Materials Science and Engineering A266 (1999) 8 – 17

Evaluation of heat damage on B4C particulate reinforced

aluminum alloy matrix composite using acoustic
emission techniques
H. Nayeb-Hashemi *, D. Shan
Department of Mechanical, Industrial and Manufacturing Engineering, Northeastern Uni6ersity, Boston, MA 02115, USA

Received 13 September 1998; received in revised form 11 January 1999

Abstract

The effects of prolonged heat damage on the tensile properties of aluminum alloys reinforced with B4C particle of average
particle size of 10.5 and 20.9 mm was investigated by both mechanical and acoustic emission monitoring during tensile test. To
ensure the matrix contribution was similar for all specimens, specimens were heat treated to the same condition for each material
after heat damage. The particle/matrix bond strength was found reduced significantly despite what was reported in the literature
of stable particle/matrix interface at temperature lower than 660°C. The micromechanisms of fracture changed from particle
fracture and debonding, for as received specimens, to particle/matrix debonding after exposure to high temperature. The
cumulative acoustic emission (AE) events during post heat exposure reduced with the increase of heat damage on the specimen.
A Weibull probability distribution function that incorporates heat damage and is related to the cumulative AE events is
developed. This model could be used to evaluate the heat damage by subjecting the heat damaged specimen to a tensile test while
collecting AE data up to a strain level o. The total events can then be used in connection with the model to predict the extent of
the heat damage to the particle/matrix interface. The results indicate that the tensile strength was little sensitive to the
particle/matrix bond strength. © 1999 Elsevier Science S.A. All rights reserved.

Keywords: B4C particulate reinforced aluminum alloy matrix composite; Heat damage; Acoustic emission

1. Introduction as pistons in car engines, it is important to predict the

heat damage and understand the associated damage
In recent years, aluminum matrix composites rein- mechanisms using a non-destructive evaluation tech-
forced by ceramic particulates have received wider ac- nique. Acoustic emission techniques are found to be
ceptance as structural materials in design of many very useful to monitor and collect information about
components. These materials have high specific strength the damage in materials.
and stiffness, low coefficient of thermal expansion and A strong interfacial bond between matrix and rein-
densities that are similar to that of matrices, which forcement enables stress transfer from the matrix to the
make them attractive materials for automotive and stronger reinforcement particles. Investigations have
aerospace industries. The most commonly used ceramic been carried out to understand the effects of interfacial
particulates as reinforcement, are SiC, B4C and Al2O3. reactions on the interfacial bond strength. In SiC parti-
While most research on these materials has been cle reinforced aluminum alloy composites, Al4C3 has
devoted to the properties of aluminum composites rein- been found to form when the composite is exposed to
forced with SiC and Al2O3, research on composites elevated temperature. B4C reacts with aluminum in a
reinforced with B4C is very limited. Since these materi- more complex manner at high temperature. Reports on
als are being used in high temperature applications such the reactivity of B4C in aluminum are not consistent
though.
* Corresponding author. Tel.: +1-617-3735515; fax: + 1-617- In a study of aluminum alloy 6061 reinforced with 23
3732921. vol.% B4C, Nieh and Karlak [1] found that B4C is

0921-5093/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S 0 9 2 1 - 5 0 9 3 ( 9 9 ) 0 0 0 3 1 - 3
 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17 9

