Low-temperature densification of TiN-TiB2 composites through reactive hot pressing with excess Ti additions

Article abstract of DOI:10.1111/j.1551-2916.2008.02854.x

Reactive hot pressing of Ti and BN powder mixtures is used to produce dense TiN_x-TiB₂ composites. The effect of excess Ti along with a small addition, ～1 wt% Ni, on the reaction and densification of the composite was investigated. A

Full text of DOI:10.1111/j.1551-2916.2008.02854.x

J. Am. Ceram. Soc., 92 [2] 311–317 (2009)
DOI: 10.1111/j.1551-2916.2008.02854.x
r 2009 The American Ceramic Society
ournal
J
Low-Temperature Densification of TiN–TiB2 Composites Through
Reactive Hot Pressing with Excess Ti Additions
Lingappa Rangarajw and Canchi Divakar
Materials Science Division, National Aerospace Laboratories (CSIR), Bengaluru 560 017, India
Vikram Jayaram
Department of Materials Engineering, Indian Institute of Science, Bengaluru 560 012, India
Reactive hot pressing of Ti and BN powder mixtures is used to
produce dense TiNx–TiB2 composites. The effect of excess Ti
along with a small addition, B1 wt% Ni, on the reaction and
densification of the composite was investigated. A composite of
B99.9% relative density (RD) was produced at 12001C at 40
MPa for 30 min with 1 wt% Ni, whereas composites produced
without Ni are porous and contain residual reactants. The
microstructural studies on composite samples with excess Ti pro-
duced at short durations indicate the presence of a transient (Ni–Ti)
phase from which Ti is finally removed to form substoichiometric
TiNx. The hardness of the dense TiNx–TiB2 composite is B22
GPa. The densification mechanism in this system is contrasted with
the role of nonstoichiometry in the Zr–B4C system.
Ti2Ni plays a dominant role in the completion of reaction and
densification at 11001–12001C as a result of high pressure (150
MPa). In the presence of large amounts of Ni, transient liquid
phases can persist below the melting point of Ni because of the
deep eutectics in the Ti–Ni system.15 However, the retention of
large amounts of metal in the final composite is deleterious for all
the major properties, including hardness, high temperature
strength, and chemical inertness. Dense (90%–97% RD) nano-
crystalline TiB2–TiN composites have also been made by me-
chanical and field activation starting with titanium, boron, and
BN powders.16 A combustion process using Ti, BN, and nitro-
gen gas pressure of 0.1–10 MPa produced TiNx–TiB2 composites
with 92.6% RD.17 In all of the above methods, either the dens-
ification temperatures are high or the amount of added Ni is
large. Thus, reducing sintering temperature without compromis-
ing important properties continues to be a challenge.
I. Introduction
In recent times, we have shown that small amounts of Ni can
enable the reactive densification of stoichiometric 3Ti12BN
mixtures to yield TiB2–2TiN composites at temperatures as
low as 16001C with 99.5% RD.18 This development represents
a significant reduction in processing temperature with only mi-
nor additions of metal that do not seriously affect hardness.
During the process of reaction, intermediate alloy compositions
produce a transient liquid phase that enables rapid reaction and
matter transport that aid in densification. The melting points of
Ti and Ni are 16701 and 14551C, respectively (Fig. 1(a)). How-
ever, liquid formation can take place at a lower temperature;
thus melting can begin once the local compositions of the metal
in the vicinity of Ni particles exceed a value that ranges from
10% Ni at the eutectic temperature of B9421C to B4% at
14001C. At the end, Ti in the reaction mixture is completely
converted to the end product phases and elemental Ni remains
randomly distributed in the matrix.
The Ti–Ni and Zr–Ni binary phase diagrams (Figs. 1(a) and
(b))15 are very similar at the Ti/Zr-rich ends (eutectics at B9421
and B9601C, respectively). In an earlier work we have shown
that excess Zr additions to a Zr–B4C mixture can result in
greatly enhanced densification of the ZrB2–ZrC composite while
retaining the excess Zr not as a metallic phase but in the form of
a substoichiometric carbide (ZrCx).19 This technique offers a
conceptual shift in the method by which a metal can aid in the
sintering of a strongly bonded compound without being retained
in the final product as a deleterious phase. The principle is anal-
ogous to the devitrification of residual glass that can improve
the creep resistance of liquid-phase-sintered ceramics. Because
titanium nitride (TiN) is also known to exist over a wide range
of stoichiometry,20 the present work seeks to examine the role of
a similar excursion into excess Ti addition into the densification
of reactive Ti–BN powder mixtures.
