Carbon 124 (2017) 348e356

Contents lists available at ScienceDirect

Carbon
journal homepage: www.elsevier.com/locate/carbon

Multiwall carbon nanotubes filled with Al4C3: Spectroscopic

signatures for electron-phonon coupling due to doping process
P.T. Araujo a, b, *, N.M. Barbosa Neto c, **, M.E.S. Sousa c, R.S. Ange

lica d, S. Simo
~ es e,
M.F.G. Vieira e, M.S. Dresselhaus f, g, M.A. Leite dos Reis h, ***
a
Department of Physics and Astronomy, University of Alabama, Tuscaloosa, AL, 35401, USA
b
Center for Materials for Information Technology (MINT Center), University of Alabama, Tuscaloosa, AL, 35401, USA
c
Programa de Po
s-graduaça~o em Física, Instituto de Ci^
, 66075-110, Bel

encias Exatas e Naturais, Universidade Federal do Para em, PA, Brazil
d
Instituto de Geoci^
encias, Programa de Po
s-graduaça~o em Geologia e Geoquímica, Universidade Federal do Para
, 66075-110, Bel
em, PA, Brazil
e
Faculdade de Engenharia da Universidade do Porto, Universidade do Porto, 4200-465, Porto, Portugal
f
Department of Electrical Engineering and Computer Sciences, Massachusetts Institute of Technology, Cambridge, MA, 02139, USA
g
Department of Physics, Massachusetts Institute of Technology, Cambridge, MA, 02139, USA
h
Programa de Po
s-graduaça ~o em Engenharia de Recursos Naturais da Amazo ^nia, Universidade Federal do Para
, 66075-110, Bel

em, PA, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: The spectroscopic signatures related to doping mechanisms in multiwall carbon nanotubes filled with
Received 9 April 2017 aluminum carbide (Al4C3@MWCNTs) were studied and interpreted relative to changes in their electronic
Received in revised form and phononic structures. Unfilled MWCNTs were used as standard samples to help interpreting the filling
15 August 2017
and the doping processes. The samples were characterized via scanning electron microscopy, trans-
Accepted 20 August 2017
Available online 23 August 2017
mission electron microscopy, X-ray diffraction and resonant Raman spectroscopy. The electron-phonon
coupling mechanisms associated to the Raman intensities, frequencies and linewidths of the G- and
G'-band Raman modes were analyzed and connected to the doping mechanism in these multi-walled
Keywords:
Al4C3-MWCNTs
systems. Our results indicate that the Al4C3 particles transfer electrons to the MWCNTs. In order to
Density of states shed light into the experimental findings, theoretical calculations were performed using two examples of
Scanning electron microscopy filled and unfilled achiral MWCNTs and the results for the density of electronic states indicate that the
Transmission electron microscopy two systems under consideration, exhibit metallic behavior, with aluminum carbide doping the carbon
Raman spectroscopy nanotubes, thereby supporting our experimental observations.
Electron-phonon coupling © 2017 Elsevier Ltd. All rights reserved.

1. Introduction nanocomposites have been applied as reinforcing materials [1e3].

In 2007, Cava and collaborators [4] developed a synthesis path
Synthesis of nanocomposites has been widely used by re- capable of producing nanocomposites based on MWCNTs filled
searchers worldwide as alternative solutions for advancing, among with metal oxides, such as iron oxides. Posteriorly, Reis et al. [5]
others, the aerospace, the automotive, and the aeronautic industry. presented results of a hybrid process based on the DC arc
Metal matrix nanocomposites (MMC) based on multiwall carbon discharge and the Chemical Vapor Deposition (CVD) techniques, in
nanotubes (MWCNT) have historically been one of the leading which nanocomposites of MWCNTs filled with aluminum carbide
composites in this field of study [1e3]. Indeed, due to their (Al4C3) were successfully synthesized. The results further suggested
outstanding mechanical properties (including high Young modulus that the formation of Al4C3 occurs through chemical reactions be-
values of about 1200.00 GPa), these MWCNT-based tween the carbon nanotubes and the aluminum matrix. In their
synthesis process [5], the nanotubes are directly extruded from the
aluminum matrix via the vapor-liquid-solid (VLS) growth process
* Corresponding author. Department of Physics and Astronomy, University of
while the sintering of the nanocomposite occurs. Carbon nanotube-
Alabama, Tuscaloosa, AL, 35401, USA. based nanocomposites have a great technological appeal due to
** Corresponding author. their potential applications which involve reinforcement of mate-
*** Corresponding author. rials [6], electronic device fabrication [7] and other applications [8].
E-mail addresses: paulo.t.araujo@ua.edu (P.T. Araujo), newtonfisico@gmail.com
Dresselhaus et al. [9] proposed that an increased charge density
(N.M. Barbosa Neto), marcosallan@ufpa.br (M.A. Leite dos Reis).

http://dx.doi.org/10.1016/j.carbon.2017.08.041
0008-6223/© 2017 Elsevier Ltd. All rights reserved.
 P.T. Araujo et al. / Carbon 124 (2017) 348e356 349

