Selectively oxidized carbon nanocatalysts for the oxidation of: Cis -cyclooctene

Article abstract of DOI:10.1039/c7nj03706f

In this work, selectively oxidized carbon nanotubes and graphene flakes were used as carbocatalysts in the mild oxidation of cis-cyclooctene. Firstly, the multi-walled carbon nanotubes and graphene flakes were oxidized using three different oxidation procedures: (i) nitric acid treatment; (ii) nitric acid followed by thermal treatment at 400 °C; (iii) gas phase oxidation with O₂, and subsequently characterized by several techniques. Both oxidized nanomaterials showed an overall increase in the oxygen content and revealed that the different oxidation procedures promoted different superficial chemical compositions: nitric acid treated carbon nanotubes and graphene flakes (respectively, MWCNTh and GFh) presented increased amounts of carboxylic acids, anhydrides and phenols up to 552, 84, and 648 μmol g^-1, respectively, for material MWCNTh, and 612, 120, and 1392 μmol g^-1, respectively, for material GFh; the nanomaterials treated with nitric acid followed by thermal treatment at 400 °C (MWCNTht and GFht) presented a loss of carboxylic acids and an increase in carboxylic anhydride content (36 and 132 μmol g^-1, respectively, for material MWCNTht, and 36 and 156 μmol g^-1, respectively, for material MWCNTht), when compared to the nitric acid treated materials MWCNTh and GFh; and finally the gas phase treated nanomaterials, MWCNTo and GFo, presented an increase in the amounts of phenols, carbonyls/quinones, and lactones up to 720, 216, and 144 μmol g^-1, respectively, for material MWCNTo, and 804, 420, and 144 μmol g^-1, respectively, for material GFo. The carbon nanocatalysts showed activity in the selective oxidation of cis-cyclooctene to epoxycyclooctane using mild conditions: hydrogen peroxide was used as oxidant at 80 °C. The modifications introduced in the pristine materials through nitric acid followed by thermal treatment at 400 °C originated the nanocatalysts with the best activity and selectivity: MWCNTht presented 47% of substrate conversion and 79% epoxycyclooctane selectivity whereas GFht showed 57% conversion and 85% epoxycyclooctane selectivity. Reusability experiments revealed no loss of catalytic activity up to three catalytic cycles. The results indicate that the fine tuning of the carbon nanomaterials through appropriate oxidation procedures is crucial in achieving metal-free carbocatalysts with enhanced performance and selectivity in selective oxidation reactions.

Full text of DOI:10.1039/c7nj03706f

View Article Online
View Journal
NJC
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: C. Freire, B.
Jarrais and A. Guedes, New J. Chem., 2018, DOI: 10.1039/C7NJ03706F.
Volume 40 Number
1
January 2016 Pages 1–846
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
NJC
New Journal of Chemistry
www.rsc.org/njc
A journal for new directions in chemistry
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1144-0546
PAPER
Jason B. Benedict et al.
The role of atropisomers on the photo-reactivity and fatigue of
diarylethene-based metal–organic frameworks
rsc.li/njc

Page 1 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
Selectively oxidized carbon nanocatalysts for the
oxidation of cisꢀcyclooctene
[a]
[b]
[a]
Bruno Jarrais, Alexandra Guedes, and Cristina Freire*
[
a]
REQUIMTE/LAQV, Departmento de Química e Bioquímica
Faculdade de Ciências da Universidade do Porto
Rua do Campo Alegre s/n, 4169ꢀ007 Porto (Portugal)
Corresponding author (C. Freire): Eꢀmail: acfreire@fc.up.pt
[
b]
Instituto de Ciências da Terra, Pólo da FCUP and Departamento de Geociências,
Ambiente e Ordenamento do Território
Faculdade de Ciências da Universidade do Porto
Rua do Campo Alegre 687, 4169ꢀ007 Porto, Portugal
Keywords: oxidized carbon nanotubes; oxidized graphene flakes, alkene catalytic
epoxidation.
1

New Journal of Chemistry
Page 2 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
Abstract
In this work, selectively oxidized carbon nanotubes and graphene flakes were
used as carbocatalysts in the mild oxidation of cisꢀcyclooctene. Firstly, the multiꢀwalled
carbon nanotubes and graphene flakes were oxidized through three different oxidation
procedures: (i) nitric acid treatment; (ii) nitric acid followed by thermal treatment at 400
C; (iii) gas phase oxidation with O
2
and subsequently characterized by several
techniques. Both oxidized nanomaterials showed an overall increase in the oxygen
content and revealed that the different oxidation procedures promoted different
superficial chemical compositions: nitric acid treated carbon nanotubes and graphene
flakes (respectively, MWCNTh and GFh), presented increased amounts of carboxylic
ꢀ1
acids, anhydrides and phenols up to 552, 84, and 648 ꢁmol g , respectively, for
ꢀ1
material MWCNTh, and 612, 120, and 1392 ꢁmol g , respectively, for material GFh;
the nanomaterials treated with nitric acid followed by thermal treatment at 400 ˚C
(
MWCNTht and GFht), presented a loss of carboxylic acids and an increase in
ꢀ1
carboxylic anhydrides content (36 and 132 ꢁmol g , respectively, for material
ꢀ1
MWCNTht, and 36 and 156 ꢁmol g , respectively, for material MWCNTht), when
compared to the nitric acid treated materials MWCNTh and GFh; and finally the gas
phase treated nanomaterials, MWCNTo and GFo, presented an increase in the
ꢀ
amounts of phenols, carbonyls/quinones, and lactones up to 720, 216, and 144 ꢁmol g
1
ꢀ1
,
respectively, for material MWCNTo, and 804, 420, and 144 ꢁmol g , respectively, for
material GFo. The carbon nanocatalysts showed to be active in the selective oxidation
of cisꢀcyclooctene to epoxycyclooctane using mild conditions: hydrogen peroxide as
oxidant at 80 ˚C. The modifications introduced in the pristine materials, through nitric
acid followed by thermal treatment at 400 ˚C, originated the nanocatalysts with best
activity and selectivity: MWCNTht presented 47% of substrate conversion and 79%
epoxycyclooctane selectivity whereas GFht showed 57% conversion and 85%
epoxycyclooctane selectivity. Reusability experiments revealed no loss of catalytic
2

