Cell wall mechanical properties of closed-cell Al foam

Insu Jeon
a,
*
, Kiyotaka Katou
b
, Tsutomu Sonoda
b
, Tadashi Asahina
b
, Ki-Ju Kang
a
a
School of Mechanical Systems Engineering, Chonnam National University, 300 Yongbong-dong, Buk-gu, Gwangju 500-757, Republic of Korea
b
Materials Research Institute for Sustainable Development, National Institute of Advanced Industrial Science and Technology (AIST), Shimoshidami, Moriyama-ku,
Nagoya 463-8560, Japan
a r t i c l e i n f o
Article history:
Received 3 March 2008
Received in revised form 11 August 2008
a b s t r a c t
The mechanical properties of the cell wall, such as the elastic modulus, 0.2% offset yield
stress and power-law hardening exponent of the closed-cell Al foam are determined using
both experimental measurements and nite element analyses. A 6 6 12 cm
3
ingot of
the cell wall base material, which is sampled from melted Al-1.5 wt.%Ca alloy before foam-
ing, is prepared, and its mechanical properties are initially measured to set the limit values
of the mechanical properties of the Al cell wall. Two 5 5 5 mm
3
Al foam specimens of
completely different structures are fabricated, and directly modeled for the nite element
analysis by using a microfocus X-ray CT system, 3D reconstruction program, 3D scanned
data processing software, and commercial mesh generation program. Subsequently, uniax-
ial compression tests are carried out on the specimens, and the numerical simulations of
these tests are performed using the nite element models. For the simulations, various
mechanical properties for the cell wall selected from the measured properties of the base
material are used. Then, the Al cell wall mechanical properties are precisely determined by
comparing the computed forcedisplacement curves with the measured ones. Finally, the
effects of each mechanical property on the compressive behavior of the foam material are
analyzed.
2008 Elsevier Ltd. All rights reserved.
1. Introduction
As a functional light weight material, closed-cell Al
foamis widely used in transportation applications, military
equipment, machine tools and so on. Therefore, the
mechanical performances of the foam material, including
its effective elastic modulus, compressive strength, and en-
ergy absorption capacity, have become an important sub-
ject of research. Much research has been done on
performance improvement of the foam material. Two main
streams of research have been developed subsequently:
rst, the analysis of the signicant effects of geometric
parameters related to the cell structure on the perfor-
mance of the foam material (Miyoshi et al., 1999, 2000;
Simone and Gibson, 1998a; Grenestedt, 1998; Grenestedt
and Tanaka, 1999; Markaki and Clyne, 2001; Jeon and
Asahina, 2005), and second, the analysis on the effects of
the cell wall material properties on the performance of
the foam material (Jeon et al., submitted for publication).
With respect to the effects of the geometric parameters,
decreased cell size and increased thickness of the cell wall
increase the compressive strength of the aluminum foam
material (Miyoshi et al.,1999, 2000). Furthermore, cell face
curvatures and corrugations decrease the elastic modulus
and plastic collapse stress (Simone and Gibson, 1998).
The wavy imperfection and cell shape variation decrease
the stiffness of foam materials (Grenestedt, 1998;
Grenestedt and Tanaka, 1999). The effects of cell wall
microstructures, which are formed differently depending
on the production processes and the composition of the
cell wall raw materials, on the compressive response of
closed-cell foams were also presented (Markaki and Clyne,
2001). Recently, the effects of intrinsic structural defects of
0167-6636/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mechmat.2008.08.002
* Corresponding author. Tel.: +82 62 530 1688; fax: +82 62 530 1689.
E-mail address: i_jeon@chonnam.ac.kr (I. Jeon).
Mechanics of Materials 41 (2009) 6073
Contents lists available at ScienceDirect
Mechanics of Materials
j our nal homepage: www. el sevi er . com/ l ocat e/ mechmat
the closed-cell Al foam on the decrease of the effective
modulus as well as the plateau stress of the foam material
have been reported (Jeon and Asahina, 2005).
