬀractometer (Rigaku, RINT-2500VHF, output 18 kW, Cu K) and an electron probe

microanalyzer (JEOL, JXA-8200), respectively. Their bulk density was measured by the Archimedean method.

3. Results and Discussion

3.1 Reaction products

X-ray diﬀraction (XRD) patterns for the sintered bodies, which were obtained from their polished surfaces perpen-dicular to the compression direction in hot pressing, showed that TiB2, TiC and Ti3SiC2 were produced as expected in

eq. (1) and that any unreacted B4C, Ti and SiC did not

remained. The XRD patterns of the composites of x¼0:3 and 1 are shown in Fig. 1. The lattice plane having the strongest reﬂection peak listed in the JCPDS ﬁle17–19)is (101) for TiB2, (200) for TiC and (008) for Ti3SiC2. The relative

intensities of diﬀraction peaks were very close to those in the JCPDS ﬁle for both TiB2and TiC, but were not for Ti3SiC2.

In the JCPDS file, the firstly, secondly and thirdly strongest intensities for Ti3SiC2 are reflected from (008), (104) and

(105) planes, respectively. In this study, the strongest intensity for Ti3SiC2 was obtained from (104) plane and

the relative intensities agreed well with calculated ones20) based on its crystal structure. Such inconsistency with the JCPDS ﬁle in the relative intensities for Ti3SiC2 was found

in many XRD analyses, e.g., Ti3SiC2 powder10,20–24) and

surfaces of pressure-sintered Ti3SiC2 polycrystals.12,26) The

pole ﬁgure analysis for surfaces of Ti3SiC2 polycrystals

prepared by the reactive hotpressing of Ti, SiC and C powders showed clearly that they had no preferential texture.27) _Tang _{et al.}28) _{examined the eﬀect of the} glass-mounting of Ti3SiC2powder samples for XRD on the change

in orientation of crystallites, and pointed out that the powder samples easily formed {001} texture in the preparation because of the planner structure of Ti3SiC2 crystallites and

that the relative intensities for a powder prepared to have randomly distributed crystal planes were in agreement with calculated ones.20)_{These facts strongly suggest that such hot} pressing as that used in this study did not bring about a particular texture in a polycrystalline Ti3SiC2due to applied

pressure. According to Murugaiahet al.,24)_{crystal grains at} free surfaces of pressureless-sintered Ti3SiC2 easily have

(008) orientation.

The relative intensity of each product, Ix=Ii, was

calculated as a measure of the product concentration, where

Ixis the peak intensity of the strongest reﬂection for a product X andIithe sum of those for the all products. The Ti3SiC2

phase increased with xin eq. (1), and correspondingly the TiB2and TiC phases decreased, as shown in Fig. 2. Although

eq. (1) demands no existence of TiC phase at x¼1, an unignorable amount of TiC was detected. TiC has wide nonstoichiometry for carbon as widely known, and Ti3SiC2

also is nonstoichiometric. Their nonstoichiometry requires a modiﬁcation in eq. (1).

TiC can have a large carbon deﬁciency: in TiC ranges

from 1 to about 0.5. Storms29) _{summarized many lattice} constant data of TiC, and concluded that the lattice constant

of oxygen-free TiC increases a little with decreasingdown to 0.8 and below this value greatly decreases with decreasing

. Oxygen soluble in TiC lowers the lattice constant. Oxygen is an inevitable impurity in TiC under such an ordinarily-controlled atmosphere as in this work. The lattice constant of TiC phase formed in the sintered bodies is shown in Fig 3. The lattice constant slightly increased withxbelowx¼0:1, and above this value decreased. The lattice constant of TiC phase at x¼0, where TiC of ¼1 must be produced, is smaller than that of oxygen-free TiC of ¼1. This fact

30 40 50 60

Intensity (a.

u.)

TiB2 TiC Ti3SiC2

x = 0.3

x = 1 (104)

(008)

(105)

2θ/ _°

Fig. 1 X-ray diﬀraction patterns of the composites ofx¼0:3and 1.

