Effect of Ni content on the products of Ni-Ti-B system via self-propagating high-temperature synthesis reaction

Article abstract of DOI:10.1016/j.jallcom.2007.03.054

The effect of Ni content on the products of Ni-Ti-B system via self-propagating high-temperature synthesis (SHS) reaction has been investigated in this research. The results show that the products of SHS reactions consist mainly of TiB₂ and Ni.

Full text of DOI:10.1016/j.jallcom.2007.03.054

Journal of Alloys and Compounds 457 (2008) 286–291
Effect of Ni content on the products of Ni–Ti–B system via
self-propagating high-temperature synthesis reaction
L. Huang, H.Y. Wang, Q. Li, S.Q. Yin, Q.C. Jiang∗
Key Laboratory of Automobile Materials of Ministry of Education and Department of Materials Science and Engineering,
Nanling Campus, Jilin University, No. 5988 Renmin Street, Changchun 130025, PR China
Received 30 January 2007; received in revised form 11 March 2007; accepted 12 March 2007
Available online 16 March 2007
Abstract
The effect of Ni content on the products of Ni–Ti–B system via self-propagating high-temperature synthesis (SHS) reaction has been investigated
in this research. The results show that the products of SHS reactions consist mainly of TiB2 and Ni. Besides, the transient phases of Ni4B3, Ni3B,
NiB and Ni3Ti also exist in the final products, which means the SHS reactions of Ni–Ti–B system are incomplete. The change in the content of Ni
within the chosen range from 30 to 70 wt.% has little effect on the phase compositions of the final products. However, the sizes of TiB2 particulates
have been greatly influenced by the Ni content. The average sizes of normal TiB2 particulates are nearly the same and about 4–6 ␮m when Ni
contents are 30, 40 and 50 wt.%. Furthermore, TiB2 particulates in the products of these three systems present exaggerated growth and their sizes
can even reach 10–15 ␮m. The size of TiB2 particulates decreases dramatically to 1–2 ␮m when Ni content increases to 60 wt.% while to 0.6 ␮m or
less at 70 wt.% Ni. The addition of Ni facilitates to form more liquid phases that are beneficial to TiB2 formation during SHS reaction process. The
formation mechanism of TiB2 in Ni–Ti–B system can be characterized by the solution, reaction and precipitation processes. This can be further
substantiated by the presence of remaining liquids, the typical hexagonal-prism morphology and the growth striation on (0 0 0 1) crystal face of
TiB2 particulates.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Ceramics; Self-propagating high-temperature synthesis (SHS); Microstructure; Field emission scanning electron microscope (FESEM); X-ray diffraction
(XRD)
1. Introduction
bustion velocity of Ti–B system (80–90 mm/s [1]) is too high to
be controlled, and the ignition temperature of that system is also
very high, which has not been measured accurately so far. In
addition, it is difficult to density the SHS products of Ti–B sys-
tem due to the low crystalline boundary diffusion coefficient of
TiB2 [4,6,7]. Therefore, it is necessary to add metal elements as
diluents to Ti–B system to reduce the heat evolved during SHS
reactions and synthesize finer TiB2 particulates. Simultaneously,
the metallic phase can be used as the bonding agent to obtain
dense bulk materials and improve the mechanical properties of
products.
Recently, most work on the SHS reactions of Ti–B sys-
tem with the addition of different metals has been focused on
Al–Ti–B [1,8–11], Fe–Ti–B [12,13], Cu–Ti–B [14,15] and so
forth. Taneoka et al. [1] produced TiB2–Al composites via SHS
reaction of Al–Ti–B system. The average size of TiB2 par-
ticulates synthesized (∼5 ␮m) was much smaller than that of
the starting Ti, and it may be due to the partial melting of Ti
behind the combustion front, which resulted in the increase of
The self-propagating high-temperature synthesis (SHS) reac-
tion of Ti–B system has attracted more and more attention during
the last several decades [1]. One of the most notable character-
istics of SHS reaction of that system is the high exothermicity,
which is because of the large formation heat release of TiB2
(ꢀHTiB = −324 kJ/mol) [1–3]. The high exothermicity result-
2
ing from the formation of TiB2 benefits the process of SHS
reaction, however, itcanalsocausesomeproblems. Forexample,
TiB2 tends to grow exaggeratedly at high-temperature, which
creates strong internal stresses due to its anisotropy effects and,
hence, induces the formation of transgranular cracks [4,5]. As a
consequence, both physical and mechanical properties of TiB2
synthesized will be reduced dramatically. Furthermore, the com-
∗
Corresponding author. Tel.: +86 431 8509 4699; fax: +86 431 8509 4699.