LTRA high temperature ceramics based on Zr- and Hf-
borides, carbides, and nitrides are part of a group of ma-
U
terials that are candidates for use at temperatures in excess of
20001C under oxidizing conditions.1–3 Similar materials based
on Ti such as TiB2, TiN, and TiC are hard, inert at moderate
temperatures, and refractory. Possible applications may be for
use as cutting tools and wear resistant parts. Because the melting
temperature of these materials are very high (B30501C), the con-
ventional processing of these materials generally requires high
temperatures in the range of 18001–21001C and/or applied pres-
sure to achieve higher density.4–7 Reactive hot pressing (RHP) of
reactant powder mixtures is known to reduce sintering temper-
atures from those needed when powders of the product phases
are used.8–10 While universal explanations are not available, it
appears likely that the fine microstructural scale of the reaction
products can generate a large density of high diffusivity paths for
matter transport.11 It must be noted that little or no densification
occurs in the absence of applied pressure. RHP of TiH2–BN with
1 wt% nickel (Ni) addition produced 99.9% RD at 30 MPa,
18501C.9 Another alternative that has been explored to reduce
sintering temperatures is the use of much larger quantities of el-
ements, such as Ni. For example, the RHP of Ti–BN with up to
25 at% Ni addition has been studied by several workers.12–14 The
3Ti–2BN powder mixture treated at 150 MPa, 12001C for 2 h
showed incomplete reaction,12 and the addition of Ni produced
B99% RD samples with TiN, TiB2, Ni3Ti, and TiB phases.13,14
The addition of Ni to the powder mixture produces low tem-
perature eutectic (B9421C) in the Ti:Ni (B2:1). The formation of
H.-J. Kleebe—contributing editor
Manuscript No. 25047. Received January 30, 2008; approved October 20, 2008.
This work was support from the Director, NAL, Head, Materials Science Division,
NAL and a grant to the Indian Institute of Science from the Defense Research and
Development Organization.
II. Experimental Procedure
The TiNxB0.8ꢀ1–TiB2 composites were produced by RHP of a
stoichiometric and nonstoichiometric titanium (Ti) and boron
wAuthor to whom correspondence should be addressed. e-mail: ranga@css.nal.res.in
311

312
Journal of the American Ceramic Society—Rangaraj et al.
Vol. 92, No. 2
Ni-Ti
1800
Atomic Percent Nickel
30 40 50 60 70 80 90 100
and particle size B5 mm (M/s Alpha Chemicals, Bengaluru, In-
dia), and Ni B99.5% pure and particle size B4 mm (M/s INCO,
London, U. K.). The required amounts of Ti and BN powders
according reactions (1)–(3) with 1 wt% Ni were mixed in a cen-
trifugal ball mill (Model Pulverisette 6, Fritsch, Oberstein, Ger-
many) for 24 h in hexane. The Ni addition is about 1 wt% of the
total mixture and corresponds to B1.26 wt% of the Ti content.
Some selected powder mixtures were also prepared without Ni.
Powder mixtures were dried at B1001C for 5 h. The dried pow-
der mixtures were hot-pressed in a high-density graphite die (in-
ternal diameter 25 mm and height 70 mm). To avoid direct
contact between the powder and the die/punch assembly, a flex-
ible graphite sheet (0.2 mm thick) was used.18,19 The RHP ex-
periments were conducted in a vacuum hot press (M/s Materials
Research Furnaces, Suncook, NH). The pressure of 40 MPa on
the sample was usually applied within 25 min before reaching
the final temperature, was held for the required time, and re-
leased within 10 min from the start of the cooling sequence. The
sample assembly was cooled to room temperature at 101C/min.
The application of pressure was initiated at B9501C and final
pressure was reached in 15–25 min. The RHP experiments were
conducted at 11001–12001C for 1–30 min. In order to understand
the role of excess Ti, composites of TBN-1 and TBN-3 with Ni
were prepared after partial densification for 1 min at 12001C.
