SCIENCE CHINA Chemistry •  ARTICLES  •

doi: 10.1007/s11426-016-0405-3

· SPECIAL ISSUE ·Green Chemistry

Recyclable bifunctional aluminum salen catalyst for CO2 fixation: the efficient formation of five-membered heterocyclic compounds Rongchang Luo, Zhi Yang, Wuying Zhang, Xiantai Zhou & Hongbing Ji* Key Laboratory of Low-Carbon Chemistry & Energy Conservation of Guangdong Province, School of Chemistry, Sun Yat-sen University, Guangzhou 510275, China Received October 3, 2016; accepted December 7, 2016; published online March 1, 2017

A variety of unique Al(salen) complexes functionalized by imidazolium-based ionic liquid (IL) moieties with the salen ligand at the two sides of 3,3′-position have been successfully prepared, rather than familiar 5,5′-position reported previously. The catalytic activity obtained by these bifunctional catalysts could be superior to those of the binary type catalysts in the formation of five-membered heterocyclic compounds from the cycloaddition reaction of CO2 and three-membered heterocyclic compounds (including terminal epoxides and N-substituted aziridines), presumably due to the distinguished intramolecularly synergistic catalysis, which might lead to perform the cycloaddition reaction at ambient conditions and retain excellent yield and unprecedented chemo- or regioselectivity. Moreover, the polyether-based trifunctional Al(salen) catalysts with the best catalytic performance could be regenerated and reused at least eight times without any obvious decreases in catalytic activity. Finally, the kinetic investigation suggested the structure of catalysts had important influences on the catalytic activity, thereby proposing the possible reaction mechanism. carbon dioxide, cyclic carbonate, salen aluminum, bifunctional catalyst, cooperative effect

Citation:

Luo R, Yang Z, Zhang W, Zhou X, Ji H. Recyclable bifunctional aluminum salen catalyst for CO2 fixation: the efficient formation of five-membered heterocyclic compounds. Sci China Chem, 2017, 60: doi: 10.1007/s11426-016-0405-3

1   Introduction In terms of the so-called “sustainable society” and “green chemistry” concepts, carbon dioxide (CO2) is one of the major greenhouse effect gases but it has been attracted much attention as a highly abundant, inexpensive, nontoxic, nonflammable and biorenewable building block (an ideal C1 resource) both from industrial and academic viewpoints in recent years [1]. Efficient strategies for the chemical fixation of CO2 are extremely difficult and challenging attributed to exceptional thermodynamic and kinetic stability. In general, the most important methodologies to transform CO2 into economically competitive products are the following two: use *Corresponding author (email: [email protected]) © Science China Press and Springer-Verlag Berlin Heidelberg 2017

the higher-energy substrates or choose the oxidized low-energy synthetic targets [1a]. Therefore, the insertion of CO2 to three-membered heterocyclic compound (such as epoxide or aziridine) is generally regarded as a preeminent route in this area because of its highly atom economy and environmental friendliness (Scheme 1) [2]. More interestingly, the corresponding five-membered cyclic products are widely used as important intermediates in the production of pharmaceuticals and fine chemicals [3]. Until now, numerous catalytic systems have been developed for promoting the transformation, including metal oxides [4], quaternary onium salt [5], alkali metal halides [6], ILs [7], organic bases [8], metalloporphyrin [9], and metallosalen complexes [10], etc. [11]. Prominent among these are a series of metalloporphyrin or metallosalen complexes (such as chromium, cobalt, zinc, and chem.scichina.com   link.springer.com

2

 

   Sci China Chem   

Luo et al.

   Protocols Scheme 1        for the preparation of five-membered heterocyclic organic compounds (color online).

aluminium) in the presence of quaternary onium salts or organic bases since these binary catalytic systems could catalyze the coupling reaction of CO2 with three-membered heterocyclic compounds under mild conditions [12]. Soon afterwards, most known bifunctional catalysts have suitable built-in functions (nucleophiles or basic sites) might also be easy to implement directly the catalytic cycles. The interor intra-molecular synergistic effect between a Lewis acid from the metal center and Lewis base from a nucleophile (X−, cocatalyst) results in remarkably improved catalytic behavior compared with a metal-free catalyst system [10b,13]. In 2014, our group [14] reported novel Al(salen) derivatives functionalized by polyether-based imidazolium ILs exhibited the built-in “CO2 capture” capability in the CO2/epoxides cycloaddition reaction. However, the higher reaction temperature (≥100 °C) or CO2 pressure (≥1.0 MPa) is often required so as to obtain the excellent catalytic properties. Additionally, these salen-type bifunctional catalysts may present lower catalytic activity than those of the similar binary catalytic system under the desired reaction conditions, presumably due to the conformational restriction aroused from the catalyst structure [10g,15]. In order to further enhance the activity of the bifunctional catalyst from the standpoint of the structure-activity relationships, there are generally four kinds of main design strategies

that can be used to implement the process intensification of the coupling reaction based on the acid-base synergistic activation [15b]: (1) regulating the Lewis acidity or basicity of metal catalyst to facilitate the ring-opening step; (2) modulating the steric configuration of metal complexes through close proximity to a substrate molecule; (3) introducing the “CO2 philic” groups into the framework of catalysts for reducing diffusion-dependent kinetic behavior so as to perform the reaction under low-pressure conditions; (4) improving their stability and the recovery performance of catalyst. Thus, we envision that the structure optimization of the catalysts can make the CO2-promoted cycloaddition reaction to carry out under ambient pressure and moderate temperature, thereby improving the acid-base coordinative capability between the intramolecular nucleophiles and metal center. In view of the above discussion and our previous studies, herein, a variety of innovative imidazolium based IL-functionalized Al(salen) derivatives were sequentially designed and synthesized successfully via covalent linkage at the 3,3′-position of salen ligand, as described in Scheme 2 [16–18]. The bifunctional Al(salen) complexes could show much higher catalytic activity than the binary catalytic systems in the cycloaddition reaction of CO2 and three-membered heterocyclic compounds owning to the improved synergistic catalysis from the intramolecular Lewis

   The Scheme 2        synthetic route of various bifunctional Al(salen) catalysts (color online).