E-mail address: jiangqc@mail.jlu.edu.cn (Q.C. Jiang).
0925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.03.054

L. Huang et al. / Journal of Alloys and Compounds 457 (2008) 286–291
287
nucleation rate of TiB2. Gotman et al. [8], Nikitin et al. [9]
and Wang et al. [10] also synthesized TiB2–Al composites via
SHS reactions of Al–Ti–B system, and very-fine-scale TiB2 par-
ticulates ranging from tens of nanometers up to 1–2 ␮m were
obtained. Moreover, they all found that Al3Ti phase existed in
the final products, which may be due to the short dwell time of
the combustion zone, and, hence, the full conversion of Ti–TiB2
had not been achieved [8,10].
Compared with the research on the SHS reactions of Al–Ti–B
system, the one on SHS reactions of Fe–Ti–B system is a little
limited, though the wettability between TiB2 and Fe is very
good [11]. Maksimov et al. [12] and Degnan and Shipway [13]
investigated the combustion synthesis of Fe–Ti–B system and
obtained fine TiB2 particles (1–5 ␮m) with a relatively homoge-
neous dispersion in the Fe matrix by lowering the melting point
of reactants.
In addition, using Cu as the diluent of Ti–B system is
also investigated. Xu et al. [14] successfully synthesized dense
TiB2–Cu composites without other interphase compounds (such
as CuTi) via SHS combined with quasi-static consolidation and
found that fine TiB2 particulates grew in near equivalent axis-
like and block-like shapes. The similar result was obtained by
Kwon et al. [15] when fabricating for the Cu-60 vol% TiB2
and Cu-70 vol% TiB2 compositions by the SHS reactions of
Cu–Ti–B system. However, the interfacial bonding of TiB2/Cu
is not good enough because the wettability angle of TiB2/Cu
is 142◦ [14,16] in vacuum, which should be improved in the
subsequent research.
Although, the research on SHS reactions of Al (Fe, Cu)–Ti–B
systems has been well developed, the one on Ni–Ti–B system is
very limited. Horlock et al. [3] deposited Ni (Cr)–TiB2 cermet
coatings onto steel substrates from a feedstock powder produced
via the SHS reaction from Ni (Cr), Ti and B powders. The
SHS reacted powder was found to comprise principally 1–5 ␮m
sized TiB2 in a crystalline, Ni-based solid solution matrix, with
approximately 60 vol% TiB2 and 40 vol% metallic binder. In
addition, XRD patterns revealed that there were around 2% of
TiB, Ni2B, NiTi and TiO phases. In the present study, the pur-
pose is to investigate the effect of Ni content on the products
of Ni–Ti–B system via SHS reactions. Results on the thermo-
dynamic calculation of the SHS reactions, phase compositions,
microstructures and formation mechanisms of the products are
reported.
2000, America) and X-ray diffraction (XRD) using Cu K␣ radiation (Rigaku-
D/Max 2500PC, Japan).
3. Results and discussion
3.1. Thermodynamic calculation
In order to determine the possible phases produced and the
direction of SHS reactions in Ni–Ti–B system, the reactions
that may take place based on 1 mol reactants are conducted as
follows:
13 Ti(s) + B(s)
→
→
→
→
→
13 TiB2(s)
12 NiB(s), ꢀG◦2
17 Ni4B3(s) ꢀG◦3
,
ꢀG◦1
(1)
(2)
(3)
(4)
(5)
2
3
12 Ni(s) + B(s)
1
2
3
47 Ni(s) + B(s)
,
7
34 Ni(s) + B(s)
14 Ni3B(s), ꢀG◦4
14 Ni3Ti(s), ꢀG◦5
1
4
1
34 Ni(s) + Ti(s)
4
The standard Gibbs free energy changes (ꢀG◦) for Eqs. (1–5)
as functions of temperature are theoretically calculated accord-
ing to thermodynamic data [17,18] and the results are shown
in Fig. 1. It can be observed that the ꢀG◦ of the five reactions
are all negative, which indicates that the above reactions are all
favorable thermodynamically. Furthermore, the result reveals
that ꢀG◦1 is far more negative than the others in the temperature
range investigated, which means Eq. (1) is easier to occur ther-
modynamically than Eqs. (2–4). In addition, it shows that Ni and
TiB2 are the thermodynamically stable phases in the products.
