Metal halides supported on mesoporous carbon nitride as efficient heterogeneous catalysts for the cycloaddition of CO2

Article abstract of DOI:10.1016/j.molcata.2015.03.024

Abstract A series of ZnCl₂/mp-C₃N₄ catalysts were prepared through a wet impregnation method using mesoporous g-C₃N₄ materials as catalytic supports. The physicochemical properties of the ZnCl₂/mp-C₃N₄ catalysts were characterized using several techniques, including N₂ adsorption-desorption, X-ray diffraction, Fourier transform infrared, UV-vis diffuse reflectance, and X-ray photoelectron spectroscopy. In the cycloaddition reactions of CO₂ with propylene oxide (PO) to propylene carbonate (PC), the ZnCl₂/mp-C₃N₄ materials revealed high catalytic performances at 140 C and 6 h. Furthermore, other metal halides including ZnBr₂, CoCl₂, FeCl₃, NiCl₂, and MgCl₂ catalysts supported on mp-C₃N₄ were fabricated and the synthesized catalysts also demonstrated high performances in the catalytic cycloaddition of CO₂. A possible catalytic mechanism has been proposed. The mp-C₃N₄ material served mainly as a basic support to adsorb CO₂ whereas the supported metal halides were active sites to promote the activation of PO and realize its transform into PC.

Full text of DOI:10.1016/j.molcata.2015.03.024

Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Metal halides supported on mesoporous carbon nitride as efficient
heterogeneous catalysts for the cycloaddition of CO2
Jie Xu∗, Fei Wu, Quan Jiang, Jie-Kun Shang, Yong-Xin Li∗
Jiangsu Key Laboratory of Advanced Catalytic Materials and Technology, School of Petrochemical Engineering, Changzhou University, Gehu Road 1,
Changzhou, Jiangsu 213164, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of ZnCl2/mp-C3N4 catalysts were prepared through a wet impregnation method using meso-
porous g-C3N4 materials as catalytic supports. The physicochemical properties of the ZnCl2/mp-C3N4
catalysts were characterized using several techniques, including N2 adsorption–desorption, X-ray diffrac-
tion, Fourier transform infrared, UV–vis diffuse reflectance, and X-ray photoelectron spectroscopy. In the
cycloaddition reactions of CO2 with propylene oxide (PO) to propylene carbonate (PC), the ZnCl2/mp-
C3N4 materials revealed high catalytic performances at 140 ◦C and 6 h. Furthermore, other metal halides
including ZnBr2, CoCl2, FeCl3, NiCl2, and MgCl2 catalysts supported on mp-C3N4 were fabricated and the
synthesized catalysts also demonstrated high performances in the catalytic cycloaddition of CO2. A pos-
sible catalytic mechanism has been proposed. The mp-C3N4 material served mainly as a basic support to
adsorb CO2 whereas the supported metal halides were active sites to promote the activation of PO and
Received 3 February 2015
Received in revised form 30 March 2015
Accepted 30 March 2015
Available online 2 April 2015
Keywords:
Carbon nitride
CO2 conversion
Cycloaddition
Propylene carbonate (PC)
realize its transform into PC.
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
cation. By contrary, heterogeneous catalysts, including grafted ILs
[12–14], modified molecular sieves [8], smectites [15,16], and
Notwithstanding being a major greenhouse gas, carbon diox-
ide (CO2) is an abundant, nontoxic, economic, and renewable
C1 building block [1,2]. Within the past decades, concerns about
anthropogenic emissions along with climate change have sparked
global interest in the chemical transformation of CO2 to valuable
chemicals [3,4]. One of the most interesting protocols utilizing CO2
as a raw material is the cycloaddition of CO2 with epoxides (e.g.
propylene oxide, PO) to cyclic carbonates (e.g. propylene carbonate,
PC) [5,6]. Such products are among the most important feedstocks
which are extensively used as aprotic polar solvents, electrolyte
components in lithium batteries, precursors for polycarbonates,
and intermediates in organic synthesis [7,8].
A wide range of catalysts, both homogeneous and heteroge-
neous systems, have been developed for the cycloaddition of CO2 to
cyclic carbonates. Among them, ionic liquids (ILs) [9], organometal-
lic complexes [2], quaternary ammonium [7] and phosphonium
[1,10] salts have been proved to be the most efficient catalysts
[3,11]. Despite the excellent catalytic activities, due to their intrin-
sic homogeneous nature, the main drawback in the use of such
catalysts is the difficulty in catalyst recovery and product purifi-
mixed metal oxides [17,18], etc, offer technical advantages in the
catalyst–product separation and reactor design [18]. However, the
solid catalysts still suffer from low catalytic activity and/or selec-
tivity. In this context, it is of high interest to explore and develop
a novel catalyst which can afford high catalytic activity as well as
convenient catalyst separation [6].
