A novel catalyst Pd@ompg-C3N4 for highly chemoselective hydrogenation of quinoline under mild conditions

Article abstract of DOI:10.1016/j.jcat.2012.10.018

Polymeric mesoporous carbon graphitic nitrides (mpg-C₃N ₄) and ordered mesoporous graphitic carbon nitrides (ompg-C ₃N₄) with different surface area and morphology were used to prepare palladium catalysts (Pd@C₃N₄) by an easy ultrasonic-assisted method. These catalysts demonstrated excellent activity and selectivity for hydrogenation of quinoline to 1,2,3,4-tetrahydroquinoline under mild temperature (30-50 °C) and H₂ pressure (1 bar). Pd@ompg-C₃N₄(r = 2.5) showed the best catalytic performance and both the activity and selectivity could be maintained for at least six reaction runs. The introduction of ordered cylindrical mesoporous structure and high concentration of surface Pd⁰ (about 70%) contribute to the high reaction activity and selectivity over Pd@ompg-C ₃N₄ catalysts.

Full text of DOI:10.1016/j.jcat.2012.10.018

Journal of Catalysis 297 (2013) 272–280
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
A novel catalyst Pd@ompg-C3N4 for highly chemoselective hydrogenation
of quinoline under mild conditions
⇑
Yutong Gong, Pengfei Zhang, Xuan Xu, Yi Li, Haoran Li, Yong Wang
Key Lab of Applied Chemistry of Zhejiang Province, Department of Chemistry, Zhejiang University, Hangzhou 310028, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Polymeric mesoporous carbon graphitic nitrides (mpg-C3N4) and ordered mesoporous graphitic carbon
nitrides (ompg-C3N4) with different surface area and morphology were used to prepare palladium cata-
lysts (Pd@C3N4) by an easy ultrasonic-assisted method. These catalysts demonstrated excellent activity
and selectivity for hydrogenation of quinoline to 1,2,3,4-tetrahydroquinoline under mild temperature
(30–50 °C) and H2 pressure (1 bar). Pd@ompg-C3N4(r = 2.5) showed the best catalytic performance and
both the activity and selectivity could be maintained for at least six reaction runs. The introduction of
ordered cylindrical mesoporous structure and high concentration of surface Pd0 (about 70%) contribute
to the high reaction activity and selectivity over Pd@ompg-C3N4 catalysts.
Received 10 June 2012
Revised 16 October 2012
Accepted 17 October 2012
Available online 22 November 2012
Keywords:
Carbon nitride
Quinoline
Chemoselective hydrogenation
Heterogeneous catalyst
Mild condition
Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction
long reaction time. For example, homogeneous Ir-based catalysts
for the room temperature hydrogenation of quinolines typically re-
The synthesis of nitrogen heterocycles, including 1,2,3,4-
tetrahydroquinoline (THQ), is of great current interest. THQ is an
important intermediate for the synthesis of drugs, agrochemicals,
dyes, alkaloids, and many other biological active molecules [1,2].
During recent decades, some catalytic methods have been estab-
lished for the synthesis of THQ, such as the catalytic cyclization
method [3,4], Beckman rearrangement [5], and the hydrogenation
of quinoline (Scheme 1). Among these processes, direct selective
hydrogenation of quinoline is widely regarded as the most conve-
nient and most promising one owing to its high atom utilization
together with easy access of the raw material.
Many homogeneous or heterogeneous metal catalysts can re-
duce isolated C@C, C@O and C@N double bonds in olefins, carbon-
yls, and imines, but hydrogenation of nitrogen-containing
aromatics is more challenging [6,7]. Successful examples of the
hydrogenation of quinoline by H2 with homogeneous metal cata-
lysts were firstly reported by Fish and coworkers in 1982 [8]. After
that, homogeneous N-heterocycle hydrogenation, including quino-
line hydrogenation, has proved possible with many Rh [9,10], Ru
[11,12], Ir [13,14], and Os [15] catalysts precursors. Unfortunately,
in most cases, the homogeneous catalysts displayed low activity
under mild reaction conditions. Almost all quinoline hydrogena-
tion catalysts require high pressures, elevated temperatures, or
quire an excess of 15 bar of H2 for high conversion, while catalysts
that operate at lower hydrogen pressures require elevated temper-
atures (ꢀ100 °C or higher) [16,17]. Very recently, Crabtree and
coworkers reported a new homogeneous Ir catalyst, which cata-
lyzed hydrogenation of quinolines under mild conditions – as
low as 1 bar of H2 and 25 °C, however, after a long reaction time
of 18 h [7]. Many homogeneous systems require the presence of
harsh additives such as I2, acid, or PPh3 for good conversion or
selectivity. In addition, the use of homogeneous metal catalysts
on an industrial scale is limited by practical problems due to the
difficulties in recovering the expensive catalyst metals, ligands,
and additives from the reaction mixtures. In this context, the
development of new and improved methods for the efficient and
selective production of THQ continues to be a challenging goal.
