The effect of solution treatment under loading on the microstructure and phase transformation behavior of porous NiTi shape memory alloy fabricated by SHS

Article abstract of DOI:10.1016/j.jallcom.2008.07.023

Porous NiTi shape memory alloy (SMA) was fabricated by self-propagating high-temperature synthesis (SHS). With this study, a new solution treatment "solution treatment under loading" was applied to porous NiTi SMA fabricated by SHS to determine microstruc

Full text of DOI:10.1016/j.jallcom.2008.07.023

Journal of Alloys and Compounds 475 (2009) 378–382
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
The effect of solution treatment under loading on the microstructure and phase
transformation behavior of porous NiTi shape memory alloy fabricated by SHS
Mehmet Kayaa,∗, Nuri Orhana,1, Bulent Kurta,2, Tahir I. Khanb,3
a Department of Metal Education, Firat University, Elazıg˘ 23119, Turkey
b Department of Mechanical Engineering, University of Calgary, Calgary, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Porous NiTi shape memory alloy (SMA) was fabricated by self-propagating high-temperature synthe-
sis (SHS). With this study, a new solution treatment “solution treatment under loading” was applied to
porous NiTi SMA fabricated by SHS to determine microstructural improvement regarding single phase
NiTi. The effect of solution treatment under load on chemical composition, constituent phases and phase
transformation behaviors of the specimens was investigated and discussed. The chemical composition of
the specimens considerably changed with solution treatment under loading. Intermetallic phases such as
Ti2Ni and Ni4Ti3 disappeared, the density of B2(NiTi) phase increased and phase transformation temper-
atures sharply decreased. Porous single phase B2(NiTi) SMA with high chemical homogeneity could be
obtained by the load applied during solution treatment at 1050 ◦C.
Received 22 April 2008
Received in revised form 3 July 2008
Accepted 7 July 2008
Available online 19 August 2008
Keywords:
NiTi shape memory alloy
Porous materials
Self-propagating high-temperature
synthesis
© 2008 Elsevier B.V. All rights reserved.
Phase transformation behaviors
1. Introduction
due to the composition fluctuation in the specimen since the raw
powders are not mixed sufficiently and the particle size of the
Recently, the production of porous NiTi shape memory alloys
(SMAs) has been received considerable interest due to their extraor-
dinary mechanical characteristics similar to those of some natural
biomaterials for hard tissue implants, quite high damping capacity
and low relative density [1,2]. In addition, they have some unique
properties like shape memory effect, superelasticity, excellent cor-
rosion resistance, and good biocompatibility. Moreover, the porous
structure of the alloys would help tissue cell in growth, nutrition
exchange and medicament transportation [3,4].
So far, porous NiTi SMAs have been fabricated with powder
metallurgy (PM) processes such as combustion synthesis with self-
propagating wave [5–7], metal injection molding (MIP) [8], hot
isostatic pressing (HIP) [8], and spark plasma sintering (SPS) [9].
These processes can avoid the problems associated with casting,
like segregation or extensive grain growth and have the added
advantages of precise control of composition and easy realization
of complex part shapes. However, they can cause oxidation and ten-
dency to form other Ti–Ni phases, such as Ni3Ti, Ti2Ni and Ni4Ti3
reactants is not small enough [2,9,10]. In some cases even elemen-
tal Ni in varying proportion occurs, which is toxic in living tissue
[10].
None of the above PM processes has allowed obtaining fully
dense high purity single phase NiTi components [11]. The mechan-
ical properties of NiTi depend on its phase state at a certain
temperature. Fully austenitic NiTi material generally has suitable
properties for surgical implantation and superelasticity [12]. Ti2Ni,
Ni3Ti and Ni4Ti3 phases existing in SHS-synthesized porous NiTi
SMA may increase the brittleness of the products. Moreover, they
can lead to the cavitation corrosion and deteriorate the biocom-
patibility of porous NiTi SMAs in the physiological environments
[1,13].
