High-efficiency and low-cost Li/ZnO catalysts for synthesis of glycerol carbonate from glycerol transesterification: The role of Li and ZnO interaction

Article abstract of DOI:10.1016/j.apcata.2016.12.019

A series of efficient and low-cost Li/ZnO catalysts were prepared by a simple impregnation method and investigated for the synthesis of glycerol carbonate (GC) from the transesterification of glycerol with dimethyl carbonate (DMC). The Li/ZnO catalysts were characterized using XRD, SEM, FT-IR, TG-DSC, TPD and XPS. It was found that the basicity of the catalysts highly depended on the Li loading and calcination temperature. The weak and moderate basic sites on the catalyst surfaces originated from the ZnO and Li⁺interaction. The strong basic sites were attributed to the substitution of Zn²⁺by Li⁺in the ZnO lattice, which led to straining of Zn-O bonds and the formation of [Li⁺O^?] species. It was the strong basic sites rather than the weak and moderate basic sites that catalyzed the transesterification of glycerol with DMC. The highest catalytic activity was observed over the ZnO loaded with 1 wt.% LiNO₃and calcined at 500 °C. Glycerol conversion of 97.40% and GC yield of 95.84% were obtained over this catalyst at 95 °C in 4 h.

Full text of DOI:10.1016/j.apcata.2016.12.019

Applied Catalysis A: General 532 (2017) 77–85
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Feature Article
High-efficiency and low-cost Li/ZnO catalysts for synthesis of glycerol
carbonate from glycerol transesterification: The role of Li and ZnO
interaction
Xianghai Song, Yuanfeng Wu, Fufeng Cai, Donghui Pan, Guomin Xiao∗
School of Chemistry and Chemical Engineering, Southeast University, Nanjing 211189, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of efficient and low-cost Li/ZnO catalysts were prepared by a simple impregnation method and
investigated for the synthesis of glycerol carbonate (GC) from the transesterification of glycerol with
dimethyl carbonate (DMC). The Li/ZnO catalysts were characterized using XRD, SEM, FT-IR, TG-DSC, TPD
and XPS. It was found that the basicity of the catalysts highly depended on the Li loading and calcination
temperature. The weak and moderate basic sites on the catalyst surfaces originated from the ZnO and
Li+ interaction. The strong basic sites were attributed to the substitution of Zn2+ by Li+ in the ZnO lattice,
which led to straining of Zn-O bonds and the formation of [Li+ O−] species. It was the strong basic sites
rather than the weak and moderate basic sites that catalyzed the transesterification of glycerol with DMC.
The highest catalytic activity was observed over the ZnO loaded with 1 wt.% LiNO3 and calcined at 500 ◦C.
Glycerol conversion of 97.40% and GC yield of 95.84% were obtained over this catalyst at 95 ◦C in 4 h.
© 2016 Elsevier B.V. All rights reserved.
Received 29 August 2016
Received in revised form
12 December 2016
Accepted 22 December 2016
Available online 23 December 2016
Keywords:
Li/ZnO
[Li+ O−] species
Transesterification
Glycerol carbonate
1. Introduction
ence of Lewis acid catalysts [13,14]. However, the byproduct NH3
needs to be removed continuously during the reaction, so as to
Biodiesel has received growing attention in recent years as a
promising alternative to fossil fuel [1]. However, rapid develop-
ment in the biodiesel industry has led to the accumulation of
abundant glycerol (a byproduct accounting for about 10 wt.% of
the total biodiesel product), which results in the dramatic decrease
of glycerol price and jeopardizes the biodiesel production econ-
omy [2]. Thus, it is urgent to find new applications of glycerol.
Recent research emphasis has been focused on converting glycerol
to other valuable chemicals [3–5]. Among various glycerol deriva-
tives, glycerol carbonate (GC) is found to be a useful product of great
importance [6]. GC, due to its low flammability, low toxicity, high
boiling point and biodegradability, is an important cosmetic ingre-
dient and used for the manufacture of coating, biolubricant, glycidol
(GD), polycarbonates, polyesters and polyurethanes [7–10].
improve the GC yield. Production of GC from glycerol and CO2
under super-critical conditions is an ideal pathway, because of the
low costs of raw materials, high atom utilization efficiency, and no
other byproducts except H2O [15,16]. However, the utilization of
this method is limited by thermodynamic disadvantages, poor CO2
activity and low product yield. The transesterification of glycerol
with dimethyl carbonate (DMC) over basic catalysts is an impor-
tant and preferred method. Typically, the reaction proceeds under
mild conditions with high GC yield, and hardly any side reaction.
Among the methods mentioned above, the transesterification of
glycerol is a promising one, and has received much research atten-
tion in recent years. A number of catalysts, including homogeneous
and heterogeneous catalysts, have been studied for the glycerol and
DMC transesterification to produce GC. Compared with heteroge-
neous catalysts, homogeneous catalysts are generally more reactive
and have a much higher TOF. Homogeneous catalysts such as K2CO3
[7], NaOH [17] and CH3OK [18] have been reported to catalyze this
reaction. However, a serious drawback accompanied with homo-
geneous catalysts is their hard separation from the product. In any
case, if the quantity of catalyst is minute, scarce problems are to
be expected if left in the product mixture. On the other side, het-
erogeneous catalysis is very interesting and of great importance.
Various heterogeneous catalysts, such as CaO [19], CaO-La2O3 [20]
KF-hydroxyapatite [21], K2CO3-MgO [22], ZnO/La2O3 [23], Mg-Al
Various approaches have been employed to synthesize GC from
glycerol (Scheme 1). The carbonylation of glycerol with phosgene
or carbon monoxide is a traditional method in the GC synthesis
[11,12], but it has been abandoned due to the high toxicity of
reactants. Another method is the glycerolysis of urea in the pres-
∗
Corresponding author at: Southeast University, Road No. 2, Jiangning District,
Nanjing 211189, China.
E-mail address: Xiaogm@seu.edu.cn (G. Xiao).
http://dx.doi.org/10.1016/j.apcata.2016.12.019
0926-860X/© 2016 Elsevier B.V. All rights reserved.