 

   Sci China Chem   

Luo et al.

acidic site and nucleophiles. It was found that the catalytic performance of the catalysts might deeply depend on their structural features. More interestingly, the polyether-based Al(salen) catalyst can present the enhanced catalytic activity and excellent chemo- or regioselectivity in both the preparation of cyclic carbonates from terminal epoxides and the formation of 5-aryl-2-oxazolidinones from N-substituted aziridines under solvent-free and mild conditions. Moreover, the used catalysts can be easily regenerated and recovered by the addition of diethyl ether while retaining the original high activity on account of the concept of “one-phase catalysis and two-phase separation”. Finally, the kinetic experiments including the calculation of apparent activation energy suggested the importance of intramolecularly acid-base synergistic effect for the CO2-cycloaddition reaction.

2   Results and discussion 2.1   Synthesis of catalysts Recently, bifunctional catalysts that show high catalytic activity by using two or more catalytic groups had attracted much attention on the CO2 activation and catalytic conversion. It was well known that the category of bifunctional catalyst has been less developed as a probable result of the more synthetically demanding characteristics of catalyst preparation [19]. Since the metallosalen complexes were easy-to-synthesis, as well as could regulate conveniently the steric and electronic properties about the metal centers [13], a series of IL-functionalized Al(salen) complexes at the 3,3′-position of salen ligand were prepared successfully according to the previously reported synthetic method (Scheme 2) [14], which could very easily introduce bromide anion (nucleophilic ability: Br−>Cl−) into the framework of salen ligand [18,20]. Moreover, the dissociative bromide anion within ILs would coordinate to the active center of metallosalen skeleton forming a highly acitve six-coordinated complex [10b]. The flexible design approach might afford the possibility to easily modulate the environment around the metal center within these salen-type catalysts so as to the proximity and activation of substrates. Therefore, the problem of the restricted molecular conformation for the bifunctional catalsyt can be well resolved. 2.2   Catalytic performance in CO2/epoxide cycloaddition reaction The synthesis of five-membered cyclic carbonates via the 100% atom-economical cycloaddition of epoxides with CO2 was one of the most promising ways because cyclic carbonate products were widely used as aprotic high-boiling polar solvents, electrolytes for lithium-ion batteries, precursors of polymeric materials, and fine chemical intermediates [21]. On the basis of our previous reports and better understand the

3

present reaction process, a series of large-scale synthesis of styrene cyclic carbonate (SC) from the solvent-free cycloaddition reaction of styrene oxide (SO) with CO2 were carried out using Mettler Toledo EasyMax™102 system (Mettlter-Toledo, Switzerland) equipped with a 100 mL stainless steel autoclave in a semi-batch operation [14,17,19,22]. In the meantime, the course of reaction was also monitored by in situ Fourier transform infrared spectroscopy (FT-IR) using a Mettler Toledo React IR ic10 reaction analysis system (Switzerland). Initial screening experiments gave very promising results and demonstrated our conjecture. Obviously, a negligible SC yield was obtained when tranditional Al(salen) complex (chloro[N,N′-ethylenebis(salicylideneiminato)]aluminum(III), SA) was used as the sole catalyst even after 48 h at 80 °C and 2.0 MPa (Table 1, entry 1). Similarly, the 0.2 mol% catalyst [BMIm]Br exhibited low catalytic activity in the absence of catalyst SA (Table 1, entry 2). For comparison, considering the structure of the bifunctional Al(salen) derivatives ISA3 that involves two imidazolium-IL moieties and only one metal center within one structure, thus, the binary catalytic system composed of SA and [BMIm]Br with the molar ratio of 1:2 afforded a 70% yield of SC within 8 h under identical conditions (Table 1, entry 3). More interestingly, the experimental result was significantly less active than bifunctional catalyst ISA3, which could obtain up to 91% yield of SC at 0.1 mol% low catalyst loading (Table 1, entry 7). Additionally, from the in situ FT-IR spectroscopy as illustrated in Figure 1, the intensity of the characteristic signals for SO at 876 and 815 cm−1 declined rapidly, which were assigned to epoxy ring skeleton vibration. While, an increasing bands appeared at 1159 and 1065 cm−1 which were attributed to C=O bending vibration, and the peak of C=O stretch vibration shifted from 1810 to 1790 cm−1 indicated that specific intermolecular interactions in the reaction mixture change [19,23]. It also indicated the formation of SC as the sole product with rather excellent selectivity (99%). There was no other product such as polycarbonate was confirmed by FT-IR, nuclear magnetic resonance of hydrogen and carbon (1H and 13C NMR) and GC-MS. Hence, the single-component bifunctional catalyst could maintain the high catalytic activity with excellent selectivity in large-scale production. For the CO2/epoxide coupling, it was well known that activation of epoxide and subsequent epoxide ring opening were vital steps. In general, the cooperative action (or dual activation) of an anion X− (nucleophilic) and a metal center M (Lewis acid) could promote the ring-opening of epoxide. Hence, in the next moment, it was necessary to further investigate the main influence factors of intramolecularly synergistic effect in the one-component catalytic system. Firstly, the catalytic activity of catalysts was also closely related to the metal active center since the coordination of the epoxide to the metal center was well known  to be one of the  key steps

 

4

   Sci China Chem   

Luo et al.

