Anda di halaman 1dari 10

JOURNAL OF MATERIALS S C I E N C E 27 (1992) 5 4 5 3 - 5 4 6 2

Failure behaviour of particulate-reinforced


aluminium alloy composites under uniaxial
tension
J. C. LEE, K. N. S U B R A M A N I A N
Department of Metallurgy, Mechanics and Materials Science, Michigan State University,
East Lansing, MI 48824, USA

Tensile tests were carried out at room temperature on 6061-aluminium alloy reinforced with SiC
and AI203 particulates. Although a significant increase in strength could be achieved by
introducing ceramic reinforcements into the aluminium alloy matrix, it is associated with
a substantial decrease in fracture strain. In order to understand the reason for the inferior ductility
of such composites, analytical solutions were obtained using a simple composite model. SEM
studies were carried out on the side surfaces of the fractured specimens to verify the proposed
failure behaviour. Failure modes observed to operate in such composites under uniaxial tension
are described.

1. I n t r o d u c t i o n number of cracked SiCp and debonded interfaces were


The addition of moderate amounts of SiC particulate counted at the fracture surface. From this analysis, the
(SiCp) and AI2Oa particulate ((A1203)p), usually less number of cracked SiCp was found to be more than
than 30 vol %, into molten aluminium alloy has been twice the number of the debonded interfaces. Exten-
found to result in significant increases in the strength sive plastic deformation in the matrix between SiCp
and elastic modulus [1-7], fatigue resistance [8-10], was also observed from the side surfaces of the tensile
and wear resistance [11], as well as improved high- specimen, at regions adjacent to the fracture surface.
temperature properties [12-14] of the composites. Based on the above observations, they proposed the
However, it also results in substantial decrease in the matrix failure of the composite due to nucleation,
ductility and, consequently, the fracture toughness of growth and coalescence of voids as a dominant failure
the composites. This has been the main drawback for mechanism of the SiCp/A1 composites. As a conse-
the wide use of such metal matrix composites rein- quence, the cracking of SiCp and debonding of
forced with ceramic reinforcements. Several studies SiCp/A1 interface were attributed to the matrix failure.
have been carried out to improve the ductility and the Based on the examination of fractographs of pulled-
fracture toughness of ceramic-reinforced metal matrix out SiCw that are coated with the matrix material,
composites by modified processing techniques [3, 8, Christman et al. [9, 25] have proposed ductile failure
15 20]. Large differences in the properties between in the matrix near the SiCw as one of the operating
the reinforced and the unreinforced aluminium alloys failure mechanisms of such composites.
still exist due to large differences in the properties of Janowski and Pletaka [26] had studied interracial
the matrix and the reinforcements. Poor ductility and microstructures of aluminium alloy composites rein-
fracture toughness of such metal matrix composites forced with SiCp and (A1203)p and their effect on the
are attributed to the operative failure mechanisms. tensile properties of such composites. Failure modes
Nutt and Duva [21], in their in situ TEM studies on present in particulate-reinforced composites were ob-
SiC whisker-reinforced aluminium matrix composites served to change depending on the characteristics of
(SiCw/A1 composites), observed that the void nu- the interfacial microstructures. Amorphous reaction
cleated at the corner of the whisker ends and grew layers formed at the (A1203)p/At interface were found
towards the centres of the whisker ends. However, the to degrade interfacial bonding, and thereby result in
void formation along the long sides of the SiCw/A1 inferior strength and ductility of such composities.
interface, which are parallel to the tensile direction, The purpose of the present investigation was to
was not observed. Based on the above experimental study the general failure behaviour of particulate-
results, void nucleation at the SiCw/A1 interface (i.e. reinforced metal matrix composites under uniaxial
interracial debonding) was proposed as the important tension.
failure mechanism in the SiCw/A1 composites
[21-23].
You et al. [24] proposed another failure mechanism 2. E x p e r i m e n t a l p r o c e d u r e
of the SiCp/A1 composite on the basis of the observa- The materials used in this study, 6061 aluminium
tions on the tensile fracture surfaces. In their study, the alloys reinforced with 10v01% SiCp and (A1203)p,

