Journal of Alloys and Compounds 490 (2010) 624–630

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier.com/locate/jallcom

Processing and surface properties of Al–AlN composites produced from

nanostructured milled powders
Hamid Abdoli a,∗ , Ehsan Saebnouri b , Seyed Khatiboleslam Sadrnezhaad a ,
Mohsen Ghanbari a , Taghi Shahrabi b
a
Materials and Energy Research Center (MERC), P.O. Box 14155-4777, Tehran, Iran
b
Department of Materials Science and Engineering, Tarbiat Modares University, P.O. Box: 14115-143, Tehran, Iran

a r t i c l e i n f o a b s t r a c t

Article history: Al–AlN nanostructured composites were fabricated via the powder metallurgy method using mechan-
Received 6 September 2009 ically milled aluminum powder mixed in a planetary ball-mill with different content of AlN (0, 2.5, 5,
Received in revised form 15 October 2009 10 wt.%) as the reinforcement. After milling for 25 h, powders were degassed and die-pressed uniaxially
Accepted 16 October 2009
in a steel die and then sintered at 650 ◦ C for different times. The sinterability, tribilogical and corro-
Available online 24 October 2009
sion behavior of composites were investigated at predefined conditions. Sinterability of composites was
degraded with increasing the reinforcement content. Wear resistance was improved by increasing vol-
Keywords:
ume fraction of AlN and this improvement was more pronounced at higher fraction of reinforcement.
Composite materials
Nanostructured materials
The potentiodynamic polarization was used for corrosion testing in 0.05 and 0.5 mol/L NaCl solutions.
Powder metallurgy According to the results of the experiment, the amount of the second phase did not exhibit any detectable
Sintering influence on the corrosion current density. In the diluted solution, Epit was reduced in the diluted solu-
Corrosion tion as a result of decrease in the amount of reinforcing particles. In the concentrated solution, Epit was
equal or lower than Ecorr for all the specimens. The SEM observations revealed that regular cracks were
distributed all over the surface and corrosion products had covered the surface.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction cold compaction and sintering could be of interest for production

Aluminum matrix composites (AMCs) have attracted much
attention in automotive, aerospace and military industries, due to
their high mechanical and physical behavior [1,2]. AMCs usually
are reinforced with SiC [3] and/or Al2 O3 [4]. In order to improve
the properties such as wear resistance, researchers have examined
other reinforcing materials such as carbide [5] and nitride ceram-
ics [6]. Aluminum nitride (AlN) is a good example with appropriate
physico-chemical, mechanical and thermal properties which sug-
gest it as a suitable reinforcement material for aluminum [7]. AlN
can hardly react with molten Al and exhibits high thermal con-
ductivity (320 W m−1 K−1 ), low dielectric constant (8.8 at 1 MHz),
high wear resistance, low density [8] and low coefficient of thermal
expansion (≈4.7 × 10−6 K−1 ), which is closed to that of silicon [8].
AlN particulates demonstrate high electrical resistivity (1013  cm)
by which the galvanic corrosion follows a reducing trend and they,
additionally, hydrolyze easily especially in an alkaline environment
[8,9].
Among many techniques applied for fabrication of aluminum
based composite materials, consolidation of milled powders by
∗ Corresponding author. Tel.: +98 912 319 2887.
E-mail address: habdoli@alum.sharif.edu (H. Abdoli).
of large batches of small parts [10]. Composite powder could be
obtained by milling of Al/AlN mixture with predefined ratio [7] or
reaction ball-milling of Al powder under nitrogen atmosphere. The
latter can be performed by ‘gas/solid reaction milling’ of elemen-
tal Al powder, i.e. attrition milling in the presence of urea [11,12]
or ‘cryomilling’ of Al, i.e. mechanical alloying in liquid nitrogen
[13–15]. However, although few reports [16,17] are available in the
literature focusing on the sintering process of mechanically milled
aluminum matrix composite powders, some studies [18–21] have
been performed on tribological and corrosion behavior of Al/AlN
composites.
