Professional Documents
Culture Documents
J. Chan, R. Kwok
Singapore Technologies Kinetics Ltd (ST Kinetics), Singapore
Abstract
Introduction
AZ31 is a very commonly used Al-containing (or Zr-free) Mg alloy and is characterized
by: (a) low cost, (b) ease of handling, (c) good strength and ductility and (d) resistance to
atmospheric corrosion. AZ31 has recently been surface-reinforced with SiC microparticulates
[1], C60 molecules [2], and multi-walled carbon nanotubes [3], using the friction stir processing
technique. In these studies, good dispersion and hardening of the base matrix were reported.
ZK60A (Mg-Zn-Zr system) is a commonly used Zr-containing (or Al-free) Mg alloy and is
characterized by: (a) high strength and ductility after aging (T5 heat treatment), (b) good creep
resistance, (c) poor arc weldability due to hot-shortness cracking and (d) excellent resistance
weldability. Superplasticity of the Mg-Zn-Zr system has been recently studied [4-6]. Here,
superplasticity was attributed to fine grain size (lesser twinning effects) and crystallographic
textural effects. Similarly, the superplasticity of Mg-Zn-Zr system reinforced with SiC particles
of micron or sub-micron size has also been reported [7-9]. Regarding aluminium borate
whiskers, the composite interface formed with the Mg-Zn-Zr alloy matrix has also been studied
and improved [10]. The carbon nanotube (CNT) possesses a multitude of unique properties [11,
12] which have rendered it useful in some applications to date. It has superior strength and
stiffness in tension. The tensile strength and modulus of the CNT is about 30 GPa and 1 TPa,
respectively, according to Ruoff et al [13]. The CNT also bends reversibly according to Blau et
al [14]. The electrical properties of the CNT are comparable to those of metals and semi-
2221
conductors according to Avouris et al [15, 16]. The thermal conductivity of the CNT has been
predicted to be unusually high according to Kwon et al [17]. The CNT has a high aspect ratio
according to Wong et al [18], implying that it also has a high surface area to volume ratio (SA/V
ratio). The unique properties and high SA/V ratio of the CNT has lead to its combination with
other materials in efforts to create useful composite materials as shown by Chen et al [19], Chen
et al [20], Coleman et al [21] and Yao et al [22]. Open literature search has revealed that limited
attempts have been made to simultaneously increase strength and ductility of magnesium alloys
with CNT nanoparticles using a high volume production spray-deposition based solidification
processing technique.
Accordingly, the primary aim of this study was to simultaneously increase strength and
ductility of AZ31 and ZK60A with CNT nanoparticles. Disintegrated melt deposition followed
by hot extrusion was used to synthesize the AZ31/CNT and ZK60A/CNT nanocomposites.
Experimental Procedures
In this study, AZ31 (nominally 2.50-3.50 wt%Al, 0.60-1.40 wt%Zn, 0.15-0.40 wt%Mn,
0.10 wt%Si, 0.05 wt%Cu, 0.01 wt%Fe, 0.01 wt%Ni, balance Mg) rod supplied by Asianovena
Pte Ltd (Singapore) and ZK60A (nominally 4.80-6.20 wt%Zn, 0.45 wt%Zr, balance Mg)
supplied by Tokyo Magnesium Co. Ltd. (Yokohama, Japan) were used as the matrix material.
AZ31 rod and ZK60A block were sectioned to smaller pieces. All oxide and scale surfaces were
removed using machining. All surfaces were washed with ethanol after machining. CNT powder
(vapor grown, 94.7% purity, 40-70nm outer diameter) supplied by Nanostructured & Amorphous
Materials Inc (Texas, USA) was used as the reinforcement phase. Monolithic AZ31 and ZK60A
were cast using the disintegrated melt deposition (DMD) method [23]. Ingots of 40mm diameter
were obtained following the deposition stage. To form the AZ31-CNT and ZK60A-CNT
nanocomposites (see Figure 1), CNT nanoparticle powder was isolated by wrapping in Al foil of
minimal weight (< 0.50 wt% with respect to AZ31 or ZK60A matrix weight) and arranged on
top of the AZ31 or ZK60A blocks, with all other DMD parameters unchanged. All ingots were
sectioned into billets. All billets were machined to 36mm diameter and hot extruded using
20.25:1 extrusion ratio on a 150 ton hydraulic press. The extrusion temperature was 350°C. The
billets were held at 400°C for 60 min in a furnace prior to extrusion. Colloidal graphite was used
as a lubricant. Rods of 8mm were obtained. Microstructural characterization studies were
Al foil packet
CNT powder
alloy block
crucible
exit hole
Figure 1 Arrangement of raw materials in crucible before casting for AZ31/CNT and ZK60A/CNT nanocomposites
2222
conducted on metallographically polished monolithic and nanocomposite extruded samples to
determine: (a) grain and intermetallic phase(s) characteristics and (b) magnesium texture (in the
longitudinal and transverse directions). Vickers microhardness measurements were made on
polished monolithic and composite extruded samples. Smooth bar tensile properties of the
monolithic and composite extruded samples were determined based on ASTM E8M-05.