With respect to the effects of the material properties of
the cell wall, research results have not been extensive. Jeon
et al. (submitted for publication), however, demonstrated
the effects of cell wall plastic properties, such as the 0.2%
yield stress and power-law hardening exponent, on the en-
ergy absorption capacity of the closed-cell Al foam. They
also compared the stressstrain curve measured from the
a uniaxial compression test on a 10 10 10 mm
3
closed-cell Al foam with the stressstrain curves obtained
from numerical simulations of the test. Although the plas-
tic properties of the cell wall contribute signicantly to the
mechanical performance of the foam material, research on
the measurement of the precise values of these properties
have been limited.
An approximation method was applied to measure the
yield stress of the Al cell wall, which takes the yield stress
values to be one-third of the cell wall hardness (Sugimura
et al., 1997; Simone and Gibson, 1998b; Andrews et al.,
1999). Another measurement method using nanoindenta-
tion has been recently introduced to measure cell wall
mechanical properties (Kim and Tunvir, 2006; Hasan
et al., 2008). This method, however, has some difculties
Fig. 1. (a) A 6 6 12 cm
3
base material ingot, (b) three cylindrical specimens of 25 mm height and 12.7 mm diameter, and (c) a specimen between
compression platens.
Fig. 2. True stress-true strain curves of the base material.
Table 1
Measured elastic moduli and elastic limit stresses of the base material
E (GPa) Elastic limit stress, r
e
(MPa)
Comp. test case #1 66.09 12
Comp. test case #2 56.24 9.9
Comp. test case #3 62.8 11.6
Average 61.7 11.16
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 61
Fig. 3. (a) Specimen A (5 5 5 mm
3
), (b) its tomographic images and (c) reconstructed model.
Fig. 4. (a) Specimen B (5 5 5 mm
3
), (b) its tomographic images and (c) reconstructed model.
62 I. Jeon et al. / Mechanics of Materials 41 (2009) 6073
in obtaining actual cell wall mechanical properties: it usu-
ally uses the indentation loaddisplacement curves, which
include considerable variations arising from the inhomo-
geneous microstructure of the cell wall cross section, and
can not account for the effect of cell wall oxidation in the
measurement of the mechanical properties.
In this research, we suggest a new approach to deter-
mine the mechanical properties of the cell wall of the
closed-cell Al foam. After measuring the mechanical prop-
erties of the cell wall base material, taken from melted Al-
1.5 wt.%Ca alloy before foaming, we perform a direct nite
element modeling (Jeon et al., submitted for publication) of
two actual 5 5 5 mm
3
foam specimens. Uniaxial com-
pression tests using the specimens and numerical simula-
tions of the tests using the nite element models are
carried out. In the simulations, the mechanical properties
for the cell wall selected from the measured properties of
the base material are used. Then, the cell wall mechanical
properties and their effects on the compressive behavior of
foam materials are precisely determined by comparing the
computed forcedisplacement curves with the measured
ones.
2. Mechanical properties of the cell wall base material
A 6 6 12 cm
3
cell wall base material ingot, taken
from melted Al-1.5 wt.%Ca alloy before foaming, was pre-
pared to measure the mechanical properties of the base
Fig. 5. (a) Modied and smoothened outer cells and inner voids, (b) NURBS surfaces of the outer cells and inner voids, (c) fabricated geometric solid and (d)
3D nite element model of Specimen A.
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 63
material, in order to establish the limit values of the
mechanical properties of the Al cell wall (see Fig. 1a).
From the ingot, three cylindrical specimens of 25 mm
height and 12.7 mm diameter were fabricated by using
the electrical discharge machine for the uniaxial com-
pression test (see Fig. 1b). The compression tests were
carried out by using a universal testing machine AG-I of
SHIMADZU Corp. with a 100 kN load cell. The specimens
were compressed between two circular steel compression
platens of 200 mm in diameter. The surfaces of the plat-
ens were lubricated to reduce the friction effect of the
contact surfaces between the specimens and the platens
on the compression test results (ASTM E9-89A, 2000).