One of the main obstacles in the use of aluminum alloy-based
composites is the influence of reinforcement on corrosion resis-
tance, where a protective oxide film imparts corrosion resistance.
The addition of a reinforcing phase can increase the frequency of
discontinuities on the film, thereby increasing the number of sites
where corrosion can be initiated and rendering the composite liable
to severe attacks [18,20,21].
Some researchers proposed that introduction of rein-
forcing material results in the preferential attack at the
reinforcement/aluminum matrix interface. Other stud-
ies [22,23], believed that the galvanic corrosion could be
restricted when a non-conductive reinforcement is exam-
ined. Researchers [24] also supposed that galvanic corrosion

0925-8388/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2009.10.121
 H. Abdoli et al. / Journal of Alloys and Compounds 490 (2010) 624–630 625

influenced by the nature of reinforcement and the interfacial

reaction.
The authors of the current manuscript have previously studied
processing of Al–AlN composite powder by mechanical milling [6],
study on its compressibility [25] and sintering [1] behaviors. In this
context, the procedure of composite fabrication and surface prop-
erties, including wear resistance and corrosion, were focused to
reveal the influence of reinforcing phase in different weight frac-
tions on the properties of produced composite.

2. Materials and methods

The specimens were provided from a powder mixture containing pure alu-
minum (Merck, Germany) and AlN (Aldrich, USA) powders, with variable weight
fractions of 0, 2.5, 5 and 10 wt.% AlN in each powder sample. 1.5 wt.% stearic acid
(Merck, Germany) was also added to the powder mixture as the process control
agent (PCA) and milling operation was conducted for 25 h in a planetary ball-mill
under the argon atmosphere. The ball-to-powder weigh ratio (BPR) and rotational
speed were chosen as 20:1 and 270 rpm, respectively. XRD analysis was performed
using X-ray diffractometer (Siemens X-ray diffractometer) with CuK␣1 radiation.
Particle size analysis was carried out by Fritsch ‘analysette 22’.
Milled powders were degassed at 400 ◦ C for 30 min under H2 atmosphere.
Cold-compacting of powders was performed at 1.5 GPa in a steel die set of 10 mm
diameter, using zinc stearat as die lubricant. Sintering process was performed at
650 ◦ C for 20, 30 and 60 min under N2 atmosphere with heating rate of 20 ◦ C min−1 .
The sintered compacts were cooled to room temperature in the sintering furnace.
Sintered density of compacts was measured by Archimedes technique. Microstruc-
ture of the specimens was examined using Cambridge S360 scanning electron
microscope (SEM).
Tribological and corrosion experiments were carried out on the samples with
different contents of AlN and sintered for 30 min using a pin-on disk wear test
machine. For the wear test experiments, four different samples of Al/AlN composite Fig. 1. Aluminum powder morphology: (a) before and (b) after milling procedure.
with 10 mm diameter as disk and 1.5 mm in diameter AISI 52100 steel ball with
hardness value of 850 HV were employed as counterpart. The sliding velocity was
kept constant for all tests to 0.05 m/s. The wear track radius was kept 4 mm and slid-
level to form low-angle grain boundaries, (iii) changing the low-
ing distance was 500 m. All tests were done at room temperature and four different
normal loads were applied between 1 and 10 N with 3 N interval. angle boundaries to higher angle types with random orientation.
