Global warming is a problem which is increasing year by year, largely caused by the emissions of greenhouse gases such as carbon dioxide (CO2), methane, nitrous oxide, chlorofluorocarbons and so on.[1-4] Among the greenhouse gases, CO2, produced by combustion of petroleum, coal and natural gas, is the main cause of global warning because of the large amounts emitted in the atmosphere and its greenhouse nature. Therefore reduction of the concentration of CO2 in the atmosphere has been an urgent issue all over the world. To overcome the problem, there are four current approaches:[6, 7] (1) reduction of energy consumption by improving efficiency; (2) replacement of fossil fuels with renewable energy sources; (3) CO2 capture/storage (CCS); and (4) CO2 capture and utilization (CCU). At present, the CCS method may be the most efficient process to reduce CO2 in air. Among these approaches, the CCU approach is promising in terms of not only consumption of CO2 but also production of value-added chemicals since CO2 is an abundant, inexpensive, nontoxic and nonflammable C1 building block in organic synthesis.[8, 9] Nevertheless, CO2 is quite stable because CO2 contains carbon in its highest oxidation state and it requires the input of a large amount of energy to use as a C1 feedstock. Two different strategies are known for the conversion of CO2 to useful chemicals.[10-12] One is reductive CO2 conversion and the other is non-reductive CO2 conversion. The reductive conversion of CO2 to target compounds such as formic acid (HCOOH)[13-15] and methanol (CH3OH)[16-18] requires a lot of energy and powerful reducing agents such as H2. On the other hand, the non-reductive transformation of CO2 maintaining the +4 oxidation state of C is moderately exothermic or endothermic. The target compounds for non-reductive conversion include carbonates, carbamates, ureas, carboxylates, polycarbonates, polyurethanes and so on.[4-19]
Among these compounds, cyclic compounds such as cyclic carbonates, cyclic carbamates and cyclic ureas are attractive chemicals. These compounds are frequently used in both organic synthesis and industry. Conventionally, these compounds are synthesized using a toxic and hazardous reagent such as phosgene. Therefore, an alternative method is desirable, and direct synthesis from CO2 is the most highly desired one from the environmental and practical viewpoints. Methods for the direct synthesis of cyclic compounds from CO2 are proposed as follows: cyclic carbonates from CO2 + diols, cyclic carbamates from CO2 + aminoalcohols, and cyclic ureas from CO2 + diamines (Scheme 1). These intramolecular cyclization reactions proceed via two addition reactions: (1) addition of one functional group to CO2 to afford the carbamate or carbonate intermediates; and (2) addition of the other functional group to the intermediates, and the second step is the rate-determining step. Chemicals derived from fossil fuels such as coal or petroleum are being substituted by biomass-derived chemicals due to the carbon neutral property of biomass.[21, 22] For example, 1,2-propanediol can be obtained in one to three steps from biomass or glycerol, which is formed by transesterification of triglyceride.[23-25] In the same way, aminoalcohols or diamines can be obtained by amination of bio-derived diols. Therefore, the cyclization reactions shown in Scheme 1 are attractive from the environmental and organic synthesis viewpoints. Various systems without catalysts or with homogeneous or heterogeneous catalysts have been developed up to now. Systems with a recyclable heterogeneous catalyst are promising from environmental and practical viewpoints such as reusability, handling and product separation.[26-28] This mini-review focuses on the direct conversion of CO2 to these cyclic compounds by catalysts, especially heterogeneous CeO2 catalysts.
Cyclic carbonates synthesis from CO2 and diols
Cyclic carbonates, in particular five-membered cyclic carbonates, are commercially important compounds and are used as electrolytes in lithium secondary batteries, aprotic polar solvents, precursors for polycarbonate synthesis and intermediates for various chemical reactions.[29-32] In addition to their biodegradability and high solvency, they have high boiling and flash points, low odor levels and evaporation rate and low toxicities. Moreover, a possible utilization of cyclic carbonates such as propylene carbonate and ethylene carbonate is the transesterification with CH3OH to give dimethyl carbonate (DMC) and the corresponding glycol. It is well known that the transesterification of propylene carbonate with CH3OH to DMC is a widely used industrial synthetic process.[33-38]
Various synthetic phosgene-free routes for the synthesis of cyclic carbonates have been reported; [39-59] reaction of CO2 and epoxides,[39-45] reaction of CO2 and diols, CO2 oxidative carboxylation of olefin,[47-54] reaction of CO2 with cyclic ketals, reaction of CO2 with propargyl alcohols and reaction of urea with diols.[57-59] Scheme 2 shows the typical reaction routes for propylene carbonate from petroleum and biomass.[60-62] The coupling reaction of CO2 and epoxide has attracted much attention as an atom economy process since by-product is not formed in the reaction, and it has been performed in industry.[38, 63] However, this process has the drawback that epoxides are produced by the oxidation of olefins using hazardous hydroperoxide and the reactive epoxide is difficult to deal with. As mentioned in the introduction, the reaction of CO2 and diols is promising (Equation (1)) because diols can be produced by the hydrogenolysis of biomass or biomass-derived materials such as glycerol, sorbitol, cellulose and so on.[25, 64-89] However, direct synthesis from CO2 and diols is quite difficult from the viewpoints of the low reactivity of OH group and CO2, and the reaction equilibrium (Equation (1)), which is the well-known problem in the synthesis of DMC from CO2 and CH3OH.[90-93]