Different procedures are required to remove undesired phases
formed during fabrication. The amount of Ni4Ti3 phase in porous
NiTi SMA decreases slightly after solution treatment. However,
Ti2Ni and Ni3Ti phases are extremely difficult to remove by subse-
quent thermal treatments since they are thermodynamically stable
[2,13,14]. Therefore, it is necessary to implement different proce-
dures for removing the undesired Ti2Ni and Ni3Ti phases in porous
NiTi SMAs.
∗
Corresponding author. Tel.: +90 424 2370000x4242; fax: +90 4242184674.
E-mail address: mehmetkaya@firat.edu.tr (M. Kaya).
Tel.: +90 424 2370000x4242.
Tel.: +90 424 2370000x4253.
Bertheville and Bidaux [11] studied an alternative powder
metallurgical process called VPCR (vapor phase calciothermic
reduction) for the fabrication of single phase NiTi shape memory
alloys. They reported that the XRD peaks of austenite B2(NiTi) and
1
2
3
Tel.: +1 403 220 5772.
0925-8388/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2008.07.023

M. Kaya et al. / Journal of Alloys and Compounds 475 (2009) 378–382
379
of CaO were visible when using the VPCR process. In another study
[14], they reported a thin calcium oxide film over the surface of
the compact which can easily be removed by polishing or leaching.
In addition, it was determined by them that the vacuum sintered
compact was predominantly B2(NiTi) phase and secondly Ti2Ni
and Ni4Ti3 at room temperature while the sintered compact under
reducing Ca vapor was the major stable B19ꢀ(NiTi) phase instead of
the B2(NiTi) phase.
It is well accepted that the shape memory effect (SME) and
superelasticity effect (SE) of NiTi SMAs depend critically on
the reversibility of the martensitic transformation [15]. Various
researchers [4,16–18] have studied the effects of porosity, aging and
solution heat treatment on the martensitic transformation behav-
ior of porous NiTi SMAs. However, the effects of solution treatment
under loading on characteristic of porous NiTi SMAs fabricated by
SHS have not been investigated in the literature.
Porous NiTi SMAs fabricated by SHS have more porosity than
porous NiTi SMAs fabricated with the other PM processes according
to literature [2,19]. So far, various ignitors such as tungsten coil [20],
ignition reagent composed of Ti and C powder [2], laser [21] or other
means have been employed for the fabrication of porous NiTi SMAs
by SHS.
In this study, high-voltage electric arc was used to ignite the
specimens in the fabrication of porous NiTi SMA by SHS. Also, a new
solution treatment, “solution treatment under loading” was applied
to porous NiTi SMA fabricated by SHS to determine microstructural
improvement regarding single phase NiTi. In addition, the phase
transformation behavior was also investigated.
Fig. 1. Optic micrograph of porous NiTi synthesized at 200 ◦C.
solution treatment under loading [23]. The pore characteristics and
porosity ratios of the specimens produced are suitable for hard tis-
sue implants because the ideal implant material should have the
porosity in the range of 30–90%, and the optimal pore size for bone
tissue ingress is 100–500 ␮m [24].
Fig. 2 shows the microstructure of the porous NiTi fabricated
by SHS. The EDS quantitative analysis result of atomic composition
was Ni–41.01 at.%Ti for the grain boundary pointed out as Ellipse1 in
Fig. 2, corresponding to Ni3Ti2 or Ni4Ti3 phase. The same composi-
tion was seen for the formations like mushrooms or water granules
in shape, too. The atomic composition of elliptic area showed as E0
in Fig. 2 was Ni–59.29 at.%Ti, which is possibly a pore. EDS anal-
ysis shows that the composition of platelike formations in Fig. 2
is Ni–43.81 at.%Ti, which can be identified as Ni4Ti3 or Ni-rich
B19ꢀ(NiTi) martensite phase by combining EDS and XRD results.
