Bimetallic Aluminum 5,6-Dihydro-7,7-dimethyl quinolin-8-olates as Pro-Initiators for the ROP of ε-CL; Probing the Nuclearity of the Active Initiator

Abstract

Six examples of aluminum 5,6-dihydro-7,7-dimethylquinolin-8-olates, [{2-R¹-7,7-Me₂-8-R²C₉H₆N-8-O}AlR³₂]₂ (R¹ = R² = H, R³ = Me C1; R¹ = R² = H, R³ = Et C2; R¹ = R² = H, R³ = i-Bu C3; R¹ = Cl, R² = H, R³ = Me C4; R¹ = H, R² = R³ = Me C5; R¹ = Cl, R² = R³ = Me C6), have been prepared by treating the corresponding pro-ligand (L1–L4) with either AlMe₃, AlEt₃ or Al(i-Bu)₃. All complexes have been characterized by ¹H and ¹³C NMR spectroscopy and in the case of C1 and C4 by single crystal X-ray diffraction; dimeric species are a feature of their molecular structures. In the presence of PhCH₂OH (BnOH), C1–C6 displayed good control and efficiency for the ROP of ε-CL with almost 100% conversion achievable in 10 min at 90 °C; the chloro-substituted C4 and C6 notably exhibited the lowest activity of the series. However, in the absence of BnOH, C1 showed only low activity with 15% conversion achieved in 30 min forming a linear polymer capped with either a methyl or a L1 group. By contrast, when one or more equivalents of BnOH was employed in combination with C1, the resulting catalyst was not only more active but gave linear polymers capped with BnO end-groups. By using ¹H and ²⁷Al NMR spectroscopy to monitor solutions of C1, C1/BnOH and C1/BnOH/10 ε-CL over a range of temperatures, some support for a monomeric species being the active initiator at the operational temperature is presented.

1. Introduction

The past 10 years or so have seen some rapid progress in the synthesis of biodegradable polymers and in particular, aliphatic polyesters such as polylactides (PLA) and poly(ε-caprolactone) (PCL). These developments can be attributed, in a large measure, to the good biodegradability and biocompatibility properties of these materials as well as to their ease of preparation [1,2,3,4]. Typically, such polyesters can be prepared by the ring-opening polymerization (ROP) of cyclic esters catalyzed by metal complexes such as those based on Al, Ca, Sn and rare earth metals. In addition, immobilized catalysts have been considered for the improvement of the mechanical properties of PCLs [5]. Several review articles have documented advances in catalyst design for the ROP of cyclic esters [6,7]. Among the numerous reports, aluminum complexes bearing multidentate ligands such as Salen-Al and Salan-Al have attracted much attention due to their relatively high Lewis acidity, good controllability, as well as their decent selectivity towards the ROP of rac-lactide (rac-LA) and ε-caprolactone (ε-CL) [8,9,10,11]. By way of contrast, there are still relatively few examples of bidentate N^O-type aluminum complexes that have been used effectively for the ROP of cyclic esters.

With regard to N^O-aluminum complexes containing six-membered chelate rings (A–G, Chart 1), the 2-iminophenolates A constitute the most studied class of pro-initiator and indeed are highly effective in the presence of benzyl alcohol for the ROP of ε-CL and LA [12,13,14,15]. They can also efficiently promote the copolymerization of rac-lactide and glycolide [16], rac-β-butyrolactone and L-lactide [17] and the random copolymerization of rac-LA and ε-CL with various degrees of control [18]. Related bimetallic aluminum complexes incorporating two linked iminophenolate units have also been evaluated in the ROP of cyclic esters [19]. The dialkylaluminum aminophenolate complexes B have been shown as effective (pro-)initiators for the ROP of ε-CL and epoxides [20,21,22,23], while their dinuclear analogues showed higher activity [22,23]; good selectivity for the ROP of rac-LA have also been noted for some B-type systems [24]. Other aluminum phenolates such as dialkylaluminum 2-imidazolylphenolates (C) and 2-benzoxazole phenolates (D) have also been reported as efficient (pro-)initiators for the ROP of rac-lactide and ε-CL [25,26]. The β-ketiminato-aluminum complex E showed good efficiency for the random copolymerization of LA and ε-CL [18,27], while the fluorinated alkoxy-imino aluminum species F are active for the ROP of ε-CL [28]. Conversely, the amido-phosphinoxide aluminum complex G showed virtually no activity for the ROP of ε-CL [29].