To overcome these problems, various homogeneous and heterogeneous catalysts for the direct synthesis of cyclic carbonate from CO2 and diols are reported as shown in Table 1.[46, 94-106] This synthetic methodology was initially practiced with the heterogeneous CeO2–ZrO2 catalyst (Table 1, entries 16 and 17), which is known to be effective in the direct synthesis of DMC from CO2 and CH3OH.[107, 108] As for homogeneous catalysts, various organotin compounds that are effective in the carbonate synthesis are reported (Table 1 entries 1–5). The reactions catalyzed by organotin complex catalysts proceed with high selectivities. Among these catalyst systems, n-Bu2SnO combined with zeolite as a dehydrating reagent affords various carbonates in comparatively high yields (35–61%), although the yield is very low without the dehydrating agents. Zinc-based catalysts, especially Zn(OAc)2, were also developed, however, they demonstrated low yields and selectivities (Table 1 entries 6–10). Systems with other catalysts such as alkali carbonates, magnesium, organic bases of TBD (1,5,7-triazabicyclo[4.4.0]dec-5-ene) and so on were also applied to cyclic carbonate synthesis, however, resulting in low yield or low selectivity. In the potassium carbonate (K2CO3) catalyst (Table 1 entry 14), benzonitrile acts as not only a solvent but also a dehydrating agent to increase the conversion (Scheme 3). Dehydration by nitriles such as CH3CN, benzonitrile or 2-cyanopyridine is developed to increase yields by shifting the chemical equilibrium in the synthesis of DMC.[109-113]
Table 1. Cyclic carbonate syntheses from CO2 + diols
|Entry||Catalyst||Polyol||Solvent||T (K)||PCO2 (MPa)||Yield (%)||Selc. (%)||Ref.|
|Homogeneous catalyst system|| || || || || || || || |
|3||n-Bu2SnO + Zeolite||EG||CH3OH||393||13.8||61||100||95|
|4|| ||1,2-PDO||CH3OH||393||13.8||42||100|| |
|5|| ||Gly||CH3OH||393||13.8||35||100|| |
|7|| ||1,2-PDO||CH3CN||443||10||24||62|| |
|9|| ||PhEG||–||453||15||1||100|| |
|10|| ||EG||–||453||15||1||100|| |
|11||RhCl3 + PPh3 + KI||Gly||CH3OH||413||4 (1)a||<1||100||98|
|13||Cs2CO3 + (NH4)2CO3||1,2-PDO||CH3CN||448||10||11||100||100|
|Heterogeneous catalyst system|| || || || || || || || |
To the best of our knowledge, two heterogeneous catalyst systems, CeO2–ZrO2 and KI/ZnO, were reported. In the KI/ZnO catalyst system, CH3CN works as not only a solvent but also a dehydrating agent (Table 1 entry 18). Taking the solubility of KI and ZnO into consideration, CeO2-based catalysts are promising for the future direct synthesis of cyclic carbonates from CO2 and diols because of the high resistance of CeO2 to dissolution or sintering by thermal treatment. Detailed results on the CeO2-based catalysts and the reaction mechanism are shown below.[103-105]
Figure 1 shows ethylene carbonate and propylene carbonate formation from CO2 with ethylene glycol and 1,2-propanediol over CeO2, ZrO2, and CeO2–ZrO2 (Ce : Zr = 1:1, 1:2 and 1:4) catalysts with various calcination temperatures (1273, 1073, 873, 673 K). The selectivity for the reaction over CeO2-ZrO2 catalyst is very high (>99%) under the present conditions. High selectivity to the cyclic carbonates means that the reaction of esterification and the formation of polycarbonates are strongly suppressed. The ethylene carbonate and propylene carbonate amount at the equilibrium level are around 1.2 mmol (1.2% conversion) and 2.0 mmol (2.0% conversion), respectively, under the reaction conditions used. These equilibrium conversion values were obtained by an experimental method. The saturation level of the conversion was verified easily by changing the amount of catalyst because of the high selectivity to the product. Generally speaking, it is very convenient to determine the equilibrium conversion by thermodynamic calculation. However, in the case of carbonate synthesis, it is difficult to predict the equilibrium conversion by thermodynamic calculation. For example, even in the case of the synthesis of DMC from CH3OH and CO2, which is the simplest and the most popular system, the thermodynamic calculation did not predict the equilibrium conversion precisely. This is probably because of the difference between the experiment and model. In addition, H2O as an impurity can also affect the experimental equilibrium conversion. Further investigation and efforts are necessary to predict the equilibrium conversion by thermodynamic calculation in other fields than catalytic chemistry. The ethylene carbonate and propylene carbonate amount shown in Fig. 1 (ethylene carbonate < 0.8 mmol and propylene carbonate < 1.5 mmol) did not reach the equilibrium level, which means that the results reflect the catalytic activities in the reactions. The trends in the results obtained for ethylene carbonate and propylene carbonate syntheses are almost the same. In the case of CeO2, the cyclic carbonate amount reached a maximum on the catalyst calcined at 873 K, and decreased on catalysts calcined at a temperature higher than 873 K. This behavior can be explained by the crystalline state of CeO2 and change of the surface area. The decrease of the activity from 873 K to 1073 K treatment can be explained by a decrease of surface area (36 m2 g−1 (873 K) to 7 m2 g−1 (1073 K)). The increase in activity from 673 K to 873 K treatment is explained by the crystalline state of CeO2. The crystalline state of CeO2 catalyst calcined at 673 K is mainly amorphous, however, that of CeO2 catalysts calcined at 873 K and above 873 K is mainly crystalline. The amorphous state of CeO2 will be less active for the reaction. No ethylene carbonate formation was detected over ZrO2 at any calcination temperatures (673, 873, 1073 and 1273 K). In addition, at 673 K and 873 K, cyclic carbonate amount was increased as the content of Ce was increased. These results strongly suggest that the main active species of the catalysts are cerium ions. For the catalysts containing Zr, the calcination temperature affording the maximum propylene carbonate amount was generally changed to higher one in comparison with pure CeO2 in spite of the decrease in catalyst surface area. The active sites are thought to be weak acid–base sites[115-121] on the plain crystalline surface. The active sites can be influenced by the bulk composition of Ce/Zr and more significantly by the surface composition. Calcination temperature can influence the number, balance, and strength of the acid–base sites through the surface composition of Ce/Zr and the surface structure intricately. The tendency in the catalyst activity of CeO2–ZrO2 was similar to that obtained in synthesis of DMC from CH3OH and CO2. The detailed reaction mechanism for the synthesis of cyclic carbonates is discussed in the final section.
