Calhoun: The NPS Institutional Archive

Faculty and Researcher Publications Faculty and Researcher Publications

2013

Synthesis of nanodiamond-reinforced

aluminum metal composite powders and

coatings using high-energy ball milling

and cold spray

Woo, D. J.

Elsevier

Carbon, v. 63, 2013, pp. 404-415

http://hdl.handle.net/10945/41321

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5

.sc iencedi rect .com

Avai lab le at www

journal homepage: www.elsev ier .com/ locate /carbon

Synthesis of nanodiamond-reinforced aluminummetal composite powders and coatings usinghigh-energy ball milling and cold spray

0008-6223/$ - see front matter Published by Elsevier Ltd.http://dx.doi.org/10.1016/j.carbon.2013.07.001

* Corresponding author: Fax: +1 831 656 2834.E-mail address: sosswald@nps.edu (S. Osswald).

D.J. Woo a, B. Sneed b, F. Peerally a, F.C. Heer b, L.N. Brewer b, J.P. Hooper a,S. Osswald a,b,*

a Department of Physics, Naval Postgraduate School, 833 Dyer Road, Monterey, CA 93943, USAb Department of Mechanical and Aerospace Engineering, Naval Postgraduate School, 700 Dyer Road, Monterey, CA 93943, USA

A R T I C L E I N F O

Article history:

Received 16 April 2013

Accepted 1 July 2013

Available online 8 July 2013

A B S T R A C T

Nanodiamond-reinforced aluminum metal matrix composites (ND–Al MMC) powders were

synthesized by means of high energy ball milling. We present a systematic study of the

effect of various milling conditions on the structure and properties of the resulting MMC

powders. The described method can be used to control important powder characteristics,

including particle size and shape, Al crystal size and residual strain, and structural integrity

and dispersion of the nanoparticle inclusions, a crucial requirement for subsequent pow-

der consolidation. Raman spectroscopy was utilized for the first time to directly verify

the structural integrity and the dispersion of ND in the Al matrix. For low ball-to-powder

ratios (BPR), average particle size and size range of the ND–Al composite powders were

found to decrease during milling, while the hardness increases. A BPR of 10:1, a milling

time of 10 h, and a ND content of 10 wt.% were most effective in obtaining small powder

particle sizes, small Al crystal sizes, and improved mechanical properties reaching a hard-

ness of 3.46 GPa, a 210% increase over the pristine, untreated Al powder (1.10 GPa). Finally,

we demonstrate that the as-produced composite powders are well-suited for low-temper-

ature consolidation processing by fabricating the first cold-sprayed ND–Al MMC coating.

Published by Elsevier Ltd.

1. Introduction

The need for lightweight, high strength materials has been

recognized in numerous fields ranging from the automotive

and aircraft industries, to shipbuilding, and a variety of mili-

tary applications. As stiffness and hardness of a material in-

crease, the mass required for a certain load-bearing

application is reduced. This leads to several advantages,

including a potential increase in the maximum payload

and/or improvements in fuel efficiency. However, for many

applications, the challenge lies in finding materials that

possess the required mechanical strength, but that are also

light weight and inexpensive. While few materials in their

elemental form can meet these requirements, the develop-

ment of nanostructure-reinforced metal matrix composites

(MMCs) has opened new pathways for the design of new

materials that exhibit a unique set of properties, such as high

strength and low weight.

The performance of MMCs depends primarily on the abil-

ity (1) to adequately disperse the reinforcement agents in the

metal matrix, (2) to maintain its structural integrity during

composite synthesis and processing, and (3) to establish a

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5 405

suitable matrix-reinforcement agent interface [1–6]. The

strength of a MMC can be further enhanced by reducing the

crystal size of the metal matrix to the lower nanoscale [7].

The majority of the studies reported in the literature involv-

ing nanocarbon-based MMC’s focused their efforts on the rein-

forcement of ductile metals, such as aluminum (Al),

magnesium (Mg), and copper (Cu). Owing to their exceptional

mechanical properties and high aspect ratio, carbon nanotubes

(CNT) have been the reinforcement agent of choice in most of

these studies [1,8–10]. Carbon nanotube-reinforced aluminum

MMCs (CNT-Al MMC) have successfully been produced using a

variety of synthesis methods, including extrusion [3], sintering

and extrusion [11,12], pressing and rolling [13], infiltration [14],

and plasma spray [15]. In nearly all of these studies, CNT were

mixed with Al powder, followed by ball milling, a crucial step

required to properly disperse the nanostructures prior to pow-

der compaction [3,16–18]. With CNT-contents ranging between

1.5 and 15 wt.%, the reported improvements in mechanical

properties were primarily increases in yield strength, ultimate

tensile strength and Young’s modulus [3,11–18]. For a detailed

overview of existing literature on CNT-metal composites, the

reader is referred to a recent review by Bakshi et al. [19]. While

CNTs are a suitable reinforcement agent for a variety of applica-

tions, nanodiamond (ND) offers the possibility of achieving

superior material hardness, well beyond that possible for

CNT-Al MMCs. Increases in hardness generally scale with in-

creases in yield strength, another critical mechanical property

for most structural applications.

ND, another member of the carbon nanomaterial family

that has recently sparked interest in the materials commu-

nity [20], is considered the hardest naturally occurring mate-

rial [21], making it an ideal candidate for the reinforcement of

Al-based MMCs. The average size of the individual nanodia-

mond crystals is around 4–5 nm; however, due to the strong

tendency of nanomaterials to agglomerate, the observed par-

ticle size typically ranges from 50 to 500 nm [22]. Current

applications include biomedical imaging material [23], nano-

scale magnetic sensors [24], biocompatible materials [25,26],

cooling fluids [27], lubricants [28], electroplating baths [29]

and composites [30–33].

Existing studies on using ND as a reinforcing agent for

MMCs include incorporation into aluminum [10,34], copper

[35,10], and aluminum–magnesium [36], aluminum–copper

[10] and copper–tungsten alloys [10]. As expected, hardness

is the most thoroughly investigated characteristic of ND-rein-

forced MMCs. Using planetary ball milling, Livramento et al.

dispersed 10 and 20 wt.% ND in a nanocrystalline copper ma-

trix, yielding a 20% and 40% increase in Vicker’s hardness,

respectively, as compared to pure nanocrystalline Cu [35]. In

their study, the saturation point of ND in the Cu-based MMC

was �30 wt.%. Bobrovnitchii et al. demonstrated the potential

of ND as reinforcement agent in Al. Using a combination of

powder mixing and high pressure sintering, they increased

the hardness of Al by incorporating up to 50 wt.% ND into

Al matrix and reported a nearly linear increase in MMC hard-

ness with ND content [10]. In their study, incorporation of

10 wt.% ND lead to composite hardness of �2.34 GPa after

mixing and sintering at 2.0 GPa and 700 �C for 3 min, a

�138% increase over pure Al (0.98 GPa). While the authors re-

ported the formation of aluminum carbide phases, as ex-

pected under these conditions, no information on the

structural integrity and dispersion of the ND was given. Hard-

ness measurements on powder particles were not conducted.

