Downloaded from http://rsos.royalsocietypublishing.

org/ on September 5, 2018

Imaging carbon
rsos.royalsocietypublishing.org
nanostructures’ reactivity:
a complementary strategy
Research
to define chemical structure
Cite this article: Pérez-Luna V, Cisneros M,
Verónica Pérez-Luna1, Mario Cisneros1,
Bittencourt C, Saucedo-Orozco I, Quintana M.
2018 Imaging carbon nanostructures’ reactivity: Carla Bittencourt2, Izcoatl Saucedo-Orozco1,3
a complementary strategy to define chemical
structure. R. Soc. open sci. 5: 180605. and Mildred Quintana1,3
http://dx.doi.org/10.1098/rsos.180605 1
Instituto de Fı́sica, Universidad Autónoma de San Luis Potosı́, Manuel Nava 6,
Zona Universitaria San Luis Potosı́, San Luis Potosı́ SLP 78290, México
Received: 19 April 2018 2
Chemie des Interactions Plasma-Surface, University of Mons, Avenue Nicolas Copernic,
Accepted: 12 July 2018 1, 7000 Mons, Belgium
3
Microscopia de Alta Resolución, Centro de Investigación en Ciencias de la Salud y
Subject Category: Biomedicina, Universidad Autónoma de San Luis Potosı́, Av. Sierra Leona 550,
Chemistry San Luis Potosı́ SLP 78210, México
MQ, 0000-0002-7036-8658
Subject Area:
nanotechnology/electron microscopy/ In the search for the integration of carbon nanostructures in com-
chemical physics posite and functional materials, covalent organic reactions are
successfully performed. This approach resulted in the construc-
Keywords: tion of tailored chemical interfaces facilitating incorporation of
carbon nanotubes, graphene, graphene oxide nanocarbons. By a combination of different characterization
techniques, such as high-resolution X-ray photo-spectroscopy,
gold nanoparticles, chemical functionalization,
thermogravimetric analysis, Raman spectroscopy, UV-vis-nIR,
transmission electron microscopy and fluorescence spectroscopies, it is possible to identify and
quantify the functional moieties covalently attached to the
Author for correspondence:
carbon frame. However, the determination of the structural con-
Mildred Quintana formation of functionalized nanostructures remains a difficult
e-mail: [email protected] task. In this work, we present a straightforward methodology
This article has been edited by the Royal Society to visualize by transmission electron microscopy the functional
moieties covalently attached to the carbon network in carbon
of Chemistry, including the commissioning, peer
nanotubes and graphene. The identification of the functionalities
review process and editorial aspects up to the occurs in colloidal dispersions by using gold nanoparticles
point of acceptance. (AuNPs) as discriminating markers by molecular recognition
or by the direct growth of AuNPs on the oxygenated moieties.
Electronic supplementary material is available online This methodology, in combination with other characteriza-
at https://dx.doi.org/10.6084/m9.figshare.c. tion analysis, is expected to improve the design of hierarchical
4173812. interfaces by the spatial localization of the functionalities
responsible for colloidal stabilization in solvents with different
polarities, different from their homogeneous incorporation into
different matrices.

& 2018 The Authors. Published by the Royal Society under the terms of the Creative
Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits
unrestricted use, provided the original author and source are credited.
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

1. Introduction 2

rsos.royalsocietypublishing.org
During the last decades the insertion of the fascinating chemical and physical properties of one-dimensional
carbon nanotubes (CNTs) and two-dimensional graphene in functional materials has attracted enormous
interest. As produced, large intermolecular forces maintain pristine carbon nanostructures forming
aggregates, making the incorporation of their remarkable properties in composites challenging. Individual
CNT and graphene sheets possess high surface areas, flexibility, electrical and thermal conductivities, prop-
erties not observed in bundles of CNTs [1,2] or in graphite nanocrystals [3,4]. To a large extent, the surface
properties like hydrophobicity, aromatic stacking and surface topology from the structural to the atomic
level direct the performance and functionality of carbon nanostructures in composite materials [5] and in
complex interacting systems, i.e. at the nano–bio boundary [6]. In this direction, surface chemical function-
alization provides an effective strategy to regulate the interface between carbon nanostructures and other