Figure 1. Cyclic carbonate synthesis from glycol and CO2 over various CeO2-ZrO2 catalysts with various calcination temperatures.[103-105] The numbers near the marks are calcination temperatures (K). Reaction conditions: ethylene glycol (or 1,2-propanediol) : CO2 : CH3CN = 100 : 200 : 120 mmol, catalyst 0.05 g, T = 423 K, t = 2 h.
Download figure to PowerPoint
As mentioned above, the synthesis of the cyclic carbonates from CO2 and the corresponding diols is strongly limited by the equilibrium to give very low conversion level. However, it is expected that the removal of water by the use of the reactive solvent and molecular sieves can improve the conversion as in the case of DMC from CH3OH and CO2. At present, research on conversion improvement is limited, and not yet ready for reviewing, however, future progress is expected.
Cyclic carbamates synthesis from CO2 and aminoalcohols
Cyclic carbamates are widely used as intermediates of agrichemical, pesticide, herbicide and pharmaceutical agents,[122-124] solvents, and chiral auxiliaries.[126, 127] In particular, acylation of cyclic carbamates is a frequently-used method for the synthesis of useful compounds such as amino acids. For the synthesis of cyclic carbamates without phosgene, the following methods are known; insertion of CO2 into aziridines,[129, 130] carbonylation of aminoalcohols with CO2,[131-141] reaction of CO2 with acetylenic amines[142-144] and three-component reactions of propargylic alcohols, primary amines and CO2.[145, 146] Among these reaction systems, carbonylation of aminoalcohols is the most desirable from the viewpoints of resources utilization and green chemistry because in comparison with aminoalcohols these reagents are unstable and/or difficult to obtain or prepare. The important points of the carbonylation of aminoalcohol with CO2 are activation of OH group and amino group, especially less reactive OH group, and removal of H2O which is produced by the reaction proceeding from the aspect of the reaction equilibrium (Equation (2)).
The reported reaction systems for cyclic carbamate synthesis from CO2 and aminoalcohols are summarized in Table 2.[131-141] To compare the activities of catalysts and the reactivities of substrates, 2-aminoethanol and 2-amino-2-phenylethanol, which have a primary OH group, are selected as substrates since the substituent at the α-position of the OH group can increase the nucleophilicity of the OH group.[136, 137] Effective homogeneous systems composed of phosphine reagents and strong bases were reported by Kubota, Kodaka, Dinsmore and Muñoz (Table 2 entries 1–5), where the reaction takes place through in situ N-carboxylation and cyclodehydration (Scheme 4). The reaction system with Ph3P/Et3N/CCl4 (Table 2 entry 1) provided the moderate yield of 2-oxazolidinone (51%), although the system provided high yield in the case of 4-phenyl-2-oxazolidone synthesis from 2-amino-1-phenylethanol (91%). CCl4 acts as an initiator of the reaction by activating PPh3. The drawback of this reaction system is use of the stoichiometric amount of CCl4. The reaction system with n-Bu3P/Et3N/Fe4S4/NPSH (4-nitrothiophenol) (Table 2 entry 2) also affords the moderate yield of 2-phenyl-2-oxazolidone (68%), where Fe4S4 works as the electron mediator with thiol and phosphine. The reagents used by Kodaka and Dinsmore are called Mitsunobu reagents, which are composed of R3P and azo compounds such as diethyl azodicarboxylate (DEADC), di-tert-butyl azodicarboxylate (DBAD). The reaction systems using Mitsunobu reagents give comparatively higher yields (Table 2 entries 3 and 4). DPPCl (diphenyl chlorophosphate)/Et3N reaction system (Table 2 entry 5) is also a good one, and gives 2-phenyl-2-oxazolidone in high yield (95%). These reaction systems are useful and are often utilized for organic synthesis because the reaction conditions are very mild (CO2 ∼ atmospheric pressure, T ≤ room temperature). As shown in Scheme 4, the stoichiometric amount of alkylphosphine oxide is formed in these reaction systems, which means that alkylphosphine plays the role of removal of H2O from the reaction media. Therefore, these reaction systems are not restricted by the reaction equilibrium (Equation (2)), leading to high yields. However, these reaction systems suffer from use of a strong base such as DBU (1,8-diazabicyclo[5.4.0]undec-7-ene), trialkylamine and guanidine, which causes the production of large amounts of incompetent salts derived from reaction with acids. In addition, the stoichiometric amount of alkylphosphine oxide is difficult to remove from the product.