Three strain gauges were attached on each specimen
surface to measure the precise strain values (see Fig. 1c).
A displacement rate of 0.5 mm/min was applied to the
top surface of the specimen. The strain was accurately
determined only up to 3.5% because of the measurement
limitation of the strain gauge. The three, measured, true
stress-true strain curves are shown in Fig. 2.
From the curves in Fig. 2, the elastic modulus E and
elastic limit stress r
e
were obtained and tabulated together
with their average values, E = 61.7 GPa and r
e
= 11.16 MPa,
in Table 1. The nonlinear part of the curves can be tted by
using the RambergOsgood model as follows:

e
p
e
e
a

r
r
e

n
for

r > r
e
1
where r, the Mises stress;

e
p
, the equivalent plastic strain;
e
e
= r
e
/E, the reference strain component; a, a non-dimen-
sional constant set to 1 for computation; n, the power-
law hardening exponent.
The properly tted curve using the RambergOsgood
model for the measured curves of the base material is plot-
Fig. 6. (a) Modied and smoothened outer cells and inner voids, (b) NURBS surfaces of the outer cells and inner voids, (c) fabricated geometric solid and (d)
3D nite element model of Specimen B.
64 I. Jeon et al. / Mechanics of Materials 41 (2009) 6073
ted together in Fig. 2. For this curve tting, the power-law
hardening exponent, n, was determined as 3 obtained from
the comparison of the measured curves with the calculated
curves after substituting E = 61.7 GPa and r
e
= 11.16 MPa,
and various n values into Eq. (1).
3. Preparation of two small foam specimens
Two 5 5 5 mm
3
specimens A and B of the closed-
cell Al foam named ALPORAS of completely different
structures were prepared for the uniaxial compression
tests and nite element modeling (see Figs. 3a and 4a).
Before preparing these specimens, several 5 5 5 mm
3
foam specimens were fabricated by using the electrical
discharge machine to ensure both surface atness and
clearness. Then, of these specimens, two specimens having
no structural defects that would degrade the compressive
performance of foam materials were carefully selected by
surface inspection: these two nal specimens were the
specimens A and B. The relative density of Specimen A
was q/q
s
= 0.0785 and that of Specimen B was
q/q
s
= 0.0837. Here, q is the density of the foam material,
and q
s
= 2.7 g/Cm
3
is the density of the Al cell wall.
4. Finite element modeling of the two foam specimens
A thorough introduction of the direct nite element
modeling procedure for a closed-cell foam was presented
(Jeon et al., submitted for publication). Following the pro-
cedure, two 5 5 5 mm
3
foam specimens A and B were
modeled independently. The microfocus X-ray CT system
of Shimadzu Corp. was used to scan the outer and inner
structures of the specimens. The scanned grayscale tomo-
graphic images of the specimens A and B are shown in Figs.
3b and 4b, respectively. For each specimen, 257 images
were taken along the specimen height of 5 mm. These
images include the structures of the cell wall and the air
region within the foam materials.
Using these images, each foam specimen was recon-
structed using a 3D reconstruction program TRI/3D-BON
of Ratoc System Engineering Co. Ltd. The prole of the
Fig. 7. (a) Specimen A between two platens and (b) their nite element
models.
Fig. 8. (a) Specimen B between two platens and (b) their nite element
models.
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 65
cellular structure and the air region of a specimen in the
tomographic images were binarized using the 3D recon-
struction program and the isolated small particles from
the cellular structure in the binarized images were re-
moved. The 3D cellular structures were then reconstructed
using the marching cubes algorithm (see Figs. 3c and 4c)
(Sone et al., 2004; Jeon et al., submitted for publication).