Table 1   The results of the cycloaddition reaction of SO and CO2 with various catalysts a)

Entry d)

Catalyst (mol%)

Conv. (%)

b)

Yield (%)

TOF (h−1)

SA (0.5)

−

2

[BMIm]Br (0.2)

<10

9

5.6

3

SA (0.1)+[BMIm]Br (0.2)

71

70

116.7

4

ISA1 (0.1)

36

35

58.3

5

ISA2 (0.1)

21

20

33.3

1

−

b)

−

6

ISA3 (0.1)

92

91

151.7

7

ISA4 (0.1)

16

15

25

8

ISA5 (0.1)

33

32

53.3

9

ISA6 (0.1)

98

97

161.7

e)

ISA2 (0.1)

98

97

32.3

11d),e)

ISA-2 (0.1)

40

40

13.3

10

c)

a) Reaction conditions: a Mettler Toledo EasyMax 102 system equipped with the 100 mL stainless steel autoclave in a semi-batch operation, SO (40 g), catalyst (0.1 mol%), reaction time (6 h), reaction temperature (80 °C), CO2 pressure (2.0 MPa), solventless; b) determined by the in situ IR spectroscopy or GC, the selectivity for the formation of SC was found to be >99%, which was identified via IR, 1H NMR, 13C NMR and GC-MS; c) turnover frequency (TOF): mole of synthesized SC per mole of catalyst per hour; d) the reaction conditions were reported in our published work [14]; e) the reaction time (30 h).

   (a) Figure 1        Three-dimensional stack plot of IR spectra collected every 1 min during the coupling reaction of SO (330 mmol) and CO2 with catalyst ISA3 (0.1 mol%) at 80 °C (Tj) and 2.0 MPa CO2 pressure under solvent-free condition; (b) decline in the intensity of the characteristic peak for SO at 815 cm−1 and increase in the intensity of the characteristic peak for SC at 1800 cm−1 (color online).

in the CO2/epoxide coupling reaction. A substantial number of literatures had been published that catalysts containing a zinc cation (Zn2+) or cobalt cation (Co2+) were the most attractive for CO2 coupling with epoxide due to their high activity [24]. However, comparing to the metal aluminum as a nontoxic, readily available and environmental benign metal, these catalysts could either lead to the low catalytic activity or cause the colorization of the product mixtures (Table 1, entries 7−8 vs. entry 6). The observation could also better explain the ring opening of activated epoxide was involved in the rate-determining step in the cycloaddition reaction. Next, it was well known that the catalytic activity was closely related to the type of halide anion of the ILs within bifunctional catalysts. Generally, the nucleophilic attack from halogen anion was the rate-determining step in the ring open-

ing of the substrate, resulting in the activation of CO2. Bromide anion (Br−) with appropriate leaving group ability and nucleophilicity was the better choice relative to chloride anion (Cl−), thereby explaining the greater catalytic activity obtained with the Br− ligand [25]. For example, using chloride-based catalyst ISA2 resulted in 20% yield of SC within 6 h, which was significantly lower than that achieved by using bromide-based catalyst ISA3 (92%) under the identical conditions (Table 1, entry 5 vs. entry 6). Additionally, the catalyst functionalized at the two sides of 3,3′-position of salen backbone ISA2 was superior to the analogue functionalized at the 5,5′-position ISA-2 [14]. From the kinetic investigation, the >97% yield of SC was obtained using ISA2 (0.1 mol%) as catalyst within 30 h at 80 °C and 2.0 MPa, while only a 40% SC yield was achieved under the identical conditions when

 

   Sci China Chem   

5

Luo et al.

the analogue bearing the same imidazolium-IL moieties in the 5,5′-position was employed (Table 1, entry 10 vs. entry 11). The enhanced reaction rate might be derived from the steric configuration (or rigidity) of catalyst because the proper distance between metal center and halogen anion was crucial to the proximity and dual activation of epoxide. Moreover, the experimental results also indicated that the catalytic activity of the bifunctional catalytic system strongly depended on the chain length of the substituent at the N-position of the imidazolium ring. The SC yield increased with the length of the alkyl chain from methyl to n-octyl in the framework of imidazolium-IL units (Table 1, entry 4 vs. entry 6). This trend was mainly attributed to the better solubility of ISA3 in epoxide substrate and the weak electrostatic interaction originated from the steric effect of large groups. In order to further enhance the activity of the bifunctional catalyst, then, in view of the “CO2-expansion” effect of polyether units reported by our previous works [14], we also designed and synthesized the trifunctional Al(salen) complexes containing a Lewis acidic metal center and a polyether-based imidazolium-IL units anchored on the salen ligand at the two sides of 3,3′-position (denoted as ISA6, using polyether units instead of n-octyl groups), which could be used as highly efficient, easy-to-handle and homogeneous catalyst for the formation of cyclic carbonates from epoxides and CO2. As anticipated, the trifunctional catalyst ISA6 with only 0.1 mol% loading exhibited the extraordinary high catalytic activity (96%) at a very high reaction rate within 5 h under conditions of 80 °C and 2.0 MPa, which was better than that of polyether-free bifunctional catalyst ISA3 under the identical conditions (85%). A turnover frequency (TOF) value of up to 343 h−1 for the inert terminal epoxide SO was achieved by catalyst ISA6. Of course, the TOF value increased to a certain degree following the increased temperature and CO2 pressure or with a further decrease of catalyst loading (Table S1, entries 1–6, Supporting Information online). It was thus clear that the pressure had a great effect on the yield of SC upon varying the CO2 pressure in the range 0.1–3.0 MPa, whereas the yield changed only slightly by increasing the CO2 pressure from 3.0 MPa to 4.0 MPa. A significant drawback of using CO2 as a reagent in organic synthesis was the potential dangers associated with high-pressure operations [19]. Surprisingly, the coupling reaction between SO and CO2 could proceed efficiently even at atmospheric CO2 pressure (0.1 MPa) in spite of prolonging the reaction time to 40 h at 80 °C with only 0.1 mol% ISA6, resulting in the excellent yield of SC (Table S1, entry 7). If using 1.0 mol% ISA6 as catalyst, a 75% yield of SC was also obtained within 24 h at 50 °C (Table S1, entry 8). Notably, the trifunctional catalyst ISA6 was capable of catalyzing the insertion of CO2 into SO at atmospheric pressure and at ambient temperature ascribed to the “CO2 capture” capability of polyether chains. In consideration of the interesting

experimental results, it was deduced that these trifunctional catalyst can improve the local concentration of CO2 around the catalytic active centers so as to perform the cycloaddition reaction at ambient conditions, which originated from the weak interaction between electron-donating oxygen functional groups within polyether chain and the electrophilic carbon atom of CO2. This was indirect proof that the special feature of trifunctional catalsyt would stand a good chance to capture and activate CO2. Rate = k obs[SO]a

