Heterometallic metal-organic framework-templated synthesis of porous Co3O4/ZnO nanocage catalysts for the carbonylation of glycerol

Article abstract of DOI:10.1016/j.jssc.2017.08.035

The efficient synthesis of glycerol carbonate (GLC) has recently received great attention due to its significance in reducing excess glycerol in biodiesel production as well as its promising applications in several industrial fields. However, the achievement of high conversion and high selectivity of GLC from glycerol in heterogeneous catalytic processes remains a challenge due to the absence of high-performance solid catalysts. Herein, highly porous nanocage catalysts composed of well-mixed Co₃O₄ and ZnO nanocrystals were successfully fabricated via a facile heterometallic metal-organic framework (MOF)-templated synthetic route. Benefiting from a high porosity and the synergistic effect between Co₃O₄ and ZnO, the as-prepared composite catalysts exhibited a significantly enhanced production efficiency of GLC in the carbonylation reaction of glycerol with urea compared to the single-component counterparts. The yield of GLC over the Co₅₀Zn₅₀-350 catalyst reached 85.2%, with 93.3% conversion and near 91% GLC selectivity, and this catalytic performance was superior to that over most heterogeneous catalysts. More importantly, the proposed templated synthetic strategy of heterometallic MOFs facilitates the regulation of catalyst composition and surface structure and can therefore be potentially extended in the tailoring of other metal oxide composite catalysts.

Full text of DOI:10.1016/j.jssc.2017.08.035

Journal of Solid State Chemistry 256 (2017) 93–100
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Heterometallic metal-organic framework-templated synthesis of porous
Co3O4/ZnO nanocage catalysts for the carbonylation of glycerol
Yinyun Lüa,b, Yating Jianga, Qi Zhoua, Yunmei Lia, Luning Chena, Qin Kuanga,⁎, Zhaoxiong Xiea,
Lansun Zhenga
a
State Key Laboratory for Physical Chemistry of Solid Surfaces & Department of Chemistry, College of Chemistry and Chemical Engineering, Xiamen
University, Xiamen 361005, China
b
National Demonstration Center for Experimental Chemistry Education, Xiamen University, Xiamen 361005, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
The efficient synthesis of glycerol carbonate (GLC) has recently received great attention due to its significance in
reducing excess glycerol in biodiesel production as well as its promising applications in several industrial fields.
However, the achievement of high conversion and high selectivity of GLC from glycerol in heterogeneous
catalytic processes remains a challenge due to the absence of high-performance solid catalysts. Herein, highly
porous nanocage catalysts composed of well-mixed Co3O4 and ZnO nanocrystals were successfully fabricated via
a facile heterometallic metal-organic framework (MOF)-templated synthetic route. Benefiting from a high
porosity and the synergistic effect between Co3O4 and ZnO, the as-prepared composite catalysts exhibited a
significantly enhanced production efficiency of GLC in the carbonylation reaction of glycerol with urea
compared to the single-component counterparts. The yield of GLC over the Co50Zn50-350 catalyst reached
85.2%, with 93.3% conversion and near 91% GLC selectivity, and this catalytic performance was superior to that
over most heterogeneous catalysts. More importantly, the proposed templated synthetic strategy of hetero-
metallic MOFs facilitates the regulation of catalyst composition and surface structure and can therefore be
potentially extended in the tailoring of other metal oxide composite catalysts.
Metal-organic frameworks
Self-sacrificial template synthesis
Porous structure
Co3O4/ZnO mixtures
Glycerol carbonate
1. Introduction
thus a large number of heterojunctions are formed between them,
which potentially improve the performance of the as-formed composite
The controlled synthesis of inorganic functional composites is
always a hot issue in the field of nanomaterials because of their wide
application in rechargeable battery, supercapacitor, catalysis, photo-
catalysis, etc [1–4]. Despite the great progress made recently, it
remains a great challenge to facilely control the composition and
structure of composite structure. In recent years, metal-organic frame-
works (MOFs) have been frequently used as sacrificial templates in the
fabrication of various porous nanostructures [5–8]. As is well known,
many MOFs are isomorphic (for example, ZIF-8 and ZIF-67), and
therefore the metal ions constituting a given MOF can be readily
substituted by a wide variety of heterogeneous metal ions [9,10]. In
addition to their optimized properties [11–13], such heterometallic
MOFs can be potentially exploited as a universal platform for the
designed fabrication of porous composite nanomaterials of controlled
composition [14–16]. The biggest advantage of this synthetic strategy
is that the heterogeneous nanoparticles produced via the decomposi-
tion of heterometallic MOFs are homogeneously mixed together, and
materials in many important applications including heterogeneous
catalysis. However, the research on the application of this synthetic
strategy is still insufficient to date.
