Metal-free activation of dioxygen by graphene/g-C3N4 nanocomposites: Functional dyads for selective oxidation of saturated hydrocarbons

Article abstract of DOI:10.1021/ja200997a

Graphene sheet/polymeric carbon nitride nanocomposite (GSCN) functions as a metal-free catalyst to activate O₂ for the selective oxidation of secondary C-H bonds of cyclohexane. By fine-tuning the weight ratio of graphene and carbon nitride components, GSCN offers good conversion and high selectivity to corresponding ketones. Besides its high stability, this catalyst also exhibits high chemoselectivity for secondary C-H bonds of various saturated alkanes and, therefore, should be useful in overcoming challenges confronted by metal-mediated catalysis.

Full text of DOI:10.1021/ja200997a

COMMUNICATION
pubs.acs.org/JACS
Metal-Free Activation of Dioxygen by Graphene/g-C3N4
Nanocomposites: Functional Dyads for Selective Oxidation of
Saturated Hydrocarbons
Xin-Hao Li,*,† Jie-Sheng Chen,‡ Xinchen Wang,† Jianhua Sun,† and Markus Antonietti†
†Department of Colloid Chemistry, Max-Planck Institute of Colloids and Interfaces, Research Campus Golm, 14424 Postdam, Germany
‡School of Chemistry and Chemical Engineering, Shanghai Jiao Tong University, Shanghai 200240, P.R. China
S Supporting Information
b
Recently, we introduced g-C3N4 as a heterogeneous, metal-
ABSTRACT: Graphene sheet/polymeric carbon nitride
free organocatalyst for FriedelꢀCrafts reactions, CO2 fixation,
and artificial photosynthesis.4 As the most stable phase of C3N4
at ambient conditions, a polymeric, slightly distorted g-C3N4 can
nanocomposite (GSCN) functions as a metal-free catalyst
to activate O2 for the selective oxidation of secondary CꢀH
bonds of cyclohexane. By fine-tuning the weight ratio of
graphene and carbon nitride components, GSCN offers
good conversion and high selectivity to corresponding
ketones. Besides its high stability, this catalyst also exhibits
high chemoselectivity for secondary CꢀH bonds of various
saturated alkanes and, therefore, should be useful in over-
coming challenges confronted by metal-mediated catalysis.
be prepared via thermal polycondensation of common mono-
mers for the price of a typical mass polymer. Thus, this abundant,
stable, and metal-free polymeric solid could fulfill the require-
ments for use as an industrial catalyst. Both the composition and
the properties of g-C3N4 can be further engineered by introdu-
cing selected heteroatoms,5 which could slightly modify the
molecular orbital shape and position, such as the HOMO
relevant for oxidation, and therefore the oxidation strength and
the selectivity. Another possibility to do so is the one applied
here, which is by environment and by charge-transfer complexa-
tion to form a nanocomposite with an electron-rich system,6 thus
moving electron density within the catalyst.
atalytic oxidation of saturated CꢀH bonds of less expensive
C
alkanes to high-value-added chemicals remains a significant
task in many important current industrial and fine-chemical
processes.1 Among various catalytic systems, selective oxidation
using environmentally benign molecular oxygen is one of the
most problematic processes to control, because the abundance
and inertness of the CꢀH bonds in organic structures present
difficulties for selectivity, and the molecular oxygen often results
in overoxidation. In nature, enzymes use binding and proximity
effects to achieve astounding rate and selectivity enhancements
for specific reactions and substrates under mild conditions.2
An important new area of bioinorganic chemistry is devoted to
the simulation of the enzymatic functions with transition-metal
complexes as homogeneous catalysts, which could lead to
unprecedented activities and selectivities.3 Typically, the transi-
tion-metal centers in the enzymes play a key role in activating the
molecular oxygen, where sacrificial catalysts are usually involved;1ꢀ3
therefore, they are blue-prints for the design of biomimetic
oxidation catalysts.
Herein, we report the activation of molecular oxygen for the
selective oxidation of sp3 CꢀH bonds catalyzed by a graphene
sheet (GS)/polymeric carbon nitride (g-C3N4) nanocomposite
in a heterogeneous situation. A synergistic effect of GS and
g-C3N4 on promoting the catalytic activation of molecular
oxygen was observed, which is not found for either bare
g-C3N4 or GS alone. This system also showed high selectivity
toward ketones and high compatibility with various substrates.
