Tuning the selectivity in the aerobic oxidation of cumene catalyzed by nitrogen-doped carbon nanotubes

Article abstract of DOI:10.1002/cctc.201300909

In this study it is demonstrated that carbon nanotubes (CNTs) with doped nitrogen atoms in graphitic domains (NCNTs) can act as a new class of metal-free catalysts exhibiting excellent activity in the aerobic oxidation of cumene. We proved that NCNTs can promote the decomposition of hydroperoxide cumene with exceptionally high activity, resulting in strongly increased cumene conversion and extraordinarily high selectivity to acetophenone and 2-benzyl-2-propanol. The incorporation of nitrogen altered the surface electron structure of the CNTs and tuned the reactivity and selectivity. DFT calculations revealed that the remarkable improvement of catalytic performance of NCNTs is caused by the strong interaction between hydroperoxide cumene and the NCNTs. NCNTs also exhibited desirable recyclability after four cycling tests. This study not only provides a novel method for the cumene oxidation to high-value-added products at moderate reaction temperatures and oxygen atmospheric pressure, but also gives new insights into the effect of surface nitrogen doping on carbon-catalyzed liquid-phase oxidation of aromatic hydrocarbons.

Full text of DOI:10.1002/cctc.201300909

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201300909
Tuning the Selectivity in the Aerobic Oxidation of Cumene
Catalyzed by Nitrogen-Doped Carbon Nanotubes
Shixia Liao, Yumei Chi, Hao Yu, Hongjuan Wang, and Feng Peng*[a]
In this study it is demonstrated that carbon nanotubes (CNTs)
with doped nitrogen atoms in graphitic domains (NCNTs) can
act as a new class of metal-free catalysts exhibiting excellent
activity in the aerobic oxidation of cumene. We proved that
NCNTs can promote the decomposition of hydroperoxide
cumene with exceptionally high activity, resulting in strongly
increased cumene conversion and extraordinarily high selectiv-
ity to acetophenone and 2-benzyl-2-propanol. The incorpora-
tion of nitrogen altered the surface electron structure of the
CNTs and tuned the reactivity and selectivity. DFT calculations
revealed that the remarkable improvement of catalytic perfor-
mance of NCNTs is caused by the strong interaction between
hydroperoxide cumene and the NCNTs. NCNTs also exhibited
desirable recyclability after four cycling tests. This study not
only provides a novel method for the cumene oxidation to
high-value-added products at moderate reaction temperatures
and oxygen atmospheric pressure, but also gives new insights
into the effect of surface nitrogen doping on carbon-catalyzed
liquid-phase oxidation of aromatic hydrocarbons.
Introduction
As one of the most interesting metal-free catalysts, carbon ma-
terials have shown enormous promise in catalysis, owing to
their physicochemical and mechanical properties, such as high
surface areas, outstanding electron conductivity, corrosion re-
sistance, and thermal stability.[1] Especially, carbon nanotubes
(CNTs) have been widely explored as a new-generation catalyst
in the oxidative dehydrogenation of hydrocarbons to corre-
sponding olefins,[2] wet oxygen oxidation of phenolic water,[3]
the selective oxidation of aldehydes, alcohols, or hydrocar-
bons,[4] and so on. One of the most prominent and high-indus-
trial-relevance examples is the oxidative dehydrogenation of
ethylbenzene. A widely accepted reaction mechanism propos-
es that rich-in-electron diketone-like carbonyl groups on the
surface of carbon materials are the active sites.[2]
aerobic oxidation of hydrocarbon.[7] In our previous work, we
have proven that pristine CNTs can directly catalyze the aero-
bic oxidation of cyclohexane to adipic acid and their activities
can be considerably improved by N doping.[7a,d] Ma and co-
workers recently reported that N-doped sp2-hybridized layered
carbons exhibits remarkably improved catalytic activity for eth-
ylbenzene oxidation as well as a strongly increased acetophe-
none selectivity.[7b] However, there is a lack of clear knowledge
on the mechanism of carbon-based liquid-phase catalysis for
different reaction systems and the tuning mechanism of N
doping on reactivity, thus, further studies are required.