0022-2461 9 1992 Chapman & Hall 5453


were obtained as extruded cylindrical bars. Sheet present study. Most of the cracks were formed at SiCp
tensile specimens were machined with the tensile di- and (AlzO3)p in the form of interfacial debonding and
rection oriented both parallel (longitudinal) and per- particulate cracking, or sometimes in the form of duc-
pendicular (transverse) to the extrusion direction, and tile failure near the reinforcement. Although severe
polished with abrasive papers and rotating laps. They plastic deformation can be observed in the matrix near
were then T6 heat treated according to the conditions the region of cracked SiCp, particulate cracking was
listed in Table I. The oxidation layer formed during always found to precede the matrix failure. The micro-
heat treatment was removed by polishing with graphs exhibiting the initiation of cracks in S i C p and
diamond abrasives. (AlzO3)p in the tensile specimens are shown in Fig. 3a
Tensile tests were carried out with an Instron oper- and b.
ated at a constant crosshead speed of 0.1 cm rain-1 in
air at room temperature. The fracture surface and the
side surfaces of tensile specimens were examined by
4. Theoretical background
optical and scanning electron microscopic techniques.
Most three-dimensional engineering problems can be
analysed using a two-dimensional approach, because
most failures are initiated at free surfaces, where the
3. Results
largest stresses develop. Consider a large thin plate
Some of tensile properties of the reinforced and
having a circular inclusion, whose elastic constants
the unreinforced aluminium alloy are presented in
and thermal expansion coefficient are different from
Table II.
those of the matrix. Uniform uniaxial loading is ap-
Although considerable increase in the elastic
plied at infinity on the composite system as shown in
modulus and strength result due to reinforcements,
Fig. 4a. One can solve this problem by superposing
they are accompanied by substantial decrease in the
the stress function due to uniaxial loading and stress
fracture strain. In order to understand the reasons for
function due to the inelastic strain caused by thermal
the inferior ductility of the composite, studies were
expansion mismatch as in Fig. 4b. Although the solu-
carried out tO characterize the interfacial debonding
tions for the problems given in Sections 4.1 and 4.2
and particulate cracking. A significant number of de-
can be found elsewhere [27], detailed calculation steps
bonded interfaces could be observed from the subsur-
with slightly different methods will be presented in
face of the fractured specimens, as shown in Fig. 1, for
order to utilize the intermediate solutions to the
which deep etching was carried out. Such interfacial
analysis.
debondings were formed in a direction perpendicular
to the tensile loading as shown in Fig. lb and c.
Finely polished tensile specimens were prepared
4.1 Large plate having a circular inclusion
and tested with an Instron to observe the particulate
with different elastic constants subjected
cracking and possible matrix failure at the composite
to uniaxial tension
surface. Observations made on the side surfaces of the
Consider a plate having a small circular elastic inclu-
fractured tensile specimens, especially in regions ad-
sion of radius a, which is subjected to uniaxial tension.
jacent to fracture, revealed significant amounts of
Under such conditions, the boundary conditions at
microcracks at SiCp, as shown in Fig. 2. Microcracks
infinity (r = oo) are given by
initiated from the matrix were not observed in the
(7,r(oo, O) = (7o(1 + cos20)/2 (la)

TABLE I T6 heat-treatment condition used in the present study (7,0(00,0) = - (7osin20/2 (lb)
(7o0(0% 0) = (7o(1 - cos20)/2 (lc)
SiCp/A1 composite (AI203)p/A1 composite
The Airy stress function, q), for this case can be given
Solution treatment 530~ 70 min 560 ~ 60 min
as a linear combination of function of polar coordin-
RT ageing 24 ~ 48 h 24 ~ 65 h
Artificial ageing 200 ~ 10 h" 170 ~ 14 h b ates r and 0. Matrix part of this function, qbm', is
(7 0
a Overaged condition. ~ml(r,O) = ~-r2(1 -- cos20)
b Peak-aged condition.

T A B L E II The tensile properties of T6 heat-treated reinforced


and unreinforced aluminium alloy
+~-
I AaZlogr + Ba2cos20

Ca4(COS 20"~]
Material Direction" E(GPa) %s(MPa) r e~(%)
+ \ ,.2 }/ (2)

SiCp/A1 T - 269.5 323.0 2.8 where A, B, and C are constants, and superscript ml
L - 302.0 368.6 5.3 denotes matrix under applied uniaxial loading. One
(A1203)p/AI T 80.6 b 285,0 345.2 2.5 can notice that the first term in Equation 2 describes
L 79.9 b 301.5 364.5 9.0
the undisturbed field (when there is no inclusion in the
6061 AI 68.0 b 265.5 310.5 20.0
matrix), and the last three terms describe the local
a T, transverse direction; L, Longitudinal direction. disturbance due to the discontinuity (i.e. inclusion) in
b Values measured by sonic resonance method. the elastic medium. However, according to Saint

5454
Figure ] (a) Scanning electron micrograph showing crack development in a tensile specimen of SiCp/Al composite (etched with HCl to reveal
the subsurface region). (b) Interfacial debonding and particulate cracking. Note the crack propagation into the matrix in front of the
particulate crack. (c) Joining of particulate crack and debonded interface.

