Structure and thermal stability of ball milled Ti-Al-H powders

Article abstract of DOI:10.1016/j.jallcom.2004.07.023

Powders with different compositions of the binary Ti-Al system were prepared by indirect reactive ball milling (IRBM). During milling, nanocrystalline powders of titanium and titanium-aluminium hydrides were obtained using methanol as a reactive control agent. The formation of these hydride compounds due to the reaction Ti-Al-methanol improved grain size reduction of the powders. Light metal hydrides were detected by X-ray diffraction (XRD) and thermogravimeter and differential thermal analysis (TG-DTA) techniques even after very short milling times (4 h). The use of methanol as a hydrogen source during ball milling seems to be an inexpensive and easy way to prepare nanocrystalline and amorphous ductile metal based alloys in considerably shorter milling times and without an atmosphere of hydrogen gas.

Full text of DOI:10.1016/j.jallcom.2004.07.023

Journal of Alloys and Compounds 388 (2005) 266–273
Structure and thermal stability of ball milled Ti–Al–H powders
a
b
b
´
´
´
´
´
J. Morales-Hernandez , J. Velazquez-Salazar , L. Garcıa-Gonzalez ,
c,∗
d
c
´
˜
˜
F.J. Espinoza-Beltran , J.D.O. Barceinas-Sanchez , J. Munoz-Saldana
a
Programa de Posgrado en Ingenier´ıa, Facultad de Ingenier´ıa, Universidad Auto´noma de Quere´taro,
Centro Universitario Cerro de las Campanas, Col. Centro, C.P. 76000, Quere´taro, Qro., Me´xico
Programa de Posgrado en Materiales del Centro de Investigacio´n y de Estudios Avanzados del IPN, Unidad Quere´taro,
Libramiento Norponiente 2000, Fracc. Real de Juriquilla, C.P. 76230, Quere´taro, Qro., Me´xico
b
c
Centro de Investigacio´n y de Estudios Avanzados del IPN, Unidad Quere´taro, Libramiento Norponiente 2000,
Fracc. Real de Juriquilla, C.P. 76230, Quere´taro, Qro., Me´xico
Centro de Tecnolog´ıa Avanzada del Estado de Quere´taro, Calz. del Retablo 150 Col. FOVISSSTE,
d
C.P. 76150, Quere´taro, Qro., Me´xico
Received 22 June 2004; received in revised form 26 July 2004; accepted 28 July 2004
Abstract
Powders with different compositions of the binary Ti–Al system were prepared by indirect reactive ball milling (IRBM). During milling,
nanocrystalline powders of titanium and titanium–aluminium hydrides were obtained using methanol as a reactive control agent. The formation
of these hydride compounds due to the reaction Ti–Al–methanol improved grain size reduction of the powders. Light metal hydrides were
detected by X-ray diffraction (XRD) and thermogravimeter and differential thermal analysis (TG–DTA) techniques even after very short
milling times (4 h). The use of methanol as a hydrogen source during ball milling seems to be an inexpensive and easy way to prepare
nanocrystalline and amorphous ductile metal based alloys in considerably shorter milling times and without an atmosphere of hydrogen gas.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Ti–Al alloys; High-energy ball milling; X-ray diffraction; Thermal analysis; Intermetallic compounds
1. Introduction
argon atmosphere for times as long as 1000 h to reach the
nanometric particle size and amorphous state, have been re-
ported [5,6]. The agglomeration of the elemental powders
can be significantly reduced by introducing of TiH2 pow-
ders as a process control agent (PCA) in the MA process,
which reduces the milling time (in one order of magnitude)
and increases milling efficiency [2,7]. The effective particle
size reduction using a hydride phase has been proven for
Zr–Ni [8] and Ti–Al [9] compounds in a hydrogen atmo-
sphere, which leads to the formation of light metal hydrides
and also prevent the oxidation. The mechanism of milling is
enhanced due to the brittleness behavior transferred to the
powders by the introduced or formed metal hydrides. Further
on, the absence of a PCA in the milling leads to a cold welding
process, which inhibits the milling of the powders [2]. The
removing of this PCA agent after milling does not represent
a problem considering its low thermal stability (unstable at
temperatures higher than 250 ◦C). However, the management
Ti–Al alloys have been of much interest in recent years
as lightweight and as structural materials for elevated tem-
perature strength, oxidation and corrosion resistance applica-
tions. Recently, a number of non-equilibrium and amorphous
phasesproducedbymechanicalalloying(MA)ofamixtureof
pure elemental powders in the solid state such as Mg–Ni have
been reported [1,2]. The MA method is particularly promis-
ing for the manufacture of homogeneous metastable alloys
such as amorphous alloys and supersaturated solid solutions
[1] enhancing in most of the cases mechanic, catalytic and
thermodynamic properties. Several compositions of the bi-
nary Ti–Al system have been already produced by MA [3,4].
