Ionic-liquid-catalyzed decarboxylation of glycerol carbonate to glycidol

Article abstract of DOI:10.1016/j.jcat.2012.10.015

Decarboxylation of glycerol carbonate (GLC) to produce 2,3-epoxy-1-propanol (glycidol) was conducted using various kinds of ionic liquids (ILs) as catalysts. ILs bearing an anion with medium hydrogen-bond basicity such as NO3- and I^- exhibited the higher glycidol yields than those having an anion with low or strong hydrogen-bond. FT-IR spectroscopic analysis shows that both GLC and glycidol interact with anions of ILs through their hydroxyl groups. It was possible to improve the yield of glycidol when a zinc salt with a medium Lewis acidity was co-present along with an IL. The yield of glycidol was greatly increased up to 98% when the decarboxylation was conducted in the presence of a high-boiling aprotic solvent. Computational calculations on the mechanism using 1-butyl-3-methylimidazolium nitrate as a catalyst revealed that the first step is the NO3 - assisted ring-opening of GLC followed by the ring closure, resulting in the formation of a 3-membered ring intermediate species.

Full text of DOI:10.1016/j.jcat.2012.10.015

Journal of Catalysis 297 (2013) 248–255
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Ionic-liquid-catalyzed decarboxylation of glycerol carbonate to glycidol
Ji Sik Choi a,b, Fidelis Stefanus Hubertson Simanjuntaka a,b, Ji Young Oh c, Keun Im Lee a, Sang Deuk Lee a,b
,
a,b,
Minserk Cheong c, Hoon Sik Kim c, , Hyunjoo Lee
⇑
⇑
a Clean Energy Center, Korea Institute of Science and Technology, 39-1 Hawolgok-dong, Sungbuk-gu, Seoul 136-791, Republic of Korea
b University of Science and Technology, 217 Gajeong-ro, Yuseong-gu. Daejeon, 305-350, Republic of Korea
c Department of Chemistry and Research Institute of Basic Sciences, Kyung Hee University, 1 Hoegi-dong, Dondaemun-gu, Seoul 130-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Decarboxylation of glycerol carbonate (GLC) to produce 2,3-epoxy-1-propanol (glycidol) was conducted
using various kinds of ionic liquids (ILs) as catalysts. ILs bearing an anion with medium hydrogen-bond
basicity such as NO3ꢀ and Iꢀ exhibited the higher glycidol yields than those having an anion with low or
strong hydrogen-bond. FT-IR spectroscopic analysis shows that both GLC and glycidol interact with
anions of ILs through their hydroxyl groups. It was possible to improve the yield of glycidol when a zinc
salt with a medium Lewis acidity was co-present along with an IL. The yield of glycidol was greatly
increased up to 98% when the decarboxylation was conducted in the presence of a high-boiling aprotic
solvent. Computational calculations on the mechanism using 1-butyl-3-methylimidazolium nitrate as a
catalyst revealed that the first step is the NOꢀ3 -assisted ring-opening of GLC followed by the ring closure,
resulting in the formation of a 3-membered ring intermediate species.
Received 5 August 2012
Revised 12 October 2012
Accepted 13 October 2012
Available online 17 November 2012
Keywords:
Glycerol
Glycerol carbonate
Glycidol
Ionic liquids
Decarboxylation
Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction
Glycidol is produced industrially through the oxidation of allyl
alcohol using hydrogen peroxide as the oxidant in the presence
The use of glycerol, produced in large quantities as a co-product
of biodiesels, as a renewable feedstock for synthesizing various va-
lue-added chemicals and clean fuels has been attracting recent
interest because the fate of the biodiesel industry is heavily depen-
dent on the development of suitable applications of glycerol [1–3].
Accordingly, tremendous effort has been devoted to convert glyc-
erol into valuable chemicals, including 1,3-propandiol, epichloro-
hydrine, acrolein, glyceric acid, dihydroxy acetone, etc., but only
few of those chemicals has been commercialized yet, possibly
due to the failure in finding active and selective catalysts. Recently,
focus has also been turned to the synthesis of glycerol carbonate
(GLC) from glycerol because GLC has many potential applications
as a high-boiling polar solvent, an intermediate of polyurethanes,
glycidol, etc. [4,5].
Of various applications of GLC, we are particularly interested in
the synthesis of glycidol because glycidol possesses favorable
properties making it suitable for use in stabilizers, plastics modifi-
ers, surfactants, and fire retardants. Thus, glycidol can be used in a
wide variety of industrial fields including the textile, plastic, phar-
maceutical, cosmetic, and photochemical industries [6,7].
of tungsten oxide-based catalyst or through the reaction of epi-
chlorohydrin with bases [8,9]. However, those conventional pro-
cesses to manufacture glycidol suffer from several drawbacks
such as the high cost of raw materials and/or the generation of
waste by-products. In this context, the synthesis of glycidol from
the direct decarboxylation of GLC, as shown in Scheme 1, is worthy
of attention in terms of economic and environmental points of
view [10,11]. However, unfortunately, only very little information
has been disclosed on the synthesis of glycidol from GLC.
In previous papers, we have shown that the ionic-liquid (IL)-
based catalysts used for the synthesis of alkylene carbonates from
the carboxylation of epoxides are also capable of catalyzing the
decarboxylation of alkylene carbonates in the absence of CO2
[12,13]. With this in mind, we have attempted to use ILs as cata-
lysts for the decarboxylation of GLC to produce glycidol.
Herein, we report that ILs can also be used as efficient catalysts
for the decarboxylation of GLC and that the catalytic activity of an
IL is closely related to the hydrogen-bond basicity of the anion of
the ILs.
2. Experimental
⇑
Corresponding authors. Address: Clean Energy Center, Korea Institute of Science
and Technology, 39-1 Hawolgok-dong, Sungbuk-gu, Seoul 136-791, Republic of
Korea (H. Lee). Fax: +82 29585809.
2.1. General
All of the chemicals used for the synthesis of ILs and glycidol
were purchased from Aldrich Chemical Co. (USA) and used as
E-mail
addresses:
khs2004@khu.ac.kr
(H.S.
Kim),
hjlee@kist.re.kr,
u04623@gmail.com (H. Lee).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.10.015

J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
249
Table 1
O
Activities of various alkali metal salts for the decarboxylation of GLC.a
O
O
- CO2
cat.
