Académique Documents
Professionnel Documents
Culture Documents
A R T I C L E I N F O A BS T RAC T
Keywords: Closed-cell aluminum foams represent a unique class of solid cellular light metals that are made by deliberately
CNT reinforced Aluminum foam introducing voids or pores during fabrication. This lightweight material is able to undergo large deformation at a
Liquid metallurgy nearly constant stress known as Plateau Stress because of which aluminum foams are good energy absorbers
Quasi-static and dynamic compression under dynamic loads such as an impact. In this investigation, carbon nanotubes (CNT) reinforced closed-cell
Plateau stress
aluminum foams were fabricated using the liquid metallurgy route through the dissociation of a foaming agent
Energy absorption
Relative density
within the liquid metal. Four different relative densities of CNT reinforced Al-foam were used: 0.16, 0.20, 0.26
and 0.30, to study the effect of strain rate on the mechanical properties. The compressive mechanical behavior
of CNT reinforced Al-foam has been studied under quasi-static and dynamic loading conditions. The high strain
rate compressive response was investigated using a Split Hopkinson Pressure Bar (SHPB) over a range of strain
rates up to 2750 s−1. Mechanical properties such as peak stress, plateau stress and energy absorption increased
with the increase in relative density; however, the densification strain decreased with the increase in relative
density. Dynamic compressive properties improved as the strain rate increased indicating that this material is
strain rate dependent. Among all the foams, the 0.30 relative density exhibited the highest mechanical
properties whereas the 0.20 relative density foam displayed the highest strain rate sensitivity.
⁎
Corresponding author.
http://dx.doi.org/10.1016/j.msea.2017.01.100
Received 10 November 2016; Received in revised form 23 January 2017; Accepted 28 January 2017
Available online 31 January 2017
0921-5093/ © 2017 Elsevier B.V. All rights reserved.
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
2. Experimental procedures
2.1. Materials
18
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
⎛A E ⎞
σs (t )=⎜ B B ⎟ εT (t )
⎝ As0 ⎠ (2)
⎛ 2C ⎞ t
εs (t )=⎜ OB ⎟
⎝ Hs0 ⎠
∫ εR (t ) dt
(3)
Fig. 5. Stress-strain response of 2 wt% CNT Al-foams with different relative densities
under quasi-static compression (0.001 s−1). where CoB, AB and EB are the wave speed, cross-sectional area and
elastic Young's modulus of SHPB bars, respectively. HS0 and AS0 are the
density of the foam, the dimensions of the specimen were accurately length, and the cross-sectional area of the specimen, respectively.
measured and weighed using a digital caliper and digital balance. The forces in incident and transmitted bars are given by Eqs. (4)
Relative density was estimated as density of foam/density of base and (5), respectively. To satisfy the 1-D assumption in SHPB, specimen
aluminum alloy. must be under dynamic stress equilibrium and deform under constant
strain rate [25,26]. Stress equilibrium is achieved when the forces on
front (F1) and rear (F2) surfaces of specimen are nearly equal, which
2.2. Mechanical characterization results in expression (6). This assumption is satisfied when long bars
(incident and transmitted) have uniform elastic, homogeneous, iso-
The conventional split Hopkinson pressure bar (SHPB) apparatus tropic characteristics across its cross-section and along its length; such
19
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
Fig. 6. Compressive stress-strain response of 2 wt% CNT Al-foams with different relative densities at a strain rate of (a) 1300 s−1, (b) 1800 s−1, (c) 2300 s−1 and (d) 2750 s−1.
Table 1
Peak stress (σPeak), plateau stress (σPl) and densification strain (εD) as a function of strain rate and relative density.