Venant's principle, the d i s t u r b a n c e caused by discon-


tinuity will be negligible at distances which are larger
c o m p a r e d to the r a d i u s of the discontinuity.
T h e Airy stress function for the inclusion can be
given as

9 i~(r, 0) = ~ - DrZ+Er2cos20+~Sr4cos20 (3)

where D, E a n d F are constants, a n d the superscript i 1


d e n o t e s inclusion u n d e r a p p l i e d uniaxial loading. T h e
stress c o m p o n e n t s , olj, can be o b t a i n e d directly from
the g i v e n A i r y stress functions, a n d the c o r r e s p o n d i n g
strain, eli, a n d d i s p l a c e m e n t c o m p o n e n t s , ui, can be
d e t e r m i n e d from the H o o k e ' s law a n d s t r a i n - d i s p l a c e -
m e n t relationships, respectively. Thus, the resultant
stresses a n d d i s p l a c e m e n t s in the m a t r i x a n d the
Figure2 Scanning electron micrograph of the fractured SiCp/A1 inclusion are
composite showing void formation due to joining of opened cracks,
indicated by an arrow. Note that cracks are formed perpendicular m t
CYrr = (CY0/2) [1 + AaZ/r 2
to the tensile direction and the number of particulate cracks is
significantly more than the debonded interfaces. + (1 -- 2Ba2/r 2 -- 3Ca4/r4)cos20] ~ (4a)

5455
U~nl = (O'0/S].tm)[ - 2r - B ( K m -- 1)a2/r
+ 2Ca4/r 3] sin20 (4e)
o~ = (Cro/2)(D -- E cos20) (4f)
il
at0 = (Cro/2)(E + 3FrZ/a 2) sin20 (4g)
it
~oo = (%/2)[D + (E + 6FrZ/a2)cos20] (4h)
u~' = (Oo/8g i) {D(K i -- 1)r
- [2Er - F(K i -- 3)r3/a 2] cos20} (4i)
u~ 1 = (Oo/8p i) [ 2 E r + f ( ~ i + 3)r3/a 2] sin20 (4j)

In order to determine the constants in E q u a t i o n 4,


a p p r o p r i a t e b o u n d a r y conditions for the composite
system should be set up at the interface, assuming
perfect interracial bonding.
Because the stresses and the displacements have to
be continuous at the interface, b o u n d a r y conditions at
the matrix-inclusion interface (r = a) are
o,,m 1 (a, O) i' , 0 ) ,
-= CYrr(a i'
o,0ml (a, 0) = Cy,o(a, O) (5a)
u, (a, O) = uri 1 (a, 0),
mI ml
u0 (a, O)
= il
u0 (a, 0) (5b)
Solving the b o u n d a r y conditions, one can obtain the
constants as
A = [-(K m -- 1) - F(K i -- l)]/

[F(~c i - 1) + 2] (6a)
B = 2(F - 1)/(F + ~:m) (6b)
C = (1 -- F)/(I ~ -I- Ic m ) (6c)

O = [-K m --t- 1 ] / [ - r ( K i - 1) -~- 2 ] (6d)

E = -- (lC m -}- 1 ) / ( F -I- K m ) (6e)