Heterogeneous catalysts especially supported precious metal cata-
lysts have been drawing more and more attention for their supe-
rior reusability together with their outstanding catalytic
performance [18]. However, only a few heterogeneous catalysts
are available for the hydrogenation of quinoline to date, for exam-
ple, Ru on poly(4-vinylpyridine) [19], Pd, Ru or Rh on Al2O3 [20], Ru
on polyorganophosphazenes [21], Rh on montmorillonite [22], Pd
on hydroxyapatites [23], and Pd on tannin-grafted collagen fibers
[24]. Unfortunately, the complete conversion of quinoline to THQ
with Ru-, Pd-, or Rh-based heterogeneous catalysts requires drastic
reaction conditions, typically a temperature in the range of
60–200 °C, hydrogen pressure between 2.0 MPa and 4.0 MPa. As
⇑
Corresponding author. Fax: +86 571 87951895.
E-mail address: chemwy@zju.edu.cn (Y. Wang).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.10.018

Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
273
mixture was filtered, washed by deionized water for several times,
and dried in the air at room temperature. Finally, the dried white
powder was calcined at 550 °C for 8 h. The above hydrothermally
prepared SBA-15 was then used as a template for the preparation
of ompg-C3N4 [35,36]. 8 g of cyanamide, 2.5 (r = 2.5, r = weight of
the used SBA-15 per 8 g cyanamide) g of SBA-15, and 20 g of deion-
ized water were mixed together at room temperature and stirred
for 1–2 h. Subsequently, the dispersion was filtrated and dried at
60 °C overnight until removal of water and formation of a white so-
lid. The powder was then grounded in a mortar, transferred into a
crucible, heated under air at 2.3 °C minÀ1 (4 h) up to 550 °C, and
then treated at 550 °C for 4 h. The as-obtained yellow powder
was ground in a mortar and then treated under stirring during
48 h in a 4 M NH4HF2 solution. The dispersion was then filtered,
and then, the precipitate was washed with distilled water and eth-
anol. After filtering, the yellow compound was dried under vacuum
at 100 °C overnight.
NH2
Cyclization
(1)
(2)
Cl
N
H
Beckman Rearrangement
N
H
N
OMs
Quinoline Hydrogenation
(3)
N
N
H
Scheme 1. Synthetic routes to 1,2,3,4-tetrahydroquinoline.
An ultrasonic-assisted method was applied to deposit Pd on mpg-
C3N4 and ompg-C3N4. Typically, 0.5 g ompg-C3N4(r = 2.5) was sus-
pended in 50 ml deionized water, and the mixture was immersed
into an ultrasonic instrument filled with water until ompg-C3N4
was dispersed to a uniform suspension. After that, 10 ml of PdCl2
aqueous solution (0.1 g/ml) was added into the ompg-C3N4 suspen-
sion allowing the chelate adsorption of Pd2+ on ompg-C3N4. 25 ml
NaBH4 aqueous solution (2 mg/ml) was added dropwise afterward,
which would result in a grayish brown suspension. The slurry at last
was filtered, washed with deionized water for several times, and
dried in the air at 70 °C overnight. The resulting Pd@ompg-C3N4
was then characterized by ICP, XRD, TEM, XPS, and BET, respectively.
a consequence, a highly effective heterogeneous catalyst under
mild conditions is still desirable.
Carbon nitrides are fascinating materials that have attracted
worldwide attention. They provide access to an even wider range
of applications than carbon materials because the incorporation
of nitrogen atoms in the carbon architecture can enhance the
chemical, electrical, and functional properties. For example, our
group and others have shown that a polymeric graphitic carbon ni-
tride (g-C3N4), the most stable allotrope in the air, exhibits extreme
chemical and thermal stability, can be chemically shaped to a vari-
ety of nanostructures, and can be directly used in heterogeneous
catalysis [25–28], for example in oxidation of hydrocarbons [28–
32]. However, few excellent applications as catalyst supports have
emerged so far. We have recently disclosed a new strategy for the
design of high-performance heterogeneous catalysts utilizing the
mpg-C3N4 as catalyst support [18]. Highly dispersed palladium
nanoparticles were introduced as a functional moiety into the
mpg-C3N4 framework. The hybrid material, Pd@mpg-C3N4, exhib-
ited promising catalytic performance for the selective hydrogena-
tion of phenol. Considering the influence of diffusion steps in
heterogeneous catalysis, metal nanoparticles loaded on ordered
support bearing cylindrical pores would probably reveal better
performance since substrates could reach the active sites more
smoothly than on the support with random connected spherical
pores. As part of our ongoing effort to develop new strategies for
chemoselective hydrogenation, herein, we present the synthesis
and characterization of the palladium-grafted ordered mesoporous
carbon nitride (Pd@ompg-C3N4) and employ it as an effective het-
erogeneous catalyst for the selective hydrogenation of quinoline by
use of hydrogen under mild reaction conditions. We also mention
the interesting aspects such as the pore structure effect and the
substrate scope as well as the recycling of the catalyst.
2.2. Characterization analyses
Elemental analysis was performed by Elementar Vario MACRO.
The Pd content was measured by ICP-AES (IRIS Intrepid II XSP,
Thermo Fisher Scientific, USA). The surface areas of all the supports
and catalyst were determined by AUTOSORB-1 instrument. BET
equation was used to calculate the surface area and pore volume.
