Thermally induced cracking in aluminum/boron carbide composite

Abstract

Thermal shock tests have been performed to determine the influence of rapid temperature change on crack nucleation and propagation in lightweight Al/B 4 C composite. The composite was produced using self-propagating high-temperature synthesis (SHS) with subsequent heat treatment at 700°C in vacuum. Quenching in water was chosen as thermal loading. The specimens have been subjected to a specified number of thermal cycles and then analyzed. Post-shocked samples have been characterized by means of microstructural and fractographical investigations to analyze the crack paths.

Full text

6th International DAAAM Baltic Conference
INDUSTRIAL ENGINEERING
24-26 April 2008, Tallinn, Estonia
THERMALLY INDUCED CRACKING IN ALUMINUM/BORON
CARBIDE COMPOSITE
Kimmari, E.; Kommel, L.
Abstract: Thermal shock tests have been
performed to determine the influence of
rapid temperature change on crack
nucleation and propagation in lightweight
Al/B4C composite. The composite was
produced using self-propagating high-
temperature synthesis (SHS) with
subsequent heat treatment at 700°C in
vacuum. Quenching in water was chosen
as thermal loading. The specimens have
been subjected to a specified number of
thermal cycles and then analyzed. Post-
shocked samples have been characterized
by means of microstructural and
fractographical investigations to analyze
the crack paths.
Key words: boron carbide composite,
thermal cycling, quenching, crack path.
1. INTRODUCTION
Reaction sintered materials in the system
aluminum/boron carbide were investigated.
Development of composites of Al-B4C
system is of interest because the both
constituents have low specific weight;
additionally the system combines high
hardness, high elastic modulus, outstanding
wear resistance of ceramic and enhanced
ductility, high strength/density ratio and
corrosion resistance of metallic phase [1, 2].
The additions of ceramic exert influence on
coefficient of thermal expansion and
thermal conductivity [1]. B4C as compare
with other ceramic reinforcement (Al2O3,
SiC) has lower density, resulting in
composites with higher specific stiffness
[2].
Thermal shock and thermal cycling are the
life-limiting factors for components
exploitable in rapidly changing thermal
conditions and high-temperature materials.
Monolithic ceramics, including B4C, is
known to be susceptible to thermal shock
failure through their low coefficient of
thermal expansion (CTE) [3]. Boron
carbide with CTE (4.5-5.0)×10-6 °C-1 is one
of the promising ceramic materials for the
design of low CTE composites. Although
the high-temperature application of Al/B4C
composites is limited because of a
requirement that the matrix and
reinforcement must be mutually chemically
inert in the operating temperature range,
the use of this class of materials is
favorable in unlubricated sliding, i.e. in
conditions where the friction temperatures
may reach up to relatively high values [4].
Due to high neutron absorption of B4C, the
composites are expected to find
applications in nuclear industry [2], where
the thermal properties play an important
role.
Composite materials processed and
exploited at elevated temperatures exhibit
thermal residual stresses when cooled to
room temperature due to differences in
CTE of the constituent phases. The CTE of
Al is about five times larger than that of
B4C. The developed residual stresses have
several undesirable effects such as
diminished fatigue strength, accelerated
stress corrosion, shape distortion [5].
From this point of view it is essential to
study the Al/B4C composites exposed to
cyclically changing thermal loading. The
understanding of thermal behavior and
nature of thermal damage of B4C-based
composites can provide additional
knowledge for their microstructural design.