Electrochemical tests were performed in 0.05 and 0.5 mol/L NaCl solution using a The acceleration of the grain refinement process by adding AlN
classical three electrodes cell with platinum as counter electrode, saturated calomel particles can be attributed to the generation of a high dislocation
electrode (SCE) as reference electrode and the samples with an exposed area of
activity caused by the interaction between hard particles and dis-
approximate 0.20 cm2 as working electrode. The electrolyte was in natural oxy-
gen containing state at ambient temperature. The potentiodynamic polarization
locations [27] (i.e. more plastic deformation via generation and
curves were obtained using an EG&G potentiostat model 273A at a constant voltage motion of dislocations).
scan rate of 1 mV/s. Polarization experiments started after the specimen had been Fig. 1 shows the micrograph of aluminum powder particles
immersed in the experimental solution for 2 h under open-circuit conditions. The before and after milling process. Quote to the figure, as-received
samples were polarized from −250 mV to 1600 mV versus OCP, and then the surface
powder possessed flake-like morphology and it is observed to have
morphologies were observed using SEM.
undergone a conversion from the flattened type to the equiaxed
one, after 25 h milling process. Change in powder morphology
3. Results and discussion was occurred as a result of repeatedly cold welding, fracture and
rewelding of particles through milling procedure.
3.1. Structural characterization In order to observe the way of particle packing during consol-
idation step, specimens were putted in the liquid nitrogen and
Table 1 lists the powder characteristics before and after milling then fractured by a hammer impact parallel to the pressing direc-
procedure, i.e. after 0 and 25 h milling. The crystallite size was cal- tion (Fig. 2). Obtained fracture surface from un-milled powders
culated by Williamson–Hall method regarding Warren correction (Fig. 2(a)) showed that flakes were oriented perpendicular to the
[6]. During milling process, particles have undergone severe plastic pressure direction and longitudinal voids are visible among the par-
deformation and subgrains or cells have been formed [26]. The con- ticles. For milled powders, (Fig. 2(b)), however, the fracture surface
version of subgrains or cells into grains during mechanical alloying is observed to reveal fine particles between the larger ones and
can be divided into three stages: (i) localization of deformation in the incidence of localized plastic deformation with particle inter-
shear bands consisting of a high dense network of dislocation, (ii) locking is noticeable. These observations are consistent with those
annihilation and recombination of dislocations at a certain strain reported by Hessabi et al. [28] in a similar work on Al–Al2 O3 system.
After powder preparation, milled powders were exposed to
Table 1 degassing treatment. Generally, degassing step is performed to
Characteristics of powders before and after milling process calculated by remove H2 and H2 O or any oxide layers surrounding the external
Williamson–Hall method. surface of the powder in milling process. Collisions during milling
Material Al Grain size Mean particle process remove this layer and, consequently, broken oxide layers
(nm) size (␮m) are dispersed into the materials. On the other hand, new oxides
As-received Al powder 343 53
could be regenerated on the fresh surface if the milling atmosphere
25 h-milled-Al powder 97 47 is not completely oxygen-free. Degassing produces crystalline ␥-
25 h-milled-Al/2.5 wt.% AlN powder 66 31 Al2 O3 that breaks up during compaction and facilitates the contacts
25 h-milled-Al/5 wt.% AlN powder 60 28 of fresh aluminum surfaces [29–31]. In other words, that non-
25 h-milled-Al/10 wt.% AlN powder 43 25
equilibrium structure of the milled powders is partially annealed
626 H. Abdoli et al. / Journal of Alloys and Compounds 490 (2010) 624–630
Fig. 2. SEM micrographs of the fracture surfaces of green compacts fabricated from:
(a) un-milled aluminum and (b) milled aluminum.
and more homogeneous microstructure is formed when powders
are degassed [32].
Fig. 3(a) depicts the effect of sintering time on relative density
of compacts. As a similar trend for all the specimens, density val-
ues demonstrate an increment at the initial stages of heating, after
which the rate of density increase endures a detectable reduction.