The selective liquid-phase aerobic oxidation of cumene to
hydroperoxide cumene (CHP) is extremely important in chemi-
cal industry, because CHP is an important intermediate in the
production of phenol and acetone.[8] Today, more than 90% of
the phenol in the world is produced by this route. Recently, it
was reported that N-hydroxyphthalimide–cobalt acetate homo-
geneous catalysts could oxidize cumene to acetophenone (AP)
and 2-benzyl-2-propanol (BP) with 80% selectivity and 35%
conversion,[9] which are highly valuable intermediates in the
manufacture of perfumes, pharmaceuticals, and resins.[10] Al-
though this catalyst system is expensive and complicated to
separate, it offers a possibility to change the existing compli-
cated and environment-unfriendly BP synthesis process.[10d,11]
The direct conversion of cumene to BP and AP is a very attrac-
tive route. To this end, it is of particular interest to know
whether it is possible to tune cumene oxidation by heteroge-
neous carbon catalyst to obtain high-value-added products at
moderate reaction temperatures and ambient pressure with
oxygen as the oxidant.
Some inorganic compounds and elements can be doped
into graphitic carbon materials and provide a convenient
means of fine-tuning the structure and reactivity. For example,
nonmetal heteroatom doping (boron, sulfur, phosphorous, and
nitrogen)[5] can change the band gap and thus influences elec-
tron (charge) mobility of carbons. In particular, substitutional N
doping of nanographitic carbons has received intensive atten-
tion, because its outstanding performance in electrocatalysis
has been theoretically predicted and experimentally ob-
served.[6] Recently, it was found that the N atoms introduced at
sp2-hybridized graphitic sites in the carbon framework could
effectively enhance the activity of nanocarbon materials in the
[a] S. Liao, Y. Chi, Dr. H. Yu, Dr. H. Wang, Prof. Dr. F. Peng
School of Chemistry and Chemical Engineering
South China University of Technology
Guangzhou, Guangdong, 510640 (China)
E-mail: cefpeng@scut.edu.cn
Herein, we demonstrate that CNTs with nitrogen atoms
doped into graphitic domains (NCNTs) can act as a new class
of metal-free catalyst that can afford excellent activity in the
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201300909.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 555 – 560 555

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
aerobic oxidation of cumene. NCNTs were proven to promote
the CHP decomposition with exceptionally high activity, result-
ing in a strongly increased cumene conversion and an extraor-
dinarily high selectivity to BP and AP. By correlation of the ac-
tivity with the surface structure of CNTs and further DFT calcu-
lations, a reasonable reaction network responsible for the
liquid-phase oxidation of cumene on CNTs was proposed. This
study not only provides a novel method for cumene oxidation
to high-value-added products at moderate reaction tempera-
tures and oxygen atmospheric pressure, but also gives new in-
sights into the effect of surface N doping on carbon-catalyzed
liquid-phase oxidation of aromatic hydrocarbons.
whereas CHP selectivity decreased rapidly from 90.0% to 3.3%,
and the distribution of products was changed completely. In
addition, with the reaction time increasing, the conversion of
cumene and total selectivity to AP and BP further increased.
NCNTs-3 catalyst afforded 86% conversion of cumene and
99% selectivity to AP and BP after 24 h. To our knowledge, it is
the first time that the cumene aerobic oxidation reaction can
be altered to one-step production of BP and AP giving the
total selectivity as high as 99% and a conversion higher than
80% under oxygen atmospheric pressure with the reaction
temperature as low as 353 K, in the presence of metal-free cat-
alyst.
To further discuss the relationship between catalytic perfor-
mance and structure of NCNTs, we correlated the reaction
rates with surface N loadings under the conversion of cumene
controlled at approximately 20%. The reaction rates were re-
spectively normalized by the BET surface area (Figure S3a) and
catalyst mass (Figure 1a). The obtained data did not display
a direct correlation between activity and surface area. Instead,
the activity appeared to be more in-line with the N content.
This phenomenon was also found in liquid-phase oxidation of
ethylbenzene to acetophenone.[7b] As revealed in Figure 1a,
the mass-normalized rate monotonously increased with total
nitrogen content increasing, however, the linear correlation of
activity with N contents was not possible because of the struc-
tural and morphological complexity of the synthesized NCNTs.