chemically stable at the highest temperature of com- minum. Oh et al. [6,7] studied the wetting of SiC and
posite processing. Vaidya et al. [2] studied the mechani- B4C particles with different liquid aluminum alloys.
cal properties of SiC and B4C reinforced aluminum The wettablity of B4C with aluminum was found to
alloy 6061 in an annealed condition and reported that increase in air. This was attributed to the formation of
no reaction product was found at the particle/matrix a layer of liquid B2O3 on the B4C particulates. Due to
interface. However, in a study of Al– B – C phase devel- its low melting point, B2O3 exists on the surface of B4C
opment in B4C/Al composites, Pyzik and Beaman [3] above 450°C as a liquid and enhances wettablity
found that the formation of Al4BC in the composite through a liquid-liquid reaction when contacted with
starts at 450°C. Heat treatments above 600°C produce aluminum liquid. Alloy elements such as Mg were
multiphase ceramic materials including AlB2, Al4BC, found to be very effective in promoting wetting of B4C.
AlB24C4 and Al4C3. Viala et al. [4] studied the chemical The purpose of this paper is to understand the effect
reactivity of boron carbide with aluminum at tempera- of heat damages on the mechanical properties of B4C
tures ranging from 627 to 1000°C. It was found that the particulate reinforced aluminum alloy. The acoustic
reaction rate was very slow at a temperature lower than emission method was used to explore its feasibility in
660°C. In the sample heated for 160 h at 627 and monitoring and understanding the extent of heat dam-
647°C, no phases other than B4C and Al were found by age on B4C particulate reinforced aluminum matrix
XRD. However, when the duration of heat exposure composites non-destructively, and helping to identify
was extended to 450 h, small amounts of Al3BC and the damage mechanisms.
AlB2 were detected using a scanning electron micro-
scope. Similarly, an increase of heating temperature to
677°C resulted in the formation of Al3BC and AlB2 2. Experimental procedure
after 160 h of heat exposure.
The thermal mismatch produces high dislocation The materials tested were 2219-T4 and 6061-T6 alu-
densities in the vicinity of the interface. Nieh and minum alloys reinforced with B4C particulates. The
Karlak [1] found that the aging response of the matrix Army Material Research and Technology Laboratory
was enhanced by addition of B4C. The effect of the B4C in Watertown, MA provided the materials. The materi-
on aging was attributed to the thermal expansion differ- als were manufactured by ALCOA by blending atom-
ence between particulate and matrix. The high disloca- ized aluminum alloy powders with B4C particles,
tion density and interfacial stresses serve as short-circuit compacting by cold isostatic pressing, densifying to
paths for pipe diffusion that accelerate the aging pro- roughly theoretical density by degassing and vacuum
cess. In the study of Vaidya et al. [2], transmission hot pressing, and finally extruding into plates to thick-
electron micrographs of the B4C/6061 interface showed ness of 2.54 mm. The specified reinforcement volume
many dislocations tangles and cells in the matrix in the fraction and average particle sizes were 15 vol.% and
vicinity of B4C particulates, which indicated that the 10.5 mm, respectively for 2219-T4 aluminum alloy and
dislocation density was very high. 20 vol.% and 20.9 mm, respectively for 6061-T6 alu-
The mechanical properties of B4C and SiC reinforced minum alloy.
6061 aluminum have been compared by a number of An Instron 1350 Material Testing System was used
researchers [1,2]. The B4C reinforced alloy has been for the tensile tests. The specimens were of a dog-bone
found to be superior to that of SiC reinforced 6061 shape with cross sections of 6.35×2.44 mm and gage
aluminum in the strength, strain to failure in tension lengths of 25.4 mm, Fig. 1. Tensile strain was measured
and strain hardening. These are attributed to the strong with an extensometer and recorded with an AET 5500
bonding between B4C and 6061 aluminum alloy matrix. system as an analog parameter. AE activities were
This can be due to good wetting between B4C and monitored during the test. Tension tests were per-
aluminum, lack of interfacial products and higher parti- formed under stroke control at a loading rate of
cle strength. In the annealed condition, SEM of the 0.00254 mm s − 1. This slow loading rate was chosen in
tensile fracture surfaces of B4C/6061 composites indi- order to increase AE sensitivity and prevent AE event
cated that the micromechanism of the failure was pileup. It is possible that several simultaneous events
mainly due to particle/matrix debonding with very few appear as one or for several subsequent events to be
fractured particles. However, TEM micrographs of the missed if the loading rate is too high.
same composite treated to T6 condition showed a num- To stimulate the heat damage, three 2219-T4-
ber of fractured B4C particles (Vaidya and Zurek [5]). 15%B4C (10.5 mm) aluminum alloy composite speci-
The particle cracking was attributed to that of overall mens were heated in an air atmosphere oven at 535°C
stresses in the material exceeding the fracture stress of for three time intervals, 53, 114 and 163 h. Then the
the B4C particles. specimens were water quenched and tempered naturally
The interfacial strength of B4C/matrix is highly influ- to T4 condition. Four 6061-T6-20%B4C (20.9 mm)
enced by the wetting characteristics of B4C in alu- aluminum alloy composite specimens were heated in an
10 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17

