Preparation of mesoporous CaO-ZrO2 catalysts without template for the continuous synthesis of glycerol carbonate in a fixed-bed reactor

Article abstract of DOI:10.1016/j.apcata.2019.117313

Recently, transformation of glycerol into glycerol carbonate (GLC) has been of critical importance. Although traditional Ca-based catalysts present high activity for the GLC synthesis, their industrial applications are hindered by the leaching of Ca²⁺. Here, a series of mesoporous CaO-ZrO₂ with high surface area were obtained without template. CO₂ temperature-programmed desorption displayed that the basic properties of the catalysts can be controlled by the content of Ca dopant. Importantly, Ca²⁺ can be introduced into t-ZrO₂ lattice, resulting in the formation of a solid solution structure. Based on systematic experiments, a correlation between their basic properties and catalytic behavior was explained. Under the optimized conditions, the catalyst with a certain ratio of Ca: Zr = 0.6: 1 showed an excellent performance with a 97% glycerol conversion and a 93% selectivity of GLC. Furthermore, the solid solution structure rendered this catalyst high resistance to reactant flushing in a fixed-bed, without leaching of active sites.

Full text of DOI:10.1016/j.apcata.2019.117313

Applied Catalysis A, General 590 (2020) 117313
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Preparation of mesoporous CaO-ZrO catalysts without template for the
2
continuous synthesis of glycerol carbonate in a fixed-bed reactor
a,
a
a
a
a
a
a
Xuelan Zhang *, Shuwei Wei , Xueying Zhao , Zheng Chen , Hongwei Wu , Ping Rong , Yu Sun ,
a
b
a,
Ya Li , Hao Yu , Dengfeng Wang *
a
College of Chemistry, Chemical Engineering and Materials Science, Zaozhuang University, Zaozhuang, 277160, China
College of Chemical and Environmental Engineering, Shandong University of Science and Technology, 266590, Qingdao, China
b
A R T I C L E I N F O
A B S T R A C T
Keywords:
Recently, transformation of glycerol into glycerol carbonate (GLC) has been of critical importance. Although
Glycerol
Glycerol carbonate
CaO-ZrO
Fixed-bed
traditional Ca-based catalysts present high activity for the GLC synthesis, their industrial applications are hin-
dered by the leaching of Ca . Here, a series of mesoporous CaO-ZrO with high surface area were obtained
2
2+
2
without template. CO
2
temperature-programmed desorption displayed that the basic properties of the catalysts
2+
can be controlled by the content of Ca dopant. Importantly, Ca can be introduced into t-ZrO
2
lattice, resulting
in the formation of a solid solution structure. Based on systematic experiments, a correlation between their basic
properties and catalytic behavior was explained. Under the optimized conditions, the catalyst with a certain ratio
of Ca: Zr = 0.6: 1 showed an excellent performance with a 97% glycerol conversion and a 93% selectivity of
GLC. Furthermore, the solid solution structure rendered this catalyst high resistance to reactant flushing in a
fixed-bed, without leaching of active sites.
1. Introduction
glycerol and phosgene or CO, which had been obsoleted due to their
high-toxic and unsafe properties of phosgene or CO [7,8]. The pre-
In the past decades, special attention has been paid to biodiesel
2
paration of GLC from CO and glycerol has been studied extensively and
production owing to depletion of fossil energy and to environmental
problems. As an important renewable energy, the utilization of bio-
diesel not only solves the issue of energy-crisis, but also generates little
negative effects on the ecosystem [1]. Considering these aspects, its
production scale would be expanded rapidly in a short time. However,
we are meeting a tremendous challenge that a great deal of glycerol
is still a current interest. However, this approach only gives a minimal
GLC yield but requires harsh reaction conditions, inhibiting its further
practical employment [9–11]. To solve this problem, urea, which is
2
regarded as an activated form of CO , has been applied to the synthesis
of GLC and its analogues [6,12–14]. Although a relatively high yield of
GLC could be achieved, the synthesis of GLC from urea and glycerol
must be conducted at low pressures below 20–30 mbar in order to re-
move the gaseous ammonia. In addition, the GLC selectivity is often far
from satisfactory because of undesired side-reactions. Compared to
aforementioned processes, the synthesis of GLC via transesterification
between glycerol and dimethyl carbonate (DMC) is more promising and
facile [15–17]. Meanwhile, the co-product methanol can be easily se-
parated because of its low boiling point (around 64.7 °C). More inter-
estingly, DMC could be synthesized directly from CO and methanol,
2
which is the by-product generated in the process, making the entire
cycle efficient and sustainable (cf. Scheme 1) [18,19].
Recently, Tudorache et al. prepared an expensive enzymatic catalyst
for the reaction of DMC and glycerol, but this catalytic system was
highly sensitive to poisoning. Worse still, a relative long reaction time
(
about 100 g/kg of biodiesel) is created as an undesired product. Con-
sequently, a surplus waste of glycerol would reduce the market com-
petitiveness of biodiesel industry, and therefore, it is an urgent need to
convert glycerol into chemicals of high-added value [2]. Among the
kinds of glycerol derivatives, glycerol carbonate (GLC) has been at-
tracted much attention due to its outstanding physic-chemical features
including non-toxicity, low-flammability, high biodegradability and
water-solubility [3,4]. Moreover, it has been reported that GLC is a
chief raw material to produce other fine chemicals such as adhesive,
surfactant, polymers and elastomer. In addition, GLC is also a green
solvent with high polarity, so it has been applied broadly in paint and
electrochemistry fields [3–6]{Ochoa-Gómez, 2012 #3;Sonnati, 2013
#
4}. Conventionally, GLC was prepared from the reaction between
⁎
Corresponding authors.
E-mail addresses: mhszxl@sxicc.ac.cn (X. Zhang), bitwdf@163.com (D. Wang).
https://doi.org/10.1016/j.apcata.2019.117313
Received 17 July 2019; Received in revised form 11 October 2019; Accepted 20 October 2019
Available online 02 November 2019
0926-860X/ © 2019 Elsevier B.V. All rights reserved.