O
Entry
Catalyst
C (%)
Y (%)
S (%)
TOF (hꢀ1
)
OH
OH
Glycerol carbonate
1
2
3
4
5
6
7
8
9
None
Na2SO4
NaF
NaCl
NaBr
NaNO3
NaN(CN)2
NaCH3CO2
NaNO2
Na2SO3
NaPF6
4.5
3.4
55.9
38.7
62.0
53.7
67.0
45.5
33.7
40.8
56.4
1.2
73.9
77.4
38.7
63.6
66.6
70.7
45.5
33.7
40.8
56.9
26.1
31.5
–
149
103
165
143
179
121
90
72.2
100
97.5
80.6
94.7
100
100
100
99.2
4.7
Glycidol
Scheme 1. Decarboxylation of glycerol carbonate (GLC) to glycidol.
received without further purifications. GLC was obtained from TCI
(Japan) and ILs including 1-butyl-3-methylimidazolium nitrate
([BMIm]NO3), 1-butyl-2,3-dimethylimidazolium nitrate, and 1-bu-
tyl-3-methylimidazolium bicarbonate ([BMIm]HCO3) were pre-
pared according to the literature procedure [14,15]. Other ILs
tested for the decarboxylation of GLC were purchased from C-Tri
Co. (Korea).
The interactions of ILs with GLC or glycidol were investigated
using a FT-IR (iS10, Nicolet) spectrometer equipped with a Smart
Omni-Transmission accessory.
109
150
3
10
11
12
NaBF4
6.3
2.0
5
TOF (hꢀ1) = moles of glycidol produced/moles of catalyst/h.
Reaction conditions: GLC = 169 mmol, molar ratio of catalyst/GLC = 0.005,
T = 175 °C, P = 2.67 kPa, t = 45 min. C = GLC conversion, Y = glycidol yield, S = glyc-
idol selectivity.
a
2.67 kPa. As can be seen in Table 1, the decarboxylation of GLC pro-
ceeded extremely slowly in the absence of a catalyst, resulting in a
conversion as low as 4.5%. Interestingly, the conversion was almost
quantitative when 0.5 mol% of NaCl with respect to GLC was used
as a catalyst, but the yield and selectivity of glycidol were only
62.0% and 63.6%, respectively, due to the formation of dimeric side
products of glycidol, including 1,4-dioxane-2,5-dimethanol, 3-
(oxrian-2-yloxy)propane-1,2-diol, and 1,4-dioxane-2,6-dimetha-
nol (see Scheme S1 in the Supporting information).
Of the sodium salts tested, only NaNO3 showed higher glycidol
yield than that of NaCl. All the other sodium salts exhibited cata-
lytic performances lower than that of NaCl in terms of either con-
version or yield. It is worthwhile to note that the catalytic activities
of NaPF6 and NaBF4 are almost negligible, strongly indicating that
anions play pivotal roles in the decarboxylation of GLC.
2.2. Decarboxylation of GLC
In a 100 mL three-necked flask equipped with a condenser and
an electrical heater, GLC was decarboxylated into glycidol in the
presence of a catalyst and/or a solvent at a specified temperature
under a reduced pressure of 2.67 kPa. Volatiles produced during
the decarboxylation reaction were collected in a cold receiver im-
mersed in a dry ice–acetonitrile bath. After the completion of the
reaction, the product mixture in the flask were analyzed using a
HPLC (Waters) equipped with an Aminex HPX-87H column (Bio-
rad) and a RI detector (Waters 410). The mobile phase used was
a 5 mM H2SO4 aqueous solution, and the flow rate was set at
0.6 mL/min. The volatiles collected in the cold receiver were ana-
lyzed by a gas chromatograph (Agilent, 6890N) equipped with a
HP-INNOWAX capillary column (30 m ꢁ 0.32 mm ꢁ 0.25
lm) and
a flame-ionization detector, and a GC–Mass spectrometer (Hewlett
3.2. Effect of cation
Packard 6890-Agilent 5973 MSD).
As NaCl exhibited higher performance than most of the other
sodium salts, it was hoped that the alkali metal chlorides, CsCl
and RbCl, which have larger-sized cations than that of Na+, would
give better performance due to the longer cation–anion distance
and consequently due to the increased nucleophilicity or basicity
of Clꢀ in these alkali metal chlorides. However, as shown in Table 2,
there was observed no distinct correlation between the catalytic
activity and the size of the alkali metal cation. In contrast, metal
chlorides with a multivalent cation such as MgCl2, ZnCl2, and AlCl3
exhibited almost no activity for the decarboxylation, implying that
2.3. Computational calculations
The mechanism for the IL-assisted decarboxylation of GLC lead-
ing to glycidol was theoretically investigated using a Gaussian 09
program [16]. The geometry optimizations and thermodynamic
corrections were performed using the hybrid Becke 3-Lee–Yang–
Parr (B3LYP) exchange–correlation functional with the 6-31+Gꢂ ba-
sis sets for C, H, N, and O. For simplicity, [EMIm]NO3 (1-butyl-3-
methylimidazolium nitrate) was used instead of [BMIm]NO3 and
the bulk solvent effect was not included in the calculation. In order
to obtain the most stable geometries, all possible interaction pat-
terns were optimized. Restrictions on symmetries were not im-
posed on the initial structures. All stationary points were verified
as minima through full calculation of the Hessian and a harmonic
frequency analysis.
Table 2
Effect of cation for the decarboxylation of GLC.a
Entry
Catalyst
C (%)
Y (%)
S (%)
TOF (hꢀ1
)
1
2
3
4
5
6
7
8
9
NaCl
LiCl
KCl
RbCl
97.5
98.8
92.6
99.6
99.9
99.8
4.9
2.4
13.4
3.5
62.0
61.2
59.7
54.7
62.4
57.1
2.5
2.1
1.5
2.0
1.8
63.6
61.9
64.5
54.7
62.4
57.3
51.0
77.8
11.3
57.1
34.6
165
163
159
146
166
152
7
6
4
5
5
3. Results and discussion
CsCl
3.1. Activities of alkali metal salts: effect of anion
[BMIm]Cl
MgCl2
AlCl3
SnCl4
ZnCl2
FeCl3
As the first step for the development of high-performance cata-
lysts for the decarboxylation of GLC, we have undertaken a com-
prehensive investigation of the activities of various sodium salts
because some sodium salts like NaCl and Na2SO4 are known to cat-
alyze the decarboxylation of GLC [11,17].