0
10
20
(a)
1670°C
1600
1455°C
1400
1200
1000
882°C
800
600
0
10
20
30
40
50
60
70 80
90 100
Ti
Weight Percent Nickel
Ni-Zr
Atomic Percent Zirconium
20 30 40 50 60 70 80 90 100
0
10
1855°C
(b)
1800
1600
1400
1200
1000
800
(2) Characterization of the Composites
The top and bottom surfaces of reactively hot-pressed compos-
ites were ground and polished using SiC abrasive paper and di-
amond paste down to 0.25 mm using an automatic polishing
machine (Ecomet 4000 with Automet 2000, M/s Buehler, Lake
Bluff, IL) followed by ultrasonic cleaning with acetone. X-ray
diffraction (XRD) patterns were recorded (Philips, Eindhoven,
the Netherlands) using CuKa radiation to identify the phases
(βZr)
(αZr)
˚
0
10
20
30
40
50
60 70
80
90 100
present in the composites. The lattice parameter (A) of TiN in
Ni
Weight Percent Zirconium
stoichiometric and nonstoichiometric composites was deter-
mined by recording the XRD patterns in the 2y range of 901–
1421 (using a graphite monochromator in the diffracted beam)
and extrapolating the calculated lattice parameter using stan-
dard functions to a Bragg angle of 901. Density measurements
were conducted using the Archimedes method after boiling the
samples in water for 1 h and cooling to room temperature. Vic-
kers hardness measurements (Model HSV-20, Shimadzu Co.,
Kyoto, Japan) on the polished surface were performed at a test
load of 4.9 N (500 g) and a holding period of 15 s. An average of
15 readings was taken for each reported value. Microstructural
characterization of the composites was carried out using optical
(Axiovert 200 M MAT, Carl Zeiss Light Microscopy, Goet-
tingen, Germany) and scanning electron microscopy (SEM:
FEI-Sirion, Eindhoven, the Netherlands) with energy dispersive
X-ray microanalysis (EDAX, super-ultra-thin window, Genesis
Spectrum, Mahwah, NJ). The volume fractions of residual metal
were measured through image analysis of the SEM micrographs
by the line intercept method.
Fig. 1. Binary phase diagrams:15 (a) Ti–Ni and (b) Zr–Ni.
nitride (BN) powder mixture according to the following reac-
tions with 1 wt% nickel (Ni). The nitrogen stoichiometry in each
reaction is the value expected based on mass balance:
3Ti þ 2BN ! 2TiN þ TiB2
(1)
(2)
(3)
3:25Ti þ 2BN ! 2:25TiNxꢁ0:89 þ TiB2
3:5Ti þ 2BN ! 2:5TiNxꢁ0:8 þ TiB2
The composites formed according to reactions (1)–(3) will be
referred as TBN-1, TBN-2, and TBN-3, respectively. The ex-
pected volume fractions of TiN and TiB2 phases are 60%–65%
and 40%–35%, respectively. The theoretical densities of the
TBN-1, TBN-2, and TBN-3 composites with 1 wt% Ni accord-
ing to the rule of mixtures are 5.08, 5.03, and 4.96 g/cm3, re-
spectively. For this the density values of TiB2 (4.52 g/cm3),21
TiNxB1 (5.43 g/cm3),21 and TiNxB0.89 (5.22 g/cm3) and
TiNxB0.8 (5.12 g/cm3) have been used. The density of TiNxB0.89
and TiNxB0.8 have been calculated using measured lattice
parameters with the appropriate stoichiometry of nitrogen.
According to thermodynamic data for the 3Ti12BN reaction
reported in the literature,22 the enthalpy of reaction (1) at stan-
dard condition is DH1298 5 ꢀ452 kJ/mol and the free enthalpy is
DG1 5 ꢀ407 kJ/mol at 12271C. It can be seen that the reactions
are exothermic and highly favorable. In order to prevent the
reactions from becoming self-sustaining, the maximum heating
rate used was 101C/min.
