Sterically controlling 2-carboxylated imidazolium salts for one-step efficient hydration of epoxides into 1,2-diols

Article abstract of DOI:10.1039/d1gc00788b

In order to overcome the disadvantages of excessive water and many byproducts in the conventional process of epoxide hydration into 1,2-diols, 2-carboxylated imidazolium salts were first adopted as efficient catalysts for one-step hydration of epoxides into 1,2-diols. By regulating the cation chain lengths, different steric structures of 2-carboxylated imidazolium salts with chain lengths from C1 to C4 were prepared. The salt with the shortest substituent chain (DMIC) exhibited better thermal stability and catalytic performance for hydration, achieving nearly 100% ethylene oxide (EO) conversion and 100% ethylene glycol (EG) selectivity at 120 °C, 0.5 h with just 5 times molar ratio of H₂O to EO. Such a tendency is further confirmed and explained by both XPS analysis and DFT calculations. Compared with other salts with longer chains, DMIC has stronger interaction of CO₂^?anions and imidazolium cations, exhibiting a lower tendency to release CO₂^?and form HO-CO₂^?, which can nucleophilically attack and synergistically activate ring-opening of epoxides with imidazolium cations. The strong huge sterically dynamic structure ring-opening transition state slows down the side reaction, and both cations and anions stabilized the transition state imidazolium-EG-HO-CO₂^?, both of which could avoid excessive hydration into byproducts, explaining the high 1,2-diol yield. Based on this, the cation-anion synergistic mechanism is then proposed.

Full text of DOI:10.1039/d1gc00788b

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Sterically controlling 2-carboxylated imidazolium
salts for one-step efficient hydration of epoxides
into 1,2-diols†
Cite this: Green Chem., 2021, 23,
992
2
a,b
a,c
a
a
a,b
a,c
Qian Su,
Xin Tan, Xiaoqian Yao, Ting Ying, Li Dong,
Mengqian Fu,
a,b
a,b
Weiguo Cheng
*
and Suojiang Zhang
*
In order to overcome the disadvantages of excessive water and many byproducts in the conventional
process of epoxide hydration into 1,2-diols, 2-carboxylated imidazolium salts were first adopted as
efficient catalysts for one-step hydration of epoxides into 1,2-diols. By regulating the cation chain lengths,
different steric structures of 2-carboxylated imidazolium salts with chain lengths from C1 to C4 were pre-
pared. The salt with the shortest substituent chain (DMIC) exhibited better thermal stability and catalytic
performance for hydration, achieving nearly 100% ethylene oxide (EO) conversion and 100% ethylene
glycol (EG) selectivity at 120 °C, 0.5 h with just 5 times molar ratio of H
further confirmed and explained by both XPS analysis and DFT calculations. Compared with other salts
with longer chains, DMIC has stronger interaction of CO −
anions and imidazolium cations, exhibiting a
lower tendency to release CO − and form HO–CO −
, which can nucleophilically attack and synergistically
2
O to EO. Such a tendency is
2
2
2
activate ring-opening of epoxides with imidazolium cations. The strong huge sterically dynamic structure
ring-opening transition state slows down the side reaction, and both cations and anions stabilized the
Received 2nd March 2021,
Accepted 16th March 2021
transition state imidazolium-EG-HO–CO −
, both of which could avoid excessive hydration into bypro-
2
DOI: 10.1039/d1gc00788b
ducts, explaining the high 1,2-diol yield. Based on this, the cation–anion synergistic mechanism is then
rsc.li/greenchem
proposed.
2
resulting from a large amount of byproducts (Scheme 1). As
one of the energy-intensive chemical processes, EG production
Introduction
1
,2-Diols, such as ethylene glycol (EG) and 1,2-propylene glycol has triggered great interest for technical innovation both in
(
PG), are important raw materials of polyester resins, anti- academia and industry, of which two-step transformation from
1
freezes, cosmetics, medicines, and other products. Especially CO2 and EO is thought to be an alternative energy-saving
in recent years, the increasing consumption of polyester pro- process for EG production. A two-step process via first fix-
3
,4
ducts has greatly promoted the demand for EG, which could ation of CO into ethylene carbonate (EC) and then hydrolysis
2
be above 17 million tons per year by 2018. However, the into EG provides an efficient way for high selectivity with a
2
current industrial production process for EG, the direct lower H O/EO molar ratio (<1.5) and realizes the efficient util-
3
,5
hydration of ethylene oxide (EO), is less competent in this task ization of CO2. However, it needs excessive CO to comple-
2
due to the huge consumption caused by a large excess use of tely convert EO and the recycling of the residual CO –EO
2
6
water (22 molar ratio of H
2
O to EO) and low EG yield (∼89%) mixture makes the process more complex and unsafe. For
simplicity, one-step catalytic hydration of EO with no introduc-
tion of CO2 is an alternative route. To date, various acid or
a
base catalysts have been developed for improving EG yield at a
Key Laboratory of Green Process and Engineering, State Key Laboratory of
Multiphase Complex Systems, Beijing Key Laboratory of Ionic Liquids Clean Process,
Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190,
China. E-mail: wgcheng@ipe.ac.cn, sjzhang@ipe.ac.cn; Fax: +86 10 8262 7080;
Tel: +86 10 8262 7080
1
,7–11
low H O/EO ratio,
e.g. carboxylic acid-carboxylate compo-
2
1
2
13
sites, NY bimetallic salts,
amines,
supported metal
b
Innovation Academy for Green Manufacture, Chinese Academy of Sciences, Beijing
1
00190, China
c
School of Chemistry and Chemical Engineering, University of Chinese Academy of
Sciences, Beijing 100049, China
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
d1gc00788b
Scheme 1 Conventional direct hydration process of EO into EG.