Samples were outgassed at 100 °C for 20 h until the residual pres-
sure was less than 10À4 Pa. The diffraction data were collected at
room temperature with 2h scan range between 5° and 90° using
a wide-angle X-ray diffraction (Model D/tex-Ultima TV, 1.6 kV, Rig-
aku, Japan) equipped with Cu Ka radiation (1.54 Å). The X-ray pho-
toelectron spectra were obtained with an ESCALAB MARK II
spherical analyzer using a magnesium anode (Mg 1253.6 eV) X-
ray source. The powder samples were pressed to pellets and fixed
to a stainless steel sample holder without further treatment. The
XPS spectrum was shifted according to C1s peak being at
288.2 eV so as to correct the charging effect [18]. TEM (Model
JEM-1230, JEOL Co. Ltd., Japan) characterization was operated at
an accelerating voltage of 80 kV. FTIR was measured by Bruker
Verctor 22. Benzene-TPD and pyridine-TPD were conducted as fol-
lows: 0.1 g of dry ompg-C3N4(r = 2.5) was mixed with quartz sand
and placed into a quartz tube located inside an electrical furnace.
Then, benzene or pyridine was brought into the tube by He for
1 h. The samples were blown by pure He for about 2 h until no
change of TCD signal next. At last, the electrical furnace was sub-
jected to a 10 °C/min heating rate up to 400 °C, under a He flow
of 40 cm3/min. The desorption of benzene and pyridine was mon-
itored by a thermal conductivity detector.
2. Experimental
2.1. Materials and catalysts preparation
Ludox HS-40 silicon dispersion and SBA-15 were used as tem-
plates for the preparation of mpg-C3N4 and ompg-C3N4, respec-
tively. Ludox HS-40 with an average diameter of 12 nm was
purchased from Sigma–Aldrich and used as received without any
further treatment. Mpg-C3N4(r = 1, r = MLudoxHS-40/Mcyanamide) was
synthesized according to previous literature [33]. SBA-15 was syn-
thesized according to a literature procedure [34]. Briefly, 8 g Plu-
ronic P123 was dissolved in 240 ml 2 M HCl solution, and 17 g
tetraethyl orthosilicate (TEOS) was added, and then the mixture
was stirred at 35 °C for 20 h. The suspension was heated at
150 °C in a hydrothermal reactor for 48 h. After that, the resulting
2.3. Catalytic procedure and recycling
The hydrogenation of quinoline using hydrogen was carried out
in a batch-type reactor operated under atmospheric conditions.
Experiments were conducted using a three necked glass flask
(capacity 25 ml) precharged with quinoline, solvent, and catalyst.

274
Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
Table 1
The mixture was stirred using a magnetic stirrer and heated in a
water bath. The system was equipped with a thermocouple to con-
trol the temperature and a reflux condenser. The reactor was
sealed and purged with H2 to remove the air for three times and
then equipped with a hydrogen balloon. Samples were taken
between certain time intervals and analyzed using a Shimadzu
GC-2014 gas chromatograph equipped with a Rtx-1071 column.
Decane was used as an internal standard to calculate quinoline
conversion and THQ selectivity. The GC conditions for the product
analysis were as follows: Injector Port Temperature: 260 °C; Col-
umn Temperature: Initial temperature: 60 °C (1 min); Gradient
Rate: 50 °C/min (3.8 min); Final Temperature: 250 °C (10 min).
The catalyst used above was recovered by centrifugation, washed
three times with ethanol, and dried at 80 °C for 6 h. The slight loss
of catalyst was replenished by new catalyst without changes of
other conditions for the next cycle.
Textural properties of mpg-C3N4, ompg-C3N4, active carbon and Pd@ompg-C3N4.
a
Entry
Support or catalyst
SBET (m2/g)
Pvb (cm3/g)
Psc (nm)
1
2
3
4
5
6
mpg-C3N4(r = 1)
148
142
198
212
899
197
0.52
0.28
0.44
0.48
0.64
0.32
9.4
ompg-C3N4(r = 1)
ompg-C3N4(r = 1.5)
ompg-C3N4(r = 2.5)
active carbon
6.07
7.17
7.42
4.58
5.68
Pd@ ompg-C3N4(r = 2.5)
All were determined from the desorption branch of the nitrogen.
a
BET surface area.
Pore volume.
Pore size.
b
c
in Fig. 1. N2 adsorption–desorption isotherm curves of these
samples show a pronounced hysteresis, typical of the existence
of mesopore structures, only weak microporosity. The Brunauer–
Emmett–Teller (BET) surface area, BJH pore sizes, and pore
volumes are summarized in Table 1. The BET surface area (as deter-
mined from the desorption branch of the nitrogen) are between
142 m2 gÀ1 and 212 m2 gÀ1, depending on the weight fraction of
the template. In Fig. 1c, nitrogen adsorption isotherms of the
parent ompg-C3N4(r = 2.5) and modified Pd@ompg-C3N4(r = 2.5)
material are also presented. After the modification, the isotherm
shape was preserved, indicating that the Pd nanoparticles do not
block or alter the pore system.
The Pd loading after modification was analyzed by ICP-AES and
listed in Table 2. All the actual Pd loadings for the prepared four
catalysts are slightly less than the expected 10 w%, but the loadings
of Pd do not differ much on the four different kinds of C3N4 as
shown in Table 2.