Kaftelen and Ovecoglu [34] synthesized Al–ND MMCs con-

taining up to 10 wt.% ND by employing a combination of high-

energy ball milling and sintering. Precursor powders were

milled for up to 2 h using 6-mm steel balls and a ball-to-pow-

der ratio (BPR) of 7:1. A maximum ND–Al MMC powder parti-

cle hardness of �1.4 GPa (10 wt.% ND) was reported. After

sintering, the hardness increased to �3.3 GPa, indicating

some discrepancies between the hardness measurements of

the precursor powders and the consolidated MMC. Similar

to the study by Bobrovnitchii et al., they reported a linear rela-

tionship between ND content and hardness in the sintered

composites. The sintering also led to substantial carbide for-

mation, suggesting strong chemical interaction between ND

and Al matrix. Again, no data on ND dispersion in powder

particles or its structural integrity after milling and sintering

was reported.

Since most MMC synthesis techniques utilize some form of

powder consolidation, the structural and compositional charac-

teristics of the precursor powder are of great importance and

strongly affect the properties of the consolidated MMC. For

example, without any treatment of the precursor powders,

the best obtainable dispersion of the reinforcement agents for

consolidation methods such as pressing, sintering, rolling,

extrusion, and thermal spray [37–39], would be limited to the

size of the powder particles. In the case of thermal spray meth-

ods, particle size and shape, hardness, and residual strain in

precursor powders control important processing parameters

such as particle velocity, deposition efficiency, and coating den-

sity [40]. The ability to fully control both structural characteris-

tics and mechanical properties of the composite powders is

therefore crucial for further processing and consolidation, par-

ticularly when employing thermal spray techniques. Control

over these powder characteristics can be achieved by ball mill-

ing, also known as mechanical alloying (MA), a versatile powder

processing method commonly utilized to mix different pow-

dered elements into a homogeneous composite [37,41].

While previous studies have demonstrated the great poten-

tial of ND as reinforcement agent for lightweight metals, these

studies focused primarily on the consolidated (typically sin-

tered) composites rather than the precursor powders and there-

fore lack a comprehensive analysis of the relationship between

synthesis conditions and resulting powder characteristics.

In this paper, we utilize MA to synthesize ND-reinforced Al-

MMC (ND–Al MMC) precursor powders, intended to be used for

cold spray deposition, a low-temperature thermal spray meth-

od. The primary focus is on investigating the effects of milling

parameters such as milling time, ball-to-powder mass ratio

(BPR) and ND concentration, on both the structural characteris-

tics and the mechanical properties of the MMC powder.

2. Experimental

2.1. Materials

The Al powder (H-10) was purchased from Valimet Inc. (Stock-

ton, CA). Particle size and purity were given by the manufac-

turer as �325 mesh and 99.7%, respectively. ND powder was

406 C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5

obtained from the International Technology Center (Research

Triangle Park, NC, USA). Average ND crystal and agglomerate

size were reported as 4–5 nm and 200 nm, respectively. The

process control agent (PCA) used in this study was stearic acid

(Sigma–Aldrich) with a purity of >95%. All chemicals were

used as-received, without further purification.

2.2. Composite synthesis

2.2.1. MMC powderMMC powders were produced using a SPEX 8000 M high-en-

ergy ball mill (1735 RPM) with a SPEX 65 mL-hardened steel

vial and 5 mm-hardened steel grinding balls. Sample vials

were sealed in an argon-filled glove box to remove oxygen

and moisture from the vial prior to a ball milling. For all

experiments, 3.0 wt.% stearic acid was added to the powder

mixture as a PCA to prevent agglomeration. The ND content

in the Al matrix was varied between 0, 5 and 10 wt.%. Milling

conditions included ball-to-powder mass ratios of 10:1, 20:1

and 30:1, and milling times of 1, 4 and 10 h, leading to a total

of 27 different MMC powders.

2.2.2. MMC coatingMMC powders were consolidated using a Series C cabinet cold

spray system from CenterLine (Windsor) Limited, Supersonic

Spray Technologies (SST), Canada. The system was operated

using a vibrational feed hopper and a computer controlled

2-axis robotic spray gun. Powders were deposited on an exter-

nally heated (�160 �C) 1018 steel substrate. Prior to the depo-

sition, the substrate surface was cleaned and roughened

using aluminum oxide grit (SST-G0002) from CenterLine.

Nitrogen was selected as spray gas with temperature and

pressure set to 450 �C and 250 psi, respectively.

2.3. Material characterization

The morphology of the ball-milled MMC powders was charac-

terized using a Zeiss Neon 40 field emission gun scanning

electron microscope–focused ion beam (FIB–SEM) instrument

(Jena, Germany) operated at a voltage of 20 kV.

A laser diffraction particle size analyzer (Partica LA-950,

Horiba, Kyoto, Japan) was used for quantitative particle size

analysis. Around 300 mg of each sample was mixed with iso-

propanol (dispersant liquid), and transferred to an analytic

glass cell. Raman spectra were recorded with an inVia Raman

Microspectrometer from Renishaw (UK) using 325-nm HeCd

laser, a 2400 l/mm grating, and 10· objective (spatial resolu-

tion �0.8 lm). For Raman imaging of the cross-section of

composites, polished samples for nanoindentation were

used. X-ray diffraction (XRD) was conducted utilizing a Phil-

lips PW1830 diffractometer with a Cu target at a current of

30 mA and voltage of 35 kV. The Williamson–Hall analysis

[42] was used to relate changes in the width of the XRD peaks,

bsample (integral breadths), to the lattice strain (e) and the crys-

tallite size (L):

