- Email: [email protected]
aluminum bulk nanolaminated composites with two-dimensional alignment of carbon nanotubes
Available online at www.sciencedirect.com
Scripta Materialia 66 (2012) 331–334 www.elsevier.com/locate/scriptamat
Strong and ductile carbon nanotube/aluminum bulk nanolaminated composites with two-dimensional alignment of carbon nanotubes Lin Jiang, Zhiqiang Li,⇑ Genlian Fan, Linlin Cao and Di Zhang⇑ State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, PR China Received 22 September 2011; accepted 18 November 2011 Available online 26 November 2011
In order to combine high tensile strength and ductile behavior, carbon nanotube (CNT)/Al nanolaminated composites with alternating layers of Al (400 nm) and CNTs (50 nm) were fabricated by flake powder metallurgy. Compared with conventional homogeneous nanocomposites composed of the same constituents, the final bulk products with high level ordered nanolaminates exhibited both greatly improved tensile strength of 375 MPa and plasticity of 12%, mainly because they enabled enhanced dislocation storage capability and two-dimensional alignment of CNTs. Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Aluminum; Carbon nanotubes; Powder consolidation; Layered structures; Ductility
With a Young’s modulus higher than 1 1012 Pa and tensile strength about 50 times that of steel, quasione-dimensional carbon nanotubes (CNTs) can still undergo plastic deformation to failure of 20–30% [1–3]. Thus, incorporating CNTs into light metals and alloys, such as Al, was supposed to make a new class of engineering nanocomposites with both high strength and good ductility, which is suitable for the automotive or aerospace industries [4]. However, for lack of effective fabrications, the as-prepared CNT/Al nanocomposites usually failed to fully realize the strengthening efficiency of CNTs and achieve satisfying ductility, which is an obvious drawback to their practical applications [3,5]. Several methods, such as high energy ball milling (HEBM) [6], an in situ approach [7] and nanoscale dispersion (NSD) [8], have been exploited to supply homogeneous CNT/Al composite powders for the powder metallurgy (PM) route. In these approaches, most emphasis has been placed on breaking up agglomeration of CNTs to homogeneously disperse them in three-dimensional (3-D) spherical Al powders [6–8], whereas little attention has been paid to the control of shape and stacking modes of composite powders, so that CNT distribution and composite structure have been left to evolve randomly. Such a homogeneous structure with 3-D randomly distributed CNTs, however, is unfavorable to enable longitudinal properties
⇑ Corresponding
authors. Tel.: +86 21 3420 2584; fax: +86 21 3420 3913; e-mail addresses: [email protected]; [email protected]
of one-dimensional (1-D) CNTs and improve dislocation storage capability, which is the prerequisite for compatible plastic strains of metals. Moreover, structural defects were caused either by the severe mechanical damage to CNTs in HEBM [9] or by the unstable growth of CNTs during in situ reactions. Therefore, the ductility of these as-fabricated nanocomposites was disappointing. For example, HEBM CNT/Al nanocomposite shows strength of 345 MPa but only 4.7% for elongation [6], while the in situ CNT/Al nanocomposite shows strength of 398 MPa but only 2% for elongation [7]. Thus, a strategy to tailor the structure of CNT/Al nanocomposites and thus endowing such materials with both high strength and good ductility is in great demand. Recently, nanolaminate architecture has become a prevailing model to develop a new generation of high strength metal/metal or metal/ceramic bimaterials with considerable ductility [10,11], due to their remarkable ability of energy-absorbing and dislocation storage [11,12]. Meanwhile, there is an increasing effort to control two-dimensional (2-D) distribution of CNTs and thus create a CNT/metal laminate structure. Zhu et al. [13] fabricated CNT/Cu laminated composites by repeated cold rolling of CNT films sandwiched between Cu thin foils and annealing, and Kim et al. [14] obtained a similar laminate microstructure by selective dipcoating of CNTs and Cu. In these attempts, both strength and ductility enhancement were realized by the laminate structure with 2-D distribution of CNTs. Nevertheless, these methods are restricted to the
1359-6462/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2011.11.023
332
L. Jiang et al. / Scripta Materialia 66 (2012) 331–334