3. Results and discussion
3.1. Catalyst characterization
Elemental analysis of mpg- and ompg-C3N4 results are shown in
Table S1 (Supporting information), and all these supports give a
C/N ratio around 0.68. The residual H contents attributable to
incomplete condensation or adsorbed water vary from 2.75 W%
to 3.10 W%. To understand the textural properties, the mpg-C3N4
and ompg-C3N4 supports prepared under different conditions
and the Pd@ompg-C3N4(r = 2.5) catalyst were characterized by N2
adsorption–desorption measurements. The isotherms and the
Barrett–Joyner–Halenda (BJH) pore size distributions are shown
Fig. 1. N2 adsorption/desorption isotherms of C3N4s (a), Pd@ompg-C3N4 (r=2.5) (c) and corresponding Barrett-Joyner-Halenda (BJH) pore-size distribution curve (b), (d)
determined from the desorption branch.

Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
275
Table 2
The small-angle XRD patterns of mpg-C3N4, ompg-C3N4, and
SBA-15 are shown in Fig. S1 (Supporting information). The pattern
of SBA-15 shows three well-resolved peaks, which can be assigned
to the (100), (110), (200) diffractions of a 2D hexagonal pore
structure with p6mm symmetry. The ompg-C3N4 pattern shows a
similar form, proving a perfect replication of the SBA-15-ordered
pore structure. However, there is no obvious peak for mpg-C3N4,
which is an evidence of a disordered pore system. The wide-angle
XRD patterns of ompg-C3N4 and Pd@ompg-C3N4 are shown in
Fig. 2a. A diffraction peak at 27.3° is observed for both of the
Pd loadings on four C3N4 supports and active carbon.
Entry
Catalyst
Pd loading (w%)
1
2
3
4
5
Pd@ompg-C3N4(r = 1)
Pd@ompg-C3N4(r = 1.5)
Pd@ompg-C3N4(r = 2.5)
Pd@mpg-C3N4(r = 1)
Pd@C
9.48
9.52
9.33
9.55
9.04
Fig. 2. XRD patterns of ompg-C3N4(r = 2.5) and Pd@ompg-C3N4(r = 2.5) (a) and XPS pattern of ompg-C3N4(r = 2.5) (b).
35
mean size=4.09 1.06
30
25
20
15
10
5
0
2
3
4
5
6
7
8
Particle Size (nm)
Fig. 3. TEM images of 10w%Pd@ompg-C3N4(r = 2.5) with different magnification (A–C) and size distribution of Pd particles (D).

276
Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
H2
Pd@C3N4
N
N
H
quinoline
1,2,3,4-tetrahydroquinoline (py-THQ)
N
N
H
decahydroquinoline (DHQ)
5,6,7,8-tetrahydroquinoline(bz-THQ)
Scheme 2. Possible reaction pathways for quinoline hydrogenation.
samples and ascribed to the characteristic peak of (002) packing of
g-C3N4. The diffraction peaks at 40.0°, 46.6°, and 68.1° in the XRD
patterns can be assigned to the (111), (200), and (220) planes
of Pd metal particles, respectively. The average size of the crystal-
lites was calculated to be 5.4 nm from the major diffraction peak
(111) using Scherrer’s formula. The morphologies of Pd@ompg-
C3N4 composite materials were also investigated by TEM (Fig. 3).
The TEM images revealed well-dispersed Pd particles with a mean
size of 4.1 nm (Fig. 3D), which is smaller than that calculated from
XRD results. XPS was then applied to identify the surface composi-
tion and oxidation state of Pd species. Fig. 2b shows the XPS spec-
tra in the Pd 3d binding energy region of Pd@ompg-C3N4(r = 2.5).
Two peaks that appear at 341.3 eV and 336.1 eV are related to
Pd0 3d3/2 and Pd0 3d5/2, while peaks at 343.1 eV and 337.6 eV are
attributed to Pd2+ 3d3/2 and Pd2+ 3d5/2. The percentage of the Pd0
and Pd2+ species was calculated from the relative areas of these
peaks. It indicates that Pd0 is the main metal species on the surface
of the as-prepared catalyst (ꢀ70%), which is consistent with
Pd@mpg-C3N4 as clarified in our previous work [18]. The catalytic
activity of precious metals supported on carbon material is highly
dependent on the nature of surface functional groups present on
the support. Indeed, support (carbon carriers) modification with
acidic oxygen-containing and basic nitrogen-containing surface
groups is already widely performed [37–39]. For example,
Radkevich and coworkers proved the presence of amine groups
in the carbon carrier helped stabilize Pd0 in a high-dispersed state
and resulted in increased proportion of Pd0, resistant to reoxida-
tion [40]. In our case, ompg-C3N4 has a high nitrogen content of
Table 3
Conversion of quinoline in different solvents.
Entry
Solvent
Conv. (%)a
Sel. (%)a
1
2
3
4
5
6
7
Ethanol
Water
82
32
99
80
86
57
76
100
100
99
100
100
100
100
Acetonitrile
Methanol
Toluene
Tetrahydrofuran
n-Hexane
Reaction conditions: quinoline (0.5 mmol), Pd (4.7 mol% relative to quinoline), 5 ml
solvent, using Pd@ompg-C3N4(r = 1).
a
Determined by GC and GC–MS.
genation, we selected Pd@ompg-C3N4(r = 1) as catalyst and tested
its reaction activity in several solvents. As demonstrated in Table 3,
Pd@ompg-C3N4 promoted the exclusive formation of py-THQ with
no evidence of benzene ring reduction or complete hydrogenation,
that is, exclusive nitrogen-containing ring hydrogenation, which
confirms the high selectivity of the as-prepared Pd@ompg-C3N4.
