Investigation of glycerolysis of urea over various ZnMeO (Me = Co, Cr, and Fe) mixed oxide catalysts

Article abstract of DOI:10.1016/j.cattod.2019.09.017

In this study, we investigated the glycerolysis of urea over various ZnMeO (Me = Co, Cr, and Fe) mixed oxide catalysts. ZnMeO mixed oxide catalysts were prepared by a co-precipitation method for two Zn/Me ratios, resulting in Zn-rich mixed oxide (Zn2MeO) and Zn-poor mixed oxide (ZnMe2O). In the glycerolysis of urea, the Zn2MeO catalysts exhibited higher glycerol conversion and glycerol carbonate yields than the ZnMe2O catalysts due to the predominance of homogeneous catalysis through Zn isocyanate (NCO) complexes from the Zn2MeO catalysts. Specifically, Zn2CrO was the best catalyst, with the highest yield of glycerol carbonate. Fourier transform infrared (FT-IR) and thermogravimetric analysis (TGA) results of the spent catalysts clearly demonstrated the dominant formation of a solid Zn NCO complex over the spent Zn2CrO catalyst, a unique feature indicating that the better catalytic performance of Zn2CrO was due to the additional heterogeneous reaction route through the solid Zn NCO complex.

Full text of DOI:10.1016/j.cattod.2019.09.017

CatalysisTodayxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Investigation of glycerolysis of urea over various ZnMeO (Me = Co, Cr, and
Fe) mixed oxide catalysts
Huy Nguyen-Phu, Lien Thi Do, Eun Woo Shin⁎
School of Chemical Engineering, University of Ulsan, 93 Daehak-ro, Nam-gu, Ulsan 44610, South Korea
A R T I C L E I N F O
A B S T R A C T
Keywords:
In this study, we investigated the glycerolysis of urea over various ZnMeO (Me = Co, Cr, and Fe) mixed oxide
catalysts. ZnMeO mixed oxide catalysts were prepared by a co-precipitation method for two Zn/Me ratios, re-
sulting in Zn-rich mixed oxide (Zn2MeO) and Zn-poor mixed oxide (ZnMe2O). In the glycerolysis of urea, the
Zn2MeO catalysts exhibited higher glycerol conversion and glycerol carbonate yields than the ZnMe2O catalysts
due to the predominance of homogeneous catalysis through Zn isocyanate (NCO) complexes from the Zn2MeO
catalysts. Specifically, Zn2CrO was the best catalyst, with the highest yield of glycerol carbonate. Fourier
transform infrared (FT-IR) and thermogravimetric analysis (TGA) results of the spent catalysts clearly demon-
strated the dominant formation of a solid Zn NCO complex over the spent Zn2CrO catalyst, a unique feature
indicating that the better catalytic performance of Zn2CrO was due to the additional heterogeneous reaction
route through the solid Zn NCO complex.
Glycerolysis
Zn-rich mixed oxide
Zn-poor mixed oxide
Zn NCO complex
Glycerol carbonate
1. Introduction
species dissolved in the liquid phase from ZnAl mixed oxide catalysts.
Moreover, the disordered ZnAl2O4 spinel structure influences the cat-
Because glycerol is a by-product of the biodiesel and biofuel man-
ufacturing industry which is considered as a solution for renewable
fuel, the crude glycerol production capacity has rapidly increased in
recent years [1]. The synthesis of glycerol carbonate (GC) has attracted
great attention as a useful route for chemically converting glycerol into
value-added products [2]. Among the various reaction pathways for
producing GC, the glycerolysis of urea (Scheme 1) is favorable because
urea is inexpensive, easily available, and recyclable from the reaction of
NH3 and CO2 [3–13].
alytic performance via the surface acidity of the catalysts and the for-
mation of a Zn-containing complex in the solid phase.
However, the investigation of dual catalysis over Zn-based mixed
oxide catalysts has not been extended to Zn-based mixed oxide catalysts
containing other metal components that are also favorable to the pre-
paration of Zn-based mixed oxide catalysts [12,15–18]. In this study,
we investigate the glycerolysis of urea over various ZnMeO (Me = Co,
Cr, and Fe) mixed oxide catalysts with two Zn/Me ratios: a high Zn ratio
(Zn2MeO) and a low Zn ratio (ZnMe2O). The Zn-rich mixed oxide
catalysts (Zn2MeO) are expected to generate more Zn-containing in-
termediates in the liquid phase and follow both homogeneous and
heterogeneous reaction routes; in contrast, the Zn-poor mixed oxide
catalysts (ZnMe2O) are expected to form a spinel lattice structure with
a lower level of dissolved Zn species in the liquid phase and follow a
heterogeneous reaction route.
Various Zn-based mixed oxide catalysts have been used for the
glycerolysis of urea [4–7,10,13,14]. Zn-based mixed oxide catalysts
follow a dual catalytic mechanism: a homogeneous reaction route
through dissolved Zn species and a heterogeneous reaction route over
solid active sites. In previous studies, it was found that ZnAl mixed
oxide catalysts exhibited better catalytic performance through the dual
mechanism than the ZnO catalyst. The catalytic performance is related
to not only the formation of Zn-containing intermediates during the
reaction but also the disordered structure of the ZnAl2O4 spinel lattice
in ZnAl mixed oxide catalysts [5,10,14]. Zn-containing intermediates
are produced through Zn species dissolved from ZnAl mixed oxide
catalysts, following the homogeneous reaction route in the liquid phase.
The homogeneous reactivity is directly related to the amount of Zn
2. Experimental
2.1. Catalyst preparation
The catalysts used in this study were prepared by a co-precipitation
method from metal nitrate salts [Zn(NO3)2∙6H2O, Cr(NO3)3∙9H2O, Co
⁎ Corresponding author.
E-mail address: ewshin@ulsan.ac.kr (E.W. Shin).
https://doi.org/10.1016/j.cattod.2019.09.017
Received 29 June 2019; Received in revised form 30 August 2019; Accepted 11 September 2019
0920-5861/©2019ElsevierB.V.Allrightsreserved.
Pleasecitethisarticleas:HuyNguyen-Phu,LienThiDoandEunWooShin,CatalysisToday,https://doi.org/10.1016/j.cattod.2019.09.017

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Scheme 1. Glycerolysis of urea to produce GC and NH3.