Among various catalysis reactions, with the development of large-
scale biodiesel as a renewable energy source, effective utilization of the
by-product glycerol is increasingly becoming an urgent economic and
environmental issue. As such, considerable research has been focused
on the exploration of green processes for the efficient conversion of
glycerol to high value-added chemicals [17–21]. Glycerol carbonate
(GLC) is one of the most valuable reaction intermediates among the
various glycerol derivatives given its further conversion to important
chemical products [22–26]. To date, a variety of catalysis routes of
converting glycerol into GLC have been developed, including transes-
terification with acyclic organic carbonates [27,28], carbodeoxygena-
tion with CO or CO2 [29–31], and glycerolysis with urea [32–36].
However, despite some progress, the conversion and selectivity of
glycerol into GLC remains unsatisfactory largely due to the lack of high
⁎
Corresponding author.
E-mail address: qkuang@xmu.edu.cn (Q. Kuang).
http://dx.doi.org/10.1016/j.jssc.2017.08.035
Received 10 June 2017; Received in revised form 6 August 2017; Accepted 27 August 2017
Available online 30 August 2017
0022-4596/ © 2017 Elsevier Inc. All rights reserved.

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
Table 1
performance catalysts. In the above heterogeneous catalysis processes,
the production efficiency of GLC has proven to be closely associated
with the acid-base property of solid catalysts and their surface area
[28,31]. Therefore, the catalytic performance of catalysts would be
remarkably enhanced by the rational regulation of their composition
and surface structure. As a result, various strategies based on pore
formation [37] or on the combination of metal oxides with different
Lewis acid/base properties [33,38,39] have been recently developed,
proving success in the carbonylation of glycerol. Therefore, it is
anticipated that the joint application of pore formation and metal
oxide combinations would potentially enable a more significant im-
provement to the performance of catalysts. Unfortunately, the con-
The feeding amounts of Co(NO3)2·6H2O (X) and Zn(NO3)2·6H2O (Y) in the preparation
of MOF precursorsa.
No.
MOF
precursors
feeding amounts of metal salts
Co:Zn
X
Y
1
2
3
4
5
ZIF-Co
100:0
75:25
50:50
25:75
0:100
0.450 g (1.55 mmol)
0.338 g (1.17 mmol)
0.225 g (0.78 mmol)
0.112 g (0.39 mmol)
0
0
ZIF-Co75Zn25
ZIF-Co50Zn50
ZIF-Co25Zn75
ZIF-Zn
0.116 g (0.39 mmol)
0.232 g (0.78 mmol)
0.348 g (1.17 mmol)
0.464 g (1.56 mmol)
a
2-methylimidazole was used as ligand for the synthesis of these Co/Zn-MOFs.
Fig. 1. Low-magnification SEM images of rhombic dodecahedral Co/Zn MOFs: (a) ZIF-Co, (b) ZIF-Co75Zn25, (c) ZIF-Co50Zn50, (d) ZIF-Co25Zn75, and (e) ZIF-Zn; (f) XRD patterns of
rhombic dodecahedral Co/Zn-MOFs.
94

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
Table 2
in a ceramic crucible, heated to 350 °C at a heating rate of 5 °C min–1
and maintained for 3 h under air.
Energy-dispersive X-ray data of different heterometallic MOF precursors.
Sample
C
N
Co
Zn
Co/Zn
2.3. Material characterization
ZIF-Co75Zn25
ZIF-Co50Zn50
ZIF-Co25Zn75
70.76%
75.66%
73.76%
25.70%
21.68%
22.87%
3.10%
2.16%
1.77%
0.44%
0.50%
1.60%
7:1
4.3:1
1.1:1
The morphology of the products was observed by scanning electron
microscope (SEM, Hitachi S4800) and transmission electron micro-
scope (TEM, JEOL 2010F) at an accelerating voltage of 200 kV. The
composition of products was identified by X-ray powder diffraction
(XRD) on a Rigaku D/max X-ray diffractometer (CuKα radiation, λ =
0.15418 nm). The high-angle annular dark-field scanning transmission
electron microscopy (HAADF-STEM) and the energy-dispersive X-ray
(EDX) spectroscopy and elemental mapping were performed with an
FEI TECNAI F30 microscope operated at 300 kV. The thermogravi-
metric (TG) and derivative thermogravimetric analysis (DTG) curves of
Co/Zn-MOF precursors were measured using an SDT Q600 thermo-
gravimetric analyzer under air atmosphere within a temperature range
of 30 °C to 700 °C, at a heating rate of 10 °C min−1. Nitrogen adsorp-
tion–desorption isotherms of the products were measured at 77 K on a
Tristar II 3020 sorption analyzer (Micromeritics). The specific surface
area was determined by the Brunauer–Emmett–Teller (BET) method,
and the pore size distribution was calculated by the Barrett–Joyner–
Halenda (BJH) method.
Fig. 2. TG curves of Co/Zn-MOF precursors under an air atmosphere. The dotted line
presents DTG curve of the representative precursor ZIF-Co50Zn50
.
2.4. Catalyst testing
struction of well-mixed composite catalysts whilst maintaining a highly
porous structure is extremely difficult by conventional synthetic
methods.