R. Soc. Open Sci. 5: 180605

materials, compounds or molecules. Thus, understanding carbon nanostructures’ chemical reactivity is
of paramount importance in material design. The main characteristic associated with most carbon nano-
materials is the presence of pentagons and heptagons in a predominantly hexagonal carbon network
leading to positive and/or negative curvature [7]. In fullerenes and CNTs, the curvature has been put in
relation with an enhanced reactivity; namely, the reactivity increases with increasing degree of curvature
[8]. In graphene, however, the edges are usually considered the most reactive sites as only a few defects
are observed on the highly crystalline lattice [9]. Then, covalent reactions are presumed to occur preferen-
tially at the graphene edges and CNT tips. The chemical attachment of functional moieties on the reactive
sites of graphitic nanostructures is determined by a number of characterization techniques. The quantitative
characterization includes thermogravimetric analysis (TGA), X-ray photo-spectroscopy (XPS), Raman,
UV-vis-nIR and photoluminescence spectroscopies. However, monitoring the spatial localization of the
functional groups covalently attached on the carbon surface has remained a difficult task. In fullerenes,
nuclear molecular resonance (NMR) [10] is used with this aim, whereas higher molecular weight and
structural conformation limit NMR characterization use in CNTs and graphene.
In the present work, we show the straightforward localization by TEM of carboxylic and amino functional
groups covalently grafted to CNTs and graphene by the discriminating attachment of AuNPs on the
introduced chemical functional groups. AuNPs are chosen as contrast markers because they are easy
to synthesize in different sizes, present high chemical stability and their facile surface functionalization
renders them perfect systems for directed self-assembly [11,12]. The identification of the reactive sites
is possible due to the selective control of non-covalent interactions at the molecular level or by inducing
the growth of AuNPs in situ assisted by UV-light irradiation [13]. The localization of the functional
groups grafted to CNTs and graphene layers by transmission electron microscopy (TEM) allows the
identification of the conformational carbon network arrangement towards common organic reactions
used for the construction of molecular functional interfaces [14].

2. Experimental
2.1. Materials and reagents
All solvents and chemicals were purchased from Sigma-Aldrich and used without further purification.
HiPCO single-walled carbon nanotubes (SWCNTs) were obtained from Carbon Nanotechnology, Inc.
(lot no. R0496, www.cnanotech.com). Multi-walled carbon nanotubes (MWCNTs) were purchased from
NANOCYL NC7000 (www.nanocyl.com). Graphite was purchased from Bay Carbon, Inc. (SP-1 graphite
powder, batch no. 04100, lot no. 011705) www.baycarbon.com). Gold (III) chloride hydrate solution
(99.99% trace metal basis, Aldrich) was used as a metal precursor (0.05 mM).

2.2. Characterization techniques

TEM images were acquired on a TEM JEOL JEM-2100, using an accelerating voltage of 200 kV. Samples
were prepared by drop casting of stable dispersions onto a TEM grid (200 mesh, Lacey carbon films).
Raman scattering was measured with a Thermo Scientific DXR Raman microscope equipped with a
diode-pumped solid-state laser at a wavelength of 532 nm as an excitation source. It had a 20 objective
with a 50 mM slit aperture, 7 s of exposure time and laser power of 10 mW. Samples were prepared by
drop casting of the dispersion on silicon oxide surfaces (Si-Mat silicon wafers, CZ) and the solvent was
allowed to evaporate. For Raman analysis 30 spectra were taken of each sample. The chemical composition
of the samples was analysed by X-ray photoelectron spectroscopy (XPS) using an XPSVERSAPROBE PHI
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

5000 from Physical Electronics, equipped with a monochromatic Al Ka X-ray source under UHV conditions. 3
The energy resolution was 0.7 eV. For the compensation of built-up charge on the sample surface during the

rsos.royalsocietypublishing.org
measurements, a dual beam charge neutralization composed of an electron gun (B1 eV) and an argon ion
gun (r10 eV) was used. The XPS spectra were deconvoluted into different chemical surroundings using
commercially available software (CASA-XPS).

2.3. Synthesis of AuNPs

400 mM HAuCl4 in deionized water –methanol solution (1M, 80 ml) was irradiated with UV light (GE. R
500 W Helios Italquartz, UVB, UVA ozone free emission 310–450 nm, lmax ¼ 360 nm) for 60 min under
magnetic stirring. The products were washed by centrifugation to remove all the by-products of the
reaction. The average size distribution is 36 + 9 nm [15].

R. Soc. Open Sci. 5: 180605

2.4. Purified CNTs ( p-SWCNTs and p-MWCNTs)
300 mg of pristine CNTs was sonicated for 5 min in 300 ml of 2.6 M solution of nitric acid and refluxed
for 48 h at 1258C under magnetic stirring. The solution was neutralized, filtered (Millipore JH 0.45 mm
filter) and washed with deionized water. p-CNTs were washed with methanol and dried in vacuum
overnight (249 mg, 83%).

2.5. CNTs oxidation (ox-SWCNTs and ox-MWCNTs)

After purification, p-CNTs devoid of metallic catalytic particles were sonicated in piranha solution for
30 min and washed thoroughly until reaching pH ¼ 4.5 [16].

2.6. Amine functionalized CNTs (f-SWCNTs and f-MWCNTs)

Boc-protected derivatives were prepared as follows: p-CNTs (120 mg) were dispersed and homogenized
in distilled water (100 ml) by sonication. Then 4-[(N-Boc) aminomethyl] aniline (2 g, 8.9 mmol) and
isoamyl nitrite (2 ml, 14.8 mmol) were added to the suspension and the reaction was refluxed at 808C
overnight. The suspension was filtered and then washed with N,N-dimethylformamide (DMF) and
methanol until the solutions were clear of any impurity. The obtained black solid was dried under
vacuum overnight (170.1 mg). The cleavage of the Boc groups was carried out by their dispersion in
4 M hydrochloric acid (HCl) using 1,4-dioxane as the solvent. The reaction was kept stirred at room
temperature overnight and then it was filtered and washed thoroughly with DMF and methanol.
Then the completely cleaned samples were dried (140.1 mg). The presence of amino terminal groups
was verified by using the Kaiser test before and after the de-protection of the Boc group, obtaining an
average of 124 + 54 and 297 + 75 mmol g21 of f-SWCNTs and f-MWCNTs, respectively [17].