Fig. 1. Schematic diagram of the tensile specimens and AE sensor placement, dimension in mm.

air atmosphere oven at 520°C for four time intervals, 3. Results and discussions
56, 80, 104, 162 h. Another 6061-T6-20%B4C (20.9
mm) aluminum alloy composite specimen was heated in The matrix microhardness of the as received speci-
a vacuum atmosphere at 520°C for 162 h. All of heated men and the heat damaged specimens were measured
6061-T6 specimens were water quenched and tempered using a TUCKON Microhardness Tester (Model 300)
at 165°C for 12 h to T6 conditions. Upon completion and BOECKLER VIA-110 Video Microhardness Sys-
of heat treatment the materials’ microstructure were tem. Ten Knoop hardness numbers were measured
evaluated using optical microscope and scanning elec- using a load of 10 g on each specimen and the average
tron microscopes. Here for the sake of clarity of discus- was taken. The results are shown in Fig. 3. The matrix
sion, we present the results of heat damage study on microhardnesses of the specimens with different heat
2219-T4-15%B4C (10.5 mm) aluminum alloy composite damages were found to have been relatively unaffected
specimens. However, the data from other materials has by the duration of heat exposure. This could be due to
helped us to better understand and effectively interpret subsequent aging procedure that was identical for all
the AE data from 2219-T4-15%B4C (10.5 mm) alu- the specimens. This insures that AE activity contribu-
minum alloy composite specimens. tions from the matrix material to be fairly identical
Acoustic emissions were monitored with an AET among various heat damaged specimens. The AE activ-
5500 computerized system. Two piezoelectric sensors ity of the matrix material could be due to movement of
with the center frequency of 175 kHz and sensitivity of
−80 dB (re 16/m Bar) was clamped to the gage section
of the specimens with a spring clamp system. The
sensors were coupled to specimens with a couplant.
Two 60 dB preamplifiers were used which with 10 dB
system amplification and band width of 10 kHz–1.5
MHz, provided the total system gain of 70 dB. Teflon-
rubber sheets were used between specimens and grips to
reduce the noise level from the system. The band pass
filters in the preamplifiers were 125 – 250 kHz and a
floating signal threshold setting of 0.3 V was used to
eliminate the background noise. Sensors were placed in
a linear location array. The tests were run in time-dif-
ference module TDM. The lead pencil fractures were
used to simulate AE signals in the calibration of each
test, according to the ASTM standard (ASTM E 976-94
[8]). A region between two sensors was specified from
which all AE data was collected and analyzed, Fig. 2.
The software rejected the other unwanted signals out- Fig. 2. Schematic diagram of the acoustic emission experimental
side of this region. setup.
 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17 11