The synthesis of GLC was performed as follows: 2.4 g of glycerol
(0.025 mol) was added to a round bottom flask (50 mL) and treated at
150 °C under a flow of dry nitrogen for 20 min, followed by addition of
excessive amounts of urea (1.5 g, 0.025 mol). Once urea was totally
dissolved, the catalyst (130 mg, 6 wt%) was added and the resulting
mixture was further stirred. During the reaction, dry nitrogen was
continuously bubbled into the reaction mixture to remove the formed
NH3. After 3 h, ethanol was added and the catalyst was separated by
filtration. The products in the flask were analyzed by gas chromato-
Herein, we propose a facile but effective route to fabricate highly
porous mixed metal oxide composite catalysts for the synthesis of GLC
through the thermolysis of heterometallic MOFs at an appropriate
temperature. In the present study, Co/Zn mixed MOFs (Co/Zn-MOFs)
were used as self-sacrificial templates to successfully fabricate highly
porous rhombic dodecahedral nanocages built with homogeneously
mixed Co3O4 and ZnO nanocrystals. As expected, the as-prepared
porous Co3O4/ZnO nanocage catalysts exhibited significantly enhanced
catalytic performance in the carbonylation of glycerol compared to the
porous counterparts prepared from single-metal MOFs. Particularly,
the yield of GLC over the Co50Zn50-350 catalyst reached 85.2%, with
93.3% conversion and near 91% GLC selectivity, and this catalytic
performance was superior to that over most heterogeneous catalysts.
graphy-mass spectrometry using
(Shimadzu, QP 2010 Plus) equipped with a 30-m long 5% phenyl
methyl siloxane capillary column and a flame ionization detector.
a RTX-5 gas chromatograph
3. Results and discussion
3.1. Template-controlled synthesis of rhombic dodecahedral Co3O4/
ZnO nanocages
2. Experimental section
2.1. Chemicals and materials
The porous rhombic dodecahedral Co3O4/ZnO nanocages were
prepared via a two-step route involving the fabrication of Co/Zn-
MOF precursors and a subsequent calcination treatment at 350 °C in
air [40]. In order to better understand the role of the two metal oxide
catalysts and their heterojunctions in the catalysis reaction, we
adjusted the ratios of Co3O4 relative to ZnO in a wide range by using
heterometallic MOFs with different Co/Zn compositions as precursors.
For convenience, the as-prepared precursors were labelled as ZIF-Co,
ZIF-Co75Zn25, ZIF-Co50Zn50, ZIF-Co25Zn75, and ZIF-Zn according to
the input ratios of the metal ions. Accordingly, the calcined products
were labelled as Co-350, Co75Zn25-350, Co50Zn50-350, Co25Zn75-350,
and Zn-350. Fig. 1a–e shows representative SEM images of the as-
prepared Co/Zn-MOF precursors at different molar ratios of Co2+ to
Zn2+, i.e. ZIF-Co, ZIF-Co75Zn25, ZIF-Co50Zn50, ZIF-Co25Zn75, and ZIF-
Zn. All of the precursors were rhombic dodecahedral crystals, with
decreasing size on increasing Zn2+ concentration in the precursor
solution – ZIF-Co crystals were 1.5 0.2 µm, while ZIF-Zn crystals
were 250 50 nm. The addition of Zinc(II) salt results in accelerating
the nucleation process of nanocrystals, and more nucleation centers
allows for growing smaller nanoparticles. A similar trend has been
reported in a previous study [41].
Cobalt nitrate hexahydrate (Co(NO3)2·6H2O, 99%), zinc nitrate
hexahydrate (Zn(NO3)2·6H2O, 99%), glycerine (99%), urea (99%),
methanol (99.9%), and ethanol (99.7%) were purchased from
Sinopharm Chemical Regent Co., Ltd., Shanghai, China and 2-methy-
limidazole (mIM, 99%) was purchased from Alfa Aesar, Lancashire,
United Kingdom. Dry N2 (99.9%) was purchased from Linde Industrial
Gases Co., Ltd, Shanghai, China. All reagents were used without further
purification.
2.2. Synthesis of porous rhombic dodecahedral Co3O4/ZnO
nanocages
The rhombic dodecahedral Co/Zn-MOF precursors were prepared
as follows: Co(NO3)2·6H2O (X) and Zn(NO3)2·6H2O (Y) at different
molar ratios were dissolved in 3 mL of methanol (Table 1). The
solution containing Co2+ and Zn2+ ions was mixed with 20 mL of a
methanol solution of mIM (3.4 mol dm−3) and stirred for 6 h at room
temperature. The produced purple precipitates were collected by
centrifugation, washed with alcohol three times, and dried at 80 °C
for 24 h. Finally, the as-prepared Co/Zn-MOF precursors were loaded
The powder XRD patterns (Fig. 1f) indicate that the three Co/Zn-
95

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
Fig. 3. SEM images (insets show high magnification image) of (a) Co-350 (b) Co75Zn25-350, and (c) Co50Zn50-350. (d) Co25Zn75-350 and (e) Zn-350 samples obtained after calcination
at 350 °C under air for 3 h. (f) XRD patterns of the five porous products prepared at 350 °C.