(NO3)2∙6H2O, and Fe(NO3)3∙9H2O] [17,19,20]. All chemicals were ob-
tained from Sigma–Aldrich Korea (Gyounggi, South Korea). Typically,
an aqueous solution of Zn(NO3)2∙6H2O and nitrate salt of Me (Cr, Co, or
Fe) with a molar ratio of 1:2 for the ZnMe2O catalysts or 2:1 for the
Zn2MeO catalysts was gradually mixed with a basic solution of am-
monia under constant pH and vigorous stirring. For the preparation of
Zn2CoO and ZnCo2O, a hydrogen peroxide solution (H2O2) was con-
tinually added to the mixture to oxidize Co2+ to Co3+. After complete
mixing, the suspension was filtered, and the remaining precipitate was
washed with deionized water several times. Finally, the solid powder
was dried at 100 °C overnight and calcined at 600 °C for 6 h.
2.3. Catalyst characterization
X-ray diffraction (XRD) patterns for fresh and spent catalysts were
obtained using a Rigaku RAD-3C diffractometer (Rigaku Corp., Tokyo,
Japan) with Cu Ka radiation (λ = 1.5418 Å) at a scattering angle (2θ)
scan rate of 2°/min, operating at 35 kV and 20 mA. The spent catalysts
were analyzed by using a Thermo Scientific Nicolet iS5 FTIR spectro-
meter (Thermo Fisher Scientific, Waltham, MA, USA). The numbers of
acidic and basic sites were measured based on the temperature-pro-
grammed desorption of NH3 and CO2 (TPD- NH3/CO2) on
a
MicrotracBEL BELCAT-M instrument (MicrotracBEL Corp., Osaka,
Japan). TPD- NH3/CO2 results were recorded for the temperature range
of 50–600 °C by a TCD detector. The detailed procedure for the TPD
analysis has been described elsewhere [5]. Thermogravimetric analysis
(TGA) of the spent catalysts was measured by a TGA Q50 apparatus (TA
Instruments, New Castle, DE, USA).
2.2. Reaction tests
Here, 0.2 mol of glycerol was added to a 100-ml round-bottom re-
actor at 80 °C under stirring by a magnetic bar. The reactor was con-
nected to a vacuum pump through an HNO3 solution trap (to remove
NH3) and a cold trap (to protect the vacuum pump). Then, 0.2 mol of
urea was added to the reactor to mix with the glycerol in the solution.
When the dissolution was complete, the catalyst (5 wt% of the glycerol
mass) was added to the reactor. Reaction tests were carried out under
vacuum pressure (3 kPa) at 140 °C with constant stirring.
3. Results and discussion
3.1. Catalyst characterization: ZnO phase and ZnMe2O4 spinel phase
Fig. 1 shows XRD patterns of the fresh catalysts. The Zn-rich cata-
lysts (Zn2MeO) are composed of two primary phases (ZnO and the
corresponding spinel ZnMe2O4 phase) while only the spinel ZnMe2O4
phase was detected in the Zn-poor catalysts (ZnMe2O). These crystal-
line phases can be readily identified in the XRD patterns of the fresh
catalysts (Fig. 1A). Typical XRD peaks for the ZnO phase with the P63mc
space group (JCPDS No. 36-1451) are observed at 31.8°, 34.4°, 36.3°,
47.5°, 56.7°, and 62.9° in the XRD patterns of the Zn2CoO, Zn2FeO, and
Zn2CrO catalysts [14]. The spinel ZnCo2O4 phase is detected in the XRD
patterns of the Zn2CoO and ZnCo2O catalysts, with characteristic XRD
peaks at 19.0°, 31.2°, 36.8°, 44.7°, 55.6°, 59.3°, and 65.1° (JCPDS No.
23-1390, Fd3m space group) [21]. Similarly, the XRD patterns of the
Zn2FeO and ZnFe2O catalysts exhibit the spinel ZnFe2O4 phase, with
typical peaks at 18.2°, 29.9°, 35.2°, 42.8°, 53.1°, 56.6°, and 62.2°
(JCPDS No. 82-1049, Fd3m space group) [22]. Characteristic XRD
peaks for the spinel ZnCr2O4 phase (JCPDS No. 22-1107, Fd3m space
group) appear at 18.8°, 30.3°, 35.7°, 43.4°, 53.9°, 57.6°, and 63.1° in the
XRD patterns of the Zn2CrO and ZnCr2O catalysts [23]. The lattice
structure of all the ZnMe2O4 phases belongs to the spinel group. Typi-
cally, in a normal spinel ZnMe2O4 lattice, the Zn2+ cations occupy the
tetrahedral sites and Me3+ cations occupy the octahedral sites. An
disordered property in the spinel structure can generate a partially in-
versed spinel structure where some Zn2+ cations occupy the octahedral
sites and some Me3+ cations occupy the tetrahedral sites [24,25].
The XRD peaks corresponding to the (311) plane of the spinel
ZnMe2O4 phases are enlarged in Figs. 1.B–D to compare the lattice
spacing between the Zn-rich Zn2MeO and Zn-poor ZnMe2O catalysts.
For the Zn2MeO catalysts, the position of the characteristic XRD peak
for the (311) plane is shifted to a lower 2θ value with respect to the
positions for the ZnMe2O catalysts, indicating that the lattice spacing of
Zn2MeO is greater than that of ZnMe2O. This peak position shift for the
Zn-rich Zn2MeO catalysts is caused by the interaction between the ZnO
and spinel ZnMe2O4 phases in the Zn-rich mixed oxide of Zn/Me
(Zn2MeO catalysts in this study), where the excess Zn (of ZnO phase)
After the reaction tests, ethanol was added to the final products, and
the liquid products were separated from the spent catalyst by filtration.
The liquid products were quantitatively analyzed using a gas chroma-
tography apparatus (Acme 6100 GC, YL Instrument Co., Ltd., Dongan-
gu, Anyang, South Korea) with a flame ionization detector and a ca-
pillary column [DB-Wax (30 m × 0.25 mm × 0.25 μm)]. The molar
amount of each component was calculated using an internal standard
method, with tetraethylene glycol as the internal standard chemical.
The glycerol conversion, GC selectivity, GC yield, and by-product se-
lectivity were calculated using the equations below.