2.7. Photodeposition of AuNPs on CNT (AuNP@p-SWCNTs, AuNP@p-MWCNTs,

AuNP@ox-SWCNTs and AuNP@ox-MWCNTs)
10 mg of the p-CNTs or ox-CNTs were sonicated in 10 ml of 1 M solution of methanol in water for
10 min. 1.5 M citric acid solution in water was used to adjust the pH to 1. HAuCl4 was used as metal
precursors (0.05 mM, 400 ml). Water dispersions were deoxygenated with Ar to avoid photo-oxidation
of the tubes. Solutions were irradiated with UV light for intervals of 60 min under magnetic stirring [13].

2.8. Exfoliation of graphite

Few-layer graphene (FLG) was produced by ultrasonication, using an ultrasonic tip processor GEX 750.
Samples were sonicated in cycles of 30/30 s on/off for periods of 3 h, at the lower power of the ultrasonic
tip (20%, 150 W). During ultrasonication, samples were kept in an ice bath to avoid overheating. After
sonication, the FLG was washed and dispersed in DMF [18].

2.9. Graphene oxide synthesis

GO was obtained by oxidizing graphite crystals using the improved Hummers’ method reported by
Marcano et al. [19]. A concentrated mixture of H2SO4 (360 ml) and H3PO4 (40 ml) was added to a
mixture of 3 g of graphite and 18 g of KMnO4. The mixture was stirred for 12 h at 508C. After this,
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

the mixture was filtered using a 0.2 mm polytetrafluoroethylene (PTFE) membrane filter. The filtrate was 4
centrifuged for 5 h at 4000 r.p.m. and the supernatant was discarded; the resultant material was then

rsos.royalsocietypublishing.org
washed with 200 ml of deionized water, 200 ml of 30% HCl and 200 ml of ethanol, and it was filtered
using a PTFE membrane filter. This washing process was repeated three times. The remaining material
was dispersed in 200 ml of ether, and the resulting suspension was filtered again with a PTFE membrane
filter. The solid obtained on the filter was finally dried in vacuum for 24 h.

2.10. Amine functionalized graphene

Boc derivative was prepared as follows: 5 ml of water was added to 25 ml of colloidal graphene
dispersion (DMF/H2O, 5 : 1); the reagents 4-[(N-Boc) aminomethyl] aniline (2 equiv. per FLG) and
isoamyl nitrite (5 equiv. per FLG) were added to the suspension and the reaction was refluxed at

R. Soc. Open Sci. 5: 180605

808C overnight. The resulting solution was cooled down to room temperature, filtered (PTFE membrane,
0.2 mm), washed thoroughly and dispersed for the cleavage of the Boc groups in 4 M hydrochloric acid
(HCl) using 1,4-dioxane as the solvent. The reaction was kept stirred at room temperature overnight and
after that it was filtered and washed thoroughly with DMF. The presence of amino terminal groups was
verified by using the Kaiser test before and after the de-protection of the Boc group, obtaining an average
of 653 + 65 mmol g21 of free amino terminal groups [20].

2.11. Photodeposition of AuNPs on GO (AuNP@GO)

Two milligram of GO was dispersed in 80 ml of a 1 M solution of methanol in water. A citric acid
solution (1.5 M) was used to adjust the pH to 4.5. AuHCl4 was used as a metal precursor (0.05 mM,
400 ml). The solution was irradiated with UV light for 60 min under magnetic stirring.

2.12. AuNP recognition (AuNP@f-SWCNTs, AuNP@f-MWCNTs, and AuNP@f-FLG)

One millilitre of f-CNTs and f-FLG dispersions in fresh DMF were mixed with 1 ml of AuNPs.
Dispersions were allowed to stabilize under magnetic stirring overnight [21].