Page 3 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
activity up to three catalytic cycles. The results indicate that the fine tuning of the
carbon nanomaterials through appropriate oxidation procedures, is crucial in achieving
metalꢀfree carbocatalysts with enhanced performance and selectivity in selective
oxidation reactions.
3

New Journal of Chemistry
Page 4 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
1
. Introduction
Epoxide formation is a central reaction in the oxyfunctionalization of many
molecules and is of great importance in the present era of increasing industrial
1
production demands, due to the versatility of epoxides as intermediates in synthesis.
For example, it has been reported that 4.5 million tons of propylene oxide and 70000
2
tons of butene oxide are produced yearly. The established catalytic processes used in
chemical industries use metals, in many cases noble metals, or metal oxides as
catalysts, which are often expensive, difficult to remove, toxic, and are frequently
3
obtained from limited natural resources. Among the oxidation of alkenes, the selective
oxidation of cyclooctene is of major relevance. Recently, various types of
heterogeneous catalysts have been used to improve the catalytic ability for the
4
5
selective oxidation of cyclooctene, including Au/C, Au/graphite, mesoporous mixed
6
7
oxides (Ga, Ga–Nb and Ga–Mo) , CuO/mesoporous SiO
2
nanocomposites , and
sandwichꢀtype polyoxometallates with various transition metals in the central belts
8
(
Ni,Co, Mn, Zn, Fe). Despite these efforts, the search for a catalyst that combines the
toxicological benefits of a metalꢀfree synthesis with the convenience of heterogeneous
workup, whilst retaining high activity and selectivity, is of critical importance.
In the quest for a system that can satisfy such demands, carbon materials appear
as excellent candidates, given their elemental natural abundance and the fact that
organic chemistry can provide the tools to modify its reactivity. Traditionally, carbon
materials have been used as supports for catalysts in heterogeneous catalytic
processes, although their use as catalysts on their own is becoming more and more
9
common, earning them the term “carbocatalysts”. These carbocatalysts have proved
to be efficient in a wide variety of reactions, which are beyond the scope of this work.
Among the large variety of carbon materials, in this work we will focus on multiꢀwalled
carbon nanotubes and graphene flakes. Specifically regarding oxidation of alkenes, it is
worth mentioning the work by Yu et al., which demonstrated that multiꢀwalled carbon
4

Page 5 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
nanotubes can be used as a metalꢀfree catalyst for the aerobic oxidation of
1
0
cyclohexene, and the pioneer work by Bielawski et al., which demonstrated the
efficient oxidation of various alcohols and cisꢀstilbene, and the hydration of various
1
1
alkynes, catalyzed by graphene oxide (GO).
Pristine MWCNT and graphene are not very reactive as the active sites are found
essentially at the edges of the nanotubes or graphene layers. The introduction of
oxygen functional groups by oxidative treatments, in either gasꢀ or liquidꢀphase, can
increase the number of active sites in various reactions. The concentration and type of
surface groups plays a very important role on determining the surface chemistry of the
carbon materials.
Herein, we report the modification of multiꢀwalled carbon nanotubes and
graphene flakes, through different oxidation procedures, in order to produce carbon
nanomaterials with distinct surface chemistry, and to further understand how their
chemistry affects their performance as mild and efficient catalysts for the selective
oxidation of cisꢀcyclooctene. Furthermore, we use hydrogen peroxide as the oxidant in
the catalytic system, which is a great advantage to the environment and industry
because of the simplicity in handling, the environmental friendliness with water
generation as a solo byꢀproduct, the high content of active oxygen species and the
1
2
cheapness compared to organic peroxides and peracids.
2
. Experimental
2.1 Materials and solvents
All the reagents and solvents used during the experimental execution of this work
were used without further purification. Multiꢀwalled carbon nanotubes (MWCNT) were
commercially obtained from Nanocyl S.A., Ref. 3100 MWCNT (>95 % carbon purity;
9.5 nm average diameter). Commercial graphene flakes (sample denoted as GF) was
from Graphene Technologies (Lot #GTXꢀ7/6ꢀ10.4.13). Cisꢀcyclooctene (95 %),
5

New Journal of Chemistry
Page 6 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
chlorobenzene (99.5 %), and hydrogen peroxide solution (30 % w/w) were from Sigmaꢀ
Aldrich. Nitric acid (65 %) was from Merck, and acetonitrile (HPLC grade) was from
Fisher Scientific. Gases used in the modification of the carbon materials were from Air
2
Liquid: N (99.999 %) and air (99,999 %).
2.2. Preparation of the modified carbocatalysts
1
3
The oxidation procedures were adapted from the work of Figueiredo et al. The
pristine multiꢀwalled carbon nanotubes (MWCNT) and graphene flakes (GF) were
submitted to three different treatments – acid oxidation, acid oxidation followed by
thermal treatment at 400 ˚C and oxidation in the gas phase (5 % O
2 2
/N ) at 500 ˚C –
described as follows and shown on Figure 1. (i) Oxidation with HNO
3
: About 3 g of
3
MWCNT or GF were refluxed in HNO 7 M (150 cm ) for 3 h, washed with distilled
3
water to neutral pH, dried in an oven at 110 ˚C for 24 h and stored in a desiccator for
later use (MWCNTh or GFh). (ii) Thermal treatment: About 1.5 g of MWCNTh or GFh
were subsequently heated to 400 ˚C at 5 ˚C/min in a fusedꢀsilica tubular reactor, under
3
ꢀ1
nitrogen flow (100 cm min ), and kept at this temperature for 1 h. The resulting
materials (MWCNTht or GFht) were cooled to room temperature under the same
atmosphere and stored in a desiccator. (iii) Gas phase oxidation: About 1.5 g of
2 2
MWCNT or GF were oxidized at 500 ˚C (5 ˚C/min) under a 5% O /N atmosphere (100
3
ꢀ1
cm min ), for 3 h in a fusedꢀsilica reactor, cooled to room temperature under nitrogen
atmosphere and stored in a desiccator (MWCNTo or GFo).
6