For geometric modeling of each foam specimen, every
cell in each reconstructed model was disjointed, and all
the disjointed cells were modied and smoothened one
by one to fabricate their nonuniform rationale B-spline
(NURBS) surfaces. Figs. 5a and 6a show the modied and
smoothened outer cells and inner voids of the specimens
A and B, respectively. The cube-shaped frames in Figs. 5a
and 6a represent the disjointed status of each cell, and
the bold solid lines in the disjointed cells indicate the
boundary of each cell. For simplicity of modeling, all inner
voids having maximum diameters that were less than
0.1 mm were disregarded. Then, the NURBS surfaces were
fabricated using each cell. Figs. 5b and 6b show the fabri-
cated NURBS surfaces for the outer cells and inner voids
of specimen A and B, respectively. Two geometric solids
that have the same structure of the two foam specimens
were fabricated using Boolean operations between two
Fig. 9. Variation of true stress-true strain curves obtained from curves of the base material according to the change of (a) the power-law hardening
exponent, (b) elastic limit stress and (c) elastic modulus.
66 I. Jeon et al. / Mechanics of Materials 41 (2009) 6073
5 5 5 mm
3
geometric solid cubes and the outer and in-
ner NURBS surfaces of each specimen (Jeon et al., submit-
ted for publication). The completed geometric solids for
the closed-cell foam specimen A and B are shown in Figs.
5c and 6c, respectively. The 3D scanned data processing
software RapidForm
TM
of INUS Tech. Inc. was used for all
geometric modeling (Shin et al., 2007; Jeon et al., submit-
ted for publication).
Subsequently, compatible 3D nite element meshes
were generated directly in each fabricated geometric solid
after iterative calculations using the commercial mesh
generation software PATRAN of MSC Software Corp. (see
Figs. 5d and 6d). In this process the volume errors,
VmVa
Va
,
where V
m
is the volume of the meshed model and V
a
is
the actual volume of the specimen, were 5.5% for Specimen
A and 4.94% for Specimen B, respectively.
Using the nite element models, two Al foamspecimens
between two platens were completely modeled (see Figs. 7
and 8). Quadratic tetrahedral elements with 10 nodes were
used for the specimens and platens. The total number of
elements and nodes used for specimen A were 79311 and
145725, and for specimen B were 75994 and 137585,
respectively. For simulating the compression test process,
a package code ABAQUS of Dassault Systems was used.
The incremental plasticity theory of isotropic-hardening
materials was selected for the elasticplastic behavior of
the specimens, which uses the objective stress rate for -
nite deformation of the Al cell wall. Moreover, the
power-law hardening rule in Eq. (1) was used to calculate
the plastic hardening index in the incremental plasticity
theory (McMeeking, 1977; Jeon and Im, 2001; Lee et al.,
2007; Jeon et al., submitted for publication). As the bound-
ary condition for the simulations, x- and y-direction de-
grees of freedom were constrained at several nodes
around the center on the lower surface of each specimen
to prevent rigid body motions, and z-direction displace-
ment was loaded at over 20% strain on the top surface of
each upper platen. The frictionless contact condition for
the contact surfaces between the specimens and the plat-
ens was applied, which represents the lubricated surface
condition of the platens.
As for the material properties of the platens,
E = 214 GPa and v = 0.3 of Tool Steel SKS3 were applied.
For the Al cell wall, various values of the power-law hard-
ening exponents, n = 3, 5, 6, 8.5, 9 and elastic limit stresses,
r
e
= 11.16, 25, 28, 30, 31, 32, 34 MPa, were selected; these
various values were the measured mechanical properties
of the cell wall base material and selected values based
on the measured values. For the cell wall elastic modulus,
E = 61.7 GPa of the base material, E = 68 GPa of the pub-
lished values of ALPORAS (Sugimura et al., 1997) and
E = 73.1 GPa of Aluminum 2024-T6 were selected for the
numerical tests. Fig. 9ac show the calculated true stress-
true strain curves after substituting the various selected
values of n, r
e
and E in Eq. (1) together with the measured
curves of the base material.