Our study provides an ideal example of heterogeneous and
sustainable catalysts that can achieve highly selective, O2-based
oxidation of nonactivated secondary CꢀH bonds without the
employment of any metal components and co-reducing agents.7
Recently, we reported the oxidation of activated methyl groups
by B-doped g-C3N4, but the catalyst gave no conversion of
nonactivated CꢀH bonds.5e
Monolayer GSs were selected as the electron-donating modi-
fier for g-C3N4 due to their layered structure similar to g-C3N4
and their suitable electronic, mechanical, and chemical
properties.8 A recently described photochemical reduction tech-
nique is applied to generate stable aqueous colloids of clean GSs
from graphene oxide.9 The GSs exhibit monolayer structure
(Figure 1a) with a thickness around 0.9 nm (also see Supporting
Information, Figure S1). The homogeneous distribution of GSs
in both the precursor (cyanamide) and the bulk phase of g-C3N4
is key for preparing high-quality nanocomposites. Figure 1b
shows the zeta potentials of the GS solution in the presence of
Thus far, efficient O2-based catalytic oxidations of saturated
alkanes in the chemical industry are only compatible with the
production of terephthalic acid and cyclohexanone using metal-
complex catalysts.1 The corresponding processes mostly apply
homogeneous metal catalysts with the assistance of a large
amount of acid to offer only moderate selectivities.1aꢀd Recycl-
able catalysts that are stable but active are still highly desirable to
further decrease the economic and environmental costs of those
processes. The ideal sustainable catalytic oxidation process would
use molecular oxygen as the only oxygen source, a recyclable and
inexpensivecatalyst in nontoxic solvents with good conversion and
selectivity, and an inexpensive energy source.1a,3
Received: February 1, 2011
Published: May 11, 2011
r
2011 American Chemical Society
8074
dx.doi.org/10.1021/ja200997a J. Am. Chem. Soc. 2011, 133, 8074–8077
|

Journal of the American Chemical Society
COMMUNICATION
Table 1. Study of Reaction Conditionsa
entry
catalyst
g-C3N4
conv (%)b
OH (%)c
CdO (%)c
1
0
0
ꢀ
ꢀ
ꢀ
16
ꢀ
6
ꢀ
ꢀ
ꢀ
2
GS (20 mg)
Figure 1. (a) TEM image of GS. (b) Zeta potential of GS in a
cyanamide solution with different concentrations. (c) Photos of mix-
tures of GS and cyanamide in forms of colloidal solution (left), dry
powder (middle), and redispersed solution (right).
3
g-C3N4þGS
GSCN-2.5
0
4
5d
2
84
67
94
87
ꢀ
93
93
92
98
ꢀ
GSCN-12.5
11
12
3
6
GSCN-20
7
GSCN-50
13
ꢀ
7
8e
9f
10f
11f
12f
13g
GSCN-20 (Ar)
GSCN-20 (third)
GSCN-20 (fourth)
GSCN-20 (fifth)
GSCN-20 (sixth)
GSCN-20
0
13
12
15
14
0
7
8
2
ꢀ
a Typical conditions: 10 mL of CH3CN, 10 mmol of cyclohexane, 50 mg
of catalyst, O2 (10 bar, note that this pressure might lead to explosion
and therefore the catalytic reaction should be conducted in a high-
pressure reactor with an explosion-proof pressure sensor), 150 °C,
4 h. b Conversions were determined by GC-MS by using anisole as an
internal standard. c Selectivities of cyclohexanol (OH) and cyclohex-
anone (CdO). d Caprolactone (23%) was detected. e Reference experi-
ments in the absence of O2. f The GSCN-20 catalyst (50 mg) was reused
for 3ꢀ6 cycles. The catalyst was recycled via centrifugation, washed with
CH3CN, and dried under vacuum oven overnight. g 10 mol % of BHT
was added.
Figure 2. (a) UVꢀvis absorbance spectra of g-C3N4 and GSCN-x. (b)
XRDpatterns of g-C3N4, GSCN-50, andagroundmixture ofg-C3N4 (3 g)
and highly reduced r-GO (Gr, 50 mg, prepared by heating r-GO powder
at 873 K for 8 h in N2). (c) SEM image of GSCN-20.
different amounts of cyanamide, unambiguously revealing that
the introduction of cyanamide never disturbs the colloidal
stability of GS colloids. After removing the water from the dark
brown solution (Figure 1c, left), we obtained gray or black
powder (Figure 1c, middle), which could be redispersed into
water to form a homogeneous dispersion without visible aggre-
gation (Figure 1c, right). This phenomenon gives direct evidence
of the homogeneous distribution of r-GO in cyanamide solid.