(1)

Rate = d[SO] / dt

(2)

ln[SO] = k obst

(3)

Rate = k obs[SO]1

(4)

k obs = Aexp( Ea / RT )

(5)

lnk obs = lnA Ea / RT

(6)

where Rate is the reaction rate; kobs is the pseudo-first order rate constant; [SO] is the SO concentration; a is the reaction order; t is the reaction time; R is the ideal gas constant (8.314 J mol−1 K−1); T is the absolute temperature (K); A is the preexponential factor; Ea is the apparent activation energy (kJ mol−1). To further explore the reason for the enhanced activity, a series of kinetic investigation of this reaction was carried out with various catalysts (such as ISA2, ISA3 and ISA6) using propylene carbonate (PC) as solvent, which might illustrate the activation energies of styrene cyclic carbonate formation. Taking catalyst ISA6 as an example, firstly, it can be assumed that the concentrations of the catalyst ISA6 and CO2 molecular were constant under the selected conditions due to the bifunctional catalyst owing high stability and CO2 was present in large excess (Eq. (1)). Thus, from Eqs. (2), (3), the observed rate constant (kobs) was determined from the slope of a linear plot of the natural logarithm of the changing SO concentration ([SO]) with time (t). The results of in situ IR spectra showed a good fit to first order kinetics, implying that the reactions were first order with respect to SO concentration, i.e. the reaction order (a) was equal to 1 (Eq. (4)). In turn, a fitting the data from a plot of the natural logarithm of the observed pseudo-first order rate constant (lnkobs) against the reciprocal absolute temperature (1/T) in the different temperature range. The apparent activation energy (Ea) for the CO2 insertion reaction into SO catalyzed by catalyst ISA6 was determined over the range 40–70 °C by using the Arrhenius equation based on the relationship between the observed rate constants and the temperature (Eqs. (5), (6)), and the results were shown in Figure S3 in the Supporting Information online material. The value of the pre-exponential factor A and activated energy Ea were calculated to be 4.42×105

 

6

   Sci China Chem   

Luo et al.

s−1 and 63.01 kJ mol−1, respectively. Thus, the kinetic equation on the synthesis of SC with catalyst ISA6 was obtained at Rate=−d[SO]/dt=4.42×105e−63.0/RT[SO]. Nevertheless, for the homogeneous catalyst [BMIm]Br alone, this value (Ea) increased to 80.74 kJ mol−1 reported by our previous work [22]. From the observation of the data, it could be speculated that metal active center may act by reducing the activation energy value by 17.73 kJ mol−1, thereby activating the epoxide effectively via acid-base pairing interaction. Similarly, as described in Figure 2, the activation energy with the catalyst ISA2 and ISA3 derived from the studies were equal to 76.56 and 62.66 kJ mol−1, respectively. Especially, the Ea value (62.66 kJ mol−1) obtained by catalyst ISA3 was very close to that of ISA6 (63.01 kJ mol−1), which indirectly verified that the introduction of polyether chains into the bifunctional catalyst could only improve the local concentration of CO2 around catalytic active center (physical properties), but could not activate CO2 molecular directly (chemical properties). In addition, the Ea value with catalyst ISA2 (76.56 kJ mol−1) was far greater than that of catalyst ISA3 (62.66 kJ mol−1), further indicating that the bromide anion as a nucleophile could reduce the reaction activation energy with respect to chloride anion

due to its nucleophilic ability. In brief, it can be concluded that from an energetic standpoint the cycloaddition reaction can carry out at ambient conditions and the essential data have a guiding significance for industrial synthesis of cyclic carbonates from CO2 and epoxides. 2.3   Catalytic performance in CO2/aziridine cycloaddition reaction The cycloaddition reaction of CO2 and aziridine was also a commercial routes using CO2 as C1 resource to afford fivemembered heterocyclic compound, and oxazolidinones had widely used as intermediates, chiral auxiliaries, pharmaceutical chemistry and antibiotic agents against various Gram-positive bacteria [26]. Hence, in the light of our efforts to further expand the scope of these unique bifunctional Al(salen) catalysts, initial studies were conducted for the regioselective synthesis of 5-aryl-2-oxazolidinone from CO2 and 1-n-butyl2-phenylaziridine as a representative reaction with a 10 mL stainless steel autoclave in a semi-batch operation [27]. At the first place, the reaction did not occur without any catalysts  (Table 2, entry 1).  Similarly,  either individual  SA or

   The Figure 2        calculated results of apparent activate energy (color online).

Table 2   The results of the cycloaddition reaction of 1-n-butyl-2-phenylaziridine and CO2 a)

Entry

Catalyst

t (h)

T (°C)

1 2 3 e) 4 f) 5 6 7 8 9 10

− SA [BMIm]Br SA+[BMIm]Br ISA3 ISA6 ISA6 ISA6 ISA6 ISA6

2 2 2 2 2 2 1.5 1 2 2

50 50 50 50 50 50 50 50 40 30

Conv. (%) b) 10 22 16 35 89 99 88 38 79 37

Yield (%) <5 17 12 28 86 95 85 35 75 32

b)

Sel. (%) 95:5 97:3 97:3 97:3 98:2 98:2 98:2 97:3 97:3 97:3

c)

TOF (h−1)

d)

− 85 60 − 430 475 567 350 375 160

a) Reaction conditions: the 10 mL stainless steel autoclave in a semi-batch operation; 1-n-butyl-2-phenylaziridine (1 mmol); catalyst (1.0 mol%); reaction time (2 h); reaction temperature (50 °C); CO2 pressure (1.0 MPa); solventless; b) determined by GC with biphenyl as an internal standard, and the structure of product was identified via 1H NMR, 13C NMR and GC-MS; c) the molar ratio of 5-position to 4-position; d) same as Table 1; e) [BMIm]Br (2.0 mol%); f) SA (1.0 mol%) and [BMIm]Br (2.0 mol%).