The initial reaction rate rapidly increased, and then slowly in-
creased after the N content was higher than 3.44%. For
NCNTs-3 catalyst, the initial reaction rates normalized by cata-
lyst mass and BET surface area were 10.4 and 3 times of those
of the undoped CNTs, respectively. Nitrogen doping dramati-
cally enhanced the catalytic activity of CNTs for cumene oxida-
tion. As expected, the value of the apparent activation energy
for oxidation of cumene was 12.8 kJmolꢀ1 using NCNTs-3 as
catalyst (see Figure S4), which is far lower than that for com-
mercial CNTs (50.4 kJmolꢀ1), demonstrating that the oxidation
of cumene is promoted by the N doping. The effects of differ-
ent N species including quaternary (NQ), pyridinic (NP) and pyr-
rolic (NPyr) nitrogen on the cata-
Results and Discussion
N doping is an effective way to improve catalytic activity of
carbon materials. To explore the special effect of N doping in
cumene oxidation system, four different NCNTs with nitrogen
contents in the range of 0.31–4.36 atom% were prepared by
the chemical vapor deposition method with changing precur-
sors and atmosphere. In Figure S1 (Supporting Information),
the typical bamboo-like structures of NCNTs prepared with Fe
catalyst are shown.[12] The concentration of N on the surface of
samples was measured by X-ray photoelectron spectroscopy
(XPS). As shown in Figure S2 and Table S1, the N/(N+C) atomic
ratios on the surfaces of catalysts varied from 0.31% to 4.36%
and different N functional groups were deconvoluted (see our
previous article[7d]).
The effect of N doping on the cumene oxidation reaction is
shown in Table 1. CNTs without N doping afforded 16.1%
cumene conversion with 90.0% selectivity to CHP. Notably, N
doping markedly promoted the activity, for example, NCNTs-3
exhibited the best catalytic performance, reaching 74.7% con-
version after 8 h of reaction, which is 4.6 times that of undop-
ed CNTs. This result strongly implies that the N doping indeed
contributes to the considerable improvement of activity. More
interestingly, with the content of nitrogen rising, the total se-
lectivity to AP and BP increased distinctly from 10% to 96.7%,
lytic activity were similar to that
of total N content (Figure 1b),
Table 1. Catalytic activity of different NCNTs for the oxidation of cumene.[a]
however, the essential effect of
Catalyst
N/(N+C) SBET
[atom%] [m2 gꢀ1
X[b]
[%]
Selectivity [%]
CHP BP AP
BP/AP[c] r[c]
r[c]
various nitrogen forms in NCNTs
on the catalytic activity remains
not clearly distinguishable. The
tuning role of nitrogen doping
on products selectivity of
cumene aerobic oxidation is also
clearly revealed in Figure 1c. The
BP/AP ratio rapidly decreased
first, and then remained con-
stant at 2.2 after the N content
was higher than 3.44%.
]
[mmolmꢀ2 hꢀ1
]
[mmolgꢀ1 hꢀ1
]
CNTs
0.00
0.31
2.21
3.44
4.36
3.44
3.44
20.9
35.1
48.9
72.6
52.4
72.6
–
–
72.6
–
16.1 90.0
30.7 71.0 23.9 5.1 5.0
53.9 15.7 57.2 27.1 2.6
8.4
1.6 5.3
0.673
0.874
2.378
2.026
14.1
30.6
116.2
147.0
NCNTs-1
NCNTs-2
NCNTs-3
NCNTs-4
NCNTs-3[d]
NCNTs-3[e]
74.7
72.1
86.4
73.9
64.9
0.1
3.1 56.1 40.8 2.2
3.3 56.2 40.5 2.2
1.0 55.8 43.2
3.3 57.0 39.7
7.4 54.0 38.6
3.004
157.5
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
FeNx/NCNTs-3 3.44
NCNTs-3[f]
blank[g]
3.44
–
–
–
–
0.4
5.6
–
0.4
0.5
2.7 99.2
3.2 93.9
blank[h]
–
[a] Reaction conditions: cumene (10 mL), catalyst (0.1 g), T=353 K, t=8 h, O2 flow rate=10 mLminꢀ1; conver-
sion and selectivity were determined by using GC and iodometry. [b] Conversion of cumene. [c] All conversions
were controlled to nearly 20%. [d] Reaction for 24 h. [e] Without HCl treatment. [f] 2 wt% p-benzoquinone was
added. [g] Without CHP and catalyst. [h] 2 wt% CHP was added without catalyst.