Compressive properties of the monolithic and nanocomposite extruded samples were determined
based on ASTM E9-89a. Fractography was performed on the tensile-fracture surfaces.
Microstructural Characteristics
Results of microstructural characterization studies are listed in Table 1. Microstructural
analysis results revealed that average grain size was reduced while aspect ratio remained
constant in the case of each nanocomposite. Nearly equiaxed grains were observed in each
monolithic material and nanocomposite. Grain size was smaller in the case of nanocomposite,
suggesting the ability of CNT nanoparticles to serve as either nucleation sites or obstacles to
grain growth during solid state cooling. X-ray diffraction (XRD) analysis revealed the presence
of β-Al12Mg17 phase and Mg-Zn based phases in AZ31 and ZK60A systems, respectively, as
shown in Figure 2. In each nanocomposite, no reaction products based on Mg and Al2O3 (such as
MgO in this case [24]) having more than 2% by volume were detected using X-ray diffraction
analysis. Unlike monolithic ZK60A, intermetallic phase(s) were not detected in the
nanocomposite by X-ray diffraction. Intermetallic phase(s) were not observed within the
resolution of XRD (less than 2% by volume) as well as FESEM in the ZK60A nanocomposite.
This indicated that the intermetallic particle size or content (or both) in the nanocomposite was
significantly lower compared to that of monolithic ZK60A. The intermetallic phase(s)
precipitation was possibly regulated at nanoscale due to the presence of: (a) dissolved Al and/or
(b) well dispersed CNT nanoparticles. In the case of (a), dissolved Al possibly altered the
intermetallic phase stabilities in the ZK60A matrix. This is similar to free Si from SiC
nanoparticles being possibly responsible for altered intermetallic phase stabilities in Mg-Zn
alloys as recently reported [25]. In the case of (b), dissolved Zn possibly segregated at the liquid-
CNT nanoparticle interface enabling intermetallic phase manipulation at the nanoscale. This
Table 1 Results of grain characteristics and microhardness of AZ31, ZK60A and related nanocomposites
2223
500 (a) + Al12Mg17
450 (all other peaks
counts per second (cps) 400 correspond to Mg)
350 AZ31 +
300
250
200
AZ31/1.0vol%CNT
150
100
50
0
0 20 40 60 80 100
2 Theta (degrees)
Figure 2 Representative X-ray diffraction scans (concerning intermetallic phases) of: (a) monolithic AZ31
and AZ31/CNT nanocomposite and (b) monolithic ZK60A and ZK60A/CNT nanocomposite
2224
(a) (b)
CNT
CNT
Texture results are shown in Figure 4. In monolithic AZ31, the dominant textures in the
transverse and longitudinal directions were (1 0 -1 0) and (1 0 -1 1) [and (0 0 0 2)], respectively.
In the AZ31 nanocomposite, the dominant textures in the transverse and longitudinal directions
was (1 0 -1 1). In monolithic and ZK60A nanocomposite materials, the dominant texture in the
transverse and longitudinal directions was (1 0 -1 1).
a a
c
c c
c a
c a
c
a
a
Mechanical Properties
Hardness
The results of microhardness measurements are listed in Table 1. The AZ31
nanocomposite exhibited significantly higher hardness than the monolithic material. The reverse
was true in the case of the ZK60A nanocomposite.