2. EXPERIMENTAL
The materials were produced by reaction
bonding of attrition mixed B4C and Al
powders. The mixed powders containing
52 vol.% B4C were heated in a furnace up
to temperature 850°C to initiate the
exothermic reaction. When reaction
became complete pressure was applied at
liquid state of Al for better densification
achievement. Synthesized material was
then heat treated at 700°C in vacuum
during 30 minutes for structural
development. Low temperature was chosen
in order to prevent the formation of brittle
high-temperature phases [6]. Samples of
parallelepipedic shape 20×12×7 (mm) were
prepared by diamond polishing down to a
diamond particle size of 1 μm.
For thermal loading samples were heated
in a furnace and kept for 20 minutes at
specified temperature in order to attain
thermal equilibrium. The temperatures lie
in the range 300-600°C. Heating was
performed in air at atmospheric pressure.
Heated samples were then dropped into
quenching medium (water maintained at
room temperature). The procedure
described above represents one thermal
cycle. Up to 20 cycles were done at each
temperature.
The microstructural observations were
done using Hitachi TM-1000 Scanning
Electron Microscope (SEM) and optical
microscopy (OM).
3. RESULTS AND DISCUSSION
3.1 Characterization of initial composite
The main microstructural character of
Al/B4C composites is the variety of
microstructural types tailored [6-8].
Depending on processing conditions the
microstructure can vary from the two-
phase typical for traditional MMCs to the
multiphase interpenetration one.
SEM-image of the microstructure of initial
composite is presented in Fig. 1, where
boron carbide particles are in dark-grey
color and binder is in light-grey color.
Fig. 1. SEM-image of initial composite.
The grains of B4C are surrounded by a
matrix composed of Al and reaction
products. Some of the large grains are
cracked due to polishing. Average size of
carbide grains is 17.3 μm.
XRD-patterns [8] revealed that the main
reaction product is ternary carbide Al3BC;
weaker lines correspond to reflections from
aluminum carbide Al4C3 were observed
also. The process of nucleation and growth
of Al3BC was described by Viala et al. [7]:
Al3BC phase is formed at the interface
between B4C and Al supersaturated in
carbon and boron, and is aggregated in
crystals surrounding the B4C particles.
Fig. 2. OM- and SEM-images showing the
absence and the presence of Al3BC-phase
around B4C particles.

Microstructural investigations suggest that
Al3BC was formed only at some B4C-
grains in given processing conditions. As it
is shown in Figs. 1 and 2, in some places
Al3BC crystals have already formed a
quasi-continuous reaction zone, but mainly
B4C particles are still in direct contact with
aluminum. Debonding at some Al-B4C
interfaces is also obvious (see Fig. 2).
The interfacial zone between reinforcement
and matrix is an essential part of
composites. The strength of the interface is
dependent on the nature of reinforcement
(added or formed in-situ), on adhesion, and
extent of the interfacial chemical reaction.
It should be noted also that adhesion work
between Al and B4C at 700°C is relatively
low (about 320 MJ/m2) [2].
3.2 Characterization of post-shocked
composites
SEM micrographs of the samples cyclically
(up to 20 cycles) quenched in water from
300°C and 400°C did not reveal of any
large crack formation. It can be concluded
that such thermal loading was not able to
generate any catastrophic failure or to
propagate the pre-existing flaws.
The formation of few large cracks in
surface area was caused by quenching of
samples from 500°C and 600°C. Some
little cracks formed were stopped in the
pools of aluminum or at B4C-grains.
Debonding of some individual particles
was caused also.
In following are investigated the paths of
large cracks formed as a result of thermal
loading and possible mechanisms of their
propagation are proposed. It must be noted
that all large cracks were initiated from
diamond pyramid indents or from surface
macrodefects such as voids.
Fig. 3 shows three different response
behaviors to indentation. There are three
Vickers pyramid indents placed in different
zones of microstructure: in a B4C grain
(Fig. 3, a), in a large pool of binder with no
evidence of reaction products (Fig. 3, b),
and in a zone with reaction products (Fig.
3, c).
Fig. 3. Vickers indents in different zones of
microstructure.
In the first two cases the crack nucleation
and propagation into the matrix is
suppressed by high ductility of Al.
Otherwise, when the indent is placed in a
region with reaction products, a crack is
originated and propagated.
Fig. 4. Thermal crack path (600°C, 10
cycles); RP – reaction products.