Page 7 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
Figure 1. Schematic representation of the oxidation procedures.
2
.3 Physicochemical characterization of the modified carbocatalysts
Xꢀray photoelectron spectroscopy (XPS) was performed at ‘‘Centro de Materiais
da Universidade do Porto’’ (CEMUP, Porto, Portugal) in a VG Scientific ESCALAB
00A spectrometer using nonꢀmonochromatised Al K radiation (1486.6 eV). The
2
α
powdered samples were pressed into pellets prior to the XPS measurements. To
correct possible deviations caused by electric charge of the samples, the C1s band at
1
4,15
284.6 eV was taken as internal standard.
The XPS spectra were deconvoluted with
the CasaXPS software, using nonꢀlinear least squares fitting routine after a Shirleyꢀtype
background subtraction. The surface atomic percentages were calculated from the
corresponding peak areas and using the sensitivity factors provided by the
manufacturer.
Temperature programmed desorption (TPD) was performed at “LCM –
Laboratório de Catálise e Materiais”, Faculdade de Engenharia da Universidade do
7

New Journal of Chemistry
Page 8 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
Porto, in an Altamira Instruments device AMIꢀ200. Typically, samples of 100 mg were
ꢀ1
3
ꢀ1
heated at 5 ˚C min up to 1100 ˚C, under a flow of He (25 cm min ); the amounts of
CO (m/z = 28) and CO (m/z = 44) released from the carbon samples were monitored
2
using a Dymaxion 200 AMU Ametek mass spectrometer.
Fourier transform infrared spectroscopy (FTIR) spectra of the compounds were
ꢀ1
collected with a Jasco FT/IRꢀ460 Plus spectrophotometer in the range 400–4000 cm ,
ꢀ1
using a resolution of 4 cm and 32 scans. The spectra of the dried samples were
obtained in KBr pellets (Merck, spectroscopic grade) containing 0.04 wt. % of carbon
material.
The Raman spectra were measured directly on the materials using a Labram
Dilor–JobinꢀYvon–Spex spectrometer, with the monochromatic green light at 532 nm of
the Nd:YAG laser as the excitation source, at a power of 25 mW. The excitation laser
frequency was maintained for all measurements, as the D band is a resonant mode
and exhibits dispersive behavior, hence both its position and shape can vary with
different excitation laser frequencies. A density filter at 75% was used to avoid thermal
decomposition of samples by the laser. The ×50 objective lens of an Olympus optical
microscope was used to focus the laser beam on the sample and to collect the
scattered radiation. A highly sensitive CCD camera was used to collect the Raman
spectra. Each Raman spectrum presented in this study corresponds to the
ꢀ1
accumulation of 10 spectra recorded from 1000 to 3400 cm over 10 s.
Transmission electron microscopy (TEM) images were collected at “Universidade
de TrásꢀosꢀMontes e Alto Douro”, with a Zeiss LEO 906 microscope operating at an
accelerating voltage of 120 kV. The samples were dispersed in dimethylformamide
under sonication, after which a carbonꢀcoated 400 mesh copper grid was immersed in
the suspension and then airꢀdried.
2
Textural characterization of the samples was carried out by N adsorption–
desorption at ꢀ196 ˚C in a Quantachrome NOVA 4200e apparatus, at “LCM –
Laboratório de Catálise e Materiais”, Faculdade de Engenharia da Universidade do
8

Page 9 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
Porto. The apparent surface area (SBET) was determined by applying the Brunauer–
16
Emmett–Teller (BET) equation.
2.4 Catalytic studies of the modified carbocatalysts in the selective oxidation of
cisꢀcyclooctene
The oxidation reaction of cisꢀcyclooctene was carried out in acetonitrile (MeCN),
3
using a borosilicate 50 cm reaction vessel, with addition of H
2
O
2
(30 % w/w), in the
presence of the carbon materials as heterogeneous catalysts, at 80 ˚C. In a typical
3
experiment, cisꢀcyclooctene (2.5 mmol) was dissolved in MeCN (25 cm ) and stirred.
Afterwards, 50 mg of carbon catalyst was added. The reaction was started with the
2 2
addition of 281 ꢁL of H O solution (2.75 mmol). The mixture was stirred at 80 ˚C for 48
hours. The reaction progress was followed by gas chromatography in a Thermo
Scientific Trace 1300 gas chromatograph, equipped with a Teknokroma TRBꢀ5 column
(
30 m × 0.25 mm × 0.25 ꢁm) and a flame ionization detector (FID). The mobile phase
was a mixture of air (21 % O in 79% N ) and N (>99,998 %), and the detector fuel gas
2
2
2
was hydrogen (>99,95 %). The oven temperature program was: start at 80 ˚C, 20 ˚C
ꢀ1
min , 200 ˚C (8 min). At 24 hours intervals, an aliquot was taken directly from the
3
reaction mixture with a hypodermic syringe (approximately 0.1 cm ), filtered through a
0
.2 ꢁm Nylon membrane filter, and 1 ꢁL was injected into the chromatograph, with a 10
L microsyringe, for analysis of starting materials and products. The reaction products
ꢁ
reported were identified by mass spectrometry coupled with gas chromatography, in a
Thermo Scientific Trace 1300 chromatograph equipped with a TGꢀ5MS Thermo
Scientific column (60 m × 0.25 mm × 0.25 ꢁm) and a MS ISQ quadrupole operating in
electron ionization mode (70 eV). At the end of reactions, the catalysts were filtered,
washed with MeCN several times to remove the remaining substrate, reaction products
and oxidant. The recovered catalysts were dried at 100 ˚C overnight under vacuum and
reused in a new reaction under identical experimental conditions, with readjustment of
9

New Journal of Chemistry
Page 10 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
all quantities, without changing the molar ratios and reaction concentrations. A blank
reaction without catalyst was also performed.
3
. Results and discussion
3.1 Carbon nanomaterials characterization
In order to study the presence of oxygen functional groups in the oxidized carbon
materials, FTIR spectra were obtained (Figure 2).
a
e
f
b
c
d
g
h
4000 3500 3000 2500 2000 1500 1000 500
4000 3500 3000 2500 2000 1500 1000 500
ꢀ
1
ꢀ1
Wavenumber (cm )
Wavenumber (cm )
Figure 2. FTIR spectra of pristine MWCNT (a), MWCNTh (b), MWCNTht (c), MWCNTo
(
d), GF (e), GFh (f), GFht (g), and GFo (h).
Because of the extensive presence of πꢀπ interactions in the pristine materials,
their FTIR spectra present very weak and poorly resolved vibration bands. The bands
observed in the original MWCNT and GF are assigned to the vibrations of the materials
ꢀ1
structure and of carbon–oxygen groups from residual oxidation: the band at 1560 cm
is assigned to C=C stretching vibrations of the aromatic carbon in nonꢀgraphitic
1
7
ꢀ1
domains, the broad band at ~1130 cm has contributions from the skeletal C–C
17
tangential motions (also observed in other carbon materials, such as C60
some C–O stretching vibrations of oxygen groups. Finally, the band at 3430 cm
corresponds to OH stretching vibrations from hydroxyl groups. The FTIR spectra of the
)
and from
ꢀ1
1
0