The possibility that residual
metal impurities are the active
sites should be considered. In
our work, all as-prepared
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 555 – 560 556

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 1. The dependences of mass-normalized reaction rates (r) on a) the total N content, b) different N groups, and c) the effect of the nitrogen content on
&
*
products distribution. NP =pyridinic group, NPyr =pyrrolic group, NQ =quaternary nitrogen. : Cumene reaction rate; ": CHP, N: BP, and : AP selectivity. Reac-
tion conditions: cumene 10 mL, catalyst 0.1 g, 353 K, O2 flow rate of 10 mLminꢀ1, the conversion and selectivity were determined by GC and iodometry. All
conversions were controlled at approximately 20%.
carbon-based catalysts were washed in concentrated hydro-
chloric acid to remove the metal. As shown in Table 1, for
NCNTs-3, no obvious changes in cumene conversion and prod-
ucts distribution were found compared with the NCNTs-3 un-
treated by HCl. In NH3 atmosphere and at high temperature,
metal nitrides could also be easily formed,[13] thus we prepared
an FeNx-loaded NCNTs-3 catalyst (Table 1). A slightly decreased
cumene conversion indicates that FeNx is not the active site,
implying that the residual metal impurities do not play an
active role in cumene oxidation under our reaction conditions.
It was widely accepted that cumene autoxidation undergoes
a radical-involved reaction mechanism.[14] To prove the domi-
nant role of radical species in our system, p-benzoquinone as
a typical radical scavenger and CHP as a free-radical initiator
were added to the reactant (see Table 1). After adding p-ben-
zoquinone, the reaction was nearly totally prevented (NCNTs-
3[f]), which indicates that cumene oxidation catalyzed by NCNTs
is also a radical-involved reaction. For blank tests with or with-
out addition of CHP, the conversion of cumene remained very
low, possibly because CHP is stable at the reaction tempera-
ture. It can be concluded that CNTs play an important role in
cumene oxidation.
The CHP decomposition process could effectively accelerate
the reaction rate of cumene and shorten the induction period.
To discover the role of CHP in our system, the effect of differ-
ent CNTs on the decomposition rate of CHP was investigated
as shown in Table 2. Compared with the blank, CNTs obviously
promoted the decomposition of CHP. Especially, for N-doped
CNTs, the conversion of CHP rapidly increased with the content
of nitrogen increasing. The decomposition products of CHP are
AP and BP. After adding p-benzoquinone, the decomposition
reaction of CHP was nearly totally prevented, indicating that
CHP decomposition catalyzed by NCNTs is also a radical-in-
volved reaction. In addition, the dependence of cumene con-
version on CHP decomposition rate was determined (Fig-
ure S5), demonstrating that the cumene conversion increases
with CHP decomposition rate increasing. Luz recently also re-
ported that the decomposition ability of hydroperoxide on dif-
ferent catalysts was responsible for their catalytic oxidation ac-
tivity of alkanes.[16] Hence, it was supposed that, if CHP decom-
C
C
position is accelerated with NCNTs, more RO and R free radi-
cals are generated, which will accelerate chain propagation re-
actions (discussed later in reaction mechanism).
Many researchers have discussed the detailed reaction pro-
cess of cumene oxidation.[14b,15] Although the reaction network
remains different in different catalysis systems, two approaches
Table 2. The performance of different CNTs for the decomposition of
CHP.[a]
C
to the generation of important cumenyl radical (R ) have been
Catalyst
N/(N+C)
[atom%]
t
[h]
Conversion
[%]
r
[mmolgꢀ1 hꢀ1
]
commonly accepted in the chain-initiation process, namely the
direct abstraction of H atom from cumene (denoted as RH)
–
–
4
4
4
1
0.4
0.4
1
5.5
13.4
52.3
80.7
97.5
97.8
0
1.6
3.9
CNTs
0.00
0.31
2.21
3.44
4.36
3.44
C
and the reaction of cumene with cumenyl oxygen radical (RO )
NCNTs-1
NCNTs-2
NCNTs-3
NCNTs-4
NCNTs-3[b]
25.0
92.8
280.3
281.1
0
generated by the decomposition of CHP (also denoted as
ROOH) as shown in Equations (1)–(3):
C
RH ! R
ð1Þ
ð2Þ
ð3Þ
[a] Reaction conditions: CHP (2 g), catalyst (0.1 g), acetonitrile as solvent
(8 mL), T=353 K. The conversion was determined by using iodometry.
The decomposition products were exclusively AP and BP. [b] 2 wt% p-
benzoquinone was added.