With Pd@ompg-C3N4(r = 1) under the primary conditions, we
could readily achieve 100% selectivity toward py-THQ, with a
conversion of quinoline of 32–86% in 4 h at 30 °C and 1 bar of
hydrogen pressure (Table 3). In homogeneous catalysis systems,
the electronic properties of the solvent can alter the mechanistic
aspects of a reaction via a change in the free energy state of the
reactants, which can alter the favorability of different reaction
paths [41]. With heterogeneous catalysts, the solvation of reacting
species and its impact on the overall reaction mechanism are not
clearly understood [42]. However, in a study of polar and nonpolar
reactants, it has been stated that hydrogenation of less polar sub-
strates in more polar solvents is preferred [43,44]. Note that in
our reaction systems, the reaction rates were affected by the choice
of solvents. However, no difference was observed in selectivity in
53 wt%. The
p-bonded planar CANAC-layers along with the
incompletely condensed amino groups (Supporting information,
Fig. S2) in the ompg-C3N4 are suitable for stabilizing highly
dispersed Pd0 particles, preventing their reoxidation and improv-
ing their activity in hydrogen activation.
3.2. Hydrogenation of quinoline
different solvents. Table
3 indicates that the conversion of
Quinoline hydrogenation ideally involves the regioselective
hydrogenation of the nitrogen-containing ring to form py-THQ.
However, the complete hydrogenation product decahydroquino-
line (DHQ) and the benzene ring hydrogenation product 5,6,7,8-
tetrahydroquinoline (bz-THQ) may also form during the reaction
process (Scheme 2). For example, commercial Pd@C catalyst only
gave an 84% selectivity toward py-THQ [24].
quinoline after 4 h in the different solvents decreased in the fol-
lowing order: acetonitrile > toluene > ethanol > methanol > n-hex-
ane > tetrahydrofuran > water. In this case, it was found that the
variation in specific activity did not correlate with either the sol-
vent dielectric constant, polarity, or the dipole moment (Support-
ing information, Fig. S3). This phenomenon was also observed in
the hydrogenation of citral using Pt/SiO2, and the reason has been
attributed to the higher multipole moments of some solvents [44].
Acetonitrile gave the highest activity but itself would be hydroge-
nated to triethylamine under harsh reaction conditions. In addi-
tion, considering the toxicity of acetonitrile and toluene, we
3.2.1. The effect of solvents
In order to investigate the catalytic response of Pd@C3N4
(including Pd@mpg-C3N4 and Pd@ompg-C3N4) in quinoline hydro-

Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
277
afterward limited the subsequent catalytic study in an ethanol sol-
vent feed at 1 bar of hydrogen pressure.
3.2.2. The effect of textural properties
In effort to compare the potential abilities of Pd@C3N4 with dif-
ferent textural properties, the conversions of quinoline as a func-
tion of reaction time over different Pd@C3N4 catalyst are
presented in Fig. 4. For comparison, the reaction result with
Pd@C as catalyst is also listed in the figure. It can be observed that
Pd@ompg-C3N4(r = 1) exhibited the highest activity among the
tested three catalysts. Yet, Pd@mpg-C3N4 showed an activity be-
tween that of the Pd@ompg-C3N4 and Pd@C. The reduction in quin-
oline using Pd@ompg-C3N4(r = 1) led to a 100% conversion in 100%
selectivity toward py-THQ at 40 °C in 3 h, and a similar Pd@mpg-
C3N4 catalyst gave a 54% conversion (100% selectivity) under the
same reaction conditions. In comparison, the reaction performed
using Pd@C as catalyst gave a 48% conversion and only 88% selec-
tivity toward py-THQ. The selectivity to py-THQ remained 100% on
Pd@C3N4 catalysts even the conversion reached 100%, but de-
creased markedly on the Pd@C catalyst. The superior activity of
Pd@C3N4 can be attributed to the high nitrogen content in the
carbon nitride carrier resulting in increased proportions of Pd0,
which was reported as the active site for the activation of H2
[18,40]. Pd@Al2O3 was reported to show very good activity and
selectivity in the hydrogenation of quinoline under relatively
harsher conditions (>100 °C, 2 MPa H2) [20], however, under our
mild reaction conditions (40 °C, 1 bar H2), only minor py-THQ
was observed (Supporting information, Fig. S6).
Fig. 4. 1,2,3,4-Tetrahydroquinoline yield as a function of time at 40 °C using 5 ml
ethanol as solvent over 25 mg Pd@mpg-C3N4(r = 1), Pd@ompg-C3N4(r = 1) and
27 mg Pd@C.
Pd@ompg-C3N4 catalysts show much higher efficiency than
Pd@mpg-C3N4. The advantage of using ompg-C3N4 can attribute
to the uniform cylindrical pores. The character allows substrates
to diffuse to the active sites with relatively less barriers. While in
the case of mpg-C3N4, the spherical pores are just connected by
random smaller bottlenecks which will hinder mass transfer in
all directions [27,35].
A comparative study of Pd@ompg-
C3N4(r = 1), Pd@ompg-C3N4(r = 1.5) and Pd@ompg-C3N4(r = 2.5)
was then carried out. As shown in Fig. 5, apparently different sur-
face area and pore size exhibit a significant influence in activity, for
example, with Pd@ompg-C3N4(r = 2.5) as catalyst, the reaction
completed in 2.5 h with 100% selectivity toward py-THQ which is
much faster than that using Pd@ompg-C3N4(r = 1) as catalyst.