F = 0 (6f)
where F = g m / f

4.2. Large plate having a circular inclusion


with different thermal expansion
coefficient
If the system involves inelastic strain caused by
thermal expansion coefficient mismatch, it will pro-
duce an inelastic stress on the body. Some i m p o r t a n t
results will be listed without derivation, because the
exact solutions for this p r o b l e m can be found else-
where [27]. The compressive stress, - P , at the inclu-
Figure 3 (a) Particulate cracks in SiCp/A1 composite which are sion/matrix interface, resulting from thermal expan-
opened up due to tensile loading. Arrows indicate the initiation of
sion mismatch, is obtained as
the crack propagation into the matrix. (b) Particulate cracks and
interfacial debonding in (AlzO3)p/A1 composite caused by tensile P = [4gmgi(1 + q ) A ~ A T ] /
loading. A, interracial debonding; B, particulate cracking; C, matrix
adhering to (A1203) p. [2g i + (K i - 1)gin] (7)
Stress c o m p o n e n t s for the matrix and inclusion are
CYrn~' = ( - - CY0/2 ) [1 4- B a Z / r 2 determined as
+ 3Ca4/r43 sin 20 (4b) ~rrm 2 (r, O) = -- PaZ/r 2 (8a)
o~ l = ( % / 2 ) [1 - AaZ/r 2 1112
cYoo (r, O) = Pa2/r z (8b)
m2
- (1 - 3Ca4/r 4) cos20] (4c) o~0 (r, 0) = 0 (8c)
u m~ = (Cyo/81am) {(K m - 1)r - 2Aa2/r cy~(r, 0) = -- P (8d)
+ [-2r + B(~: m + 1)a2/r cy~;(r, 0) = - P (8e)
+ 2Ca4/r s] cos 20} (4d) cr~;(r, 0) = 0 (8f)
5456
%

++lf++
,1r

r162162162162162
1o) %

% %

++++++ ++++++
+<mi.tm

~
x k'

AT &r

+
@
r 4, z

r162162162162162
r162162162162
if0 (TO
(b)
Figure 4 (a) Two-dimensional composite model; a large thin plate having a circular inclusion with different elastic constants, (~c,Ix), and
thermal expansion coefficients, c(, subjected to uniaxial tension, c%, where la is the shear modulus, c( is the thermal expansion coefficient,
l~ = (3 - 4v) for plane strain, • = (3 - v)/(1 + v) for plane stress. (b) Schematics drawings of the superposition of stresses caused by external
loading and thermal expansion coefficient mismatch.

5457
where superscripts m 2 and i 2 represent matrix and Normatizedstress, ~x/%
inclusion parts under inelastic contribution due to 1.6-
thermal expansion coefficient mismatch.
Hence, from the results in Sections 4.1 and 4.2, the
total stress and displacement should be the sum of the 1.2-
elastic and the inelastic terms
~ ...........
13.1j
=
i1 i2
(~ij -JF (~ij, ~M =
rn~ m2
i~ij + {~ij
(9a)
/
I i1 i2 m s m2
0.8-
it

ui = ui + ui, u~ = ui + ui 9 (9b) iJ

Above stress components evaluated in the polar


coordinate can be transformed into the stress 0.4-
C ,'
components in the Cartesian coordinate using the
transformation laws
CYxx = CYrr C 0 S 2 0 - - 2tYrO cosOsinO 24 23 22 -1 0 ~ ~ ) ;
(b) Normalized distance, q/r
+ ~oo sin 20 (lOa)
oyy = ~,sinZO + 2~,ocosOsinO NormaLized sires% crx / %
1.6-
+ qoo cos 20 (lOb)
~y = (~,~ - eroo)cosOsinO
1.2
+ ~,o(COS20 - sinZO) (lOc)

3.8
5. Analysis and discussion
5.1. Stress concentration and load transfer
When a discontinuity is present in a body, local stress 0.4
disturbances will be developed near the discontinuity.
However, the extent and shape of stress disturbance
near the discontinuity depend on the geometry of the /•D

t}}{r
(el Normalized distance

Figure 5 Normalized stress (~,,/go) distribution in the region of


O"0
circular SiCp and 6061 AI alloy (a) Schematic illustration of the
loading configuration and the trajectory along which stress distri-
butions are drawn. (b) Along the line ABODE, ( ) inclusion,
( - - - ) matrix. (c) Along the interface BCD. The shaded semi-circle
in (b) and (c) indicates the location of the particulate.