Page 11 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
oxidized carbon materials present more resolved vibration bands, since the oxidation
introduces some disruption of the electronic conjugation and leads to the rupture of
some π–π interactions. In addition to the three vibrational bands stated above, these
ꢀ1
spectra also show a typical vibrational band of OH groups at 1340 cm , due to the O–H
ꢀ1
bending mode, and a vibrational band at 1740 cm due to the C=O stretching mode of
18–20
carboxylic acid, carbonyl, ketone and quinone groups.
Also worth of mention is the
ꢀ1
shoulder at 1220 cm , which is considerably more intense in the spectra of the
oxidized carbon materials, when compared with the pristine materials (negligible in the
21
case of GF), which suggests the introduction of new phenol/hydroxyl groups.
XPS analysis was carried out to assess the type and relative amount of functional
groups on the pristine and oxidized carbon materials. The relative amounts of the
different elements, C, O and N, were calculated from the corresponding peak areas
and are shown in Table 1.
Table 1. XPS surface atomic percentages and C/O atomic ratios for the carbon
materials.
%
At
Material
C/O ratio
C1s
98.7
94.5
96.6
97.4
92.9
91.5
93.9
95.3
O1s
1.3
5.4
3.2
2.6
6.8
8.2
5.8
4.5
N1s
ꢀ
MWCNT
MWCNTh
MWCNTht
MWCNTo
GF
75.9
17.2
30.2
37.5
13.7
11.2
16.2
21.2
0.1
0.2
ꢀ
0.3
0.3
0.3
0.2
GFh
GFht
GFo
Pristine MWCNT is composed by 98.7 % of carbon and 1.3 % of oxygen,
whereas GF is composed by 92.9 % of carbon and a considerable amount of oxygen ꢀ
6.8 % ꢀ and 0.3 % of residual N. After the oxidation treatments applied to MWCNT, the
amount of oxygen increased in all the materials, with the percentage varying between
1
1

New Journal of Chemistry
Page 12 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
the lowest value of 2.6 % for MWCNTo and the highest value of 5.4 % for MWCNTh.
Material MWCNTht presents an oxygen content of 3.2 %, as it results from a thermal
treatment on material MWCNTh, at 400 ˚C, under an inert atmosphere. Under these
conditions several deoxygenation reactions occur, which explain this oxygen content
decrease. C/O atomic ratios have considerably decreased, as a consequence of the
oxygen percentage increase, for MWCNTh, MWCNTht, and MWCNTo (C/O ratios of
17.2, 30.2, 37.5 and 13.0, respectively), when compared with the C/O atomic ratio of
MWCNT (C/O ratio of 75.9). As for the graphene flakes, there was a modest increase
in oxygen content for material GFh (8.2 %), oxidized with nitric acid, when compared
with the pristine GF (6.8 %). Subsequent oxygen amount decrease in GFht (5.8 %)
follows the same trend observed for the carbon nanotubes, as the thermal treatment
employed was identical. Interestingly, material GFo, oxidized in the gas phase (5 %
2 2
O /N at 500 ˚C), revealed a decrease in the oxygen content (4.5 %). C/O atomic ratios
for the oxidized graphene flakes remained close to that of the starting material GF (C/O
ratio of 13.7), with material GFh showing a slight decrease (C/O ratio of 11.2), and
materials GFht and GFo revealing an increase upon the different modification
procedures (C/O ratios of 16.2 and 21.7, respectively). These results indicate a milder
surface oxidation extent in the graphene flakes when compared to the carbon
nanotubes counterparts. However, it should be noted the higher oxygen content of the
starting material GF, which is similar to values found for reduced graphene oxides
22,23
prepared by the Hummers method (6.0ꢀ13.5).
The relative atomic percentages of carbonꢀbased functional groups, which were
determined based on the deconvolution of C1s high resolution spectra, are shown in
Table 2, and the deconvoluted C1s and O1s high resolution spectra of carbon
nanotubes and graphene flakes are presented in Figures S1 and S2, respectively. The
profile of the high resolution C1s spectra for all the studied materials was similar to
each other, with a predominant asymmetric peak, and a smaller contribution of C in
surface oxygen functional groups on the high energy side of the main peak. For the
1
2

Page 13 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
deconvolution of the C1s spectra, it has been generally considered an asymmetric
peak for carbon atoms in polyaromatic structures, and the presence of the various
types of oxygen groups.
Table 2. Relative atomic percentages of carbonꢀbased functional groups presented in
the XPS high resolution C1s spectra.
%
C
2
84.6 eV
Material
2
86.3 eV
287.4 eV
(C=O)
288.9 eV
(COOH)
290.3
(πꢀπ*)
(
graphitic
(
CꢀO)
structure)
MWCNT
MWCNTh
MWCNTht
MWCNTo
GF
74.5
5.9
5.6
4.6
5.1
6.6
6.0
5.6
5.1
2.8
5.1
6.7
6.1
4.8
5.8
6.4
6.7
1.2
2.7
3.2
3.1
0.9
3.1
3.0
2.8
15.5
13.3
10.2
11.3
12.6
10.4
9.2
73.3
75.3
74.4
75.2
74.6
75.8
75.5
GFh
GFht
GFo
9.8
Therefore, the C1s spectra were deconvoluted into five components corresponding to
graphite (284.6 eV), carbon in single CꢀO bonds as ethers, alcohols, phenols (286.3
eV), carbon in carbonyl groups (287.4 eV), carbon in carboxyl or ester groups (288.9
2
4,25
eV) and the πꢀπ* shakeꢀup peak (290.3 eV).
All the oxidized carbon materials
presented a decrease of the πꢀπ* satellite peak, indicating that the oxidation procedures
gave rise to a decrease in the electronic delocalization. In fact, there is a similar trend
in both multiꢀwalled carbon nanotubes and graphene flakes: a decrease in electronic
delocalization upon nitric acid oxidation (MWCNTh and GFh), followed by a further
decrease upon thermal treatment at 400 ˚C, under an inert atmosphere (MWCNTht and
GFht). Inversely, the oxygen amount does not accompany this trend: while it increases
upon nitric acid oxidation, after the referred thermal treatment, there is a decrease in
oxygen content in both materials, indicating that the thermal treatment did not promote
the rearrengement of further graphitic domains. As for the gas phase oxidized
1
3