3. Results and discussion

The lack of solubility of most carbon nanostructures renders difficult their manipulation, imposing limit-
ations to their use. The introduction of functional moieties such as amino, carboxylic or thiol groups on
the surface of carbon nanostructures has aided to solve this problem [22,23]. Amino and carboxylic
groups can be positively or negatively charged, introducing hydrophilic characteristics. In addition, a rela-
tively easy coupling chemistry can be applied on these terminal functions, linking nearly any desired
functionality [24]. In a similar approach, functional small molecules covalently attached to carbon
nanostructures are used to stabilize AuNPs [21]. In this work, terminal amino or carboxylic functional
groups are introduced on CNTs and graphene by applying two chemical reactions: a diazonium-based
arylation reaction and oxidation in strong acids, respectively. Pristine and functionalized systems were
then conjugated with AuNPs in colloidal dispersions and characterized by Raman and XPS spectroscopies
before the visualization of the samples by TEM. The interaction of AuNPs with purified nanomaterials
(p-SWCNTs, p-MWCNTs and FLG) are compared with amino functionalized nanostructures
(f-SWCNTs, f-MWCNTs and f-FLG) and oxidized ones (ox-SWCNTs, ox-MWCNTs and GO). A schematic
representation of the functionalized carbon nanostructures is shown in figure 1. Purified and functionalized
carbon nanostructures were carefully characterized by Raman and XPS spectroscopies.
Raman spectroscopy is used as a powerful tool in the structural characterization of carbon nanomater-
ials. The strongest Raman band of graphitic materials is the G-band (tangential mode at around 1600 cm21);
weaker bands are the D-band (disorder or defect-induced band at around 1350 cm21) and the 2D band at
around 2600–2800 cm21. The intensity of the D-band relative to the G-band is often used as a measure of the
quality of the carbon nanostructure and as an indicator of chemical functionalization. The higher the D-band
intensity, the stronger is the functionalization. In pure graphene the D-band originates from a hybridized
vibrational mode associated with the edges, defects and adatoms [25].
The chemical composition of the samples was analysed by XPS. The overlap between C–O and C–N
features makes difficult the clear identification of the amine funtionalites covalently grafted to the carbon
networks. For this reason, the amount of free NH2 groups in functionalized nanostructures is identified
by the Kaiser test.
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

rsos.royalsocietypublishing.org
p-SWCNTs f-SWCNTs ox-SWCNTs

R. Soc. Open Sci. 5: 180605

p-MWCNTs f-MWCNTs ox-MWCNTs

FLG f-FLG GO
Figure 1. Schematic of carbon nanostrutures.
(a) (b)
intensity (arb. units)

ox-SWCNT
f-SWCMT
p-SWCNT
intensity (arb. units)

295 290 285 280

(c)
intensity (arb. units)

500 1000 1500 2000 2500 3000 295 290 285 280
Raman shift (cm–1) binding energy (eV)
Figure 2. (a) Raman spectroscopy of p-SWCNTs, f-SWCNTs and ox-SWCNTs. (b) C 1s core-level photoemission line for p-SWCNTs.
(c) C 1s core-level photoemission line for ox-SWCNTs.
In figure 2a, the Raman spectra of p-SWCNTs, f-SWCNTs and ox-SWCNTs are displayed. The two
most prominent peaks are at 1584 and 2700 cm21. These signals are attributed to the G and 2D
bands, respectively. The radial breathing modes (RBMs) for SWCNTs appear around 250 cm21.
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

6
(a) (b)

rsos.royalsocietypublishing.org
R. Soc. Open Sci. 5: 180605
(c) (d)

(e) (f)

Figure 3. TEM micrographs of (a) p-SWCNT, (b) AuNP@p-SWCNT, (c) f-SWCNT, (d) AuNP@f-SWCNT, (e) ox-SWCNTs and
(f ) AuNP@ox-SWNCTs.

The signal at 1350 cm21 is associated to the D band. For p-SWCNTs the signal corresponding to the D
band has a low intensity, which indicates a very low amount of defects, while f-SWCNTs and
ox-SWCNTs display broader and higher D band intensities; the ID/IG ratios are 0.01, 0.14 and 0.16,
respectively. In f-SWCNTs an aryl diazonium-based reaction in water introduces amino terminal
groups by the transformation of a diazonium group into a radical capable of grafting covalently to the
CNTs, while oxidation in ox-SWCNTs attacks the surface of the tubes, breaking the graphitic lattice
and introducing carboxylic, ketone, alcohol and aldehyde functionalities among others. The narrowed
2D band of f-SWCNTs accounts for better dispersibility of the tubes compared with p-SWCNTs and
ox-SWCNTs.
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

(a) (b) 7

intensity (arb. units)

ox-MWCNT

rsos.royalsocietypublishing.org
f-MWCNT
p-MWCNT
intensity (arb. units)

295 290 285 280

(c)

intensity (arb. units)

R. Soc. Open Sci. 5: 180605

500 1000 1500 2000 2500 3000 295 290 285 280
Raman shift (cm–1) binding energy (eV)
Figure 4. (a) Raman spectroscopy of p-MWCNTs, f-MWCNTs and ox-MWCNTs. (b) C 1s core-level photoemission line for p-MWCNTs.
(c) C 1s core-level photoemission line for ox-MWCNTs.