2992 | Green Chem., 2021, 23, 2992–3000
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
1
1
10
oxides, cation- or anion-exchange resins, etc. Although sig-
nificant progress has been made, unsatisfactory activity of
heterogeneous catalysts and low selectivity of homogeneous
catalysts, environmental problems, and a high H
ratio (6–8) are still the disadvantages needed to be overcome.
Thus, the development of an efficient and environmentally
benign catalyst to produce 1,2-diols at a lower H O/epoxide
2
O/EO molar
7
,8
2
molar ratio with both high activity and selectivity is still a huge
challenge.
Scheme 3 2-Carboxylated imidazolium salts with chain length from C1
To combine the advantages of two-step transformation and to C4.
avoid the disadvantages of direct hydration, inspired by the
two-step process wherein CO acts as an intermediate, from
2
the perspective of atom-economy, the synthesis of EG might be cessively investigated. Then, reaction conditions were opti-
simplified as one-step catalytic hydration of EO with a catalyst mized and the suitability of the optimal catalyst for different
2
functioning as CO (Scheme 2). 2-Carboxylated imidazolium epoxides was tested. The catalytic mechanism was finally pro-
salts are very appropriate choices in this catalyst design due to posed based on the XPS analysis and DFT calculation results.
their adjustable cation and anion structures, and the environ- This work not only pioneers the catalytic hydration of epoxides
1
4
ment-friendly properties. To the best of our knowledge, few by using 2-carboxylated imidazolium salts, but also provides a
research studies have been reported associated with 2-carboxy- new idea in turning the multi-step reaction to one step by
lated imidazolium salts as catalysts for one-step hydration of designing a catalyst similar to intermediates.
epoxides.
2
-Carboxylated imidazolium salts, also called N-heterocyclic
carbene CO adducts (NHC–CO ), have gained great attention
2
2
in homogeneous catalysis which are generally synthesized by
Results and discussion
Characterization of 2-carboxylated imidazolium salts
addition of imidazolium chlorides with CO
formation from imidazole and dimethyl carbonates.
particular, the zwitterionic CO
convenient CO carrier to accomplish CO
nucleophilic incorporation of the OvCvO unit as proposed thetic procedures. Both H NMR and C NMR analyses were
for various Lewis base catalyst systems, such as CO cyclo- conducted for all the samples, as listed in the following experi-
addition with epoxides, in which CO was thought to be added mental section. For all the samples, the NMR together with
to the epoxide via nucleophilic attack of the carboxylate group ESI-MS (Fig. S1†) showed consistent results with those
2
or one-step trans-
1
5–17
In
2
adduct could be utilized as a 2-Carboxylated imidazolium salts with chain lengths from C1
fixation through the to C4 were successively prepared according to the typical syn-
2
2
2
1
1
13
2
2
1
8–20
on NHC–CO
used as a ring-opening catalyst for epoxides in a one-step lium salts with C1 to C4 chains were successfully synthesized.
hydration reaction, which has not been reported yet. Elemental analysis results of all the prepared samples (DMIC,
2
.
Given this feature, it is promising to be reported previously, indicating that 2-carboxylated imidazo-
In this work, four kinds of 2-carboxylated imidazolium salts EMIC, PMIC, BMIC) are listed in Table S1.† As shown in
containing only clean element C/N/O/H with different chain Table S1,† the measured C/N molar ratio for these four
lengths were prepared and first used for one-step hydration, samples was respectively 3.01, 3.53, 4.01, and 4.56, which was
namely
1,3-dimethyl-imidazolium-2-carboxylate
(DMIC), well consistent with that calculated from theoretical molecular
1
-ethyl-3-methyl-2-carboxylate (EMIC), 1-propyl-3-methyl-2-car- weight. FT-IR spectra were obtained for confirmation of the
boxylate (PMIC), and 1-butyl-3-methyl-2-carboxylate (BMIC), chemical structure of the as-prepared samples. The spectra of
respectively (Scheme 3). Firstly, the effect of different chain all samples are displayed in Fig. 1. The characteristic peaks of
lengths on the properties and catalytic performance was suc- CvO feature near 1625–1675 cm−
1 22,23
,
and the peak repre-
senting C–C stretch (corresponding to the bond between the
imidazolium carbon and the carboxylate carbon theoretically
to occur around 1300 cm− ) were observed in all the samples,
1
−
which indicates the formation of the carboxylate (O–CvO ).
