Composites: Part A 40 (2009) 1482–1489

Contents lists available at ScienceDirect

Composites: Part A

journal homepage: www.elsevier .com/locate /composi tesa

Structural characterization of a mechanically milled carbonnanotube/aluminum mixture

Dominique Poirier a,*, Raynald Gauvin a, Robin A.L. Drew b

a Mining and Materials Engineering, McGill University, 3610, University Street, Montreal, Canada H3A 2B2b Engineering and Computer Science, Concordia University, 1515, Ste-Catherine Ouest, Montreal, Canada H3G 2W1

a r t i c l e i n f o

Article history:Received 14 November 2008Received in revised form 25 May 2009Accepted 28 May 2009

Keywords:A. Metal-matrix compositesA. NanostructuresE. Powder processing

1359-835X/$ - see front matter � 2009 Elsevier Ltd.doi:10.1016/j.compositesa.2009.05.025

* Corresponding author. Tel.: +1 514 398 4755x032E-mail address: Dominique.Poirier@mail.mcgill.ca

1 www.cientifica.com, abstract of the ‘‘Nanotubes for(Augest 2005).

a b s t r a c t

The structural evolution of carbon nanotubes (CNTs) during mechanical milling was investigated usingSEM, TEM, XRD, XPS and Raman spectroscopy. The study showed that milling of the CNTs alone intro-duces defects but preserves the tubular structure. When milling the CNTs with aluminum (Al) powderin order to produce a composite, Raman spectroscopy has shown that most of the nanotubes aredestroyed. During sintering of the CNT/Al milled mixture, the carbon atoms available from the destruc-tion of the nanotubes react with the Al to form aluminum carbide (Al4C3). The effect of milling on the Almatrix was also studied.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Since the landmark paper on carbon nanotubes (CNTs) by Iijima15 years ago [1], researchers have demonstrated the exceptionalpotential and diversity of these materials. A single-wall carbonnanotube, i.e. a nanotube made of only one graphite sheet rolled-up in cylinder, has a Young’s modulus as high as 1.8 TPa and a ten-sile strength as high as 63 GPa [2], which is one to two orders ofmagnitude superior to the best known steels. Multi-wall nano-tubes display lower but still exceptional mechanical properties,and they are easier to synthesize. These superior mechanical prop-erties combined with a low density generate several outlets for acomposite reinforced by CNTs. The current prohibitive cost of CNTs(as high as $500 g�1 and higher for single-wall purified carbonnanotubes) is predicted to significantly decrease in the next fewyears due to an increased efficiency of production combined withthe development of new production routes (10–100� decrease inthe next 5 years according to Cientifica1). At some point, it can beexpected that an economical composite made of aluminum (Al)and CNTs could be suitable for the automotive or aerospace indus-tries, where the decrease of fuel consumption by weight reductionis a priority.

CNTs have a strong tendency to form bundles due to their highaspect ratio and their van der Waals bondings [3]. This extensiveagglomeration, detrimental for composite mechanical properties,

All rights reserved.

0; fax: +1 514 398 4492.(D. Poirier).the composite market” report

is a main issue when trying to use them as reinforcement [4].Mechanical milling, a solid state high-energy ball milling processwhere particles are repeatedly fractured and welded [5], has beenused successfully to disperse uniformly a variety of reinforcementswithin Al matrix [6–8]. Furthermore, the mechanical properties ofa CNT/Al composite made by mechanical milling would be furtherimproved by Al grain refinement up to the nanoscale due to theintensive plastic deformation and by the incorporation and disper-sion of the oxide layer initially present on the surface of Al powders[5]. However, different studies show that CNTs can be modifiedwhen milled alone, going from simple shortening [9–12] to amor-phisation [13]. Also, studies of CNT milling with metals, such asiron [14] or magnesium for hydrogen storage [15,16] indicateaccelerated CNT damages. Only few research groups have investi-gated the dispersion of CNTs in an Al matrix by mechanical milling[17–19], and their investigation of the effect of milling on thenanotube structure has been very limited.

In this present work, the possibility of producing CNT/Al nano-composites by mechanical milling is studied, the emphasis beingon the structural evolution of the CNTs and of the Al upon milling.Characterization of the milled mixtures was done using X-ray dif-fraction (XRD), Raman spectroscopy, scanning and transmissionelectron microscopy (SEM and TEM) and X-ray photoelectron spec-troscopy (XPS).

