Open Journal of Composite Materials, 2013, 3, 97-106
http://dx.doi.org/10.4236/ojcm.2013.34010 Published Online October 2013 (http://www.scirp.org/journal/ojcm)
97
Effect of Reinforcement Clustering on Crack Initiation
Mechanism in a Cast Hybrid Metal Matrix Composite
during Low Cycle Fatigue
A. K. M. Asif Iqbal, Yoshio Arai*, Wakako Araki
Graduate School of Science and Engineering, Saitama University, Saitama, Japan.
Received July 3rd, 2013; revised August 3rd, 2013; accepted August 13th, 2013
Copyright © 2013 A. K. M. Asif Iqbal et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
ABSTRACT
The reinforcement distribution of metal matrix composites (MMCs) plays an important role in low cycle fatigue. Thus,
it is essential to study the effect of reinforcement clustering on the crack initiation mechanism of MMCs. In this study,
the effect of reinforcement clustering on the microcrack initiation mechanism in a cast hybrid MMC reinforced with
SiC particles and Al2O3 whiskers was investigated experimentally and numerically. Experimental results showed that
microcracks always initiated in the particle-matrix interface, located in the cluster of the reinforcements. The interface
debonding occurred in the fracture which created additional secondary microcracks due to continued fatigue cycling.
The microcrack coalesced with other nearby microcracks caused the final fracture. To validate the experimental results
on the microcrack initiation, three dimensional unit cell models using finite element method (FEM) were developed.
The stress distribution in both the reinforcement clustering and non-clustering regions was analyzed. The numerical
analysis showed that high stresses were developed on the reinforcements located in the clustering region and stress
concentration occurred on the particle-matrix interface. The high volume fraction reinforced hybrid clustering region
experienced greater stresses than that of the SiC particulate reinforced clustering region and low volume fraction rein-
forced hybrid clustering region. Besides, the stresses developed on the non-clustering region with particle-whisker se-
ries orientation were reasonably higher than that of the non-clustering region with particle-whisker parallel orientation.
The high volume fraction reinforced hybrid clustering region is found to be highly vulnerable to initiate crack in cast
hybrid MMC during low cycle fatigue.
Keywords: Cast Metal Matrix Composites; Crack Initiation; Reinforcement Clustering; Low Cycle Fatigue
1. Introduction
The metal matrix composites (MMCs) provide a combi-
nation of the metallic properties of the matrix (high
toughness) with the ceramic properties of the reinforce-
ment (high strength and high modulus) to give a material
greater strength and stiffness, higher temperature capa-
bilities and more excellent wear resistance than a similar
monolithic material [1-5]. Therefore, MMCs are particu-
larly attractive for structural applications such as aero-
space and automotive industries and wear applications,
especially in the frictional area of braking system [1].
The production techniques of MMCs have been well
advanced in recent years, such as powder metallurgy, the
extrusion process and liquid infiltration. However in
practice, it is often difficult to obtain a homogeneous
distribution of reinforced particles or whiskers. Further, it
has been found that the non-uniformity in the reinforce-
ment arrangement can have significant effects on the me-
chanical properties of the MMCs [6,7]. Existing experi-
mental and theoretical evidences suggest that the homo-
geneity of particles or whiskers spatial arrangement plays
a key role in controlling the yield strength, ductility, fa-
tigue and fracture behaviors of MMCs [8]. It is generally
agreed that the yield strength and the work hardening
increase with increased clustering of reinforcements [9].
However, the failure strain is significantly reduced in a
clustered microstructure [10]. This is often attributed to
the stress concentration in the reinforcement clusters [11],
which may lead to preferential nucleation and propaga-
*Corresponding author.

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
98
tion of damage in the clusters. Davidson [12] has ex-
perimentally observed crack initiation in particle clus-
ters in the AA2014 + 15% SiCp MMC, and reported that
the preferential site for crack propagation is the regions
of higher particle volume fractions. In addition, Boyd et
al. [13] observed that damage develops in the particle-
rich zones in SiCp reinforced Al alloy. This damage was
caused by particle rupture with interfacial decohesion.
They also observed that on the fracture surface of tested
specimens, the particle density was higher than the mate-
rial mean value. Bourgeois et al. [14] showed experi-
mentally on an Al-SiC composite that damage grows first
in the whole composite but localization occurs in a parti-
cle-rich zone. Besides, Prangnell et al. [15] studied com-
pression of an Al-Si/SiCp MMC. They observed that the
porosity increases more in the clustered zone than in the
matrix zone, and particle cracking is the less dominant
fracture mode in the clustered region than in the other
region. Murphy and Clyne [16] pointed out in the case of
as-cast Al-Si/SiCp composite, the preferential occurrence
of porosity growth and particle rupture in clustered re-
gion. Furthermore, Yoshimura et al. [17] found that par-
ticle clustering decreases composite ductility and ulti-
mate strength. Numerical simulation has also been done
by many researchers which allowed the study of cluster-
ing more intensively. Li et al. [18] showed experimen-
tally and verified by simulation on an Al/SiCp composite
that rupture localizes preferentially in particle-rich re-
gions. Moreover, Tszeng [19] modeled MMC in order to
point out the influence of some cluster characteristics.