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
Scheme 1. Synthesis of GLC via transesterification between
2
glycerol and DMC: indirect utilization of CO .
was required to achieve good catalytic performance [20]. It is well
accepted that basic catalysts exhibit superior catalytic ability over
acidic samples towards GLC synthesis. Homogeneous basic catalytic
nitrate as Zr source [39]. Nevertheless, few studies investigated the
preparation of CaO-ZrO samples via a similar manner. Meanwhile, the
effect of Ca2+ ion doping on zirconia and the catalytic performance of
2
samples like NaOH and K
2
CO
3
could promote GLC synthesis efficiently,
the CaO-ZrO
unknown.
2
catalysts in the continuous GLC synthesis are remained
but they suffered from the inconvenience of separating catalysts and
products [21]. With the aim of avoiding this disadvantage, several
heterogeneous basic materials have been explored. For instance, basic
In the present work, a series of mesoporous CaO-ZrO
2
catalysts with
different contents of calcium were prepared via a co-precipitation
procedure which is free of a template. The physicochemical properties
of the prepared catalysts were characterized by XRD, BET, Raman, XPS,
catalysts containing alkali metal elements like NaOH/γ-Al
2 3
O [22],
K
2
CO [23], Li/ZnO [24] and Na-modified
3
2 3
/MgO [22], KF/γ-Al O
zeolite [25] have been used for the GLC formation. Unfortunately, some
basic species from these catalysts would be leaching into the reaction
systems, leading to a poor catalytic stability. Diverse mixed oxides such
2 2
TEM and by the CO temperature-programmed desorption (CO -TPD)
measurements. Aim at obtaining a stable Ca-modified solid base system,
we will demonstrate that these catalysts are particularly efficient for the
continuous GLC synthesis in a fixed-bed reactor, and moreover Ca
species are not leaching anymore.
2 3 2 3
as ZnO/La O [26], MgO/La O [27], Mg/Zr/Sr mixed oxides [28] and
fluorine-doped Mg/Al mixed oxides [29] were also evaluated. Despite
their high reactivity, the GLC selectivity was often suppressed owing to
severe side-reactions.
2. Experimental
In addition to these basic materials mentioned above, the Ca-based
samples have attracted a broadly attention. Ochoa-Gómez et al. found
that CaO possessed the highest activity among several alkali earth metal
oxides [21]. However, CaO was easily poisoned and deactivated by
water and carbon dioxide. Later, Li concluded that the catalytic ability
of CaO originated from the homogeneous Ca-species, which was gen-
erated due to its interaction with DMC and glycerol [30]. This con-
clusion was also ascertained by Simanjuntak and his co-workers [31].
Zheng et al. disclosed that Ca-Al-O hydrocalumite could give a 93%
conversion of glycerol. Meanwhile, Ca/Al mixed oxides derived this
precursor were also active for the GLC production [32]. Besides,
Granados-Reyes et al. suggested that Ca/Al layered double hydroxide
could be applied to catalyze the GLC manufacture [33]. In addition,
CaO-MgO was also selected as a catalyst [34]. Nevertheless, further
investigations revealed that the active CaO phase could dissolve into
the reacting glycerol, generating a calcium glycerol species [31,32].
Even worse, traditional Ca-based catalysts suffered from limited specific
area, lowering their reactivity to a certain degree [32]. Thus, the de-
velopment a Ca-based solid basic catalyst with high activity and sta-
bility for the GLC synthesis from glycerol and DMC is still highly de-
sirable and remains as a great challenge.
2
.1. Preparation of catalysts
CaO-ZrO
using metal nitrates as starting materials. Briefly, a CaO-ZrO
catalyst was used as an example to demonstrate the preparation pro-
cess: 0.7 g Ca(NO ·4H (3 mmol) and 2.15 g Zr(NO ·5H
5 mmol) were dissolved into 50 mL deionized water under stirring
2
catalytic samples were synthesized via a co-precipitation
2
(0.6:1)
3
)
2
2
O
3
)
4
2
O
(
until completely dissolved. Meanwhile, 1.04 g NaOH (26 mmol) was
dissolved into another 50 mL deionized water as precipitant.
Subsequently, the two solutions were slowly added into a 500 mL
beaker under vigorous stirring with pH of 10.0. After the titration, the
suspension was stirred for 2 h at room temperature. Then, the pre-
+
cipitate was washed with an excess of deionized water to remove Na
ions. The filter cake was refluxed in deionized water at 80 °C for 48 h.
Once the slurry cooled down to room temperature, the solid products
were collected and then dried at 80 °C for 12 h. Finally, the dried pro-
ducts were calcined at 600 °C for 7 h under a flowing steam of nitrogen.
For convenience, the mixed oxide catalysts were designated as nCaO-
ZrO
, where n presented the starting Ca2+: Zr
4+
atomic ratio in the
2
mother solution (n = 0, 0.2, 0.4, 0.6 or 0.8). For a comparison, pure
ZrO was also prepared in the similar manner without adding Ca
·4H O in the starting mixture.
On the other hand, zirconia was regarded as a promising catalytic
promoter and supporter for numerous base-catalyzed organic reactions,
benefiting from the distinct thermal stability and high surface area of
zirconia [35]. Many researchers have proved that an insertion of alkali
metal ions and alkaline earth ions into zirconia could enhance both
surficial basicity and structural stability of the resulting catalyst, which
in turn promotes their catalytic activity and reusability significantly.
2
(
NO
3
)
2
2
2.2. Samples characterization
Structural characterization of the catalysts was carried out on a
Rigaku Miniflex diffractometer with Ni-filtered Cu Kα radiation using a
For instance, mesoporous Na
the efficient manufacture of DMC and propylene glycol [36]. Liu et al.
and Varkolu et al. found that mesoporous MgO-ZrO could promote the
2 2
O-ZrO basic catalysts were acquired for
step of 5°/min (in a 2θ range of 0-8° for low-angle XRD patterns and 10-
8
0° for wide-angle XRD patterns).
X-ray photoelectron spectroscopy (XPS) experiments were per-
2
production of propylene glycol methyl ether and glycerol carbonate
with high reactivity, respectively [37,38]. In their approaches, organic
zirconium (IV) and pluronic P123 were chosen as a starting material
and a template respectively to obtain a mesoporous framework. Un-
doubtedly, this brought about a high cost by this type of catalysts
formed by a Thermo Scientific k-Alpha spectrometer with Al Kα
1486.6 eV) radiation. The binding energy of the adventitious carbon (C
s) was set at 284.6 eV for correction. Prior to the measurements, all the
(
1
samples were treated under vacuum for 3 h.
The Raman spectra were recorded on a Horiba-Jobin-Yvon spec-
trometer equipped with a He-Ne laser light source excited by a wave-
length of 632.8 nm.
[
36–38]. Coupled with the weakness that these catalysts were prepared
via a complicated sol-gel procedure, a more cheap and convenient
method should be established. Actually, we have prepared mesoporous
MgO-ZrO catalysts via simple co-precipitation method using zirconium
2
The textural properties including surface area (BET), pore volume
and pore size distribution of the catalysts were derived from N
2
2

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
adsorption-desorption technique using a Micromeritics ASAP-2020 in-
strument at −196 °C. Prior to the collection of adsorption isotherms,
catalysts were pretreated in vacuum at 300 °C for 5 h.
In order to determine the elemental composition, an inductively
coupled plasma-optical (ICP) emission spectroscopy (PerkinElmer ICP
OPTIMA-3000) was applied.
High resolution transmission electron microscopic (TEM) images
were observed on a transmission electron microscope (FEI Tecnai F20)
with a field emission gun at 200 kV.
CO
2
-TPD experiments were employed to study the basic properties
of the catalysts. Firstly, the sample (0.20 g) was put in a quartz reactor
−
1
bed and pre-treated in a flow of Ar (40 mL min ) at 600 °C for 2 h, and
−1
then decreased to 30 °C. After the absorption of CO
h using a six-way valve, the catalyst was swept with Ar (40 mL min
for 3 h to eliminate all physisorbed CO at surface. Finally, the CO -TPD
experiment was carried out between 30–700 °C with a heating rate of
2
(40 mL min ) for
−
1
2
)
2
2
−
1
10 °C min
2
and the desorbed CO was measured using an on-line gas
chromatography (GC) with a TCD detector.
2.3. Catalytic reaction
In order to investigate the catalyst life more practically, the trans-
esterification of DMC and glycerol was carried out in a continuous-flow,
vertical stainless fixed-bed tubular reactor (8 mm inner diameter,
600 mm length) equipped with a stuff vessel of 300 mL, a product
condenser and a product vessel of 3000 mL. The reactor was heated by a
furnace with a thermocouple and a temperature controller. Typically, a
3-mL catalyst (2.6 g, 40–60 mesh) was placed in the thermostatic zone
of the reactor. After the catalyst was heated to the desired temperature
in a short time, the reactant mixture of DMC and glycerol (n (DMC): n
(
glycerol) = 3:1) was continuously fed into the reactor using a Gilson
pump. The reaction products were cooled in the product condenser.
Once all the parameters remained unchanged, a regular sampling was
collected for quantitative analysis by a GC equipped with a HP-5 ca-
pillary column and a flame ionization detector. The column tempera-
ture ramp was: 70 °C for 3 min, then heating up to 240 °C at
−
1
3
0 °C min , and 240 °C for 4 min. The products were identified by a
GC–MS (Agilent 7890-700B) with HP-5MS capillary column (30 m).
Glycerol conversion (Xglycerol), GLC selectivity (SGLC) were calculated by
the following expressions. GLC yields (YGLC) were determined by mul-
tiplying conversion and selectivity.
Fig. 1. XRD patterns of nCaO-ZrO
2
catalysts: (a), wide angle; (b), low angle.
3
1.4° were related to monoclinic ZrO
and tetragonal ZrO (t-ZrO , JCPDS No. 50-1089) was detected at 30.2°,
4.9°, 50.3° and 59.6°. This differed from the result reported by Wang
et al., who found that m-ZrO phase was mainly found for ZrO using
NH CO as precipitant [40]. In this study, we selected NaOH to
2 2
(m-ZrO JCPDS No. 37-1484),
moles of initial glycerol − moles of final glycerol
Xglycerol =
× 100%
2
2
moles of initial glycerol
3
2
2
moles of GLC formed
S
GLC
=
× 100%
(
4
)
2
3
moles of initial glycerol − moles of final glycerol
synthesize the catalyst precursors, and thus the anion on the surface of
such samples would be partly neutralized by Na+ ion during aging. This
conferred precursors to have a tetragonal structure, and in turn bene-
YGLC = Xglycerol × SGLC × 100%
To better understand the correlation between the number of basic
sites and glycerol conversion, the turnover frequencies (TOF) were
determined by the following principle:
ficiating for the creation of t-ZrO
2
in the final catalysts [41]. Upon the
addition of Ca , the peaks to m-ZrO disappeared and all the mixed
oxides only retained t-ZrO . This finding was attributed to the fact that
2+
2
2
4
the crystal structure of t-ZrO2 could be stabilized selectively after the
1
0 × Xglycrol × LHSV
TOF =
insertion of foreign atoms [36–38]. Herein, benefiting from the doping
(
Mglycerol + 3 × MDMC) × TB
of Ca2 , the structure of t-ZrO
+
was extremely stable even after thermal
2
in which, LHSV is liquid hourly space velocity (h−1); Mglycerol and MDMC
treatment at 600 °C. Importantly, it should be noted that the diffraction
reflection related to calcium oxide was not observed when n was less
than 0.8. Meanwhile, the peak assigned to t-ZrO2 at about 30.2° con-
tinuously moved from higher to lower 2θ values with the rise of n,
being in good agreement with those reported by Pradhan et al. [42].
Apart from this, the intensity of characteristic peaks assigned to t-ZrO
became weaker and weaker. Furthermore, the data listed in Table 1
demonstrated that the lattice parameters based on this peak expanded
−
1
B
are molecular weight of glycerol and DMC, respectively (g mol ); T is
−
1
the total number of basic sites present on the catalysts (μmol g ).
. Results and discussion
.1. Textual and structural properties of the prepared samples
3
2
3
2+
The wide-angle XRD patterns of CaO-ZrO
Fig. 1(a). In this work, pure ZrO was observed in a mixture of mono-
clinic and tetragonal phases. Specifically, the peaks at 23.9°, 28.1° and
2
solid bases are shown in
gradually because the ionic radius of Ca (1.0 Å) was longer than that
of Zr4 (0.84 Å) [42]. Therefore, all these obtained results excluded the
possibility of generation of amorphous form of CaO for samples with n
+
2
3