Recently, owing to its unique combination of multiple physico-
chemical properties, graphitic carbon nitride (g-C3N4) has attracted
a great deal of attention and also been proposed as the most promis-
ing candidate to complement traditional carbon materials [19]. Up
to now, g-C3N4, especially mesoporous g-C3N4 having large surface
areas and rich pores, has been widely applied in numerous research
fields, including photocatalysis [20,21], fuel cells [22,23], gas stor-
age [24], and heterogeneous catalysis [25–28]. As an analogue to
graphite, g-C3N4 possesses two-dimensional stacked structures
with ␲-conjugated planar layers. In particular, at the edges of the
graphitic sheets, there exist a large number of uncondensed termi-
nal amine and imine groups [26,29]. The characteristic structures
thus enable g-C3N4 as a typical solid base material. Ansari et al.
[30,31] synthesized a series of mesoporous g-C3N4 materials using
nanocasting approaches, and found the synthesized g-C3N4 could
catalyze various CO2-invloving transformations. Likewise, Huang
et al. [32] prepared g-C3N4/SBA-15 via a chemical vapor deposition
method and investigated the catalytic performance in cycloaddi-
∗
Corresponding authors. Tel.: +86 519 86 330 135.
E-mail addresses: shine6832@163.com (J. Xu), liyxluck@163.com (Y.-X. Li).
http://dx.doi.org/10.1016/j.molcata.2015.03.024
1381-1169/© 2015 Elsevier B.V. All rights reserved.

78
J. Xu et al. / Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
tion of CO2 to cyclic carbonates. Despite the above pioneer work, it
should be noted that the basic sites of g-C3N4 (or even mesoporous
g-C3N4) belong to weak base [33,34]. Hence, in the base-centered
CO2-activiating reactions, the catalytic activity was limited, and
thus rigorous reactions conditions were demanded.
UV–vis diffuse reflectance spectra (DRS) were recorded on a
Shimizu UV-3600 spectrophotometer. BaSO4 was used as a stan-
dard reference. Each sample was pressed into a thin tablet and
tested under ambient conditions. The absorption spectrum was
calculated from the reflectance data with Kubelka–Munk function.
Fourier transform infrared (FT-IR) spectra of the samples
were collected in transmission mode from KBr pellets at room
temperature on a Bruker Tensor 27 spectrometer with a reso-
lution of 4 cm−1, using 32 scans per spectrum in the region of
400–4000 cm−1. The mass ratio of every sample to KBr was constant
at 1:100.
X-ray photoelectron spectroscopy (XPS) measurements were
performed using a Perkin–Elmer PHI 5000C spectrometer work-
ing in the constant analyzer energy mode with Mg K˛ radiation as
the excitation source. The carbonaceous C 1s line (284.6 eV) was
used as the reference to calibrate the binding energies.
Zinc halides (e.g. ZnCl2) are also high-performance catalysts for
the cycloaddition of CO2 with epoxide as it has been revealed that
Zn2+ cations could activate the epoxide molecule by the coordi-
nation of oxygen atom thereof [1,3]. On the other hand, Zhu et al.
[35] reported that Zn-doped g-C3N4 sample demonstrated superior
activity to the bare g-C3N4 in the NO decomposition. Inspired by
the reports, in the present contribution, we prepared mesoporous-
g-C3N4-supported ZnCl2 catalysts via a simple wet impregnation
method. In the cycloaddition reactions of CO2 with PO to PC, the
ZnCl2/mp-C3N4 catalysts demonstrated high activity, affording a
maximum a PO conversion of 73%. Moreover, it has been also found
that other metal halides catalysts supported on mp-C3N4 could
catalyze the reactions with high performance.
2.4. Catalytic test
2. Experiment
The cycloaddition of CO2 with PO was carried out in 80 mL stain-
less steel autoclave equipped with a magnetic stirrer. In a typical
reaction process, 7 mL of PO, 3 mL N,N-dimethyl formamide (DMF),
and 0.2 g of the catalyst were added into the reactor. Then, the reac-
tor was pressurized with CO2 to a desired pressure and heated to
140 ◦C under stirring for 6 h. After the reaction, the autoclave was
cooled down to the room temperature in ice water and the excess
of CO2 was vented. The liquid product was separated by centrifu-
gation and analyzed using a GC equipped with a SE-54 capillary
column and FID. The liquid mixture consisted of DMF, PO, PC, and
1,2-propylene glycol (PG) as the byproduct originating from the
hydrolysis of PO with trace H2O. No other substance was detected.
The carbon balance was nearly 100%. Their quantitative calcula-
tion (i.e. conversion of PO, and selectivity to PC) was based on an
area-normalization method.
2.1. Preparation of mp-C3N4
The mesoporous C3N4 was prepared according to an established
nanocasting method reported previously [25]. 4 g of cyanamide was
dissolved in 16 g of aqueous suspensions of 12 nm silica spheres
(Ludox HS40, Aldrich) under vigorous stirring. The mixture was
heated in an oil bath at 50 ◦C under stirring overnight to remove
water. The resultant white solid was ground in a mortar, transferred
into a covered crucible, and heated at 3 ◦C min−1 up to 550 ◦C and
then treated for further 4 h. Afterwards, the as-synthesized yellow
powder was ground and immersed into 200 mL of NH4HF2 aqueous
solution (4 mol L−1) for 2 days to remove the template. Then, the
dispersion was centrifuged and the yellow precipitate was washed
using distilled water and ethanol for several times. Finally, the yel-
low sample was dried at 50 ◦C under vacuum overnight and the
mass of the obtained C3N4 was ca. 1.8 g. The resulting C3N4 sample
was designated as mp-C3N4.