distribution can be attained by filling carbon nanotubes with metal from 1 to 10 (in order to detect the nanotubes alignment/orien-
nanoparticles. In their report [9] they explain that the increase in tation) and the normal 5 to 70 range.
the charge density is mediated by electron-phonon coupling The simulations were performed with two types of triple wall
mechanisms available in such systems. It is also interesting to note carbon nanotubes (TWCNT), which are simpler versions of MWCNT
that the strong many-body interactions observed in these systems, i.e., armchair (chirality 3.3@6.6@9.9) and zigzag TWCNTs
nanotube-based composites provide an additional route for (chirality 6.0@11.0@16.0), in which the innermost tube of each
exploring superconductivity phenomena [10]. system exhibits electronic properties of a metallic and semi-
The mutual interactions between Al4C3 and MWCNTs could be conductor CNT, respectively. The MWCNT systems were designed
observed through the electronic and vibrational structure of the with three walls each in order to shorten the time spend in the
MWCNT systems. Resonant Raman spectroscopy (RRS) is a suitable calculation, and they were filled with Al4C3 so that the electronic
technique to study the behavior of electron and phonons through density of states of the MWCNTs filled with Al4C3 could be
the analysis of photons inelastically scattered by the materials. compared with the equivalent unfilled MWCNT systems. In order to
These inelastically scattered photons most always carry important calculate the density of states (DOS), the model systems were
information about the interactions between the electrons and separated into three regions: right and left electrodes (semi-infin-
phonons in or associated with these materials [11]. These electron- ite); a sequence of the unit cell of tube equivalent to an aromatic
phonon (e-ph) interactions carry, among other things, valuable ring; and a central region (scattering center for the transport calcu-
information related to the charge transfer mechanisms in doped lations). The systems were optimized via the MMþ (Molecular
systems [12e16]. The e-ph coupling mechanism can also be mechanics) force field, and then the DOS calculations were per-
responsible for generating electron-hole (e-h) pairs which leads to formed within the Extended Hückel Theory (ETH), which is a semi-
the phenomena of phonon self-energy renormalizations, first empirical method to describe the wave function of the electron that
explained by Walter Kohn [17]. Nowadays, these self-energy has very good agreement with experimental results. This method
renormalizations, which are often observed as a phonon fre- has been successfully employed in many other problems [22e24].
quency softening, are baptized as Kohn anomalies and these The Hamiltonian of these systems were constructed via the H-
anomalies have been observed in several different situations matrix method in which the Hamiltonian is formed by the energy of
[12e16,18e21]. In the present work, we provide an analysis of the molecule's ionized valence orbital and by the coating matrix S
multiwall carbon nanotubes (MWCNTs) filled with aluminum car- [25], which is a correction accounting for any loss coming from the
bide (Al4C3) particles. Our experiments, which are supported by approximations made by using the semi-empirical method.
theoretical calculations, demonstrate that the innermost tubes in Additionally, ETH utilizes a self-consistent Hamiltonian
the MWCNT systems host the Al4C3 nanoparticles and that these formalism of matrix Green functions out of balance (NEGF), so that
nanoparticles are efficient electron donors to the MWCNTs. Our the electric currents simulated can give important information
analysis also allow for the identification of Raman scattering about electronic transport. Also, the external potential applied
spectroscopic signatures which are valuable to recognize the suc- gives us information about the interactions between electrons and
cessful filling and doping of these complex multiwall carbon phonons. Moreover, the Green's function formalism uses the
nanotube systems. spectral transmission method to describe the electronic distribu-
tion of the system as well as to characterize the electronic prop-
2. Materials and methods erties of the target systems, i.e., density of states (DOS), MPSH
(Molecular Projected Self-consistent Hamiltonian) states, differential
The Al4C3@MWCNTs were synthesized through a hybrid process conductance, among others [26]. In the present work the same
which combines the DC arc discharge method with the chemical isovalue (~0.03) was used to all the plots and visual grid sampling
vapor deposition (CVD) method as described in Ref. [5]. The unfilled was set to 1.
MWCNTs were synthesized via the CVD method. A Jobin Yvon
T64000 spectrometer with a charge coupled device (CCD) for signal 3. Results and discussion
detection was used to acquire the Raman spectra. All the spectra
were obtained at room temperature in a backscattering geometry. Fig. 1(a) and (b) show the morphology of the unfilled CNTs used
The samples were excited at 514.5 nm (2.41 eV) and the power in this work as control samples. The SEM image (Fig. 1(a)) shows a
density measured from the objective was kept low enough to avoid dense agglomerate of MWCNTs while the TEM image (Fig. 1(b))
heating of the samples. A 20
objective lens with a focal distance reveals CNTs with several walls with the diameters of the outer-
f ¼ 20:5 mm and numerical aperture NA ¼ 0.35 was used to focus most tubes ranging from approximately 10 nme60 nm. For the
the laser beam on the sample surface and also to collect the back- MWCNTs filled with Al4C3, the SEM images show MWCNTs with
scattered signals. The morphologies of the MWCNTs and of the diameters of at most 90 nm, as shown in Fig. 2(a) and (b). Also from
Al4C3@MWCNT samples were characterized with a Scanning Elec- TEM image, we are able to estimate that for both samples the
tron Microscopy (SEM) model VEGA3 SB e TESCAN tuned at 20 kV majority of the tubes have external diameters around 10 nm and
and with a Transmission Electron Microscopy (TEM) model TECNAI that most of the filled MWCNTs present closed tips, as shown in
G2 20 S-TWIN tuned at 100 kV. Fig. 2(c) and Fig. S2 (supplementary information), where the main
XRD analysis were carried out in a divergent beam diffractom- morphological difference between the two samples (filled and
eter (Empyrean, PANalytical) equipped with a PW3050/60 (q/q) unfilled MWCNTs) was the presence of Al4C3 inside the filled
goniometer, a PIXel3D 2
2 area detector, a Co-sealed X-ray tube MWCNTs (Fig. 2(d)). As previously shown by Colbert et al. [27],
(Ka1 ¼ 0.178901 nm), and a kb iron metallic filter. The instrument open-tip CNTs are expected under applied voltages of about 75 V
conditions were as follows: 40 kV and 40 mA; 0.02 2q step size and and high temperatures (around 1200  C). However, when the
an acquisition time of 20 s per step in a step-scan mode; 1/2 voltage is turned-off, the tip is closed again. These observations
divergent slit and 1 anti-scattering slit. The instrument resolution imply that dome-closed tip is much easier to be found in the arc
was obtained by using the LaB6 NIST/SRM 660b standard. discharge synthesis method since higher voltages are applied and
The sample preparation included deposition of the materials on at the end of the synthesis procedure the voltage is also turned off
a silicon plate (zero background sample holder) with a few drops of at high temperatures. In our case, the Al4C3-filled nanotubes were
ethyl alcohol. The samples were scanned in two main 2q ranges: synthesized by the arc discharge apparatus [5] and nucleated by the
350 P.T. Araujo et al. / Carbon 124 (2017) 348e356

Fig. 1. Micrographs of unfilled CNTs. (a) SEM of a dense agglomerate of MWCNTs, and (b) TEM image showing the MWCNTs with several different diameters.

metal cluster. These conditions favor the extrusion or base-growth Raman spectra from the samples do not show any Raman mode
mechanism with dome-closed tip at the tubes' end. These obser- originating from the Al4C3 after the final purification of the sam-
vations are in agreement with previous electron diffraction in- ples. The samples were purified by thermal oxidation at 500  C [5].
vestigations obtained for this nanocomposite by one of the co- The purification process is necessary in order to avoid possible
authors of this manuscript [28]. Additionally, it is also obtained decoration of the external walls of the MWCNTs. All the TEM im-
from the TEM images that the filling ratio of aluminum carbide in ages acquired show no evidence of Aluminum Carbide nano-
our nanocomposite is around 23%. particles in the outer walls of the tubes as expected after
In studying the composition of the filling particles, reference purification.
[28] presents the results of X-ray diffraction acquired for our In order to better understand the influence of aluminum carbide
samples, before and after purification, which clearly show diffrac- on the hybrid Al4C3@MWCNT systems, resonant Raman spectros-
tion peaks that are assigned to Al4C3 and Al. Additionally, the copy (RRS) was carried out to obtain spectral features related to the

Fig. 2. Micrographs of filled CNTs (Al4C3@MWCNTs). (a) SEM of the MWCNT species; (b) SEM magnification shows examples of MWCNTs with a variety of diameters; (c) TEM image
shows closed tip MWCNTs (white arrow); and (d) TEM magnification shows aluminum carbide inside the MWCNT (red arrow). (A colour version of this figure can be viewed online.)
 P.T. Araujo et al. / Carbon 124 (2017) 348e356 351

Fig. 3. G-band (a) and G0 -band (b) Raman features of unfilled multiwall carbon nanotubes (without aluminum carbide - Al4C3). (c) and (d) show, respectively, the G-band and G0 -
band Raman features of filled multiwall carbon nanotubes (Al4C3@MWCNTs). In the experimental plots, black solid dots are experimental data and the gray solid lines are fittings
obtained from the deconvolution of two Lorentzian curves (dashed black lines).

