Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Metallic syntactic foams synthesis, characterization and mechnical properties
(USC Thesis Other)
Metallic syntactic foams synthesis, characterization and mechnical properties
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
METALLIC SYNTACTIC FOAMS
SYNTHESIS, CHARACTERIZATION
AND
MECHANICAL PROPERTIES
by
Gerhard Castro
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirement for the Degree
DOCTOR OF PHILOSOPHY
(Materials Science)
March 2013
ii
Acknowledgements
I am deeply grateful to my advisor, Professor Steve Nutt, for his continuous
support and valuable guidance through the program. His enthusiasm and
encouragement made my achievements today possible, and it will continue
to be an inspiration throughout my career ahead.
I would like to thank to my thesis committee for taking the time to review
and evaluate my work. The committee members include Professor Edward
Goo and Professor Charles G. Sammy. I wish to thank to Mr. Roy Tsai, Mr.
Ezra Pryor, Mr. Josue Vega, Mr. Bo Yin and Mr. Wenchen Xu for their
assistance with the experiments.
Also, I want to thank all the members of my family for their unflagging love,
encouragement and support throughout my life; specially my wife Georgina
and my son Geramie.
This work was supported by M.C.Gill Co.
iii
Table of contents
Page
Acknowledgements ii
List of figures vii
List of tables xiv
Glossary xv
Abstract xvi
1. – Introduction 1
2. - Background. 4
2.1. - Cellular metal. 4
2.2. - Syntactic metal foams. 6
2.3. – Composite metal foams using liquid metallurgy method. 8
2.3.1. – Aluminum matrix-ceramic microsphere syntactic foam. 8
2.3.2. – Aluminum matrix-ceramic microsphere syntactic foam
through stir-casting technique.
9
2.3.3. – Aluminum matrix–steel microsphere syntactic foam -
North Carolina State University.
10
2.3.4. – Aluminum/magnesium matrix–steel microsphere syntactic
foam- Georgia Institute of Technology.
12
2.3.5. – Magnesium matrix – ceramic microsphere syntactic foam. 13
2.3.6. – Lead matrix – fly ash microsphere syntactic foam. 14
iv
2.3.7. – Aluminum matrix- hollow SiC spheres (Deep Springs
Technologies- CPS Technologies)
15
2.4. – Open cell metal foams using replication casting method. 16
2.4.1. – Cast open-cell aluminum foams using inorganic granules. 16
2.4.2. – Open-cell brass foam using silica-gel beads. 17
2.4.3. – Open-cell aluminum foams using soft alumina balls. 18
2.4.4. – Porous Gray Iron Foams Castings using sand balls. 20
2.4.5. – Dual phase steel foam using SiC performs. 21
2.4.6. – Ni-Mo-Cr open-cell foams using sodium aluminate preform. 22
2.4.7. – Amorphous open-cell metallic foams using sintered salt
preforms.
23
2.5. – Mechanical properties of metal foams. 24
2.6. – Energy absorption. 28
2.7. – Applications of metal foam. 30
3. - Synthesis of syntactic steel foam using mechanical pressure
infiltration.
32
3.1. - Experimental procedures. 32
3.1.1. – Materials. 32
3.1.2. - Foam production. 33
3.1.3. - Density calculation. 34
3.1.4. - Microstructure of the syntactic foam. 35
v
3.1.5. – Compression. 38
3.2. - Results and Discussion. 38
3.2.1. - Microsphere preheatment temperature. 38
3.2.2. - Plunger infiltration travel. 40
3.2.3. - Amount of microspheres. 40
3.2.4. - Interfacial reactions. 41
3.2.5. - Mechanical properties. 43
3.2.6. - Compression strength. 45
3.2.7. - Energy absorption of steel syntactic foams. 46
4.- Synthesis of syntactic steel foam using gravity-fed infiltration. 50
4.1. – Experimental. 50
4.1.1. – Materials. 50
4.1.2. - Foam production. 51
4.1.3. – Density. 52
4.1.4. – Microstructure. 53
4.1.5. – Compression. 55
4.2. - Results and Discussion. 55
4.2.1. - Microsphere preheatment temperature. 55
4.2.2. - Interfacial reactions. 56
4.2.3. - Mechanical properties. 57
4.2.4. - TRIP steel syntactic foam. 62
vi
4.2.5. - Effect of relative density on the mechanical properties. 63
5. - Synthesis of syntactic aluminum foam using gravity-fed
infiltration.
69
5.1. Experiment. 69
5.1. 1. Materials. 69
5.1. 2. Foam production. 70
5.1. 3. Compression. 71
5.1. 4. Impact test. 71
5.2. Results and discussion. 73
5.2.1. Quasistatic compression testing. 73
5.2.2. Impact response of aluminum syntactic foam. 75
5.2.3. Impact response of aluminum syntactic foam with facesheet. 82
5.2.4. The effect of microsphere size. 86
6. – Conclusions. 90
7. - Future actions. 93
8. – References. 95
vii
List of figures
Page
Figure 1. - 2D slices of X-ray tomograms of AlSi6Cu4 (a) and
AlSi7 (b) foams. [4]
5
Figure 2. - Honeycomb core structures: (A) hexagonal, (B) square
and (C) triangular shaped. [11]
5
Figure 3. – Lattice architectures: (A) tetrahedral, (B) pyramidal
and (C) kagome. [12]
6
Figure 4. – (a) Typical morphologies of the hollow CM, (b) Optical
micrograph showing the cross section of typical aluminum matrix
syntactic foam. [13]
9
Figure 5. – Optical micrographs of the as cast composite foam:
(a) 20 vol %, (b) 50 vol %. [21]
10
Figure 6. – Digital images showing the cells structure in (a)
aluminum–low carbon steel composite metal foam and (b)
aluminum-stainless steel composite metal foam with ϕ 3.7 mm
microsphere and sphere wall thickness of 200 μm. [24]
11
Figure 7. – (a) Schematics of aspiration casting apparatus. (b)
Fe-Cr hollow spheres in AZ31B Mg allow matrix. Inset: Diffusion
bonded preform, ~ 2 cm diameter. [25]
13
Figure 8. – Closed spherical cell structure of a syntactic magnesium
viii
foam revealing the thin walls of the hollow ceramic spheres. [26] 13
Figure 9. – (a) Schematics of infiltration casting apparatus used to
fabricate syntactic metal foams. [26]
14
Figure 10. – (a) Schematics illustration of the experimental setup.
(b) Microstructure of the central portion of a 50% vol. fly-ash
composite foam. [27]
15
Figure 11. –Panels made with 3mm diameter hollow SiC spheres. 16
Figure 12. – (a) Process for making porous metal foams. (b) Pore
structure of open-cell aluminum foam. [28]
17
Figure 13. – (a) Scheme of the infiltration process. (b) Final piece
of foam after final machining and SiO2 dissolution
(specimen diameter= 57.3 mm). [33]
18
Figure 14. – (a) A representative picture showing ceramic balls after
random packing (original ball size= 5 mm). (b) A representative
picture of foam produced by this method. [34]
19
Figure 15. – (a) Sand balls. (b) Sample with porosity of 72.08%. [35] 20
Figure 16. – Dual phase steel foams having different porosity. [36] 21
Figure 17. – Metallographic cross-section of Ni-Mo-Cr replicated
(46% relative density) after space-holder removal. [37]
22
Figure 18. –Uniform macrostructure of the amorphous open-cell
metallic foams. [38].
23
ix
Figure 19. – Schematics compressive stress-strain curve for foams,
showing the three regimes of linear elasticity, plastic yielding
and densification. [39]
24
Figure 20. – Cubic unit cell as provided in cellular solids by
Gibson and Ashby. [39]
25
Figure 21. – Stress-strain response for and elastic slid and a foam
made from the same solid, showing the energy absorbed
at stress σp. [39]
28
Figure 22. – Stress-strain responses are measured at a single
strain rate. [39]
29
Figure 23. – Infiltration process: (a) Melting, (b) Adding the
microspheres, (c) Initiation of the infiltration, (d) Infiltration finished.
34
Figure 24. – Medium carbon syntactic steel foam. 36
Figure 25. – (a) Microstructure of the low carbon steel syntactic
foams (left). (b) Microstructure of the medium carbon steel syntactic
foams (right).
37
Figure 26. – (a) The wall of an alumina microsphere observed
in high magnification using back-scattered electron (BSE) imaging,
also indicating the line of scan. (b) EDS line scan at the interface
between the steel and alumina microsphere.
43
Figure 27. –Stress-strain response of low carbon (bottom)
x
and medium carbon (top) steel foam. 44
Figure 28. – Compression of medium carbon syntactic steel foam:
showing different stages of compression: 0%, 10%, 20%, 30%,
40% and 50% strain.
46
Figure 29. – Stress-strain curves of our foams and steel
foams from reference 1 and 2.
47
Figure 30. – Gravity pressure infiltration process. 52
Figure 31. – Samples of steel syntactic foam produced by
gravity-pressure infiltration.
54
Figure 32. –Microstructure of the steel syntactic foam samples
(A, B, C, D, E).
54
Figure 33. – (a) The wall of an alumina microsphere observed
in high magnification using back-scattered electron (BSE)
imaging, also indicating the line of scan. (b) EDS line scan
at the interface between the steel and alumina microsphere.
56
Figure 34. – Stress-strain curves for steel foams of different C
content. (a) Samples A, B and C. (b) Samples C, D and E.
58
Figure 35. – Compression of sample B: showing different
stages of compression: 0%, 10%, 20%, 30%, 40% and 50% strain.
61
Figure 36. – Compression of sample E: showing different
stages of compression: 0%, 10%, 20%, 30%, 40% and 50% strain.
61
xi
Figure 37. – Stress-strain curves of samples A, B, C and TRIP steel. 62
Figure 38. – (a) Schematics showing the use of wire mesh
in the manufacture of syntactic foams. (b) Sample of steel
syntactic foam with relative density of 0.60.
64
Figure 39. – (a) Stress-strain curves for steel syntactic foam samples
with relative density of 0.46, 0.60, 0.68, 0.75 and 1.
66
Figure 40. – Energy absorption per unit mass vs. density: aluminum
egg-box [57], aluminum foam [50], hollow truss metallic lattice [56]
and 0.6 RD steel syntactic foam.
67
Figure 41. – Sample of aluminum syntactic foam produced by
gravity-fed infiltration.
71
Figure 42. – (a) A picture and (b) schematic view of the DynaTup
9250HV Instron testing machine.
73
Figure 43. – Stress –strain curves for aluminum syntactic. (a) Al
6061-4.45 mm microsphere diameter. (b) Al 1100-4.45 mm
microsphere diameter. (c) Al 1100-3.05 mm microsphere diameter.
75
Figure 44. – Comparison of load –displacement behavior of ASF
(1100 aluminum matrix- 4.45 mm microsphere diameter) tested
with different impact energies: (a) 60 J (b) 120 J (c) 180 J.
79
Figure 45. – Side, top and bottom views of the impacted specimens of
ASF (1100 aluminum matrix- 4.45 mm microsphere diameter)
xii
tested with different impact energies: (a) 60 J (b) 120 J (c) 180 J.
Samples size is 93 mm x 93 mm x 12.7 mm.
79
Figure 46. – Comparison of load –displacement behavior of
ASF (6061 aluminum matrix- 4.45 mm microsphere diameter)
tested with different impact energies: (a) 60 J (b) 120 J (c) 180 J.
81
Figure 47. – Side, top and bottom views of the impacted specimens
of ASF (6061 aluminum matrix- 4.45 mm microsphere diameter)
tested with different impact energies: (a) 60 J (b) 120 J (c) 180 J.
Samples size is 93 mm x 93 mm x 12.7 mm.
81
Figure 48. – Comparison of load –displacement behavior of
ASF -1100 aluminum matrix- 4.45 mm microsphere diameter
+ aluminum face sheet tested with different impact energies:
(a) 60 J (b) 120 J (c) 180 J.
83
Figure 49. – Side, top and bottom views of the impacted specimens
of ASF -1100 aluminum matrix- 4.45 mm microsphere diameter
+ aluminum face sheet tested with different impact energies:
(a) 60 J (b) 120 J (c) 180 J. Samples size is 93 mm x 93 mm x 12.7 mm.
84
Figure 50. – Comparison of load –penetration depth behavior of
ASF (1100 aluminum matrix- 3.05 mm microsphere diameter) with
different impact energies: (a) 60 J (b) 120 J (c) 180 J.
85
Figure 51. – Side, top and bottom views of the impacted specimens
xiii
of ASF (1100 aluminum matrix- 3.05 mm microsphere diameter)
with different impact energies: (a) 60 J (b) 120 J (c) 180 J.
Samples size is 93 mm x 93 mm x 12.7 mm.
86
Figure 52. – Stress-strain curves of steel syntactic foam, aluminum
syntactic foam and conventional aluminum foam.
88
Figure 53. – (a) CT scan of a group of packed plastic spheres
simulating the hollow alumina microspheres (diameter 4.45 mm).
(b) Model of the steel syntactic foam in ABAQUS.
94
Figure 54. – Applications of cellular material according to the
degree of openness needed and whether the application is more
functional or structural. [69]
94
xiv
List of tables
Page
Table 1. - Potential applications for metal foams. [39] 31
Table 2. - Typical chemical analysis of alumina microballoons. 33
Table 3. – Comparison of physical properties of our samples with
other steel foams.
49
Table 4. – Chemical composition of the investigated ferritic-pearlitic
steels.
50
Table 5. – Typical chemical analysis of alumina microballoons. 51
Table 6. – Energy absorption of the steel syntactic foam with different
steel matrices.
59
Table 7. – Energy absorption of the steel syntactic foam with
different relative densities and aluminum foam with relative density
of 0.148.
65
Table 8. – ASF samples parameters, testing condition and impact
results.
82
xv
Glossary
Symbols
t : Cell wall thickness.
l : Cell edge length.
*
: Density of foamed material.
s
: Density of the cell wall material.
I : Second moment of area of cell edge.
F : Compression force.
Es : Young’s modulus of cell wall material.
σ : Yield strength of the cell wall material.
ε : Strain.
δ : Deflection of cell edge.
E* : Young’s modulus of foamed material.
C1 : Constant.
K IC : Fracture toughness.
x : Thickness of metallic foam.
Ev : Energy absorption per unit of volume.
A : Contact area of the metal foam.
M : Mass.
v : Velocity.
xvi
Abstract
In this study, we report two procedures for producing lab-scale syntactic
steel by melt infiltration of millimeter-sized alumina microspheres:
mechanical pressure infiltration and gravity-fed infiltration. Both methods
yield foam with uniform distributions of microspheres and negligible
unintended porosity. The most critical parameters in the manufacture of the
syntactic steel foams are the melt temperature and the preheat temperature of
the microspheres prior to infiltration. The preheatment temperature of the
microspheres must be close to the melting temperature of steel.
Syntactic steel foams with relative density of about half of solid steel
densities were produced using monosized microspheres randomly situated in
a mold. Microspheres with a diameter of 1.27 mm were used for the
mechanical pressure infiltration method and microspheres with a diameter of
4.45 mm for the gravity-fed infiltration method. Different steel chemical
compositions were selected to produce steel foams of different inherent yield
strength: including several ferritic-pearlitic steels and one TRIP steel
(TRansformation-Induced Plasticity). The resultant foams were
characterized by chemical and microstructural analysis. The microstructure
of the samples consisted of blends of ferritic and pearlitic constituents in
varying proportions for the ferritic-pearlitic steels, while the cast TRIP steel
matrix presented an austenitic microstructure.
xvii
The basic mechanical properties of the steel syntactic foams were studied
under compression loading. The pearlitic syntactic foams have greater
compression strength and energy absorption capacity than the ferritic
syntactic foams, but the TRIP steel syntactic foam exhibited the highest
compression strength and highest energy absorption capacity. The
properties of the steel syntactic foams were compared to those of other steel
foams, aluminum foams and other cellular structures reported in the
literature.
