- Email: [email protected]
Effect of milling time on dual-nanoparticulate-reinforced aluminum alloy matrix composite materials
Materials Science & Engineering A 590 (2014) 338–345
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Effect of milling time on dual-nanoparticulate-reinforced aluminum alloy matrix composite materials Hansang Kwon a,b,n, Mart Saarna c, Songhak Yoon d, Anke Weidenkaff d, Marc Leparoux b a
Pukyong National University, Department of Materials System Engineering, San 100, Yongdang-dong, Nam-gu, 608-739 Busan, Korea Empa, Swiss Federal Laboratories for Materials Science and Technology, Advanced Materials Processing, Feuerwerkerstrasse. 39, CH-3602 Thun, Switzerland c Tallinn University of Technology, Department of Materials Engineering, Ehitajate tee 5, 19086 Tallinn, Estonia d Empa, Swiss Federal Laboratories for Materials Science and Technology, Solid State Chemistry and Catalysis, CH-8600 Dübendorf, Switzerland b
art ic l e i nf o
a b s t r a c t
Article history: Received 5 May 2013 Received in revised form 13 October 2013 Accepted 15 October 2013 Available online 25 October 2013
Carbon nanotubes (CNT) and nano-silicon carbide (nSiC)-reinforced aluminum (Al)-6061 alloy matrix composite materials were fabricated using high-energy ball milling and hot-pressing processes. The nSiC was used not only as a solid mixing agent to better disperse the CNTs in the Al powder, but also as a mean of inducing fine particle strengthening. The densification behavior of the dual-nanoparticulate-reinforced composites varied with the milling time. The crystallite sizes of Al in composites became significantly smaller when the milling time was increased. Moreover, the high-energy ball milling time significantly affected the microstructure and mechanical properties of the composites. We believe that the dualnanoparticulate-reinforced composites can be used in a variety of applications as industrial component materials with precisely controlled properties. & 2013 Elsevier B.V. All rights reserved.
Keywords: Carbon nanotube (CNT) Nano-silicon carbide (nSiC) Metal-matrix composites (MMCs) Mechanical properties Powder processing
1. Introduction Engineering materials are continually required because they exhibit high performances and possess reliable characteristics and flexibility for multiple applications. Aluminum (Al) matrix composite materials have developed greatly since Duralumin was first fabricated by German metallurgists and then improved upon by American and Japanese scientists to become super and extra super Duralumin. These materials are widely used in many industrial fields, especially in automobile and aviation technology [1]. However, following this sensational materials development, only a few modified Al or Al matrix composite products have been introduced, such as Al cans and Al composite panels. Recently, a new generation of Al matrix composites with highly anticipated materials properties has been introduced by combining Al and carbon nanotube (CNT) materials. Since they were discovered in 1991 [2], CNTs have been in the limelight as one of the strongest candidates for developing highly efficient nextgeneration materials due to their particularly high mechanical and electrical properties, chemical stability, and high thermal conductivity [3,4]. Many researchers are attempting to fabricate
n Corresponding author at: Empa, Swiss Federal Laboratories for Materials Science and Technology, Advanced Materials Processing, Feuerwerkerstrasse. 39, CH-3602 Thun, Switzerland. Tel.: þ 41 58 765 6227; fax: þ41 58 765 6990. E-mail addresses: [email protected], [email protected], [email protected] (H. Kwon).
0921-5093/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2013.10.046
metal matrix composite materials reinforced with CNTs [5–8]. Kuzumaki et al. are pioneers of Al-CNT composite material fabrication, production composites with mechanical properties similar to those of pure Al bulk [9]. Esawi et al. have demonstrated the fabrication of Al-CNT composites by rolling and extrusion methods [10,11]. Liao et al. have produced Al-CNT composites by ultrasonic dispersion and spark plasma sintering processes and achieved an approximately 10% enhancement in tensile strength and Vickers hardness with respect to pure Al [12]. Deng et al. have also attempted to fabricate Al-CNT composite materials by a powder metallurgical route [13]. Bakshi et al. have tried thermal spraying coating of an Al-CNT composite powder to achieve better properties than those of ordinary coated materials [14]. Despite this enthusiastic line of research, CNT-reinforced Al matrix composite materials are still far away from being commercialized due to difficulties associated with the following issues: dispersing the CNTs, developing a suitable processing scheme and controlling the Al-CNT interface [5,15]. Our previous study demonstrated that the Vickers hardness of dual-nanoparticulatereinforced Al matrix composites is eight times higher than that of pure Al bulk. Moreover, we observed that agglomerated CNTs could be well dispersed in an Al powder using nano-SiC (nSiC) as a solid mixing agent by mechanical ball milling and hot-pressing processes [16]. In this study, the effect of mechanical ball milling time on dual-nanoparticulate-reinforced Al6061 alloy matrix composite materials was investigated. CNTs, nSiC, and Al powders were
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
mixed for five different milling times and then hot-pressed. The mechanical properties of the hot-pressed composites were measured using indentation and bending test equipments. The crystallite size and Raman spectra were rationalized in terms of the observed mechanical properties. Furthermore, microstructural analysis was performed to better understand the mechanical behavior of the dual-nanoparticulate-reinforced Al matrix composites as a function of the milling time.
2. Experimental procedure Materials: Multiwalled CNTs (Baytubes C150P, Bayer Material Science, purity 99.5%, mean diameter: 20 nm, length: 10 mm) and gas-atomized Al6061 alloy powder (ECKA Granules, purity 99.5%, particle size below 63 mm) were used as starting materials. The SiC nanoparticles were produced in our laboratory by an inductively coupled plasma (ICP) process and is described in detail elsewhere [17]. The average particle size was between 20 and 30 nm. Producing the composite powders: The powder composition was adjusted to Al6061 powder-1 vol% nSiC, then 6 wt% CNT was added to the mixture. The aluminum alloy powder and the nanoparticulate materials (CNTs and/or SiC) were mixed in a planetary ball mill (Retsch GmbH, PM400) for 30, 60, 120, 180, and 360 min in an argon atmosphere at 360 rpm using Ø10 mm balls, a 10:1 ball-to-powder weight ratio, and 20 wt% heptane as a process control agent. At the end of the process, the tight bowl containing the powder blend was transferred into a glove-box with a controlled inert argon atmosphere where the powder was passivated for almost 1 week. The milled powder was indeed in a highly activated energy state and could have easily oxidized/ burned in contact with air. Consolidation of the composite powders: After passivation, the blend was placed in a high-temperature steel mold. The mold was then heated up in air at the desired pressing temperature for 1.5 h and transferred rapidly (less than 5 s) to a uniaxial press (Walterþ Bai AG, Switzerland). A pressure of 400 kN was then applied onto the composite powder for 4 s. The samples measured 30 mm in diameter and approximately 5 mm in thickness. Characterizations: The density of the composites was measured by Archimedes's method according to ISO 3369:1975. The macro Vickers hardness of the composites were measured according to EN ISO 6507-1 with loads of 20 kg, applied for 15 s (220, GNEHM Härteprüfer AG and Paar MTH4 microhardness tester). At least five measurements were performed per sample. A four-point bending test was carried out using a WalterþBai 150 kN servo-hydraulic
339
test machine and DionSTAT software using a loading speed of 1 mm/min. The sample dimensions were approximately 3 5 22 mm3. The composites' microstructure was observed by optical microscopy (Zeiss Axioplan light microscope), high-resolution cold field emission scanning electron microscopy (Hitachi, HRCFE-SEM S-4800), and high-resolution transmission electron microscopy (Jeol, HR TEM JEM-2200FS). The XRD patterns were measured using an X'Pert Pro diffractometer (PANAlytical) with Cu-Kα radiation (λ¼1.54056 Å, 45 kV and 40 mA) in the 2θ range of 20–801 using a linear detector (X'Celerator). A step size of 0.01671 and a scan rate of 0.051/s were used. The crystallite size was calculated by the Scherrer equation [18]. Raman spectroscopy was performed on the bulks using a red He–Ne ion laser with a wavelength of 633 nm (Leica microsystems).