0 0.2 0.4 0.6 0.8 1

x

TiC Ti3SiC2

Σ

Ii

Relativ

e

intensity

,

Ix

/ TiB2

1500°C 1600°C 2000°C

[image:2.595.313.541.74.278.2] [image:2.595.55.283.558.766.2]

(3)

means that TiC phase contained oxygen and its carbon content decreased with x. The lattice constants a and cof Ti3SiC2 phase formed in the sintered bodies were not

changed withxas shown in Fig. 4. The constantsaandcare in agreement with the values reported by Jeitschko and Nowtony.30)

3.2 Sintering behavior

The bulk density of sintered bodies was shown in Fig. 5. Fully dense bodies were prepared at 2000_{C for} _x_¼_{0, at} 1600_{C for}_x_¼₀_:₁_{and 0.3 and at 1500}_{C for}_x_¼₀_:₅_{to 1. In} the previous study,5) _{the powder mixture of} _x_¼_{0, which} produced only TiB2and TiC, requires a sintering temperature

above 1800_{C for full densiﬁcation by the reactive hot} pressing. The formation of Ti3SiC2 signiﬁcantly improves

the sinterability of TiB2and TiC.

In every shrinkage coeﬃcient curve during heating, several peaks were observed. Shrinkage coeﬃcient curves

during heating are shown in Fig. 6. There are two local maximum points A and B in the coeﬃcient curve forx¼0, where TiB2 and TiC were only produced as reaction

products, and three local maximum points C, D and E in the curves for x0:3, where TiB2, TiC and Ti3SiC2 were

produced. The points A and C are located at the same temperature. According to the study of the solid-state reactions between B4C and Ti powders,31) TiB2 and TiC

are substantially produced at 1000_{C. In our study,} temper-ature was measured for a graphite die. Although real sample temperature becomes higher than die temperature with increasing temperature in the resistance-heated hot-pressing system used in this study, the temperature diﬀerences is small below 1200C, e.g., 1231C at a die temperatures of 1200C.32) Points A and C are both located near 960C. The reaction of eq. (1) accompanies a volume reduction.

0 0.2 0.4 0.6 0.8 1

0.431 0.432 0.433

x

Lattice constant,

a

/ nm

2000°C 1500°C 1600°C

Fig. 3 Lattice constants of TiC phase formed in the composites.

0.2 0.4 0.6 0.8 1

1.76 1.765 1.77

x

Lattice constant,

c

/ nm c-axis 0.3 0.305 0.31

a-axis

Lattice constant,

a

/ nm

1500°C 1600°C

Fig. 4 Lattice constants of Ti3SiC2phase formed in the composites.

0 0.2 0.4 0.6 0.8 1

3.5 4.0 4.5 5.0

x

Bulk density

,

d

/ 10

3 kg

m

-3

1500°C 1600°C 2000°C

Fig. 5 Bulk densities of the composites.

1000 1500 2000

0 2 4 6 8 10 12

Temperature, T / °C

Shr

inkage coefficient,

χ

x = 0.1

x = 0.3

x = 0.5

x = 0.8

x = 1 C

D 0

2 4 6 8 10 12

x = 0

shr

inkage

A

E B

/ 10

-3 K

-1

Fig. 6 Shrinkage coeﬃcient curves for the composites during heating. A and C: formation of TiB2and TiC, B: densiﬁcation of TiB2and TiC,

[image:3.595.56.285.71.273.2] [image:3.595.313.540.76.278.2] [image:3.595.314.538.320.533.2] [image:3.595.56.283.320.530.2]

(4)

Although TiC was in fact produced more than in eq. (1), a volume reduction must occur because the molar volume of the product TiC is the smallest in the reactants and products. Hence, the shrinkage in points A and C was caused by the production of TiB2and TiC mixtures, and that in point B near

1700_{C resulted from their densiﬁcation.}

In the synthesis of Ti3SiC2 from a Ti-Si-TiC powder

mixture, Ti3SiC2is not produced at 1000C, but at 1100C.23)

Gaoet al.24)_{reported that, in the synthesis of Ti}

3SiC2from a

Ti-Si-C powder mixture by the same hot-pressing system as in this study, a large sample volume change due to the production of Ti3SiC2 occurred at 1170C during heating.

These facts suggest that the shrinkage in point D near 1200C is due to the formation of Ti3SiC2. The samples ofx¼0:1

and 0.3 were fully consolidated at 1600_{C, and those of}_x_¼ 0:5to 1 at 1500_{C. Consequently, the shrinkage near point E} is caused by the densiﬁcation of TiB2, TiC and Ti3SiC2

mixtures. Ti3SiC2 contributes to the densiﬁcation of the

composites above 1450_C.