Thermogravimetric analysis (TGA) gave the decomposition
temperatures of these four salts. All their thermal decompo-
sition temperatures were labelled on the TGA curves in
Fig. S2,† which were 225 °C, 190 °C, 170 °C and 160 °C,
respectively, for DMIC, EMIC, PMIC and BMIC. As shown by
the curves, the decomposition temperature exhibited
a
decreasing trend with the increasing chain length of alkyl sub-
stituents from C1 to C4. DMIC which has the shortest chain
length and the best symmetry of the nitrogen substituents had
Scheme 2 One-step catalytic hydration of epoxides into 1,2-diols by
-carboxylated imidazolium salts functioning as CO
2
2
.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 2992–3000 | 2993

View Article Online
Paper
Green Chemistry
Fig. 1 FT-IR spectra of 2-carboxylated imidazolium salts with chain
lengths from C1 to C4.
the highest thermal stability with the highest decomposition
temperature.
Analysis of active groups
To further figure out main active sites of catalysts, the effect of
different substituent chain lengths on the binding energy was
investigated by XPS analysis. Typically, two extremes with C1
and C4 chain length (DMIC and BMIC) were chosen for O 1s
XPS analysis. As illustrated in Fig. 2, those peaks in O 1s
spectra representing C–O and CvO were found respectively at
Fig. 2 XPS spectra of 2-carboxylated imidazolium salts with chain
lengths of C1 and C4.
5
30.00–530.13 eV and 529.28–529.70 eV. Compared with
DMIC, all the peaks of BMIC (530.00 eV and 529.28 eV) shifted
to higher binding energies (530.13 eV and 529.70 eV),
suggesting the stronger interior interaction of C–O and CvO,
Table 1 Catalyst screening for the EO hydration reaction
a
which might lead to weaker binding interaction with the imi- Entry
Catalyst
Blk-N
Blk-CO2
DMC
N-Methyl imidazole
N-Ethyl imidazole
N-Propyl imidazole
N-Butyl imidazole
EO conv. (%)
EG selec. (%)
2
4
dazolium ring. Such a tendency was consistent with the
thermal stability obtained from TGA. Furthermore, the
strength in the interaction with the imidazole ring reflects the
1
2
3
4
5
6
7
8
9
2
72.2
97.5
79.1
99.1
98.7
99.0
99.6
99.9
99.9
75.2
77.8
85.2
89.1
77.8
83.3
70.9
40.1
33.8
33.7
22.6
69.5
82.2
86.8
87.4
85.0
83.1
b
−
−
ability of the OvC–O (CO
2
) group to react with other mole-
−
cules. Thus, CO
2
might be the main active group.
2 3
Na CO
Catalyst screening
NaHCO
DMIC
EMIC
PMIC
BMIC
3
With EO as a model reactant, the hydration reaction was con- 10
1
1
1
1
2
3
ducted under identical conditions by using the prepared
samples as catalysts. In order to ensure the efficient conver-
sion of EO, it is necessary to keep EO in the liquid phase
system by giving certain external pressure to prevent EO from
Reaction conditions: EO: 1.0 mL, 120 °C, 0.5 h, H
Cat: 2.1 mol% of EO. 2.0 MPa N2. 2.0 MPa CO . All yields and selec-
2
2
O/EO = 5 : 1, and
a
b
gasification in the reaction process. To avoid the influence of tivities were determined by GC.
the atmosphere on the reaction, inert gas N is used for press-
2
urization during the hydration reaction. First of all, all the cat-
alysts were tested under modest conditions (120 °C, 2.0 MPa
N
2
, 0.5 h) with equal molar quantity. The results listed in (entries 10 and 11), which is much higher than that of Blk-N
Table 1 show an increase in EO conversion but a decrease in without any catalyst, together with conventional hydration cat-
EG selectivity as the chain length grows. By comparing the alysts Na CO (entry 8) and NaHCO (entry 9). Furthermore, to
results in Table 1, DMIC and EMIC with a shorter chain are explore the function of DMIC, the Blk-CO with no catalyst
found to have higher selectivity for EG (∼87.0%) than others under a CO atmosphere instead of the inert N was tested. It
2
2
3
3
2
2
2
2994 | Green Chem., 2021, 23, 2992–3000
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
turns out that even 2.0 MPa CO
2
(0.02 mol according to the curves shown in Fig. S3,† 2.0 MPa of N
2
pressure is optimal for
formula “n = PV/RT” which is almost 50 times DMIC) is used, good EO conversion. According to the EO saturated pressure at
and no obvious advantages in improving EG selectivity were different temperatures calculated from the Aspen Plus TM soft-
found compared with DMIC (0.0004 mol). This concluded that ware (Table S1†), lower N
2
pressure than 2.0 MPa is not
DMIC functioned more than CO itself. In addition, other cata- enough to keep EO in the liquid phase, explaining the worse
2
lysts with similar function in ring-opening of epoxides were EO conversion therein. However, higher N
also tested for EO hydration under identical conditions (EO: good for EO transformation which might be due to mass trans-
.0 mL, 120 °C, 0.5 h, H O/EO = 15 : 1, cat: 5.4 mol% of EO). A fer resistance caused by high gas solubility in the liquid phase.