III. Results and Discussion
The conditions of RHP, phases observed, lattice parameter of
TiN, and hardness of the composites are given in Table I. Ini-
tially, we present densification results for TiN–TiB2 composites
produced with stoichiometric and nonstoichiometric Ti–BN
powder mixtures with Ni (1 wt%) addition. Subsequently, we
compare the results with those for ZrB2–ZrC composites pro-
duced with the Zr–B4C powder mixture reported earlier.19
(1) Reaction of Ti–BN Powder Mixtures
The XRD patterns of the TiN–TiB2 (TBN-1) and TiN0.8–TiB2
(TBN-3) composites produced with 1 wt% Ni at 40 MPa,
11001C after 1 min are shown in Fig. 2. The formation of
TiN, TiB2, and TiB phases with starting Ti and BN lines may be
observed in Figs. 2(b) and (c). The XRD patterns of the TBN-1,
(1) Materials and Processing
The raw materials used were powders of Ti B99.5% pure and
particle size B44 mm (M/s Alfa-Aesar, MA), BN B99.5% pure

February 2009
Low-Temperature Densification of TiN–TiB2
313
Table I. TiN–TiB2 Composites Produced at Different Experimental Conditions
Experimental conditions
(MPa ꢃ (1C ꢃ min)ꢀ1
Lattice parameter
Density (g/cc)
Hardness
˚
Composites
)
Phases
(A)
(%RD)
(GPa)
TBN-1
TBN-1
TBN-1
TBN-2
TBN-3
TBN-3
TBN-3
TBN-3
40/1100/1 (1 wt% Ni)
40/1200/1 (1 wt% Ni)
40/1200/30 (1 wt% Ni)
40/1200/30 (1 wt% Ni)
40/1100/1 (1 wt% Ni)
40/1200/1 (1 wt% Ni)
40/1200/30 (1 wt% Ni)
40/1200/30
TiN, TiB2, TiB, Ti, BN
TiN, TiB2
TiN, TiB2
4.24570.001
4.25170.001
4.25070.001
4.25070.001
4.23270.001
4.24470.001
4.24670.001
4.24170.001
3.56
4.37 (86)
4.71 (93)
4.92 (97.8)
3.98
4.44 (89)
4.94 (99.9)
3.52
—
12.2170.74
20.1370.7
20.9670.72
—
16.9371.32
22.5770.52
—
TiNx, TiB2
TiNx, TiB2, TiB, Ti, BN
TiNx, TiB2
TiNx, TiB2
TiNx, TiB2, Ti, BN
RD, relative density.
TBN-2, and TBN-3 composites produced with 1 wt% Ni at 40
MPa, 12001C after 30 min, TBN-3 for 1 min, and TBN-3 with-
out Ni after 30 min are shown in Fig. 3. Completion of reaction
can be seen with Ni addition (Figs. 3(b)–(e)) and the observed
phases are TiN and TiB2, whereas the composite produced
without Ni at 12001C after 30 min in Fig. 3(f) showed unreact-
ed Ti and BN in addition to TiN, TiB2 and TiB phases similar to
those seen in Figs. 2(b) and (c).
order of B0.002, which translates to a lattice parameter increase
of B0.008 A, in good agreement with the observed shift.
˚
Micrographs of the polished surfaces of the composites pro-
duced with 1 wt% Ni at 12001C for 30 min (Fig. 5) reveal that
the stoichiometric composition (TBN-1) remains highly porous
while one of the nonstoichiometric (TBN-3) compositions has
an almost fully dense microstructure. Most significantly there is
no evidence for any residual Ti in TBN-3. Higher magnification
SEM micrographs of TBN-3 composites produced without and
with 1 wt% Ni at 12001C for 30 min (Fig. 6) reveal the signifi-
cant increase in density when both Ni and excess Ti are added.
The EDAX spectra clearly distinguish the two phases (Figs. 6(c)
and (d)) in the composite, revealing TiN grains of B2–3 mi-
crometers and TiB2 of submicrometer dimensions. Elemental
analysis using EDAX of the Ti/N ratio in TiN grains of TBN-1
yielded a value of 51.670.9/48.470.9, whereas in TBN-3 the
value was 58.571.9/41.471.9. The accurate quantification of
light elements such as N is prone to systematic errors, especially
when the grains are approximately micrometer dimensions.
However, if we take the stoichiometric composite (TBN-1) as
representing a 1:1 standard (Ti:N), the increase in the Ti:N ratio
in the nonstoichiometric composite (TBN-3) is a significant in-
dicator of nonstoichiometry. Coupled with the XRD (Fig. 3(d))
and SEM evidence, indicating the absence of free Ti, these ob-
servations strongly suggest that the excess Ti has led to the for-
mation of nitrogen-deficient TiNx, as originally hypothesized in
reaction (3).
(2) Nonstoichiometric TiN in the Composites
It is known that TiN can form over a wide range of nonstoi-
chiometric compositions with different N/Ti ratios (0.6–
1.1).20,23,24 The lattice parameter of TiN has a maximum of
4.24 A in the stoichiometric composition.20 The extrapolation of
˚
the data to a Bragg angle of 901 reveals (Fig. 4) that the lattice
parameter of TiNx drops from 4.250 (x 5 1) to 4.246 (x 5 0.8).