NiTi matrix is seen grey; Ni3Ti2 or Ni4Ti3 phases are seen as light
grey in Fig. 2, and NiTi2 phase in Fig. 3a is seen as cornered shapes
in dark grey.
Although the phases such as Ni4Ti3 and NiTi2 usually form in
porous NiTi alloys fabricated by SHS [2], the phases such as Ni3Ti,
pure Ni and pure Ti rarely occur [10]. Ni3Ti, pure Ni and pure Ti
were not observed in this study. The presence of Ni4Ti3 and Ni3Ti2
in the porous NiTi SMA synthesized by SHS is due to the overall
stoichiometry of the Ni-rich (50.5 at.%Ni) specimen. Moreover, the
presence of Ti2Ni phase, since Ti powder is more flammable than Ni
powder, could make the parent NiTi phase richer in Ni. If the mix-
ing is not homogeneous, the amount of undesired phases becomes
larger [2,9,10].
2. Experimental procedures
The raw materials used were Ni (99.8 wt.%) and Ti (99.5 wt.%) powders with an
average size of 44 ␮m (Alfa Aesar). Firstly, the powders of Ni and Ti with 50.5 at.%Ni
were blended in a rotating container for 24 h for a homogenous mixture, and then
the blended powder was pressed into cylindrical compacts of 10 mm in diameter and
15 mm in height using a hydraulic press at a cold compaction pressure of 100 MPa.
The green samples after compacting were preheated up to 200 ◦C with a heating rate
of 15 ◦C/min in a furnace under the protection of high purity argon gas, then were
subjected to electrical discharge pulse (14 kV and 30 mA) for about 2 s [22,23]. The
temperature of the green samples was increased in a short time at the beginning
of current application and ignition started. Once ignited, combustion wave self-
propagated along the axis of the specimen to the other end in a very short time, thus
porous NiTi SMA was synthesized. To investigate the effects of solution treatment
under loading on microstructure and phase transformation behavior, synthesized
specimens were separately solution treated under a load of 25 kg (3.2 MPa) and 50 kg
(6.41 MPa) at 1050 ◦C for 1 h in a furnace under the protection of high purity argon
gas. Finally, the specimens were quenched into water at room temperature.
To investigate the chemical composition, the surfaces of specimens were etched
by a mixture of 10% HF, 5% HNO3 in water [13]. An energy-dispersive X-ray spectrom-
eter (EDS) coupled with the scanning electron microscopy (SEM, LEO Evo-40VP) was
used to locally measure the chemical composition of the specimens. The phase con-
stituents were determined by X-ray diffraction (XRD, Rigako Rad-B D-Max 2000
XRD) analysis using CuK␣ radiation with 1.54046 Å. The phase transformation tem-
peratures and energies were measured by differential scanning calorimeter (DSC,
PerkinElmer Pyris 6). The samples, typically 14–17 mg in weight, were placed in alu-
minium pans. The measurements were carried out under nitrogen gas flow. In the
DSC thermal analysis, specimens were heated from −50 to 140 ◦C and kept isother-
mally for 2 min to establish thermal equilibrium, then cooled down to −50 ◦C, and
also kept isothermally for 2 min and then ramped back to 140 ◦C again. Finally, they
were cooled to −50 ◦C to finish the thermal cycle. Heating and cooling rates were
kept at 5 and 10 ◦C/min for the first and second thermal cycles respectively.
3. Results and discussions
Fig. 1 shows the general morphology of porous NiTi fabricated by
SHS which consists of the combustion channels and pores. The dis-
tribution of the pores is uniform and most of the pores are isolated
and rarely interconnected. The pores have fairly low size in ␮m
and various shapes. The specimens fabricated by SHS has a general
porosity of 55.5 vol.% and the porosity decreases to 40.6 vol.% with
Fig. 2. SEM micrograph and EDX regions of the porous NiTi SMA fabricated by SHS.