bsample cos h

k¼ e

ksin hþ k

Lð1Þ

where k (Scherrer constant) and k (wavelength of X-ray) were

set to 0.94 and 0.1542 nm, respectively. In the linear plot of

bsample cosh vs. sinh, the reciprocal of the y-intercept multi-

plied by the constant k yields the crystallite size, while the

slope of the curve divided by k equals the microstrain. The

peak broadening related to the instrument (binst) was deter-

mined using a LaB6 standard (SRM 660b from National Insti-

tute of Standards & Technology). XRD peaks recorded from

the samples were fitted using a mixed Lorentzian–Gaussian

function. The binst were then removed from the measured

Lorentzian (bobs-L) and Gaussian (bobs-G) contributions using a

geometrical average [42,43]:

bsample ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðbobs�L � binstÞ �

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffib2

obs�G � b2inst

qrð2Þ

the integral breadths of the Lorentzian and Gaussian profile

were calculated from the measured full-width at half maxi-

mum (FWHM) values of the peaks according to:

bobs�L ¼p � FWHM

2ð3Þ

bobs�G ¼ffiffiffiffiffiffiffiffiffiffiffiffi

plnð2Þ

r� FWHM

2ð4Þ

For nanoindentation measurements, sample powders

were dispersed in an epoxy resin. EpoFix resin and hardener

from Struers (Cleveland, OH) were mixed together in a mass

ratio of 5:1 and transferred to a silicone rubber mold contain-

ing the composite powders. After curing at ambient tempera-

ture for 24 h, the embedded powder sample was mechanically

polished using 400, 800, 1200, and 4000 grit paper with 0.1 and

0.05 lm-aluminum oxide suspensions. The hardness of the

composites was characterized by a G200 nanoindenter from

Agilent Technologies (Santa Clara, CA, USA). Indentation data

was collected using the continuous stiffness monitoring

(CSM) method. A Poisson’s ratio of 0.33, a measuring depth

of 100–200 nm from the depth limit of 500 nm, an allowable

thermal drift rate of 0.05 nm/s, a strain rate of 0.05 1/s, a har-

monic displacement target of 2 nm, a frequency of 45 Hz, and

a surface approach velocity of 10 nm/s and an approach dis-

tance of 1000 nm were set for the test condition. The Berko-

vich tip shape consisting of diamond was calibrated using a

fused silica reference standard. Dependent upon the sample’s

surface, a 60· or 100· objective lens was used to locate nano-

indentation points on the center of individual powder particle

cross sections; 10–15 different points were measured on each

sample. Because the powder particles were suspended in

epoxy, the effective machine compliance for each particle var-

ied and was corrected using the approach of Buchheit and

Vogler [44]. In this approach, the effective machine compli-

ance is measured for each powder particle and then used to

adjust the indentation measurement such that the stiffness-

square divided by the load is constant for all depths.

3. Results and discussion

3.1. Particle morphology

In the scope of this study, a full-factorial series of 27 different

Al–ND composites was produced by varying ND content and

critical milling parameters, including milling time and ball-

to-powder ratio, all at three levels. Fig. 1 shows the SEM

images of three MMC samples that contain 5 wt.% ND, in

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5 407

comparison to the as-received Al powder. All of these pow-

ders were milled using a ball-to-powder ration of 10:1, but dif-

ferent milling times ranging from 1 (b) to 4 (c) to 10 h (d). It can

be seen that both particle size and shape change notably

upon milling, as expected. The average particle size decreases

with milling time, but remains larger for all milled samples as

compared to the as-received powder. The size distribution of

the powder particles also narrows significantly as the milling

time increases. During the initial phase of the milling process,

powder particles are cold-welded together and flattened by

the milling media, leading to a substantial increase in particle

size. With increasing milling time, particles are work hard-

ened, start to fracture, and become more rounded, leading

to a decrease in particle size and a more homogeneous pow-

der with respect to shape and size distribution. Both particle

size and shape are crucial process parameters for many pow-

der-based consolidation methods. For example, in cold spray

deposition, particle shape and size are directly linked to the

impact energies upon deposition [45], which in turn deter-

mine deposition efficiency and bond strength.

In order to obtain more quantitative information on

changes in particle size and size distribution of the MMC

powders upon milling, we analyzed all samples using dy-

namic light scattering. Fig. 2 illustrates the particle size distri-

butions of the composites powders shown in Fig. 1. The

particle size distribution curves (Fig. 2a) were used to deter-

mine the average (mean) particle size and the size range of

powder particles in each sample. The size range, marked as

the distance between D10 and D90, is referring to the 10%

and 90% diameter of the size distribution, respectively

(Fig. 2a). The average particle sizes and size ranges of the

powders shown in Fig. 1 are 10.6, 63.4, 52.3 and 51.1 lm and

8.3, 117.0, 99.1, and 86.5 lm with the order of as-received Al,

1, 4 and 10 h of milling, respectively. These values were deter-

mined by averaging the results of three consecutive measure-

Fig. 1 – SEM micrographs of the (a) as-received Al in comparison

4 h and (d) 10 h using a BPR of 10:1.

ments. The error bar in Fig. 2b reflects the obtained standard

deviation. The size ranges of the composites narrow with

increasing milling time (Fig. 2b), suggesting that longer mill-

ing times yield more uniform powders and lower overall par-

ticle sizes. Thus the size analysis using laser diffraction is in

good agreement with SEM observations.

Fig. 3 shows the effects of ND content and BPR on the par-

ticle size distribution. Both particle size and size range of the

composites were found to increase with increasing BPR, sug-

gesting that smaller BPRs are more effective in reducing the

particle size of the composite powders for the given vial size.

Although higher BPRs generate more collisions between mill-

ing media and powder, the larger number of balls increases

the number of ball-to-ball collisions and reduce the total vol-

ume available for ball acceleration, both of which reduce the

impact energy of the milling media. The larger number of

collisions also results in enhanced sample heating, which

promotes particle agglomeration [46]. In addition, a higher

BPR means that less powder material can be processed per

batch by the MA approach. It was observed that with

increasing ND content, both the particle size and size range

decrease regardless BPR. Similar results were previously re-

ported for other carbon nanostructure-reinforced Al-MMCs

[34,47]. As the ND alters the mechanical properties of the

MMC (particularly hardness and fracture toughness) it af-

fects the milling behavior of the powders. The hardness of

the MMC particles increases with increasing ND content.

The harder particles are more brittle and tend to fracture

easier, leading to smaller particles size for a given milling

time. This effect will be discussed in more detail later. The

mean particle sizes of the MMC powders containing 5 and

10 wt.% ND range from �26 to 60 lm. This is a suitable size

range for thermal spray applications, allowing the as-pro-

duced MMC to be used as feedstock powders, for example,

for cold spray processing.

to ND–Al MMC powders (5 wt.% ND) ball-milled for, (b) 1 h, (c)

Fig. 2 – Particle size analysis of ND–Al MMC (5 wt.% ND) powders ball-milled for 1, 4 and 10 h using a BPR of 10:1, in

comparison to pure as-received Al. The recorded (a) particle size distributions were used to (b) determine important powder

characteristics, such as average particle size and size range.