2. Experimental

Multi-wall carbon nanotubes used in this work were purchasedfrom Nanostructured and Amorphous Materials Inc. They have a

D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489 1483

purity higher than 95%, their inner diameter is between 10–30 nmand their length between 0.5 and 50 lm. As shown in Fig. 1a, mostnanotubes were found to be stacked cones and to contain bendsand other defects, which is typical of CNTs produced by chemicalvapor deposition (CVD). While higher quality nanotubes are avail-able, commercial composite applications will required a cheap andlarge scale production of CNTs, as it is the case with CVD [20]. Twotypes of Al powder were used: either spherical Al powder of A.P.S.3.0–4.5 lm, 97.5% purity or equiaxed Al powder of A.P.S. 7–15 lm,99.5% purity. In both cases, the main impurity is iron. It was statedelsewhere that the initial powder size does not affect the milling[5] and no differences were indeed found in this study while vary-ing the Al powder size. Fig. 1 shows the initial constituents.

As a reference test, CNTs were first milled alone for timesvarying from 0 to 5 h in a high energy Spex 8000 mill. In this typeof mill, a ball-containing vial is shaken in a figure of 8 motion at afixed speed of 1200 rpm and it is the impact of the balls againstthe sample and the end of the vial that induces milling and mix-ing. Tungsten 7/1600 carbide balls and container with a 10:1 ball-to-powder ratio were used. A mixture of 10 vol% CNT/Al was thenalso milled using the same procedures. In that case, the millingwas performed under argon atmosphere and 2 wt% stearic acidwas added as a process control agent to avoid excessive sticking

Fig. 1. The raw materials: (a) FEG-TEM picture of CNTs, (b) FEG-SEM micrographs of

and agglomeration of the Al [21]. After 5 h of milling, some of thepowder was heat treated at 630 �C in a tubular furnace undervacuum for 1 h to simulate the heat treatment induced duringthe sintering process. Also, the 10 vol% CNT/Al mixture was coldcompacted in a 2 cm diameter cylinder and sintered at 630 �Cin a tubular furnace under vacuum for 1 h to obtain a bulkmaterial.

Morphology of the nanotubes and the powders was studiedusing a Field Emission SEM Hitachi S-4700. XPS technique per-formed on a VG Escalab MII with Mg source and 300 W powder al-lowed an evaluation of the bonding changes with milling. XRDcharacterization of the powders was performed with a RigakuRotaflex Ru-200B type using Cu Ka1 radiation at 4.8 kW. Ramanspectra were recorded with a Renishaw Invia Raman spectrometerwith an argon laser light (514 nm). Nanotube structure was stud-ied with a Field Emission TEM Jeol JEM-2100F. A thin film fromthe sintered 10 vol% CNT/Al compact, produced by focus ion beam(FIB) Hitachi FB2000A, was also observed by TEM. Phase transfor-mation during heat treatment of the sintering process was studiedwith a Setaram Setsys 1750 differential scanning calorimeter(DSC). Hardness of the powders was evaluated using a Vickermicrohardness indentor with a load of 50 g and a dwell time of15 s.

Al powder, 3.0–4.5 lm and (c) FEG-SEM micrographs of Al powder, 7.0–15.0 lm.

1484 D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489

3. Results and discussion

3.1. Reference test: milling of CNTs alone

Fig. 2 presents the evolution of nanotube microstructure of ref-erence test with milling time. The initial nanotubes are long,smooth and do not display open tips, while milled nanotubes seemshorter, show open tips and are more broken and damaged. This isconsistent with literature findings [9–13]. After few hours of mill-ing, the nanotubes become compacted and evaluation of the struc-tural evolution with SEM becomes limited. Damage to the tubularstructure of the nanotubes should be minimized to maintain theirmechanical properties. However, the shortening of the nanotubescould be beneficial depending on applications. Shorten nanotubesshould be easier to disperse.