In purified samples, shown in figure 2b, the C 1s spectrum presents a main feature at 284.3 eV
and a secondary feature corresponding to a p plasmon excitation centred at 291 eV. The shoulder at
the high-energy side of the C 1s peak for ox-SWCNTs (shown in figure 2c) is generated by photoelectrons
emitted from C atoms in oxygenated groups: hydroxyl (component centred at 286.2 eV), carbonyl
(287.2 eV) and carboxyl groups (288.9 eV). The chemical composition of p-SWNCTs is C (99.3%) and
O (0.7%). For f-SWCNTs, the Kaiser test value of 125 mmol g21 of free NH2 in combination with
the XPS of p-SWCNTs corresponds to a chemical composition of C (99%), O (0.7%) and N (0.15%),
while ox-SWCNTs contains C (84%) and O (16%). As we have previously reported [13], AuNP@ox-
SWNCT shows a reduction in the intensity of the C 1 s satellite peaks due to oxygenated groups. This
signal reduction demonstrates that AuNPs are selectively grown on the oxygen-rich areas forming
C-O-Au bonds.
TEM micrographs of p-SWCNTs, AuNP@p-SWCNTs, f-SWCNTs, AuNP@f-SWCNTs, ox-SWCNTs
and AuNP@ox-SWCNTs are shown in figure 3. Figure 3a shows aggregates of p-SWCNTs formed by
p– p stacking interactions. In figure 3c, f-SWCNTs are dispersed in smaller bundles than those formed
by ox-SWCNTs as shown in 3e. In figure 3b, a small number of AuNPs can be observed on the
p-SWCNTs. At closer inspection, AuNPs appear mostly at the tips, which are in principle regions of
defect abundance (oxygenated groups) [16]. In addition, AuNPs are observed dispersed on the TEM
grid. The lack of defects or functional groups in the surface of the p-SWCNTs results in a very low
amount of AuNPs on the sample. Figure 3c shows a higher density of AuNPs distributed on the
f-SWCNTs’ small bundles. In figure 3f, a similar quantity of AuNPs is observed for ox-SWCNTs,
although the aggregation is larger in the last one. These results are consistent with the Raman ID/IG
analysis. The formation of smaller bundles for f-SWCNTs compared to ox-SWCNTs might result from
the fact that oxidation mainly occurs close to defects, amorphous carbon or on the tips of SWCNTs,
while the organic arylation reaction attacks the aromatic carbon network [26].
Figure 4 shows the Raman spectra of p-MWCNTs, f-MWCNTs and ox-MWCNTs. The Raman spec-
troscopy of MWCNTs is totally different from that of SWCNTs; the RBM region is completely absent
because nested tubes of different ratios are analysed. This fact resulted in the D band at 1350 cm21 as
the most prominent signal for MWCNTs making the Raman analysis difficult; the ID/IG ratios are
1.15, 1.26 and 1.13 for p-MWCNTs, f-MWCNTs and ox-MWCNTs, respectively.
The XPS analysis gives similar information as the one reported for SWCNTs. Additional bands com-
pared with the purified material were obtained for ox-MWCNTs. These bands correspond to hydroxyl
(286.2 eV), carbonyl (287.2 eV) and carboxyl groups (288.9 eV). The composition of p-MWCNTs is C
(99.4%) and O (0.6%). The Kaiser test value of 297 mmol g21 for f-MWCNTs combined with the XPS
of p-MWCNTs relate to a composition of C (99.05%), O (0.6%) and N (0.35%). After the oxidation
treatment, the chemical composition of ox-MWCNTs changes to C (90.8%) and O (9.2%).
TEM micrographs of p-MWCNTs, AuNP@p-MWCNTs, f-MWCNTs, AuNP@f-MWCNTs,
ox-MWCNTs and AuNP@ox-MWCNTs are reported in figure 5. Figure 5a shows large aggregates of
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

(a) 8
(b)

rsos.royalsocietypublishing.org
R. Soc. Open Sci. 5: 180605
(c) (d)

(e) (f)

Figure 5. TEM micrographs of (a) p-MWCNTs, (b) AuNP@p-MWCNT, (c) f-MWCNTs, (d ) AuNP@f-MWCNTs, (e) ox-MWCNT and
(f ) AuNP@ox-MWCNT.

p-MWCNTs; the growth of AuNP on these samples resulted in spherical AuNPs completely separated
from p-MWCNTs, as observed in figure 5b. The lack of defects on the p-MWCNTs prevents the nucle-
ation of the AuNPs on their surface. Instead, f-MWCNTs produced completely dispersed individual
tubes; as observed in figure 5c, the terminal amino functional groups induced the whole stabilization
of the tubes in DMF, allowing the further recognition of AuNPs. In f-MWCNTs, higher degree of funtion-
alization is produced compared with f-SWCNTs; the results are confirmed by the Kaiser test. The higher
functionalization is responsible for the higher dispersibility and the higher amount of AuNPs deposited
on the surface of f-MWCNTs. The strong chemical treatment in ox-MWCNTs caused the oxidation of
impurities in the sample and the addition of functional groups on the surface of the CNTs, like
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

(a) (b) 9

intensity (arb. units)

GO

rsos.royalsocietypublishing.org
f-FLG
FLG
intensity (arb. units)

295 290 285 280

(c)

intensity (arb. units)

R. Soc. Open Sci. 5: 180605

500 1000 1500 2000 2500 3000 295 290 285 280
Raman shift (cm–1) binding energy (eV)
Figure 6. (a) Raman spectroscopy of FLG, f-FLG and GO. (b) C 1s core-level photoemission line for FLG. (c) C 1s core-level
photoemission line for GO.