These include the existence of prior fractured particle

produced during manufacturing of composites, the par-
ticle/matrix interface decohesion and particle fracture
due to sufficient local stresses at the particle/matrix
interface. The fracture of the particle initiates the crack
when the local stress is equal to or greater than the
particle strength. In this case, the strength of the rein-
forcement controls the strength of composite in the
stage of crack initiation. With the interfacial reaction
zone growing during heat exposure, the bond strength
is decreasing. Micro cracks may be initiated in the
low-strength reaction zone under the thermal stress or
local stresses. Therefore the debonding of particle/ma-
trix interface is more likely after heat exposure. In this
case, the particle strength will not contribute to the
overall strength of the material. The strength of com-
posite is dominated by matrix properties during the
void/crack linkage stage. In our experiments, the initial
decrease of the strength of the material observed is due
to the interface weakening by the heat damage on the
Fig. 3. Microhardness of matrix versus heating time. material. However, after the interface bond strength is
degraded to certain level, further exposure to high
dislocations in the matrix, fracture of inclusions and temperature will not influence the strength of the com-
void formation at the matrix/inclusions interface. Since posite significantly for the above stated reasons. This is
all heat damaged specimens were subjected to an identi- supported by the results of scanning electron mi-
cal heat treatment at the end of heat exposure duration, croscopy evaluation of the tensile fracture surfaces.
we expected their matrix microstructure to be identical. Both fractured and debonded particles were found on
Therefore the matrix contribution to the AE activity is the SEM micrographs of the fracture surfaces of as
expected to be similar among different specimens. received and 53-h heat damaged specimens, but only
Optical microscopy observation of the specimen mi- debonded particles were identified on the SEM mi-
crostructure with different heat damage levels showed crograph of the fracture surfaces of 114- and 163-h heat
no obvious grain growth after prolonged heat exposure. damaged specimens.
Ferry et al. [9,10] have shown the same results in the In the tensile test of this study, the number of
recrystallization studies of an aluminum alloy 2014 cumulative AE events increased exponentially after ma-
reinforced with 20 vol.% of Al2O3 particulates. After terial yielding. The results show that the cumulative
exposing the material for 150 h at 500°C, they found events decreased after specimens were heat damaged
the grain size only changed from 15 to 17 mm. Because
the particles in a composite with high volume fraction
are very closely spaced, the grain growth in the recrys-
tallization process is impeded and the grain size is
stabilized.
Fig. 4 shows the stress -strain curves of specimens
with different heat damage level. The result show that
the tensile strength of the material decreased after heat
damages (Table 1). After 53 h exposure to 535°C, the
strength of the material decreased from 441 to 372
MPa. Then the extended exposure time did not further
degrade the strength of the composite. Similar results
were obtained by repeating experiments.
Fig. 5 shows photomicrographs of tensile fracture
surface of as-received and heat damaged specimens.
The failure of the composite material can be character-
ized as two stages: void formation/crack initiation and
void linkage/crack propagation. Voids could be ini- Fig. 4. Stress – strain curves and cumulative events versus strain of the
tiated around particles due to several mechanisms. as-received specimen and specimens with different heat damage levels.
12 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17

Table 1
Residual tensile strength

Specimen condition As-received (MPa) 53 h heated (MPa) 114 h heated (MPa) 163 h heated (MPa)

Ultimate Strength 441 372 355 377

(Fig. 4). There are many suggested mechanisms for the tion increased upon material yielding and slows down
observed acoustic emission activity in aluminum alloys due to dislocation pinning. Furthermore, there was very
and particulate reinforced aluminum alloys [11–17]. little activity from unreinforced matrix when subjected
While Heiple et al. [12] and Scruby et al. [13,14] relate to the tensile test, Fig. 7.
the high rise in AE activity on the onset of material Based on the results from previous research and our
yielding to dislocation motion, Cousland and Scala experimental observations, the majority of the events
[15–17] and McBride et al. [18] relate it to fracture of should be attributed to the particle fracture and the
large inclusions in the material. Cousland and Scala failure of strong particle/matrix bond. Since exposing
studied AE activity of AA 7075, 7050, 2024, 2124 and material to high temperature weakened the particle/ma-
6061-T651, under both monotonic and cyclic loading, trix interface, it was expected that the number of
and found that the AE depend on the inclusion size and debonded particles to increase and the number of bro-
number of inclusions. In AA 6061-T651, which has ken particles to decrease. The SEM micrographs of
fewer and smaller inclusions than the others, fewer AE fractured surfaces showed these changes. This had a
activity was observed. Furthermore no AE was detected direct effect on the observed AE activities. This was
during compression deformation. McBride et al. [18] further confirmed by halting tensile tests prior to failure
did not detect any AE from inclusion-free AA 7075-T6 and polishing and studying the damage initiation and
during slow fatigue crack growth. Therefore the AE growth in the gage length. We observed both debonded
events were attributed to the inclusion fracture. particles as well as fractured particles. This shows the
With the same AE measurement equipment and AE activity is related to both particle/matrix debonding
parameter setup, earlier we studied the acoustic emis- and fracture.
sion activity of the aluminum alloy 6061 reinforced The interface region in a given composite is very
with 15 vol.% of SiC with average particle size of 5 mm important in determining the ultimate properties of the
and the same matrix reinforced with 20 vol.% of B4C composite. A strong interfacial bond is usually achieved
with average particle size of 20.9 mm. Comparing the by the formation of an adequately thin layer at the
result of three tests, we found that the particles size (5, interface under optimum wetting conditions of the
10.5 and 20.9 mm) and volume fraction (15 and 20 molten metal on the reinforcement. Addition of alloy-
vol.%) had a quite obvious affect on the AE activity ing elements, such as lithium and magnesium, has
(Fig. 6). Total number of cumulative events was about proved to be an effective method to enhance the wet-
700 for the composite with 15 vol.% of 5 mm particles tability of the reinforcement by the molten matrix.
and about 1100 for the composite with 15 vol.% of 10.5 Some of these alloying elements can react with the
mm particles, while the total events exceeded 30 000 for reinforcement to produce chemical reaction products at
the composite with 20 vol.% of 20.9 mm particles. The the interface. Carotenuto et al. [19] showed that the
fracture surface of the composite with 5 mm particle SiC/Al interface is chemically stable at temperatures
size was covered with voids and with a very few frac- below 650°C. At higher temperatures SiC reacts with
tured particles. Thus major AE source was attributed to aluminum leading to formation of Al4C3 and Si at the
particle/matrix interface debonding and linkage of particle/matrix interface. The reaction involves solution
voids. These results are in agreement with those re- of SiC, diffusion of silicon and carbon away from the
ported by Mummery et al. [11]. In a study of SiC interface into the matrix and reaction between dissolved
particle reinforced aluminum alloy composite, they carbon and aluminum with precipitation of Al4C3,
found that there was a critical particle size, 3 mm, to when the solubility is exceeded. The aluminum carbide
initiate AE activity in the composite. Then the AE forms as a discontinuous layer on the SiC substrate and
activity increased with the increase of particle size. free silicon precipitate in between the carbide crystal.
The AE activity of our specimens started well into Wang et al. [20] used electron microscopy techniques to
the plastic regime. Therefore the dislocation activity study the interface in cast aluminum alloy A356 rein-
could not be considered as an AE source for these forced with 15 vol.% SiCp. They observed formation of
group of specimens. The AE activity from the disloca- magnesium aluminate MgAl2O4, a spinel at the inter-
tions depends on the dislocation density and the dis- face. In addition, fractographic observation by Vedani
tance that dislocations can move. It has been shown et al. [21] on the same material indicated preferential
that the AE activity corresponding to dislocation mo- cracking of the reinforcement particles in the underaged
 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17 13

