Poly(4-vinylimidazolium)s/Diazabicyclo[5.4.0]undec-7-ene/Zinc(II) Bromide-Catalyzed Cycloaddition of Carbon Dioxide to Epoxides

Poly(4 ‐ vinylimidazolium)s, derived from the self ‐ immobilization of 4 ‐ vinylimidazoliums, with diazabicyclo[5.4.0]undec ‐ 7 ‐ ene (DBU) and zinc bromide (ZnBr2) are used as a highly efficient catalyst for the chemical fixation of carbon dioxide. This catalytic system has been applied for the preparation of cyclic carbonates from terminal epoxides and carbon dioxide. Many functional groups, including chloro, vinyl, ether, and hydroxy groups are well tolerated in the reactions. Moreover, the catalytic system was found to catalyze the conversion of more sterically congested epoxides which are generally considered to be challenging substrates for fabricating the cyclic organic carbonates. In addition, the disubstituted epoxides are found to react with retention of configuration. The polymer precatalyst is easily recovered and reused. A plausible reaction mechanism is proposed.

carbon dioxide fixation; cyclic carbonates; cycloaddition; organocatalysis; polymers

Carbon dioxide (CO2) has been attracting much attention as it is considered to a major cause of climate change, because of its greenhouse effect.1 To reduce the continuous accumulation of CO2 in the atmosphere, the scientific and industrial initiatives have, in the recent decades, been focused on the chemical conversion of CO2 to useful chemicals.2 However, the chemistry of CO2 is limited because of its high stability. One of the potentially useful ways of reducing CO2 is its reaction with epoxides to form cyclic carbonates.3 This reaction has tremendous potential because cyclic carbonates are widely used as the precursors of polycarbonates and other polymers, serve as excellent aprotic polar solvents, electrolytes in rechargeable batteries, and intermediates in the production of pharmaceuticals and fine chemicals.4 Therefore, a variety of diverse homogeneous and heterogeneous catalyst systems5 has been developed for catalyzing the reactions of CO2 with epoxides. However, most of them suffer from low catalytic activity, water ‐ or air ‐ sensitivity of catalyst, and harsh reaction conditions.

Organocatalysts are usually robust, inexpensive, readily available, and non ‐ toxic. Most of them are also inert towards moisture and oxygen. Although a catalyst is not consumed during the process, its separation from the final products may sometimes be difficult. Therefore, the recovery and reuse of the catalyst molecules can be a scientific challenge due to economic and environmental relevance. Recently, the polymerization of catalysts has been developed as one of the heterogenization methods.6 In catalyst polymerization, the active centres of a catalyst are distributed over the polymer chain, which helps to prevent the loss of activity; therefore, polymerized catalysts show high activities in certain reactions. The structure of 4 ‐ vinylimidazolium is similar to that of styrene, and its polymer product, poly(4 ‐ vinylimidazolium)s, seems to be similar to that of polystyrenes. However, reports on the synthesis and use of poly(4 ‐ vinylimidazolium)s are rare.7 Very recently, we found that poly(4 ‐ vinylimidazolium)s derived from the self ‐ immobilization of 4 ‐ vinylimidazoliums acted as a precatalyst in the carboxylation of epoxides with CO2. Herein we communicate a very active, stable, recyclable and cost ‐ effective organic precatalyst, poly(4 ‐ vinylimidazolium)s, for the carboxylation of epoxides with CO2. The use of 1 ‐ vinyl ‐ and 1,3 ‐ divinylimidazole ‐ based cross ‐ linked polymers as the catalyst in the carboxylation of epoxides with CO2 has been reported.8 Moreover, the use of poly[1,3 ‐ bis(4 ‐ vinylbenzyl)imidazolium]Cl ‐ based cross ‐ linked polymers as the catalyst in carboxylation of epoxides with CO2 has also been reported.9 However, all the reactions were carried out at high temperatures and high pressures of CO2.

Previously, the 4 ‐ vinylimidazolium salt was produced by the decarboxylation of urocanic acid.7 However, histidine can also be used to obtain 4 ‐ vinylimidazolium as shown in Scheme 1.