3. Results and discussion The gas-atomized raw Al6061 alloy particles were highly spherical in shape and exhibited a wide distribution of sizes, as shown in Fig. 1a and b. The grains of the Al6061 alloy particles also showed a wide distribution of sizes ranging from 1 to 6 mm (Fig. 1c). The high-aspect-ratio CNTs were highly agglomerated, and many opened single CNT's tips were observed as shown in Fig. 1d and e. Parts of the CNT surfaces were covered with amorphous impurities. Fig. 1f shows the typical morphology of the nSiC particles fabricated by our own designed and modified inductively coupled plasma equipment [17]. The large size difference between the matrix particles and the reinforcement particulate materials make their mixing very challenging. However, it is very important to produce homogeneously well-mixed composite powders to create high-performance materials. Thus, high-energy ball milling was performed to disperse the CNTs in the Al6061 alloy powders together with the nSiC particles. The nSiC particles were used as a solid dispersion agent for the CNTs in the Al powders and play an important, synergistic role in enhancing the mechanical properties of such composites by, for example, inducing dispersion and fine particle effects [16]. Fig. 2 shows SEM micrographs of the dual-nanoparticulate-reinforced Al alloy composite powder as a function of the milling time. After only 30 min ball milling, several CNT and nSiC clusters were easily identified in the composite powders (Fig. 2a and b). The powder that was ball milled for 60 min showed many flake-like particles, but many CNT and nSiC clusters were still observed in the Al6061 alloy particles (Fig. 2c). These two powders showed a similar mean size distribution even though different ratios of particle shapes were observed.
Fig. 1. Raw (a–c) Al 6061 alloy particle, (d and e) CNTs, and (f) nano-SiC particles.
340
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
Fig. 2. SEM micrographs of dual-nanoparticulate-reinforced Al 6061 alloy composite powders after various ball milling times. (a and b) 30, (c) 60, (d) 120, (e) 180, and (f) 360 min.
Table 1 Some properties of the Al 6061–1 vol%SiC–6 wt% CNT composites as a function of the milling time. Milling time (min)
30 60 120 180 360
Density (g/cm3)
(%)7 5
2.581 2.602 2.638 2.723 2.765
96.9 97.7 99.0 102.2 103.2
Vickers hardness (HV)
Bending s trength (MPa)
Indentation modulus (KN/mm2)
Crystallite size (nm)7 10%
ID/IG 7 0.5
63.6 7 2.1 86.8 7 1.9 190.0 7 7.3 257.8 7 12 333.77 1.9
1177 4.7 1147 19.3 230 7 11.0 2647 8.7 286 7 22.3
46.9 46.9 70.2 83.4 99.7
135.7 129.4 55.4 46.5 46.2
1.25 1.12 0.98 1.09 1.01
The mean size of the Al particles in the 120-min-milled powder was smaller than the sizes of the Al particles observed in the powders that were milled for 30 and 60 min (Fig. 2d). It was difficult to find any CNTs in the composite powder. Overall, the Al particle size distribution tended to decrease as the milling time increased (Fig. 2d–f). Nevertheless, the 180-min-milled powder showed a finer size distribution compared to the 360-min-milled powder. Probably, the 360-min-milled powder agglomerated again during milling due to the production of very fine particles and, in part, the cold-welding effect. Furthermore, it was very difficult to detect CNTs in the 180-min-milled powder. The CNTs are broken down or implanted and trapped in the Al6061 alloy grains. Nevertheless, it was possible to produce dualnanoparticulate-reinforced Al6061 alloy composite powders by high-energy ball milling for various milling times. The hot-pressed dual-nanoparticulate-reinforced Al6061 alloy matrix composites showed a relative density that varied from 97% to 103% with milling time, as indicated in Table 1. These high densities are due to the large surface area of the fine particles, which induced high reactivity, i.e., particles with small mean sizes are easily densified as bigger particles under the same compaction conditions. The calculated crystallite size (Table 1) is in good agreement with the observed particle size (Fig. 2). Furthermore, the hot-pressed composites made with the 180-, 360-min-milled powders showed relative a density higher than 100%. This may be explained by a partial oxidation of aluminum (density of Al2O3: 3.9–4.1 g/cm 3) and also by the presence of some steel impurities coming from the jar and the balls. Nevertheless, none of these species could be identified by XRD measurements. However, the dual-nanoparticulate-reinforced Al6061 composite powders produced by high-energy ball milling were well densified, showing a lower porosity of 0–3%, by a hot-pressing method.
Fig. 3 shows an SEM micrograph of the longitudinal cross section of the dual-nanoparticulate-reinforced Al6061 matrix composites as a function of the milling time after chemical etching. Two morphologies, a circular and an elongated microstructure, were observed in the 30-min-milled and hot-pressed composite (Fig. 3a). These circular and elongated microstructures are induced by the spherical and flaky particles in the milled composite powders. The CNT and nSiC mixture was clearly observed in the boundary layer observed between the Al grains and the chemical composition of the boundary region was also confirmed by EDAX analysis (Fig. 3b). This layer is smaller than 500 nm on average (Fig. 3b and c). Most of the grooves at the grain boundaries were however created during the etching process as they were not observed before. A more pronounced elongated microstructure of the perpendicular to pressing direction was observed in the case of the 60-min-milled and hot-pressed composite, as shown in Fig. 3d–f. The 60 min milled powder is indeed relatively more flaky than the 30 min milled powder. The composites that were milled for 120, 180, and 360 min and subsequently hot-pressed (Fig. 3g, j, and m) showed similar microstructure (difficult recognized continuous boundary in low magnification compared to 30 and 60 min). A morphology characterized by many disconnected boundaries was observed at the same magnification (Fig. 3h, k, and n). In particular, very fine Al grains were observed in the 360-min-milled composite sample (Fig. 3o). These fine Al grains were produced during the ballmilling process and revealed by etching. However, the microstructure of the composites is significantly attributed to the original shape of the composite powder particles. The phases of the composites were determined using X-ray diffraction (XRD). Whatever the milling time, the XRD patterns of the composites were similar as shown in Fig. 4. The Al and SiC
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
341
Fig. 3. SEM micrographs of the perpendicular to pressing direction of dual-nanoparticulate-reinforced Al 6061 alloy hot-pressed composites after various ball milling times. (a and b) 30, (c) 60, (d) 120, (e) 180, and (f) 360 min.
peaks were clearly observed in all the composites. Moreover, some aluminum carbide (Al4C3) was also detected in the samples, except in case of the hot-pressed composite based on 30 min milled powders. In general, graphene (CNT surface) is very stable against chemical reactions, but disordered CNTs, such as tubes with opened tips, and CNTs featuring amorphous impurities easily react with Al [19]. The detected Al4C3 was produced during the hotpressing process because the mechanically ball-milled powders are highly unstable in terms of energy, making the powders more reactive. In other words, a processing temperature of 500 1C delivers a sufficient driving force for the reaction between Al and C derived from amorphous carbon and defects in the CNTs. The 30-min-milled powder seems not so unstable, at least in terms of activation energy, compared to the other powders. The processing temperature of 500 1C did not allow surmounting of the energy barrier for the reaction between carbon and aluminum, explaining why no Al4C3 was detected in the 30-min-milled and hot-pressed composite.