Table 2. Cyclic carbamate syntheses from CO2 + 2-aminoethanols
|Entry||Catalyst||Solvent||R||R||T (K)||PCO2 (MPa)||Yield (%)||Selc. (%)||Ref.|
|Stoichiometric reagent system|| || || || || || || || || |
|2||n-Bu3P/Et3N/ Fe4S4/NPSH||CH3CN||H||Ph||r.t.||0.1(O2 (10%))||68||–||132|
|Homogeneous catalyst system|| || || || || || || || || |
|6||Ph3SbO + Molecular Sieves||Benzene||H||H||433||5||trace||–||136|
|Non catalyst system|| || || || || || || || || |
|Heterogeneous catalyst system|| || || || || || || || || |
As for homogeneous catalyst systems (Table 2 entries 6–8), systems with organotin or organoantimon complex and ionic liquid with alkali metal promoter have been developed under severe conditions (403–473 K, 3–10 MPa and 6–24 h). As shown in Table 2 entry 6, Ph3SbO catalyst hardly showed any activity in the case of 2-aminoethanol, although it afforded cyclic carbamates in high yields for substrates having dialkyl or 2-hydroxyethyl group (∼90%). On the other hand, n-Bu2SnO catalyst demonstrated higher yield for the transformation of 2-aminoethanol although at a high temperature of 453 K (Table 2 entry 7). In these catalyst systems, the organometals activated the OH group by coordination as shown in Scheme 5. Ionic liquid with alkali metal promoter is effective for the reaction, where the cation of the ionic liquid acts as the activator to the carbonyl moiety of the carbamate produced by amino group + CO2.
From the environmental and practical viewpoints as mentioned above, non-catalytic systems or systems with heterogeneous catalysts are preferable. Arai and co-workers and Tominaga and co-workers reported non-catalytic systems of cyclic carbamates synthesis from CO2 and aminoalcohols at 423 K and 453 K, respectively (Table 2 entries 9 and 10), which, however, have many problems such as low yields (<56%) and narrow scope of aminoalcohols. As for heterogeneous catalysts, Garcia and co-workers and our group reported that CeO2 is an effective catalyst for the reaction (Table 2 entries 11 and 12). CeO2 nanoparticle (np-CeO2) reported by Garcia catalyzed the reaction in ethanol solvent, which suffered from low yield (<60%) and narrow substrate scope. In addition, CeO2 has also been reported to be an effective catalyst for the synthesis of methyl benzylcarbamate from benzylamine, CO2 and CH3OH. Moreover, CeO2 catalyst system with a CH3CN solvent gave excellent results in our laboratory. The detailed results on the scope, the solvent effect and the reaction mechanism over CeO2 are summarized below.
Various metal oxides were tested as catalysts for the formation of 2-oxazolidinone from 2 MPa CO2 and 2-aminoethanol at 423 K and the reaction rates on various metal oxides were compared as TOF in Fig. 2. Formation of 2-oxazolidinone hardly proceeded without a catalyst (yield 0.7% in 2 h). Among these metal oxides, CeO2 showed the highest TOF and the TOF for CeO2 was more than thirty times as high as those for the other metal oxides. In addition, CeO2 catalyst can be reused at least three times without loss of selectivity and activity.
Figure 2. Comparison of formation rate of 2-oxazolidinone from CO2 and 2-aminoethanol over various metal oxides Reaction conditions: 2-aminoethanol 10 mmol, acetonitrile 41 g, catalyst (metal = 0.2 mmol), PCO2 = 2 MPa (at r.t.), T = 423 K, t = 2 h. TOF (Turnover frequency) was measured under the conditions where the conversion was below 40%. TOF (h−1) = Amount of product (mol) / Total amount of metal atom (mol) / Time (h).
Download figure to PowerPoint
The solvent effect is an important factor for the synthesis of cyclic carbamates from CO2 and aminoalcohols. The reaction of 2-aminoethanol + CO2 over CeO2 at 423 K was examined with various solvents (CH3CN, ethanol (EtOH), 1,4-dioxane, tetrahydrofuran (THF), dimethylformamide (DMF), dimethylsulfoxide (DMSO), N-methyl-2-pyrrolidone (NMP)) (Fig. 3). CH3CN, DMSO and NMP were found to be better solvents than the others in terms of conversion and selectivity of 2-oxazolidinone. The solvents affect the reaction of 2-aminoethanol and CO2 from two aspects. One is the side reaction involved in the solvents. In fact, the side reaction in EtOH solvent proceeded more preferably than that in CH3CN solvent, which is derived from the reactivity of the solvent with the substrate or products. The other aspect is the solvent effect on the catalytic activity. The activity in EtOH solvent was lower than that in CH3CN solvent. Taking into consideration that generally alcohol is strongly adsorbed on CeO2 surface,[115, 117, 140] EtOH would block the adsorption of the substrate. These solvent effects can explain the difference in the yield and selectivity between our results and np-CeO2 in the literature.