fabrication of miniature sized sheets and, more importantly, they can hardly reduce the laminate thickness to submicron or nanoscale. In the current study, CNT/Al bulk nanolaminated composites were fabricated by a flake powder metallurgy (flake PM) route. Through morphology transformation from 3-D sphere to 2-D nanoflake and surface modification with poly(vinyl alcohol) (PVA), the Al nanoflakes can uniformly anchor 1-D CNTs onto their surface. The as-obtained composite powders could be forced to stack in an orderly way into nanolaminates with 2-D alignment of CNTs through powder compacting, followed by hot extrusion. The high level ordering nanolaminate architecture and the structure integrity of CNTs were achieved in the final products. Thus, excellent mechanical performance of the final products was promised. In a typical flake PM, three steps were involved, as shown in Figure 1b. (i) Nanoflake powder preparation. The near-spherical powder (10 lm in diameter, and 99.5 wt.% in purity) with 1 wt.% stearic acid can quickly reach a flake thickness of several hundred nanometers by ball-milling in an attritor at 423 rpm at room temperature for only 1–2 h. (ii) Adsorption of CNTs. Mutil-wall CNTs (30– 50 nm in diameter, 3 lm in length, functionalized with carboxyl groups (–COOH)) were dispersed into water by ultrasonicating for 2 h and the as-prepared Al nanoflakes were surface modified by PVA with 1700–1800 repeat units. The PVA-coated nanoflakes were added into water to make powder slurry, and then the CNT aqueous dispersion was added in drop by drop. The mixed slurry was mechanically stirred until its color changed from black to transparent, then filtered and rinsed with deionized water to get the CNT/Al nanocomposite powders. (iii) Nanoflake powder alignment and consolidation. Before compacting, the CNT/Al nanocomposite powders were heated in flowing Ar atmosphere at 500 °C for 2 h to remove PVA from the nanocomposite powder. Compacting was used to align the nanoflakes into column (U40 mm 30 mm) under 500 MPa pressure. Sintering in flowing Ar atmosphere at 550 °C for 2 h and hot extrusion at 440 °C with an extrusion ratio of 20:1 at a ram speed of 0.5 mm min1 were conducted to consolidate the nanoflakes. For comparison, we also prepared the CNT/Al nanocomposites
by conventional PM, in which slurry blending is first used to disperse CNT in spherical Al powders with a 3-D spherical morphology (Fig. 2a) and then HEBM (initial ball-to-powder weight ratio of 20:1, 300 rpm/6 h) was used to co-mill CNTs and Al powders, as illustrated in Figure 1a. The consolidation parameters of convention PM were the same as that of flake PM. The structure parameters of samples were characterized by field emission scanning electron microscopy (FESEM) using a LEO Supra 55 FESEM and transmission electron microscopy (TEM) in a Philips CM200 microscope operated at 200 kV. To evaluate the tensile strength, specimens were machined from the extruded rods with the tensile axis parallel to the extrusion direction. The gauge length of the specimens was 25 mm, and the diameter was 5 mm. The tensile strength was measured by a universal testing machine at an initial strain rate of 5 104 s1 at room temperature (AUTOGRAPH AG-I 50 KN, Shimadzu Co. Ltd., Japan). The dislocation density of samples was analyzed by X-ray diffraction (XRD, Rigaku D/max-2550/PC) with a Cu Ka radiation source and calculated as a function of the contribution of micro-strain and crystallite size using the following relationships [15]: qe ¼ ð3Ke2 =D2 b2 Þ1=2
ð1Þ
where e is micro-strain, K is the factor related to Gaussian strain distribution, D is crystallite size and b is the Burgers vector. As seen in Figure 2b, the as-prepared nanoflakes have a 2-D planar morphology with an average diameter of 45 lm and thickness of 300–500 nm, thus assuring a large aspect ratio (diameter to thickness), which is helpful for the alignment of flake powders under compacting. After PVA modification, these 2-D nanoflakes can uniformly adsorb CNTs on their surface (Fig. 2c) due to the formation of hydrogen bonding between the – OH group of the PVA membrane and the –COOH group of the functionized CNTs [16]. After compacting, the CNT/Al powder compact exhibits a structure with a strikingly strong nanoflake alignment (Fig. 2d) and 2-D distributed CNTs are confined by the nanoflakes (inset of Fig. 2d). After extrusion, the dense nanocomposites fabricated by flake PM show that all the Al platelets with aligned CNTs are organized with their faces parallel to extrusion direction (Fig. 3a). The extruded multilayer structure with alternating Al (400 nm) and CNT (50 nm) layers can be clearly observed in Figure 3b, while, as obviously seen in Figure 3c and d, CNTs
Figure 1. Fabrication procedures for CNT/Al nanocomposites: (a) conventional PM and (b) flake PM.