New Journal of Chemistry
Page 14 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
materials, MWCNTo and GFo, there is also a decrease in the electronic delocalization
when compared to their pristine counterparts. These observations show that in both
oxidation procedures there is a decrease in the electronic delocalization of the
materials. The peaks corresponding to carbon atoms bound to oxygen (CꢀO, C=O, and
OꢀC=O) present similar variations upon nitric acid oxidation for both carbon nanotubes
and graphene flakes. There is an increase in the relative areas of the peaks
corresponding to C=O and OꢀC=O, and a slight decrease in the relative area of the
peak assigned to CꢀO, which indicates the formation of new carboxilic acids, carboxylic
anhydrides, carbonyls and esters groups in MWCNTh and GFh. After the thermal
treatment applied to materials MWCNTh and GFh, there is a decrease in the relative
area of the peak assigned to CꢀO for both materials (MWCNTht and GFht), while the
relative area of the peak assigned to OꢀC=O increases in MWCNTht and remains
roughly the same in GFht. Taking into account the decrease in oxygen content
observed after the thermal treatment, these results show a rearrengement of the
surface oxygen functional groups relative contents. In the gas phase oxidations, a
similar trend to nitric acid oxidation is observed for the both carbon nanotubes
(
MWCNTo) and the graphene flakes (GFo) counterpart. In these materials, there is an
increase in the relative areas of the peaks assigned to C=O and OꢀC=O, and a
decrease in the relative area of the peak due to CꢀO, indicating that this treatment
induced the formation of new carboxilic acids, carboxylic anhydrides, carbonyls and
esters groups. The relative atomic percentages of oxygenꢀbased functional groups
obtained from the deconvolution of the O1s high resolution spectra are presented in
Table 3. The broad band in the O1s high resolution spectra was deconvoluted into four
2
5
peaks in all materials: a peak at 531.3 eV, assigned to O=C in ketones, quinones and
aldehydes, a peak at 532.7 eV, associated to O–C from phenol and epoxy groups, a
peak at 533.9 eV due to O–C=O in esters and carboxylic acids, and a peak at 536.0,
assigned to oxygen atoms in water molecules. Analysis of the relative percentages of
each type of oxygen types shown in Table 3 and in Figure 3 (normalized without taking
1
4

Page 15 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
into account the area of the peak atributted to water) shows an increase in the
percentage of the O–C=O components, in comparison to OꢀC component, for all
materials, revealing the significant introduction of acidic groups such as carboxylic
acids, promoted by the different oxidation treatments.
Table 3. Relative atomic percentages of oxygenꢀbased functional groups presented in
the XPS high resolution O1s spectra.
%
O
Material
5
31.3 eV (O=C)
532.7 eV (OꢀC)
46.8
533.9 eV (OꢀC=O) 536.0 eV (HꢀOꢀH)
MWCNT
MWCNTh
MWCNTht
MWCNTo
GF
22.4
19.5
32.7
35.4
30.0
12.6
31.9
30.7
26.1
11.3
2.4
2.7
3.0
6.7
7.6
7.7
7.4
31.2
33.7
36.9
25.0
36.4
30.6
34.0
46.7
GFh
29.7
30.8
GFht
32.7
28.9
GFo
37.6
28.9
The relative percentage of the O=C component for the carbon nanotube
materials also show an increase upon oxidation, while for the graphene flake materials
it shows a modest increase in GFo, and a slight decrease in GFh and GFht. Taking into
account the O=C/OꢀC ratio, this indicates that the introduction of ketones, quinones
and aldehyde functional groups in the carbon materials, was preferred over the
formation of groups such as epoxides, hydroxyls and phenols (albeit to a greater extent
in the carbon nanotube materials). These tendencies indicate that the type and the
amount of oxygen groups introduced onto the carbon surface are highly dependent on
the oxidation procedure used.
1
5

New Journal of Chemistry
Page 16 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
MWCNT
GF
50
40
30
20
10
MWCNTh
MWCNTht
MWCNTo
GFh
GFht
GFo
5
4
3
2
0
0
0
0
O=C
OꢀC
OꢀC=O
O=C
OꢀC
OꢀC=O
Oxygen type
Oxygen type
Figure 3. XPS relative atomic percentage of oxygen for each contribution of O1s peak
components (normalized with the withdrawal of the peak assigned to water molecules).
It is also noteworthy that the percentages depicted in Figure 3 follow the same
tendencies for both the carbon nanotube and graphene flake materials, indicating that
the oxidation procedures proceeded similarly, regardless of the nature of the initial
carbon material.
TPD was used to estimate the content and nature of oxygen functional groups
present in the pristine and modified carbon materials. Figure 4 shows the TPD profiles
of carbon nanotube and graphene flake materials, and the amounts of CO and CO
2
released obtained by integration of the areas under the respective TPD profiles,
together with the total mass percentage of oxygen are presented in Table 4.
1
6

Page 17 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
0
0
0
.5
.4
.3
(
a)
0.2
(b)
0
0
.1
.0
0.2
.1
0
0.0
MWCNT
MWCNT
MWCNTh
MWCNTht
MWCNTo
MWCNTh
MWCNTht
MWCNTo
0
200
400
600
T (ºC)
800
1000
0
200
400
600
800
1000
T (ºC)
0
.5
.4
.3
0.2
.1
0.0
(
c)
0.2
(d)
0
0
0
.1
.0
0
0
GF
GF
GFh
GFht
GFh
GFht
GFo
GFo
0
200
400
600
800
1000
0
200
400
600
800
1000
T (ºC)
T (ºC)
Figure 4. TPD profiles of carbon nanotube and graphene flake materials: (a), (c) ꢀ CO
2
evolution and (b), (d) ꢀ CO evolution.
All the carbon material TPD profiles reveal a substantial increase in the amounts
of released CO and CO, upon the different oxidation procedures, the only exception
being the CO amount released by material GFo, which released less CO than pristine
2
2
2
GF. Moreover, when comparing the total mass percentage of oxygen obtained from the
integration of the TPD profiles with the percentages obtained from XPS data, one
clearly indentifies the same patterns in the oxygen content variations: a large increase
upon nitric acid oxidation, followed by a subsequent decrease upon thermal treatment
at 400 ˚C.
1
7