Figure 3. Solvent effect for the cyclic carbamate synthesis from CO2 and 2-aminoethanol over CeO2 Reaction conditions: 2-aminoethanol 10 mmol, solvent 1000 mmol, CeO2 0.17 g, PCO2 = 2 MPa (at r.t.), T = 423 K, t = 2 h.
Download figure to PowerPoint
The substrate scope was examined with various aminoalcohols over CeO2 as shown in Scheme 6. β-Aminoalcohols were transformed almost quantitatively to the corresponding five-membered ring cyclic carbamates in high yields (91– > 99%). γ-Aminopropanols, which are difficult to synthesize because of the longer alkyl chain, were also converted to six-membered ring cyclic carbamates in high yields (95– > 99%), although these substrates required longer reaction times than 2-aminoalcohols. In the case of 2-ethanolamine, the maximum yield by the reaction equilibrium is determined to be 97% by the experimental method. Therefore, the yields for the other substrates will be controlled by the reaction equilibrium. The conversion at equilibrium in the reaction of 2-aminoethanols with CO2 to the corresponding carbamates was much higher than that in the reaction of diols with CO2 to the corresponding cyclic carbonates. The reaction of 2-aminoethanol with CO2 is more exothermic than that of diols with CO2 because of the higher basicity of the amino group than that of the hydroxyl group, and this can explain the tendency in the equilibrium conversion. This interpretation is also applied to the case of the reaction of diamines with CO2.
Scheme 6. Scope of CeO2-catalyzed cyclic carbamates synthesis from CO2 and aminoalcohols. The results for the various cyclic carbamates are shown under the products as Time, Yield (selectivity). Reaction conditions: substrate 10 mmol, acetonitrile 41 g, CeO2 0.17 g, PCO2 = 5 MPa (at r.t.), T = 423 K.
Download figure to PowerPoint
Moreover, optically pure (R)-1-amino-2-propanol was applied to the cyclic carbamate synthesis over CeO2 (Scheme 6). The configuration of the substrate was retained to afford (R)-5-methyl-2-oxazolidinone in 99% ee. These results demonstrated that the reaction system composed of CeO2 and CH3CN can be applied to various aminoalcohols and is very useful in organic and synthetic chemistry.
The reaction mechanism for the synthesis of cyclic carbamates from CO2 and aminoalcohol over CeO2 is discussed in the final section.
Cyclic ureas synthesis from CO2 and diamines
Cyclic ureas are heterocyclic motifs frequently observed in biologically active molecules, and in particular six-membered ring ureas are important[148-152] because of their antineoplastic, anti-viral[150, 151] and anti-arrhythmic activity. For the synthetic methods of cyclic ureas without toxic reagents including phosgene, the following methods are known; the reaction of epoxide with CO2 and diamines, the reaction of cyclic carbonates with diamines,[154, 155] and the reaction of CO2 with diamines.[156-163] From the environmental viewpoint, reaction of CO2 with diamines will provide an attractive alternative method for the direct synthesis of cyclic ureas (Equation (3)).
Compared with the syntheses of cyclic carbonates or carbamates, the synthesis of cyclic ureas is comparatively facile because nucleophilicity of the amino group is much higher than that of the OH group. Table 3 summarizes catalysts and reagents for the direct reaction of CO2 with diamines to cyclic ureas.[156-163] As mentioned in the section on cyclic carbamates synthesis, the reaction system composed of PhTMG and DPPA is effective for this reaction (Table 3 entry 1), however, there are many issues such as production of the salt derived from the neutralizations of the strong base and the stoichiometric amount of phosphine oxide.
Table 3. Cyclic urea syntheses from CO2 + diamines
|Entry||Catalyst||Solvent||Diamine||T (K)||PCO2 (MPa)||Yield (%)||Selc. (%)||Ref.|
|Stoichiometric reagent system|| || || || || || || || |
|Homogeneous catalyst system|| || || || || || || || |
|Non catalyst system|| || || || || || || || |
|Heterogeneous catalyst system|| || || || || || || || |
As for homogeneous catalyst systems for this reaction, Ph3SbO/P4S10 and TBA2[WO4] catalysts were reported (Table 3 entries 2 and 3). The reaction by Ph3SbO/P4S10 proceeds on the same reaction route mentioned in carbamate synthesis by organometal catalysts. The TBA2[WO4] catalyst reported by Mizuno is an efficient homogeneous one for the synthesis of cyclic ureas from diamines and atmospheric CO2. The high activity of the TBA2[WO4] catalyst is achieved by activation of CO2 by the weak Lewis basic property and capture of diamine by hydrogen bonding.
