Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
Scripta Materialia 186 (2020) 387–391
Contents lists available at ScienceDirect
Scripta Materialia
journal homepage: www.elsevier.com/locate/scriptamat
Direct co-deposition of mono-sized nanoparticles during sputtering
Mikhail N. Polyakov, Rachel L. Schoeppner , Laszlo Pethö, Thomas E.J. Edwards
∗,
Keith Thomas, Bence Könny
˝
u, Xavier Maeder, Johann Michler
Empa - Swiss Federal Laboratories for Materials Science and Technology, Feuerwerkerstrasse 39, 3602 Thun, Switzerland
a r t i c l e i n f o
Article history:
Received 4 March 2020
Revised 13 May 2020
Accepted 13 May 2020
Keywo rds:
Nanocomposites
Nanoparticles
Synthesis
Thermal stability
Sputtering
a b s t r a c t
Nanoparticle-reinforced thin films synthesis is often limited by precipitation thermodynamics and kinet-
ics to certain material combinations and nanoparticle size distributions. We demonstrate a new method
that allows direct co-deposition of mono-sized nanoparticles with various matrix materials, independent
particle size and composition control, yielding flexible material selection, and nanoparticle density spatial
variations laterally and depth-wise. Tungsten nanoparticles were co-deposited into magnetron-sputtered
copper, giving a uniform distribution of nanoparticles in transmission electron microscopy. To demon-
strate the application potential, W nanoparticles stabilized nanograined Cu at 500 °C; pure Cu reference
showed significant grain growth. This method opens new possibilities in tailored nanocomposite fabrica-
tion.
©2020 Acta Materialia Inc. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license.
( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
The properties of a material can be improved by altering the
structure at the nanoscale in many different ways. For example, the
microstructure can be modified either during or after fabrication
by means of grain refinement or the incorporation of nanotwins,
which will generally increase the strength of the material [ 1 , 2 ]. A
nanoscale architecture can be introduced into the sample structure
to modify the density, surface area, and mechanical behavior of a
material [ 3–5 ]. One or more additional phases can also be formed
in the structure, often with the goal of increasing the strength or
thermal stability of the material [ 6 , 7 ].
Nanoparticles are commonly incorporated into bulk materials
as an alternate phase. When nanoparticles are freestanding, their
nanoscale dimensions, surfaces, and coatings can be tuned to pro-
duce vivid colors, strong sensing capabilities, magnetic effects, and
other attractive material properties [ 8–15 ]. Their nanoscale fea-
tures cause them to exhibit physical and chemical properties which
are very different from their bulk counterparts, spurring much re-
search into their fabrication and utilization. When nanoparticles
are incorporated into a matrix, they can enhance the strength,
thermal stability, and optical properties of the composite material;
thus, a variety of methods have been developed for incorporating
nanoparticles into a matrix material [ 16–19 ] One common method
is supersaturation of a material with another element and subse-
∗Corresponding author.
E-mail address: thomas.edwards@empa.ch (T.E.J. Edwards).
quent heat treatments to promote the formation of precipitates, of-
ten used in nickel superalloys and age-hardened aluminum alloys.
Such an approach requires that the materials have significant mis-
cibility at reasonably low temperatures and that the precipitates
form with the desired shapes and size distributions.
To avoid the requirement of miscibility, the materials can also
be deposited in a non-equilibrium state (e.g. by physical vapor
deposition or ion implantation) prior to annealing [ 20–23 ]. How-
ever, precipitate shapes and spatial distributions are not easily con-
trolled by annealing [ 24 ]. Other methods consist of incorporat-
ing existing nanoparticles or agglomerated metal ions into a ma-
trix, using methods such as sol-gel suspension, CVD, etc [ 25–33 ].
However, such methods are limited in nanoparticle/matrix mate-
rial selection, due to diffusion caused by the elevated temperatures
of deposition, chemical reactivity considerations, and the limited
availability of deposition routes for certain materials [ 12 , 25 , 32 , 34 ].
In addition, it is often difficult to control the nanoparticle sizes in-
dependently of other deposition parameters and avoid agglomera-
tion of the nanoparticles.