For instance, Ti–Al alloys with high Al content milled in
∗
Corresponding author. Tel.: +52-442-4414913; fax: +52-442-4414938.
E-mail address: fespinoza@qro.cinvestav.mx (F.J. Espinoza-Beltra´n).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.07.023

J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
267
of either metal hydrides or hydrogen atmospheres, to in situ
form these hydrides, both as a PCA in the milling process
represent a challenge due to its high reactivity when exposed
to the atmosphere. This, in fact increases the necessity of use
of glove boxes with controlled atmospheres even during the
preparation of powders for milling.
We used a novel method to synthesize titanium–aluminum
hydride by a modification of the reactive ball milling process
(RBM) [10]. Our modification, which will be called hereafter
indirect reactive ball milling (IRBM), consists in the replace-
ment of the hydrogen atmosphere with a small amount of a
reactive liquid (methanol) during milling, keeping argon as
atmosphere. The methanol will be called the reactive process
control agent (RPCA) in this procedure. A systematic study
of the Ti–Al system prepared by this process of IRBM has
not been reported yet.
The aim of this paper is to study the structural evolution
and thermal stability of nanocrystalline and amorphous Ti–Al
alloy powders obtained by IRBM process, where aluminum
content was varied from 5 to 90 wt.%. In this work, results on
the monitoring of formation of (Ti, Al)H brittle phases in the
different compositions of Ti–Al powders are presented. The
behavior of the particle sizes reduction and the formation of
amorphous phases at different milling times are analyzed.
410, with diameters of 3.3 mm and 8.0 mm, were introduced
in a hardened steel vial, evacuated and sealed in a glove box,
which was filled with 99.9% of pure argon. The total pow-
der weight was 8.0 g and the ball to powder ratio was 5:1.
After milling procedure, the samples were kept in an argon
gas atmosphere for periods longer than 12 h to reduce their
reactivity with the ambient, and to use them for the several
analyses, especially for mixtures with higher aluminum con-
tent.
The structure and phases content of the milled pow-
der was examined by X-ray diffraction (XRD) (Rigaku-
DMAX/1200) with Ni-filtered and Cu K␣ radiation and the
powder morphology by scanning electron microscopy (SEM)
(XL30 ESEM Philips) operated at 20 kV. Thermogravimeter
and differential thermal analysis (TG–DTA) (Mettler Toledo
TGA/SDTA851e) were performed under nitrogen flow with a
heating rate of 10 ◦C/min. The microstructure of the powders
also was investigated by a transmission electron microscope
(TEM) (JEOL-JEM-2000FX II).
3. Results and discussion
3.1. Powder preparation
Fig. 1 shows a series of XRD patterns indicating the evo-
lution of phases as a function of milling time (0–16 h) of the
sample Ti–8.5 at.% Al (Ti–5 wt.% Al). Before milling, the
XRD pattern shows characteristic peaks from the starting el-
ements. After 4 h of milling, aluminum atoms diffuse into
Ti lattice, which can be observed in the XRD pattern by the
vanishing of Al peaks. On the other hand, we observed an
increasing of the Ti peaks intensity. This has been already
reported as the formation of a solid solution of Al in Ti with
2. Experimental details
Elemental powders of Al and Ti (99.8 and 99.5% pure,
respectively and average particle size 50 ␮m) were mixed
to prepare several Ti–Al compositions (Ti100−xAlx, x = 8.5,
16.5, 30.7, 43.2, 72.7 y 94.2 at.% Al). IRBM was carried out
for 4, 8, 12, 16, 24, 30 and 36 h in a SPEX 8000 Mixer/Mill.