L. Jiang et al. / Scripta Materialia 66 (2012) 331–334
333
enabled the improved longitudinal properties of CNTs in matrix. The strengthening efficiency of CNTs (R) can be expressed as: R ¼ ðrc rm Þ=V f rm
Figure 2. FESEM of: (a) spherical Al powders; (b) Al nanoflakes, inset shows the thickness of the nanoflakes; (c) nanoflake with uniformly adsorbed CNT on its surface; (d) CNT/Al nanolaminates after compacting, inset shows the CNTs confined by nanoflakes.
exhibit disorientation and form a 3-D random distribution in the conventional PM nanocomposites. The resulting tensile properties of CNT/Al nanocomposites fabricated by conventional PM and flake PM are shown in Figure 4a. The flake PM CNT/Al nanocomposite shows a plasticity of 12% at a tensile strength of 375 MPa, while the conventional PM nanocomposite exhibits only a plasticity of 6% at a tensile strength of 330 MPa. Therefore, nanocomposites with the high level ordering nanolaminates exhibited greatly improved tensile strength and plasticity compared with nanocomposites composed of the same constituents but randomly distributed CNTs. The increased strength was mainly attributed to the nanolaminate structure, which enhanced the 2-D alignment of CNTs through geometric confinement and thus
ð2Þ
where rc and rm are the tensile strength of nanocomposites and matrix respectively, and Vf is volume fraction of CNTs. Using the generalized shear-lag model, when fibrous reinforcement has perfect alignment in the loading direction, reinforcement has the strengthening efficiency R, which can be expressed as S/2, where S is the aspect ratio of reinforcement, i.e. the ratio between the diameter and length of reinforcements [17]. In the CNT/Al nanocomposite by flake PM, the enhanced alignment of CNTs enables the R of 27, which is close to the S/2 (35 in this study) and much higher than the R (7.5) of conventional PM nanocomposites with 3-D randomly distributed CNTs. Moreover, the other factor responsible for the enhanced strength is the maintaining of structure integrity of CNTs in flake PM nanocomposites. As shown in Figure 3e, the relative intensity ratio of D band to G band (ID/IG) of conventional PM nanocomposite increased to 1.91 while was only 0.7 for raw CNTs, which implies that the amount of defects in the CNTs apparently increased. However, there was no differences in either the magnitude or the shape of the peaks between the raw CNTs and flake PM CNT/Al nanocomposite. This implies that well-maintained integrity of CNTs appeared in flake PM nanocomposites and as a result, the potential of CNTs as reinforcements in such materials is promising. A better appreciation of the unique mechanical properties of the CNT/Al nanolaminated composites
Figure 3. Microstructure of CNT/Al nanocomposites: (a) optical microscopy and (b) TEM of flake PM material; (c) optical microscopy and (d) TEM of conventional PM material; (e) Raman spectrum of raw CNTs, flake PM and conventional PM CNT/Al nanocomposites.
Figure 4. Tensile properties of CNT/Al nanocomposites (a) loading with 1 vol.% CNTs and fabricated by conventional PM and flake PM (inset shows the relevant strengthening efficiencies of CNTs); (b) fabricated by various methods; the data was drawn based on recent reviews (Refs. [4,5]).
334
L. Jiang et al. / Scripta Materialia 66 (2012) 331–334
Table 1. Structure parameters and tensile properties of pure Al and CNT/Al nanocomposites. Ref.
Component
Grain size (nm)
Strength (MPa)
Uniform elongation (%)
Total elongation (%)
Dislocation density (m2)
Process
[18] [18] This study This study
Pure Al Pure Al CNT/Al CNT/Al
180–200 180–200 215 230
334 350 330 375
1.8 <1 4 8
7 <1 6 12
1.33 1014 0.33 1014 5.7 1014 11.5 1014
ARB ARB & annealing Conventional PM Flake PM
can be gained by comparing with those fabricated by other processes. As shown in Figure 4b, most CNT/Al nanocomposites with randomly distributed CNTs have a strength–ductility trade-off; that is, high strength accompanied with low ductility (points inside the purple region). However, the flake PM nanocomposites (a red point outside the purple region) exhibit both high strength and good ductility, indicating the possibility of retaining good ductility in CNT/Al nanolaminated composites. This is mainly due to the nanolaminate structure, which enables the full potential of CNTs to hinder recovery and thus keeps dislocations inside tiny grains. As shown in Table 1, the grain size of conventional PM and flake