Fig. 3 – Mean particle size and size range of ND–Al MMC powders ball-milled for 10 h using different BPRs and ND contents.

408 C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5

3.2. Structural integrity and dispersion of ND

In this study, ND–Al MMC powders were synthesized by high-

energy ball-milling, employing stearic acid (3 wt.%) as a pro-

cess control agent (PCA). Different BPRs and milling times

were used to investigate the dispersion of ND in the Al matrix

and its structural integrity during milling.

Raman spectroscopy was utilized to verify the presence of

ND in the composite particles and to evaluate the dispersion

of the reinforcement agent upon milling. Raman spectroscopy

is a powerful tool for the analysis of carbon nanostructures

and carbon nanostructure-containing composites as it can

provide information on both the structural integrity and the

dispersion of the nanostructures [48,49]. Fig. 4 shows the Ra-

man spectra of two MMC powders containing 5 and 10 wt.%

ND, in comparison to Al and as-received ND. Both MMC and

Al powders were milled for 1 h using a BPR of 10:1. The as-re-

ceived ND powder exhibits two distinct Raman features: a

downshifted and asymmetrically broadened diamond peak

around 1325 cm�1, and a broad band between 1500 and

1800 cm�1 often referred to as the ‘‘G band’’ [50,51]. This band

consists of at least three different Raman peaks resulting

from the contributions of sp2 carbon (1590 cm�1), O–H

(�1640 cm�1), and C@O (1740 cm�1) species [52]. The milled

Al sample does not show any significant Raman signal, sug-

gesting that contributions of the stearic acid to the Raman

signal of the MMCs can be neglected.

The Raman spectra of the milled MMCs show several dis-

tinct changes as compared to the as-received ND, indicating

structural changes in the sample upon milling. For both sam-

ples, the diamond Raman peaks are absent, while the G band

is noticeably narrowed and downshifted towards lower wave-

numbers. In addition, two new Raman bands appear around

850 and 1370 cm�1. The latter one is referred to as D band, a

defect-activated, double-resonant Raman feature correspond-

ing to the breathing vibration of the hexagonal rings of gra-

phitic carbon [53]. The Raman peak around 850 cm�1 is

assigned to aluminum carbide (Al4C3) [54]. The observed

changes can be explained as follows, in agreement with pre-

vious studies [55]. During milling, surface atoms are con-

Fig. 4 – UV Raman spectra (325 nm excitation) of ND–Al

MMCs containing 5 and 10 wt.% ND, in comparison to as-

received ND and Al milled with 10:1 BPR for 1 h. Spectra

were normalized with respect to the maximum intensity in

the plotted spectral window, but were recorded using

different measurement times.

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5 409

verted from sp3 to sp2 carbon, leading to the surface graphiti-

zation of the ND crystals. As the hexagonal ring structure of

the sp2-phase develops, the D band appears in the Raman

spectra. At the same time, the sp2-contribution to the G band

increases, resulting in the observed shape changes that ap-

pear as a downshift. The graphitic surface of the ND reacts

Fig. 5 – Raman image of cross-sectioned ND–Al MMC powder pa

and b) and 10 h (c and d). (left: optical microscope image, right: d

range 1100–1800 cm�1).

partially with the surrounding aluminum to form aluminum

carbide. The formation of a thin layer of aluminum carbide

at the ND–Al interface may be beneficial to the overall perfor-

mance of the MMC as it improves the bonding strength be-

tween the reinforcement agent and the metal matrix. Thus,

the Raman spectra recorded from the milled samples do not

provide direct evidence for the existence of the diamond core.

However, previous studies investigating the graphitization

of ND showed that the observed changes in the Raman spec-

tra can be ascribed to surface graphitization and the related

shielding effects rather than a complete sp3-to-sp2 conversion

[55]. It is well known that the Raman scattering cross section

of sp2 carbon is much larger than that of sp3 species [55].

Thus, the diamond core is effectively shielded from Raman

measurements, leading to the disappearance of the diamond

Raman peak in the ND–Al MMC powders, suggesting that the

observed changes are only surface effects and the ND core

maintains its diamond structure.

Since Al metal is not Raman active [59], the Raman fea-

tures of ND can also be used to evaluate the dispersion of

the reinforcement agent in the metal matrix. We therefore re-

corded a Raman map from the cross-section of single MMC

particle and plotted the absolute integrated intensity (area

under curve) between 1100 and 1800 wave numbers, relative

to the Al background (baseline). Because the cross sectional

surface was finely polished, the entire mapping area was in

the focal plane of the objective and intensity changes can

be assigned solely to the content of ND in the Al matrix.

Fig. 5a–d show the cross-sectional optical photograph and

the corresponding integrated intensity Raman maps of a

MMC powder particle (10 wt.% ND) after milling (BPR of 10:1)

for 1 and 10 h, respectively. It can be seen that the recorded

ND Raman intensity distributions match the optical images

recorded through the microscope, suggesting that the ND

rticles containing 10 wt.% ND, milled with 10:1 BPR for 1 (a

istribution of integrated intensity of ND Raman signal in the

410 C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5

agglomerates have been dispersed homogenously within the

Al particles, even after short milling times. However, it should

be noted that since the spatial resolution of the Raman spec-

trometer is limited to about 0.8 lm, statements about ND dis-

persion below this limit cannot be made. These results

demonstrate that high-energy ball milling is an effective

means for size reduction and dispersion of ND agglomerates

in an Al matrix, in analogy to CNT-Al composites [2,3].

In order to identify changes in the crystal size of the Al

matrix and the extend of residual strain after milling, we re-

corded the XRD patterns of all MMC powders. Fig. 6a shows

the diffraction pattern of two MMC samples containing 5

and 10 wt.% ND and pure Al powder after milling (BPR of

30:1) for 4 h. The XRD pattern of as-received ND and as-re-

ceived Al are shown for comparison. While both MMC sam-

ples exhibit the diffraction peaks of Al, the characteristic

(111) peak of ND is not observed. Unlike Raman spectroscopy,

which is highly surface sensitive in solid samples, the X-ray

diffraction signal requires constructive interference from

many atomic layers and is thus proportional to the volume

of the crystals. The intensity of diffraction signals coming

from surface layers and nanocrystals is relatively low and typ-

ically overshadowed by the much stronger XRD signal from

the metal matrix. Therefore, while both ND and Al4C3 were

detected by Raman spectroscopy, they are not observed in

the XRD pattern of the MMCs (Fig. 6b).