2
pressure is also not
1
2
higher H
2
O/EO ratio of 15 : 1 was used to fully stimulate the To give consideration of EO conversion and EG selectivity,
O/EO ratio of 5 are chosen for further
catalyst performance. As displayed in Fig. 3, these common 30 min, 120 °C, and H
2
catalysts with a ring-opening function, such as 1-ethyl-3- investigations based on the results in Fig. S3.† When the cata-
methylimidazolium bromide (EmimBr) and 1-butyl-3-methyl- lyst quantity reaches over 500 mg, namely 18.0 mol% of EO,
imidazolium bromide (BmimBr), exhibited almost the same both the EO conversion and EG selectivity will be nearly up to
catalytic activity for EO conversion (99.9%) and EG selectivity 100%. Therefore, the optimal conditions for EO hydration are
(
84.8%). Among all the evaluated catalysts, DMIC and EMIC 18.0 mol% DMIC of EO at a H
exhibited the best catalytic activity at a value of 99.9% in both MPa N , 0.5 h.
EO conversion and EG selectivity, higher than conventional The catalytic suitability of various epoxide substrates was
carboxylic imidazolium ionic liquids (EmimAc, BmimAc, and then tested with DMIC as the catalyst under the optimal con-
N(Et) Ac). In comparison with Blk-CO which achieved only ditions. All the catalytic results are listed in Table 2, DMIC
9.1% in EO conversion and 94.1% in EG selectivity, DMIC could catalyze these epoxides with high 1,2-diol selectivity,
and EMIC were further confirmed to be a catalyst functioning most of which reached nearly 100%. For small-molecule reac-
better than only the CO bulk phase. Considering the better tants, such as EO and propylene oxide (PO), high conversion
stability than that of EMIC, DMIC was preferred for this reac- was also obtained, which might be due to the easier mass
tion and used for further investigation. transfer of reactant molecules to active sites. Hence, DMIC
2
O/EO ratio of 5 at 120 °C, 2.0
2
4
2
9
2
With DMIC as the catalyst and EO as the model substrate, could be widely used in the production of many kinds of 1,2-
the reaction conditions including temperature, N2 pressure, diols from various epoxides.
time, H O/EO molar ratio and catalyst amount were optimized
2
successively. All the results are displayed in Fig. S3.† For this
EO hydration reaction, the by-products (DEG, TEG, etc.) were
Confirmation of active sites
generated from the condensation of EO with the glycol To confirm the possible active sites, the density functional
product in the last step. Thus, once the product is generated theory (DFT) calculations were carried out to analyze the
and accumulates to a certain concentration, the next step reac- Natural Population Analysis (NPA) charge and bond energy of
tion is about to start. To obtain both high EO conversion and the structures with different substituent chain lengths (C1 to
EG selectivity, the reaction rate must be controlled in an appro- C4). All calculations were carried out using the Gaussian 09
priate stage by adjusting the reaction conditions. From the program. The B3LYP/6-311++G(d,p) method has been used for
structure optimizations, and subsequent frequency calcu-
lations at the same level verify the optimized structures to be
Table 2 Catalytic suitability test for different epoxides
Time
(h)
EO conv.
(%)
EG selec.
(%)
Entry
Substrate
Product
1
2
0.5
0.5
99.9
94.6
99.9
94.3
3
0.5
41.5
99.9
4
5
0.5
2.0
85.6
57.2
99.9
99.9
6
2.0
63.7
99.9
Fig. 3 Comparison of DMIC and EMIC with common catalysts for ring- Reaction conditions: EO: 1.0 mL, 120 °C, 2 MPa N , H O/EO = 5 : 1,
2
2
opening.
and cat: 18.0 mol% of EO.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 2992–3000 | 2995

View Article Online
Paper
Green Chemistry
Proposed catalytic mechanism
Considering the good activity and selectivity of DMIC, the reac-
tion process was further investigated by using DMIC as the
catalyst. At first, the possible intermediate was detected by
ESI-MS. As shown in the spectra in Fig. S4,† two kinds of inter-
mediates were obtained as I and II. Based on the intermedi-
ates, four possible reaction routes were proposed based on XPS
and DFT results. The routes were respectively displayed as A,
B, C and D in Scheme 4.
Fig. 4 Optimized structure with different chain lengths from C1 to C4.
ground states without imaginary frequencies (NImag = 0)
(
Fig. 4).