The sinterability of compacts can be evaluated by the following
equation:
s − g
= (1)
th − g
where s , g , th are sintered, green, and theoretical density,
respectively. As observed in Fig. 3(b), densification of Al and
Al–2.5 wt.% AlN compacts was completed after 20 min, while com-
plete sintering of composite compacts with 5 and 10 wt.% AlN was
implemented after 30 min. Therefore, prolonging the sintering time
does not seem to be a promising method to enhance the density
level significantly. On the other hand, it may be led to the grain
growth phenomenon which normally takes place in polycrystalline
materials so as to decrease the system energy by the reduction
of total grain boundary energy [33–35]. Nanocrystalline materials
are known to be thermodynamically unstable and success in their
consolidation is intimately related to the control of the competi-
tion between densification and coarsening to maintain superior
properties of the material generated by nanocrystallization
[36–38].
Table 2 gives the average crystallite size of the Al and Al–AlN
compacts after 30 min of sintering process. The noticeable role
played by AlN particles in deriving finer grains is attributed to the
pinning effect of AlN particles that lie in the migration path of the
boundaries and hinder their movements.
Fig. 3(b) shows that sinterability of composite pellets degrades
by increasing reinforcement weight percent. This is mainly resulted
from reducing green density of composite compacts due to higher
strength and hardness of composite powder, which leads to lower
contacts between adjacent particles. The hardness increase of alu-
Fig. 3. Variation of (a) relative density and (b) sinterability as a function of sintering
time.
minum powder after milling is originated from formation of refined
microstructure with random interfaces, distribution of fine oxides,
and high-density of dislocations resulted in higher hardness of
mechanically milled Al powder [1,6]. Higher hardness levels for
composite powder is mainly due to the presence of hard and non-
deformable AlN particles and their effective role as milling agent
[5] which results in smaller mean particle size of milled pow-
der (as summarized in Table 1). It is, therefore, reasonable that
degradation of compressibility occurs when the mean particle size
declines.
Fig. 4 compares the polished sections of the sintered Al–2.5 wt.%
AlN composite compacts after sintering for 20 and 60 min. Pore-free
surface of unreinforced matrix is consistent with Fig. 3(a), which
shows that 20 min of sintering is sufficient to reach near-full den-
sity.
3.2. Surface characterization
3.2.1. Wear behavior
The results obtained from sliding wear tests are shown in Fig. 5.
In this figure, the variation of wear factor, k (mm3 /N m), is plotted
Table 2
Crystallite size (nm) of Al and Al–AlN compacts sintered at
650 ◦ C for 30 min under N2 atmosphere.
Samples Grain size (nm)
Al 693
Al–2.5 wt.% AlN 414
Al–5 wt.% AlN 328
Al–10 wt.% AlN 287
 H. Abdoli et al. / Journal of Alloys and Compounds 490 (2010) 624–630
Fig. 4. Polished surface of Al–AlN composites sintered for (a) 20 min and (b) 60 min.
vs. applied normal load for different samples, where k (m2 /N) is the
volume loss material (m3 )/applied load (N) × sliding distance (m).
It is observed that the wear resistance increases for all applied
loads by increasing the volume fraction of reinforcement. Previ-
ously Bermúdez et al. [39] reported that under relatively mild
conditions, the main wear mechanism is oxidative and adhesive
wear, where only surface layers are being removed. By the presence
of hard AlN particles, wear resistance is improved because that the
existence of these particles at the surface of the material reduces the
oxidative areas which cause oxidative wear at low applied loads.
Fig. 5. Variation of wear factor (k) vs. applied load for samples.
627
Fig. 6. Potentiodynamic polarization curves of Al–AlN composite with different
weight fraction of the reinforcing particles in 0.05 mol/L NaCl solution.
This effect is more pronounced at higher volume fraction of AlN and
so at higher volume fraction of AlN the wear factor is low at mild
conditions.