ketone, carboxyl and hydroxyl groups. This procedure reduces the hydrophobicity of the CNTs, enabling
their dispersion in aqueous media. However, as observed in figure 5e, ox-MWCNTs are less dispersed
than f-MWCNTs; the most probable explanation for this observation is again that oxidation reactions
mostly occur on the amorphous carbon residues, close to defects on the carbon skeleton and on the
tips of the tubes with higher curvature. The ox-MWCNTs show higher concentration of AuNPs distrib-
uted on the tip of the tubes and in the zones where amorphous carbon is still present, as seen in figure 5f.
More images are reported in the electronic supplementary material. Although f-MWCNTs exhibit a
lower functionalization degree than ox-MWCNTs, the first appear as well-dispersed tubes (figure 5c)
completely covered by AuNPs (figure 5d), while the oxidized ones remain as smaller bundles, confirm-
ing that oxidation occurs at the residual amorphous carbon (figure 5f ) mostly close to the defects.
The same trend was observed for SWCNTs.
In figure 6a, the Raman spectra of FLG, f-FLG and GO are diplayed. A clear difference between the
spectra is observed. Two characteristic peaks are located approximately at 1350 and 1590 cm21; these
peaks are attributed to the D and G band, respectively. In the Raman spectra, it is clearly observed
that the peak associated with the D band considerably increases its intensity from FLG to f-FLG and
finally to GO, indicating a higher concentration of functional groups. The ID/IG ratios are 0.12, 0.53
and 1.04 for FLG, f-FLG and GO, respectively. The 2D band is located at approximately 2700 cm21 .
The position and shape of this band are highly sensitive to the number and thickness of the graphene
layers [20]. The signal associated with the 2D band obtained for FLG and f-FLG exhibited well-dispersed
sheets. In GO, the 2D band is completely absent as a result of the strong degree of functionalization
consistent with the broad and intense D band.
The XPS characterization for FLG and GO are reported in figure 6. In FLG (figure 6b), the main peak
at 284.8 eV is attributed to the binding energy of photoelectrons emitted from sp2 C atoms. A secondary
peak appears at 286.2 eV related to C–O and C –N bonds. The chemical composition is C (96.7%),
O (2.5%) and N (0.7%). The presence of N in the sample derives from the non-covalent adsorption of
the solvent (DMF) on FLG produced by the sonication process [20]. In GO, the presence of oxidized
groups increased considerably, with three components corresponding to the following: (1) C–O and
C–N are revealed at 286.2 eV, (2) C¼O bonds shown by the peak at 287 eV and (3) O– C¼O species
at 288.0 eV. As expected, the GO spectrum shows two main peaks due to the high amount of oxygenated
functionalities. In f-FLG, the chemical composition obtained by the combination of the Kaiser test value
and XPS analysis of FLG is C (95.8%), O (2.5%) and N (0.8%). The percentage of N corresponded only to
the covalently attached free NH2 functionalities because the Kaiser test blank was FLG. The chemical
composition is C (69.3%) and O (29.3%).
TEM analysis clearly corroborates the Raman spectroscopy. In figure 7, TEM images of FLG,
AuNP@FLG, f-FLG, AuNP@FLG, GO and AuNP@GO are displayed. FLG relatively free of defects is
observed in figure 7a. The UV-light irradiation produced the growth of AuNPs mainly at the edges of
FLG, as shown in figure 7. At a closer view, the borders of smaller and superficial sheets are responsible
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

10
(a) (b)

rsos.royalsocietypublishing.org
R. Soc. Open Sci. 5: 180605
(c) (d)

(e) (f)

Figure 7. TEM micrographs of (a) FLG, (b) AuNP@FLG, (c) f-FLG, (d ) AuNP@FLG, (e) GO and (f ) AuNP@GO.

for the observation of AuNPs on the inner part of GO. The functionalization with amino terminal groups
of f-FLG induces higher dispersibility of smaller sheets; figure 7c. As observed in figure 7d, AuNPs are
completely dispersed on f-FLG. Finally, the higher oxidation of GO produced the growth of AuNPs on
the whole sheets in the form of aggregated NPs, denoting higher and localized chemical function-
alization. More images are reported in the electronic supplementary material. These results are in
complete agreement with Raman and XPS analysis. TEM images confirm that defects and oxygenated
functions in FLG are mostly distributed at the edges. As for SWCNTs and MWCNTs, pointed covalent
chemical functionalization results in materials with less amount of attached molecules randomly distrib-
uted on the complete surface. The homogeneous distribution of the functionalities allows to obtain
well-dispersed f-FLG in organic solvents. However, aggressive oxidation reactions produce highly
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

oxidized clusters of carbon atoms, resulting in lower dispersibility of GO compared with f-FLG. Impor- 11
tantly, FLG is a more chemically reactive surface compared to SWCNTs and MWCNTs, most probably as

rsos.royalsocietypublishing.org
a consequence of concerted cooperative reactions occurring on the 2D surface [27].