10
11
5.2
a
Reaction conditions: GLC = 169 mmol, molar ratio of catalyst/GLC = 0.005,
T = 175 °C, P = 2.67 kPa, t = 45 min. C = GLC conversion, Y = glycidol yield, S = glyc-
Decarboxylation of GLC was conducted in the presence of an al-
kali metal salt at 175 °C for 45 min under a reduced pressure of
idol selectivity, TOF (hꢀ1) = moles of glycidol produced/moles of catalyst/h.

250
J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
the Lewis acidity of cations is also an important factor in determin-
ing the catalytic activity of metal halides. From these results, it is
cautiously concluded that both the Lewis acidity of the cation
and the nucleophilicity of the anion should be considered together
for a metal salt catalyst to be active and selective for the decarbox-
ylation of GLC. However, as can be seen in Tables 1 and 2, none of
the metal catalysts tested were very selective toward the forma-
tion of glycidol.
of ILs, a Kamlet–Taft solvent parameter determined by the nature
of the anion of an IL, was introduced. Based on the reported b val-
ues of the ILs [18–21], the correlation of the yield of glycidol with
the b value was plotted. As can be seen in Fig. 1, the yield of glyc-
idol increased with increasing b values up to 0.8, but decreased
thereafter. On the whole, ILs with a b value in the range 0.6–0.8
showed higher glycidol yields than those with b values higher than
0.8 or lower than 0.6. On the contrary, conversions of GLC re-
mained almost constant at over 95% as long as the b value was
above 0.6. From these results, it is clear that, for an IL to be active
and selective, the b value of the anion should be in the range of
approximately 0.6–0.8.
3.3. Ionic-liquid-catalyzed decarboxylation
As a means of improving the selectivity and yield of glycidol, we
have attempted to use ILs because the acidity of the cation and the
nucleophilicity of the anion can be tailored, and some ILs bearing a
halide anion are known to catalyze the decomposition of ethylene
carbonate into ethylene oxide and CO2 [18]. Furthermore, in the
presence of an IL, the decarboxylation can be carried out in a
homogeneous way. The catalytic activity of 1-butyl-3-methylimi-
dazolium chloride ([BMIm]Cl) was evaluated first because
[BMIm]Cl, like NaCl, is also an ionic compound consisting of a cat-
ion and an anion in a 1:1 ratio, and because the Clꢀ of [BMIm]Cl is a
strong nucleophile. However, contrary to our expectation, no
noticeable improvement in the selectivity or yield of glycidol was
observed with the use of [BMIm]Cl, implying that either the acidity
of the cation or the basicity of the anion for the IL is not optimized
for the decarboxylation of GLC. To find more selective ILs through a
suitable combination of cations and anions, the decarboxylation of
GLC was undertaken in the presence of various [BMIm]-based ILs.
As can be seen in Fig. 1, [BMIm]NO3 and [BMIm]I showed consid-
erably higher yields and selectivities of glycidol than did other
ILs. As in the case of sodium metal salts, 1-butyl-3-methylimidazo-
In order to investigate the effect of cation, the catalytic
performance of four different ILs with a nitrate anion, including
[BMIm]NO3, 1-butyl-2,3-dimethylimidazolium nitrate ([BDMIm]
NO3), N-methyl-N-propylpyrrolidinum nitrate ([MPPyr]NO3), and
tetra-n-butylammonium nitrate ([Bu4N]NO3), was compared. As
can be seen in Table 3, the conversion of GLC was almost quantita-
tive irrespective of the ILs, implying that the ILs with a NOꢀ3 is suf-
ficiently active enough to complete the decarboxylation of GLC,
irrespective of the types of cations. In contrast, the yield and selec-
tivity of glycidol were slightly affected by the type of the cation.
[BDMIm]NO3 afforded the higher yield and selectivity of glycidol
than [BMIm]NO3, whereas [MPPyr]NO3 and [Bu4N]NO3 exhibited
lower glycidol yields and selectivity than [BDMIm]NO3. It is partic-
ularly notable that [BDMIm]NO3 showed higher selectivity to glyc-
idol than [BMIm]NO3. The reason for the higher selectivity with
[BDMIm]NO3 is not clear, but the reduction of the acidity of the
imidazolium cation by the substitution of the acidic C(2)–H with
the electron donating methyl group seems to decrease the cat-
ion–anion interaction, consequently resulting in the increase in
the b value of NOꢀ3 . However, considering the higher glycidol yield
and selectivity with [BDMIm]NO3 than those with [BMIm]NO3 hav-
ing the b value in the range 0.6–0.8, it is likely that the increase in
the anion b value is small, because the b value should not exceed
0.8 for the IL to be selective toward the decarboxylation of GLC
(vide supra). One might suspect that the formation of side products
could be related to the interaction between the cation and glycidol
through the epoxy oxygen atom. However, comparison of the FT-IR
spectra for the interactions of glycidol with [BMIm]NO3 and
[BDMIm]NO3 clearly indicated that the interactions of cations with
glycidol were almost negligible because the C–O (epoxy oxygen)
stretching frequency of glycidol was not changed appreciably by
the substitution of the acidic C(2)–H with the methyl group (see
Fig. S1 in Supporting information).
lium acetate ([BMIm]AcO) with a strongly basic anion, CH3COꢀ,
2
afforded glycidol in a much lower yield of 39.3% compared with
that of [BMIm]Cl. Similar to NaBF4 and NaPF6, [BMIm]BF4 and
[BMIm]PF6 with a non-coordinating anion showed negligible activ-
ity toward decarboxylation. Such a strong dependence of glycidol
yield and selectivity on the type of anion again suggests that the
degree of interaction between the anion and the hydroxyl groups
of GLC and glycidol through the hydrogen-bond is the most impor-
tant factor in determining the activity of ILs and the selectivity of
glycidol.