At higher loads, as reported by Da Costa et al. [40] for Al/AlNi
composites, predominant wear mechanism in the material is abra-
sive wear. Therefore, it is reasonable to conclude that under this
condition, AlN particles cause hardening effect on the neighboring
matrix and so reduce its permanent deformation. Also, as a result
of abrasion resistance of hard AlN particles, abrasive attack of the
counter part is reduced. These effects are pronounced by increas-
ing AlN weight fraction. It is clearly seen that severe wear occurs in
the load of 7 N for composite materials containing 5 and 10 wt.% of
AlN. For monolithic-Al and composite containing 2.5 wt.% AlN, this
condition occurs at lower loads.
3.2.2. Corrosion behavior
The potentiodynamic polarization curves of monolithic Al
and Al–AlN-nanostructured composite specimens immersed in
0.05 mol/L NaCl solution are shows in Fig. 6. The results of Tafel
measurements for the specimens investigated are summarized in
Table 3. All potential values were measured vs. a saturated calomel
electrode (SCE). The values for Ecorr , Epit , ˇa , ˇc and icorr were
extracted from polarization curves where icorr is corrosion cur-
rent density, Ecorr is corrosion potential, Epit is pitting potential
(or breakdown), ˇa and ˇc are Tafel slopes and B is Stearn–Grey
coefficient.
Apparently, there are few differences between Ecorr of the sam-
ples. The corrosion current density values for Al–2.5 wt.% AlN and
Fig. 7. Potentiodynamic polarization curves of Al–AlN composite with different
weight fraction of the reinforcing particles in 0.5 mol/L NaCl solution.
628 H. Abdoli et al. / Journal of Alloys and Compounds 490 (2010) 624–630
Table 3
Corrosion parameters of samples extracted from Fig. 6.
Material Ecorr (mV) Epit (mV)
Al −705 −231
Al–2.5 wt.% AlN −713 −454
Al–5 wt.% AlN −720 −300
Al–10 wt.% AlN −728 −161
Al–5 wt.% AlN composites were smaller than that of monolithic Al
and sample Al–10 wt.% AlN. However, the calculated icorr for differ-
ent samples are not much far from others and, hence, the corrosion
current density values of the investigated composite materials
seem to be independent on the weight fraction of the reinforcing
particles.
Susceptibility of aluminum and aluminum alloys (as well as
other passive metals; like stainless steels) to pitting corrosion in
chloride environments, is studied commonly by evaluating Epit ,
namely the potential value which above it pre-existing passive film
breaks down and pits will be initiate [41,42]. Epit depends upon
the chloride content in the environment, and typically, a semilog-
arithmic relationship is found against the chloride concentration
[41–44]. Meantime, it also depends on the alloy composition, the
presence of other species in the solution (especially metal ions),
and the temperature (although the influence of temperature was
found to be negligible in the range from 0 to 30 ◦ C) [43–45]. Further-
more, appreciable differences in Epit can be introduced depending
upon the electrochemical method used for its evaluation (which
can be based on potential control, current control, or on open-
circuit measurements), and on the parameters adopted (e.g., scan
rate in potentiodynamic polarization tests) [44].
According to Table 3, adding the reinforcing particles caused to
decreasing of breakdown potential. It was expected because the
presence of second phase in a metal or alloy will result in forma-
Fig. 8. SEM micrographs showing surface morphology of Al–AlN composite with (a) 0, (b) 2.5, (C) 5 and (d) 10 weight fraction of the reinforcing particles after polarization
test in 0.5 mol/L NaCl solution.
icorr (␮A/cm2 ) ˇa (mV/decade) ˇC (mV/decade) B (mV/decade)
1.39 0.30 0.35 0.071
0.95 0.32 0.27 0.064
0.84 0.31 0.26 0.061
1.05 0.33 0.23 0.059
tion of microgalvanic cells and breakdown will happen at negative
potentials. But, an inverse effect was observed when AlN con-
tent was increased and Epit experienced an abrupt increment from
−454 mV for Al–2.5 wt.% AlN to −161 mV for Al–10 wt.% AlN sam-
ple. This may to be resulted from none electroconductivity of AlN
material and its monotonous dispersion in the aluminum matrix.