380
M. Kaya et al. / Journal of Alloys and Compounds 475 (2009) 378–382
Fig. 4. XRD patterns of porous NiTi SMAs, (a) the specimen synthesized, (b) the
specimen applied load of 3.2 MPa at 1050 ◦C for 1 h and (c) the specimen applied
load of 6.41 MPa at 1050 ◦C for 1 h.
load on the specimen was raised up to 6.41 MPa during solution
treatment, Ni4Ti3, Ti2Ni and B19ꢀ(NiTi) phases disappeared and the
microstructure transformed to B2(NiTi) phase (Fig. 4 (curve-c)).
Zhu et al. [26] reported free nickel beside TiNi and intermetallics
such as Ti2Ni and Ni3Ti from XRD results when sintering time was
short. They also reported that free Ni disappeared, and the speci-
men was composed mainly of NiTi phase a small amount of other
intermetallics when the sintering time exceeded 8 h.
Recently, Biswas [27] has proposed that the second phase (Ni3Ti,
NiTi2) can be eliminated completely by a post-reaction heat treat-
ment. While unreacted Ni disappeared the Ni14 Ti11 and Ni3Ti are
found distributed within NiTi after post-reaction heat treatment at
1050 ◦C, the specimen post-heat treated at 1150 ◦C is clearly single
phase. This is because of the eutectic temperature of NiTi2 and N3Ti
is 984 and 1118 ◦C, respectively. The higher post-treatment temper-
ature of 1150 ◦C induced the dissolution of undesired phases [28].
During the formation of NiTi, NiTi2 and Ni3Ti phases exothermic
reactions occur as given below [26]:
Fig. 3. SEM micrographs and EDX regions of the specimens. (a) Solution treatment
under 3.2 MPa and (b) solution treatment under 6.41 MPa.
Ni + Ti → NiTi + 67 kJ/mol
Ni + Ti → NiTi2 + 83 kJ/mol
Ni + Ti → Ni3Ti + 140 kJ/mol
(1)
(2)
(3)
Fig. 3 shows the microstructures of solution treated specimens
under different loads at 1050 ◦C for 1 h. where the microstructure
is predominantly NiTi phase as desired. The result of the atomic
composition from the EDS quantitative analysis in Fig. 3a was
Ni–64.19 at.%Ti for the cornered dark grey area (Ellipse0, E0), pos-
sibly corresponding to NiTi2. Ni4Ti3 and Ni3Ti2 phases were not
observed in Fig. 3. The load applied accelerated the dissolution
of these phases by increasing local stress and thus diffusion [25].
There is some amount of NiTi2 phase in the solution treated spec-
imen under 3.2 MPa (Fig. 3a) but it disappeared completely after
the application of a load of 6.41 MPa during the solution treatment
(Fig. 3b). The EDS quantitative analyses results of atomic composi-
tion were Ni–46.41 at.%Ti and Ni–46.23 at.%Ti for the areas pointed
out as Ellipse0 and Ellipse1 in Fig. 3b respectively, corresponding
to NiTi phase richer in Ni.
Fig. 4 shows XRD patterns of the synthesized (curve-a) and solu-
tion treated specimens under loading. It is seen that the desired
products such as B2(NiTi) and B19ꢀ(NiTi) are the predominant
phases in the Fig. 4 (curve-a). In addition, the SHS process resulted
in the formation of several second phases such as Ti3Ni4 and Ti2Ni.
Chu et al. [13] also determined the same phases. In addition, they
reported that the amount of metastable Ni4Ti3 phase decreased
sharply after solution heat treatment at 1050 ◦C for 4 h, and Ti2Ni
phase could not be removed by solution treatment [13]. The results
in our study where load and solution treatment were implemented
simultaneously showed that when the load on the specimen was
3.2 MPa during solution treatment, it can be seen in Fig. 4 (curve-b)
that the amount of B2(NiTi) phase increased, B19ꢀ(NiTi) decreased,
Ni4Ti3 was not observed but still there was little NiTi2. When the
In the binary Ni–Ti phase diagram, NiTi, NiTi2 and Ni3Ti phases
are stable phases. Moreover, according to the amount of energy
exerted during the formation of phases, Ni3Ti and NiTi2 phases are
more thermodynamically favoured than NiTi. Consequently, it is
difficult to completely remove Ni3Ti and NiTi2 phases from the syn-
thesized sample only by altering heat treatment [26]. In our study,
a single phase porous B2(NiTi) was obtained by solution treatment
under loading at 1050 ◦C. This situation is harmonious with EDS
and XRD results.
