Alternating copolymerization of carbon dioxide and cyclohexene oxide catalyzed by salen Co^III(acetate) complexes

Substituted derivatives of the parent ligand salen were prepared by condensation of salicylaldehyde derivatives with ethylenediamine. Substituted salicylaldehydes were prepared via the Reimer–Tiemann reaction or through lithium aluminum hydride reduction followed by manganese dioxide oxidation of a substituted salicylic acid [37‐39]. Stirring the Schiff base ligands with cobalt^II acetate tetrahydrate under aerobic conditions afforded the requisite Co^III complexes [38]. The nature of the R₁, R₂, and R₃ substituents, the hydrocarbonated chain R, and their abbreviations are also shown in Fig. 1.

The product was sufficiently dried under reduced pressure. Complexes 1–6 were isolated as a solid powder. Numerous attempts to isolate complexes 1–5 in a mixture of methylene chloride and hexane or other solvents were proved to be unsuccessful. Fortunately, the structure of a closely related complex 6 could be elucidated by single crystal X-ray analysis. Crystal suitable for X-ray diffraction was grown from a mixture of methylene chloride and hexane at 20 °C. A single crystal was selected for X-ray analysis (Table 1). As shown in Fig. 2, the structure is indeed a six-coordinate octahedral as the structure of the salen Co^III(2,4-dinitrophenoxy) [35, 36]. X-ray diffraction data showed that the crystal was analogous monomeric structures. The overall geometry around the metal cobalt was a six-coordinate octahedral with an average compressed axial O(2)–Co(1)–O(3) bond angle of 167.20(9)°, and equatorial O(1)–Co(1)–N(1), N(1)–Co(1)–N(2), O(2)–Co(1)–O(4), and O(4)–Co(1)–O(1) bond angles of 91.22(9)°, 92.17(10)°, 100.69(9)°, and 86.57(9)°, respectively. The distances from the Co atom to O(1), O(2), O(3), O(4), N(1), and N(2) were 1.871(2), 1.871(2), 1.962(2), 1.988(2), 1.880(2), and 1.899(3) Å, respectively (Table 2). The average Co–N and Co–O bond lengths are in the same range as corresponding values in similar octahedral Co^III systems [40, 41]. The molecular structure of complex 6 is shown in Fig. 2.

Different Co^III complexes were evaluated as catalysts for the copolymerization of CHO and CO₂ in the presence of [PPN]Cl under various reaction conditions. These complexes exhibit different catalytic activities. As shown in Table 3, complex 1 bearing ligand with the R₁ being a donor electronic methyl group displays the highest catalytic activity (see entry 1, in Table 3). The reason that the catalytic activity depended on the structure of ligand should be attributed to that donor electronic methyl could stabilize the active salen Co^III species against decomposition to inactive salen Co^II by reversibly intramolecular Co–O bond formation and dissociation. When the R₂ or R₃ group is replaced by an iodine atom, complexes 3 and 5 present a moderate catalytic activity (see entries 3 and 5, in Table 3). Complex 6 bearing the sterically less bulky ligands shows the lowest activity among its analogues complexes 1–6. Under the same conditions, when the R₁ group is replaced by a hydrogen atom, complex 2 exhibits a lower catalytic activity (see entry 2, in Table 3). While diimine-bridges with the two carbon atoms are superseded by that with the three carbon atoms, the catalytic activity is dramatically decreased. The catalytic activity is influenced markedly by the nature of the substituents grafted on the ligands molecules and the length of the diimine bridges between the two nitrogen atoms in the ligands, respectively. These results might be mainly ascribed to the steric bulk effect of the substituents R on the phenolates of the ligand. This agrees well with the literature data [5]. These differences in ligands framework and the length of the diimine bridges between the two nitrogen atoms in the ligands also reflect on the selectivity formation of poly(cyclohexenylene carbonate) (PCHC) over cyclohexene carbonate (CHC) and the resulting copolymer number average molecular weights (M _n ). Complex 1 displays the highest PCHC/CHC selectivity and yields polymer with the highest M _n . However, complex 6 presents the lowest PCHC/CHC selectivity and produces polymer with the lowest M _n . Both the length of the diimine bridges and the substituents on the ligands do not have a clear effect on the microstructure of the resultant polymer. With complexes 1–6 the copolymerization afford products possessing >99 % carbonate linkages.