The PO conversion and selectivity to PC were calculated as fol-
lows:
APC × fPC + APG × fPG
APC × fPC
Conv. =
, Sel. =
APO + APC × fPC + APG × fPG
APC × fPC + APG × fPG
(1)
2.2. Preparation of ZnCl2/mp-C3N4
where A, and f were the peak area of GC, and response factor for
each product.
0.1 g of ZnCl2 (0.73 mmol) was dissolved into 20 mL of absolute
ethanol. 1 g mp-C3N4 was then added into the solution and stirred
for 1 h. Next, the mixture was heated in a bath at 50 ◦C for several
hours to remove ethanol. The obtained yellow solid was calcinated
at 300 ◦C for 2 h under N2 atmosphere. The resultant material was
labeledas mZnCl2/mp-C3N4, where m indicatedthe loadingamount
(wt%).
3. Results and discussions
3.1. Structure characterization
Fig. 1 displays the XRD patterns of mp-C3N4 and ZnCl2/mp-
C3N4 materials with various loading amounts. The XRD pattern of
mp-C3N4 revealed a pronounced peak at 2ꢁ = 27.6◦ (d = 0.323 nm),
indexed as (0 0 2) planes, which was characteristic stacking of the
conjugated aromatic system [25,36]. In addition, a minor peak with
low intensity was also observed at 2ꢁ = 13.2◦, which was attributed
to the in-plane structural packing motif, e.g. the hole-to-hole array
of nitride pores [37]. Likewise, whereas the whole intensity of the
diffraction peaks decreased upon introducing ZnCl2, the ZnCl2/mp-
C3N4 demonstrated similar XRD patterns to the pure mp-C3N4
sample, suggesting that the overall graphitic structures of mp-C3N4
have undergone no significant variation.
N2 adsorption–desorption isotherms of mp-C3N4 and
ZnCl2/mp-C3N4 materials are presented in Fig. 2. All materi-
als showed type IV isothermal curves with H2 hysteresis loop
at p/p0 = 0.6–0.95, indicating that the samples had typical meso-
porous structures, along with relatively concentrated pore size
distributions. Furthermore, the adsorption curves proceeded fur-
ther ascent at p/p0 > 0.95, suggesting that these materials contained
2.3. Sample characterization
X-ray diffraction patterns were recorded with
a Rigaku
D/max 2500 PC X-ray diffractometer equipped with a graphite
monochromator (40 kV, 40 mA) using Ni-filtered Cu-K˛ radiation
(ꢀ = 1.5418 Å).
Nitrogen adsorption–desorption isotherms were measured at
−196 ◦C using a Micromeritics ASAP 2020 analyzer. Prior to the
analysis, the samples were degassed (10 ␮mHg) at 150 ◦C for at
least 4 h. The specific surface area was calculated according to the
Brunauer–Emmet–Teller (BET) method, and pore size distribution
was determined by the Barret–Joyner–Halenda method.
Transmission electron microscopy (TEM) experiments were
conducted on a JEOL 2010 electron microscope operating at 200 kV.
Before being transferred into the TEM chamber, the samples
dispersed in ethanol were deposited onto holey carbon films sup-
ported on Cu grids.

J. Xu et al. / Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
79
Fig. 1. XRD patterns of mp-C3N4 and mZnCl2/mp-C3N4 materials (a: mp-C3N4, b–e:
m = 2.5–10 wt%).
Fig. 2. N2 adsorption–desorption isotherms of mp-C3N4 and mZnCl2/mp-C3N4
materials (a: mp-C3N4, b–f: m = 2.5–12.5 wt%).
Table 1
Textual parameters of mp-C3N4 and ZnCl2/mp-C3N4 materials.
The information of chemical bonding was investigated by FT-
IR technique. As shown in Fig. 3A, the FT-IR spectra of both mp-
C3N4 and ZnCl2/mp-C3N4 exhibited very similar bands. The broad
bands located at 3450–3250 cm−1 were indicative of primary (i.e.
terminal –NH2 groups linked to the edges of graphitic CN sheets)
Sample
SBET (m2 g−1
)
Pore size (nm) Pore volume (cm3 g−1
)
mp-C3N4
298
244
232
231
206
198
15.3
15.2
15.0
14.5
12.6
10.7
0.93
0.95
0.93
0.80
0.72
0.67
2.5ZnCl2/mp-C3N4
5.0ZnCl2/mp-C3N4
7.5ZnCl2/mp-C3N4
10ZnCl2/mp-C3N4
12.5ZnCl2/mp-C3N4
and secondary amines, and physically adsorbed water (H
O H)
molecules [38]. The multiple bands in the range of 1600–1200 cm−1
were assigned to the stretching modes of aromatic amines [39].