Fig. 5 shows the result of DSC measurement for the non-solution
treated specimen. The broad peak in the heating curve corresponds
to the martensite (monoclinic) to austenite (cubic) transformation,
and the broad peak on the cooling curve corresponds to the trans-
formations from austenite to martensite. There are two peaks in the
first heating curve corresponding to B19ꢀ → R (intermediate trigo-
nal martensite) → B2 transformation. The first peak disappeared in
the second heating curve. A similar transformation, B19ꢀ → R → B2
transformation due to the formation of Ni4Ti3 after aged at 400
and 450 ◦C and B19ꢀ → B2 transformation after aged at 475, 500
and 550 ◦C, was determined by Chu et al. [18]. It is thought that the
incoherent stress fields in the structure disappeared after the first
heating–cooling cycle, thus one-step (B19ꢀ → B2) transformation
occured.
A one-step transformation from the high-temperature parent
phase B2 to the low-temperature monoclinic martensite phase B19ꢀ
in the cooling curves is seen in Fig. 5. For equiatomic dense NiTi

M. Kaya et al. / Journal of Alloys and Compounds 475 (2009) 378–382
381
Table 1
Measured DSC transformation temperatures and associated energies for porous NiTi SMA synthesized.
Heating rate
(◦C/min)
As (◦C)
Ap (◦C)
Af (◦C)
Ms (◦C)
Mp (◦C)
Mf (◦C)
ꢀHM → A
(J/g) heating
ꢀHA → M
(J/g) cooling
ꢀSM → A
(×10−3 J/g K)
ꢀSA → M
(×10−3 J/g K)
5
10
91.86
85.81
99.63
96.11
107.26
103.90
67.64
67.13
65.10
62.54
60.58
53.97
−0.744
−0.985
0.782
1.053
−2.064
−2.747
2.169
2.937
alloys, the similar martensitic transformation usually is seen in the
cooling curve. Two-steps or multi-stage martensitic transformation
(MST) can also occur in dense NiTi alloys after certain thermal or
mechanical treatments, which can be described as B2 → R, R → B19ꢀ,
or B2 → B19ꢀ [15].
Therefore, we did not attempt more than two cycles in this study.
The transformation temperatures of the specimens solution treated
under loading could not be determined in the working range of DSC
device since they are lower than −50 ◦C.
Prymak et al. [29] determined that Ap and Mp (peak tempera-
tures in heating and cooling curves, respectively) for porous NiTi
fabricated by SHS were 62.5 and 22.8 ◦C, respectively, which are
lower than those of our study, but the width of the peaks was low
(nearly 0.05 m W/mg). That is, the shape memory effect of speci-
mens fabricated by them was little. In addition, it was determined
by Jiang and Rong [5] that As and Af points for austenitic transfor-
mation were 92 and 103 ◦C, respectively, and also Mp was 67 ◦C for
porous NiTi fabricated by SHS. The transformation temperatures
determined by them are closer to the transformation temperatures
determined in our study.
After solution treatment under loading no evident peak was
seen in the temperature range (−50 and 140 ◦C) in DSC analysis.
This situation shows that the parent NiTi phase becomes richer
in Ni due to the disappearance of Ni4Ti3 with solution treatment
under loading decreased the transformation temperatures. If Ni
ratio decreases in parent phase due to the formation of Ni-rich
Ni4Ti3 phase with aging at lower temperature, the transformation
temperature again increases.