PM nanocomposite is respectively 215 and 230 nm. Generally, such ultrafine-grained Al has a very low uniform elongation (strain before necking). Huang et al. [18] observed that ultrafine-grained Al obtained by accumulative roll bonding (ARB) has a strength of 334 MPa, but a very low uniform elongation less than 2%, and when such ARB Al was annealed at 150 °C for 30 min, the total elongation decreased markedly to less than 1%, making the material almost brittle. This is mainly because of a result of diminishing strain hardening capacity due to the intrinsic difficulty in keeping dislocations inside the tiny grains after annealing or hot-deformation [11,18]. Our strategy is to make full use of the CNTs that homogeneously distributed within and between metal nanolaminates to initiate, drag and pin dislocations, such that dynamic recovery could be reduced after hot extrusion. As shown in Table 1, the dislocation density of flake PM nanocomposite (11.5 1014 m2) is much higher than that of pure Al (1.33 1014 m2) with similar grain size and about twice that of convention PM nanocomposite, indicating that the nanolaminate composites with 2-D aligned distribution of CNTs have a higher recovery-hindering efficiency than that of composites with 3-D random distribution of CNTs. Significant dislocation storage is required for compatible plastic strains of metal materials, allowing a high strain-hardening rate, which leads to larger uniform strains while maintaining a high level of strength [18,19]. Thus, the increased ductility was achieved in the flake PM CNT/Al nanocomposites. In conclusion, CNT/Al nanocomposites with nanolaminate structure over large-scale dimensions were fabricated by flake PM. The as-fabricated nanolaminates with CNTs aligned between interlayers enabled longitudinal properties of 1-D CNTs and improved dislocation storage capability, which assured both enhanced strength and ductility over the nanocomposites composed of the same constituents but randomly distributed CNTs. Thus, the flake PM and relevant nanolaminate design were proved to be a practical and effective solu-
tion for strong CNT/Al nanocomposites with remarkable ductility. Finally, it is worth mentioning that the flake PM strategy toward creating nanolaminates can be applied to other metallic or ceramic based bimaterials reinforced with a variety of nanofibers or nanosheets such as graphene, while it can still be used as a reference and for the enlightenment on the current advocation of microstructural reinforcement. The authors would like to acknowledge the financial support of the National Basic Research Program of China (973 Program, No. 2012CB619600), the National High-Tech R&D Program (863 Program, No. 2012AA030611), the National Natural Science Foundation (Nos. 51071100, 51131004, 50890174), the International S&T Cooperation Program of China (Nos. 2010DFA52550, 2009DFA52410) and Shanghai Science & Technology Committee (No. 11JC1405500). [1] G.D. Zhan, J.D. Kuntz, J. Wan, A.K. Mukherjee, Nat. Mater. 2 (2003) 38. [2] R. George, K. Kashyap, R. Rahul, S. Yamdagni, Scripta Mater. 53 (2005) 1159. [3] H.J. Choi, G.B. Kwon, G.Y. Lee, D.H. Baea, Scripta Mater. 59 (2008) 360. [4] S.R. Bakshi, D. Lahiri, A. Agarwal, Int. Mater. Rev. 55 (2010) 41. [5] S.R. Bakshi, A. Agarwal, Carbon 49 (2011) 533. [6] A.M.K. Esawi, K. Morsi, A. Sayed, A.A. Gawad, P. Borah, Mater. Sci. Eng. A 508 (2009) 167. [7] C. He, N. Zhao, C. Shi, X. Du, J. Li, H. Li, Q. Cui, Adv. Mater. 19 (2007) 1128. [8] H. Kwon, M. Estili, K. Takagi, T. Miyazaki, A. Kawasaki, Carbon 47 (2009) 570. [9] D. Poirier, R. Gauvin, R.A.L. Drew, Composites: Part A 40 (2009) 1482. [10] N.A. Mara, D. Bhattacharyya, A. Misra, R.G. Hoagland, Scripta Mater. 58 (2008) 874. [11] L. Jiang, Z.Q. Li, G.L. Fan, D. Zhang, Scripta Mater. 65 (2011) 412. [12] H.A. Hassana, J.J. Lewandowskia, Scripta Mater. 61 (2009) 1072. [13] Y.H. Li, W. Housten, Y.M. Zhao, Y.Q. Zhu, Nanotechnology 18 (2007) 205607. [14] T.J. Kang, J.W. Yoon, D.I. Kim, S.S. Kum, Y.H. Huh, J.H. Hahn, et al., Adv. Mater. 19 (2007) 427. [15] S.K. Halder, M. De, S.P. Sengupta, J. Appl. Phys. 48 (1977) 3560. [16] L. Jiang, G.L. Fan, Z.Q. Li, X.Z. Kai, D. Zhang, Z.X. Chen, Carbon 49 (2011) 1965. [17] S.I. Cha, K.T. Kim, S.N. Arshad, C.B. Mo, S.H. Hong, Adv. Mater. 17 (2005) 1377. [18] X.X. Huang, N. Hansen, N.H. Tsuji, Science 312 (2006) 249. [19] C.M. Hu, C.M. Lai, X.H. Du, N.J. Ho, J.C. Huang, Scripta Mater. 59 (2008) 1163.