MOF precursors exhibited the same XRD patterns as pure ZIF-8 (ZIF-
Zn) and pure ZIF-67 (ZIF-Co), indicating that all the samples were
composed of pure-phase crystals and had the same crystal structure.
Interestingly, the color of ZIF-Co75Zn25, ZIF-Co50Zn50, and ZIF-
Co25Zn75 varied from deep purple, to light purple, to light yellow,
reflecting a change in the Zn or Co content in the three Co/Zn-MOF
precursors (insets of Fig. 1a–e). The exact Co2+ and Zn2+ ion content in
the Co/Zn-MOF crystals was assessed by energy-dispersive X-ray
note, the true Co/Zn ratios for the three Co/Zn-MOF crystals are all
higher than the raw ratios in the precursor solutions, indicating that
the mIM ligand prefers to coordinate with Co2+ over Zn2+ in the given
synthetic conditions.
The conversion from Co/Zn-MOF precursors into porous Co3O4/
ZnO nanocages was achieved via thermal decomposition under an air
atmosphere. To determine the appropriate decomposition tempera-
ture, TG curves of Co/Zn-MOF precursors were measured using pure
ZIF-8 and ZIF-67 as references. As shown in Fig. 2, all Co/Zn-MOF
precursors, as well as pure ZIF-8 and ZIF-67, were highly stable below
300 °C, with a weight loss of less than 3% corresponding to the removal
analysis (Table 2). As shown, Co and Zn coexist in the ZIF-Co75Zn25
,
ZIF-Co50Zn50, and ZIF-Co25Zn75 composites and the Co/Zn ratios
decrease with the increasing input of Zn(NO3)2 in the reaction. Of
96

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
precursors are essentially porous, hollow architectures built with
homogeneously mixed Co3O4 and ZnO nanocrystals, i.e., porous
Co3O4/ZnO nanocages.
The unique nanocage structural feature presented by the porous
Co3O4/ZnO materials was confirmed by TEM. As shown in the low-
magnification TEM images (Fig. 5a1, b1, c1), the rhombic dodecahedral
particles from all Co/Zn-MOF precursor calcination products were
composed of primary crystals of 5–10 nm in size. Given the uniform
distribution of Co3O4 and ZnO crystals and their small size, hetero-
interfaces are bound to frequently appear between two crystallites.
However, it was difficult to directly distinguish Co3O4 from ZnO in the
nanocages in low-magnification TEM images since Co3O4 and ZnO
have an extremely similar electron density and contrast. Thus, high-
resolution TEM analysis of lattice fringe spacings enabled identification
of the ZnO and Co3O4 phases. Indeed, a large amount of hetero-
interfaces formed by Co3O4 and ZnO were observed (Fig. 5a2, b2, c2).
The selected area electronic diffraction (SAED) patterns recorded from
the whole porous rhombic dodecahedral nanocages (Fig. 5a3, b3, c3)
presented a series of concentric rings, including both Co3O4 and ZnO
patterns, indicating that the primary crystallites of Co3O4 and ZnO are
randomly attached with each other.
Fig. 4. (a) The EDX spectrum of Co75Zn25-350 rhombic dodecahedra. (b) HAADF-
STEM image and element mapping images of an individual Co75Zn25-350 rhombic
dodecahedral nanocage.
To evaluate the specific surface area and the porosity, the N2
adsorption–desorption isotherms of the as-synthesized porous nano-
cages with different compositions were further performed (Fig. 6). The
specific surface areas of Co-350, Co75Zn25-350, Co50Zn50-350,
Co25Zn75-350 and Zn-350 were calculated by the BET method to be
59.0, 102.9, 100.2, 70.6, and 283.3 m2·g−1, respectively. Moreover, all
of the samples displayed a typical type IV adsorption isotherm with a
H3-type hysteresis loop at a relative pressure of 0.4–1.0, indicating the
presence of the mesoporous structure and the synthesized Co3O4/ZnO
nanocages possessed approximate pore size [7]. This deduction was
further demonstrated by the results of pore size distribution measure-
ments. The pore size of Co-350, Co75Zn25-350, Co50Zn50-350,
Co25Zn75-350 and Zn-350 were calculated by the BJH method to be
9.7, 7.2, 6.2, 8.1, and 8.4 nm, respectively. The pore size distribution
results are well consistent with TEM observations (Fig. 5).
of guest molecules in the MOF pores. Upon increasing temperature, all
the samples began to decompose, resulting in a sharp drop of their
weight at approximately 350–400 °C, where an obvious exothermic
peak can be seen in the corresponding DTG curve (the dotted dark
yellow curve in Fig. 2). Moreover the starting decomposition tempera-
ture being closely related to their Co/Zn ratio despite their similar
structures. The three Co/Zn-MOF precursors began to decompose at
approximately 350–400 °C, which is higher than the starting decom-
position temperature of pure Co-MOFs (ZIF-67) but lower than that of
pure Zn-MOFs (ZIF-8). Further, the higher the Co2+ content, the lower
the starting decomposition temperature, indicating that Co2+ ions in
the framework act as catalysts to promote the decomposition of MOFs.