3.3 Microstructure

The sintered bodies had characteristic morphology for microstructure, in which the shapes of starting large Ti particles remained strongly. The microstructure is shown in Fig. 7, where dark phases are TiB2, grey ones TiC and white

ones Ti3SiC2. There are Ti3SiC2 phases outside and inside

of rings composed of very ﬁne TiB2 grains. As shown in

Fig. 7(b), contacts between grains in TiB2 rings are not

necessarily complete. There are very thin Ti3SiC2 layers at

interfaces between grains in many places. Many angular TiB2grains are also formed inside of TiB2rings. Most TiC

phases are formed near centers of TiB2rings. As mentioned

above, TiB2 and TiC phases were ﬁrst produced. The TiB2

phases were probably formed ﬁrst on the peripheries of Ti particles, and the TiC phases inside the TiB2phases forming

rings. The outer diﬀusion of Ti through the TiB2 rings, the

inner diﬀusion of C and Si through the rings and the inner diﬀusion of B through the rings are needed for the respective formation of Ti3SiC2 outside the rings, Ti3SiC2 inside the

rings and angular TiB2grains inside the rings.

The analysis of TiB2 phases by EPMA scarcely showed

solubility of Si and C. The nonstoichometry of TiB2is small:

65.6 to 66.7 mol% B.33)Assuming that TiB2 is

stoichiomet-ric, the average composition of the total TiC and Ti3SiC2

phases is given by the rational formula Ti0:5Six=ð2þ4xÞ Cð1þxÞ=ð2þ4xÞ because the equation

B4Cþ(3þ2x)TiþxSiC

¼2TiB2þ(1þ2x)TiþxSiþ(1þx)C

ð2Þ

holds. The experimental points estimated in the Ti-Si-C phase diagram34) _{are shown in Fig. 8. This diagram well} explains the fact that a large amount of Ti3SiC2and a small

amount of TiC were formed atx¼1. In the diagram, a very slight amount of SiC should coexist in the composite of

x¼0:1. In fact, any SiC phases were not detected by XRD. As thexvalue approaches 1 in the diagram, the correspond-ing tie line between the Ti3SiC2 and TiC phases at

equi-librium moves to the high carbon-deﬁcient TiC side: the

10

(a)

µ

m

10

(c)

µ

m

10

(d)

µ

m

5

TiB

₂

Ti

₃

SiC

2

TiC

(b)

µ

m

Fig. 7 Microstructures (backscattered electron images) of the composites. (a):x¼0:5, 1500_{C; (b): enlarged view of (a); (c):}_x_¼₀_:₁_,

[image:4.595.78.517.72.386.2]

(5)

carbon content of TiC phase formed in a sintered body decreases with increasing x. This is consistent with the experimental results. The phase diagrams35,36)diﬀerent in the constitutional range of Ti3SiC2 phase from the diagram

shown in Fig. 8 were also presented. In these diagrams, Ti3SiC2 phase only must be produced at x¼1. This is

incoherent from the results shown in Fig. 2.

4. Conclusions

The Ti3SiC2 phase in the TiB2-TiC-Ti3SiC2 composites

which were prepared by the reactive hotpressing of B4 C-xSiC-(3þ2x)Ti (x¼0 to 1) powder mixtures increased withx, and correspondingly the TiB2and TiC phases in the

composites decreased fundamentally in accordance with the reaction B4CþxSiCþ(3þ2x)Ti!2TiB2þ(1x)TiCþ xTi3SiC2. At x¼1, however, TiC was formed due to the

nonstoichiometry of TiC and Ti3SiC2 phases. The powder

mixtures of x0:1 which produced Ti3SiC2 were fully

densiﬁed by hot pressing at much lower temperatures than that producing TiB2 and TiC only. No preferential texture

existed in the hot-pressed composites containing Ti3SiC2

phase. Both TiB2 and TiC were ﬁrst produced near 960C

during heating, and their densiﬁcation without Ti3SiC2

formation occurred near 1700_{C. Ti}

3SiC2 phase was

pro-duced near 1200_{C, and it contributed to the densiﬁcation of} the composites above 1450_{C. The carbon deﬁciency of} the product TiC phase was increased with x because the tie line between the Ti3SiC2 and TiC phases at equilibrium

in the Ti-Si-C phase diagram moves to the high carbon-deﬁcient TiC side.

REFERENCES

1) M. W. Barsoum and B. Houng: J. Am. Ceram. Soc.76(1993) 1445– 1451.

2) D. Brodkin, S. R. Kalidindi, M. W. Barsoum and A. Zavaliangos: J. Am. Ceram. Soc.79(1996) 1945–1952.

3) D. Brodkin, A. Zavaliangos, S. R. Kalidindi and M. W. Barsoum: J. Am. Ceram. Soc.82(1999) 665–672.

4) I. I. Spivac, R. A. Andrievskii, V. V. Klimenko and V. D. Lazarenko: Porosh. Metall.140(1974) 17–21.