In the present study, a novel method for incorporating nanopar-
ticles in a controlled manner into an arbitrary matrix material is
demonstrated. Here, the Cu matrix was deposited by magnetron
sputtering, while W nanoparticles were synthesized using termi-
nated gas condensation and subsequently ejected onto the sub-
strate during the concurrent Cu deposition. Terminated gas con-
densation has previously been used to produce single element (and
occasionally two element) metal and semiconductor nanoparticles
https://doi.org/10.1016/j.scriptamat.2020.05.032
1359-6462/© 2020 Acta Materialia Inc. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license.
( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
388 M.N. Polyakov, R.L. Schoeppner and L. Pethö et al. / Scripta Materialia 186 (2020) 387–391
Fig. 1. Co-deposition schematic. The nanoparticle source and sputtering source de-
posit simultaneously onto the substrate, allowing for tailored film architectures.
by DC or radio frequency plasma sputtering followed by extraction
of grown clusters through a differential pumping chamber; source
parameter optimization has recently been extensively studied [ 35–
39 ]. The W nanoparticles here were uniformly distributed into the
Cu matrix and the sample demonstrated significantly improved re-
sistance to grain growth compared to a pure Cu film deposited
under identical conditions. With this deposition method, the ma-
trix and nanoparticle material selection is free from many of the
phase diagram and thermodynamic considerations that can domi-
nate for precipitation-formed particles. In addition, the size of the
nanoparticles can be controlled by mass selection using a setup
akin to a mass spectrometer, which is critical for optical applica-
tions [ 9 , 11 , 17 , 27 , 40 ]. Thus, this study demonstrates a new level of
control that is now available for the fabrication of nanocomposites
and control of their properties.
The samples were deposited using a custom QPREP 500 PVD
chamber from Mantis Deposition Ltd, allowing co-deposition of
nanoparticles during physical vapor deposition; a schematic of the
deposition chamber is shown in Fig. 1 . Two types of films were
deposited: pure Cu reference films and a second with incorporated
W nanoparticles. Sample films were deposited onto carbon film
Transmission Electron Microscopy (TEM) grids and Si (100) wafers.
Copper was magnetron sputtered from a 76 mm high purity target
using a HiPIMS power source (HiPSTER 1, Ionautics AB); tuning the
pulse properties allows for unique grain-size control and obtaining
a non-porous copper layer. The developed process parameter set
keeps twin formation to a minimum. In addition, the highly ener-
gized sputtered copper adatoms have a higher mobility upon land-
ing onto the substrate and enhance the embedding of the tungsten
particles into the matrix. The average driving current was set to
25 mA and the plasma power to 20 W. The pulse length was 40
μs, with a 300 Hz repetition rate. This reached a peak current of
2 A and a pulse charge of 140 μC. An 8 standard cubic centimeter
per minute (sccm) argon flow was regulated to produce a 1.7 e-3
mbar process pressure using a throttle valve.
W nanoparticles were produced by a 51 mm magnetron head
driven by a DC power source. The nanoparticles are formed in
an aggregation zone and then a pressure difference extracts the
nanoparticles through the exit aperture [ 41 ]. To successfully incor-
porate the tungsten nanoparticles into the matrix, the pressure set-
points of various regions of the vacuum system were optimized.
These included the nanoparticle aggregation zone, the substrate
zone, the magnetron sputtering of the matrix material, and the tra-
jectory of the particles starting from the aggregator exit aperture.
The aperture width was matched so that the pressure difference
at the exit of the aggregation zone enabled both the extraction of
the particles towards the substrate and maintained the terminated
gas condensation mechanism, resulting in the desired particle di-
ameters. The current was set to 250 mA, resulting in an average
plasma power of 75 W. The magnetron was inserted 80 mm deep
into the nanoparticle aggregation zone and maintained at 20 °C
by external cooling. An argon flow of 65 sccm builds 2e-1 mbar
pressure within the aggregation zone. The exiting argon flow cre-
ates 1e-3 mbar pressure at the aperture, and a residual 5e-4 mbar
pressure within the main chamber. This background pressure due
to the nanoparticle source operation is compensated by regulat-
ing the main chamber pressure with a throttle valve and does not
significantly influence the magnetron sputtering of the matrix ma-
terial.
During deposition, the extracted nanoparticle beam was also
driven through a quadrupole mass spectrometer (NanoGen 50,
Mantis Deposition Ltd.) to measure the mass distribution of the
nanoparticle beam, which is then converted to a nanoparticle di-
ameter distribution. The mass spectrometer also provides the op-
tion of nanoparticles size-filtering, which is a limitation for many
other techniques, wherein the nanoparticle size is generally cou-
pled with another parameter such as volume fraction [ 31 ]. Some
possible film architectures which can be fabricated with such a
system are diagrammatized in Fig. 1 and illustrated in the supple-
mentary material Figure S1.