On the basis of the above TG results, the thermal conversion
temperature of Co/Zn-MOF precursors was determined as 350 °C.
The representative SEM images of the products after thermal
decomposition are shown in Fig. 3a–e. The decomposed products
apparently maintained their original rhombic dodecahedral shape, yet
with obviously concave faces. Further, a proportion of the particles
became near-transparent under the electron beam, suggesting a highly
porous and hollow architecture, much like a cage. High magnification
SEM images (insets of Fig. 3a–e) revealed the rhombic dodecahedra to
be composed of numerous primary nanocrystals. The formation of
Co3O4/ZnO hybrids with nanocage structure is greatly affected by the
calcination temperature. When the calcination temperature reaches
and exceeds the decomposition temperature of the precursor, organic
ligands and other guest molecules in the ZIF-Co/Zn rhombic dodeca-
hedral precursors would be burned out in air atmosphere and released
in the form of CO2, H2O, and NOx, while Co(II) ions and Zn(II) ions in
the precursors react with O2 to form new porous Co3O4/ZnO frame-
works [7]. In order to assess the composition of the calcined products,
powder XRD analysis and elemental mapping were further conducted.
The XRD patterns of the calcination products from the three Co/Zn-
MOF precursors were indexed on the basis of a phase mixture
constituted by spinel-type Co3O4 (JCPDS Card No. 43-1003) and
wurtzite-type ZnO (JCPDS Card No. 5-664) (Fig. 3f). Additionally,
the intensity of the diffraction peaks assigned to the ZnO phase became
stronger with the increasing Zn(NO3)2·6H2O content. Represented by
Co75Zn25-350, the EDX spectrum (Fig. 4a) confirmed the presence of
Zn, Co, and O in the sample. The corresponding elemental mapping
(Fig. 4b) further revealed that the distribution of the three elements in
the product to be completely consistent, but the respective concentra-
tions in the central zone were distinctly lower than those in the outer
zone. The above results illustrate that the rhombic dodecahedral
particles generated via the thermal decomposition of Co/Zn-MOF
3.2. Catalysis performance of rhombic dodecahedral Co3O4/ZnO
nanocages
Previous studies have demonstrated that the production efficiency
of GLC could be significantly enhanced by depositing Co3O4 onto ZnO
particles due to a Zn–Co interaction at the interface, despite both
Co3O4 and ZnO exhibiting minimal performance in the catalysis of
carbonylation of glycerol [33,38,42]. As indicated in the TEM images,
pores and phase-interfaces were formed between Co3O4 and ZnO in the
composite nanocages. Indeed, these rich hetero-interfaces would
considerably promote the synergistic role of Co3O4 and ZnO in
catalyzing the reaction and the highly porous structure of the Co3O4/
ZnO nanocages would facilitate the diffusion of substrates in the
catalysis process. Considering these structural advantages, it is reason-
able to expect that the as-synthesized porous Co3O4/ZnO nanocages
would possess superior heterogeneous catalytic performance in the
synthesis of GLC from glycerol and urea.
Herein, the conversion of GLC from glycerol was achieved via the
glycerolysis reaction of glycerol with urea. Given the more moderate
reaction conditions, easy availability and low cost of urea, and easy
recycling of the by-product (ammonia), the carbonylation reaction of
glycerol with urea is regarded as a more plausible alternative to other
GLC conversion methods proposed [36]. The above synthetic route
involves a complicated process including the elimination of various
reactive groups (dehydration and deamination) [33]. As shown in
Scheme 1, a well-accepted mechanism for the production of GLC
includes two consecutive steps: (1) the carbamylation of glycerol to
glycerol carbamate (A), liberating a moles of ammonia, and (2) the
carbonylation of the glycerol carbamate to GLC (B), with elimination of
97

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
Fig. 5. Low magnification TEM, high resolution TEM and SAED images of (a1, a2, a3) Co75Zn25-350, (b1, b2, b3) Co50Zn50-350 and (c1, c2, c3) Co25Zn75-350. The SAED patterns were
recorded from the corresponding whole particles.
second lot of a moles of ammonia; the first step occurs at a faster
reaction rate than the second step [43]. However, the by-product, 4-
(hydroxymethyl) oxazolidin-2-one (C), may be simultaneously formed
in the second step due to the dehydration reaction of glycerol
carbamate. Additionally, the target product GLC is likely to further
react with urea, thereby producing (2-oxo-1,3-dioxolan-4-yl)methyl
carbamate (D). Clearly, in order to achieve optimal GLC selectivity, the
desired catalysts should be able to successfully suppress the dehydra-
tion of glycerol carbamate and the over-reaction of GLC with urea.