2225
A significant increase in microhardness by 48% was observed in the AZ31
nanocomposite when compared to monolithic material as listed in Table I. This was consistent
with earlier observations made on Mg/Al2O3, AZ31/C60 and AZ31/MWCNT nanocomposites
[24, 31, 33]. The increase in hardness of the nanocomposite in the present study can be attributed
to: (a) reasonably uniform distribution of harder CNT in the matrix, and (b) higher constraint to
localized matrix deformation during indentation due to the presence of intermetallic phase and
nanoparticles [24, 31, 32].
A significant decrease in microhardness by 17% was observed in the ZK60A
nanocomposite when compared to monolithic material as listed in Table 1. There was lower
constraint to localized matrix deformation in the ZK60A nanocomposite compared to monolithic
material. This was despite the: (a) lower grain size and (b) reasonably uniform distribution of
harder CNT nanoparticles in the nanocomposite (see Figure 3b). The decrease in hardness can be
primarily attributed to significantly reduced precipitation of intermetallic phase(s) in the matrix
of the nanocomposite compared to monolithic material (see Figure 2b).
2226
Table 3 Results of compressive testing of AZ31, ZK60A and related nanocomposites
350.00 (b)
(a)
300.00 CNT pull-out
AZ31/1.0vol%CNT
250.00
AZ31
stress (MPa)
200.00
150.00
100.00
50.00
0.00 TENSILE
0.000 0.050 0.100 0.150 0.200 0.250
strain
600.00
(c)
500.00
AZ31/1.0vol%CNT
400.00 AZ31
stress (MPa)
300.00
200.00
100.00
0.00
COMPRESSIVE
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350
strain
Figure 5 (a) Representative tensile stress-strain curves of monolithic AZ31 and AZ31/CNT nanocomposite (b)
tensile CNT pull-out in AZ31/CNT nanocomposite (c) representative compressive stress-strain
curves of monolithic AZ31 and AZ31/CNT nanocomposite [35]
2227
350.00 (a)
(b)
300.00
ZK60A/1.0vol%CNT
CNT pull-out
250.00
stress (MPa)
200.00 ZK60
A
150.00
100.00
50.00
TENSILE
0.00
0.000 0.020 0.040 0.060 0.080 0.100 0.120 0.140 0.160 0.180 0.200
strain
600.00
(c)
500.00
400.00
ZK60 ZK60A/1.0vol%CNT
A
stress (MPa)
300.00
200.00
100.00
0.00
COMPRESSIVE
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450 0.500
strain
Figure 6 (a) Representative tensile stress-strain curves of monolithic ZK60A and ZK60A /CNT nanocomposite (b)
tensile CNT pull-out in ZK60A /CNT nanocomposite (c) representative compressive stress-strain
curves of monolithic ZK60A and ZK60A /CNT nanocomposite
2228
the 45° angle between the (0 0 0 2) basal plane and the vertical axis as shown in Figure 4 [36,
37]. The tensile/compressive strength increase in AZ31/1.0vol%CNT compared to monolithic
AZ31 can also be attributed to the following factors (pertaining to reinforcement): (a) dislocation
generation due to elastic modulus mismatch and coefficient of thermal expansion mismatch
between the matrix and reinforcement [31, 32, 38, 39], (b) Orowan strengthening mechanism
[36, 38, 39] and (c) load transfer from matrix to reinforcement [31, 38].
In monolithic AZ31, 0.2%TYS was about 1.8 times the 0.2%CYS. This can be attributed
generally to: (a) half the strain rate used (less strain hardening) in compressive testing compared
to tensile testing and (b) {1 0 1 -2} <1 0 1 -1>-type twinning being more easily activated along
the c-axis (of the HCP unit cell in Figure 4) in tension than in compression along the c-axis [40,
41]. With the c-axis of the HCP unit cell oriented 90° to the force axis (extrusion direction) in
this case, {1 0 1 -2} <1 0 1 -1>-type twinning was less easily activated in compression along the
c-axis during tensile testing. In AZ31-1.0vol%CNT nanocomposite, 0.2%TYS was only about
1.3 times the 0.2%CYS. Here, the tensile/compressive yield stress anisotropy was significantly
reduced despite half the strain rate used (less strain hardening) in compressive testing compared
to tensile testing. This can be similarly attributed to {1 0 1 -2} <1 0 1 -1>-type twinning being
activated along the c-axis of the HCP unit cell in Figure 4 with comparatively similar ease in
both tension and compression along the c-axis, based on the 45° angle between the c-axis and the
vertical axis [29, 40, 41].