We present also the compression and impact behavior of aluminum syntactic
foams (ASF) produced by gravity-fed infiltration of millimeter-sized
ceramic microspheres. Aluminum syntactic foams with relative density of
0.46 were produced using monosized microspheres (4.45 mm and 3.05 mm)
randomly situated in a mold and two types of aluminum alloy matrices: 1100
and 6061. The impact behavior was experimentally investigated using a
drop-weight testing machine. The impact tests were carried out using a
hemispherical indenter (16.1 mm diameter) on ASF plates (93 mm x 93 mm
x 12.7 mm thick). We have studied the influence of the type of aluminum
matrix, size of microspheres and the addition of a face sheet into the impact
behavior of ASF. Results show that 1100 Al alloy outperforms 6061 Al
alloy, it can absorb higher amount of energy at higher velocities
(penetration); at lower velocities both absorb the same amount of energy
(equal-energy interval). The use of smaller microspheres decreases the
amount of energy absorbed compared to larger microspheres. The use of
face sheet increases significantly the energy absorption capacity aluminum
syntactic foams.
xviii
Metal syntactic foam (steel and aluminum) offers potential advantages over
conventional aluminum foams and other metallic cellular structures, the
inherent strength of metal combined with the reduced density and lack of
defects in the regular structure of the syntactic foam presents an attractive
material with excellent strength, modulus and energy absorption. However,
thus far there have been few reports describing efforts to produce steel and
aluminum syntactic foams, and these have relied on powder metallurgical
approaches as opposed to molten state processing. Problems and challenges
for achieving metal syntactic foam with lower relative densities, higher
energy absorption capacity and the scaling up of the synthesis processes are
discussed.
1. – Introduction.
Prior efforts to produce steel foams have involved the use of powder
metallurgy techniques [1, 2]. One such technique involved the use of hollow
ceramic/metallic microspheres to produce syntactic foams. Interstitial spaces
between close-packed microspheres were filled with steel powder, and the
mixture was subsequently sintered to produce the final syntactic foam [2]. A
second approach involved the use of granular carbonate compounds as
foaming agents, which were blended and compacted with steel powder, then
melted to expand the foam (stochastic foams) [1].
Synthesis of syntactic steel foams using liquid metal presents processing
challenges because of the high temperatures involved and the need for a
force to promote melt infiltration into ceramic microspheres. Despite these
challenges, attempts have been made to produce steel foams from molten
steel via the Gasar process [3], which is based on the nucleation of gas
during solidification of a supersaturated steel melt. Syntactic steel foams
represent a new class of metal foams. There have been only a few reports
describing the liquid-processing approach and mostly employed to produce
aluminum and magnesium syntactic foams. The procedures used to make
these syntactic foams (e.g., vacuum, gas pressure, centrifugal force, stir
casting technique, etc) cannot be readily used to process ferrous alloys. The
proposed methods of producing syntactic steel foam is based on simple
methods: mechanical pressure infiltration and gravity-fed pressure
infiltration - no vacuum, gas pressure, centrifugal force or stir casting
2
methods are involved – and the method requires only low-cost materials
(low-carbon steel and alumina microspheres). Hollow alumina microspheres
are used because of the inherent resistance to high temperatures (molten
steel) and high strength.
A customized induction furnace for melting small amounts of steel (less than
one kilogram) was designed and built for this purpose. One of the limitations
of using an induction furnace for melting steel (as opposed to a resistance
furnace) is that the temperature of the melt is not easily measured, either
intermittently or continuously. Using a common immersion pyrometer was
not possible because of the small interior diameter of our crucible (38.1 mm).
Despite the limitations associated with the lab-scale setup, the proposed
synthesis method, when sufficiently developed, permits lab-scale trials of
melting steel to produce steel syntactic foams [1]. Aluminum syntactic
foams were also manufactured by the gravity-fed pressure infiltration
method in a resistance furnace. Although they are not as energy absorbent
than steel syntactic foams, they are certainly more energy absorbent than
conventional aluminum foams and they represent an attractive material as an
impact energy absorbent material.
The object of the investigation is to study the feasibility to produce metal
foams using liquid processing with higher levels of strength and energy
absorption capacity than the metal foams currently reported in the literature.
Specifically, syntactic steel foams based on steel-hollow alumina
microspheres and also aluminum-hollow alumina microspheres. After
characterization and evaluation of the mechanical properties of the produced
3
metal syntactic foams, the performance of the syntactic metal foam will be
compared to currently published data of other cellular materials. Syntactic
metal foam are designed to improve on the recognized deficiencies of
conventional metal foam. Syntactic metal foams are suitable candidate
materials in motor vehicles to improve crashworthiness, for armor
applications in military vehicles and structures, and for blast resistant
structures. In order to fulfill this goal, the energy absorbed per unit weight
and volume of the metallic syntactic foam have to be maximized to comply
with the strict requirements of minimum weight and volume in most motor
vehicles. Thus, deploying syntactic metal foam in energy absorber
applications would produce space and weight savings in vehicles.
4
2. – Background.
2.1. - Cellular metal.
A cellular solid is one composed of an interconnected network of solid struts
which form the edges or faces of a cell. One class of cellular metallic
structures is known as metal foams. Generally there is no a clear accepted
definition for the term “metal foam”. These metal foams can exist in two
kind of structures that are open-cell and closed-cell foams. The open cell
structure is contained in the cell edges only so those cell connect through
open faces. However, the closed-cell structure is consisted of cell faces and
edges so that each cell is sealed off from its neighbor.
Typical metal foams (stochastic metal foams) are the result of blowing
bubbles in liquid metal using a gas release agent, it is well know these
metallic foams often contain many structural defects and a irregular structure
[4]. The fundamental deficiencies that stochastic foams present are wavy
distortions of the cell walls, cell wall thickness variations, broken cell walls
and non-uniform shape and size of the cells. This heterogeneity leads to non-
uniform, variable, anisotropic material properties, which are problematic
when attempting to design metal foams into new applications (Figure 1).
5
Figure 1. - 2D slices of X-ray tomograms of AlSi6Cu4 (a) and AlSi7 (b) foams. Images
represent approximately the central section of the foams along the foaming direction.
Some broken cell walls are indicated by arrows. The broken lines indicate
interconnection of cells. [4]
Another broad class of cellular solids corresponds to periodic cellular metals.
Examples of periodic cellular structures are honeycomb, prismatic
(corrugation) and lattice truss [5, 6, 7, 8, 9, 10, 11]. The closed-cell
honeycomb configurations can be hexagonal, triangular, and square shaped
(examples shown in Figure 2).
Figure 2. - Honeycomb core structures: (A) hexagonal, (B) square and (C) triangular
shaped. [11]
6
The open space volume in periodic cellular metal can be further increased by
aligning slender beams instead of plates in different configuration as the core.
Figure 3 shows tetrahedral, pyramidal and kagome lattice structures. These
structures have free flowing channels in two or three directions. They
become very mechanically efficient as the relative density decreases [12].
Figure 3. – Lattice architectures: (A) tetrahedral, (B) pyramidal and (C) kagome. [12]
2.2. - Syntactic metal foams.
Syntactic metal foams are the alternative to solve the problems encountered
in stochastic foams. The syntactic metal foams often contain few or none
structural defects and they posses a regular structure. Metal matrix syntactic
foams are a relatively new class of materials. Metal matrix syntactic foams
are composite materials consisting of hollow or porous ceramic
microspheres embedded in a continuous metal matrix. They have the ability
to absorb impact energy by plastic deformation of the metallic matrix and
the collapse of the hollow microspheres at a relatively high plateau stresses.
Compared to stochastic metal foams and polymer foams, they have the
advantage of having higher compressive strength and more homogenous
mechanical properties and better energy-absorbing capability due to
extensive strain accumulation. They can be used at high temperatures and
7
harsh environments. One disadvantage is that syntactic metal foams usually
have higher relative densities than other types of cellular structures.
There are only few reports on steel syntactic foams and these are based on
powder metallurgy methods. Most of the work on metal matrix syntactic
foams has been done on systems of aluminum and magnesium matrices. The
main reason is their low melting points which make their manufacture easier
compared to steel matrices. Processing of metal syntactic foams with high
melting temperature metals presents numerous challenges. There is a need to
develop steel syntactic foams using liquid metallurgy to widen the range of
applications of this class of materials. Steel syntactic foams could have
applications as lightweight steel structures and as energy absorbers against
impact and crash in commercial and military vehicles.
Casting syntactic metal foams is done by pouring hot liquid metal into and
around a space-holder material. If the space-holder material (a group of
hollow microspheres) remains embedded in the cellular solid, the result is a
metallic syntactic foam. The space holder material is different from the
matrix material; usually the melting point of the space-holder material must
be higher than the metal matrix. This kind of syntactic foams have received
the name of composite metal foams because they consist of three phases:
metal matrix, the space-holder material and the gas phase. If the space-
holder material (a preform or a group of solid microspheres) is removed
after infiltration of the liquid metal, the result is an open-cell metal foam.
This method is known as replication casting, this method has been
8
successfully used to produce open cell foams using different kind of
preforms and solid microspheres.
2.3. – Syntactic metal foams using liquid metallurgy method.
2.3.1. – Aluminum matrix-ceramic microsphere syntactic foam.
In the Engineering Department at the University of Liverpool (UK), Prof. Y.
Zhao used 6082 aluminum alloy and ceramic microsphere (CM) powder to
fabricate Al matrix syntactic foams samples [13]. The particles of the CM
powder were either porous or hollow (Figure 4a).The samples were
manufactured by a melt infiltration process: a block of 6082 aluminum alloy
was placed at the top of predetermined amount of CM powder contained in
a steel tube and was heated in a electric furnace at 700 °C for 30 min. The
assembly was removed from the furnace and the molten Al alloy was
pressed into the CM powder. Figure 4b shows the micrograph of the cross
section of a typical aluminum matrix syntactic foam. In all the syntactic
foams, the CMs are randomly and homogenously distributed in the
aluminum matrix and account for around 60% in volume. The measured
densities of the syntactic foam samples varied slightly in a narrow range of
1.4-1.45 g/cm3.
9
Figure 4. – (a) Typical morphologies of the hollow CM, (b) Optical micrograph showing
the cross section of typical aluminum matrix syntactic foam. [13]
These are additional studies that have manufactured aluminum syntactic
foams using ceramic microspheres by melt infiltration process: G.H. Wu
[14], P.K. Rohatgi [15], I. N. Orbulov [16], D.K. Balch [17], Q.Zhang [18],
R.A. Palmer [19] and X.F. Tao [20].
2.3.2. – Aluminum matrix-ceramic microsphere syntactic foam through stir-
casting technique.
New syntactic foam composites comprising ZnAl22 eutectoid alloy and
varying volume fraction (6, 15, 20, 25, 35, 50 vol %) of fly ash microsphere
were successfully prepared through stir-casting method [21]. The
microspheres used were Ni-coated fly ash with an average diameter of
150μm and a wall thicknesses of 10 μm. A stirrer was introduced into the
melt when the temperature reached 600 °C. The fly ash microballoons were
added manually to the vortex formed. Just before pouring, the melt was
again stirred for 1 min and immediately cast into cast iron mold. For
10
metallographic examination, samples cut from the foams were machined,
polished as per conventional polishing techniques and etched to reveal the
distribution of the microballoons and phase constituents (figure 5).
Figure 5. – Optical micrographs of the as cast composite foam: (a) 20 vol %, (b) 50 vol %.
[21]
These are additional studies that have manufactured aluminum syntactic
foams using ceramic microspheres through stir-casting technique: D. P.
Mondal [22], A. Daoud [23].
2.3.3. – Aluminum matrix–steel microsphere syntactic foam- North Carolina
State University.
In the Department of Mechanical and Aerospace Engineering at the North
Carolina State University, a new kind of syntactic foams was developed:
aluminum–steel composite metal foam [24]. The gravity casting technique
11
was used to infiltrate steel hollow spheres (low carbon steel and stainless
steel hollow spheres) with an aluminum alloy matrix A356 (Al–7% Si alloy).
The hollow spheres were produced by a powder metallurgy process at
Fraunhofer Institute, Germany. The outer diameter of the hollow spheres
used in this gravity casting experiment was 3.7 mm with 200 um wall
thickness. The spheres were placed in a mild steel permanent casting mold,
vibrated to pack in a maximum dense arrangement and pre-heated to either
700 °C or 740 °C. The preheating of the hollow spheres prevented premature
solidification of aluminum on contacting the spheres and mold during
casting (Figure 6).
Figure 6. – Digital images showing the cells structure in (a) aluminum–low carbon steel
composite metal foam and (b) aluminum-stainless steel composite metal foam with ϕ 3.7
mm microsphere and sphere wall thickness of 200 μm. [24]
12
2.3.4. – Aluminum and magnesium matrix–steel microsphere syntactic
foam- Georgia Institute of Technology.
The Georgia Institute of Technology has produced composite syntactic foam
consisting of thin-wall, hollow Fe-Cr stainless steel spheres cast in various
metal matrices: including aluminum alloys 6061, 7075, 413, magnesium
alloy AZ31B and unalloyed aluminum and magnesium [25]. Hollow spheres
were formed from slurries of fine particle size iron and chromium oxides via
the coaxial nozzle process (1-10 μm diameter range, 6 um average diameter).
The oxides were reduced and sintered in hydrogen to the corresponding Fe-
Cr alloy. The preform consisted of packed individual spheres or a rigid,
diffusion bonded structure.
The aspiration casting apparatus is shown schematically in Figure 7a,
consisted of a fused silica casting tube, an intermediate vacuum chamber,
and a rough vacuum pump. The casting tube contains the preform to be
infiltrated. During the casting process, the matrix material is superheated in a
crucible to approximately 100 °C above the liquidus temperature. The end of
the casting tube is then lowered into the melt. When the vacuum pump is
turned on, molten metal flows into the tube and infiltrates the preform as the
tube is evacuated. Microstructural analysis revealed no porosity in any of the
castings (i.e. complete infiltration), and complete wetting of the spheres by
the matrix as shown in Figure 7b.
13
Figure 7. – (a) Schematics of aspiration casting apparatus. (b) Fe-Cr hollow spheres in
AZ31B Mg allow matrix. Inset: Diffusion bonded preform, ~ 2 cm diameter. [25]
2.3.5. – Magnesium matrix – ceramic microsphere syntactic foam.
Scientists in the Department of Material Science at the University of
Erlangen in Germany have developed magnesium syntactic foam utilizing
hollow alumina spheres [26]. In this technology, thin walled hollow alumina
spheres are embedded in a magnesium matrix. Liquid magnesium infiltrates
an alumina sphere array using a pressure casting technique.