 

   Sci China Chem   

Luo et al.

[BMIm]Br alone could not catalyze the solvent-free cycloaddition effectively for 2 h at 50 °C and 1.0 MPa CO2 pressure, which was acquired at a low product yield (Table 2, entries 2 and 3). As expected, polyether-based trifunctional catalyst ISA4 afforded the highest yield of 95%, and was found to exhibit up to 98% regioselectivity for the 5-aryl-2-oxazolidinones as the desired products with a catalyst loading of 1.0 mol% in the absence of any cocatalyst (Table 2, entry 6). Thereafter, the decreased yield (ca. 86%) was achieved with the polyether-free bifunctional catalyst ISA3 under the described conditions (Table 2, entry 5). To our delight, the observed phenomenon was also demonstrating the significant importance of acid-base bifunctionality within catalyst in comparison with previously discussed on the CO2/epoxide coupling reaction, synergistically leading to the almost quantitative conversion (99%), yield (95%), and regioselectivity (98:2) under relatively milder conditions. As anticipated, while shortening the reaction time or dropping the reaction temperature, the desired yield decreased drastically. A shortening of the reaction time from 2 to 1.5 h maintained the yield of 85%; however, a further decrease in the reaction time to 1 h lowered the yield to 35% (Table 2, entries 7 and 8). Additionally, upon lowering the temperature to 40 °C, a moderate yield of 79% (Table 2, entry 9) was obtained. Certainly, further decrease of the temperature from 40 to 30 °C led to an unsatisfactory yield of 37% (Table 2, entry 10). Thus, the present bifunctional catalytic system could catalyze the cycloaddition reaction from the three-membered heterocyclic compounds and CO2 to produce the corresponding five-membered heterocyclic compounds in excellent yields and good selectivity under relatively milder conditions. Additionally, a relatively long induction period (>10 h) appeared in the initial reaction stage at ambient condition (0.1 MPa), while this period disappeared completely with increasing the CO2 pressure to 1.0 MPa at the same temperature, which was elaborated in Figure 3. The unexpected phenomenon hinted the concentration of CO2 had a significant impact on the activation of substrate. Hence, it can be suspected that the nitrogen atoms within aziridine having additional Brønsted basic site could directly lead to activate CO2 molecular or present “CO2 capture” capability due to the weak interaction, which was likely to be one of the rate-determining steps and enhanced the catalytic activity of cycloaddition reaction. In other words, the activated CO2 by organic base (i.e. substrate molecular) exhibited an enhanced electrophilicity which made it much easier to insert into the Al–N bond of the intermediate.

7

   Dependence Figure 3        of the conversion on reaction time using 1-n-butyl2-phenylaziridine as substrate. Reaction conditions: substrate (1.0 mmol), temperaure (50 °C), ISA4 (1.0 mol%), CO2 (1.0 MPa for A, and 0.1 MPa for B), soventless (color online).

reaction was carried out at 50 °C and 1.0 MPa CO2 pressure, while the reaction from 1-n-butyl-2-phenylaziridine and CO2 was also performed under identical conditions. Both used 1.0 mol% ISA6 as catalyst and the same operational approach. The catalyst was easily recovered by adding a certain amount of diethyl ether in the end of the cycloaddition reaction. From the Figure 4, the catalyst ISA6 could be reused for eight successive runs without any significant loss in its catalytic activity, which was also maintained at a high level, and the chemo- and regioselectivity still remained at 98%. Nevertheless, the obvious decrease of activity obtained by the regenerated polyether-free catalyst ISA3 may be due to the drastically loss of catalyst during the recovery process. Therefore, the introduction of polyether chain into the bifunctional catalyst could improve the recyclability of catalyst derived from immiscibility of PEG in diethyl ether; potentially endow the

2.4   Recycling experiment It was worth noting that the reusability and stability of a catalyst system were the two key factors that identify whether it found potentially practical application in industry. Accordingly, to test catalyst reusability,  the SO/CO2  cycloaddition

   The Figure 4        results of recyclability and reusability of catalyst ISA6 in the coupling reaction of three-membered heterocyclic compounds to CO2. Reaction conditions: 10 mL stainless steel autoclave, ISA6 (1.0 mol%), SO (1 mmol) or 1-n-butyl-2-phenylaziridine (1 mmol), reaction temperature (50 °C), CO2 pressure (1.0 MPa) (color online).

8

 

   Sci China Chem   

Luo et al.

novel complexes with the feature of solvent-regulated separation. Also, various characteristic techniques could demonstrate the catalyst ISA6 has the excellent chemical- and thermostability in the catalytic cycle. For example, the results of FT-IR spectroscopy between fresh ISA6 and recovered ISA6 after eight cyclic runs of cycloaddition reaction were described carefully. It might be revealing the basical similarity in intensities of peaks at 1635 and 1099 cm−1, which were associated with the stretching vibration modes of C=N and C−O−C in the framework of salen ligand, respectively. The 27 Al NMR also confirmed that the recovered catalyst ISA6 presented here does not change the structure of catalytic active sites. Thermogravimetric (TG) analysis results proved that the catalyst ISA6 with high thermo-stability could endure about 215 °C with little loss of its weight ascribed to its stable skeletal composition. The non-removable residue belonged to the formation of aluminum oxide in air atmosphere at high temperature. Thus, the polyether-based trifunctional Al(salen) could reflect the excellent advantages on the catalyst recyclability and catalytic activity compared to the reported work, such as the bifunctional M(salen) catalysts bearing quaternary ammonium salts [10b,10d] or quaternary phosphonium salts [28], which generally brought about vari-

ous problems including the leaching loss of catalyst in recyclability and the relatively lower catalytic activity. 2.5   Substrate scope To evaluate the substrate scope and generality of the developed catalytic system, both terminal epoxides and N-substituted aziridines were used as the substrate for the CO2-cycloaddition reaction system using 1.0 mol% ISA6 as a catalyst at 50 °C and 1.0 MPa under solvent-free conditions, and the results were summarized in Figure 5. Obviously, all of investigated terminal epoxides (1a–1k) could be smoothly converted to the corresponding cyclic carbonates with good-to-excellent yields and extremely excellent chemoselectivities (>98%) [2c]. Among these, the catalytic activity of aliphatic epoxides decreased as the chain length of electron-donating alkyl group increased due to its bulky steric hindrance and apolar nature which might adversely affect overall kinetics or solubility (1a>1b>1c>1d). Moreover, the electron-withdrawing nature of the chloromethyl or bromomethyl group within aliphatic epoxides could be driven the cycloaddition reaction favorably (1e>1f). Notably, the catalytic activity decreased generally for chloro-substituted aromatic SO derivatives with