G- and G0 -band (also known as 2D-band) of the unfilled and filled since these tubes would have a higher curvature and, therefore,
MWCNT samples, as shown in Fig. 3. The G-band is related to lower frequencies. Following the same argument, the G2 peak could
tangential vibrations of carbon atoms on the tube's surface and the be related to the distribution of diameters for the outermost tubes
G0 -band is related to the breathing of the hexagons formed by the since for those tubes the curvatures are smaller [9,18,31,40e46]. It
carbon atoms on the tube's structure [29e32]. The full spectra with is worth to explain why we consider this hypothesis: MWCNTs can
all the Raman features observed for both samples are shown in be realized as a rolled up multilayer graphene. In this sense, by
Fig. S1 in the supplementary material. The G-, G0 - and D-bands taking the multilayer graphene as an example, Ferrari et al. [47]
(which is an inter-valley double-resonance process related to dis- have shown that for multilayer graphene systems with more than
order on carbon materials) [33] are observed but no signals coming five layers it is not trivial to distinguish differences among their G
from the RBM modes are observed as expected for bundled- and G0 Raman bands, which makes it difficult to distinguish the
samples in MWCNTs with large diameters in average. Addition- contributions from intermediates sheets. It is also known that if
ally, they also show that the nanocomposite presents a lower modifications are made to the bottom or to the top layer of
density of defects when compared with the unfilled MWCNTs, multilayer graphenes, one starts distinguishing responses from
which means that the filled MWCNTs improve in quality along the such bottom or top layers and their adjacent layers [14,48,49].
process. This observation is expected due to the high temperatures Again, it is difficult to precisely access how the layers between the
involved in the filling process. bottom and top layers are behaving but different trends can be seen
First, we will discuss the G-band feature. Our spectra show that for the bottommost layers and for the topmost layers [14,48,49]. At
both filled and unfilled MWCNT systems present three peaks in the this point, it is helpful to recall that our multi-walled nanotubes
G-band feature, two of them were baptized G1 and G2 peaks and the present, in average, six to seven concentric nanotubes according to
third one is the D0 -band, as shown in Fig. 3(a) and (c). These G-band our TEM images, with most of the diameters ranging from 6 to
profiles are expected for MWCNTs systems, presenting low in- 15 nm. In this case, we could think of the innermost nanotube as
tensity and smeared out Raman peaks due to the diameter distri- the equivalent to the bottommost layer in the multilayer graphene,
bution of the multi-walled structures [34e37]. Each peak forming while the outermost nanotube would be equivalent to the topmost
the G-band features as displayed in both Fig. 3(a) and (c), are well layer in the multilayer graphene. On top of that, even though the
fitted with Lorentzian curves. The D0 -band is a disorder-related diameters of our tubes are relatively large, curvature effects still
intra-valley double resonance process [38,39], while the other exist and will help to further distinguish sets of inner tubes and
two modes, named G1 and G2, clearly delineate two spectral re- outer tubes. Indeed, Ni et al. [50] clearly demonstrated this. In their
gimes, which we hypothesize to be associated to diameter distri- study, they show how the frequencies for the D-, G- and G0 -bands
butions comprising innermost and outermost tubes, as previously depend on the diameter of the constituent tubes. They also show
reported in the literature [34]. We suggest that the G1 peak is how the wall-to-wall distances vary with the diameters and that
related to the distribution of diameters for the innermost tubes meaningful changes to the frequencies only exist all the way to
352 P.T. Araujo et al. / Carbon 124 (2017) 348e356

15 nm of diameter beyond which no big changes in frequencies are is also worth noting that there is no inversion in the intensity of the
observed anymore. It is also relevant to comment that, large peaks: the G1 feature is always more intense than the G2 feature. At
diameter tubes do not present curvature effects which are as strong a first glance, the absence of a frequency-shift for the G-band due to
as the curvature effects for small diameter tubes. By considering the presence of aluminum carbide, suggests that no charge transfer
that the average wall-to-wall distances in multi-wall system is is taking place in our nanocomposite. However, it is well estab-
about 0.34 nm, the diameters from the smallest to the largest tube lished that modifications in the G-band frequency are not sub-
in the MWCNT systems increase in average 0.68 nm. Consequently, stantial under low levels of doping. Therefore, we do not expect to
as the diameters of the constituent tubes in the MWCNT system see changes for the G1 and G2 frequencies. The sensitivity of the
increase, the relative changes in the frequencies of the modes for frequency to doping increases substantially for those bands which
each tube decrease. It is well established that for diameters larger originate from second order processes, as it is the case for the G0 -
than 2 nm, curvature effects will start fading. This creates a cur- band discussed further in the text. It is important to comment,
vature effect gradient (and therefore a gradient of frequencies), however, that both the FWHM and the intensity of the G-band will
which justifies once again our diameter distributions as proposed change even under low dopant levels. Namely, it is expected that as
in the manuscript and it also explains why we are not able to we dope the carbon material the FWHM and the intensity should
resolve the frequencies of the intermediate tubes. Now, like it is the decrease [52]. Therefore, even though we cannot use the G1 and G2
case for graphene, by modifying the innermost tubes and/or the frequencies to monitor the doping level, the changes we observed
outermost tubes we are able to distinguish between sets of tubes for the FWHM and intensities could suggest that the outer tubes'
closest to the innermost tube and sets of tubes closest to the distribution would be getting undoped while the inner tubes' dis-
outermost tubes in the MWCNTs. Through such modifications we tribution would be getting doped. If this assumption was truly valid,
can phenomenologically suggest what is happening to the tubes in the intensity for G1 would have to decrease while the intensity for
the middle of the MWCNTs. Our TEM images clearly show that the the G2 peak would have to increase. In the plots presented in
innermost and outermost nanotubes in our MWCNT systems are Fig. 3(a) and (c) we observe otherwise. Moreover, this G1 doping/G2
perturbed differently (see Fig. S2 available in the supplementary undoping conclusion is in contradiction with the fact that Al4C3 is a
information). On top of that, as shown by Araujo et al. [29], envi- charge donor. Additionally, it is noticeable that the presence of
ronmental perturbations, which are diameter dependent, will defects and strain, which may be caused by the presence of
affect much more small diameter tubes than large diameter tubes. aluminum carbide, is not able to explain the behavior observed for
These arguments, all together, make very plausible to assign to the the G-band features. It is known that the increase (decrease) of
G and G0 bands of the MWCNTs systems two types of tubes distri- defect densities is related to a broadening (narrowing) of every
bution: inner and outer diameter distributions. Raman band in carbon materials such as nanotubes, graphene
Our deconvolution analysis indicated that no appreciable fre- layers and graphite (frequencies for the Raman bands essentially do
quency shift is observed for the spectroscopic features of the G- not change) [53e55] and the larger (smaller) the strain, stress,
band due to the presence of Al4C3 in the MWCNTs with the first transverse pressure or hydrostatic pressure, the broader (narrower)
peak located around 1577 cm1 (hereafter named G1 peak), the the FWHM regardless the Raman mode (for stress/strain Raman
second peak located around 1609 cm1 (G2 peak) and the D0 -band bands downshift in frequency while for hydrostatic and transverse
located around 1620 cm1. Furthermore, the lack of RBM modes pressures the Raman bands upshift in frequency) [56e60]. These
signals (see Fig. S1 in the supplementary information) and the behaviors are in clear disagreement with the ones observed in our
linewidths obtained from the Lorentzian deconvolution indicates work when the Al4C3 is present, see Table 1.
that the samples investigated comprise tubes with large innermost This is why the authors hypothesize that the major behaviors of
diameters [34e37]. This result is in accordance with our SEM im- the G-band features (G1 and G2) are not quite related to the doping
ages, as shown in Figs. 1 and 2. Although no modifications are (undoping) phenomena but instead, it is a strong indicative that,
observed for the G-band frequencies, a significant change, caused beyond doping, a reorganization of tube diameter distributions for
by the presence of Al4C3, is observed in the full width at half the innermost and outermost distributions is taking place. In fact,
maximum (FWHM) and in the intensity ratio of the two G1 and G2 the narrowing (broadening) of the FWHM for G1 (G2) bands, see
features in the G-band. Table 1 summarizes the FWHM and in- Table 1, indicates a possible decrease of the innermost tubes'
tensity ratio values as extracted from the spectral fittings. diameter distribution followed by an increase of the outermost
It is also observed that the spectral features related to the D0 - tubes' diameter distribution when the MWCNTs are filled with
band are in good agreement with those observed in the literature aluminum carbide [34]. This reorganization-related assumption is
[38,39,51] and do not present any significant change when corroborated by the TEM micrographs previously reported for the
comparing the D0 -band fitting result for the nanocomposite and for samples investigated [28] and now depicted in Fig. S2 in the sup-
the unfilled tubes. plementary information. Moreover, the observation of a decrease in
Moreover, for the Al4C3-filled tubes the lower frequency peak the I(D)/I(G) ratio, from ~1.27 for unfilled CNTs to ~0.51 for the
IG
(G1) becomes narrower than the G2 peak and the ratio IG1 between nanocomposites (see Fig. S1 in the supplementary information)
2
the intensities for the observed peaks is increased from 2.5 to 3.0. It also corroborates with this hypothesis since it indicates that the
number of defects decreases in filled tubes.
In order to endorse or refute our hypothesis, we now discuss the
Table 1
Spectroscopic parameters for the filled and unfilled MWCNTs obtained from the G0 -band for the filled and unfilled MWCNT systems. It is known
Lorentzian fittings of the G-band. Du ¼ u(G2) - u(G1). that, when compared with the G-band, the G0 -band is more sen-
sitive to doping [21,52,61]. Indeed, MWCNTs filled with Al4C3, when
G-bands FWHM Frequency (cm1) I(G1)/I(G2) Du (cm1)
(cm1) compared with the unfilled tubes, clearly show important varia-
tions in the G0 -band frequency, FWHM and intensity [12e16,61,62].
unfilled G1 51 ± 1.0 1576 ± 0.7 2.5 33
unfilled G2 33 ± 1.0 1609 ± 0.7 Our G0 -band analysis for filled and unfilled MWCNTs furnishes us
unfilled D' 16 ± 1.0 1619 ± 0.7 with strong evidences of charge transfer from the Al4C3 to the
filled G1 30 ± 1.0 1577 ± 0.7 3.0 31 MWCNTs and help us understand the changes observed for the G-
filled G2 54 ± 1.0 1608 ± 0.7 band feature. Similar to the G-band case, it is clear that the G0 -band
filled D' 14 ± 1.0 1621 ± 0.7
presents two different spectral regimes, which we call G0 1 and G0 2
 P.T. Araujo et al. / Carbon 124 (2017) 348e356 353