Figure 8. – Closed spherical cell structure of a syntactic magnesium foam revealing the
thin walls of the hollow ceramic spheres. [26]
14
The alumina spheres were processed using a patented powder slurry
sintering process. Four different spheres geometries were used 2.8 mm
outside diameter, 133 μm wall thickness; 2.8 mm outside diameter, 181 μm
wall thickness; 3.7 mm outside diameter, 115 μm wall thickness; 3.7 mm
outside diameter, 150 μm wall thickness. Four magnesium matrix alloys
were used. A schematic of the infiltration casting technique is shown in
Figure 9. The resulting composite foam contains closed cells of homogenous
and isotropic morphology. The densities of the syntactic magnesium foams
were between 1.0 and 1.4 g/cm
3
. The volume fraction of the spheres kept
constant at approximately 63%.
Figure 9. – (a) Schematics of infiltration casting apparatus used to fabricate syntactic
metal foams. [26]
2.3.6. – Lead matrix – fly ash microsphere syntactic foam.
In this work, the squeeze infiltration process was employed to synthesize a
lead-fly ash composite foam [27]. The mean size of the fly-ash particles was
170 μm. The fly ash performs were prepared by wet forming, which
15
consisted of compressing, within a split die, a slurry mixture containing the
fly-ash particles, urea, birch sawdust and a binder. The lead-fly ash
composite was produced by infiltration of molten lead into the porous
preform under pressure. The preform was located in the mold chamber and
then preheated to 360 °C. Molten lead was then poured into the infiltration
mold. After pouring, a squeeze pressure was applied to the melt using a ram.
This pressure was quickly raised to and held at 200 MPa to force the melt to
totally infiltrate the porous preform.
Figure 10. – (a) Schematics illustration of the experimental setup. (b) Microstructure of
the central portion of a 50% vol. fly-ash composite foam. [27]
2.3.7. – Aluminum matrix- hollow SiC spheres (Deep Springs Technologies-
CPS Technologies)
They have produced Al matrix filled with hollow ceramic ‘bubbles’ to create
structures that weigh 40% less than aluminum. This panel was made with
3mm diameter hollow SiC spheres. The process encapsulates hybrid ceramic
16
modules using a combination of materials to encapsulate a tile or matrix of
tiles, into a by infiltrating with aluminum at high pressure. The infiltration
process used resembles Resin Transfer Molding (RTM) used in polymer
composites.
Figure 11. –Panels made with 3mm diameter hollow SiC spheres.
2.4. – Open cell metal foams using replication casting method.
2.4.1. – Cast open-cell aluminum foams using inorganic granules.
Prof. J. Banhart at the Fraunhofer-Institute for Manufacturing and Advanced
Materials developed a casting method for producing open-cell aluminum
foams [28]. The process has three steps: preparation of space-holding filler
17
(inorganic granules); infiltration of the filler with hot liquid metal, an
external pressure has to be applied to ensure a complete filling of the
interstices between the granules; removal of filler granules by leaching or
thermal treatment (Figure 12a). The materials have open pores that can be
seen in figure 12b. The density of the open-cell aluminum foams described
usually range from 900 to 1200 kg/m
3
. This corresponds to a porosity of 55
to 67%.
Figure 12. – (a) Process for making porous metal foams. (b) Pore structure of open-cell
aluminum foam. [28]
These are additional studies that have manufactured aluminum open-cell
foams using the casting replication technique: Q. Fabrizio 29], S. Yu [30],
W. Lucai [31] and B.Y. Hur [32].
2.4.2. – Open-cell brass foam using silica-gel beads.
This work introduces the use of amorphous SiO2 (silica gel) particles as
space holder and wet solutions of HF as solvent for Cu-based open cell foam
production [33]. Commercial silica-gel beads have been used as space
holders. The beads have been randomly piled-up in the crucible and then
18
gently shaken for improving the packing. The infiltration process is
performed in an induction vacuum melt casting machine. A small cylinder of
solid brass is placed at the bottom of a graphite crucible. Then, the bed of
silica-gel beads is placed on top of the brass cylinder. After reaching the
superheat of 100 °C above the meting temperature, the induction power is
turned off and the silica –gel bed is pressed down by a steel cylinder (Figure
13a). The infiltrated metal can solidify between the silica particles. Finally,
the piece is submerged in aqueous HF bath (25% vol.) till complete SiO2
dissolution (Figure 13b).
Figure 13. – (a) Scheme of the infiltration process. (b) Final piece of foam after final
machining and SiO 2 dissolution (specimen diameter= 57.3 mm). [33]
2.4.3. – Open-cell aluminum foams using soft alumina balls.
Open-cell aluminum foams were made from a novel process using soft
alumina balls [34]. Since the ceramic balls are deformable, they can be
compressed into a high dense packing. As a result, the product porosity after
casting reached around 90 %. These ceramic balls can withstand the casting
temperature and pressure and are easily removed by running water. The soft
19
ceramic balls were made of coarse α-alumina particles, polymeric additives,
small amounts of bentonite as the sintering aid and water. The ceramic balls
were the put into a cylindrical mold in a layer-by-layer manner (Figure 14a).
Finally, the casting was carried out putting a pre-determined weight of
aluminum in circular shape on top of the ceramic bed. The whole mold was
then put into the top-open furnace and its temperature was raised to 740 °C,
held for about 1.5 h to completely melt the aluminum. Then a small pressure
was applied to push the molten aluminum though the bed. After removing
the cast from the mold, it was then put in a water bath. The ceramic balls
were removed with the aid of ultrasonic vibration, an open-cell aluminum
foam was therefore obtained (figure 14b).
Figure 14. – (a) A representative picture showing ceramic balls after random packing
(original ball size= 5 mm). (b) A representative picture of foam produced by this method
(size of pore is about 4mm). [34]
20
2.4.4. – Porous Gray Iron Foams Castings using sand balls.
A technique has been developed for preparation of porous gray cast iron
foam [35]. It was carried out in the Institute of Technology Coimbatore
(India). The technique is based on casting metal around sand balls. This
process has the following three steps: preparation of space-holder filler (sand
balls), infiltration of the space-holder filler with liquid gray iron and the
removal of space-holder filler. Sand balls (sand ball diameters: 15, 30 and 45
mm) were prepared manually, using a core box with the mixture consisting
of silica sand, bentonite, dextrin and sodium silicate as filler materials
(figure 15a). The sand balls were filled in to the mould cavity. Gray cast iron
was melted in an induction furnace and poured into the mould cavity at
various temperatures as 1380 °C, 1385 °C and 1390° C. The porous gray
iron samples were subjected to radiographic inspection for analyzing the
pores formed in the metal. They achieved maximum percentage porosity of
72.03% (figure 14b).
Figure 15. – (a) Sand balls. (b) Sample with porosity of 72.08%. [35]
21
2.4.5. – Dual phase steel foam using SiC performs.
In the Mechanical Department of the Milan Polytechnic, a technique that
uses SiC preforms as space holders was successfully applied for dual steel
foam processing [36]. The infiltration process was performed in a 10 kW
centrifugal induction casting machine having an alumina crucible. The space
holder (SiC filters) was placed into a graphite mold. The molten metal was
forced into the mold by centrifugal force. After complete solidification and
cooling, the solid steel and filter were submerged in an aqueous HF (25%
vol) bath till complete SiC dissolution. The filters were available in three
different porosities: 10, 20 and 30 ppi (figure 16).
Figure 16. – Dual phase steel foams having different porosity. [36]
22
2.4.6. – Ni-Mo-Cr open-cell foams using sodium aluminate preform.
In this work, it was demonstrated that sodium aluminate (NaAlO2) can be
used successfully as a space-holder for the casting of Ni-base alloy foam by
the casting replication method [37]. The nickel base alloy (Ni-22.5 Mo-
12.5Cr-1Ti-0.5Mn-0.1Al-0.1Y, in wt %) was used: melting temperature of
1350 °C and a density of 8.6 g/cm
3
. Sodium aluminate was selected as
space-holder because it exhibits a melting point of 1650 °C, well above the
casting temperature of Ni-super alloy and it has good solubility in water and
acids which it is of importance for the removal step.
Figure 17. – Metallographic cross-section of Ni-Mo-Cr replicated (46% relative density)
after space-holder removal. [37]
Preforms of lightly sintered powders of sodium aluminate were infiltrated at
low pressure by liquid nickel base alloy. The sodium aluminate phase was
23
then removed by dissolution in a 10% HCl solution, resulting in nickel base
alloy foams with average relative density of 46% (figure 17). Sodium
aluminate is thus an excellent candidate for creation of foam using nickel
alloys and other high-melting point metals.
2.4.7. – Amorphous open-cell metallic foams using sintered salt performs.
Amorphous metallic foams with an open-cell structure are processed with
the salt replication method by infiltration of a sintered salt pattern (SrF2 and
BaF2) with liquid Vit106, a Zr-based bulk metallic glass [38]. After pattern
removal in nitric acid, the Vit106 foams exhibit highly uniform pores, about
250μm in size, and relative densities in the range 15–22% (figure 18).
Figure 18. –Uniform macrostructure of the amorphous open-cell metallic foams. [38].
24
2.5. – Mechanical properties of metal foams.
The single most important feature of a cellular solid is its relative de density.
The definition of relative density is the density of the cellular material
divided by that of the solid from which the cell walls are made. A typical
stress-strain curve is shown in figure 19. It shows a linear elastic
deformation (region I) followed by a long deformation plateau of almost
constant stress. Linear elasticity is governed by cell wall deformation due to
bending and axial forces.
Figure 19. – Schematics compressive stress-strain curve for foams, showing the three
regimes of linear elasticity, plastic yielding and densification. [39]
25
The elastic module of the metallic foam can be determined by the initial
slope of the stress–strain curve. The long plateau (region II) is the result of
the collapse of the cells by elastic buckling, plastic collapse or brittle
crushing. Finally, the plateau end and the stress begin to rise sharply (region
III- densification region), as the flattened cell walls are completed collapse
and impinge. The metal foams are examples of elastic-plastic materials.
Gibson and Ashby [39] formulated expressions for the mechanical properties
of metal foams based on the compression behavior. The expressions are
derived using basic mechanics and a simple geometry assuming a cubic unit
cell with ligaments of length l and square cross sections of side t, as shown
in figure 20.
Figure 20. – Cubic unit cell as provided in cellular solids by Gibson and Ashby. [39]
26
Figure 20 does not provide a realistic representation of physical syntactic
foam cell geometry, but it is useful in that the deformation of the ligaments
is easily understood and provides relatively straightforward geometry with
which to derive equations. Cell structure in actual syntactic metal foams is
more complex and typically not that uniform throughout the material. While
other options for unit cell definitions are available and equations could be
obtained from more complicated, representative geometry, if the
deformation behavior of the cell ligaments is consistent amongst different
geometries, the properties can be adequately understood using this
representation.
Rather than properties derived explicitly, the expression are presented as
proportionalities that remain valid if the deformation mechanism in real
foam cells remain consistent with those assumed for the derivation. Using
this representation of a unit cell, the relative density and second moment of
area of a ligament can be related to those dimensions by
2
*
l
t
s
and
4
t I
Gibson and Ashby derive their expressions for elastic properties using
standard beam theory and the stress and stress relation of the entire cell.
Using beam theory, the deflection of the edge of a unit cell is proportional to
Fl
3
/E
s
I. The global compressive stress is proportional to the force transmitted
to the ligament as σ α F/l
2
while the global strain is proportional to the
27
displacement as ε α δ/l. These relationships are then combined using
Hooke’s law of elasticity to determine the expression for the elastic modulus,
l l
F
E
. *
2
or
4
1 *
l
I E C
E
S
Equations (1) and (2) can then be substituted into equation 4 to obtain
2
*
1
*
S S
C
E
E
for open cell foams. The constant C1 includes the constant of proportionality
and is determined from test data to be approximately equal to one.
Gibson and Ashby derive the above expression for the elastic properties of
there dimensional foams based on the deformation mechanism of a unit cell
that is not a truly representative of the cell structure found in real foams.
While this method may be valid due to expressions being written as
proportionalities that include constant arising from specific geometries, the
only deformation considered in the derivation is the deformation as a result
of ligament bending. Gibson and Ashby argue that ligament bending is the
governing mechanism for failure of the entire network and exclude the
deformation resulting from axial and shear forces.
28
2.6. – Energy absorption.
The main idea of packaging using a foam is the ability to absorb energy
while keeping the peak force on the packaged object below the limit that will
cause damage or injury. Foams convert kinetic energy into energy of some
other sort like deformation energy and heat. Figure 21 shows that stress-
strain curve of the metal foam compared to that of the equivalent solid
matrix material. The foamed material usually shows a lower peak force than
the solid material for the same energy absorption. [39, 40]
Figure 21. – Stress-strain response for and elastic slid and a foam made from the same
solid, showing the energy absorbed at stress σ p. [39]
The energy absorbed by the meal foam per unit volume at the strain ε is
simply the area under the stress-strain curve up to ε. As figure 21 shows that
the foamed materials have the long plateau in the stress-strain curve which
29
allows large energy absorption at a nearly constant load. It is very important
to understand that the effect of the relative density on the mechanical
behavior of the metal foam (Figure 22). If the density is too low, foamed
material will be crushed before enough energy has been absorbed. If the
density is too high, the force exceeds the critical value before sufficient
energy can be dissipated. The performance depends on the density of the
foamed material.
Figure 22. – Stress-strain responses are measured at a single strain rate [39]
30
2.7. – Applications of metal foam.
Cellular metallic materials are used in wide range of applications. Metal
foams offer significant performance gains in light, stiff structures, for the
efficient absorption of energy, for thermal management and perhaps for
acoustic control and other, more specialized applications. They are
recyclable and nontoxic. Table 1 shows a summary of some of the structural
applications of metal foams [39]. In many applications a medium, liquid or
gaseous, is required to be able to pass through the cellular material. Various
degrees of openness are needed, ranging from very open for high rate fluid
flow to completely closed for load-bearing structural applications, and
appropriate materials should be chosen to satisfy these conditions.
31
Table 1. Potential applications for metal foams. [39]
Application Comment
Lightweight structures Excellent stiffness-to-weight ratio when loaded in
bending.
Mechanical Damping The damping capacity of metal foams is larger than
that of solid metals by up to a factor of 10.
Vibration Control Foamed panels have higher natural flexural frequencies
than solid sheet of the same mass per unit area.
Acoustic absorption Reticulated metal foams have sound-absorbing
capacity.
Energy management Metal foams have exceptional ability to absorb energy
at almost constant pressure.
Packaging Ability to absorb impact at constant load, coupled with
thermal stability above room temperature.
Thermal Management Metal foams are non-flammable; oxidation of cell faces
of closed-cell aluminum foams appears to impart
exceptional resistance to direct flame.
32
3. - Synthesis of syntactic steel foam using mechanical pressure
infiltration.
3.1. - Experimental procedures.
3.1.1. – Materials.
The materials selected to produce the syntactic steel foams were steel and
hollow alumina microspheres (Washington Mills Company). Medium
carbon steel with a target composition of 0.5 % C and 1.4 % Si was selected
for initial trials. This composition was chosen because an increase in the
level of silicon in steel reportedly increases melt fluidity by lowering the
liquidus temperature [41]. A low carbon steel with a composition of 0.2 % C
and 0.1 % Si was also employed and tested. The composition of the alumina
microspheres is shown in Table 2. The hollow microspheres (diameter
1.27 mm) were approximately spherical with a surface texture that was
rough and irregular. Alumina microspheres were sorted and classified for the
infiltration experiments. Additionally, microspheres that were broken or
defective were separated from whole microspheres using a water buoyancy
method. Typically, broken or defective spheres sank, while whole
microspheres floated.