Methanol was added as RPCA. Stainless steel balls AISI-
Fig. 1. XRD patterns of Ti–8.5 at.% Al (Ti–05 wt.% Al) powders ball milled by IRBM for different times until amorphous state is reached.

268
J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
a structure similar to that of hex-Ti [11]. The solid solution
Ti(Al) shows an asymmetric broadening with a light shifting
of the XRD peaks as a function of milling time compared
to the original titanium positions. The TiH1.924 phase for-
mation is already observed after 4 h of milling as a more
intense peak at about 35◦ (2θ position), very close to one of
the reference lines for Ti(Al) (denoted as Ti). The formation
of this metal hydride must be the result of the reaction be-
tween the hydrogen arising from the RPCA with Ti(Al) in the
current milling condition. This mechanochemical process is
milling time dependent. As the milling time increases, this
reaction leads to a gradual increasing of the relative intensity
of the peak related to TiH1.924. After 8 h of milling the main
diffraction peak of the titanium hydride reaches its highest in-
tensity, while a broadening and an intensity reduction of the
XRD peaks of Ti(Al) was observed. This intensity reduction
is correlated with the particle size reduction and formation
of an amorphous phase. These results coincide with those re-
ported by Itsukaichi et al. [12] for the same alloy system. It
is well known that nitrogen contamination takes place after
the milling procedure, when the Ti–Al powder is in an amor-
phous state [13] and it is exposed to the room atmosphere.
Contamination with nitrogen was here observed for samples
without or low aluminum content as superimposed broad and
low intensity peaks of TiN with TiH1.924. This contamination
does not represent a problem and can be easily avoided by
increasing either the aluminum content or the resting time
in argon atmosphere of the samples. On the other hand, as
mentioned before, the presence of a brittle component such
as TiH1.924, has a great contribution in the reduction of parti-
cle size and also for producing amorphous phases [8]. After
16 h of milling, a single broad peak of an amorphous phase is
observed with the maximum diffraction about 2θ ≈ 42◦, and
a remnant signal due to TiH1.924, is still observed (Fig. 1). A
similar behavior was observed for the samples with composi-
tions 16.5, 30.7, and 43.2 at.% Al (not shown here). Thus, the
sequence of phase transformations observed in Ti100−xAlx
milled samples at the compositions x ≤ 43.2 at.% Al can be
represented as follows:
Ti + Al + RPCA → Ti(Al)solid sol. + TiH1.924
→ (Ti, Al)Hamorphous + TiN
(1)
Burgio et al. [14] reported the formation of an amorphous
phase after long time of MA (longer than 100 h) for similar
Ti–Al compositions, without PCA additions.
Fig. 2 shows the diffraction patterns for a sample with
Al content of 72.7 at.% (Ti–60 wt.% Al) for the milling time
values of 4, 12, 24 and 30 h. After 4 h of milling, diffraction
peaks due to hex-Ti are observed as small and broad peaks,
some of them coinciding with peaks of TiH1.924. This clearly
indicates that one part of the Ti content reacts with hydro-
gen to form titanium hydride. The main diffraction peak of
Al decreases with milling time, which is due to an alloying
with TiH1.924 to form a (Ti, Al)H ternary alloy. Hashi et al.
[15] reported (Ti, Al)H diffraction patterns similar to that of
TiH1.924 with a peak shifting depending on the Al content. We
additionally measured the interplanar distance of the sample
Ti–72.7 at.% Al. A progressive shifting of the interplanar dis-
˚
tance of (1 1 1) TiH1.924 peak with milling time (from 2.569 A
˚
for 4 h to 2.575 A for 30 h) and simultaneous decreasing of
(1 1 1) Al peak intensity were observed. Both behaviors are
due to the formation of (Ti, Al)H alloy. After 30 h of milling
there is no evidence of grain size reduction. This is due to
the high aluminum content in the sample, which inhibits the
fragility effect of the hydride on the powder. We observed
a similar behavior for samples with higher aluminum con-
tent (as high as 94.1 at.% Al), where the diffraction peaks
Fig. 2. XRD patterns of Ti–72.7 at.% Al (Ti–60 wt.% Al) powders ball milled by IRBM for different times until amorphous state is reached.

