Aerobic oxidation of α-pinene catalyzed by carbon nanotubes

Article abstract of DOI:10.1039/c5cy00136f

Carbon nanotubes (CNTs) and nitrogen-doped CNTs (NCNTs) as metal-free catalysts exhibited an excellent activity in the selective oxidation of α-pinene with molecular oxygen as the terminal oxidant. Two distinct pathways, i.e. epoxidation and allylic oxidation, were active in this reaction. Enhancement of epoxidation was observed over CNTs, yielding the highest epoxidation/allylic oxidation products ratio. Excellent activity was achieved over NCNTs, giving 54.5% α-pinene conversion and 272.4 mmol g^-1 h^-1 mass-normalized activity, which compete with that of the state-of-the-art metal catalysts. Allylic oxidation was enhanced over NCNTs, using which equimolar amounts of epoxide and allylic products were produced. Thus, N-doping boosted the overall conversion and the yields of both epoxidation and allylic oxidation products, which was supported by the results of theoretical simulation.

Full text of DOI:10.1039/c5cy00136f

View Article Online
View Journal
Catalysis
Sc ience &
Technology
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: Y. Cao, Y. Li, H.
Yu, F. Peng and H. Wang, Catal. Sci. Technol., 2015, DOI: 10.1039/C5CY00136F.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
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
Information for Authors.
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 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.
www.rsc.org/catalysis

Page 1 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
Aerobic Oxidation of αꢀPinene Catalyzed by Carbon Nanotubes
Yonghai Cao, Yuhang Li, Hao Yu∗, Feng Peng∗, Hongjuan Wang
School of Chemistry and Chemical Engineering, South China University of Technology,
Guangzhou, Guangdong, 510640 (China)
Abstract
Carbon nanotubes (CNTs) and nitrogenꢀdoped CNTs (NCNTs) as metalꢀfree
catalysts exhibited excellent activity in the selective oxidation of αꢀpinene with molecular
oxygen as the terminal oxidant. Two distinct pathways, i.e. epoxidation and allylic
oxidation, were active in this reaction. An enhancement of epoxidation was observed over
CNTs, affording the highest epoxidation/allylic oxidation products ratio. Excellent
activity were achieved over NCNTs, displaying 54.5% αꢀpinene conversion and 272.4
ꢀ1
ꢀ1
mmol g h massꢀnormalized activity, which is competing with the stateꢀofꢀtheꢀart metal
catalysts. The allylic oxidation was enhanced over NCNTs, through which equimolar
amounts of epoxide and allylic products are produced. Thus, Nꢀdoping boosted the
overall conversion and the yields of both of the epoxidation and allylic oxidation
products, which are supported by theoretical simulation results.
Keywords: αꢀPinene; Carbon catalysis; Nitrogen doping; Epoxidation; Molecular
dynamics
∗
Tel. & Fax: +86 20 8711 4916.
Eꢀmail address: yuhao@scut.edu.cn (H. Yu); cefpeng@scut.edu.cn (F. Peng).
1

Catalysis Science & Technology
Page 2 of 31
View Article Online
DOI: 10.1039/C5CY00136F
1
. Introduction
Selective oxidation of hydrocarbons towards oxygenated derivatives is a significant
chemical valueꢀchain as the relatively cheap molecules can be converted into valueꢀadded
1
ꢀ3
products. Molecular oxygen is the most promising oxidant from the viewpoint of green
chemistry and economy. Most of the aerobic oxidations have been proved to be the
radical autoxidation, which can be accelerated by metal catalysts through the
2
, 4
HarberꢀWeiss cycle.
The oxidation of cycloolefins yields the corresponding epoxides and allylic
3, 5
oxidation products.
The allylic oxidation is usually dominant for cycloolefins with
6
monocyclic structure, e.g. cyclohexene, in which 2ꢀcyclohexeneꢀ1ꢀone and
2
ꢀcyclohexeneꢀ1ꢀol are produced through the decomposition of cyclohexene
7
hydroperoxide, an important stable intermedium. Meanwhile, trace amounts of epoxide
7
products may be produced as well. Cycloolefins with complex structure, e.g. terpene
substrates, tends to be oxidized to form epoxidation products, which has been revealed in
3
, 8
αꢀpinene oxidation. The oxidation of αꢀpinene displays complex product distributions,
since the oxidation may occur at four possible sites in the molecule as shown in Scheme 1.
3
O addition to allyl radicals will generate peroxyl radicals. After that, peroxyl radical
2
added to the C=C double bond of substrate moleculer was proceeded, yielding a dialkyl
peroxide, which can eliminate an alkoxyl radical to favorably form αꢀpinene oxide, an
3
origin of the epoxidation products.
2

Page 3 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
Scheme 1 Molecular structure of αꢀpinene and its four possible oxidation sites (denoted
8
as a–d).
The aerobic oxidation of αꢀpinene is of great interest for the research and industrial
communities. The allylic oxidation of αꢀpinene is industrially relevant, because it
produces α, β unsaturated alcohol and ketone, i.e. verbenone and verbenol, which are
9
important intermediates in fragrance industry and organic synthesis. The epoxidation
product, αꢀpinene oxide, can be isomerized to campholenic aldehyde, an intermediate in
the synthesis of sandalwoodꢀlike fragrances, such as Sandalores (Givaudan) or
1
0
Polysantols (Firmenich). The products can be isolated from the autoꢀoxidation of
11
αꢀpinene with air, but the yields are usually low. Numerous studies have revealed that
the oxidation of αꢀpinene can be catalyzed by solid catalysts containing Co, Au, Mn, Rh,
1
2ꢀ21
Ti, V, etc..
0%) to verbenone for the allylic oxidation of αꢀpinene. Timofeeva et al. used
vanadiumꢀcontaining nickel phosphate molecular sieves to catalyze the αꢀpinene
For example, the silica–titania coꢀgel catalyst afforded high selectivity (ca.
2
1
6
2
2
oxidation, giving about 48% allylic oxidation products. Also, molecular sieves are
23
promising solid catalysts for the epoxidation. Trong et al. have utilized modified
TiꢀMCMꢀ41 for the oxidation of αꢀpinene to αꢀpinene oxide with 100% selectivity using
2
3
tꢀbutyl hydroperoxide as oxidant. So far, metal catalysts played central role in the
3