In comparison with the six-membered ring examples highlighted above (A–G, Chart 1), those based on five-membered N^O-chelate rings have been considerably less studied as ROP (pro-)initiators, though a tendency to form dimeric species is a feature of their structural chemistry (H–M, Chart 1). Nevertheless, a recent report has shown that five-membered ring aluminum complexes can display significantly higher polymerization rates when compared to their six-membered ring counterparts (e.g., two to three-fold increase for ε-CL polymerization) [30]. Elsewhere, N^O-bidentate aluminum complexes of type H [31,32] and J have been reported but have not been the subject of ROP studies though H has been discussed in terms of the relative stability of its mono- and dimeric forms [33]. On the other hand, dimeric J was shown to catalyze the cycloaddition of CO₂ with epoxides [34]. Moreover, the anilinotropone-based dimeric aluminum complexes K exhibited, in the presence of BnOH, high activity in the ROP of rac-lactide [35]. Nonetheless, common to J and K it remains uncertain whether the dimeric structural forms are maintained during the polymerization or undergo dissociation to their monomeric forms. Interestingly, we have found the di- and monomeric complexes of dialkylaluminum 2-(arylimino)-quinolin-8-olates (L) display very different catalytic performance for the ROP of ε-CL. For example, the binuclear systems showed very low activity, while the mononuclear examples gave much higher efficiency and in a more controlled manner [36]. Furthermore, our group has recently synthesized a series of bimetallic dialkylaluminum hydroquinolin-8-olates (M: R¹ = H, Chart 1) that showed good activity only at high temperature for the ROP of ε-CL. By using variable temperature ¹H and ²⁷Al NMR spectroscopy, we proposed that the di-aluminum complexes partly dissociated into mononuclear species at high temperature [37]. However, these complexes displayed only poor solubility at room temperature which precluded full assignments of their peaks in their ¹H NMR spectra.

With a view to improve the solubility of M (Chart 1) and to re-investigate the nuclearity of the active initiator, we report herein the introduction of two methyl groups at the 7-position of the ligand framework. In particular, we report a series of bimetallic aluminum 5,6-dihydro-7,7-dimethylquinolin-8-olates, [{2-R¹-7,7-Me₂-8-R²C₉H₆N-8-O}AlR³₂]₂ (R¹ = R² = H, R³ = Me; R¹ = R² = H, R³ = Et; R¹ = R² = H, R³ = i-Bu; R¹ = Cl, R² = H, R³ = Me; R¹ = H, R² = R³ = Me; R¹ = Cl, R² = R³ = Me), that differ in the substitution pattern at the R¹, R² and R³ positions. A full evaluation of these complexes as either initiators or pro-initiators for the ROP of ε-CL is conducted and these results compared with those observed using parent M. In addition, the pathway by which pro-initiator is transformed in to the initiator is probed using both ¹H and ²⁷Al NMR spectroscopy.

2. Materials and Methods

2.1. General Considerations and Materials

All manipulations of air or moisture-sensitive compounds were performed using standard Schlenk techniques under an atmosphere of high-purity nitrogen or using glove box techniques. Toluene was dried by refluxing it over sodium/benzophenone and distilled under nitrogen and stored over activated molecular sieves (4 Å) for 24 h in a glove box prior to use. n-Hexane and CDCl₃ were dried over CaH₂ for 48 h, distilled under nitrogen and stored over activated molecular sieves (4 Å) in a glove box prior to use. Solutions of Me₃Al (1.0 M in toluene), Et₃Al (1.0 M in toluene), i-Bu₃Al (1.0 M in toluene) and methyllithium (1.0 M in toluene) were purchased from Aldrich and used as received. Elemental analyses were performed using a PE2400II Series instrument (Perkin-Elmer Co., Shanghai, China). The ¹H and ¹³C NMR spectra were recorded on a Bruker DMX-400/300 MHz (Karlsruhe, Germany) spectrometer using TMS as an internal standard; δ values are given in ppm and J values in Hz. The ²⁷Al NMR spectra were recorded on Bruker 500 MHZ (Beijing, China) spectrometer. The NMR spectra of the complexes and ligands were recorded in CDCl₃ at room temperature. IR spectra were recorded on a Perkin-Elmer System 2000 (Shanghai, China) FT-IR spectrometer. The GPC measurements were performed using a set-up based on a Waters-1515 HPLC pump, a Waters 2414 (Beijng, China) refractive index detector and a combination of Styragel HT-2, HT-3 and HT-4 columns, the effective molar mass ranges of which are 100–10,000, 500–30,000 and 5000–600,000, respectively. THF was used as the eluent (flow rate: 1 mL/min, at 35 °C). Molecular weights and molecular weight distributions were calculated using polystyrene as a standard.