J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
269
Fig. 3. Variation in grain size of Ti–8.5 at.% Al (Ti–05 wt.% Al) powders as a function of milling time in IRBM.
of TiH1.924 phase show lower intensity. Thus, in a system
with these aluminum contents, larger time periods of milling
are required for an effective grain size reduction. We con-
firmed the existence of a compromise between brittleness of
hydrides and welding of aluminum, both processes showing
a strong dependence on the titanium and aluminum contents.
The sequence of phase transformations for Ti100−xAlx sam-
ples with compositions x > 43.2 at.% Al, and for long milling
time (which is longer when x increases), can be represented
as follows:
The average grain size of the IRBM milled powders was
determined from XRD data by applying a profile analysis
to all diffraction peaks (Ti, Al, TiH1.924, and TiN) in ac-
cord with the Warren–Averbach method [16]. Results as a
function of milling time for Ti–8.5 at.% Al (Ti–5 wt.% Al)
and Ti–72.7 at.% Al (Ti–60 wt.% Al) powders are shown
in Figs. 3 and 4, respectively. The advantage of using the
Warren–Averbach analysis to obtain average grain size is that
it proves more consistent grain size values, even for samples
with XRD patterns with superimposed peaks. Nanometric
grain sizes (less than 15 nm) were already attained even af-
ter 4 h ball milling, regardless of the Al content by using the
Ti + Al + RPCA → Al(Ti)solid sol. + (Ti, Al)Hamorphous (2)
Fig. 4. Variation in grain size of Ti–72.7 at.% Al (Ti–60 wt.% Al) powders as a function of milling time in IRBM.

270
J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
IRBM process. The grain size was reduced to about 7 nm after
a milling time of 8 h and to about 5 nm after 12 h and approx.
3 nmafter16 h, reachinganamorphousstateforpowderswith
small amount of Al (Ti–5 wt.% Al). Again, it is important to
remark the vanishing of the Al peaks from the XRD data af-
ter 8 h, which is indicative of the dissolution of Al into the Ti
lattice leading to the formation of a metastable Ti(Al) solid
solution before the development of the metal hydride. The
measurement of grain size particularly from the TiN peaks
revealed both nanocrystallites with size of less than 3 nm and
powders in an amorphous state.
For compositions with high Al content (Ti–60 wt.% Al),
an average grain size of less than 10 nm was measured for
the Al(Ti) peaks after a milling time of 12 h and thereafter no
further reduction was measured. For the same powders, the
Ti(Al) and TiH1.924 showed constant reduction in grain size
as a function of milling time up to reach sizes of less than
5 nm. In general, as the Al content increases longer milling
time is needed to reach particle sizes of less than 5 nm or
amorphous powders. Using the IRBM process we were able
to produce powders with these characteristics of the system
Ti–Al by milling times as short as 16 h for alloys with low
and middle Al contents and 36 h for alloys with high Al con-
tents. As mentioned before, the main effect in the mechanism
of reduction in particle size comes from the formation of a
titanium hydride during milling.
Results on ball milling of the Ti–Al system were reported
by different authors. For instance, Moon and Lee [17] and
Moon et al. [18] prepared Al–Ti alloys with high Al-contents
by conventional RBM (using hydrogen atmosphere). These
authors report results of grain size as a function of the milling
time between 10 and 15 nm for milling times as long as 100 h.
No clear evidence on the formation of the titanium hydrides
are observed from the XRD patterns, which is possibly the
reason for the long milling times. In this same way, Takasaki
and Fruya [19] reported results about the Ti–Al system milled
in a planetary ball mill, using hydrogen atmosphere for times
of about 30 h to obtain amorphous structure. In that work, the
authors did not observe the formation of the hydride phase.
Further on, Zhang et al. [20] studied the thermal stability
Fig. 5. TEM micrographs (bright and dark field images) and SAD pattern of: (a) Ti–5 wt.% Al and, (b) Ti–30 wt.% Al ball milled for 16 h by IRBM, both
showing nanoparticles of the Ti–Al alloy.