Table 3 summarizes the catalytic performance of three porous
catalysts, the selectivity for the target product B (i.e., GLC) was only
57.5%. The two main by-products were the intermediate product
glycerol carbamate (A) and its dehydrated product 4-(hydroxymethyl)
oxazolidin-2-one (C), accounting for 12.8% and 20.9% of the final
product, respectively. However, the product distribution upon the use
of the various catalysts significantly changed compared to the blank
reaction. Using Co-350 as the catalyst, glycerol conversion increased to
86.4%, but selectivity for the target product B significantly decreased
(37.6%) and that for the intermediate product A increased to 36.5%.
Zn-350 achieved
a worse glycerol conversion (55.5%) and GLC
mixed-oxide nanocages (Co75Zn25-350, Co50Zn50-350, and Co25Zn75
350) and two pure-oxide nanocages (Co-350 and Zn-350) in the
carbonylation reaction of glycerol with urea. Without the use of
-
selectivity (6.8%), with the main products being the dehydrated
product C (53.8% of selectivity) and the over-reaction product D
(16.8% of selectivity). The above results indicate that single Co3O4 or
98

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
ZnO catalysts performed badly in the reaction. Interestingly, the
combination of the two oxides led to remarkable enhancement in
catalytic performance. In stark contrast to the two single oxide
catalysts, the selectivity of the target product B over Co75Zn25-350,
Co50Zn50-350, and Co25Zn75-350 reached 78–91%, which is more than
two- and 10-fold that for the Co-350 and Zn-350, respectively,
although the increase in glycerol conversion was not as dramatic. Of
note, the intermediate product A and the dehydrated by-product C
were suppressed to a very low degree in the three cases. Thus, the
mixed metal catalysts converted more of the intermediate product A
into the desired GLC, albeit with a decrease in the dehydration
reactions to form C. Clearly, such significant promotion in catalytic
performance is closely associated with the formation of abundant
Co3O4/ZnO hetero-interfaces. According to the conversion and selec-
tivity values using Co-350 and Zn-350, we can deduce that Co3O4 and
ZnO play distinct roles as catalysts in the reaction process. Co3O4
promotes the conversion of glycerol into the intermediate product A
and suppress the dehydration side reaction, while ZnO facilitates the
conversion of the intermediate product A. Therefore, the establishment
of Co3O4/ZnO hetero-interfaces in the porous nanocages may enable
the two catalysts to collaborate more effectively, remarkably promoting
the selectivity of the target product B in the catalysis process.
Fig. 6. N2 adsorption-desorption isotherms of the porous nanocages with different
compositions (i.e., Co-350, Co75Zn25-350, Co50Zn50-350, Co25Zn75-350, Zn-350).
Of note, the catalytic performance of the three porous Co3O4/ZnO
nanocages were somewhat different, despite all of them being en-
hanced. With the Co75Zn25-350 catalyst, glycerol conversion was
observed to be 88.9%, with 78.3% selectivity. With the Co50Zn50-350
catalyst, the conversion of glycerol increased to 93.3%, and the
selectivity for GLC increased to 91.3%. Although the conversion of
glycerol over the Co50Zn50-350 catalyst decreased to 70% after 3
recycles due to the loss in the recovery process of catalyst with
centrifugation, the selectivity still maintained ~91% as before.
Moreover, the Co25Zn75-350 catalyst with higher ZnO content exhib-
ited lower selectivity (79.0%). Thus, the results do not seem predictable
and the underlying mechanism cannot be adequately described in the
present work. However, it is certain that the catalytic performance of
these porous Co3O4/ZnO nanocages is inherently correlated to the
number of hetero-interfaces as well as the presence of the two oxides,
since Co3O4 and ZnO undergo different functions.
Scheme 1. Synthesis of glycerol carbonate from glycerol and urea.
Table 3
Conversion of glycerol and selectivity of every product over different catalysts (130 mg)
after reaction at 150 °C for 3 h.
The catalytic performance of Co50Zn50-350, which was overall the
best among the three mixed metal nanocages, was further compared
with that of other catalysts reported in the literature (Table 4). In the
homogeneous process catalyzed by Zn2+-containing catalysts (e.g.,
ZnCl2), the selectivity of GLC is usually extremely high (near 100%)
and the conversion of glycerol can reach above 80% [32,35]. Recently,
Marakatti et al. achieved 96.4% conversion and 97% selectivity to GLC
over the Zn2+-zeolite catalyst obtained by ion exchange [37]. In fact, it
is very difficult to achieve high selectivity (> 90%) while fulfilling high
conversion (> 80%) in the heterogeneous process of carbonylation of
glycerol. For example, a GLC selectivity of above 97% was achieved
over using a Co3O4 nanoparticle catalyst supported on ZnO, whereas
the conversion of glycerol was less than 70%, thereby leading to a less
desirable GLC yield [38]. In this work, the yield of GLC over the
Co50Zn50-350 reached 85.2%, which was superior to most of other
catalyst (i.e., Co3O4/ZnO-dry mixing, Co3O4/ZnO-IMP, AuPd/MgO-
SOL-110 °C, AuPd/MgO-IMP-110 °C, 30% SiW12/MCM-41) in refer-
ences [34,38,44]. The GLC selectivity over the Co50Zn50-350 reached
near 91%, comparable to most of the best catalyst [35,37,38]. And for
Co50Zn50-350, TOF was calculated to be 54.6 mmol/g/h based on the
GLC yield per gram catalyst per total reaction time. Overall, the
catalytic performance of the Co50Zn50-350 is highly adequate for the
production of GLC.