Scripta Materialia 66 (2012) 331–334 www.elsevier.com/locate/scriptamat
Strong and ductile carbon nanotube/aluminum bulk nanolaminated composites with two-dimensional alignment of carbon nanotubes Lin Jiang, Zhiqiang Li,⇑ Genlian Fan, Linlin Cao and Di Zhang⇑ State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, PR China Received 22 September 2011; accepted 18 November 2011 Available online 26 November 2011
In order to combine high tensile strength and ductile behavior, carbon nanotube (CNT)/Al nanolaminated composites with alternating layers of Al (400 nm) and CNTs (50 nm) were fabricated by flake powder metallurgy. Compared with conventional homogeneous nanocomposites composed of the same constituents, the final bulk products with high level ordered nanolaminates exhibited both greatly improved tensile strength of 375 MPa and plasticity of 12%, mainly because they enabled enhanced dislocation storage capability and two-dimensional alignment of CNTs. Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Aluminum; Carbon nanotubes; Powder consolidation; Layered structures; Ductility
With a Young’s modulus higher than 1 1012 Pa and tensile strength about 50 times that of steel, quasione-dimensional carbon nanotubes (CNTs) can still undergo plastic deformation to failure of 20–30% [1–3]. Thus, incorporating CNTs into light metals and alloys, such as Al, was supposed to make a new class of engineering nanocomposites with both high strength and good ductility, which is suitable for the automotive or aerospace industries [4]. However, for lack of effective fabrications, the as-prepared CNT/Al nanocomposites usually failed to fully realize the strengthening efficiency of CNTs and achieve satisfying ductility, which is an obvious drawback to their practical applications [3,5]. Several methods, such as high energy ball milling (HEBM) [6], an in situ approach [7] and nanoscale dispersion (NSD) [8], have been exploited to supply homogeneous CNT/Al composite powders for the powder metallurgy (PM) route. In these approaches, most emphasis has been placed on breaking up agglomeration of CNTs to homogeneously disperse them in three-dimensional (3-D) spherical Al powders [6–8], whereas little attention has been paid to the control of shape and stacking modes of composite powders, so that CNT distribution and composite structure have been left to evolve randomly. Such a homogeneous structure with 3-D randomly distributed CNTs, however, is unfavorable to enable longitudinal properties
⇑ Corresponding
authors. Tel.: +86 21 3420 2584; fax: +86 21 3420 3913; e-mail addresses: [email protected]; [email protected]
of one-dimensional (1-D) CNTs and improve dislocation storage capability, which is the prerequisite for compatible plastic strains of metals. Moreover, structural defects were caused either by the severe mechanical damage to CNTs in HEBM [9] or by the unstable growth of CNTs during in situ reactions. Therefore, the ductility of these as-fabricated nanocomposites was disappointing. For example, HEBM CNT/Al nanocomposite shows strength of 345 MPa but only 4.7% for elongation [6], while the in situ CNT/Al nanocomposite shows strength of 398 MPa but only 2% for elongation [7]. Thus, a strategy to tailor the structure of CNT/Al nanocomposites and thus endowing such materials with both high strength and good ductility is in great demand. Recently, nanolaminate architecture has become a prevailing model to develop a new generation of high strength metal/metal or metal/ceramic bimaterials with considerable ductility [10,11], due to their remarkable ability of energy-absorbing and dislocation storage [11,12]. Meanwhile, there is an increasing effort to control two-dimensional (2-D) distribution of CNTs and thus create a CNT/metal laminate structure. Zhu et al. [13] fabricated CNT/Cu laminated composites by repeated cold rolling of CNT films sandwiched between Cu thin foils and annealing, and Kim et al. [14] obtained a similar laminate microstructure by selective dipcoating of CNTs and Cu. In these attempts, both strength and ductility enhancement were realized by the laminate structure with 2-D distribution of CNTs. Nevertheless, these methods are restricted to the
1359-6462/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2011.11.023
332
L. Jiang et al. / Scripta Materialia 66 (2012) 331–334