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
Table 1
Textural chemical and basic properties of various CaO-ZrO
2
catalysts.
Pore volume
Sample
Ca2+:Zr4+ atomic ratio
d
XRD (Å)c
S
BET (m2 g−1
)
Average pore diameter
(nm)
Number of basic sites (μmol g−1) and contributiond
3
−1
(
cm
g
)
Initiala
In solidb
Total
W.
M.
S.
ZrO
2
–
–
2.9453
2.9511
2.9513
2.9587
2.9607
252
243
212
167
87
0.28
0.33
0.37
0.32
0.49
4.6
5.4
8.3
9.2
10.1
43
88
113
138
104
18.3 (42.6)
37.4 (42.5)
35.8 (31.7)
36.1 (26.1)
34.3 (33.0)
21.7 (50.4)
40.6 (46.1)
51.5 (45.6)
63.6 (46.1)
34.1 (32.8)
3.0 (7.0)
0.2CaO-ZrO
0.4CaO-ZrO
0.6CaO-ZrO
0.8CaO-ZrO
2
2
2
2
0.2: 1
0.4: 1
0.6: 1
0.8: 1
0.08:1
0.36:1
0.55:1
0.73:1
10.0 (11.4)
25.7 (22.7)
38.3 (27.8)
35.6 (34.2)
a
b
c
The nominal atomic ratio in the mother solution.
Measured by ICP for the final solid catalysts.
Based on the XRD peak at 30.2°.
d
The value in parentheses is the contribution of single basic sites to the total number of total basic sites. W.: Weak sites; M.: Medium basic sites; S.: strong basic
sites.
from 0.2 to 0.6. In light of the fact that the divalent alkali-earth metal
2
minor carbonate species [45]. As illustrated by the following CO -TPD
4
+
ions could substitute Zr
solution, the absence of separated CaO peak could be explained by the
formation of CaO-ZrO solid solution when n was varied from 0.2 to 0.6
37–39]. Nevertheless, due to the limited capacity of ZrO for in-
in the host lattice to generate a stable solid
characterization, their basic strength was determined by the amount of
Ca content. Considering that a stronger basic strength would result in
the formation of carbonate species with a higher stability, and thus, the
2
[
2
2
catalyst 0.8CaO-ZrO showed the most intense peak at 289.5 eV, owing
corporating CaO dissolution into its lattice, two featured diffraction
to its superior basic strength than other samples. Apart from this, the
peaks assigned to CaO were observed (JCPDS No. 48-1467) at 37.3° and
gradual rise of binding energy from 289.1 to 289.5 eV with a rise of n
5
3.8° as n was further increased to 0.8. Noting that the width of peaks
also pointed the incorporation of Ca2+ into ZrO
2
lattice. The typical Zr
3d spectra are presented in Fig. 3 (b). For pristine ZrO , there appeared
2
+
became smaller after insertion of Ca , the average crystallite sizes
2
were thus calculated based on 2θ of 30.2° by the Scherrer formula. The
two peaks at 184.2 and 181.8 eV with a high intensity, which were
associated with Zr 3d5/2 and Zr 3d3/2 energy states of Zr(IV) oxide
species, respectively [46]. The intensity of these two reflections gra-
dually decreased with an increase of Ca content. Meanwhile, it is worth
crystallite sizes of ZrO
2 2 2 2
, 0.2CaO-ZrO , 0.5CaO-ZrO , 0.6CaO-ZrO , and
0
.8CaO-ZrO were 12.8, 16.9, 15.5, 14.8 and 11.5 nm, respectively. At
2
the same time, the following TEM measurement further affirmed that
these samples consisted mainly of crystals with a size of 5–20 nm,
suggesting that they were a class of nano-catalysts. Fig. 1(b) gives the
noting that adding Ca into ZrO
2
support could give rise to a continue
increase of Zr 3d binding energy. These observations also support that
2+
low-angle XRD patterns of nCaO-ZrO
2
catalysts. It could be seen that all
Ca
2
had entered into the t-ZrO lattice, creating a solid solution.
the catalysts show a sharp and symmetric peak at about 2θ = 1°, which
Fig. 3(c) displays the Ca 2p spectra of these catalysts, and the XPS
spectrum of single CaO is also presented for a comparison. When
n < 0.8, all the catalysts exhibited a broad and intense band centered at
346.1 eV related to the emission from Ca 2p3/2 of Ca2 in oxide state
was corresponding to their mesoporous structure [39]. As discussed
2
+
above, the addition of excessive Ca
pearance of free CaO phase, which was adverse to the generation of a
uniformly mesoporous framework. As a result, ZrO exhibited the
highest intensity of diffraction, whereas Ca-rich sample (0.8CaO-ZrO
into the system induced the ap-
+
2
2
[46,47]. In the case of sample 0.8CaO-ZrO , its spectral profile was
2
)
similar to that of CaO, in which an additional peak was observed at
about 350.5 eV. According to the previous work, this satellite peak can
be assigned to Ca 2p1/2 state of Ca2+ [46,47]. More importantly, the
binding energies of Ca 2p2/3 in all the mixed oxides were lower than
presented the lowest one. This is consistent with the results of these
kinds of Zr-based solid bases in previous literature [37,39].
It was accepted that Raman analysis was an effective technique to
study the microstructure of Zr-containing materials. Raman spectra of
that of pure CaO, because the nCaO-ZrO
sessed a solid solution structure. Fig. 3(d) depicts the O1 spectra of
these materials. For samples ZrO , 0.2CaO-ZrO , 0.4CaO-ZrO and
0.6CaO-ZrO , the Ca content did not influence the shape of the O1s
2
-containing materials pos-
nCaO-ZrO
2
, pure ZrO
2
and CaO are shown in Fig. 2. Apparently, bare
at
69, 314, 460 and 645 cm , but also the typical reflections to
ZrO support presented not only the characteristic peaks of t-ZrO
2
2
2
2
2
−1
2
2
−1
monoclinic ZrO
2
phase at 331, 380, 536 and 559 cm
[40,42]. This
cures largely. Precisely, only a sharp peak appeared at about 529.7 eV.
According to Pradhan et al., this peak was ascribed to the lattice oxygen
fact is in line with its XRD result in Fig. 1(a). In contrast, samples of
.2CaO-ZrO , 0.4CaO-ZrO and 0.6CaO-ZrO demonstrated the peaks
attributed to t-ZrO only. No Raman lines due to m-ZrO or CaO were
detected. These also suggested that doping a foreign Ca atom could
covert ZrO from a monoclinic to a tetragonal phase. In addition, it
0
2
2
2
of the CaO-ZrO
O1s peak also implied that strong Lewis basic sites were present at their
surfaces [6], being in line with the following CO -TPD characterization
results. As to the Ca-rich sample (0.8CaO-ZrO ), beside this peak, a new
2
solid solution [42]. Moreover, the appearance of this
2
2
2
2
2
should be noted that the peaks of these Ca-modified samples widened
one with a mild intensity emerged at 531.2 eV, which was consistent
obviously, due to the existence of oxygen vacancy which was resulted
with the O1 s spectrum of pure CaO [48]. This result can be explained
as follows: all the Ca ions were introduced into t-ZrO lattice without
2
2
+
2+
from the incorporation of Ca
more, the peaks at 269 and 645 cm
venumbers for mixed oxide catalysts with a rise of Ca content, further
affirming that they possessed a stable structure of CaO-ZrO solid so-
lution. Nevertheless, a new reflection corresponding to CaO emerged at
into the ZrO
2
lattice [43,44]. Further-
−1
had a slight shift to lower wa-
creation of any free CaO when n < 0.8, and thus only the O1S peak
related to the lattice oxygen appeared. The catalyst 0.8CaO-ZrO also
2
2
had a counterpart of CaO oxygen due to the existence of separated CaO,
judging from the XRD results. Thus, these two O1S reflections were
clearly detected. At the same time, this also provided a strong support
−1
2
2
89 cm for 0.8CaO-ZrO sample due to that some free CaO particles
were formed on its surface (see Fig. 1(a)).
To further verify the generation of CaO-ZrO
2
for the conclusion that CaO-ZrO was in a solid solution state for
2
solid solution, all the
samples of 0.2 ≤ n ≤ 0.8. Additionally, in the case of Ca concentration
for these mixed oxide samples, the measured XPS values were higher
than those from the ICP measurements, implying that their surface was
enriched in Ca (see Fig. S1 in Supplementary Material).
samples were analyzed in-depth by the XPS measurements. For pure
ZrO , its C 1s spectra gave an intense reflection at 284.5 eV and a
2
shoulder peak at 289.1 eV (see Fig. 3(a)). The former one was attributed
to the adventitious carbon, while the latter one was originated from the
2
Textural properties of the nCaO-ZrO catalysts are evaluated by an
4