This simulation showed that the load at which crack nu-
cleation occurs is lower in the cluster than outside. At
present, numbers of researchers have employed the nu-
merical analysis to predict the effects of reinforcements
on the MMCs. By and large, these analyses approached
the problem by considering the unit cell model, where
one particle or whisker was embedded in matrix [20,21].
In addition, the shape of the particle or whisker was as-
sumed to be cylindrical, spherical, rectangular or cubical
[22,23]. Zhang et al. [24] have analyzed the effect of par-
ticle clustering on the flow behaviour of SiC particulate
reinforced Al-MMC by using the microstructure based
cell model. They successfully predicted the flow behav-
iour and revealed that the percentage of the particle
cracking in the particle clustering model is higher com-
pared with that in the particle random distribution model.
These results concerning the damage initiation and frac-
ture weigh in favor of either particulate- or whisker-re-
inforced MMCs. The experimental and numerical ana-
lysis for hybrid MMC is rare. At present, studies of cast
hybrid MMCs are limited in the investigation of fracture
mechanisms and wear properties [25]. Besides, the au-
thors of this article have previously investigated the mi-
crocrack initiation and stage by stage growth of the crack
to final failure in cast hybrid MMC [26]. They experi-
mentally investigated the hybrid effect on the crack ini-
tiation mechanism. However, the effect of reinforce-
ment clustering on the initiation of microcracks in hybrid
MMC is very complicated due to the presence of parti-
cles and whiskers together. Due to the complicated mi-
crostructure, various experimental and numerical inves-
tigations are needed to be explained to clarify the damage
nucleation and fracture mechanism of the composite.
Thus, in the present work, the effect of reinforcement
clustering on the crack initiation mechanism in cast hy-
brid MMC reinforced with SiC particles and Al2O3
whiskers was investigated. Fractographic analysis was
used to explain the clustering dependency in microcrack
nucleation. A three-dimensional unit-cell model in the
periodic boundary condition was developed using finite
element method (FEM) to analyze the stress distribution
in the clustering and non-clustering regions which could
lead to microcrack initiation in cast hybrid MMC.
2. Materials and Experimental Procedures
The cast hybrid metal matrix composite was fabricated
with 21 vol% SiC particles and 9 vol% Al2O3 whiskers as
reinforcements and the aluminium alloy JIS-AC4CH as
matrix [27]. The material was fabricated by the squeeze
casting process with a 100 MPa maximum pressure, us-
ing a hybrid preform made of SiC particles and Al2O3
whiskers. The squeeze casting pressure of 100 MPa was
adequate to overcome the resistance against flow and to
press the melt into all the open pores of the hybrid pre-
form. The materials were heat treated using the T7 proc-
ess. The chemical composition of AC4CH alloy is listed
in Table 1. As described in the previous report [26], due
to the formation process of the preform the whiskers in
the hybrid MMC are randomly oriented in a plane. “Lon-
gitudinal cross section” is defined as the plane and “lat-
eral cross section” is defined as the one perpendicular to
the plane in this article.
Figure 1 shows the lateral and longitudinal micro-
structures of the hybrid MMC. Most of the SiC particles
in the hybrid MMC were rectangular with sharp corners
(Figures 1(a) and (b)), and most of the Al2O3 whiskers
were roller-shaped (Figures 1(a) and (b)). The average
length of SiC particles and the Al2O3 whiskers was 23
µm and 35 µm respectively. The average diameter of the
Al2O3 whiskers was 2 µm. The Al2O3 whiskers were ran-
domly oriented in the same plane as the longitudinal cross
section of the specimen. At frequent intervals, clusters of
SiC particles and Al2O3 whiskers were observed in the hy-
brid MMC as indicated by the broken lines in Figures

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
99
Table 1. Chemical composition of AC4CH alloy, (wt%).
Si Fe Mg Ti Al
7.99 0.2 (max) 0.57 0.07 Bal
(a)
(b)
Figure 1. Microstructure in (a) lateral and (b) longitudinal
cross section of hybrid MMC showing cluster of reinforce-
ments.
1(a) and (b), respectively. The local volume fraction of
the reinforcement in the clustering region is found to be
60%. The mechanical properties of reinforcement mate-
rials and hybrid MMC are shown in Table 2. The listed
properties for the hybrid MMC are along the direction in
the longitudinal cross section. Conventional three point
bending tests were carried out on rectangular bar smooth
specimens to reduce the observation area of crack initia-
tion. The specimen dimensions were as follows: length of
100 mm, thickness of 6 mm and width of 8 mm. The ma-
chined surfaces of the specimens were polished by using
a polishing machine with 15, 3, and 1 µm diamond parti-
cles sequentially until all scratches and surface machin-
ing marks were removed. Three point bending tests were
performed using special bending fixtures equipped with a
5 kN load cell in a Shimazu Servo Pulser. The span dis-
tance was 60 mm. Load and deflection data were re-
Table 2. Mechanical properties of reinforcement and tested
materials.
