Production of diethyl carbonate from ethylene carbonate and ethanol over supported fluoro-perovskite catalysts

Article abstract of DOI:10.1016/j.catcom.2018.01.019

The KCaF₃/C (K-Ca(A)) catalyst was shown to be effective as heterogeneous basic catalysts for the transesterification of ethylene carbonate (EC) and ethanol among fluoro-perovskite (XYF₃)/C (X = K, Cs, Y = Mg, Ca) catalysts. Although the KMgF₃/C (K-Mg (A)) catalyst exhibited the highest catalytic activity among the XYF₃/C catalysts studied, the potassium leaching was observed on K-Mg (A). CO₂ temperature programmed desorption (TPD) revealed that the superior catalytic activity of the XYF₃/C was due to its strong basic sites. CO₂-TPD and XPS measurements indicated that strong basic sites are generated by an increase in the electron density of fluorine.

Full text of DOI:10.1016/j.catcom.2018.01.019

CatalysisCommunications108(2018)7–11
Contents lists available at ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Short communication
Production of diethyl carbonate from ethylene carbonate and ethanol over
supported fluoro-perovskite catalysts
Hajime Iida⁎, Ryuhei Kawaguchi, Kazu Okumura
Depart of Applied Chemistry, School of Advanced Engineering, Kogakuin University, 2665-1, Nakano-machi, Hachioji-shi, Tokyo 192-0015, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords:
The KCaF3/C (K-Ca(A)) catalyst was shown to be effective as heterogeneous basic catalysts for the transester-
ification of ethylene carbonate (EC) and ethanol among fluoro-perovskite (XYF3)/C (X = K, Cs, Y = Mg, Ca)
catalysts. Although the KMgF3/C (K-Mg (A)) catalyst exhibited the highest catalytic activity among the XYF3/C
catalysts studied, the potassium leaching was observed on K-Mg (A). CO2 temperature programmed desorption
(TPD) revealed that the superior catalytic activity of the XYF3/C was due to its strong basic sites. CO2-TPD and
XPS measurements indicated that strong basic sites are generated by an increase in the electron density of
fluorine.
Fluoro-perovskite
Transesterification
Diethyl carbonate
1. Introduction
A mechano-chemical technique has been often used for various
complexes synthesis at lower temperature [19–21]. The application to
Diethyl carbonate (DEC) has been applied as a solvent for paint and
fragrances, the electrolyte in Li-ion batteries, and as an ethylization or
carbonylation reagent in organic synthesis [1–3]. DEC has also received
attention as an oxygen-containing fuel additive [3–5]. The transester-
ification of ethylene carbonate (EC) and ethanol (Eq. (1)) for the pro-
duction of DEC is an environmentally harmonized process because the
EC reactant is produced through the reaction of ethylene oxide and
carbon dioxide, which is a greenhouse gas (Eq. (2)) [2,6].
supported catalyst preparation is expected to highly disperse active
component. Therefore, the present study was applied the mechano-
chemical treatment to the preparation of supported fluoro-perovskite
catalysts.
The purpose of the present study was to explore the effects of the
alkali and alkali earth metals and catalyst preparation procedure on the
catalytic performance of activated carbon-supported fluoro-perovskite
(XYF3/C, X = K, Cs, Y = Mg, Ca) catalysts during the transesterifica-
tion of EC and ethanol.
2. Experimental
(1)
(2)
Supported fluoro-perovskite catalysts were prepared using a me-
chano-chemical
method.
Activated
carbon
(FY-1,
Cataler,
S
BET = 800 m2 g−1), SiO2 (Q-6, Fuji Silysia, SBET = 275 m2 g−1), or
Basic catalysts are effective for catalyzing the transesterification
reaction. Various solid basic catalysts have been reported for this pur-
pose, including alkaline earth metal oxides [7–13], polymer resins [14],
and basic zeolites [15], because the separation and recovery of these
catalysts is easier than that of conventional homogeneous basic cata-
lysts. Among them, alkaline fluorides are an attractive choice due to
their high catalytic activity in transesterification reactions
[2,6,12,16,17]. For instance, we recently reported that supported
fluoro-perovskite catalysts, which are prepared by an impregnation
method, are effective for the transesterification of an edible oil doped
with fatty acids and methanol [18].