Catalyst
Conversion of
glycerol (%)
Selectivity of products (%)
A
B
C
D
Blank
Co−350
76.8
86.4
88.9
93.3
87.4
70.0
78.0
55.5
12.8
36.5
1.7
0
0
0
57.5
37.6
78.3
91.3
90.7
90.7
79.0
6.8
20.9
3.3
4.0
0
0
0
6.5
11.4
15.0
9.3
8.7
9.3
Co75Zn25-350
Co50Zn50-350 (1st)
Co50Zn50-350 (2nd)
Co50Zn50-350 (3rd)
Co25Zn75-350
Zn-350
0
0
6.7
53.8
14.3
16.8
Table 4
The catalytic performance of Co50Zn50-350 and other catalysts reported in the literature.
Catalyst
Conversion GLC selectivity GLC yield
(%)
Refs
(%)
(%)
Co50Zn50-350
Co3O4/ZnO-dry mixing
Co3O4/ZnO-IMP
AuPd/MgO-SOL-110 °C
AuPd/MgO-IMP-110 °C
Zn2+-zeolite-ion exchanged 96.4
93.3
69
29
87
86
91.3
97
> 99
77
85.2
66.9
~29
67
This work
[38]
[38]
[34]
[34]
[37]
[35]
[44]
64
55
97
99.7
77
93.5
80.2
57.8
4. Conclusions
ZnCl2
30% SiW12/MCM-41
80.4
75
Rhombic dodecahedral Co3O4/ZnO nanocages built from well-
mixed Co3O4 and ZnO nanocrystals were fabricated via a simple Co/
99

Y. Lü et al.
Journal of Solid State Chemistry 256 (2017) 93–100
[14] J. Xu, S. Liu, Y. Liu, RSC Adv. 6 (2016) 52137–52142.
[15] S.K. Rajabi, S. Sohrabnezhad, S. Ghafourian, J. Solid State Chem. 244 (2016)
160–163.
[16] S. Chen, M. Xue, Y. Li, Y. Pan, L. Zhu, D. Zhang, Q. Fang, S. Qiu, Inorg. Chem.
Front. 2 (2015) 177–183.
[17] C. Avelino, I. Sara, V. Alexandra, Chem. Rev. 107 (2017) 2411–2502.
[18] J.N. Chheda, G.W. Huber, J.A. Dumesic, Angew. Chem. Int. Ed. 46 (2007)
7164–7183.
[19] F. Jerome, Y. Pouilloux, J. Barrault, ChemSusChem 1 (2008) 586–613.
[20] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem. 10 (2008)
13–30.
[21] A. Dibenedetto, F. Nocito, A. Angelini, I. Papai, M. Aresta, R. Mancuso,
ChemSusChem 6 (2013) 345–352.
[22] G. Rokicki, P. Rakoczy, P. Parzuchowski, M. Sobiecki, Green Chem. 7 (2005)
529–539.
[23] L. Ubaghs, N. Fricke, H. Keul, H. Höcker, Macromol. Rapid Commun. 25 (2004)
517–521.
Zn-MOF-templated chemical transformation at 350 °C. Excitingly, they
exhibited remarkably enhanced conversion and selectivity in the
catalytic synthesis of GLC via the glycerolysis reaction of glycerol with
urea. The efficient catalytic activity was due to (1) the as-synthesized
nanocages possessing massive pores, which provide rich active sites
and significantly facilitate diffusion of the reactants and (2) the
establishment of Co3O4/ZnO hetero-interfaces allowing the function
of the two oxide catalysts to collaborate more effectively. It should be
noted that the Co3O4/ZnO ratio in the composite catalysts is easily
adjusted by tuning the Co/Zn ratios in the hetero-metallic MOF
precursors. Consequently, the performance of the Co3O4/ZnO nanoc-
age catalysts may still be considerably improved. More importantly, it
can be also anticipated to extend the hetero-metallic MOF-templated
synthetic strategy to the fabrication of other bimetal oxides in view of
the diversity of MOFs.
[24] C. Vieville, J.W. Yoo, S. Pelet, Z. Mouloungui, Catal. Lett. 56 (1998) 245–247.
[25] A. Mhanna, F. Sadaka, G. Boni, C. Brachais, L. Brachais, J. Couvercelle,
L. Plasseraud, L. Lecamp, J. Am. Oil Chem. Soc. 91 (2013) 337–348.
[26] H. Mei, Z. Zhong, F. Long, R. Zhuo, Macromol. Rapid Commun. 27 (2006)
1894–1899.
Acknowledgments
[27] J. Esteban, E. Domínguez, M. Ladero, F. Garcia-Ochoa, Fuel Process. Technol. 138
(2015) 243–251.