Parameters Al2O3 SiC Al alloy
AC4CH
Hybrid
MMC
Young’s modulus (GPa)380 450 70.0 142
Poisson’s ratio 0.27 0.20 0.33 0.28
Yield strength (MPa)- - 131 166
Tensile strength (MPa)- - 262 228
Tensile elongation (%) 9.22 2.77
corded by a computer data acquisition system. First, the
Monotonic bending tests were conducted at a displace-
ment rate of 0.0025 mm·s−1. The strength of the hybrid
MMC was calculated from the maximum load at failure
as a nominal bend stress σc = 386 MPa. Cyclic fatigue
tests were conducted in the load control mode under the
load ratio R = 0.1 at the frequency of 0.5 Hz. Three
specimens of hybrid MMC which are mentioned in this
article TP-1, TP-2 and TP-3 were tested with the maxi-
mum stresses of 0.7σc, 0.8σc, and 0.9σc. All tests were
carried out at room temperature. The number of cycles to
failure was considered as the fatigue life Nf. To observe
the initiation and growth of microcracks, the plastic rep-
lica technique was used at various times during the fa-
tigue life. During replication, the specimen was held at
mean load to ensure that any cracks that might be present
would be fully opened. Replicas were taken using Bioden
replicating films softened in acetone. Finally, an optical
microscope was used to examine the replicas. Prior to
testing, no cracks were seen. In this article, “crack initia-
tion” is defined as the point at which a black line of sev-
eral micrometers is first observed in the magnified rep-
lica image, during the cyclic loading test. The tensile and
fracture surfaces were comprehensively examined using
scanning electron microscopy (SEM) and energy-disper-
sive x-ray spectroscopy (EDS) to characterize the crack
initiation site.
3. Experimental Results
The initiation and early propagation of the microcracks
on the smooth surface of the hybrid MMC are observed
from optical micrographs of the same areas on replicas
during fatigue testing. Figure 2 shows the optical micro-
graphs of replicas obtained at various stages of fatigue
testing of the hybrid MMC specimen TP-2. For TP-2, at
14% of the fatigue life, several cracks of length around
10 - 15 µm were initiated (indicated by arrows in Figure
2(a)). These cracks then coalesced together, and at 26%
of the fatigue life, the crack extended to 160 µm in length
(Figure 2(b)). A few cracks 15 - 25 µm in length were
formed ahead of the main crack tip at 60% of the fatigue

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
100
life, (arrows in Figure 2(c)). At 90% of the fatigue life,
all of these cracks coalesced and a fatal crack was pro-
duced (Figure 2(d)). The size of the fatal crack was
around 500 µm on the specimen surface. The final failure
took place at 2110 cycles. Similar phenomena were ob-
served in the specimen TP-1 and TP-3. Figure 3 shows
the optical micrograph of the crack initiation site at the
matching tensile surface of the fractured specimen TP-2.
It is apparent that the microcracks initiated at the parti-
cle-matrix interface (as indicated by the “Particle1” and
the “Matrix1” arrows on the left in Figure 3 which cor-
respond to the microcracks in Figure 2(a)). The crack
initiated SiC particle was located in the cluster of SiC
particles and Al2O3 whiskers (as indicated by the circles
“cluster-1” in Figure 3). Moreover, another two SiC par-
ticles were found located in parallel and in series to the
crack initiated SiC particle in the clustering region (As
indicated by the “Particle2” and “Particle3” arrows re-
spectively in Figure 3). Furthermore, one Al2O3 whisker
was found located parallel and another one located in
series (as indicated by the “Whisker1” and “Whisker2”
arrows respectively in Figure 3) very close to the crack
initiated SiC particle. The cluster-2 region of Figure 3
indicates the secondary microcrack initiation site. It is
also observed that the secondary microcracks were initi-
ated at the particle-matrix interface which was located in
the cluster of SiC particles and Al2O3 whiskers (as indi-
cated by the “Particle4” and the “Matrix2” arrows in Fig-
ure 3, which correspond to the microcracks in Figure
2(c)). Figure 4 shows the SEM image of the matching
fracture surfaces of the microcrack initiation site at the
cluster of SiC particles and Al2O3 whiskers in the hybrid
MMC specimen TP-2. The cluster region is indicated by
the circle “cluster-1” in Figure 4(a) which is corre-
sponding to the “cluster-1” in Figure 3 where the inner
particle clustering is shown in Figure 4. The dark flat
area indicated by P1 in Figure 4(a) corresponds to the
location indicated by the “Particle1” arrow in Figure 3,
and the area M1 in Figure 4(a) corresponds to the loca-
tion indicated by the “Matrix1” arrow in Figure 3. Fig-
ure 4(b) shows the EDS mapping analysis results on the
areas corresponding to Figure 4(a). The presence of Al,
Si, and O on the fracture surfaces is indicated by the
green, blue, and red colors respectively in Figure 4(b).