New Journal of Chemistry
Page 18 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
2
Table 4. Total amounts of CO and CO released by TPD and total mass percentage of
oxygen obtained from TPD data (assuming that all the surface oxygen is released as
CO and/or CO ).
2
ꢀ
1
ꢀ1
Material
MWCNT
MWCNTh
MWCNTht
MWCNTo
GF
[CO
2
] ꢁmol g
[CO] ꢁmol g
228
% O
0.71
4.11
2.38
2.09
2.69
5.80
3.86
2.90
% O (XPS)
1.72
108
696
1076
7.07
240
1008
4.22
180
948
3.43
312
1056
8.85
GFh
852
1920
10.63
7.58
GFht
336
1740
GFo
216
1380
5.90
However, for the gas phase treated materials, the TPD data shows a increase in
the oxygen mass percentage for both MWCNTo and GFo materials, while the XPS
data shows an increase for material MWCNTo and a decrease for material GFo. Sure
enough, the oxygen mass percentages obtained from the TPD data, are all nearly
halved when compared to their XPS counterparts, indicating that the location of the
oxygen functional groups within the carbon materials lies preferably at the outer
surfaces of the materials, e.g., at the carbon nanotube edges and outer walls and
graphene flake edges and exposed basal planes. Interestingly, the TPD profiles are
quite similar between the two different types of carbon materials. Nitric acid treated
materials MWCNTh and GFh, clearly show two distinct peaks in the 250 – 500 ˚C
2
region of the CO TPD profile, a temperature range ascribed to the decomposition of
carboxylic acid groups, corroborating the XPS results that revealed a high proportion of
carboxylic acids. All the materials present peaks in the 550 – 650 ˚C and 650 – 750 ˚C
2
regions of the CO TPD profile, which are assigned to carboxylic anhydrides and
lactones, respectively. In the CO TPD profiles it is possible to identify a large peak in
the 700 – 800 ˚C region, assigned to the decomposition of phenol groups. It is also
possible to observe two shoulders on this peak: one at lower temperatures (550 – 650
1
8

Page 19 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
˚
C), corresponding to the decomposition of carboxylic anhydrides, and another at
higher temperatures (850 – 1000 ˚C), ascribed to the decomposition of carbonyl and
quinone groups. For some materials, namely MWCNT, MWCNTh, MWCNTht, and
GFh, it is possible to observe that there is a small CO evolution in the low temperature
region 200 – 450 ˚C. The identification of oxygen functional groups that decompose in
this temperature range is still a subject of debate. However, there is a growing
consensus that the decomposition of epoxyl and hydroxyl functionalities occur in this
2
5–27
temperature range.
In order to estimate the amount of each oxygen surface group,
and CO TPD profiles was performed. A multiple Gaussian
the deconvolution of the CO
2
function was used for fitting each of the TPD profiles, and the numerical calculations
were based on a nonlinear routine, which minimized the square of the deviations, using
the LevenbergꢀMarquardt method to perform the iterations. The deconvoluted TPD
profiles of the carbon nanotube materials and graphene flake materials are shown in in
the Supplementary Information, in Figures S3 and S4, respectively, and the detailed
data obtained from the deconvolutions is shown in Tables S1 and S2 for the carbon
nanotube materials, and in Tables S3 and S4 for the graphene flake materials. For the
13,26–28
TPD profiles deconvolution, the following assumptions were adopted:
a) The CO profile results from the contribution of peak #1, peak #2, peak #3
(
2
and peak #4, attributed to strongly carboxylic acids, weakly acidic carboxylic acids,
carboxylic anhydrides and lactones, respectively; the same width at halfheight (W) was
imposed for peaks #2, #3 and #4;
(
b) Materials GF and GFht present a fifth peak in the CO
00 ˚C, attributed to lactones located in energetically different sites and its width at
halfheight (W) was imposed the same as for peaks #2, #3 and #4;
c) CO spectra are decomposed into five components: peak #1, attibuted to the
2
profile peak at around
9
(
decomposition of epoxyl and/or hydroxyl functionalities; peak #2, a peak not yet fully
understood in the literature, that has been attributed to the decomposition of carbonyl
1
9

New Journal of Chemistry
Page 20 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
groups in αꢀsubstituted ketones and aldehydes, that occurs at roughly the same
temperature of the CO peak #2; the same width at halfꢀheight (W) was imposed for
2
peaks #1 and #2; peak #3 that results from the releasing of one CO molecule during
carboxylic anhydride decomposition – this peak has the same component peak
M 2
maximum T , W and magnitude A, as the CO peak #3; peak #4 and peak #5,
ascribed, respectively, to the decomposition of phenols and carbonyl/quinones, in
which the W was taken the same for both peaks.
(
d) A new CO peak around 1020 ˚C was included (CO peak #6) in the TPD
profiles of materials MWCNT and MWCNTh, based on the observation that the CO
profile does not return to the baseline at high temperatures, and it seems that a new
peak is appearing at those temperatures. This peak can be assigned to basic groups,
such as pyrones and chromenes, or may result from the rearrangement of the
28
surface. This peak was used for the deconvolution, but it was not taken into
consideration for quantitative analysis.
It can be observed that all of the fits are in reasonable agreement, having
M
approximately the same values of T and W for each component, when applied to the
carbon nanotube materials or to the graphene flake materials containing different types
and amounts of surface groups. Thus, the peak maxima found were between 270 and
310 ˚C for the strongly acidic carboxylic groups, 360 and 430 ˚C for the less acidic
carboxylic groups, 510 and 575 ˚C for carboxylic anhydrides, 630 and 730 ˚C for
lactones, 880 and 920 ˚C for lactones located in energetically different sites, 690 and
750 ˚C for phenols, and 830 and 950 ˚C for carbonyls/quinones. The areas and the
relative percentages of each peak of the deconvoluted CO and CO profiles allow us to
2
estimate the amount of each oxygen functional group present in the different materials.
Not taking into account the CO peaks #1 and #2, as these peaks are still not fully
understood, and their relative area percentages are generally small, the estimated
amounts of different oxygen functional groups are shown in Figure 5.
2
0