Relative Density Strain Rate (s−1) Strain Rate (s−1) Strain Rate (s−1) Strain Rate (s−1) Strain Rate (s−1)
0.001 1300 1800 2300 2750
σPeak (MPa) σPl (MPa) εD σPeak (MPa) σPl (MPa) σPeak (MPa) σPl (MPa) σPeak (MPa) σPl (MPa) σPeak (MPa) σPl (MPa)
0.16 3.51 4.23 0.71 6.21 4.22 6.53 5.54 7.13 4.82 7.5 5.76
0.20 8.15 7.52 0.59 7.53 6.71 9.15 8.52 12.04 12.24 13.1 13.67
0.26 12.03 10.54 0.55 11.12 11.54 11.85 12.11 14.33 13.56 15.2 15.8
0.30 18.06 16.60 0.46 17.57 17.14 18.66 18.13 20.54 21.14 21 21.4
Fig. 7. Dynamic compressive stress-strain response of 2 wt% CNT Al-foams of RD=0.20 Fig. 8. Dynamic compressive stress-strain response of 2 wt% CNT Al-foams of RD=0.30
as a function of strain rate. as a function of strain rate.
as EB, AB and COB. is achieved in the specimen. The stress equilibrium in the specimen
takes less than 40 μs to be achieved. In every test we ensured that a
F1=AB EB (εI +εR ) (4)
fairly constant strain rate was achieved over the duration of the test.
F2=AB EB εT (5) Quasi-static compression testing was carried out at strain rate of
1×10−3 s−1 using ADMET material testing machine.
εI +εR=εT (6)
To assure the 1-D wave theory assumption of SHPB holds, the 3. Results and discussion
specimen aspect ratio (length/width) is selected to be 1, and a copper
(C14500) pulse shaper was used for this work, which helped to achieve The stress-strain response of 2 wt% CNT reinforced foam under
stress equilibrium in the specimens [27]. The detailed process for quasi-static compression is shown in Fig. 5, which is similar in nature
selecting specimen aspect ratio and pulse shaper are reported else- to a typical compression behavior shown in Fig. 1. Peak stress is
where [18]. Fig. 4 shows that the stress equilibrium condition of Eq. (6) reported as the maximum stress value attained just before the start of
20
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
Fig. 12. Peak stress against strain rate at different relative densities.
Fig. 9. Peak stress as a function of relative density at different strain rates.
Fig. 13. Plateau stress against strain rate at different relative densities.
Fig. 10. Plateau stress as a function of relative density at different strain rates.
Table 3
Table 2 Variation of “m” with relative density.
Variation of “A” and “b” with strain rate.
# Relative Density mPeak mPlateau
# Strain Rate (s−1) Peak Stress Plateau Stress
1 0.16 0.24 0.40
A b A b 2 0.20 0.65 0.69
3 0.26 0.45 0.45
1 1300 104.80 1.58 208.8 2.13 4 0.30 0.26 0.37
2 1800 133.02 1.61 216.2 2.04
3 2300 135.50 1.55 261.8 2.12
4 2750 131.10 1.53 284.6 2.07
density varies from 0.16 to 0.30.
Fig. 6a-d show engineering stress-strain plots under dynamic
compression at different strain rates and relative densities. At
1300 s−1, it is found that compressive stress significantly increases as
the relative density increases. Hence, the highest peak stress and
plateau stress of 17.5 MPa and 17.0 MPa, respectively, are observed for
the relative density of 0.30. For RD=0.16 foam, the peak stress and
plateau stress increase up to a strain rate of 1800 s−1 and no further
increase with strain rate is observed. The higher relative density foams
(0.20, 0.26 and 0.30) exhibit an increase in peak and plateau stresses
with strain rate up to 2300 s−1 with nearly no gain beyond this strain
rate. It should be noted that higher relative density foam is stiffer due
to the effect of having thicker cell walls, which in turn increases the
stress levels. To better visualize the combined effect of strain rate and
relative density, a summary of the compressive stresses of this
investigation is presented in Table 1.