The NPA charge distributions by Natural Bond Orbital
NBO) calculations of two O atoms in C–O and CvO on
For Route A and Route C with I as the intermediate, DMIC
(
−
−
2
first reacts with the H in H O to form DMICH with OH as
different structures are depicted in Table 3. A decreasing trend
in bond energy of the imidazole ring and carbonate anion
2
−
anions, and the OH nucleophilically attacks C–O bonds to
induce ring-opening of EO together with a hydrogen-bonding
effect from IMC –COOH of the cation. The whole process is a
2
(
IMC
were examined for NPA charge of O both in C–O and CvO. As
shown in Table 3, the IMC –C bond energy decreases from
91.4 to 189.2 with the increase of the chain length. The NPA
2
−C) was observed from C1 to C4, while fewer changes
strong alkali catalytic process, which is less possible to occur
2
−
because of low stability of strong base OH from weak base
1
−
CO
2
. According to previous research, there exists a reversible
analysis of O shows that with the increase of the chain length,
the charge density increases slightly from −0.701 to −0.703.
The increase of negative charge density of O atoms might con-
tribute to the increase of the binding interaction of CO2−, as
obtained from XPS analysis. Therein, the IMC –C bond was
2
more sensitive to changes of substituent chain lengths than
transformation between CO2 and HO–CO2− (II) in water solu-
−
tion (Route B and Route D). Considering the steric hindrance
−
caused by the imidazole skeleton, HO–CO
2
free from the
cation connection effect was thought to be the main site for
activating reactant molecules, which could attack C–O bonds
to induce ring-opening of EO together with a hydrogen-
others, which might lead to different thermal stabilities and
2
0
−
bonding effect from IMC −H of the cation. Thus, the following
2
even the catalytic activities. Therefore, CO
2
may be the main
discussion is mainly focused on Route B and Route D.
active site in the reaction process. The highest EG selectivity
For Route B and Route D with II as the intermediate, DMIC
for DMIC might be attributed to the strongest IMC –C bond
2
−
−
first interacts with H
2
O to form DMIHC with HO–CO
2
as the
energy, wherein the active site CO
2
is strongly bound with the
2
5,26
and the HO–CO − nucleophilically attacks C–O
anion,
bonds to induce ring-opening of EO and meanwhile the IMC
2
imidazole ring, slowing down the rate of the whole reaction,
especially the next side reaction. The lower bond energy of
2
–
H from the cation could synergistically stabilize the transition
state via the hydrogen-bonding effect. The whole process is a
weak alkali catalytic process. Considering the nearly neutral
reaction system, Route B and Route D are more reasonable.
2
IMC –C from C1 to C4 leads to easier activation of its adjacent
−
CO2 , hence higher EO conversion was obtained and continu-
ous side reaction was ineluctable, explaining the lower EG
selectivity from C1 to C4.
Table 3 NPA charge distributions and the bond energy of C–C bonds (Kcal mol− ) for different structures
1
Parameters
C1
C2
C3
C4
NPA charge of O in C–O and CvO
IMC2−C bond energy
−0.701/−0.701
191.4
−0.701/−0.702
190.1
−0.702/−0.703
189.5
−0.702/−0.703
189.2
Scheme 4 Possible routes for EO activation by DMIC.
2996 | Green Chem., 2021, 23, 2992–3000
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
Fig. 5 Potential energy surface profiles of the DMIC-catalyzed process according to Route B (Mechanism B) and Route D (Mechanism D), and the
optimized geometries for the intermediates and transition states.
2
To compare the relative energy of Route B and Route D and TS-AB and reactant H O could form a relatively stable inter-
screen the most reasonable route, the quantum chemistry cal- mediate which would not easily be broken for further side
culations were conducted. The reactants, products, intermedi- reaction. This might explain the higher EG selectivity catalyzed
ates, and transition state equilibrium geometries were opti- by DMIC than NaHCO
3
, which attributes to the DMIC cation
mized using the B3LYP method and 6-31+G(d,p) basis set inhibition. Furthermore, there exists a balance between DMIC
−
(
2
B3LYP/6-31+G(d,p)). The vibrational frequency calculations and DMIHC with HO–CO to be main active sites for EO acti-
were carried out at the same level to confirm the nature of the vation. The difficulty in shifting the balance to DMIHC slows
extreme points of the potential energy surface. Vibrational fre- down the speed for the reaction. Compared with other 2-car-
quency with only one imaginary frequency stands for the tran- boxylated imidazolium salts with longer chain lengths, DMIC
sition state. Then, the intrinsic reaction coordinate theory was has higher interior binding energy which is more difficult in
employed to construct the minimum-energy path (MEP). All bonding with H O molecules, therein the dragging effect inhi-
2
the above calculations were performed using the Gaussian 09 bits the side reaction and leads to higher EG selectivity.
program.
To further understand the side reaction process for two
The mechanisms of the DMIC-catalyzed process according different mechanisms, the charges of different transition
to Route B (Mechanism B) and Route D (Mechanism D) were states, I in Mechanism B and D in Mechanism D, were
respectively analyzed by calculation (Fig. 5). The main differ- obtained and analyzed. As shown in Fig. 6, for I, the H of the
ence between Mechanism B and Mechanism D is the source of H O molecule swayed between itself and EG, causing a rela-
2
the hydrogen donor. For Mechanism B, H
2
O acts as a hydro- tively high charge (−0.930) of the O compared to the charge
gen donor while for D it is the IMC –H. From the calculation (−0.823) of O in D. The exposed O of EG tended to attack the C
2
results in Fig. 5, it is found that different donors cause atom of another EG, leading to a side reaction and formed the
different relative energies for transition states. For Mechanism byproduct condensed glycol. Thus, considering the cation
B with H
2
O as the donor, the relative energy of transition state effect in catalytic performance, the Mechanism D is more
−
1
TS-FG is 17.7 kcal mol , and for Mechanism D with IMC –H reasonable to explain the DMIC-catalyzed process.