A closer inspection of the (111) Al peak of the as-received

and the milled Al powders (Fig. 6c) reveals a commonly ob-

served increase in peak width upon milling, which indicates

a reduction in crystal size and/or presence of residual lattice

strain. The same changes were observed for the milled ND–

Al MMCs.

3.3. Residual strain and crystallite size

The observed changes in the position and width of the Al XRD

peaks can be used to determine changes in crystal size and

residual strain in the metal matrix upon milling. Peak broad-

ening in XRD is caused by three primary factors: (1) reduction

Fig. 6 – X-ray diffraction pattern of ND–Al MMCs containing 5 a

received and milled Al. MMCs and Al were milled for 4 h using

in crystallite size, (2) presence of micro-strain, and (3) instru-

mental broadening [56]. There exist numerous methods that

utilize the observed broadening to quantify both crystal size

and residual strain, the most common being the William-

son–Hall [42] and the Warren–Averbach [57] methods. In this

study we employed the Williamson–Hall method to deter-

mine changes in crystal size and lattice strain upon milling.

The results of the XRD peak analysis for the ND-Al MMC with

10wt% ND are depicted in Fig. 7. For comparison, the crystal

size (and lattice strain) of the as-received powder was above

the measurable size range of the Williamson–Hall analysis,

suggesting crystal sizes of >150 nm.

A continuous reduction in crystal size with increasing

milling time was observed for a BPR of 10:1 BPR. This BPR pro-

vides suitable kinetic energy to break up crystallites and to

deform the lattice due to the high impact velocity of the mill-

ing media. In contrast, in the case of the highest BPR (30:1),

the crystal size increases with increasing milling time. How-

ever, after only 1 h of milling, the crystal size is already

slightly lower than that measured after 10 h of milling using

a BPR of 10:1. Higher BPRs reduce the vial volume available

for acceleration and increase the number of ball-to-ball colli-

sions, both of which lower the effective speed of the grinding

balls and reduce the impact energy [46]. Higher BPRs also lead

to increased friction between grinding balls [58], which can

cause sample heating and heat-induced crystal growth. High

BPRs and long milling times thus undermine the crystal

reduction in the Al matrix. One can therefore picture the mill-

ing process as follows: The reduction of the crystal size to a

certain value requires a certain number of grinding ball-pow-

der collisions. As higher BPRs lead to an increased number of

collisions per time interval, crystal size reductions can be

achieved in shorter milling periods as compared to low BPRs.

Due to increased heating, longer milling times result in crys-

tal subsequent crystal growth. Thus, higher BPR accelerates

the crystal size reduction, but is also subject to increased

sample heating that can cause crystal growth for longer mill-

ing times. The increased temperature during milling with

high BPR may also lead to enhanced agglomeration and car-

nd 10 wt.% ND in comparison to as-received ND, and as-

a BPR of 30:1.

Fig. 7 – Crystal size (a) and residual strain (b) of ND–Al MMC containing 10 wt.% ND.

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5 411

bide formation. For example, while high BPRs and short mill-

ing times lead to small crystal sizes, the mean particle size

and size range of the corresponding samples are much larger

than that of samples produced using low BPRs and long mill-

ing times (see Fig. 3).

The residual strain was not strongly affected by the varia-

tions in milling conditions, but does appear to be more sensi-

tive to the impact velocity rather than the number of

collisions in ball-milling, in spite of more sample heating, as

short milling times (1 h) at a BPR of 30:1 cause less residual

strain than a BPR of 10:1.

3.4. Hardness

The hardness of the MMCs was evaluated by nanoindentation

on the polished cross-section of individual Al-ND MMC pow-

der particles. The highest hardness value (3.46 GPa) was mea-

sured for the ND–Al MMC containing 10 wt.% ND, after milling

for 10 h using a BPR of 10:1, a more than 210% increase over

the pristine, unmilled Al (1.10 GPa) and an almost 100% in-

crease over pure Al milled under the same conditions.

Fig. 8a shows the increases in hardness as a function of ND

Fig. 8 – (a) Hardness of ND–Al MMC with different ND content m

received Al powder). (b) Linear relationship between hardness o

for different milling times.

content for different BPRs (10 h milling time). The smallest

BPR (10:1) was found to yield the highest hardness after 10 h

milling for powders containing 10 wt.% ND. Although high

BPRs (30:1) and short milling times (1 h) were found to yield

similar crystal size reductions, the resulting MMC hardness

(not shown) was only 1.68 GPa, suggesting that low BPRs are

more effective in dispersing ND in Al while maintaining small

crystal sizes in the metal matrix [58]. In order to separate the

effects of the ND inclusions and ball milling on the hardness,

we measured the hardness of the pristine Al before and after

milling. In the case of the lowest BPR (10:1), milling for 10 h

leads to a hardness increase of �67%. The inclusion of

10 wt.% ND results in an additional �150 % increase in hard-

ness, clearly demonstrating the effect of ND reinforcement of

Al and providing further evidence for the survival and disper-

sion of the ND agglomerates in the Al matrix.

The increases in hardness in the present work were com-

parable to those observed in earlier studies on ND–Al MMCs.

Bobrovnitchii et al. [10] reported a hardness of 2.34 GPa after

sintering an Al powder containing 10 wt.% ND. Kaftelen and

Ovecoglu [34] synthesized ND–Al MMC composite powders

containing up to 10 wt.% ND using a combination of ball mill-

illed for 10 h using BPR of 10:1, 20:1, and 30:1 (cross bar: as-

f ND–Al MMC, produced using a BPR of 10:1, and ND content

Fig. 9 – Relationship between measured hardness and metal

matrix crystal size for as-produced ND–Al MMC powders, in

comparison to milled and as-received pure Al.

412 C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5

ing and sintering. The hardness of the powder particle was re-

ported to reach 1.4 GPa at 10 wt.% ND after milling for 1.5 h

using a BPR of 7:1, which is much lower than the values

achieved in this study. Subsequent sintering yielded a hard-

ness of �3.3 GPa. Previous studies revealed a nearly linear

relationship between MMC hardness and ND content [10],

which is in good agreement with our experiments (Fig. 8b).

By combining the measured hardness data with the crystal

size data (Fig. 9), we suggest that both the dispersion

strengthening and the crystal refinement of the metal matrix

control the hardness of the ND–Al MMC composite powders.