Furthermore, the sharp transmission peaks with strong intensity at
ca. 810 cm−1 were evidence of the presence of triazine units [39,40].
Despite these analogies, the spectra of ZnCl2/mp-C3N4 materials
demonstrated weak bands at ca. 2360 cm−1 which have not been
detected in that of mp-C3N4. The unique bands were attributed to
the trapped CO2 molecules [31]. As mentioned above, owing to the
presence of plenty of amine species at the edge of graphitic sheets,
mp-C3N4 was a typical solid base and could adsorb/absorb ambient
acidic CO2. However, the basic intensity of mp-C3N4 was very weak
and meanwhile, due to the low detection sensitivity, the adsorbed
a small proportion of macropores which should originate from
interparticle void. Compared with the pure mp-C3N4 sample, the
ZnCl2-loaded mp-C3N4 materials displayed no apparent change
in terms of isotherms, only lower adsorbed quantity. Correspond-
ingly, as listed in Table 1, upon increasing the loading amounts,
the textual parameter including surfaces and pore volumes
decreased monotonously. However, the pore sizes underwent no
obvious decline. This confirmed that after the loading of ZnCl2, the
mesostructures of mp-C3N4 have well remained.
Fig. 3. FT-IR (A) and UV–vis (B) spectra of mp-C3N4 and mZnCl2/mp-C3N4 materials (A: a: mp-C3N4, b–e: m = 2.5–10 wt%).

80
J. Xu et al. / Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
B
A
Zn 2p 1/2
Zn 2p 3/2
Zn 2p3
N 1s
NKLL Mg
OKLL Mg
O 1s
C 1s
500
Cl 2p3
1000
1000
800
600
400
200
1055 1050 1045 1040 1035 1030 1025 1020
Binding energy (eV)
Binding energy (eV)
Fig. 4. XPS survey (A) and Zn 2p (B) spectra of 10ZnCl2/mp-C3N4.
CO2 by g-C3N4 was hardly detected by the present FT-IR technique.
Given this point, it was implied that after the loading of ZnCl2, the
total basic intensity of mp-C3N4 had been enhanced.
10ZnCl2/mp-C3N4 sample. The Zn 2p spectrum (Fig. 4B) showed
two independent peaks located at 1023 and 1046 eV, correspond-
ing to the signals of Zn 2p 3/2 and 1/2, respectively. The values
verified that the chemical valence of Zn element was +2, similar to
the results involving ZnCl2 catalysts supported on other materials
[43,44].
The effect of metal loading on the electronic structure of mp-
C3N4 was analyzed by UV–vis DRS characterization. As shown
in Fig. 3B, the spectrum of mp-C3N4 presented a remarkable
absorption peak centered at ca. 410 nm, which was derived from
bandgap between HOMO and LUMO in the polymeric melon units
of g-C3N4 [35]. Similarly, in the case of 10ZnCl2/mp-C3N4, its
spectrum showed a major absorption in UV region. However,
the UV absorption edge of 10ZnCl2/mp-C3N4 was moved towards
longer wavelengths in comparison with that of mp-C3N4. The phe-
nomenon has also been previously reported over g-C3N4 supported
Fe [36], Zn [41], and V [42] catalysts. The shift of absorption peak
after the loading of ZnCl2 probably resulted from the d–p repul-
sion of the Zn 3d and N 2p orbits, namely a host–guest interaction
between the ZnCl2 and g-C3N4 [36].
XPS technique was further employed to probe the surface
chemical environment of the ZnCl2/mp-C3N4 materials and a rep-
resentative survey spectrum of 10ZnCl2/mp-C3N4 was presented in
Fig. 4A. The surface of 10ZnCl2/mp-C3N4 was mainly constituted of
C, N, O, Zn, and Cl species. Wherein, the O element should originate
from the adsorbed water due to the mesostructures of mp-C3N4.
Based on the peak areas calculated, the weight percentage of ZnCl2
was ca. 9.2%, very close to the nominal loading amount of the
3.2. Catalyst activity
The ZnCl2/mp-C3N4 materials were employed as solid catalysts
for cycloaddition of CO2 with PO. As listed in Table 2, the catalytic
activity acquired over the bare mp-C3N4 sample was limited; only
a conversion of 10% was received. After the introduction of a small
amount of ZnCl2, the catalytic conversion of PO was improved. In
the case of the product distribution, the selectivity to the desired PC
was over 99%. Further elevating the loading amounts, the activity
increased progressively. The highest PO conversion was achieved
over the 10ZnCl2/mp-C3N4 catalyst, affording a maximum PC yield
as much as 72.6%. On the other hand, under the same reaction
conditions, the unload ZnCl2 sample exhibited an outstanding PO
conversion. It is well documented that zinc halides can efficiently
promote the cycloaddition of CO2 with epoxides to cyclic carbon-
ates [1,10,45]. Compared with homogeneous catalysts, almost all
heterogeneous catalysts suffer from lower reaction rates due to
mass transfer and diffusion of the reactants to the catalytically
active sites. Therefore, it is reasonable that the catalytic activity
Table 2
Catalytic performances of different ZnCl2/mp-C3N4 catalystsa.