4. Conclusion
In this work, two well-characterized covalent organic reactions, namely diazonium-based arylation reac-
tion and chemical oxidation in solution were successfully performed on SWCNTs, MWCNTs and FLG.
The Kaiser test, and Raman and XPS spectroscopies were used to corroborate the degree of functionali-
zation and chemical composition achieved for each reaction on each type of nanocarbon. We have
demonstrated that the conjugation of the systems with AuNPs allows the localization of the functional

R. Soc. Open Sci. 5: 180605

groups on the carbon frame, giving information of the structural configuration of the material. Pristine
materials show the lowest density of AuNPs, consistent with the low amount of defects and functional
groups mostly concentrated at the tips and edges. Functionalization of carbon nanostructures by a dia-
zonium-based arylation reaction, however, introduces amino terminal groups randomly distributed on
the whole surface of f-SWCNTs, f-MWCNTs and f-FLG, yielding higher dispersibility in polar organic
solvents. Finally, in ox-SWCNTs, ox-MWCNTs and GO, the strong oxidation reaction generates clusters
of oxygenated moieties close to defects, amorphous carbon, tips and edges, resulting in the growth of
AuNPs at localized zones. In oxidized materials, the increase in the amount of AuNPs on carbon nano-
structures is related with the higher concentration of oxygenated moieties as nucleation points for AuNP
growth. The diazonium-based arylation reaction produced nanostructures presenting AuNPs well
dispersed on the whole carbon network, while oxidation reactions caused the aggregation of AuNPs
mostly close to the tips and edges of the carbon lattices where a higher concentration of defects
is expected.
In summary, the final concentration and dispersion of AuNPs on carbon nanostructures is only
dependent on the number of functionalities present in the carbon lattice, but not on the type of
carbon nanostructure. We have shown that TEM characterization of carbon nanostructures conjugated
with AuNPs is an important complementary tool for a better interpretation of the chemical structure
of functionalized carbon nanomaterials. This strategy allows the direct observation and quantification
of AuNPs attached to chemically oxygenated and aminated moieties. It is expected that this intepretation
allows the tailored synthesis of novel functionalized carbon nanostructures for advanced applications in
the biomedicine and energy fields.
Data accessibility. Additional TEM images of AuNP@CNS supporting this investigation have been uploaded as part of the
electronic supporting information.
Authors’ contribution. M.Q. designed the study. V.P-L. and M.C. synthesized the nanohybrids and prepared all the
samples. I.S-O. performed TEM characterization, and collected and analysed the data. C.B. performed XPS analysis.
V.P-L., M.C., C.B., I.S-O. and M.Q. interpreted the results and wrote the manuscript. All the authors revised the
manuscript and gave their final approval for publication.
Competing interests. The authors declared no competing interests.
Funding. This research was supported by CONACYT through the projects I-225984 and PDCPN-1767. V.P-L. and M.C.
thank CONACYT for the PhD and MC scholarships number 326331 and 22481, respectively. C.B. and M.Q. are grateful
to FNRS for financial support U.N.003.18.
Acknowledgements. We are grateful to Dr Aurora Robledo for technical support in Raman spectroscopy.

References
1. Jeantet A, Chassagneux Y, Claude T, Lauret JS, readout. Nat. Nanotechnol. 4, 861– 867. 6. Nel AE, Mädler L, Velegol D, Xia T, Hoek EMV,
Voisin C. 2018 Interplay of spectral diffusion and (doi:10.1038/nnano.2009.267) Somasundaran P, Klaessig F, Castranova V,
phono-broadening in individual photo-emitters: 4. Yin J, Li X, Yu J, Zhang Z, Zhou J, Guo W. Thompson M. 2009 Understanding
the case of carbon nanotubes. Nanoscale 10, 2014 Generating electricity by moving a biophysicochemical interactions at the nano-bio
683–689. (doi:10.1039/c7nr05861f ) droplet of ionic liquid along graphene. Nat. interface. Nat. Mater. 8, 543–557. (doi:10.
2. Graydon O. 2018 Nanowire detectors. Nat. Nanotechnol. 9, 378– 383. (doi:10.1038/nnano. 1038/nmat2442)
Photonics 12, 58. (doi:10.1038/s41566-018- 2014.56) 7. Mackay AL, Terrones H. 1991 Diamond from
0099-2) 5. Wu Y, Cheng G, Katsov K, Sides SW, Wang J, graphite. Nature 352, 762. (doi:10.1038/
3. Changyao C, Rosenblatt S, Bolotin KI, Kalb W, Tang J, Fredrickson GH, Moskovits M, Stucky GD. 352762a0)
Kim P, Kymissis I, Stormer HL, Heinz TF, Hone 2004 Composite mesostructures by nano- 8. Lin T, Zhang WD, Huang J, He C. 2005 A DFT
H. 2009 Performance of monolayer graphene confinement. Nat. Mater. 3, 816–822. (doi:10. study of the amination of fullerenes and carbon
nanomechanical resonators with electrical 1038/nmat1230) nanotubes: reactivity and curvature. J. Phys.
 Downloaded from http://rsos.royalsocietypublishing.org/ on September 5, 2018

