Size Effects in Ductile Cellular Solid 2
Size Effects in Ductile Cellular Solid 2
Size Effects in Ductile Cellular Solid 2
1!2n
d
!2p
d
#4nm
d
1!2n
d
!2p
d
#4nm
, (1)
where m is the reduced sti!ness factor for the boundary layer of thickness nd and the stress free
layer of zero sti!ness is of thickness pd. Eq. (1) is plotted on Fig. 2 with m"0.85, n"0.5 and
p"0.25; it follows the general trend of the data but reaches the plateau value for the modulus at
higher values of /d.
The normalized moduli for the open-cell foam decrease more rapidly with decreasing /d than
those for the closed-cell foam. Two factors contribute to this. The cell faces in the closed-cell foam
may increase the sti!ness of its boundary layer compared with that for the open-cell foam. And the
cut cell edges of the open-cell foam are stress-free, like those of the honeycomb, while the cut cell
edges and faces at the boundary of the closed cell foam are not completely stress-free, due to the
presence of the faces.
E.W. Andrews et al. / International Journal of Mechanical Sciences 43 (2001) 701}713 709
The size e!ect for the uniaxial compressive strength of a honeycomb arises from the stress-free
walls at the boundary of the honeycomb. Analysis of the same phenomenon in foams gives [14]
'
H
'
"
(/d!1/2)
(/d)
. (2)
Eq. (2) is plotted along with the data shown in Fig. 3; it describes the data well. We note that the
normalized strength decreases more rapidly with decreasing /d for the open-cell than the
closed-cell foam, again, due to the e!ect of the cell faces.
Both our data and models suggest that the bulk values of both Young's modulus and strength
are reached at values of /d of about 5}8, lower than the value of 15 suggested by Brezny and
Green [12]. The discrepancy may be due to a di!erence in test methods. They used three point
bend specimens which subject the more compliant top and bottom surfaces of the beam to the
highest normal stresses, magnifying the size e!ect in smaller beams and increasing the critical value
of /d. Their tests were performed on a brittle reticulated vitreous carbon foam which exhibits
a Weibull size e!ect, confounding the specimen size/cell size e!ect.
The trends we observe in sti!ness and strength are similar to those of Bastawros et al. [13]. In
that study, uniaxial compression tests were performed on prismatic aluminum foam specimens of
constant length and width and varying depth. They found that sti!ness and strength became
essentially constant when the depth was greater than about 4 times the cell size.
4.2. Shear strength
The normalized shear strength of the closed-cell foam decays rapidly to a plateau value at a ratio
of the specimen thickness to cell size, t/d, of 2. Since there are no stress or strain gradients in the
out-of-plane direction in the shear test, the two-dimensional honeycomb analysis of size e!ects on
shear strength can be applied directly to foams. Assuming that, on average, the cell edges and faces
are cut at the midpoint along their length, the honeycomb analysis indicates that H
'
/H
'"''
"2 for
t/d"1, and that H
'
/H
'"''
decays to a value of unity for t/d*2 [14], in close agreement with our
experimental results (Fig. 5).
4.3. Indentation
The indentation stress varies with the inverse of the indenter diameter, D. In an indentation test,
the foam immediately below the indenter crushes and the foam at the perimeter of the indentation
yields in shear and then ruptures (Fig. 8). The total force on the indenter is the sum of the crushing
force (equal to the plastic collapse strength of the foam times the cross sectional area) and the
shearing force (equal to the shear strength of the foam times the perimeter)
P
'"
"P
'"'"
#P
'"'
"H
'
D
4
#kH
'
Dd
or
H
'"
H
'
"1#k
H
'
H
'
d
D
. (3)
710 E.W. Andrews et al. / International Journal of Mechanical Sciences 43 (2001) 701}713
Fig. 8. Optical photographs of indentation specimens. (a) Specimen loaded to the initial peak stress and then unloaded.
View from above the indentation showing torn cell walls at perimeter of indentation. (b, c) Longitudinal section of
a specimen loaded beyond the initial peak stress and then unloaded, showing the indentation, the crushed zone beneath
the indentation and torn cell walls on the back wall of the perimeter of the indentation.
Eq. (3) is compared with the indentation data in Fig. 7 using our measured values for
H
'
"1.42 MPa and H
'
"1.03 MPa. We "nd a value of k"3.98 from the "t of Eq. (3) to the data,
giving k"1. This suggests that yielding at the boundary of the indenter extends to a depth of about
1 cell into the foam. The indentation data are well described by Eq. (3).