Table 3
Catalytic performances of cycloaddition reactions of CO2 under various DMF
Catalyst
Conv. (%)
Sel. (%)
Yield (%)
volumesa.
mp-C3N4
10.4
27.9
43.6
58.2
73.0
71.5
97.3
99.7
99.8
99.4
99.6
99.4
99.7
99.5
10.4
27.8
43.3
58.0
72.6
71.3
96.8
Entry
VPO (mL)
10
8
7
6
5
VDMF (mL)
Conv. (%)
Sel. (%)
Yield (%)
2.5ZnCl2/mp-C3N4
5.0ZnCl2/mp-C3N4
7.5ZnCl2/mp-C3N4
10ZnCl2/mp-C3N4
12.5ZnCl2/mp-C3N4
ZnCl2
1
2
3
4
5
0
2
3
4
5
25.2
55.2
73.0
72.9
65.3
99.1
99.6
99.4
99.6
99.5
25.0
55.0
72.6
72.6
65.0
a
a
Reaction conditions: VPO = 7 mL, VDMF = 3 mL, pCO2 = 2.5 MPa, T = 140 ◦C, t = 6 h,
Evaluated over 10ZnCl2/mp-C3N4. Reaction conditions: pCO2 = 2.5 MPa,
T = 140 ◦C, t = 6 h, and Wcatal. = 0.2 g.
and Wcatal. = 0.2 g.

J. Xu et al. / Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
81
Fig. 5. Effects of reaction time and temperature on the catalytic performances of 10ZnCl2/mp-C3N4. Reaction conditions: VPO = 7 mL, VDMF = 3 mL, pCO2 = 2.5 MPa, and
Wcatal. = 0.2 g.
Table 4
decreased when ZnCl2 were supported on mp-C3N4. However, it
a
Recycling tests for 10ZnCl2/mp-C3N4 in the cycloaddition reactions of CO2
.
should be noted that the reaction catalyzed by ZnCl2 is homoge-
neous, suffering from difficulty in catalyst separation as well as
product purification.
Catalytic run
Conv. (%)
Sel. (%)
1
2
3
4
73.0
59.3
58.7
57.9
99.4
99.2
99.4
99.5
Table 3 shows the catalysts results of cycloaddition reactions of
CO2 catalyzed by the 10ZnCl2/mp-C3N4 catalyst with various DMF
volumes. Under the DMF-free condition, the PO conversion was
25.2% (entry 1), while after the addition of DMF, the obtained cat-
alytic activity increased noticeably. According to the work reported
by Zhong et al., in the CO2-involving heterogeneous reactions,
besides being a solvent, DMF could activate CO2 molecule by form-
ing carboxylate ion [45]. As expected, with the increase of DMF
volume (entries 2–4), the catalytic conversions increased progres-
sively. However, excessive employment of DMF would deteriorate
the catalytic activity (entry 5). One possible reason accounting
for this phenomenon was that redundant DMF might impede the
adsorption of reactants on the surface of the present heterogeneous
catalyst, thus undermining the catalytic efficiency. To elucidate the
effect of reaction conditions on the catalytic performances, the
variations of the catalytic conversions at different reaction time
and temperatures were further studied (Fig. 5). At the first 2 h,
the catalytic reaction proceeded with a low PO yield of ca. 22%
(Fig. 5A). Upon prolonging the evaluation, the catalytic activity
increased remarkably but leveled off at 6 h. The reaction tempera-
ture was also a key parameter to affect the catalytic performances.
As shown in Fig. 5B, the cycloaddition of CO2 with PO almost hardly
reacted under mild temperatures below 110 ◦C; as the tempera-
ture was increased, the PC yield increased drastically. It is widely
reported that the cycloaddition reactions of CO2 with epoxides are
typically exothermic processes. Therefore, in viewpoint of its ther-
modynamic equilibrium, higher temperatures would inhibit the
formation of cyclic carbonates [6]. In addition, higher temperatures
are liable to result in the polymerization of the cyclic carbonates,
and therein deteriorate the catalytic productivity [46]. Considering
the above results, a temperature of 140 ◦C and 6 h were selected
as optimal reaction conditions for realizing the highest catalytic
activity.
a
Reactionconditions:VPO = 7 mL, VDMF = 3 mL, pCO2 = 2.5 MPa, Wcatal. = 0.2 g, t = 6 h,
and T = 140 ◦C.
alytic performances. Under a low pressure of 0.5 MPa (Fig. 6A),
the PC yield was ca. 40%. As the pressure was elevated, the PC
yield increased drastically. However, when the CO2 pressure was
above 2.5 MPa, the productivity showed no significant improve-
ment. Likewise, it has been found that the catalytic performance
depended on the catalytic amounts (Fig. 6B). In the absence of any
catalyst, the blank test indicated that the PC yield was ca. 5%. After
the addition of 10ZnCl2/mp-C3N4 catalyst with 0.1 g, the PC yield
reached 56.3%. Further increasing the catalytic amount, the PC yield
increased progressively while leveled off as the catalyst weight was
over 0.2 g.