Glycerol conversion (%)
Initial amount of glycerol Residual amount of glycerol
=
× 100
Initial amount of glycerol
Amount of GC
Initial amount of glycerol
GC yield (%) =
GC selectivity (%) =
× 100
GC yield (%)
Glycerol conversion (%)
× 100
Byproduct (except ZnGly) selectivity (%)
Amount of byproduct
Initial amount of glycerol Residual amount of glycerol
=
× 100
Fourier transform infrared (FTIR) spectra of the liquid products
were obtained using a Thermo Scientific™ Nicolet™ iS™5 FTIR spec-
trometer (Thermo Fisher Scientific, Waltham, MA, USA). The levels of
metal atoms (Zn, Co, Cr, and Fe) in the liquid phase were measured
using an Agilent Technologies 5110 inductively coupled plasma optical
emission spectrometer (ICP-OES, Agilent, Santa Clara, CA, USA).
2

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 1. A) XRD patterns of fresh catalysts. a) Zn2CoO, b) Zn2FeO, c) Zn2CrO, d) ZnCo2O, e) ZnFe2O, and f) ZnCr2O. B, C, D) Enlarged XRD regions of the (311) peak
(spinel phase ZnMe2O4) for the Zn-rich Zn2MeO catalyst and corresponding Zn-poor ZnMe2O catalyst.
can generate a layer of amorphous ZnO on the surface of the spinel
ZnMe2O4 phase. The excess Zn cations in the ZnO phase can migrate
into the lattice of the spinel phase to produce a non-stoichiometric
spinel phase (Zn/Me ratio > 0.5), implying more Zn amount in the
spinel structure [5,17,26]. In detail, the excessive Zn2+ cations can
replace the Me3+ cations in the octahedral sites of the spinel lattice
structure. The Zn migration and substitution cause a distortion in the
spinel lattice structure, with consequent changes in the interplanar
lattice spacing of the spinel phase due to the different cation radii. For
the octahedral sites, the cation radii of the metal cations can be esti-
mated as 88 pm for Zn2+, 68.5 pm for Co3+, 78.5 pm for Fe3+, and
75.5 pm for Cr3+ [27]. The radius of the Zn2+ cation is much larger
than the radii of the other Me3+ cations. Therefore, the migration of
Zn2+ to the octahedral sites of the spinel ZnMe2O4 lattice structure can
expand the lattice spacing of the spinel structure, with a shift in the
corresponding XRD peak position to a lower value [17,26,28]. In
summary, the XRD peak shift of the (311) plane indicates an interaction
between the ZnO and ZnMe2O4 phases in the Zn-rich Zn2MeO catalysts.
FTIR spectra of the fresh catalysts are displayed in Fig. 2. Two vi-
bration bands (ν1, and ν2) of MOe bonds, related to the metal cations at
the octahedral sites of the spinel ZnMe2O4 lattice, are observed in the
FTIR spectra of both Zn2MeO and ZnMe2O in the range of
552–690 cm−1 (ν1) and 425–582 cm−1 (ν2) [15,29]. Because the ma-
jority of the octahedral sites in a normal spinel lattice structure are
occupied by trivalent Me3+ cations [24], the positions of ν1 and ν2 may
change depending on the nature of Me3+ cations. For the Zn2CoO and
ZnCoO2 catalysts (Figs. 2a and b), the ν1 and ν2 modes of the spinel
ZnCo2O4 lattice are found at 676 and 590 cm−1, respectively. For the
Zn2CrO and ZnCr2O catalysts, two bands of the spinel ZnCr2O4 lattice
appear at 623 and 505 cm−1 (Figs. 2e and f). However, due to the
limitation of our FTIR instrument for measurements below 400 cm−1
,
the FTIR spectra of the Zn2FeO and ZnFe2O catalysts only show the full
band of ν1 vibration of the spinel ZnFe2O4 lattice at 544 cm−1 (Figs. 2c
and d). Beside the vibration modes ν1 and ν2 of the spinel phase, vi-
bration of the ZnO lattice is detected at 434 cm-1 in the FTIR spectra of
all of the Zn-rich catalysts (Zn2MeO) [30], providing further evidence
for the occurrence of the ZnO phase in these fresh catalysts. Therefore,
the FTIR results are consistent with the XRD data. The Zn-poor ZnMe2O
catalysts are composed of the pure spinel ZnMe2O4 phase, whereas both
the ZnMe2O4 and ZnO phases are dominant in the Zn-rich Zn2MeO
catalysts.
To investigate the chemisorption properties (acidity and basicity) of
these fresh catalysts, we determined the number of surface acidic sites
and surface basic sites using the TPD- NH3 and TPD−CO2 methods,
respectively. TPD- NH3 and TPD−CO2 profiles of the catalysts are
plotted in Figs. S1 and S2 (see Supplementary Data), and the total
numbers of acidic and basic sites are summarized in Table 1. Among the
Zn-rich Zn2MeO catalysts, Zn2CrO has the greatest number of both
acidic sites (0.186 mmol/g) and basic sites (0.157 mmol/g), and the
numbers of acidic and basic sites follow the same trend: Zn2CrO >
Zn2CoO > Zn2FeO. For the Zn-poor ZnMe2O catalysts, the numbers
Table 1
Acidity and basicity of fresh catalysts.
Sample
Acidic Sites (mmol NH3 /gcat.
)
Basic Sites (mmol CO2/gcat)
Zn2CoO
Zn2FeO
Zn2CrO
ZnCo2O
ZnFe2O
ZnCr2O
0.144
0.081
0.186
0.202
0.225
0.259
0.113
0.066
0.157
0.082
0.084
0.140
Fig. 2. FTIR spectra of fresh catalysts: a) Zn2CoO, b) ZnCo2O, c) Zn2FeO, d)
ZnFe2O, e) Zn2CrO, and f) ZnCr2O.
3

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Table 2
Analysis of liquid products obtained from the glycerolysis of urea over various ZnMe (Me = Co, Cr, and Fe) mixed oxide catalysts at various reaction times. (Reaction
temperature = 140 °C, reaction pressure = 3 kPa, glycerol/urea ratio = 1:1). (2): 2,3-dihydroxypropyl carbamate, (4): 4-(hydroxymethyl)oxazolidin-2-one, and (5):
(2-oxo-1,3-dioxolan-4-yl)methyl carbamate.