The blue area indicated by P1 in Figure 4(b) contains a
significant amount of Si (94%) and a small amount of Al
(6%), identifying the area as a SiC particle (correspond-
ing to P1 in Figure 4(a)). The green area indicated by M1
contains a large amount of Al (92%) and a small amount
of Si (8%), indicating that this area is Al matrix (corre-
sponding to M1 in Figure 4(a)). Therefore, the blue and
green area indicated by the P1-M1 pair in the matching
halves denoted the crack initiation site in the cluster of
SiC particles and Al2O3 whiskers (Figure 2(a)) where
SiC particle-matrix interfacial debonding occurred. The
(a) (b)
(c) (d)
Figure 2. Crack initiation at the cluster of reinforcement
and crack propagation at various stages of fatigue life of
hybrid MMC specimen TP-2: σmax = 308 MPa, Nf = 2110
cycles. (a) N1/Nf = 0.14; (b) N1/Nf = 0.26; (c) N1/Nf = 0.6; (d)
N1/Nf = 0.9.
Figure 3. Optical micrograph of the crack initiation site at
the matching tensile surface of the fractured specimen
TP-2.

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
101
(a)
(b)
Figure 4. Matching fracture surface of microcrack initiation
site in TP-2: (a) SEM micrograph (b) EDS mapping analy-
sis.
blue P2-P2 and P3-P3 pair in Figure 4(b) indicated the
presence of SiC particles on both sides of the fractured
surface, meaning that interface debonding was followed
by transgranular fracture in this clustering region (corre-
sponding to the P2-P2 and P3-P3 pair in Figure 4(a)). The
coexistence of green and red, indicating the presence of
both Al and O, identifies this area as an Al2O3 whisker,
denoted by W1 in Figure 4(b) (corresponding to W1 in
Figure 4(a)). This Al2O3 whisker was located very close
to the debonded SiC particle. Interfacial debonding was
also found in this Al2O3 whisker, as indicated by the
W1-M2 pair in Figure 4(b) (corresponding to the W1-M2
pair in Figure 4(a)). Between the P1-M1 pair and the
neighboring SiC particle on the specimen surface in Fig-
ure 4(a), a number of dimples were nucleated (indicated
by the D arrow in Figure 4(a)) in the aluminum alloy
matrix. EDS mapping analysis confirmed the presence of
a few Si particles on the opposite side of the dimples (as
indicated by the Si arrow in Figure 4(b)). Dimple forma-
tion indicated the occurrence of void nucleation, which
was induced by plastic deformation of the Al matrix at
the second phase Si particles. However, the edge of the
dimples was not as clear as those in the unstable fracture
region are, likely due to the mutual contact effect due to
cyclic loading. Similar observations were made for the
fracture surface of the specimens TP-1 and TP-3. The
above observations clearly demonstrated the effect of
reinforcement cluster on microcrack initiation mecha-
nism in hybrid MMC. Microcracks were always initiated
at the particle-matrix interfaces where an Al2O3 whisker
was also located very near to the crack initiated SiC par-
ticle (marked by W1 in Figure 4). Previous researches
reported that the main reason of microcrack initiation in
the particulate or whisker reinforced system is debonding
at the reinforcement-matrix interface [6,28,29]. The re-
sults obtained for hybrid MMC are consistent with these
observations. Moreover, the results showed that the crack
initiated particles always existed in the cluster of rein-
forcement, indicating the vulnerability of the clustering
region in microcrack initiation.
Figure 5 illustrates the surface view and the cross sec-
tional view for describing the crack initiation mechanism
in the cluster of the hybrid MMC. It appears that the
crack initiation sites of hybrid MMC is at the interface of
the particle-matrix which is located in the cluster of the
reinforcement. This clustering effect might be occurred
due to the elastic-plastic interaction between the rein-
forcing particles, whisker and matrix. Furthermore, the
local volume fraction of the reinforcements in the clus-
tering region is found reasonably higher than that of the
non-clustering region which could lead to initiate micro-
cracks in the reinforcement clustering region. Besides,
few SiC particles and Al2O3 whiskers were found located
in parallel and series orientation to the crack initiated SiC
particle which might affect the crack initiation mecha-
nism of hybrid MMC. Therefore, to understand the char-
acteristics of the elastic-plastic stress fields, a numerical
analysis is carried out on the reinforcing particles, whisk-
(a) (b)
Figure 5. Schematic diagram of fatigue crack initiation
mechanism at the cluster of reinforcement in hybrid MMC
(a) surface view (b) cross sectional view.