By regarding to the results of the Tafel polarization, the values of
cathodic Tafel slopes decrease by increasing the weight percent of
reinforcing particles, while the anodic Tafel slopes shows no signif-
icant changes. Therefore, the Stearn–Gray coefficient is decreased
by the increasing the weigh percent of reinforcement particulates.
The cathodic Tafel slope increment may be related to an enlarge-
ment of the corroded specimen area due to the aggressive attack,
while the anodic Tafel slope decrease may be related to the effect
of a partially protective layer, which also determines the slope rise
[45].
Fig. 7 plots the typical polarization curves for Al–AlN com-
posite with different weight fraction of the reinforcing particles
in 0.5 mol/L NaCl solution and Table 4 shows the corresponded
corrosion parameters of samples extracted from Fig. 7. The cor-
rosion current density values (icorr ) were decreased by increasing
the reinforcing particle content. Ecorr was equal or higher than Epit
measured in this solution for all samples. In all samples there is a
blunt change in current density on increasing potential, emanating
from weak points on the surface, which results in corrosive attack.
 H. Abdoli et al. / Journal of Alloys and Compounds 490 (2010) 624–630
Table 4
Corrosion parameters of samples extracted from Fig. 7.
Material Ecorr (mV) icorr (␮A/cm2 )
Al −808 16.60
Al–2.5 wt.% AlN −794 15.83
Al–5 wt.% AlN −783 15.04
Al–10 wt.% AlN −779 11.12
Fig. 9. SEM micrograph of a pit propagated around a reinforcement particle.
In comparison with polarization curve in lower chloride con-
tent solutions (Fig. 6), the Ecorr of all samples are shifted toward
the negative value. This change is more pronounced in the case of
monolithic aluminum, where the decrease of anodic slope (ˇa ) is
greater. For all samples, slopes of the anodic area are decreased
significantly while for cathodic area the slopes are increased. The
corrosion current densities (icorr ), calculated from polarization
curves, are greater in the high chloride content solution.
Fig. 8 shows the SEM images of the sample after potentiody-
namic polarization tests in 0.5 mol/L solution. After the polarization
tests, localized small pitting corrosion is observed on the surface of
the specimens. Regular cracks distributed all over the surface and
corrosion products covered the surface of the specimens are shown
in this figure. Fig. 8(c) shows the crack containing passive layer
which facilitates the reaching of attacking ions and raises the corro-
sion rate. This phenomenon is appeared as mud-cake type cracking
due to drying stage. White corrosion products are more revealed in
the micrograph prepared by back-scattered electron.
It is known that the corrosion properties of a milled powder is
influenced by its poor compressibility as well as the high stored
energy regions originated by accumulating the crystal defects [46].
Although the level of relative density reached to near the unity,
but it seems that the manner of powder compacting imposed its
effect on corrosion behavior of samples. Applied uniaxial pres-
sure during forming of green samples led particles to arrange in a
preferential direction perpendicular to applied pressure axis. Inten-
sive localized attack on the surface is worthy to note which is
corresponded to a pit propagated around a reinforcement parti-
cle (Fig. 9). Similar selective attack along preferential directions
was reported previously by Bertolini et al. [47] for extruded alu-
minum matrix composites. They also reported that the ceramic
particles which were fractured during sample preparation could
be supposed as “pit nucleation sites”.
4. Conclusions
Al–AlN-nanostructured composites were made via 25 h high-
energy ball-milling of powder mixture with different content of
AlN, followed by cold press and sintering. It was observed that the
629
ˇa (mV/decade) ˇC (mV/decade) B (mV/decade)
0.01 0.29 0.006
0.03 0.37 0.013
0.15 0.34 0.045
0.16 0.36 0.049
densification was occurred within 20 min for Al and Al–2.5 wt.%
AlN compacts, while it prolonged to 30 min for composites with 5
and 10 wt.% AlN. After that, continued sintering procedure did not
seem to be a promising method to enhance the density level signif-
icantly. Wear resistance of the sintered composites was improved
by increasing weight percent of AlN content as a result of increasing
the strength of the material against oxidative and abrasive wear at
low and high applied loads, respectively. Improvement in the wear
resistance was more appreciable for composites containing 5 and
10 wt.% AlN rather than 2.5 wt.% AlN and monolithic aluminum.