J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
271
Fig. 6. DTA and TGA curves of Ti–72.7 at.% Al and Ti–94.1 at.% Al powders ball milled by IRBM for 30 and 36 h of milling, respectively.
of mechanically alloyed Ti–Al powders with 75 at.% of Al in
Ar-atmosphere, in order to prepare Al3Ti. After milling times
of 40 h a nanocrystalline alloy of Al(Ti) was prepared, which
was thereafter heat treated by 800 and 900 ◦C to successfully
Fig. 5 shows TEM micrographs (bright and dark field im-
ages) and the correspondent selected area diffraction (SAD)
patterns of Ti–5 wt.% Al and Ti–30 wt.% Al powders, which
were ball milled for 16 h by IRBM both showing nanopar-
ticles of the Ti–Al alloy. The grain size observed from the
dark field image was less than 20 nm, which confirms our
results from the Warren–Averbach analysis of the XRD pat-
terns. TEM results reported elsewhere by Moon and Lee
[17] showed the tendency of multimodal particle distribu-
tions from the multiple components of the milled powder.
For instance, Moon and Lee [17] observed differences in
pure Al powders SAD pattern showed that the area con-
´
prepare D022-Al3Ti. Calderon et al. [21] used MA experi-
ments to produce nanocrystalline powders of Ti–Al–X (X =
Cr, Fe, Mn) intermetallic alloys getting particle sizes in the
range of 120–280 nm milling as long as 700 h in high energy
horizontal ball mills. We conclude that the proposed IRBM
process offers a higher efficiency in powder size reduction
and high energy MA compared to the RBM or conventional
high energy MA process.
Fig. 7. XRD patterns of Ti–72.7 at.% Al (Ti–60 wt.% Al) powders ball milled by IRBM for 30 h of milling and heated to various temperatures.

272
J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
sisted of TiH1.924, Ti(Al), and in some cases Al(Ti), TiN
or Al3Ti.
This sample was annealed by 1 h in a continuous argon flux at
the temperatures of 500, 600 and 650 ◦C; close to the points
A, B and C of this figure. Fig. 7 shows the XRD patterns
obtained from the sample Ti–72.7 at.% Al (Ti–60 wt.% Al)
as prepared after 30 h of milling and annealed at mentioned
temperatures. After annealing at 500 ◦C the diffraction pat-
tern shows a decomposition of the (Ti, Al)H compound into
Ti(Al) solid solution. It is not surprising that at this temper-
ature there is not signal from titanium hydride, which is due
to hydrogen desorption as was observed in the TGA curve.
After annealing at 600 ◦C the diffraction pattern does not
show significant changes respective to the sample annealed
at 500 ◦C. The thermal stability of the mixture apparently
was not affected by the temperature near the aluminum fu-
sion. However, for annealing at 650 ◦C, a clear change is
shown in the diffractogram, where crystalline phase of Al2Ti
is identified. On the other side, low intensity and broadened
peaks from Al or Ti(Al) phase are still observed. We assume
that two exothermic events (B and C) are taking place; one
corresponds to the formation of Al2Ti phase, which is main-
taining a compromise with a second endothermic event, fu-
sion of remnant aluminum. This behavior can be described
by the next relation:
3.2. Thermal stability of the powders
As mentioned before, the light metal hydrides are unstable
at temperatures higher than 250 ◦C. By doing the characteri-
zation of thermal stability of the powders we were able both
to eliminate the hydride compound from the powders and to
confirm their formation during the IRBM process. Addition-
ally, we were interested in studying the structural evolution
of the powders during the heat treatment in order to identify
intermetallic phases.
Fig. 6 shows the TG–DTA curves (dashed lines) for the
sample Ti–72.7 at.% Al (Ti–60 wt.% Al) after 30 h of milling,
measuredinanitrogenfluxwithaheatingrateof10.0 ◦C/min.
Simultaneously, structural changes of the milled and subse-
quently annealed sample were analyzed by XRD measure-
ments. TG curve shows a mass decreasing associated with the
desorption of hydrogen and some volatile species from 8.32 g
at 200 ◦C until 8.00 g at 600 ◦C, approximately 3.8 wt.%.