In Fig. 4 is presented a typical path of a
large crack in samples quenched from
temperatures 500°C and 600°C.
Analyzing Fig. 4 it can be concluded that
all fracture is going through the matrix near
boron carbide grains, where the reaction
product (Al3BC) are formed and partly
through the interface. The fraction of
cracked B4C particles along the crack paths
is neglible. Voids formation at the Al-B4C
interface indicates the destruction of
bonding.
No crack’s widening was observed after 20
quenching cycles, as distinct from evident
oxidation.
Based on microstructural observations, the
toughening mechanism of crack deflection
by boron carbide particles can be supposed,
that was firstly proposed by Lange [11].
This approach means that when a crack
intersects an obstacle, it is bent to some
angle before it can move on. Tough boron
carbide particles are expected act as
obstacles.
An example of crack deflection by B4C-
particle is presented in Fig. 5. The stress
needed to bypass the B4C-particle must be
greater than the fracture stress of the matrix
[12].
The crack path near a particle will depend
on the residual stress fields in this zone
resulting from CTE mismatch and phase
transformation during processing.
Fig. 5. Crack deflection by B4C grain.
Table 1 gives an overview about the
properties of composite’s phases.
Table 1. Properties of phases; E – Young’s
modulus, G – shear modulus, B – bulk
modulus, α − CTE, ν − Poisson’s ratio.
Al B4C Al3BC
E, GPa
G, GPa
B, GPa
α×10−6,Κ−1
ν
69 [9]
27 [9]
75 [9]
23.6[9]
0.35[9]
460[2]
180[2]
245[2]
5.0 [2]
0.16[2]
129 [10]
175 [10]
The spherical shell solution method based
on Hashin’s composite sphere assemblage
[13, 14] was used to predict the thermal
expansion stresses developed in composite
after cooling from heat-treatment
temperature. Under purely elastic
conditions, after cooling from heat
treatment temperature the radial stress at
interface in particle and in matrix was
found to be compressive (–445 MPa). At
the same time the matrix undergoes to
large tensile residual stress when cooled
from heat-treatment temperature [14, 15].
While the spherical particle was concerned,
the predicted equivalent stress distributes
uniformly. Actually the particles are
angular. Therefore there is a stress
concentration in the pointed particle
corner. With decreasing pointed corner
degree of the particle, this concentration
increases rapidly and becomes more and
more intense [16].
The misfit strain induced by CTE
difference can be accommodated in
different ways: elastically, by localized
plastic deformation of the matrix or
interface rupture [17].
Surey et al. [18] have found that further
reheating changes the sign of the radial
stress at the interface, i.e. radial stress
becomes tensile. This phenomenon is a
matter of great importance since in the case
of weak interface and/or large radial stress
it may lead to decohesion. The absolute
values of radial stresses caused by CTE

mismatch in the case of purely elastic
behavior are presented in Fig. 6.
0
100
200
300
400
0,95 1,00 1,05 1,10 1,15 1,20
r/a
Radial stress, MPa
(absolute value)
280°C
380°C
480°C
580°C
Fig. 6. Predicted radial stress evolution in
the matrix for different temperature drop;
conditions are purely elastic. (a – radius of
the particle, r – distance from the center of
particle).
The results were calculated using spherical
shell solution [13]. The values of stresses
are expected to be high enough to cause the
plastic deformation of the matrix. No
cooling rate effect and influence of
formation of reaction product were taken
into account.
The effect of cooling rate on residual stress
in composites with different volume
fraction of ceramic was studied by Ho and
Saigal [19]. They found that the composites
subjected to a higher cooling rate
experience higher residual stresses in the
matrix.
Further heating of composite cause
reducing of residual tensile stresses in
matrix and may generate the compressive
stresses, which magnitude is limited due to
viscoplastic relaxation [15]. The following
cooling cause the tensile stresses build up
again. Due to high volume fraction and
angular shape of boron carbide particles
the accumulation of tensile stresses may
reach high values exceeding the yield
stress of Al [15]. The large stresses
interacting with neighbouring defects of
the microstructure are able to nucleate the
crack.
SEM-micrograph (Fig. 7) of the fracture
surface of a non-shocked sample subjected
to three-point bending indicates that the
crack front is propagated partly through the
interface of B4C-particles and partly
through the matrix.
Fig. 7. Fracture surface of a non-shocked
sample.
The overwhelming majority of B4C-
particles visible in fracture surface are
located flatways. It can be proposed that
there is a weak interface between large
surfaces of B4C and binder, caused by poor
wettability of boron carbide by the liquid
aluminum under present processing
conditions [2]. There is also evidence of
non-planar crack front that confirms above
presupposed toughening mechanism of
crack deflection by B4C-particles.
4. CONCLUSION
Al/B4C composites with low volume of
reaction products are expected to be
sustainable to low cycle thermal quenching
up to 400°C.
The thermal shock failure is dominated by
one or few major cracks. In a
discontinuously reinforced matrix the rate
of fracture extension is reduced by crack
deflection. The crack initiation sites are
mainly located in regions with higher
reaction product volume. Thus, the reaction
products are expected to act as embrittling
species. The interface reaction is supposed
to have harmful effects on the crack growth
resistance of the composite.
Fractography shows the interface failures
at large surfaces of B4C-grains and
evidence of non-planar crack, suggesting
the poor adhesion at large surfaces of B4C.