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
102
ers and matrix in the clustering and non-clustering re-
gions as described in the following section.
4. Numerical Model
To characterize the effect of reinforcement clustering on
stress distribution in hybrid MMC (reinforced by SiC
particles and Al2O3 whiskers), three dimensional (3-D)
unit cell models using finite element method (FEM) were
developed. ABAQUS software [30] was used for the cal-
culation. The models consist of SiC particle, Al2O3
whisker and Al alloy matrix. The schematic illustrations
and finite element mesh of the models are shown in Fig-
ures 6-8. Figure 6 represents the reinforcement cluster-
ing region of hybrid MMC with three different SiC parti-
cle and Al2O3 whisker arrangement. In model-1 (Figure
6(c)), SiC particulate reinforced clustering region is
shown. In Model-2 and 3 (Figures 6(d) and (e)), parti-
cles are located around a whisker, showing the clustering
of both the SiC particles and Al2O3 whisker, representing
the hybrid clustering regions. Whereas, Model-4 and 5
(Figures 7 and 8) includes a particle and a whisker
which are located in series and parallel orientation to
each other, representing the non-clustering region of hy-
brid MMC. In this numerical analysis, it is assumed that
all the SiC particles other than the center particles in
Model-1 are cubic shape of length b and the whiskers are
rectangular shape of length l and width d respectively.
However, the center SiC particles of model-1 (Figure
6(c)) are considered as the rectangular shape of length a,
width w and height h. Only 1/8 of one unit cell is treated
because of the symmetry of the cell. For models-3, 4 and
5, reinforcement volume fractions is modeled as real mi-
crostructure of hybrid MMC of 9 vol% Al2O3 whisker
and 21 vol% SiC particles in an Al alloy matrix. On the
other hand, the reinforcement volume fraction of model-
1 and model-2 is kept 58 vol% and 51 vol% respectively
as the local volume fraction of the reinforcements usually
increases in the clustering region as shown in Figure 1.
Size determination of the model was made by the fol-
lowing formulae: b3/LH2 = Vp (for cubic shaped particles)
and ld2/LH2 = Vw (for whisker), where, Vp and Vw is the
particle and whisker volume fraction respectively. The
SiC particles and Al2O3 whiskers were modeled as linear
elastic. The behaviour of Al alloy matrix was modeled as
isotropic elastic-plastic response. Symmetric boundary
condition is applied to the Y-Z plane at the left, X-Y
plane at the front and X-Z plane at the top of the models.
Moreover, periodic boundary condition is applied to the
Figure 6. Schematic illustration of reinforcement clustering region of hybrid MMC: (a) periodic particle and whisker ar-
rangement (b) 1/8 of one unit cell, analysis based on symmetry (c) finite element mesh of Model-1 (d) Model-2 (e) Model-3.

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
103
(a) (b)
Figure 7. Model-4 representing the non-cl ustering region of hybrid MMC where particle and whisker is placed in series: (a)
schematic illustration of the periodic particle and whisker arrangement (b) finite element mesh.
Figure 8. Model-5 representing the non-clustering region of
hybrid MMC where particle and whisker is placed in par-
allel: (a) schematic illustration of the periodic particle and
whisker arrangement (b) finite element mesh.
X-Z plane at the bottom and X-Y plane at the back of all
the models. A uniform displacement of 0.04 µm for the
half length of the unit cell, 20 µm is applied to the Y-Z
plane at the right of all the models which is correspond-
ing to a 308 MPa nominal stress for the hybrid MMC.
Numerical Results and Discussion
In order to study the effects of reinforcement clustering
on crack initiation mechanism, the stress distribution in
hybrid MMC are highlighted in Figure 9. Figures 9(a),
(b) and (c) represent normal stresses acting along the
loading direction in the reinforcement clustering regions
(Model-1, 2 and 3) whereas Figure 9(d) and e represent
the normal stresses developed in the non-clustering re-
gions (Model-4 and 5). It can be apparently found that
the normal stresses developed in the reinforcement clus-
tering regions (Figures 9(a), (b) and (c)) are significantly
greater than that of the non-clustering regions (Figures
9(d) and (e)). It is noteworthy that the clustering regions
reinforced with both SiC particles and Al2O3 whiskers,
i.e. hybrid clustering region (Model-2 and 3, Figures 9(b)
and (c)) experiences reasonably higher normal stresses
than that of the clustering regions which is reinforced
with only SiC particles (Model-1, Figure 9(a)), indicat-
ing the hybridization effect on the stress concentration
and vulnerability of crack initiation in this region. More-
over, the hybrid clustering region with high volume frac-
tion reinforcements (Model-2, Figure 9(b)) have greater
stresses than that of the clustering region with low vol-
ume fraction reinforcement (Model-3, Figure 9(c)), in-
dicating the influence of local volume fraction on the
stress development in the clustering region of hybrid
MMC. Besides, in all the models, the SiC particles-Al
alloy interface edges experience higher stresses. This is
attributed to the fact of the stress concentration in the
vicinity of the reinforcement. Furthermore, the maximum
stress concentration at the particle-matrix and whisker-
matrix interface located in the clustering regions (Fig-
ures 9(b) and (c)) is much higher than that in the non-
clustering regions (Figures 9(d) and (e)), suggesting the
early interface debonding at the clustering regions and
initiation of crack. In addition, it can be seen from Fig-
ures 9(b) and (c) that the stresses developed in the areas
indicated by “m” in Model-2 and Model-3 in Figure 6(b)
are reasonably higher than the stresses developed in other
sides of the clustering regions. This clearly indicates the
hybrid effect of the hybrid MMC. The SiC particle and
Al2O3 whisker deform elastically within the plastically
deforming Al alloy matrix. Once the particle and whisker
existed very close to one another, the elastic-plastic in-
teraction occurs between these three materials, results the
higher stress concentration at this location and cracks
likely to initiate in this place. Therefore, the reinforce-
ment clustering region where SiC particles and Al2O3
whisker exist very close to one another is highly vulner-
able for crack initiation.