In order to better evaluate the evolution of nanotubes structureupon milling, X-ray diffraction patterns were obtained for the ini-tial nanotubes and for the nanotubes milled for 2 and 5 h (Fig. 3).The first peak at 26� corresponds to the interplanar spacing of0.34 nm between the sheets of the MWNTs, and is similar to thatof graphite. Two other peaks are visible at 43� and 53� correspond-ing to the (1 0 0) and (0 0 4) planes of multi-walled nanotubes.Magnification of the first peak on Fig. 3b shows that the peaks after2 and 5 h of milling are slightly broader than the initial peak. Thisis due to an increase in defects in the nanotube structure aftermilling. Since the difference is small, it is likely that most of theCNT tubular structure is conserved after 5 h of milling.

XPS confirms that milling damages the nanotube structure bydecreasing the amount of C–C bonds from 75% to 63% (Table 1).

Fig. 2. FEG-SEM micrographs showing the effects of milling time on n

Milling also leads to some oxidation by the formation of extra C–O and C@O bonds: the amount of oxygen present in the CNT, foundfrom the area under the oxygen peak, is increased, as shown in Ta-ble 2.

Overall, the different tests performed indicate that the millingof CNTs alone preserved the tubular structure even though CNTsbecame shorter and an increase in defects was observed.

3.2. Al evolution upon milling

Fig. 4 shows the evolution in morphology of the Al powder(10 vol% CNT/Al mixture) with milling time. After 30 min of mill-ing, Al particles are slightly flattened. At this stage, the deformationmechanism dominates. Agglomerates of CNTs appear as shown bythe arrow. After 2 h of milling, Al is in the shape of sheets com-posed of a few welded layers (Fig. 4b). These sheets will tend toweld together until bigger, harder and more equiaxed particlesare formed, as seen in our case after 5 h of milling. According tothe literature [6] these particles are more prone to fracture becausethey are larger and are more brittle due to strain hardening anddispersion of the second phase. This creates an equilibrium be-tween welding and fracture mechanisms. The process is at steadystate and the particle morphology and size stabilize. Fig. 4d stillshows the existence of a few CNTs on the surface of Al particlesafter 5 h of milling.

A XRD spectrum was obtained from the 10 vol% CNT/Al mix-tures milled for 2 and 5 h and from an Al powder standard in orderto evaluate the grain size refinement and the lattice strain. Thespectrum of the 10 vol% CNT/Al mixture milled for 5 h is also used

anotube microstructure (a) initial (b) 30 min, (c) 2 h and (d) 5 h.

15 25 35 45 55Degree

Inte

nsity

CNTs, (002)(a)

(b)

CNTs, (100) CNTs, (004)

20 22 24 26 28 30

Degree

Inte

nsity

Initial nanotubes

Nanotubes, 2h milling

Nanotubes, 5h milling

Initial nanotubes

Nanotubes, 2h milling

Nanotubes, 5h milling

Fig. 3. (a) X-ray diffraction peaks of initial CNTs and of CNTs milled for 2 and 5 h (b)magnification of the first peak (normalized intensities).

Table 1Bonding types of initial and 5 h milled CNTs from XPS.

Bonding type Bonding energy (eV) % Rel. at.

Initial CNTs Milled CNTs

C–C 284.8 75.4 ± 0.8 62.9 ± 0.6C–C–O or C–N 285.8 13.3 ± 0.1 20.3 ± 0.2C–O 286.9 5.5 ± 0.6 6.6 ± 0.7C@O 287.5 2.7 ± 0.3 3 ± 0.3C@O aromatic 531.4 1 ± 0.5 1.6 ± 0.8C@O 532.8 1.2 ± 0.6 2.4 ± 0.2C–O 533.9 0.9 ± 0.5 2.7 ± 0.3

Table 2Composition of initial and 5 h milled CNTs from XPS.