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
2 2
Fig. 2. Raman spectra of CaO, ZrO and nCaO-ZrO mixed oxides.
2 2
Fig. 3. C 1s (a), Zr 3d (b), Ca 2p (c) and O 1s (d) XPS spectra of CaO, ZrO and nCaO-ZrO samples.
5

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
2 2
N adsorption-desorption technique and the N adsorption-desorption
isotherms are displayed in Fig. 4. Apparently, all the isotherms were of
type IV with clear hysteresis loops at higher relative pressure. Following
the IUPAC principle, these catalytic materials were of mesoporous solid
catalysts [49,50]. Moreover, it was notable that their hysteresis loops
were greatly affected by the Ca content. For the pure ZrO
samples (0.2CaO-ZrO and 0.4 CaO-ZrO ), hysteresis loops demon-
strated an H2 type with P/P = 0.5–0.9, suggesting the formation of an
ordered structure of mesoporous [51]. In the case of the Ca-rich sample
0.8CaO-ZrO ), H3 shaped hysteresis loops (P/P = 0.7–1.0) were ob-
served, indicating an existence of slit-shaped pores [51]. For 0.6CaO-
ZrO , although it mainly presented H2 shaped hysteresis loops, the
higher opening and closing P/P position resembled that of H3 type.
2
and Ca-poor
2
2
0
(
2
0
2
0
Collectively, all the above results indicate that an increase of Ca content
makes their mesoporous structure unordered, which is also confirmed
by the results in Fig. 1(b). The pore size distributions of the nCaO-ZrO
2
catalysts are also given in Fig. 4(b). It could be found that all the
samples exhibited substantially developed porous structures, and their
pore sizes were widely distributed between 5–50 nm. Moreover, the
2 2
catalysts ZrO and 0.2CaO-ZrO presented a sharp pore volume dis-
tribution in a narrow and smaller pore diameter range, while a broad
distribution peak of porous structure with a large pore diameter was
observed for the catalytic materials with a high Ca content. This cor-
roborated again that an excessive Ca amount did not facilitate forma-
tion of mesoporous structure. From the results reported in Table 1, it
could be observed that the surface areas of these samples were sig-
nificantly high, which was beneficial for them exposing more active
sites. Simultaneously, the obtained data revealed that a decrease in
surface area with an increasing n was ascribed mainly to the negative
effect of a high Ca content on their mesoporous frameworks. Moreover,
the pore diameters of the catalysts increased with a rise of the n value;
clearly, the sample of 0.8CaO-ZrO
TEM images of catalysts are shown in Fig. 5. All the samples except
for 0.8CaO-ZrO are composed mainly of nano-crystals with diameters
of 10–20 nm, clustering together to generate a sponge-like network. In
particular, the 0.6CaO-ZrO catalyst had a continuous pore/channel
2
possessed the largest pore diameter.
2
2
with a 4-10-nm width. Interestingly, amorphous structures were
dominant without appearance of obvious crystalline phase for this
catalyst. However, some separated particles were detected for 0.8CaO-
Fig. 4. N
of nCaO-ZrO
2
adsorption–desorption isotherms (a) and pore-size distributions (b)
catalysts.
2
2
ZrO , which might be owing to the existence of free CaO particle at its
2
Fig. 5. TEM images of various nCaO-ZrO catalysts.
6