In metals, hardness is closely related to yield strength and

can be increased by several factors, including a reduction in

crystal/crystallite size, an increase in the volume fraction of

a non-shearable second phase (dispersion strengthening),

an increase in dislocation density, and the addition of a solid

solution strengthening agent. To first order, the strengthening

mechanism for the ND–Al MMC is dominated by dispersion

strengthening and crystal refinement. Solid solution

strengthening is irrelevant for MMC materials studies in this

Fig. 10 – SEM image of cold sprayed (a) ND–Al MMC (10 wt.% ND)

for 0.5 h using a BPR of 10:1. The inset in (b) shows the as-rece

paper. However, it is difficult to separate the contributions

of dislocation density and crystal size for nanoscale metallic

structures. The influence of crystal size on hardness is quite

strong for all of the ball milled MMC samples, with or without

addition of ND. The hardness of the as-received Al powders is

much lower than that of the milled, pure Al. This difference in

hardness is likely due to the combined effects of crystal

refinement and increased dislocation density that take place

during ball milling. While the hardness increases significantly

as ND is added, the crystallite size stays in the range 30–

60 nm (Fig. 9). This observation strongly suggests that ND is

well dispersed, well below the �1 lm spatial resolution limit

of Raman spectroscopy and is causing the increase in hard-

ness through a dispersion strengthening mechanism. This

also indicates that, unlike in polymer-based composites, ball

milling of MMC effectively breaks up larger ND agglomerates,

present in the as-received ND powders. While the crystal size

of the metal matrix remains the same with increasing milling

time, ND agglomerate size is reduced, resulting in a finer dis-

persion of the reinforcement agent.

Therefore, while a linear relationship between MMC hard-

ness and ND content can be justified within the experimental

error of this study (Fig. 8b), we would like to point out that the

milling behavior of Al-ND composites is more complex and

that the resulting hardness does not necessarily scale linearly

with ND content. The hardness of the composite powder var-

ies for different ND contents and similar milling times will re-

sult in different metal matrix grain sizes and residual strain,

both of which directly affect the hardness of the composite,

in addition to the dispersion strengthening of ND, as dis-

cussed above.

We can conclude that for milled ND–Al MMC powders,

both dispersion strengthening and crystal refinement play

an important role in the strengthening of the material, with

dispersion strengthening becoming more dominant as the

amount of ND increases. Finally, it should be noted that

the formation of third phases, such as aluminum carbide,

at the metal matrix-reinforcement agent interface can sig-

nificantly alter the mechanical properties of the composite.

Further research will be conducted to investigate the forma-

tion of carbide phases and to study their effect on the

mechanical properties of the composite, for example, by

controlling the level of graphitization of the ND surface prior

to milling.

and (b) pure Al. In both cases, precursor powders were milled

ived ND particles.

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5 413

3.5. Young’s modulus

We also measured the Young’s modulus (E) of the ND–Al com-

posites using nanoindentation. However, it should be noted

that while E shows a relative increase with increasing dia-

mond content, the absolute values for E of the MMC powders

are not reliable, given the experimental conditions of the

measurement (powder dispersed in epoxy matrix). Therefore,

only relative increases are given. In this study, it was found

that the addition of 5 and 10 wt.% ND resulted in a �7 and

�45% increase in Young’s modulus, respectively, relative to

the value of the pure Al powder milled under the same condi-

tions (3 h, BPR of 10:1).

3.6. Cold spray deposition

In order to demonstrate the suitability of the produced ND–Al

MMC powders as feed-stock for low-temperature consolida-

tion methods, we cold spray deposited two ND–Al MMC pow-

ders (10 wt.% ND) that were milled (BPR of 10:1) for 0.5 and

3 h, respectively. A milling time of 0.5 h was selected because

the cold spray deposition efficiency was significantly higher

as compared to longer milling times (>1 h). Since both Raman

imaging and hardness measurements revealed good ND dis-

persion after 1 h of ball milling, we decided to further reduce

the milling time. Fig. 10 shows the SEM image of a cold

sprayed ND–Al MMC (0.5 h milling time) deposit in compari-

son to cold sprayed pure Al. The composite film shows ND

agglomerates in the size range 30–200 nm, well-dispersed in

the Al matrix. The average size of ND agglomerate in the

cold-sprayed composite measures 75–100 nm, which is simi-

lar to the value of the as-received ND powder (inset in

Fig. 10b). The estimated average inter-particle distance is

around 100–200 nm. While this data was obtained from the

composite film, the ND dispersion is believed to be similar

in the as-produced MMC powder as cold spray only consoli-

dates the powder particles (average size �28 lm). Moreover,

since good ND dispersion was achieved for relatively short

milling times (�0.5 h), samples milled for longer periods are

expected to exhibit at least similar dispersion levels.

The highest hardness, (3.26 ± 0.25) GPa, was found for ND–Al

MMC (10 wt.% ND) after milling the precursor powder for 3 h, as

expected. For comparison, the hardness of cold sprayed pure Al

was found as (1.47 ± 0.34) and (2.41 ± 0.16) GPa for as-received

and ball-milled powders (Fig. 10b), respectively. The corre-

sponding values for the Young’s modulus were measured as

(92.3 ± 7.1), (72.9 ± 4.2), and (74.2 ± 6.7) GPa for the ND–Al MMC,

the as-received Al, and the ball-milled Al powders, respectively.

While a detailed analysis of the relationship between MMC

powder characteristics and cold spray performance is outside

the scope of this study, the results clearly demonstrate that

ball milling is a powerful method for the synthesis of MMC

powders and their application as feed-stock for cold spray

consolidation.

4. Conclusions

We have successfully synthesized a series of nanodiamond-

reinforced aluminum metal matrix composite (ND–Al MMC)

powders using high energy ball milling, and have presented

a systematic study of the effects of the various milling param-

eters and the ND content on the structural characteristics and

mechanical properties of the as-produced MMC powders. The

results presented in this study demonstrate that high energy

ball milling is a feasible means for the synthesis of ND–Al

MMC powders as it provides excellent control over (1) particle

size distribution, (2) crystal size and residual strain of the me-

tal matrix, and (3) structural integrity and dispersion of the

ND reinforcement agent.

It was found that while higher BPRs accelerate the crystal-

lite size reduction for short milling times, during milling, they

lead to crystallite growth for longer milling, likely due to in-

creased sample heating. Moreover, higher BPRs are less effec-

tive in ND dispersion and particle size reduction of ND–Al

composites, rendering low BPRs and longer milling times

more suitable for MMC synthesis. Using a BPR of 10:1 and a

ND content of 10 wt.%, the hardness of the Al MMC increased

by �210%, reaching 3.46 GPa after 10 h of milling, the highest

ever reported value for ND–Al MMC with similar ND content.