The nanocomposite films deposited onto carbon film grids were
cleaned in an oxygen plasma cleaner (PlasmaPrep2, GaLa instru-
mente) and subsequently annealed in vacuum (starting vacuum:
< 2.5e-5 mbar), first to 300 °C for 3 h and then to 500 °C for
2 h, using a Carbolite STF 16/450 tube furnace. The films deposited
onto carbon film grids were directly imaged in plan-view by TEM
(JEM220 0fs, JEOL) at 20 0 kV; additionally, TEM cross-sections of an
as-deposited Cu film with W nanoparticles on a Si substrate and
an annealed Cu film with W nanoparticles on a carbon film TEM
grid were prepared using a Tescan-Vela focused ion beam micro-
scope and subsequently imaged by TEM. The cross-sectional TEM
sample of the annealed film was taken from the carbon film TEM
grid sample rather than the film on the Si substrate, because sili-
cides formed upon annealing of Cu on Si. Chemical analysis was
performed by STEM-EDX (Titan Themis, FEI) at 300 kV using a Su-
perEDX detector system.
Bright-field scanning TEM (BF-STEM) and Dark-field STEM (DF-
STEM) images of the as-deposited samples with and without W
nanoparticles are shown in Fig. 2 . Fig. 2 a–c show the sample with
W nanoparticles, with the nanoparticles generally appearing darker
than the Cu in BF-STEM images and brighter in DF-STEM images
through atomic contrast. Further EDX-based chemical analysis is
reported in the supplementary material Figure S2. The nanoparticle
density by volume, n , was determined by counting more than 600
nanoparticles in DF-STEM images; particle spacing measurement
assumed an ideal gas law distribution, where the center-to-center
spacing is approximated by 0 . 893 (
3
4 πn
)
1 / 3
. The average nanoparti-
cle spacing was determined to be 9.5 nm and the nanoparticles
were measured to have an average diameter of 4.0 nm, with a
standard deviation of 0.7 nm. This particle size matches that mea-
sured by the mass spectrometer during deposition, where a 4.0 nm
average diameter was measured, with a standard deviation of
1.3 nm.
The samples did not show agglomeration of nanoparticles. In
addition, in Fig. 2 c, the nanoparticles are scattered throughout the
Cu grains, not just at grain boundaries. Therefore, the nanoparti-
cles do not cause renucleation events; rather, the Cu grains appear
to be simply growing around the W nanoparticles. This is in con-
trast to the renucleation due to nanoparticles that is sometimes
observed for other deposition methods, such as Pt and Au nanopar-
ticles in CVD ZnO [ 42 ].
A plan view of as-deposited pure Cu is shown in Fig. 2 d. The
grain shapes are qualitatively similar for the samples with and
without W nanoparticles, and the average Cu grain diameters are
28 ±11 nm for the sample with W nanoparticles and 25 ±11 nm
for the sample without W nanoparticles. Grain sizes were deter-
M.N. Polyakov, R.L. Schoeppner and L. Pethö et al. / Scripta Materialia 186 (2020) 387–391 389
Fig. 2. TEM images of as-deposited Cu films with and without W nanoparticles (np). a) BF-STEM plan-view image showing the grain structure of the Cu and the distribution
of W nanoparticles (dark dots). b) DF-STEM image of the same area as a); here W nanoparticles appear as bright dots. c) Cross-sectional BF-STEM image of a Cu film with W
nanoparticles. d) BF-STEM plan-view image of a pure Cu film. For both compositions, the average Cu grain size is similar at 28 and 25 nm, with and without W, respectively.
All scale bars are 50 nm.
mined from TEM images and are “column diameter” values, since
the Cu grains are columnar rather than spherical. Overall, the two
samples start with similar microstructures and grain sizes.
TEM images of the films after annealing to 500 °C in vacuum
for 2 h are shown in Fig. 3 . The grain structure of the annealed
sample with W nanoparticles remains similar to that before an-
nealing. The Cu grain diameter increased slightly from 28 nm to
36 ±13 nm after annealing, and the W nanoparticles are still dis-
persed: many particles remain at the grain interiors rather than
the grain boundaries. This retention of nanoparticle dispersion is
similarly seen in the cross-section images, Fig. 3 d.