Fig. 5. 1,2,3,4-Tetrahydroquinoline yield as a function of time at 40 °C using 5 ml
ethanol as solvent over Pd@ompg-C3N4.
3.2.3. The influence of temperature
Variation of the reaction temperature had a considerable effect
on the reaction conversion, as illustrated in Fig. 6. Conversion of
Fig. 6. 1,2,3,4-Tetrahydroquinoline yield as a function of time (a) and temperature (b) using 5 ml ethanol as solvent over 25 mg Pd@ompg-C3N4(r = 2.5) at different
temperature and time.

278
Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
investigated the recyclability of the Pd@ompg-C3N4(r = 2.5) cata-
lyst. Fig. 7 gives the results of catalyst recycling at 40 °C in ethanol.
The Pd@ompg-C3N4(r = 2.5) was able to recycle at least six times
without significant loss of activity, which is a prerequisite for prac-
tical applications. Highly dispersed Pd0 clusters are capable of
reoxidation during air (or oxygen) contact which may lead to the
deactivation of the catalysts [45]. However, our catalysts are very
stable in the air and the samples can be stored in air at room tem-
perature for several months without loss of their activities.
3.2.5. Possible mechanism
Py-THQ was the only reaction product observed over the entire
range of conditions studied, which suggests that the hydrogenation
of quinoline is going via the exclusive adsorption of N-heterocycle
part [46–48]. To specifically explain the exclusive adsorption
behaviors is difficult, but depending on the nature of the catalyst’s
active site, quinoline can interact with the surface through the N-
heterocycle to form strong NÁ Á ÁHAN interactions or
p p interac-
Á Á Á
Fig. 7. Reuse of Pd@ompg-C3N4(r = 2.5) under 40 °C for 2.5 h.
tions. Both our group and others have demonstrated the efficient
use of the NH and NH2 groups on the surface of carbon nitride
for the effective adsorption of phenol via OÁ Á ÁHAN or O-HÁ Á ÁN
interactions recently [18,49]. Similar to this result, although an
interaction between the aromatic ring of quinoline and the
tron of the carbon nitride could not be excluded, we expect that
such a interaction is weaker and quinoline adsorbs on the
Á Á Á
p elec-
p
p
carbon nitride mainly via a NÁ Á ÁHAN interaction. This conclusion
is supported by the TPD experiments. TPD experiments of benzene
and pyridine were conducted, and the results were displayed in
Fig. 8. Desorption of pyridine from C3N4 started at around 75 °C,
and it exhibited a maximum desorption rate at about 161 °C. While
for benzene, the temperatures for initial desorption and maximum
desorption rate located at 50 °C and 130 °C, respectively. The re-
sults demonstrate that the binding energy of pyridine with C3N4
is higher than that of benzene with C3N4. It indicates that the incor-
poration of nitrogen atom increases the interaction between pyri-
dine and C3N4. Similarly, pyridine ring of quinoline in principal
possesses higher bonding energy with C3N4 than benzene ring. This
stronger interaction may account for the high selectivity for
1,2,3,4-tetrahydroquinoline.
Fig. 8. TPD profiles of benzene and pyridine.
The transfer of electron density in the heterojunction from the
semiconductor C3N4 enriches the electron density of the metallic
Pd and accelerates the hydrogenation reaction, as compared to
Pd@C. Scheme 3 illustrates the proposed step sequence of the
hydrogenation of quinoline. In the initial stage, the adsorption of
N-heterocycle of quinoline onto the carbon nitride support is sig-
nificantly promoted by the formation of NÁ Á ÁHAN hydrogen bonds
between the N-heterocycle of quinoline molecule and the NH and
NH2 groups on the surface of carbon nitride. Subsequently, those
quinoline molecules interacted with the active H produced over
Pd nanoparticles, resulting in the formation of py-THQ. That the
product of quinoline hydrogenation is depended on the adsorption
types of quinoline over the catalyst had already been studied in
detail by Fish et al. [46], and improved selectivity toward the
100% was achieved with a 100% selectivity within 2.5 h at 1 bar of
hydrogen pressure and 40 °C. The reaction was accelerated at high-
er temperature but without any losing of the selectivity, for exam-
ple, full conversion in 1.5 h was obtained at 50 °C with a high
selectivity of 100%. But, even at a low temperature of 30 °C, 100%
conversion and 100% selectivity could be reached after a minor
longer reaction time of 4 h. Hydrogenation under mild conditions
is extremely desirable in industry.
3.2.4. The reusability of the catalyst
It has been reported that the commercial Pd@C lost its activity
in quinoline hydrogenation after used for 3 times [24]. We then
N
H
H
H
H
H
H
H
H
N
H
N
N
e
Pd
Pd
Pd
Carbon Nitride
Scheme 3. Possible reaction mechanism of quinoline hydrogenation over Pd@C3N4.

Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
279
Table 4
Hydrogenation of other quinolines and 2,3-benzofuran.
Substrates
Product
Solvent
Ethanol
Temperature (°C)
Time (h)
5
Conv.a (%)
>99
Sel.a (%)
60
95
Ethanol
Toluene
Toluene
60
100
100
2
2
1
>99
>99
>99
>99
>99
>99
Ethanol
Ethanol
40
40
3
5
80
95
>99
>99
Reaction conditions: 0.5 mmol substrate, 4.7 mol%Pd, 5 ml solvent, unless otherwise noted.
a
Determined by GC and GC–MS.
formation of py-THQ via hydrogen bonds was also observed by Shi
and coworkers [24].