2.2. Syntheses of 2-R¹-7,7-Me₂,8-R²C₉H₆N-8-OH (L)

R¹ = R² = H L1. A mixture of 5,6-dihydro-7,7-dimethylquinolin-8-one (0.88 g, 5.0 mmol) and sodium borohydride (0.19 g, 5.0 mmol) was dissolved in methanol (30 mL) in a 100 mL round flask and stirred for 5 h at room temperature. The reaction was then quenched with water, extracted with dichloromethane and the organic phase dried over anhydrous magnesium sulfate and filtered. The filtrate was collected and the solvent evaporated under reduced pressure to give L1 as a light yellow solid. Yield: 0.86 g (96%). ¹H NMR (400 MHz, CDCl₃, ppm): δ 8.40 (d, J = 6.0 Hz, 1H, Py–H), 7.41 (d, J = 8.0 Hz, 1H, Py–H), 7.14–7.11 (m, 1H, Py–H), 4.38 (s, 1H, CH–OH), 4.31 (s, 1H, OH), 2.92–2.79 (m, 2H, CH₂), 1.75–1.65 (m, 2H, CH₂), 1.19 (s, 3H, CH₃), 0.85 (s, 3H, CH₃). ¹³C NMR (100 MHz, CDCl₃, ppm): δ 157.3, 146.3, 136.5, 130.4, 122.0, 76.2, 34.0, 33.3, 27.5, 24.8, 18.9. IR (cm⁻¹): 3117 (m), 2953 (m), 2922 (m), 2859 (w), 2714 (w), 1583 (m), 1446 (m), 1377 (m), 1359 (m), 1307 (w), 1260 (m), 1160 (m), 1107 (m), 1042 (s), 981 (m), 936 (w), 904 (w), 857 (w), 788 (s), 751 (m), 713 (m). Anal. Calcd for C₁₁H₁₅NO: C, 74.54; H, 8.53; N, 7.90%. Found: C, 74.40; H, 8.53; N, 7.81%.

R¹ = Cl, R² = H L2. Using a similar procedure to that described for L1, but with 2-chloro-5,6-dihydro-7,7-dimethylquinolin-8-one as the ketone, gave L2 as a white powder. Yield: 1.03 g (97%). ¹H NMR (400 MHz, CDCl₃, ppm): δ 7.38 (d, J = 8.0 Hz, 1H, Py–H), 7.14 (d, J = 8.0 Hz, 1H, Py–H), 4.27 (s, 1H, CH–OH), 3.76 (s, 1H, OH), 2.85–2.69 (m, 2H, CH₂), 1.71–1.67 (m, 2H, CH₂), 1.16 (s, 3H, CH₃), 0.87 (s, 3H, CH₃). ¹³C NMR (100 MHz, CDCl₃, ppm): δ 158.3, 148.6, 139.6, 129.0, 123.0, 76.0, 34.0, 33.3, 27.5, 24.3, 19.0. IR (cm⁻¹): 3281 (m), 2943 (m), 2905 (m), 2865 (w), 1570 (m), 1469 (m), 1440 (m), 1416 (w), 1377 (m), 1363 (m), 1256 (m), 1230 (w), 1197 (w), 1164 (m), 1131 (w), 1090 (m), 1033 (s), 946 (m), 904 (w), 870 (w), 842 (m), 789 (s), 756 (m), 727 (m). Anal. Calcd for C₁₁H₁₄ClNO: C, 62.41; H, 6.67; N, 6.62%. Found: C, 62.58; H, 6.59; N, 6.43%.

R¹ = H, R² = Me L3. Methyllithium (5.0 mL, 5.0 mmol, 1.0 M solution in toluene) was added dropwise to a stirred solution of 5,6-dihydro-7,7-dimethylquinolin-8-one (0.88 g, 5.0 mmol) in toluene (10.0 mL) at −30 °C. The reaction mixture was allowed to warm slowly to room temperature and stirred overnight. Following quenching with water and extraction with dichloromethane, the organic phase was dried over anhydrous magnesium sulfate and filtered. The filtrate was collected and the solvent removed under reduced pressure to give L3 as a white solid. Yield: 0.93 g (97%). ¹H NMR (400 MHz, CDCl₃, ppm): δ 8.39 (d, J = 4.0 Hz, 1H, Py–H), 7.38 (d, J = 8.0 Hz, 1H, Py–H), 7.11–7.08 (m, 1H, Py–H), 4.25 (s, 1H, OH), 2.90–2.74 (m, 2H, CH₂), 1.94–1.86 (m, 1H, CH₂), 1.63–1.58 (m, 1H, CH₂), 1.37 (s, 1H, CH₃), 1.15 (s, 1H, CH₃), 0.93 (s, 1H, CH₃). ¹³C NMR (100 MHz, CDCl₃, ppm): δ 162.3, 148.3, 139.6, 128.2, 122.4, 74.7, 36.1, 33.0, 26.7, 24.5, 23.9, 21.9. IR (cm⁻¹): 3413 (m), 2959 (m), 2872 (w), 1702 (w), 1580 (m), 1447 (m), 1424 (m), 1386 (m), 1359 (m), 1336 (w), 1283 (m), 1260 (m), 1169 (m), 1131 (m), 1077 (s), 1018 (m), 978 (m), 934 (w), 907 (w), 789 (s). Anal. Calcd for C₁₂H₁₇NO: C, 75.35; H, 8.96; N, 7.32%. Found: C, 75.26; H, 8.82; N, 7.46%.