78
X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
ambient temperature for 12 h. After removing the water on a rotary
evaporator, the obtained white solid was dried at 100 ◦C for 10 h and
then calcined at an appropriate temperature for 5 h to get the corre-
sponding catalyst. The obtained catalyst was named as nLi/ZnO-T,
where n and T represented LiNO3 loading and calcination temper-
ature, respectively. For example, 5 wt.% LiNO3 supported on ZnO
calcined at 300 ◦C was denoted as 0.05Li/ZnO-300.
Other catalysts, such as Li/ZrO2, Li/␥-Al2O3, K/ZnO and Cs/ZnO,
were prepared by loading 10 wt.% alkaline metal salts on the sup-
port in the same way as Li/ZnO and calcined at 500 ◦C for 5 h.
Scheme 1. Methods of preparing glycerol carbonate from glycerol.
2.3. Catalyst characterization
hydrotalcite [24] and ionic liquids [25] have been used to syn-
thesize GC from glycerol and DMC. However, there are still some
problems need to be addressed for heterogeneous catalysts, such as
poor reusability, high reaction temperature, requirement of addi-
tional solvent, long reaction time, and requirement of high glycerol
to DMC molar ratio. A new strategy needs to be fostered, with the
aim of getting higher activities and stabilities at moderate operating
conditions. And new catalysts still need to be explored.
It was reported that, doping with lithium could improve the
basicity of solid base catalysts, therefore enhancing the catalytic
activity in base-catalyzed reactions [26,27]. Recently, synthesis of
GC from glycerol and dimethyl carbonate over LiNO3/Mg4AlO5.5
was reported by Liu et al. [28]. Full glycerol conversion and 96.28%
GC yield were achieved in this reaction, which was far superior to
that of Mg4AlO5.5 with only 52.09% glycerol conversion. Lithium
incorporation onto MgO was reported to create the strongest basic
sites compared with other alkali or alkaline earth metal ions,
because of the higher Li molar concentration and the ion size effect
[29,30].
X-ray diffraction (XRD) patterns of the catalysts were recorded
on an Ultima IV X-ray diffractometer using Cu K␣ radiation (40 kV
and 40 mA) as the X-ray source. The scanning range (2␪) was from
20 to 75◦, with a scanning rate of 20◦ min−1 and a step function of
0.02.
The morphologies of the synthesized catalysts were studied by
scanning electron microscopy (SEM) images using a Philips XL-30
ESEM operating at a voltage of 10 kV.
The surface functional groups of the catalysts were determined
by fourier transform infrared spectrometer (FT-IR) using a Nicolet
5700 spectrometer. FT-IR spectra were recorded in the wavenum-
ber range of 400–4000 cm−1 with KBr pellets as a reference for the
measurements.
Thermogravimetric and differential scanning calorimeter (TG-
DSC) were conducted with a SDT 2960 DSC instrument. The
catalysts (about 5 mg) were placed into an aluminum pan, and the
experiment was performed under air atmosphere ranging from 50
to 800 ◦C at a heating rate of 20 ◦C per minute.
Li-doped ZnO has found its application in many fields, such as
conductive coating, electrodes for dye-sensitized solar cells, and
field emission materials. But only few authors have studied it as
a catalyst for various reactions [31]. As far as we know, Li-doped
ZnO catalyst has not been studied in GC synthesis from glycerol
transesterification. Thus, in this work, Li-modified ZnO as solid base
catalysts were prepared by impregnating ZnO in LiNO3 solution,
and their catalytic activities were investigated. The properties of
the Li/ZnO catalysts were characterized by X-ray diffractometry
(XRD), thermogravimetric-differential scanning calorimetry (TG-
DSC), Fourier transform infrared spectrometry (FT-IR), scanning
electron microscopy (SEM), CO2-temperature-programmed des-
orption (TPD) and X-ray photoelectron spectroscopy (XPS). Besides,
the structure and activity correlations of Li/ZnO catalysts for the GC
synthesis were also discussed.
The basicity of the catalysts was measured by CO2 temperature
programmed desorption (CO2-TPD). In a typical experiment, the
catalyst (100 mg) was pretreated in a flow of He (30 mL min−1) at
200 ◦C for 1 h to remove moisture and other adsorbed gases. After
cooling to 100 ◦C, the catalyst was exposed to pure CO2 for 0.5 h and
then purged with He flow (30 mL min−1) for 1 h to exclude physi-
cally adsorbed CO2. Subsequently, the sample was heated to 800 ◦C
at a rate of 10 ◦C min−1 and the desorbed CO2 was detected using a
thermal conductivity detector.
X-ray photoelectron spectroscopy (XPS) measurements were
performed using a Thermo Fisher Scientific ESCALAB 250Xi X-ray
photoelectron spectrometer. Mg K␣ radiation (1253.6 eV) was used
as the X-ray source at an ultrahigh pressure of 3.0 × 10−7 mbar. The
collected elements binding energy (BE) were calibrated by refer-
encing the C 1s peak at 284.6 eV. The precision of BE was within
0.1 eV.
2. Experimental section
2.1. Materials
2.4. Catalytic activity test
ZnO (AR), was purchased from Xilong Chemical Co., Ltd. LiNO3
(99.9 wt%), ␥-Al2O3 (99.9 wt.%), anhydrous KF (99.5 wt%), CsF
(99.0 wt%) were obtained from Aladdin Industrial Corporation,
Shanghai, China. ZrO2 (AR), dimethyl carbonate (DMC) (99.0 wt%),
glycerol (99.0 wt.%), were purchased from Sinopharm Chemical
Reagent Co., Ltd., Shanghai, China.
The transesterification of Glycerol with DMC was performed
in a 50 mL round bottomed flask equipped with a condenser at
atmospheric pressure. In a typical experiment, 50 mmol (4.6 g) of
glycerol and 100 mmol (9.0 g) of DMC were placed into the flask,
followed by 0.23 g (5 wt.%) of catalyst. Subsequently, the reaction
system was heated to the desired temperature for a designed time.
After the reaction, the catalyst was separated from the reaction
system by centrifugation. The product was analyzed using a gas
chromatograph (GC-6890, China), equipped with a flame ionization
detector and a capillary column (SE-30, 30 m × 0.25 mm). Ethylene
glycol monobutyl ether was added to the product as an internal
standard substance for the quantification analysis.