Both histidine and histamine are commercially available. We chose histamine as the starting material. The reaction of histamine with NaNO2, followed by chlorination reaction with SOCl2, afforded 4 ‐ (2 ‐ chloroethyl)imidazole in 74% yield.10 Dehydrochlorination and dimethylation of 4 ‐ (2 ‐ chloroethyl)imidazole in the presence of NaHCO3, K2CO3 and MeI in acetonitrile afforded an 4 ‐ vinylimidazolium salt, 1, in 70% yield.11 Polymerization of 1 in the presence of AIBN afforded poly(4 ‐ vinylimidazolium)s, 2, the precursor of an organocatalyst, in high yield.12 Broad 1H NMR signals were observed, as expected for a high molecular weight polymer. The polymer was soluble in DMSO and DMF, and freely soluble in water, but insoluble in chloroform, dichloromethane, and acetone. The weight average molecular weight (MW) of 2 was found to be ∼30000 from light scattering experiments.13 The polymer decomposed at 350 °C. The abilities of 1 and 2 to catalyze the carboxylation of epoxides with CO2 in the presence of a base were examined [Eq. (1)].

First, the reaction conditions were screened, including the CO2 pressure, reaction temperature, reaction time, solvent, base, and the catalyst amount to optimize the yield of the cyclic carbonate (Table 1). Initially, the study was performed under the reaction conditions (2 mol% catalyst, ZnI2 as the Lewis acid, K2CO3 as the base, DMSO as the solvent, at 80 °C reaction temperature, and 24 h of reaction time) adopted from the previous work14 on the NHC ‐ catalyzed carboxylation of epoxides. The coupling of styrene oxide (SO) with CO2 to afford styrene carbonate (SC) was chosen as the model reaction. When the reaction was carried out under the adopted reaction conditions, the SC yield was 86% (Table 1, entry 1). Changing the solvent to DMF afforded a slightly higher SC yield (entry 2: 88% yield). When diazabicyclo[5.4.0]undec ‐ 7 ‐ ene (DBU) was used instead of K2CO3 in DMF, SC was obtained quantitatively (entries 3 and 4). When the amount of catalyst 2 was reduced to 1 mol%, the yield was still high (entry 5, 90% yield).

1 Screening for the reaction conditions. [a]

1 [a]Reaction conditions: Styrene oxide (SO) (5 mmol), 2, Lewis acid (the same equiv. of 2), solvent (3 mL), base, and CO2 (1 atm) at 80 °C for 24 h.

• 2 [b] Isolated yield.
• 3 [c] Reaction was performed at 60 °C.
• 4 [d] Reaction was performed at 100 °C.
• 5 [e] Reaction time: 10 h.

Thus, in the presence of 1 mol% of 2, the effect of metal ions, such as ZnX2 (X=Cl, Br, and I), CrCl3, LiI, and FeCl3, on the reaction was also examined (entries 5 – 10). The SC yield was greatly affected by the different metal ions. Among them, the use of ZnBr2 afforded the best result (entry 6, 94% yield), because of its strong Lewis acidity. The reactivity of different anions in the zinc salts decreased in the following order: Br−>I−>Cl−.15 The different reactivity may be due to the different electrophilicity of Zn depending upon the anion.13 The effect of the reaction temperature on the product yield was also examined (entries 6, 11, and 12). In the lower temperature range, the yield increased with the increasing temperature (entry 6 vs 11); however, this trend was not observed at higher temperatures. The yield decreased to 87% at 100 °C (entry 6 vs. 12). The reaction proceeded well with 1 mol% of catalyst at 80 °C. When the reaction time was reduced to 10 h, the excellent yield was maintained (entry 13, 94% yield). The necessity of a combination of 2, DBU, and a Lewis acid in the reaction was confirmed by the observation that in the presence of solely 2, DBU, or ZnBr2 a negligible reaction was observed (entries 14 – 16).

Thus, the optimum reaction conditions were established as follows: 1 mol% catalyst, 2 mol% DBU, and 1 mol% ZnBr2 in DMF at 80 °C under 1 atm CO2 for 10 h. Under the optimized reaction conditions, the maximum turnover number was 320 (see the Supporting Information). When 1 was used as the precatalyst under the optimized reaction conditions, SC was obtained in 88% yield. The experiments were also carried out to examine the recyclability of the catalyst using SO as the substrate. The amount of 2 was too small to confirm the recycling test; therefore, twice the amounts of 2, ZnBr2, and DBU as those used during optimization were used. After performing the reaction in DMF, excess MeOH was added to the reaction mixture to precipitate the polymer catalyst. The recovered polymer catalyst was dried and reused for the next run. The SC yields for the eight consecutive runs were 99%, 98%, 91%, 97%, 97%, 94%, 94%, and 91%. Considering the weight loss in the catalyst purification process, no considerable decrease in the SC yield was observed (Figure 1).