R¹ = Cl, R² = Me L4. Using a similar procedure to that described for L3, but with 2-chloro-5,6-dihydro-7,7-dimethylquinolin-8-one as the ketone, gave L4 as a white powder. Yield: 1.09 g (97%). ¹H NMR (400 MHz, CDCl₃, ppm): 7.34 (d, J = 8.0 Hz, 1H, Py–H), 7.11 (d, J = 8.0 Hz, 1H, Py–H), 3.65 (s, 1H, OH), 2.87–2.70 (m, 2H, CH₂), 1.89–1.81 (m, 1H, CH₂), 1.65–1.59 (m, 1H, CH₂), 1.38 (s, 3H, CH₃), 1.12 (s, 3H, CH₃), 0.94 (s, 3H, CH₃). ¹³C NMR (100 MHz, CDCl₃, ppm): δ 161.4, 146.5, 136.5, 129.2, 121.8, 74.9, 36.2, 33.0, 27.1, 24.5, 24.4, 22.0. IR (cm⁻¹): 3477 (m), 2963 (m), 2924 (m), 1697 (m), 1572 (m), 1444 (m), 1427 (m), 1385 (m), 1360 (m), 1328 (m), 1316 (w), 1261 (m), 1191 (w), 1166 (m), 1127 (m), 1069 (s), 1016 (m), 932 (w), 854 (w), 815 (s), 748 (m). Anal. Calcd for C₁₂H₁₆ClNO: C, 63.86; H, 7.15; N, 6.21%. Found: C, 63.45; H, 7.28; N, 6.46%.

2.3. Syntheses of [{2-R¹-7,7-Me₂-8-R²C₉H₆N-8-O}AlR₃²]₂ (C)

R¹ = R² = H, R³ = Me C1. Me₃Al (5.0 mL, 5.0 mmol, 1.0 M solution in toluene) was added dropwise to a stirred solution of L1 (0.89 g, 5.0 mmol) in toluene (10 mL) at −78 °C. The resulting solution was allowed to warm slowly to room temperature and stirred for 3 h. Following concentration of the reaction mixture to ca. 1 mL, hexane (10 mL) was added to induce precipitation. The precipitate was filtered affording C1 as a white powder. Yield: 0.75 g, 65%. ¹H NMR (400 MHz, toluene-d₈, ppm): δ 7.44 (d, J = 8.0 Hz, 1H, Py–H), 6.63 (d, J = 8.0 Hz, 1H, Py–H), 6.47–6.41 (m, 1H, Py–H), 4.58 (s, 1H, CH–O), 2.07–1.98 (m, 2H, CH₂), 1.52 (s, 3H, CH₃), 1.21–1.13 (m, 2H, CH₂), 0.68 (s, 3H, CH₃), −0.12 (s, 3H, Al–CH₃), −0.26 (s, 3H, Al–CH₃). ¹³C NMR (100 MHz, toluene-d₈, ppm): δ 157.0, 141.7, 140.1, 134.2, 124.0, 82.1, 36.2, 35.8, 30.1, 24.2, 17.7, −3.26. Anal. Calcd for C₁₃H₂₀AlNO: C, 66.93; H, 8.64; N, 6.00%. Found: C, 67.13; H, 8.77; N, 5.88%.