The increased hardness is caused primarily by the ND rein-

forcement, accounting for approximately three-quarters of

the total observed enhancement. Raman spectroscopy re-

vealed a homogeneous dispersion of ND in the individual

MMC powder particles, at a maximum resolution of 0.8 lm,

and the formation of a thin Al4C3 film at the ND surface,

which may serve as an interfacial bonding layer between

the reinforcement agent and the metal matrix.

Finally, we employed cold spray deposition to consolidate

the as-produced MMC powders and demonstrated their feasi-

bility as feed-stock material for thermal spray applications.

The produced MMC coatings showed homogenous dispersion

of ND agglomerates, yielding hardness values exceeding 3 GPa.

Acknowledgments

The authors would like to thank Dr. Sarath Menon (NPS) for

his assistance with SEM analysis and Mr. Douglas Seivwright

(NPS) for his help with cold spray deposition. This work was

supported by the Office of Naval Research.

R E F E R E N C E S

[1] Kuzumaki T, Miyazawa K, Ichinose H, Ito K. Processing ofcarbon nanotube reinforced aluminum composite. J MaterRes 1998;13(9):2445–9.

[2] Esawi AMK, Morsi K. Dispersion of carbon nanotubes inaluminum powder. Composites: Part A 2007;38:646–50.

[3] Esawi AMK, Morsi K, Sayed A, Gawad AA, Borah P. Fabricationand properties of dispersed carbon nanotube–aluminumcomposites. Mater Sci Eng A 2009;508(1–2):167–73.

[4] Wang L, Choi H, Myoung J-M, Lee W. Mechanical alloying ofmulti-walled carbon nanotubes and aluminium powders forthe preparation of carbon/metal composites. Carbon2009;47(15):3427–33.

[5] Sridhar I, Narayanan KR. Processing and characterization ofMWCNT reinforced aluminum matrix composites. J Mater Sci2009;44:1750–6.

[6] Choi H, Shin J, Min B, Park J. Reinforcing effects of carbonnanotubes in structural aluminum matrix nanocomposites. JMater Res 2009;24(8).

414 C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5

[7] Myers M, Chawla K. Grain boundaries in plastic deformation(Grain-size strengthening). In: Mechanical behavior ofmaterials. Cambridge UK: Cambridge Univ Press; 2009. p.345–8.

[8] Baker AA, Braddick DM, Jackson PW. Fatigue of boron–aluminum and carbon–aluminum fiber composites. J MaterSci 1972;7(7):747.

[9] Lamotte ED, Phillips K, Perry AJ, Killias HR. Continuously castaluminum–carbon fiber composites and their tensileproperties. J Mater Sci 1972;7(3):346–8.

[10] Bobrovnitchii GS, Skury ALD, Monteiro SN, Tardim RC. Effectof nanodiamond addition on the mechanical properties ofpolycrystalline metallic and polymeric composites. Mater SciForum 2010;660–661:848–53.

[11] Benjamin JS, editor. New materials by mechanical alloyingtechniques. Oberursel, Germany: DGMInformationgeselschaft; 1989.

[12] Kwon H, Estili M, Takagi K, Miyazaki T, Kawasaki A.Combination of hot extrusion and spark plasma sintering forproducing carbon nanotube reinforced aluminum matrixcomposites. Carbon 2009;47(3):570–7.

[13] Esawi AMK, El Borady MA. Carbon nanotube-reinforcedaluminium strips. Compos Sci Technol 2008;68(2):486–92.

[14] Zhou S, Zhang X, Ding Z, Min C, Xu G, Zhu W. Fabrication andtribological properties of carbon nanotubes reinforced Alcomposites prepared by pressureless infiltration technique.Compos Part A: Appl Sci Manuf 2007;38(2):301–6.

[15] Laha T, Chen Y, Lahiri D, Agarwal A. Tensile properties ofcarbon nanotube reinforced aluminum nanocompositefabricated by plasma spray forming. Compos Part A: Appl SciManuf 2009;40(5):589–94.

[16] Stein J, Lenczowski B, Frety N, Anglaret E. Mechanicalreinforcement of a high-performance aluminium alloyAA5083 with homogeneously dispersed multi-walled carbonnanotubes. Carbon 2012;50(6):2264–72.

[17] George R, Kashyap K, Rahul R, Yamdagni S. Strengthening incarbon nanotube/aluminium (CNT/Al) composites. ScriptaMater 2005;53(10):1159–63.

[18] Bartolucci SF, Paras J, Rafiee MA, Rafiee J, Lee S, Kapoor D,et al. Graphene–aluminum nanocomposites. Mater Sci Eng A2011;528(27):7933–7.

[19] Bakshi SR, Lahiri D, Agarwal A. Carbon nanotube reinforcedmetal matrix composites – a review. Int Mater Rev2010;55(1):41–64.

[20] Mochalin VN, Shenderova O, Ho D, Gogotsi Y. The propertiesand applications of nanodiamonds. Nat Nanotechnol2011;7(1):11–23.

[21] Maitra U, Prasad KE, Ramamurty U, Rao CNR. Mechanicalproperties of nanodiamond-reinforced polymer-matrixcomposites. Solid State Commun 2009;149(39–40):1693–7.

[22] Shenderova O, McGuire G. Nanocrystalline diamond. In:Gogotsi Y, editor. Nanomaterials handbook. Boca Raton,FL: CRC Press; 2006. p. 203–37.

[23] Mochalin VN, Gogotsi Y. Wet chemistry route to hydrophobicblue fluorescent nanodiamond. J Am Chem Soc2009;131(13):4594–5.

[24] Maze JR, Stanwix PL, Hodges JS, Hong S, Taylor JM, CappellaroP, et al. Nanoscale magnetic sensing with an individualelectronic spin in diamond. Nature 2008;455(7213):644–7.

[25] Zhang QW, Mochalin VN, Neitzel I, Knoke IY, Han JJ, Klug CA,et al. Fluorescent PLLA-nanodiamond composites for bonetissue engineering. Biomaterials 2011;32(1):87–94.

[26] Ho D. Nanodiamonds: applications in biology and nanoscalemedicine. New York: Springer; 2009.

[27] Davidson JL, Bradshaw DT. Compositions with nano-particlesize diamond powder and methods of using same fortransferring heat between a heat source and a heat sink. USpatent 6858157B2, 2005.