   The Figure 5        results of synthesis of cyclic carbonates from various terminal epoxides and formation of 5-substituted 2-oxazolidinones from N-substituted aziridines using CO2 as a chemical feedstock. Reaction conditions: catalyst (1.0 mol% ISA6), CO2 pressure (1.0 MPa), reaction temperature (50 °C) (color online).

 

   Sci China Chem   

Luo et al.

electron-withdrawing substituents on the para-position of aromatic ring (1h<1g). Additionally, the trifunctional catalyst ISA6 could also obtained excellent yield and exhibited a certain degree of tolerance to various functionalities, including olefin group, ester group, ether group, halogen, etc. (1i and 1j). Unfortunately, it should be noted that cyclohexene oxide (internal epoxide, 1k) presented lower activity with a certain amount of generated polycarbonate. Steric congestion around the disubstituted epoxide ring tends to cause a decrease in conversion rate. Therefore, it can be concluded that the reaction was favorable at the less hindered side of the epoxide, and both steric and electronic effects played an important role in the ISA6-catalyzed CO2/epoxide coupling reaction [2c]. For N-substituted aziridines, the substrates bearing linear alkyl groups at the nitrogen atom gave the corresponding 2-oxazolidinones in good-to-excellent yields all with above 95% regioselectivities under identical conditions. Notably, the catalytic activity and regioselectivities decreased obviously following the decreased length of alkyl groups at the nitrogen atom (n-propyl>n-butyl>n-pentyl: 2a>2b>2c or 2e>2f>2g) [29]. However, the substrates with a branched alkyl group at the nitrogen atom showed slow reaction rate probably due to the steric effect, and the catalytic activity decreased in the following order: iso-pentyl2a or 2f>2b or 2g>2c) [27e]. As a consequence, it was deduced that the activity and regioselectivity for N-substituted aziridines might greatly depend on the electron nature of the phenyl group and the steric hindrance of the substituent group at the nitrogen atom.

9

2.6   Mechanism consideration On the basis of the current experimental results and the features of the new catalysts, a plausible reaction mechanism including two reaction paths for the CO2-based cycloaddition catalyzed by bifunctional Al(salen) catalyst had some differences between epoxide and aziridine although both were high active three-membered heterocyclic compounds, and the reaction process were described in Scheme 3 [2c]. In Path A for SO, the initial coordination of the epoxide to an aluminum cation within bifunctional catalyst was followed by ring-opening process derived from intramolecular nucleophilic attacking from the bromide anion, which was the rate-determining step. The flexible configuration of catalyst could cause in close proximity to facilitate the ring-opening step. Simultaneously, CO2 molecular activated by ILs moieties was inserted expediently into the generated active Al−O bond, thereby leading to form metal carbonate species. Finally, an intramolecular cyclization reaction could lead to the formation of the cyclic carbonate and regeneration of the catalyst. Thus, these bifunctional catalysts can make the ring-opening procedure less energetically demanding and the subsequent CO2 activation easier. However, in Path B for N-n-butyl-2-phenylaziridine, initially, the electrophilic carbon atom of a CO2 molecular might interact with the nitrogen of aziridines to form a zwitterionic adduct of base-CO2 complex. Subsequently, Al-based catalyst owning built-in a nucleophile could promote the ring-opening step of the activated aziridine. Accordingly, the well-designed bifunctional Al(salen) catalysts can fulfill the role of cocatalyst: activate CO2 and activate aziridine simultaneously. In addition, from the reported work of Lu’s group [10d], the results suggested that the nucleophilic ring-opening occurred mainly at the β-

   Possible Scheme 3        mechanism for ISA6-catalyzed CO2 cycloaddition reaction (color online).

10

 

   Sci China Chem   

Luo et al.

carbon atom (less substituted carbon) relative to the α-carbon atom (most substituted carbon) for styrene oxide bearing an electron-withdrawing group, which was different from the aromatic aziridines under identical conditions. Indeed, the ring-opening of aromatic aziridines more easily occurs at the methine carbon to give the 5-aryl-2-oxazolidinones. On the other hand, since the IL-functionalized Al(salen) catalysts (our work) were similar to the Al(salen) catalysts bearing quaternary ammonium salts (Lu’s work) on the catalyst structure, we speculated the ring-opening at the methylene C−O bond for styrene oxide was also achieved in our catalytic system. Thus, when (R)-(+)-styrene oxide was employed as an enantiopure substrate, the nearly 100% yield for (R)-(+)-styrene cyclic carbonate was obtained with the configuration retention. The results were in accordance with our hypothesis. Thanks to the acid-base synergistic effect to activate both the substrate and CO2, the cycloaddition reaction could proceed under atmospheric CO2 pressure. Thus, this experimental evidence indicated a monometallic reaction mechanism for the bifunctional catalytic system.