composites, whereas overaging led to more frequent for the cavity formation from inclusions. Critical local
interface debonding. The B4C exhibits better wettability elastic energy is found to be necessary but not sufficient
by molten aluminum than that of SiC, by formation of for cavity formation. The interfacial stress at the inclu-
liquid B2O3 on the surface of B4C particles. After sion/matrix interface must reach to the bond strength
exposing a 6061 aluminum alloy reinforced with B4C for the cavity formation. For particles larger than 100
particles to 535°C for a long period, based on the Å, the energy criterion is satisfied when the stress at the
previous published results [2 – 7], it is believed that a inclusion/matrix interface is equal to the bond strength
layer of Al4BC, Al2B and aluminum carbide forms, [22]. Since in our metal matrix composite the particles
which weaken the interface of B4C/Al. are larger than 100 Å, we can assume that the energy
criterion is satisfied. The tensile stress field at the inclu-
sion/matrix interface was obtained by Argon et al. [22]
4. Theoretical study in a study of ductile fracture of materials having rigid
inclusions. The tensile stress at interface of inclusion
Reinforcement/matrix interface debonding is a main and matrix is given by
damage mechanism of particulate reinforced metal ma-
trix composites. Many theoretical and experimental
investigations have been undertaken to find the criteria srr : s(op)+ sT (1)

Fig. 5. Scanning electron micrographs of the tensile fracture surfaces. Arrows indicate the locations of particle fracture or particle matrix bonding.
14 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17

(o − ou)n
−
P(o)= 1−e o0 , o] ou

P(o)= 0, oB ou (2)
where ou is the minimum strain to initiate debonding
and o0 is a normalization factor.
Many investigators have found that the increase in
reaction zone size reduces the bond strength between
matrix and reinforcement considerably [23]. The thick-
ness, x, of the reaction products after exposing to
temperature T for duration of t can be found from the
classical diffusion equations as
Q
−
x=a
te 2kT (3)