In contrast, there was significant grain growth for the sample
without W nanoparticles, as in Fig. 3 e. The average grain diame-
ter increased from 25 nm to 118 ±54 nm after annealing. This
stark difference in grain growth is evident in the DF-TEM images
in Fig. 3 c and f (with and without W nanoparticles), noting the
scale-bar change. In Fig. 3 e, twin boundaries are evident; these are
also present in the as-sputtered Cu grains for both samples, but are
less evident as the as-sputtered grains are smaller.
The overall W composition of the sample, based on an aver-
age particle diameter of 4 nm spaced at 9.5 nm, is ~0.75 vol.%.
Thus, this small volume of nanoparticles was sufficient to signifi-
cantly inhibit the grain growth of the nanocrystalline Cu, even at
0.57 homologous temperature for 2 h. In fact, one may consider
the pinning potential of the W nanoparticles with radius r on the
growth of Cu grains, radius R
c
, according to the generalized Zener
drag equation
43
:
R
c
=
Kr
f
m
where the prefactor K , and the exponent m of the particle volume
fraction, f , have been determined experimentally for a range of in-
soluable and precipitate particles in engineering alloys to be best
valued at 0.17 and 1, respectively, for f < 0.05, as here. For 4 nm
diameter W particles at 0.7 vol.% this yields Cu grains of 91 nm
diameter –somewhat larger than the measured values following
heat treatment, but nevertheless lower than the Cu grain size in
the absence of W nanoparticles. In fact, further investigations into
Zener drag [38] have shown that, as here, additional grain refine-
ment occurs when the particles are stable, and the initial grain size
is both refined and homogeneous.
390 M.N. Polyakov, R.L. Schoeppner and L. Pethö et al. / Scripta Materialia 186 (2020) 387–391
Fig. 3. TEM images of annealed (500 °C, 2 h) Cu films with W nanoparticles (plan view - a, b, c; cross-sectional view - d) and without W nanoparticles (plan view - e, f).
Note the change in scale-bar in e and f: the grain growth is considerable for the
pure Cu sample.
The stability of the nanoparticles results from two important
considerations: W is immiscible in Cu, and W has a high melt-
ing point (3695 K vs. 135 8 K for Cu). These two properties are
of note for demonstrating the flexibility of this method of co-
deposition. To alternatively have W distributed in Cu would re-
quire co-sputtering of both materials (since they are immiscible),
as has been previously performed for a variety of compositions
[23] . However, subsequent annealing of such co-sputtered mate-
rial may not result in the desired structure, since annealing affects
both phases and independent control of the two structural com-
ponents (matrix and nanoparticles) is hence lost, i.e. Cu grains can
grow before the W nanoparticles are completely formed by seg-
regation. On the other hand, direct deposition of the W nanopar-
ticles into the Cu as performed here allows for a controlled and
thermally-stable as-deposited structure.
The small amount of W that was required to effect this in-
creased thermal stability also holds promise for producing mate-
rials with certain significantly improved properties (thermal sta-
bility, optical properties, strength), while causing minimal alter-
ations to other properties (electrical and thermal conductivity, den-
sity, corrosion resistance), as the majority phase remains unalloyed.
Further control of the nanoparticle distribution can also be im-
plemented by having nanoparticles embedded both randomly and
in layers, with lateral and depth gradients in particle density also
possible. This allows for additional tuning of the material proper-
ties, see diagram, Fig. 1 , and micrographs, Figure S1.
The flexibility of the deposition method also lends itself favor-
ably to the possibility of creating model systems to verify funda-
mental theories in Materials Science, such as Zener drag evoked
here, but also particle strengthening in materials mechanics, and
the kinetics of elemental interdiffusion. Indeed, an idealized start-
ing state for diffusion studies can be generated with this method:
a mechanical mixture with a uniform 3D distribution of point
sources of pure elements in a pure matrix. Similarly in the case
of the Zener theory [43] , the core assumptions are fulfilled by the
present co-deposited system: the W particles are reasonably ap-
proximated as spherical, although the Cu grains are themselves
columnar, which facilitates use of a 2D model, the particles are in-
coherent in the matrix and are uniformly distributed, and finally
the particle sizes are narrowly distributed and are small enough
relative to the initial grain size to interact with only one grain
boundary at a time.