33
Table 2: Typical chemical analysis of alumina microballoons.
Component Percentage
Al2O3 (by difference)
99.2
SiO2
0.60
Fe2O3
0.02
Na2O
0.15
CaO 0.01
Other oxides 0.02
3.1.2. - Foam production.
The pressure infiltration procedure involved melting a prescribed quantity of
steel in a free-standing crucible situated in a customized induction furnace
(12.5 KW, 10-50 KHz). The steel melting capacity of the crucible was 580 g
(7.44 10-5 m
3
). Once the steel melted completely, alumina microspheres
were added to the surface of the melt. Because of the high buoyancy and low
wettability, microspheres floated on the melt surface. The graphite crucible
(with the molten steel and microspheres) was then removed from the
induction furnace and placed in an infiltration press. A mechanical
infiltration press with a graphite plunger was applied to the microspheres,
forcing the melt into the interstitial spaces between the packed microspheres
(Figure 23). Force was applied to overcome the inherent surface tension
forces resisting infiltration. The resultant casting consisted of two regions: a
dense layer of excess at the bottom, covered by a thick layer of syntactic
34
steel foam. During the melting and infiltration process, the duration of each
step was controlled to produce repeatable and consistent results. Several
experiments were carried out to determine how process variables influenced
the quality of the foam. Key variables included the temperature at which
microspheres were preheated, the plunger travel distance, and the volume
ratio of microspheres to steel. This will be discussed in more detail in the
Results and Discussion section.
Figure 23. – Infiltration process: (a) Melting, (b) Adding the microspheres,
(c) Initiation of the infiltration, (d) Infiltration finished.
3.1.3. - Density calculation.
Syntactic foam samples were sectioned to final dimensions of ϕ 33.8 × 10
mm. The density of the syntactic foam was determined by measuring the
weight and dimensions of the sample. Thus, the density of the steel syntactic
foam sample was 4220 kg/m
3
, and the relative density of the syntactic foam
sample was 0.54 (compared to solid steel-7800 Kg/m
3
). Image analysis of
35
polished sections revealed a microsphere volume fraction of 0.53. A random
collection of spheres can be expected to yield a minimum packing density of
52.36% [42]. Therefore, a volume fraction of 0.53 corresponds roughly to
random loose packing of microspheres. Usually, random dense packing
occurs when a loose collection of spheres is vibrated to achieve greater
packing efficiency [42].
3.1.4. - Microstructure of the syntactic foam.
Polished sections of cast foam samples were prepared to determine the
distribution of microspheres and the quality of the infiltration. A typical
distribution of microspheres in the steel matrix is shown in Figure 24. Using
a nearest neighbor analysis, an index value of 2 was determined, indicating
that the distribution was “regular” - a value of 0 is characteristic of a
clustered distribution, while a value of 2.15 indicates a regular distribution
[43]. There was no evidence of microsphere clustering or unintended
porosity (unfilled regions between the microspheres). Furthermore, there
was no sign of broken or cracked microspheres that were infiltrated with
steel in interior regions, indicating that total infiltration was obtained. Note
that minor defects were occasionally observed. When two microspheres
were nearly touching, molten steel did not flow around both microspheres
completely, leaving a small gap at the point of contact.
36
Figure 24. – Medium carbon syntactic steel foam.
Polished and etched sections of the medium carbon steel syntactic foam
revealed that the matrix consisted of two microconstituents: a large amount
of a gray-colored phase (pearlite) and a small amount of a light-colored
phase (pro-eutectoid ferrite). Ferrite is usually present in steel as a solid
solution of iron containing carbon or one or more alloying elements such as
silicon and manganese, while pearlite is present as a mixture of alternate
strips of ferrite and cementite in a single grain. In contrast, sections of the
low carbon steel syntactic foam revealed a large amount of ferrite and a
small amount of pearlite (Figure 25). At the top of figures 24a and 24b, there
is an inset showing the region where the microstructure is observed.
37
Figure 25. – (a) Microstructure of the low carbon steel syntactic foams (left). (b)
Microstructure of the medium carbon steel syntactic foams (right).
Hardness measurements were performed using a nanoindenter (Agilent -
MTS XP) to determine the properties of the microconstituents of the carbon
syntactic steel foams (ferrite and pearlite). The average hardness values of
the microconstituents in the medium carbon steel syntactic foam sample
were 2.55 GPa and 3.55 GPa for ferrite and pearlite, respectively. Each
measurement was performed ten times at different locations of the sample.
Comparing measured values with those reported in the literature, the values
are in the same range for both phases. A.J.Mian reported a value of 2.43 GPa
for AISI 1005 steel (predominantly ferrite) and 3.49 GPa for AISI 1045 steel
(ferrite/pearlite). [44]. Material hardness is directly related to yield strength
and resistance to plastic deformation [45]. Ferrite, the softer constituent,
plastically deforms more readily than pearlite.
38
3.1.5. - Compression
Compression tests of steel syntactic foam were conducted at a strain rate of 1
mm/min. The specimen dimensions were 8.38 8.38 11.68 mm
3
. Four
samples for each composition were prepared by polishing prior to testing,
and all foam samples tests had a relative density of 0.54. Load-displacement
data were recorded during testing and subsequently converted to stress–
strain data. The specimen size ensured that each sample tested contained at
least six microspheres in each direction, thus minimizing edge effects [46].
3.2. - Results and Discussion.
Production of syntactic steel foams using alumina microspheres presented
challenges because of the poor wettability between the constituents and the
high melt temperature of steel, which often caused rapid cooling and
premature solidification. Another difficulty was the entrapment of air during
melt infiltration. Key process variables for producing high quality foams
were the preheatment temperature of the microspheres, the travel of the
plunger, and the ratio of microspheres/melt. The effects of these parameters
are considered below.
3.2.1. - Microsphere preheatment temperature.
For complete infiltration of the microspheres, the melt must remain liquid
during the entire process. Our infiltration experiments were non-isothermal,
because the molten steel was always hotter than the preheated microspheres
and the plunger. Thus, as the liquid steel flowed between the microspheres,
39
some cooling invariably occurred. A completely isothermal infiltration
process would eliminate the possibility of premature liquid steel
solidification, but this is difficult to achieve in practice because of the high
melting point of steel.
In initial experiments, microspheres were preheated in a resistance furnace
to either 300 C or 500 C. However, the microspheres tended to cool rapidly
once removed from the furnace and poured on the steel melt just prior to
infiltration because of the low mass and high surface area of the
microspheres. Similar difficulties were encountered during the infiltration of
SiC preforms with cast iron [47]. The resulting samples showed incomplete
infiltration associated with premature solidification of the melt.
Superior infiltration was achieved by adding microspheres directly to the
melt surface in the induction furnace and holding for three minutes in the
induction furnace to heat the microspheres. No attempt was made to measure
the temperature of steel or the microspheres inside the induction furnace
using an immersion or optical pyrometer because the liquid steel was
covered by the alumina microspheres and a ceramic lid covered the crucible
to avoid excessive heat losses and promote higher preheatment temperature
of the microspheres. The temperature of the microspheres prior to infiltration
was undoubtedly higher using this technique than during initial trials.
Increasing the time the microspheres were exposed to the heated melt
increased the temperature of the microspheres (and of the melt, also), but
three minutes was sufficient for equilibration, and extending the preheating
time (more than three minutes) did not appear to benefit the infiltration
40
process. The graphite plunger also was preheated to 600 C to reduce the risk
of premature solidification.
3.2.2. - Plunger infiltration travel.
Another challenge associated with pressure infiltration was air entrapment,
which typically prevents complete infiltration of the syntactic foam.
Multiple trials revealed that the rate of descent of the graphite plunger was
critical to achieve infiltration without air pockets. The descent of the heated
plunger was thus achieved in two stages. In the first stage, the plunger was
lowered to touch the microspheres in the crucible, allowing air to gradually
escape from interstitial cavities as microsphere packing increased slightly.
Melt infiltration during this stage was negligible, and some equilibration of
plunger, microspheres, and melt occurred. In the second stage, a threshold
stress of ~0.5MPa was applied quickly to achieve melt infiltration of the
microspheres. Interestingly, sub-threshold plunger stresses produced no
infiltration, as the melt tended to “race-track” along crucible walls instead.
Such experiments demonstrated the existence of this minimal stress required
to achieve infiltration.
3.2.3. - Amount of microspheres.
The graphite crucibles provided a finite volume for the steel melt and the
microspheres (0.0381 m interior diameter and 0.05 m height). Infiltration
experiments were performed to determine the sample size and proportions of
constituents to achieve complete infiltration. The following quantities of
microspheres were employed: 6 10-6 m
3
, 12 10-6 m
3
, 18 10-6 m
3
and
41
24 10-6 m
3
(measured in a graduated tube). With 6 10-6 m
3
and 12 10-
6 m
3
, complete infiltration was achieved with minimal unintended porosity.
However, when using 18 10-6 m
3
and 24 10-6 m
3
, the quality of the
resulting samples was poor, and foams contained substantial unintended
porosity. The unintended porosity was attributed to premature solidification
of the melt and associated inhibition of infiltration. While the melt volume
was sufficient for infiltration, the thermal energy was insufficient to prevent
premature freezing. We concluded that 12 10-6 m
3
of microspheres was
roughly the maximum volume of microspheres that could be used to produce
a high-quality foam sample with our crucible size. Larger samples can be
achieved using larger crucibles. It is interesting to note that samples
produced with 6 10-6 m
3
and 12 10-6 m
3
presented similar microsphere
distribution and the same relative density (0.54). This is explained by the
fact that the microspheres tended to float and aggregate at the top of the
sample in the same pattern, producing similar microsphere distributions in
those samples and subsequently equal relative density.
3.2.4. - Interfacial reactions.
Alumina and steel is a non-wetting system in which iron is the oxidizable
phase. At high temperatures, interfacial reactions between alumina and steel
are possible; including alumina dissolution in molten steel have been
reported [48]. These reactions can affect interface structure and thus the
mechanical behavior of syntactic steel foams [48]. The possible formation of
an interface product requires diffusion, and thus the reaction depends
strongly on the infiltration temperature and exposure time. Usually long
42
exposure times (in the order of hours) and the high melting temperature of
steel contribute to the development of the chemical reaction between liquid
steel and alumina [49].
To determine if chemical dissolution of the alumina occurred, composition
profiles were measured by x-ray spectroscopy (EDS) on polished syntactic
foam sections. The probe was stepped across the interface between an
alumina microsphere and the medium carbon steel, providing an opportunity
to examine the region for evidence of interdiffusion and intercalation. Figure
26 shows the microsphere-steel interface. Although the edge of the
microsphere wall was not smooth, the border was clear and discrete. There
was no visible interface layer between the microsphere and the steel matrix.
The composition profile across the steel-alumina interface shows the
distribution of elements in the region (Figure 26). The Al and O content
were constant within the microsphere wall (to the left of the line in figure
26(b)), then both decreased to almost zero in the steel matrix. Likewise, the
Fe and Mn chemical contents were constant in the steel matrix (to the right
of the line in figure 26(b), then both decreased to almost zero at the
microsphere wall. Therefore, there was no evidence that chemical reaction
or diffusion between steel and alumina occurred at the interface. This could
explained by the fact that the time of contact between the molten steel and
the alumina microspheres in our experiments were in the order of minutes.
Because of the short solidification time, there was insufficient time for a
dissolution reaction between the two phases.
43
Figure 26. – (a) The wall of an alumina microsphere observed in high magnification
using back-scattered electron (BSE) imaging, also indicating the line of scan. (b) EDS
line scan at the interface between the steel and alumina microsphere.
3.2.5. - Mechanical properties.
Compared to conventional metallic foams, syntactic foams typically feature
higher compression strengths, isotropic mechanical properties, and superior
energy-absorbing capabilities due to extensive strain accumulation at
relatively high stresses. The syntactic steel foam produced in this study
exhibited a characteristic ductile stress–strain behavior when tested in
compression (figure 27). The first stage was a linear elastic region, which
was followed by a distinctive knee where the slope of the curve dropped to
almost zero. The stress at this knee was taken as the compression strength
for the foam. The knee was followed by a long stress plateau, during which
the cell walls buckled and collapsed. Note that the slope in this region
gradually increased as the deformation progressed and was not completely
flat, as in an ideal energy absorber. The duration of the plateau depended on
the relative density of the foam – normally foams with lower relative density
44
present longer stress plateaus. The duration of the plateau region of our
samples extended to approximately 50 % strain, which is consistent with the
relative density of our samples 0.54. Finally, the plateau ends and the stress
began to rise sharply, as the microspheres were completely collapsed, and
the densification stage began.
Figure 27. –Stress-strain response of low carbon (bottom) and medium carbon (top) steel
foam.
3.2.6. - Compression strength.
The compression strength of metal foams depends strongly on the relative
density and the yield strength of the matrix material [39]. As the yield
strength of the matrix increases, so does the compression strength of the
resultant syntactic foam. This is apparent in the stress-strain curves of low
45
carbon and medium carbon steel syntactic foams, shown in Figure 27. The
low-carbon steel foam exhibits a lower compression strength than the
medium-carbon steel foam. Increasing the carbon content increases the yield
strength of steel, and thus the foam. The low-carbon steel syntactic foam
contained a largely ferritic microstructure which was not as strong as the
pearlitic microstructure of the medium-carbon steel foam.
Insight into the mechanisms of deformation is provided by images of the
foam sample recorded at progressively increasing strains, as shown in Figure
28. The sample deformed smoothly throughout most of the compression
range (up to 50% strain). Cell walls buckled continuously and collapsing
cells were uniformly distributed through the sample during compression.
After the cell collapse mechanism was exhausted, the ceramic microspheres
were completely crushed, and some pulverized fragments exfoliated from
the sample. Steel foams with high carbon content (2 % - 3 % C: gray iron
matrix) normally present serrated stress–strain curves, with large stress
drops in the stress plateau [1]. This behavior is considered undesirable for
the energy absorption function of metallic foams because the serrations are
associated with cell wall cracking and fragmentation in a brittle manner, as
opposed to plastic bending [1]. Fragmentation behavior has not been
observed in any of the syntactic steel foams produced here – the foams
remain intact and in one piece after compression loading.
46
Figure 28. – Compression of medium carbon syntactic steel foam: showing different
stages of compression: 0%, 10%, 20%, 30%, 40% and 50% strain.
3.2.7. - Energy absorption of steel syntactic foams.
During compression of the syntactic steel foam, the work done is
irreversibly absorbed as plastic deformation. The energy absorbed per unit
volume is given by the area under the stress–strain curve up to the onset of
densification. A common objective in producing metal foam is to maximize
the amount of energy absorbed by increasing the height of the stress plateau
(compression strength) as well as the duration of the stress plateau
(densification strain).
The compression behavior of the steel syntactic foam produced here is
compared with other steel foams in Table 3 (Fig. 29). The table includes
properties for a stochastic steel foam produced by a powder metallurgical
method [1], as well as properties for a composite syntactic foam produced by
47
filling the interstitial spaces between densely packed steel microspheres with
steel powder and sintering them into a solid cellular structure [2].
Figure 29. – Stress-strain curves of our foams and steel foams from reference 1 and 2.
Table 3 shows that the low-carbon syntactic steel foam exhibits an energy
absorption capacity of 69.45 MJ/m
3
, while the medium-carbon syntactic
steel foam has a value of 122.68 MJ/m
3
. The low-carbon steel foam thus has
an energy absorption capacity comparable to the steel foam produced in
references 1 and 2. On the other hand, the medium carbon steel foam shows
greater energy absorption per unit volume and also greater energy absorption
per unit mass than any of the other steel foams.