Above 700 ◦C, in TG curve, an increasing of mass, possi-
bly due to absorption of nitrogen from the atmosphere of
the measurement system to form Ti or Ti(Al) nitrides is ob-
served. DTA curve in the range from 300 to 700 ◦C shows
threeexothermicevents(pointsA, BandC, respectively). The
first exothermic peak (A) with a maximum at 474.1 ◦C, is due
to the hydrogen desorption and to a structure reordering such
as the formation of some crystalline phase of Ti(Al) solid so-
lution. The event between 550 and 700 ◦C is associated with
the formation of the crystalline phase of Al2Ti (long period
superstructure), as was proved by XRD measurements. The
next phase transformation was identified as an endothermic
event at about 658 ◦C corresponding to the aluminum fusion.
TiH1.924 + Al(Ti) + Ti
500 ◦C
−→ Ti(Al) + Al + H2 ↑ Ti(Al) + Al
650 ◦C
−→ Ti + Al2Ti + Al
(3)
TG–DTA curves from the sample Ti–94.1 at.% Al
(Ti–90 wt.% Al) after 36 h of milling time are shown also
in Fig. 6 (solid lines). For this sample the Ti content is low,
and consequently its capacity to form TiH1.942 is limited, as
already shown by the TG curve. The loss of mass in the range
Fig. 8. XRD patterns of the Ti–94.1 at.% Al (Ti–90 wt.% Al) powders ball milled by IRBM for 30 h of milling and heated to various temperatures.

J. Morales-Herna´ndez et al. / Journal of Alloys and Compounds 388 (2005) 266–273
273
from 200 to 600 ◦C due to hydrogen desorption is less than
0.1 g (∼0.9 wt.%). The DTA curve does not show notable
thermal events before 600 ◦C, except a wide exothermic peak
centered at about 450 ◦C. This released energy is associated
with both mass loss and reordering events of the system. For
higher temperatures, three events are marked: an endother-
mic (point Aꢀ, 659 ◦C) and two exothermic events (point Bꢀ,
726 and Cꢀ, 853 ◦C). Point Aꢀ corresponds to fusion of Al,
while Bꢀ and Cꢀ are phase transformations that can be un-
derstood from XRD measurements on annealed samples as
shown in Fig. 8. This figure was constructed with plots of
diffractograms obtained with samples annealed at 500 and
850 ◦C by 1 h under argon atmosphere. The plot of the sam-
ple “as prepared” defines the presence of TiH1.942 and Al(Ti)
phases, while the patterns for the sample annealed at 500 ◦C
showed peaks of elemental Al and the intermetallic Al3Ti
that can be formed according to the next reaction:
oxygen atmosphere. The corresponding TiN contaminant for-
mation was detected when the powders were exposed to the
ambient environment.
For the maximum milling time, about 36 h, samples with
high Al content were amorphous or nanostructured with a
crystalline grain less than 10 nm embedded in an amorphous
matrix. DTA and TGA results showed the hydrogen desorp-
tion and some structure reordering in the temperature range of
200 to 600 ◦C. As temperature increased, amorphous and/or
metastable phases changed to more stable phases such as
Al2Ti and Al3Ti intermetallic compounds.
Acknowledgements
This work was supported by CONACYT under the re-
search project G33178-U, and partially by CONCyTEQ
´
´
(Queretaro, Mexico). The authors recognize the helpful par-
◦
TiH1.924 + Al(Ti) 5−00→C Al3Ti + Al(Ti) + H2+ ↑ Al3Ti
´
´
ticipation of L.L. Dıaz Flores, M.A. Hernandez Landaverde,
R. Flores Farıas, M.C. Delgado Cruz, J.E. Urbina Alvarez,
and R. Martinez (CIMAV).
◦
´
+ Al(Ti) 8−50→C Al2O3 + Al(Ti)
(4)
XRD results for the sample annealed at 850 ◦C proved that
some oxygen is absorbed by Al to form Al2O3 in addition
to formation of the intermetallic Al3Ti with respect to the
sample annealed at 500 ◦C. Oxygen absorption could occur
due to the reaction of powders with residual oxygen in the
annealing atmosphere. Thus, the exothermic peaks Bꢀ and
Cꢀ, in Fig. 8, can be associated with the phase transformation
needed to form intermetallic Al3Ti and Al2O3, respectively.