The main Al peaks in the composites are slightly shifted to low 2θ values compared to the positions given in JCPDS card no. 89-4037 and these peaks become broader when the milling time is increased, as shown in Fig. 4b. This phenomenon is due to a stress-induced change in the lattice constant of Al in the composites. The stress level was observed to increase significantly with milling time, which indicates that the stress was created mainly during the ball-milling process. Furthermore, the shapes of the peaks suddenly changed for the composites that were milled for longer than 60 min. This suggests an evolution mainly by changes on particle shape and size (Fig. 2c and d). The crystallite size of the Al in the composites was calculated from the XRD patterns using the Scherrer equation (ignoring then any strain effect), as shown in Eq. (1). d ¼ 0:9
Kλ B cos θ
ð1Þ
where d is the crystallite size and λ, θ, and B represent the X-ray wavelength, the Bragg scattering angle, and the full width at half
342
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
Fig. 4. XRD diffraction of (a) the hot-pressed dual-nanoparticulate-reinforced Al 6061 alloy composites ball milled for various times. (b) Highly magnified main peaks of the composites shown in (a).
Fig. 5. Crystallite size and Al4C3 area calculated from the XRD diffraction patterns of the composites after ball milling times of 30, 60, 120, 180, and 360 min.
maximum (FWHM), respectively. K is a constant that depends on the crystallite shape (0.9). The values of FWHM measured from the XRD patterns (Fig. 4b) were calibrated using a standard sample (CeO2 NIST SRM-674b). The crystallite size of the Al in the composites decreased with increasing milling time (Table 1 and Fig. 5). A sudden change in crystallite size was observed between the composites made with powders milled for 60 min and 120 min. This is similar to the tendency observed as in the peak shape described above. Moreover, the crystallite size of the 360min-milled composite is three times as fine as that of the 30-minmilled composite. We performed a quantitative analysis of the Al4C3 formed in the composites based on the XRD diffraction patterns obtained. It is very difficult to precisely define the quantity of Al4C3 formed from the XRD results, but some tendencies (i.e., area of Al4C3) can be discerned, at least under the same conditions. Therefore, the area of the 2θ peak at 55.81 was calculated for the different composites and are plotted in Fig. 5 as a function of the powder milling time. These areas are differing by a factor of 10 approximately. An inverse correlation is observed between the Al crystallite size and the relative quantity of aluminum carbide. The Al4C3 content indeed increase more or less linearly until a powder milling time of 120 min while the crystallite size decreases. Then
a pronounced slope increase is seen for the carbide content while the crystallite size of Al drastically decrease after about 180 min, the crystallite size reaches a plateau while the carbide content further slightly increase. The particle size seems then to play a more important role in the formation of Al4C3 than the particle shape. The particle shape completely changed from a spherical and flaky morphology for the 30- and 60-min-milled based samples to a solely spherical morphology for the samples milled for over 120 min. This means that the particle shape did not significantly provide any driving force for the formation of Al4C3. However, the large surface area of the Al particles provided a higher reaction potential. The formation of aluminum carbide in Al-metal matrix composites reinforced with CNTs is carefully handled as on one hand Al4C3 is well known as a brittle and hygroscopic compound [20], but on the other hand if confined at the interface, it can efficiently transfer stress from the matrix to the CNT reinforcement material [5,15,21]. Regarding the potential role of aluminum carbide and according to the above results, 120 min milling time seems to be a good compromise solution for producing high performance Al-based nanocomposites with the composition Al6061–1vol% nSiC–6wt% CNT. The composites made with the powders milled for different times were also analyzed using Raman. The resulting spectra are presented in Fig. 6 showing the characteristic D (defect) and G (graphitic) bands of carbon. The relative peak intensities of these two bands decreased with increasing milling time. Therefore, the fraction of trapped and implanted CNTs in the soft Al matrix was relatively higher than the fraction of CNTs broken down during the ball-milling process due to high impact energy. This phenomenon caused fewer CNTs to be exposed to the laser beam during Raman spectroscopy, resulting in low peak intensities. In general, the ratio between the relative intensities of these two bands ID/IG is indicative of the CNT quality [22]. The ID/IG ratio of the composites was slightly reduced for long milling times (Table 1). However this decrease may not be significant as the very low intensities measured especially for the longer milling times may induce large errors too. Furthermore, the 60- and 120-min-milled samples also featured an inflection area, as also observed with XRD and SEM analyses. However, the destruction of the CNTs in the dualnanoparticulate-reinforced Al6061 alloy matrix composites submitted to ball milling and hot-pressing did not seriously affect the composites' material properties. Finally, no shifting or shape change of the typical carbon peaks occurred even though aspect ratio of the CNT changed and some Al4C3 formed, indicating that
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
the CNTs in the composites remained approximately the same condition. The composites were characterized mechanically with a fourpoint bending test; the Al6061 alloy without any reinforcement but hot pressed under the same conditions as the composites were also tested for comparison. The results of the bending tests are summarized also in Table 1. The bending strength of the composites increased with the milling time and reached a maximum value of approximately 300 MPa. Fig. 7 shows a digital photograph of the Al6061 alloy bulk before and after the bending test. The different composites did not show any cracks on their surface after the bending tests, but they were completely destroyed, whatever the powder milling time (Fig. 7b). The bending strength of the
Fig. 6. Raman spectra of the dual-nanoparticulate-reinforced Al 6061 matrix composite materials after ball milling times of 30, 60, 120, 180, and 360 min.
343
pure Al6061 alloy was impossible to measure under the test conditions due to the material's high plasticity; it deformed too much without breaking. Nevertheless, its yield strength was measured around 60 MPa, being the lowest strength value. After bending tests, the fracture surfaces of the samples were characterized by SEM observations. As for the composite powders (Fig. 2), many CNTs were found relatively easily in the boundary layer between the particles in the 30-min-milled composite sample (Fig. 8a). Most of the cracks propagated along the lamellar structure of the composite. The composites made out of short time milled powders (30 and 60 min) significantly showed a smooth fracture surface with much less dimpling than that observed in ductile materials such as Al; however, the composites made with powders milled for longer times showed clear dimples, as shown in Fig. 8c–e. This means that a longer milling time leads to a better bonding surface. In particular, the height of the elongated dimples was different in the samples milled for 120, 180, and 360 min. The length of the elongated dimples is directly related to the elongation or toughness of the composites. In general, formation of the proper Al4C3 also can offer better chemical link between the reinforcement and the matrix, resulting in leading to strong interface bonding [5,15,21]. According to the previously discussed XRD results, the produced composites with milling time 60 min are containing some of the small amount of Al4C3 and this formed carbide supported enhanced interface bonding. However, the longest elongated dimple was observed in the 180-min-milled composite and was about a few micrometers (Fig. 8d). The Vickers hardness and the bending strength were plotted together in Fig. 9 and showed similar enhancement as a function of the milling time. According to the Hall–Petch relationship, the strength of a material increases with decreasing grain size, and this relationship is valid down to a grain size of 10 nm. Below this grain size, the strength decreases due to grain sliding [23]. The mechanical properties are then in agreement with this principle as it has been showed before based on XRD analyses that the Al grain
Fig. 7. Digital photographs of a composite sample before and after bending test.