2229
crystallographic texture compared to monolithic ZK60A. Here, the significant
tensile/compressive yield stress anisotropy can be attributed to compressive shear buckling of
CNT in ZK60A/1.0vol%CNT. The CNT is prone to buckling followed by fracture during
compressive deformation unlike during tensile deformation.
2230
Conclusions
AZ31 Based Material
1. Monolithic AZ31 and AZ31/1.0vol%CNT nanocomposite can be successfully
synthesized using the DMD technique followed by hot extrusion.
2. Compared to monolithic AZ31, tensile and compressive strengths of AZ31/1.0vol%CNT
were enhanced. This was despite AZ31/1.0vol%CNT exhibiting (1 0 -1 1) dominant
texture in the transverse and longitudinal directions (or (0 0 0 2) basal plane at about 45°
angle to force axis), unlike monolithic AZ31.
3. Compared to monolithic AZ31, tensile failure strain and compressive average failure
strain of AZ31/1.0vol%CNT were enhanced and slightly higher, respectively. This can be
commonly attributed to AZ31/1.0vol%CNT: (a) exhibiting (1 0 -1 1) dominant texture in
the transverse and longitudinal directions (or (0 0 0 2) basal plane at about 45° angle to
force axis) and (b) having reasonably uniform distribution of CNT.
4. Compared to monolithic AZ31, AZ31/1.0vol%CNT exhibited significantly high
increment in tensile WOF and high increment in compressive WOF.
Acknowledgements
Authors wish to acknowledge NUS and Temasek Defence Systems Institute (TDSI) for
funding this research (TDSI/09-011/1A and WBS# R265000349).
References
[1] Y. Morisada, H. Fujii, T. Nagaoka and M. Fukusumi, Mater. Sci. Eng. A, Vol 433, 2006, p 50-54
[2] Y. Morisada, H. Fujii, T. Nagaoka and M. Fukusumi, Scripta Mater., Vol 55, 2006, p 1067-1070
[3] Y. Morisada, H. Fujii, T. Nagaoka and M. Fukusumi, Mater. Sci. Eng. A, Vol 419, 2006, p 344-348
[4] R. Lapovok, P.F. Thomson, R. Cottam and Y. Estrin, Mater. Sci. Eng. A, Vol 410-411, 2005, p 390-393
[5] W.J. Kim, M.J. Kim and J.Y. Wang, Mater. Sci. Eng. A, Doi:10.1016/j.msea.2009.08.064, 2009
[6] H. Watanabe, T. Mukai and K. Higashi, Scripta Mater., Vol 40 (No. 4), 1999, p 477-484
[7] T.G. Nieh, A.J. Schwartz and J. Wadsworth, Mater. Sci. Eng. A, Vol 208, 1996, p 30-36
2231
[8] F. Yan, K. Wu, G.L. Wu, B.L. Lee and M. Zhao, Mater. Lett., Vol 57, 2003, p 1992-1996
[9] Y. Feng, X. Zhou, Z. Min and W. Kun, Scripta Mater., Vol 53, 2005, p 361-365
[10] G. Sasaki, W.G. Wang, Y. Hasegawa, Y.B. Choi, N. Fuyama, K. Matsugi and O. Yanagisawa, J. Mater.