The mole ratio of [PPN]Cl to the complex 1 played an important role in the copolymerization of CHO with CO₂, as listed in Table 4 (entries 1–4). Although almost negligible yield was observed in the absence of cocatalyst (see entry 1 in Table 4), the addition of cocatalyst [PPN]Cl resulted in the production of the alternating copolymer (see entries 2–4, in Table 4). Optimization of the molar ratio between complex 1 and [PPN]Cl revealed that the highest yield was achieved using 1 equiv. of [PPN]Cl. However, the increase in the ratio of [PPN]Cl to complex 1 above 2 equiv. could increase the reaction activities. The CHC only was obtained. The CHC formation might be based on the presence of a depolymerization reaction that is promoted at higher [PPN]Cl concentrations. It is most likely due to a backbiting reaction starting from an alcoholate chain end that is displaced from the Co^III center in a coordination equilibrium with [PPN]Cl after epoxide ring opening [18].

The influence of reaction temperature on the copolymerization of CHO and CO₂ was investigated using complex 1/[PPN]Cl (see entries 3 and 5–7, in Table 4). Dramatic increase in catalytic activity with temperature was observed, and the turnover frequency (TOF) at 80 °C (48.7 h⁻¹) almost doubles that at 60 °C (25.6 h⁻¹) (see entries 3 and 5, in Table 4), which reaches to 52.2 h⁻¹ when the copolymerization is performed at 100 °C (see entry 6, in Table 4). The highest catalytic activity (54.5 h⁻¹) is achieved at 120 °C (see Entry 7, in Table 4). While the yield of copolymer increases with increasing temperature up to about 80 °C, and then decreases with the further increase. Similar results have been observed previously [33]. In agreement with the increase in yield, M _n of the obtained copolymer increases correspondingly from 21.6 kg mol⁻¹ at 60 °C to 38.7 kg mol⁻¹ at 80 °C. In the meantime, polymer dispersity index remains narrow and almost constant (PDI = 1.2). It is generally known that the homopolymerization of CHO easily takes place at elevated temperatures and is normally observed at catalyst systems with regard to much-studied zinc complexes [10]. Moreover, the polycarbonates obtained at the range of 60–120 °C all have >99 % carbonate linkages.

Obviously, the mole ratio of monomer to complex plays an important role on the TOF of copolymerization, as listed in Table 4 (entries 3 and 8–10). For example, in the experiment of entry 3, TOF (10 h) is 48.7 h⁻¹ when 1,000 equiv. of CHO are used. As the ratio of CHO to complex 1 decreases to 200 equiv. in entry 8, the reaction activity decreases by almost thrice, compared with entry 3. Meanwhile, PDI also slightly decreases from 1.2 to 1.1, and M _n decreases significantly from 38.7 to 14.9 kg mol⁻¹. However, the further increases in the ratio of CHO to complex 1 above 1,000 equiv. could reduce the reaction activities (see entries 3 and 10, in Table 3). Especially, when over 1,000 equiv. of CHO are selected, both drops in catalytic activity and yield of the resultant product are observed (entry 10), which could be attributed to the lowing of the catalytic concentration when overloading monomer.

Besides operation temperature and ratio of monomer to complex, initial CHO concentration is another important issue for the reaction. In our results, it is observed that the catalytic activity increases as the initial CHO concentration increases from 30 to 50 mol l⁻¹ (see entries 3, 11, and 12, in Table 4) and decreases then with the further increase in CHO concentration (entry 13 in Table 4). The prolonged reaction does not lead to an observable decrease in selectivity for PCHC formation, which can be explained by the so-called solvent effect. A similar solvent effect was reported for the copolymerization with binary catalyst systems [32, 42].