Further heat treatment at 873 K could simultaneously trigger the
thermal condensation of cyanamide to polymeric g-C3N4 and
further the deep reduction of GS. For simplicity, the graphene/
g-C3N4 composites are denoted as GSCN-x, where x (2.5, 12.5,
25, or 50 mg) refers to the weighed-in amount of GS per batch.
The presence of GS in GSCN can be directly observed from
the color change from yellow (g-C3N4) to dark green. UVꢀvis
spectra (Figure 2a) indicate that the optical band gap and thereby
the semiconductor properties of GSCN are maintained with
gradually enhanced absorption over the whole range of visible
light with increasing amounts of GS included. According to the
elemental analysis results (Figure S2), the C/N molar ratios
gradually increase with increasing amount of GS, from 0.682 for
GSCN-2.5 to 0.691 for GSCN-50. Analysis of thermogravimetric
(TG-DTG) curves (Figure S3) has shown that the residual mass
of GSCN samples, as expected, depends on the amount of GS
added to the solution of cyanamide. The measurements suggest
that up to 5 wt % of GS is incorporated into the melon-based
carbon nitride structures.
introduction of GS does not disturb the layered structure of
polymeric g-C3N4. This can be further proved by the infrared
(FT-IR) spectrum of GSCN (Figure S5), which exhibits typical
CꢀN heterocycle stretches of the extended network connection
in the 1100ꢀ1600 cmꢀ1 region. The surface areas of the
modified g-C3N4 also remained nearly the same, even when
50 mg of GS was introduced (see Figure S6). The typical XRD
peak of aggregated GS, which appears in the XRD pattern of the
mechanical mixture of g-C3N4 and Gr (aggregates of highly
reduced GO, bar in Figure 2b), is not observed in the XRD
patterns of GSCN. All these observations further reveal that
monolayer GSs rather than stacks of homogeneous GSs are
distributed in the layered g-C3N4 host, and it is believed that the
intensely dispersed state is relevant for the electron exchange
between the components.
Our initial studies focused on the oxidation of cyclohexane
(CHA), a typical molecule that contains only secondary CꢀH
bonds. CHA is typically oxidized by oxygen gas in the presence of
cobalt naphthenate, offering only 4% conversion with 80%
selectivity for cyclohexanone (CHONE) and cyclohexanol
(CHOL).1c The oxidation product CHONE is the key raw
material in the synthesis of many useful chemical intermediates,
such as caprolactame for nylon 6 and adipic acid for nylon 66.1c,d
From Table 1, we can see that bare g-C3N4 or GS or the physical
mixture of them (g-C3N4þGS) gave no conversion of CHA.
Surprisingly, GSCN-x could favorably activate the molecular
oxygen for the oxidation of CHA under the same conditions, the
catalytic activity of which can be optimized by adjusting the
weight ratio of GS and g-C3N4 components. Low conversion
The SEM and TEM images of GSCN (e.g., Figures 2c and S4
for GSCN-20) clearly show the lamellar structure of the nano-
composite. The X-ray patterns (Figure 2b) and electron diffrac-
tion patterns (Figure S4) recorded for GSCN do not differ
significantly from those of the pure g-C3N4, indicating that the
8075
dx.doi.org/10.1021/ja200997a |J. Am. Chem. Soc. 2011, 133, 8074–8077

Journal of the American Chemical Society
COMMUNICATION
Table 2. Selective Oxidation of Alkanesa
Scheme 1. Schematic Representation of the Oxidation
Mechanism
high pressure of O2) are currently underway before these
materials can be commercially applied.
All these results indicate that GSCN-20 is an effective but mild
catalyst for hydrocarbon oxidation. Under the standard condi-
tions, GSCN-20 gave no conversion of n-hexane and DMSO.
Hexane has the same secondary CH2 groups, but this molecule
adsorbed much more weakly on the surface of GSCN. This
proves that preadsorption of the substrate is a key step. These
results revealed that GSCN-20 exhibits moderate to high che-
moselectivity for secondary CꢀH bonds (e.g., entries 2, 3, and 8,
Table 2), while substrate selectivity can be adjusted by adsorp-
tion equilibria and, therefore, should be useful in overcoming the
challenges confronted by metal-mediated catalysis.
a Typical conditions: 10 mL of CH3CN, 10 mmol of substrate, 50 mg of
GSCN-20, O2 10 bar, 150 °C, 4 h. b Temperature = 130 °C. c 0.1 g of
adamantane was used.