discontinuity, and the difference in the elastic con-


stants and thermal expansion coefficients between the
matrix and the discontinuity. For example, when
a large thin plate having a discontinuity with smaller
elastic modulus than the matrix (such as a hole) is
subjected to uniaxial tension, a stress concentration
will occur at the matrix near the equator of the hole,
and decrease rapidly at distances away from the hole.
D
However, in the case of a large thin plate having
a discontinuity with an elastic modulus higher than
that of the matrix, a stress concentration occurs at the
discontinuity rather than in the matrix. Such stress
disturbances for an SiC inclusion in an aluminium
alloy matrix are illustrated in Figs 5 and 6. The elastic
constants and thermal expansion coefficients of 6061
aluminium alloy and SiC used for the calculations are

r162162162 given in Table III. Although, computations were


carried out for SiCp/AI composite only, trends exhib-
ited by (AlzO3)p/Al composite are expected to be
similar. The stress acting on the matrix near the equa-
(o) % tor of the inclusion is found to be lower than the
applied tensile stress due to the load transfer to the
inclusion through the interface. However, the stress

5458
acting on the matrix near the pole of the inclusion is Norrnotized stress, ~y/%
calculated ~to be about 1.5 times that of the applied 1.0
tensile stress. This stress reaches a maximum value in
the matrix at a point slightly away from the pole along
the direction of the applied stress, and finally de- 0.6
creases to the level of applied tensile stress at regions C
away from the pole. Such a stress distribution at the
pole region of the inclusion, along the direction of the 8 0
tensile stress, appears to be responsible for some of the Normalizedciisto'nce'
matrix material to be adherent to the particulate rein-
forcement, as has been observed occasionally at the
debonded interfaces (Fig. 3b). This stress distribution
is illustrated in Fig. 7. -0"61
(el -1.0

Figure 6 Normalized stress (%/%) distribution in the region of


circular SiCp and 6061 A1 alloy. (a) Schematic illustration of the
loading configuration and the trajectory along which stress distri-
% butions are drawn. (b) Along the line ABODE, ( - - ) inclusion,
( - - - ) matrix. (c)Along the interface BCD, ( - - ) interface.
The shaded semi-circle in (b) and (c) indicates the location of the
particulate.

,JA TABLE III Selected material properties for SiC and AlzO 3 1-28]

Material

6061 AI
E(GPa)

68.0
p(GPa)

29.3
v

0.33
:z (10 -6 ~

28.0
1)

O B SiC 423.0 178.0 0.19 3.0


A1203 396.9 160.2 0.24 6.5

5.2. Interfacial debonding and particulate


cracking
The axial stress developed in the matrix along the
matrix/inclusion interface, (~xm, increases significantly
at the region of the pole of the inclusion, and decreases
very rapidly as the angle, 0, measured from the pole
rotates and approaches 90 ~, while the stress inside the
inclusion remains constant, as shown in Fig. 5. Such
a stress concentration at the inclusion and the
matrix/inclusion interface can cause particulate crack-
ing and interfacial debonding, respectively. Using
Equations 4 and 10, the maximum stress in the matrix
(o) % near the pole of SiCp is calculated to be about 1.5
times the applied tension, assuming plane stress
Normatizedstress, Ey/% conditions. With a tensile stress of 300-400 MPa, the
1.0 typical strength of such composites, the stress at the
pole of SiCp reaches about 450-600 MPa. Although
such a stress is not large enough compared to the
0"6]C , interracial bonding strength of SiC and aluminium
reported by Flora and Arsenault [29], interfacial
debonding may occur due to the imperfect interface
present in the composites. The formation of an inter-
24 J3 22 -1 t
-0.2- metallic compound [30] or amorphous phase [26] can
cause degradation of the interface. Sharp corners [21]
Normatized distonc% q/r
will result in higher stress concentration. Incipient
-0.6- debonding at the interface as a consequence of the
manufacture results in severe stress concentration, as
(b) -1.0 shown in Fig. 8. Once interface is debonded at the pole
of SiCp, the crack so formed will propagate along the
SiCp/A1 interface to some extent and then deviate into

5459
O"0

tftttt

Figure 8 Scanning electron micrograph showing the incipient de-


bonding and compound layer observed in (A12Oa)p/A1 composite.
(a) % The micrograph was taken from the surface perpendicular to the
4-
extrusion direction.
A