Fig. 11. Ln‘Apl’ vs. Ln'Strain Rate’. It is observed that compressive stresses at any fixed relative density
increase with strain rate. In a broad prospective, higher increase in
plateau region. Plateau stress is reported as the average stress plateau stress is observed for higher relative densities as the strain rate
throughout the plateau region (see Fig. 1). It is observed that stress- increases. This shows an agreement with the reported results of Raj
strain curves show small oscillations in the plateau region. This is et al. [12] that higher density foam, RD > 0.15, exhibits significant
attributed to the localized cell deformation and sequential compaction increase in compressive stress under dynamic compression. The
of layers throughout the plateau region [9,12]. It should be noted that compressive stress–strain curves of foam at various strain rates for
peak stress and plateau stress significantly increase as the relative relative density of 0.20 and 0.30 are shown in Fig. 7 and Fig. 8,
21
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
Fig. 14. Cross-sectional view of closed cell Al-foam composite with different relative densities (a) 0.15, (b) 0.20, (c) 0.26, and (d) 0.30.
Fig. 15. Variations of densification strain as a function of relative density at strain rate Fig. 16. Energy absorption as a function of strain rate and relative density.
of 0.001 s−1.
respectively. For the relative density of 0.20, peak and plateau stresses Table 4
displayed significant increases as the strain rate increased, whereas Energy absorption (WV) as a function of strain rate and relative density.
those with relative density of 0.30 exhibited smaller increase with the
Relative Strain Strain Strain Strain Strain
increase in strain rate. Hence, this indicates that foam with relative Density Rate Rate Rate Rate Rate
density of 0.20 is more sensitive to strain rate. (s−1) (s−1) (s−1) (s−1) (s−1)
In this study, peak stress (σPeak) also termed as the elastic collapse 0.001 1300 1800 2300 2750
stress (σel) is shown in Fig. 1. Generally, the variation of peak stress WV (MJ/ WV (MJ/ WV (MJ/ WV (MJ/ WV (MJ/
m3) m3) m3) m3) m3)
(σPeak) and plateau stress (σPl) with relative density follow a power
law relationship presented in Eqs. (7) and (8) [1]. 0.16 0.52 1.26 1.37 1.41 1.62
0.20 2.34 2.42 2.54 3.39 3.64
⎛ ρ ⎞bP 0.26 3.02 3.32 3.66 4.28 4.39
σPeak =AP ⎜ f ⎟ 0.30 4.87 5.24 5.78 6.35 6.27
⎝ ρs ⎠ (7)
22
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
Fig. 17. Deformation of closed cell Al-foam (RD=0.20) at different strain (a) 0%, (b) 5%, (c) 10%, and (d) 20%.
⎛ ρ ⎞bPl sensitivity observed in 0.20 relative density foam. For foam with
σPl =APl ⎜ f ⎟ RD=0.26, more small size pores appeared leading to less uniform cell
⎝ ρs ⎠ (8)
size. As the relative density increased to 0.30, cell size becomes smaller
where ‘Ap’ and ‘Apl’ are the respective strengthening coefficient, ρf is with an average cell size of 0.8 ± 0.2 mm and the foam has more solid
the density of foam, ρs is the density of solid metal and ‘bp’ and ‘bpl’ are areas.
the respective exponent to relative density. Fig. 9 shows the variation of Fig. 15 shows the variations of densification strain as a function of
peak stress as a function of relative density at different strain rates. The relative density of specimens tested under quasi-static compression
variation of plateau stress as a function of relative density at different (strain rate of 0.001 s−1). The densification strain is considered as the
strain rates is shown in Fig. 10. The data for peak stress and plateau strain corresponding to the intersection of tangents drawn on the
stress as a function of relative density were curve fitted using a power densified and the plateau regions. It is noted that densification strain
low relationship y=Axb. The values of ‘A’ and ‘b’ are presented in decreases as the relative density increase. The curve fit line follows the
Table 2. The value of ‘b’ in this present study shows a good agreement equation of a linear line, y=mx+b in which m equals −1.5 and b equals
with the reported values (1.5–3.0) for aluminum foams [1]. Fig. 11 0.94. The constants of this straight-line equation are within ± 6% of the
presents a the variation of Ln‘Apl’ with Ln'Strain Rate’. constants in Eq. (9) provided by Gibson et al. [1]. It may be noted that
Peak stress and plateau stress are plotted against strain rate for as the relative density varies from 0.16 to 0.30 the corresponding
different relative densities on a log-log plot (see Figs. 12 and 13) to densification strain varies from 0.71 to 0.46. For RD less than 0.16, we
estimate the strain rate sensitivity. The strain rate sensitivity ‘m’ is hypothesize that the densification strain would be invariant to the
calculated as the slope of each line (in Figs. 12 and 13) for each relative change in relative density, following the study by Mondal et al. [3] even
density presented in Table 3. It is noted that peak stress for relative though their foam system had a different reinforcement, namely fly-
densities 0.2 and 0.26 show stronger strain rate sensitivity as compared ash. The densification strain cannot be determined for dynamic tests as
to others. For plateau stress, the strain rate sensitivity of relative the impact pulse duration is not long enough to produce sufficient
densities 0.16, 0.26 and 0.30 varies in the range 0.37–0.45, while a densification.