This work was supported by the National Basic Research Program
of China (2015CB932301), the National Natural Science Foundation of
China (2171163 and 21333008) and the Fundamental Research Funds
for the Central Universities (20720160026).
[28] M.J. Climent, A. Corma, P. De Frutos, S. Iborra, M. Noy, A. Velty, P. Concepción, J.
Catal. 269 (2010) 140–149.
[29] T. Coskun, C.M. Conifer, L.C. Stevenson, G.J. Britovsek, Chemistry 19 (2013)
6840–6844.
[30] J. Liu, Y. Li, J. Zhang, D. He, Appl. Catal. A 513 (2016) 9–18.
[31] H. Li, X. Jiao, L. Li, N. Zhao, F. Xiao, W. Wei, Y. Sun, B. Zhang, Catal. Sci. Technol.
5 (2015) 989–1005.
[32] S. Fujita, Y. Yamanishi, M. Arai, J. Catal. 297 (2013) 137–141.
[33] F. Rubio-Marcos, V. Calvino-Casilda, M.A. Bañares, J.F. Fernandez, J. Catal. 275
(2010) 288–293.
References
[1] N. Reddy, S. Winkler, N. Koch, N. Pinna, ACS Appl. Mater. Interfaces 8 (2016)
3226–3232.
[2] N. Reddy, K. Lee, A. Gopalan, M. Kim, A. Showkat, Y. Nho, J. Polym. Sci. A: Polym.
Chem. 44 (2006) 3355–3364.
[34] M. Hasbi Ab Rahim, Q. He, J.A. Lopez-Sanchez, C. Hammond, N. Dimitratos,
M. Sankar, A.F. Carley, C.J. Kiely, D.W. Knight, G.J. Hutchings, Catal. Sci. Technol.
2 (2012) 1914–1924.
[35] J. Park, J. Choi, S. Woo, S. Lee, M. Cheong, H. Kim, H. Lee, Appl. Catal. A 35
(2012) 433–440.
[36] A. Galadima, O. Muraza, Waste Biomass- Valoriz. 8 (2016) 141–152.
[37] V.S. Marakatti, A.B. Halgeri, RSC Adv. 5 (2015) 14286–14293.
[38] F. Rubio-Marcos, V. Calvino-Casilda, M.A. Bañares, J.F. Fernandez, ChemCatChem
5 (2013) 1431–1440.
[39] P. Manjunathan, R. Ravishankar, G. Shanbhag, ChemCatChem 8 (2016) 631–639.
[40] Y. Lü, Q. Zhou, L. Chen, W. Zhan, Z. Xie, Q. Kuang, L. Zheng, CrystEngComm 18
(2016) 4121–4126.
[3] K. Reddy, M. Hassan, V. Gomes, Appl. Catal. A: Gen. 489 (2015) 1–16.
[4] M. Cakici, K. Reddy, F. Alonso-Marroquin, Chem. Eng. J. 309 (2017) 151–158.
[5] Y. Chen, L. Hu, J. Solid State Chem. 239 (2016) 23–29.
[6] Y. Wang, Y. Lu, W. Zhan, Z. Xie, Q. Kuang, L. Zheng, J. Mater. Chem. A 3 (2015)
12796–12803.
[7] Y. Lu, W. Zhan, Y. He, Y. Wang, X. Kong, Q. Kuang, Z. Xie, L. Zheng, ACS Appl.
Mater. Interfaces 6 (2014) 4186–4195.
[8] Y. Lu, Y. Wang, H. Li, Y. Lin, Z. Jiang, Z. Xie, Q. Kuang, L. Zheng, ACS Appl. Mater.
Interfaces 7 (2015) 13604–13611.
[41] J. Zaręba, M. Nyk, M. Samoć, Cryst. Growth Des. 16 (2016) 6419–6425.
[42] P. Zhang, L. Liu, M. Fan, Y. Dong, P. Jiang, RSC Adv. 6 (2016) 76223–76230.
[43] V. Calvino-Casilda, G. Mul, J.F. Fernández, F. Rubio-Marcos, M.A. Bańares, Appl.
Catal. A 409 (2011) 106–112.
[9] Q. Shi, Z. Chen, Z. Song, J. Li, J. Dong, Angew. Chem. Int. Ed. 50 (2011) 672–675.
[10] M. Timofeeva, V. Panchenko, N. Khan, Z. Hasan, I. Prosvirin, S. Tsybulya, S. Jhung,
Appl. Catal. A 529 (2017) 167–174.
[11] D. Sun, F. Sun, X. Deng, Z. Li, Inorg. Chem. 54 (2015) 8639–8643.
[12] R. Li, X. Ren, X. Feng, X. Li, C. Hu, B. Wang, Chem. Commun. 50 (2014)
6894–6897.
[44] N. Narkhede, A. Patel, RSC Adv. 5 (2015) 52801–52808.
[13] K. Zhou, B. Mousavi, Z. Luo, P.S. hatanasri, S. Chaemchuen, F. Verpoort, J. Mater.
Chem. A 5 (2017) 952–957.
100