2.2. Catalyst preparation
Li-doped ZnO catalysts used in this work were prepared by a
simple wet impregnation method. In a typical preparation, 5 g of
ZnO was impregnated using 50 mL of LiNO3 aqueous solution of
appropriate concentration. The result mixture was then stirred at

X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
79
Fig. 1. The XRD patterns of ZnO and ZnO with different LiNO3 loading calcined at 500 ◦C: (A) 20–75◦, and (B) 34–37◦; The XRD patterns of ZnO and 1 wt.% LiNO3 doped ZnO
calcined at different temperatures: (C) 20–75◦, and (D) 34–37◦.
3. Results and discussion
3.1. Catalyst characterization
XRD patterns of ZnO and Li-doped ZnO powders with differ-
ent Li+ contents and calcined at 500 ◦C for 5 h are presented in
Fig. 1. Clearly, all samples are crystalline materials with hexago-
nal wurtzite crystal structures (JCPDS card No. 36-1451) (Fig. 1A).
However, no signals of LiNO3 or lithium oxides due to Li doping are
detected. This is probably attributed to the high dispersion degree
of LiNO3 on the ZnO surface, which leads to the Li concentration
below the instrumental detection limitation. Since the scattering
factor of an atom is affected by its atomic number, the lighter atoms
with weak diffraction are undetectable in the presence of heavier
atoms [32]. Thus, the high scattering factor of Zn2+ to Li+ may also
Scheme 2. Incorporation of Li atoms into the ZnO structure.
account for the absence of diffraction peaks of lithium compounds.
To study the doping effect of Li+ into the ZnO lattice, the XRD
34–37◦ were also acquired and presented in Fig. 1D. Interestingly,
patterns in the 2␪ region 34–37◦ was obtained and shown in Fig. 1B.
compared with pure ZnO, the (101) peak at 36.46◦ of Li-doped
The (002) and (101) peak of Li-doped ZnO shift to lower diffraction
ZnO firstly shifts to lower diffraction angle, minimizes to 36.36◦
angles with increasing Li+ content compared with pure ZnO. This
at 300 ◦C, then it shifts to higher 2␪, maximizes to 36.48◦ at 600 ◦C,
may result from the substitution of Zn2+ (74 pm) by Li+ (76 pm)
and finally returns back to 36.38◦ at 700 ◦C. This phenomenon may
in the ZnO crystal lattice, which causes asymmetry in the crystal
be ascribed to the interaction of ZnO with the supported Li+. At
structure [33]. The biggest peak shift appears in the powder doped
low temperature, Li+ is inactive enough to substitute the Zn2+ in
with 5 wt.% LiNO3, but the peak shift does not change with further
ZnO lattice. In comparison, the Li+ and ZnO interaction may lead
increase of Li content, even at the load of 20 wt.% LiNO3. The same
to the formation of an intermediate state, resulting in the increase
result was also obtained by Wang et al. [34]. Based on these facts,
of the Zn-O bond length and asymmetry in ZnO lattice. The longer
we assume that Li atoms enter ZnO crystals and substitute partial
Zn-O bond and asymmetric ZnO lattice may account for the lower
Zn atoms up to a certain concentration, beyond which Li atoms
diffraction angle. After the substitution of Zn2+ by Li+ at 600 ◦C, the
aggregate on the boundaries of ZnO grains as a secondary phase
diffraction angle returns to the similar position with ZnO due to
(Scheme 2).
the similar ion radii between Zn2+ and Li+. The (101) peak moves to
Fig. 1C shows the XRD patterns of ZnO loaded with 1 wt.% LiNO3
lower diffraction angle again after treatment at 700 ◦C, which may
and calcined at different temperatures. The diffraction peaks of
be assigned to the appearance of single phase of ternary Li2xZn1-x
[35].
O
ZnO catalysts almost remain unchanged after treatment at various
temperatures from 100 to 700 ◦C, except that the peaks are inten-
sified with the temperature rise. This may be due to the growth of
ZnO crystals at high temperature. In order to gain a deeper insight
into the effect of calcination temperature, the XRD patterns within
The SEM graphs of Li-doped ZnO powder are presented in Fig. 2.
Clearly, all the samples present uniform cobblestone-like mor-
phology with varying grain sizes. The average particle sizes of
0.01Li/ZnO-500 and 0.1Li/ZnO-500 are 278 and 306 nm, respec-

80
X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
Fig. 2. SEM micrographs of Li doped ZnO (A) 0.01Li/ZnO-500; (B) 0.1Li/ZnO-500; (C) 0.01Li/ZnO-700; (D) 0.01Li/ZnO-Re.
tively (Fig. 2A and B), suggesting that the particle size slightly
grows with the increasing Li content. However, the particle size
for 0.01Li/ZnO-700 reaches 1.77 ␮m (Fig. 2C), implying the calcina-
tion temperature has significant impact on particle size. The result
indirectly proves the XRD results: a new phase with larger crystals
is formed at higher temperature. There are no remarkable changes
in 0.01Li/ZnO-500 after four runs (Fig. 2D), which means the mor-
phology of the catalyst is rather stable under the tested reaction
conditions.
The FT-IR spectra of the catalysts are shown in Fig. 3. The intense
broad band within 400–600 cm−1 observed in all the samples is
assigned to the oxygen sublattice vibration in ZnO [36], while the
peaks at around 1385–1405 cm−1 in pure ZnO (Fig. 3a) may be
ascribed to the carbonate (CO32−) [30,37]. The absorption band
at 1385 cm−1, growing in line with the increasing LiNO3 c−ontent
(Fig. 3b–d), is attributed to the vibration of N O in NO3 [38].
Despite the similar locations with N O, it is easy to tell from the
absorption band of CO32− due to its double peak. After doping with
2−
−
LiNO3, the double peak of CO3 is covered by N O peak of NO3
.