Therefore, a method for co-deposition of nanoparticles into
a matrix has been demonstrated that has the same material-
selection versatility, which is provided by conventional mag-
netron sputtering. This technique avoids the thermodynamic con-
siderations, which are inherent to most other nanocomposite
co-deposition methods, resulting in many new possible matrix-
nanoparticle material combinations. The nanoparticles can be ran-
domly distributed throughout the material, as was done here, or
the spatial distribution can be controlled by varying the deposi-
tion parameters and interrupting the nanoparticle or matrix depo-
sition as needed. As was demonstrated for Cu in this study, ther-
mal stability could be immediately improved using such a deposi-
tion method. This technique also holds promise for producing ma-
terials with tuned optical properties, where nanoparticle sizes and
distributions need to be well controlled. Finally, this technique al-
lows for the fabrication of tailored microstructures for fundamental
investigations of nanocomposite behavior during mechanical defor-
mation and in other conditions. Overall, this deposition method
M.N. Polyakov, R.L. Schoeppner and L. Pethö et al. / Scripta Materialia 186 (2020) 387–391 391
lends significant flexibility and control to the field of nanocompos-
ite fabrication.
Declaration of Competing Interests
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared to
influence the work reported in this paper.
Acknowledgements
T.E.J.E. acknowledges the EMPAPOSTDOCS-II program which has
received funding from the European Union’s Horizon 2020 research
and innovation program under the Marie Skłodowska-Curie grant
agreement number 754364.
Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at doi:10.1016/j.scriptamat.2020.05.
032 .
References
[1] L. Lu , Y. Shen , X. Chen , L. Qian , K. Lu , Science 304 (2004) 422–426 .
[2] S. Zheng , I.J. Beyerlein , J.S. Carpenter , K. Kang , J. Wang , W. Han , N.A. Mara , Nat.
Commun. 4 (2013) 1696–1698 .
[3] T.A . Schaedler , A .J. Jacobsen , A . Tor ren ts , A .E. Sorensen , J. Lian , J.R. Greer ,
L. Valdevit , W.B . Carter , Q. Ge , J.A. Jackson , S.O. Ku cheyev , N.X. Fang ,
C.M. Spadaccini , Science 334 (2011) 962–965 .
[4] L.R. Meza
, S. Das , J.R. Greer , Science 345 (2014) 1322–1326 .
[5] J. Erlebacher , M.J. Aziz , A. Karma , N. Dimitrov , K. Sieradzki , Nature 410 (2001)
450–453 .
[6] T. Chookajorn , H.a Murdoch , C.a Schuh , C.A.S. To ngj ai Chookajorn , Heather
A. Murdoch , Science (80-.). 337 (2012) 951–954 .
[7] T.J. Rupert , Curr. Opin. Solid State Mater. Sci. 20 (2016) 257–267 .
[8] G. Walters , I.P. Parkin , J. Mater. Chem. 19 (2009) 574–590 .
[9] H.A. Atwater , A. Polman , Nat. Mater. 9 (2010) 205–213 .
[10] M. Kerk er ,
Scatt. Light Other Electromagn. Radiat. 16 (1969) 27–96 .
[11] J.A. Scholl , A.L. Koh , J.A. Dionne , Nature 483 (2012) 421–427 .
[12] J.N. Anker , W.P. Hall , O. Lyandre s , N.C. Shah , J. Zhao , R.P. Van Duyne , Nat. Mater.
7 (2008) 442–453 .
[13] Y. Qiang , Y. Qiang , R.F. Sabiryanov , R.F. Sabiryanov , S.S. Jaswal , S.S. Jaswal ,
Y. Liu , Y. Liu , H. Haberland , D.J. Sellmyer , D.J. Sellmyer ,Phys. Rev. B -Condens.
Matter Mater. Phys. 66 (2002) 644041–644044 .
[14] S.H. Baker , M. Roy , S.C. Thornton , C. Binns , J. Phys. Condens. Matter 24 (2012)
176 0 01 .
[15] D.A. Eastham, Y. Qiang, T. H. Maddock, J. Kraft, J.P. Schille, G.S. Thompson and
H. Haberland, J. Phys. Condens. Matter, doi: 10.1088/0953-8984/9/37/001 .
[16] I.A. Ibrahim , F.A. Mohamed , E.J. Lavernia , J. Mater. Sci. 26
(1991) 1137–1156 .