344
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
Fig. 8. SEM micrographs of fractured surfaces of the dual-nanoparticulate-reinforced Al 6061 alloy hot-pressed composites after ball milling times of (a and b) 30, (c) 60, (d) 120, (e) 180, and (f) 360 min.
Fig. 9. Vickers hardness and bending strength of the dual-nanoparticulate-reinforced composite materials as a function of the milling time.
size of the composites decreased with increasing milling time (Table 1). However, the composite performances were not only enhanced by the addition of nSiC and CNT but also by grain refinement. The Vickers hardness of the pure Al6061 alloy bulk used for comparison was 45 HV. The maximum hardness value of 330 HV was achieved for the 360-min-milled sample. This value is almost six times higher than that of the pure Al6061 alloy bulk. Additional samples reinforced with only 1 vol% nSiC or 6 wt% CNT were also elaborated under same conditions for comparison. The hardness values were however reaching a plateau after 360 min up to a maximum milling time of 1200 min. Vickers hardness values of approximately 100 HV and 300 HV were measured for the Al6061–1 vol% nSiC and Al6061-6 wt% CNT composites, respectively. Thus, the hardness of the dual-nanoparticulate-reinforced composite was enhanced by 300% compared to that of the composite with only nSiC added and by more than 10% compared to that of the composite with only CNT added. In other words, the hardness value of 300 HV measured for the 360-min-milled dualnanoparticulate-reinforced composite (approximately 334 HV in Table 1) was mainly due to the above-mentioned grain refinement and reinforcement effects. The remaining 34 HV (10%) can be
assumed by reasonable errors. Because the low-aspect-ratio (almost spherical in shape) nSiC (nanosized) particles easily infiltrate the agglomerated high-aspect-ratio CNTs during the ball-milling process and the nSiC particles function as nano-balls, the CNTs are well dispersed in the Al powders. This synergistic effect of the two nano-materials provided a sub-enhancement of approximately 34 HV of the total 334 HV enhancement observed for the 360-min-milled composite. The indentation modulus (almost equivalent to the elastic modulus) of each composite increased with the milling time. A sudden increase in the indentation modulus was observed between the 60- and 120-min-milled composites, as indicated in Table 1. This tendency was already observed and discussed within the context of the microstructural, XRD and Raman spectroscopy analyses. The indentation moduli of the 30- and 60-min-milled composites were lower than the modulus of the Al6061 alloy bulk (100% densified), mainly due to the porosity (Table 1). However, overall, a milling time of 120– 180 min is considered to be optimal for the employed dualnanoparticulate-reinforced Al6061 alloy matrix composite.
4. Conclusions Dual-nanoparticulate-reinforced Al 6061 alloy matrix composites made out of blends milled for various times were successfully fabricated by a high-energy ball milling and simple hot-pressing process. The shape of the Al particles in the composite powders varied with the milling time, varying from spherical and flaky to flaky only or finely spherical. This tendency was followed by the microstructure and fracture surface of the composites. The crystallite size of the Al particles decreased with increasing milling time. A small amount of Al4C3 was detected in the composites, except for the 30-min-milled composite, and was observed to increase with increasing milling time. We were able to greatly enhance the Vickers hardness of the dual-nanoparticulate-reinforced composites relative to the hardness of the composites featuring only nSiC or CNTs. The Vickers hardness and bending strength of the dualnanoparticulate-reinforced composites were enhanced with increasing milling time and showed maximum values that were approximately five times higher than those of pure Al 6061 alloy bulk. However, most of the results showed a high-enhancement region from 60 to 120 min of milling time, which occurred mainly due to particle-shape and -size effects. The high-energy ball
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
milling time significantly affected the microstructure and mechanical properties of the composites, which suggests that the characteristic properties of the composites can be controlled by adjusting the milling time. We believe that dual-nanoparticulatereinforced composites have many potential applications as industrial component materials with precisely controlled properties.
Acknowledgments The authors would like to acknowledge Bernhard von Gunten at EMPA for his technical support. References 〈http://en.wikipedia.org/wiki/Duralumin〉. S. Iijima, Nature 354 (1991) 56–58. A.P. Ramirez, Bell Labs Tech. J. 10 (2005) 171–185. J. Tersoff, R.S. Ruoff, Phys. Rev. Lett. 73 (1994) 676–679. H. Kwon, D.H. Park, J.F. Silvain, A. Kawasaki, Compos. Sci. Technol. 70 (2010) 546–550. [6] S. Cho, K. Kikuchi, A. Kawasaki, Acta Mater. 60 (2012) 726–736. [1] [2] [3] [4] [5]
345
[7] H. Fukuda, K. Kondoh, J. Umeda, B. Fugetsu, Compos. Sci. Technol. 71 (2011) 705–709. [8] K.T. Kim, S.I. Cha, T. Gemming, J. Eckert, S.H. Hong, Small 4 (200) (1936–1940). [9] T. Kuzumaki, K. Miyazawa, H. Ichinose, K. Ito, J. Mater. Res. 13 (1998) 2445–2449. [10] A.M.K. Esawi, M.A. El Borady, Compos. Sci. Technol. 68 (2008) 486–492. [11] A.M.K. Esawi, K. Morsi, A. Sayed, M. Taher, S. Lanka, Compos. Part A Appl. Sci. Manuf. 42 (2011) 234–243. [12] J.Z. Liao, M.J. Tan, I. Sridhar, Mater. Des. 31 (2010) 96–100. [13] C.F. Deng, D.Z. Wang, X.X. Zhang, A.B. Li, Mater. Sci. Eng. A 444 (2007) 138–145. [14] S.R. Bakshi, A.K. Keshri, A. Agarwal, Mater. Sci. Eng. A 528 (2011) 3375–3384. [15] H. Kwon, M. Estili, K. Takagi, T. Miyazaki, A. Kawasaki, Carbon 47 (2009) 570–577. [16] H. Kwon, S. Cho, M. Leparoux, A. Kawasaki, Nanotechnology 23 (2012) 225704. [17] Y. Leconte, M. Leparoux, X. Portier, N. Herlin-Boime, Plasma Chem. Plasma Process. 28 (2008) 233–248. [18] A.W. Burton, K. Ong, T. Rea, I.Y. Chan, Microporous Mesoporous Mater. 117 (2009) 75–90. [19] L. Ci, Z. Ryu, N.Y. Jin-Phillipp, M. Rühle, Acta Mater. 54 (2006) 5367–5375. [20] T. Etter, P. Schulz, M. Weber, J. Metz, M. Wimmler, J.F. Loffler, P.J. Uggowitzer, Mater. Sci. Eng. A 448 (2007) 1–6. [21] S. Cho, k. Kikuchi, A. Kawasaki, H. Kwon, Y. Kim, Nanotechnology 23 (2012) 315705. [22] C. Thomsen, S. Reich, Top. Appl. Phys. 108 (2007) 115–232. [23] T. Shanmugasundaram, M. Heilmaier, B.S. Murty, V.S. Sarma, Mater. Sci. Eng. A 527 (2010) 7821–7825.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Effect of milling time on dual-nanoparticulate-reinforced aluminum alloy matrix composite materials Hansang Kwon a,b,n, Mart Saarna c, Songhak Yoon d, Anke Weidenkaff d, Marc Leparoux b a
Pukyong National University, Department of Materials System Engineering, San 100, Yongdang-dong, Nam-gu, 608-739 Busan, Korea Empa, Swiss Federal Laboratories for Materials Science and Technology, Advanced Materials Processing, Feuerwerkerstrasse. 39, CH-3602 Thun, Switzerland c Tallinn University of Technology, Department of Materials Engineering, Ehitajate tee 5, 19086 Tallinn, Estonia d Empa, Swiss Federal Laboratories for Materials Science and Technology, Solid State Chemistry and Catalysis, CH-8600 Dübendorf, Switzerland b
art ic l e i nf o
a b s t r a c t
Article history: Received 5 May 2013 Received in revised form 13 October 2013 Accepted 15 October 2013 Available online 25 October 2013
Carbon nanotubes (CNT) and nano-silicon carbide (nSiC)-reinforced aluminum (Al)-6061 alloy matrix composite materials were fabricated using high-energy ball milling and hot-pressing processes. The nSiC was used not only as a solid mixing agent to better disperse the CNTs in the Al powder, but also as a mean of inducing fine particle strengthening. The densification behavior of the dual-nanoparticulate-reinforced composites varied with the milling time. The crystallite sizes of Al in composites became significantly smaller when the milling time was increased. Moreover, the high-energy ball milling time significantly affected the microstructure and mechanical properties of the composites. We believe that the dualnanoparticulate-reinforced composites can be used in a variety of applications as industrial component materials with precisely controlled properties. & 2013 Elsevier B.V. All rights reserved.