According to the above analyses, it is supposed that the pos-
sible SHS reactions in Ni–Ti–B system can be summarized as
an overall reaction:
1
3
2
3
1
3
xNi + (1 − x)Ti + (1 − x)B → xNi + (1 − x)TiB2
(6)
The adiabatic combustion temperature (Tad) of the Ni–Ti–B
system according to Eq. (6) can be calculated based on the rela-
tionship given by Ref. [19], which assumes that the propagating
2. Experimental
The starting materials were made of commercial powders of 30, 40, 50,
60 and 70 wt.% Ni (99.8% purity, ∼5 ␮m), Ti (99.5% purity, ∼38 ␮m) and B
(98.0% purity, ∼3 ␮m), respectively, with a molar ratio of B/Ti = 2.0. Powder
blends were dry mixed sufficiently by ball milling for 6 h and then pressed
into cylindrical preforms (20 mm diameter and 10 mm length) under pres-
sures ranging from 80 to 85 MPa to obtain densities of 59 4% theoretical
density.
The SHS experiments were conducted in a combustion chamber under a pro-
tective atmosphere of industrial argon (99.9%). The preforms were placed on the
graphite-flat, which was fixed above a tungsten electrode and ignited by the arc
heat. Microstructures and phase analyses of the SHS products were investigated
by using field emission scanning electron microscope (FESEM) (FEI-XL30,
America) equipped with energy-dispersive spectrum (EDS) (EDAX-Genesis
Fig. 1. Change of standard Gibbs free energy as temperature for ꢀG◦1 (TiB2),
ꢀG◦2 (NiB), ꢀG◦3 (Ni4B3), ꢀG4◦ (Ni3B) and ꢀG5◦ (Ni3Ti).

288
L. Huang et al. / Journal of Alloys and Compounds 457 (2008) 286–291
Fig. 2. Adiabatic combustion temperature calculated according to different Ni
contents.
mode is initiated at room temperature without any preheat, as
shown in Eq. (7):
ꢀ
Tad(298)
ꢁ
ꢀH(298) +
njCp(Pj) dT
298
ꢁ
+
njL(Pj) = 0
(7)
298−Tad(298)
where ꢀH(298) is the reaction enthalpy at 298 K, CP(Pj) and
L(Pj) the heat capacity and phase transformation enthalpy (if
the products go through a phase change, e.g. solid to liquid or
liquid to gas) of the products, respectively, nj the stoichiomet-
ric coefficient of the products, and Tad refers to the adiabatic
combustion temperature.
The calculated result of Eq. (7) is shown in Fig. 2. It can be
seen that the adiabatic temperature of 1800 K corresponds to the
Ni content of 72.20 wt.%. Furthermore, the Tad increases with
Ni content decreasing except the phase transition ranges (hori-
zontal line), which indicates that Ni serves as a diluent. Based
on the empirical criterion proposed by Merzhanov [19,20], the
SHS reaction can be self-sustaining only when Tad ≥ 1800 K.
Therefore, the Ni content range of Ni–Ti–B system is selected
from 30 to 70 wt.% in the present study.
Fig. 3. XRD patterns of the SHS reaction products synthesized by green pre-
forms with a molar ratio of B/Ti = 2.0 mixed with (a) 30, (b) 40, (c) 50, (d) 60
and (e) 70 wt.% Ni, respectively, compared with (f) the standard pattern for TiB2
[21].
tion products, though the Tad decreases obviously with the
increase of Ni content (Fig. 2).
(ii) The existence of transient phases indicates that the SHS
reactions under this experiment condition are incomplete,
which is mainly due to the short dwell time in the com-
bustion zone resulting from the high combustion velocity
of the SHS reactions, and, hence, there is not enough time
for achieving the equilibrium composition as designed. In
addition, the amounts of transient phases tend to increase
with Ni content increasing. The work on how to reduce the
amount of transient phases in the SHS products of Ni–T–B
system will be continued.