(see Fig. 3(b) and (d)). It is true that the G0 -band frequency lowers as contributing more to the Raman scattering than the innermost
the tube diameter becomes smaller and, therefore, the G0 1 peak, tubes' diameter distribution. Next, we explain the doping mecha-
which presents a lower frequency, is assigned to that same inner- nism we believe is taking place, which is supported by theoretical
most tubes' diameter distribution as that for the G-band G1 feature calculations presented later in the text. In short, Al4C3 particles are
[50]. The G0 2 peak, which presents a higher frequency, is assigned to enclosed by the innermost tubes of the MWCNT systems. These Al4C3
the same outermost tubes' diameter distribution as that for the G- particles are doping the innermost tubes' distribution with negative
band G2 feature [36,37]. When compared with the behavior charges. In the sequence, these negative charges are migrating to the
observed for the G-band features, it is evident that the G0 -band tubes in the outermost tubes' distribution.
features undergo considerable frequencyI G0
shifts, changes in the Apparently, this migration of charges seems not to be reasonable
FWHM values and changes in the I 10 intensity ratio. Noticeably, this because it is expected that the transfer of negative charges from the
G
ratio decreases when the MWCNTs 2are filled with Al4C3, see Table 2. Al4C3 to the innermost tubes should be followed by a softening of
More precisely, as shown in Fig. 3(b) and (d), the G0 1 intensity de- the carbon-carbon (CeC) bonding connecting carbon atoms on the
creases while the G0 2 intensity increases for the Al4C3 - filled tube's surface of the innermost tubes distribution. As widely
MWCNT samples. accepted, these bonding are like springs associated with the
As regards the frequency shifts observed for the G0 1 and for the vibrational mode. Therefore, the softening of the CeC bonding can
G0 2 peaks, it is observed that the frequency of the G0 1 peak, which be interpreted as a softening of the spring constant, which would
we suggest to be related to the innermost tubes' distribution, un- redshift the G0 1 peak frequency. The same behavior was previously
dergoes a blueshift of about 4 cm1, while the frequency of the G0 2 observed for other carbon materials, such as graphene sheets
peak, which we suggest to be related to the outermost tubes' [21,52,61,63e69]. Furthermore, the electronic states of the Al4C3
diameter distribution, undergoes a redshift of about 7 cm1. The particles would be expected to hybridize with the electronic states
shifts observed for the G0 -band indicate that the Al4C3 particles, of the innermost tubes' distribution. As a consequence, an increase
which are hosted by the innermost tubes, work as electron donors of the electronic density of states of the hosting tubes would be
to the MWCNT systems [15,16,61,62], and the magnitudes of the expected. This enhancement in the density of electronic states
shifts corroborate the hypothesis raised earlier that we are working would be expected to enhance the Raman scattering cross-section
in a low doping regime. Moreover, the blue (red) shift observed for associated to the G0 1 peak, which would imply an increase of the
the G0 1 (G0 2) band indicates that the tubes in the innermost dis- intensity observed for this peak. On the other hand, the Raman
tribution are positively doped while the outermost tubes are intensity for the G0 2 peak would decrease.
negatively doped [52,61,62]. It is worth to mention that the However, it is clear that the spectral changes observed here are
decrease observed in both G0 1 and G0 2 FWHMs is an additional not in agreement to the phenomenology described above. Our data
confirmation that doping is taking place [52,61] but it cannot be shows that the G0 1 peak blueshifts in frequency and presents a
taken by itself as an evidence to distinguish which MWCNT decrease in its intensity, while the G0 2 peak redshifts in frequency
diameter distributions (innermost or outermost clusters of and presents an increase in its intensity. We claim that these results
MWCNT) are receiving or losing charges. The G0 -band frequencies, are consistent with an avalanche/self-doping mechanism, which
on the other hand can be taken to precisely assess which distri- starts in the Al4C3 particles and ends in the outermost tubes' dis-
bution is receiving or losing charge. Differently from the fre- tribution of the MWCNTs structures. These mechanisms are un-
quencies, the FWHM will decrease for both cases: positive and derstood as follows: the unfilled MWCNT systems are in
negative doping [52,61]. equilibrium and, therefore, neutral. When the Al4C3 particles are
The behavior of the G0 -band intensities related to peaks 1 and 2, enclosed by the MWNCTs, their electronic states will indeed hy-
as shown in Fig. 3(b) and (d), seems to indicate doping as well. In bridize with the electronic states belonging to innermost tubes in
2006, Souza Filho et al. studied double wall carbon nanotubes the MWCNTs. The Al4C3 particles are charge donors and they also
(DWCNT) with dopant Br2 molecules adsorbed at the outermost work as a reservoir of charges, whose chemical potential is higher
tube of the DWCNT systems [16]. One of their observations is that relative to the chemical potentials of the neutral tubes in the
the G0 -band have two peaks associated to inner and outer tubes as MWCNTs' configuration. The presence of these Al4C3 particles
well, whose intensity magnitudes invert when going from the therefore leads the new Al4C3@MWCNT hybrid systems to find a
undoped to the doped DWCNT systems. In their case, when the new equilibrium configuration. This new equilibrium condition
DWCNTs are doped, the intensity of the peak related to the inner starts with charges migrating to the innermost tube that will
wall decreases while the intensity of the peak related to the outer eventually acquire a higher chemical potential than the rest of the
wall increases. We are observing the I same effect in our measure- tubes in the MWCNT system, which would trigger a migration of
G0
ments: we observe a decrease in the I 10 ratio when going from the charges to tubes adjacent to the innermost tubes and so forth. This
G
unfilled to the filled MWCNT systems,2 in which the G0 1 peak in- migration mechanism would propagate until the chemical poten-
tensity decreases while the G0 2 peak intensity increases (note that it tial for the outermost tubes gets in equilibrium with the reservoir. It
increases substantially more than its equivalent for the DWCNT is also reasonable to assume that during the process for reaching
case). This inversion seems to support the hypothesis that the equilibrium, electrons belonging to the innermost tubes will be
outermost tubes' diameter distribution are being doped and are transferred to the outermost tubes as well, which would generate
regions of positive doping, which is in agreement to the signatures
observed for the G0 1 peak (note that the reservoir, even after
Table 2 equilibrium, will repel negative charges as well). A similar mech-
Spectroscopic parameters for the filled and unfilled MWCNTs obtained from the anism happens in p-n junctions in which the depletion area of the
Lorentzian fit of the G0 -band. Du ¼ u(G0 2) - u(G0 1).
negative material is positively charged, while the depletion area for
G'-bands FWHM Position I(G0 1)/I(G0 2) Du (cm1) the positive material is negatively charged [70,71]. Such processes
(cm1) (cm1) can be understood as self-doping mechanisms and following this
unfilled G'1 60 ± 1.0 2676 ± 0.7 1.8 37 reasoning, the Al4C3@MWCNTs hybrid systems are potential nano-
unfilled G'2 52 ± 1.0 2713 ± 0.7 sized n-p-n junctions, which are vastly utilized in transistors' en-
filled G'1 56 ± 1.0 2680 ± 0.7 0.7 26 gineering [70,71]. The negative doping of the outermost tubes
filled G'2 48 ± 1.0 2706 ± 0.7
agrees well with the signatures observed for the G0 2 peak.
354 P.T. Araujo et al. / Carbon 124 (2017) 348e356