Besides, the N O peak becomes weaker with temperature rises and
disappears from the catalyst calcine−d at 700 ◦C (Fig. 3e), implying
the complete decomposition of NO3 upon calcination at elevated
temperature. The absorption band at 1636 cm−1, which is ascribed
to the O H bending vibration of H2O [38], is intensified with the
rising LiNO3 loading. This may be ascribed to the easy deliquesce
of LiNO3 and also indicates the incomplete decomposition of LiNO3
calcined at 500 ◦C. The broad band at 3423 cm−1 can be assigned
to the O H stretching mode of hydroxyl group. No clear absorp-
Fig. 3. FT-IR spectra of (a) ZnO; (b) 0.01Li/ZnO-500; (c) 0.05Li/ZnO-500; (d)
0.1Li/ZnO-500; (e) 0.01Li/ZnO-700; (f) 0.01Li/ZnO-500 after four cycles.
−
Besides, no NO3 absorption band is observed in the recovered
catalyst (Fig. 3f), indicating the leaching of NO3− from ZnO surfaces.
The catalysts were analyzed via TG-DSC to study the thermal
behavior of ZnO loaded with LiNO3. The 0.05Li/ZnO is quite stable
and shows no weight loss up to 483 ◦C (Fig. 4A). However, there
are two endothermic peaks at 153 and 255 ◦C respectively during
this process, which may be attributed to the Li+ and ZnO interac-
tion that leads to the appearance of basic sites on ZnO surfaces. At
tion bands ascribed to NO2 at 1550 and 1320 cm−1 are observed,
−
indicating the product of LiNO3 decomposition is not LiNO2 [39].

X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
81
Fig. 5. CO2-TPD profiles of (a) ZnO (b) 0.01Li/ZnO-300 (c) 0.01Li/ZnO-500 (d)
0.01Li/ZnO-500 without CO2 adsorption (e) 0.01Li/ZnO-700 (f) 0.03Li/ZnO-500 (g)
0.05Li/ZnO-500 (h) 0.05Li/ZnO-500 without CO2 adsorption.
orption peak of strong basic sites grows larger (Fig. 5c). However,
after calcination at 700 ◦C, none of the desorption peaks is left
(Fig. 5e), indicating that the basicity of the catalyst highly depends
on the processing temperature. As reported, the presence of Mg2+
-
O2− pair is the main cause for the basic sites in magnesium oxide
[41,42]. Cosimo et al. discovered that MgO loaded with Li+ shows
more and stronger basic sites, and ascribed this to the strained
Mg-O bonds and the formation of [Li+ O−] species resulting from
the substitution of Mg2+ by Li+ in the MgO lattice [29]. The CO2
desorption peak within 650–800 ◦C grows with the increasing Li+
concentration (Fig. 5c, f, g), implying the number of basic sites is
in accordance with the Li+ concentration[30]. Together with the
characterizations, it is reasonable to make four hypotheses. (1) The
addition of Li+ improves the partial negative charge of oxygen ions
in ZnO due to the lower ionization potential of Li compared with
Zn, which leads to the presence of weak and moderate basic sites
on ZnO surfaces. (2) The strong basic sites originate from [Li+ O−]
which is resulted from the substitution of Zn2+ by Li+ in the ZnO lat-
tice at higher temperature. (3) The weak and moderate basic sties
convert to strong basic sites through the substitution of Zn2+ by Li+
with temperature rise, as confirmed by the disappearance of the
peak at around 199 ◦C, the decrease of peak area at 355 ◦C and the
enlargement of peak area at 715 ◦C (Fig. 5b, c). (4) The strong basic
sites are destroyed at high temperature.
Fig. 4. TG-DSC curves of (A) 0.05Li/ZnO without calcination; (B) a: 0.01Li/ZnO-500,
b: 0.01Li/ZnO-500 after four runs. The solid lines indicate the TG curves and dashed
lines indicate DSC curves.
483–645 ◦C, the 0.05Li/ZnO undergoes a remarkable weight loss
and exhibits a drastic−DSC endothermic peak, which is mainly
attributed to the NO3 decomposition. Besides, the 0.01Li/ZnO-
500 shows slight weight loss within 300–800 ◦C, which may be
assigned to the decomposition of residual NO3− (curve a in Fig. 4B).
Since the catalytic reactions were performed at ambient conditions,
these catalysts were expected to be quite stable under the tested
reaction conditions. The recovered catalyst exhibits two stages
of drastic weight loss at 155–250 and 330–467 ◦C, respectively,
which are assigned to the loss of loosely-adsorbed low-boiling-
point compounds, and the decomposition of high-boiling-point
organic compounds adsorbed on catalyst surfaces, respectively.
The CO2-TPD measurements determining the basic strength dis-
tribution and total basicity of the catalysts are shown in Fig. 5. ZnO
exhibits almost no CO2 desorption peak, indicating there is no basic
site on the surfaces of pure ZnO (Fig. 5a). To exclude the influence of
NO3− decomposition, blank run of 0.01Li/ZnO-500 and 0.05Li/ZnO-
500 without CO2 adsorption was studied as control. Compared with
the catalysts with CO2 adsorption, both of these two blank runs
show a peak at around 550 ◦C, besides, no other peak was observed
(Fig. 5c, d, g and h). Hence, the peak at around 550 ◦C should be
excluded. The catalyst calcined at 300 ◦C shows three peaks at 199,
355 and 715 ◦C, which represents the desorption of CO2 from weak,
moderate and strong basic sites, respectively (Fig. 5b) [40].
The wide XPS spectra for ZnO and 10 wt.% LiNO3-impregnated
ZnO sample (0.1Li/ZnO-500) are presented in Fig. 6. The ZnO sur-
faces comprise of Zn, O and a trace amount of C. The Li 1s peak
can only be seen in the zoomed-in spectra, which may be due to
the incorporation of Li+ into ZnO crystal lattices or to the low Li
content in ZnO structure. N and C signals are obviously observed
upon impregnation with LiNO3, accompanied by an increase of O 1s
signal intensity. The C signal may originate from the CO2 adsorbed
on the basic sites or the formation of Li2CO3. The remarkable peak
intensity of N 2p, with respect to the Li peak, implies that the major-
−
ity of NO3 remains on the ZnO surfaces.