C
C
ROOH ! RO þ HO
C
C
RO þ RH ! R þ ROH
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 555 – 560 557

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
tion between CHP and CNTs is strengthened largely, which is
consistent with the better catalytic performance of NCNTs in
C
CHP decomposition. Similarly, the oxygenous free radicals (RO
C
and ROO ) are also more easily adsorbed on the NCNT, as
shown in Figure S6. The appropriate interaction would be ben-
eficial for the chain-propagation process and the formation of
AP and BP. This can well explain the role of N doping.
Based on above data, a possible reaction network is pro-
posed as shown in Scheme 1. 1) At the initial stage, cumene is
adsorbed on the surface of CNTs by the p–p conjugation effect
between the hexatomic ring of CNTs and the benzene ring of
C
cumene. As no initiator is added, R radical is formed through
H abstraction of cumene at a low rate of k1. Then the chain-
propagation process starts: the direct interaction between
C
oxygen and the formed R radical to produce the cumyl peroxy
C
radical (ROO ) (k2), followed by hydrogen abstraction from
C
cumene to produce CHP and R radical (k3). 2) Once CHP is pro-
duced, some of the CHP molecules are absorbed on the sur-
face of CNTs and the others are desorbed as final product (k6).
As CNTs can accelerate the electron-transfer-induced decompo-
sition of hydroperoxides,[19] the absorbed CHP molecules are
Figure 2. Optimized structures and adsorption energies of a–c) cumene and
d–f) CHP on the surface of a,d) CNT; b,e) N site of NCNT; and c,f) adjacent C
site of NCNT. *: C, *: O, *: N, *: H.
C
more inclined to decompose into RO (k4), which can easily ab-
C
stract a H atom of cumene to form R and BP (k5). The formed
C
Electron-donating nitrogen doped into in graphitic domains
was reported to facilitate the adsorption of reactive intermedi-
ates and thus improve the electrocatalysis of oxygen reduc-
tion.[17] To get more insights into the structure–property rela-
tionship, the first-principle DFT was used to calculate the ad-
sorption energies of reactant and radicals on CNTs and NCNTs.
The CNT model is in perfect arm-chair structure without de-
fects. NCNT represents the CNT with one carbon atom replaced
by quaternary N. As shown in Figure 2, the adsorption of reac-
tant and radicals on both CNT and NCNT is energetically favor-
able. This can be ascribed to the conjugation effect between
the benzene ring of the aromatic hydrocarbon and the gra-
phene p-system.[18] After N doping, the adsorption energies of
R continues to enter the chain-propagation process. At the
C
same time, AP is produced through the b-scission of RO (k7).
For pristine CNTs, the formed CHP is more inclined to
desorb from the surface of tubes, and only a small proportion
is adsorbed and further decomposed for the new cycle. As an
incorporated N atom has one more electron than carbon,
which can increase the charge density in addition to the conju-
gated p-bond system,[6a,20] the formed CHP is easily adsorbed
on the surface of NCNTs, which is demonstrated by the lower
adsorption energy. At the same time, the electron-transfer-in-
duced decomposition reaction of CHP is improved dramatically
C
with the enhancement of charge density. Therefore, more RO
C
and R free radicals are generated, leading to a higher reaction
C
C
cumene and R are evidently unchanged. However, the adsorp-
rate and BP concentration. With the accumulation of RO and
tion energy of CHP prominently declines from ꢀ1.73 kJmolꢀ1
possible enhanced interaction between RO and NCNTs, more
C
to ꢀ12.67 kJmolꢀ1 after N doping, indicating that the interac-
AP is also produced. The concentration of CHP remains at
a low level owing to its rapid decomposition. These
results well explain the differences of catalytic activity
and products distribution between CNTs and NCNTs.
In addition to the remarkable catalytic perfor-
mance, the good stability of CNTs as catalysts is also
an advantage (Figure 3). Taking the NCNTs-3 catalyst,
for example, after four cycles, no obvious difference
in both cumene conversion and product distributions
was found. The ID/IG ratio in Raman spectroscopy
(Figure S7) and the morphology (Figure 4) of the cat-
alyst were obviously unchanged after recycling.
These results strongly indicate that NCNTs are very
active and stable catalysts for the oxidation of
cumene. Therefore, it is promising to produce AP
and BP with NCNTs catalysts through cumene liquid
aerobic oxidation in industry.