Transformation temperatures and associated energies mea-
sured by DSC for porous NiTi SMA synthesized are given in Table 1.
The transformation temperatures and transformation enthalpies
were evaluated with DSC. Transformation entropies were cal-
culated with Eqs. (4) and (5) by taking into consideration the
equilibrium condition in a system at constant temperature and
pressure [28].
Barrabes et al. [30] fabricated porous NiTi by SHS and deter-
mined that As and Af points for austenitic transformation with
a broad peak (30 ◦C) were 73 and 103 ◦C, respectively and the
enthalpy of austenitic transformation was 2.28 J/g for the untreated
NiTi. At the same time, they determined that Ms and Mf for
martensitic transformation were 69 and 49.6 ◦C, respectively, and
transformation enthalpy was 2.26 J/g. The transformation tem-
peratures determined by them are closer to the transformation
temperatures in our study and it is seen that the transforma-
tion hysteresis (between difference Ap and Mp) and enthalpies
explained by them are higher.
ꢀG = ꢁH − T0 ꢁS = 0 ⇒ ꢁH = T0 ꢁS
Ms + Af
(4)
(5)
T0
=
2
ꢀG, ꢀH, ꢀS and T0 are Gibbs energy change, enthalpy change,
entropy change and equilibrium temperature, respectively. As and
Af are austenite start and finish, Ms and Mf are martensite start
and finish temperatures, respectively. The transformation temper-
atures of porous NiTi fabricated by SHS are higher for the implants.
The phase transformation temperatures decrease with the second
thermal cycling, but the enthalpy and entropy of the transformation
increase in absolute value. The changes in enthalpy and entropy for
B19ꢀ → B2 transformation are lower than those for B2 → B19ꢀ trans-
formation. This result reveals that the heat required for B2 → B19ꢀ
transformation is greater than that of B19ꢀ → B2 transformation.
Jiang and Rong [5] determined that the transformation temper-
atures decreased very little with thermal cycling up to 10 times.
4. Conclusions
1. Porus NiTi SMAs free from pure Ni, pure Ti and Ni3Ti phases were
fabricated by SHS. The distribution of the pores was uniform and
most of the pores were isolated and rarely interconnected.
2. In case of loading during the solution treatment, undesired
intermetallics disappear nearly completely. Porous single phase
B2(NiTi) SMA with high chemical homogeneity was obtained by
the load applied during solution treatment, but still more studies
are needed to determine the effects of the load applied during
solution treatment on dissolution of intermetallics.
3. The phase transformation temperature of synthesized specimen
is higher for implant material. The phase transformation tem-
peratures of specimen solution treated under loading decrease
due to richer parent phase in Ni.
Acknowledgement
This work was supported by doctoral program in Firat University,
Turkey (Project No. 1043).
References
[1] P. Sevilla, C. Aparicio, J.A. Planell, F.J. Gil, J. Alloys Compd. 439 (2007) 67–73.
[2] C.L. Chu, C.Y. Chung, P.H. Lin, S.D. Wang, J. Mater. Process. Technol. 169 (2005)
103–107.
[3] F.T. Cheng, K.H. Lo, H.C. Man, J. Alloys Compd. 437 (2007) 322–328.
[4] F.T. Cheng, P. Shi, G.K.H. Pang, M.H. Wong, H.C. Man, J. Alloys Compd. 438 (2007)
238–242.
[5] H.C. Jiang, L.J. Rong, Mater. Sci. Eng. A 438–440 (2006) 883–886.
[6] M. Whitney, S.F. Corbin, R.B. Gorbet, Acta Mater. 56 (2008) 559–570.
[7] D.E. Burkes, J.J. Moore, J. Alloys Compd. 430 (2007) 274–281.
Fig. 5. DSC curves of porous NiTi SMA synthesized at 200 ◦C after green pressure of
100 MPa.

382
M. Kaya et al. / Journal of Alloys and Compounds 475 (2009) 378–382
[8] L. Krone, E. Schüller, M. Bram, O. Hamed, H.P. Buchkremer, D. Stöver, Mater. Sci.