In order have a better insight into the anion effect on the decar-
boxylation of GLC, the concept of hydrogen-bond basicity (b value)
It was anticipated that [Bu4N]NO3 and [MPPyr]NO3 could exhi-
bit similar or higher glycidol selectivity than [BDMIm]NO3 because
the acidities of [Bu4N]+ and [MPPyr]+ are weaker than that of
[BDMIm]+, due to the absence of acidic hydrogen atom. However,
contrary to our expectation, [Bu4N]NO3 and [MPPyr]NO3 showed
considerably lower glycidol yields and selectivities than
1.2
1.0
0.8
0.6
0.4
0.2
0.0
100
80
60
40
20
0
Yield of Glycidol (%)
Conv. of GLC (%)
β value of anion
[BDMIm]NO3. It is presumed that the b values of NOꢀ in these ILs
3
are higher than 0.8.
Table 3
Effect of cation of ILs on the decarboxylation of GLC.a
Entry
Cation
C (%)
Y (%)
S (%)
TOF (hꢀ1
)
1
2
3
4
[BMIm]
[BDMIm]
[Bu4N]
99.8
99.1
99.3
100
68.7
73.2
64.2
58.9
68.9
73.9
63.8
58.9
183
195
171
157
[BMIm]X
[MPPyr]
Fig. 1. Correlation of Kamlet–Taft hydrogen-bond basicity (b) of the anion of ILs
with their activity for the decarboxylation of GLC. Reaction conditions:
GLC = 169 mmol, molar ratio of IL/GLC = 0.005, T = 175 °C, P = 2.67 kPa,
t = 45 min. ꢂ b values were obtained from the references [19–22].
a
Reaction conditions: GLC = 169 mmol, molar ratio of IL/GLC = 0.005, T = 175 °C,
P = 2.67 kPa, t = 45 min. C = GLC conversion, Y = glycidol yield, S = glycidol selec-
tivity, TOF (hꢀ1) = moles of glycidol produced/moles of catalyst/h.

J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
251
(g)
(f)
(e)
(d)
l
3150
l
3284
(e)
(d)
l
3242
l
3394
l
3358
(c)
(b)
l
l
l
3363
3420
(c)
(b)
l
3459
(a)
3417
3420
l
3467
(a)
l
l
1790
3450
3500
3000
2500
2000
1500
4000 3500 3000 2500
2000
1750
1500
Wavenumbers (cm-1
)
-1
)
Wavenumbers (cm
Fig. 3. FT-IR spectra showing the interactions of glycidol with ILs: (a) glycidol, (b)
glycidol–[BMIm]BF4, (c) glycidol–[BMIm]NO3, (d) glycidol–[BMIm]I, and (e) glyc-
idol–[BMIm]Cl. The molar ratio of glycidol to IL was set at 1.
Fig. 2. FT-IR spectra showing the interactions of GLC with ILs: (a) GLC, (b) GLC–
[BMIm]BF4, (c) GLC–[BMIm]OTf, (d) GLC–[BMIm]NO3, (e) GLC–[BMIm]I, (f) GLC–
[BMIm]Cl, and (g) GLC–[BMIm]CH3CO2. The molar ratio of GLC to IL was set at 1.
investigated by FT-IR spectroscopy. As can be seen in Fig. 3, sub-
stantial interactions were observed between ILs and glycidol,
although the interactions were much weaker than those with GLC.
Therefore, for an IL to be active and selective, the interaction of
the anion of the IL with GLC should be strong enough, and the
interaction of the anion of the IL with glycidol must be sufficiently
weak. This is probably the reason why [BMIm]NO3 and [BMIm]I
bearing an anion with a medium b value exhibit selectivity higher
than that of other ILs.
3.4. FT-IR studies
The effect of anions on the activity of ILs toward the decarbox-
ylation of GLC was further demonstrated by FT-IR studies. As can
be seen in Fig. 2, upon contact of GLC with [BMIm]NO3 and [BMI-
m]I bearing an anion with a b value of around 0.70, the absorption
band centered at 3450 cmꢀ1 corresponding to the hydroxyl group
of free GLC moved to lower frequencies of 3363 and 3358 cmꢀ1
,
respectively. The O–H stretching frequency shift was even more
pronounced when GLC was interacted with [BMIm]Cl and [BMI-
m]AcO with a more basic anions having a b values of 0.93 and
0.99, respectively. The peaks corresponding to the hydroxyl group
of GLC appeared at 3242 cmꢀ1 for [BMIm]Cl and at 3150 cmꢀ1 for
[BMIm]AcO, suggesting that the degree of interaction between
the hydroxyl group and the anions of an ILs is of pivotal impor-
tance in determining the activity of the ILs and the yield and selec-
tivity of glycidol.
On the contrary, the O–H stretching frequency of GLC at
3450 cmꢀ1 shifted to a higher frequency of 3467 cmꢀ1 upon inter-
action with [BMIm]BF4, indicating that there is no interaction be-
tween the hydroxyl group of GLC and BFꢀ4 . The higher frequency
shift can be ascribed to the weakening of hydrogen-bonding inter-
actions between GLC molecules due to the dilution with
[BMIm]BF4, thereby resulting in the increase in the O–H bond
strength.
3.5. Effect of catalyst loading
The effect of catalyst loading on the decarboxylation of GLC was
investigated for the [BMIm]NO3- and [BMIm]Cl-catalyzed reac-
tions. As can be seen in Fig. 4, for the decarboxylation in the pres-
ence of [BMIm]NO3, the conversion of GLC quantitative at the
[BMIm]NO3/GLC molar ratio of 0.005. On the contrary, the glycidol
yield and selectivity reached the maximum values of approxi-
mately 69% at the molar ratio of 0.0025. Any change in yield and
selectivity was not observed on further increase in the molar ratio.
For the [BMIm]Cl-catalyzed decarboxylation shown in Fig. 5, the
conversion of GLC reached 100% at the [BMIm]Cl/GLC molar ratio
of 0.0025 and remained quantitative on further increase in the mo-
lar ratio. Unlike the case of the [BMIm]NO3-catalyzed reaction,
however, the yield of glycidol decreased gradually above the molar
ratio of 0.0025. It is likely that the interaction of glycidol with Clꢀ,
one of the strongest basic anions, is more accelerated with the
increasing concentration of [BMIm]Cl, thereby promoting the
In consideration of the experimental and FT-IR results, it is obvi-
ous that the activation of the O–H bond of GLC by the anion of an IL
through hydrogen-bonding interaction is a prerequisite for GLC to
be decarboxylated as depicted in Scheme 2. The stronger the inter-
action is, the easier the decarboxylation is.