3   Conclusions In summary, unique acid-base bifunctional Al(salen) complexes have been successfully prepared from the introduction of imidazolium-based IL units in the framework of the salen ligand at the two sides of 3,3′-position. These experimental results confirm that the structure of the catalyst could affect its catalytic activity. An intramolecularly synergistic effect or the improved double activation method could strongly encourage the metal-catalyzed CO2-coupling reaction to perform at ambient conditions. The polyether-based trifunctional Al(salen) complex could present the enhanced catalytic activity with excellent yield and chemo- or regioselectivity in comparison with polyether-free bifunctional catalyst or the similar binary catalytic system for the conversion of terminal epoxides into the corresponding cyclic carbonates and the formation of 5-aryl-2-oxazolidinones from a wide variety of N-substituted aziridines. Moreover, these catalysts could be regenerated and reused at least eight times without any obvious decreases in catalytic activity. Overall, it is still worth exploring the novel and extraordinary promising metallosalen catalyst in CO2 fixation because they are facile to design and functionalize on the salen skeleton, and further studies are underway in our laboratory.     This Acknowledgments          work was supported by the National Science for Distinguished Young Scholars of China (21425627), the National Natural Science Foundation of China (21676306), the Natural Science Foundation of Guangdong Province (2016A030310211, 2015A030313104) and the Fundamental Research Funds for the Central Universities of Sun Yat-sen University.

    The authors declare that they have no conflict of Conflict of interest      interest.     The Supporting information          supporting information is available online at http://chem.scichina.com and http://link.springer.com/journal/11426. The supporting materials are published as submitted, without typesetting or editing. The responsibility for scientific accuracy and content remains entirely with the authors. 1

2

3

4

5

6

7

8

9

10

(a) Sakakura T, Choi JC, Yasuda H. Chem Rev, 2007, 107: 2365–2387; (b) Mikkelsen M, Jørgensen M, Krebs FC. Energy Environ Sci, 2010, 3: 43–81; (c) Omae I. Coordin Chem Rev, 2012, 256: 1384–1405; (d) Maeda C, Miyazaki Y, Ema T. Catal Sci Technol, 2014, 4: 1482–1497; (e) Liu Q, Wu LP, Jackstell R, Beller M. Nat Commun, 2015, 6: 5933–5947 (a) Sakakura T, Kohno K. Chem Commun, 2009, 1312–1330; (b) Decortes A, Castilla AM, Kleij AW. Angew Chem Int Ed, 2010, 49: 9822–9837; (c) Pescarmona PP, Taherimehr M. Catal Sci Technol, 2012, 2: 2169–2187; (d) He Q, O’Brien JW, Kitselman KA, Tompkins LE, Curtis GCT, Kerton FM. Catal Sci Technol, 2014, 4: 1513–1528; (e) Fiorani G, Guo W, Kleij AW. Green Chem, 2015, 17: 1375–1389; (f) Bobbink FD, Dyson PJ. J Catal, 2016, 343: 52–61 (a) Yang ZZ, He LN, Gao J, Liu AH, Yu B. Energy Environ Sci, 2012, 5: 6602–6639; (b) Zhao YN, Yang ZZ, Luo SH, He LN. Catal Today, 2013, 200: 2–8 (a) Adeleye AI, Kellici S, Heil T, Morgan D, Vickers M, Saha B. Catal Today, 2015, 256: 347–357; (b) Chen A, Zhao T, Gao H, Chen L, Chen J, Yu Y. Catal Sci Technol, 2016, 6: 780–790 (a) Dai W, Jin B, Luo S, Luo X, Tu X, An CT. Appl Catal A-Gen, 2014, 470: 183–188; (b) Alves M, Grignard B, Gennen S, Mereau R, Detrembleur C, Jerome C, Tassaing T. Catal Sci Technol, 2015, 5: 4636–4643; (c) Barthel A, Saih Y, Gimenez M, Pelletier JDA, Kühn FE, D’Elia V, Basset JM. Green Chem, 2016, 18: 3116–3123; (d) Zhou Y, Hu S, Ma X, Liang S, Jiang T, Han B. J Mol Catal A-Chem, 2008, 284: 52–57 (a) Chen W, Zhong L, Peng X, Sun R, Lu F. ACS Sustain Chem Eng, 2015, 3: 147–152; (b) Song J, Zhang B, Zhang P, Ma J, Liu J, Fan H, Jiang T, Han B. Catal Today, 2012, 183: 130–135; (c) Yang Z, Sun J, Cheng W, Wang J, Li Q, Zhang S. Catal Commun, 2014, 44: 6–9 (a) Palgunadi J, Kwon OS, Lee H, Bae JY, Ahn BS, Min NY, Kim HS. Catal Today, 2004, 98: 511–514; (b) Zhang S, Chen Y, Li F, Lu X, Dai W, Mori R. Catal Today, 2006, 115: 61–69; (c) Wilhelm ME, Anthofer MH, Reich RM, D’Elia V, Basset JM, Herrmann WA, Cokoja M, Kühn FE. Catal Sci Technol, 2014, 4: 1638–1643; (d) Liu M, Liang L, Li X, Gao X, Sun J. Green Chem, 2016, 18: 2851–2863; (e) Liu M, Li X, Lin X, Liang L, Gao X, Sun J. J Mol Catal A-Chem, 2016, 412: 20–26; (f) Liu M, Liang L, Liang T, Lin X, Shi L, Wang F, Sun J. J Mol Catal A-Chem, 2015, 408: 242–249 (a) Jing H, Nguyen SBT. J Mol Catal A-Chem, 2007, 261: 12–15; (b) Wang L, Lin L, Zhang G, Kodama K, Yasutake M, Hirose T. Chem Commun, 2014, 50: 14813–14816; (c) Chen A, Chen C, Xiu Y, Liu X, Chen J, Guo L, Zhang R, Hou Z. Green Chem, 2015, 17: 1842–1852; (d) Wang L, Kodama K, Hirose T. Catal Sci Technol, 2016, 6: 3872–3877 (a) Ema T, Miyazaki Y, Koyama S, Yano Y, Sakai T. Chem Commun, 2012, 48: 4489–4491; (b) Qin Y, Guo H, Sheng X, Wang X, Wang F. Green Chem, 2015, 17: 2853–2858; (c) Maeda C, Taniguchi T, Ogawa K, Ema T. Angew Chem Int Ed, 2015, 54: 134–138; (d) Jiang X, Gou F, Chen F, Jing H. Green Chem, 2016, 18: 3567–3576 (a) North M, Pasquale R. Angew Chem Int Ed, 2009, 48: 2946–2948;

 

   Sci China Chem   

Luo et al.