Keywords: Carbon nanotube (CNT) Nano-silicon carbide (nSiC) Metal-matrix composites (MMCs) Mechanical properties Powder processing
1. Introduction Engineering materials are continually required because they exhibit high performances and possess reliable characteristics and flexibility for multiple applications. Aluminum (Al) matrix composite materials have developed greatly since Duralumin was first fabricated by German metallurgists and then improved upon by American and Japanese scientists to become super and extra super Duralumin. These materials are widely used in many industrial fields, especially in automobile and aviation technology [1]. However, following this sensational materials development, only a few modified Al or Al matrix composite products have been introduced, such as Al cans and Al composite panels. Recently, a new generation of Al matrix composites with highly anticipated materials properties has been introduced by combining Al and carbon nanotube (CNT) materials. Since they were discovered in 1991 [2], CNTs have been in the limelight as one of the strongest candidates for developing highly efficient nextgeneration materials due to their particularly high mechanical and electrical properties, chemical stability, and high thermal conductivity [3,4]. Many researchers are attempting to fabricate
n Corresponding author at: Empa, Swiss Federal Laboratories for Materials Science and Technology, Advanced Materials Processing, Feuerwerkerstrasse. 39, CH-3602 Thun, Switzerland. Tel.: þ 41 58 765 6227; fax: þ41 58 765 6990. E-mail addresses: [email protected], [email protected], [email protected] (H. Kwon).
0921-5093/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msea.2013.10.046
metal matrix composite materials reinforced with CNTs [5–8]. Kuzumaki et al. are pioneers of Al-CNT composite material fabrication, production composites with mechanical properties similar to those of pure Al bulk [9]. Esawi et al. have demonstrated the fabrication of Al-CNT composites by rolling and extrusion methods [10,11]. Liao et al. have produced Al-CNT composites by ultrasonic dispersion and spark plasma sintering processes and achieved an approximately 10% enhancement in tensile strength and Vickers hardness with respect to pure Al [12]. Deng et al. have also attempted to fabricate Al-CNT composite materials by a powder metallurgical route [13]. Bakshi et al. have tried thermal spraying coating of an Al-CNT composite powder to achieve better properties than those of ordinary coated materials [14]. Despite this enthusiastic line of research, CNT-reinforced Al matrix composite materials are still far away from being commercialized due to difficulties associated with the following issues: dispersing the CNTs, developing a suitable processing scheme and controlling the Al-CNT interface [5,15]. Our previous study demonstrated that the Vickers hardness of dual-nanoparticulatereinforced Al matrix composites is eight times higher than that of pure Al bulk. Moreover, we observed that agglomerated CNTs could be well dispersed in an Al powder using nano-SiC (nSiC) as a solid mixing agent by mechanical ball milling and hot-pressing processes [16]. In this study, the effect of mechanical ball milling time on dual-nanoparticulate-reinforced Al6061 alloy matrix composite materials was investigated. CNTs, nSiC, and Al powders were
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
mixed for five different milling times and then hot-pressed. The mechanical properties of the hot-pressed composites were measured using indentation and bending test equipments. The crystallite size and Raman spectra were rationalized in terms of the observed mechanical properties. Furthermore, microstructural analysis was performed to better understand the mechanical behavior of the dual-nanoparticulate-reinforced Al matrix composites as a function of the milling time.
2. Experimental procedure Materials: Multiwalled CNTs (Baytubes C150P, Bayer Material Science, purity 99.5%, mean diameter: 20 nm, length: 10 mm) and gas-atomized Al6061 alloy powder (ECKA Granules, purity 99.5%, particle size below 63 mm) were used as starting materials. The SiC nanoparticles were produced in our laboratory by an inductively coupled plasma (ICP) process and is described in detail elsewhere [17]. The average particle size was between 20 and 30 nm. Producing the composite powders: The powder composition was adjusted to Al6061 powder-1 vol% nSiC, then 6 wt% CNT was added to the mixture. The aluminum alloy powder and the nanoparticulate materials (CNTs and/or SiC) were mixed in a planetary ball mill (Retsch GmbH, PM400) for 30, 60, 120, 180, and 360 min in an argon atmosphere at 360 rpm using Ø10 mm balls, a 10:1 ball-to-powder weight ratio, and 20 wt% heptane as a process control agent. At the end of the process, the tight bowl containing the powder blend was transferred into a glove-box with a controlled inert argon atmosphere where the powder was passivated for almost 1 week. The milled powder was indeed in a highly activated energy state and could have easily oxidized/ burned in contact with air. Consolidation of the composite powders: After passivation, the blend was placed in a high-temperature steel mold. The mold was then heated up in air at the desired pressing temperature for 1.5 h and transferred rapidly (less than 5 s) to a uniaxial press (Walterþ Bai AG, Switzerland). A pressure of 400 kN was then applied onto the composite powder for 4 s. The samples measured 30 mm in diameter and approximately 5 mm in thickness. Characterizations: The density of the composites was measured by Archimedes's method according to ISO 3369:1975. The macro Vickers hardness of the composites were measured according to EN ISO 6507-1 with loads of 20 kg, applied for 15 s (220, GNEHM Härteprüfer AG and Paar MTH4 microhardness tester). At least five measurements were performed per sample. A four-point bending test was carried out using a WalterþBai 150 kN servo-hydraulic
339
test machine and DionSTAT software using a loading speed of 1 mm/min. The sample dimensions were approximately 3 5 22 mm3. The composites' microstructure was observed by optical microscopy (Zeiss Axioplan light microscope), high-resolution cold field emission scanning electron microscopy (Hitachi, HRCFE-SEM S-4800), and high-resolution transmission electron microscopy (Jeol, HR TEM JEM-2200FS). The XRD patterns were measured using an X'Pert Pro diffractometer (PANAlytical) with Cu-Kα radiation (λ¼1.54056 Å, 45 kV and 40 mA) in the 2θ range of 20–801 using a linear detector (X'Celerator). A step size of 0.01671 and a scan rate of 0.051/s were used. The crystallite size was calculated by the Scherrer equation [18]. Raman spectroscopy was performed on the bulks using a red He–Ne ion laser with a wavelength of 633 nm (Leica microsystems).