The BE of O 1s is shown in Fig. 7A. The O 1s spectra for ZnO
samples exhibit an asymmetric peak at 529.8 eV corresponding to
lattice O2− [31], the asymmetry may arise from the existence of
the surface defects [43]. After doping with LiNO3, 0.01Li/ZnO-500
exhibits a peak at around 531.7 eV (Fig. 7A-b), which may be a
mixed peak of various O species, such as surface defects, NO3− and
[Li+ O−]. The intense peak at 532.8 eV for 0.1Li/ZnO-500 is assigned
−
The desorption peaks representing weak and moderate basic
to NO3 (Fig. 7A-d) [26], implying the ZnO surfaces are covered by
sites almost disappear after calcination at 500 ◦C, while the des-
abundant NO3−, as confirmed by the IR spectra. Besides, the O 1s

82
X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
The Zn 2p BE in the recovered 0.01Li/ZnO-500 is consistent with a
fresh catalyst (Fig. 7B-e). Given the low activity of the recovered cat-
alyst, it is clear the active center is not related to Zn. Also the Zn 2p
peak intensity is sharply decreased due to the organic compounds
adsorbed on ZnO surfaces.
No Li 1s XPS peak is observed for 0.01Li/ZnO calcined at 500 or
700 ◦C (Fig. 7C-b, c), implying the incorporation of Li into ZnO crys-
tal lattices. The 0.1Li/ZnO-500 shows an obvious Li peak (Fig. 7C-d),
suggesting that Li is modestly incorporated into the ZnO crystal lat-
tice, and beyond some amount, the excessive Li is dispersed on the
ZnO surfaces. The small N 1s XPS peak at 407.21 eV for 0.01Li/ZnO-
500 disappears after calcination at 700 ◦C (Fig. 7D-b, c), indicating
the complete decomposition of NO3− at 700 ◦C, which is consistent
with the IR results. An intense N 1 s XPS peak appears when the
LiNO3 loading increases to 10 wt.% (Fig. 7D-d). This may originate
from the accumulation of excessive LiNO3 on the edge of ZnO parti-
cles, which complicates the decomposition of LiNO3, in accordance
with the XRD and O 1s XPS results.
Fig. 6. Full XPS spectra for (a) ZnO (b) 0.1Li/ZnO-500.
BE for 0.01Li/ZnO calcined at 700 ◦C increases compared with pure
ZnO (Fig. 7A-c), indicating the formation of new species, as con-
firmed by the XRD patterns. The recovered 0.01Li/ZnO-500 presents
a broad peak at around 531.6 eV, and the O 1s BE is 0.2 eV higher
than pure ZnO (Fig. 7A-e). This may be assigned to the organic com-
pounds adsorbed on catalyst surfaces, which is consistent with the
TG results.
The BE of Zn 2p1/2 for ZnO is centered at 1044.16 eV, but it
increases to 1044.42 eV for 0.01Li/ZnO-500 (Fig. 7B-a, b), imply-
ing the change of Zn chemical environment and the occurrence of
Zn and Li interaction. However, the BE of Zn 2p1/2 decreases to
1044.26 eV with the Li loading increase to 10 wt.%, accompanied by
the dramatic decrease of the Zn 2p peak intensity (Fig. 7B-d), indi-
cating complex Li and ZnO interaction at high Li+ loading. Moreover,
the Zn 2p BE of 0.01Li/ZnO remains unchanged after thermal treat-
ment at 700 ◦C compared with the one calcined at 500 ◦C (Fig. 7B-c).
This result together with the BE of O 1s suggests that the interaction
between Li atoms and O atoms mainly occurs at high temperature.
3.2. Catalytic activities
3.2.1. Screening of catalysts
In the present work, the catalytic activity of various catalysts
was tested through the reaction of glycerol with DMC, by means
of glycerol conversion and GC selectivity. As reported, alkali metal
oxides could improve the base strength of catalyst [26,29,44]. Based
on this, a serious of alkali metals loaded on different supports were
prepared and investigated for their catalytic activity towards glyc-
erol transesterification (Table 1). Clearly, among all the catalysts
investigated, Li/ZnO gave the best result with glycerol conversion of
92.12% and GC selectivity of 96.69%. Li/␥-Al2O3 and K/ZnO showed a
lower glycerol conversion compared with Li/ZnO. Despite the high
glycerol conversion, Cs/ZnO presented a quite low GC selectivity
with a GC yield of only 77.66%. Li/ZrO exhibited a slightly lower
glycerol conversion and GC selectivity compared with Li/ZnO, with
GC yield of 83.75%. Hence, Li/ZnO was selected as the catalyst.
Fig. 7. XPS spectrum of (A) O 1s, (B) Zn 2p, (C) Li 1s, (D) N 1s. (a) ZnO (b) 0.01Li/ZnO-500 (c) 0.01Li/ZnO-700 (d) 0.1Li/ZnO-500 (e) 0.01Li/ZnO-Re.

X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
83
Table 1
with glycerol conversion of 90.99% and GC selectivity of 96.35%.
The catalyst calcined at 600 ◦C presents almost the same glycerol
conversion of 89.27% with the catalyst calcined at 500 ◦C, but with a
lower GC selectivity of only 90.72%. This may be due to the forma-
tion of a few stronger basic sites (Li2O) at elevated temperature,
which leads to the decarbonylation of GC as mentioned above.
However, after treatment at 700 ◦C, the catalytic activity is dra-
matically reduced with glycerol conversion of only 29.36%, which is
ascribed to the disappearance of basic sites on the catalyst as shown
by TPD. Besides, the sinter of the catalyst may also contribute to the
low activity as depicted by SEM.
Screening of the catalysts for preparation of glycerol carbonate.a
Catalystb
Conv. (%)
Sel. (%)
GD
GC Yield (%)
GC
Li/␥-Al2O3
Li/ZrO2
Li/ZnO
K/ZnO
Cs/ZnO
83.39
87.65
92.12
84.64
90.24
4.84
3.90
3.31
7.56
13.94
95.16
95.55
96.69
92.44
86.06
79.35
83.75
89.07
78.24
77.66
a
Reaction conditions: glycerol: 50 mmol (4.6 g), DMC: 100 mmol (9.0 g), catalyst:
0.23 g, temperature: 95 ◦C, time: 4 h.
b
Though the catalyst calcined at 300 ◦C presents a large amount
of moderate basic sites and a few strong basic sites, it gave a glycerol
conversion of only 52.13%. The catalyst calcined at 500 ◦C shows a
few strong basic sites and almost no moderate basic sites but far
better activity compared with 0.01Li/ZnO-300. These results fur-
ther confirm our hypothesis: it is the strong basic sites rather than
moderate basic sites that catalyze the transesterification of glycerol
with DMC.