The stochastic high density steel foam produced in reference 1 has one of the
highest energy absorption capacities among all metal foams, 85 MJ/m3, but
48
this value is significantly less than the medium-carbon steel foams produced
in the present study (122 MJ/m3). The stochastic foam of reference 1 is
characterized by irregular, non-circular pores of different size [1]. In contrast,
the composite syntactic foam described in reference 2 exhibits uniform pore
sizes characteristic of syntactic foams. However, the high degree of matrix
porosity may be responsible for the relatively low energy absorption
capacity of these syntactic foams (67.80 MJ/m3) [2]. In contrast, the foam
described here presents a more uniform cell structure free of unintended
porosity, largely because of the use of monosized microspheres and the
absence of unintended porosity. Comparing the foam produced here with
aluminum closed-cell foam [50] reveals that the aluminum foam absorbs 2.6
MJ/m3 (6.5 KJ/Kg) at 50% strain, a value much lower compared to the
values reported for the present syntactic steel foams. Syntactic steel foams
have much greater energy absorption values compared to traditional closed–
cell aluminum foams (348 % greater energy absorption per mass) and are
well-suited to applications requiring high impact energy absorption.
49
Table 3: Comparison of physical properties of our samples with other steel foams.
Powder
Metallurgy
Carbon
Steel
Foam [2]
Powder
Metallurgy
Stainless Steel
Foam [2]
Blowing
Agent
Steel Foam
High
density[1]
Blowing
Agent
Steel Foam
Low
density[1]
Medium
Carbon
Syntactic
Metal
Foam
Low
Carbon
Syntactic
Metal
Foam
Sphere OD
(mm) 1.4 2 * * 1.27 1.27
Sphere wall
thickness
(mm) 0.05 0.1 * * 0.04 0.04
Measured
density(g/cm
3
) 2.55 2.95 4.34 3.52 4.21 4.21
Relative
density (%) 32.40 37.50 55.60 45.12 54.0 54.0
Densification
Strain (%) 57.00 54.00 50.00 50.00 50.00 50.00
Energy
absorption per
volume at
densification
(MJ/m
3
) 37.60 67.80 85.00 50.00 122.68 69.45
Energy
absorption per
mass at
densification
(KJ/Kg) 14.75 22.98 19.59 14.20 29.14 16.50
50
4.-Synthesis of syntactic steel foam using gravity-fed infiltration
4.1. - Experimental.
4.1.1. - Materials.
The syntactic steel foams were produced from two primary constituent
materials - low carbon steel (AISI 1018) and hollow alumina microspheres
(Washington Mills Company). Five ferritic-pearlitic steel compositions were
obtained by adding specific amounts of carbon and ferrosilicon, as shown in
Table 4, yielding foams with different degrees of intrinsic strength and
ductility. A sixth foam was produced from a high-alloy TRIP-steel (chemical
composition 0.03 % C, 15.5 % Cr, 6.1 % Mn, 6.1 % Ni, 0.9 % Si and 0.1 %
Al) [51]. The chemical composition was obtained by adding specific
amounts of electrolytic iron, carbon, ferrochromium, ferromanganese, nickel,
ferrosilicon and aluminum.
Table 4: Chemical composition of the investigated ferritic-pearlitic steels.
Sample % C % Si
A 0.20 0.10
B 0.26 0.40
C 0.32 0.70
D 0.38 1.00
E 0.44 1.30
51
The microspheres were sorted and classified according to size. Mono-sized
alumina microspheres with a diameter of 4.45 mm were used for the
infiltration experiments. The microspheres were approximately spherical
with a slight surface texture. The chemical composition of the alumina
microspheres is shown in Table 5.
Table 5: Typical chemical analysis of alumina microballoons.
Component Percentage
Al2O3 (by difference)
99.2
SiO2
0.60
Fe2O3
0.02
Na2O
0.15
CaO 0.01
Other oxides 0.02
4.1.2. - Foam production.
Steel charges were melted using a customized induction furnace (12.5 KW,
10-50 KHz), and a special-designed crucible was used for the gravity
pressure infiltration procedure (Figure 30). The crucible consisted of two
chambers: an upper chamber for melting the steel, and a lower chamber
(mold) for packing and infiltrating the alumina microspheres. After melting
the steel charge in the upper chamber (Figure 29a), the melt infiltrated the
packed microspheres in the lower chamber, filling interstitial spaces (Figure
52
30b and 30c). Vents were machined in the lower chamber to accommodate
displaced gas during melt infiltration. During the melt and infiltration
process, the duration of each step was controlled to produce repeatable and
consistent samples, and experiments were carried out to determine how
process variables influenced foam quality. The most important parameters
controlling the infiltration were the preheatment temperature of the
microspheres and the melt temperature prior to infiltration.
Figure 30. – Gravity pressure infiltration process.
4.1.3. - Density.
Syntactic foam samples were sectioned to final dimensions of ϕ 23 × 23 mm
(Figure 31). The density of the syntactic foam was determined by measuring
the weight and dimensions of the sample. Thus, the density of the syntactic
steel foam sample was 4220 kg/m
3
, and the relative density of the steel
syntactic foam sample was 0.46 (steel density = 7800 Kg/m
3
). Image
analysis of polished sections revealed that the syntactic steel foam consisted
of 63% alumina microspheres and 37% steel matrix by volume. The
53
difference between the relative density (46 %) and the steel matrix volume
fraction (37 %) is due to the weight of the hollow microspheres that is
included in the density of the syntactic foam. A random collection of mono-
sized spheres can be expected to yield a minimum packing density of 52.36%
[42] and a maximum of 74% when ordered in a FCC fashion. A volume
fraction of 0.63 corresponds roughly to random dense packing of mono-
sized microspheres [42].
4.1.4. - Microstructure.
Polished sections of cast syntactic foam samples were prepared to determine
the distribution of microspheres and the quality of the infiltration. A typical
distribution of microspheres in the steel matrix is shown in Figure 31. No
evidence of microsphere clustering or unintended porosity (unfilled regions
between the microspheres). Furthermore, there was no sign of broken or
cracked microspheres that were infiltrated with steel in interior regions.
Despite mutual contact between adjacent microspheres, melt infiltration
occurred at the contact points, and infiltration gaps were not observed.
Polished and etched sections of the samples A, B, C, D, and E revealed that
the matrix consisted of a gray-colored constituent (pearlite) and a light-
colored constituent (pro-eutectoid ferrite). These specimens present different
ratios of pearlite/ferrite, and the ratio increases with increasing alloying
content (C, Si). The steel matrix of sample A consisted of a largely ferritic
microstructure, while the steel matrix of sample E contained a largely
pearlitic microstructure (Figure 32). Ferrite is relatively more ductile, while
pearlite is relatively stronger.
54
Figure 31. – Samples of steel syntactic foam produced by gravity-pressure infiltration.
Figure 32. –Microstructure of the steel syntactic foam samples (A, B, C, D, E).
55
4.1.5. - Compression.
Compression tests were conducted at a strain rate of 1 mm/min using
specimens with dimensions ϕ 22.9 × 22.9 mm. Three samples for each
composition were prepared by polishing prior to testing, and all foam
samples tests had a relative density of 0.46. The compression tests were
performed using a load frame (Instron 5585H) and load-displacement data
were converted to stress–strain data. The specimen size ensured that each
sample tested contained at least six microspheres in each direction, thus
minimizing edge effects [46].
4.2. - Results and Discussion.
4.2.1. - Microsphere preheatment temperature.
For complete infiltration of the microspheres, the melt must remain liquid
and fluid during the entire infiltration process. To achieve this, sources of
heat loss must be minimized or eliminated. The heat loss sources come from
liquid steel pouring and from cooling of the microspheres when exposed to
open air. For this reason, the crucible and the mold were joined in one
assembly and only the induction furnace was used to heat this assembly.
This method preheated the microspheres to 1100°C prior to infiltration. The
melt flow into the mold was gravity fed and occurred without applied
pressure, resulting in complete infiltration. Mold vents allowed for gas
escape during the infiltration, eliminating gas entrapment in the cast samples.
56
4.2.2. - Interfacial reactions
To determine if chemical dissolution of the alumina occurred, composition
profiles were measured by x-ray spectroscopy (EDS) on polished syntactic
foam sections. The probe was stepped across the interface between an
alumina microsphere and the medium carbon steel. Figure 32a shows the
microsphere-steel interface. There was no visible interface layer between the
microsphere and the steel matrix. The composition profile across the steel-
alumina interface shows the distribution of elements in the region (Figure
33b). The Al and O content were constant within the microsphere wall (to
the left of the line in figure 33(b)), then both decreased to almost zero in the
steel matrix. Likewise, the Fe and Mn chemical contents were constant in
the steel matrix (to the right of the line in figure 33(b), then both decreased
to almost zero at the microsphere wall. Therefore, there was no evidence that
chemical reaction or diffusion between steel and alumina occurred at the
interface.
Figure 33. – (a) The wall of an alumina microsphere observed in high magnification
using back-scattered electron (BSE) imaging, also indicating the line of scan. (b) EDS
line scan at the interface between the steel and alumina microsphere.
0
20
40
60
80
100
120
140
160
180
200
0 20 40 60 80
Distance (microns)
Counts
O
Mn
Fe
Al
Si
57
4.2.3. - Mechanical properties.
Quasistatic compression tests were conducted on foam samples at a strain
rate of 1mm/min. The stress-strain curves of samples A, B, C, D and E
exhibit elastic-plastic stress–strain behavior characteristic of foams and other
cellular materials (Figures 34a and 34b). The first stage consists of a linear
elastic region, which is followed by a distinct knee where the slope of the
curve drops to almost zero. The stress at this knee is taken as the
compression strength for the syntactic foam. The knee is followed by a long
stress plateau, during which the cell walls bend, buckle and collapse. Finally,
the plateau ends and the stress began to rise sharply, as the microspheres are
completely collapsed and the densification stage begins.
Figure 34a (samples A, B and C) shows that the compression strength
increases as the carbon content increases (composition A through
composition C). Increasing the alloy content increases the yield strength of
the steel, and thus the foam strength. Note that the slope in this region
gradually increases as the deformation progresses and is not perfectly flat
(plastic), as in an ideal energy absorber. Normally, strain hardening in
stochastic metallic foams is insignificant, but the strain hardening that occurs
during the compression of samples A, B and C could be the combined result
of the steel matrix ductility and the uniform structure of the steel syntactic
foam.
58
Figure 34. – Stress-strain curves for steel foams of different C content. (a) Samples A, B
and C. (b) Samples C, D and E.
In samples D and E, cell walls developed cracks as they bent, buckled and
collapsed. As a result, the plateau stress did not increase with increasing
strain (Figure 34b). When cracking occurred in the cell walls, the
relationship between the alloy content and the compression strength changed.
Unstable crack propagation causes brittle fragmentation of cell walls (as
opposed to plastic bending) and is undesirable for the energy absorption
function of metallic foams [1]. Fragmentation occurs when the foam
material lacks sufficient fracture toughness (K IC) to plastically arrest cracks
nucleated on the tension side of bending struts [52]. To resist strut
fragmentation failure, the size of the plastic zone ahead of an opening crack
tip must be comparable to the strut width. The plastic zone size, K
IC
2
/πσ
y
2
,
is employed here as a characteristic length scale to quantify the material
resistance to fracture. Steel foams with high carbon contents (2 - 3 % C: gray
iron matrix) normally present serrated stress–strain curves with large stress
drops in the stress plateau caused by strut cracking [1]. If struts undergo
plastic yielding during collapse (as opposed to fragmentation), the foam will
59
exhibit the maximum plateau stress and absorb the maximum energy
possible for the architecture, metal matrix and relative density [52].
During compression of foam samples, the work done is irreversibly absorbed
as plastic deformation. The energy absorbed per unit volume is given by the
area under the stress–strain curve up to the onset of densification. Table 6
shows that sample C exhibits greater energy absorption than the other
compositions: 1.12 times more energy than sample B and 1.26 times more
energy than sample D. Sample C has the strongest matrix of foam samples in
which cracking does not occur.
Table 6: Energy absorption of the steel syntactic foam with different steel
matrices.
TRIP
STEEL
A B
C
D
E
Densification Strain (%)
50 50 50
50
50
50
Energy absorption per
volume at densification
(MJ/m
3
)
138.30 78.98 94.02
104.78
85.50
67.82
Energy absorption per
mass at densification
(KJ/Kg)
38.55 22.01 26.20
29.20
23.83
18.90
60
Insight into deformation mechanisms was provided by images recorded at
progressively larger strains, as shown in Figures 35 and 36. Sample B
deformed smoothly (without fragmentation) up to 50% strain. Cell walls
buckled continuously, and collapsing cells were uniformly distributed
through the sample during compression. Uniform (non-localized)
deformation is reportedly enhanced by foam homogeneity, while
heterogeneous foam structures promote deformation bands [53]. After the
cell collapse mechanism was exhausted and the ceramic microspheres were
completely crushed, some pulverized fragments exfoliated from the sample
periphery. Deformation of sample E (Fig. 36) resulted in the development of
cracks and associated fragmentation. Such macroscopic cracks have been
attributed to the collapse of weak struts and the subsequent formation of
deformation bands (manifest as stress oscillations in the stress-strain curve)
[53]. Deformation bands reduce compressive strength and compromise the
ability to absorb energy. Total fragmentation behavior has not been observed
in any of the syntactic steel foams produced here – the foams remain intact
and in one piece after compression loading.
61
Figure 35. – Compression of sample B: showing different stages of compression: 0%,
10%, 20%, 30%, 40% and 50% strain..
Figure 36. – Compression of sample E: showing different stages of compression: 0%,
10%, 20%, 30%, 40% and 50% strain.
62
4.2.4. - TRIP steel syntactic foam.
One of the primary objectives of foam design is to increase the energy
absorbed per unit weight. This can be achieved by increasing the yield
strength and toughness of the foam material without compromising foam
density or densification strain. To pursue this approach, TRIP steel foams
(transformation-induced plasticity) were produced. The TRIP acronym
derives from the deformation behavior in which the austenitic matrix
transforms to martensite, allowing for increased strength and ductility [54].
A casting composition of TRIP-steel was selected (CrMnNi) as opposed to a
wrought alloy composition [55]. Figure 36 shows the stress-strain curve for
the TRIP steel foam in as-cast condition. The TRIP steel foam shows higher
compression strength relative to the ferritic-pearlitic foams and absorbs
approximately 32 % more energy.
Figure 37. – Stress-strain curves of samples A, B, C and TRIP steel.
63
4.2.5. - Effect of relative density on the mechanical properties.
The effects of relative density on mechanical behavior were explored by
preparing a set of foams with different microsphere fractions. While the
densification strain can be increased by increasing the microsphere fraction
(for example using a bimodal distribution of microspheres) [20], the
compression strength will generally decrease. Conversely, reducing the
microsphere content generally increases the compression strength, although
the densification strain will decrease. The primary challenge associated with
changing the relative density of syntactic foams is maintaining a uniform
distribution of microspheres (without clusters). Mechanical stirring has been
used to produce syntactic aluminum foams with variable relative density, but
stirring molten steel poses special difficulties, and alternative approaches are
required [21].