relative density of 0.20 displays the highest strain rate sensitivity of ⎛ρ ⎞
0.69. Fig. 14a-d show the cross-sectional view of closed cell Al-foam εD=1 − 1.4 ⎜ f ⎟
⎝ ρs ⎠ (9)
composite with different relative densities 0.16, 0.20, 0.26, and 0.30,
respectively. Foam with RD=0.16 has irregular cell size and shape The energy needed under compression to deform any foam speci-
whereas foam with RD=0.20 has more uniform cell size, cell wall men up to particular strain is defined as the energy absorption capacity.
thickness and cell distribution an average cell size of 1.3 ± 0.3 mm. Under uniaxial compressive loading, energy absorption, WV, per unit
Therefore, this might be the plausible reason for the higher strain rate volume is given as the area under stress-strain curve up to a specific
23
A. Aldoshan, S. Khanna Materials Science & Engineering A 689 (2017) 17–24
•
[20] D.P. Mondal, M.D. Goel, S. Das, Effect of strain rate and relative density on
At all applied strain rates, mechanical properties significantly compressive deformation behaviour of closed cell aluminum-fly ash composite
increase as the relative density increases in terms of peak stress, foam, Mater. Des. 30 (4) (2009) 1268–1274.
[21] ASTM, Standard test methods for determining average grain size, ASTMP, E112-
plateau stress and energy absorption.
•
10, p. Pennsylvania, 2011.
For similar strain rates, plateau stress is approximately 230–300% [22] T. Mukai, T. Miyoshi, S. Nakano, H. Somekawa, K. Higashi, Compressive response
higher for relative density of 0.30 compared to the low relative of a closed-cell aluminum foam at high strain rate, Scr. Mater. 54 (2006) 533–537.
[23] T. Hamada, H. Kanahashi, T. Miyoshi, N. Kanetake, Effects of the strain rate and
density foam (RD=0.16). Energy absorption in the foam of RD=0.30
alloying on the compression characteristics of closed cell aluminum foams, Mater.
is also 240–270% higher than low relative density foam RD=0.16. Trans. 50 (6) (2009) 1418–1425.
• For all relative densities, compressive stresses and energy absorp- [24] H. Kolsky, An investigation of the mechanical properties of materials at very high
rates of loading, Proc. Phys. Soc.: Sect. B,1949, pp. 676–700.
tion displayed a strain rate dependence.
•
[25] B.A. Gama, S.L. Lopatnikov, J.W. Gillespie, Hopkinson bar experimental techni-
The CNT foam of RD=0.20 exhibits the highest strain rate sensitivity que: A critical review, Appl. Mech. Rev. 57 (4) (2004) 223–250.
among all the foams studied. It is postulated that this behavior [26] W.W. Chen, B. Song, Spilt Hopkinson (Kolsky)Bar: Design, Testing and
relates to the more uniform cell structure in the 0.20 relative density Applications, Springer, New York, 2011.
[27] K.S. Vecchio, F. Jiang, Improved pulse shaping to achieve constant strain rate and
foam. stress equilibrium in split-hopkinson pressure bar testing, Metall. Mater. Trans. A
38 (11) (2007) 2655–2665.
24