Catalysis Science & Technology
Page 4 of 31
View Article Online
DOI: 10.1039/C5CY00136F
catalytic oxidation of αꢀpinene.
Nanocarbons have attracted great attentions from the research and industrial
communities in the past two decades as metalꢀfree catalysts, due to their unique chemical
2
4ꢀ26
and physical properties.
It has been proved that the defects and surface functionalities
of carbon materials play significant roles in the oxidative dehydrogenation (ODH) of
2
7ꢀ29
30, 31
32
hydrocarbons,
the activation of methane,
the hydrogenation of nitrobenzene, the
3
3
34
hydroxylation of benzene to phenol, the oxidation of arylalkanes, the dehydration and
35ꢀ38
39
oxidation of alcohols,
the Baeyer−Villiger oxidation and the wet air oxidation of
4
0
phenol. Our previous works have revealed that carbon materials and its nitrogenꢀdoped
derivatives showed unexpected activities in the selective oxidation of hydrocarbons in the
4
1ꢀ46
liquidꢀphase with oxygen as oxidant.
Especially, nitrogen doped CNTs (NCNTs)
displayed excellent performances in the allylic oxidation of cyclohexene, resulting in the
6
high 2ꢀcyclohexeneꢀ1ꢀnone selectivity. The remarkable activity of carbons is due to the
charge transfer between CNTs and radicals, facilitated by delocalized electrons in
4
7
graphene layers. It is believed that they promote the radical chain propagation via
6
stabilizing peroxyl and cycloxyl radicals. Recently, nitrogen doped nanocarbons were
discovered as efficient catalysts for the epoxidation of transꢀstilbene and styrene using
4
8, 49
tertꢀbutyl hydroperoxide (TBHP) as oxidant.
The quaternary nitrogen species is
responsible for the epoxidation. These studies proposed a new approach for the alkene
oxidation catalyzed by nanocarbons.
In this work, CNTs and NCNTs were used as metalꢀfree catalysts, which displayed
4

Page 5 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
excellent activity in the oxidation of αꢀpinene in the liquid phase using oxygen as oxidant.
To our best knowledge, it is the first time to demonstrate the αꢀpinene oxidation can be
metalꢀfree catalyzed by nanocarbons with considerable activity. The carbon skeleton is
the key essential for the epoxidation, based on the unique structure of αꢀpinene molecule.
Nitrogen dopants in carbon skeletons can efficiently enhance the allylic oxidation through
stabilizing peroxyl and cycloxyl radicals. A plausible mechanism was proposed for the
metalꢀfree catalytic oxidation according to catalytic results and theoretical calculation.
2
. Experimental
2
.1 Preparation of catalysts
4
2, 43
CNTs were synthesized according to our previous works.
In brief, CNTs were
produced by a chemical vapor deposition (CVD) method with liquefied petroleum gas as
carbon source over a FeMo/Al O catalyst in a horizontal tubular quartz furnace with 4
2
3
5
0
cm inner diameter (i.d.). The details of FeMo/Al O catalyst can be found in Ref.
2
3
Before the growth of CNTs, the catalyst was activated by a mixture of H and N (25 and
2
2
3
ꢀ1
o
2
5 Ncm min ) for 30 min. The growth of CNTs was carried out at 700 C for 130 min
3
ꢀ1
3
ꢀ1
3
ꢀ1
with 20 Ncm min liquefied petroleum gas, 10 Ncm min H , and 50 Ncm min N .
2
2
Nꢀdoped CNTs were synthesized by a similar CVD method using aniline (AN) and
xylene as carbon and nitrogen sources. To grow NCNTs, 10 mL mixtures of AN and
xylene, with 0, 10, 50 and 100 v% AN, were injected by a syringe pump at a rate of 3 mL
ꢀ1
o
h . The liquid mixtures were vaporized in the quartz tube at about 180 C. The growth of
o
3
ꢀ1
NCNTs was carried out at 800 C in Ar or NH at 500 Ncm min . The residual
3
5

Catalysis Science & Technology
Page 6 of 31
View Article Online
DOI: 10.1039/C5CY00136F
FeMo/Al O catalyst in the asꢀsynthesized CNTs or NCNTs was removed by 12 mol/L
2
3
concentrated HCl aqueous solution for 4 h before characterizations and catalytic tests.
4
2, 43
FeO /NCNTs were prepared as follows.
FeSO ·7H O was dissolved into 20 mL
4 2
x
distilled water, and then 3 g HCl washed NCNTs were added to form suspension. After
ultrasonicated for 5 min, 45 mL hydrogen peroxide (30%) was added dropwise with
o
vigorous stirring. The resulting suspension was refluxed at 80 C for 4 h. The solids were
collected by filtration and washed with distilled water, and subsequently dried in air at
o
o
1
20 C for 12 h. After that, the catalyst was treated with N at 450 C for 2 h. The
2
o
FeN /NCNTs was synthesized by the same method but treated in NH at 800 C for 4 h.
x
3
To study the effect of Mo impurities, the NCNT was firstly purified in concentrated
HCl for 4 h, then impregnated with ammonium molybdate ((NH ) Mo O
ꢀ
24
4H O). The
4
6
7
2
o
o
solids were washed and dried at 60 C overnight, then were heated in Ar at 500 C for 3 h
o
and NH at 800 C for 3 h to obtain MoOx/NCNTs and MoNx/NCNTs, respectively.
3
2
.2 Catalyst characterization
Brunauer–Emmett–Teller (BET) specific surface areas (SSA) were measured by N2
adsorption at liquid N temperature in an ASAP 2010 analyzer. The concentrations of
2
metal impurities in the catalysts were determined by ICPꢀAES (Prodigy SPEC Leeman).
51
The surface oxygeneous groups were analyzed by Boehm titration. Xꢀray diffraction
(XRD) patterns were recorded on a Bruker D8 ADVANCE diffractometer equipped with
a rotating anode using Cu Kα radiation (40KV, 40mA). Raman spectra were obtained in a
6