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
to measure the catalyst basicity. Clearly, all the profiles in Fig. 6 con-
sisted of three Gaussian peaks, which stood for weak basic sites (below
−
2
00 °C) and were attributed to OH groups, for moderate basic sites
(
between 200 and 400 °C) related to metal-oxygen pairs (including Ca-
O and Zr-O), and for strong basic sites (over 400 °C) associated with an
2-
isolated O , respectively [39]. As expected, the strong basic sites of
these samples shifted gradually towards higher orientations with a rise
of n, illustrating that the introduction of Ca2+ could promote the basic
strength of the resulted nCaO-ZrO
these catalysts gave decreasing order of ZrO
ZrO < 0.4CaO-ZrO < 0.6CaO-ZrO < 0.8CaO-ZrO
2
catalysts. In terms of basic strength,
a
2
< 0.2CaO-
Meanwhile,
2
2
2
2
.
the basic site distributions were calculated and are summarized in
Table 1 for all the catalysts. No doubt, the distributions of these samples
varied greatly depending on the n value. Particularly, the strong basic
peaks became more and more intense with a rise of n, at expense of
weak and moderate basic sites belonging to ZrO
percentage of strong basic sites for 0.6CaO-ZrO was 27.8%, which was
about 4 times that of Ca-free sample (ZrO ). The total number of basic
2
. Consequently, the
2
2
sites increased sharply with an increasing of the Ca content when
n ≤ 0.8, and then decreased with a further increase of n. This trend was
caused by the double-edged effect of Ca2+ on the surface basicity. On
the one hand, the amount of metal-oxygen (Ca-O) would increase with
the addition of Ca2+, resulting in an improvement of the number of
2
-
moderate basic sites. Simultaneously, the charge density of O would
increase with the decline of O1s binding energy, owing to the sub-
stitution of Zr4+ by Ca
2+
and to the reduced positive charges in the
resulted catalysts, which had been proved by the foregoing XRD,
Raman, and XPS techniques. Consequently, more and strong basic sites
were created. On the other hand, the increasing of the Ca content led to
a decreased surficial area, which might in turn reduce the exposed basic
sites. The positive effect caused by the addition of Ca2+ on their basic
property was predominant when n was in the range of 0-0.6. However,
the negative effect caused by doping of Ca2+ surpassed its positive
influence when n > 0.6. Coupled with the fact that the number of weak
sites was almost independent of n except for pure ZrO
the change of total basic sites for these samples exhibited a typical
volcanic shape, following a decreasing order: 0.6CaO-ZrO > 0.4CaO-
ZrO > 0.8CaO-ZrO > 0.2CaO-ZrO > ZrO
2
(see Table 1),
Fig. 6. CO
2
-TPD of nCaO-ZrO
2
and ZrO
2
catalysts.
2
2
2
2
2
.
surface.
3.3. Catalytic performance
3
.2. Surface basic properties
3.3.1. Reaction system for the transesterification of DMC and glycerol
For glycerol, its two terminal hydroxyl groups are more active than
the middle one. After the transesterification between DMC and the
terminal hydroxyl groups, the intra-molecular transesterification in-
volving the middle hydroxyl group and the neighboring carbonyl group
was facilitated. Thus, the target product GLC was manufactured with a
high yield. However, when the GLC concentration was increased, the
undesired product glycerol dicarbonate (GDC) would be generated via
The basic properties of the nCaO-ZrO
2
catalysts were investigated
desorption curves are
using the CO -TPD measurement, and the CO
2
2
shown in Fig. 6. Generally speaking, the temperature for desorption
peak was served to determine the basic species, and the basic strength
could be reflected by the maximum desorption temperature. The
2
amount of desorpted CO , corresponding to the peak area, was applied
Scheme 2. Reaction system for the reaction of DMC and glycerol.
7

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
Moreover, the total glycerol conversion could be obtained at a low
−1
LHSV value (0.5 h ). Apparently, the rate of this reaction was not
dependent on extra-particle mass transfer. On the other hand, following
-
5
these conditions, the reaction rate was calculated as 5 × 10 mol
−1
−1
s
mL . The intra-particle diffusion degree could be evaluated using
the Thiele modulus revised by Weisz [52]:
νR2
C D
0 eff
φ =
In this formula, ν is the rate of reaction (mol s mL−1); R stands for
−1
the radius of the catalyst particle (about 0.02 cm in this work); C
0
is
-
4
−1
referred to the initial concentration of glycerol (4.3 × 10 mol mL );
-
4
2 −1
and Deff is the diffusion coefficient (selected as 10 cm s ). The value
of φ was calculated to be about 0.004, which is much less than 1.0. It
was believed that the overall reaction would be restricted by diffusion
when φ was 1.0 [52], thus, the intra-particle diffusion hardly limited
the GLC synthesis in this work. In addition, the catalytic performances
2 2 2 2
of ZrO , 0.2CaO-ZrO , 0.4CaO-ZrO and 0.4CaO-ZrO under different
reaction conditions are also presented in Supplementary Material.
The effect of LHSV on the GLC production is shown in Fig. 7(b). The
glycerol conversion declined slightly with an increased LHSV value
−
1
from 0.5 to 1.5 h . Nevertheless, a sharp decrease of glycerol con-
−
1
version was observed when LHSV was further increased to 2.0 h
Rationally, a higher LHSV value indicated that more reactant was em-
ployed. Considering that the catalytic ability of this 0.6CaO-ZrO
.
2
sample was constant, the residence time was too short for converting
glycerol to GLC completely. Thus, the suitable LHSV was chosen as
−
1
1.5 h
in this work.
3.3.3. Effect of reaction temperature
The effect of varying the temperature from 70 to 100 °C on the re-
action was studied in the case of 0.6CaO-ZrO
sults presented that the glycerol conversion increased rapidly until
0 °C. This might be explained by the vaporization loss of co-produced
2
(see Fig. 7(c)). The re-
9
methanol from reaction system, driving the reaction towards the GLC
production [53]. However, higher temperatures also accelerated the
rate of further reaction to generate GDC. Consequently, the GLC se-
lectivity declined constantly with the increase of the reaction tem-
perature. Based on these facts, the GLC yield reached its maximal value
of 90% at 90 °C, but showed a slight decrease when the reaction
reached at 100 °C. Therefore, the optimal reaction temperature was
selected as 90 °C.
3.3.4. Effect of Ca content and basic property
The catalytic performance of various catalysts with a different Ca
content after reaction of 8 h is illustrated in Fig. 8(a). Without catalyst,
no significant glycerol conversion was obtained. All the Ca-containing
2
catalysts gave higher glycerol conversion when compared to pure ZrO ,
pointing out the positive effect of Ca on the reaction. Moreover, the
glycerol conversion was related highly to the value of n. The catalyst
Fig. 7. Effect on reaction parameters on the GLC synthesis over 0.6CaO-ZrO
catalyst; error bars give the standard deviation.
2
ZrO
high conversion of 78% could be achieved for the catalyst 0.8CaO-ZrO
Specially, the catalyst 0.6CaO-ZrO reached the highest glycerol con-
2
exhibited a low reactivity of 50% glycerol conversion, while a
Effect of diffusion (a). Reaction conditions: reaction temperature, 90 °C; cata-
lyst, 3 mL; n (DMC): n (glycerol) = 3:1. Effect of LHSV (b). Reaction conditions:
reaction temperature, 90 °C; catalyst, 3 mL; n (DMC): n (glycerol) = 3:1. Effect
of reaction temperature (c). Reaction conditions: catalyst, 3 mL; n (DMC): n
2
.
2
version of 97%. Taking into account of the enhancement of surface
basicity, it seemed that the high number of basic sites could induce a
high catalytic ability towards the transesterification reaction. Indeed,
this hypothesis was supported by Fig. 8(a), where the sequence of
glycerol conversion was consistent roughly with that of total amounts
of basic sites. Similar conclusions were also drawn in previous works
−
1
(glycerol) = 3:1; LHSV, 1.5 h
.
the transesterification reaction between the hydroxyl group of GLC and
another DMC molecule (Scheme 2).
[
24,26,28,54]. In addition, TOF values of catalysts in the kinetic regions
(
not at a very high glycerol conversion) are a useful clue to unravel the
3
.3.2. Effect of liquid hourly space velocity (LHSV)
Firstly, two different LHSV values were selected to estimate the
role of basic property to their activity; where TOF denotes as the
number of glycerol molecules converted on a unit basic site per hour
−1
effect of diffusion (Fig. 7(a)). Notably, the glycerol conversion in-
creased linearly at first 2 h when the LHSV value was set at 2 h
(h ). Thus, we demonstrate the glycerol conversion, total number of
basic sites on the catalysts and TOF values for the nCaO-ZrO at 90 °C
−1
.
2
8