Additionally, it is expected that these doping mechanisms will as shown in Fig. 4. This upshift is measured to be about 0.5 and
lead to a screening effect in avalanche, in which the wall-to-wall such behavior indicates a decrease in the average wall-to-wall
interactions, mediated by van der Walls forces, decrease in distances, which aligns well with the hypothesis raised above.
magnitude. When applied to our systems, this means that tubes in Furthermore, the narrowing observed for x-ray diffraction peak
the outermost tubes distribution will have their mutual wall-to- from the filled MWCNTs indicates an average decrease in the wall-
wall interactions gradually decreased in magnitude. This decrease to-wall distances' distribution between the concentric tubes
in the van der Waals interactions, as well established in the liter- constituting the MWCNTs. The X-ray data is in agreement with the
ature [29e32], would contribute with an extra frequency redshift results obtained in the Raman spectra analysis and in the TEM
for the G0 2 peak describing the outer tubes' diameter distribution. It images (see Fig. S2 in the supplementary information). The big
is important to emphasize that this screening effect has been re- picture indicates that the presence of the Al4C3 inside the MWCNT
ported elsewhere for triple wall carbon nanotubes [40e42]. structures makes the diameter distribution for the inner (outer)
Moreover, we hypothesize that, in the new equilibrium condition, distribution decrease (increase) with a net result indicating tubes of
as discussed earlier, it is likely that the innermost tubes and the larger diameters. Therefore our MWCNTs, now larger in diameters
outermost tubes are attracting each other since they have average and with less curvature effects, are more similar to a multi-layer
charge distributions of opposite signs. This attraction would lead to graphene system with well-defined interlayer distances and,
a further decrease in the average wall-to-wall distances as well as a consequently, a narrower x-ray diffraction peak. On top of that, the
change in the tube diameter distribution, which would host tubes samples improve in quality after the filling, which leads to less
with slightly larger diameters. These changes would naturally uncertainty in the value of the wall-to-wall distances. Our X-ray
manifest as a modification in the inner tubes configuration, which results are aligned with the observations reported by Hiroyuki et al.
would lose some small diameter species. These modifications in the [50], which brings the wall-to-wall distances in MWCNTs as a
carbon nanotubes' population in the inner and outer tubes' diam- function of the diameters of the constituent tubes.
eter distributions are coherent with the FWHM narrowing In order to endorse our experimental conclusions, we per-
observed for the G1 peak and with the FWHM broadening observed formed theoretical calculations with two types of TWCNTs: one
for the G2 peak in the G-band (see Table 1). Also, the doping armchair and one zigzag, which are simpler versions of MWCNT
mechanism as explained thus far covers the decrease in the in- systems. The innermost tube of each system exhibits electronic
tensity observed for the G0 1 peak followed by an increase in the properties of a metallic and semiconductor CNT, respectively (the
intensity observed for the G0 2 peak when going from unfilled to details of the calculational methods are described earlier in the
filled MWCNTs. Indeed, the negative doping of the outer tubes will text). The metallicity (and, therefore, the conduction/insulating
enhance their Raman cross-sections and enhance, consequently, properties) of the innermost tube is ensured by its chirality choice,
their Raman intensity. which was set before the calculations started. Interestingly, as
In order to get more information about the wall-to-wall dis- shown in Fig. S3 (supplementary information), the density of
tances in the MWCNT systems and endorse our analysis related to electronic states calculated for the zigzag tube indicates, in fact, a
the self-doping and shielding mechanisms, X-ray diffraction mea- metallic behavior. This result stems from the fact that for a MWCNT
surements were carried out in the filled and unfilled MWCNTs the inter-tube junctions may cause an enhancement of the bulk
samples studied here. Fig. 4 presents the X-ray diffractograms ob- transport properties [75]. Furthermore, it is observed that the
tained for both samples. These X-ray experiments provide precise armchair CNT in the filled TWCNT system, presents a substantial
information related to the wall-to-wall distances and our X-ray increase of its density of electronic states in comparison with the
results show a typical diffractogram that reflects the distribution of unfilled system. This increase is measured by the area that is
wall-to-wall distances in MWCNTs [72]. It is seen that the peaks in quantitatively obtained by integrating the curve density of states
Fig. 4 are formed by broad bands that are consistent to the distri- (DOS) vs. energy, as shown in Fig. 5(a). For the Al4C3-filled armchair
bution of several slightly distinct wall-to-wall distances between TWCNT it is found the value 576.2 for the integrated area. This value
the concentric tubes present in our MWCNTs samples. Our unfilled is about 3.5 times greater than that obtained for the unfilled tube,
MWCNT samples present a mean wall-to-wall distance of the order which is found to be 166. The calculations show, therefore, that
of 0.34 nm, in agreement with results previously reported in the electrons are leaving the Al4C3 particles and going to the MWCNT.
literature [8,9,34e36,40e42,44e46,50,72e74]. It is noticeable that We also performed the calculations for different excitation energies
the presence of Al4C3 in the MWCNT systems generates an average and it is seen that the higher the energy, the more enhanced the
upshift (narrowing) of the center (FWHM) of the diffraction band, transfer of charges, which leads us to a self-doping mechanism
driven by light absorption. The zigzag system presents a smaller
variation in its DOS, being the area under the curve DOS vs energy
equal to 268.5 for the Al4C3-filled system and 238.9 for the unfilled
system, as seen in Fig. S3(a) (supplementary information). It is
important to note that a transfer of negative charges still happens
from the Al4C3 particles to the MWCNTs, although the transfer is
smaller in magnitude.
The Molecular Projected Self-consistent Hamiltonian (MPSH)
states in our calculations allow us to access the Van Hove singu-
larities of the MWCNT systems and it is possible to conclude that
the Al4C3@MWCNT systems show much more electronic de-
generacies in the highest occupied molecular orbital (HOMO) levels
than the unfilled MWCNT systems, i.e., the singularities 1, 2, 3 and 4
of the Al4C3-filled MWCNT systems exhibit a greater DOS when
compared with the respective singularities of the unfilled MWCNT
Fig. 4. X-ray diffraction pattern of multiwall carbon nanotubes filled with Al4C3 (red
systems, as seen Fig. 5(a). It is also observed that the HOMO level is
solid lines) and unfilled (black solid lines). The black vertical lines indicate the center of localized in the hybrid system composed by the Al4C3 and the
X-ray diffraction peaks. (A colour version of this figure can be viewed online.) TWCNT innermost walls while the lowest unoccupied molecular
 P.T. Araujo et al. / Carbon 124 (2017) 348e356 355