Page 7 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
LabRAM Aramis micro Raman spectrometer with an excitation wavelength at 633 nm
with 2 ꢁm spot size. Xꢀray photoelectron spectroscopy (XPS) was performed in a Kratos
Axis ultra (DLD) spectrometer equipped with an Al Ka Xꢀray source in ultrahigh vacuum
ꢀ10
(
UHV) (<10 Torr). Binding energies (±0.2 eV) were referenced to the C peak at 284.6
1s
eV as graphite. Transmission electron microscope (TEM) images were obtained with a
FEI Tecnai G2 12 microscope operated at 100 kV. The specimens for TEM were prepared
by ultrasonically suspending the sample in acetone and depositing a drop of the
suspension onto a grid.
2
.3 αꢀPinene oxidation
The αꢀpinene oxidation reaction was carried out in a magnetically stirred 50 mL
Teflonꢀlined autoclave in batch mode. Typically, 10 mL αꢀpinene, 20 mL solvent, 2 mL
oꢀdichlorobenzene (oꢀDCB) as internal standard and 70 mg catalyst were placed into the
autoclave. 1100 rpm stirring speed was used to minimize the effect of mass transfer and
homogenize the mixture. Before reaction, the reactor was flushed with N . Then, the
2
reactor was heated to a stable operational temperature, and subsequently pure O was fed
2
into the reactor (defining t=0). To determine verbenylꢀhydroperoxide, the samples were
double analyzed, before and after reducing verbenylꢀhydroperoxide to verbenol with
5
2
triphenyl phosphine (Acros, 99%) . Conditions of GC: KBꢀ1701 capillary column (30 m,
DF=0.51 mm, 0.1 um i.d.), a flame ionization detector (FID). Temperature program was
o
o
carried out: initial temperature = 170 C, 4 min; final temperature = 250 C, heating rate =
7

Catalysis Science & Technology
Page 8 of 31
View Article Online
DOI: 10.1039/C5CY00136F
o
o
o
1
0 C/min, temperature of injector = 280 C, temperature of detector = 280 C.
2
.4 Theoretical calculation
Molecular dynamics simulations were performed with Forcite package in Materials
Studio. A (5,5) CNT, extended to a supercell of (1*1*20), was employed in
threeꢀdimensional period boundary conditions with a size of 30 Å × 30 Å × 49.19 Å.
Optimized radical molecules were set into the box, so as to investigate interactions in the
system. Graphitic nitrogen doped CNTs (NꢀCNT) was selected as the comparative
material. In details, 15 nitrogen atoms dispersedly in the CNT skeleton, displaying an N
atom ratio of 3.75%.
The system was simulated with NVT ensemble at 353 K through
5
3ꢀ55
NoséꢀHooverꢀLangevin (NHL)
thermostat with the decay constant of 1.0 ps. All
56
structures were assigned with COMPASS force field. Simultaneously, for an exception,
carbon radical was set to c3 type (generic sp2 hybridization with 3 bonds) to maintain the
lone electron. The summation of electrostatic potential and van der Waals potential were
ꢀ5
both calculated through Ewald method with the accuracy of 1.0*10 kcal/mol. Total
simulation time was 400 ps, where 100 ps was used for equilibrium of system and 300 ps
for dynamics analysis. The simulation step was 1.0 fs and outputting a trajectory frame at
each 2.0 ps.
8

Page 9 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
3
. Results and discussion
3
.1. Oxidation of αꢀpinene
Scheme 2 Catalytic oxidation of αꢀpinene.
Table 1 Catalytic performance of the carbons and metal catalysts in the αꢀpinene
[
a]
oxidation
[c]
[d]
[e]
rw
rs
Selectivity (%)
[b]
SBET
X
E/A
Catalyst
(mmol (mmol
2
[f]
(
m /g)
(%)
Ratio
ꢀ1
ꢀ1
ꢀ2 ꢀ1
g h ) m h )
[
g]
Blank
－
10.3
19.1
24.6
54.5
34.2
59
－
45.0
43.0
272.4
3.6
－
24.5
38.2
33.8
37.8
34
11.7
6.8
7.9
14.1
22
15.0
5.9
6.1
17.5
16
40.4
26.7
37.9
15.3
－
0.9
3.0
2.4
1.2
0.9
－
AC
731.5
127.8
155.1
1071
－
0.03
0.34
1.76
0.3
CNT
[
h]
NCNT
[
i]
CoꢀHMS
[j]
MnAlPOꢀ36
75.8
1.6
－
91
－
－
－
[k]
CoꢀPOM/MILꢀ101
2050
45
0.08
－
27
29
－
－
o
[
a] Conditions: 80 C, 1.5 MPa O , 10 mL αꢀpinene, 20 mL CH CN, 2 mL oꢀDCB, 70 mg catalyst, 4 h.
2
3
[b] Conversion. [c] Initial reaction rate of αꢀpinene consumption normalized by catalyst mass. [d]
Initial reaction rate of αꢀpinene consumption normalized by catalyst surface. [e] Selectivity of major
products. The byꢀproducts include pinocarveol, pinenol, myrtenal and others. [f] E/A ratio means the
9