R¹ = R² = H, R³ = Et C2. Using a similar procedure to that described for C1, but with Et₃Al (1.0 M solution in toluene) as the alkyl aluminum reagent, gave C2 as a white powder. Yield: 0.55 g, 42%. ¹H NMR (400 MHz, toluene-d₈, ppm): δ 8.11 (d, J = 4.0 Hz, 1H, Py–H), 6.74 (d, J = 5.6 Hz, 1H, Py–H), 6.57 (t, J = 6.0 Hz, 1H, Py–H), 5.01 (s, 1H, CH–O), 2.34–2.22 (m, 2H, CH₂), 1.54 (s, 3H, CH₃), 1.45–1.37 (m, 1H, Al–CH₂CH₃), 1.30–1.25 (m, 2H, CH₂), 1.21–1.15 (m, 3H, Al–CH₂CH₃), 0.94 (s, 3H, CH₃ and Al–CH₂CH₃), 0.51–0.44 (m, 2H, Al–CH₂CH₃), 0.35–0.26 (m, 2H, Al–CH₂CH₃). ¹³C NMR (100 MHz, toluene-d₈, ppm): δ 164.4, 158.5, 142.5, 132.1, 122.7, 80.9, 36.7, 36.5, 30.2, 24.9, 18.6, 11.4, 11.3, 1.9. Anal. Calcd for C₁₅H₂₄AlNO: C, 68.94; H, 9.26; N, 5.36%. Found: C, 69.17; H, 9.44; N, 5.23%.

R¹ = R² = H, R³ = i-Bu C3. Using a similar procedure to that described for C1, but with i-Bu₃Al (1.0 M solution in toluene) as the alkyl aluminum reagent, gave C3 as a white powder. Yield: 0.33 g, 21%. ¹H NMR (400 MHz, CDCl₃, ppm): δ 8.23 (d, J = 8.0 Hz, 1H, Py–H), 7.67 (d, J = 4.0 Hz, 1H, Py–H), 7.38–7.34 (m, 1H, Py–H), 4.61 (s, 1H, CH–O), 2.94–2.80 (m, 2H, CH₂), 1.86–1.67 (m, 3H, CH₂ and Al–CH₂CH(CH₃)₂), 1.23 (s, 3H, CH₃), 0.90–0.75 (m, 12H, Al–CH₂CH(CH₃)₂), 0.73 (s, 3H, CH₃), −0.07–−0.09 (m, 4H, Al–CH₂CH(CH₃)₂). ¹³C NMR (100 MHz, CDCl₃): δ 164.0, 142.2, 139.6, 133.6, 122.8, 80.4, 35.3, 34.3, 28.2, 28.1, 28.0, 27.9, 26.2, 26.0, 24.5, 22.6, 17.4. Anal. Calcd for C₁₉H₃₂AlNO: C, 71.89; H, 10.16; N, 4.41%. Found: C, 71.56; H, 10.52; N, 4.67%.

R¹ = Cl, R² = H, R³ = Me C4. Using a similar procedure to that described for C1, but with L2 as the pro-ligand, gave C4 as a white powder. Yield: 1.14 g, 85%. ¹H NMR (400 MHz, CDCl₃, ppm): 7.74 (d, J = 8.0 Hz, 1H, Py–H), 7.44 (d, J = 8.0 Hz, 1H, Py–H), 4.92 (s, 1H, CH–O), 2.85–2.81 (m, 2H, CH₂), 1.92–1.84 (m, 1H, CH₂), 1.71–1.65 (m, 1H, CH₂), 1.51 (s, 3H, CH₃), 0.84 (s, 3H, CH₃), -0.58 (s, 6H, Al–CH₃). ¹³C NMR (100 MHz, CDCl₃, ppm): δ 158.4, 145.5, 142.4, 132.8, 126.7, 81.5, 36.1, 35.8, 31.6, 29.5, 23.9, 18.1, 14.3, −4.5. Anal. Calcd for C₁₃H₁₉AlClNO: C, 58.32; H, 7.15; N, 5.23%. Found: C, 58.11; H, 7.24; N, 5.38%.

R¹ = H, R² = R³ = Me C5. Using a similar procedure to that described for C1, but with L3 as the pro-ligand, gave C5 as a white powder. Yield: 0.36 g, 29%. ¹H NMR (400 MHz, Toluene-d₈, ppm): δ 7.49 (d, J = 4.0 Hz, 1H, Py–H), 6.74 (d, J = 8.0 Hz, 1H, Py–H), 6.51–6.48 (m, 1H, Py–H), 2.21–2.18 (m, 2H, CH₂), 1.56–1.51 (m, 1H, CH₂), 1.49 (s, 3H, CH₃), 1.43 (s, 3H, CH₃), 1.22–1.15 (m, 1H, CH₂), 0.83 (s, 3H, CH₃), −0.10 (s, 3H, Al–(CH₃)₂), −0.30 (s, 3H, Al–(CH₃)₂). ¹³C NMR (100 MHz, toluene-d₈, ppm): δ 163.2, 143.4, 142.5, 134.3, 126.0, 87.7, 40.3, 36.72, 28.9, 27.8, 25.2, 23.3, −2.09, −2.92. Anal. Calcd. for C₁₄H₂₂AlNO: C, 67.99; H, 8.97; N, 5.66%. Found: C, 68.20; H, 8.83; N, 5.79%.