Scheme 1. Reaction network of cumene with CNTs as catalysts.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 555 – 560 558

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Experimental Section
Preparation
The NCNTs were synthesized by a chemical vapor deposition
method in a horizontal tubular quartz furnace with 4 cm inner di-
ameter over FeMo/Al2O3 catalyst. The details of the FeMo/Al2O3 cat-
alyst can be found in Ref. [21]. To obtain NCNTs with different N
contents, 10 mL mixtures of xylene and aniline with different ani-
line volume percentage were injected by a syringe pump at a rate
of 3 mLhꢀ1. The liquid mixtures were vaporized in the quartz tube
at approximately 353 K. The growth of NCNTs was performed at
1073 K in Ar or NH3 at 500 cm3 minꢀ1 [8d]
NCNT-1 and NCNT-2 were
.
denoted as N-doped carbon nanotubes prepared in Ar with 10%
and 100% (v/v) of aniline in the precursors, respectively. NCNTs-3
and NCNTs-4 were denoted as N-doped carbon nanotubes pre-
pared in NH3 with 0% and 100% (v/v) of aniline in precursors, re-
spectively. As a comparison, we also prepared nitrogen-free carbon
nanotubes under the same conditions with xylene as precursor in
Ar, denoted as CNTs. The residual FeMo/Al2O3 catalyst in all ob-
tained CNTs was removed by washing with 12 molLꢀ1 HCl aqueous
solution before catalytic tests and characterizations. The commer-
cial CNTs were purchased from Shenzhen Nanotech Port Co. Ltd
(NTP) and also purified with above method.
Figure 3. Reuse of NCNTs-3 catalyst. Reaction conditions: cumene 10 mL, O2
flow rate 10 mLminꢀ1, 353 K, 8 h. &: Cumene conversion; &: BP, &: AP, and
&: CHP selectivity.
Characterizations
BET specific surface areas were measured by N2 adsorption at
liquid N2 temperature in an ASAP 2010 analyzer. Raman spectra
were obtained in a LabRAM Aramis micro Raman spectrometer
with an excitation wavelength at 532 nm with 2 mm spot size. TEM
images were obtained with a FEI Tecnai G2 12 microscope operat-
ed at 100 kV and a JEOL JEM2010 microscope operated at 200 kV.
Specimens for TEM and HRTEM were prepared by ultrasonically
suspending the sample in acetone and depositing a drop of the
suspension onto a grid. XPS analysis was performed in a Kratos
Axis ultra (DLD) spectrometer equipped with an AlKa X-ray source
in ultrahigh vacuum (<133ꢁ10ꢀ10 Pa). The binding energies (ꢁ
0.2 eV) were referenced to the C1s peak at 284.6 eV. The surfaces
of samples were cleaned by heat treatment at 373 K in ultrahigh
vacuum prior to the measurements.
Figure 4. TEM images of NCNTs-3 a) before reaction and b) after the 4th cy-
cle.
Conclusions
We have demonstrated that carbon nanotubes with the nitro-
gen atoms doped into graphitic domains as a new class of
metal-free catalyst can afford an excellent activity in the aero-
bic oxidation of cumene. We proved that nitrogen-doped
carbon nanotubes can promote the hydroperoxide cumene
decomposition with exceptionally high activity, resulting in
a highly increased cumene conversion and an extraordinarily
high selectivity to 2-benzyl-2-propanol and acetophenone. The
incorporation of nitrogen altered the surface electron structure
of the carbon nanotubes and tuned the reactivity and selectivi-
ty. DFT calculations revealed that the remarkable improvement
of catalytic performance of nitrogen-doped carbon nanotubes
is caused by the strong interaction between hydroperoxide
cumene and carbon nanotubes. Based on these results, a possi-
ble reaction network was given. In addition, nitrogen-doped
carbon nanotubes showed good recyclability in cumene oxida-
tion. This study not only provides a novel method for cumene
oxidation to high-value-added products at a moderate reaction
temperature and oxygen atmospheric pressure, but also gives
new insight into the effect of surface nitrogen doping on
carbon-catalyzed liquid-phase oxidation reaction of aromatic
hydrocarbons.
Calculations
All results were calculated by using Cambridge Sequential Total
Energy Package codes, implemented in Materials Studio environ-
ment.[22] Exchange–correlation function was the popular Perdew–
Burke–Ernzerhof generalized gradient approximation.[23] The inter-
actions were described by norm-conserving pseudopotentials. Lat-
tice constants of the cell were 18, 18, and 14 ꢂ, respectively. A
plane wave basis set with a cutoff energy of 300 eV was used. K-
point sampling was restricted to a single point. Models used in this
work were single-walled arm-chair nanotubes with 96 atoms. The
adsorption energy (Eads) was defined as the total energy gained by
molecule adsorption at equilibrium distance, as shown in Equa-
tion (4):[24]
Eads ¼ EtubeþmoleculeꢀEtubeꢀEmolecule
ð4Þ
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 555 – 560 559

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[5] a) V. Strelko, V. Kuts, P. Thrower, Carbon 2000, 38, 1499–1503; b) C. H.