Page 21 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
800
600
400
200
0
MWCNT
MWCNTh
MWCNTht
MWCNTo
1
500
000
GF
GFh
GFht
GFo
1
500
0
Figure 5. Amounts of oxygen functional groups estimated from the deconvolution of
the TPD profiles of the carbon materials.
Hence, it is possible to conclude that: (i) the oxidation of MWCNT with HNO
3
promotes the increase in all the oxygen functional group amounts, most notably
phenols, carboxylic acids, and carbonyls/quinones, whereas for GF, this increase is
observed only for phenols, carboxylic acids and anhydrides, while the
carbonyls/quinones amount is nearly depleted; (ii) the subsequent thermal treatment at
400 ˚C, under an inert atmosphere, of materials MWCNTh and GFh, resulted in the
almost complete removal of carboxylic acid groups, while the amounts of carboxylic
anhydrides and lactones showed a slight increase, indicating some degree of
rearrangemnt of the oxygen surface groups; (iii) the gas phase oxidation treatment (5%
2
1

New Journal of Chemistry
Page 22 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
2 2
O /N ) favoured the introduction of phenols, carbonyls/quinones, and lactones; more so
in the carbon naotube materials (MWCNTo), while the GFo material presented only a
slight increase in phenol and carbonyl/quinone quantities; (iv) the oxidation procedures
used rendered materials with different types and amounts of oxygen surface groups.
Raman spectroscopy was employed in order to gain some general structural
information on the studied nanomaterials. Figure 6 (left) shows first and second order
Raman spectra of the pristine and modified MWCNT and GF materials. The firstꢀorder
Raman spectrum of the studied materials is characterized by strong bands at ∼1350
−1
and ∼1580 cm , known as D and G bands, respectively. Defectꢀfree graphite would
2
only shows the G band, common to all sp carbon systems and corresponds to the firstꢀ
order scattering of the E2g mode, whereas the D band corresponds to a phonon mode
whose intensity is strictly connected to the presence of sixꢀfold aromatic rings close to
29
local lattice distortions (defects) of the graphitic network.
2
2

Page 23 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
1
1
1
.30
.25
.20
d
c
1.15
1
1
.10
.05
b
a
1
000
1500
2000
2500
3000
MWCNT MWCNTh MWCNTht MWCNTo
ꢀ1
Raman shift (cm )
1
1
1
.2
.1
.0
h
g
0.9
0
0
0
.8
.7
.6
f
e
1
000
1500
2000
2500
3000
GF
GFh
GFht
GFo
ꢀ1
Raman shift (cm )
Figure 6. Left: Raman spectra of pristine MWCNT (a), MWCNTh (b), MWCNTht (c),
MWCNTo (d), GF (e), GFh (f), GFht (g), and GFo (h). Right: calculated I /I ratio for the
D
G
pristine and modified MWCNT and GF materials.
These distortions can be due, for example, to the presence of edges of graphitic
planes, atomic vacancies or oxygenated groups. In the secondꢀorder spectra of original
ꢀ1
and oxidized graphene flakes, there is a strong peak at ~2700 cm , assigned as the
ꢀ1
2
D mode and that is the overtone of the D peak. At ~1620 cm it is also possible to
identify a weak shoulder that corresponds to the D’ mode. The intensity ratio of the D
and G bands is commonly used to estimate the disorder degree of graphitic
2
3,30–32
materials.
MWCNT and GF materials. The oxidized carbon nanotubes presented an increase in
the I /I for all materials, indicating the introduction of defects on the nanotubes
Figure 6 (right) presents the I /I ratio for the pristine and modified
D G
D
G
2
3

New Journal of Chemistry
Page 24 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
2
structure. This result is concomitant with the N adsorption isotherms, which showed
that the oxidation procedures, increased the nanotube specific surface areas, throught
the increase of the access to the inner cavities promoted by the formation of
holes/defects.
D G 3
As for the GF materials, the I /I ratio increased considerably after HNO
oxidation for GFh, followed by a decrease after the thermal treatment at 400 ˚C in
GFht, whereas the GFo material remained practically the same as the pristine material.
These observations indicate the disorder effect of the introduction of oxygen
functionalities on the structure of GF and the occurrence of structural defects. The
results indicate that nitric acid, being a potent oxidant, introduced defects and oxygen
functional groups to a greater extent in GFh, while the subsequent thermal treatment
promoted the removal of part of the oxygen functionalities and possibly some degree of
D G 2
graphitization in GFht. As for the GFo, the unaltered I /I ratio indicates that the O
oxidation did not promote significant structural defects These results are in good
agreement with the data obtained by XPS and TPD, concerning the variation of oxygen
content upon the different oxidation procedures.
The morphology of the pristine and modified carbon materials was analyzed by
TEM and some examples of the obtained TEM micrographs are shown in Figure 7. The
micrograph obtained for material MWCNT (Figure 7a) shows that the thin multiꢀwalled
carbon nanotubes exist in wellꢀdefined tubular structures with lengths in the micron
range. The image obtained for material MWCNTht (Figure 7b), subjected to nitric acid
treatment and subsequent thermal treatment at 400 ˚C, shows peculiar sferoidal
structures, which appear to be crossꢀlinking different nanotubes. The acquired image of
GF (Figure 7c) suggests that it is composed by wrinkled fewꢀlayer graphene sheets,
whose sizes are in the range of hundreds of nm to one ꢁm. The micrograph obtained
for material GFht (Figure 7d) shows that the ball milling and subsequent thermal
treatment at 600 ˚C, did not promote any apparent significant structural changes in the
2
4

Page 25 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
oxidized material. However, it is possible to observe some defective regions (marked
with blue arrows), which may be attributed to the oxidation procedure.
Figure 7. TEM micrographs of materials MWCNT (a), MWCNTht (b), GF (c), and GFht
(
d).
Textural properties of the pristine and modified carbon material were evaluated
by the corresponding N adsorption–desorption isotherms performed at ꢀ196 ˚C, in
2
order to assess the modifications promoted by the different oxidation procedures.
Figure S5 in Supplementary Information depicts the N2 adsorption–desorption
2
5