References
[1] R. Janot, A. Rougier, L. Aymard, C. Lenain, R. Herrera-Urbina, G.A.
Nazri, J.M. Tarascon, J. Alloys Compd. 438 (2003) 356.
[2] C. Suryanarayama, Prog. Mater. Sci. 46 (2001) 1.
[3] P.L. Martin, D.A. Hardwick, in: J.H. Westbrook, R.L. Fleischer
(Eds.), Intermetallic Compounds, Wiley, New York, 1994, p. 637.
[4] A. Takasaki, Y. Furuya, Nanostruct. Mater. 11 (1999) 1205.
[5] T. Itsukaichi, K. Masuyama, M. Umemoto, I. Okane, J.G. Cabanas-
Moreno, J. Mater. Res. 8 (1993) 1817.
[6] M. Oehring, T. Klassen, R. Bormann, J. Mater. Res. 8 (1993) 2819.
[7] K.Y. Wang, J.G. Wang, G.L. Chen, J. Mater. Res. 10 (1995) 1247.
[8] S. Orimo, H. Fujii, T. Yoshino, J. Alloys Compd. 217 (1995)
287.
[9] T. Fuji, M. Kuzutani, N. Nakabo, K. Ameyama, J. Jpn. Inst. Met.
62 (1998) 462.
[10] A. Calka, J.S. Williams, Mater. Sci. Forum 88 (1992) 787.
[11] D.L. Zhang, D.Y. Ying, Mater. Lett. 50 (2001) 149.
[12] T. Itsukaichi, S. Shiga, K. Masuyama, M. Umemoto, I. Okane, Mater.
Sci. Forum 88 (1992) 631.
[13] J.S. Benjamin, M.J. Bomford, Metall. Trans. 8A (1977) 1301.
[14] N. Burgio, W. Guo, M. Mag´ın, F. Padella, F. Martelli, I. Soletta,
Proc. ASM Int. Conf. Struct. Appl. Mech. Alloying (1990) 175.
[15] K. Hashi, K. Ishikawa, K. Suzuki, K. Aoki, J. Alloys Compd. 330
(2002) 547.
[16] F.J. Espinoza-Beltra´n, O. Ceh-Soberanis, L. Garc´ıa-Gonza´lez, J.
Morales-Herna´ndez, Thin Solid Films 437 (2003) 170–175.
[17] K.I. Moon, K.S. Lee, J. Alloys Compd. 264 (1998) 258.
[18] K.I. Moon, H.S. Park, K.S. Lee, J. Alloys Compd. 325 (2001)
236.
[19] A. Takasaki, Y. Furuya, Nanostruct. Mater. 11 (1999) 1205.
[20] F. Zhang, L. Lu, M.O. Lai, J. Alloys Comp. 297 (2000) 211.
[21] H.A. Caldero´n, V. Garibay-Febles, A. Cabrera, M. Umemoto, J.G.
Caban˜as-Moreno, K. Tsuchiya, Mater. Res. Soc. Symp. Proc., vol.
552, 1999, Materials Research Society, KK5.6.1.
4. Conclusions
Indirect reactive ball milling was successfully applied to
obtain different nanocrystalline and amorphous powder al-
loys of the Ti–Al binary system. The effect of the milling
time on the structural evolution and thermal stability of the
samples was studied. A TiH1.924 phase was invariably de-
tected by XRD in all compositions even after 4 h of milling
by IRBM. We confirmed that the in situ formation of titanium
hydride material lead to an efficient and effective grain size
reduction in the samples regardless of the aluminium content
during milling. Compositions with Al contents less than x ≤
43.2 at.% showed faster grain size reduction. The grain size
observedfromTEMmicrographsandconfirmedwiththepro-
file analysis of XRD patterns was less than 15 nm after very
short milling times (4 h) and less than 5 and 10 nm for al-
loys with low (milled by 8 h) and high Al contents (milled by
36 h), respectively. Samples with high Ti content need special
care after milling, due to the high reactivity of the powders
of TiH1.924 and Ti(Al) when they are exposed to nitrogen and