fabrication of miniature sized sheets and, more importantly, they can hardly reduce the laminate thickness to submicron or nanoscale. In the current study, CNT/Al bulk nanolaminated composites were fabricated by a flake powder metallurgy (flake PM) route. Through morphology transformation from 3-D sphere to 2-D nanoflake and surface modification with poly(vinyl alcohol) (PVA), the Al nanoflakes can uniformly anchor 1-D CNTs onto their surface. The as-obtained composite powders could be forced to stack in an orderly way into nanolaminates with 2-D alignment of CNTs through powder compacting, followed by hot extrusion. The high level ordering nanolaminate architecture and the structure integrity of CNTs were achieved in the final products. Thus, excellent mechanical performance of the final products was promised. In a typical flake PM, three steps were involved, as shown in Figure 1b. (i) Nanoflake powder preparation. The near-spherical powder (10 lm in diameter, and 99.5 wt.% in purity) with 1 wt.% stearic acid can quickly reach a flake thickness of several hundred nanometers by ball-milling in an attritor at 423 rpm at room temperature for only 1–2 h. (ii) Adsorption of CNTs. Mutil-wall CNTs (30– 50 nm in diameter, 3 lm in length, functionalized with carboxyl groups (–COOH)) were dispersed into water by ultrasonicating for 2 h and the as-prepared Al nanoflakes were surface modified by PVA with 1700–1800 repeat units. The PVA-coated nanoflakes were added into water to make powder slurry, and then the CNT aqueous dispersion was added in drop by drop. The mixed slurry was mechanically stirred until its color changed from black to transparent, then filtered and rinsed with deionized water to get the CNT/Al nanocomposite powders. (iii) Nanoflake powder alignment and consolidation. Before compacting, the CNT/Al nanocomposite powders were heated in flowing Ar atmosphere at 500 °C for 2 h to remove PVA from the nanocomposite powder. Compacting was used to align the nanoflakes into column (U40 mm 30 mm) under 500 MPa pressure. Sintering in flowing Ar atmosphere at 550 °C for 2 h and hot extrusion at 440 °C with an extrusion ratio of 20:1 at a ram speed of 0.5 mm min1 were conducted to consolidate the nanoflakes. For comparison, we also prepared the CNT/Al nanocomposites
by conventional PM, in which slurry blending is first used to disperse CNT in spherical Al powders with a 3-D spherical morphology (Fig. 2a) and then HEBM (initial ball-to-powder weight ratio of 20:1, 300 rpm/6 h) was used to co-mill CNTs and Al powders, as illustrated in Figure 1a. The consolidation parameters of convention PM were the same as that of flake PM. The structure parameters of samples were characterized by field emission scanning electron microscopy (FESEM) using a LEO Supra 55 FESEM and transmission electron microscopy (TEM) in a Philips CM200 microscope operated at 200 kV. To evaluate the tensile strength, specimens were machined from the extruded rods with the tensile axis parallel to the extrusion direction. The gauge length of the specimens was 25 mm, and the diameter was 5 mm. The tensile strength was measured by a universal testing machine at an initial strain rate of 5 104 s1 at room temperature (AUTOGRAPH AG-I 50 KN, Shimadzu Co. Ltd., Japan). The dislocation density of samples was analyzed by X-ray diffraction (XRD, Rigaku D/max-2550/PC) with a Cu Ka radiation source and calculated as a function of the contribution of micro-strain and crystallite size using the following relationships [15]: qe ¼ ð3Ke2 =D2 b2 Þ1=2
ð1Þ
where e is micro-strain, K is the factor related to Gaussian strain distribution, D is crystallite size and b is the Burgers vector. As seen in Figure 2b, the as-prepared nanoflakes have a 2-D planar morphology with an average diameter of 45 lm and thickness of 300–500 nm, thus assuring a large aspect ratio (diameter to thickness), which is helpful for the alignment of flake powders under compacting. After PVA modification, these 2-D nanoflakes can uniformly adsorb CNTs on their surface (Fig. 2c) due to the formation of hydrogen bonding between the – OH group of the PVA membrane and the –COOH group of the functionized CNTs [16]. After compacting, the CNT/Al powder compact exhibits a structure with a strikingly strong nanoflake alignment (Fig. 2d) and 2-D distributed CNTs are confined by the nanoflakes (inset of Fig. 2d). After extrusion, the dense nanocomposites fabricated by flake PM show that all the Al platelets with aligned CNTs are organized with their faces parallel to extrusion direction (Fig. 3a). The extruded multilayer structure with alternating Al (400 nm) and CNT (50 nm) layers can be clearly observed in Figure 3b, while, as obviously seen in Figure 3c and d, CNTs
Figure 1. Fabrication procedures for CNT/Al nanocomposites: (a) conventional PM and (b) flake PM.