Figure 10 represents the variation of the maximum
and minimum normal stresses developed on the interface
of the reinforcement-matrix in all the models of both the

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
104
Figure 9. Numerical results normal stress along loading direction (a) in the SiC particulate reinforcement clustering region
Model-1, (b) in the SiC particle and Al2O3 whisker reinforcement clustering region Model-2, (c) in the SiC particle and Al2O3
whisker reinforcement clustering region Model-3, (d) in the reinforcement non-clustering region Model-4 and (e) in the rein-
forcement non-clustering region Model-5.
Figure 10. Comparison of the maximum and minimum
normal stresses developed on the reinforcement-matrix in-
terfaces in the clustering and non-clustering regions.
clustering and non-clustering regions. It is obvious from
Figure 10 that the maximum and minimum normal
stresses developed on the reinforcement-matrix interface
in the clustering regions are extensively higher than that
of the non-clustering regions. The normal stresses de-
veloped in the SiC particle and Al-matrix in the particu-
late clustering region (Model-1) is relatively lower than
that of the hybrid clustering regions (Model-2 and
Model-3). Moreover, from Figure 10, it is significant
that the maximum normal stress on the particle-matrix
interface in the reinforcement non-clustering region
where SiC particle and Al2O3 whisker is placed in series
(Model-4) are reasonably higher than those of the non-
clustering region of Model-5, where SiC particle and
Al2O3 whisker is placed in parallel orientation. Thus, it
can be concluded that the reinforcements located in the
clustering region experience higher stress than that of the
non-clustering region and stress concentration at the in-
terface of reinforcement-matrix is very high in the clus-
tering region. In elastic state this clustering effect occurs
because the ceramics have elastic stiffness one order
higher than that of the Al alloy. In low cycle fatigue,
elastic deformation occurred in the reinforcing SiC parti-
cles and Al2O3 whiskers whereas the matrix alloy de-
formed plastically during cyclic loading. As the rein-
forcements did not experience plastic deformation, the
stress on the particle-matrix or whisker-matrix interfaces
was higher in the hybrid MMC. In addition, the edge of
the stiff ceramic reinforcements acted as stress concen-
trators that localizing the plastic strain between the parti-

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
105
cles and the whiskers. Thus, a large strain mismatch oc-
curred between these two reinforcement materials and
the Al alloy. For this large strain mismatch, the stress
became too high on the particle-matrix interface, and
cracks initiated at these locations. Moreover, it can also
be concluded that the stress concentration at the rein-
forcements has the dependency on the reinforcement
volume fraction as well as at the reinforcement-matrix
interface, indicating the vulnerability of the reinforce-
ment cluster on fatigue crack initiation.
5. Conclusions
The effect of reinforcement clustering on crack initiation
mechanism in cast hybrid MMC reinforced with SiC par-
ticles and Al2O3 whiskers was investigated experimen-
tally and numerically. The following conclusions were
made:
1) The experimental results showed that the crack ini-
tiated at the particle-matrix interface which was located
in the cluster of the reinforcements in the cast hybrid
MMC. The interface debonding occurred in the fracture
which created additional secondary microcracks due to
continued fatigue cycling. Numerous voids were formed
ahead of the crack tip and the microcrack coalesced with
other nearby microcracks;
2) The numerical analysis confirmed that the high
stress was developed in the reinforcements located in the
clustering region and stress concentration occurred on the
particle-matrix interfaces. The stress concentration on the
particle-matrix interface located in the high volume frac-
tion reinforced hybrid clustering region is found to be
very high compared to that in the SiC particulate rein-
forced clustering region and low volume fraction rein-
forced hybrid clustering region. Moreover, the non-clus-
tering region where SiC particle and Al2O3 whisker lo-
cated in series experienced higher stresses than that of
the non-clustering region where SiC particle and Al2O3
whisker located in parallel orientation;
3) The high volume fraction reinforced hybrid cluster-
ing region is found to be highly vulnerable to initiate
crack in the cast hybrid MMC during low cycle fatigue.