In addition to the catalytic activity, recyclability is also a very
important criterion to evaluate a heterogeneous catalyst. Regard-
ing this issue, the spent 10ZnCl2/mp-C3N4 catalysts was then rinsed
by ethanol for several times, dried overnight, and reused directly
for another run. The corresponding catalytic results are listed in
Table 4. The PO conversion acquired within the first recycle (i.e. the
second run) was ca. 60%, appreciably lower than the value obtained
in its first run. In the case of the product distribution, the selec-
tivity to PC was higher than 99%. As mentioned above, mp-C3N4
materials were prepared using cyanamide as a precursor. The trans-
formation from cyanamide to the final g-C3N4 underwent a series of
co-condensation steps, and therefore there existed a large amount
of uncondensed N-containing fragments [6,47], which thereafter
reacted with metal halides after the loading. However, due to the
intrinsically unstable disadvantage of the N-containing fragments
in mp-C3N4, a small proportion of such species would inevitably
leach out and diffuse into the liquid phase during the reactions,
resulting in the noticeable decline in the recycle test. Despite the
Besides reaction time and temperatures, we also investigated
the influence of CO2 pressure and catalytic amount on the cat-

82
J. Xu et al. / Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
Fig. 6. Influence of CO2 pressure and catalyst amount on the catalytic performance of 10ZnCl2/mp-C3N4. Reaction conditions: VPO = 7 mL, VDMF = 3 mL, t = 6 h, and T = 140 ◦C.
Table 5
Mn+
II
Mn+
Catalytic performances of various mp-C3N4-supporting catalystsa.
X-
O
O
Entry
Catalyst
Conv. (%)
Sel. (%)
Yield (%)
1
2
3
4
5
6
ZnCl2/mp-C3N4
ZnBr2/mp-C3N4
CoCl2/mp-C3N4
FeCl3/mp-C3N4
NiCl2/mp-C3N4
MgCl2/mp-C3N4
73.0
99.2
70.4
77.2
63.7
71.7
99.6
99.7
99.6
99.8
99.6
99.5
72.7
98.9
70.1
77.0
63.4
71.3
III
H3C
H3C
N
CO2
I
N
N
CO2*
O C O*
N
N
N
O
H2N
a
The molar amounts of halide compounds were all 0.73 mmol. Reaction condi-
NH2
N
N
N
N
tions: VPO = 7 mL, VDMF = 3 mL, pCO2 = 2.5 MPa, T = 140 ◦C, t = 6 h, and Wcatal. = 0.2 g.
O
N
N
N
N
H3C
X
N
N
N
N
N
N
loss of catalytic activity in the first recycle, the following two
catalytic tests, demonstrated highly stable PO conversions. This
phenomenon evidenced that the ZnCl2/mp-C3N4, subjected with
the initial run under high temperature and pressure, would become
chemically and physically stable.
H2N
N
N
N
N
N
N
IV
O
C
O
C
V
As revealed above, the PO conversion obtained over the pristine
mp-C3N4 material was only 10%; by contrast, the mp-C3N4-
supported ZnCl2 samples could catalyze the cycloaddition reactions
of CO2 with much higher yields. Undoubtedly, in such catalytic sys-
tem, mp-C3N4 mainly acted as a catalytic support whereas ZnCl2
was the catalytically active compound. In this study, besides ZnCl2,
we have also prepared a series of transition metal halide cata-
lysts supported on mp-C3N4 and submitted them to cycloaddition
reactions of CO2. It should be noted that the loading amounts
of these metal halides were fixed as 0.73 mmol, the same to the
10ZnCl2/mp-C3N4 catalyst. As summarized in Table 5, the sup-
ported ZnBr2 catalyst (entry 2) displayed superior activity to the
supported ZnCl2 one, affording a 99% PC yield. Other supported
metal halides, including CoCl2, FeCl3, NiCl2, and MgCl2 (entries 3–6)
also exhibited notably high catalytic activity (>60%). Obviously, the
supported metal halides play a crucial role in promoting the cat-
alytic cycloaddition.
O
O
O
O
H3C
X
H3C
Scheme 1. A possible mechanism of cycloaddition of CO2 with PO catalyzed by
metal halide supported on mp-C3N4.
Especially, the high surface area and accessible mesopores enabled
mp-C3N4 to expose much more defect amines on its surface. With
this unique surface environment of mp-C3N4, and DMF as a Lewis
base [45], CO2 molecules were adsorbed by the N atoms through an
acid–base interaction (step I) and transformed into activated CO2
species. On the other hand, the metal cation coordinated with oxy-
gen atom of PO and led to the polarization of C O bonds (step II).