Acknowledgments
We thank Prof. M. Antonietti for helpful discussion. Financial
support from the Joint Petroleum and Petrochemical Funds of the
National Natural Science Foundation of China and China National
Petroleum Corporation (U1162124), the Fundamental Reasearch
Funds for the Central Universities, the Program for Zhejiang Lead-
ing Team of S&T Innovation (2011R50007), and the Partner Group
Program of the Zhejiang University and the Max-Planck Society are
greatly appreciated.
3.2.6. Test for other quinolines
Furthermore, the applicability of Pd@ompg-C3N4(r = 2.5) cata-
lyst to different heterocyclic compounds was investigated, and
the results are summarized in Table 4. The reactions were retarded
when methyl group was introduced. Therefore, a little higher tem-
perature is needed for the hydrogenation of substituted quinolines
to corresponding 1,2,3,4-tetrahydroquinolines smoothly. 2,3-
Benzofuran was used to test if this kind of catalyst is also highly
effective for hydrogenation of oxygen heterocyclic compounds.
To our delight, the catalyst actually showed high activity and
selectivity toward the reduction of 2,3-benzofuran as demon-
strated in Table 4.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.10.018.
These data include MOL files and InChiKeys of the most important
compounds described in this article.
4. Conclusions
References
Mpg-C3N4 and ompg-C3N4 were successfully synthesized, and
Pd was subsequently deposited by an easy ultrasonic-assisted
method, leading to the formation of highly dispersed Pd nanopar-
ticles. The polymeric C3N4 supports, especially the ompg-C3N4
remarkably enhanced both the catalytic activity and the selectivity
in the hydrogenation of quinoline with molecular hydrogen as a
reductant, compared with the active carbon. Recycling experiment
of Pd@ompg-C3N4 indicated that the catalyst could be reused for
more than six times without significant loss of catalytic activity
and selectivity. The enhanced catalytic activity can be attributed
[1] R.T. Shuman, P.L. Ornstein, J.W. Paschal, P.D. Gesellchen, J. Org. Chem. 55
(1990) 738.
[2] A.R. Katritzky, S. Rachwal, B. Rachwal, Tetrahedron 52 (1996) 15031.
[3] R. Omar-Amrani, A. Thomas, E. Brenner, R. Schneider, Y. Fort, Org. Lett. 5 (2003)
2311.
[4] T. Kubo, C. Katoh, K. Yamada, K. Okano, H. Tokuyama, T. Fukuyama,
Tetrahedron 64 (2008) 11230.
[5] K. Maruoka, T. Miyazaki, M. Ando, Y. Matsumura, S. Sakane, K. Hattori, H.
Yamamoto, J. Am. Chem. Soc. 105 (1983) 2831.
[6] Y.G. Zhou, Acc. Chem. Res. 40 (2007) 1357.
[7] G.E. Dobereiner, A. Nova, N.D. Schley, N. Hazari, S.J. Miller, O. Eisenstein, R.H.
Crabtree, J. Am. Chem. Soc. 133 (2011) 7547.
to synergetic interaction between Pd and C3N4 supports, the
p-
[8] R.H. Fish, A.D. Thormodsen, G.A. Cremer, J. Am. Chem. Soc. 104 (1982) 5234.
[9] Y. Alvarado, M. Busolo, F. Lopez-Linares, J. Mol. Catal. A: Chem. 142 (1999) 163.
[10] M. Rosales, L.J. Bastidas, B. Gonzalez, R. Vallejo, P.J. Baricelli, Catal. Lett. 141
(2011) 1305.
[11] S.M. Lu, X.W. Han, Y.G. Zhou, J. Organomet. Chem. 692 (2007) 3065.
[12] Z.J. Wang, H.F. Zhou, T.L. Wang, Y.M. He, Q.H. Fan, Green Chem. 11 (2009) 767.
[13] W.B. Wang, S.M. Lu, P.Y. Yang, X.W. Han, Y.G. Zhou, J. Am. Chem. Soc. 125
(2003) 10536.
bonded planar CANAC-layers along with the incompletely
condensed amino groups in the carbon nitride are suitable for
stabilizing highly dispersed Pd0 particles. Good accessibility of Pd
active sites and high concentration of surface Pd0 contribute to
the high reaction activity over C3N4-supported Pd catalysts. The
introduction of ordered cylindrical mesoporous structure may
facilitate mass transfer and therefore lead to better catalytic
performance.
[14] S.M. Lu, X.W. Han, Y.G. Zhou, Adv. Synth. Catal. 346 (2004) 909.
[15] M. Rosales, S. Castillo, A. Gonzalez, L. Gonzalez, K. Molina, J. Navarro, I.
Pacheco, H. Perez, Transition Met. Chem. 29 (2004) 221.

280
Y. Gong et al. / Journal of Catalysis 297 (2013) 272–280
[16] R. Yamaguchi, C. Ikeda, Y. Takahashi, K. Fujita, J. Am. Chem. Soc. 131 (2009)
8410.
[33] F. Goettmann, A. Fischer, M. Antonietti, A. Thomas, Angew. Chem. Int. Ed. 45
(2006) 4467.