4. Conclusion

In summary, we studied the doping mechanism of MWCNTs

filled with Al4C3 nanoparticles. SEM and TEM studies confirm that
the Al4C3 enter inside the innermost tubes in the MWCNT systems.
The Raman spectroscopy results provide important signatures for
the G-band and the G0 -band, which are intimately related to the
doping of MWCNTs triggered by the Al4C3 nanoparticles. The main
findings, which are supported by theoretical calculations, can be
summarized as follows: (1) the G- and G0 -bands are constituted of
two distinct spectral regions. The G-band (G0 -band) has two peaks,
which we baptized G1 (G0 1) and G2 (G0 2). We hypothesize that these
peaks are related, respectively, to a diameter distribution for
innermost tubes in the MWCNT and to a diameter distribution for
outermost tubes in the MWCNT; (2) the spectroscopic signatures,
summarized in Tables 1 and 2, reveal that electrons are being
transferred from the Al4C3 nanoparticles (which serve as reservoir
of charges) to the MWCNT systems in an avalanche/self-doping
scheme that culminates in an excess of negative charges in the
outer tubes' diameter distribution and an excess of positive charges
in the innermost tubes' diameter distribution; and (3) the Raman
spectroscopic signatures when merged with the X-ray spectroscopy
results and the TEM images show that a re-distribution in the di-
ameters present in the inner and outer tubes' distributions is taking
place, in which the distribution of diameters for the innermost
tubes seems to be decreasing while the distribution of diameters
for the outermost tubes seems to be increasing.