2
as the donor, the relative energy of TS-AB is 29.8 kcal mol−
1
.
Above all, with EO hydration taken as an example, the poss-
Thus, the initially activated step of Mechanism B occurred ible reaction mechanism for the DMIC-catalyzed process is
−
more easily than Mechanism D. Meanwhile, a more stable depicted as follows (Scheme 5). The CO in DMIC first trans-
2
−
ending state J was obtained according to Mechanism B forms into HO–CO
2
in the aqueous phase. With the synergis-
−
1
(
(
−30.4 kcal mol ) rather than that E in Mechanism D tic effect of hydrogen bonding from IMC
2
–H and nucleophilic
−26.5 kcal mol− ). Therefore, the main reaction step from EO
1
to EG during the hydration process without the DMIC cation
effect occurs more easily than that with the DMIC cation
effect. This tendency is in accordance with the above catalytic
test results where NaHCO
than DMIC, which indicates that HO–CO
3
exhibited higher EO conversion
−
2
functions more for
the main reaction. However, the faster reaction rate leads to
uncontrollable EG selectivity because of the quick side reac-
−
tion in the presence of HO–CO
2
. For DMIC, due to the steric
Fig. 6 The charge distribution of different transition states I and D in
interaction of imidazolium cations, the stable transition state Mechanism B and Mechanism D.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 2992–3000 | 2997

View Article Online
Paper
Green Chemistry
increases initially and then remains constant, together with
ESI-MS (Fig. S4†) this explains the transformation process
−
−
from CO
2 2
to HO–CO .
Conclusions
Four kinds of 2-carboxylated imidazolium salts with different
substituent chain lengths (C1-C4) were prepared. 1,3-
Dimethylimidazolium-2-carbonate (DMIC) with the shortest
substituent chain length was found to be the most efficient
catalyst both in activity and selectivity. Both XPS analysis and
DFT calculations proved that CO2− is the active site for EO acti-
vation and higher binding energy of its adjacent C–C bond of
DMIC contributes to higher thermal stability and catalytic
Scheme 5 Proposed mechanism for EO hydration catalyzed by DMIC.
selectivity for EG. The catalytic mechanism was finally pro-
−
posed that CO
2
functions as the nucleophilic agent by reversi-
action from the HO–CO
2
−, the EO ring could be opened bly forming HO–CO2− with water, which induced ring-opening
(TS-AB). The strong binding effect of interior DMIC causes the of EO synergistically with a hydrogen-bonding effect from
slower ion transforming step, which slows down the sub- IMC –H. The high EG selectivity resulted from the huge steric
2
−
sequent HO–CO
2
formation, C–O activation and hydration, dynamic structure and the stabilizing structure of imidazo-
which might exist temporarily as an interacted steric dynamic lium-EG-HO–CO2−, both of which avoid excessive hydration
−
structure. Meanwhile, the IMC –H and HO–CO could stabil- into byproducts. This work provides a new approach for EG
2
2
ize the formed EG as a stabilizing structure of imidazolium- selectivity improvement, which might be a good reference for
−
EG-HO–CO
2
by a hydrogen-bonding effect (TS-CD), which effective product yield in the reaction involving intermediates.
also prevented the pre-formed EG molecule from contacting
EO for excessive hydration into byproduct, so that the formed
intermediate could go through an intramolecular conden-
Experimental section
Preparation of 2-carboxylated imidazolium salts
−
sation to produce EG and in the meantime released HO–CO
.
2
−
The HO–CO
2
then dehydrated to restore DMIC.
The whole reaction process was further confirmed by in situ First of all, 5 g of N-methylimidazole and certain amounts of
−
1
FT-IR (500–3000 cm ), with propylene oxide (PO) as the dimethyl carbonates (1.7 molar ratios of N-methylimidazole)
model reactant. As shown in Fig. 7, the characteristic peaks of were successively added to a 50 mL stainless-steel reactor. The
C–O, –OH, CvC, CvN and CvO were detected. The C–O peak reactor was sealed and stirred at 120 °C for 24 h. Thereafter,
and –OH peak might be respectively attributed to reactant PO the reactor was cooled to room temperature and the mixture in
and product propylene glycol (PG), which respectively showed the reactor was filtered. The filter cake was transferred into a
a decreasing trend and increasing trend as the reaction pro- beaker with methanol solution in the refrigerator below 0 °C.
ceeded. The CvC and CvN peaks might be attributed to DMIC was obtained after drying the recrystallized white solid
imidazole of the catalyst DMIC, of which the peak intensity at 50 °C in the vacuum oven overnight. According to the above
remains almost constant in the whole process. The CvO peak similar steps, a series of 2-carboxylated imidazolium salts with
might be attributed to the CO2 of DMIC and the HO–CO2− of different chain lengths (DMIC, EMIC, PMIC, BMIC) were
the catalyst intermediate DMIHC, of which the peak intensity obtained (see ESI-MS in Fig. S1†).