(iii) As shown in Fig. 3(f) as well as the dotted line, the peaks of
TiB2 in situ synthesized from the SHS reactions in Ni–Ti–B
systems make no measurable differences from the standard
peaks of TiB2 from the PDF Card (No. 35-0741) [21] under
this measurement accuracy of XRD (4◦min−1). Therefore,
the TiB2 synthesized is of the designed stoichiometric ratio
(B/Ti = 2.0). This is different from TiC1−x in situ synthe-
sized from M (M = Al [22], Cu [23], Fe [24], Ni [25],
Cr [26], Ti [27])–Ti–C systems, and the value of x can
be changed over a wide composition range, even with a
designed molar ratio of C/Ti = 1.0.
3.2. XRD analyses
Fig. 3(a–e) shows the XRD patterns of the SHS reaction
products synthesized by green preforms with a molar ratio of
B/Ti = 2.0 with (a) 30, (b) 40, (c) 50, (d) 60 and (e) 70 wt.%
Ni, respectively, compared with the standard pattern for TiB2
(Fig. 3(f)) [21]. It can be observed that the phase compositions
of the final products in the five systems are nearly the same,
consisting mainly of TiB2 and Ni as designed, as well as the
transient phases Ni4B3, Ni3B, NiB and Ni3Ti. The above XRD
result is consistent with the thermodynamic calculation (Fig. 1)
and it reveals several points as follows:
(i) Ni content within the chosen range from 30 to 70 wt.% has
little effect on the phase compositions of the SHS reac-

L. Huang et al. / Journal of Alloys and Compounds 457 (2008) 286–291
289
Fig. 4. Typical FESEM of the SHS reaction products synthesized by green preforms with a molar ratio of B/Ti = 2.0 mixed with (a) 30, (b) 40, (c) 50, (d) 60 and (e)
70 wt.% Ni, respectively, (the insets in (d) and (e) are the local magnification of the central frames of (d) and (e), respectively).
3.3. Microstructure
Ni–Ti–B systems. In addition, it is obvious to see that some
TiB2 particulates exhibit exaggerated growth in products of the
three systems, and the sizes of the TiB2 grains can even reach
10–15 ␮m, which is different from the result of the study of
Horlock et al. [3] that the size of TiB2 particles synthesized
from 35 wt.% Ni (Cr)–Ti–B system ranged from around 1 to
6 ␮m without the exaggerated growth phenomenon. Further-
more, the TiB2 particulates even present transgranular cracks in
the product of 30 wt.% Ni–Ti–B system (as shown by the arrows
in the top right corner in Fig. 4(a)). These phenomena can be
attributed to the strong trend of TiB2 to grow anisotropically at
Fig. 4(a–e) shows the typical FESEM of the SHS reaction
products synthesized by green preforms with a molar ratio of
B/Ti = 2.0 with (a) 30, (b) 40, (c) 50, (d) 60 and (e) 70 wt.%
Ni, respectively. It can be observed that large amounts of TiB2
particulates with the morphology of hexagonal-prism have been
synthesized as designed, which is accordant with the results of
the study of Horlock et al. [3] and our previous study [28,29].
The average sizes of normal TiB2 particulates are nearly the
same and about 4–6 ␮m in the products of 30, 40 and 50 wt.%

290
L. Huang et al. / Journal of Alloys and Compounds 457 (2008) 286–291
high-temperature. When the Ni content is relatively low, the con-
tents of Ti and B per unit volume are relatively high, and, thus,
much heat can be evolved from the SHS reactions of Ni–Ti–B
system, which creates high-temperature to accelerate the growth
of TiB2. Strong internal stresses have been generated during the
anisotropic growth of TiB2, which can even cause the formation
of transgranular cracks [4,5]. Besides, it can be seen that there
are small amounts of remaining liquids (bonding agents) on the
surface of or among TiB2 particulates, which will be discussed
in the following part.
When Ni content increases to 60 wt.%, the average size of
TiB2 particulates dramatically decreases to only about 1–2 ␮m,
and there are large amounts of remaining liquids surround-
ing TiB2 particulates, as shown in Fig. 4(d). Furthermore, the
exaggerated growth phenomenon of TiB2 particulates has disap-
peared. When Ni content increases to 70 wt.%, the average size
of TiB2 particulates decreases further. The insert in Fig. 4(e) is
the local magnification as schematically illustrated, from which
it can be seen that the average size of TiB2 is only about 0.6 ␮m
or less, and the prismatic morphology of TiB2 is still complete
and clear.