5) S. Sugiyama, M. Kimura, K. Asari, T. Yoshida, T. Atsumi and H. Taimatsu: J. Jpn. Soc. Powder Powder Met.45(1998) 1065–1070. 6) S. Sugiyama, K. Asari and H. Taimatsu: J. Jpn. Soc. Powder Powder

Met.47(2000) 308–314.

7) S. Sugiyama, K. Asari and H. Taimatsu: J. Ceram. Soc. Jpn.108(2000) 747–752.

8) R. Pampuch, J. Lis, J. Piekarczyk and L. Stobierski: J. Mater. Synth. Process.1(1993) 93–100.

9) J. Lis, R. Pampuch, J. Piekarczyk and L. Stobierski: Ceram. Intern.19 (1993) 219–222.

10) M. W. Barsoum and T. El-Raghy: J. Am. Ceram. Soc. 79 (1996) 1953–1956.

11) M. W. Barsoum, D. Brodkin and T. El-Raghy: Scripta Mater.36(1997) 535–541.

12) M. W. Barsoum and T. El-Raghy: J. Mater. Synth. Process.5(1997) 197–216.

13) Y. M. Luo, S. Q. Li, J. Chen, R. G. Wang, J. Q. Li and W. Pan: Am. Ceram. Soc.85(2002) 3099–3101.

14) W. Pan and S. Shi: J. Euro. Ceram. Soc.27(2007) 413–417. 15) C. Qin, L. Wang, S. Bai, W. Jiang and L. Chen: Key Eng. Mater.336–

338(2007) 1383–1385.

16) M. W. Barsoum, T. El-Raghy and L. U. J. T. Ogbuji: J. Electrochem. Soc.144(1997) 508–2516.

17) Powder Diﬀraction File, (JCPDS-International Center for Diﬀraction Data, Pennsylvania, 1994) No. 35-741.

20) S. Arunajatesan and A. H. Carim: Mater. Lett.203(1994) 19–324. 21) Z. Sun, S. Yang and H. Hashimoto: J. Alloys Comp. 368 (2004)

312–317.

22) H. Li, D. Chen, J. Zhou, J. H. Zhao and L. H. He: Mater. Lett.58(2004) 1741–1744.

23) S. Yang, Z. M. Sun and H. Hashimoto: J. Alloys Comp.368(2004) 312–317.

24) A. Murugaiah, A. Souchet, T. El-Ragy, M. Radovic, M. Sundburg and

C

m

o

l%

C

Ti5Si3C_δ

Ti5Si4

TiSi

Si

TiSi2

Ti3SiC2

Ti

mol% Ti

mo

l%

S

i

TiC_δ

60

40

20

40

Ti3SiC2

x

= 0

x

= 1

20

6

₀

4

0

[image:5.595.142.454.77.310.2]

(6)

M. W. Barsoum: J. Am. Ceram. Soc.87(2004) 550–556.

25) N. F. Gao, J. T. Li, D. Znag and Y. Miyake: J. Euro. Ceram. Soc.22 (2000) 2365–2370.

26) Z. Zhang, Z. Sun, H. Hashimoto and T. Abe: J. Am. Ceram.86(2003) 431–436.

27) B. J. Kooi, R. J. Poppen, N. J. M. Carvalho, J. Th. M. De Hosson and M. W. Barsoum: Acta Mater.51(2003) 2859–2872.

28) K. Tang, C. Wang, X. Xu and Y. Huang: Mater. Lett. 55 (2002) 50–54.

29) E. K. Storms:The Refractory Carbides, (Academic Press, New York, 1967) pp. 1–17.

30) W. Jeitschko and H. Nowtony: Monatsch. Chem.98(1967) 329–337. 31) H. Taimatsu, T. Atsumi, T. Okaniwa and K. Kodama: J. Ceram. Soc.

Jpn.107(1999) 1041–1045.

32) H. Taimatsu, Y. Kodaira and S. Sugiyama: Mater. Trans.46(2008) 1256–1261.

33) T. B. Massalsky:Binary Alloy Phase Diagrams, Vol. 1, (The Materials Information Society, Materials Park, Ohio) pp. 544–548.

34) L. Ratiﬀ and G. W. Powell: Research on Diﬀusion in Multiphase Ternary Systems, AFML Tech. Rep. 70-42, (National Technical Information Service, Alexandria, Virginia, 1970).

35) W. J. J. Wakelkamp, F. J. J. van Loo and R. Metselaar: J. Euro. Ceram. Soc.8(1991) 135–139.

[doi:10.2320/matertrans.MRA2008174]