11

12

13 14 15

16 17 18

(b) Tian D, Liu B, Gan Q, Li H, Darensbourg DJ. ACS Catal, 2012, 2: 2029–2035; (c) Supasitmongkol S, Styring P. Catal Sci Technol, 2014, 4: 1622–1630; (d) Ren WM, Liu Y, Lu XB. J Org Chem, 2014, 79: 9771–9777; (e) Castro-Osma JA, North M, Wu X. Chem Eur J, 2016, 22: 2100–2107; (f) Rulev YA, Larionov VA, Lokutova AV, Moskalenko MA, Lependina OL, Maleev VI, North M, Belokon YN. ChemSusChem, 2016, 9: 216–222; (g) He S, Wang F, Tong WL, Yiu SM, Chan MCW. Chem Commun, 2016, 52: 1017–1020 (a) Saptal VB, Sasaki T, Harada K, Nishio-Hamane D, Bhanage BM. ChemSusChem, 2016, 9: 644–650; (b) Li Y, Zou B, Hu C, Cao M. Carbon, 2016, 99: 79–89; (c) Buyukcakir O, Je SH, Choi DS, Talapaneni SN, Seo Y, Jung Y, Polychronopoulou K, Coskun A. Chem Commun, 2016, 52: 934–937; (d) Babu R, Kathalikkattil AC, Roshan R, Tharun J, Kim DW, Park DW. Green Chem, 2016, 18: 232–242 (a) Miller AW, Nguyen SBT. Org Lett, 2004, 6: 2301–2304; (b) Lu XB, Zhang YJ, Jin K, Luo LM, Wang H. J Catal, 2004, 227: 537–541; (c) Paddock RL, Nguyen SBT. Chem Commun, 2004, 14: 1622–1623; (d) Decortes A, Martínez Belmonte M, Benet-Buchholz J, Kleij AW. Chem Commun, 2010, 46: 4580–4582; (e) Adhikari D, Miller AW, Baik MH, Nguyen SBT. Chem Sci, 2015, 6: 1293–1300 Ren Y, Chen J, Qi C, Jiang H. ChemCatChem, 2015, 7: 1535–1538 Luo R, Zhou X, Chen S, Li Y, Zhou L, Ji H. Green Chem, 2014, 16: 1496–1506 (a) Martín C, Whiteoak CJ, Martin E, Martínez Belmonte M, Escudero-Adán EC, Kleij AW. Catal Sci Technol, 2014, 4: 1615–1621; (b) Rintjema J, Peña Carrodeguas L, Laserna V, Sopeña S, Kleij AW. Top Organomet Chem, 2015, 53: 39–71 Du Y, Wu Y, Liu AH, He LN. J Org Chem, 2008, 73: 4709–4712 Luo R, Tan R, Peng Z, Zheng W, Kong Y, Yin D. J Catal, 2012, 287: 170–177 Wang Q, Wilson C, Blake AJ, Collinson SR, Tasker PA, Schröder M.

19 20 21 22 23

24 25 26

27

28

29

11

Tetrahedron Lett, 2006, 47: 8983–8987 Luo R, Zhou X, Zhang W, Liang Z, Jiang J, Ji H. Green Chem, 2014, 16: 4179–4189 Yang Y, Zhang Y, Hao S, Kan Q. Chem Eng J, 2011, 171: 1356–1366 (a) Zhang H, Liu HB, Yue JM. Chem Rev, 2014, 114: 883–898; (b) Shaikh AAG, Sivaram S. Chem Rev, 1996, 96: 951–976 Luo R, Zhou X, Fang Y, Ji H. Carbon, 2015, 82: 1–11 (a) Darensbourg DJ, Horn Jr A, Moncada AI. Green Chem, 2010, 12: 1376–1379; (b) Foltran S, Alsarraf J, Robert F, Landais Y, Cloutet E, Cramail H, Tassaing T. Catal Sci Technol, 2013, 3: 1046–1055; (c) Werner T, Büttner H. ChemSusChem, 2014, 7: 3268–3271; (d) Elmas S, Subhani MA, Harrer M, Leitner W, Sundermeyer J, Müller TE. Catal Sci Technol, 2014, 4: 1652–1657; (e) Dutta B, Sofack-Kreutzer J, Ghani AA, D’Elia V, Pelletier JDA, Cokoja M, Kühn FE, Basset JM. Catal Sci Technol, 2014, 4: 1534–1538; (f) Liu X, Zhang S, Song QW, Liu XF, Ma R, He LN. Green Chem, 2016, 18: 2871–2876 Bai D, Duan S, Hai L, Jing H. ChemCatChem, 2012, 4: 1752–1758 Sun J, Han L, Cheng W, Wang J, Zhang X, Zhang S. ChemSusChem, 2011, 4: 502–507 (a) Kathalikkattil AC, Tharun J, Roshan R, Soek HG, Park DW. Appl Catal A-Gen, 2012, 447–448: 107–114; (b) Nale DB, Rana S, Parida K, Bhanage BM. Appl Catal A-Gen, 2014, 469: 340–349 (a) Yang ZZ, He LN, Peng SY, Liu AH. Green Chem, 2010, 12: 1850–1854; (b) Qi C, Ye J, Zeng W, Jiang H. Adv Synth Catal, 2010, 352: 1925–1933; (c) Yang ZZ, Li YN, Wei YY, He LN. Green Chem, 2011, 13: 2351–2353; (d) Lin XZ, Yang ZZ, He LN, Yuan ZY. Green Chem, 2015, 17: 795–798; (e) Ma TY, Qiao SZ. ACS Catal, 2014, 4: 3847–3855 (a) Tian D, Liu B, Zhang L, Wang X, Zhang W, Han L, Park DW. J Ind Eng Chem, 2012, 18: 1332–1338; (b) Miao CX, Wang JQ, Wu Y, Du Y, He LN. ChemSusChem, 2008, 1: 236–241 Kumar S, Jain SL. Ind Eng Chem Res, 2014, 53: 541–546