The uniformity, volume fraction, aspect
ratio, shape and spacing of the B4C
particles are expected to have a strong
effect on thermal stress distributions and
thermal damage of particulate composite.
Additional experimental approaches are
needed to validate the predicted radial
stresses, to clarify the actual stress
distributions and to take into account the
cooling rate effect and the influence of
interface reaction.
5. REFERENCES
1. Miserez, A., Müller, R., Rossol, A.,
Weber, L., Mortensen, A. Particle
reinforced metals of high ceramic content.
Mat. Sci. Eng. A, 2004, 387-389, 822-831.
2. Kislij, P., Kuzenkova, M., Bodnaruk,
N., Grabchuk, B. Boron Carbide. Nauk.
Dumka, Kiev, 1988. (in Russian).
3. Smagorinski, M.E., Tsantrizos, P.G.
Development of light composite materials
with low coefficients of thermal expansion.
Mater. Sci. Tech., 2000, 16, 853- 861.
4. Chapman, T.R., Niesz, D.E., Fox, R.T.,
Fawcett, T. Wear-resistant aluminum-
boron-carbide cermets for automotive
brake applications. Wear, 1999, 236, 81-87.
5. Ho, S., Lavernia, J. Thermal residual
stresses in metal matrix composites: a
review. Appl. Compos. Mater., 1995, 2, 1-30.
6. Pyzik, A., Beaman, R. Al-B-C phase
development and effects on mechanical
properties of B4C/Al-derived composites.
J. Am. Ceram. Soc., 1995, 78(2), 305-312.
7. Viala, J.C., Bouix, J., Gonzalez, G.,
Esinouf, C. Chemical reactivity of
aluminium with boron carbide. J. Mater.
Sci., 1997, 32, 4559-4573.
8. Kommel, L., Kimmari, E. Solid phase’s
transformations in boron carbide based
composites during heat treatment. Solid
State Phenom., 2008, 138, 175-180.
9. Schwartz, M. Encyclopedia of
Materials, Parts, and Finishes, 2nd Ed.
CRC Press, New York, 2002.
10. Torquato, S., Yeong, C.L.Y., Rintoul,
M., Milius, D., Aksay, I. Elastic properties
and structure of interpenetrating boron
carbide/aluminium multiphase composites.
J. Am. Ceram. Soc., 1999, 82(5), 1263-
1268.
11. Lange, F.F. The interaction of a crack
front with a second-phase dispersion. Phil.
Mag., 1970, 22, 983-992.
12. Green, D.J., Nicholson, P.S. Fracture of
brittle particulate composites. In Fracture
Mechanics of Ceramics, vol. 4 (Bradt, R.C.,
Hasselman, D.P.H., Lange, F.F., eds.).
Plenum Press, New York, 1978, 945-960.
13. Hashin, Z. The elastic moduli of
heterogeneous materials. J. Appl. Mech.-T.
ASME, 1962, 29, 143-150.
14. Shu, K.-M., Tu, G.C. The
microstructure and the thermal expansion
characteristics of Cu/SiCp composites. Mat.
Sci. Eng. A, 2003, 349, 236-247.
15. Zhang, Q., Wu, G., Jiang, L., Luan, B.
Thermal properties of a high volume
fraction SiC particle-reinforced pure
aluminum composite. Phys. Stat. Sol. A,
2005, 6, 1033-1040.
16. Ya, M., Lu, J. Residual Stress in Metal
Matrix Composites. In Handbook of
Residual Stress: Manufacturing and
Materials Processing (Lu, J., ed.). Society of
Experimental Mechanics, Inc., USA, 2005.
17. Kustov, S., Goljandin, S.,
Sapozhnikov, K., Vincent, A., Maire, E.,
Lormand, G. Structural and transient
internal friction due to thermal expansion
mismatch between matrix and
reinforcement in Al-SiC particulate
composite. Mat. Sci. Eng. A, 2001, 313,
218-226.
18. Surey, M. Teodosiu, C., Menezes, L.F.
Thermal residual stresses in particle-
reinforced viscoplastic metal matrix
composites. Mat. Sci. Eng. A, 1993, 167,
97-105.
19. Ho, S., Saigal, A. Three-dimensional
modeling of thermal residual stresses and
mechanical behavior of cast SiC/Al
particulate composites. Acta Metall.
Mater., 1994, 42, 3253-3262.