Addition of the AlN particles as reinforcement material to alu-
minum matrix had no significant effect on corrosion current density
in both 0.05 and 0.5 mol/L NaCl solutions, but Epit was reduced in
the diluted solution as a result of decrease in amount of reinforce-
ment particles from 10 to 2.5 wt.%. Moreover, based on the tests
carried out in concentrated solution, Epit was equal or lower than
Ecorr for all samples. SEM images of the sample after potentiody-
namic polarization showed uniformly distributed localized attack
over the surface.
Acknowledgment
The authors would like to acknowledge the Material and Energy
Research Center (MERC) for financial support of this research.
References
[1] H. Abdoli, H. Asgharzadeh, E. Salahi, J. Alloys Compd. 473 (2009) 116–122.
[2] M.E. Smagorinski, P.G. Tsantrizos, S. Grenier, A. Cavasin, T. Brzezinski, G. Kim,
Mater. Sci. Eng. A 244 (1998) 86–90.
[3] M.S. El-Eskandarany, J. Alloys Compd. 279 (1998) 263–271.
[4] B. Prabhu, C. Suryanarayana, L. Vaidyanathan, R. An, Mater. Sci. Eng. A 425
(2006) 192–200.
[5] E.M. Ruiz-Navasa, J.B. Fogagnolo, F. Velasco, J.M. Ruiz-Prieto, L. Froyen, Compos.
Part A 37 (2006) 2114–2120.
[6] H. Abdoli, E. Salahi, H. Farnoush, K. Pourazrang, J. Alloys Compd. 461 (2008)
166–172.
[7] C. Troadec, P. Goeuriot, P. Verdieq, Y. Laurent, J. Vicenqc, G. Boitier, J.L. Cher-
mantc, B.L. Mordike, J. Eur. Ceram. Soc. 17 (1997) 1867–1875.
[8] E. Salahi, J.G. Heinrich, Br. Ceram. Trans. 102 (2003) 161–168.
[9] Z. Liu, G. Huang, M. Gu, Mater. Chem. Phys. 99 (2006) 75–79.
[10] J. Cintas, F.G. Cuevas, J.M. Montes, E.J. Herrera, Scripta Mater. 52 (2005)
341–345.
[11] J. Cintas, F.G. Cuevas, J.M. Montes, E.J. Herrera, Scripta Mater. 53 (2005)
1165–1170.
[12] J. Cintas, J.M. Montes, F.G. Cuevas, E.J. Herrera, J. Alloys Compd. 458 (2008)
282–285.
[13] C. Goujon, P. Goeuriot, M. Chedru, J. Vicens, J.L. Chermant, F. Bernard, J.C. Niepce,
P. Verdier, Y. Laurent, Powder Technol. 105 (1999) 328–336.
[14] E.J. Lavernia, B.Q. Han, J.M. Schoenung, Mater. Sci. Eng. A 493 (2008) 207–214.
[15] D.B. Witkin, E.J. Lavernia, Prog. Mater. Sci. 51 (2006) 1–60.
[16] S.M. Zebarjad, S.A. Sajjadi, Mater. Des. 28 (2007) 2113–2120.
[17] N. Zhao, P. Nash, X. Yang, J. Mater. Process. Technol. 170 (2005) 586–592.
[18] A.J. Griffiths, A. Turnbull, Corros. Sci. 36 (1994) 23–35.