100
80
Although most of the ILs tested was highly active for the decar-
boxylation of GLC, the performance of these ILs still needs to be
greatly improved in terms of yield and selectivity of glycidol. It is
postulated that the low yield and selectivity of glycidol are resulted
from the activation of the hydroxyl group of glycidol by the anions
of ILs, which could lead to the formation of side products through
the dimerization and oligomerization of glycidol [23,24]. To con-
firm our postulate, the interactions of glycidol and ILs were
60
Conv. of GLC
Yield of Glycidol
Selec. to Glycidol
40
20
Me
0
O
N+
X-
O
0.000
0.005
0.010
0.015
0.020
H
O
O
H
Molar ratio ([BMIm]NO3 /GLC)
N
Bu
Fig. 4. Effect of [BMIm]NO3/GLC molar ratio on the decarboxylation of GLC.
Reaction conditions: GLC = 169 mmol, T = 175 °C, P = 2.67 kPa, t = 45 min.
Scheme 2. Hydrogen-bonding interaction between [BMIm]X and GLC.

252
J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
100
80
60
40
20
0
100
Conv. of GLC
Yield of Glycidol
Selec. to Glycidol
80
60
40
20
0
Conv. of GLC
Yield of Glycidol
Selec. to Glycidol
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Molar ratio of Zn(NO3)2/[bmim]NO3
0.000
0.005
0.010
0.015
0.020
Molar ratio ([BMIm]Cl/GLC)
Fig. 7. Effect of added Zn(NO3)2 on the [BMIm]NO3-catalyzed decarboxylation of
GLC. Reaction conditions: GLC = 169 mmol, molar ratio of [BMIm]NO3/GLC = 0.005,
T = 175 °C, P = 2.67 kPa, t = 45 min.
Fig. 5. Effect of [BMIm]Cl/GLC molar ratio on the decarboxylation of GLC. Reaction
conditions: GLC = 169 mmol, T = 175 °C, P = 2.67 kPa, t = 45 min.
the experimental results shown in Fig. 4, which shows that the
yield of glycidol at the [BMIm]NO3/GLC molar ratio of 0.0025 is
higher than that at the molar ratio of 0.005. Considering that the
maximum amount of [BMIm]NO3 to be consumed in the formation
of the completely inactive species, [BMIm]2Zn(NO3)2Cl2 by the
reaction with 20 mol% of ZnCl2 is 40 mol% of the initial loading of
[BMIm]NO3, it is conceivable that 60 mol% of [BMIm]NO3 is sur-
vived. Therefore, the glycidol yield obtained at the [BMIm]NO3/
GLC of 0.005 and at the ZnCl2/[BMIm]NO3 molar ratio of 0.2 would
be very similar to that achieved at the [BMIm]NO3/GLC of 0.003 in
the absence of ZnCl2.
As can be seen in Fig. 7, a synergy effect was observed with the
addition of Zn(NO3)2 up to the molar ratio of Zn(NO3)2/[BMIm]NO3
at 0.2, resulting in the increase in the glycidol yield from 68.7% to
77.2%. However, unlike the addition of ZnCl2, the drop in the yield
of glycidol with the increase in the Zn(NO3)2/[BMIm]NO3 molar ra-
tio was not pronounced. The yield of glycidol decreased from 77.2%
to 71.9% with the variation of the molar ratio from 0.2 to 2. This can
be attributed to the much weaker Lewis acidity of Zn(NO3)2 than
that of ZnCl2. The formation of [BMIm]2Zn(NO3)4 can also be con-
ceivable from the interaction of [BMIm]NO3 with Zn(NO3)2, but
the change in b value of NOꢀ3 upon complexation seems not be
significant.
secondary reactions of glycidol such as dimerization and
oligomerization.
3.6. Effect of added zinc salt
Kim et al. reported that [EMIm]Br promotes the decomposition
of ethylene carbonate into ethylene oxide and CO2, and that the
rate of decomposition increased greatly when ZnCl2 was co-pres-
ent with [EMIm]Br at a 1:2 M ratio because of the formation of a
Zn-containing IL, [EMIm]2ZnCl2Br2 [18]. It was anticipated that
the co-use of a zinc salt like ZnCl2 or Zn(NO3)2 with [BMIm]NO3
would give similar synergy effect to that observed in the decompo-
sition of ethylene carbonate. As can be seen in Fig. 6, the yield of
glycidol increased from 68.7% to 74.3% when 0.2 M equivalent of
ZnCl2 was used along with [BMIm]NO3. However, on further in-
crease in the molar ratio of ZnCl2/[BMIm]NO3 above 0.2, the yield
of glycidol and the conversion of GLC dropped rapidly. Moreover,
the catalytic activity of [BMIm]NO3 was completely quenched at
the ZnCl2/[BMIm]NO3 molar ratio of 2. As in the case of [EMIm]Br
and ZnCl2, there seems to be a reaction between [BMIm]NO3 and
Lewis acidic ZnCl2 under the experimental condition, forming
[BMIm]2Zn(NO3)2Cl2-like species. Once [BMIm]NO3 is complexed
with ZnCl2, the b value of NOꢀ gets greatly reduced due to the
The reason for the increase in the glycidol yield with the addi-
tion of 0.2 M equivalent of Zn(NO3)2 to the amount of [BMIm]NO3
can be explained by the higher frequency shift of the O–H stretch-
ing band. As can be seen in Fig. 8, the O–H stretching frequency
shifted from 3363 to 3400 cmꢀ1 with the addition of 0.2 M equiv-
alent of Zn(NO3)2, suggesting that the hydrogen-bonding interac-
tions between NOꢀ3 of [BMIm]NO3 with the hydroxyl groups of
GLC and glycidol are weakened by the interaction of NOꢀ3 with
the Lewis acidic Zn(NO3)2.