−
1,3-Dimethylimidazolium-2-carboxylate (DMIC)
Synthesized by the reaction of 1-methylimidazole and dimethyl
1
carbonate. H NMR (600 MHz, D O) δ (ppm): 3.92 (s, 6H, N–
2
3
1
CH
3
), 7.63 (s, 2H, imidazole). C NMR (600 MHz, D
2
O) δ
(
ppm): 35.68 (d, N-CH
3
), 123.23 (dm, C4 and C5), 138.14 (s,
−
C2), 163.42 (s, COO ).
1-Ethyl-3-methylimidazolium-2-carboxylate (EMIC)
Synthesized by the reaction of 1-ethylimidazole and dimethyl
1
carbonate. H NMR (600 MHz, D O) δ (ppm): 1.38 (t, 3H, CH ),
2
3
3
.775 (s, 3H, N–CH
3
), 4.110 (m, 2H, N–CH
2
), 7.30 (d, 1H, imid-
1
3
azole), 7.37 (d, 1H, imidazole). C NMR (600 MHz, D
2
O) δ
Fig. 7 In situ FT-IR spectra for PO hydration catalyzed by DMIC.
(ppm): 14.24, 35.61, 44.91, 124.78, 126.53, 134.26, 163.57.
2998 | Green Chem., 2021, 23, 2992–3000
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
1
-Propyl-3-methylimidazolium-2-carboxylate (PMIC)
Conflicts of interest
Synthesized by the reaction of 1-propylimidazole and dimethyl
carbonate. H NMR (600 MHz, D
1
There are no conflicts to declare.
2
O) δ (ppm): 0.82 (t, 3H, CH
3
),
1
CH
.80 (m, 2H, N–CH ), 3.805 (s, 3H, N–CH ), 4.06 (t, 2H, N–
2 3
−
13
2
), 7.34 (d, 1H, imidazole), 7.38 (d, 1H, imidazole).
C
NMR (600 MHz, D
2
O) δ (ppm): 9.76, 22.78, 35.52, 51.01, Acknowledgements
1
22.11, 123.40, 135.60, 163.50.
This work was supported by Innovation Academy for Green
Manufacture, Chinese Academy of Sciences (no. IAGM-2019-
A01), the Transformational Technologies for Clean Energy and
Demonstration, Strategic Priority Research Program of the
Chinese Academy of Sciences, Grant No. XDA 21030500, and
the National Natural Science Foundation of China (no.
1
-Butyl-3-methylimidazolium-2-carboxylate (BMIC)
Synthesized by the reaction of 1-butylimidazole and dimethyl
carbonate. H NMR (600 MHz, D O) δ (ppm): 0.82 (m, 3H,
2
CH ), 1.20 (m, 2H, CH ), 1.72 (m, 2H, CH ), 3.88 (s, 3H, N–
CH
imidazole). C NMR (600 MHz, D O) δ (ppm): 12.60 (s, C–
2
CH ), 18.76 (s, CH ), 31.83 (s, CH ) 36.39 (s, N–CH ), 49.30 (s,
1
3
2
2
21890763, no. 21908226, and no. 22078329). The authors are
3
), 4.31 (t, 2H, N–CH
2
), 7.30 (d, 1H, imidazole), 7.35 (d, 1H,
grateful for the assistance from teacher Ling Wang of Analysis
and Test Center, Institute of Process Engineering, Chinese
Academy of Sciences in ESI-MS (electrospray ionization mass
spectrometry).
1
3
3
2
2
3
−
2
N–CH ), 121.71 (s, C4), 123.07 (s, C5), 139.96 (s, C2), 158.48
(
s, COO−).
Characterization of all 2-carboxylated imidazolium salts
Notes and references
The chemical structures of the materials were analyzed by
Fourier transform infrared (FTIR) spectroscopy (Thermo
Nicolet 380) with anhydrous KBr as the standard. Elemental
analysis (C/H/N) was conducted on a Vario EL Cube elemental
analyzer. The analysis of active groups was performed on an
1
2
3
4
5
6
B. Li, S. Bai, X. Wang, M. Zhong, Q. Yang and C. Li, Angew.
Chem., Int. Ed., 2012, 51, 11517–11521.
J. Sun, X. Yao, W. Cheng and S. Zhang, Green Chem., 2014,
16, 3297–3304.
B. Xu, J. Wang, J. Sun, Y. Huang, J. Zhang, X. Zhang and
S. Zhang, Green Chem., 2015, 17, 108–122.
W. Cheng, Q. Su, J. Wang, J. Sun and F. T. T. Ng, Catalysts,
1
3
1
X-ray photoelectron spectroscope (ESCALAB250Xi). C and H
NMR spectra were recorded on a Bruker Advance III HD 600 M
NMR instrument. TG-DTA measurements were conducted on a
DTG-60H thermal analyzer (Shimadzu, Japan) at a heating rate
2013, 3, 878–901.