To ensure uniform dispersions of microspheres, we deployed wire mesh
inserts in the microsphere beds to create effective “preforms” for melt
infiltration. The mold was filled with alternating layers of wire mesh and
packed microspheres, as shown in Fig. 38a. The mesh inserts fixed the
microspheres in position, preventing redistribution during melt infiltration.
The resultant material was an anisotropic structure of syntactic foam layers
alternating with solid steel slabs (from the mesh inserts, Fig. 38b). Wire
meshes of different dimensions were used to produce syntactic foams with
relative density values of 0.60, 0.68 and 0.75.
64
Figure 38. – (a) Schematics showing the use of wire mesh in the manufacture of syntactic
foams. (b) Sample of steel syntactic foam with relative density of 0.60.
An alloy composition of 0.32 % C, 0.70 % Si was used for these foam
samples to maximize energy absorption during compression. Compression
test results shown in Figure 39 include the syntactic foam with relative
density values of 0.46 and 1 (fully dense) for comparison purposes. Table 7
shows the values of energy absorbed per unit volume and per unit weight for
the samples with different relative densities. The values in Table 7 were
calculated using the densification strain shown in the same table, and
densification strains were obtained from the stress-strain curves. The energy
absorption capability of a typical aluminum closed-cell foam with relative
density of 0.148 [50] is included in Table 7 for comparison.
65
Table 7: Energy absorption of the steel syntactic foam with different relative densities
and aluminum foam with relative density of 0.148.
0.46 RD 0.60 RD 0.68 RD 0.75RD
1.00 RD 0.148 RD
Energy
absorption
per volume
at
densification
(MJ/m
3
)
104.78 183.23 154.98 161.11
41.27 2.6
Energy
absorption
per mass at
densification
(KJ/Kg) 29.20 39.15 29.21 27.50
5.29 6.5
Densification
Strain (%) 0.50 0.49 0.42 0.38
0.11 0.5
All of the syntactic steel foams produced in this study have much greater
energy absorption than the closed-cell aluminum foam. The aluminum
stochastic foam absorbs 6.5 KJ/Kg (2.6 MJ/m
3
) while the syntactic steel
foam of 0.6% RD absorbs 39.15 KJ/Kg (183.23 MJ/m
3
), a six-fold increase.
In fact, this particular steel foam absorbs six times more energy per unit of
mass, and seventy times more per unit volume.
66
Figure 39. – (a) Stress-strain curves for steel syntactic foam samples with relative density
of 0.46, 0.60, 0.68, 0.75 and 1.
The energy absorption capacity of syntactic steel foam also compares
favorably to other types of cellular structures, such as egg-box structures and
truss core structures. For example, D.T. Queheillalt [56] reported an average
value of 8 KJ/Kg (4 MJ/m
3
) for hollow lattice truss structures with relative
densities ranging from 0.031 to 0.233. These truss structures consisted of
304 stainless steel solid wires and tubes. In other work, J. Mzupan [57]
reported a value of 1.44 KJ/Kg (0.17 MJ/m
3
) for an aluminum egg-box panel
with a relative density of 0.044. These values of 8 and 1.44 kJ/Kg are far
less than the values of 27-39 kJ/Kg for steel foams.
67
Figure 40. – Energy absorption per unit mass vs. density: aluminum egg-box [57],
aluminum foam [50], hollow truss metallic lattice [56] and 0.6 RD steel syntactic foam.
To better understand the practical value of syntactic steel foams, one can
calculate the thickness of metallic foam (x) necessary to fully absorb the
kinetic energy of a moving vehicle of mass m and velocity v [58]
x =
mv
2
2E
v
A
where Ev is the energy absorption per unit of volume and A is the contact
area of the metal foam. Assuming an average value of m = 1000 Kg, v =
27.78 m/s and A = 0.2 m
2
, a piece of aluminum foam 0.74 m thick would
absorb all the kinetic energy, while a piece of syntactic steel foam (RD =
0.60) would require a thickness of only 0.011 m. Thus, the steel foams
68
produced here are well-suited to applications requiring high impact energy
absorption.
The strain-stress curves show that increasing the relative density of the foam
increases the compression strength and decreases the duration of the stress
plateau. Therefore, by increasing the relative density, the ability to absorb
energy at low stress levels decreases. Usually there is a tradeoff between
foam strength and densification strain. For example, samples with relative
density 0.75 absorb a large amount of energy per unit volume because of the
high compression strength rather than the large stress plateau (densification
strain of approximately 0.38). A different behavior is exhibited by samples
of relative density 0.46, which show lower compression strength but longer
plateaus (densification strain of approximately 0.5). The foams absorb
similar amounts of energy per unit volume but at different stress levels,
thereby expanding the design space for materials selection.
Comparing the energy absorbed per unit mass for the different foams, a
syntactic foam with a low compression strength and a long stress plateau can
absorb more energy per unit mass than one with a high compression strength
and a short stress plateau. Among all the samples produced, the foam with
relative density 0.6 showed the greatest energy absorption per unit mass.
Foam selection will depend on the application requirements, such as a high
compression strength and short stress plateau or a relatively low
compression strength and longer stress plateau.
69
5.-Synthesis of syntactic aluminum foam using gravity-fed infiltration
5.1. Experiment
5.1. 1. Materials
Syntactic aluminum foams were produced from two primary constituent
materials: aluminum (1100 and 6061 aluminum alloy) and hollow alumina
microspheres (Washington Mills Company). These two aluminum alloys
(with different chemical composition) were chosen to study aluminum
syntactic foams with different degrees of intrinsic strength and ductility. The
microspheres were sorted and classified according to size. Mono-sized
alumina microspheres with diameters of 4.45 ± 0.15 mm and 3.05 ±0.1 4
mm were used for the infiltration experiments. These sizes of microspheres
were selected to minimize the resistance to melt infiltration. Microspheres
with a diameter smaller than 3 mm offer greater resistance to infiltration [19].
The microspheres were approximately spherical and showed a slight surface
texture. One type of face sheet was used to reinforce the ASF plates: high-
strength aluminum alloy 2024 - 0.5 mm thick. The ASF plates were bonded
to the facesheet using rubber-toughened epoxy adhesive film.
70
5.1. 2. Foam production
A custom-designed stainless steel mold was used for the gravity-fed
infiltration procedure. The mold consisted of two chambers: an upper
chamber for melting the aluminum charge, and a lower chamber for packing
and infiltrating the alumina microspheres. Aluminum charges were melted
using a resistance furnace. After melting the aluminum charge in the upper
chamber, the melt flows through an opening in the bottom of the upper
chamber, passes through a ceramic filter and then enters the mold,
infiltrating the packed microspheres from bottom to top. Vents were
machined in the mold to accommodate displaced gas during melt infiltration,
avoiding entrapment of air in the cast sample. This method has been used to
produce steel syntactic foam by the same authors [59 ]. During the melt and
infiltration process, the duration of each step was controlled to produce
repeatable and consistent samples, and experiments were carried out to
determine how process variables influenced foam quality. Syntactic
aluminum foams plates (190 mm x 190 mm x 25.4 mm) were prepared using
gravity-fed infiltration process (Figure 41). The gravity-fed infiltration
method yields ASF with uniform distributions of microspheres and
negligible unintended porosity. The density of the syntactic foam was
determined by measuring the weight and dimensions of the sample. Thus,
the density of the ASF sample was 1242 kg/m3, and the relative density of
the ASF sample was 0.46 (aluminum density = 2700 kg/m3).
71
Figure 41. – Sample of aluminum syntactic foam produced
by gravity-fed infiltration.
5.1. 3. Compression
Quasistatic compression tests were conducted at a strain rate of 1 mm/min
using specimens with dimensions 24 mm x 24 mm x 24 mm. Six samples for
each type of ASF plates were prepared by polishing prior to testing. The
samples were tested at room temperature and in the as-cast condition. The
compression tests were performed using a load frame (Instron 5567) and
load-displacement data were converted to stress –strain data.
5.1. 4. Impact test
ASF plates were sectioned to final dimensions of 93 mm x 93 mm x 12.7
mm for the impact testing. The samples were tested at room temperature and
in the as-cast condition. Each impact test was done at least three times to
ensure reliability and repeatability of the test results. A DynaTup 9250HV
drop weight testing machine was used to perform the impact testing of the
72
ASF plates. Figure 42 shows the drop weight testing machine and a
schematic view of the machine.
A hemispherical steel tup of 16.1 mm diameter was used as the indenter. The
ASF plates were placed between the top and bottom clamp plates, making
sure that the mid-point of the plate is positioned directly underneath the
impactor (central impact). The impact weight (6.4 kg) is released from a
preset height and falls freely along two smooth guided columns and through
the center hole of the clamp plates (Figure 42.b). Once impact began, the
contact forces are acquired by the transducer mounted on the impactor. The
force-time history is directly recorded in a computer. The corresponding
force-displacement history of the impactor could be calculated by integration
using the force-time history. Assuming the impactor is perfectly rigid, the
force-displacement history of the impactor could be considered as the force-
displacement curve of the ASF plate. Using this curve the amount of impact
energy absorbed by the ASF plate is calculated. We tested plates made with
1100 Al alloy and 4.45 mm diameter microspheres, 1100 Al alloy and 3.05
mm diameter microspheres, 6061 Al alloy and 4.45 mm diameter
microspheres. Also, 1100 aluminum and 4.45 mm diameter microspheres
reinforced with a face sheet of high-strength aluminum alloy 2024. The
absorbed energy factor (absorbed energy normalized by the impact energy)
is a common parameter that was be used to evaluate the performance of the
samples at different impact energies (Table 8). The values presented in table
8 are averages values. Pictures of specimens after impact test were taken.
73
Figure 42. – (a) A picture and (b) schematic view of the DynaTup 9250HV Instron
testing machine.
5.2. Results and discussion
5.2.1. Quasistatic compression testing
The stress –strain curves of all tested ASF samples exhibit an elastic-plastic
stress –strain behavior which is characteristic of metal foams and other
cellular materials (Figure 43). The first stage consists of a linear elastic
region, which is followed by a distinct knee. The stress at this knee is taken
as the compression strength for the syntactic foam. The knee is followed by
a long hill plateau, during which the cell walls bend, buckle and collapse.
Note that the slope in this region gradually increases as the deformation
74
progresses and is not perfectly flat (plastic), as in an ideal energy absorber.
Normally, strain hardening in metallic stochastic foams is insignificant, but
the strain hardening that occurs during the compression of the ASF samples
could be the combined result of the aluminum matrix ductility and the
uniform structure of the ASF plate [59]. Finally, the plateau ends and the
stress begin to rise sharply, as the microspheres are completely collapsed
and the densification stage begins.
Figure 43 shows that the compression strength (and energy absorption
capacity) of ASF Al 1100 is lower than ASF Al 6061 when using
microspheres of size 4.45 mm. The solid-solution strengthening mechanism
in the 6061 aluminum alloy matrix is responsible for this behavior [60].
Figure 42 also shows that the compression strength of ASF 1100 Al alloy
with microsphere diameter of 4.45 mm is higher than with microsphere
diameter of 3.05 mm. The reported results on the relationship between
microsphere size and foam compression strength are contradictory. For
example, Wu et al studied effects of the hollow sphere diameter on the
compressive strength. They found that smaller spheres produce higher
compressive strength than the larger ones [14]. Rohatgi et al reported the
yield strength of syntactic foams containing 65 vol.% of cenospheres of
different size decreases from 34 to 9 MPa as the cenosphere size decreases
from 150–250 to 75–105 mm [15]. More discussion on this topic will be
presented later in the paper.
75
Figure 43. – Stress –strain curves for aluminum syntactic. (a) Al 6061-4.45 mm
microsphere diameter. (b) Al 1100-4.45 mm microsphere diameter. (c) Al 1100-3.05 mm
microsphere diameter.
5.2.2. Impact response of aluminum syntactic foam
The effect of the aluminum alloy matrix in the impact behavior of the ASF
was analyzed by comparing ASF plates manufactured with 1100 and 6061
aluminum alloys using the same size of microspheres (4.45 mm). In order to
interpret our results, we will review the results of a study on the impact
response of composite laminates presented by Liu [61]. They found that the
absorbed energy was quite lower than the impact energy when the impact
energy was low. As the impact energy increased, they found an interval in
which the absorbed energy became closer to impact energy (equal-energy
interval). The equal-energy interval was bounded by two points. The point of
lower bound was named penetration threshold, indicating the onset of
penetration and the upper bound was named perforation. Increasing the
impact energy even more, the absorbed energy would remain somewhat
76
constant but now it would represent only a portion of the impact energy used
(perforation zone) [61].
Figure 44 shows the force –displacement curves of the ASF-1100 aluminum
alloy tested with three different impact energies: 60, 120 and 180 J. When
the ASF plates are impacted with 60 and 120 J, partial penetration occurs. In
this case, the force-displacement curves rise linearly, reach a maximum level
and return back to the origin, forming a closed curve that represents the
impactor moving along with the ASF plate, stopping at the point of
maximum displacement and then moving back in the opposite direction
(rebounding). The area enclosed by the closed curve is the energy that the
ASF plate absorbs during the test. If the impact energy increases (180 J),
complete perforation takes place and any excess impact energy that is not
absorbed by the sample would be retained in the traveling impactor as the
indenter moves past the ASF plate. When impacted with 180 J, first the load
increases with displacement linearly up to an initial peak load. A plateau of
around 12 –14 mm displacement is observed after the initial peak load.
Finally a decrease in the load is observed as the sample is complete
perforated. In this case, the force-displacement curve is no longer a closed
curve. The area bounded by the force-displacement curve and the
displacement axis constitute the energy absorbed by the perforated ASF
plate.
For the 60 and 120 J impact energy, the absorbed energy factor is
approximately 91 %, the ASF plates absorb almost all the impact energy in
the form of damage and leaving very limited amount of energy for a rebound
77
(equal-energy interval). This behavior represents what a good energy
absorbing material is expected to do. When impacted with 180 J, the
absorbed energy factor is approximately 82 %. The lower normalized
absorbed energy shows that the ASF samples are no longer able to absorb
the entire impact energy at increasing impact velocities (perforation zone).
Then, the samples manufactured with 1100 aluminum alloy and
microspheres with diameter of 4.45 mm require a minimum absorbed energy
of approximately 148.41 J to perforate the samples.
Inspecting the impacted samples, it was found that the amount of damage
increases with increasing impact energy and the damage is not only
restricted to the area beneath the impactor but also the area around the direct
area of impact (Figure 44). The failure resembles a ductile failure, no
complete fracture or cracking of the ASF plates occurred during the impact
testing. At 60 J of impact energy, radial cracks can be observed at the back
of the ASF plate. At 120 J, larger cracks and an evident bulging formation
can be seen At 180 J, the cracks lead to a kind of petaling kind of failure and
some pieces of foam are pushed off by the indenter. The resistance to impact
of the ASF plates is derived from collapsing and bending of the aluminum
wall cells and crushing of the alumina microspheres. Tearing of wall cells at
the periphery of impactor is another factor contributing to the resistance of
material [62]. The visual examination of the aluminum syntactic foam
“plugs” that results from the complete penetration provides clear evidence
for the failure mechanisms (collapsing, bending and tearing).
Figure 46 shows the force –displacement curves of ASF-6061 Al alloy
tested with three different impact energies: 60, 120 and 180 J. When the
78
samples were tested with impact energy of 60 J, partial penetration occurs.