5. Uniaxial compression tests of the two foam
specimens
A universal testing machine AG-I of SHIMADZU Corp.
with a 5 kN load cell was used for the uniaxial compression
tests of the two 5 5 5 mm
3
foam specimens. The spec-
imens were compressed between two circular platens of
100 mm in diameter, which was sufciently larger than
the specimens (see Figs. 7a and 8a). The surfaces of the
platens were lubricated to reduce the friction effect of
the contact surfaces between the foam specimens and
the platens (ASTM E9-89A, 2000).
A displacement rate of 0.15 mm/min was applied to
the top surface of the specimens up to 60% strain to
Fig. 9 (continued)
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 67
obtain the entire compressive behavior of the foam spec-
imens. Because of the difculty in attaching the strain
gauge to the nonuniform surfaces of the foam specimens,
the machine compliance, which represents the relation
between the displacement of the testing machine and
the applied loading, was used to determine precise spec-
imen displacements (Jeon and Asahina, 2005).
6. Results and discussion
The forcedisplacement curves of the specimens A and
B obtained from the experimental measurements and the
numerical simulations are plotted in Figs. 10 and 11,
respectively. To dene the changes in the obtained force
displacement curves, three evaluation parameters were
introduced: the peak force in the curves, f
peak
, the displace-
ment at the peak force, d
peak
, and the elastic stiffness in the
elastic interval of the curves, f/d.
Figs. 10a and 11a showthe dependence of the forcedis-
placement curve on the power-lawhardening exponents of
the cell wall. Five different computations for each speci-
men were carried out: the same measured values of
E = 61.7 GPa and r
e
= 11.16 MPa of the cell wall base mate-
rial were used, but ve different selected values of n: n = 3,
Fig. 10. Effects of (a) the power-law hardening exponent, (b) elastic limit stress and (c) elastic modulus on the forcedisplacement curve of Specimen A.
68 I. Jeon et al. / Mechanics of Materials 41 (2009) 6073
5, 6, 8.5, 9 were used. From the two gures, it is found that
f
peak
and d
peak
of the computed curves rapidly decrease
with the increase in the power-law hardening exponent.
The % errors of the evaluation parameters,
Exp: oneNum: one
Exp: one
100, which were obtained from Figs. 10a
and 11a, are tabulated in Table 2. In Table 2, the % errors
of d
peak
obtained from the computation using n = 8.5 are
6.8692% and 8.42798% for the specimens A and B, respec-
tively. These % error values are close to the volumetric er-
rors of the specimens. Moreover, these small % error values
mean that the peak force positions in the computed force
displacement curves are horizontally very close to those of
the experimentally measured curves. For these reasons, the
power-law hardening exponent of the Al cell wall was
determined as n = 8.5 in this research.
Figs. 10b and 11b show the effects of the elastic limit
stress of the cell wall on the forcedisplacement curve.
For these computations, the same values of E = 61.7 GPa
and n = 8.5 but seven different selected values of r
e
:
r
e
= 11.16, 25, 28.5, 30, 31, 32, 34 MPa were used for each
specimen. In the two gures, d
peak
is insensitive to the
change of the r
e
but f
peak
and f/d dramatically increase with
the increase of r
e
. Table 3 shows the % errors of the evalu-
ation parameters according to the change of r
e
. The closest
values of each % error to each volume error of the speci-
mens A and B were obtained from computation using
r
e
= 31 MPa. In this case, the congurations of the com-
puted forcedisplacement curves of each specimen closely
resemble the measured ones. Therefore, r
e
= 31 MPa is
determined for the value of the elastic limit stress of the
Al cell wall.
The computed forcedisplacement curves of each spec-
imen using the same values of n = 8.5 and r
e
= 31 MPa, and
three different elastic moduli, E = 61.7, 68, and 73.1 GPa
are shown compared with the measured curves in Figs.
10c and 11c. In these gures, it is hard to nd signicant
differences among the computed curves. This means that
the effect of the elastic moduli of the cell wall in the range
of a general Al alloy, E = 61.7 73.1 GPa, on the forcedis-
placement curve is very small. Table 4 shows the % errors
of the evaluation parameters obtained from the curves. In
Table 4, all of the % error values stay within 10% and close
to the volumetric errors of the specimens. Therefore, we
selected the elastic modulus of the Al cell wall to be
E = 68 GPa, which was also measured by Sugimura et al.