Acknowledgements

All authors are grateful to Brazilian agencies for financial sup-

port. Particularly, M. A. L. R. is grateful to ELETROBRAS ELE-
TRONORTE under the contract number (4500075173) and CNPq,
Fig. 5. (a) Density of states and (b) projections of electronic orbitals for triple-wall under the contract number (475659/2013-9). M. E. S. S. is grateful to
armchair carbon nanotubes. In (a), the peaks (1), (2), (3) and (4) label the Van Hove
CAPES for the financial support and to the Graduate Program in
singularities found for both HOMO and LUMO states of the system. In figure (b) the
blue parts of the orbitals stand for the positive nodes of the wavefunctions, while the Physics at the Federal University of Par
a for making its experi-
red parts stand for the negative nodes. (A colour version of this figure can be viewed mental facilities available for this research. P. T. A. and N. M. B.N. are
online.) grateful to CNPq under the contract number (401453/2014-6). M. S.
D. is grateful to MIT for the travel grant to Brazil, and for NSF grant
DMR 1507.806.
orbital (LUMO) level does not show molecular orbitals related to
the Al4C3, which is a signature that a negative charge transfer
Appendix A. Supplementary data
mechanism from the Al4C3 to the MWCNT walls is taking place, see
Fig. 5(b). Similar results were obtained by Louis et al. [76], in which
Supplementary data related to this article can be found at http://
a transfer of negative charges from the titanium particles to the CNT
dx.doi.org/10.1016/j.carbon.2017.08.041.
system is reported. Finally, it is important to comment that the
electronic calculations indicate that both Al4C3-filled armchair and
References
Al4C3-filled zigzag systems show metallic behaviors, which are
originating from the doping of the MWCNTs triggered by the Al4C3 [1] T. Laha, Y. Liu, A. Agarwal, J. Nanosci. Nanotechnol. 7 (2007) 515.
nanoparticles. [2] J.-z. Liao, M.-J. Tan, I. Sridhar, Mater. Des. 31 (Supplement 1) (2010) S96.
It is interesting to say that our theoretical calculations also [3] J. Liao, M.-J. Tan, Powder Technol. 208 (2011) 42.
[4] C.E. Cava, R. Possagno, M.C. Schnitzler, P.C. Roman, M.M. Oliveira,
suggest that the charge transfer that occurs from the HOMO to the C.M. Lepiensky, A.J.G. Zarbin, L.S. Roman, Chem. Phys. Lett. 444 (2007) 304.
LUMO states of the Al4C3@MWCNT armchair system could be [5] M.A.L. dos Reis, A.F. Saraiva, M.F.G. Vieira, J.D. Nero, J. Nanosci. Nanotechnol.
triggered by light, in a light induced charge transfer mechanism 12 (2012) 6955.
[6] P.M. Ajayan, J.M. Tour, Nature 447 (2007) 1066.
(see Fig. 5(b) to compare the HOMO and LUMO states of the filled [7] X. Zhang, D. Luo, Y. Niu, Y. Wang, G. Cui, J. Comput. Theor. Nanosci. 9 (2012)
MWCNTs). As a consequence of the injection of electrons in the 1680.
innermost CNTs due to the light induced mechanism, the self- [8] H. Li, W.Y. Yin, K. Banerjee, J.F. Mao, IEEE Trans. Electron Dev. 55 (2008) 1328.
[9] M.S. Dresselhaus, A. Jorio, M. Hofmann, G. Dresselhaus, R. Saito, Nano Lett. 10
doping process can be enhanced since the excitation of electrons
(2010) 751.
between the van-Hove singularities perturbs the equilibrium of the [10] M. Tinkham, Introduction to Superconductivity, Dover publications Inc., 1996.
system, allowing the injection of extra electron from Al4C3 reservoir [11] A. Jorio, M.A. Pimenta, A.G.S. Filho, R. Saito, G. Dresselhaus, M.S. Dresselhaus,
New J. Phys. 5 (2003) 139.
to the MWCNTs, as previously reported elsewhere [76]. The veri-
[12] P.T. Araujo, D.L. Mafra, K. Sato, R. Saito, J. Kong, M.S. Dresselhaus, Phys. Rev.
fication of such self-doping process is of great importance for the Lett. 109 (2012) 046801.
understanding of doping mechanisms on carbon based materials [13] P.T. Araujo, D.L. Mafra, K. Sato, R. Saito, J. Kong, M.S. Dresselhaus, Sci. Rep. 2
and it is a topic to be further investigated in the future. (2012) 1017.
[14] P.T. Araujo, O. Frank, D.L. Mafra, W. Fang, J. Kong, M.S. Dresselhaus, M. Kalbac,
Sci. Rep. 3 (2013) 2061.
356 P.T. Araujo et al. / Carbon 124 (2017) 348e356