To show the catalytic activity of 3 [3=poly(NHC)s; NHC=1,3 ‐ dimethyl ‐ 4 ‐ vinylimidazol ‐ 2 ‐ ylidene], the coupling reactions of CO2 to various substituted epoxides were also conducted at 80 °C and 1 atm CO2 (Table 2). To our delight, many functional groups, including chloro, vinyl, ether, and hydroxy groups were well tolerated in the reactions. Thus, all the epoxides used except propylene oxide (entry 6) could be transformed to the corresponding carbonates in almost quantitative yields. In the case of propylene oxide, a slightly lower yield (56%) was observed because of its low boiling point (33 °C).

2 Cycloaddition of CO 2 with various terminal epoxides. [a]

• 6 [a] 1 mol% of 2, epoxide (5 mmol), ZnBr2 (1 mol%), DBU (2 mol%), CO2 (1 atm), 80 °C, 10 h.
• 7 [b] Isolated yields.

Notably, most of the substrates afforded the desired products under mild reaction conditions, and the functional groups were stable in the reactions, indicating the outstanding efficiency of the catalyst. This catalytic reaction may be a good method for preparing functional monomers.

Owing to the steric hindrance and electronic effect, 1,1 ‐ disubstituted and internal disubstituted epoxides are often considered to be more challenging substrates for fabricating the cyclic organic carbonates.16 A series of these substrates (Table 3) was investigated using the catalyst system. The amounts of 2, ZnBr2, and DBU were used four times more than those used for the cycloaddition of CO2 with terminal epoxides (see the Supporting Information). As expected, higher pressures and temperatures were needed to convert the substrates to the corresponding organic carbonates; e.g., the cycloaddition of 1,1 ‐ dimethyloxirane with CO2 at 90 °C and 1 atm of CO2 was unsuccessful. However, under 5 atm of CO2 pressure, 1,1 ‐ dimethyloxirane was converted into 4,4 ‐ dimethyl ‐ 1,3 ‐ dioxolan ‐ 2 ‐ one in 46% yield after 24 h of reaction time. When the CO2 pressure was increased to 7 atm and 10 atm, the yields of 4,4 ‐ dimethyl ‐ 1,3 ‐ dioxolan ‐ 2 ‐ one increased to 53% and 83% yields, respectively. Thus, the reactions of other substrates with CO2 were carried out at 90 °C under 10 atm of CO2. When the CO2 pressure was increased, most of the substrates afforded the corresponding cyclic carbonates in 68 – 87% yields. In the case of stilbene oxide (entry 5), a relatively low yield (35%) was observed after 24 h of reaction time. The yield increased to 55% when the reaction time was increased to 48 h. Moreover, ethyl 2 ‐ oxo ‐ 5 ‐ phenyl ‐ 1,3 ‐ dioxalane ‐ 4 ‐ carboxylate (entry 6) was found to be a good substrate.17

3 Cycloaddition of CO 2 with various internal epoxides. [a]

• 8 [a] The catalyst (4 mol%), ZnBr2 (4 mol%), DBU (8 mol%), expoxide (5 mmol), DMF (4 mL), CO2 (10 atm), 90 °C, 24 h.
• 9 [b] Isolated yields.
• 10 [c] 48 h

However, in the case of 2,2,3,3 ‐ tetramethyloxirane, no reaction was observed under the reaction conditions (90 °C, 10 atm CO2, 24 h). When the reaction time was prolonged to 72 h, no reaction was observed.

The internal epoxides provided an opportunity to study the stereochemistry of this cyclic carbonate synthesis. According to their 1H and 13C NMR spectra, the cyclic carbonates, (±) ‐ trans ‐ 4,5 ‐ dimethyl ‐ 1,3 ‐ dioxolan ‐ 2 ‐ one and (±) ‐ trans ‐ 4,5 ‐ diphenyl ‐ 1,3 ‐ dioxolan ‐ 2 ‐ one, had the same stereochemistry as that of the starting epoxides. Thus, the disubstituted epoxides were found to react with retention of configuration.