After that, the halogen anion attacked a carbon atom of PO, resulting
in a ring opening and generating a haloalkoxy anion (step III). The
subsequent haloalkoxy anion attacked the activated CO2 molecule
located on the surface of mp-C3N4 and generated a linear halocar-
bonate (step IV). Subsequently, the halocarbonate transformed into
PC via a ring closing (step V), and meanwhile yielded the halogen
anion. In this sense, it is reasonable to explain the higher catalytic
activity obtained in entry 2 of Table 5 than that of entry 1, as the
According to the previously published work [1,29,48,49], a pos-
sible reaction mechanism for cycloaddition of CO2 catalyzed by
metal halide supported on mp-C3N4 was proposed, (Scheme 1).
First, due to its insufficient polycondensation, there existed
abundant uncondensed species, in the forms of primary and/or
secondary amines, at the edges of graphitic sheets of mp-C3N4.

J. Xu et al. / Journal of Molecular Catalysis A: Chemical 403 (2015) 77–83
83
leaving ability of Br− was higher than Cl−, thereby much more easily
actuating the ring-closing reaction [50].
[14] W.-L. Dai, L. Chen, S.-F. Yin, S.-L. Luo, C.-T. Au, Catal. Lett. 135 (2010) 295–304.
[15] B.M. Bhanage, S.-I. Fujita, Y. He, Y. Ikushima, M. Shirai, K. Torii, M. Arai, Catal.
Lett. 83 (2002) 137–141.
[16] B.M. Bhanage, S. Fujita, Y. Ikushima, K. Torii, M. Arai, Green Chem. 5 (2002)
71–75.
[17] K. Yamaguchi, K. Ebitani, T. Yoshida, H. Yoshida, K. Kaneda, J. Am. Chem. Soc.
121 (1999) 4526–4527.
4. Conclusion
In summary, mesoporous g-C3N4 was employed as a cat-
alytic support to load ZnCl2. The synthesized ZnCl2/mp-C3N4
demonstrated superior catalytic activity to the bare mp-C3N4 in
the cycloaddition of PO to PC. The highest PC yield was 72.6%
obtained over 10ZnCl2/mp-C3N4. In addition, other metal halide
catalysts supported on mp-C3N4 exhibited high PO conversion
(>63%). In the proposed mechanism, mp-C3N4 having high sur-
face area and mesostructures mainly served as a solid base to
adsorb CO2 while metal halide was a key formulation to activate
PO molecule via polarizing O atom and opening its ring by metal
cation and halide anion, respectively. Although the catalytic activi-
ties received over the supported metal halide catalysts were lower
than those obtained over the pure ones, the utilization of mp-C3N4
solid enabled metal halides to be heterogeneous catalysts, thus
rendering convenience in catalyst–product separation. It is antic-
ipated that mp-C3N4 would be widely applied in heterogeneous
organocatalysis involving CO2 conversion.
[18] W.-L. Dai, S.-F. Yin, R. Guo, S.-L. Luo, X. Du, C.-T. Au, Catal. Lett. 136 (2010)
35–44.
[19] J. Zhu, P. Xiao, H. Li, S.A.C. Carabineiro, ACS Appl. Mater. Inter. 6 (2014)
16449–16465.
[20] F. Su, S.C. Mathew, L. Möhlmann, M. Antonietti, X. Wang, S. Blechert, Angew.
Chem. Int. Ed. 50 (2011) 657–660.
[21] P. Zhang, Y. Gong, H. Li, Z. Chen, Y. Wang, Nature Commun. 4 (2013) 1593.
[22] Y. Zheng, J. Liu, J. Liang, M. Jaroniec, S.Z. Qiao, Energy Environ. Sci. 5 (2012)
6717–6731.
[23] M. Kim, S. Hwang, J.-S. Yu, J. Mater. Chem. 17 (2007) 1656–1659.
[24] S.S. Park, S.-W. Chu, C. Xue, D. Zhao, C.-S. Ha, J. Mater. Chem. 21 (2011)
10801–10807.
[25] F. Goettmann, A. Fischer, M. Antonietti, A. Thomas, Angew. Chem. Int. Ed. 45
(2006) 4467–4471.
[26] 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) 7884–7887.
[27] Z. Long, Y. Zhou, G. Chen, W. Ge, J. Wang, Sci. Rep. 4 (2014) 3651.
[28] Y. Wang, J. Yao, H. Li, D. Su, M. Antonietti, J. Am. Chem. Soc. 133 (2011)
2362–2365.
[29] Q. Su, J. Sun, J. Wang, Z. Yang, W. Cheng, S. Zhang, Catal. Sci. Technol. 4 (2014)
1136–1138.
[30] M.B. Ansari, B.-H. Min, Y.-H. Mo, S.-E. Park, Green Chem. 13 (2011) 1416–1421.
[31] M.B. Ansari, H. Jin, S.-E. Park, Catal. Sci. Technol. 3 (2013) 1261–1266.
[32] Z. Huang, F. Li, B. Chen, T. Lu, Y. Yuan, G. Yuan, Appl. Catal. B 136 (2013)
269–277.
Acknowledgments
[33] F. Su, M. Antonietti, X. Wang, Catal. Sci. Technol. 2 (2012) 1005–1009.
[34] J. Xu, T. Chen, Q. Jiang, Y.-X. Li, Chem. -Asian J. 9 (2014) 3269–3277.