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
Fig. 8. Effect of Ca content and basic property on the GLC synthesis; error bars give the standard deviation.
The relationship between glycerol conversion (a); TOF values (b); GLC selectivity (c); GLC yield (d); and total number basic sites.
−
1
Reaction conditions for (a), (c) and (d): 90 °C, catalyst, 3 mL; n (DMC): n (glycerol) = 3:1; LHSV, 1.5 h
.
−
1
Reaction conditions for (b): 80 °C, catalyst, 3 mL; n (DMC): n (glycerol) = 3:1; LHSV, 2.0 h
.
−
1
with 2 h
catalysts containing Ca are higher than that of bare ZrO
TOF values presented a typical volcano-shaped profile in terms of Ca
loadings and the catalyst 0.6CaO-ZrO possessed the highest TOF value
of LHSV in Fig. 8(b). Apparently, the TOF values of the
the highest glycerol conversion of 46% (see Fig. S2(a) in Supplementary
material). When the calcination temperature was at 400 °C, the cata-
2
. Moreover, the
lytic ability increased in an order: 0.6CaO-ZrO
2
> 0.4CaO-
2
ZrO > 0.2CaO-ZrO > 0.8CaO-ZrO > ZrO . It was noted that this
2
2
2
2
−
1
at 16.4 h . This fact further highlights the importance of basic sites for
samples towards the transformation of glycerol during reaction.
Aiming at further investigating the role of every basic species (weak,
moderate and strong basic sites) towards the reaction of DMC between
glycerol separately, all the precursors were treated at variable tem-
peratures ranging from 200 to 600 °C. Generally, after calcination at a
relatively low temperature (200 °C), the catalytic ability of mixed oxide
order is agreement with the variation in the number of both weak and
moderate sites (see Fig. S2(b) in Supplementary Material). All these
findings ascertained that the surficial basic sites of catalysts played a
vital role for the transestification reaction. In other words, each basic
species might have a similar activity towards the glycerol conversion
for GLC synthesis.
Fig. 8(c) shows the plot of product selectivity against the total
amount of basic sites for all the catalysts. It is clearly shown that the
selectivity of GLC was depressed when more basic sites were present on
the surfaces of the catalysts. Correspondingly, the selectivity of by-
product GDC increased with the increasing of basic density. For in-
−
mostly originated from weak basic sites associated with OH , created
by decomposition of surface bicarbonate species [55]. Upon calcination
at 400 °C, only partial decomposition occurred, thus, the catalytic
ability was related to both weak and moderate basic sites. In case of
thorough decomposition at higher temperatures like 600 °C, the cata-
lytic ability was ascribed to the united force of the three types of basic
stance, the strongly basic catalyst 0.6CaO-ZrO
selectivity in comparison with 0.2CaO-ZrO . It should be noted that the
catalyst 0.8CaO-ZrO also showed a high reactivity for the GDC
synthesis, despite that it had smaller basic density than that of 0.6CaO-
ZrO . This might be attributed to its strong basic strength, as illustrated
by the CO -TPD analysis. Finally, the catalytic behaviors of these ma-
terials with respect to GLC yield are illustrated in Fig. 8(d). As expected,
0.6CaO-ZrO showed the superior catalytic performance than the other
2
exhibited inferior GLC
2
−
2-
sites (weak OH , moderate Ca/Zr-O pairs and strong O sites) [55].
The effect of 600 °C on the catalytic performances is illustrated in
Fig. 8(a), so, we only show the catalytic results of various samples that
were obtained at 200 and 400 °C, respectively (see Fig S(2) in Supple-
mentary material). When the calcination temperature was at 200 °C, the
catalytic ability of samples, expressed by glycerol conversion, altered
2
2
2
2
only a little except for pure ZrO
these nCaO-ZrO catalysts exhibited similar amounts of weak basic sites
see Table 1). Importantly, their activity order was in good agreement
with that of weak basic sites density and the catalyst 0.2 CaO-ZrO gave
2
. It might be attributed to the fact that
catalysts with a 90%-yield of GLC. This highlighted that a right basic
property was vital afford a high GLC yield.
Based on the experimental results and discussion aforementioned, a
plausible reaction mechanism was proposed (see Scheme 3(a)). First,
2
(
2
9

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
2
Scheme 3. Possible reaction mechanism for (a), the GLC synthesis; and (b), GDC synthesis over nCaO-ZrO catalysts.
both the terminal hydroxyl groups and the middle one in glycerol were
activated by basic sites with the abstraction of Hδ+. Consequently, the
glycerol became a bidentate ligand, chelating to the catalyst surface.
This point was supported by the fact that no mono-transesterification
product was detected. As a result, the nucleophilic ability of glycerol
would be strengthened and the oxygen atom in terminal hydroxyl group
would attack the carbon atom of carbonyl group in DMC molecule, with
a release of a methanol molecule. Similar to this procedure, another
activated oxygen atom of the hydroxyl group continues attacking the
nearby carbonylic carbon atom, finishing the intra-molecular transes-
terification. Finally, the target product GLC was formed with liberation
of another methanol molecule. Herein, it was possible that a higher
basic density could give glycerol more nucleophilicity, which favored
consecutive transesterification would take place between GLC and an-
other DMC molecule, giving rise to the manufacture of GDC and a new
molecule of methanol (see Scheme 3(b)). Reasonably, the production of
GDC was probably determined by the equilibrium between desorption
and absorption of GLC on the catalyst. Higher basic concentration or a
stronger basic strength induced a stronger adsorption of GLC, facil-
itating the conversion of GLC to GDC. As a consequence, the selectivity
of GDC increased greatly with consumption of GLC, which was affirmed
by the foregoing experimental results.
3.4. Stability of catalyst
It was widely acknowledged that the main problem concerning the
Ca-based catalyst for the GLC synthesis was its poor catalytic stability
during the recycling [27–29]. So, it was very important to examine the
2
the transesterification. Thus, the catalyst 0.6CaO-ZrO showed high
reactivity owing to its high ability for activating glycerol. In addition,
similar mechanisms that glycerol was activated on basic sites were also
proposed previously [24,26,28,54].
lifetime of the CaO-ZrO
Fig. 9 depicts a comparative performance of the 0.6CaO-ZrO
.8CaO-ZrO catalysts over 15 h. To our delight, apart from its high
catalytic ability, the catalyst 0.6CaO-ZrO was also very stable for a
2
catalysts for the continuous synthesis of GLC.
2
and
2
However, if the catalyst (such as 0.6CaO-ZrO ) contained too many
0
2
basic sites, the production of GDC would be aggravated, leading to a
much lower GLC selectivity. This behavior could be explained as fol-
lows: GLC also has a free hydroxyl group, and it thus could also link to
2
long time reactant flushing. Remarkably, the glycerol conversion re-
mained almost unaltered even after 15 h in the stream. On the contrary,
although the catalyst 0.8CaO-ZrO displayed a good activity with a
2
the basic sites providing nucleophilic oxygen atoms. Then,
a
10