6. Acknowledgements
The authors express gratitude to the Ministry of Educa-
tion, Science, Sports and Culture of the Government of
Japan for providing financial support during this research
work.
REFERENCES
[1] S. Suresh, A. Mortensen and A. Needleman, “Fundamen-
tals of Metal Matrix Composites,” Butterworh/Heine-
mann, London, 1993.
[2] J. Hemanth, “Quartz (SiO2p) Reinforced Chilled Metal
Matrix Composite (CMMC) for Automotive Applica-
tions,” Materials and Design, Vol. 30, No. 2, 2009, pp.
323-329. http://dx.doi.org/10.1016/j.matdes.2008.04.064
[3] D. B. Miracle, “Metal Matrix Composite—From Science
to Technological Significance,” Composite Science and
Technology, Vol. 65, No. 15, 2005, pp. 2526-2540.
http://dx.doi.org/10.1016/j.compscitech.2005.05.027
[4] V. K. Lindroos and M. J. Talvitie, “Recent Advances in
Metal Matrix Composites,” Journal of Materials Proc-
essing Technology, Vol. 53, No. 1, 1995, pp. 273-284.
http://dx.doi.org/10.1016/0924-0136(95)01985-N
[5] S. Basavarajappa, G. Chandramohan and D. J. Paulo, “Ap-
plication of Taguchi Technique to Study Dry-Sliding
Wear Behaviour of Metal Matrix Composites,” Materials
and Design, Vol. 28, No. 4, 2007, pp. 1393-1398.
http://dx.doi.org/10.1016/j.matdes.2006.01.006
[6] N. Chawla, C. Andres, J. W. Jones and J. E. Allison, “Ef-
fect of SiC Volume Fraction and Particle Size on the Fa-
tigue Resistance of a 2080 Al/SiCp Composite,” Metal-
lurgical Materials Transactions A, Vol. 29, No. 11, 1998,
pp. 2843-2854.
[7] A. Ayyar, G. A. Crawford, J. J. Williams and N. Chawla,
“Numerical Simulation of the Effect of Particle Spatial
Distribution and Strength on Tensile Behavior of Particle
Reinforced Composites,” Computational Materials Sci-
ence, Vol. 44, No. 2, 2008, pp. 496-506.
http://dx.doi.org/10.1016/j.commatsci.2008.04.009
[8] J. E. Spowart, B. Maruyama and D. Miracle, “Multi-Scale
Characterization of Spatially Heterogeneous Systems:
Implications for Discontinuously Reinforced Metal Ma-
trix Composite Microstructures,” Materials Science and
Engineering A, Vol. 307, No. 1, 2001, pp. 51-66.
http://dx.doi.org/10.1016/S0921-5093(00)01962-6
[9] S. F. Corbin and D. S. Wilkinson, “The Influence of Par-
ticle Distribution on the Mechanical Response of Particu-
late Metal Matrix Composite,” Acta Metallurgica et Ma-
terialia, Vol. 42, No. 4, 1994, pp. 1311-1318.
http://dx.doi.org/10.1016/0956-7151(94)90147-3
[10] D. L. McDanels, “Analysis of Stress-Strain, Fracture and
Ductility of Aluminium Matrix Composites Containing
Discontinuous Silicon Carbide Reinforcement,” Metal-
lurgical Transaction A, Vol. 16, No. 6, 1985, pp. 1105-
1115.
[11] X. Q. Xu and D. F. Watt, “A Numerical Analysis of the
Effects of Reinforcement Content on Strength and Ductil-
ity in Al(SiC)p MMCs,” Acta Materialia, Vol. 44, No. 11,
1996, pp. 4501-4511.
http://dx.doi.org/10.1016/1359-6454(96)00063-8
[12] D. L. Davidson, “The Effect of Particulate SiC on Fatigue
Crack Growth in a Cast Extruded Aluminum Alloy
Composite,” Metallurgical Transcations A, Vol. 22, No.
1, 1991, pp. 97-112.
[13] J. D. Boyd, R. D. Evans, G. Klein and S. Tao, “Damage
Accumulation and Fracture in Metal Matrix Composites,”
Proceedings International Conference on Composite Ma-
terials, Gold Coast, 14-18 July 1997, pp. 409-417.
[14] N. Bourgeois, K. Derrien and D. Baptiste, “Observation

Effect of Reinforcement Clustering on Crack Initiation Mechanism in
a Cast Hybrid Metal Matrix Composite during Low Cycle Fatigue
106
and Modeling of Damage Evolution in Particulate Rein-
forced MMC,” Proceedings International Conference on
Composite Materials, Whistler, 14-18 August 1995, pp.
241-248.
[15] P. B. Prangnell, S. J. Barnes, S. M. Roberts and P. J.