When impacted with 120 and 180 J, perforation occurs. For the 60 J impact ,
the samples are able to absorb almost all the impact energy-91% (equal-
energy interval) and for the 120 J impact energy, it still absorb a great part of
the impact energy and produces complete perforation. Therefore, we can
deduce that the samples manufactured with 6061 aluminum alloys and
microspheres with diameter of 4.45 mm require minimum absorbed energy
of around 111.5 J to perforate the samples. It is very important to notice that
the force –displacement curve for the 180 J impact test has different features
than the other two curves. First, there is a noticeable decrease in the maxium
impact force and this is reflected in the low absorbed energy. Second, there
is fracture event (brittle behavior) that is visible as a fluctuation on the force
–displacement curve (Figure 46c). This proves that the Al 6061 Al alloy
does not behave properly at 180 J of impact energy (or even higher impact
energies).
Figure. 44. – Comparison of load –displacement behavior of ASF (1100 aluminum
matrix- 4.45 mm microsphere diameter) tested with different impact energies: (a) 60 J (b)
120 J (c) 180 J.
79
Figure. 45. – Side, top and bottom views of the impacted specimens of ASF (1100
aluminum matrix- 4.45 mm microsphere diameter) tested with different impact energies:
(a) 60 J (b) 120 J (c) 180 J. Samples size is 93 mm x 93 mm x 12.7 mm.
We have found that the damage is localized mostly in the area beneath the
impactor but there is also a small deformed area around the hole made by the
impactor but this is not as pronounced as in the case of ASF 1100 Al alloy
plates. The type of failure can be qualified as a ductile-brittle failure. At 60 J
of impact energy, small radial cracks can be observed at the back of the ASF
plate (Figure 47). At 120 J, sample was perforated and some pieces of the
“petaling” structure can be seen in the impacted sample. At 180 J, complete
perforation and all the deformed foam has been removed by the impactor
(plugging). It can be seen that the damage in the 1100 Al alloy samples is
spread over a larger area compared to the ASF 6061 Al alloy which could
help explain its ability to absorb greater amount of energy. We can conclude
this section saying that 1100 ASF aluminum alloy outperform 6061 ASF
80
aluminum alloy because it can absorb higher amount of energy before it is
perforated. This is opposite to what it happens when the ASF are tested
under quasistatic conditions where the ASF 6061 Al alloy has greater
compression strength than the ASF 1100 Al alloy.
Figure 46. – Comparison of load –displacement behavior of ASF (6061 aluminum
matrix- 4.45 mm microsphere diameter) tested with different impact energies: (a) 60 J (b)
120 J (c) 180 J.
81
Figure 47. – Side, top and bottom views of the impacted specimens of ASF (6061
aluminum matrix- 4.45 mm microsphere diameter) tested with different impact energies:
(a) 60 J (b) 120 J (c) 180 J. Samples size is 93 mm x 93 mm x 12.7 mm.
Table 8: ASF samples parameters, testing condition and impact results.
Aluminum matrix
Microsphere
diameter (mm)
Impact
energy (J)
Impact
velocity (m/s)
Absorbed
energy (J)
Absorbed
energy factor
Perforatio
n
6061 4,45 60 4,16 54,38 0,91 Partial
6061 4,45 120 5,92 111,46 0,93 Complete
6061 4,45 180 7,37 100,03 0,56 Complete
1100 4,45 60 4,18 54,26 0,90 Partial
1100 4,45 120 5,91 110,66 0,92 Partial
1100 4,45 180 7,30 148,41 0,82 Complete
1100+Al facesheet 4,45 60 4,19 53,13 0,89 Partial
1100+Al facesheet 4,45 120 5,92 111,66 0,93 Partial
1100+Al facesheet 4,45 180 7,38 175,47 0,97 Partial
1100 3,05 60 4,17 55,68 0,93 Partial
1100 3,05 120 5,94 92,95 0,77 Complete
1100 3,05 180 7,31 89,18 0,50 Complete
82
5.2.3. Impact response of aluminum syntactic foam with face sheet
The effect of adding a face sheet was analyzed by comparing the impact
behavior of the ASF 1100 Al alloy matrix-4.45 mm microsphere diameter
with and without face sheet. Figure 8 shows the load –displacement curves
of ASF 1100 Al alloy matrix and 4.45 mm microsphere diameter reinforced
with a high-strength aluminum alloy 2024 tested with three different impact
energies: 60, 120 and 180 J. None of the tested specimens were completely
perforated as shown in Figure 48.
The absorbed energy factor is approximately the same for all the three
impact energies (equal-energy interval). Adding the face sheet helps to
increase minimum absorbed energy necessary for perforation. The samples
manufactured with 1100 Al alloy- microspheres diameter of 4.45 mm and
reinforced with Al facesheet require a minimum absorbed energy greater
than 175 J to perforate the sample while the samples without facesheet only
require 148.41 J . Moreover, the final deflection is always lower for the face
sheet reinforced ASF plates than the ASF plates without facesheet. The
maximum impact force is always lower for the face sheet reinforced ASF
plates than the ASF plates without facesheet. This increase in the impact
force is due to additional resistance of the facesheet by bending and
stretching (Figure 49). The use of facesheet is a efficient method to increase
the capacity of the ASF plates without adding excessive amount of weight,
alternatives material for the facesheet are being considered for future studies
[62].
83
Figure 48. – Comparison of load –displacement behavior of ASF -1100 aluminum
matrix- 4.45 mm microsphere diameter + aluminum face sheet tested with different
impact energies: (a) 60 J (b) 120 J (c) 180 J.
Figure. 49. – Side, top and bottom views of the impacted specimens of ASF -1100
aluminum matrix- 4.45 mm microsphere diameter + aluminum face sheet tested with
different impact energies: (a) 60 J (b) 120 J (c) 180 J. Samples size is 93 mm x 93 mm x
12.7 mm
84
5.2.4. The effect of microsphere size
The effect of the microsphere size was analyzed by comparing the impact
behavior of the ASF 1100 Al alloy matrix with 4.45 mm and 3.05 mm
microsphere diameter. Figure 50 shows the load –displacement curves of
ASF 1100 Al alloy- 3.05 mm microsphere diameter tested with three
different impact energies: 60, 120 and 180 J. Only the samples that were
tested with 60 J of impact energy were partially penetrated as shown in
Figure 49. For impact energies of 120 and 180 J, perforation occurs.
When impacted with 60 J, the absorbed energy factor is approximately 93 %
(equal-energy interval). After that, the absorbed energy factor keeps
decreasing as the impact energy increases (perforation zone) as can be seen
in table 1. It can de deduced the samples manufactured with 1100 Al alloys
and microspheres with diameter of 3.05 mm require a minimum absorbed
energy of approximately 90 J to perforate the samples. Then, 1100 ASF Al
alloy with microspheres diameter of 4.45 mm outperform 1100 ASF Al alloy
with microspheres diameter of 3.05 mm. This is in concordance to what
happens when the ASF are tested under quasistatic conditions where the
ASF with smaller microspheres has lower compression strength than the
ASF with larger microspheres. Inspecting the impacted samples, it was
found that the damage is somewhat similar to the samples of 1100 ASF
aluminum alloy with microspheres diameter of 4.45 mm: ductile failure
occurs in these samples (Figure 51).
85
Figure 50. – Comparison of load –penetration depth behavior of ASF (1100 aluminum
matrix- 3.05 mm microsphere diameter) with different impact energies: (a) 60 J (b) 120 J
(c) 180 J.
Figure 51. – Side, top and bottom views of the impacted specimens of ASF (1100
aluminum matrix- 3.05 mm microsphere diameter) with different impact energies: (a) 60
J (b) 120 J (c) 180 J. Samples size is 93 mm x 93 mm x 12.7 mm.
86
Returning to the relationship between microsphere diameter and
compression strength, there are contradictory results. For example, P.K.
Rohatgi reported decreasing compression strength with decreasing
cenosphere size [15]. In that study it is mentioned that the higher void
content in the syntactic foams containing smaller size microspheres may be
responsible for this type of trend. But after visual inspection of the cast
samples, we have found no evidence of void content in our samples with
microspheres of 3.05 mm of diameter. On the other hand, there are several
reports on increasing compression strength with decreasing microsphere size
[60, 14, 63]. A. Rabiei published a paper on composite metal foams in which
two different size of low carbon steel microsphere (1.4 mm and 3.7 mm)
were used to produce aluminum syntactic foams and the aluminum foam
with the smaller steel microspheres showed higher compression strength
[63]. Wu et al studied effects of the hollow sphere diameter on the
compressive strength. They found that smaller spheres produce higher
compressive strength than the larger ones [14]. Orbulov et al found that the
smaller SL150 type microballoons gave higher compressive strength
compared to the larger SL300 type. They mentioned that the SL150
microballoons are significantly smaller and they also have thinner walls. The
smaller diameter and higher curvature give higher compressive strength and
mechanical stability to the microballoons [60]. It is known that the strength
of the microspheres exerts a great influence in the resulting strength of the
foam. Therefore, it is necessary to have an idea of the strength of the
microspheres in order to explain the compression strength of the resulting
foams. The uniaxial compression method was used: random groups of 30
microspheres of each size were selected and compressed until failure [64].
The results show that the larger micrspheres are slightly stronger (31.85 N)
87
than the smaller microspheres (28.78 N). This could explain at least in part
why the the larger microspheres produce stronger foams. Possible causes of
the lower strength of the smaller microsphere could be related to the
morphology of the microspheres, the way they are manufactured, variability
of wall thickness, etc [65]. Further work is required to make more general
conclusions about the effect of microsphere size on the mechanical
properties of the ASF plates. There has been suggestions on the literature
advocating for the use of high quality engineered hollow microspheres
produced by powder metallurgy (instead of melt atomization), but they
usually are most costly than the microspheres used in this study [66].
As it can be seen in figure 52 aluminum syntactic foams are not as strong
and energy absorbent that steel syntactic foams, but they are certainly
stronger and more energy absorbent than conventional aluminum foams.
Currently produced aluminum foams are too weak for automotive or military
applications. Metal syntactic foams provide designers of protection systems
with foam materials that have a wide range of energy absorbing capacities.
The selection of a specific metal foam material will depend on the
application: maximum allowable stress and amount of energy to be absorbed.
88
Figure 52. – Stress-strain curves of steel syntactic foam, aluminum syntactic foam and
conventional aluminum foam.
Also, it is interesting to compare the performance of the aluminum syntactic
foam produced in his study with the performance of aluminum “stochastic”
foam under impact loading. Mohan et al published a study in which it is
reported that closed cell Al foam with an average relative density of 9.5%
and 20 mm thick has energy absorption factor of 21 % when impacted with
almost 60 J (completely perforation using 2.65 kg impact weight). In this
case, we can estimate that the minimum amount of energy necessary to
perforate those samples is around 12.6 J. Although, our samples were tested
at different conditions (6.4 kg and samples 12.7 mm thick), the minimum
energy absorbed necessary to perforation is almost twelve times higher than
the aluminum “stochastic” foam (148.41 J).
There are different areas that can be investigated to improve their
mechanical properties of aluminum syntactic foams: incorporation of wire
89
meshes into the metal foam structure to control their density and therefore
the compression strength and energy absorption capacity, this approach has
already been study in steel syntactic foams [67]; the addition of
nanoparticles (nanosized SiC reinforcements) in the liquid aluminum before
infiltration could play a big role in improving the mechanical performance
[68] and finally the effect of heat treatments should be considered too.
90
5. – Conclusions.
The production of syntactic steel foams by pressure infiltration and by
gravity-fed infiltration of hollow alumina microspheres with molten steel
was demonstrated using different steel compositions. The relative density of
the mechanical pressure infiltration steel syntactic foams was 0.54
(corresponding to 4.3 g/m
3
) with monosized microspheres 1.27 mm in
diameter. The relative density of the gravity-fed infiltration steel syntactic
foams was 0.46 (corresponding to 3.59 g/m
3
) with monosized microspheres
4.45 mm in diameter. There was no evidence of clustering, unintended
porosity, or breakage of microspheres. The critical parameters in the
manufacturing of the steel syntactic foams were the steel melt temperature
and the preheat temperature of the microspheres prior to infiltration.
Syntactic steel foams produced by both methods exhibited elastic-plastic
foam behavior in compression - an initial linear elastic region was followed
by a long stress plateau, and ultimately by densification. As expected, the
compression strength and energy absorption was greater for the pearlitic
syntactic steel foam compared to the ferritic syntactic steel foam. Increasing
the intrinsic strength and toughness of the steel (as in TRIP steel) yielded the
foam with the greatest compression strength and energy absorption capacity.
Syntactic steel foams with different relative densities were also produced
and tested, showing that the foam with RD of 0.60 exhibited the maximum
energy absorption. The results demonstrate that maximizing energy
absorption capacity requires selecting an optimal relative density, as well as
balancing intrinsic strength and toughness of the foam metal. Note that all of
91
the steel foams surpassed the energy absorption capacity of Al foams, some
by as much as a factor of 6 per unit weight and by a factor of 70 per unit
volume. When compared to other energy absorbing cellular structures, such
as metallic trusses or egg-box structures, the energy absorption values of the
prototype steel foams compare favorably. The superior performance is
achieved with relatively low-cost constituent materials and a scale-able, low-
cost method of manufacture based on gravity-fed casting. The gravity-fed
infiltration method employed here can also be extended to molds with
different geometries and to other variants of the casting technique.
Syntactic aluminum foams produced by by gravity-fed infiltration were
tested by low-velocity impact testing. The Al 1100 alloy matrix (ductile
aluminum alloy) performs better than the Al 6061 alloy matrix (strong
aluminum alloy) under the impact speeds used in this study: it requires more
impact energy to achieve perforation in the Al 1100 alloy matrix (148.41 J
compared to 111.5 J) This is opposite to what it happens when the ASF are
tested under quasistatic conditions where the ASF 1100 Al alloy has lower
compression strength than the ASF 6061 Al alloy. The addition of the
facesheet helps to increase the absorbed factor energy in the case of 180 J
impact energy from 83% (without facesheet) to 92 % (with facesheet and to
delay the initiation of perforation. This increase in absorbed energy and the
impact load is due to additional resistance of the facesheet. The ASF
samples manufactured with 1100 aluminum alloys and microspheres
diameter of 4.45 mm are stronger and requires more impact energy to
achieve perforation than samples manufactured with 1100 aluminum alloys
and microspheres diameter of 3.05 mm (148.41 J compared to 90 J) Further
92
work is required to explain the effect of microsphere size and wall thickness
on the mechanical properties of the ASF plates.
Potential applications of syntactic metal foams include lightweight energy
absorbing components for improving crashworthiness of motor vehicles,
enhancing armor in military vehicles and structures, and new blast-resistant
structures. They could be implemented in bumpers or crumples zones for
suburban cars, railcars or trams and in vehicles that requires protection
against blast protection from explosives.
93
6. - Future actions
1 - Production of larger steel syntactic foam samples using the gravity-fed
infiltration method (scaling-up). As a first step, large samples of aluminum
foams have being manufactured to help identify the challenges producing
larger samples. The technique known as counter-gravity casting could be
also a good alternative to produce steel syntactic foam because of the
reduced heat losses during the transfer of the liquid steel from the furnace to
the preheated mold.
2 - Studying the mechanical properties of the steel syntactic foams under
impact testing and blast testing.
3 - Studying the influence of the microsphere size and wall thickness in the
mechanical properties of the steel syntactic foams. There are powder-
metallurgy techniques to produce customized ceramic microspheres with
specific size and wall thickness according to requirements. [24, 25, 26].