(1997).
From these results, we determined the mechanical
properties of the Al cell wall as E = 68 GPa, r
e
= 31 MPa
and n = 8.5. Finally, the 0.2% offset yield stress, r
Y
, was
determined as 35.5 MPa by using the stress corresponding
to the intersection of the calculated true stress-true strain
curve, which was obtained after substituting the deter-
mined properties in Eq. (1), and the line parallel to the elas-
tic part of the curve offset by a 0.2% strain. In particular,
this r
Y
is remarkably smaller than the published values,
r
Y
= 120170 MPa (Sugimura et al., 1997; Simone and
Gibson, 1998b; Andrews et al., 1999), which were obtained
from the approximation method. However, the mechanical
properties determined in this research were veried
through the computation for the stressstrain behavior of
a 10 10 10 mm
3
closed-cell Al foam specimen under
compression and were successfully used to analyze the
plastic collapse mechanism of the foam material (see Jeon
et al., submitted for publication).
Further, we demonstrated that the power-law harden-
ing exponent of the cell wall had a signicant effect on
the magnitude of the peak force and the displacement at
the peak force in the forcedisplacement curve, and that
Fig. 10 (continued)
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 69
the elastic limit stress of the cell wall had an important ef-
fect on the magnitude of the peak force as well as the elas-
tic stiffness. However, the elastic limit stress did not affect
the displacement at the peak force. Moreover, we showed
that the elastic modulus in the range of a general Al alloy,
E = 61.7 73.1 GPa did not have any signicant effects on
the forcedisplacement curve.
Using the determined mechanical properties of the cell
wall, the compressive behavior of the two foam specimens
was numerically investigated. Fig. 12ac and Fig. 13ac
show the deformation shapes of two specimens with the
distributions of the equivalent plastic strain eld greater
than 0.2%. Note that the equivalent plastic strain less than
0.2% are regarded as elastic strain for clear evaluation of
the actual plastic strain in the specimen.
Figs. 12a and 13a represent the initial state of the spec-
imen A and B. In this stage, no equivalent plastic strain
elds were detected in the specimens. Figs. 12b and 13b
show the plastically buckled shapes of the cell wall in the
two specimens (see the regions within the closed curves)
Fig. 11. Effects of (a) the power-law hardening exponent, (b) elastic limit stress and (c) elastic modulus on the forcedisplacement curve of Specimen B.
70 I. Jeon et al. / Mechanics of Materials 41 (2009) 6073
Fig. 11 (continued)
Table 2
% Errors of the evaluation parameters according to the changes in the power-law hardening exponent, n
E (GPa) r
e
(MPa) n d
peak
error(%) f
peak
error (%) Stiffness (f/d) error(%)
Spec. A Spec.B Spec.A Spec. B Spec. A Spec. B
Cal. #1 61.7 11.16 3 97.97169 94.44165 33.73116 18.29626 46.26194 45.73207
Cal. #2 61.7 11.16 5 31.29902 25.9021 33.04793 42.87535 61.50028 60.37315
Cal. #3 61.7 11.16 6 26.23276 15.65987 44.07231 51.92538 64.59853 64.85735
Cal. #4 61.7 11.16 8.5 6.8692 8.42798 55.82837 62.49337 68.53542 72.39752
Cal. #5 61.7 11.16 9 3.39854 9.21198 57.17856 63.6846 68.60759 73.20666
Table 3
% Errors of the evaluation parameters according to the changes in the elastic limit stress, r
e
E (GPa) r
e
(MPa) n d
peak
error(%) f
peak
error(%) Stiffness (f/d) error(%)
Spec. A Spec. B Spec. A Spec. B Spec. A Spec. B
Cal. #1 61.7 11.16 8.5 6.8692 8.42798 55.82837 62.49337 68.53542 72.39752