Choi, S. H. Park, S. I. Woo, Green Chem. 2011, 13, 406–412.
Catalytic tests
The liquid oxidation reactions were performed in a three-necked
flask (20 mL), supplied with a magnetic stirrer, reflux condenser,
and an oil bath. Cumene (10 mL) and catalyst (100 mg) were put
into the flask, sonicated for 5 min, and heated to the preconcerted
temperature. Then oxygen was passed through the solution by
bubbling at the temperature until the reaction was over. The CHP
concentration was determined according to the iodometric
method.[25] After the reduction of generated CHP to BP through
the triphenylphosphine reaction,[26] the other products in the liquid
phase were detected by gas chromatography (Agilent GC-6820)
equipped with a 30 mꢁ0.25 mmꢁ0.25 mm HP-5 capillary column
and a flame ionization detector (detector temperature 553 K, injec-
tor temperature 553 K, and oven temperature 413 K) using toluene
as external standard.
[6] a) S. Maldonado, S. Morin, K. J. Stevenson, Carbon 2006, 44, 1429–1437;
b) S. Kundu, T. C. Nagaiah, W. Xia, Y. Wang, S. V. Dommele, J. H. Bitter, M.
Santa, G. Grundmeier, M. Bron, W. Schuhmann, J. Phys. Chem. C 2009,
113, 14302–14310; c) R. Li, W. Ji, J. Lin, Phys. Rev. B 2007, 76, 195406;
d) X. Hu, Y. Wu, H. Li, Z. Zhang, J. Phys. Chem. C 2010, 114, 9603–9607.
[7] a) H. Yu, F. Peng, J. Tan, X. Hu, H. Wang, J. Yang, W. Zheng, Angew.
Chem. 2011, 123, 4064–4068; Angew. Chem. Int. Ed. 2011, 50, 3978–
3982; b) Y. Gao, G. Hu, J. Zhong, Z. Shi, Y. Zhu, D. S. Su, J. Wang, X. Bao,
D. Ma, Angew. Chem. 2013, 125, 2163–2167; Angew. Chem. Int. Ed.
2013, 52, 2109–2113; c) J. Long, X. Xie, J. Xu, Q. Gu, L. Chen, X. Wang,
ACS Catal. 2012, 2, 622–631; d) Y. Cao, H. Yu, J. Tan, F. Peng, H. Wang, J.
Li, W. Zheng, N.-B. Wong, Carbon 2013, 57, 433–442; e) J. Luo, F. Peng,
H. Wang, H. Yu, Catal. Commun. 2013, 39, 44–49.
[8] S. Matsui, T. Fujita, Catal. Today 2001, 71, 145–152.
[9] K. Novikova, M. Kompanets, O. Kushch, S. Kobzev, M. Khliestov, I.
Opeida, React. Kinet. Mech. Catal. 2011, 103, 31–40.
[10] a) S. Devika, M. Palanichamy, V. Murugesan, Appl. Catal. A 2011, 407,
76–84; b) G. Yang, Y. Ma, J. Xu, J. Am. Chem. Soc. 2004, 126, 10542–
10543; c) J.-Y. Qi, H.-X. Ma, X.-J. Li, Z.-Y. Zhou, M. C. Choi, A. S. Chan, Q.-
Y. Yang, Chem. Commun. 2003, 1294–1295; d) H.-Y. Hou, T.-S. Liao, Y.-S.
Duh, C.-M. Shu, J. Therm. Anal. Calorim. 2006, 83, 167–171.
[11] C. E. Frank, W. E. Foster, Ind. Eng. Chem. 1954, 46, 1019–1021.
[12] S. Van Dommele, A. Romero-Izquirdo, R. Brydson, K. De Jong, J. Bitter,
Carbon 2008, 46, 138–148.
The thermolysis tests of CHP were conducted according to refer-
ence.[19] Acetonitrile (8 mL) and catalyst (0.1 g) were added in
a flask, sonicated for 5 min, and then placed in an oil bath at
353 K. After flushing with N2 for 5 min, CHP (2 g) was added and
the reaction was conducted in a closed system. The unreacted CHP
was tested by the aforementioned iodometric method.