Al2O3 (JRC-ALO-6, a reference catalyst of the Catalysis Society of
Japan, SBET = 161 m2 g−1) was used as the support. Alkaline fluoride
(XF, where X = K, Cs) and alkaline earth hydroxide (Y(OH)2, where
Y = Mg, Ca) were combined with the support powder and zirconia balls
in zirconia vessels. In the case of an unsupported KMgF3 catalyst, KF
and Mg(OH)2 were combined with zirconia balls in zirconia vessels. The
molar ratio of the alkaline metal to the alkaline earth metal was 3:1.
The molar amount of the alkaline earth metal relative to the support
powder was 20 mmol gsupport−1. The zirconium vessels were placed in a
planetary ball mill (Frisch, P-7), and the mechano-chemical treatment
was performed at a rotation rate of 800 rpm for 30 min. As shown in the
⁎
Corresponding author.
E-mail address: iida@cc.kogakuin.ac.jp (H. Iida).
https://doi.org/10.1016/j.catcom.2018.01.019
Received 22 September 2017; Received in revised form 13 January 2018; Accepted 15 January 2018
1566-7367/©2018ElsevierB.V.Allrightsreserved.

H. Iida et al.
CatalysisCommunications108(2018)7–11
Fig. 1. XRD profiles of as-prepared catalysts.
: KCaF3 (JCPDS 21-1341)
: KMgF3 (JCPDS 18-1033)
: CsCaF3 (JCPDS 21-0817) : CaF2 (JCPDS 35-0516)
: K2MgF4 (JCPDS 23-0469) : MgO (JCPDS 45-0946)
500
Cs-Ca (A)
d= 24.5 nm
SBET= 11 m2g-1
K-Ca (A)
d= 30.7 nm
SBET= 153 m2g-1
K-Mg (A)
d= 10.7 nm
SBET= 36 m2g-1
K-Mg (B)
d= 19.8 nm
SBET= 196 m2g-1
20
25
30
35
2 / deg.
40
*
45
50
Crystallite size XYF3(110), K2MgF4(103).
BET: BET surface area
S
chromatograph (FID-GC; DB-1 capillary column, 60 m, 0.52 mm i.d.,
GC-14B, Shimadzu). The DEC yield was defined as:
90
80
70
60
50
40
30
20
10
0
: K-Mg (A)
: K-Mg (B)
: Cs-Ca (A)
: K-Ca (A)
DEC yield = molar amount of DEC formed/molar amount of EC charged
(4)
To examine the reusability of catalyst, 2nd run test was performed.
After the 1st run test, the catalyst was separated from the reaction so-
lution by filtration, the separated catalyst was washed with ethanol and
then dried at 373 K, calcined at 773 K. the resultant material was used
to the 2nd run test.
The Brunauer–Emmett–Teller (BET) surface areas of the newly-
prepared catalysts were determined by N2 adsorption at 77 K using a
flow absorption apparatus (Flow sorb II 2300, Micromeritics). The
composition of the flow gas was N2: He = 30: 70 on a volume basis.
Each catalyst was degassed at 473 K for 15 min, prior to measurement.
X-ray diffraction (XRD; RINT2000, Rigaku) analysis was performed
using Cu Kα radiation. The size of fluoride crystallites was calculated
using the Scherrer equation. CO2 temperature programmed desorption
(TPD) measurements were carried out with a pulse reactor in con-
junction with a mass spectrometer (M-200GA-DM, Canon Anelva) to
detect desorbed carbon dioxide at m/z = 44. Catalyst samples were
pretreated at 773 K for 1 h in a stream of helium. After cooling, carbon
dioxide was adsorbed on the catalysts at ambient temperature and CO2-
TPD data were acquired at a heating rate of 10 K min−1 in a steam of
helium up to a maximum temperature of 1073 K. X-ray photoelectron
spectroscopy (XPS) was performed using a Quantum 2000 (Ulvac Phi)
instrument with an Al anode, a beam diameter of 100 μm and an ac-
celeration voltage of 15 kV. Leaching tests for alkaline metal were
carried out with an atomic absorption spectrometry (AAS; AA-7000,
Shimadzu). Absorption wavelengths for K and Cs were 767 nm and
827 nm, respectively. After the separation of catalyst by the filtration,
the alkaline metal concentration in the reaction solution was analyzed.