E.W. Andrews et al. / International Journal of Mechanical Sciences 43 (2001) 701}713 711
5. Conclusions
The Young's modulus and plastic collapse strength of both the closed-cell Alporas and the
open-cell Duocel foams increased to a plateau level as the ratio of specimen size to cell size
increased. The plateau values were reached at /d"6 for the Young's modulus for both foams and
at /d"8 and 5 for the compressive strength of the open- and closed-cell foams, respectively. The
modulus and strength of the open-cell foam decreased more rapidly with decreasing /d than those
of the closed-cell foam due to the e!ect of the cell faces. Extrapolation of the analytical models for
size e!ects in honeycombs to foams gives a good description of the data for uniaxial loading.
The size e!ect for the shear strength vanishes for specimens with a thickness of at least twice the
cell size. The two-dimensional honeycomb model for the e!ect of t/d on the shear strength (which is
also applicable to foams) describes the data for the closed-cell aluminum foam well.
The indentation strength decreases as the size of the indenter increases relative to the cell size and
approaches a limiting value slightly higher than the uniaxial plastic collapse strength of the foam.
The decrease varies as the inverse of the indenter diameter. This can be understood in terms of the
observation that during an indentation test the foam immediately below the indenter crushes and
the cell walls at the perimeter of the indentation yield and then rupture. A simple model gives
a good description of the data.
Acknowledgements
We are grateful for "nancial support of ARPA (Contract number N00014-96-1-1028). Some of
the tests on the e!ect of specimen size on the compressive modulus and strength of the open-cell
foam were performed by Dr. A.E. Simone, whose assistance is appreciated. The research of Dr. P.
Onck has been made possible by a fellowship of the Royal Netherlands Academy of Arts and
Sciences. The assistance of Prof. J.W. Hutchinson is highly appreciated. Figs. 2, 4 and 5 appear with
permission, from the Annual Review of Material Science, volume 30 copyright 2000 by Annual
Reviews www.AnnualReviews.org.
References
[1] Elliot JC. Method of producing metal foam. U.S. Patent No 2,751,289, 1956.
[2] Gradinger R, Simancik F, Degischer HP. Determination of mechanical properties of foamed metals. International
Conference Welding Technology, Materials and Materials Testing, Fracture Mechanics and Quality Management,
Vienna University of Technology, 1997.
[3] Sugimura Y, Meyer J, He MY, Bart-Smith H, Grenestedt J, Evans AG. On the mechanical performance of closed
cell foams. Acta Materiala 1997;45:5245}59.
[4] Bart-Smith H, Bastawros A-F, Mumm DR, Evans AG, Sypeck DJ, Wadley HNG. Compressive deformation and
yielding mechanisms in cellular Al alloys determined using X-ray tomography and surface strain mapping. Acta
Materiala 1998;46:3583}92.
[5] Evans AG, Hutchinson JW, Ashby MF. Multifunctionality of cellular metal systems. Progress in Materials Science
1998;43:171}221.
[6] Simone AE, Gibson LJ. Aluminum foams produced by liquid state processes. Acta Materiala 1998;46:3109}23.
712 E.W. Andrews et al. / International Journal of Mechanical Sciences 43 (2001) 701}713
[7] von Hagen H, Bleck W. Compressive, tensile and shear testing of melt-foamed aluminum. Symposium R: Porous
and Cellular Materials for Structural Applications, Materials Research Society Spring Meeting 1998, San Francisco.
[8] Andrews EW, Sanders W, Gibson LJ. Compressive and tensile behaviour of aluminum foams. Materials Science
and Engineering A 1999;270:113}24.
[9] Deshpande, Fleck NA. Isotropic constitutive models for metallic foams. Journal of the Mechanics and Physics of
Solids 2000;48:1253}83.
[10] Harte A-M, Fleck NA, Ashby MF. Fatigue failure of an open cell and a closed cell aluminum alloy foam. Acta
Materiala 1999;47:2511}24.
[11] Gioux G, McCormack TM, Gibson LJ. Failure of aluminum foams under multiaxial loads. International Journal
of Mechanical Sciences 2000;42:1097}117.
[12] Brezny R, Green DJ. Characterization of edge e!ects in cellular materials. Journal of Materials Science
1990;25:4571}8.
[13] Bastawros A-F, Bart-Smith H, Evans AG. Experimental analysis of deformation mechanisms in a closed-cell Al
alloy foam. Journal of the Mechanics and Physics of Solids 2000;48:301}22.
[14] Onck PR, Andrews EW, Gibson LJ. Size e!ects in ductile cellular solids Part I: Modelling. International Journal of
Mechanical Sciences 2001;43:681}99.
E.W. Andrews et al. / International Journal of Mechanical Sciences 43 (2001) 701}713 713