R¹ = Cl, R² = R³ = Me C6. Using a similar procedure to that described for C1, but with L4 as the pro-ligand, gave C6 as a white powder. Yield: 0.48 g, 34%. ¹H NMR (300 MHz, CDCl₃, ppm): δ 7.52 (d, J = 9.0 Hz, 1H, Py–H), 7.28 (t, J = 9.0 Hz, 1H, Py–H), 2.88–2.83 (m, 2H, CH₂), 2.02–1.91 (m, 1H, CH₂), 1.70–1.62 (m, 1H, CH₂), 1.44 (s, 3H, CH₃), 1.29 (s, 3H, CH₃), 0.91 (s, 3H, CH₃), −0.69 (s, 6H, Al–(CH₃)₂). ¹³C NMR (100 MHz, CDCl₃, ppm): δ 161.1, 146.7, 141.1, 128.2, 125.0, 81.0, 37.4, 33.8, 27.3, 25.1, 23.2, 21.0, 14.4, −8.3. Anal. Calcd for C₁₄H₂₁AlClNO: C, 59.68; H, 7.51; N, 4.97%. Found: C, 59.55; H, 7.32; N, 4.85%.

2.4. X-ray Crystallographic Studies

Single crystal X-ray diffraction data for C1 and C4 were collected on a Rigaku Sealed Tube CCD (Saturn 724+) diffractometer (Tokyo, Japan) with graphite monochromated Mo-Kα radiation (λ = 0.71073 Å) at 293(2) K. Cell parameters were obtained by global refinement of the positions of all collected reflections. Intensities were corrected for Lorentz and polarization effects and empirical absorption. The structures were solved by direct methods and refined by full-matrix least squares on F². All hydrogen atoms were placed in calculated positions. Structure solution and refinement were performed using the SHELXTL-97 package [38,39]. Details of the X-ray structure determinations and refinements are provided in

Table 1. Crystal data and structure refinement for C1 and C4.

	C1	C4
Empirical formula	C26H40Al2N2O2	C13H19AlNOCl
Formula weight	466.56	267.72
Crystal color	colorless	colorless
Temperature (K)	293(2)	293(2)
Wavelength (Å)	0.71073	0.71073
Crystal system	monoclinic	monoclinic
space group	P2/n	P21/n
a (Å)	9.762(2)	10.934(5)
b (Å)	9.773(2)	9.902(4)
c (Å)	13.676(3)	13.019(6)
α (°)	90.00	90.00
β (°)	90.39(3)	100.391(5)
γ (°)	90.00	90.00
Volume (Å3)	1304.7(5)	1386.5(11)
Z	2	4
Dcalc (Mg m−3)	1.188	1.283
µ (mm−1)	0.136	0.323
F (000)	504.0	568.0
Crystal size (mm)	0.49 × 0.29 × 0.21	0.64 × 0.38 × 0.16
θ range (°)	5.14–54.96	4.49–54.94
Limiting indices	−12 ≤ h ≤ 11 −12 ≤ k ≤ 12 −17 ≤ l ≤ 17	−14 ≤ h ≤ 14 −12 ≤ k ≤ 12 −16 ≤ l ≤ 16
No.of rflns collected	13159	14473
No. of unique rflns Rint	2963 (0.0301)	3144 (0.0526)
Completeness to θ (%)	99.0%	99.3%
Goodness-of-fit on F2	1.198	1.137
Final R indices [I > 2σ (I)]	R1 = 0.0372, wR2 = 0.1078	R1 = 0.0652, wR2 = 0.1592
R indices (all data)	R1 = 0.0429, wR2 = 0.1246	R1 = 0.0678, wR2 = 0.1621
Largest diff. peak and hole (eÅ−3)	0.44/−0.48	0.36/−0.34

. The Cambridge Crystallographic Data Centre CCDC 1851289 and 1851290 for C1 and C4, respectively, contain the supplementary crystallographic data for the paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre (Cambridge, UK).