L. Deng and H. Kvamsdal, Green Energy Environ., 2016, 1,
179.
Y. Wu, Y. Ding, J. Xu, Y. Wang, K. Mumford, G. W. Stevens
and W. Fei, Green Energy Environ., 2020, DOI: 10.1016/j.
gee.2020.04.016.
Z. Dai, Q. Sun, F. Chen, S. Pan, L. Wang, X. Meng, J. Li and
F. S. Xiao, ChemCatChem, 2016, 8, 812–817.
M. Zhong, Y. Zhao, Q. Yang and C. Li, J. Catal., 2016, 338,
of 10 °C min− under a N atmosphere using an empty cruci-
1
2
ble as the reference. The molecular weight of the synthesized
samples was verified using electrospray ionization mass spec-
trometry (ESI-MS, micrOTOF-Q II, Bruker, Germany). All reac-
tant conversion and product selectivity were recorded and ana-
lyzed via a GC (Agilent 6820) with a TCD detector. The compo-
sition of the products was identified by using GC–MS
7
8
9
(GCMS-QP2020).
184–191.
W. Dai, C. M. Wang, B. Tang, G. Wu, N. Guan, Z. Xie,
M. Hunger and L. Li, ACS Catal., 2016, 6, 2955–2964.
0 M. R. Altiokka and S. Akyalçin, Ind. Eng. Chem. Res., 2009,
Catalytic test of 2-carboxylated imidazolium salts
1
With EO as the model reactant, the hydration reaction was
48, 10840–10844.
conducted under identical conditions using 2-carboxylated 11 Z. J. Yang, Y. F. Li, Q. B. Wu, N. Ren, Y. H. Zhang, Z. P. Liu
imidazolium salts as catalysts. The hydration reaction of EO and Y. Tang, J. Catal., 2011, 280, 247–254.
and H O was carried out in a 25 ml stainless steel reactor 12 Z. J. Yang, N. Ren, Y. H. Zhang and Y. Tang, Catal.
equipped with a magnetic stirrer and an automatic tempera- Commun., 2010, 11, 447–450.
ture control system. First the reactor was filled with 1 mL of 13 J. W. V. Hal, J. S. Ledford and X. Zhang, Catal. Today, 2007,
EO and certain amounts of catalysts at room temperature, and 123, 310–315.
then charged with N to target pressure from a high-pressure 14 S. Zhang, J. Sun, X. Zhang, J. Xin, Q. Miao and J. Wang,
2
2
reservoir tank before being heated. After that, the reactor was
heated to the target temperature and maintained at this con- 15 H. Zhou, G. Wang, W. Zhang and X. Lu, ACS Catal., 2015, 5,
stant pressure for a desired period. When the reaction was 6773–6779.
over, the reactor was cooled down to room temperature and 16 J. A. Stewart, R. Drexel, B. Arstad, E. Reubsaet,
Chem. Soc. Rev., 2014, 43, 7838–7869.
the N
by GC-MS and GC.
2
was released slowly. The organic products were analyzed
B. M. Weckhuysen and P. C. A. Bruijnincx, Green Chem.,
2016, 18, 1605–1618.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 2992–3000 | 2999

View Article Online
Paper
Green Chemistry
1
7 L. F. B. Wilm, T. Eder, C. Mück-Lichtenfeld, P. Mehlmann, 22 A. M. Voutchkova, L. N. Appelhans, A. R. Chianese and
M. Wünsche, F. Buß and F. Dielmann, Green Chem., 2019,
1, 640–648.
R. H. Crabtree, J. Am. Chem. Soc., 2005, 127, 17624–
17625.
2
1
1
2
2
8 J. Wang, J. Sun, W. Cheng, C. Shi, K. Dong, X. Zhang and 23 L. H. Finger, J. Guschlbauer, K. Harms and J. Sundermeyer,
S. Zhang, Catal. Sci. Technol., 2012, 2, 600–605. Chem. – Eur. J., 2016, 22, 16292–16303.
9 J. Wang, J. Sun, C. Shi, W. Cheng, X. Zhang and S. Zhang, 24 M. J. Ajitha and C. H. Suresh, J. Org. Chem., 2012, 77, 1087–
Green Chem., 2011, 13, 3213–3217. 1094.
0 H. Zhou, W. Zhang, C. Liu, J. Qu and X. Lu, J. Org. Chem., 25 N. J. Bridges, C. C. Hines, M. Smiglak and R. D. Rogers,
008, 73, 8039–8044. Chem. – Eur. J., 2007, 13, 5207–5212.
1 J. Li, J. Peng, B. Ying, G. Lai and X. Li, J. Organomet. Chem., 26 C. Rijksen and R. D. Rogers, J. Org. Chem., 2008, 73, 5582–
011, 696, 2116–2121. 5584.
2
2
3000 | Green Chem., 2021, 23, 2992–3000
This journal is © The Royal Society of Chemistry 2021