Heterogeneous catalysts or non-catalytic systems have been developed as more preferable systems to homogeneous ones. Arai and co-workers and Zhao and co-workers reported non-catalytic systems for the direct CO2 conversion to cyclic ureas (Table 3 entries 4 and 5). However, these systems were conducted under harsh reaction conditions such as high pressure (≥ 6.0 MPa) and high temperature (≥ 473 K), and cannot achieve a satisfactory yield of six-membered-ring urea. As for heterogeneous catalysts, polyethylene-glycol-supported potassium hydroxide (KOH/PEG1000) and CeO2 catalysts were reported (Table 3 entries 6–8). KOH/PEG1000 was an efficient catalyst at 8 MPa CO2 pressure and 423 K, which suffered from high pressure, narrow scope of substrates and poor yields of cyclic ureas (≤ 82%), which has room for improvement. Nanoparticulate CeO2 prepared by biopolymer-template process using mesoporous alginate aerogel was an efficient catalyst at 0.7 MPa CO2 pressure and 433 K. However, this catalyst has drawbacks of narrow scope of substrates and low yields of cyclic ureas (≤ 37%). On the other hand, it has recently been demonstrated that pure cerium oxide (CeO2) acts as an effective heterogeneous catalyst for the reaction even at low CO2 pressure of 0.3 MPa in isopropyl alcohol (IPA) solvent to afford cyclic ureas in high yields and selectivities. The detailed results on the scope and the reaction mechanism over CeO2 are summarized in a later section.
Various metal oxides were tested as catalysts for the formation of 2-imidazolidinone from 0.5 MPa CO2 and ethylenediamine at 433 K and the results of TOF and selectivity are shown in Fig. 4. Formation of 2-imidazolidinone hardly proceeded without a catalyst under these conditions (yield 0.8% in 1 h). ZnO and CeO2 gave higher conversion and TOF than the other metal oxides. As for selectivity, CeO2 shows higher value of 93%, however, the value shown by ZnO is clearly lower (68%). In addition, at longer reaction times, pure CeO2 afforded high yield (94%), however, ZnO provided lower yield (< 40%) than in the case of 1 h reaction time. This is because the produced polymer may cover the surface of ZnO, leading to deactivation of ZnO. Therefore, pure CeO2 is the most effective catalyst for the reaction among the various metal oxides tested. The lower yield associated with nanoparticulate CeO2, which is prepared using sodium alginate, may arise from the presence of residual impurities such as sodium. From the results of the reusability and leaching tests, CeO2 catalyst can be reused at least three times without loss of selectivity and activity and regarded as a heterogeneous catalyst.
Figure 4. Comparison of the formation rate of 2-imidazolidinone and selectivity in the reaction of ethylenediamine and CO2 over various metal oxides Reaction conditions: ethylenediamine 10 mmol, methanol 6.4 g, metal oxide (metal : 0.2 mmol), PCO2 = 0.5 MPa, T = 433 K, t = 1 h. TOF (Turnover frequency) was measured under the conditions where the conversion was below 40 %. TOF (h−1) = Amount of product (mol) / Total amount of metal atom (mol) / Time (h).
Download figure to PowerPoint
Solvent effects for the synthesis of 2-imidazolidinone from CO2 and ethylenediamine are shown in Fig. 5. At low conversion conditions, the catalytic performance is not influenced so significantly by the solvents. On comparing the detail of the solvent effect, 1-propanol and IPA provided slightly higher conversion and selectivity, and CH3CN and NMP provided lower selectivity.
Figure 5. Solvent effect for the cyclic urea synthesis from CO2 and ethylenediamine over CeO2 Reaction conditions: ethylenediamine 10 mmol, solvent 200 mmol, CeO2 0.086 g, PCO2 = 0.5 MPa, T = 433 K, t = 1 h.
Download figure to PowerPoint
Cyclic urea syntheses from CO2 and various diamines over CeO2 using CH3OH and IPA are shown in Scheme 7. In the case of CH3OH as solvent, ethylenediamine, 1,2-propanediamine and 2-methyl-1,2-propanedimiane reacted to afford the corresponding cyclic ureas in high yields (86–94%). N-alkyl ethylenediamines were also converted selectively. 1,3-Propanediamine and N-methyl-1,3-propanediamine were applied to provide the six-membered-ring ureas. However, N,N′-dimethylethylenediamine was converted in low yield (43%). Using IPA as solvent instead of CH3OH provided high yield (78%) with high selectivity (94%). Other diamines were also transformed using IPA solvent to afford the corresponding cyclic ureas in higher yields (88–96%). These results indicate that IPA is a preferred solvent for the synthesis of various cyclic ureas. It should be noted that 1,3-propanediamines, which are known to be more difficult substrates to convert,[160, 162] can be converted to the corresponding six-membered-ring ureas, and they are core structures of important medicines for antineoplastic, HIV-1 and anti-arrhythmic.[148-152] These yields are higher than the previously reported yields (max yield: 86%). Therefore, this catalytic system composed of CeO2 and 2-propanol solvent is applicable to the various diamines to afford the corresponding cyclic ureas in high yields. From the result of the solvent effect (Fig. 5), the solvent effect for the activity and selectivity is small when ethylenediamine is used as a substrate. On the other hand, from the result for the substrate scope (Scheme 7), N-alkylated diamines or 1,3-propanediamines were subject to the solvent effect, leading to the low activity and selectivity in the case of CH3OH solvent. The decrease of activity can be explained by the adsorption strength of these diamines and the solvents. Adsorption of these diamines on CeO2 is weaker than that of ethylenediamine because of the bulky alkyl chain or stability of adspecies, which results in the situation that CH3OH can be adsorbed on CeO2. Therefore, CH3OH interfered with the adsorption of diamines and the activity is decreased. In contrast, the adsorption of IPA on CeO2 is much weaker than that of CH3OH owing to the bulky alkyl chain of IPA, and hence IPA cannot affect the activity of CeO2. At the same time, this behavior of weak IPA adsorption is also connected to high selectivity. The main side reaction is the reaction of the produced ureas with the solvent, which is activated on CeO2. Therefore, adsorption of the solvent on CeO2 and nucleophilicity of the solvent have an influence on the selectivity. Adsorption of IPA is weaker than CH3OH on CeO2, and at the same time, IPA is less reactive than CH3OH because IPA has a bulky alkyl chain. Totally, the selectivity in IPA solvent is higher than that in CH3OH solvent. The reaction mechanism for the synthesis of cyclic ureas from CO2 and diamines over CeO2 is discussed in the following section.