Fig. 6. Number of cumulative AE events versus strain for composites where a is a constant, Q is the activation energy, k is
with particle size of 5, 10.5 and 20.9 mm. the Boltzmann’s constant.
In the study of the effect of fiber/matrix reaction
where s(op) is the flow stress due to the local plastic zone on the interfacial bond strength, Hashemi and
strain in the region of particles and sT is the negative Seyyedi [24] showed that there was virtually little or no
pressure in the neighborhood area of the particles. The mechanical bonding between fiber and matrix, after the
stress field around the particles is not necessarily equal specimen was exposed to 545°C for 1 week.
during a tensile test. Argon et al. [20] have shown that Defining the reaction zone size which causes a signifi-
while the interfacial stress is independent of the particle cant degradation of the particle/matrix bond strength
size at an applied plastic strain of o for small particle as x0,
volume fraction, the interfacial stress field is expected to −
Q

vary for larger volume fraction of inclusions due to x0 = a

t0e 2kT (4)
their interactions. This may lead to the inclusion/matrix
debonding at different strain levels. we can define the damage h to the interfacial bond
Even though it is difficult to find the exact relation strength on a specimen after its exposure to tempera-


ture T for a period of time t as
between local stress field and far field tensile strain, we
m
know that the variation in particle size, orientation and x
h= b* , (5)
local volume fraction leads to the variation in the local x0
strain and stress field at the particle/matrix interfaces. Constant m reflects the non-linear relation between the
Assuming that this variation obeys Weibull distribu- interface reaction zone growth and the damage to the
tion, the probability of a particle debonding at an bond strength. b* is a normalization constant. Substi-
applied far field strain level of o can be defined as tuting for x and x0, damage h can be expressed as
m
h= bt 2 (6)

where b= b*/t (m/2)

0 is a constant related to activation
energy Q, temperature T, and time limit t0, which
normalizes h into a range between 0 and 1.
During the high temperature exposure, the bond
strength between particle and matrix is weakened.
Defining the bond strength before high temperature
exposure as si, the residual bond strength after speci-
men subjected to high temperature exposure could be
expressed as
s*=
i (1− h)si (7)
Substituting for h, results in a bond strength after time

 
t of
Fig. 7. Stress and cumulative events versus strain for 6061-T6 alu- m

minum alloy. s*=

i 1− bt 2 si (8)
  
H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17 15
m q
c6g
Because the bond strength of particles is weakened Npt = 1− bt 2 (14)
6p
during high temperature exposure, it is easier to initiate
debonding during subsequent tensile test. The minimum In the tensile test following high temperature expo-
strain level, o*u , to initiate AE activity on the specimen sure experiment, the number of particles debonded at
subjected to high temperature during subsequent tensile strain o, can be expressed as
test, is less than ou of the as received specimen (with no
Nd = NptP(o) (15)
heat damage). This assumption has been proved with
experimental results. The strain levels that AE activities Assuming that the number of AE events during
started were 0.050, 0.039, 0.019 and 0.011% for as tensile test, Nae is proportional to the number of the
received, 53-h heat damaged, 114-h heat damaged and particles debonded, the cumulative events can be writ-
163-h heat damaged specimen, respectively. Relating ten as
this reduction to the heat damage on the specimen, a
Nae = kNd (16)
relation between o*u and ou could be expressed as
 m
 Substituting for Nd and using Eqs. (10), (14) and (16)

< =
o*u = 1− bt 2 ou (9) can be stated as

After the bond strength weakened during high tem-

Nae = k
c6g  m
1− bt 2
q
1− e
−

(o − ou
*)n
m r
 (17)
perature exposure, the probability of a particle debond- 6p 1 − bt 2 o0

ing at a certain strain level during subsequent tensile

test will increase, compared to that of as received
material. Probability of particle debonding after high Substituting from Eq. (9) into Eq. (15), the cumulative
temperature exposure can be defined by modifying Eq. AE events at strain o can be given as
o − 1 − btm2 ounnÂ
(2) using Eq. (8).
c6g 
1− bt
m q
 Á
Ã
Í1− e
−
1 − btm2 ro0 Ì
Ã

(o − ou

*)n
− Nae = k 2 (18)
P(o)= 1 − e m r
(10) 6p
1 − bt 2 o0 Ã Ã
Ä Å
In order to evaluate the above model, number of
The related probability density function is: constants in the equation should be evaluated. These