0
100
200
300
400
Reaction time / min
Fig. 2. Catalytic activities of XYF3/C catalysts.
XRD profile of K-Mg(A) after the milling process (Fig. S1), the desired
fluoro-perovskites were formed during this step according to the fol-
lowing equation,
3XF + Y(OH)2 → XYF3 + 2 XOH
(3)
In order to eliminate the alkaline hydroxide byproduct from the
catalyst, the material obtained from the mechano-chemical treatment
was washed three times with methanol and then dried at 373 K for 12 h,
and finally calcined at 773 K for 1 h in a stream of N2 (termed catalyst
(A)). In addition, the K-Mg catalyst was also prepared with the calci-
nation and washing steps reversed (termed catalyst (B)). The flowcharts
for the both preparation procedures were described in supplementary
material (Fig. S2).
3. Results and discussion
The transesterification of EC was performed using a batch reactor.
As shown in Fig. S3, K-Ca(A)/C (Activated carbon) exhibited the
highest catalytic activity among the supported K-Ca (A) catalysts. In
addition, the catalytic activity of K-Mg (A)/C was higher than that of
unsupported K-Mg (A) catalyst (Fig. S4). Hereinafter, the activated
carbon was used as the support. As shown in Fig. 1, diffraction peaks
The reaction conditions were as follows: 10 mmol of EC, 400 mmol of
−1
ethanol, 0.5 g of catalyst (0.57 g-cat g-EC
of the amount of catalyst to
EC), reaction temperature of 343 K. The amounts of reactants and
products were determined using flame ionization detector/gas
a
8

H. Iida et al.
CatalysisCommunications108(2018)7–11
(A)
(B)
-1
Cs-Ca(A) Total: 0.63 mmol-CO2 gcat
-1
Cs-Ca(A) Total: 0.55 mmol-CO2 gcat
0.37
0.37
-1
S peak: 0.46 mmol-CO2 gcat
-1
S peak: 0.37 mmol-CO2 gcat
-1
W peak: 0.17 mmol-CO2 cat
g
-1
W peak: 0.18 mmol-CO2 gcat
-1
K-Ca(A) Total: 0.75 mmol-C-1O2 gcat
-1
K-Ca(A) Total: 0.63 mmol-C-1O2 gcat
S peak: 0.42 mmol-CO2 gcat
S peak: 0.34 mmol-CO2 gcat
W peak: 0.29 mmol-CO2 gcat
-1
W peak: 0.33 mmol-CO2 gcat
-1
-1
K-Mg(A) Total: 0.89 mmol--C1O2 gcat
-1
-1
K-Mg(A) Total: 0.83 mmol--C1O2 gcat
S peak: 0.74 mmol-CO2 gcat
S peak: 0.62 mmol-CO2 gcat
-1
W peak: 0.15 mmol-CO2 gcat
-1
W peak: 0.21 mmol-CO2 gcat
-1
K-Mg(B) Total: 0.79mmol-C-1O2 gcat
K-Mg(B) Total: 0.62 mmol--C1O2 gcat
S peak: 0.38 mmol-CO2 gcat
S peak: 0.50 mmol-CO2 gcat
W peak: 0.29 mmol-CO2 gcat
-1
-1
W peak: 0.24 mmol-CO2 gcat
350 450 550 650 750 850 950 1050
Temperature / K
350 450 550 650 750 850 950 1050
Temperature / K
Fig. 3. CO2-TPD profiles of as-prepared and used catalysts: (A) as-prepared, (B) used.