3. Results and discussion The gas-atomized raw Al6061 alloy particles were highly spherical in shape and exhibited a wide distribution of sizes, as shown in Fig. 1a and b. The grains of the Al6061 alloy particles also showed a wide distribution of sizes ranging from 1 to 6 mm (Fig. 1c). The high-aspect-ratio CNTs were highly agglomerated, and many opened single CNT's tips were observed as shown in Fig. 1d and e. Parts of the CNT surfaces were covered with amorphous impurities. Fig. 1f shows the typical morphology of the nSiC particles fabricated by our own designed and modified inductively coupled plasma equipment [17]. The large size difference between the matrix particles and the reinforcement particulate materials make their mixing very challenging. However, it is very important to produce homogeneously well-mixed composite powders to create high-performance materials. Thus, high-energy ball milling was performed to disperse the CNTs in the Al6061 alloy powders together with the nSiC particles. The nSiC particles were used as a solid dispersion agent for the CNTs in the Al powders and play an important, synergistic role in enhancing the mechanical properties of such composites by, for example, inducing dispersion and fine particle effects [16]. Fig. 2 shows SEM micrographs of the dual-nanoparticulate-reinforced Al alloy composite powder as a function of the milling time. After only 30 min ball milling, several CNT and nSiC clusters were easily identified in the composite powders (Fig. 2a and b). The powder that was ball milled for 60 min showed many flake-like particles, but many CNT and nSiC clusters were still observed in the Al6061 alloy particles (Fig. 2c). These two powders showed a similar mean size distribution even though different ratios of particle shapes were observed.
Fig. 1. Raw (a–c) Al 6061 alloy particle, (d and e) CNTs, and (f) nano-SiC particles.
340
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
Fig. 2. SEM micrographs of dual-nanoparticulate-reinforced Al 6061 alloy composite powders after various ball milling times. (a and b) 30, (c) 60, (d) 120, (e) 180, and (f) 360 min.
Table 1 Some properties of the Al 6061–1 vol%SiC–6 wt% CNT composites as a function of the milling time. Milling time (min)
30 60 120 180 360
Density (g/cm3)
(%)7 5
2.581 2.602 2.638 2.723 2.765
96.9 97.7 99.0 102.2 103.2
Vickers hardness (HV)
Bending s trength (MPa)
Indentation modulus (KN/mm2)
Crystallite size (nm)7 10%
ID/IG 7 0.5
63.6 7 2.1 86.8 7 1.9 190.0 7 7.3 257.8 7 12 333.77 1.9
1177 4.7 1147 19.3 230 7 11.0 2647 8.7 286 7 22.3
46.9 46.9 70.2 83.4 99.7
135.7 129.4 55.4 46.5 46.2
1.25 1.12 0.98 1.09 1.01
The mean size of the Al particles in the 120-min-milled powder was smaller than the sizes of the Al particles observed in the powders that were milled for 30 and 60 min (Fig. 2d). It was difficult to find any CNTs in the composite powder. Overall, the Al particle size distribution tended to decrease as the milling time increased (Fig. 2d–f). Nevertheless, the 180-min-milled powder showed a finer size distribution compared to the 360-min-milled powder. Probably, the 360-min-milled powder agglomerated again during milling due to the production of very fine particles and, in part, the cold-welding effect. Furthermore, it was very difficult to detect CNTs in the 180-min-milled powder. The CNTs are broken down or implanted and trapped in the Al6061 alloy grains. Nevertheless, it was possible to produce dualnanoparticulate-reinforced Al6061 alloy composite powders by high-energy ball milling for various milling times. The hot-pressed dual-nanoparticulate-reinforced Al6061 alloy matrix composites showed a relative density that varied from 97% to 103% with milling time, as indicated in Table 1. These high densities are due to the large surface area of the fine particles, which induced high reactivity, i.e., particles with small mean sizes are easily densified as bigger particles under the same compaction conditions. The calculated crystallite size (Table 1) is in good agreement with the observed particle size (Fig. 2). Furthermore, the hot-pressed composites made with the 180-, 360-min-milled powders showed relative a density higher than 100%. This may be explained by a partial oxidation of aluminum (density of Al2O3: 3.9–4.1 g/cm 3) and also by the presence of some steel impurities coming from the jar and the balls. Nevertheless, none of these species could be identified by XRD measurements. However, the dual-nanoparticulate-reinforced Al6061 composite powders produced by high-energy ball milling were well densified, showing a lower porosity of 0–3%, by a hot-pressing method.
Fig. 3 shows an SEM micrograph of the longitudinal cross section of the dual-nanoparticulate-reinforced Al6061 matrix composites as a function of the milling time after chemical etching. Two morphologies, a circular and an elongated microstructure, were observed in the 30-min-milled and hot-pressed composite (Fig. 3a). These circular and elongated microstructures are induced by the spherical and flaky particles in the milled composite powders. The CNT and nSiC mixture was clearly observed in the boundary layer observed between the Al grains and the chemical composition of the boundary region was also confirmed by EDAX analysis (Fig. 3b). This layer is smaller than 500 nm on average (Fig. 3b and c). Most of the grooves at the grain boundaries were however created during the etching process as they were not observed before. A more pronounced elongated microstructure of the perpendicular to pressing direction was observed in the case of the 60-min-milled and hot-pressed composite, as shown in Fig. 3d–f. The 60 min milled powder is indeed relatively more flaky than the 30 min milled powder. The composites that were milled for 120, 180, and 360 min and subsequently hot-pressed (Fig. 3g, j, and m) showed similar microstructure (difficult recognized continuous boundary in low magnification compared to 30 and 60 min). A morphology characterized by many disconnected boundaries was observed at the same magnification (Fig. 3h, k, and n). In particular, very fine Al grains were observed in the 360-min-milled composite sample (Fig. 3o). These fine Al grains were produced during the ballmilling process and revealed by etching. However, the microstructure of the composites is significantly attributed to the original shape of the composite powder particles. The phases of the composites were determined using X-ray diffraction (XRD). Whatever the milling time, the XRD patterns of the composites were similar as shown in Fig. 4. The Al and SiC
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
341
Fig. 3. SEM micrographs of the perpendicular to pressing direction of dual-nanoparticulate-reinforced Al 6061 alloy hot-pressed composites after various ball milling times. (a and b) 30, (c) 60, (d) 120, (e) 180, and (f) 360 min.
peaks were clearly observed in all the composites. Moreover, some aluminum carbide (Al4C3) was also detected in the samples, except in case of the hot-pressed composite based on 30 min milled powders. In general, graphene (CNT surface) is very stable against chemical reactions, but disordered CNTs, such as tubes with opened tips, and CNTs featuring amorphous impurities easily react with Al [19]. The detected Al4C3 was produced during the hotpressing process because the mechanically ball-milled powders are highly unstable in terms of energy, making the powders more reactive. In other words, a processing temperature of 500 1C delivers a sufficient driving force for the reaction between Al and C derived from amorphous carbon and defects in the CNTs. The 30-min-milled powder seems not so unstable, at least in terms of activation energy, compared to the other powders. The processing temperature of 500 1C did not allow surmounting of the energy barrier for the reaction between carbon and aluminum, explaining why no Al4C3 was detected in the 30-min-milled and hot-pressed composite.