The precision in Bragg angle is B0.021, which translates to a
precision of B2 parts in 104 in the lattice parameter. When
compared to the literature, the magnitude of the shift because of
nonstoichiometry is lower, while in the present work the abso-
lute lattice parameters are higher. For example, the lattice pa-
rameter for monolithic TiN in the powder diffraction standard
˚
JCPDS: 38–1420 is 4.2417 A, while for the metal-nitrogen gas
reaction the lattice parameters of the produced TiN0.792 and
TiN0.999 are 4.2294 A and 4.2389 A, respectively.23 However, the
values for TiN in the RHP composites are smaller than those
˚
˚
˚
˚
reported for TiNx–TiB2 composites (4.26 A TiN0.79 and 4.29 A
TiN) produced by combustion synthesis under nitrogen gas
pressure.17 Part of the reason for the discrepancy may be attrib-
uted to the residual stress present because of the thermal ex-
(3) Densification of the TiN–TiB2 Composites
A comparison of the densities achieved after 30 min at 12001C
(Table I) reveals that in the presence of Ni there is a substantial
influence of nonstoichiometry in raising the RD from 93%
(TiN–TiB2) to 99.9% (TiN0.8–TiB2). It may be noted that
when Ni is absent, the RD for TBN-3 is only B70% (Table
I), which is similar to the 78% that has been reported for a sto-
ichiometric composite produced previously at 14001C.18 Thus,
pansion coefficient (TEC, a) mismatch between TiB2 (5ꢂ 10ꢀ6
/
K) and TiN (9ꢂ 10ꢀ6/K). Using a volume rule of mixtures for
the TEC of the composite (TBN-1: 7.48 ꢂ 10ꢀ6/K and TBN-3:
7.67 ꢂ 10ꢀ6/K) and treating the TiN as an embedded inclusion,
it can be shown that for a processing temperature of 12001C, the
TiN would be subject to a tensile residual strain, DaDT, of the
Fig. 3. The X-ray diffraction patterns of the composites produced at 40
MPa/12001C: (a) TBN-3 starting mixture, (b) TBN-1 with Ni for 30 min,
(c) TBN-2 with Ni for 30 min, (d) TBN-3 with Ni for 30 min, (e) TBN-3
with Ni for 1 min, and (f) TBN-3 without Ni for 30 min (m, hBN; }, Ti;
& , TiN; J, TiB2; and D, TiB).
Fig. 2. The X-ray diffraction patterns of the composites produced with
1 wt% Ni at 40 MPa/11001C after 1 min: (a) TBN-3 starting mixture, (b)
TBN-1, and (c) TBN-3 (m, hBN; }, Ti; & , TiN; J, TiB2; and n, TiB).

314
Journal of the American Ceramic Society—Rangaraj et al.
Vol. 92, No. 2
randomly (Fig. 7(a)). A bright core and dark rim are observed in
the particles in TBN-3 (Fig. 7(b)). The EDAX spectra shown in
Figs. 7(c) and (d) reveal the core to be enriched with Ni relative
to the rim. The compositional analysis suggests that Ti is being
removed from the Ni–Ti region toward the TiN matrix at the
rim, and the contrast of the rim is similar to that of the TiN
matrix. Image analysis reveals that the volume fractions of Ni-
rich regions are significantly greater in TBN-3 (1.3%) as com-
pared with TBN-1 (0.7%). It was also observed that TBN-1 is
characterized by an inhomogeneous pore distribution with large
pores (Fig. 8(a)) and large pore-free regions, while the pores in
TBN-3 are more uniformly distributed and smaller in size
(Fig. 8(b)). Thus, when the RHP duration is extended to 30
min, TBN-1 retains a significant amount of porosity (Fig. 5(a)),
while TBN-3 is almost fully dense (Fig. 5(b)). SEM at higher
magnification reveals that after densification is complete in
TBN-3 (Fig. 5(b)), the residual Ni is similar to that seen in the
TBN-1 that was fully densified at 16001C.18 Thus, there is a
strong suggestion that during the intermediate stage of sintering
at comparable relative densities, the presence of excess Ti leads
to larger volume fraction of liquid when compared to the sto-
ichiometric composition and also that the Ti in this liquid is
subsequently removed through the creation of a nonstoichio-
metric nitride. The high hardness (B22 GPa) achieved in the fi-
nal composite is a testimony to the microstructural refinement
that accompanies the low processing temperature and the
absence of any residual metallic phase other than the B0.5
vol% Ni.