(A) < K-Mg(A) ≪ K-Ca (A) < K-Mg (B), whereas the crystallite size
100
80
XYF3/C ( : 1st
K2MgF4/C ( : 1st, : 2nd)
: 2nd)
of fluoride was increased in the following order: K-Mg (A) < K-Mg
(B) < Cs-Ca (A) < K-Ca (A). The relationship between the both was
not confirmed, although the supported active component dispersion
was generally increased with an increase in the surface areas of catalyst.
Fig. 2 shows the catalytic activities of the K-Mg (A, B), K-Ca (A), and
Cs-Ca (A) catalysts for transesterification of EC and ethanol. By the way,
another product was not detected except for DEC and ethylene glycol as
products for transesterification (Eq. (1)). Therefore, the selectivity for
DEC was 100% and the conversion of EC is equal to the yield of DEC.
The catalytic activity K-Mg (A) with a perovskite structure was re-
markably higher than that of K-Mg (B) with a layered perovskite-like
structure. Although the formation of MgO was observed for K-Mg (A
and B) in Fig. 1, the DEC yield for MgO/C (5.3% @ 6 h.) which was
prepared using Mg (OH)2 only, was smaller than those for K-Mg (A and
B). This result indicates that MgO little contribute to the reaction. The
catalytic activity of XYF3/C with perovskite structure decreased in the
following order: K-Mg > Cs-Ca > K-Ca. The order of the catalytic
activity agreed with the order of crystallite size of fluoride, not relating
to the BET surface area. This indicates that the fluoride dispersion af-
fects to the catalytic activity.
60
40
20
0.00 0.20 0.40 0.60 0.80 1.00 1.20
Strong basic site density
(Amount of CO2 desorbed (S peak))
-1
/ mmol-CO2/gcat
Fig. 3 (A) shows the CO2-TPD profiles of as-prepared catalysts.
Desorption peaks assignable to strong basic sites (S peak) were observed
at temperatures over 750 K. and desorption peaks assignable to weak
basic sites (W peak) were observed at temperatures below 750 K. As
shown in Fig. 3(A), the total amounts of CO2 desorbed on the as-pre-
pared catalysts were increased in the following order: Cs-Ca (A) = K-Ca
(A) < K-Mg (B) < K-Mg (A). This indicates that the total number of
basic sites do not relate to the catalytic activity so much. Fig. 4 shows
the relationship between the catalytic activity and number of strong
basic sites. The DEC yields for XYF3/C almost correlated to the number
of strong basic sites. Whereas the K-Mg (B) with layered perovskite like
structure did not agree with the correlation for XYF3/C. The difference
of catalytic activity between perovskite and layered perovskite like
structures cannot explained by the number of strong basic sites. Fig. 5
shows the XP spectra of as-prepared and used catalysts. As shown in
Fig. 5, the binding energy for F1s of as-prepared catalyst was decreased
in the following order: K-Ca (A) > K-Mg (A), K-Mg (B) > Cs-Ca (A).
The shift toward to lower energy indicates an increase in the electron
density of fluorine. As shown in Fig. 3 (A), the peak temperature of
Fig. 4. Relationship between DEC yield and strong basic site density of catalysts (1st and
2nd).
assignable to KMgF3 (perovskite, JCPDS 18-1033) were observed for K-
Mg (A), whereas peaks assignable to K2MgF4 (layered perovskite-like,
JCPDS 23-0469) were observed for K-Mg (B). In addition, the formation
of MgO as a byproduct was observed for both catalysts. The difference
in the washing and calcination procedure during the catalyst prepara-
tion greatly affected the crystalline structure of the K-Mg fluoride. The
residual potassium probably caused the transition from a perovskite
structure to a layered perovskite-like structure during the calcination.
In the case of the K-Ca (A) and Cs-Ca (A) catalysts, peaks assignable to a
perovskite structure were observed for the K-Ca (A) and Cs-Ca (A)
catalysts together with the peaks assignable to CaF2 (JCPDS 35-0516).
The formation of CaF2 would be attributed to the loss of some of the
XCaF3 (X = K, Cs) perovskite from the catalysts via leaching of po-
tassium during the washing step after the milling. The BET surface area
and crystallite size of fluoride of the prepared catalysts were shown in
Fig. 1. The BET surface area was increased in the following order: Cs-Ca
9

H. Iida et al.
CatalysisCommunications108(2018)7–11
Fig. 5. XP spectra of as-prepared and used catalysts.