The catalysts were prepared by loading 10 wt.% of alkali metal salt on support
and calcined at 500 ◦C for 5 h.
Table 2
Effect of LiNO3 loading and catalyst calcination temperature on catalytic
performance.a
Catalyst
Conv. (%)
Sel. (%)
GD
GC Yield (%)
GC
ZnO
LiNO3
2.53
4.56
4.14
3.65
4.11
5.79
7.13
7.29
9.28
1.41
95.44
95.86
96.35
95.89
94.21
92.87
92.71
90.72
98.59
2.42
15.81
90.99
91.80
92.12
14.34
52.14
89.27
29.36
15.16
87.67
88.03
86.79
13.32
48.34
80.98
28.95
3.3. Effect of reaction parameters
0.01Li/ZnO-500
0.05Li/ZnO-500
0.1Li/ZnO-500
0.01Li/ZnO-100
0.01Li/ZnO-300
0.01Li/ZnO-600
0.01Li/ZnO-700
3.3.1. Effect of reaction temperature
As the catalyst 0.01Li/ZnO pretreatment at 500 ◦C exhibited the
best catalytic activity, it was further studied to evaluate the reac-
tion parameters. Fig. 8A shows the effect of reaction temperature,
ranging from 65 to 105 ◦C, on the transesterification of glycerol
with DMC. The glycerol conversion was only 16.96% at 65 ◦C, and
it increased with an increase in temperature and reached a maxi-
mum of 97.4% at 95 ◦C. There was no appreciable improvement in
glycerol conversion beyond this temperature, but the GD selectiv-
ity was improved slightly. Accordingly, it was proposed that high
reaction temperature might favor the decarbonylation of GC, which
led to the formation of GD.
a
Reaction conditions: glycerol: 50 mmol (4.6 g), DMC: 100 mmol (9.0 g), catalyst:
0.23 g, temperature: 95 ◦C, time: 4 h.
3.2.2. Effect of LiNO3 loading on the catalysts
The effect of LiNO3 loading is shown in Table 2. Clearly, the
glycerol conversions over pure ZnO and LiNO3 were only 2.53%
and 15.81%, respectively. However, the glycerol conversion of
0.01Li/ZnO was markedly improved to 90.99% with GC selectiv-
ity of 96.35%. The glycerol conversion was not obviously increased
any more with further increase of Li content. However, the GC
selectivity gradually decreased with the increasing Li amount,
accompanied by the increased selectivity in GD, which was formed
from the GC decarbonylation.
According to the TPD results, 0.01Li/ZnO-500 has only a small
amount of strong basic sites, and almost no moderate basic sites.
Besides, both the strength and amount of basic sites increase
with the rising Li content. Together with the catalytic activity of
0.01Li/ZnO-500, we suggest that strong basic sites may mainly
account for the transesterification of glycerol with DMC. However,
a larger amount of the byproduct GD would be produced with the
increasing amount and strength of strong basic sites, which means
the decarbonylation is facilitated by the basicity of catalyst [40]. As
a result, the catalyst composition was set as 1 wt.% LiNO3 supported
on ZnO.
3.3.2. Effect of reaction time
The effect of reaction time from 1 to 5 h on the reaction of glyc-
erol with DMC was investigated (Fig. 8B). The glycerol conversion
is enhanced with the reaction time prolonged from 1 to 4 h, and
it is maximized to 97.63% in 4 h. Both the glycerol conversion and
GC selectivity decrease when the reaction time is prolonged to 5 h.
Moreover, the GD selectivity increases with time, which implies the
newly-formed GC is gradually converted to GD [40]. These results
suggest that the catalyst could catalyze both glycerol transesterifi-
cation, which is absolutely dominant, and GC decarbonylation. As
the glycerol transesterification reaches equilibrium, the GC decar-
bonylation becomes dominant and the amount of the byproduct GD
gradually increases with time. Thus, the optimum reaction time is
4 h.
3.3.3. Effect of catalyst amount
3.2.3. Effect of calcination temperature on the catalysts
The effect of catalyst amount on the catalytic activity and selec-
tivity is illustrated in Fig. 8C. The glycerol conversion is steadily
promoted with an increase in the catalyst amount up to 5 wt.%,
beyond which it does not further increase. Meanwhile, the GC selec-
tivity is gradually decreased from 97.36% (1 wt.%) to 94.69% (9 wt.%)
with the increase of catalyst amount. On the contrary, the selec-
tivity towards GD is slowly elevated with an increase in catalyst
amount. This might be attributed to the increase in strong basic
sites with the rising catalyst amount, which leads to the decar-
bonylation of GC. These results indicate that for maximization of
GC yield, the amount of the catalyst should be appropriate in order
to prevent the formation of GD.
As known to all, the basic and acidic properties of mixed metal
oxides alter with a change in calcination temperature, in addi-
tion to their composition [45]. Due to the 0.01Li/ZnO-500 catalyst
showed the best catalytic performance, it was further studied to
know the relationship between its calcination temperature and
catalytic activity. The 0.01Li/ZnO was calcined within100–700 ◦C,
and investigated for the transesterification activity of glycerol with
DMC (Table 2). The 0.01Li/ZnO-100 shows a very low catalytic activ-
ity with a glycerol conversion of only 14.34%. The activity is rapidly
improved with the rising calcination temperature up to 500 ◦C, but
beyond that, it is reduced by further temperature rise. The catalyst
pretreated at 500 ◦C has the best activity among all the catalysts,

84
X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
Fig. 8. Effects of reaction parameters: (A) temperature, (B) time, (C) catalyst loading, (D) DMC to glycerol molar ratio. Reaction conditions: glycerol: 50 mmol (4.6 g), DMC:
100 mmol (9.0 g), catalyst: 5 wt.% (0.23 g), temperature: 95 ◦C, time: 4 h.