Sample
Reaction time (h)
Glycerol conv. (%)
GC yield (%)
Selectivity (%)
Metallic content in liquid phase (mmol)a
GC
(2)
(4)
(5)
Co
Fe
Cr
Zn
Zn2CoO
Zn2FeO
Zn2CrO
ZnCo2O
ZnFe2O
ZnCr2O
1
3
1
3
1
3
3
5
3
5
3
5
50
74
57
74
50
76
50
63
34
47
37
53
34
52
41
52
38
57
29
40
18
27
21
33
67
71
72
70
75
74
57
63
53
57
56
62
24
9
7
2
–
–
–
–
–
–
–
–
–
–
–
–
–
–
1.50
0.66
1.43
0.75
2.07
0.99
0.38
0.41
–
7
14
1
–
–
17
7
10
15
7
–
–
8
–
–
16
5
1
–
0.02
13
10
6
8
–
–
28
17
32
20
32
21
4
0.09
–
13
4
0.13
–
11
8
–
–
–
–
–
15
4
–
–
7
0.02
0.03
–
4
13
–
a
Amounts of metals dissolved in the liquid phase as measured by the ICP-OES method.
of both acidic and basic sites follow the order of ZnCr2O >
ZnFe2O > ZnCo2O, and the highest values (acidic site
amount = 0.259 mmol/g, and basic site amount = 0.140 mmol/g) are
observed for ZnCr2O. In the literature, high amounts of acidic and basic
sites in ZnO/ZnAl2O4 mixed oxides are attributed to the disordered
structure of a partially inversed spinel lattice prepared by the poly-
styrene-template method [10,14] or citrate complex method [5]. In this
study, all catalysts were prepared by the co-precipitation method,
where the structure of the prepared spinel phase is less disordered. As a
result, the amounts of acidic and basic sites are lower than those of
catalysts prepared by other methods. The TPD profiles were deconvo-
luted to check the relative distributions of acidic sites and basic sites for
each catalyst (Figs. S1 and S2, see the Supplementary Data). The
strength of deconvoluted acidic and basic sites can be classified as
follows: weak strength sites (< 300 °C), medium strength sites
(300–550 °C) and strong strength sites (> 550 °C) [31–33]. The dis-
tributions of individual acidic and basic sites on each fresh catalyst are
summarized in Table S1 (see the Supplementary Data).
Fig. 3. Relationship between glycerol conversion (%) and GC yield (%) in the
glycerolysis of urea over various ZnMe (Me = Co, Cr, and Fe) mixed oxide
catalysts.
3.2. Glycerolysis of urea over various ZnMeO catalysts: the formation of Zn
isocyanate complex in the liquid phase and on the solid catalysts
generate an isocyanate (NCO) complex of Zn, which is an active site for
the reaction. This phenomenon was also confirmed in our previous
studies [5,10,14]. Furthermore, we found that the NCO complexes ad-
sorbed on the surface acidic sites of the catalyst (ZnAl2O4 phase) act as
additional active sites to catalyze the reaction. For this reason, we
measured the metal cation levels using an ICP-OES instrument (Table 2)
and collected FTIR spectra of both the liquid products (Fig. 4) and the
spent catalysts (Fig. 5) to identify the existence of an NCO functional
vibration.
The reaction results for the glycerolysis of urea over various ZnMe
mixed oxide catalysts are summarized in Table 2. The Zn-rich Zn2MeO
catalysts were tested at reaction times of 1 and 3 h while the Zn-poor
ZnMe2O catalysts were investigated at longer reaction times (3 and 5 h)
since the ZnMe2O catalysts exhibit lower catalytic reaction rates than
the Zn2MeO catalysts. The relationship between the glycerol conver-
sion and GC yield for each catalyst is depicted in Fig. 3. Clearly, the data
for the Zn2MeO catalysts correspond to higher positions with a higher
glycerol conversion compared to the results for the ZnMe2O catalysts.
In Fig. 3, data points at a higher position indicate that a higher GC yield
or selectivity can be achieved at a given glycerol conversion. For all of
the Zn2MeO catalysts, the highest glycerol conversion and GC yield
(glycerol conversion = 76%, GC yield = 57%) was observed for
Zn2CrO at a reaction time of 3 h; meanwhile, the results for the Zn2CoO
and Zn2FeO catalysts are similar, with glycerol conversion = 74% and
GC yield = 52% at a reaction time of 3 h. Among the ZnMe2O catalysts,
the glycerol conversion and GC yield follow the order of ZnFe2O <
ZnCr2O < ZnCo2O; the best reaction performance was observed for
the ZnCr2O catalyst at a reaction time of 5 h, with glycerol conver-
sion = 63% and GC yield = 40%.
The ICP results in Table 2 show that Zn atoms are detected in the
liquid products for all of the Zn-rich Zn2MeO catalysts. After a reaction
time of 1 h, high levels of Zn are detected: 2.07 mmol for Zn2CrO,
1.50 mmol for Zn2CoO, and 1.43 mmol for Zn2FeO. After a reaction
time of 3 h, the level of Zn in the liquid phase for the Zn-rich Zn2MeO
catalysts is lower, resulting from the formation of zinc glycerolate.
However, no Fe or Co atoms are detected in the liquid phases of the
Zn2CoO or Zn2FeO catalysts. Moreover, a negligible amount of Cr
(0.02 mmol) is observed in the liquid product of the Zn2CrO catalyst
after a reaction time of 1 h. The level of Cr in the liquid is 100-fold times
smaller than the corresponding level of Zn (2.07 mmol) in the liquid
and falls to zero after a reaction time of 3 h. The FTIR spectra of the
liquid products obtained from the Zn2MeO catalysts (Fig. 4A) are in
good agreement with the ICP results. The peak at 2210 cm−1 can be
According to Fujita et al. [13] and Turney et al. [9], in the glycer-
olysis of urea over Zn-containing catalysts, Zn atoms can leach from the
ZnO phase into the liquid phase through a reaction with urea to
4

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Fig. 4. FTIR spectra of liquid products obtained from the glycerolysis of urea
over different catalysts at various reaction times. A): a) Zn2CoO-1 h, b)
Zn2CoO-3 h, c) Zn2FeO-1 h, d) Zn2FeO-3 h, e) Zn2CrO-1 h, and f) Zn2CrO-3 h.
B): a) ZnCo2O-3 h, b) ZnCo2O-5 h, c) ZnFe2O-3 h, d) ZnFe2O-5 h, e) ZnCr2O-
3 h, and f) ZnCr2O-5 h.