[20] B.Y. Li, L.J. Rong, Y.Y. Li, V.E. Gjunter, Acta Mater. 48 (2000) 3895–3904.
[21] C. Zanotti, P. Giuliani, A. Terrosu, S. Gennari, F. Maglia, Intermetallics 15 (2007)
404–412.
Eng. A 378 (2004) 185–190.
[9] C. Shearwood, Y.Q. Fu, L. Yu, K.A. Khor, Scr. Mater. 52 (2005) 455–460.
[10] B.Y. Tay, C.W. Goh, M.S. Yong, A.M. Soutar, Q. Li, M.K. Ho, M.H. Myint, Y.W. Gu,
C.S. Lim, SIMTech Technical Reports 7 (2006) 21–25.
[22] M. Kaya, N. Orhan, B. Kurt, Dog˘u Anadolu Bölgesi Aras¸ tırmaları 6 (2008)
154–158.
[11] B. Bertheville, J.E. Bidaux, J. Alloys Compd. 387 (2005) 211–216.
[12] J. Ryhanen, Biocompatibility evaluation of nickel titanium shape memory metal
alloy, Ph.D. Thesis, Oulu University, Oulu, 1999.
[23] M. Kaya, The investigation of the metallurgical and mechanical characteristics
of NiTi shape memory alloys produced with powder metallurgy, Ph.D. Thesis,
Firat University, Elazig, Turkey, 2008.
[13] C.L. Chu, J.C.Y. Chung, P.K. Chu, Trans. Nonferrous Met. Soc. China 16 (2006)
49–53.
[14] B. Bertheville, J.E. Bidaux, Scr. Mater. 52 (2005) 507–512.
[15] S.L. Wu, X.M. Liu, P.K. Chu, C.Y. Chung, C.L. Chu, K.W.K. Yeung, J. Alloys Compd.
449 (2008) 139–143.
[24] C.Y. Chung, C.L. Chu, S.D. Wang, Mater. Lett. 58 (2004) 1683–1686.
[25] N. Özdemir, M. Aksoy, N. Orhan, J. Mater. Process. Technol. 141 (2003) 228–233.
[26] S.L. Zhu, X.J. Yang, D.H. Fu, L.Y. Zhang, C.Y. Li, Z.D. Cui, Mater. Sci. Eng. A 408
(2005) 264–268.
[27] A. Biswas, Acta Mater. 53 (2005) 1415–1425.
[16] B.Y. Li, L.J. Rong, Y.Y. Li, V.E. Gjunter, Scr. Mater. 44 (2001) 823–827.
[17] S.L. Zhu, X.J. Yang, F. Hu, S.H. Deng, Z.D. Cui, Mater. Lett. 58 (2004) 2369–2373.
[18] C.L. Chu, C.Y. Chung, P.H. Lin, Mater. Sci. Eng. A 392 (2005) 106–111.
[19] S. Wu, C.Y. Chung, X. Liu, P.K. Chu, J.P.Y. Ho, C.L. Chu, Y.L. Chan, K.W.K. Yeung,
W.W. Lu, K.M.C. Cheung, K.D.K. Luk, Acta Mater. 55 (2007) 3437–3451.
[28] X.T. Zu, C.F. Zhang, S. Zhu, Y. Huo, Z.G. Wang, L.M. Wang, Mater. Lett. 57 (2003)
2099–2103.
[29] O. Prymaka, D. Bogdanskib, M. Kollerb, S.A. Esenweinb, G. Muhrb, F. Beckmannc,
T. Donathc, M. Assadd, M. Epple, Biomaterials 26 (2005) 5801–5807.
[30] M. Barrabes, P. Sevilla, J.A. Planell, F.J. Gil, Mater. Sci. Eng. C 28 (2008) 23–27.