[17] K. Awazu , M. Fujimaki , C. Rockst uhl , J. Tomin aga , H. Murakami , Y. Ohki ,
N. Yoshida , T. Watanabe , J. Am. Chem. Soc. 130 (2008) 1676–1680 .
[18] R.G. Palgrave , I.P. Parkin , Chem. Mater. 19 (2007) 4639–4647 .
[19] M.K. Kwon , J.Y. Kim , B.H. Kim , I.K. Park , C.Y. Cho , C.C. Byeon , S.J. Park , Adv.
Mater. 20 (2008) 1253–1257 .
[20] H.B. Liao , R.F. Xiao , J.S. Fu , P. Yu , G.K.L. Won g , P. Sheng , Appl. Phys. Lett. 70
(1997) 1–3
.
[21] R.H. Magruder Iii , L. Yang , R.F. Haglund , C.W. White , R. Dorsinville , R.R. Alfano ,
Appl. Phys. Lett. 62 (1993) 1730–1732 .
[22] K. Fukumi , A. Chayahara , K. Kadono , T. Sakaguchi , Y. Horino , M. Miya ,
J. Hayakawa , M. Satou , Jpn. J. Appl. Phys. 30 (1991) L742–L744 .
[23] F.T.N. Vüllers , R. Spolenak , Acta Mater. 99 (2015) 213–227 .
[24] S. Johnsen , J. He , J. Androulakis , V.P. Dravid , I. To dorov , D.Y. Chung ,
M.G. Kanatzidis , J. Am. Chem. Soc. 133 (2011) 3460–3470 .
[25] R. Serna , C.N. Afonso , C. Ricolleau , Y. Wang , Y. Zheng , M. Gandais , I. Vickridge ,
Appl. Phys. A 71 (20 0 0) 583–586 .
[26] W. Wang , L. Qu , G. Ya ng , Z. Chen , Appl. Surf. Sci. 218 (2003) 24–28 .
[27] R. Serna , A. Suárez-García , C.N. Afonso , D. Babonneau , Nanotechnology 17
(2006) 4588–4593 .
[28] Z. Konstantinovi
´
c , M.G. Del Muro , M. Varela , X. Batlle , A. Labarta , Nanotechnol-
ogy 17 (2006) 4106–4111 .
[29] N. Bahlawane , K. Kohse-Höinghaus , T. Weimann , P. Hinze , S. Röhe , M. Bäumer ,
Angew. Chemie - Int. Ed. 50 (2011) 9957–9960 .
[30] S.K. Medda , S. De , G. De , J. Mater. Chem. 15 (2005) 3278 .
[31] S.F. Hassan , M. Gupta , Metall. Mater. Trans.
A 36 (2005) 2253–2258 .
[32] U. Kreibig , C. v. Fragstein , Z. Phys 224 (1969) 307–323 .
[33] R.G. Palgrave , I.P. Parkin , J. Am. Chem. Soc. 128 (2006) 1587–1597 .
[34] D. Bekermann , D. Barreca , A. Gasparotto , C. Maccato , CrystEngComm 14 (2012)
6347 .
[35] T. Hihara , K. Sumiyama , J. Appl. Phys. 84 (1998) 5270–5276 .
[36] S. Yamamuro , K. Sumiyama , K. Suzuki , J. Appl. Phys. 85 (1999) 4 83–4 89 .
[37] S. Sawa , N. Tanaka , R. Katoh , K. Sumiyama , Mater. Trans. 49
(2008) 1219–1222 .
[38] K. Sumiyama , T. Hihara , D.L. Peng , R. Katoh , Sci. Tech . Adv. Mater. 6 (2005)
18–26 .
[39] M. Khojasteh , V.V. Kresin , Appl. Nanosci. 7 (2017) 875–883 .
[40] Y. Wang , J. Hu , Y. Lin , C.W. Nan , NPG Asia Mater 2 (2010) 61–68 .
[41] A.H. Kean , L. Allers , NSTI Nanotechnol. Conf. Trade Show - NSTI Nanotech 2006
Tech . Proc. 1 (2006) 749–752 .
[42] A . Salaün , J.A . Hamilton , D. Iacopino , S.B. Newcomb , M.G. Nolan , S.C. Pad-
manabhan ,
I.M. Povey , M. Salaün , M.E. Pemble , Thin Solid Films 518 (2010)
6921–6926 .
[43] P.A . Manoharm , M. Ferry , T. Chandra , ISIJ Int. 38 (1998) 913–924 .