Fig. 5. Local magnifications of the frame in Fig. 1(a).
Ni–Ti–B system, from which it can be observed the clear growth
striation on the (0 0 0 1) crystal face of TiB2. Hence, the solution,
reaction and precipitation mechanism is further confirmed by the
presence of growth striation.
The decrease of the average size of TiB2 particulates is
mainly due to the decrease of combustion temperature as Ni
content increases (Fig. 2), because grain growth is an exponen-
tial function of the combustion temperature [30]. The decrease
of combustion temperature can be attributed to the following
two reasons: (i) the heat release from Ti–B reaction per unit vol-
ume of the reactant decreases when Ni content increases and (ii)
the more Ni content is, the more heat absorption of Ni melt is.
Besides, another important factor influencing the size of TiB2
grains is the decrease of Ti and B concentration per unit with Ni
content increasing, which reduces the probability of Ti and B
atoms to diffuse continuously onto the TiB2 grains synthesized.
In addition, the excess Ni phase also serves as a diluent and pre-
vents the diffusion of Ti and B atoms during the SHS reaction,
which also restrains the growth of TiB2.
4. Conclusions
(1) The main products of Ni–Ti–B system synthesized via SHS
reactions are TiB2 and Ni. In addition, the transient phases
Ni4B3, Ni3B, NiB and Ni3Ti are also synthesized, which
indicates the SHS reactions in Ni–Ti–B system are incom-
plete. And the change in the content of Ni within the chosen
range from 30 to 70 wt.% has little effect on the phase
compositions of the products.
(2) The average sizes of normal TiB2 particulates are nearly
the same and about 4–6 ␮m when Ni contents are 30, 40
and 50 wt.%. Furthermore, TiB2 particulates in the prod-
ucts of these three systems present exaggerated growth and
their sizes can even reach 10–15 ␮m. In addition, the size of
TiB2 particulates decreases dramatically to 1–2 ␮m when
Ni content increases to 60 wt.% while to 0.6 ␮m or less at
70 wt.% Ni.
(3) The addition of Ni facilitates to form more liquid phases that
are beneficial to TiB2 formation during SHS reaction pro-
cess. The formation mechanism of TiB2 in Ni–Ti–B system
can be characterized by the solution, reaction and precip-
itation processes. This can be further substantiated by the
presence of remaining liquids, the typical hexagonal-prism
morphology and the growth striation on (0 0 0 1) crystal face
of TiB2 particulates.
3.4. Solution, reaction and precipitation
In the present study, the addition of Ni to Ti–B system
is an aid to form liquids during the SHS reaction, because
the Ni–Ti and Ni–B liquids can be formed at lower temper-
ature compared with the higher melting points of Ni, Ti and
B. Then further dissolution of Ti and B is greatly accelerated
with temperature rising and, hence, the Ni–Ti–B liquids tend
to be formed. When the concentrations of Ti and B in the
liquids reach supersaturation, the nucleation of TiB2 becomes
thermodynamically feasible, and then TiB2 will be formed
and precipitated out from the liquids. Therefore, the formation
mechanism of TiB2 can be characterized by the solution, reac-
tion and precipitation processes, which can be evidenced by
the presence of remaining liquids and the typical hexagonal-
prism morphology of TiB2 (Fig. 4(a–e)). Horlock et al. [3] also
reached the similar conclusion that the formation of TiB2 was a
melting–solution–precipitation reaction in their study.
Acknowledgements
This work is supported by The National Natural Science
Foundation of China (no. 50531030) and The Ministry of Sci-
ence and Technology of the People’s Republic of China (nos.
2005CCA00300 and 2006AA03Z566) as well as The Project
985-Automotive Engineering of Jilin University.
Fig. 5 is the local magnification of Fig. 4(a), and it shows
the typical morphology of TiB2 in situ synthesized by 30 wt.%

L. Huang et al. / Journal of Alloys and Compounds 457 (2008) 286–291
291
References
[15] Y.J. Kwon, M. Kobashi, T. Choh, N. Kanetake, Scripta Mater. 20 (2004)
577–581.