Withers, “The Effect of Clustering on Damage Formation
in Particulate Reinforced MMCs Deformed in Compres-
sion,” Key Engineering Materials, Vol. 47, 1996, pp.
937-944.
[16] A. M. Murphy and T. W. Clyne, “The Effect of Initial
Porosity and Particle Clustering on the Tensile Failure of
Cast Particulate MMC,” Proceedings International Con-
ference on Composite Materials, Whistler, 14-18 August
1995, pp. 35-42.
[17] H. N. Yoshimura, M. Goncalves and H. Goldenstein,
“The Effect of SiCp Clusters and Porosity on the Me-
chanical Properties of PM Al Matrix Composites,” Key
Engineering Materials, Vol. 127-131, 1996, pp. 985-992.
http://dx.doi.org/10.4028/www.scientific.net/KEM.127-1
31.985
[18] M. Li, S. Ghosh and O, Richmond, “An Experimental-
Computational Approach to the Investigation of Damage
Evolution in Discontinuously Reinforced Aluminium
Matrix Composite,” Acta Materialia, Vol. 47, No. 12,
1999, pp. 3515-3532.
http://dx.doi.org/10.1016/S1359-6454(99)00148-2
[19] T. C. Tszeng, “The Effect of Particle Clustering on the
Mechanical Behaviour of Particle Reinforced Compos-
ites,” Composites Part B, Vol. 29, No. 3, 1998, pp. 299-
308. http://dx.doi.org/10.1016/S1359-8368(97)00031-0
[20] W. X. Zhang, L. X. Li and T. J. Wang, “Interphase Effect
on the Strengthening Behaviour of Particle-Reinforced
Metal Matrix Composites,” Computational Materials Sci-
ence, Vol. 41, No. 2, 2007, pp. 145-155.
http://dx.doi.org/10.1016/j.commatsci.2007.03.011
[21] Y. W. Yan, L. Geng and A. B. Li, “Experimental and Nu-
merical Studies of the Effect of Particle Size on the De-
formation Behavior of the Metal Matrix Composites,”
Materials Science and Engineering A, Vol. 448, No. 1,
2007, pp. 315-325.
http://dx.doi.org/10.1016/j.msea.2006.10.158
[22] D. F. Watt, X. Q. Xu and D. J. Llyod, “Effect of Particle
Morphology and Spacing on the Strain Fields in a Plasti-
cally Deforming Matrix,” Acta Materialia, Vol. 44, No. 2,
1996, pp. 789-799.
http://dx.doi.org/10.1016/1359-6454(95)00209-X
[23] X. Q. Xu and D. F. Watt, “A Finite Element Analysis of
Plastic Relaxation and Plastic Accumulation at Second
Phase Particles,” Acta Materialia, Vol. 44, No. 2, 1996,
pp. 801-811.
http://dx.doi.org/10.1016/1359-6454(95)00210-3
[24] P. Zhang and F. Li, “Effects of Particle Clustering on the
Flow Behaviour of SiC Particle Reinforced Al Metal Ma-
trix Composites,” Rare Metal Materials and Engineering,
Vol. 39, No. 9, 2010, pp. 1525-1531.
http://dx.doi.org/10.1016/S1875-5372(10)60123-3
[25] Md. Rafiquzzaman and Y. Arai, “Fracture Mechanism of
Aluminium Cast Alloy Locally Reinforced by SiC Parti-
cles and Al2O3 Whiskers under Monotonic and Cyclic
Load,” Journal of Materials Science and Technology, Vol.
24, No. 3, 2008, pp. 273-280.
http://dx.doi.org/10.1179/174328407X239028
[26] A. A. Iqbal, Y. Arai and W. Araki, “Effect of Hybrid
Reinforcement on Crack Initiation and Early Propagation
Mechanisms in Cast Metal Matrix Composites during
Low Cycle Fatigue,” Materials and Design, Vol. 45,
2013, pp. 241-252.
http://dx.doi.org/10.1016/j.matdes.2012.09.002
[27] “Aluminium Alloy Castings,” JIS H5202, Japan Indus-
trial Standard, 2002. (In Japanese)
[28] E. Y. Chen, L. Lawson and M. Meshii, “The Effect of
Fatigue Microcracks on Rapid Catastrophic Failure in
Al-SiC Composites,” Materials Science and Engineering
A, Vol. 200, No. 1, 1995, pp. 192-206.
http://dx.doi.org/10.1016/0921-5093(95)07009-5
[29] Z. Z. Chen and K. Tokaji, “Effects of Particle Size on
Fatigue Crack Initiation and Small Crack Growth in SiC-
Particulate Reinforced Aluminium Matrix Composites,”
Materials Letters, Vol. 58, No. 17, 2004, pp. 2314-2321.
http://dx.doi.org/10.1016/j.matlet.2004.02.034
[30] ABAQUS 6.9, Dassault Systems, 2009. (Software)