Catalysis Science & Technology
Page 10 of 31
View Article Online
DOI: 10.1039/C5CY00136F
molar ratio of epoxidation products, i.e. αꢀpinene oxide, to allylic oxidation products, i.e. verbenol and
verbenone. [g] Without catalyst. [h] The NCNTs were synthesized by 100% AN in NH atmosphere. [i]
3
57
o
Ref. . Conditions: 3.97 mL αꢀpinene, 100 mg catalyst, 40 mL 1,4ꢀdioxane, 80 C, atmospheric
58
o
pressure, 24 h. [j] Ref. . Conditions: 40 mL αꢀpinene, 250 mg catalyst, 50 C, 3 MPa O , 8 h. [k] Ref.
2
5
9
o
.
Conditions: 0.016 mL αꢀpinene, 14 mg catalyst, 1 mL CH CN, 50 C, 0.1 MPa O , 2 h.
3 2
Activated carbons (AC), CNTs and NCNTs (see Fig. S1) were evaluated in the
αꢀpinene oxidation, as summarized in Table 1. The main products, such as αꢀpinene oxide,
verbenol, verbenone and verbenylꢀhydroperoxide, were produced through the aerobic
oxidation of αꢀpinene (Scheme 2). 10.3% αꢀpinene conversion was obtained after 4 h
3
without any catalysts, which can be attributed to the radical autoxidation process. The
selectivity pattern may be obviously influenced by the oxygen pressure in the case of
8
autoxidation. However, no obvious difference was observed on the selectivity pattern in
our study from 0.5 to 2.5 MPa O in the presence of carbon catalysts (Fig. S2). Utilizing
2
AC and CNTs as catalysts, the conversions were significantly improved to 19.1% and
2
4.6% under the same conditions, respectively. The activity of CNTs was higher than AC,
probably due to the diffusion resistance in micropores and relatively low graphitization
6
, 43, 44
degree of AC.
When NCNTs (with 4.36% N content, see Table S1) were used as
catalysts, the conversion was improved by about twoꢀfolds compared with CNTs,
affording high up to 54.5% conversion, corresponding to a mass normalized activity of
ꢀ1
ꢀ1
ꢀ2 ꢀ1
2
72.4 mmol g h and a surface area normalized activity of 1.76 mmol m h . This
result was even higher than typical stateꢀofꢀtheꢀart metallic solid catalysts as listed in
1
0

Page 11 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
Table 1, indicating NCNTs is promising for the αꢀpinene oxidation. The influence of
residual metals in NCNTs was ruled out by washing with 12 mol/L HCl, postꢀloading
metal oxides and nitrides over NCNTs. The results showed that the effect of metal was
negligible and NCNTs were intrinsically a metalꢀfree catalyst. (see Supporting
Information, Fig. S3, S4 and Table S2)
Two types of products can be generated by the oxidation of αꢀpinene through two
3
, 8
distinct pathways, i.e. epoxidation and allylic oxidation.
Compared with the blank
experiment, the introduction of AC or CNTs significantly improved the epoxidation,
resulting in 38.2% and 33.8% αꢀpinene oxide selectivity, respectively. Accordingly, the
epoxidation/allylic oxidation (E/A) ratios were 3.0 and 2.4, indicating the selective
production of αꢀpinene oxide. In this work, the E/A ratio was calculated from the molar
ratio of αꢀpinene oxide to the sum of verbenol and verbenone, and
verbenylꢀhydroperoxide was not counted in the allylic products, because the epoxide
comes from it, as discussed below. When NCNTs were used as catalysts, the selectivity of
αꢀpinene oxide remained high at 37.8%, meanwhile the selectivities to verbenone and
verbenol were enhanced, resulting in the lower E/A ratio of 1.2.
11

Catalysis Science & Technology
Page 12 of 31
View Article Online
DOI: 10.1039/C5CY00136F
(
a)100
(
b)100
80
80
60
40
20
0
OH
60
40
20
0
0
2
4
6
8
0
2
4
6
8
Time (h)
Time (h)
Fig. 2 Conversions and product yields of (a) CNTꢀ and (b) NCNT (100%AN+NH ) ꢀ
3
o
catalyzed aerobic oxidation of αꢀpinene. Conditions: 80 C, 1.5 MPa O , 10 mL αꢀPinene,
2
2
0 mL CH CN, 2 mL oꢀDCB, 70 mg catalyst.
3
NCNT
CNT
-7
-8
-9
39.2 kJ/mol
64.0 kJ/mol
-
10
11
-
2.6
2.7
2.8
2.9
3.0
3.1
1000/T (K)
Fig. 3 Arrhenius plots of catalytic oxidation of αꢀpinene with CNTs or NCNTs
100%AN+NH ) as catalyst.
(
3
1
2

Page 13 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
Fig. 2 shows the time courses of αꢀpinene oxidation catalyzed by CNTs and NCNTs.
In the presence of CNTs or NCNTs, the reaction starts up rapidly without induction
ꢀ1
period. NCNTs displayed the higher pseudo firstꢀorder reaction rate constant of 0.149 h
ꢀ1
than that of CNTs, 0.0592 h , indicating the much higher activity than CNTs. The
improvement of activity was also evidenced by the Arrhenius plots shown in Fig. 3. The
ꢀ1
apparent activation energy (E ) was calculated as 39.2 kJ mol for NCNTs, which is
a
ꢀ1
much lower than that of CNTs (64.0 kJ mol ), indicating the positive effect of NCNTs on
reducing the reaction energetic barrier.
80
70
60
50
40
30
20
10
0
NCNTs
CNTs
0
10
20
30
40
50
60
70
Conversion (%)
Fig. 4 Product distributions of CNTs and NCNTs (100%AN+NH ) as functions of
3
αꢀpinene conversion.
As shown in Fig. 2, when the nanocarbons were used as catalysts, epoxide was the
main product. It should be noted that NCNTs afforded the obviously higher yields to
verbenone and verbenol in equimolarity than CNTs. Fig. 4 shows the dependences of
selectivity of oxygenated derivatives on αꢀpinene conversion. Three main products were
1
3

Catalysis Science & Technology
Page 14 of 31
View Article Online
DOI: 10.1039/C5CY00136F
selected, αꢀpinene oxide, verbenone and verbenol, representing epoxide and allylic oxides,
respectively. Compared with CNTs, NCNTs provided much higher selectivity to allylic
oxides and almost constant selectivity of epoxide. Although the relatively high epoxide
selectivity can be reached at low conversions on CNTs, the epoxide selectivity decreased
significantly at high conversions. Benefiting from the reduced selectivity of
verbenylꢀhydroperoxide, NCNTs afforded the higher selectivity of allylic oxides, because
the decomposition of verbenylꢀhydroperoxide generates verbenone and verbenol. To
summarize, NCNTs selectively oxidize αꢀpinene to produce αꢀpinene oxide at about 37%
selectivity even at high conversion level up to 65% (Fig. 4), while CNTs cannot achieve a
balance of selectivity and conversion, due to the overꢀoxidation in the prolonged reaction
duration.
1.8
40
1.6
30
1.4
1.2
20
1.0
10
0.8
0
1
2
3
4
N/(N+C) (%)
1
4

Page 15 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
Fig. 5 Dependences of activity and selectivities of four main products on the gross
nitrogen content in NCNTs. The selectivities were obtained at similar αꢀpinene
conversion of about 30%.
To further understand the impact of nitrogen doping on catalytic performance,
NCNTs with different nitrogen contents were synthesized by tuning the composition of
precursor and reaction atmosphere. In this work, the nitrogen contents were in the range
4
2
from 0.3% to 4.36%, measured by XPS, as reported by our previous work. (see also
Table S1 for quantitative XPS results) The introduction of nitrogen heteroatoms
effectively enhanced the catalytic activity of the NCNTs and had the significant influence
on the product distribution as well (Table S3). Fig. 5 shows the dependences of specific
activity and selectivity of oxygenated derivates on the nitrogen content on NCNTs. It is
clearly shown that the nitrogen dopant is beneficial for the specific activity of NCNTs,
demonstrating the positive effect of nitrogen heteroatoms. On the aspect of selectivity, the
most apparent impact of Nꢀdoping was the decrease of verbenylꢀhydroperoxide
selectivity with the increase of nitrogen content, attributed to the decomposition of
6
, 41
peroxide over carbons and nitrogen dopants.
A slight increase of the selectivity of
epoxide by about 8 percent was observed as the nitrogen content changed from 0.3% to
4
.36%. It was worth noting that the selectivity of verbenone considerably increased,
6
indicating the selective production of ketone products by NCNTs.
Table 2 Properties and performances of NCNTs (AN+NH ) in the aerobic oxidation of
3
[
a]
αꢀpinene with different HNO oxidation durations and annealing temperatures.
3
1
5

Catalysis Science & Technology
Page 16 of 31
View Article Online
DOI: 10.1039/C5CY00136F
ꢀ1
[c]
Boehm titration (mmol g )
Selectivity (%)
HNO
3
Annealing
(K)
Raman
S
BET
X
2
(
m /g)
[b]
(%)
reflux (h)
I /I
D G
ꢀOH
ꢀC=O
ꢀCOOH
[
[
d]
d]
0
7
7
333
88.3
179.7
207.1
0.93
1.67
1.48
0.26
0.74
n.a.
0.19
1.22
n.a.
0.10
1.03
n.a.
54.5
51.0
61.4
37.8
40.6
41.1
14.2
17.1
18.7
17.5
13.8
13.4
15.3
16.4
15.9
333
1173
o
[
a] The NCNTs synthesized with 100% AN in NH were used as the base case. Conditions: 80 C, 1.5
3
o
MPa O , 10 mL αꢀPinene, 20 mL CH CN, 2 mL oꢀDCB, 70 mg catalyst, 4 h. [b] 9 M HNO , 110 C.
2
3
3
[c] In Ar gas for 4 h. [d] Vacuum drying at 333 K overnight. [e] Not applied.
To further tune the selectivity, oxygenꢀcontaining groups were introduced on the
carbon catalysts by treating with HNO . A negative effect of oxygen functionalities on
3
catalytic performance was observed over CNTs (Table S4). It is consistent with the
4
3
6
46
previous works concerning the oxidations of cyclohexane, cyclohexene, ethylbenzene
6
0
and cumene catalyzed by carbons, indicating that the oxygen functional groups and
6, 43, 46
defects are unfavorable for the radical oxidation of hydrocarbon in the liquid phase.
As shown in Table 2, after treated by HNO for 7 h, abundant oxygenꢀcontaining groups
3
were introduced on NCNTs, meanwhile the conversion slightly decreased from 54.5% to
5
1.0%. This result indicated that the oxygen groups were unfavorable the specific activity,
taking into consideration the significantly improved specific surface area of NCNTs after
o
43
HNO treatment. After removed the groups by annealing at 900 C, the conversion was
3
further elevated to 61.4%, while the selectivity of epoxide increased from 37.8% to
1
6

Page 17 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
4
1.1%. Hence, by the oxidationꢀannealing process, the performance of NCNTs was
improved, because of either the improved surface area or the unique interaction between
NCNTs and αꢀpinene.
[
a]
Table 3 Effect of solvents on αꢀpinene oxidation catalyzed by NCNTs
[b]
[d]
Polarity
Selectivity (%)
[
c]
Solvent
X (%)
(kcal/mol)
Acetonitrile
Acetone
0.46
54.5
28.6
26.7
13.9
21.9
37.8
29.2
23.0
35.6
26.9
14.1
17.5
10.3
14.8
16.1
11.6
15.3
23.2
12.4
18.9
18.4
0.355
0.164
0.099
11.4
13.1
16.5
15.9
1
,4ꢀDioxane
Toluene
nꢀHexane
0.009
o
[
a] Reaction Conditions: 80 C, 1.5 MPa O , 10 mL αꢀpinene, 20 mL CH CN, 2 mL oꢀDCB, 70 mg
2
3
catalyst, 4 h. The NCNTs were synthesized by xylene in the NH atmosphere. [b] The measure of
3
ꢂ
polarity used is the empirical solvent parameter ꢀ , derived from the solvatochromism of a
ꢁ
6
1
pyridinium Nꢀphenolate betaine. The polarity data of solvents were obtained from reference . [c]
Conversion. [d] Selectivity of major products. The byꢀproducts include pinocarveol, pinenol, myrtenal
and others.
Solvent has significant influences on the outcome of organic reaction. In this study,
five solvents with different polarity were selected to investigate the solvent effect. As
summarized in Table 3, the conversion of αꢀpinene increased with the solvent polarity
1
7

Catalysis Science & Technology
Page 18 of 31
View Article Online
DOI: 10.1039/C5CY00136F
ascending. The highest conversion, 54.5%, was obtained when acetonitrile was used as
solvent, which has the strongest polarity among the solvents tested. It is probably caused
by the polar solvent facilitates the desorption of polar oxide products, therefore speeds up
surface renewal of catalyst. For example, because acetonitrile is immiscible with
αꢀpinene, the formed intermediate products on catalyst surfaces are rapidly transferred to
solvent instead of contacting and reacting with neighboring αꢀpinene molecules, avoiding
3
3
the overꢀoxidation. It was noticed that, although the polarity of toluene is higher than
nꢀhexane, the lower conversion of αꢀpinene was observed in toluene. It may be attributed
to the competitive adsorption between substrate and toluene upon active sites of NCNTs,
since toluene can strongly interact with and block active sites, as observed in the
6
2
metalꢀfree catalysis system with toluene involved.
80
O
70
OH
O
OOH
60
50
40
30
20
10
0
0
1
2
3
4
5
6
Cycle
Fig. 6 The recyclability of NCNTs (100%AN+NH ) in the oxidation of αꢀpinene in the liquidꢀphase.
3
o
Reaction conditions: 80 C, 1.5 MPa O , 10 mL αꢀpinene, 20 mL CH CN, 2 mL oꢀDCB, 70 mg
2
3
1
8

Page 19 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
catalyst, 4 h.
The recyclability of catalyst is very important for practical applications. Here, we
selected NCNTs (100%AN+NH ) to evaluate the stability for 5 runs. After each run, the
3
recycled catalyst was repeatedly washed by deionized water and acetone, and then dried
o
at 110 C overnight. The catalytic results are shown in Fig. 6. NCNTs showed
outstanding recyclability. After five recycles, there were no noticeable changes in both
the conversion of αꢀpinene and product distributions, demonstrating the structural
stability of NCNTs under the reaction conditions (Fig.S5).
3
.2 Reaction mechanism
It is widely accepted that the aerobic oxidation of αꢀpinene is a radical chain
reaction. The addition of carbon catalysts can boost the overall reaction rate of the radical
process, which has been proved in the cases of cyclohexane, cyclohexene, cumene and
6
, 43, 46, 60
ethylbenzene.
As shown in Table 4, when pꢀbenzoquinone as a radical scavenger
was added to the reaction, the conversion of αꢀpinene dramatically suppressed to 2.7%,
even lower than that of blank experiment (Table 1). This result demonstrates that the
oxidation of αꢀpinene catalyzed by carbons is a radical reaction as well. Furthermore,
NCNTs cannot catalyze the conversion of verbenol, verbenone and αꢀpinene oxide,
indicating that the radical reaction is initiated from reactive radicals derived from
substrate αꢀpinene, e.g. alkyl, alkyol and dialkyl peroxide radicals. Thus, the
6
, 41ꢀ45
carbonꢀradical interaction should be the essential for the process.
Table 4 Catalytic activities of oxidations of various substrates catalyzed by NCNTs
1
9

Catalysis Science & Technology
Page 20 of 31
View Article Online
DOI: 10.1039/C5CY00136F
[
a]
(100%AN+NH3).
Substrate
Additive
X (%)
2.7
[b]
αꢀPinene
pꢀbenzoquinone
αꢀPinene oxide
Verbenol
―
―
―
<1
<1
Verbenone
<1
[a] Conditions: 353 K, 1.5 MPa O , 10 mmol substrates, 20 mL CH CN, 2 mL oꢀDCB, 70 mg catalyst.
2 3
[b] 10 mL αꢀpinene.
Scheme 3. Proposed reaction pathways of the NCNTꢀcatalyzed oxidation of αꢀpinene.
6
Compared with the cyclohexene oxidation in the liquid phase, allylic regioisomers
of αꢀpinene lead to the formation of the more complex radical intermediates, depending
2
0

Page 21 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
on the sites “b”, “c” and “d” involved (Scheme 1). A plausible mechanism of the aerobic
αꢀpinene oxidation was proposed to elucidate the observed catalytic behaviors of CNTs
and NCNTs. As shown in Scheme 3, αꢀpinene is oxidized by molecular oxygen via allylic
oxidation and epoxidation routes. Alkyl radicals are formed through pyrolysis of αꢀpinene,
and then peroxyl radicals are formed by the insertion of oxygen, through which the chain
reaction propagates. Similar actions were also observed in the Baeyer−Villiger oxidation
3
9
catalyzed by carbons. In the case of αꢀpinene oxidation, peroxyl radicals are preferably
added to the unsubstituted Cꢀatom, i.e. the bꢀsite, forming an epoxide and alcoxyl radical
3
(
Pathway I), resulting in the selective production of epoxide. The stabilization effect of
•
CNTs or NCNTs on radicals can improve the lifetime of R ꢀOOꢀ R radicals, which
(a)
(b) (c)
are supported by the theoretical calculations as detailed later, thereby NCNT can
4
4
efficiently increase the overall conversion of substrate. Alkoxyl radicals undergo further
propagation by abstracting a hydrogen atom from the substrate (pathway II). Finally, the
alkoxyl radical may be converted to ketone product through the abstraction of the weakly
3
bonded αH atom by O . Meanwhile, the termination of two peroxyl radicals on surfaces
2
of CNTs also formed verbenol and verbenone (pathway IV).
Pathways II and IV are similar to the autoxidation process. However, the
autoxidation cannot afford the selective formation of αꢀpinene oxide, because pathway II
produces equimolar amounts of epoxidation and allylic oxidation products. Hence, the
more allylic oxides are usually obtained through autoxidation (Table 1), taking into
8
consideration the contribution of pathway IV.
2
1

Catalysis Science & Technology
Page 22 of 31
View Article Online
DOI: 10.1039/C5CY00136F
In the cases of carbon catalyzed oxidation, the higher epoxide selectivity was
reached, indicated by the E/A ratio more than unity (Table 1 and Fig. 3). This result
suggests the unique catalytic role of carbons. Being similar to peroxyl radicals, alkoxy
ꢀ
radicals might interact with substrate to form R ꢀOꢀ R
(b) (c)
(Pathway III). Its
(a)
8
propagation may be different from the autoxidation in the presence of carbon materials.
ꢀ
In the autoxidation, R ꢀOꢀ R is very vulnerable and disintegrated quickly, releasing
(a)
(b) (c)
an epoxide and an alkyl radical. The alkyl radical may propagate the chain reaction. In
ꢀ
the presence of CNTs or NCNTs, the stability of R ꢀOꢀ R might be improved, thus
(a)
(b) (c)
additional αꢀpinene oxide can be obtained, resulting in the higher selectivity of epoxide.
The role of NCNTs in the reaction is dualfunctional. In addition to stabilize radicals
as discussed above, NCNTs have a strong ability to decompose peroxides, which has
6
, 43, 46, 60
been demonstrated in our previous works.
Hence, the pathway II or IV may be
enhanced as the same time, resulting in the enhanced selectivity of allylic oxides,
verbenol and verbenone (Fig. 4). As a result, the formations of epoxide and allylic oxides
were both improved over NCNTs. It explains the highly efficient conversion of αꢀpinene
and the low selectivity of verbenylꢀhydroperoxide over NCNTs.
2
2

Page 23 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
•
•
Fig.7 − Optimized structures of (a, b) R ꢀOOꢀ R and (c, d) R ꢀOꢀ R on the surface of (a, c)
(a)
(b) (c)
(a)
(b) (c)
CNT and (b, d) NCNT.
Table 5 The adsorption energies (ꢂE ) of freeꢀradicals on the CNTs and NCNTs from
a
theoretical calculations.
ꢂE (kJ/mol)
a
Catalyst
ꢀ
ꢀ
R ꢀOOꢀ R
R ꢀOꢀ R
(a) (b) (c)
(
a)
(b) (c)
CNT
ꢀ43.09
ꢀ60.56
ꢀ68.52
ꢀ30.16
NCNT
2
3

Catalysis Science & Technology
Page 24 of 31
View Article Online
DOI: 10.1039/C5CY00136F
The aforementioned discussion indicates that the stabilization of radicals, i,e,
•
•
R ꢀOOꢀ R and R ꢀOꢀ R , plays important roles in the αꢀpinene oxidation upon
(a)
(b) (c)
(a)
(b) (c)
CNTs and NCNTs. Molecular dynamics simulations were carried out to further
understand the interaction between the radicals and the pristine CNTs and NCNTs
(represented by a nitrogen atom embedded in a graphene sheet). Fig. 7 shows the
calculated configurations of radicalꢀnanotube complexes. The relative position between
radicals and nanotube demonstrated that an adsorption occurred (Fig. S6), supporting the
scenario that CNTs and NCNTs stabilize the radicals. The adsorption energies between
the radicals and CNT or NCNT displayed significant difference. As shown in Table 5, the
•
adsorption energy of R ꢀOOꢀ R radicals on the NCNT is more negative than CNT,
(a)
(b) (c)
suggesting a stronger interaction between the radical and NCNT. However, for the
ꢀ
R ꢀOꢀ R , CNT displays a much more negative adsorption energy than NCNT. These
(a)
(b) (c)
results offered an excellent explanation that the E/A ratios of αꢀpinene oxidation were
efficiently modulated upon CNTs and NCNTs (Table 1). On CNTs, although the
•
relatively weak adsorption of R ꢀOOꢀ R may lead to the lower overall production of
(a)
(b) (c)
ꢀ
αꢀpinene oxide, the higher adsorption energy of R ꢀOꢀ R may be favorable for the
(a)
(b) (c)
production of αꢀpinene oxide through pathway III, resulting the higher E/A ratio. On
NCNTs, actually pathway I and III are enhanced for the production of αꢀpinene oxide,
ꢀ
meanwhile, the relatively weak adsorption between R ꢀOꢀ R and NCNTs may be
(a)
(b) (c)
4
3
favorable for the sequential allylic oxide production through pathway II or IV. Thus,
NCNT improved not only the catalytic efficiency but also the selectivities of both the
2
4

Page 25 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
epoxide and allylic oxides.
4
. Conclusions
In summary, we have demonstrated that excellent performances in the epoxidation
of αꢀpinene can be achieved in the presence of CNTs and NCNTs as metalꢀfree catalysts.
Nitrogen dopant significantly enhances not only the mass specific activity in the
oxidation of αꢀpinene, but also the selectivity of allylic oxidation products, featured by
the considerably decreased E/A ratio. An improved selectivity of epoxide was obtained as
NCNTs treated by the HNO and thermal reduction. As revealed in the mechanism
3
proposed, the interaction between radicals and carbon catalysts plays central role in this
•
•
reaction. The radicals, i.e., the R ꢀOOꢀ R and R ꢀOꢀ R , played important role in
(a)
(b) (c)
(a)
(b) (c)
the radical propagation process for producing αꢀpinene oxide. The theoretical simulations
displayed that the remarkable improvement of catalytic performance is caused by the
•
strong interaction between CNTs/NCNTs with R ꢀOOꢀ R and alkoxy radicals. The
(a)
(b) (c)
results presented here demonstrate that NCNTs is a promising alternative catalyst for the
selective production of epoxide products in the terpene oxidation by using molecular
oxygen as ultimate oxidant.
Acknowledgements
This work was supported by the National Science Foundation of China (Nos.
2
1133010, 21273079), the Guangdong Provincial National Science Foundation (Nos.
2
5

Catalysis Science & Technology
Page 26 of 31
View Article Online
DOI: 10.1039/C5CY00136F
9
251064101000020, S20120011275, 2014A030310447), Program for New Century
Excellent Talents in University (NCETꢀ12ꢀ0190) and the Fundamental Research Funds
for the Central Universities of China (No. 2014ZG0005). We thank the Information and
Network Engineering and Research Center in South China University of Technology for
the theoretical calculations.
2
6

Page 27 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
References
1
1
2
. G. B. Sul'pin, Organic Reactions Catalyzed by Metal Complexes, Nauka, Moscow,
988.
. R. A. Sheldon and J. K. Kochi, Metalꢀcatalyzed oxidations of organic compounds,
Academic Press, New York, 1981.
3
4
5
6
7
3
8
1
9
1
. U. Neuenschwander, F. Guignard and I. Hermans, ChemSusChem, 2010, 3, 75ꢀ84.
. B. P. C. Hereijgers and B. M. Weckhuysen, J. Catal., 2010, 270, 16ꢀ25.
. U. Neuenschwander and I. Hermans, J. Org. Chem., 2011, 76, 10236ꢀ10240.
. Y. Cao, H. Yu, F. Peng and H. Wang, ACS. Catal., 2014, 4, 1617ꢀ1625.
. S. M. Mahajani, M. M. Sharma and T. Sridhar, Chem. Eng. Sci., 1999, 54,
967ꢀ3976.
. U. Neuenschwander and I. Hermans, Phys. Chem. Chem. Phys., 2010, 12,
0542ꢀ10549.
. P. A. Wender and T. P. Mucciaro, J. Am. Chem. Soc., 1992, 114, 5878ꢀ5879.
0. C. Brocke, M. Eh and A. Finke, Chemistry & biodiversity, 2008, 5, 1000ꢀ1010.
11. A. Blumann and O. Zeitschel, Ber. Dtsch. Chem. Ges., 1913, 46, 1178ꢀ1198.
1
2
1
1
1
2. T. Joseph, D. P. Sawant, C. S. Gopinath and S. B. Halligudi, J. Mol. Catal. A: Chem.,
002, 184, 289ꢀ299.
3. E. Błaż and J. Pielichowski, Molecules, 2006, 11, 115ꢀ120.
4. M. V. Patil, M. K. Yadav and R. V. Jasra, J. Mol. Catal. A: Chem., 2007, 277, 72ꢀ80.
5. C. Della Pina, E. Falletta and M. Rossi, Chem. Soc. Rev., 2012, 41, 350ꢀ369.
2
7

Catalysis Science & Technology
Page 28 of 31
View Article Online
DOI: 10.1039/C5CY00136F
1
1
1
1
1
2
2
2
6. J. Thomas, Chem. Commun., 1999, 829ꢀ830.
7. S. Bhattacharjee, T. J. Dines and J. A. Anderson, J. Catal., 2004, 225, 398ꢀ407.
8. W. J. Yang, C. C. Guo, Z. Y. Li and N. Y. Tao, J. Porphyrins Phthalocyanines, 2009,
3, 973ꢀ979.
9. E. N. dos Santos, C. U. Pittman Jr and H. Toghiani, J. Mol. Catal., 1993, 83, 51ꢀ65.
0. P. Selvam, V. M. Ravat and V. Krishna, J. Phys. Chem. C, 2011, 115, 1922ꢀ1931.
1. P. McMorn, G. Roberts and G. J. Hutchings, Catal. Lett., 2000, 67, 203ꢀ206.
2. M. N. Timofeeva, Z. Hasan, V. N. Panchenko, I. P. Prosvirin and S. H. Jhung, J. Mol.
Catal. A: Chem., 2012, 363, 328ꢀ334.
2
1
2
3. D. Trong On, M. P. Kapoor, P. N. Joshi, L. Bonneviot and S. Kaliaguine, Catal. Lett.,
997, 44, 171ꢀ176.
4. D. S. Su, J. Zhang, B. Frank, A. Thomas, X. Wang, J. Paraknowitsch and R. Schlögl,
ChemSusChem, 2010, 3, 169ꢀ180.
25. Y. Wang, X. Wang and M. Antonietti, Angew. Chem. Int. Ed., 2012, 51, 68ꢀ89.
26. D. S. Su, S. Perathoner and G. Centi, Chem. Rev., 2013, 113, 5782ꢀ5816.
27. D. S. Su, N. Maksimova, J. J. Delgado, N. Keller, G. Mestl, M. J. Ledoux and R.
Schlögl, Catal. Today, 2005, 102, 110ꢀ114.
2
3
2
3
8. J. Zhang, X. Liu, R. Blume, A. H. Zhang, R. Schlögl and D. S. Su, Science, 2008,
22, 73ꢀ77.
9. M. F. R. Pereira, J. J. M. Orfão and J. L. Figueiredo, Appl. Catal. A: Gen, 2001, 218,
07ꢀ318.
2
8

Page 29 of 31
Catalysis Science & Technology
View Article Online
DOI: 10.1039/C5CY00136F
3
3
3
3
2
3
0. N. Muradov, F. Smith and A. TꢀRaissi, Catal. Today 2005, 102–103, 225ꢀ233.
1. S. Y. Lee, J. H. Kwak, G. Y. Han, T. J. Lee and K. J. Yoon, Carbon, 2008, 46,
42ꢀ348.
2. Y. J. Gao, D. Ma, C. L. Wang, J. Guan and X. H. Bao, Chem. Commun., 2011, 47,
432ꢀ2434.
3. J.ꢀH. Yang, G. Sun, Y. J. Gao, H. B. Zhao, P. Tang, J. Tan, A.ꢀH. Lu and D. Ma,
Energy Environ. Sci., 2013, 6, 793ꢀ798.
4. Y. Gao, G. Hu, J. Zhong, Z. Shi, Y. Zhu, D. S. Su, J. Wang, X. Bao and D. Ma,
Angew. Chem. Int. Ed., 2013, 52, 2109ꢀ2113.
3
3
2
3
5. G. S. Szymański, Z. Karpiński, S. Biniak and A. Świa
627ꢀ2639.
6. C. MorenoꢀCastilla, F. CarrascoꢀMarın, C. ParejoꢀPérez and M. López Ramón,
̧t kowski, Carbon, 2002, 40,
́
Carbon, 2001, 39, 869ꢀ875.
3
7. M. Hayashi, Chem. Rec., 2008, 8, 252ꢀ267.
3
8. Y. Kuang, N. M. Islam, Y. Nabae, T. Hayakawa and M.ꢀa. Kakimoto, Angew. Chem.
Int. Ed., 2010, 49, 436ꢀ440.
9. Y. Nabae, H. Rokubuichi, M. Mikuni, Y. Kuang, T. Hayakawa and M.ꢀa. Kakimoto,
ACS. Catal., 2013, 3, 230ꢀ236.
3
40. S. Yang, X. Li, W. Zhu, J. Wang and C. Descorme, Carbon, 2008, 46, 445ꢀ452.
41. Y. H. Cao, X. Y. Luo, H. Yu, F. Peng, H. J. Wang and G. Q. Ning, Catal. Sci. Tech.,
2013, 3, 2654ꢀ2660.
2
9