K2p
Cs3d5/2
F1s
684.9 eV
683.9 eV
685.8 eV
K-Ca (A)
Cs-Ca (A)
686.9 eV
K-Mg (A)
K-Mg (B)
K-Ca (A)
685.5 eV
684.9 eV
K-Mg (A)
K-Mg (B)
Cs-Ca (A)
685.3 eV
685.3 eV
As prepared
Used
As prepared
Used
As prepared
Used
300298296294292290 735 730 725 720 715
695 690 685 680
Binding energy / eV
Binding energy / eV
Ca2p
Binding energy/ eV
Mg2p
As prepared
Used
Cs-Ca (A)
K-Mg (A)
As prepared
Used
K-Ca (A)
K-Mg (B)
370 360 350 340
Binding energy / eV
65
55
45
Binding energy / eV
Table 1
Catalyst reusability and leaching of alkaline metals.
of binding energy for F1s and the peak temperature of main S peak.
Therefore, the strong basic site would be generated by an increase in
the electron density of fluorine. The strong Lewis basic site would be
assigned to active sites for the reaction based on the correlation in
Fig. 4. Catalyst reusability and leaching of alkaline metals in the reac-
tion solution were summarized in Table 1. It is clear from Table 1 that
XYF3/C exhibited superior reusability compared to KF/C, although
leaching of alkaline metal was confirmed. The catalytic activity of K-Mg
(A) in the 2nd run was slightly lower than that in 1st run. The catalytic
activity of K-Ca (A) was increased in the 2nd run. Fig. 3 (B) shows the
CO2-TPD profiles of used catalysts (after 1st run). As shown in Fig. 3
(B), the strong basic site density on the K-Mg (A), K-Mg (B), and Cs-Ca
(A) catalysts decreased as a result of after the 1st run. The percentage
decrease of strong basic sites increased in the following order: K-Mg (A)
(16.2%) < Cs-Ca (A) (19.6%) < K-Mg (B) (24.0%). In contrast, the
strong basic site density on K-Ca (A) increased as a result of the reaction
(by 23.5%). The variation of catalytic activity between 1st and 2nd runs
Catalyst
DEC yield/%a
Concentration of alkaline metals in reaction
solution
1st
2nd
2nd/1st /wt%
K-Mg (A)
K-Ca (A)
Cs-Ca (A) 72.7 54.9 0.76
K-Mg (B)
KF/C
78.1 76.4 0.98
64.3 73.1 1.14
0.22
0.09
0.47
0.02
0.41
32.6 23.1 0.71
69.2 2.3 0.03
a
Reaction time: 6 h.
main S peak (the largest peak among S peaks) was increased with in the
following order: K-Ca (A) (855 K) < K-Mg (A) (865 K), K-Mg (B)
(880 K) > Cs-Ca (A) (952 K). There is a correlation between the shift
10

H. Iida et al.
CatalysisCommunications108(2018)7–11
of K-Ca (A) agrees with the variation of strong basic site density of CO2-
TPD and the peak shift of F1s of XP spectra. In addition, there are also a
correlation between the DEC yield at 2nd run and strong basic site
density as shown in Fig. 4. Although the details are unknown, the strong
basic site density was increased by the change in the structure during
the reaction, resulting in an increase in the catalytic activity of K-Ca
(A). The concentration of alkaline metals in the reaction solution was
increased in following order: K-Mg(B) < K-Ca (A) < K-Mg(A) < KF/
C < Cs-Ca (A) as shown in Table 1. In contrast, it is clear from Fig. 5
that the peak intensities for K 2p, and F1s, and Mg 2p (or Ca 2p)
changed little as a result of the reaction in the cases of the K-Mg (A), K-
Mg (B), and K-Ca (A) catalysts, although the peak intensities for Cs 3d5/
2, Ca 2p, and F1s of the Cs-Ca (A) decreased remarkably as a result of
the reaction. This result indicates that most of alkaline metal was re-
insured on K-Mg (A, B) and K-Ca (A). From the above facts, K-Ca (A)
was indicated to exhibit superior catalytic activity and reusability, re-
sistance against alkaline metal leaching.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.catcom.2018.01.019.