[17] R.A. Sanchezdelgado, D. Rondon, A. Andriollo, V. Herrera, G. Martin, B.
Chaudret, Organometallics 12 (1993) 4291.
[34] D.Y. Zhao, Q.S. Huo, J.L. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120
(1998) 6024.
[18] Y. Wang, J. Yao, H.R. Li, D.S. Su, M. Antonietti, J. Am. Chem. Soc. 133 (2011)
2362.
[35] X.F. Chen, Y.S. Jun, K. Takanabe, K. Maeda, K. Domen, X.Z. Fu, M. Antonietti, X.C.
Wang, Chem. Mater. 21 (2009) 4093.
[19] R.A. Sanchez-Delgado, N. Machalaba, N. Ng-A-Qui, Catal. Commun. 8 (2007)
2115.
[36] Y.S. Jun, W.H. Hong, M. Antonietti, A. Thomas, Adv. Mater. 21 (2009)
4270.
[20] M. Campanati, A. Vaccari, O. Piccolo, J. Mol. Catal. A: Chem. 179 (2002) 287.
[21] A. Spitaleri, P. Pertici, N. Scalera, G. Vitulli, M. Hoang, T.W. Turney, M. Gleria,
Inorg. Chim. Acta 352 (2003) 61.
[22] M. Campanati, M. Casagrande, I. Fagiolino, M. Lenarda, L. Storaro, M.
Battagliarin, A. Vaccari, J. Mol. Catal. A: Chem. 184 (2002) 267.
[23] N. Hashimoto, Y. Takahashi, T. Hara, S. Shimazu, T. Mitsudome, T. Mizugaki, K.
Jitsukawa, K. Kaneda, Chem. Lett. 39 (2010) 832.
[24] H. Mao, C. Chen, X.P. Liao, B. Shi, J. Mol. Catal. A: Chem. 341 (2011) 51.
[25] Y. Wang, X.C. Wang, M. Antonietti, Angew. Chem. Int. Ed. 51 (2012) 68.
[26] A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J.O. Muller, R. Schlogl, J.M.
Carlsson, J. Mater. Chem. 18 (2008) 4893.
[27] X. Jin, V.V. Balasubramanian, S.T. Selvan, D.P. Sawant, M.A. Chari, G.Q. Lu, A.
Vinu, Angew. Chem. Int. Ed. 48 (2009) 7884.
[37] D.A. Bulushev, I. Yuranov, E.I. Suvorova, P.A. Buffat, L. Kiwi-Minsker, J. Catal.
224 (2004) 8.
[38] S. Hermans, C. Diverchy, O. Demoulin, V. Dubois, E.M. Gaigneaux, M. Devillers,
J. Catal. 243 (2006) 239.
[39] M. Gurrath, T. Kuretzky, H.P. Boehm, L.B. Okhlopkova, A.S. Lisitsyn, V.A.
Likholobov, Carbon 38 (2000) 1241.
[40] V.Z. Radkevich, T.L. Senko, K. Wilson, L.M. Grishenko, A.N. Zaderko, V.Y. Diyuk,
Appl. Catal. A – Gen. 335 (2008) 241.
[41] M.H. Abraham, P.L. Grellier, J.L.M. Abboud, R.M. Doherty, R.W. Taft, Can. J.
Chem. – Rev. Can. Chim. 66 (1988) 2673.
[42] B.S. Akpa, C. D’Agostino, L.F. Gadden, K. Hindle, H. Manyar, J. McGregor, R. Li,
M. Neurock, N. Sinha, E.H. Stitt, D. Weber, J.A. Zeitler, D.W. Rooney, J. Catal. 289
(2012) 30.
[28] J.J. Zhu, S.A.C. Carabineiro, D. Shan, J.L. Faria, Y.J. Zhu, J.L. Figueiredo, J. Catal.
274 (2010) 207.
[29] Y. Wang, Y. Di, M. Antonietti, H.R. Li, X.F. Chen, X.C. Wang, Chem. Mater. 22
(2010) 5119.
[30] Y. Wang, J.S. Zhang, X.C. Wang, M. Antonietti, H.R. Li, Angew. Chem. Int. Ed. 49
(2010) 3356.
[31] Y. Wang, H.R. Li, J. Yao, X.C. Wang, M. Antonietti, Chem. Sci. 2 (2011) 446.
[32] P.F. Zhang, Y. Wang, J. Yao, C.M. Wang, C. Yan, M. Antonietti, H.R. Li, Adv. Synth.
Catal. 353 (2011) 1447.
[43] R.A. Rajadhyaksha, S.L. Karwa, Chem. Eng. Sci. 41 (1986) 1765.
[44] S. Mukherjee, M.A. Vannice, J. Catal. 243 (2006) 108.
[45] M.V. Klyuev, A.A. Nasibulin, Kinet. Catal. (Russia) 37 (1996) 231.
[46] E. Baralt, S.J. Smith, J. Hurwitz, I.T. Horvath, R.H. Fish, J. Am. Chem. Soc. 114
(1992) 5187.
[47] R.H. Fish, E. Baralt, H.S. Kim, Organometallics 10 (1991) 1965.
[48] R.H. Fish, J.N. Michaels, R.S. Moore, H. Heinemann, J. Catal. 123 (1990) 74.
[49] E. Haque, J.W. Jun, S.N. Talapaneni, A. Vinu, S.H. Jhung, J. Mater. Chem. 20
(2010) 10801.