3
strong electron-withdrawing effect of Cl ligands, thereby resulting
in the decrease in GLC conversion and glycidol yield. The increase
in glycidol yield with the addition of 0.2 equivalent of ZnCl2 with
respect to [BMIm]NO3 can be rationalized to a certain extent from
100
Conv. of GLC
This is in contrast to the previously reported results on the
decarboxylation of ethylene carbonate. The highest decomposition
was achieved at the ZnCl2/[BMIm]Br molar ratio of 0.5, at which le-
vel the basicity of Brꢀ or Clꢀ becomes the highest [18]. Such a dis-
crepancy seems to have originated from the lack of an additional
hydroxyl group in the ethylene oxide produced. In the decomposi-
tion of ethylene carbonate, secondary reactions of ethylene oxide
would not be possible because any functional group to interact
with basic Brꢀ or Clꢀ is not present in the ethylene oxide molecule.
As listed in Table 4, other Lewis acids like MgCl2, FeCl3, and
AlCl3 were found to exhibit similar synergy effect to that observed
with ZnCl2 when they were used together along with [BMIm]NO3
with the molar ratio of Lewis acid/[BMIm]NO3 at 0.2. As mentioned
above, the anion basicity of NOꢀ3 seems to be weakened by the
presence of a Lewis acid, thereby resulting in the decrease in the
interaction between NOꢀ3 and the hydroxyl group of glycidol.
Yield of Glycidol
Selec. to Glycidol
80
60
40
20
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Molar ratio of ZnCl2/[BMIm][NO3]
Fig. 6. Effect of added ZnCl2 on the [BMIm]NO3-catalyzed decarboxylation of GLC.
Reaction conditions: GLC = 169 mmol, molar ratio of [BMIm]NO3/GLC = 0.005,
T = 175 °C, P = 2.67 kPa, t = 45 min.

J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
253
100
(d)
Conv. of GLC
Yield of Glycidol
Selec. to Glycidol
l
80
60
40
20
3446
(c)
(b)
l
3400
l
3363
(a)
l
0
3450
140
150
160
170
180
190
Temperature (oC)
3500
3000
2500
2000
)
1500
Fig. 9. Effect of reaction temperature on the decarboxylation of GLC. Reaction
conditions: GLC = 169 mmol, molar ratio of [BMIm]NO3/GLC = 0.005, molar ratio of
Zn(NO3)2/[BMIm]NO3 = 0.2, P = 2.67 kPa, t = 45 min.
cm-1
Wavenumbers (
Fig. 8. FT-IR spectra of showing the effect of added ZnCl2 on the [BMIm]NO3-
catalyzed decarboxylation of GLC: (a) GLC, (b) GLC/[BMIm]NO3 (1/1), (c) GLC–
[BMIm]NO3/Zn(NO3)2 (1/1/0.2), and (d) GLC/[BMIm]NO3/Zn(NO3)2 (1/1/0.5). The
numbers in the parentheses are the molar ratios.
100
80
Table 4
Effect of added metal salts on the [BMIm]NO3-catalyzed decarboxylation of GLC.a
60
Entry
Catalyst
C (%)
Y (%)
S (%)
TOF (hꢀ1
)
40
1
2
3
4
5
ZnNO3
ZnCl2
FeCl3
MgCl2
AlCl3
99.0
97.5
99.8
99.4
99.9
77.2
74.3
73.2
74.5
73.8
78.0
76.2
73.3
74.9
73.9
206
198
195
199
197
Conv. of GLC
Yield of Glycidol
Selec. to Glycidol
20
0
a
0
10
20
30
40
50
60
Reaction conditions: GLC = 169 mmol, molar ratio of [BMIm]NO3/GLC = 0.005,
molar ratio of metal chloride/[BMIm]NO3 = 0.2, T = 175 °C, P = 2.67 kPa, t = 45 min.
C = GLC conversion, Y = glycidol yield, S = glycidol selectivity, TOF (hꢀ1) = moles of
glycidol produced/moles of [BMIm]NO3/h.
Reacion time (min)
Fig. 10. Effect of reaction time on the decarboxylation of GLC. Reaction conditions:
GLC = 169 mmol, molar ratio of [BMIm]NO3/GLC = 0.005, molar ratio of Zn(NO3)2/
[BMIm]NO3 = 0.2, P = 2.67 kPa, T = 175 °C.
3.7. Effects of reaction temperature and time
(0.025 g) were loaded and then heated to 175 °C at 2.67 kPa. For
the batch reaction, GLC (33.6 g) was added from the beginning
along with the solvent and was decarboxylated at 175 °C for 2 h.
For the continuous reactions, GLC was introduced into the reaction
flask at 175 °C at a flow rate of 0.2 mL (0.28 g, 2.37 mmol/min) over
a period of 2 h using an HPLC pump. In order to complete the reac-
tion, the mixture was reacted further at that temperature for
30 min. Glycidol was removed from the reaction system as soon
as it formed by performing the reaction under a reduced pressure.
The effects of reaction temperature and time were also investi-
gated using a catalyst system, Zn(NO3)2/[BMIm]NO3 (0.2/1). As can
be seen in Fig. 9, both the conversion of GLC and the yield of glyc-
idol increased with the temperature rise and reached a maximum
at 175 °C. However, further increases in the temperature only re-
sulted in the decrease in the selectivity to glycidol.
The effect of the reaction time was also investigated at 175 °C.
As can be seen in Fig. 10, the decarboxylation was completed in
30 min, indicating that the decarboxylation of GLC proceeded very
rapidly in the presence of Zn(NO3)2/[BMIm]NO3 (0.2/1).
Table 5
Effect of solvent on the decarboxylation of GLC for the batch and continuous
reactions.
3.8. Effect of solvent on the decarboxylation of GLC for the batch and
continuous reactions
Entry
Solvent
C (%)
Y (%)
S (%)
TOF (hꢀ1
)
1a
2b
3a
4b
5a
5b
DMPEG 350
DMPEG 350
Dibenzyl ether
Dibenzyl ether
Dibutyl phthalate
Dibutyl phthalate
99.7
100
100
99.4
99.8
100
83.2
98.2
82.1
97.3
84.4
96.9
83.5
98.2
82.1
97.9
84.6
96.9
128
150
123
148
129
148
In principle, the formation of side products can be suppressed to
a great extent through a suitable combination of cations and an-
ions of ILs. However, the optimal combination seems hard to find.