References
[1] A.G. Shaikh, S. Sivaram, Chem. Rev. 96 (1996) 951–976.
[2] C. Murugan, H.C. Bajaj, Fuel Process. Technol. 92 (2011) 77–82.
[3] T. Kawamura, A. Kimura, M. Egashira, S. Okada, J. Yamaki, J. Power Sources 104
(2002) 260–264.
[4] D. Li, W. Fang, Y. Xing, Y. Guo, R. Lin, J. Hazard. Mater. 161 (2009) 1193–1201.
[5] B.C. Dunn, C. Guenneau, S.A. Hilton, J. Pahnke, J. Dworzanski, H.L.C. Meuzelaar,
J.Z. Hu, M.S. Solum, R.J. Pugmire, Energy Fuel 16 (2002) 177–181.
[6] C. Murugan, H.C. Bajaj, R.V. Jasra, Catal. Lett. 137 (2010) 224–231.
[7] M. Kouzu, T. Kasuno, M. Tajika, Y. Sugimoto, S. Yamanaka, J. Hidaka, Fuel 87
(2008) 2798–2806.
[8] M. Kouzu, J. Hidaka, Fuel 93 (2012) 1–12.
[9] X. Liu, H. He, Y. Wang, S. Zhu, X. Piao, Fuel 87 (2008) 216–221.
[10] S. Yan, H. Lu, B. Liang, Energy Fuel 22 (2008) 646–651.
[11] S. Mahesh, A. Ramanathan, K.M.M.S. Begum, A. Narayanan, Energy Convers.
Manag. 91 (2015) 442–450.
[12] H. Liu, L. Su, Y. Shao, L. Zou, Fuel 97 (2012) 651–657.
[13] X. Yu, Z. Wen, H. Li, S. Tu, J. Yan, Fuel 90 (2011) 1868–1874.
[14] Y. Ren, B. He, F. Yan, H. Wang, Y. Cheng, L. Lin, Y. Feng, J. Li, Bioresour. Technol.
113 (2012) 19–22.
4. Conclusion
KCaF3/C catalyst were found to be relatively effective as hetero-
geneous catalysts for the transesterification of EC and ethanol com-
pared to KF/C. Although K-Mg (A) had superior catalytic activity
among the XYF3/C catalysts studied, the potassium leaching is re-
markably observed in the reaction solution.
There is a correlation between the DEC yield and the strong basic
site density of XYF3/C catalysts. Therefore, the strong basic sites were
indicated to contribute to the catalytic activity for transesterification of
EC. In addition, the shift of the binding energy for F1s was correlated to
the peak intensity of CO2 desorbed at the highest temperature.
Therefore, the strong basic site would be generated by the electron
donation from alkaline metal to fluorine.
[15] R. Sree, N.S. Babu, P.S.S. Prasad, N. Lingaiah, Fuel Process. Technol. 90 (2009)
152–157.
[16] M. Verziu, M. Florea, S. Simon, V. Simon, P. Filip, V.I. Parvulescu, C. Hardacre, J.
Catal. 263 (2009) 56–66.
[17] P. Qiu, B. Yang, C. Yi, S. Qi, Catal. Lett. 137 (2010) 232–238.
[18] H. Iida, K. Fukasawa, D. Sekine, A. Igarashi, Fuel Process. Technol. 163 (2017)
16–19.
[19] A. Perejón, N. Murafa, P.E. Sánchez-Jiménez, J.M. Criado, J. Subrt, M.J. Diánez,
L.A. Pérez-Maqueda, J. Mater. Chem. C 1 (2013) 3551–3562.
[20] V. Manivannan, P. Parhi, J.W. Kramer, Bull. Mater. Sci. 31 (2008) 987–993.
[21] G. Branković, V. Vukotić, Z. Branković, J.A. Varela, J. Eur. Ceram. Soc. 27 (2007)
729–732.
11