Cal. #2 61.7 25 8.5 6.49058 8.73957 11.05809 23.88477 36.29876 27.43667
Cal. #3 61.7 28 8.5 6.64383 0.40493 22.11082
Cal. #4 61.7 30 8.5 5.24655 8.77475 4.07765 10.73255 13.1409 14.40955
Cal. #5 61.7 31 8.5 5.24655 8.78983 7.08129 8.14667 5.40935 7.87728
Cal. #6 61.7 32 8.5 6.52664 9.41803 10.04247 5.5482 0.54771 7.63744
Cal. #7 61.7 34 8.5 9.59895 0.42946 1.33613
Table 4
% Errors of the evaluation parameters according to the changes in the elastic modulus, E
E (GPa) r
e
(MPa) n d
peak
error (%) f
peak
error(%) Stiffness(f/d) error(%)
Spec. A Spec. B Spec. A Spec. B Spec. A Spec. B
Cal. #1 61.7 31 8.5 5.24655 8.78983 7.08129 8.14667 5.40935 7.87728
Cal. #2 68
a
31 8.5 8.86144 8.70942 8.51777 7.01161 1.19502 8.66801
Cal. #3 73.1 31 8.5 9.86208 9.05619 9.57739 6.14852 0.2301 6.65899
a
See (Sugimura et al., 1997).
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 71
at the compressive displacements of 0.14 mm for specimen
A and 0.155 mm for specimen B, respectively, which are
the deformation state before the peak force in Figs. 10c
and 11c. In these gures, localized plastic strain regions
around the buckled cell wall were found. Finally, Figs.
12c and 13c show the plastically collapsed cell walls of
each specimen (see the regions within the closed curves).
These gures were taken at the compressive displacements
of 0.30 mm for specimen A and 0.34 mm for specimen B,
respectively, which are the deformation state after the
peak force in Figs. 10c and 11c. At this stage, the fully
developed plastic zones were also observed in the
specimens.
7. Conclusions
A new approach was attempted to determine the
mechanical properties of the cell wall of the closed-cell
Al foam, such as the elastic modulus, 0.2% offset yield
stress, and power-law hardening exponent. To set the limit
values for the mechanical properties of the Al cell wall, the
properties of the cell wall base material was initially mea-
sured using a 6 6 6 cm
3
base material ingot. Two
5 5 5 mm
3
Al foam specimens of completely different
structures were prepared and used for uniaxial compres-
sion tests as well as for direct nite element modeling.
Numerical simulations were carried out by using the two
models and various mechanical properties selected from
the measured properties of the base material. The Al cell
wall mechanical properties were precisely determined by
comparing the computed forcedisplacement curves and
the measured ones.
Based on the results of this research, the mechanical
properties were determined as E = 68 GPa, r
Y
= 35.5 MPa
and n = 8.5. Further, the effects of each mechanical prop-
erty on the forcedisplacement curve of the foam material
were elucidated. An increase in the power-law hardening
exponent of the cell wall rapidly decreased both the mag-
nitude of the peak force and the displacement at the peak
force in the forcedisplacement curves. An increase in the
elastic limit stress of the cell wall dramatically increased
the magnitude of the peak force and the elastic stiffness
but did not change the displacement at the peak force.
Also, the elastic modulus of the cell wall in the range of a
general Al alloy, E = 61.7 73.1 GPa, did not give any sig-
nicant effects on the forcedisplacement behavior of the
foam material.
Fig. 12. The computed deformation shape of the Specimen A with its equivalent plastic strain eld at the state of (a) the initiation, (b) the cell wall buckling
and (c) the cell wall collapse.
72 I. Jeon et al. / Mechanics of Materials 41 (2009) 6073
Acknowledgement
I. Jeon thanks Dr. Tetsuji Miyoshi of Shinko Wire Co.
LTD. for supplying the base material ingot of the ALPORAS.