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
2 2
Fig. 9. Conversion of glycerol over 0.6CaO-ZrO and 0.8CaO-ZrO catalysts in a long-term test.
Reaction reactions: 90 °C, catalyst, 3 mL; n (DMC): n (glycerol) = 3:1; LHSV, 1.5 h−1. Error bars give the standard deviation.
Table 2
2
Catalytic behavior comparison of current 0.6CaO-ZrO catalyst with other reported Ca-based catalysts.
Catalyst
Time (h)
Glycerol conversion (%)
GLC selectivity (%)
Production mode
Reference
0
0
CaO
CaO
.6CaO-ZrO
.6CaO-ZrO
2
2
–
2
1.5
3
3
3
3
1.5
97
93
90
86.0
87
98
61
97
100
Continuous
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Present work
Present work
[21]
[33]
[32]
[33]
[32]
[34]
a
100
74.4
77
75
82
Ca-Al hydrocalumite
CaAl-layerd hydroxide
CaO-Al
2
O
3
93
100
CaO-MgO
Reaction conditions: glycerol 40 mmol, DMC 120 mmol, 0.3 g catalyst, 90 °C, 2 h.
glycerol conversion of 87% for the reaction time of 2 h from the start,
the conversion decreased gradually to 78% after a reaction time of 8 h.
Then, its activity remained at a certain level. As discussed above, the
times, dried at 100 °C, and was then characterized by means of XRD and
TEM (see Fig. S3 and Fig. S4 in Supplementary Material). It was found
that the used catalyst still retained the solid solution structure and
basic sites of the catalyst 0.6CaO-ZrO
2
mainly originated from the ad-
. Consequently, all the basic species were firmly
anchored onto the t-ZrO framework thanks to the formation of solid
solution. These basic sites were highly dispersed in the ZrO lattice,
rather than that CaO simply coated on ZrO , which was seen for the
wormhole-like ordered mesoporous frameworks. Moreover, N
2
ad-
2
+
dition of Ca into ZrO
2
sorption/desorption technique further revealed that the used catalyst
2
−1
2
possessed a high BET surface area (164 m g ). It is thus concluded
2
that the catalyst 0.6CaO-ZrO
structural stability.
2
has an excellent reusability and a good
2
supported bases obtained via impregnation. Therefore, the solid-solu-
tion structure endowed the nature of high structural stability for the
catalyst 0.6CaO-ZrO . For the catalyst 0.8CaO-ZrO , a fraction of its
2 2
3.5. Activity comparisons
basic sites stems from the separated CaO phase, and thus they are
weakly bound to catalyst surface. When it was employed to catalyze
GLC synthesis, this fraction of basic sites was easily lost due to the
generation of dissolved Ca-species [30–32]. Moreover, the two catalysts
It is difficult to directly compare the present catalyst 0.6CaO-ZrO
with the reported Ca-containing catalysts since they were evaluated in a
different way from the present work. Herein, we further tested catalyst
2
.6CaO-ZrO with a batch reactor. The typical catalytic assessment is
illustrated in Supplementary Material, and the comparison results are
given in Table 2. It was worth mentioning that the glycerol conversion
obtained in the batch reactor was higher than that achieved in the
fixed-bed, whereas the GLC selectivity showed a slight decrease with
2
0
were analyzed by elemental analysis and CO
reactions. The analytic results demonstrate no obvious changes for the
Ca content for the catalyst 0.6CaO-ZrO . Nevertheless, the Ca:Zr ratio
initially used in the catalyst 0.8CaO-ZrO decreased to a certain extent.
Furthermore, the deviation of basic concentration between fresh and
2
-TPD techniques after the
2
2
−
1
2
0.6CaO-ZrO . The above changes might be attributed to the relatively
used 0.6CaO-ZrO
2
(135 μmol g ) could be negligible, while the
−1
long residual time (2 h) for reagent on catalyst in batch reactor com-
pared to that in fixed-bed reactor. Thus, the accessibility of the reactant
to the active sits was greatly enhanced. However, a long residual time
also induced more probability for side reactions, reducing GLC se-
lectivity to some extent. Despite of this negative effect, the catalyst
number of basic sites of used 0.8CaO-ZrO
2
(93 μmol g ) had a no-
ticeable decline. Once the leaching of these free CaO phase terminated,
the residual basic species originated from the solid solution structure
would give a stable glycerol conversion with flushing of reactant.
Additionally, the used 0.6CaO-ZrO
2
catalyst after a time of 15 h in
0
.6CaO-ZrO
2
still afforded a GLC yield of 90% within a reaction of 2 h
the fixed-bed was recovered by centrifuge, washed by ethanol for three
at 90 °C. Taking into account the fact that bare CaO only gave about
11