Proc. Technol., Vol 187-188, 2007, p 429-432
[11] R.F. Service, Sci., Vol 281, 1998, p 940-942
[12] P.G. Collins and P. Avouris, Scient. Amer., Vol 283 (No. 6), 2000, p 62-69
[13] Y. Min-Feng, B.S. Files, S. Arepalli and R.S. Rouff, Phys. Rev. Lett., Vol 84 (No. 24), 2000, p 5552-5555
[14] W.H. Knechtel, G.S. Dusberg, W.J. Blau, E. Hernandez and A. Rubio, Appl. Phys. Lett., Vol 73 (No. 14),
1998, p 1961-1963
[15] P.G. Collins, M.S. Arnold and P. Avouris, Sci., Vol 292, 2001, p 706-709
[16] H. Stahl, J. Appenzeller, R. Martel, P. Avouris and B. Lengeler, Phys. Rev. Lett., Vol 85 (No. 24), 2000, p
5186-5189
[17] S. Berber, Y.K. Kwon and D. Tomanek, Phys. Rev. Lett., Vol 84 (No. 20), 2000, p 4613-4616
[18] K.K.H. Wong, M. Zinke-Allmang, J.L. Hutter, S. Hrapovic, J.H.T Luong and W. Wan, Carbon, Vol 47,
2009, 2571-2578
[19] P. Chen, X. Wu, J. Lin and K.L. Tan, J. Phys. Chem. B, Vol 103, 1999, p 4559-4561
[20] X. Chen, J. Xia, J. Peng, W. Li and S. Xie, Comp. Sci. Technol., Vol 60, 2000, p 301-306
[21] J.N. Coleman, S. Curran, A.B. Dalton, A.P. Harvey, B. McCarthy, W. Blau and R.C. Barklie, Phys. Rev. B,
Vol 58, 1998, p R7492-R7495
[22] V. Lordi and N. Yao, J. Mater. Res., Vol 15 (No. 12), 2000, p 2770-2779
[23] L.M. Tham, M. Gupta and L. Cheng, Mater. Sci. Technol., Vol 15, 1999, p 1139-1146
[24] S.F. Hassan and M. Gupta, J. Alloys Compd., Vol 419, 2006, p 84-90
[25] M. De Cicco, H. Konishi, G. Cao, H.S. Choi, L.-S. Turng, J.H. Perepezko, S. Kou, R. Lakes and X. Li,
Metall. Mater. Trans. A, Vol 40A, 2009, p 3038-3045
[26] B.Q. Han and D.C. Dunand, Mater. Sci. Eng. A, Vol 277, 2000, p 297-304
[27] N. Eustathopoulos et al., Wettability at High Temperatures (volume 3), Pergamon, 1999
[28] J.D. Gilchrist, Extraction Metallurgy (third edition), Pergamon Press, 1989
[29] M. Paramsothy, S.F. Hassan, N. Srikanth and M. Gupta, Comp. Part A, Vol 40, 2009, p 1490–1500
[30] M. Gupta, M.O. Lai and C.Y. Soo, Mater. Sci. Eng. A, Vol 210, 1996, p 114-122
[31] S.F. Hassan and M. Gupta, J. Mater. Sci., Vol 41, 2006, p 2229-2236
[32] C.S. Goh, J. Wei, L.C. Lee and M. Gupta, Nanotechnol., Vol 17, 2006, p 7-12
[33] S.F. Hassan and M. Gupta, Metall. Mater. Trans. A, Vol 36 (No. 8), 2005, p 2253-2258
[34] M.H. Al-Saleh and U. Sundararaj, Carbon, Doi:10.1016/j.carbon.2008.09.039, 2008
[35] M. Paramsothy, S.F. Hassan, N. Srikanth and M. Gupta, J. Nanosci. Nanotechnol., Vol 10, 2010, p 956-964
[36] D. Hull et al., Introduction to Dislocations (fourth edition), Butterworth-Heinemann, 2002
[37] T. Mukai, M. Yamanoi, H. Watanabe and K. Higashi, Scripta Mater., Vol 45, 2001, p 89-94
[38] Z. Szaraz, Z. Trojanova, M. Cabbibo and E. Evangelista, Mater. Sci. Eng. A, Vol 462, 2007, p 225-229
[39] L.H. Dai, Z. Ling and Y.L. Bai, Comp. Sci. Technol., Vol 61, 2001, p 1057-1063
[40] T. Laser, C. Hartig, M.R. Nurnberg, D. Letzig and R. Bormann, Acta Mater., Vol 56, 2008, p 2791-2798
[41] J. Bohlen, S.B. Yi, J. Swiostek, D. Letzig, H.G. Brokmeier and K.U. Kainer, Scripta Mater., Vol 53, 2005,
p 259-264
[42] S.F. Hassan and M. Gupta, J. Alloys Compd., Vol 429, 2007, p 176-183
[43] R.E. Reed-Hill, Physical Metallurgy Principles (second edition), D Van Nostrand Company, 1964
[44] G.E. Dieter, Mechanical Metallurgy (SI Metric Edition), McGraw-Hill Book Company, 1988
2232