As shown in entry 13 of Table 1, GSCN-20 does not catalyze
oxidation of CHA in the presence of butylated hydroxytoluene
(BHT), a terminating agent that suppresses autoxidation (a free
radical chain process). This result strongly suggests that the
reaction in our catalytic system progresses via superoxide radical
anion (•O2ꢀ), as summarized in Scheme 1. Our previous results
also show that the excited electrons from the conduction band
(LUMO) of the C3N4 could reduce molecular oxygen to form
•O2ꢀ, which stays surface-bound to the C3N4 to compensate the
positive charge of the hole.4d At the same time, the substrates are
oxidized by the HOMO (positive hole) of C3N4 and then react
(2ꢀ3%) was achieved by GSCN-2.5 and GSCN-50 samples with
too little or too much GS. GSCN-12.5 provided a conversion
(11%) of CHA with a moderated selectivity (67%) to CHONE.
Notably, GSCN-20 gave a good conversion (12%) and a high
selectivity (94%) toward CHONE. No overoxidation bypro-
ducts, such as adipic acid or valeric acid, were detected in the GC-
MS analysis for all the reactions, demonstrating that the selective
CꢀH oxidation is provided from the HOMO of the catalytic
dyad of GS layers and g-C3N4 layers in a synergistic fashion.
To test the reusability of GSCN in the catalytic reaction, the
used GSCN catalyst was recovered from the reaction solution
via centrifugation or filtration. The GSCN-20, for instance,
was reused for more than six cycles without any obvious loss of
its catalytic activity in the catalytic oxidation of CHA (entries
5ꢀ8, Table 1); the conversion of CHA was 14% in the sixth
cycle with 98% selectivity to CHONE. Since GSCN-20 was the
best catalyst for activation of O2, it was used in subsequent
experiments.
In a further set of experiments, we examined the efficacy of
transformation of CꢀH to CdO under the same reaction
conditions for a series of substrates, which contain primary,
secondary, or tertiary CꢀH bonds, or two of them. GSCN-20
could promote the oxidation of various secondary CꢀH bonds
(entries 1ꢀ4, Table 2) with moderate to high conversions and
excellent selectivities toward the corresponding ketones. Tertiary
CꢀH bonds can also be oxidized to corresponding alcohols
(entry 5, Table 2). In principle, the conversions of these
substrates could be further improved by optimizing the catalytic
conditions, recyclizing of the reactants, or introducing cocata-
lysts, as currently only standard conditions for evaluating the
catalytic performance of GSCN-20 were used. Further improve-
ments in both conversions and critical reaction conditions (e.g.,
with the surface-bound O2ꢀ, ensuring the selectivity of the
•
catalytic reaction. As a semimetal with a band gap of about zero,
graphene has conduction and valence bands that meet at the
Dirac point,8 so a direct catalytic influence of GSs can be
excluded.8d A catalytic influence of the valence bond between
GS and C3N4 can also be excluded from the solid-state 13C NMR
analysis (Figure S7) of GSCN. The introduction of GS, however,
could obviously shift the HOMO of C3N4 to lower energies via,
e.g., πꢀπ* or charge-transfer interactions, which can be con-
firmed by the significant decrease in photoluminescence inten-
sity (Figure S8) as compared with that of g-C3N4. Those
interactions can be controlled by varying the weight ratio of
the components.6 We assume that this is why the weight ratio of
GS and g-C3N4 layers is important for the catalytic activity and
selectivity of the hybrid. The as-formed cyclohexyl hydroper-
oxide could either be directly decomposed or further initiate the
autoxidation of CHA (see Supporting Information) to CHOL
and CHONE in the presence of GSCN dyads. As expected, the
physical mixture, with weak interactions between g-C3N4 and
GS, offers low catalytic activity.
In summary, we have prepared a series of polymeric carbon
nitride composites, GSCN-x, using monolayer GS as a nanofiller.
8076
dx.doi.org/10.1021/ja200997a |J. Am. Chem. Soc. 2011, 133, 8074–8077

Journal of the American Chemical Society
COMMUNICATION
The catalytic reactions presented above clearly indicate that
GSCN-20 is a high-performance catalyst to activate molecular
oxygen for selective oxidation of secondary CꢀH bonds of
saturated alkanes with good conversion and high selectivity to
the corresponding ketones. The synergistic effect of GS and
carbon nitride on facilitating these transformations has been
unambiguously elucidated. Ease of mass production with high
stability, and good activity and selectivity by using molecule
oxygen without the assistance of any sacrificial cocatalyst are
merits stimulating a wider range of applications for industrially
important compounds and intermediates.
(7) (a) Adam, W.; Saha-M€oller, C. R.; Ganeshpure, P. A. Chem. Rev.
2001, 101, 3499. (b) Recupero, F.; Punta, C. Chem. Rev. 2007, 107, 3800
and references therein.