4 - Modeling of mechanical behavior of steel syntactic foams: modeling the
geometry of the foam and numerical simulation of the compression behavior
(elastic region and densification region).
94
Figure 53. – (a) CT scan of a group of packed plastic spheres simulating the hollow
alumina microspheres (diameter 4.45 mm). (b) Model of the steel syntactic foam in
ABAQUS.
5 – Producing open-cell steel foam using the casting replication method.
Sodium aluminate (NaAlO2) has been shown to be a suitable space holder to
handle the high temperatures of molten steel [37]. Open-cell steel foam
could have different functional applications as shown in Figure 54. Also,
reducing the relative density of the steel foam is attractive because it will
allow the steel foam to absorbent energy at lower stress compared to the
syntactic steel foam produced in this study.
Figure 54. – Applications of cellular material according to the degree of openness needed
and whether the application is more functional or structural. [69]
95
7. – References
[1] C. Park, S.R. Nutt, Mater. Sci. Eng. A 288 (1) (2000) 111–118.
[2] B.P. Neville, A. Rabiei, Mater. Des. 29 (2) (2008) 388–396.
[3] S.K. Hyun, H. Nakajima, Mater. Trans. (JIM) 43 (3) (2002) 526–531
[4] M. Mukherjee, F. Garcia-Moreno, J.Banhart. Scripta Materialia 63(2010)
235-238.
[5] A.G.Evans, J.W. Hutchinson, N.A.Fleck, M.F.Ashby, H.N.G Wadley.
Progress in Materials Science 46(2001) 309-327.
[6] D. J. Sypeck, H.N.G. Wadley. Advanced Engineering Materials 2002, 4,
10.
[7] C.Lim, I. Jeon, K.Kang. Materials and Design. 30 (2009) 3082-3093.
[8] D.T.Queheillalt, H.N.G.Wadley. Acta Materialia 53(2005) 303-313.
[9] H.N.G.Wadly, N.A Fleck, A.G.Evans. Composite Science and
Technology 63(2003) 2331-2343.
[10] H.N.G.Wadly. Phil. Trans.R. Soc. A (2006) 364, 31-68.
[11] H.N.G. Wadley. 2006. Philosophical Transactions of the Royal Society
of London A 364: 31-68.
[12] S. S. Chung Ho. Lost foam casting of periodic cellular materials with
aluminum and magnesium alloys. University of Toronto. 2009
[13]Y. Zhao, X. Tao, X.Xue. Mater. Sci. and Tech. (MS&T) 2008. Oct 5-9,
2008, 2607-2615.
[14] G.H. Wu, Z.Y. Dou,* D.L. Sun, L.T. Jiang, B.S. Ding and B.F. He.
Scripta Materialia 56 (2007) 221–224.
[15] P.K. Rohatgia, J.K. Kima, N. Guptab, S. Alarajc, A. Daoud.
Composites: Part A 37 (2006) 430–437.
96
[16] I. N. Orbulov, Á. Németh, J. Dobránszky. Gépészet. 2008. Budapest,
29-30. May 2008.
[17] D. K. Balch, D. C. Dunand. Acta Materialia 54 (2006) 1501–1511.
[18] Q. Zhang, P. D. Lee, R. Singh, G. Wua, T. C. Lindley. Acta Materialia
57 (2009) 3003–3011.
[19] R.A. Palmer, K. Gao, T.M. Doan, L. Green, G. Cavallaro. Materials
Science and Engineering A 464 (2007) 85–92
[20] X.F. Tao, L.P. Zhang, Y.Y. Zhao. Materials and Design 30 (2009)
2732–2736.
[21] A. Daoud. Materials Science and Engineering A 488 (2008) 281–295.
[22] D.P. Mondal, S. Das, N. Ramakrishnan, K. U. Bhasker. Composites:
Part A 40 (2009) 279–288.
[23] A. Daoud, M.T. A. El-khair, M. Abdel-Aziz, P. Rohatgi. Composites
Science and Technology 67 (2007) 1842–1853.
[24] L. J. Vendra, A. Rabiei. Materials Science and Engineering A 465
(2007) 59–67.
[25] Nadler, J. L. Clark, J. K. Cochran, K. J. Lee, T. H. Sanders. Proceedings
MetFoam 99, Bremen Germany, June 13-15, 1999.
[26] M. Hartmann, K. Reindel, R.F. Singer. Materials Research Symposium
Proceedings, pp. 211-216, Vol. 521, 1998.
[27] W. Deqing, S. Ziyan, G.Hong, H.F.Lopez. Journal of Materials
Synthesis and Processing. Vol. 9, N. 5. September 2001, 247-251.
[28] J. Banhart, M.F.Ashby, N.A. Fleck. Metal foams and porous metal
structures. (1999), 159-162.
[29] Q. Fabrizio, A. Boschetto, L. Rovatti, L. Santo. Materials Letters 65
(2011) 2558–2561.
97
[30] S. Yu, J. Liu, M. Wei, Y. Luo, X. Zhu, Y. Liu. Materials and Design 30
(2009) 87–90.
[31] W. Lucai, C. Yuyong, W. Fang, W. Jianguo, Y. Xiaohong. China
Foundry. Vol. 5. N-4. 240-243.
[32] B.Y. Hur, Y.S. Um, J.S. Cho, B.K. Park. International Conference
“Advanced Metallic Materials”. 5 − 7 November, 2003, Smolenice, Slovakia.
[33] E. M. Castrodeza, C. Mapellia. Journal of Materials Processing
Technology 209 (2009) 4958–4962.
[34] K.S. Chou, M.A. Song. Scripta Materialia. 46(2002) 379-382.
[35] A. K. Shaikdawood, S. S. M. Nazirudeen. Jordan Journal of Mechanical
and Industrial Engineering. Volume 4, Number 2, March. 2010 ISSN 1995-
6665 Pages 292 - 299
[36] A. Gruttadaria, D. Mombelli, E.M.Castrodeza, C.Mapelli. Revista
Materia, V.15, N.2, 192-198, 2010.
[37] Y. Boonyongmaneerat, D.C.Dunand. Advanced Engineering Materials,
2008, 10, 4.
[38] A.H. Brothers, R. Scheunemann, J.D. DeFouw, D.C. Dunand. Scripta
Materialia 52 (2005) 335–339.
[39] L.J. Gibson, M.F. Ashby, 1997. Cellular Solids: Structure and
Properties, 2nd ed. Cambridge University Press, Cambridge.
[40] J. Zhang, M.F. Ashby, Int. J. Mech. Sci., 34(1992) 475.
[41] B. Shah, K. Parsons, L. Swenson, Mod. Cast. 92-2 (2002).
[42] A. Rabiei, A.T. O’ Neill. Mater. Sci. Eng. A. 404, 1-2 (2005)159–164.
[43] T. F. Cox, Stat. & Prob. Letters. 1-4 (1983) 161-166.
[44] A. J. Mian, N. Driver, P. T. Mativenga, Int. J. Adv. Manuf .Technol. 50,
1-4 (2010) 163–174.
98
[45] J.F.Shackelford. Introduction of Material Science for Engineers. 6th
edition (2005) 220-221.
[46] Y. Sirong, L. Jiaan, L. Yanru, L. Yaohui. Mater. Sci. Eng. A. 457, 1-2
(2007) 325–328.
[47] G. Standke, T. Müller, A. Neubrand, J. Weise, M. Göpfert. Advanc.
Eng. Mater. 12-3 (2010) 189–196.
[48] M. Bahrainia, T. Minghettib, M. Zoelligc, J. Schubertd, K. Berrothd, C.
Schelleb, T. Graulea, J. Kuebler, Compos. Part A 40-10 (2009) 1566–1572.
[49] K. Nakashima, K. Takihira, K. Mori, N. Shinozaki, Mater. Trans. (JIM).
33-10 (1992) 918-926.
[50] D. Ruan, G. Lu, F.L. Chen, E. Siores, Compos. Struct. 57, 1-4 (2002)
331–336.
[51] S. Martin, S. Wolf, U. Martin, L. Krüger, A. Jahn. ESOMAT 2009,
05022 (2009), doi:10.1051/esomat/200905022.
[52] J. P. Schramm, D. C. Hofmann, M. D. Demetriou, W. L. Johnson. Appl.
Phys. L. 97, 241910 (2010).
[53] A. Ohrndorf, U. Krupp (2005), "Correlation between fracture behavior
and ductility of the cell strut material in case of metallic foams", ICF XI -
11th International Conference on Fracture.
[54] M. Weider, K. Eigenfeld. Steel Res. Int. 82 (2011) No. 9.1064-1069.
doi: 10.1002/srin.201100071
[55] L. Krüger, S. Wolf, U. Martin, S. Martin, P. R. Scheller, A. Jahn, A.
Weiß. J. Phys 240 (2010) 012098.
[56] D.T. Queheillalt, H.N.G. Wadley. Acta Mater 2005; 53(2):303-313.
[57] M.Zupan, C.Chen, N.A.Fleck. Int.J.Mech.Sci. 2003; 45(5):851-871.
[58] E.M.A. Maine, M.F.Ashby. Materi. Des. 23(2002) 307-319.
[59] G. Castro, S.R. Nutt. Mater. Sci. Eng. A 553(2012) 89–95.
99
[60] I. N. Orbulov, J. Ginsztler. Composites: Part A 43 (2012) 553–561.
[61] D. Liu, B. B. Raju, X. Danga. Int. J. Imp. Eng.24(2000) 733–746.
[62] K. Mohan, T. H. Yipa, S. Idapalapatib, Z. Chena, Mater. Sci. Eng. A
529 (2011) 94– 101.
[63] A. Rabiei, L.J. Vendra. Materials Letters 63 (2009) 533–536.
[64] K. Yan, X. Xie, B. Li, J. Yuan, J. Zhang. Mat. Sci. Eng. A 528 (2011)
8671– 8675.
[65] M. Koopman, G. Gouadec, K. Carlisle, K.K. Chawla, G. Gladysz.
Scripta Materialia 50 (2004) 593–596.
[66] D. D. Luong, O. M. Strbik III, V. H. Hammond, N. Gupta, K. Cho J.
Alloys and Comp. 550 (2013),412-422.
[67] D. K. Balch, J. G. O’Dwyer, G. R. Davis, C. M. Cady, G. T. Gray III, D.
C. Dunand, Mater. Sci. Eng. A 391 (1 –2) (2005) 408-417.
[68] X. Li, Y. Yang. AFS Transactions (2007)133(02)1-12.
[69] J. Banhart, Progress in Materials Science 46 (2001).
Abstract (if available)
Abstract
In this study, we report two procedures for producing lab-scale syntactic steel by melt infiltration of millimeter-sized alumina microspheres: mechanical pressure infiltration and gravity-fed infiltration. Both methods yield foam with uniform distributions of microspheres and negligible unintended porosity. The most critical parameters in the manufacture of the syntactic steel foams are the melt temperature and the preheat temperature of the microspheres prior to infiltration. The preheatment temperature of the microspheres must be close to the melting temperature of steel. ❧ Syntactic steel foams with relative density of about half of solid steel densities were produced using monosized microspheres randomly situated in a mold. Microspheres with a diameter of 1.27 mm were used for the mechanical pressure infiltration method and microspheres with a diameter of 4.45 mm for the gravity-fed infiltration method. Different steel chemical compositions were selected to produce steel foams of different inherent yield strength: including several ferritic-pearlitic steels and one TRIP steel (TRansformation-Induced Plasticity). The resultant foams were characterized by chemical and microstructural analysis. The microstructure of the samples consisted of blends of ferritic and pearlitic constituents in varying proportions for the ferritic-pearlitic steels, while the cast TRIP steel matrix presented an austenitic microstructure. ❧ The basic mechanical properties of the steel syntactic foams were studied under compression loading. The pearlitic syntactic foams have greater compression strength and energy absorption capacity than the ferritic syntactic foams, but the TRIP steel syntactic foam exhibited the highest compression strength and highest energy absorption capacity. The properties of the steel syntactic foams were compared to those of other steel foams, aluminum foams and other cellular structures reported in the literature. ❧ We present also the compression and impact behavior of aluminum syntactic foams (ASF) produced by gravity-fed infiltration of millimeter-sized ceramic microspheres. Aluminum syntactic foams with relative density of 0.46 were produced using monosized microspheres (4.45 mm and 3.05 mm) randomly situated in a mold and two types of aluminum alloy matrices: 1100 and 6061. The impact behavior was experimentally investigated using a drop-weight testing machine. The impact tests were carried out using a hemispherical indenter (16.1 mm diameter) on ASF plates (93 mm x 93 mm x 12.7 mm thick). We have studied the influence of the type of aluminum matrix, size of microspheres and the addition of a face sheet into the impact behavior of ASF. Results show that 1100 Al alloy outperforms 6061 Al alloy, it can absorb higher amount of energy at higher velocities (penetration)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Fiber-reinforced syntactic foams
PDF
Synthesis, characterization, and mechanical properties of nanoporous foams
PDF
Fiber reinforced hybrid phenolic foam
PDF
Deformation behavior and microstructural evolution of nanocrystalline aluminum alloys and composites
PDF
Mechanical behavior and microstructure optimization of ultrafine-grained aluminum alloys and nanocomposites
PDF
Fatigue and fracture of pultruded composite rods developed for overhead conductor cables
PDF
Synthesis and mechanical behavior of highly nanotwinned metals
PDF
Mechanical behavior of materials in extreme conditions: a focus on creep plasticity
PDF
Electrodeposition of platinum-iridium coatings and nanowires for neurostimulating applications: fabrication, characterization and in-vivo retinal stimulation/recording EIS studies of hexavalent a...
PDF
Development of high frequency focused transducers for single beam acoustic tweezers
PDF
Processing, mechanical behavior and biocompatibility of ultrafine grained zirconium fabricated by accumulative roll bonding
PDF
Chemical vapor deposition of graphene and two-dimensional materials: synthesis, characterization, and applications
PDF
Microstructural evolution and mechanical properties of a copper-zirconium alloy processed by severe plastic deformation
PDF
Ageing and mechanical failure of fiber reinforced polymers
PDF
The extension of selective inhibition sintering (SIS) to high temperature alloys
PDF
Shock wave response of in situ iron-based metallic glass matrix composites
PDF
Processing and properties of phenylethynyl-terminated PMDA-type asymmetric polyimide and composites
PDF
High temperature creep behaviors of additively manufactured IN625
PDF
Nanomaterials under extreme environments: a study of structural and dynamic properties using reactive molecular dynamics simulations
PDF
Microstructural evolution and mechanical properties of an aluminum alloy processed by equal-channel angular pressing and high-pressure torsion
Asset Metadata
Creator
Castro, Gerhard
(author)
Core Title
Metallic syntactic foams synthesis, characterization and mechnical properties
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Materials Science
Publication Date
04/23/2013
Defense Date
04/23/2013
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
compression,impact test,infiltration,metal foam,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Nutt, Steven R. (
committee chair
), Goo, Edward K. (
committee member
), Sammis, Charles G. (
committee member
)
Creator Email
gercasgil001@hotmail.com,gerhardc@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-242857
Unique identifier
UC11288131
Identifier
etd-CastroGerh-1584.pdf (filename),usctheses-c3-242857 (legacy record id)
Legacy Identifier
etd-CastroGerh-1584.pdf
Dmrecord
242857
Document Type
Dissertation
Rights
Castro, Gerhard
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
compression
impact test
infiltration
metal foam