[15] I.O. Maciel, J. Campos-Delgado, E. Cruz-Silva, M.A. Pimenta, B.G. Sumpter, S. Piscanec, D. Jiang, K.S. Novoselov, S. Roth, A.K. Geim, Phys. Rev. Lett. 97
V. Meunier, F. Lo
pez-Urías, E. Mun ~ oz-Sandoval, H. Terrones, M. Terrones, (2006) 187401.
A. Jorio, Nano Lett. 9 (2009) 2267. [48] W. Fang, A.L. Hsu, R. Caudillo, Y. Song, A.G. Birdwell, E. Zakar, M. Kalbac,
[16] A.G. Souza Filho, M. Endo, H. Muramatsu, T. Hayashi, Y.A. Kim, E.B. Barros, M. Dubey, T. Palacios, M.S. Dresselhaus, P.T. Araujo, J. Kong, Nano Lett. 13
N. Akuzawa, G.G. Samsonidze, R. Saito, M.S. Dresselhaus, Phys. Rev. B 73 (2013) 1541.
(2006) 235413. [49] L.C. Campos, T. Taychatanapat, M. Serbyn, K. Surakitbovorn, K. Watanabe,
[17] W. Kohn, Phys. Rev. Lett. 2 (1959) 393. T. Taniguchi, D.A. Abanin, P. Jarillo-Herrero, Phys. Rev. Lett. 117 (2016)
[18] R. Saito, G. Dresselhaus, M.S. Dresselhaus, Physical Properties of Carbon, Im- 066601.
perial College Press, London, 1998. [50] N. Hiroyuki, S. Yoshiyuki, N. Hamazo, K. Atsuhiro, Appl. Phys. Express 1 (2008)
[19] B.N. Brockhouse, T. Arase, G. Caglioti, K.R. Rao, A.D.B. Woods, Phys. Rev. 128 064005.
(1962) 1099. [51] K. Sato, R. Saito, Y. Oyama, J. Jiang, L.G. Cançado, M.A. Pimenta, A. Jorio,
[20] L.M. Malard, M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus, Phys. Rep. 473 G.G. Samsonidze, G. Dresselhaus, M.S. Dresselhaus, Chem. Phys. Lett. 427
(2009) 51. (2006) 117.
[21] J. Yan, E.A. Henriksen, P. Kim, A. Pinczuk, Phys. Rev. Lett. 101 (2008) 136804. [52] M. Lazzeri, F. Mauri, Phys. Rev. Lett. 97 (2006) 266407.
[22] D. Kienle, J.I. Cerdda, A.W. Ghosh, J. Appl. Phys. 100 (2006) 043714. [53] G.D. Saraiva, A.G. Souza Filho, G. Braunstein, E.B. Barros, J. Mendes Filho,
[23] A. Zienert, J. Schuster, T. Gessner, J. Phys. Chem. A 117 (2013) 3650. E.C. Moreira, S.B. Fagan, D.L. Baptista, Y.A. Kim, H. Muramatsu, M. Endo,
[24] C. Kergueris, J.P. Bourgoin, S. Palacin, D. Esteve, C. Urbina, M. Magoga, M.S. Dresselhaus, Phys. Rev. B 80 (2009) 155452.
C. Joachim, Phys. Rev. B 59 (1999) 12505. [54] M.S. Dresselhaus, A. Jorio, A.G. Souza Filho, R. Saito, Philosophical transactions
[25] R. Hoffmann, J. Chem. Phys. 39 (1963) 1397. of the royal society a: mathematical, Phys. Eng. Sci. 368 (2010) 5355.
[26] F. Guinea, C. Tejedor, F. Flores, E. Louis, Phys. Rev. B 28 (1983) 4397. [55] L.G. Cançado, A. Jorio, E.H. Martins Ferreira, F. Stavale, C.A. Achete, R.B. Capaz,
[27] D.T. Colbert, R.E. Smalley, Carbon 33 (1995) 921. M.V.O. Moutinho, A. Lombardo, T.S. Kulmala, A.C. Ferrari, Nano Lett. 11 (2011)
[28] M.A.L. dos Reis, E.M.S. Rodrigues, J. Del Nero, S. Simo~es, F. Viana, M.F.G. Vieira, 3190.
M.T.F. Vieira, J. Mater. Sci. Eng. B 5 (2015) 9. [56] C. Thomsen, S. Reich, H. Jantoljak, I. Loa, K. Syassen, M. Burghard,
[29] P.T. Araujo, I.O. Maciel, P.B.C. Pesce, M.A. Pimenta, S.K. Doorn, H. Qian, G.S. Duesberg, S. Roth, Appl. Phys. A 69 (1999) 309.
A. Hartschuh, M. Steiner, L. Grigorian, K. Hata, A. Jorio, Phys. Rev. B 77 (2008) [57] P.V. Teredesai, A.K. Sood, D.V.S. Muthu, R. Sen, A. Govindaraj, C.N.R. Rao,
241403. Chem. Phys. Lett. 319 (2000) 296.
[30] P.T. Araujo, A. Jorio, Phys. Status Solidi (b) 245 (2008) 2201. [58] R. Kumar, S.B. Cronin, Phys. Rev. B 75 (2007) 155421.
[31] P.T. Araujo, C. Fantini, M.M. Lucchese, M.S. Dresselhaus, A. Jorio, Appl. Phys. [59] T.M.G. Mohiuddin, A. Lombardo, R.R. Nair, A. Bonetti, G. Savini, R. Jalil,
Lett. 95 (2009) 261902. N. Bonini, D.M. Basko, C. Galiotis, N. Marzari, K.S. Novoselov, A.K. Geim,
[32] J.S. Soares, A.P.M. Barboza, P.T. Araujo, N.M. Barbosa Neto, D. Nakabayashi, A.C. Ferrari, Phys. Rev. B 79 (2009) 205433.
N. Shadmi, T.S. Yarden, A. Ismach, N. Geblinger, E. Joselevich, C. Vilani, [60] A.L. Aguiar, E.B. Barros, R.B. Capaz, A.G. Souza Filho, P.T.C. Freire, J.M. Filho,
L.G. Cançado, L. Novotny, G. Dresselhaus, M.S. Dresselhaus, B.R.A. Neves, D. Machon, C. Caillier, Y.A. Kim, H. Muramatsu, M. Endo, A. San-Miguel, J. Phys.
M.S.C. Mazzoni, A. Jorio, Nano Lett. 10 (2010) 5043. Chem. C 115 (2011) 5378.
[33] S.B. Cronin, A.K. Swan, M.S. Ünlü, B.B. Goldberg, M.S. Dresselhaus, M. Tinkham, [61] A. Das, B. Chakraborty, S. Piscanec, S. Pisana, A.K. Sood, A.C. Ferrari, Phys. Rev.
Phys. Rev. B 72 (2005) 035425. B 79 (2009) 155417.
[34] X. Zhao, Y. Ando, L.-C. Qin, H. Kataura, Y. Maniwa, R. Saito, Appl. Phys. Lett. 81 [62] D.L. Mafra, J. Kong, K. Sato, R. Saito, M.S. Dresselhaus, P.T. Araujo, Phys. Rev. B
(2002) 2550. 86 (2012) 195434.
[35] E.F. Antunes, A.O. Lobo, E.J. Corat, V.J. Trava-Airoldi, A.A. Martin, C. Veríssimo, [63] A. Das, S. Pisana, B. Chakraborty, S. Piscanec, S.K. Saha, U.V. Waghmare,
Carbon 44 (2006) 2202. K.S. Novoselov, H.R. Krishnamurthy, A.K. Geim, A.C. Ferrari, A.K. Sood, Nat.
[36] M.S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio, Phys. Rep. 409 (2005) 47. Nano 3 (2008) 210.
[37] R. Saito, M. Hofmann, G. Dresselhaus, A. Jorio, M.S. Dresselhaus, Adv. Phys. 60 [64] J. Yan, Y. Zhang, P. Kim, A. Pinczuk, Phys. Rev. Lett. 98 (2007) 166802.
(2011) 413. [65] A.B. Kuzmenko, L. Benfatto, E. Cappelluti, I. Crassee, D. van der Marel, P. Blake,
[38] L.G. Cançado, K. Takai, T. Enoki, M. Endo, Y.A. Kim, H. Mizusaki, A. Jorio, K.S. Novoselov, A.K. Geim, Phys. Rev. Lett. 103 (2009) 116804.
L.N. Coelho, R. Magalha ~es-Paniago, M.A. Pimenta, Appl. Phys. Lett. 88 (2006) [66] L.-C. Yin, H.-M. Cheng, R. Saito, M.S. Dresselhaus, Carbon 49 (2011) 4774.
163106. [67] N. Caudal, A.M. Saitta, M. Lazzeri, F. Mauri, Phys. Rev. B 75 (2007) 115423.
[39] S. Osswald, M. Havel, Y. Gogotsi, J. Raman Spectrosc. 38 (2007) 728. [68] K.-i. Sasaki, H. Farhat, R. Saito, M.S. Dresselhaus, Phys. E Low-dimensional Syst.
[40] T.C. Hirschmann, P.T. Araujo, H. Muramatsu, X. Zhang, K. Nielsch, Y.A. Kim, Nanostructures 42 (2010) 2005.
M.S. Dresselhaus, ACS Nano 7 (2013) 2381. [69] H. Farhat, K. Sasaki, M. Kalbac, M. Hofmann, R. Saito, M.S. Dresselhaus, J. Kong,
[41] T.C. Hirschmann, M.S. Dresselhaus, H. Muramatsu, M. Seifert, U. Wurstbauer, Phys. Rev. Lett. 102 (2009) 126804.
E. Parzinger, K. Nielsch, Y.A. Kim, P.T. Araujo, Phys. Rev. B 91 (2015) 075402. [70] N.W. Ashcroft, N.D. Mermin, Solid State Physics, Harcourt College Publishers,
[42] T.C. Hirschmann, P.T. Araujo, H. Muramatsu, J.F. Rodriguez-Nieva, M. Seifert, Orlando, 1976.
K. Nielsch, Y.A. Kim, M.S. Dresselhaus, ACS Nano 8 (2014) 1330. [71] C. Kittel, Introduction to Solid State Physics, John Wiley & Sons Inc., New york,
[43] N.M. Barbosa Neto, J.R. Alves da Cunha, K. Pereira da Silva, M. Terrones, W. 1996.
Paraguassu, P.T. Araujo, (2017). [72] H. Yusa, T. Watanuki, Carbon 43 (2005) 519.
[44] R. Pfeiffer, F. Simon, H. Kuzmany, V.N. Popov, Phys. Rev. B 72 (2005) 161404. [73] S. Bandow, M. Takizawa, K. Hirahara, M. Yudasaka, S. Iijima, Chem. Phys. Lett.
[45] F. Villalpando-Paez, L.G. Moura, C. Fantini, H. Muramatsu, T. Hayashi, Y.A. Kim, 337 (2001) 48.
M. Endo, M. Terrones, M.A. Pimenta, M.S. Dresselhaus, Phys. Rev. B 82 (2010) [74] R. Pfeiffer, C. Kramberger, F. Simon, H. Kuzmany, V.N. Popov, H. Kataura, Eur.
155416. Phys. J. B - Condens. Matter Complex Syst. 42 (2004) 345.
[46] F. Villalpando-Paez, H. Muramatsu, Y.A. Kim, H. Farhat, M. Endo, M. Terrones, [75] R. Bhatia, V. Prasad, R. Menon, J. Appl. Phys. 109 (2011) 053713.
M.S. Dresselhaus, Nanoscale 2 (2010) 406. [76] C.N. Louis, G. Jayam, R.A. A, Mater. Phys. Mech. 10 (2010) 10.
[47] A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri,