New Understanding of Selective Aerobic Oxidation of Ethylbenzene Catalyzed by Nitrogen-doped Carbon Nanotubes

Article abstract of DOI:10.1002/cctc.202001503

Selective aerobic oxidation of hydrocarbons undergoes a free-radical chain reaction to yield corresponding value-added products is the significant process in the chemical industry. Nanocarbons with heteroatoms doping as free-metal catalysts have been prov

Full text of DOI:10.1002/cctc.202001503

The European Society Journal for Catalysis
Supported by
Accepted Article
Title: New Understanding of Selective Aerobic Oxidation of
Ethylbenzene Catalyzed by Nitrogen-doped Carbon Nanotubes
Authors: Yongzhao Su, Yuhang Li, Zhicheng Chen, Jiangnan Huang,
Hongjuan Wang, Hao Yu, Yonghai Cao, and Feng Peng
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.202001503
Link to VoR: https://doi.org/10.1002/cctc.202001503
0
1/2020

ChemCatChem
10.1002/cctc.202001503
New Understanding of Selective Aerobic Oxidation of Ethylbenzene
Catalyzed by Nitrogen-doped Carbon Nanotubes
a†
b, c†
b
a
a
a
Yongzhao Su , Yuhang Li , Zhicheng Chen , Jiangnan Huang , Hongjuan Wang , Hao Yu ,
a*
b*
Yonghai Cao , Feng Peng
a
School of Chemistry and Chemical Engineering, Guangdong Provincial Key Lab of Green
Chemical Product Technology, South China University of Technology, Guangzhou 510640,
China
bSchool of Chemistry and Chemical Engineering, Guangzhou Key Laboratory for New Energy
and Green Catalysis, Guangzhou University, Guangzhou 510006, China
c
School of Chemistry, Sun Yat-sen University, Guangzhou 510275, China
†
These authors contributed equally.
*
Corresponding author:
E-mail: fpeng@gzhu.edu.cn (F. Peng)
meyhcao@scut.edu.cn (Y. Cao)
1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
Abstract
Selective aerobic oxidation of hydrocarbons undergoes a free-radical chain reaction to
yield corresponding value-added products is the significant process in the chemical industry.
Nanocarbons with heteroatoms doping as free-metal catalysts have been proved to be excellent
alternatives in various fields of reactions. In this work, nitrogen doped carbon nanotubes
(NCNTs) were applied for the aerobic oxidation of ethylbenzene (EB) in the liquid-phase. The
catalytic performance was unexpectedly suppressed, even lower than the control experiment,
which is totally different from earlier publications. Mechanism study demonstrated that N
doping would inhibit the abstraction of α-H from EB molecule and end the radical propagation,
thus suppressing the overall activity. Addition of TBHP would be helpful for the α-H abstraction
and forming alkyl radicals, which start the Franck-Rabinowitch cage reaction and promote
radical propagation in the presence of carbon catalysts. Herein, a higher catalytic efficiency
with 46% EB conversion and 72% AcPO selectivity were obtained over NCNTs with O
2
and
TBHP as dual-oxidants.
Keyword: Carbon catalysis; nitrogen doped carbon nanotubes; ethylbenzene oxidation;
Oxygen; tert-butyl hydroperoxide
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
1
. Introduction
Selective oxidation of alkanes for producing oxygenated derivates such as ketone, alcohol,
aldehydes etc is the significant process to the downstream chemicals in the manufacture
industry.[1] Various classic reactions are implemented in the liquid phase, for instance, the
6
oxidation of cyclohexane (CyH) to KA oil (cyclohexanone and cyclohexanol, 6×10 tons per
7
year), p-xylene to terephthalic acid (capacity 3×10 tons/year), and ethylbenzene (EB) to 1-
6
[2]
phenylethyl hydroperoxide (PEHP, capacity 5×10 tons/year), etc. The complex radical
interactions (Eqs. 1-6) usually occur in the processes when molecular oxygen is used as the
ultimate oxidant.[2a, 3] Among these elementary reactions, the homolytic dissociation of
peroxide is considered to be the major step for the chain propagation, leading to a high catalytic
performance.[2a, 3b, 4]
Initiation:
ROOH → RO + OH
(1)
(2)
(3)
(4)
(5)
(6)
RO + RH → ROH + R

OH + RH → H
2
O + R
Propagation:
Termination:
R + O → ROO
2
ROO + RH → ROOH + R
ROO+ ROO → ROH + R=O +O
2
The aerobic oxidation of aryl alkanes (typically toluene, EB and cumene etc.) to yield
corresponding ketones or alcohols still face significant challenges. Although numerous metal-
based catalysts are developed, the high cost, toxic property and complex preparation procedure
greatly hinder their commercial applications. Recently, nanocarbons (carbon nanotubes (CNTs),
carbon fiber, graphene, carbon black, diamond, etc.) as metal-free catalysts have been
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
intensively investigated and proved to be alternatives to the state of art metal-based catalysts in
[5]
various reactions. Our previous work demonstrated that CNTs as a catalyst displayed high
[
6]
activities in aerobic oxidation of cumene and EB. An excellent catalytic activity with 38.2%
EB conversion and 60.9% AcPO selectivity over CNTs could be obtained.[6b] Mechanism study
showed that a weak interaction between carbon atoms and active intermediate radicals
happened, which facilitated the decomposition of PEHP to improve the catalytic activity and
yield desired products.[6b, 7]
N doping is a traditional protocol to modify nanocarbons, which would greatly improve
the catalytic activity through enhancing the interaction between the surface carbons and
intermediate radicals.[5c, 7-8] Meanwhile, N doping is also believed to facilitate the
decomposition of peroxides to improve the catalytic performance subsequently.[5a, 5b, 5g, 6a]
However, there is few work focused on aerobic oxidation of EB catalyzed by N doped carbons.
A number of works from other groups preferentially utilized tert-butyl hydroperoxide (TBHP)
as terminal oxidant instead of O
2
in the presence of N doped carbons for the EB oxidation.[9]
2
For instance, Ma and co-worker reported that N doped sp -hybridized carbon as catalysts were
used for EB oxidation with high mole ratio TBHP (3 times as substrate) as the terminal
oxidant.[9a, 10] As high as 98.6% EB conversion and 91.3% AcPO yield are obtained. They
believed that surface carbons would host reactive oxygen species during EB oxidation after N
doping, thus improving the activity.[9a] In fact, when TBHP as oxidant for the hydrocarbon
oxidation is performed, it could be decomposed into different radicals via Eqs. 7-9. These
radicals can be involved to active C-H bond of hydrocarbons to obtain oxygenated products.[3b]
This mechanism is similar to the reaction system with O as oxidation. Nevertheless, for the
2
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
bulk chemical industry, O
2
is still the priority choice for the liquid phase reaction. Herein, we
applied NCNTs as a catalyst for EB oxidation with O as terminal oxidant and found N doping
2
could inhibit carbon catalyst activity unexpectedly, which has never been explored before.
t-BuOOH → t-BuO• + •OH
(7)
(8)
(9)
t-BuOOH + t-BuO• → t-BuOH + t-BuOO•
2
t-BuOO• → 2t-BuO• + O
2
In this work, NCNTs were selected as the catalyst for the EB oxidation in the liquid-phase
using O as the terminal oxidant. A new phenomenal of activity suppression was observed,
2
which is totally different from the work reported. N dopant acts as an inhibitor for the
production of alkyl radical at the initiate stage. Detailed investigations were then carried out to
explore the effect of N doping on the aerobic oxidation of EB. This work provides the new
understanding for the heteroatoms doped carbon-catalysis in the liquid-phase reaction.
2
2
. Experimental sections
.1 Materials and chemical reagents
Pristine CNTs (CNT) was obtained from Zhongshan CNM Plastic Co., Ltd, and the metal
impurities in CNT was tested by ICP-AES (Table S1). CNTs was washed by concentrated HCl
to exclude metal impurity before use. TBHP, EB, acetonitrile (CH
3
CN) and anisole were
purchased from Shanghai Lingfeng Chemical Reagent Co., LTD.
2
.2 NCNTs preparation
NCNTs was synthesized through a chemical vapor deposition of pyridine on the washed
CNTs.[8a, 11] Briefly, CNTs was placed into a quartz tube and heated from room temperature to
o
o
7
60 C with a rate of 10 C/min under N
2
atmosphere. Pyridine solution was charged into the
5
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
quartz tube with 2 mL/h for 1 h. After that, the furnace was cooled down in N
2
atmosphere and
NCNTs was obtained. The obtained NCNTs was noted as NCNTs@1H. Similar process was
carried out with pyridine solution for 0.3 h, which was noted as NCNTs@0.3H.
2
.3 Characterization
The Brunauer-Emmett-Teller (BET) surface areas and pore size of different materials were
investigated by N
2
adsorption at liquid N temperature in an ASAP 2010 analyzer. TEM and
2
HRTEM images were recorded in a JEM-2100 microscope operating at 200 kV. A small amount
of catalyst was ultrasonically suspended in absolute ethanol, and depositing a drop of the
suspension onto a grid with amorphous copper film for the measurement of TEM and HRTEM.
Raman spectra were obtained with a LabRAM Aramaic micro Raman spectrometer with an
excitation wavelength at 633 nm with 2 μm spot size to investigate the structure of catalysts.
FTIR spectra were measured on a Perkin-Elmer model 783IR spectrophotometer. XPS analysis
was performed with an ESCALAB 250 spectrometer equipped with an Al K
binding energies were referenced to the C1s peak at 284.6 eV.
α
X-ray source, the
2
.4 EB oxidation catalyzed by NCNTs
The EB oxidation reaction to obtain value-added products (Scheme 1) was carried out in
a 50 mL stainless-steel autoclave with a Teflon-liner. Typically, EB (3 mL), CH CN (20 mL)
and anisole (C OCH , 650 μL) as reactant, solvent and internal standard, respectively, were
sonicated for 15 min in a conical flask, and then transferred to the autoclave, which was charged
times with argon to remove air. The reaction was conducted at 155 °C with 1000 rpm, and O
3
6
H
5
3
3
2
was kept uninterrupted to maintain the reaction pressure at 1.5 MPa. After reaction, the reactor
was cooled down to room temperature and the mixed liquid products were obtained. For
6
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
comparation of the different oxidation system, the mixed reactants followed the aforementioned
procedure expect TBHP as oxidant with or without O . The liquid mixture was analyzed by gas
chromatography, which was performed using an Agilent GC-7890B equipped with a 30 m ×
.32 mm × 0.25 μm 123-0732E capillary column and a flame ionization detector. To determine
the amount of PEHP, one of the products of the reaction, the excess triphenylphosphine (PPh
2
0
3
)
was added into the obtained liquid sample to convert PEHP to stoichiometric 1-phenylethyl
alcohol (PEA) at room temperature. [12]
Scheme 1 Selective aerobic oxidation of EB to oxygenated products catalyzed by CNTs.
2
.5 TBHP decomposition
5 mg catalyst and 10 mL TBHP aqueous solution were added into a 25 mL three-necked
3
flask. Nitrogen gas was introduced to replace the air for 5 minutes, and then the flask was sealed.
After that, the temperature was heated to 80°C and kept for 0.5 hours. When the reaction was
finished, the solution was cool down to the room temperature and the concentration of TBHP
was determined by iodometry method.
2
.6 Computational details
Density functional theory (DFT) calculations were performed with DMol3 package.[13]
Herein, the graphene nanoribbon model was built to mimic the CNTs surface, as shown in Fig.
7
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
S1. Furthermore, three carbon atoms were replaced by nitrogen to represent the NCNTs surface,
as shown in Fig. S2. All the reaction-related configurations were optimized with PBE functional
under DNP basis set. All electron was adopted as the core treatment. DFT-D correction based
on Grimme method was considered.[14] Electronic spin was unrestricted. Global orbital cutoff
was set to 4.8 Å for all atoms. The k-point was set to 8×1×1. Besides, the COSMO solvation
model was involved with dielectric constant 37.5 of acetonitrile. All calculations were tightly
converged without setting electronic smearing.
3
. Results and discussion
Table 1 Catalytic performance over nanocarbons in the selective oxidation of EB.[a]
Conversion
Selectivity (%)
(%)
S
BET
R [
w
b]
R
s
[c]
Entry
Catalysts
2
2
(
m /g)
(mmol/g/h)
(mmol/m /h)
others[d]
1
2
3
4
5
6
7
8
Blank
CNTs
-
-
-
0.3
-
9
12
25
-
8
12
-
86
41
-
4
22
-
178
133
133
161
133
132
150
42
22
~0
~0
~0
23
6
NCNTs@1H[e]
NCNTs@1H[f]
NCNTs@0.3H[g]
NCNTs[h]
-
-
-
-
-
-
-
-
186
-
-
-
-
-
-
1.4
-
45
65
-
13
18
-
30
2
13
15
NCNTs-r[i]
NCNTs-r-600[j]
-
-
-
-
-
[a]
o
[b]
Reaction conditions: 20 mL CH
3
CN, 3 mL EB, 35 mg catalyst, 1.5 MPa O
2
, 155 C, 4 h.
Initial rate of EB consumption per gram of catalyst. [c] Initial rate of EB consumption per m of
2
catalyst surface. [d] The other products include benzoic acid, benzaldehyde and other esters. [e]
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
NCNTs@1H. [f] 65 mg NCNTs@1H, 12h. [g] NCNTs@0.3H, 1 at.% N content. [h] 5 mg
NCNTs@1H. [i] NCNTs@1H after one reaction was denoted as NCNTs-r. NCNTs@1H after
one reaction and then heat treatment at 600 ℃ with argon for 2 h was denoted as NCNTs-r-600.
The selective oxidation of EB catalyzed by different carbon catalysts were performed and
summarized in Table 1. CNTs and NCNTs (3 at. % N, Fig. S3) were selected as model catalysts
for this reaction (texture property could be found in Fig. S4 and Table S2). As shown in Figure
[j]
1
, CNTs with typically cylindrical graphitic walls was observed. NCNTs was synthesized via a
surface N modification method.[11, 15] Small amount of amorphous carbons outside of carbon
wall was observed in NCNTs, which probably come from the pyridine pyrolysis.[15] In the
control experiment, 9% EB conversion from the self-autoxidation was obtained.[3b] Adding
CNTs (Table 1, entry 2) into the reaction system improved EB conversion to 22% with 25%
AcPO selectivity. This catalytic performance of CNTs is the result of the interaction between
radical species and graphene sheets, which accelerates the decomposition of PEHP.[6b]
Meanwhile, Fe impurities have been verified to be negligible effect on EB conversion,
indicating that this process is intrinsically the metal-free catalysis.[6b] Reported studies
demonstrated that N doping could greatly enhance the catalytic activity of carbon catalysts,
which was ascribed to the capability of facilitating the decomposition of peroxides and the
propagation of free radical chains in the selective oxidation of hydrocarbons.[5b, 5c, 16]
Unexpectedly, when NCNTs@1H (Table 1, entry 3) was used as the catalyst, EB conversion
was ~0%, much lower than that of blank or CNTs experiment. This result indicated that N
doping could not improve the catalytic performance of nanocarbon catalysts but suppressed the
activity of EB oxidation, which is rarely observed in other hydrocarbon oxidation processes.
9
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
On the basis of reported literatures, some metal-based catalysts could usually facilitate the
radical propagation in the hydrocarbons oxidation in the liquid phase.[17] Meanwhile, the typical
induction period would occur before the reaction starts. The addition of metal-based catalysts
varies the induced chain lengths and shortens the induction period.[17b-e] In current work, we
found that N doping could efficiently suppress the overall activity totally, even we prolonged
reaction time to 12 h (Table 1, entry 4). At the same time, we shorted the time to 0.3 h for
pyridine decomposition on NCNTs synthesis (noted as NCNTs@0.3H). 1 at.% N content on
NCNTs@0.3H was obtained (Fig. S5). As predicted, no activity was observed as well (Table
1
, entry 5). These results confirmed that N dopant likely acted as an inhibitor for the production
of alkyl radical at the initiate stage in this work. Herein, we believed that the intrinsic
mechanism is totally different from metal-catalysis systems.
Furthermore, the used NCNTs@1H was filtered after the reaction and reapplied as a
catalyst (noted as NCNTs-r) for EB oxidation directly. A considerable EB conversion (6%) with
6
5% selectivity of AcPO were obtained. Herein, we suspected that some sites on carbon surface
for prohibiting EB conversion were covered, restoring low EB conversion rate. The recycled
o
NCNTs@1H was then annealed at 600 C under Ar atmosphere (noted as NCNTs-r-600) and
utilized as a catalyst for EB oxidation. As expected, EB conversion was totally suppressed once
again. Hitherto, the inhibition of selective aerobic oxidation by nitrogen-doped carbon materials
has not been reported, and the surface sites for the activity inhibition of NCNTs have not been
defined, which need to be revealed. For the sake of simplification, NCNTs@1H is represented
by NCNTs in the following discussion.
10
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
Figure 1 TEM and HRTEM images of CNTs (a, b) and NCNTs@1H (c, d)
(a)
OOH
O
(b)
80
60
40
20
0
80
60
40
20
0
OH
0
20
40
60
80
100
120
0
10
20
30
40
CNTs amount (mg)
NCNTs amount (mg)
Fig. 2 The effect of CNTs (a) and NCNTs@1H (b) amount on the selective oxidation of EB
o
with O
2
, conditions: 20 mL CH
3
CN, 3 mL EB, 1.5 MPa O
2
, 155 C, 4 h. EB conversion (●),
PEA selectivity (○), AcPO selectivity (□), PEHP selectivity (△).
11
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
In current work, N atoms probably acted as the significant sites for the inhibition of EB
oxidation. Reducing the gross N sites might be helpful to recover NCNTs activity. Comparably,
CNTs without N doping displayed a feasible enhancement for the EB conversion according to
the increasing catalyst dosage (Fig. 2a), which is consistent with our previous study.[6b] Herein,
NCNTs with different dosage in the reaction was investigated. As shown in Fig. 2b, a
considerable EB conversion was observed when NCNTs dosage was less than 25 mg. The
highest EB conversion (23%) was obtained using 5 mg NCNTs. Ketone selectivity increased
alone with NCNTs dosage increasing and then declined sharply (35 mg NCNTs). Since EB
oxidation with O
certain radicals on carbon surface should be produced to propagate the chain reaction (Eqs.1-
), we speculate that certain elementary reactions might be inhibited on the surface of NCNTs.
2
as terminal oxidant is the classic free-radical chain reaction and numerous
6
The catalytic performances of NCNTs-r and NCNTs-r-600 further suggested that the inhibition
effect on NCNTs surface probably exist. Comprehensive characterizations were then carried
out to verify the physical and chemistry variation of NCNTs catalysts. TEM images showed
that no significant change in NCNTs was obtained (Fig. S6). Raman results revealed that
surface defect degree of NCNTs increased slightly after reaction (Fig. S7), which might be
come from the effect of oxidation procedure. XPS results further confirmed that the stability of
NCNTs during the different steps could be obtained (Fig. S8, Table S3 and S4). The content of
C, N elements were not changed significantly. The minor increase of O content could be
ascribed to the oxidation procedure. Overall, no significant change over NCNTs was observed.
Generally, hydrocarbons oxidation catalyzed by NCNTs with O
2
as an oxidant undergoes
two significant steps: the homolytic cleavage of peroxides and α-H abstraction of substrate
12
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
molecule.[3b, 18] Aforementioned results show that NCNTs with 5 mg provided 30% selectivity
for PEHP, much lower than that of CNTs, demonstrated that N doping still could enhance the
dissociation of PEHP molecule to propagate the chain reaction in this reaction (Eq. 1).[5c, 7] The
higher dosage of NCNTs could decrease the selectivity of PEHP apparently, but the dissociation
of PEHP did not improve EB conversion while suppressed the activity. Meanwhile, if EB
molecule activation would be inhibited in the presence of NCNTs, the overall activity would be
totally suppressed even at low doses of NCNTs. However, EB conversion was still slightly
enhanced when NCNTs dosage was quite low (Fig. 2b). These results indicated that the reaction
mechanism of EB oxidation catalyzed by NCNTs is totally different from the CNTs-catalyzed
system. N doping probably changed the radical reaction routing. In fact, Hermans et al. have
demonstrated both experimentally and theoretically that α-H abstraction by ROO, RO or HO
from substrate to produce carbon-centered radicals (R) are significant steps for chain
propagation (Eqs. 2, 3, 5) in the hydrocarbon oxidation.[3b, 7] Herein, in this study, we believe
that the homolytic cleavage of the existed PEHP was slow enough to produce EBO and HO
radicals when NCNTs dosage was very low, and the considerable EB conversion could be
obtained. When the amount of NCNTs was relatively high (more than 5 mg), the homolytic
cleavage rate of PEHP was too fast to generate active radicals effectively, and EB conversion
rate gradually decreased. When the amount of NCNTs reached 35mg, PEHP was quickly
decomposed and disappeared, and there was no time for the reactive radical to react with EB
molecule, resulting in a zero conversion rate of EB, as expected.
Some of the arguments might focus on the effect of solvent. In this work, EB oxidation
without solvent catalyzed by NCNTs was performed (Fig. 3). Results demonstrated that the
13
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
similar activity trend was observed. The activity first increased and then decreased alone with
the increase of NCNTs dosage. When 300 mg NCNTs was used, EB activity was totally
suppressed. Also, N functionalities might be a key role for suppressing activity. It is of great
significance for us to verify the role of N functionalities such as pyridinic N, graphitic N,
pyrrolic N etc. in this reaction. However, based on the data from Table 1, we can conclude that
N doping with low content (1 at.%, Fig. S5) still can suppress activity of CNTs. Meanwhile,
even we tuned the ratio of N functionalities (Fig. S5 and Table S4), the activity of CNTs was
still not changed. This result further confirms that N doping could only enhance activity in the
limited mole ratio range between NCNTs and EB in the liquid-phase oxidation. Moreover, in
current work, we still can not figure out the intrinsic role of different types of N and further
studies are in progress in our group to verify this idea.
100
O
80
60
40
20
0
O OH
O H
0
50
100
150
200
250
300
NCNTs amount (mg)
Fig. 3 EB oxidation without solvent catalyzed by NCNTs, conditions: 25 mL EB, 1.5 MPa
o
O
2
, 155 C, 4 h. EB conversion (●), PEA selectivity (○), AcPO selectivity (□), PEHP
selectivity (△).
Herein, there is a question that where is PEHP from before the reaction. As we know that
14
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
commercial products of EB might contain a trace amount of PEHP from the oxidized EB by O
2
in the bottle before use. When the reaction started, the existed PEHP could promote the reaction
more quickly. Under a higher dosage of NCNTs, although there was no conversion of EB after
4
hours of reaction, the EB after reaction was exactly the same as the original EB because the
trace PEHP in EB was quickly decomposed. To confirm our speculation, the recycled EB after
reaction catalyzed by 35 mg NCNTs was obtained. The recycled EB was then used for the
reaction catalyzed by CNTs, EB conversion decreased to 16%, much lower than the pristine
one (34%, Table 2), indicating that the oxidation activity of recycled EB catalyzed by CNTs is
indeed much lower than that of the pristine EB due to the disappearance of trace PEHP. This
result also implied that the considerable EB conversion over NCNTs with low dosage might be
ascribed to the existed PEHP.
Table 2 Catalytic performance of CNTs in the oxidation of treated EB.[a]
Conversion (%)
Selectivity (%)
Entry
EB solution
Catalysts
Others [b]
1
2
Pristine
CNTs
CNTs
34
16
55
22
15
10
13
44
17
14
After reaction[c]
[a]
Reaction conditions: 20 mL CH
3
CN, 3 mL EB, 1.5 MPa O
2
, 65 mg catalyst, 1.5 MPa O , 155
2
oC, 4 h. [b] The other products include benzoic acid, benzaldehyde and other esters. [c] EB
solution was obtained by filtration after the reaction catalyzed by 35 mg NCNTs.
For the effect of N doping on CNTs in the EB oxidation, DFT simulations were carried out
to reveal the interactions between the carbons and intermediate radicals on the surface of
different catalysts. CNTs and NCNTs with 3% N concentration were selected as model catalysts.
15
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
In the initial stages of EB oxidation, several radicals such as EB, EBO, EBOO radicals, etc.,
were involved in this process.[6b] As mentioned, EB oxidation in the liquid phase undergoes a
free radical chain process (Eqs. 1-9), the initial stage catalyzed by nanocarbon plays a crucial
role for the overall activity.[5b, 5e, 6b, 19] In this work, EB molecule would disperse onto the surface
of nanocarbon catalyst and be adsorbed on the surface. And then, EB molecule would be
activated by π–π interaction between the graphene sheets of CNTs and the benzene ring of
EB,[6b, 20] where α-H would be abstracted by peroxide or nanocarbons to form EB•. The insertion
of O into EB• to produce EBOO• occures at the following step. Therefore, the reaction would
2
be led into the Franck-Rabinowitch cage and radical propagation is promoted.[3b] During these
steps, CNTs and NCNTs display the different performances from the calculation results (Fig.
4
). As shown that ∆E of different steps on CNTs are lower than those on NCNTs, implying that
CNTs is a better catalyst in this reaction. Among these steps, the α-H abstraction from EB
molecule is vital procedure for the following steps. The ∆E for this process in the presence of
CNTs is found to be -1.356 eV, demonstrating that CNTs could facilitate the α-H abstraction
from EB molecule. It is worth noting that ∆E with the presence of NCNTs is found to be +0.758
2
2
eV, indicating that N doping is not favarable to promoting α-H abstraction, but inhibits the
production of these crucial active radicals, thereby hindering the overall activity. This results is
consistent with our experiment results.
16
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
Fig. 4 Reaction energy variations (∆E) of key intermediate radical conversions on CNTs and
NCNTs surface from DFT calculations.
In order to further verify our hypothesis and explain the mechanism of catalytic oxidation
of ethylbenzene by NCNTs, we try to bypass initiate stage (Eq. 1-3) and directly get into the
chain-propagation process (Eq. 4-6). Currently, many researches used TBHP as initiator or
oxidant in the EB oxidation, just because TBHP could be decomposed into CH
3
COCH
3
and
•OH, which could activate EB molecule more easily and thus it launches a radical reaction (Eqs.
7
-9).[21] Additionally, TBHP as solo oxidant (Fig. 5a) displayed the improved activity without
inhibition behavior in EB oxidation in the presence of NCNTs, which is also observed in other
groups work.[9a, 9b]
17
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
(a)
(b) 80
80
60
40
20
0
60
40
20
0
0
20
40
60
80
100
120
0
10
20
30
40
50
60
NCNTs dosage (mg)
TBHP dosage (wt%)
(c)
80
60
40
20
0
(d) 80
60
40
20
0
0
10
20
30
40
50
60
70
0
20
40
60
80
100
CNTs dosage (mg)
NCNTs dosage (mg)
Fig. 5 (a) The effect of NCNTs dosage on EB oxidation with TBHP as oxidant, conditions: 20
o
mL CH
3
CN, 3 mL EB, MEB : MTBHP = 1:2, 80 C, 4 h. (b) The effect of TBHP amount (based
on EB dosage) on EB oxidation by NCNTs, conditions: 20 mL CH
3
CN, 3 mLEB, 35 mg catalyst,
and
, 155 C, 4 h, 30 wt%
o
1
.5 MPa O
2
, 155 C, 4 h. The EB oxidation catalyzed by CNTs (c) and NCNTs (d) with O
2
o
TBHP as oxidants, conditions: 20 mL CH
3
CN, 3 mL EB, 1.5 MPa O
2
TBHP (based on EB dosage). EB conversion (●), PEA selectivity (○), AcPO selectivity (□),
PEHP selectivity (△).
On the basis of aforementioned discussion, TBHP was added into the reaction system to
bypass the prohibition behavior in EB oxidation catalyzed NCNTs with 1.5 MPa O . Before
2
evaluation, the effect of TBHP amount on EB oxidation was investigated (Fig. 5b). EB
conversion increased from 0% to 47% alone with the increase of TBHP dosage (0 to 60 wt.%),
where the highest selectivity of AcPO is 75%. This exciting result further confirms our
18
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
conjecture that the initiation of α-H abstraction from EB molecules forming alkyl radicals is
probably a key step in the inhibition of EB oxidation activity. The decomposition of TBHP to
produce t-BuO• and HO• radicals, which would react with EB molecule (Eqs. 10-12). The
produced EB• radicals then were inserted with O and the Franck-Rabinowitch cage reaction
2
[7]
could be initiated (Eq.13). Furthermore, it is found that the low-dosage of TBHP as an initiator
in the EB oxidation did not significantly improve the reaction activity (Table S5), indicating
that TBHP acts as an oxidant rather than an initiator in this study. To verify the role of N doping,
CNTs and NCNTs as catalysts with different dosages for EB oxidation were performed. When
CNTs was used as the catalyst, EB conversion slightly increased from 35% to 46% (Fig. 5c),
showing a small influence on the improvement of EB oxidation. In this process, PEHP (~35%)
and AcPO (~45%) are the majority in the product distribution. EB conversion firstly increased
and then decreased alone with the increasing of NCNTs dosage. The highest EB conversion is
4
6% with 72% AcPO selectivity, which is the main products in this process (Fig. 5d). The
activity decline is probably ascribed to the rapid decomposition of TBHP at high dosage of
NCNTs, which is also confirmed by the experiment of TBHP decomposition promoted by
NCNTs (Table S6). Noting that 35% AcPO and 39% PEHP selectivity was obtained even
without catalyst. After the addition of NCNTs, the selectivity for PEHP decreased significantly
while the selectivity for AcPO increased accordingly, suggesting that NCNTs still enhanced the
transformation of PEHP to yield AcPO in this study.
t-BuO• + EB → EB• + t-BuOH
(10)
(11)
(12)
t-BuOO• + EB → EB• + t-BuOOH
HO• + EB → EB• + H O
2
19
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
EB + O
As mentioned, the addition of TBHP is helpful for α-H abstraction process. The insertion
to accelerate the propagation of intermediate radicals is also important for this radical
reaction. It might be argued that the improved EB conversion rate was only resulted from the
effect of TBHP. To verify the important role of O for the enhanced activity, several experiments
with different oxidants were carried out. For comparison, solo TBHP as oxidant under 1.5 MPa
atmosphere was conducted (Fig. 6). Relatively low activities were observed for blank (9%),
2
→ EBOO
(13)
of O
2
2
N
2
CNTs (17%), NCNTs (13%), respectively. It is notable that EB conversion over NCNTs is lower
than CNTs, because that NCNTs could promote the decomposition of peroxides.[5e] After that,
experiments were carried out at O
blank (35%), CNTs (41%), NCNTs (46%), respectively, were obtained. These results confirmed
that the improved EB conversion was originated from the synergistic promotion of O and
2
atmosphere. Significant improvement on EB conversion for
2
TBHP as oxidants. Meanwhile, the radical reactions in the presence of molecular oxygen still
dominate the reaction catalyzed by NCNTs.
80
60
40
20
0
Blank
CNTs
NCNTs
1
8
6
4
2
0
00
0
0
0
0
N +TBHP O +TBHP N +TBHP O +TBHP N +TBHP O +TBHP
2
2
2
2
2
2
Fig. 6 The selective oxidation of EB catalyzed by carbon materials with different oxidants.
o
Conditions: 20 mL CH
3
CN, 3 mL EB, 15 mg catalyst (if applicable), 155 C, 4 h, 1.5 MPa N
2
20
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
or O
Furthermore, high selectivity of AcPO could be obtained in the presence of TBHP. In the
blank experiment, only 12% selectivity for AcPO was obtained with O as the oxidant (Table
, entry 1). The addition of TBHP improved the selectivity of AcPO to 35% (Fig. 5b),
2
, 30 wt% TBHP (based on EB dosage).
2
1
demonstrating that key elementary reaction for the production of AcPO was facilitated.
According to Fig. 2b, we can find that the selectivity for AcPO would increase with the increase
of NCNT dosage (below 30 mg). And when 25 mg NCNTs were used without TBHP, as high
as 74% selectivity for AcPO could be obtained. Moreover, the introduction of TBHP at smaller
dosage of NCNTs (15 mg) could also afford high selectivity for AcPO (74%) (Fig. 5d). On the
basis of mechanism study from Hermans’ group, the alkoxy radicals (EBO) can abstract H-
atoms from the alkane substrate or decompose via β-scission to form alcohol (Route Ⅰ) or ketone
(
Route II) (Fig. S9).[3b] In this work, nanocarbons could accelerate the route II to generate more
ketone with the absence of TBHP. The addition of TBHP could enhance the formation of EBO
through Eqs. 13, 5 and 1. Therefore, we believe that the high selectivity of AcPO is mainly
originated from the catalytic ability of NCNTs, while the addition of TBHP may accelerate the
route II for the production of AcPO.
4
. Conclusion
In this study, the aerobic oxidation of EB in the liquid-phase catalyzed by CNTs and
NCNTs was evaluated. CNTs displayed a considerable activity in EB oxidation, which is
consistent with our previous work. While, EB oxidation activity was totally suppressed when
NCNTs was used at higher doses. Mechanism study and DFT calculations demonstrated that N
dopant probably played the vital role for hindering the abstraction of α-H from EB molecule,
21
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
which ended the initial stage of radical propagation in EB oxidation. By adding the TBHP as
dual-oxidant would be helpful to skip the suppressed step and produce the peroxides radicals.
Therefore, the radical propagation was launched and EB conversion could be improved
subsequently. This study provides a new understanding of the catalytic oxidation of EB by
nitrogen doped nanocarbons.
Supporting Information
The additional characterizations and discussions can be found in the Supporting
Information.
Conflicts of interest
There are no conflicts to declare.
Acknowledgments
We would like to acknowledge the financial support from the National Natural Science
Foundation of China (Nos. 21872041, 21906058 and 21978107), Fundamental Research Funds
for the Central Universities of China (No. 2019MS037) and Science and Technology Program
of Guangzhou City (No. 201804010112 and 202002030287).
References
[
1] J. G. Calvert, R. G. Derwent, J. J. Orlando, T. J. Wallington, G. S. Tyndall, Mechanisms of
atmospheric oxidation of the alkanes, OUP USA, 2008.
[
2] a) I. Hermans, P. A. Jacobs, J. Peeters, Chem. Eur. J. 2006, 12, 4229-4240; b) R. A. Sheldon,
J. K. Kochi, Academic Press, New York 1981; c) G. Franz, R. A. Sheldon, Ullmann,
Wiley-VCH, Weinheim 2000; d) B. Sumit, M. Doble, Homogeneous catalysis:
mechanisms and industrial applications, John Wiley & Sons, 2000.
22
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
[
[
[
3] a) I. Hermans, P. Jacobs, J. Peeters, Chem. Eur. J. 2007, 13, 754-761; b) I. Hermans, J.
Peeters, P. A. Jacobs, J. Org. Chem. 2007, 72, 3057-3064.
4] a) U. Neuenschwander, F. Guignard, I. Hermans, ChemSusChem 2010, 3, 75-84; b) U.
Neuenschwander, E. Meier, I. Hermans, ChemSusChem 2011, 4, 1613-1621.
5] a) Y. Cao, H. Yu, J. Tan, F. Peng, H. Wang, J. Li, W. Zheng, N.-B. Wong, Carbon 2013, 57,
4
33-442; b) Y. Cao, H. Yu, F. Peng, H. Wang, ACS Catal. 2014, 4, 1617-1625; c) X.
Yang, Y. Cao, H. Yu, H. Huang, H. Wang, F. Peng, Catal. Sci. Tech. 2017, 7, 4431-4436;
d) Y. Cao, B. Li, G. Zhong, Y. Li, H. Wang, H. Yu, F. Peng, Carbon 2018, 133, 464-473;
e) C. Mu, Y. Cao, H. Wang, H. Yu, F. Peng, Chem. Eng. Sci. 2018, 177, 391-398; f) G.
Zhong, H. Wang, H. Yu, H. Wang, F. Peng, Electrochim. Acta 2016, 190, 49-56; g) H.
Yu, F. Peng, J. Tan, X. W. Hu, H. J. Wang, J. A. Yang, W. X. Zheng, Angew. Chem. Int.
Ed. 2011, 50, 3978-3982; h) P. Q. Yan, B. S. Zhang, K. H. Wu, D. S. Su, W. Qi, Carbon
2
2
019, 143, 915-936; i) C. Wang, W. Liu, S. M. Wei, D. S. Su, W. Qi, ChemCatChem
019, 11, 397-400; j) G. Wen, Q. Gu, Y. Liu, R. Schloegl, C. Wang, Z. Tian, D. S. Su,
Angew. Chem. Int. Ed. 2018, 57, 16898-16902; k) D. Su, Y. Ding, X. Huang, X. Yi, Y.
Qiao, X. Sun, A. Zheng, Angew. Chem. Int. Ed. 2018, 57, 13800-13804; l) Y. Nabae, H.
Rokubuichi, M. Mikuni, Y. Kuang, T. Hayakawa, M.-a. Kakimoto, ACS Catal. 2013, 3,
2
5
3
30-236; m) Y. Lin, B. Li, Z. Feng, Y. A. Kim, M. Endo, D. S. Su, ACS Catal. 2015, 5,
921-5926; n) N. Gupta, O. Khavryuchenko, A. Villa, D. Su, ChemSusChem 2017, 10,
030-3034; o) D. S. Su, J. Zhang, B. Frank, A. Thomas, X. Wang, J. Paraknowitsch, R.
Schloegl, ChemSusChem 2010, 3, 169-180.
[6] a) S. Liao, Y. Chi, H. Yu, H. Wang, F. Peng, ChemCatChem 2014, 6, 555-560; b) J. Luo, F.
23
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
Peng, H. Yu, H. J. Wang, W. X. Zheng, ChemCatChem 2013, 5, 1578-1586; c) S. Liao,
F. Peng, H. Yu, H. Wang, Appl. Catal. A: Gen 2014, 478, 1-8.
[
7] X. X. Yang, H. J. Wang, J. Li, W. X. Zheng, R. Xiang, Z. K. Tang, H. Yu, F. Peng, Chem.
Eur. J. 2013, 19, 9818-9824.
[8] a) H. Fu, K. Huang, G. Yang, Y. Cao, H. Wang, F. Peng, Q. Wang, H. Yu, ACS Catal. 2020,
1
0, 129-137; b) X. Duan, H. Sun, Y. Wang, J. Kang, S. Wang, ACS Catal. 2015, 5, 553-
59; c) W. Li, Y. Gao, W. Chen, P. Tang, W. Li, Z. Shi, D. Su, J. Wang, D. Ma, ACS
5
Catal. 2014, 1261-1266; d) H. Watanabe, S. Asano, S.-i. Fujita, H. Yoshida, M. Arai,
ACS Catal. 2015, 5, 2886-2894; e) A. Villa, N. Gupta, O. Khavryuchenko, D. Su,
ChemSusChem 2017, 10, 3030-3034.
[
9] a) Y. Gao, G. Hu, J. Zhong, Z. Shi, Y. Zhu, D. S. Su, J. Wang, X. Bao, D. Ma, Angew. Chem.
Int. Ed. 2013, 52, 2109-2113; b) S. Yang, L. Peng, P. Huang, X. Wang, Y. Sun, C. Cao,
W. Song, Angew. Chem. Int. Ed. 2016, 55, 4016-4020; c) Z. Ma, H. Zhang, Z. Yang, G.
Ji, B. Yu, X. Liu, Z. Liu, Green Chem. 2016, 18, 1976-1982; d) S. Balasubramanyan, S.
Sugunan, B. N. Narayanan, J. Mater. Sci 2018, 53, 12079-12090.
[
10] P. Tang, Y. Gao, J. Yang, W. Li, H. Zhao, D. Ma, Chin. J. Catal. 2014, 35, 922-928.
11] Z. He, B. Dong, W. Wang, G. Yang, Y. Cao, H. Wang, Y. Yang, Q. Wang, F. Peng, H. Yu,
ACS Catal. 2019, 9, 2893-2901.
[
[
12] a) G. Yang, Y. Ma, J. Xu, J. Am. Chem. Soc. 2004, 126, 10542-10543; b) H. Ma, J. Xu, Q.
Zhang, H. Miao, W. Wu, Catal. Commun. 2007, 8, 27-30; c) S. Evans, J. R. L. Smith, J.
Chem. Soc., Perkin Trans. 2 2000, 1541-1551; d) S. Evans, J. R. L. Smith, J. Chem.
Soc., Perkin Trans. 2 2001, 174-180.
24
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
[
[
[
13] B. Delley, J. Chem. Phys. 2000, 113, 7756-7764.
14] S. Grimme, J. Comput. Chem. 2006, 27, 1787-1799.
15] X. Ning, Y. Li, J. Ming, Q. Wang, H. Wang, Y. Cao, F. Peng, Y. Yang, H. Yu, Chem. Sci.
019, 10, 1589-1596.
2
[
16] a) P. Tang, G. Hu, M. Li, D. Ma, ACS Catal. 2016, 6, 6948-6958; b) D. S. Su, S. Perathoner,
G. Centi, Chem. Rev. 2013, 113, 5782-5816.
[
17] a) E. Spier, U. Neuenschwander, I. Hermans, Angew. Chem. Int. Ed. 2012, 52, 1581-1585;
b) O. A. Kholdeeva, M. N. Timofeeva, G. M. Maksimov, R. I. Maksimovskaya, W. A.
Neiwert, C. L. Hill, Inorg. Chem. 2005, 44, 666-672; c) O. Kholdeeva, V. Grigoriev, G.
Maksimov, M. Fedotov, A. Golovin, K. Zamaraev, J. Mol. Catal. A: Chem. 1996, 114,
1
23-130; d) R. R. Hiatt, K. C. Irwin, C. W. Gould, J. Org. Chem. 1968, 33, 1430-1435;
e) J. Black, J. Am. Chem. Soc. 1978, 100, 527-535.
[18] I. Hermans, J. Peeters, P. A. Jacobs, Top. Catal. 2008, 50, 124-132.
[19] Y. Cao, Y. Li, H. Yu, F. Peng, H. Wang, Catal. Sci. Tech. 2015, 5, 3935-3944.
[20] W. Chen, L. Duan, D. Zhu, Environ. Sci. Technol. 2007, 41, 8295-8300.
[21] a) R. R. Hiatt, K. C. Irwin, J. Org. Chem. 1968, 33, 1436-1441; b) R. R. Hiatt, T. Mill, K.
C. Irwin, J. K. Castleman, J. Org. Chem. 1968, 33, 1428-1430; c) R. R. Hiatt, T. Mill,
K. C. Irwin, J. K. Castleman, J. Org. Chem. 1968, 33, 1421-1428.
25
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001503
TOC
N doped carbon tubes block the abstraction of α-H from EB molecule, thus ending the initial
stage of free radical propagation in EB oxidation process. By adding TBHP as a co-oxidant, it
will help to skip the inhibition step and generate peroxide free radicals.
26
This article is protected by copyright. All rights reserved.