(o − ou

*)n constants could be obtained either from the acoustic
−
n(o− o*)
u
n−1
e 1 − bt 2
m r
o0 emission activities of a single test at different strain
levels or from the AE results of multiple tests on
f(o)=
 1−bt

m r
2 o0
(11) specimens with different damage level. In a study of
boron fiber/titanium composite, it was found that the
fiber fracture strain is proportional to 1/
x [23]. So the
For a material reinforced with an average particle constant m is chosen to be 0.5. For our material to be
size of D, volume fraction of c and particle volume of exposed to solution treatment temperature 535°C, we
6p, the number of particles in the gauge section of define our longest exposure time, 163 h, as one half of
specimen used to pick up AE activity can be expressed the time interval which causes the significant bond
as strength degradation. The time t when the heat damage
h equals to 1 is therefore chosen to be 326 h and the
Np =
c6g
(12) constant b is evaluated to be 0.2353 from Eq. (6).
6p Further research in form of a detailed metallographic
study is needed to justify the selection of time interval
where 6g is the volume of the gauge section of the for complete bond strength degradation. Based on the
specimen. preliminary study, we believe the 326 h is a reasonable
During exposure to high temperature, some of the estimate at this time. Using average particle volume of
particles may lose their bond strength with the matrix 6p = 523.6 mm3 and composite volume fraction of c=
due to reaction growth and developed thermal residual 0.15, the number of particles in the gauge section Np is
stresses upon cooling. After exposure to high tempera- found to be 1.127 × 108. The constants k, n, o0, and ou
ture, the number of particles left without debonding, are obtained using the AE results of as received mate-
Npt, can be expressed as a function of heat damage as rial. These constant are found to be 9.649× 10 − 6, 2.2,
Npt = Np(1−h)q (13) 14 and 0.05, respectively. The constant q and r are
obtained using the AE result from 163-h damaged
where q is a constant. Substituting for h from Eq. (6) specimen. These constants are found to be 0.755 and
into Eq. (13) results in 0.2, respectively.
16 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17

Fig. 8. (a) Probability and (b) probability density of a particle Fig. 9. Model prediction for 53 h heat damaged specimen (a) cumula-
debonding at different strain level for 53 and 163 h damaged speci- tive events versus strain; (b) AE events rate versus strain.
men.

Eqs. (10) and (11) are plotted for different heat collected. For an unknown heat damaged specimen, the
damage level, Fig. 8. Fig. 8 shows that at an applied cumulative events at strain o should be an indication of
strain level o the probability of particle debonding on heat damage. The cumulative events and strain o can be
heat damaged specimen is higher than that of as re- substituted in to Eq. (18), and Eq. (18) can be solved
ceived material. Furthermore the mean far field strain for heat damage h (or exposure time t, h= bt m/2).
for debonding decrease with the increase of exposure
time.
The cumulative events versus strain for as received,
53-, 114- and 163-h heat damaged specimens were
plotted and compared with the experimental data. Figs.
9 and 10 are the comparison for 53- and 114-h dam-
aged specimen. The results show a good agreement
between the theoretical model and experimental data in
most of the region. Even though the model is not
sensitive enough for estimating heat damage on the
material applying small tensile strain (proof test), the
results indicate of possible usage of the model to predict
the extent of the heat damage to the particle/matrix
interface, if specimen can be subjected to higher strain
level while detecting and collecting acoustic emission
data. To use the Eq. (18) to predict the extent of the
heat damage to the particle/matrix interface, the con-
stants in the equation should be found from prior
tensile test on specimens subjected to a known heat
damage duration and their AE activities data collected
during tensile tests. To predict the extent of the heat
damage to the particle/matrix interface in a specimen,
heat damaged material is subjected to a tensile strain o,
which is not necessarily the failure strain of the speci- Fig. 10. Model predication for 114 h heat damaged specimen; (a)
men, and cumulative acoustic emission events Nae is cumulative events versus strain; (b) AE events rate versus strain.
 H. Nayeb-Hashemi, D. Shan / Materials Science and Engineering A266 (1999) 8–17 17