The main Al peaks in the composites are slightly shifted to low 2θ values compared to the positions given in JCPDS card no. 89-4037 and these peaks become broader when the milling time is increased, as shown in Fig. 4b. This phenomenon is due to a stress-induced change in the lattice constant of Al in the composites. The stress level was observed to increase significantly with milling time, which indicates that the stress was created mainly during the ball-milling process. Furthermore, the shapes of the peaks suddenly changed for the composites that were milled for longer than 60 min. This suggests an evolution mainly by changes on particle shape and size (Fig. 2c and d). The crystallite size of the Al in the composites was calculated from the XRD patterns using the Scherrer equation (ignoring then any strain effect), as shown in Eq. (1). d ¼ 0:9
Kλ B cos θ
ð1Þ
where d is the crystallite size and λ, θ, and B represent the X-ray wavelength, the Bragg scattering angle, and the full width at half
342
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
Fig. 4. XRD diffraction of (a) the hot-pressed dual-nanoparticulate-reinforced Al 6061 alloy composites ball milled for various times. (b) Highly magnified main peaks of the composites shown in (a).
Fig. 5. Crystallite size and Al4C3 area calculated from the XRD diffraction patterns of the composites after ball milling times of 30, 60, 120, 180, and 360 min.
maximum (FWHM), respectively. K is a constant that depends on the crystallite shape (0.9). The values of FWHM measured from the XRD patterns (Fig. 4b) were calibrated using a standard sample (CeO2 NIST SRM-674b). The crystallite size of the Al in the composites decreased with increasing milling time (Table 1 and Fig. 5). A sudden change in crystallite size was observed between the composites made with powders milled for 60 min and 120 min. This is similar to the tendency observed as in the peak shape described above. Moreover, the crystallite size of the 360min-milled composite is three times as fine as that of the 30-minmilled composite. We performed a quantitative analysis of the Al4C3 formed in the composites based on the XRD diffraction patterns obtained. It is very difficult to precisely define the quantity of Al4C3 formed from the XRD results, but some tendencies (i.e., area of Al4C3) can be discerned, at least under the same conditions. Therefore, the area of the 2θ peak at 55.81 was calculated for the different composites and are plotted in Fig. 5 as a function of the powder milling time. These areas are differing by a factor of 10 approximately. An inverse correlation is observed between the Al crystallite size and the relative quantity of aluminum carbide. The Al4C3 content indeed increase more or less linearly until a powder milling time of 120 min while the crystallite size decreases. Then
a pronounced slope increase is seen for the carbide content while the crystallite size of Al drastically decrease after about 180 min, the crystallite size reaches a plateau while the carbide content further slightly increase. The particle size seems then to play a more important role in the formation of Al4C3 than the particle shape. The particle shape completely changed from a spherical and flaky morphology for the 30- and 60-min-milled based samples to a solely spherical morphology for the samples milled for over 120 min. This means that the particle shape did not significantly provide any driving force for the formation of Al4C3. However, the large surface area of the Al particles provided a higher reaction potential. The formation of aluminum carbide in Al-metal matrix composites reinforced with CNTs is carefully handled as on one hand Al4C3 is well known as a brittle and hygroscopic compound [20], but on the other hand if confined at the interface, it can efficiently transfer stress from the matrix to the CNT reinforcement material [5,15,21]. Regarding the potential role of aluminum carbide and according to the above results, 120 min milling time seems to be a good compromise solution for producing high performance Al-based nanocomposites with the composition Al6061–1vol% nSiC–6wt% CNT. The composites made with the powders milled for different times were also analyzed using Raman. The resulting spectra are presented in Fig. 6 showing the characteristic D (defect) and G (graphitic) bands of carbon. The relative peak intensities of these two bands decreased with increasing milling time. Therefore, the fraction of trapped and implanted CNTs in the soft Al matrix was relatively higher than the fraction of CNTs broken down during the ball-milling process due to high impact energy. This phenomenon caused fewer CNTs to be exposed to the laser beam during Raman spectroscopy, resulting in low peak intensities. In general, the ratio between the relative intensities of these two bands ID/IG is indicative of the CNT quality [22]. The ID/IG ratio of the composites was slightly reduced for long milling times (Table 1). However this decrease may not be significant as the very low intensities measured especially for the longer milling times may induce large errors too. Furthermore, the 60- and 120-min-milled samples also featured an inflection area, as also observed with XRD and SEM analyses. However, the destruction of the CNTs in the dualnanoparticulate-reinforced Al6061 alloy matrix composites submitted to ball milling and hot-pressing did not seriously affect the composites' material properties. Finally, no shifting or shape change of the typical carbon peaks occurred even though aspect ratio of the CNT changed and some Al4C3 formed, indicating that
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
the CNTs in the composites remained approximately the same condition. The composites were characterized mechanically with a fourpoint bending test; the Al6061 alloy without any reinforcement but hot pressed under the same conditions as the composites were also tested for comparison. The results of the bending tests are summarized also in Table 1. The bending strength of the composites increased with the milling time and reached a maximum value of approximately 300 MPa. Fig. 7 shows a digital photograph of the Al6061 alloy bulk before and after the bending test. The different composites did not show any cracks on their surface after the bending tests, but they were completely destroyed, whatever the powder milling time (Fig. 7b). The bending strength of the
Fig. 6. Raman spectra of the dual-nanoparticulate-reinforced Al 6061 matrix composite materials after ball milling times of 30, 60, 120, 180, and 360 min.
343
pure Al6061 alloy was impossible to measure under the test conditions due to the material's high plasticity; it deformed too much without breaking. Nevertheless, its yield strength was measured around 60 MPa, being the lowest strength value. After bending tests, the fracture surfaces of the samples were characterized by SEM observations. As for the composite powders (Fig. 2), many CNTs were found relatively easily in the boundary layer between the particles in the 30-min-milled composite sample (Fig. 8a). Most of the cracks propagated along the lamellar structure of the composite. The composites made out of short time milled powders (30 and 60 min) significantly showed a smooth fracture surface with much less dimpling than that observed in ductile materials such as Al; however, the composites made with powders milled for longer times showed clear dimples, as shown in Fig. 8c–e. This means that a longer milling time leads to a better bonding surface. In particular, the height of the elongated dimples was different in the samples milled for 120, 180, and 360 min. The length of the elongated dimples is directly related to the elongation or toughness of the composites. In general, formation of the proper Al4C3 also can offer better chemical link between the reinforcement and the matrix, resulting in leading to strong interface bonding [5,15,21]. According to the previously discussed XRD results, the produced composites with milling time 60 min are containing some of the small amount of Al4C3 and this formed carbide supported enhanced interface bonding. However, the longest elongated dimple was observed in the 180-min-milled composite and was about a few micrometers (Fig. 8d). The Vickers hardness and the bending strength were plotted together in Fig. 9 and showed similar enhancement as a function of the milling time. According to the Hall–Petch relationship, the strength of a material increases with decreasing grain size, and this relationship is valid down to a grain size of 10 nm. Below this grain size, the strength decreases due to grain sliding [23]. The mechanical properties are then in agreement with this principle as it has been showed before based on XRD analyses that the Al grain
Fig. 7. Digital photographs of a composite sample before and after bending test.