[28] Red’kin V. Lubricants with ultradisperse diamond–graphitepowder. Chem Technol Fuels Oils 2004;40(3):164–70.

[29] Dolmatov VY. Detonation synthesis ultradisperseddiamonds: properties and applications. Russ Chem Rev2001;70(7):607.

[30] Behler KD, Stravato A, Mochalin V, Korneva G, Yushin G,Gogotsi Y. Nanodiamond-polymer composite fibers andcoatings. ACS Nano 2009;3(2):363–9.

[31] Wang DH, Tan LS, Huang H, Dai L, Osawa E. In-situnanocomposite synthesis: arylcarbonylation and grafting ofprimary diamond nanoparticles with a poly (ether–ketone) inpolyphosphoric acid. Macromolecules 2008;42(1):114–24.

[32] Cheng JL, He JP, Li CX, Yang YL. Facile approach tofunctionalize nanodiamond particles with V-shaped polymerbrushes. Chem Mater 2008;20(13):4224–30.

[33] Mochalin VN, Neitzel I, Etzold BJM, Peterson A, Palmese G,Gogotsi Y. Covalent incorporation of aminated nanodiamondinto an epoxy polymer network. ACS Nano2011;5(9):7494–502.

[34] Kaftelen H, Ovecoglu M. Microstructural characterization andwear properties of ultra-dispersed nanodiamond (UDD)reinforced Al matrix composites fabricated by ball-millingand sintering. J Compos Mater 2012;46(13):1521–34.

[35] Livramento V, Correia JB, Shohoji N, �Osawa E. Nanodiamondas an effective reinforcing component for nano-copper. DiamRelat Mater 2007;16(2):202–4.

[36] Filho SL, Rodriguez RM, Earthman JC, Lavernia EJ, editors.Synthesis of diamond reinforced Al–Mg nanocrystallinecomposite powder using ball milling. Materials scienceforum. Trans Tech Publ.; 2003.

[37] Suryanarayana C. Non-equilibrium processing ofmaterials. Pergamon; 1999.

[38] Lavernia E, Han B, Schoenung J. Cryomilled nanostructuredmaterials: processing and properties. Mater Sci Eng A2008;493(1):207–14.

[39] Han B, Ye J, Tang F, Schoenung J, Lavernia E. Processing andbehavior of nanostructured metallic alloys and compositesby cryomilling. J Mater Sci 2007;42(5):1660–72.

[40] Wu J, Fang H, Yoon S, Kim HJ, Lee C. Measurement of particlevelocity and characterization of deposition in aluminumalloy kinetic spraying process. Appl Surf Sci2005;252(5):1368–77.

[41] De Castro CL, Mitchell BS. Nanoparticles from mechanicalattrition. In: Baraton MI, editor. Synthesis functionalizationand surface treatment of nanoparticles. Valencia,CA: American Scientific Publishers; 2002. p. 1–15.

[42] Suryanarayana C, Norton MG. X-ray diffraction: a practicalapproach. New York: Plenum; 1998.

[43] Poirier D, Gauvin R, Drew RAL. Structural characterization ofa mechanically milled carbon nanotube/aluminum mixture.Compos Part A: Appl Sci Manuf 2009;40(9):1482–9.

[44] Buchheit TE, Vogler TJ. Measurement of ceramic powdersusing instrumented indentation and correlation with theirdynamic response. Mech Mater 2010;42(6):599–614.

[45] Jodoin B, Ajdelsztajn L, Sansoucy E, Zuniga A, Richer P,Lavernia EJ. Effect of particle size, morphology, and hardnesson cold gas dynamic sprayed aluminum alloy coatings. SurfCoat Technol 2006;201(6):3422–9.

[46] Magini M, Iasonna A. Energy transfer in mechanical alloying:overview. Mater Trans 1995;36(2):123–33.

[47] Morsi K, Esawi A. Effect of mechanical alloying time andcarbon nanotube (CNT) content on the evolution ofaluminum (Al)–CNT composite powders. J Mater Sci2007;42(13):4954–9.

[48] Gupta ML, Sydlik SA, Schnorr JM, Woo DJ, Osswald S, SwagerTM, et al. The effect of mixing methods on the dispersion ofcarbon nanotubes during the solvent-free processing of

C A R B O N 6 3 ( 2 0 1 3 ) 4 0 4 – 4 1 5 415

multiwalled carbon nanotube/epoxy composites. J Polym SciPart B: Polym Phys 2013;51(6):410–20.

[49] Mu MF, Osswald S, Gogotsi Y, Winey KI. An in situ Ramanspectroscopy study of stress transfer between carbonnanotubes and polymer. Nanotechnology 2009;20(33):335703.

[50] Ferrari AC, Robertson J. Raman spectroscopy of amorphous,nanostructured, diamond-like carbon, and nanodiamond.Philos Trans R Soc Lond Ser A: Math Phys Eng Sci2004;362(1824):2477–512.

[51] Osswald S, Mochalin VN, Havel M, Yushin G, Gogotsi Y.Phonon confinement effects in the Raman spectrum ofnanodiamond. Phys Rev B 2009;80(7):075419.

[52] Mochalin V, Osswald S, Gogotsi Y. Contribution of functionalgroups to the Raman spectrum of nanodiamond powders.Chem Mater 2009;21(2):273–9.

[53] Reich S, Thomsen C. Raman spectroscopy of graphite. PhilosTrans R Soc Lond Ser A: Math Phys Eng Sci2004;362(1824):2271–88.

[54] Zameshin A, Popov M, Medvedev V, Perfilov S, Lomakin R,Buga S, et al. Electrical conductivity of nanostructured andC60-modified aluminum. Appl Phys A 2012;107(4):863–9.

[55] Cebik J, McDonough J, Peerally F, Medrano R, Neitzel I,Gogotsi Y, et al. Raman Spectroscopy of the Transformationof Detonation Nanodiamond to Carbon Onions during LowTemperature Annealing. Nanotechnology2013;24(20):205703.

[56] He HH. Two-dimensional X-ray diffraction. Hoboken,NJ: Wiley Publishing; 2011.

[57] Warren B, Averbach B. The Effect of cold-work distortion onX-Ray patterns. J Appl Phys 1950;21(6):595–9.

[58] Suryanarayana C. Mechanical alloying and milling. ProgMater Sci 2001;46:1–184.

[59] Nafie LA. Theory of Raman Scattering. In: Lewis IR, EdwardsHGM, editors. Handbook of Raman spectroscopy. CRCPress: New York; 2001. p. 1–10.