Fig. 5. FTIR spectra of spent catalysts. A): a) Zn2CoO-1 h-spent, b) Zn2CoO-3 h-
spent, c) Zn2FeO-1 h-spent, d) Zn2FeO-3 h-spent, e) Zn2CrO-1 h-spent, and f)
Zn2CrO-3 h-spent. B): a) ZnCo2O-3 h-spent, b) ZnCo2O-5 h-spent, c) ZnFe2O-
3 h-spent, d) ZnFe2O-5 h-spent, e) ZnCr2O-3 h-spent, and f) ZnCr2O-5 h-spent.
the surface acidic sites of the catalyst to form a Zn NCO complex on the
surface, which generates another heterogeneous reaction route for the
glycerolysis of urea. This interpretation explains the reaction behavior
of the Zn-rich Zn2MeO catalysts (ZnO/ZnMe2O4 mixed oxide). The
existence of a Zn NCO complex in the liquid phase for all Zn2MeO
catalysts leads to a homogeneous reaction in the liquid phase. In con-
trast, an additional heterogeneous reaction occurs only over Zn2CrO via
the formation of the Zn NCO complex on the solid because the acidity of
Zn2CrO is higher than that of the other Zn2MeO catalysts (Table 1), and
a high surface acidity is necessary for adsorption of the Zn NCO com-
plex on the catalyst surface.
assigned to the NCO vibration [13,34]. The intensities of the NCO peaks
in Fig. 4A follow the same trend as the amount of Zn in the liquid
products for the Zn2MeO catalysts. A peak at 2210 cm−1 is clearly
present in the FTIR spectra after a reaction time of 1 h and almost
disappears after a reaction time of 3 h. Therefore, we can conclude that
for the Zn2MeO catalysts, the Zn NCO complex exists in the liquid
phase for at least 1 h of reaction time.
Fig. 5A displays the FTIR spectra of the spent Zn2MeO catalysts. All
of the typical FTIR peaks of M–O bonding in the spinel ZnMe2O4 lattice
are detected in the range of 400–680 cm−1. However, the typical peak
of Zn-O bonding in the ZnO lattice almost disappears due to the Zn
dissolution. Additionally, characteristic peaks for zinc glycerolate ap-
pear at 1950 cm−1 (C–O stretching modes with oxygen in hydrogen
bonding), 655 cm−1, and 514 cm−1 (Zn-O stretching mode) [35],
confirming the formation of zinc glycerolate. Interestingly, the vibra-
tion peak of an NCO functional group is detected only in the FTIR
spectrum of the spent ZnCr2O catalyst at 2220 cm−1, which is assigned
to the NCO complex adsorbed on the solid surface. These results are
consistent with our previous reports on the reaction over ZnO/ZnAl2O4
mixed oxides [5,10,14]. Zn atoms from the ZnO phase can dissolve and
react with urea to produce a Zn NCO complex in the liquid phase,
providing an active site for a homogeneous reaction route in the gly-
cerolysis of urea. Furthermore, the Zn NCO complex can be adsorbed on
The FTIR spectra of the liquid products from the Zn-poor catalysts
(ZnMe2O) are displayed in Fig. 4B. A vibration band for the NCO
functional group at 2210 cm−1 is detected only for the ZnCo2O catalyst
after reaction times of 3 and 5 h. Meanwhile, there is no evidence of an
NCO functional group in the FTIR spectra of the liquid products from
the ZnFe2O and ZnCr2O catalysts. The ICP results in Table 2 reveal an
interesting appearance of both Zn and Co atoms in the liquid phase of
ZnCo2O. While the Zn level in the liquid phase of the Zn2MeO catalysts
decreases for longer reaction times, the levels of Zn and Co in the liquid
phase of the ZnCo2O catalyst increase slightly with increasing reaction
time. In detail, after a reaction time of 3 h, the levels of Zn and Co in the
liquid phase of ZnCo2O are 0.09 and 0.38 mmol, respectively, and in-
crease to 0.13 and 0.41 mmol for a 5 h reaction time, respectively. In
5

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
the liquid phase of the ZnFe2O catalyst, there is no evidence of Zn or Fe
atoms in the ICP measurements. For the ZnCr2O catalyst, only a trace
level of Cr atoms is measured in the liquid phase (0.02 mmol at 3 h and
0.03 mmol at 5 h).
Based on the ICP and FTIR results for the liquid products, the
ZnFe2O catalyst does not induce the generation of an NCO complex in
the liquid phase. In contrast, for the ZnCo2O catalyst, both NCO func-
tional groups and Zn/Co atoms are present in the liquid phase, implying
NCO complexes of Co and Zn in the liquid phase. In previous works,
NCO complexes of Co2+ and Co3+ have been reported as intermediates
of the reduction reaction [36,37] or glycerolysis reaction [9]. Herein,
the Co NCO complex can form in the liquid phase in the case of the
ZnCo2O catalyst but not in the case of the Zn2CoO catalyst because no
Co atoms are found in the liquid phase of the Zn2CoO catalyst due to
the difference in the structures of ZnCo2O and Zn2CoO. As previously
mentioned, the ZnCo2O catalyst is composed of a pure normal spinel
phase of ZnCo2O4 with a predominance of Co3+ cations on the outset
surface layer. According to the model by Wachs and Routray [38], other
layers of Co3+ and Zn2+ cations can be found below the outset surface
layer. In the reaction process over ZnCo2O, the outer layer of Co3+
cations can react with urea and dissolve to the liquid phase, resulting in
the further dissolution of Co3+ and Zn2+ cations in the lower layers.
This process renders the ZnCo2O catalyst unstable and causes Zn (or
Co) NCO complex to appear in the liquid phase. In contrast, the
structure of the Zn2CoO catalyst includes an amorphous layer of ZnO
above the spinel phase of ZnCo2O4, which can prevent the Co3+ cations
from leaching into the liquid phase.
Fig. 6. XRD patterns of spent catalysts: a) Zn2CoO-3 h-spent, b) Zn2FeO-3 h-
spent, c) Zn2CrO-3 h-spent, d) ZnCo2O-3 h-spent, e) ZnFe2O-3 h-spent, and f)
ZnCr2O-3 h-spent.