Element % Rel. at

Initial CNTs Milled CNTs

O 2.3 ± 0.2 6.5 ± 0.7C 97.7 ± 1 93.5 ± 0.9

D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489 1485

later on for phase characterization (Fig. 11). The Williamson andHall method [22] was used to relate the broadening of the peak(B) to the crystallite size (L) and the lattice strain (g), in the form:

B cos h ¼ kkLþ g sin h ð1Þ

k being the Scherrer’s constant. It is also identified as geometricalfactor, because it can vary from 0.89 to 1.39 depending on parame-ters such as XRD line profile shape and grain shapes and size distri-bution. The total broadening (BT) was found by dividing the peak

area by its height. The broadening due to the instrument (BI), foundwith the annealed Al standard, was removed from each peak usingthe geometrical mean defined as:

B ¼ ðBT � BIÞ � ðB2T � B2

I Þ1=2

� �1=2ð2Þ

For each milled mixture, a plot of B � cos h as a function of sin hwas drawn for the first six peaks of the Al XRD spectrum as shownin Figs. 5 and 6. According to (Eq. (1)), the slope of the line obtainedby a regression of the data points gives the lattice strain, and thecrystallite size was calculated from the intersection with the Y-axis, assuming a Scherrer constant of 1. Table 3 shows the resultsobtained. As expected, crystallite size is decreasing with millingto reach a value of around 40 nm after 5 h of milling. However,lattice strain is lower after 5 h milling than after 2 h milling. Theextra deformation could have induced recrystallization. TEMobservations of the 5 h milled mixture corroborate the grain sizevalues obtained. Fig. 7 is an example of the dark field micrographsobtained with the FEG-TEM showing some grains in the milledpowder. Size measurements of around 30 grains indicate an aver-age grain size of around 20 nm, which is in the same order as thevalues obtained from XRD patterns, but smaller. This could bedue to sampling statistics from TEM measurements, as well asinaccuracies introduced by the numerous assumptions of theWilliamson and Hall method. It is also interesting to notice thatthe latter method leads to volume-weighted mean of distributionof size while TEM measurements give arithmetic mean which istypically smaller [23].

3.3. Nanotube structure evolution upon milling with Al

Nanotube structure evolution upon milling with and without Alwas studied with Raman spectroscopy. Fig. 8 shows the first orderpeaks composed of the D band at around 1350 cm�1 and the Gband that is typically located at 1580 cm�1 [24]. D and G bandsare due to sp2 sites. D band occurs due to the breathing modes ofsp2 atoms in rings. G band is due to the bond stretching of all pairsof sp2 atoms in both rings and chains [25]. G band is observed forperfect hexagonal graphite, and will therefore appear wider andsmaller for badly structured and defect containing nanotubes. Onthe other hand, D band is attributed to lattice defects and finitecrystal size which induce a breaking of the 2D translationalsymmetry, so that it will increase and broadened with increasingnumber of defects in the nanotubes [24]. The change in the twoband intensities can be quantified with the ratio ID/IG, calculatedin Table 4. Higher ratios are obtained after milling since thecreation of lattice defects increases the D band intensity whiledecreasing the G band intensity. The broadening of the D and Gpeaks with milling observed here also indicates an amorphisationof the nanotubes with milling, this amorphisation being accentu-ated with the presence of Al. Other indications of an increase indefects causing changes in electronic structure are the shift ofthe G bands toward higher frequencies with milling and the shifttoward lower frequencies of the D band [26].

Fig. 9 shows G0 band, an overtone of the D band, at around2700 cm�1. This band is decreasing with milling, again indicatinga certain amorphisation [24]. It is barely present for the CNTsmilled with Al, probably meaning that the carbon present in thismixture is mostly non graphitic. The band at 2900 cm�1 is due tothe presence of C–H bonds in the sample [25] or is representativeof D + G band [27].

Summarizing Raman results, milling has a drastic effect onnanotube structure. The milling of the nanotubes alone introducessignificant defects in the structure and reduces the crystallite size.Damages are accentuated if the CNTs are milled with Al. The Alpowder can be seen as many mini balls, increasing the number

Fig. 4. FEG-SEM micrographs of the mixture 10 vol% CNT/Al after (a) 30 min milling (b) 2 h milling, (c) 5 h milling and (d) 5 h milling at higher magnification.

Fig. 5. Williamson and Hall plot of the 10 vol% CNT/Al powder after 2 h milling. Fig. 6. Williamson and Hall plot of the 10 vol% CNT/Al powder after 5 h milling.

Table 3Crystallite size and lattice strain values for the 2 h sand 5 h milled 10 vol% CNT/Alsamples.