[19] K. Lucas, H. Clarke, Corrosion of Aluminum-based Metal Matrix Composites,
Research Studies Press Ltd., 1993.
[20] A.K. Mishra, R. Balasubramaniam, S. Tiwari, Anti-Cor. Meth. Mater. 54 (2007)
37–46.
[21] A.J. Trowsdale, B. Noble, S.J. Harris, I.S.R. Gibbins, G.E. Thompson, G.C. Wood,
Corros. Sci. 38 (1996) 177–191.
[22] M.M. Buarzaiga, S.J. Thorpe, Corrosion 50 (1994) 176–185.
[23] P.P Trzaskoma, Corrosion 46 (1990) 402–409.
[24] S.M.A. Shibli, B. Jabeera, R. Manu, Mater. Lett. 61 (2007) 3000–3004.
[25] H. Abdoli, H. Farnoush, E. Salahi, K. Pourazrang, Mater. Sci. Eng. A 486 (2008)
580–584.
[26] D.L. Zhang, Prog. Mater. Sci. 49 (2004) 537–560.
[27] K.H. Chung, J. He, D.H. Shin, J.M. Schoenung, Mater. Sci. Eng. A 356 (2003) 23–31.
630 H. Abdoli et al. / Journal of Alloys and Compounds 490 (2010) 624–630

[28] Z.R. Hessabi, H.R. Hafizpour, A. Simchi, Mater. Sci. Eng. A 454 (2007) 89–98. [39] M.D. Bermúdez, G. Martínez-Nicolás, F.J. Carrión, I. Martínez-Mateo, J.A.
[29] S.J. Hong, P.W. Kao, Mater. Sci. Eng. A 119 (1989) 153–159. Rodríguez, E.J. Herrera, Wear 248 (2001) 178–186.
[30] S.J. Hong, P.W. Kao, C.P. Chang, Mater. Sci. Eng. A 148 (1992) 189–195. [40] C.E. Da Costa, F. Velasco, J.M. Torralba, Mater. Trans. A 33 (2002) 3541–3553.
[31] Y.C. Kang, S.L. Chan, Mater. Chem. Phys. 85 (2004) 438–443. [41] E.H. Hollingsworth, H.Y. Hunsicker, Metals Handbook, ASM International, Met-
[32] N. Zhao, P. Nash, Powder Met. 44 (2001) 333. als Park, OH, 1987, pp. 583–609.
[33] C.L.D. Castro, B.S. Mitchell, Mater. Sci. Eng. A 396 (2005) 124–128. [42] Z. Szklarska-Smialowska, Pitting Corrosion of Metals, NACE, Houston, TX, 1986.
[34] M.C. Iordache, S.H. Whang, Z. Jiao, Z.M. Wang, Nanostruct. Mater. 11 (1999) [43] H. Bohni, H.H. Uhlig, J. Electrochem. Soc. 7 (1969) 906–910.
1343–1349. [44] K. Nisancioglu, H. Holtan, Corros. Sci. 18 (1978) 835–849.
[35] F. Liu, R. Kirchheim, J. Cryst. Growth 264 (2004) 385–391. [45] A.R. Olszyna, A. Sokolowsk, J. Szmidt, A. Werbowy, M. Bakowski, Int. J. Inorg.
[36] A. Kazaryan, Y. Wang, S.A. Dregia, B.R. Patton, Acta Mater. 50 (2002) 2491–2502. Mater. 3 (2001) 1311–1313.
[37] P.C. Millett, R.P. Selvam, A. Saxena, Acta Mater. 55 (2007) 2329–2336. [46] T.G. Durai, K. Das, S. Das, J. Alloys Compd. 462 (2008) 410–415.
[38] L. Shaw, H. Luo, J. Villegas, D. Miracle, Acta Mater. 51 (2003) 2647–2663. [47] L. Bertolini, M.F. Brunella, S. Candiani, Corrosion 55 (1999) 422–431.