L. Jiang et al. / Scripta Materialia 66 (2012) 331–334
333
enabled the improved longitudinal properties of CNTs in matrix. The strengthening efficiency of CNTs (R) can be expressed as: R ¼ ðrc rm Þ=V f rm
Figure 2. FESEM of: (a) spherical Al powders; (b) Al nanoflakes, inset shows the thickness of the nanoflakes; (c) nanoflake with uniformly adsorbed CNT on its surface; (d) CNT/Al nanolaminates after compacting, inset shows the CNTs confined by nanoflakes.
exhibit disorientation and form a 3-D random distribution in the conventional PM nanocomposites. The resulting tensile properties of CNT/Al nanocomposites fabricated by conventional PM and flake PM are shown in Figure 4a. The flake PM CNT/Al nanocomposite shows a plasticity of 12% at a tensile strength of 375 MPa, while the conventional PM nanocomposite exhibits only a plasticity of 6% at a tensile strength of 330 MPa. Therefore, nanocomposites with the high level ordering nanolaminates exhibited greatly improved tensile strength and plasticity compared with nanocomposites composed of the same constituents but randomly distributed CNTs. The increased strength was mainly attributed to the nanolaminate structure, which enhanced the 2-D alignment of CNTs through geometric confinement and thus
ð2Þ
where rc and rm are the tensile strength of nanocomposites and matrix respectively, and Vf is volume fraction of CNTs. Using the generalized shear-lag model, when fibrous reinforcement has perfect alignment in the loading direction, reinforcement has the strengthening efficiency R, which can be expressed as S/2, where S is the aspect ratio of reinforcement, i.e. the ratio between the diameter and length of reinforcements [17]. In the CNT/Al nanocomposite by flake PM, the enhanced alignment of CNTs enables the R of 27, which is close to the S/2 (35 in this study) and much higher than the R (7.5) of conventional PM nanocomposites with 3-D randomly distributed CNTs. Moreover, the other factor responsible for the enhanced strength is the maintaining of structure integrity of CNTs in flake PM nanocomposites. As shown in Figure 3e, the relative intensity ratio of D band to G band (ID/IG) of conventional PM nanocomposite increased to 1.91 while was only 0.7 for raw CNTs, which implies that the amount of defects in the CNTs apparently increased. However, there was no differences in either the magnitude or the shape of the peaks between the raw CNTs and flake PM CNT/Al nanocomposite. This implies that well-maintained integrity of CNTs appeared in flake PM nanocomposites and as a result, the potential of CNTs as reinforcements in such materials is promising. A better appreciation of the unique mechanical properties of the CNT/Al nanolaminated composites
Figure 3. Microstructure of CNT/Al nanocomposites: (a) optical microscopy and (b) TEM of flake PM material; (c) optical microscopy and (d) TEM of conventional PM material; (e) Raman spectrum of raw CNTs, flake PM and conventional PM CNT/Al nanocomposites.
Figure 4. Tensile properties of CNT/Al nanocomposites (a) loading with 1 vol.% CNTs and fabricated by conventional PM and flake PM (inset shows the relevant strengthening efficiencies of CNTs); (b) fabricated by various methods; the data was drawn based on recent reviews (Refs. [4,5]).
334
L. Jiang et al. / Scripta Materialia 66 (2012) 331–334
Table 1. Structure parameters and tensile properties of pure Al and CNT/Al nanocomposites. Ref.
Component
Grain size (nm)
Strength (MPa)
Uniform elongation (%)
Total elongation (%)
Dislocation density (m2)
Process
[18] [18] This study This study
Pure Al Pure Al CNT/Al CNT/Al
180–200 180–200 215 230
334 350 330 375
1.8 <1 4 8
7 <1 6 12
1.33 1014 0.33 1014 5.7 1014 11.5 1014
ARB ARB & annealing Conventional PM Flake PM
can be gained by comparing with those fabricated by other processes. As shown in Figure 4b, most CNT/Al nanocomposites with randomly distributed CNTs have a strength–ductility trade-off; that is, high strength accompanied with low ductility (points inside the purple region). However, the flake PM nanocomposites (a red point outside the purple region) exhibit both high strength and good ductility, indicating the possibility of retaining good ductility in CNT/Al nanolaminated composites. This is mainly due to the nanolaminate structure, which enables the full potential of CNTs to hinder recovery and thus keeps dislocations inside tiny grains. As shown in Table 1, the grain size of conventional PM and flake