Values 2 h Milling 5 h Milling

Crystallite size (nm) 190 40Lattice strain 0.0039 0.0021

1486 D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489

of collisions sustained by the CNTs and thus the extent of defects. ACNT ball milling study has shown that 1 h of CNT milling with MgOpowder was equivalent to about 90 h of milling without MgO [9].Al powder, even if softer and more ductile than MgO, similarlyaccelerates milling.

Fig. 10 shows FEG-TEM images of CNTs milled with Al for30 min. CNTs are clearly visible in the Al matrix and minimalstructural changes seem to have occurred. After 5 h, it was notpossible to find any nanotubes in the Al matrix. A typical micro-structure is shown on Fig. 11. It seems that after 5 h of milling,the majority of the nanotubes are so damaged that their tubularstructure is lost.

3.4. Study of the Al4C3 formation

Fig. 12 shows the XRD patterns of the mixture of 10 vol% CNT/Alinitially, after milling for 5 h only and milled for 5 h, then followed

Fig. 7. FEG-TEM dark field micrograph of the 5 h milled 10 vol% CNT/Al powder.

0

500

1000

1500

2000

2500

3000

3500

1000 1200 1400 1600 1800 2000

Raman Shift, cm-1

Ram

an In

tens

ity

0

2000

4000

6000

8000

10000

12000

14000

16000CNTs iniMilled CNTsCNTs+Al milled

DG

Fig. 8. Raman spectra taken from the initial CNTs, the 5 h milled CNTs and the 5 hmilled 10 vol% CNT/Al mixture.

Table 4Raman spectra characteristics of CNTs for different milling states.

Milling states Peak positions (cm�1) Ratio ID/IG

D G

Initial CNTs 1355 1585 1.1Milled CNTs 1353 1589 1.3CNTs milled with Al 1343 1599 1.9

0

500

1000

1500

2000

2500

2500 2600 2700 2800 2900 3000 3100

Raman Shift, cm-1

Ram

an In

tens

ity

60006500700075008000850090009500100001050011000CNTs ini

Milled CNTsCNTs+Al milled

G'

Fig. 9. Raman spectra at higher energies taken from the initial CNTs, the 5 h milledCNTs and the 5 h milled 10 vol% CNT/Al mixture.

Fig. 10. FEG-TEM micrograph of the 10 vol% CNT/Al mixture milled for 30 min.

D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489 1487

by the 630 �C heat treatment. It can be observed that after milling,the CNT peak is significantly reduced. After heat treatment, it to-tally disappears. This could indicate CNT amorphisation and prob-ably some supersaturated solid solution of the carbon. Since thedetection limit of the XRD significantly increases with refined sec-ond phase [28], as it is the case with the CNTs getting dispersed, afirm conclusion is not possible. What is definitely visible, though, isthe predominance of the Al4C3 peaks after heat treatment. Observ-ing the 10 vol% CNT/Al sintered compact under TEM, Al4C3 precip-itates have been found, as seen on Fig. 13. The diffraction pattern

taken from the same compact confirms the presence of Al4C3,and also of aluminum oxide (Al2O3) due to Al oxidation (Fig. 14).Another indirect proof of carbide formation is the significant in-crease in hardness of the powder after milling and heat treatment,as shown in Table 5.

Two DSC curves are shown in Fig. 15. The first peak observed isassociated to water release. As the heat flow is in arbitrary unit, thepeak intensities are relative and can not be compared. The mixture10 vol% CNT/Al before milling does not show any exothermic peakranging from 450 to 650 �C. However, the milled mixture displays asignificant peak that has been identified as the exothermic Al4C3

formation elsewhere [29]. With DSC, it was found that Al4C3 doesnot form at all with unmilled CNTs, thanks to the high stabilityof their structure, but will start forming at around 450 �C aftermilling. It is therefore believed that the milling of the CNTs withthe Al lead to the destruction of the majority of the nanotubes.They either form amorphous carbon clusters or dissolve in the Almatrix. Upon heating, the carbon atoms in solution transform intoAl4C3 releasing heat.

Fig. 11. FEG-TEM micrograph of the 10 vol% CNT/Al mixture milled for 5 h.

Fig. 12. XRD patterns of the initial, milled and milled followed by heat treatment10 vol% CNT/Al mixtures.

Fig. 13. FEG-TEM micrograph of a Al4C3 precipitate observed in the sintered 10 vol%CNT/Al compact.