Fig. 4. Cos2 y/sin y—lattice parameter plot of TiNx in TBN-1, TBN-2,
and TBN-3 composites produced with 1 wt% Ni at 40 MPa, 12001C for 30
min: (a) y5ꢀ0.0035x14.2497 (R2 50.8195), (b) y5ꢀ0.0045x14.2492
(R2 50.8135), and (c) y5ꢀ0.002x14.2464 (R2 50.9081).
nonstoichiometry (excess- Ti) benefits densification only in the
presence of Ni.
The RDs of TBN-1 and TBN-3 composites with 1 wt% Ni
after partial densification at 12001C for 1 min are 86% and 89%,
respectively (Table I). The SEM micrographs of these TBN-1
and TBN-3 composites are shown in Fig. 7. The microstructure
shows the TiB2 and TiN phases, with Ni particles distributed
Reference to the Ti–Ni binary phase diagram (Fig. 1(a)) sug-
gests a mechanism for the role of excess Ti in promoting
the densification. Reaction and densification proceed in paral-
lel. During the conversion of a stoichiometric reactant mixture
to the product phases, the Ni-rich melt follows a composi-
tional trajectory at 12001C that leads to solidification once the
alloys reaches B40 wt% Ni. Thus, excess Ti ensures that even
after the BN is fully reacted, the alloy composition remains Ti-
rich and, as a result, molten, thereby allowing densification to
proceed.
(a)
Thus, in summary, the traditional route to reducing densifi-
cation temperatures relies on increasing metal (Ni, Fe, etc.) ad-
ditions that remain in the metallic phase but degrade the
properties of the composites. In contrast, excess Ti additions
serve the purpose of producing larger amounts of liquid phase
which can persist longer during densification, even at a low
temperature, and which converts into a nonstoichiometric nit-
ride. The existence of a compositional stability range for TiN is
crucial in enabling the excess Ti to be eventually eliminated from
the transient liquid.
(b)
(4) Comparison of Densification Mechanisms for TiN–TiB2
and ZrB2–ZrC Composites
We conclude with a comparison of TiNx–TiB2 behavior with
that of a ZrB2–ZrC composite obtained by RHP of Zr–B4C
mixtures.19 In the Zr–B4C system, RHP of mixtures with excess
Zr was, analogous to the present system, able to bring down the
temperature for densification from 16001 to 12001C. A compar-
ison of the densification behavior of the two systems as a func-
tion of stoichiometry and nickel addition is shown in Fig. 9. The
domain of an accelerated densification in Zr–B4C is character-
ized by Zr nonstoichiometry, while in Ti–BN the same domain is
distinguished by the presence of Ni and is further assisted by Ti
nonstoichiometry. In the case of the Ti–BN mixture with excess
Ti reaction (3) and without Ni, the reaction and densification do
not complete at 12001C (Fig. 3(e)), owing to the absence of a
liquid phase. RD increases monotonically with deviation from
Ti stoichiometry as long as Ni is present. However, mixtures
containing excess Zr densify even when Ni is absent, and the
enhancement of reaction and densification in the presence of Ni
in ZrB2–ZrC composites is small. The accelerated densification
in nonstoichiometric Zr–B4C mixtures has been attributed to the
enhanced plasticity and diffusion in a carbon-deficient ZrCx
Fig. 5. Optical micrographs of the composites produced with 1 wt% Ni
at 40 MPa, 12001C for 30 min: (a) TBN-1 (stoichiometric composite)
showing significant porosity and (b) TBN-3 (nonstoichiometric compos-
ite) showing a nearly fully dense microstructure.

February 2009
Low-Temperature Densification of TiN–TiB2
315
(a)
(b)
5 μm
5 μm
(c)
(d)
Fig. 6. Scanning electron microscopic micrographs of the TBN-3 composite produced at 40 MPa, 12001C for 30 min: (a) without Ni and (b) with 1 wt%
Ni. In (b) the light gray regions are TiN and the dark regions are TiB2, as shown by the EDAX spectra in (c) and (d), respectively.
phase.19 The reason why Ni is relatively less important in this
case appears to lie in the natural rate of the displacement reac-
tion in the two systems. We note from the solidus curves of the
Zr/Ti-rich end of the respective phase diagrams (Figs. 1(a) and
(b)), that a liquid phase can form only if there is significant Zr or
Ti remaining unreacted, above the eutectic temperature. In the
(a)
(b)
Region A
Ni
2 μm
2 μm
(c)
(d)
Fig. 7. Scanning electron microscopic micrographs with energy dispersive X-ray (EDAX) analysis of the composite produced with 1 wt% Ni at 40
MPa, 12001C for 1 min: (a) TBN-1, white regions are Ni, (b) TBN-3, showing the Ni-rich region (A), (c) EDAX spectrum of Ni-rich region in A, and (d)
EDAX spectrum of Ti-rich rim of the B50 nm region that surrounds A.