[13] S. Kurian, N. Gajbhiye, Mater. Lett. 2011, 65, 3089–3091.
[14] a) K. Hattori, Y. Tanaka, H. Suzuki, T. Ikawa, H. Kubota, J. Chem. Eng. Jpn.
1970, 3, 72–78; b) A. Bhattacharya, Chem. Eng. J. 2008, 137, 308–319.
[15] Y. F. Hsu, C. P. Cheng, J. Mol. Catal. A 1998, 136, 1–11.
[16] I. Luz, A. Leꢃn, M. Boronat, F. X. Llabrꢄs i Xamena, A. Corma, Catal. Sci.
Technol. 2013, 3, 371–379.
[17] a) R. Liu, D. Wu, X. Feng, K. Mꢅllen, Angew. Chem. 2010, 122, 2619–
2623; b) K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 2009, 323,
760–764.
[18] W. Chen, L. Duan, D. Zhu, Environ. Sci. Technol. 2007, 41, 8295–8300.
[19] P. S. Engel, W. E. Billups, D. W. Abmayr, K. Tsvaygboym, R. Wang, J. Phys.
Chem. C 2008, 112, 695–700.
[20] C. V. Rao, C. R. Cabrera, Y. Ishikawa, J. Phys. Chem. Lett. 2010, 1, 2622–
2627.
[21] W. Qian, H. Yu, F. Wei, Q. Zhang, Z. Wang, Carbon 2002, 40, 2968–2970.
[22] M. Segall, P. J. Lindan, M. Probert, C. Pickard, P. Hasnip, S. Clark, M.
Payne, J. Phys. Condens. Matter 2002, 14, 2717.
[23] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.
[24] J. Zhao, J. P. Lu, J. Han, C.-K. Yang, Appl. Phys. Lett. 2003, 82, 3746.
[25] J. Zawadiak, D. Gilner, Z. Kulicki, S. Baj, Analyst 1993, 118, 1081–1083.
[26] V. Y. Kugel, M. Tsodikov, G. Bondarenko, Y. V. Slivinskii, D. Kochubey, M.
Hidalgo, J. Navio, Langmuir 1999, 15, 463–468.
Acknowledgements
This work was supported by the National Natural Science Foun-
dation of China (No. 21133010, 21273079), the Guangdong Pro-
vincial Natural Science Foundation (No. S20120011275) and Pro-
gram for New Century Excellent Talents in Universities of China
(NCET-12-0190).
Keywords: carbon · doping · oxidation · nanotubes · nitrogen
[1] a) J. Planeix, N. Coustel, B. Coq, V. Brotons, P. Kumbhar, R. Dutartre, P.
Geneste, P. Bernier, P. Ajayan, J. Am. Chem. Soc. 1994, 116, 7935–7936;
b) P. Serp, E. Castillejos, ChemCatChem 2010, 2, 41–47; c) D. S. Su, J.
Zhang, B. Frank, A. Thomas, X. Wang, J. Paraknowitsch, R. Schlçgl,
ChemSusChem 2010, 3, 169–180.
[2] a) D. Su, N. I. Maksimova, G. Mestl, V. L. Kuznetsov, V. Keller, R. Schlçgl,
N. Keller, Carbon 2007, 45, 2145–2151; b) X. Liu, D. S. Su, R. Schlçgl,
Carbon 2008, 46, 547–549; c) J. Zhang, X. Liu, R. Blume, A. Zhang, R.
Schlçgl, D. S. Su, Science 2008, 322, 73–77.
[3] S. Yang, X. Li, W. Zhu, J. Wang, C. Descorme, Carbon 2008, 46, 445–452.
[4] a) B. Frank, R. Blume, A. Rinaldi, A. Trunschke, R. Schlçgl, Angew. Chem.
2011, 123, 10408–10413; Angew. Chem. Int. Ed. 2011, 50, 10226–10230;
b) J. Luo, F. Peng, H. Yu, H. Wang, Chem. Eng. J. 2012, 204–206, 98–
106; c) J. Luo, F. Peng, H. Yu, H. Wang, W. Zheng, ChemCatChem 2013,
5, 1578–1586.
Received: October 23, 2013
Revised: November 26, 2013
Published online on January 17, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 555 – 560 560