Based on the TGA analysis, DTGA profiles of the spent catalysts are
plotted in Fig. S3 (see Supplementary Data). The DTGA results for the
spent catalysts match the XRD and FTIR results. Decomposition peaks of
zinc glycerolate are observed in the range of 350–400 °C for the spent
Zn2MeO catalysts [44], and decomposition peaks of metal NCO com-
plex in the range of 200–350 °C are detected for the spent Zn2CrO,
ZnCr2O, and ZnCo2O catalysts [5].
The FTIR spectra of the spent Zn-poor ZnMe2O catalysts are dis-
played in Fig. 5B. For all three catalysts, the typical vibration peaks of
M–O bonding in the spinel ZnMe2O4 are unaltered compared to those of
the fresh catalysts (even for spent ZnCo2O, where Zn2+ and Co3+ ca-
tions are dissolved in the liquid phase). However, in the FTIR spectra of
spent Zn2CoO, a small additional vibration band for the NCO functional
group is observed at 2192 cm−1, which is redshifted in comparison to
the vibration for the spent Zn2MeO catalysts (2220 cm−1). Turney et al.
[9] found that the FTIR vibration of NCO functional groups related to
the Co NCO complex occurred at a lower position compared to that of
the NCO functional group related to the Zn NCO complex. Because the
NCO functional group of the spent ZnCo2O catalyst is related to the
NCO complexes of both Zn and Co, its vibrational wavenumber is lower
than that of Zn2MeO (NCO functional group of only the Zn NCO
complex). In the FTIR spectrum of the spent ZnCr2O catalyst, a small
additional vibration band for the NCO functional group appears at
2211 cm−1, which may be associated with a Cr NCO complex. A reac-
tion between Cr3+ and urea to generate an insoluble Cr NCO complex
has been previously reported [39,40].
3.3. Catalytic reaction routes through different NCO complexes
We can summarize the catalytic performance of the various ZnMe
mixed oxide catalysts in the glycerolysis of urea in terms of the for-
mation of the NCO complex. Different types of NCO complexes are
detected in the liquid phase and on the solid surface of the catalysts,
except for the ZnFe2O catalyst, for which no NCO complex is formed
because no metal cations dissolve from the catalyst. In the reaction over
the ZnCr2O catalyst, there is a small amount of Cr NCO complex on the
solid phase. For the ZnCo2O catalyst, NCO complexes of Zn and Co are
found in the liquid phase, with a low amount of complexes on the solid
surface of the spent catalyst. In contrast, for all of the Zn-rich catalysts,
a high level of Zn NCO complexes is clearly detected in the liquid phase
for a reaction time of 1 h, indicating Zn dissolution from the ZnO phase
in the Zn2FeO, Zn2CoO, and Zn2CrO catalysts. Interestingly, a higher
level of Zn NCO complexes is observed on the solid surface only for the
Zn2CrO catalyst, due to its higher surface acidity/basicity and the oc-
currence of ZnO phases.
The XRD patterns of the spent catalysts after a reaction time of 3 h
are displayed in Fig. 6. For the spent Zn-rich Zn2MeO catalysts, the
characteristic peaks for the ZnO phase almost disappear while the peaks
for the spinel phase ZnMe2O4 remain unchanged. Moreover, typical
peaks for the zinc glycerolate phase (JCPDS No. 23–1975, P21/c space
group) are clearly detected in the XRD patterns of the spent Zn2MeO
catalysts at 10.9°, 17.1°, 20.5°, 23.8°, 24.7°, 27.6°, 27.7°, 36.4°, and
38.2° [41]. Conversely, the XRD patterns of the spent Zn-poor ZnMe2O
catalysts are almost identical to those of the fresh ZnMe2O catalysts;
only typical peaks for the spinel phase ZnMe2O4 are detected, without
any other phases. The XRD results of the spent catalysts are in good
agreement with the FTIR results. That is, the ZnO phase of the Zn2MeO
catalysts disappears with the additional formation of the zinc glycer-
olate phase after the reaction, whereas the spinel ZnMe2O4 phase of the
ZnMe2O catalysts remains unchanged. For the spent ZnCo2O catalyst,
metal glycerolate is not found due to the different oxidation states of
the Co cations: Co3+ cations in the ZnMe2O4 lattice and Co NCO
complex and Co2+ cations in cobalt glycerolate [42,43].
Based on the reaction mechanism of urea glycerolysis, which has
been described elsewhere [5,10,14], we illustrate three primary cata-
lytic reaction routes (r1, r2, and r3) in Scheme 2.
i) r1 is a heterogeneous reaction route via the adsorption of glycerol
and urea on the surface acidic and basic sites of the catalysts [4].
This reaction is the slowest reaction route, resulting in the lowest
catalytic performance.
adsorbed glycerol + adsorbed urea → GC (r1)
i) r2 is a homogeneous reaction route via NCO complexes of Zn or Co in
the liquid phase [5,13]. This homogenous reaction occurs more
rapidly than r1; however, the formation of zinc glycerolate as a by-
product from the reaction between the Zn NCO complex and
6

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
Scheme 2. Catalytic reaction routes for urea glycerolysis over various ZnMe (Me = Co, Cr, and Fe) mixed oxide catalysts.
glycerol in the liquid phase reduces the number of Zn NCO active
sites and the GC selectivity.
catalytic route in urea glycerolysis. The Zn2CrO catalyst achieved the
best catalytic reaction performance because of the high level of NCO
complex in the liquid phase and the unique formation of the adsorbed
NCO complex due to the high surface acidity.
Me NCO complex (liquid) + glycerol → GC (r2)
Zn NCO complex (liquid) + glycerol → zinc glycerolate
Acknowledgments
i) r3 is another heterogeneous reaction route via NCO complexes ad-
sorbed on the surface acidic sites of the catalysts [5,14]. This reac-
tion route can enhance the catalytic reaction performance because
there is no formation of zinc glycerolate.
This work was financially supported by the Research Fund of the
University of Ulsan.
Appendix A. Supplementary data
Me NCO complex (solid) + glycerol → GC (r3)
The reaction performance of the catalysts can be evaluated on the
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.cattod.2019.09.017.
basis
of
GC
yields,
which
follow
the
order
of
ZnFe2O < ZnCr2O < ZnCo2O < Zn2FeO < Zn2CoO < Zn2CrO in
this study. Depending on the NCO complex, different catalytic reaction
routes occur for each catalyst. Scheme 2 presents the performance of
the catalysts in terms of the catalytic reaction routes (r1, r2, and r3) as
follows:
References
[1] C.A.G. Quispe, C.J.R. Coronado, J.A. Carvalho Jr., Glycerol: production, con-
sumption, prices, characterization and new trends in combustion, Renewable
Sustainable Energy Rev. 27 (2013) 475–493.