Scheme 7. Scope of CeO2-catalyzed cyclic ureas synthesis from CO2 and diamines. The results for the various cyclic ureas are shown under the products as Time, Yield (selectivity). The upper row is the result of the reaction in methanol and the lower row is that in IPA. Reaction conditions: substrate 10 mmol, solvent 200 mmol, CeO2 0.34 g, PCO2 = 0.5 MPa (at r.t.), T = 433 K.
Download figure to PowerPoint
Reaction mechanism over CeO2103–105,140,161
Kinetic analysis in combination with in situ spectroscopic analysis such as FTIR is an effective method to clarify the reaction mechanisms of CeO2. Taking the synthesis of cyclic ureas as an example, the reaction mechanism over CeO2 and catalytic properties of CeO2 are discussed in this section.
Initially, FTIR studies on adsorption of ethylenediamine and/or CO2 at 303 and 433 K were conducted. Ethylenediamine is adsorbed on Lewis acid sites of CeO2 and did not desorb from CeO2 even at 433 K, which indicates that ethylenediamine is strongly adsorbed on the Lewis acid sites of CeO2 in the reaction conditions. Introduction of CO2 to ethylenediamine-treated CeO2 at the low temperature of 303 K demonstrates that ethylenediamine reacted with CO2 on CeO2 to afford carbamate species on CeO2 at 303 K as shown below.
On the other hand, in general, ethylenediamine easily reacts with CO2 at room temperature to provide dicarbamic acid. In the actual reaction media, adsorption of the carbamic acid on CeO2 probably forms the carbamate species on CeO2. Production of 2-imidazolidinone was observed at the reaction temperature of 433 K by introduction of CO2 to ethylenediamine-treated CeO2 on FTIR. Considering that at 303 K the carbamate species was formed but 2-imidazolidinone was not formed on CeO2, 2-imidazolidinone should be produced by reaction of the carbamate species on CeO2. In general, the carbamate species is decomposed to CO2 and amine at above 383 K, which is much lower than the reaction temperature (433 K). Since the nucleophilicity of the N atom of carbamate species is lower than that of amine, the reaction proceeds through nucleophilic addition of a free amino group that is produced by the decomposition of carbamate species, which is the rate-determining step.
The effects of concentration of ethylenediamine and CO2 pressure were investigated. The reaction order with respect to ethylenediamine concentration was almost zero, which indicates that the diamine is strongly adsorbed on CeO2. Taking into account that diamine and CO2 easily formed carbamate species on CeO2 at 303 K, the carbamate species will cover the CeO2 surface. The reaction rate shows a small negative dependence on CO2 pressure. The adsorbed carbamic acid will be more stable at higher CO2 pressure. High CO2 pressure can suppress the decomposition of carbamic acid. This result also supports the finding that nucleophilic addition of a free amino group to another carbamate moiety, which is carbonyl C atom of carbamate adspecies on CeO2, is the rate-determining step.
In addition, from comparison with the reactivity of N-alkylated diamines, it is found that the reactivity of N-alkylated substrates is lower than ethylenediamine, which will be derived from steric hindrance and/or basicity of the N-alkylated amino group. This result suggests that the formation of a free amino group is important for the reaction, and this supports the proposal that nucleophilic addition of a free amino group to carbamate moiety on CeO2 is the rate-determining step.
Based on the above discussions, the proposed reaction mechanism for the cyclic urea synthesis from CO2 and ethylenediamine is shown in Scheme 8(c): (1) adsorption of amine and CO2 to afford carbamate adspecies on CeO2,; (2) decomposition of the carbamate species to amine and CO2,; (3) nucleophilic addition of the amino group to the carbamate moiety on CeO2, providing 2-imidazolidinone; and (4) desorption of 2-imidazolidinone and regeneration of CeO2. The third step is the rate-determining step.
Similar kinetic and spectroscopic studies were conducted to determine the reaction pathway of the cyclic carbamate synthesis at 423 K. The important point of the reaction mechanism in the synthesis of cyclic carbamates is whether the nucleophilic species is hydroxyl group or amino group at the rate-determining step. From the kinetic studies and in situ FTIR analyses using isotopically labeled substrates of aminoalcohols, it was determined that the reaction proceeds via nucleophilic addition of the hydroxyl group to carbamate adspecies on CeO2. The proposed reaction mechanism for the synthesis of 2-oxazolidinone from CO2 and 2-aminoethanol is shown in Scheme 8(b), which is composed of four steps: (1) adsorption of 2-aminoalcohol and CO2 to afford both carbonate and carbamate adspecies on CeO2; (2) decomposition of the carbonate to the hydroxyl group and CO2; (3) nucleophilic addition of the hydroxyl group to the carbamate moiety on CeO2, providing 2-oxazolidinone; (4) desorption of 2-oxazolidinone and regeneration of CeO2. The reaction mechanism for the formation of cyclic carbonate is suggested to be a similar one to those for the formations of cyclic carbamates and cyclic ureas (Scheme 8(a)) including (1) adsorption of ethylene glycol and CO2 to afford carbonate adspecies on CeO2, (2) decomposition of the carbonate to the hydroxyl group and CO2, (3) nucleophilic addition of the hydroxyl group to the carbonate adspecies on CeO2, providing ethylene carbonate, and (4) desorption of the ethylene carbonate and regeneration of CeO2.