(8) (a) Geim, A. K.; Novoselov, K. S. Nat. Mater. 2007, 6, 183–191.
(b) Park, S.; Ruoff, R. S. Nat. Nanotechnol. 2009, 4, 217–224. (c) Allen,
M. J.; Tung, V. C.; Kaner, R. B. Chem. Rev. 2010, 110, 132–145. (d)
Dreyer, D. R.; Jia, H.; Bielawski, C. W. Angew. Chem., Int. Ed. 2010,
49, 6813.
(9) Li, X. H.; Chen, J. S.; Wang, X. C.; Schuster, M. E.; Schl€ogl, R.;
Antonietti, M., unpublished result.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental details, more
b
characterization results, and detailed discussions. This material
is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
Xin-Hao.Li@mpikg.mpg.de
’ ACKNOWLEDGMENT
This work was supported by ENERCHEM project of the
Max Planck Society, the AvH foundation, the NSFC (20731003),
and the National Basic Research Program (2007CB613303)
of China.
’ REFERENCES
(1) For examples, see: (a) Shilov, A. E.; Shul’pin, G. B. Chem. Rev.
1997, 97, 2879and references therein. (b) Labinger, J. A.; Bercaw, J. E.
Nature 2002, 417, 507. (c) Maschmeyer, T.; Oldroyd, R. D.; Sankar,
G.; Thomas, J. M.; Shannon, I. J.; Kleptko, J.; Masters, A. F.; Catlow,
C. R. Angew. Chem., Int. Ed. 1997, 36, 1639. (d) Kesavan, V.; Sivanand,
P. S.; Chandrasekaran, S.; Koltypin, Y.; Gedanken, A. Angew. Chem.,
Int. Ed. 1999, 38, 3521. (e) Chen, M. S.; White, M. C. Science 2010,
327, 566.
(2) (a) Guengerich, F. P. Chem. Res. Toxicol. 2001, 14, 611. (b) Nam,
W. Acc. Chem. Res. 2007, 40, 465and review articles in the special issue.
(3) For example, see: (a) Das, S.; Incarvito, C. D.; Crabtree, R. H.;
Brudvig, G. W. Science 2006, 312, 1941. (b) Lee, S. J.; Cho, S.-H.;
Mulfort, K. L.; Tiede, D. M.; Hupp, J. T.; Nguyen, S. T. J. Am. Chem. Soc.
2008, 130, 16828. (c) Meeuwissen, J.; Reek, J. N. H. Nat. Chem. 2010,
2, 615.
(4) (a) Goettmann, F.; Fischer, A.; Antonietti, M.; Thomas, A.
Angew. Chem., Int. Ed. 2006, 45, 4467. (b) Goettmann, F.; Thomas,
A.; Antonietti, M. Angew. Chem., Int. Ed. 2007, 46, 2717. (c) Wang, X. C.;
Maeda, K.; Thomas, A.; Takanabe, K.; Xin, G.; Carlsson, J. M.; Domen,
K.; Antonietti, M. Nat. Mater. 2009, 8, 76. (d) Su, F. Z.; Mathew, S. C.;
Lipner, G.; Fu, X. Z.; Antonietti, M.; Blechert, S.; Wang, X. C. J. Am.
Chem. Soc. 2010, 132, 16299.
(5) (a) Zhang, J. S.; Chen, X. F.; Takanabe, K.; Maeda, K.; Domen,
K.; Epping, J. D.; Fu, X. Z.; Antonietti, M.; Wang, X. C. Angew. Chem., Int.
Ed. 2010, 49, 441. (b) Zhang, Y. J.; Mori, T.; Ye, J. H.; Antonietti, M.
J. Am. Chem. Soc. 2010, 132, 6294. (c) Wang, Y.; Zhang, J. S.; Wang,
X. C.; Antonietti, M.; Li, H. R. Angew. Chem., Int. Ed. 2010, 49, 3356. (d)
Liu, G.; Niu, P.; Sun, C. H.; Smith, S. C.; Chen, Z. G.; Lu, G. Q.; Cheng,
H.-M. J. Am. Chem. Soc. 2010, 132, 11642. (e) Wang, Y.; Li, H. R.; Yao, J.;
Wang, X. C.; Antonietti, M. Chem. Sci. 2011, 2, 446.
(6) (a) Beenken, W. J. D. Chem. Phys. 2009, 357, 144. (b) Sun, Y.; Li,
C.; Xu, Y.; Bai, H.; Yao, Z.; Shi, G. Chem. Commun. 2010, 46, 4740.
8077
dx.doi.org/10.1021/ja200997a |J. Am. Chem. Soc. 2011, 133, 8074–8077