[2] M.O. Sonnati, S. Amigoni, E.P. Taffin de Givenchy, T. Darmanin, O. Choulet,
F. Guittard, Glycerol carbonate as a versatile building block for tomorrow: synth-
esis, reactivity, properties and applications, Green Chem. 15 (2013) 283–306.
[3] M. Aresta, A. Dibenedetto, F. Nocito, C. Ferragina, Valorization of bio-glycerol: new
catalytic materials for the synthesis of glycerol carbonate via glycerolysis of urea, J.
Catal. 268 (2009) 106–114.
- ZnFe2O (glycerol conversion = 34%, GC yield = 18%): r1 route
only.
- ZnCr2O (glycerol conversion = 37%, GC yield = 21%): r1 (main
route) + r3 (minor route).
[4] M.J. Climent, A. Corma, P. De Frutos, S. Iborra, M. Noy, A. Velty, P. Concepción,
Chemicals from biomass: synthesis of glycerol carbonate by transesterification and
carbonylation with urea with hydrotalcite catalysts. The role of acid-base pairs, J.
Catal. 269 (2010) 140–149.
- ZnCo2O (glycerol conversion = 50%, GC yield = 28%): r1 (main
route) + r2 + r3 (minor route).
- Zn2FeO and Zn2CrO (glycerol conversion = 74%, GC yield = 52%):
r1 (minor route) + r2 (main route).
[5] H. Nguyen-Phu, E.W. Shin, Disordered structure of ZnAl2O4 phase and the forma-
tion of a Zn NCO complex in ZnAl mixed oxide catalysts for glycerol carbonylation
with urea, J. Catal. 373 (2019) 147–160.
- Zn2CrO (glycerol conversion = 76%, GC yield = 57%): r1 (minor
route) + r2 (main route) + r3.
[6] J.-H. Park, J.S. Choi, S.K. Woo, S.D. Lee, M. Cheong, H.S. Kim, H. Lee, Isolation and
characterization of intermediate catalytic species in the Zn-catalyzed glycerolysis of
urea, Appl. Catal. A Gen. 433–434 (2012) 35–40.
[7] H. Nguyen-Phu, C. Park, E.W. Shin, Activated red mud-supported Zn/Al oxide
catalysts for catalytic conversion of glycerol to glycerol carbonate: FTIR analysis,
Catal. Commun. 85 (2016) 52–56.
4. Conclusion
Various Zn-containing mixed oxide catalysts with different sec-
ondary metals (Me = Cr, Co, Fe) were successfully prepared by the co-
precipitation method with two Zn/Me ratios, resulting in Zn-rich and
Zn-poor catalysts. Depending on the Zn content and the nature of the
metal ions, the structures of the Zn-containing phases differed, resulting
in various types of NCO complexes in the liquid phase and on the solid
surface. The formation of the NCO complex strongly influenced the
[8] C. Hammond, J.A. Lopez-Sanchez, M. Hasbi Ab, N. Rahim, R.L. Dimitratos,
A.F. Jenkins, Q. Carley, C.J. He, D.W. Kiely, G.J. Knight, Hutchings, Synthesis of
glycerol carbonate from glycerol and urea with gold-based catalysts, Dalton Trans.
40 (2011) 3927.
[9] T.W. Turney, A. Patti, W. Gates, U. Shaheen, S. Kulasegaram, Formation of glycerol
carbonate from glycerol and urea catalysed by metal monoglycerolates, Green
Chem. 15 (2013) 1925.
[10] H. Nguyen-Phu, C. Park, E.W. Shin, Dual catalysis over ZnAl mixed oxides in the
7

H. Nguyen-Phu, et al.
CatalysisTodayxxx(xxxx)xxx–xxx
glycerolysis of urea: homogeneous and heterogeneous reaction routes, Appl. Catal.
distances in halides and chalcogenides, Acta Crystallogr. Sect. A. 32 (1976)
A Gen. 552 (2018) 1–10.
751–767.
[11] L. Wang, Y. Ma, Y. Wang, S. Liu, Y. Deng, Efficient synthesis of glycerol carbonate
from glycerol and urea with lanthanum oxide as a solid base catalyst, Catal.
Commun. 12 (2011) 1458–1462.
[28] G. Del Piero, F. Trifiro, A. Vaccari, Non-stoicheiometric Zn–Cr spinel as active phase
in the catalytic synthesis of methanol, J. Chem. Soc. Chem. Commun. 0 (1984)
656–658.
[12] F. Rubio-Marcos, V. Calvino-Casilda, M.A. Bañares, J.F. Fernandez, Novel hier-
archical Co3O4/ZnO mixtures by dry nanodispersion and their catalytic application
in the carbonylation of glycerol, J. Catal. 275 (2010) 288–293.
[13] S.I. Fujita, Y. Yamanishi, M. Arai, Synthesis of glycerol carbonate from glycerol and
urea using zinc-containing solid catalysts: a homogeneous reaction, J. Catal. 297
(2013) 137–141.
[29] J. Preudhomme, P. Tarte, Infrared studies of spinels-III. The normal II-III spinels,
Spectrochim, Acta Part A Mol. Spectrosc. 27 (1971) 1817–1835.
[30] M. Samadi, H.A. Shivaee, M. Zanetti, A. Pourjavadi, A. Moshfegh, Visible light
photocatalytic activity of novel MWCNT-doped ZnO electrospun nanofibers, J. Mol.
Catal. A Chem. 359 (2012) 42–48.
[31] Š. Hajduk, V.D.B.C. Dasireddy, B. Likozar, G. Dražić, Z.C. Orel, COx-free hydrogen
production via decomposition of ammonia over Cu–Zn-based heterogeneous cata-
lysts and their activity/stability, Appl. Catal. B Environ. 211 (2017) 57–67.
[32] M. Chen, J. Xu, F.-Z. Su, Y.-M. Liu, Y. Cao, H.-Y. He, K.-N. Fan, Dehydrogenation of
propane over spinel-type gallia–alumina solid solution catalysts, J. Catal. 256
(2008) 293–300.