2.5. General Procedure for ε-Caprolactone Polymerization

A typical polymerization procedure in the presence of one equivalent of benzyl alcohol is outlined as follows. Precatalyst C1 (0.0047 g, 0.020 mmol) was dissolved in toluene (2 mL) in a Schlenk flask at room temperature and a solution of benzyl alcohol (0.020 mmol) in toluene added and the mixture stirred at room temperature for 5 min. The flask was then placed in a temperature-controlled oil bath preheated to 90 °C and ε-CL (0.571 g, 5.0 mmol) was injected. After the solution was stirred for the pre-determined time, the polymerization was terminated by the addition of glacial acetic acid (ca. 0.2 mL). The resulting viscous solution was diluted with dichloromethane and then transferred to a beaker containing cold methanol (100 mL) with stirring. The resultant polymer was collected on filter paper and dried under reduced pressure to give a white solid.

3. Results and Discussion

3.1. Synthesis and Characterization of C1–C6

Based on previous synthetic approaches, reduction of the 5,6-dihydro-7,7-dimethylquinolin-8-ones, 2-R¹-7,7-Me₂C₉H₆N-8-O (R¹ = H; R¹ = Cl), with NaBH₄ or LiMe gave the dihydroquinolin-8-ols, 2-R¹-7,7-Me₂,8-R₂C₉H₆N-8-OH (R¹ = R² = H L1; R¹ = Cl, R² = H L2; R¹ = H, R² = Me L3; R¹ = Cl, R² = Me L4), in high yield (Scheme 1). All four compounds have been fully characterized by ¹H/¹³C NMR, FT-IR spectroscopy as well as by elemental analysis. Treatment of L1 with one equivalent of (R³)₃Al (R³ = Me, Et, i-Bu) in toluene at −78 °C afforded [{2-R¹-7,7-Me₂-8-R₂C₉H₆N-8-O}AlR³₂]₂ (R¹ = R² = H, R³ = Me C1; R¹ = R² = H, R³ = Et C2; R¹ = R² = H, R³ = i-Bu C3), while reaction of L2, L3 and L4 with solely Me₃Al gave [{2-R¹-7,7-Me₂-8-R²C₉H₆N-8-O}AlMe₂]₂ (R¹ = Cl, R² = H C4; R¹ = H, R² = Me C5; R¹ = Cl, R² = Me C6), in moderate to good yields. All six complexes have been characterized by ¹H and ¹³C NMR spectroscopy as well as by elemental analysis. In addition, C1 and C4 have been the subject of single crystal X-ray diffraction studies.

Single crystals of C1 and C4 suitable for the X-ray determinations were grown by the slow diffusion of n-hexane into their toluene solutions at room temperature under a nitrogen atmosphere. Perspective views of the structures of C1 and C4 are shown in Figure 1 and Figure 2; selected bond lengths and bond angles are collected in

Table 2. Selected bond lengths (Å) and angles (°) for C1 and C4.

	C1	C4
Bond lengths (Å)
Al(1)–N(1)	2.122(13)	2.194(3)
Al(1)–O(1)	1.876(11)	1.856(18)
Al(1)–O(1i)	1.953(11)	1.966(2)
Al(1)–C(12)	1.997(15)	1.977(3)
Al(1)–C(13)	1.983(16)	1.985(3)
O(1)–C(8)	1.427(15)	1.398(5)
Bond angles (°)
N(1)–Al(1)–O(1)	79.14(5)	78.73(9)
N(1)–Al(1)–O(1i)	154.49(5)	158.62(8)
Al(1)–O(1)–C(8)	121.91(8)	121.8(3)
Al(1)–N(1)–C(9)	113.13(9)	110.97(18)
O(1)–Al(1)–O(1i)	77.25(5)	79.93(8)
C(12)–Al(1)–C(13)	123.79(7)	134.10(13)

The atoms labeled with ‘i’ have been generated by symmetry.

. Both C1 and C4 have been generated by symmetry and adopt similar arrangements and will be discussed together. Their structures consist of two aluminum centers that are each chelated by an N^O-bound 5,6-dihydro-7,7-dimethylquinolin-8-olate which also acts as a bridging ligand via O1. The five-coordinate geometry of each metal center is completed by two methyl ligands. The main difference between the structures relates to the substituent located on the 2-position of the N^O-chelate viz. H (C1) and Cl (C4). Indeed, this dimeric arrangement is similar to that observed in the unsubstituted 5,6-dihydroquinolin-8-olates M (Chart 1) [37]. The Al-N bond length in C1 [2.122(13) Å] is slightly shorter than that in C4 [2.194(3) Å] likely reflecting the steric properties of the more bulky chloride substituent positioned at C1. There are no intermolecular contacts of note.