316
Journal of the American Ceramic Society—Rangaraj et al.
Vol. 92, No. 2
(a)
Pore
50 μm
(b)
Pore
Fig. 9. The relative density-temperature plot of the composites: (a) ~,
TiN–TiB2 without Ni18; & , TiN–TiB2 with Ni18; m, TiN0.89 –TiB2 with
Ã
Ni; , TiN0.8–TiB2 with Ni; and ꢄ, TiN0.8–TiB2 without Ni; and (b) ~,
ZrB2–ZrC with out Ni19; & , ZrB2–ZrC with Ni19; J, ZrB2–ZrC0.67
with Ni19; and m, ZrB2–ZrC0.67 without Ni.19
50 μm
that has been invoked in substoichiometric ZrCx-containing
composites.
Fig. 8. Optical micrographs of the composites produced with 1 wt% Ni
at 40 MPa, 12001C for 1 min: (a) TBN-1 and (b) TBN-3. Note the larger
and more inhomogeneous pore distribution in (a).
Acknowledgments
The authors are indebted to Mr. V. Babu for assistance in the hot-pressing
experiments, Dr. S. Usha Devi for XRD measurements, and Mr. Keshab Barai for
assistance in the SEM work.
case of the Zr–B4C mixture with excess Zr, reaction as well as
densification appears to be completed readily at low tempera-
tures even when Ni is absent.19
References
1K. Upadhya, J. M. Yang, and W. P. Hoffman, ‘‘Materials for Ultrahigh Tem-
perature Structural Applications,’’ Am. Ceram. Soc. Bull., 76 [12] 51–6 (1997).
2F. Monteverde, S. Guicciardi, and A. Bellosi, ‘‘Advances in Microstructure and
Mechanical Properties of ZrB2 Based Ceramics,’’ Mater. Sci. Eng., A, A346 [1–2]
310–9 (2003).
IV. Conclusions
3C. Mroz, ‘‘Annual Mineral Review, Zirconium Diboride,’’ Am. Ceram. Soc.
Bull., 74 [6] 164–6 (1995).
(1) It is possible to use metallic additives in reactive hot-
pressing of Ti–BN in order to promote low temperature dens-
ification of TiB2–TiN composites without compromising the
harder and refractory nature of the product.
(2) Excess (up to 2.96 wt%) Ti, in the presence of 1 wt% Ni,
in a reactive mixture of Ti–BN leads to full density at 12001C,
which is a reduction of B4001–6501C compared with previously
reported temperatures for stoichiometric mixtures.
(3) The excess Ti is present in the form of nitrogen-deficient
TiN in the composite, while the volume fraction of metal in the
final composite remains at B0.5% Ni. The microstructural re-
finement from low temperature densification and the absence of
significant residual metal lead to a hardness of B22 GPa, which
is higher than that reported earlier.
(4) Ni aids in the formation of a transient liquid, which
persists for a longer duration in the presence of excess Ti.
(5) While the consequences of nonstoichiometry on densifi-
cation are nominally similar to those in the Zr–B4C system, the
mechanism involved in the present case is dominated by the
persistence of a transient liquid phase in contrast to the plasticity
4K. Shobu and T. Watanabe, ‘‘Hot Pressing of TiN–TiB2 System,’’ J. Powder
Metall. Soc. Jpn., 32 [6] 215–8 (1995).
5T. Watanabe, H. Yamamoto, K. Shobu, and T. Sakamoto, ‘‘Factors Affecting
the Porosity and Bending Strength of Ti(CN)–TiB2 Materials,’’ J. Am. Ceram.
Soc., 71 [4] C-202–4 (1987).
6K. Shobu, T. Watanabe, Y. Enomoto, K. Umeda, and Y. Tsuya, ‘‘Frictional
Properties of Sintered TiN–TiB2 and Ti(CN)–TiB2 Ceramics at High Tempera-
ture,’’ J. Am. Ceram. Soc., 70 [5] C-103–4 (1987).