[14] H. Nguyen-Phu, E.W. Shin, Investigating time-dependent Zn species over Zn-based
catalysts in glycerol carbonylation with urea and their roles in the reaction me-
chanism, Appl. Catal. A Gen. 561 (2018) 28–40.
[15] G.N. Kustova, E.B. Burgina, G.G. Volkova, T.M. Yurieva, L.M. Plyasova, IR spec-
troscopic investigation of cation distribution in Zn–Co oxide catalysts with spinel
type structure, J. Mol. Catal. A Chem. 158 (2000) 293–296.
[33] J.C. Lavalley, Infrared spectrometric studies of the surface basicity of metal oxides
and zeolites using adsorbed probe molecules, Catal. Today 27 (1996) 377–401.
[34] W. Zhao, W. Peng, D. Wang, N. Zhao, J. Li, F. Xiao, W. Wei, Y. Sun, Zinc oxide as the
precursor of homogenous catalyst for synthesis of dialkyl carbonate from urea and
alcohols, Catal. Commun. 10 (2009) 655–658.
[16] D. Sibera, J. Kaszewski, D. Moszyński, E. Borowiak-Paleń, W. Łojkowski,
U. Narkiewicz, ZnFe2O4/ZnO nanoparticles obtained by coprecipitation route, XPS
and TEM study, Phys. Status Solidi 7 (2010) 1420–1423.
[17] H. Song, D. Laudenschleger, J.J. Carey, H. Ruland, M. Nolan, M. Muhler, Spinel-
structured ZnCr2O4 with excess Zn is the active ZnO/Cr2O3 catalyst for high-tem-
perature methanol synthesis, ACS Catal. 7 (2017) 7610–7622.
[35] J. Das, D. Khushalani, Nonhydrolytic route for synthesis of ZnO and its use as a
recyclable photocatalyst, J. Phys. Chem. C. 114 (2010) 2544–2550.
[36] T. Sun, M.D. Fokema, J.Y. Ying, Mechanistic study of NO reduction with methane
over Co2+ modified ZSM-5 catalysts, Catal. Today 33 (1997) 251–261.
[37] T. Montanari, O. Marie, M. Daturi, G. Busca, Searching for the active sites of Co-H-
MFI catalyst for the selective catalytic reduction of NO by methane: a FT-IR in situ
and operando study, Appl. Catal. B Environ. 71 (2007) 216–222.
[38] I.E. Wachs, K. Routray, Catalysis science of bulk mixed oxides, ACS Catal. 2 (2012)
1235–1246.
[18] S. Chen, Y. Wu, P. Cui, W. Chu, X. Chen, Z. Wu, Cation distribution in ZnCr2O4
nanocrystals investigated by X-ray absorption fine structure spectroscopy, J. Phys.
Chem. C. 117 (2013) 25019–25025.
[19] C. Xiangfeng, L. Xingqin, M. Guangyao, Preparation and gas sensitivity properties of
ZnFe2O4 semiconductors, Sensors Actuators B Chem. 55 (1999) 19–22.
[20] X. Song, Q. Ru, B. Zhang, S. Hu, B. An, Flake-by-flake ZnCo2O4 as a high capacity
anode material for lithium-ion battery, J. Alloys. Compd. 585 (2014) 518–522.
[21] Y. Sharma, N. Sharma, G.V. Subba Rao, B.V.R. Chowdari, Nanophase ZnCo2O4 as a
high performance anode material for Li-Ion batteries, Adv. Funct. Mater. 17 (2007)
2855–2861.
[39] H. ‐H Schmidtke, T. Schönherr, Darstellung von Isocyanatopentamminchromat(III)
mittels thermischer Zersetzung von Harnstoff, ZAAC ‐, J. Inorg. Gen. Chem. 443
(1978) 225–230.
[22] F. Liu, X. Chu, Y. Dong, W. Zhang, W. Sun, L. Shen, Acetone gas sensors based on
graphene-ZnFe2O4 composite prepared by solvothermal method, Sensors Actuators
B Chem. 188 (2013) 469–474.
[40] T. Schönherr, The cyanate ion as bridging ligand: synthesis, spectroscopic char-
acterization, and magnetic properties of (μ-Cyanato)bis(pentaamminechromium
(III)) chloride, Inorg. Chem. 25 (1986) 171–175.
[23] J.-H. Kim, H.-M. Jeong, C.W. Na, J.-W. Yoon, F. Abdel-Hady, A.A. Wazzan, J.-
H. Lee, Highly selective and sensitive xylene sensors using Cr2O3-ZnCr2O4 hetero-
nanostructures prepared by galvanic replacement, Sensors Actuators B Chem. 235
(2016) 498–506.
[41] L. Pan, S. Wang, W. Mi, J. Song, J.-J. Zou, L. Wang, X. Zhang, Undoped ZnO
abundant with metal vacancies, Nano Energy 9 (2014) 71–79.
[42] S. Gudyka, G. Grzybek, J. Gryboś, P. Indyka, B. Leszczyński, A. Kotarba, Z. Sojka,
Enhancing the deN2O activity of the supported Co3O4|α-Al2O3 catalyst by glycerol-
assisted shape engineering of the active phase at the nanoscale, Appl. Catal. B
Environ. 201 (2017) 339–347.
[24] K.E. Sickafus, J.M. Wills, N.W. Grimes, Structure of spinel, J. Am. Ceram. Soc. 82
(2004) 3279–3292.
[25] J.K. Burdett, G.D. Price, S.L. Price, Role of the crystal-field theory in determining
the structures of spinels, J. Am. Chem. Soc. 104 (1982) 92–95.
[26] D. Makovec, A. Kodre, I. Arčon, M. Drofenik, The structure of compositionally
constrained zinc-ferrite spinel nanoparticles, J. Nanopart. Res. 13 (2011)
1781–1790.
[43] E. Radoslovich, M. Raupach, P. Slade, R. Taylor, Crystalline cobalt, zinc, manga-
nese, and iron alkoxides of glycerol, Aust. J. Chem. 23 (1970) 1963.
[44] D.M. Reinoso, D.E. Damiani, G.M. Tonetto, Zinc glycerolate as a novel hetero-
geneous catalyst for the synthesis of fatty acid methyl esters, Appl. Catal. B Environ.
144 (2014) 308–316.
[27] R.D. Shannon, Revised effective ionic radii and systematic studies of interatomic
8