Therefore, the best way to reduce the formation of side products is
probably to carry out the decarboxylation reaction in a continuous
way using a solvent to minimize the contact between the IL and
glycidol while simultaneously removing glycidol as soon as it
forms. With this in mind, batch and continuous reactions for the
decarboxylation of GLC were conducted in a high-boiling solvent
using a catalytic system consisting of Zn(NO3)2 and [BMIm]NO3
at the molar ratio of 0.2:1. In a 250 mL three-necked flask equipped
with a condenser, electrical heater, and a thermocouple, a high-
boiling solvent solvent (50 g), [BMIm]NO3 (0.13 g), and Zn(NO3)2
C = GLC conversion, Y = glycidol yield, S = glycidol selectivity, TOF (hꢀ1) = moles of
glycidol produced/moles of [BMIm]NO3/h.
Batch reaction: GLC = 284 mmol, solvent 50 g, catalyst = Zn(NO3)2–[BMIm]NO3,
[BMIm]NO3 (0.64 mmol, 0.13 g), molar ratio of Zn(NO3)2/[BMIm]NO3 = 0.2, molar
a
ratio of [BMIm]NO3/GLC = 0.0025, T = 175 °C, P = 2.67 kPa, t = 2.5 h.
b
Continuous reaction: total GLC feed = 284 mmol, GLC feed rate = 0.2 -
mL(2.37 mmol)/min, solvent 50 g, catalyst = Zn(NO3)2–[BMIm]NO3, [BMIm]NO3
(0.64 mmol, 0.13 g), molar ratio of Zn(NO3)2/[BMIm]NO3 = 0.2, T = 175 °C,
P = 2.67 kPa, t = 2.5 h.

254
J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
As can be seen in Table 5, the use of high-boiling solvents such as
glycol dimethyl ether (DMPEG, MW = 350), dibenzyl ether, and
dibutyl phthalate increased the yield and selectivity of glycidol
for both batch and continuous reactions. However, the effect of sol-
vent was more effective for the continuous reactions. For example,
the analysis of the products in a receiver and the flask for the con-
tinuous reaction in DMPEG demonstrated that the decarboxylation
of GLC was quantitative and the yield and selectivity of glycidol
were as high as 98% (entry 2). Among solvents tested, DMPEG
afforded the highest glycidol yield, but difference in glycidol yield
with the solvent was not great.
system consisting of [BMIm]NO3 (0.26 g) and Zn(NO3)2 (0.049 g).
The flow rate of GLC was controlled at 0.2 mL/min (0.28 g,
2.37 mmol/min) throughout the reaction. Every 4 h, the product
in the receiver was transferred into the 100-mL sampling cylinder
by opening the valve between them and was weighed and ana-
lyzed. Fig. 11 shows the change of the molar ratio of the glycidol
recovered/GLC injected for 4 h. As can be seen in Fig. 11, the molar
ratio of glycidol/GLC was maintained in the range 0.97–0.98 for the
first seven samples taken every 4 h, but the molar ratio started to
decrease gradually thereafter down to 0.92. The analysis of the
reaction mixture in the flask after the 40 h reaction showed that
approximately 4.7 mol% of GLC was converted to the side products.
These results demonstrate that the catalytic system, [BMIm]NO3/
Zn(NO3)2, is quite stable under the experimental condition. The
TOF (moles of glycidol produced/moles of [BMIm]NO3/h) and
TON (moles of glycidol produced/moles of [BMIm]NO3) after 40 h
of the continuous reaction were calculated to be 107 and
4.3 ꢁ 103, respectively.
In order to test the catalyst longevity, the continuous reaction
was conducted 175 °C for 40 h in DMPEG (50 g) using the catalytic
1.0
0.8
0.6
0.4
0.2
0.0
3.9. Computational calculations on the mechanism for the formation of
glycidol
The mechanism for the decarboxylation of GLC was theoreti-
cally investigated using [EMIm]NO3 as the catalyst. Fig. 12 shows
the optimized structures for the interactions of [EMIm]NO3 with
GLC. The computational calculations show that the decarboxyl-
ation of GLC to glycidol proceeds in two steps. The first step is
the NO3-assisted ring-opening of GLC and the ring closure, forming
a 3-membered ring. The driving force for the ring-opening seems
to be the strong hydrogen-bonding interaction between the O
atom of the NOꢀ3 and the H atom of the hydroxyl group of GLC,
as revealed from the short H–O distance of 1.40 Å in the optimized
structure of the first transition state (TS1). With such strong inter-
action, the O–H bond can be cleaved and the resulting CH2Oꢀ be-
1
2
3
4
5
6
7
8
9
10
Sample number
Fig. 11. Change of the molar ratio of glycidol collected/GLC injected with the
samples taken every 4 h. Reaction conditions: GLC feed rate = 0.2 mL (2.37 mmol)/
min, [BMIm]NO3 = 0.26 g(1.29 mmol), molar ratio of Zn(NO3)2/[BMIm]NO3 = 0.2,
P = 2.67 kPa, T = 175 °C, t = 40 h, solvent = DMPEG 350. Glycidol collected in the
receiver was withdrawn every 4 h for analysis.
Fig. 12. Optimized structures of reaction intermediates and transition states involved in the [EMIm]NO3-assisted decarboxylation reaction of GLC: (a) reactant
Gr = 0 kcal molꢀ1), (b) TS1 ( GTzS1 = 37.4 kcal molꢀ1), (c) intermediate ( Gi = 26.3 kcal molꢀ1), (d) TS2 ( GzTS2 = 28.8 kcal molꢀ1), and (e) product ( Gp = 10.2 kcal molꢀ1). The
Gibbs free energy of formation ( G) and transition state energy (
Gà) are relative values with respect to that of the reactant.
(D
D
D
D
D
D
D

J.S. Choi et al. / Journal of Catalysis 297 (2013) 248–255
255
comes highly nucleophilic enough to interact with the neighboring
carbon atom, forming C–O bond as manifested by the C–O bond
distance of 1.71 Å. In the optimized structure of the intermediate
species, it is worth to note that there is also a strong interaction be-
tween the C(2)–H of the [EMIm]+ and the carbonate group (O–H
bond: 1.73 Å), suggesting that the intermediate species is stabi-
lized to a certain extent by the hydrogen-bonding interaction with
the imidazolium ring. The intermediate species seems to be further
stabilized via the interaction between the H atom of the dissoci-
ated HNO3 and the ether oxygen atom. The activation energy of
and Technology of Korean government as well as the ‘‘National
Agenda Project’’ program of Korea Research Council of Fundamen-
tal Science & Technology.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.10.015.