7T. Graziani, C. Melandri, and A. Bellosi, ‘‘Fabrication of Monolithic TiN and
TiN–20 vol%TiB2 Composites,’’ J. Hard Mater., 4, 29–36 (1993).
8G. J. Zhang, ‘‘Preparation of TiB2–TiC0.5N0.5 Ceramic Composite by Reactive
Hot Pressing and its Microstructure,’’ Ceram. Int., 21 [1] 29–31 (1995).
9G. J. Zhang, Z. Z. Jin, and X. Yue, ‘‘TiN–TiB2 Composites Prepared by Re-
active Hot Pressing and Effects of Ni Addition,’’ J. Am. Ceram. Soc., 78 [10] 2831–
3 (1995).
10R. Tomoshige, A. Murayama, and T. Matsushita, ‘‘Production of TiB2–TiN
Composites by Combustion Synthesis and Their Properties,’’ J. Am. Ceram. Soc.,
80 [3] 761–4 (1997).
11K. Wang and V. D. Kristic, ‘‘Reaction Sintering of TiN–TiB2 Ceramics,’’ Act.
Mater., 51, 1809–19 (2003).
12F. Olevsky, P. Mogilevsky, E. Y. Gutmanas, and I. Gotman, ‘‘Synthesis of In
Situ TiB2/TiN Ceramic Matrix Composites from Dense BN–Ti and BN–Ti–Ni
Powder Blends,’’ Metall. Mater. Trans., 27A, 2071–9 (1996).

February 2009
Low-Temperature Densification of TiN–TiB2
317
13I. Gotman, N. A. Travitzky, and E. Y. Gutmanas, ‘‘Dense In Situ TiB2/TiN
and TiB2/TiC Ceramic Matrix Composites: Reactive Synthesis and Properties,’’
Mater. Sci. Eng., A244, 127–37 (1998).
19L. Rangaraj, S. J. Suresha, C. Divakar, and V. Jayaram, ‘‘Low-Temperature
Processing of ZrB2–ZrC Composites by Reactive Hot Pressing,’’ Metall. Mater.
Trans. A, 39 [7] 1496–05 (2008).
14E. Y. Gutmanas and I. Gotman, ‘‘Reactive Synthesis of Ceramic-Matrix
Composites Under Pressure,’’ Ceram. Int.., 26, 699–707 (2000).
15Alloy Phase Diagrams , ASM Handbook, Vol. 3. ASM International, Mate-
rials Park, OH, 1992.
20L. E. Toth, Transition-Metal Carbides and Nitrides, Vol. 7, p. 87. Academic
Press, New York, 1971.
21A. W. Weimer, Carbide, Nitride and Boride Materials Synthesis and Process-
ing: Table A.1 Crystallographic Properties A. ISBN:0 412 54060 6. Chapman and
Hall, U.K., 1997.
16J. W. Lee, Z. A. Munir, M. Shibuya, and M. Ohyanagi, ‘‘Synthesis of Dense
TiB2–TiN Nanocrystalline Composites Through Mechanical and Filed Activa-
tion,’’ J. Am. Ceram. Soc., 84 [6] 1209–16 (2001).
22G. J. Zhang, M. Ando, J. F. Yang, T. Ohji, and S. Kanzaki, ‘‘Boron Carbide
and Nitride as Reactants for In Situ Synthesis of Boride-Containing Ceramic
Composites,’’ J. Eur. Ceram. Soc., 24, 171–8 (2004).
17M. Shibuya, M. Ohyangi, and Z. A. Munir, ‘‘Simultaneous Synthesis and
Densification of Titanium Nitride/Titanium Diboride Composites by High Nitro-
gen Pressure Combustion,’’ J. Am. Ceram. Soc., 85 [12] 2965–70 (2002).
18L. Rangaraj, C. Divakar, and V. Jayaram, ‘‘Reactive Hot Pressing of Tita-
nium Nitride–Titanium Diboride Composites at Moderate Pressures and Tem-
peratures,’’ J. Am. Ceram. Soc., 87 [10] 1872–8 (2004).
23S. Nagakura, T. Kusunoki, F. Kakimoto, and Y. Hirotsu, ‘‘Lattice
Parameter of the Non-Stoichiometric Compound TiNx,’’ J. Appl. Cryst., 8, 65–6
(1975).
24H. A. Wriedt and J. L. Murray, ‘‘The N–Ti (Nitrogen–Titanium) System,’’
Bull. Alloy Phase Diagrams, 8 [4] 378–88 (1987).
&