Fig. 14. Diffraction pattern of the sintered 10 vol% CNT/Al compact.

Table 5Vickers microhardness of the 10 vol% CNT/Al powders after milling and heattreatment.

Material Powder microhardness

After milling After milling and HT

10 vol% CNT 168 ± 10 329 ± 11

1488 D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489

CNTs are stable in Al and do not form carbide as long as theydisplay high-quality and low defect tubes [30–32]. However, sev-eral composite fabrication processes involve either initial CNTs oflow quality, CNT purification by acid washing, CNT functionaliza-tion or ball milling. All those steps create defects detrimental toCNT stability. Practically, a trade-off will most probably have tobe made between CNT dispersion vs. CNT damages.

4. Conclusions

In this paper, mechanical milling was investigated as a way todisperse CNTs in an Al powder. It was found from FEG-SEM, XRDand XPS observations that milling of the CNTs alone introduces de-fects in the nanotubes but preserve the tubular structure. Whenmilling the nanotubes with the Al, the metal undergoes grainrefinement, an increase in lattice strain and formation of disper-soids. Hence, a beneficial strengthening of the Al matrix is ex-pected. However, with the tested milling conditions, Ramanspectroscopy has shown that most of the nanotubes milled withthe Al are destroyed. Upon heating to 450 �C or higher, during sin-

tering, the free carbon atoms react with the Al to form Al4C3. There-fore, even if the high energy milling induced by the SPEX 8000 M isvery efficient for reinforcement dispersion and Al matrix strength-

Fig. 15. DSC curves of 10 vol% CNT/Al mixtures before and after milling.

D. Poirier et al. / Composites: Part A 40 (2009) 1482–1489 1489

ening, it can be too energetic to preserve the nanotube structure.Optimization between dispersion extent and CNT structure preser-vation is currently being investigated.

Acknowledgements

The authors thank Dr. Céline Goujon from CNRS for the earlywork in the project, Suzy Poulin from École Polytechnique for theXPS measurements, Matthieu Paillet from Université de Montréalfor Raman spectroscopy and Slavek Poplawski from McGill Univer-sity for the XRD measurements. The work was funded by NSERC,FQRNT and Nanoquébec, whose support is gratefullyacknowledged.

References

[1] Iijima S. Helical microtubules of graphitic carbon. Nature 1991;354:56–8.[2] Popov VN. Carbon nanotubes: properties and application. Mater Sci Eng

2004;R43:61–102.[3] Kim HS, Park W, Kang M, Jin HJ. Multiple light scattering measurement and

stability analysis of aqueous carbon nanotube dispersions. J Phys Chem Solids2008;69:1209–12.

[4] Robertson J. Realistic applications of CNT. Mater Today 2004:46–52.[5] Suryanarayana C. Mechanical alloying and milling. Prog Mater Sci

2001;46:1–184.[6] Fogagnolo JB, Velasco F, Robert MH, Torralba JM. Effect of mechanical alloying

on the morphology, microstructure and properties of aluminum matrixcomposite powders. Mater Sci Eng 2003;A342:131–43.

[7] Son HT, Kim TS, Suryanarayana C, Chun BS. Homogeneous dispersion ofgraphite in a 6061 aluminum alloy by ball milling. Mater Sci Eng2003;A348:163–9.

[8] Ruiz-Navas EM, Fogagnolo JB, Velasco F, Ruiz-Prieto JM, Froyen L. One stepproduction of aluminium matrix composite powders by mechanical alloying.Composites: Part A 2006;37(11):2114–20.

[9] Liu F, Zhang X, Cheng J, Tu J, Kong F, Huang W, et al. Preparation of short carbonnanotubes by mechanical ball milling and their hydrogen adsorption behavior.Carbon 2003;41:2527–32.

[10] Kim YA, Hayashi T, Fukai Y, Endo M, Yanagisawa T. Effect of ball milling onmorphology of cup-stacked carbon nanotubes. Chem Phys Lett2002;355:279–84.

[11] Pierard N, Fonseca A, Konya Z, Willems I, Van Tendeloo G, Nagy JB. Productionof short carbon nanotubes with open tips by ball milling. Chem Phys Lett2001;335:1–8.