X. Zhang, et al.
Applied Catalysis A, General 590 (2020) 117313
8
0% glycerol conversion [21,33], the carrier ZrO
enhance the reactivity remarkably. Moreover, the major shortcomings
like low surface area and rapid deactivation caused by CO con-
2
was thus efficient to
1012–1029.
[
[
[
3] B. Das, K. Mohanty, Ind. Eng. Chem. Res. 58 (2019) 15803–15817.
4] Z.I. Ishak, N.A. Sairi, Y. Alias, M.K.T. Aroua, R. Yusoff, 297 (2016) 128–138.
5] M.O. Sonnati, S. Amigoni, E.P. Taffin de Givenchy, T. Darmanin, O. Choulet,
F. Guittard, Green Chem. 15 (2013) 283–306.
2
tamination hindered industrial utilization of pure CaO. Furthermore,
many researchers had discovered that its activity mostly ascribed to the
homogeneous Ca-species [30,31]. Other Ca-based catalysts such as Ca-
[
6] W. Luo, L. Sun, Y. Yang, Y. Chen, Z. Zhou, J. Liu, F. Wang, Catal. Sci. Technol. 8
(2018) 6468–6477.
[
7] A.-A.G. Shaikh, S. Sivaram, Chem. Rev. 96 (1996) 951–976.
Al hydrocalumite [32], Ca-Al layerd double hydroxide [33], CaO-Al
2
O
3
[8] J. Hu, Y. Gu, Z. Guan, J. Li, W. Mo, T. Li, G. Li, ChemSusChem 4 (2011) 1767–1772.
[9] A.J. Kamphuis, F. Picchioni, P.P. Pescarmona, Green Chem. 21 (2019) 406–448.
10] X. Su, W. Lin, H. Cheng, C. Zhang, Y. Wang, X. Yu, Z. Wu, F. Zhao, Green Chem. 19
[
32] and CaO-MgO [34] possessed higher activity than that of CaO, but
[
they suffered from the leaching of active CaO site into reaction medium,
severely inhibiting their practical applications [30,31]. In addition, the
comparative TOF results of the present 0.6CaO-ZrO
based samples are listed in Table S2 in Supplementary Material. We
calculated the TOF value following the method proposed by Kumar
et al. [56] (see Supplementary Material). To our delight, the catalyst
(2017) 1775–1781.
[
11] S. Dabral, T. Schaub, Adv. Synth. Catal. 361 (2019) 223–246.
[12] D. Wang, X. Zhang, X. Cong, S. Liu, D. Zhou, Appl. Catal. A Gen. 555 (2018) 36–46.
13] D. Fakhrnasova, R.J. Chimentão, F. Medina, A. Urakawa, ACS Catal. 5 (2015)
284–6295.
14] H. Nguyen-Phu, E.W. Shin, J. Catal. 373 (2019) 147–160.
[15] P. Devi, U. Das, A.K. Dalai, Chem. Eng. J. 346 (2018) 477–488.
[16] M. Marimuthu, P. Marimuthu, S. Palanivelu, V. Rajagopalan, Mol. Catal. 460 (2018)
3–62.
17] Z.I. Ishak, N.A. Sairi, Y. Alias, M.K.T. Aroua, R. Yusoff, Catal. Rev. 59 (2017) 44–93.
[18] B. Liu, C. Li, G. Zhang, X. Yao, S.S.C. Chuang, Z. Li, ACS Catal. 8 (2018)
0446–10456.
2
and reported Ca-
[
6
[
−
1
0
.6CaO-ZrO
ferior to CaO-MgO mixed oxide. When the poor stability of CaO-MgO is
taken into account, the catalyst 0.6CaO-ZrO is of highly practical one
for continuous and efficient production of GLC.
2
gave a high TOF value of 12.2 h , which was only in-
5
[
2
1
[
19] A. Bansode, A. Urakawa, ACS Catal. 4 (2014) 3877–3880.
[
20] M. Tudorache, A. Negoi, V.I. Parvulescu, Catal. Today 279 (2017) 71–76.
4
. Conclusion
[21] J.R. Ochoa-Gómez, O. Gómez-Jiménez-Aberasturi, B. Maestro-Madurga,
A. Pesquera-Rodríguez, C. Ramírez-López, L. Lorenzo-Ibarreta, J. Torrecilla-Soria,
M.C. Villarán-Velasco, Appl. Catal. A Gen. 366 (2009) 315–324.
In this work, we have developed a method for preparations of me-
[
22] R. Bai, Y. Wang, S. Wang, F. Mei, T. Li, G. Li, Fuel Process. Technol. 106 (2013)
soporous CaO-ZrO
cipitation procedure. The CaO-ZrO
possessed a uniformly mesoporous structure and a large specific surface
area, which was favorable for exposure of more basic sites. Moreover,
the basic density of these samples was highly related to the Ca:Zr ratio
and exhibited
Importantly, after the incorporation of Ca
characterization results proved that the basic sites were firmly dis-
persed in the substrate thanks to the formation of solid solution. The
2
solid bases via a simple surfactant-free co-pre-
catalysts prepared by this method
209–214.
[23] J. Zhen, M. Liu, M. Kang, Ning Yin, X. Wang, Y. Tan, Y. Zhu, J. Brazil. Chem. Soc. 25
2014) 152–160.
[
2
(
24] X. Song, Y. Wu, F. Cai, D. Pan, G. Xiao, Appl. Catal. A Gen. 532 (2017) 77–85.
25] S. Pan, L. Zheng, R. Nie, S. Xia, P. Chen, Z. Hou, Chinese J. Catal. 33 (2012)
1772–1777.
[
[
[
[
[
26] D. Singh, B. Reddy, A. Ganesh, S. Mahajani, Ind. Eng. Chem. Res. 53 (2014)
a
volcano trend with the rise of
a
Ca content.
1
8786–18795.
2
+
into ZrO
2
lattice, various
27] F.S.H. Simanjuntak, V.T. Widyaya, C.S. Kim, B.S. Ahn, Y.J. Kim, H. Lee, Chem. Eng.
Sci. 94 (2013) 265–270.
28] G. Parameswaram, M. Srinivas, B. Hari Babu, P.S. Sai Prasad, N. Lingaiah, Catal.
Sci. Technol. 3 (2013) 3242–3249.
2
catalyst 0.6CaO-ZrO showed a striking activity and a good stability for
29] X. Zhang, D. Wang, J. Ma, W. Wei, Catal. Lett. 147 (2017) 1181–1196.
the continuous synthesis of GLC in a fixed-bed reactor. Furthermore, a
good correlation between the catalytic activity and the total amount of
surface basic sites has been established. However, the consecutive
transestification between GLC and DMC would take place severely over
basic materials with too much basic sites/strong basic strength. Among
[30] J. Li, T. Wang, React. Kinet. Mech. Catal. 102 (2011) 113–126.
[
31] F.S.H. Simanjuntak, T.K. Kim, S.D. Lee, B.S. Ahn, H.S. Kim, H. Lee, Appl. Catal. A
Gen. 401 (2011) 220–225.
[
32] L. Zheng, S. Xia, X. Lu, Z. Hou, Chinese J. Catal. 36 (2015) 1759–1765.
[33] J. Granados-Reyes, P. Salagre, Y. Cesteros, Appl. Clay Sci. 132–133 (2016)
2
16–222.
[
[
34] M.S. Khayoon, B.H. Hameed, Appl. Catal. A Gen. 466 (2013) 272–281.
35] T. Yamaguchi, Catal. Today 20 (1994) 199–217.
2
all the catalysts, 0.6CaO-ZrO showed the best catalytic performance
with a 97% glycerol conversion and a 93% GLC selectivity. Thus, this
work provides a new avenue for fabricating Ca-based catalyst with a
good reactivity and a long durability for the continuous synthesis of
GLC via transesterification between DMC and glycerol.
[36] Z. Liu, W. Zhao, F. Xiao, W. Wei, Y. Sun, Catal. Commun. 11 (2010) 675–678.
[
[
[
[
37] S. Liu, X. Zhang, J. Li, N. Zhao, W. Wei, Y. Sun, Catal. Commun. 9 (2008)
1
527–1532.
38] M. Varkolu, D.R. Burri, S.R.R. Kamaraju, S.B. Jonnalagadda, W.E. Van Zyl, J.
Porous Mat. 23 (2016) 185–193.
39] X. Zhang, D. Wang, G. Wu, X. Wang, X. Jiang, S. Liu, D. Zhou, D. Xu, J. Gao, Appl.
Catal. A Gen. 555 (2018) 130–137.
Acknowledgements
40] J. Wang, G. Li, Z. Li, C.H. Tang, Z.H. Feng, H. An, H. Liu, T. Liu, C. Li, Sci. Adv. 3
(
2017) e1701290.
[41] Z.H. Ma, R. Xu, C.H. Yang, W. Wei, W. Li, Y. Sun, Acta Phys. Chim. Sin. 20 (2004)
221–1225.
This work is financially supported by the Key Research and
Development Program of Shandong Province (2018GGX107010),
1
[
[
42] S. Pradhan, V. Sahu, B.G. Mishra, J. Mol. Catal. A Chem. 425 (2016) 297–309.
43] S. Figueroa, J. Desimoni, P.C. Rivas, M.M. Cervera, M.C. Caracoche, O. Sanctis,
Chem. Mater. 17 (2005) 3486–3491.
Natural
ZR2016BL21,ZR2018LB030), Science and Technology Research
Program for Colleges and Universities in Shandong Province
J18KA107), Zaozhuang independent innovation and achievements
Science
Foundation
of
Shandong
Province
(
[
[
44] L. Chen, J. Hu, R.M. Richards, ChemPhysChem 9 (2008) 1069–1078.
45] M.L. Granados, M.D.Z. Poves, D.M. Alonso, R. Mariscal, F.C. Galisteo, R. Moreno-
Tost, J. Santamaría, J.L.G. Fierro, Appl. Catal. B Environ. 73 (2007) 317–326.
(
transformation projects (201502) and National Training Programs of
Innovation and Entrepreneurship for Undergraduates (201710904063,
[46] R. Koirala, K.R. Gunugunuri, S.E. Pratsinis, P.G. Smirniotis, J. Phys. Chem. C 115
2011) 24804–24812.
(
[
[
[
47] S. Luz Martínez, R. Romero, J.C. López, A. Romero, V. Sánchez Mendieta,
R. Natividad, Ind. Eng. Chem. Res. 50 (2011) 2665–2670.
48] M.I. Sosulnikov, Y.A. Teterin, J. Electron Spectros. Relat. Phenomena 59 (1992)
111–126.
201710904064).
Appendix A. Supplementary data
49] F. Jiao, A. Harrison, J.-C. Jumas, A.V. Chadwick, W. Kockelmann, P.G. Bruce, J.
Am. Chem. Soc. 128 (2006) 5468–5474.
[50] M. Rezaei, S.M. Alavi, S. Sahebdelfar, P. Bai, X. Liu, Z.-F. Yan, Appl. Catal. B
Environ. 77 (2008) 346–354.
51] Q.J. Chen, J. Zhang, Q.W. Jin, B.R. Pan, W.B. Kong, T.J. Zhao, Y.H. Sun, Catal.
Today 215 (2013) 251–259.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2019.117313.
[
References
[
[
52] P.B. Weisz, Science 179 (1973) 433–440.
53] F.S.H. Simanjuntak, J.S. Choi, G. Lee, H.J. Lee, S.D. Lee, M. Cheong, H.S. Kim,
H. Lee, Appl. Catal. B Environ. 165 (2015) 642–650.
[
1] A. Taghizadeh-Alisaraei, H.A. Assar, B. Ghobadian, A. Motevali, Renew. Sust.
Energ. Rev. 72 (2017) 510–522.
[
[
[
54] K. Hu, H. Wang, Y. Liu, C. Yang, J. Ind. Eng. Chem. 28 (2015) 334–343.
55] P. Liu, M. Derchi, E.J.M. Hensen, Appl. Catal. B Environ. 144 (2014) 135–143.
56] P. Kumar, P. With, V. Srivastava, R. Gläser, I. Mishra, Ind. Eng. Chem. Res. 54
[2] G.M. Lari, G. Pastore, M. Haus, Y. Ding, S. Papadokonstantakis, C. Mondelli,
J. Pérez-Ramírez, Synth. Lect. Energy Environ. Technol. Sci. Soc. 11 (2018)
(2015) 12543–12552.
12