PM nanocomposite is respectively 215 and 230 nm. Generally, such ultrafine-grained Al has a very low uniform elongation (strain before necking). Huang et al. [18] observed that ultrafine-grained Al obtained by accumulative roll bonding (ARB) has a strength of 334 MPa, but a very low uniform elongation less than 2%, and when such ARB Al was annealed at 150 °C for 30 min, the total elongation decreased markedly to less than 1%, making the material almost brittle. This is mainly because of a result of diminishing strain hardening capacity due to the intrinsic difficulty in keeping dislocations inside the tiny grains after annealing or hot-deformation [11,18]. Our strategy is to make full use of the CNTs that homogeneously distributed within and between metal nanolaminates to initiate, drag and pin dislocations, such that dynamic recovery could be reduced after hot extrusion. As shown in Table 1, the dislocation density of flake PM nanocomposite (11.5 1014 m2) is much higher than that of pure Al (1.33 1014 m2) with similar grain size and about twice that of convention PM nanocomposite, indicating that the nanolaminate composites with 2-D aligned distribution of CNTs have a higher recovery-hindering efficiency than that of composites with 3-D random distribution of CNTs. Significant dislocation storage is required for compatible plastic strains of metal materials, allowing a high strain-hardening rate, which leads to larger uniform strains while maintaining a high level of strength [18,19]. Thus, the increased ductility was achieved in the flake PM CNT/Al nanocomposites. In conclusion, CNT/Al nanocomposites with nanolaminate structure over large-scale dimensions were fabricated by flake PM. The as-fabricated nanolaminates with CNTs aligned between interlayers enabled longitudinal properties of 1-D CNTs and improved dislocation storage capability, which assured both enhanced strength and ductility over the nanocomposites composed of the same constituents but randomly distributed CNTs. Thus, the flake PM and relevant nanolaminate design were proved to be a practical and effective solu-
tion for strong CNT/Al nanocomposites with remarkable ductility. Finally, it is worth mentioning that the flake PM strategy toward creating nanolaminates can be applied to other metallic or ceramic based bimaterials reinforced with a variety of nanofibers or nanosheets such as graphene, while it can still be used as a reference and for the enlightenment on the current advocation of microstructural reinforcement. The authors would like to acknowledge the financial support of the National Basic Research Program of China (973 Program, No. 2012CB619600), the National High-Tech R&D Program (863 Program, No. 2012AA030611), the National Natural Science Foundation (Nos. 51071100, 51131004, 50890174), the International S&T Cooperation Program of China (Nos. 2010DFA52550, 2009DFA52410) and Shanghai Science & Technology Committee (No. 11JC1405500). [1] G.D. Zhan, J.D. Kuntz, J. Wan, A.K. Mukherjee, Nat. Mater. 2 (2003) 38. [2] R. George, K. Kashyap, R. Rahul, S. Yamdagni, Scripta Mater. 53 (2005) 1159. [3] H.J. Choi, G.B. Kwon, G.Y. Lee, D.H. Baea, Scripta Mater. 59 (2008) 360. [4] S.R. Bakshi, D. Lahiri, A. Agarwal, Int. Mater. Rev. 55 (2010) 41. [5] S.R. Bakshi, A. Agarwal, Carbon 49 (2011) 533. [6] A.M.K. Esawi, K. Morsi, A. Sayed, A.A. Gawad, P. Borah, Mater. Sci. Eng. A 508 (2009) 167. [7] C. He, N. Zhao, C. Shi, X. Du, J. Li, H. Li, Q. Cui, Adv. Mater. 19 (2007) 1128. [8] H. Kwon, M. Estili, K. Takagi, T. Miyazaki, A. Kawasaki, Carbon 47 (2009) 570. [9] D. Poirier, R. Gauvin, R.A.L. Drew, Composites: Part A 40 (2009) 1482. [10] N.A. Mara, D. Bhattacharyya, A. Misra, R.G. Hoagland, Scripta Mater. 58 (2008) 874. [11] L. Jiang, Z.Q. Li, G.L. Fan, D. Zhang, Scripta Mater. 65 (2011) 412. [12] H.A. Hassana, J.J. Lewandowskia, Scripta Mater. 61 (2009) 1072. [13] Y.H. Li, W. Housten, Y.M. Zhao, Y.Q. Zhu, Nanotechnology 18 (2007) 205607. [14] T.J. Kang, J.W. Yoon, D.I. Kim, S.S. Kum, Y.H. Huh, J.H. Hahn, et al., Adv. Mater. 19 (2007) 427. [15] S.K. Halder, M. De, S.P. Sengupta, J. Appl. Phys. 48 (1977) 3560. [16] L. Jiang, G.L. Fan, Z.Q. Li, X.Z. Kai, D. Zhang, Z.X. Chen, Carbon 49 (2011) 1965. [17] S.I. Cha, K.T. Kim, S.N. Arshad, C.B. Mo, S.H. Hong, Adv. Mater. 17 (2005) 1377. [18] X.X. Huang, N. Hansen, N.H. Tsuji, Science 312 (2006) 249. [19] C.M. Hu, C.M. Lai, X.H. Du, N.J. Ho, J.C. Huang, Scripta Mater. 59 (2008) 1163.