the matrix. This is due to the fact that the axial stress,
Oxx, at the interface decreases and that in the matrix
increases abruptly as can be seen in Fig. 5.
Tensile loading also gives rise to stress concentra-
tion within SiCp, causing particulate cracking. From
Equations 4 and 10, the stress exerted on SiCp is
o
z B calculated to be about 1.5 times that of the applied
tensile stress. Although, this calculated stress is
smaller than the fracture strength of SiC, its failure can
occur even under a small stress value, because most
-4.5 -0'.9 -0'.3 0'.3 0'.9 '
1.5 ceramic particles have flaws, grain boundaries, and
sharp points where higher stress is concentrated.
(b) Normatized stress~ (~x/O'o
From Fig. 3, which illustrated typical particulate
2.0 cracking under tension, one can notice that particulate
cracking precedes matrix failure, although severe plas-
1.8- tic deformation could be observed in the matrix near
the region of cracked SiCp. Once SiCp has cracked, the
= 1,6- constraint on the plastic deformation of the matrix
._m will disappear. Then matrix near SiCp can easily
undergo plastic deformation. Such a situation can be
"~ 1.4-
o
considered to be similar to the matrix alloy with
E
o
a crack in it, providing a higher stress concentration at
z 1.2-
the crack tip. Under such conditions, the matrix fails
easily, and the ductility of the composite will decrease
1.0 t , , significantly.
1.0 111 1.2 1.3 1.4
(c) NormO.izedstres% O'x/Oo
Figure 7 (a) Schematic illustration of the loading configuration and 5.3. Effect of thermal residual stress on the
the trajectory along which stress distributions are drawn. (b) Nor- failure mode
malized stress (ox/o0) distribution from the pole of SiCp along
the line AB (tensile direction). (c) Detailed stress distribution
In this specific case, the pressure acting on the
pattern within the enclosed rectangle. The shaded quarter-circle in matrix/inclusion interface ( - P in Equation 7) is com-
(b) indicates the location of the particulate. puted to be about - 800 M P a for the temperature

5460
drop from the composite fabrication temperature to
room temperature (AT,~600~ However, upon )~ T~
cooling from the fabrication temperature, most of this
inelastic stress will be relieved by the plastic deforma-

0
tion of the matrix near SiCp (i.e. generation of disloca-
tion around SiCp) and only a small portion of about
- 30 MPa, as estimated by Arsenault and Taya [31],
will remain in the form of residual stress at the inter-
face. These inelastic stresses acting on SiCp and the
0
matrix will not change the state of stress significantly, ,ol $
because values are small compared to those of the
applied stress. Consequently, there should be no sub-
stantial changes in the mode of particulate cracking
and interfacial debonding as a result of the thermal
residual stress.

5.4. Proposed failure modes


0 0
(b)
Based on the present study, the failure mechanism of
the particulate-reinforced aluminium alloy composites
can be summarized as follows. When a composite
system is subjected to uniaxial tension, the maximum
tensile stress is generated at the reinforcement and
matrix near the pole of the reinforcement, in the same
direction to the applied tension. Such a stress concen-
tration may cause particulate cracking or interracial
debonding. As loading continues, the particulate
cracks and debonded interfaces easily open up due to Ir
0
the plastic deformation of the matrix near the rein-
forcement. New cracks will develop in the matrix at
the tip of the opened-up crack and propagate into the
matrix until they are joined with nearby cracks, result-
ing in a large void. Schematic representations of
various steps for this failure mode are illustrated
in Fig. 9.
The micrographs of the side surfaces of tensile speci-
mens adjacent to the fractured region are given in (d)
C7
Figs 1, 2 and 3. Extensive particulate crackings and
debonded interfaces can be observed in these micro-
graphs. Joining of nearby cracks into a large crack can
also be observed in Fig. 1. The void formation, as
a result of such coalescences, can be seen in Fig. 2.
Based on these observations, particulate cracking, in
addition to interface debonding, can be proposed as
a major contributor to the failure of the particulate-
reinforced aluminium alloy composites, (e)

6. Conclusions
Stress concentration in the matrix near the pole of the
particulate, and stress concentration within the partic-
ulate, appear to result in interfacial debonding and
particulate cracking, respectively. Increase in applied
stress causes such microcracks to open. New cracks
develop in the matrix at the tip of the opened-up (f)
crack, propagate into the matrix and join with nearby
cracks, and form a large void. Based on the micro- Figure 9 Schematicdrawings of the failuremechanism of the partic-
scopic examinations and analytical study, interfacial ulate-reinforced aluminium alloy composites. (a) Loading config-
debonding and particulate cracking together were uration. (b) Formation of particulate cracking and interfacial
debonding. (c) Opening-up of cracked plane and debonded inter-
considered to be responsible for the substantial de- face due to plastic flow of the aluminium matrix. (d) Crack propa-
crease in the ductility of the particulate-reinforced gation into the aluminium matrix due to stress concentration
aluminium alloy composites. build-up at the crack tip. (e) Joining of cracks. (f) ~oid formation.