3.3.4. Effect of the molar ratio of DMC to glycerol
In the transesterification of glycerol with DMC, the product dis-
tribution is greatly affected by the molar ratio of reactants. As
shown in Fig. 8D, the glycerol conversion is rather low (70.50%)
at a low DMC to glycerol molar ratio (1:1). It rises with increase in
the DMC to glycerol molar ratio and maximizes to 97.40% at a molar
ratio of 3:1. The GC selectivity is enhanced while the selectivity for
GD declines as the DMC to glycerol molar ratio increases, which
might result from the low catalyst concentration at high DMC to
glycerol molar ratio. This result agrees with the studies on the effect
of catalyst loading. It also indicates high DMC to glycerol molar ratio
favors the formation of GC.
3.4. Reusability of the catalyst
Reusability is one major criterion to assess a heterogeneous cat-
alyst. The reusability of the 0.01Li/ZnO-500 is showed in Fig. 9. The
catalyst was recovered by simple centrifugation, and washed with
Fig. 9. Reusability of the 0.01Li/ZnO-500 for glycerol transesterification. Reaction
methanol to remove the adsorbed organic compounds, then oven-
conditions: glycerol: 4.6 g (50 mmol), DMC: 13.5 g (150 mmol), catalyst: 0.23 g, tem-
dried at 60 ◦C for 2 h and used for the next run. The results show
perature: 95 ◦C, time: 4 h.
that there is no obvious loss of catalytic activity after 3 cycles, but
GC selectivity increases from 97.40% to 99.49%. However, the glyc-
erol conversion significantly drops to 60.57% after the fourth cycle,
which may result from the leaching of the active component. It is
interesting that the selectivity to GD is decreased with an increase
in cycle number from 2.60% to 0.51%. As mentioned above, GD might
derive from the fewer stronger basic sites of the catalyst. Thus, it
is reasonable to hypothesize that the stronger basic sites are more
leachable than strong basic sites, which consequently leads to the
decrease of GD selectivity.
Table 3
Deactivation test of catalyst.a
Entry
Cycle
Conv. (%)
Sel. (%)
GD
GC Yield (%)
GC
1
1
2
3
4
–
–
96.12
94.36
91.86
46.95
58.71
83.18
2.46
1.43
1.15
0.00
2.07
8.76
97.54
98.57
98.85
100.00
97.93
91.24
93.76
93.00
90.81
46.95
57.49
75.89
2b
3b
4b
5c
6d
3.5. Research of catalyst deactivation
a
Reaction conditions: glycerol: 50 mmol (4.6 g), DMC: 100 mmol (9.0 g), catalyst:
0.23 g, temperature: 95 ◦C, time: 4 h.
To figure out the reason for catalyst deactivation, two groups
of experiment were performed (Table 3). In one group, the reused
catalyst was calcined at 500 ◦C for 5 h after each cycle to remove
the adsorbed product species (Entry 1–4). Clearly, no significant
improvement in both product yield and cycle number was observed
b
Reused catalyst calcined at 500 ◦C for 5 h.
c
Sample after 1 h reaction.
d
After 1 h reaction, the blank reaction without catalyst proceeded for another 3 h.

X. Song et al. / Applied Catalysis A: General 532 (2017) 77–85
85
compared with the reused catalyst without calcination (Fig. 9).
In the other group, the catalyst was filtered out after 1 h reac-
tion by hot filtration, and the blank reaction was continued for
another 3 h (Entry 5–6). The results showed that the reaction pro-
ceeds smoothly even after removing the catalyst. This means there
are still active sties in the reaction system, indicating the leach-
ing of active sites. Moreover, the selectivity towards GD is up to
8.76%, which confirms our hypotheses that stronger basic sites are
more leachable. Thus, it should be active sites leaching rather than
adsorbed product species that lead to the deactivation of catalyst.
[8] U. Chandrakala, R.B.N. Prasad, B.L.A. Prabhavathi Devi, Ind. Eng. Chem. Res. 53
(2014) 16164–16169.
[9] S. Sandesh, G.V. Shanbhag, A.B. Halgeri, Catal. Lett. 143 (2013) 1226–1234.
[10] H. Mei, Z. Zhong, F. Long, R. Zhuo, Macromol. Rapid Commun. 27 (2006)
1894–1899.
[11] A.G. Shaikh, S. Sivaram, Chem. Rev. 96 (1996) 951–976.
[12] M.O. Sonnati, S. Amigoni, E.P. Taffin de Givenchy, T. Darmanin, O. Choulet, F.
Guittard, Green Chem. 15 (2013) 283–306.
[13] S. Fujita, Y. Yamanishi, M. Arai, J. Catal. 297 (2013) 137–141.
[14] M.H.A. Rahim, Q. He, J.A. Lopez-Sanchez, C. Hammond, N. Dimitratos, M.
Sankar, A.F. Carley, C.J. Kiely, D.W. Knighta, G.J. Hutchings, Catal. Sci. Technol.
2 (2012) 1914–1924.
[15] J.X. Liu, Y.M. Li, J. Zhang, D.H. He, Appl. Catal. A 513 (2016) 9–18.
[16] H.G. Li, X. Jiao, L. Li, N. Zhao, F.K. Xiao, W. Wei, Y.H. Sun, B.S. Zhang, Catal. Sci.
Technol. 5 (2015) 989–1005.
[17] 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 336 (2009) 315–324.
[18] J. Esteban, E. Domínguez, M. Ladero, F. Garcia-Ochoa, Fuel Process. Technol.
138 (2015) 243–251.
[19] F.S.H. Simanjuntak, T.K. Kim, S.D. Lee, B.S. Ahn, H.S. Kim, L. Hyunjoo, Appl.
Catal. A 401 (2011) 220–225.
[20] P. Kumar, P. With, V.C. Srivastava, R. Gläser, I.M. Mishra, Ind. Eng. Chem. Res.
54 (2015) 12543–12552.
[21] R. Bai, S. Wang, F. Mei, T. Li, G.J. Li, J. Ind. Eng. Chem. 17 (2011) 777–781.
[22] M. Du, Q. Li, W. Dong, T. Geng, Y. Jiang, Res. Chem. Intermed. 38 (2012)
1069–1077.