[16] R.J. Li, Ceramic–Metal Composites, Metallurgic Industry Press, China,
1995.
[17] Y.J. Liang, Y.C. Chen, DataHandbookofMineralThermodynamics, North-
eastern Press, China, 1993.
[18] I. Barin, Thermochemical Data of Pure Substance, third ed., Copyright
1995 by WILEY-VCH Verlag Gmbh.
[1] Y. Taneoka, O. Odawara, Y. Kaieda, J. Am. Ceram. Soc. 72 (6) (1989)
1047–1049.
[2] L.L. Ye, Z.G. Liu, M.X. Quan, Z.Q. Hu, J. Appl. Phys. 80 (3) (1996)
1910–1912.
[3] A.J. Horlock, D.G. Mccartney, P.H. Shipway, J.V. Wood, Mater. Sci. Eng.
A336 (2002) 88–98.
[19] J.J. Moore, H.J. Feng, Prog. Mater. Sci. 39 (1995) 243–273.
[20] Z.A. Munir, U.A. Tamburini, Mater. Sci. Rep. 3 (1989) 277–365.
[21] Powder Diffraction File Card No. 35-0741 (CD ROM), International Center
for Diffraction Data (ICDD), Newtown Square, PA.
[22] P.J. Li, E.G. Kandalova, V.I. Nikitin, A.G. Makarenko, A.R. Luts, Y.F.
Zhang, Scripta Mater. 49 (2003) 699–703.
[23] N. Zarrinfar, P.H. Shipway, A.R. Kennedy, A. Saidi, Scripta Mater. 46
(2002) 121–126.
[24] A. Saidi, A. Chrysanthou, J.V. Wood, J.L.F. Kellie, Ceram. Int. 23 (1997)
185–189.
[25] L. Huang, H.Y. Wang, F. Qiu, Q.C. Jiang, Mater. Sci. Eng. A422 (2006)
309–315.
[26] W.N. Zhang, H.Y. Wang, S.Q. Yin, Q.C. Jiang, Mater. Lett. 61 (2007)
3075–3078.
[27] P. Wanjara, R.A.L. Drew, J. Root, S. Yue, ActaMater. 48(2000)1443–1450.
[28] H.Y. Wang, L. Huang, Q.C. Jiang, Mater. Sci. Eng. A407 (2005) 98–104.
[29] H.Y. Wang, Q.C. Jiang, B.X. Ma, Y. Wang, F. Zhao, Adv. Eng. Mater. 7
(2005) 58–63.
[4] M.A. Einarsrud, E. Hagen, G. Pettersen, T. Grande, J. Am. Ceram. Soc. 80
(12) (1997) 3013–3020.
[5] L.H. Li, H.E. Kim, E.S. Kang, J. Eur. Ceram. Soc. 22 (2002) 973–977.
[6] B. Basu, G.B. Raju, A.K. Suri, Int. Mater. Rev. 51 (2006) 352–374.
[7] W.M. Wang, Z.Y. Fu, H. Wang, R.Z. Yuan, J. Eur. Ceram. Soc. 22 (2002)
1045–1049.
[8] I. Gotman, M.J. Koczak, Mater. Sci. Eng. A187 (1994) 189–199.
[9] V.I. Nikitin, J.I.E. Wanqi, E.G. Kandalova, A.G. Makarenko, L. Yong,
Scripta Mater. 42 (2000) 561–566.
[10] Y. Wang, H.Y. Wang, B.X. Ma, K. Xiu, Q.C. Jiang, J. Alloys Compd. 422
(2006) 178–183.
[11] M.G. Barandika, J.M. Sanchez, T. Roojo, R. Cortes, F. Castro, Scripta
Mater. 39 (1998) 1395–1400.
[12] Y.M. Maksimov, A.G. Merzanov, A.T. Pak, L.G. Raskolenko, B.S. Braver-
man, Iav. Akad. Nauk SSSR. Met. 2 (1985) 219–223.
[13] C.C. Degnan, P.H. Shipway, Metall. Mater. Trans. A33 (2002) 2973–
2983.
[14] Q. Xu, X.H. Zhang, J.C. Han, X.D. He, V.L. Kvanin, Mater. Lett. 57 (2003)
4439–4444.
[30] Y. Choi, S.W. Rhee, J. Mater. Sci. 28 (1993) 6669–6675.