To better understand the reaction mechanism, various reaction conditions were tested (Table 4). The reaction of SO in the presence of ZnBr2 alone afforded SC in 11% yield (entry 1). The same reaction in the presence of ZnBr2 and DBU18 afforded SC in 43% isolated yield (entry 2). We expected that the reaction of 2 with DBU would afford 3; however, SC was obtained in 8% yield (entry 3). This observation suggested that 3 would not be generated in situ. The rapid CO2 fixation by DBU and the coupling of aziridine to CO2 in the presence of DBU have been well documented.19 However, DBU by itself did not catalyze the reaction (entry 4). The use of imidazolium ‐ based polymeric ionic liquids as catalyst in the reaction of the coupling of CO2 with epoxides has been reported.9,20 Thus, the reaction of SO with CO2 in the presence of 2 and ZnBr2 was examined; and SC was obtained in 45% yield (entry 5). The NHC ‐ CO2 adduct has been reported to be a potent organocatalyst at high temperatures and high pressures (100 – 120 °C and 20 – 100 atm of CO2 and).9,20a,21 Thus, the 3 ‐ CO2 adduct was prepared (see the Supporting Information) and used as the catalyst under the above ‐ mentioned reaction conditions. However, no reaction was observed in the presence of 3 ‐ CO2 adduct (entry 6). Furthermore, the IR spectrum of the recovered poly(NHC) ‐ CO2 adduct showed that almost all of the CO2 was lost during the reaction. Thus, the poly(NHC) ‐ CO2 adduct itself did not play an important role in the catalytic reaction. Interestingly, when the same reaction was carried out in the presence of 3 ‐ CO2 adduct and ZnBr2 (entry 7), SC was obtained in 33% yield. Thus, the presence of both ZnBr2 and DBU is necessary for achieving good results similar to those obtained with other Lewis acid and Lewis base co ‐ catalyzed coupling reactions of CO2 with epoxides.3,22 Furthermore, the best result was obtained when the reaction was carried out in the presence of 2, DBU, and ZnBr2 (2 ‐ DBU ‐ ZnBr2) catalyst system (entry 8, 94% yield). The 1H NMR spectra of 2, 2/DBU, and 2/DBU/ZnBr2 in [D7]DMF were taken at room temperature and 80 °C (see the Supporting Information), the imidazolium CH peak of 2 appeared at δ=9.15 ppm and was observed in the presence of DBU even at 80 °C. However, in the presence of DBU and ZnBr2, it disappeared at 80 °C and reappeared at room temperature. Thus, we envision the formation of 3 ‐ ZrBr2 at 80 °C.

4 Cycloaddition under various reaction conditions. [a]

11 [a] Styrene oxide (SO) (5 mmol), DMF (3 mL), CO2 (1 atm), 80 °C, 10 h.

Although the exact mechanism of this transformation is not clear at the moment, a plausible reaction mechanism was proposed on the basis of the experimental observations (Figure 2).

We envision that the NHC ‐ ZnBr2 species may play a major role in the coupling reaction at 80 °C, i.e., the reaction may be co ‐ catalyzed by a Lewis base, DBU, and Lewis acid, NHC ‐ ZnBr2 or ZnBr2. The Lewis base and Lewis acid act together to open the epoxy ring and then react with CO2 to afford the corresponding cyclic carbonates via a ring ‐ opening and recyclization process. When a pre ‐ synthesized polymer ‐ NHC ‐ ZnBr2 complex was used as a catalyst, SC was obtained in 93% yield. This observation also supports the mechanism shown in Figure 2.

In conclusion, we have examined an efficient poly(4 ‐ vinylimidazolium)s ‐ DBU ‐ ZnBr2 catalyst system for the synthesis of cyclic carbonates by reacting terminal epoxides and internal epoxides with CO2. Cyclic carbonate is the sole product in this reaction. Efforts are underway to elucidate the mechanistic details of the reaction and extend the applications of the catalyst system.

Reactions were performed in a Schlenk tube equipped with a stirring bar and capped with a rubber cap and the following were placed in the tube in order: 1 mol% of catalyst (13 mg, 0.05 mmol), 1 mol% of ZnBr2 (12 mg, 0.05 mmol), 2 mol% of DBU (15 µL, 0.1 mmol) and 1 mL of DMF. while they were mixing together, tube was charged with CO2 by balloon for 30 seconds. Then, mono ‐ substituted epoxide (5 mmol) and 2 mL of DMF were put into the Schlenk tube. The mixture was stirred at 80 °C for 10 h and CO2 was provided by a balloon (1 atm). The reaction mixture was taken up in methanol and catalysts were filtered and filtrate was concentrated under reduced pressure. Purification by flash chromatography on silica gel with n ‐ hexane and ethyl acetate afford the cyclic carbonates.

This work was supported by the National Research Foundation of Korea (NRF) (2007 ‐ 0093864) and the Basic Science Research Program through the NRF funded by the Ministry of Education, Science and Technology (R11 ‐ 2005 ‐ 065). URS thanks the Brain Korea 21 Plus Fellowships.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re ‐ organized for online delivery, but are not copy ‐ edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Graph: 1 Synthesis of 2 from histamine.

Graph: 1 Recycling of poly(NHC ‐ Zn) complex in the cycloaddition of CO2 to SC.

Graph: 2 Plausible mechanism of cycloaddition of epoxide to CO2

Graph: miscellaneous_information