344
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
Fig. 8. SEM micrographs of fractured surfaces of the dual-nanoparticulate-reinforced Al 6061 alloy hot-pressed composites after ball milling times of (a and b) 30, (c) 60, (d) 120, (e) 180, and (f) 360 min.
Fig. 9. Vickers hardness and bending strength of the dual-nanoparticulate-reinforced composite materials as a function of the milling time.
size of the composites decreased with increasing milling time (Table 1). However, the composite performances were not only enhanced by the addition of nSiC and CNT but also by grain refinement. The Vickers hardness of the pure Al6061 alloy bulk used for comparison was 45 HV. The maximum hardness value of 330 HV was achieved for the 360-min-milled sample. This value is almost six times higher than that of the pure Al6061 alloy bulk. Additional samples reinforced with only 1 vol% nSiC or 6 wt% CNT were also elaborated under same conditions for comparison. The hardness values were however reaching a plateau after 360 min up to a maximum milling time of 1200 min. Vickers hardness values of approximately 100 HV and 300 HV were measured for the Al6061–1 vol% nSiC and Al6061-6 wt% CNT composites, respectively. Thus, the hardness of the dual-nanoparticulate-reinforced composite was enhanced by 300% compared to that of the composite with only nSiC added and by more than 10% compared to that of the composite with only CNT added. In other words, the hardness value of 300 HV measured for the 360-min-milled dualnanoparticulate-reinforced composite (approximately 334 HV in Table 1) was mainly due to the above-mentioned grain refinement and reinforcement effects. The remaining 34 HV (10%) can be
assumed by reasonable errors. Because the low-aspect-ratio (almost spherical in shape) nSiC (nanosized) particles easily infiltrate the agglomerated high-aspect-ratio CNTs during the ball-milling process and the nSiC particles function as nano-balls, the CNTs are well dispersed in the Al powders. This synergistic effect of the two nano-materials provided a sub-enhancement of approximately 34 HV of the total 334 HV enhancement observed for the 360-min-milled composite. The indentation modulus (almost equivalent to the elastic modulus) of each composite increased with the milling time. A sudden increase in the indentation modulus was observed between the 60- and 120-min-milled composites, as indicated in Table 1. This tendency was already observed and discussed within the context of the microstructural, XRD and Raman spectroscopy analyses. The indentation moduli of the 30- and 60-min-milled composites were lower than the modulus of the Al6061 alloy bulk (100% densified), mainly due to the porosity (Table 1). However, overall, a milling time of 120– 180 min is considered to be optimal for the employed dualnanoparticulate-reinforced Al6061 alloy matrix composite.
4. Conclusions Dual-nanoparticulate-reinforced Al 6061 alloy matrix composites made out of blends milled for various times were successfully fabricated by a high-energy ball milling and simple hot-pressing process. The shape of the Al particles in the composite powders varied with the milling time, varying from spherical and flaky to flaky only or finely spherical. This tendency was followed by the microstructure and fracture surface of the composites. The crystallite size of the Al particles decreased with increasing milling time. A small amount of Al4C3 was detected in the composites, except for the 30-min-milled composite, and was observed to increase with increasing milling time. We were able to greatly enhance the Vickers hardness of the dual-nanoparticulate-reinforced composites relative to the hardness of the composites featuring only nSiC or CNTs. The Vickers hardness and bending strength of the dualnanoparticulate-reinforced composites were enhanced with increasing milling time and showed maximum values that were approximately five times higher than those of pure Al 6061 alloy bulk. However, most of the results showed a high-enhancement region from 60 to 120 min of milling time, which occurred mainly due to particle-shape and -size effects. The high-energy ball
H. Kwon et al. / Materials Science & Engineering A 590 (2014) 338–345
milling time significantly affected the microstructure and mechanical properties of the composites, which suggests that the characteristic properties of the composites can be controlled by adjusting the milling time. We believe that dual-nanoparticulatereinforced composites have many potential applications as industrial component materials with precisely controlled properties.
Acknowledgments The authors would like to acknowledge Bernhard von Gunten at EMPA for his technical support. References 〈http://en.wikipedia.org/wiki/Duralumin〉. S. Iijima, Nature 354 (1991) 56–58. A.P. Ramirez, Bell Labs Tech. J. 10 (2005) 171–185. J. Tersoff, R.S. Ruoff, Phys. Rev. Lett. 73 (1994) 676–679. H. Kwon, D.H. Park, J.F. Silvain, A. Kawasaki, Compos. Sci. Technol. 70 (2010) 546–550. [6] S. Cho, K. Kikuchi, A. Kawasaki, Acta Mater. 60 (2012) 726–736. [1] [2] [3] [4] [5]
345
[7] H. Fukuda, K. Kondoh, J. Umeda, B. Fugetsu, Compos. Sci. Technol. 71 (2011) 705–709. [8] K.T. Kim, S.I. Cha, T. Gemming, J. Eckert, S.H. Hong, Small 4 (200) (1936–1940). [9] T. Kuzumaki, K. Miyazawa, H. Ichinose, K. Ito, J. Mater. Res. 13 (1998) 2445–2449. [10] A.M.K. Esawi, M.A. El Borady, Compos. Sci. Technol. 68 (2008) 486–492. [11] A.M.K. Esawi, K. Morsi, A. Sayed, M. Taher, S. Lanka, Compos. Part A Appl. Sci. Manuf. 42 (2011) 234–243. [12] J.Z. Liao, M.J. Tan, I. Sridhar, Mater. Des. 31 (2010) 96–100. [13] C.F. Deng, D.Z. Wang, X.X. Zhang, A.B. Li, Mater. Sci. Eng. A 444 (2007) 138–145. [14] S.R. Bakshi, A.K. Keshri, A. Agarwal, Mater. Sci. Eng. A 528 (2011) 3375–3384. [15] H. Kwon, M. Estili, K. Takagi, T. Miyazaki, A. Kawasaki, Carbon 47 (2009) 570–577. [16] H. Kwon, S. Cho, M. Leparoux, A. Kawasaki, Nanotechnology 23 (2012) 225704. [17] Y. Leconte, M. Leparoux, X. Portier, N. Herlin-Boime, Plasma Chem. Plasma Process. 28 (2008) 233–248. [18] A.W. Burton, K. Ong, T. Rea, I.Y. Chan, Microporous Mesoporous Mater. 117 (2009) 75–90. [19] L. Ci, Z. Ryu, N.Y. Jin-Phillipp, M. Rühle, Acta Mater. 54 (2006) 5367–5375. [20] T. Etter, P. Schulz, M. Weber, J. Metz, M. Wimmler, J.F. Loffler, P.J. Uggowitzer, Mater. Sci. Eng. A 448 (2007) 1–6. [21] S. Cho, k. Kikuchi, A. Kawasaki, H. Kwon, Y. Kim, Nanotechnology 23 (2012) 315705. [22] C. Thomsen, S. Reich, Top. Appl. Phys. 108 (2007) 115–232. [23] T. Shanmugasundaram, M. Heilmaier, B.S. Murty, V.S. Sarma, Mater. Sci. Eng. A 527 (2010) 7821–7825.