[23] D. Singh, B. Reddy, A. Ganesh, S. Mahajani, Ind. Eng. Chem. Res. 53 (2014)
18786–18795.
[24] A. Takagaki, K. Iwatani, S. Nishimura, K. Ebitani, Green Chem. 12 (2010)
578–581.
[25] S.M. Gade, M.K. Munshi, B.M. Chherawalla, V.H. Rane, A.A. Kelkar, Catal.
Commun. 27 (2012) 184–188.
[26] R.S. Watkins, A.F. Lee, K. Wilson, Green Chem. 6 (2004) 335–340.
[27] A. Corma, S.B.A. Hamid, S. Iborra, A. Velty, J. Catal. 234 (2005) 340–347.
[28] Z.M. Liu, J.W. Wang, M.Q. Kang, N. Yin, X.K. Wang, Y.S. Tan, Y.L. Zhu, J. Ind. Eng.
Chem. 21 (2015) 394–399.
4. Conclusion
High-efficiency and low-cost Li/ZnO catalysts were prepared
for synthesis of glycerol carbonate from the glycerol transester-
ification. The catalytic activity of the catalysts depended on the
composition of the catalyst and the treatment temperature. The
0.01Li/ZnO calcined at 500 ◦C was more active compared with cata-
lysts with other composition and treated at other temperature. The
high catalytic activity was ascribed to the strong basic sites origi-
nating from the formation of [Li+ O−]. The weak and moderate basic
sites, which showed poor activity towards glycerol transesterifica-
tion, could be converted to strong basic sites by the substitution
of Zn+ by Li+ in the ZnO lattice at high temperature. Only a cer-
tain amount of Li+ can enter the ZnO lattices, beyond which the Li+
stays on the ZnO surfaces. The byproduct glycidol (GD) formed by
the decarbonylation of GC was also detected in the reaction cat-
alyzed by Li/ZnO. The glycerol conversion and product distribution
were greatly influenced by the reaction parameters.
[29] J.I. Di Cosimo, V.K. Díez, C.R. Apesteguía, Appl. Catal. A 137 (1996) 149–166.
[30] V.K. Díez, C.R. Apesteguía, J.I. Di Cosimo, Catal. Today 63 (2000) 53–62.
[31] W.L. Xie, Z.Q. Yang, H. Chun, Ind. Eng. Chem. Res. 46 (2007) 7942–7949.
[32] B.G. Mishra, G.R. Rao, J. Mol. Catal. A: Chem. 243 (2006) 204–213.
[33] C.Y. Kung, C.C. Lin, S.L. Young, L. Horng, Y.T. Shih, M.C. Kao, H.Z. Chen, H.H. Lin,
J.H. Lin, S.J. Wang, J.M. Li, Thin Solid Films 529 (2013) 181–184.
[34] B. Wang, L. Tang, J. Qi, H. Du, Z. Zhang, J. Alloys Compd. 503 (2010) 436–438.
[35] V. Pavlyuk, R. Misztal, W. Ciesielski, Eur. J. Inorg. Chem. (2014) 925–931.
[36] Z.J. Li, K.X. Zhu, Q. Zhao, A. Meng, Appl. Surf. Sci. 377 (2016) 23–29.
[37] S. Polarz, A. Orlov, A. Hoffmann, M.R. Wagner, C. Rauch, R. Kirste, W. Gehlhoff,
Y. Aksu, M. Driess, M.W.E. van den Berg, M. Lehmann, Chem. Mater. 21 (2009)
3889–3897.
Acknowledgements
This work was financially supported by the National Natu-
ral Science Foundation of China (Nos. 21276050 and 21406034),
Fundamental Research Funds for the central Universities (No.
3207045414), Key Laboratory Open Fund of Jiangsu Province
(JSBEM201409) and the Priority Academic Program Development
of Jiangsu Higher Education Institutions.
[38] K. Nakamoto, Infrared Spectra of Inorganic and Coordination Compounds,
Wiley, New York, 1970.
[39] I. Ganesh, P.S. Chandra Sekhar, G. Padmanabham, G. Sundararajan, Appl. Surf.
Sci. 259 (2012) 524–537.
References
[40] G. Parameswaram, M. Srinivas, B. Hari Babu, P.S. Sai Prasad, N. Lingaiah, Catal.
Sci. Technol. 3 (2013) 3242–3249.
[41] M. DiSrio, M. Ledda, M. Cozzolino, G. Minutillo, R. Tesser, E. Santacesaria, Ind.
Eng. Chem. Res. 45 (2006) 3009–3014.
[42] M.J. Climent, A. Corma, P.D. Frutos, S. Iborra, M. Noy, A. Velty, P. Concepctión,
J. Catal. 269 (2010) 140–149.
[43] R.K. Sahu, K. Ganguly, T. Mishra, M. Mishra, R.S. Ningthoujam, S.K. Roy, L.C.
Pathak, J. Colloid Interface Sci. 366 (2012) 8–15.
[1] A. Abbaszaadeh, B. Ghobadian, M.R. Omidkhah, G. Najafi, Energy Convers.
Manage. 63 (2012) 138–148.
[2] P.P. Oh, H. Lau, J.H. Chen, M.F. Chong, Y.M. Choo, Renew. Sustain. Energy Rev.
16 (2012) 5131–5145.
[3] H.C. Zhou, J.N. Beltramini, Y.X. Fan, G.Q.M. Lu, Chem. Soc. Rev. 37 (2008)
527–549.
[4] P.U. Okoye, B.H. Hameed, Renew. Sustain. Energy Rev. 53 (2016) 558–574.
[5] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C.D. Pina, Angew. Chem. Int. Ed.
46 (2007) 4434–4440.
[44] Y.T. Algoufi, U.G. Akpan, M. Asif, B.H. Hameed, Appl. Catal. A 487 (2014)
181–188.
[45] W. Xie, H. Peng, L. Chen, J. Mol. Catal. A: Chem. 246 (2006) 24–32.
[6] N. Rahmat, A.Z. Abdullah, A.R. Mohamed, Renew. Sustain. Energy Rev. 14
(2010) 987–1000.
[7] G. Rokicki, P. Rakoczy, P. Parzuchowski, M. Sobiecki, Green Chem. 7 (2005)
529–539.