References
the first transition state (
tion of the intermediate species (
D
GzTS1) and the Gibbs free energy of forma-
[1] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C.D. Pina, Angew. Chem. Int. Ed.
46 (2007) 4434.
[2] J.A. Kenar, Lipid Technol. 19 (2007) 249.
[3] C.H.C. Zhou, J.N. Beltramini, Y.X. Fan, G.Q.M. Lu, Chem. Soc. Rev. 37 (2008) 527.
[4] C. Hammond, J.A. Lopez-Sanchez, M.H. Ab Rahim, N. Dimitratos, R.L. Jenkins,
A.F. Carley, Q. He, C.J. Kiely, D.W. Knight, G.J. Hutchings, Dalton Trans. 40
(2011) 3927.
[5] J.H. Park, J.S. Choi, S.K. Woo, S.D. Lee, M. Cheong, H.S. Kim, H. Lee, Appl. Catal. A:
Gen. 433 (2012) 35.
[6] M. Pagliaro, M. Rossi, Esterification, in: The Future of Glycerol, New Usages for
a Versatile Raw Material, RSC Green Chemistry Book Series, RSC Publishing,
Cambridge, 2010, p. 108.
[7] O. Gómez-Jiménez-Aberasturi, J.R. Ochoa-Gómez, A. Pesquera-Rodríguez, C.
Ramírez-López, A. Alonso-Vicario, J. Torrecilla-Soria, J. Chem. Technol.
Biotechnol. 85 (2010) 1663.
[8] N. Mizuno, S. Hikichi, K. Yamaguchi, S. Uchida, Y. Nakagawa, K. Uehara, K.
Kamata, Catal. Today 117 (2006) 32.
[9] W.F Richey, Chlorohydrins, in: J.I. Kroschwitz, M. Howe-Grant (Eds.), Kirk-
Othmer Encyclopedia of Chemical Technology, 4th ed., vol. 6, John Wiley, New
York, 1993, p. 140.
D
Gi) were calculated as 37.4 and
26.3 kcal molꢀ1, respectively. It is interesting to note that the inter-
mediate species has an ionic-liquid-type structure consisting of
[BMIm]+ and carbonate anion.
The second step is the proton transfer from HNO3 to an oxygen
atom of the carbonate group along with the simultaneous loss of
CO2. It is likely that the transformation of the intermediate species
proceeds immediately becauzse the activation energy for the second
step is only 2.5 kcal/mol ðDGz TS2
ꢀ GiÞ. The activation energy of the
D
second transition state ð GTS2Þ and the Gibbs free energy of forma-
D
tion of the product
(DGp) were calculated as 28.8 and
10.2 kcal molꢀ1, respectively. As a whole, the computational re-
sults are in good agreement with those obtained from the decar-
boxylation experiments and spectroscopic analysis.
[10] J.W. Yoo, Z. Mouloungui, A. Gaset, Organisation Nationale Interprofessionelle
des Oleagineux, US 6316641 B1, 2001.
4. Conclusions
[11] Y. Seki, T. Sasa, H. Takeuchi, M. Uno, M. Namba, Kao Corp., US 2011/0015414
A1, 2011.
[12] H.S. Kim, J.J. Kim, H. Kim, H.G. Jang, J. Catal. 220 (2003) 44.
[13] J. Peng, Y. Deng, New J. Chem. 25 (2001) 639.
[14] B. Mokhtarani, A. Sharifi, H.R. Mortaheb, M. Mirzaei, M. Mafi, F. Sadeghian, J.
Chem. Thermodyn. 41 (2009) 1432.
[15] Y.S. Choi, Y.N. Shim, J. Lee, C.S. Hong, M. Cheong, H.S. Kim, H.G. Jang, J.S. Lee,
Appl. Catal. A: Gen. 404 (2011) 87.
[16] M.J. Frisch, et al., Gaussian 09. Revision A.1, Gaussian Inc., Wallingford, CT,
USA, 2009 (see SI for details).
[17] M. Uno, M. Okutsu, Kao Corp., EP 2,028,181 A1, 2009.
[18] H.S. Kim, P. Jelliarko, J.S. Lee, S.Y. Lee, H. Kim, S.D. Lee, B.S. Ahn, Appl. Catal. A:
Gen. 288 (2005) 48.
[19] P.G. Jessop, D.A. Jessop, D. Fu, L. Phan, Green Chem. 14 (2012) 1245.
[20] R. Lungwitx, S. Spange, New J. Chem. 32 (2008) 392.
[21] L. Crowhurst, P.R. Mawdsley, J.M. Perez-Arlandis, P.A. Salter, T. Welton, Phys.
Chem. Chem. Phys. 5 (2003) 2790.
In summary, decarboxylation of GLC was conducted using var-
ious kinds of sodium salt and ILs. The experimental and spectro-
scopic results clearly showed that the effect of anions was more
pronounced than that of cations. The correlation of the catalytic
performance of an IL with the b value of the anion demonstrated
that the b value of the anion should be in the range of 0.60–0.80
for the IL to be active and selective for the decarboxylation of
GLC. It was possible to enhance the selectivity of glycidol to a cer-
tain extent by the co-use of Lewis acidic Zn(NO3)2 with
[BMIm]NO3. The yield of glycidol was greatly increased up to
98% when a decarboxylation was carried out using a high-boiling
solvent in a continuous way.
[22] M.A.A. Rani, A. Brant, L. Crowhurst, A. Dolan, M. Lui, N.H. Hassan, J.P. Hallett,
P.A. Hunt, H. Niedermeyer, J.M. Perez-Arlandis, M. Schrems, T. Welton, R.
Wilding, Phys. Chem. Chem. Phys. 13 (2011) 16831.
[23] G. Rokicki, P. Rakoczy, P. Parzuchowski, M. Sobiecki, Green Chem. 7 (2005)
529.
Acknowledgments
We acknowledge the financial support by grants from Korea
CCS R&D Center, funded by the Ministry of Education, Science
[24] S.R. Sandler, F.R. Berg, J. Polym. Sci. 4 (1966) 1253.