5. Conclusions References

Particle reinforced aluminum matrix composites have [1] T.G. Nieh, R.F. Karlak, Scripta Metall. 18 (1984) 25–28.
[2] R.U. Vaidya, S.G. Song, A.K. Zurek, Philos. Mag. A 70 (1994)
been considered for many structural applications. In
819 – 836.
these applications, they may be subjected to prolonged [3] A.J. Pyzik, D.R. Beaman, J. Am. Ceram. Soc. 78 (1994) 305–312.
heat damage, which may degrade their mechanical [4] J.C. Viala, J. Bouix, G. Gonzalez, C. Esnouf, J. Mater. Sci. 32
properties. Aluminum 2219-T4 reinforced with B4C was (1997) 4559 – 4573.
subjected to 535°C for various duration. It was found [5] R.U. Vaidya, A.K. Zurek, J. Mater. Sci. Lett. 15 (1996) 385–387.
[6] S.-Y. Oh, J.A. Cornie, K.C. Russell, Metall. Trans. A 20A (1989)
that despite what is believed to be stable interface
527 – 532.
between B4C and aluminum, the experimental results [7] S.-Y. Oh, J.A. Cornie, K.C. Russell, Metall. Trans. A 20A (1989)
indicated that there was significant degradation of par- 533 – 541.
ticle/matrix interfacial strength. The tensile strength of [8] ASTME 976-94, Standard Guide for Determining the Reproduci-
the material decreased after 53 h exposure to 535°C. bility of Acoustic Emission Sensor Response, ASTM, Philadel-
phia, PA 19103.
Then the extended exposure time did not further de-
[9] M. Ferry, P.R. Munroe, A. Crosky, Scripta Metall. Mater. 28
grade the strength of the composite. The micromecha- (1993) 1235 – 1240.
nism of failure changed from particle fracture to [10] M. Ferry, P.R. Munroe, A. Crosky, Scripta Metall. Mater. 29
particle/matrix debonding with increase in the heat (1993) 741 – 746.
exposure duration. [11] M. Mummery, B. Derby, C.B. Scruby, Acta Metall. Mater. 41
(1993) 1431 – 1445.
The AE activities from the particulate reinforced
[12] C.R. Heiple, S.H. Carpenter, M.J. Carr, Metal Sci. 15 (1981)
metal matrix composites were found to be affected by 587 – 598.
the particle size and volume fraction. The number of [13] C.B. Scruby, H.N.G. Wadley, K. Rushbridge, D. Stockham-
cumulative AE events during tensile test from the com- Jones, Metal Sci. 15 (1981) 599 – 609.
posite reinforced with 20 vol.% of 20.9 mm B4C parti- [14] C.B. Scruby, H.N.G. Wadley, Philos. Mag. A 44 (1981) 249–274.
[15] S.McK. Cousland, C.M. Scala, Metal Sci. 15 (1981) 609–614.
cles was over 30 000, while it was only about 700 and
[16] S.McK. Cousland, C.M. Scala, Mater. Sci. Eng. 57 (1983) 23–29.
1100 from the composites reinforced with 15 vol.% of 5 [17] S.McK. Cousland, C.M. Scala, J. Mater. Sci. Lett. 3 (1984)
mm SiC particles and 10.5 mm B4C particles, 268 – 270.
respectively. [18] S.L. McBride, J.W. MacLachlan, B.P. Paradis, J. Nondestructive
A Weibull probability distribution function that in- Evaluation 2 (1981) 35 – 41.
[19] G. Carotenuto, A. Gallo, L. Nicolas, J. Mater. Sci. 29 (1994)
corporates heat damage and relates it to the cumulative
4967 – 4974.
events is developed. This model could be used to evalu- [20] N. Wang, Z. Wang, G.C. Weartherly, Metall. Trans. 23A (1992)
ate the heat damage by subjecting the heat damaged 1423 – 1430.
specimen to a tensile test while collecting AE data up to [21] M. Vedani, E. Gariboldi, G. Silva, C. DiGregorio, Mater. Sci.
a strain level o. The total events can then be used in Technol. 10 (1994).
[22] A.S. Argon et al., Metall. Trans. A 6A (1975) 825.
connection with the model to predict the extent of the
[23] M. Taya, R.J. Arsenault, Metal Matrix Composites, Thermome-
heat damage to the particle/matrix interface of the chanical Behavior, Pergamon Press, NY, 1989.
material. [24] H.N. Hashemi, J. Seyyedi, Metall. Trans. A 20A (1989) 727–739.