[35] J. Zhu, Y. Wei, W. Chen, Z. Zhao, A. Thomas, Chem. Commun. 46 (2010)
6867–6965.
[36] X. Wang, X. Chen, A. Thomas, X. Fu, M. Antonietti, Adv. Mater. 21 (2009)
1609–1612.
This work was supported by National Natural Science Foun-
dation of China (21203014), Jiangsu Key Laboratory of Advanced
Catalytic Materials and Technology (BM2012110), and the Project
Funded by the Priority Academic Program Development of Jiangsu
Higher Education Institutions.
[37] A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J.-O. Müller, R. Schlögl, J.M.
Carlsson, J. Mater. Chem. 18 (2008) 4893–4908.
[38] J. Xu, H.-T. Wu, X. Wang, B. Xue, Y.-X. Li, Y. Cao, Phys. Chem. Chem. Phys. 15
(2013) 4510–4517.
[39] M.J. Bojdys, J.-O. Müller, M. Antonietti, A. Thomas, Chem. -Eur. J. 14 (2008)
8177–8182.
[40] J. Liu, T. Zhang, Z. Wang, G. Dawson, W. Chen, J. Mater. Chem. 21 (2011)
14398–14401.
[41] B. Yue, Q. Li, H. Iwai, T. Kako, J. Ye, Sci. Technol. Adv. Mater. 12 (2011) 034401.
[42] J. Xu, Q. Jiang, T. Chen, F. Wu, Y.-X. Li, Catal. Sci. Technol. 5 (2015) 1504–1513.
[43] S.A. Schmidt, N. Kumar, A. Shchukarev, K. Eränen, J.-P. Mikkola, D.Y. Murzin, T.
Salmi, Appl. Catal. A 468 (2013) 120–134.
[44] S.K. Pillai, S. Hamoudi, K. Belkacemi, Appl. Catal. A 455 (2013) 155–163.
[45] S. Zhong, L. Liang, B. Liu, J. Sun, J. CO2 Util. 6 (2014) 75–79.
[46] L. Han, H. Li, S.-J. Choi, M.-S. Park, S.-M. Lee, Y.-J. Kim, D.-W. Park, Appl. Catal.
A 429 (2012) 67–72.
[47] K.K. Han, C.C. Wang, Y.Y. Li, M.M. Wan, Y. Wang, J.H. Zhu, RSC Adv. 3 (2013)
9465–9469.
[48] J. Sun, W. Cheng, Z. Yang, J. Wang, T. Xu, J. Xin, S. Zhang, Green Chem. 16
(2014) 3071–3078.
[49] J. Sun, J. Wang, W. Cheng, J. Zhang, X. Li, S. Zhang, Y. She, Green Chem. 14
(2012) 654–660.
References
[1] S.-S. Wu, X.-W. Zhang, W.-L. Dai, S.-F. Yin, W.-S. Li, Y.-Q. Ren, C.-T. Au, Appl.
Catal. A 341 (2008) 106–111.
[2] C.-X. Miao, J.-Q. Wang, Y. Wu, Y. Du, L.-N. He, ChemSusChem 1 (2008)
236–241.
[3] M. North, R. Pasquale, C. Young, Green Chem. 12 (2010) 1514–1539.
[4] K.M.K. Yu, I. Curcic, J. Gabriel, S.C.E. Tsang, ChemSusChem 1 (2008) 893–899.
[5] W.-L. Dai, B. Jin, S.-L. Luo, S.-F. Yin, X.-B. Luo, C.-T. Au, J. CO2 Util. 3–4 (2013)
7–13.
[6] J. Xu, F. Wu, Q. Jiang, Y.-X. Li, Catal. Sci. Technol. 5 (2015) 447–454.
[7] X.-B. Lu, Y.-J. Zhang, B. Liang, X. Li, H. Wang, J. Mol. Catal. A 210 (2004) 31–34.
[8] R. Srivastava, D. Srinivas, P. Ratnasamy, J. Catal. 233 (2005) 1–15.
[9] J.-Q. Wang, J. Sun, W.-G. Cheng, C.-Y. Shi, K. Dong, X.-P. Zhang, S.-J. Zhang,
Catal. Sci. Technol. 2 (2012) 600–605.
[10] J. Sun, L. Wang, S. Zhang, Z. Li, X. Zhang, W. Dai, R. Mori, J. Mol. Catal. A 256
(2006) 295–300.
[11] W.-L. Dai, S.-L. Luo, S.-F. Yin, C.-T. Au, Appl. Catal. A 366 (2009) 2–12.
[12] L.-F. Xiao, F.-W. Li, J.-J. Peng, C.-G. Xia, J. Mol. Catal. A 253 (2006) 265–269.
[13] X. Zhang, D. Wang, N. Zhao, A.S. Al-Arifi, T. Aouak, Z.A. Al-Othman, W. Wei, Y.
Sun, Catal. Commun. 11 (2009) 43–46.
[50] Y. Zhou, S. Hu, X. Ma, S. Liang, T. Jiang, B. Han, J. Mol. Catal. A 284 (2008) 52–57.