This study was supported by National Research Lab pro-
gram of the Korea Science & Engineering Foundation
(R0A-2006-000-10249-0).
References
Andrews, E., Sanders, W., Gibson, L.J., 1999. Compressive and tensile
behavior of aluminum foams. Mater. Sci. Eng. A 270, 113124.
E 9-89A, 2000. Standard test methods of compression testing of metallic
materials at room temperature. ASTM, Philadelphia.
Grenestedt, J.L., 1998. Inuence of wavy imperfections in cell walls on
elastic stiffness of cellular solids. J. Mech. Phys. Solids 46,
2950.
Grenestedt, J.L., Tanaka, K., 1999. Inuenceof cell shapevariations onelastic
stiffness of closed cell cellular solids. Scripta Mater. 40, 7177.
Hasan, M.A., Kim, A., Lee, H.J., 2008. Measuring the cell wall mechanical
properties of Al-alloy foams using the nanoindentation method.
Compos. Struct. 83, 180188.
Jeon, I., Asahina, T., 2005. The effect of structural defects onthe compressive
behavior of closed-cell Al foam. Acta Mater. 53, 34153423.
Jeon, I., Asahina, T., Kang, K.J., Im, S., Lu, T.J., submitted for publication.
Finite element simulation of the plastic collapse of closed-cell
aluminum foams with X-ray computed tomography.
Jeon, I., Im, S., 2001. The role of higher order eigenelds in elasticplastic
cracks. J. Mech. Phys. Solids 49, 27892818.
Kim, A., Tunvir, K., 2006. Study of Al-alloy foam compressive behavior
based on instrumented sharp indentation technology. J. Mech. Sci.
Technol. 20, 819827.
Lee, Y.H., Lee, B.K., Jeon, I., Kang, K.J., 2007. Wire-woven Bulk Kagome
(WBK) truss cores. Acta Mater. 55, 60846094.
Markaki, A.E., Clyne, T.W., 2001. The effect of cell wall microstructure on
the deformation and fracture of aluminum-based foams. Acta Mater.
49, 16771686.
McMeeking, R.M., 1977. Finite deformation analysis of crack-tip opening
in elasticplastic materials and implications for fracture. J. Mech.
Phys. Solids 25, 357381.
Miyoshi, T., Itoh, M., Akiyama, S., Kitahara, A., 2000. ALPORAS aluminum
foam: Production process, properties, and applications. Adv. Eng.
Mater. 2, 179183.
Miyoshi, T., Itoh, M., Mukai, T., Kanahashi, H., Kohzu, H., Tanabe, S.,
Higashi, K., 1999. Enhancement of energy absorption in a closed-cell
aluminum by the modication of cellular structures. Scripta Mater.
41, 10551060.
Shin, D.S., Lee, K., Kim, D., 2007. Biomechanical study of lumbar spine
with dynamic stabilization device using nite element method.
Comput. Aided Des. 39, 559567.
Simone, A.E., Gibson, L.J., 1998a. The effects of cell face curvature and
corrugations on the stiffness and strength of metallic foams. Acta
Mater. 46, 39293935.
Simone, A.E., Gibson, L.J., 1998b. Aluminum foams produced by liquid-
state processes. Acta Mater. 46, 31093123.
Sone, T., Tamada, T., Jo, Y., Miyoshi, H., Fukunaga, M., 2004. Analysis of
three-dimensional microarchitecture and degree of mineralization in
bone metastases from prostate cancer using synchrotron
microcomputed tomography. Bone 35, 432438.
Sugimura, Y., Meyer, J., He, M.Y., Bart-Smith, H., Grenstedt, J., Evans, A.G.,
1997. On the mechanical performance of closed cell Al alloy foams.
Acta Mater. 45, 52455259.
Fig. 13. The computed deformation shape of the Specimen B with its equivalent plastic strain eld at the state of (a) the initiation, (b) the cell wall buckling
and (c) the cell wall collapse.
I. Jeon et al. / Mechanics of Materials 41 (2009) 6073 73