5461
Because particulate cracking and interracial debond- 16. D . L . M c D A N E L S , Met. Trans. 16A (1985) 1105.
ing were observed to always precede the matrix 17. R.J. ARSENAULTandS. B. WU,Mater. Sci. Engng96(1987)
77.
failure, such phenomena appear to be more respon-
18, A . K . VASUDEVAN, O. R I C H M O N D , F. Z O K and J. D.
sible for the failure than a mechanism based on the E M B U R G , ibid. A107 (1989) 63.
matrix failure 19. H. O H T S U , "Design and manufacturing of advanced com-
posites" (ASM, Detroit, 1989) p. 187.
20. D. J. LLOYD, H. LAGACE, A. M c L E O D and P. L.
References MORRIS, Mater. Sci. Engng A107 (1989) 73.
21. S.R. N U T T and J. M. DUVA, Scripta Metall. 20 (1986) 1055.
1. N. T S A N G A R A K I S , B. O. A N D R E W S and C. CAVAL-
22. S.R. N U T T and A. N E E D L E M A N , ibid. 21 (1987) 705.
LARO, J. Compos. Mater. 21 (1987) 481.
23. J . J . L O W A N D O W S K I , C. LIU and W. H. H U N T , Mater.
2. T . G . NIEH and D. J. C H E L L M A N , Scripta Metall. 18 (1984)
Sci. Engng 107A (1989) 241.
925.
24. C. P. YOU, A. W. T H O M P S O N and I. M. BERNSTEIN,
3. R . J . A R S E N A U L T and S. B. WU, ibid. 22 (1988) 767.
Scripta Metall. 21 (1987) 181.
4. R . J . A R S E N A U L T , Mater. Sei. Engng 64 (1981) 171.
25. T. CHR1STMAN, A. N E E D L E M A N , S. N U T T and
5. Y. F L O M and R. J. A R S E N A U L T , J. Metals 38 (1986) 31.
S. SURESH, Mater. Sci. Engng 107A (1989) 49.
6. V.C. N A R D O N E , Scripta Metall. 21 (1987) 1313.
26. G . M . J A N O W S K I and B. J. P L E T A K A , ibid. A129 (1990) 65.
7. R . H . JONES, C. A. L A V E N D E R and M. T. SMITH, ibid. 21
27. N. I. M U S K H E L I S H V I L I , "Some basic problems of the
(1987) 1565.
mathematical theory of elasticity" (Noordhoff, Groningen,
8. W . A . L O N G S D O N and P. K. LIAW, Engng Fract. Mech. 24
Holland, 1953) pp. 215-16.
(1986) 737.
28. Y. S. T O U L O U K I A N (ed.) "Thermophysical properties of
9. T. C H R I S T M A N and S. SURESH, Mater. Sci. Engng 102
high temperature solid materials, Vol. 6, part I (Thermophysi-
(1988) 211.
cal properties research center, Purdue University, 1967).
10. J . K . SHANG, W. YU and R. O. RITCHIE, ibid. 102 (1988)
29. Y. F L O M and R. J. A R S E N A U L T , Mater. Sei. Engng 77
818.
(1986) 191.
11. F . M . H O S K I N G , F. F. P O R T I L L O , R. W U N D E R L I N and
30. T. ISEKI, T. K A M E D A and T. M A R U Y A M A , J. Mater. Sci.
R. M E H R A B I A N , J. Mater. Sci. 17 (1982) 477.
19 (1984) 1692.
12. F. A. G I R O T , J. M. Q U E N I S S E T and R. NASLAIN,
31. R.J. ARSENAULTandM.TAYA,ActaMetall. 35(1987)651.
Compos. Sci. Technot. 30 (1987) 155.
13. V. C. N A R D O N E and J. R. STRIFE, Metall. Trans. 18A
(1987) 109.
14. K . S . R A V I C H A N D R A N , J. Metals 39 (1987) 28.
Received 17 June
15. J.J. STEPHENS, J.P. LUCASandF. M. HOSKING, Scripta
Metall. 22 (1988) 1307. and accepted 16 December 1991

5462

Anda mungkin juga menyukai