[12] Konya Z, Zhu J, Niesz K, Mehn D, Kiricsi I. End morphology of ball milled carbonnanotubes. Carbon 2004;42:2001–8.

[13] Pierard N, Fonseca A, Colomer JF, Bossuot C, Benoit JM, Van Tendeloo G, et al.Ball milling effect on the structure of single-wall carbon nanotubes. Carbon2004;42(8–9):1691–7.

[14] Li YB, Wei BQ, Liang J, Yu Q, Wu DH. Transformation of carbon nanotubes tonanoparticles by ball milling process. Carbon 1999;37:493–7.

[15] Wu CZ, Wang P, Yao X, Liu C, Chen DM, Lu GQ, et al. Hydrogen storageproperties of MgH2/SWNT composite prepared by ball milling. J Alloys Compd2006;420:278–82.

[16] Chen D, Chen L, Liu S, Ma CX, Chen DM, Wang LB. Microstructure and hydrogenstorage property of Mg/MWNTs composites. J Alloys Compd 2004;372(1–2):231–7.

[17] Esawi A, Morsi K. Dispersion of carbon nanotubes (CNT) in aluminum powder.Composites: Part A 2007;38(2):646–50.

[18] George R, Kashyap KT, Rahul R, Yamdagni S. Strengthening in carbonnanotube/aluminium (CNT/Al) composites. Scripta Mater 2005;53:1159–63.

[19] Perez-Bustamante R, Estrada-Guel I, Antunez-Flores W, Miki-Yoshida M,Ferreira PJ, Martinez-Sanchez R. Novel Al-matrix nanocomposites reinforcedwith multi-walled carbon nanotubes. J Alloys Compd 2008;450:323–6.

[20] Ci L, Ryu Z, Jin-Phillipp NY, Ruhle M. Investigation of the interfacial reactionbetween multi-walled carbon nanotubes and aluminum. Acta Mater2006;54:5367–75.

[21] Rocha CJ, Leal Neto RM, Goncalves VS, Carvalho LL, Ambrozio Filho F. Aninvestigation of the use of stearic acid as a process control agent in high energyball milling of Nb–Al and N–Al powder mixtures. Mater Sci Forum 2003;416–418:144–9.

[22] Suryanarayana C, Norton MG. X-ray diffraction: a practical approach. NewYork: Plenum; 1998. pp. 207–21.

[23] Langford JI, Louer D, Scardi P. Effect of a crystallite size distribution on X-raydiffraction lime profiles and whole-powder-pattern fitting. J Appl Cryst2000;33:964–74.

[24] Delhaes P, Couzi M, Trinquecoste M, Dentzer J, Hamidou H, Vix-Guterl C. Acomparison between Raman spectroscopy and surface characterizations ofmultiwall carbon nanotubes. Carbon 2006;44:3005–13.

[25] Casiraghi C. Raman spectroscopy of hydrogenated amorphous carbons. PhysRev 2005;B 72(8):085401/1–085401/14.

[26] Mcguire K, Gothard N, Gai PL, Dresselhaus MS, Sumanasekera G, Rao AM.Synthesis and Raman characterization of boron-doped single-walled carbonnanotubes. Carbon 2005;43:219–27.

[27] Antunes EF. Comparative study of first- and second-order Raman spectra ofMWCNT at visible and infrared laser excitation. Carbon 2006;44:2202–11.

[28] Kim HS, Suhr DS, Kim GH, Kum DW. Analysis of X-ray diffraction patterns frommechanically alloyed Al–Ti powders. Metals Mater 1996;2(1):15–21.

[29] Zhou Y, Li ZQ. Structural characterization of a mechanical alloyed Al–Cmixture. J Alloys Compd 2006;414:107–12.

[30] Kuzumaki T, Miyazawa K, Ichinose H, Ito K. Processing of carbon nanotubesreinforced aluminum composite. J Mater Res 1998;13(1):2445–9.

[31] Xu CL, Wei BQ, Ma RZ, Liang J, Ma XK, Wu DH. Fabrication of aluminum–carbon nanotube composite and their electrical properties. Carbon1999;37:855–8.

[32] Ci L, Ryu Z, Jin-Phillipp NY, Ruhle M. Investigation of the interfacial reactionbetween multi-walled carbon nanotubes and aluminum. Acta Mater2006;54:5367–75.