¹³C NMR spectroscopy has been extensively used to investigate CHO incorporation into polymer backbones, which can be used to estimate the tacticity of the copolymer [2, 14]. Figure 3 presents a ¹³C NMR spectrum of the PCHC (entry 3 in Table 4). The spectrum is very similar to those of PCHC prepared with (salen)CoX catalysts [5]. Clearly, the carbonyl regions at δ = 153.80–153.66 and 153.25–153.30 ppm can be assigned to the isotactic and syndiotactic structures of the PCHC chain, respectively [2]. The peaks at 153.25–153.30 ppm are assigned to the central carbonyl carbons of r-centered tetrads, whereas the peaks at 153.80–153.66 ppm are attributed to the central carbonyl carbons of m-centered tetrads [2, 14]. PCHCs revealed a strong contribution correlating to r-centered tetrads (78 %), indicative of syndiotactic PCHC.

MALDI-TOF mass spectrometry is specifically suitable for structural and end-group analysis [5]. Figure 4 displays a spectrum of the resultant PCHC obtained (entry 3 in Table 4). At least two series of signals with regular repeat intervals (repeating units of 142.2 mass units) were observed in the mass spectra. For the first series of peaks at the m/z of 1,931.6, 2,073.8, 2,216.0, 2,358.2, 2,500.4, and 2,642.6, there was a difference of 142.2 in m/z between every two neighboring peaks, which was also seen in the second series at the m/z of 1,947.7, 2,089.9, 2,232.1, 2,374.3, 2,516.5, and 2,658.7. Moreover, in the second series peaks in m/z were correspondingly 16 larger than those in the first series. Additionally, the weight number of propylene carbonate unit is 142.2. Thus, it could be known by calculation that the first series peaks resulted from the copolymerization reaction was initiated by insertion of CHO into a Co-acetate bond of the catalytic species attached to a Co center, and terminated with acetate groups and Na⁺ ions, and K⁺ ions instead of Na⁺ ions for the second series.

On the basis of the information presented so far, some proposed mechanism illustrated that the monomer was inserted to the cobalt–oxygen bond of the propagating end, as shown in Scheme 1. [PPN]Cl coordinated to the metal center in the axial site transferred to the propagating metal–polymer chain, thereby labilizing the metal alkoxide bond and facilitating the insertion of CO₂ [3, 6, 17, 18, 21, 24]. The CHO monomer might be first favorably inserted into the Co–OR bond of the Schiff base cobalt complex, followed by the insertion of the CO₂ monomer. Subsequent alternating copolymerization of CHO and CO₂ affords the alternating repeating unit structure to yield the polycarbonate. The side reaction to yield the cyclic carbonate could take place through the backbiting degradation of the growing polymer–catalyst complex, which may lead to the lower polymer yield. A mechanism similar to that proposed by Darensbourg and Yarbrough using chromium salen and N-MeIm in the copolymerization of CO₂ and cyclohexene oxide [14].

By use of complex 1 and [PPN]Cl catalyst system, the copolymerization of CO₂ and PO ([PO]₀/[[PPN]Cl]₀/[complex 1]₀ = 500/1/1) selectively gave the alternating copolymer in 1,4-dioxane at 60 °C in 24 h (carbonate linkages = 99 %, polycarbonate/cyclic carbonate = 99/1, M _n  = 6,400, M _w/M _n  = 1.21; Table 5, run 2), although the reaction proceeded rather slowly (36.8 % yield). Also at 80 °C, the copolymeriation proceeded more rapidly to attain 42.1 % polymer yield in the same reaction time, slightly lowering the selectivity (carbonate linkages = 95.8 %, polycarbonate/cyclic carbonate = 92/8, M _n  = 7,800, M _w/M _n  = 1.31; Table 5, run 3). When the copolymerization was conducted at an elevated temperature, such as 100 °C, a significant amount of cyclic carbonate (polycarbonate/cyclic carbonate = 30/70) was generated under otherwise similar conditions (Table 5, run 3).

The high selectivity of polycarbonate over cyclic carbonate at 60 °C was maintained after a longer reaction time to attain quantitative formation of the alternating copolymer. The alternating copolymeriation of CO₂ and PO by using complex 1 and [PPN]Cl catalyst system was carried out at 60 °C for 48 h and afforded the alternating copolymer almost quantitatively (polycarbonate/ cyclic carbonate = 93/7, carbonate linkages = 99 %, M _n  = 13,900, M _w/M _n  = 1.26, Table 5, run 5).