New Journal of Chemistry
Page 26 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
isotherms for MWCNT and GF materials, and Table 5 presents the calculated specific
surface area, SBET
The N adsorption isotherms of the carbon nanomaterials are typically type II,
.
2
3
3
accordingly to IUPAC classification, which are characteristic of nonꢀporous materials
and represent unrestricted monolayerꢀmultilayer adsorption. For the GF materials
however, a distinct hysteresis loop is observed, which is characteristic of nonꢀrigid
aggregates of plateꢀlike particles.
Table 5. Apparent BET surface areas of the carbon materials derived from the nitrogen
adsorption–desorption isotherms at −196 ˚C.
2
ꢀ1
Material
S
BET (m g )
MWCNT
MWCNTh
MWCNTht
MWCNTo
277
348
344
456
GF
352
237
247
309
GFh
GFht
GFo
The differences in the nanomaterial specific surface areas due to the different
oxidation procedures between the carbon nanotubes and graphene flakes are striking:
while the different oxidation treatments led to an increase in the nanotubes SBET, they
promoted an overall decrease in the surface areas of the GF counterparts. The SBET
2
ꢀ1
2
ꢀ1
increased from 277 m g in pristine MWCNT to 348, 344 and 456 m g in MWCNTh,
MWCNTht, and MWCNTo, respectively. These results suggest that the oxidation
procedures increased the access to the nanotubes inner cavities, through the creation
of defects and holes in the outer walls; an effect more pronounced in the gas phase
treated material MWCNTo. For the graphene flake materials, the SBET decreased from
2
6

Page 27 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
2
ꢀ1
2
ꢀ1
3
52 m g in GF to 237, 247 and 309 m g , in GFh, GFht, and GFo, respectively. This
decrease may be due a higher degree of agglomeration within the oxidized graphene
flake crystallites, induced by the introduction of oxygen surface groups. Moreover, this
decrease in the GF surface areas, follows the increase of the oxygen content present
in the materials, as estimated by TPD analysis. The specific surfaces areas obtained
for the GF materials allow the estimation of the average number of layers per
2
ꢀ1
crystallite, when comparing them the theoretical value of 2630 m g calculated for
3
4
single layer graphene. As such, the average number of layers are 7, 11, 11, and 8 for
the GF, GFh, GFht, and GFo materials, respectively.
3.2 Catalytic studies
The catalytic oxidation of cisꢀcyclooctene (CyO) with 30% H O in acetonitrile, at
2
2
80 ˚C, was studied in heterogeneous phase, using the pristine and modified carbon
nanomaterials as catalysts, as schematized in Figure 8.
Figure 8. Reaction scheme of the studied catalytic experiments.
Under these reaction conditions, three main oxidation compounds are produced,
namely, epoxycyclooctane, cyclooctꢀ2ꢀenol, and cyclooctꢀ2ꢀenone, as confirmed by
GCꢀMS analysis. Small percentages of other oxidation products were also found, but
their identification was not possible, hence these will be further referred to as Others.
Results for the catalytic oxidation of cyclooctene with 30% H O (molar ratio H O /CyO
2
2
2
2
=1.1) are summarized in Table 6.
2
7

New Journal of Chemistry
Page 28 of 37
View Article Online
DOI: 10.1039/C7NJ03706F
Table 6. Catalytic activity of the carbocatalysts in the selective oxidation of cisꢀ
cyclooctene.
b
Product selectivity (%)
Conversion
Material
Cycle
a
(
%)
Others
st
No catalyst
1
18
73
27
0
st
1
43
49
68
81
64
64
13
24
25
6
nd
MWCNT
MWCNTh
MWCNTht
MWCNTo
GF
2
12
11
rd
3
st
1
39
47
42
89
86
81
9
2
3
4
nd
2
11
15
rd
3
st
1
47
49
69
79
77
74
18
18
20
3
5
6
nd
2
rd
3
st
1
35
41
47
51
47
48
41
42
42
8
nd
2
11
10
rd
3
st
1
25
50
54
76
77
77
21
21
20
3
2
3
nd
2
rd
3
st
1
28
39
38
92
95
92
6
4
7
2
1
1
nd
GFh
2
rd
3
st
1
57
64
74
85
94
90
12
6
3
0
2
nd
GFht
2
rd
3
8
st
1
27
46
55
78
73
74
20
23
2
2
4
3
nd
GFo
2
rd
3
a
Conversion refers to GC yields and was calculated through the internal standard method, using
b
chlorobenzene as the internal standard, with 48 h of reaction time. Selectivities are with respect to
epoxide, enol + enone, and others, and obtained as molar percentages of the indicated products of the
sum of moles of all products.
2
8

Page 29 of 37
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ03706F
All the carbon nanomaterials act as catalysts in the oxidation of cisꢀcyclooctene.
Their catalytic conversions are between 25 and 57 % after 48 h of reaction. In the
absence of the carbocatalysts, the oxidation of cisꢀcyclooctene yields 18 %. Pristine
MWCNT material yields a 43 % conversion, and the oxidized nanomaterials present
similar conversions of 39, 47, and 35 % for nanomaterials MWCNTh, MWCNTht, and
MWCNTo, respectively. Selectivity to epoxycyclooctane, however, increases from 81 %
in pristine MWCNT to 89 % in MWCNTh, and decreases to 48 % in MWCNTo. This
indicates that the types and amounts of surface groups introduced by the oxidation
treatments influence the reaction pathway, with the material having more acidic groups
(
MWCNTh) yielding higher selectivity to epoxycyclooctane than those bearing more
neutral groups (MWCNT, MWCNTo), which have a higher preference for the allylic
oxidation.
For the graphene flake materials, conversion was 25 % for pristine GF, remaining
roughly the same at 28 and 27 % for materials GFh and GFo, and increasing to 57 %
for GFht. As with the nanotube materials, selectivity to epoxycyclooctane increased
from 76 % in pristine GF to 92 and 85 % in GFh and GFht, respectively, and remained
roughly the same in GFo (78 %). This corroborates the observation that the presence
of more acidic groups, e.g. carboxylic acids, carboxylic anhydrides, favour the
selectivity to epoxycyclooctane, while the existence of more neutral groups like
carbonyls and phenols increase the allylic oxidation. Interestingly, the conversion for
GFht (57 %) is significantly higher than for GFh (28 %), while their selectivities to
epoxycyclooctane remain similar (92 and 85 %, respectively). From the TPD data, it
was shown that there was a considerable removal of carboxylic acid groups for GFht,
and an increase in carboxylic anhydrides, during the applied thermal treatment. Thus,
the catalytic results indicate that the surface rearrangement of the oxygen groups
during the thermal treatment led to an optimized ratio of acidic groups, which yielded a
more active catalyst, whilst preserving the desired selectivity. These results are
consistent with previously published results by Zhou et al. and Li et al., which used
2
9