Chem. B 109, 13 755– 13 760. (doi:10.1021/ graphene oxide-gold nanoparticles platforms for bulk or graphene edges. Chem. Commun. 47, 12
jp051022g) biosensing applications. Phys. Chem. Chem. 9330 –9332. (doi:10.1039/C1CC13254G)
9. Radovic LR. 2009 Active sites in graphene and Phys. 20, 1685– 1692. (doi:10.1039/ 22. Georgakilas V, Kordatos K, Prato M, Guldi DM,

rsos.royalsocietypublishing.org
the mechanism of CO2 formation in carbon C7CP04817C) Holzinger M, Hirsch A. 2002 Organic
oxidation. J. Am. Chem. Soc. 131, 17 166 – 16. Ziegler KJ, Gu Z, Peng H, Flor EL, Hauge RH, functionalization of carbon nanotubes. J. Am.
17 175. (doi:10.1021/ja904731q) Smalley RE. 2005 Controlled oxidative cutting of Chem. Soc. 124, 760– 761. (doi:10.1021/
10. Milic D, Prato M. 2010 Fullerene unsymmetrical single-walled carbon nanotubes. J. Am. Chem. ja016954m)
bis-adducts as models for novel Soc. 127, 1541 –1547. (doi:10.1021/ja044537e) 23. Georgakilas V, Otyepka M, Bourlinos AB,
peptidomimetics. Eur. J. Org. Chem. 2010, 17. Quintana M, Prato M. 2009 Supramolecular Chandra V, Kim N, Kemp KC, Hobza P, Zboril R,
476–483. (doi:10.1002/ejoc.200900791) aggregation of functionalized carbon nanotubes. Kim KS. 2012 Functionalization of graphene:
11. Grzelczak M, Pérez-Juste J, Mulvaney P, Chem. Commun. 2009, 6005 – 6007. (doi:10. covalent and non-covalent approaches,
Liz-Marzan LM. 2008 Shape control in gold 1039/B915126E) derivatives and applications. Chem. Rev. 112,
nanoparticle synthesis. Chem. Soc. Rev. 37, 18. Quintana M, Tapia JI, Prato M. 2014 Liquid- 6156 –6214. (doi:10.1021/cr3000412)
1783 –1791. (doi:10.1039/B711490G) phase exfoliated graphene: functionalization, 24. Tasis D, Tagmatarchis N, Bianco A, Prato M.

R. Soc. Open Sci. 5: 180605

12. Grzelczak M, Vermant J, Furst EM, Liz-Marzan characterization, and applications. Beilstein 2006 Chemistry of carbon nanotubes. Chem.
LM. 2010 Direct self-assembly of nanoparticles. J. Nanotechnol. 5, 2328. (doi:10.3762/bjnano. Rev. 106, 1105 –1136. (doi:10.1021/cr050569o)
ACS Nano 4, 3591– 3605. (doi:10.1021nn/ 5.242) 25. Ferrari AC, Robertson J. 2004 Raman
100869j) 19. Marcano DC, Kosynkin DV, Berlin JM, Sinitskii A, spectroscopy of amorphous, nanostructured,
13. Quintana M, Ke X, Van Tendeloo G, Meneghetti Sun Z, Slesarev A, Alemany LB, Lu W, Tour JM. diamond-like carbon, and nanodiamond. Phil.
M, Bittencourt C, Prato M. 2010 Light-induced 2010 Improved synthesis of graphene oxide. Trans. R. Soc. Lond. A 362, 2477 –2512. (doi:10.
selective deposition of Au nanoparticles on ACS Nano 4, 4806– 4814. (doi:10.1021/ 1098/rsta.2004.1452)
single-wall carbon nanotubes. ACS Nano 4, nn1006368) 26. Dyke CA, Tour JM. 2003 Unbundled and highly
6105 –6113. (doi:10.1021/nn101183y) 20. Mata-Cruz I, Vargas-Caamal A, Yañez-Soto B, functionalized carbon nanotubes from aqueous
14. Quintana M et al. 2013 Knitting the catalytic López-Valdivieso A, Merino G, Quintana M. 2017 reactions. Nano Lett. 3, 1215– 1218. (doi:10.
pattern of artificial photosynthesis to a hybrid Mimicking rose petal wettability by chemical 1021/nl034537x)
graphene nanotexture. ACS Nano 7, 811– 817. modification of graphene films. Carbon 121, 27. Denis PA, Iribarne F. 2012 Cooperative behavior
(doi:10.1021/nn305313q) 472–478. (doi:10.1016/j.carbon.2017.06.018) in functionalized graphene: explaining the
15. Hernández-Sánchez D, Villabona-Leal G, 21. Quintana M, Montellano A, Del Rio AE, Van occurrence of 1,3 cycloaddition of azomethine
Saucedo-Orozco I, Bracamonte V, Pérez E, Tendeloo G, Bittencourt C, Prato M. 2011 ylides onto graphene. Chem. Phys. Lett. 550,
Bittencourt C, Quintana M. 2018 Stable Selective organic functionalization of graphene 111–117. (doi:1016/j.cplett.2012.08062)