From the data obtained above, we cannot compare the reactivity between diols, aminoalcohols and diamines owing to differences of the reaction conditions and the limitations of reaction equilibrium. To compare the reactivity, the reactions of CO2 with ethylene glycol, 2-aminoethanol and ethylenediamine under the same reaction conditions were tested at various reaction temperatures (383–443 K) and the results are shown in Table 4. The initial reaction rate was calculated at low conversion conditions (<3%), where the reactions are controlled by the kinetics. The initial formation rate of 2-imidazolidinone is lowest in these cyclic compounds. For example, at temperatures below 413 K, the order of the initial rate for these substrates was ethylene glycol > 2-aminoethanol > ethylenediamine. Generally speaking, thermodynamically more stable products tend to be formed more preferably. However, the observed reactivity order is different from that expected from the thermodynamic stability of the products (cyclic ureas > cyclic carbamates > cyclic carbonates).
Table 4. Comparison of the reactivity for cyclic compounds synthesis over CeO2
|T (K)||Reaction rate (mmol g−1 h−1)|
This high reactivity of the substrates with at least one OH group on CeO2 is very interesting. Two possible reasons can be considered. One is the stability of the adspecies of these substrates. From the in situ FTIR studies on the reaction of alcohols and amines with CO2, amines readily reacted with CO2 at lower CO2 pressure (0.1 MPa) to produce the stable carbamate adspecies on CeO2 than alcohols. Therefore, stabilization of the adspecies may increase the activation barrier, resulting in lowering the reactivity of diamines. The second reason is the degree of the activation of hydroxyl group and amino group at the rate-determining step (Scheme 8). Taking the basicity of the functional groups (amino group and hydroxyl group) and basicity of CeO2 into consideration, hydroxyl group is activated more easily than amino group on CeO2, which is suggested by the reported FTIR adsorption results.[116, 118] The activation of the hydroxyl group on CeO2 may decrease the activation barrier of the reaction, leading to high reactivity of diols and aminoalcohols.
On inspecting the initial rates in Table 4, the formation rate of ethylene carbonate at temperatures lower than 403 K is higher than that of 2-oxazolidinone, however, at temperatures higher than 423 K, the order of the formation rates became opposite, which indicates a difference in the activation energy and/or a change of reaction mechanism. Figure 6 shows the Arrhenius plots for these reactions. The activation energies (Ea) for the formation of ethylene carbonate and 2-imidazolidinone were calculated to be 43 and 81 kJ mol−1, respectively. On the other hand, the Arrhenius plot for the formation of 2-oxazolidinone was not linear and showed a major change around at 413 K, suggesting a change of the reaction mechanism above and below about 413 K. In the Arrhenius plot for 2-oxazolidinone synthesis, the activation energy calculated from the straight line below 403 K was 68 kJ mol−1, which is close to that of 2-imidazolidinone synthesis, and that above 423 K was 56 kJ mol−1, which is close to that of ethylene carbonate synthesis. One possible interpretation is that 2-oxazolidinone is formed through the nucleophilic addition of the amino group below 413 K, and 2-oxazolidinone is formed through the nucleophilic addition of the hydroxyl group above 413 K, which supports the above-mentioned reaction mechanism for cyclic carbamate synthesis at 423 K. Therefore, the difference of the formation rate of 2-oxazolidinone between above and below 403 K can be explained by the stabilization of adspecies of the amino group and/or activation of the hydroxyl group as discussed above.
Figure 6. Arrhenius plot for the synthesis of ethylene carbonate, 2-oxazolidinone and 2-imidazolidinone. Reaction conditions: substrate 10 mmol, CH3CN 200 mmol, CeO2 0.001–0.05 g, PCO2 = 5.0 MPa, T = 383–443 K. The reaction rate was determined under the conditions where the conversion was below 3%.
Download figure to PowerPoint
As above, the acid–base pair sites on CeO2 works as a stabilizer of adspecies of amino group and/or activator of the hydroxyl group of diols or aminoalcohols. It should be noted that CeO2 adsorbs and activates CO2, but is not deactivated by CO2, indicating that the weak base sites on CeO2 undoubtedly play an important role on the activation of CO2 in the synthesis of these cyclic compounds. Therefore, the acid–base pair sites on CeO2 and weak base property of CeO2 act effectively on the reactions of CO2 with diols, aminoalcohols or diamines, in particular diols and aminoalcohols with a hydroxyl group.
Most results introduced in this mini-review were carried out at laboratory scale using pure-CO2, and the topics are limited to the development of the catalysts. On the other hand, in the practical application of the catalytic conversion of CO2, the utilization of waste CO2 should be considered as well as the heat and pressure management of the reaction systems.[164-166] For example, future work on the effect of the composition of the CO2-containing gas on the catalytic performance will be necessary.