The ¹H NMR spectra of C1–C6 reveal characteristic peaks for the corresponding 5,6-dihydro-7,7-dimethylquinolin-8-olate ligands with the inequivalent gem-dimethyl groups giving rise to two distinct signals. Likewise, the Al-CH₃ (C1, C4–C6) or Al-CH₂ (C2, C3) resonances are seen, in most cases, as separate signals and appear most upfield in their spectra. Interestingly, the ¹H NMR spectrum of C1, recorded in three different deuterated solvents, indicates that the solvent polarity has some effect on the chemical shift of the resonances (

Table 3. The ¹H NMR chemical shifts for C1 recorded in different deuterated solvents.

Proton Type	δ (ppm) in Toluene-d8	δ (ppm) in Benzene-d6	δ (ppm) in Chloroform-d
Py-CH	7.44 (1H), 6.63 (1H),6.43 (1H)	7.29 (1H), 6.43 (1H),6.23 (1H)	8.24 (1H), 7.79 (1H), 7.50 (1H)
Cy-CH	4.58 (1H)	4.49 (1H)	4.92 (1H)
Cy-CH2	2.04 (2H), 1.20–1.11 (2H)	1.86 (2H), 1.08 (1H),0.96 (1H)	2.88 (2H), 1.94–1.84 (1H),1.74–1.66 (1H)
Cy-CH3	1.51 (3H), 0.67 (3H)	1.45 (3H), 0.58 (3H)	1.50 (3H), 0.82 (3H)
Al-CH3	−0.12 (3H), −0.26 (3H)	−0.12 (3H), −0.25 (3H)	−0.62, −0.66, −0.67

). In an apolar solvent such as toluene-d₈ and benzene-d₆, each proton displays very similar chemical shifts. However, the shift in chloroform-d is greatly affected, which is best exemplified by the most upfield Cy-CH₃ signal which appears at δ 0.58 in C₆D₆ while in CDCl₃ at δ 0.82. In addition, the number of Al-CH₃ signals shows some variation with two peaks (δ −0.12, −0.26) in C₆D₆ and three peaks (δ −0.62, −0.66, −0.67) in CDCl₃. It is uncertain as to the origin of these differences but it may be due to the ability of the more polar chloroform to interrupt the apparent monomeric/dimeric equilibrium in solution (vide infra).

3.2. Ring Opening Polymerization of ε-Caprolactone by C1–C6

To explore the capacity of C1–C6 to serve as pro-initiators for the ring-opening polymerization of ε-CL, C1 was firstly selected as the test system; the results of the catalytic evaluations are collected in Table 3. In the first instance, the study focused on establishing the optimal temperature by performing the polymerization runs between 25 and 110 °C. Typically, each run was conducted in toluene with one equivalent of PhCH₂OH (BnOH) as activator and 250 equivalents of the monomer over a run time of 10 min (runs 1–7,

Table 4. The ROP of ε-CL using C1 ^a.

Run	ɛ-CL:Al:BnOH	T (°C)	t (min)	Solvent	Conv. (%) b	Mn (×104) c	Mn (calc) (×104) d	PDI c	TOF (h−1)
1	250:1:1	25	10	Toluene	0
2	250:1:1	40	10	Toluene	0
3	250:1:1	60	10	Toluene	47	1.11	1.35	1.27	705
4	250:1:1	80	10	Toluene	85	1.61	2.43	1.47	1275
5	250:1:1	90	10	Toluene	>99	1.89	2.83	1.61	1485
6	250:1:1	100	10	Toluene	95	1.84	2.72	1.72	1425
7	250:1:1	110	10	Toluene	92	1.84	2.63	1.76	1380
8	250:1:1	90	3	Toluene	43	1.04	1.24	1.21	2150
9	250:1:1	90	5	Toluene	63	1.35	1.81	1.41	1890
10	250:1:1	90	7	Toluene	86	1.18	2.46	1.43	1843
11	250:1:0	90	10	Toluene	0
12	250:1:0	90	20	Toluene	0
13	250:1:0	90	30	Toluene	15	0.27	0.43	1.15	75
14	250:1:2	90	10	Toluene	96	1.02	2.75	1.47	1440
15	250:1:4	90	10	Toluene	85	0.61	2.43	1.37	1275
16	250:1:5	90	10	Toluene	82	0.51	2.35	1.47	1230
17	250:1:10	90	10	Toluene	69	0.28	1.98	1.15	1035
18	100:1:1	90	10	Toluene	>99	0.75	2.83	1.40	594
19	300:1:1	90	10	Toluene	93	1.58	2.66	1.47	1674
20	400:1:1	90	10	Toluene	63	2.11	1.81	1.57	1512
21	500:1:1	90	10	Toluene	51	2.23	1.46	1.49	1530
22	750:1:1	90	10	Toluene	23	2.51	0.67	1.35	1035
23	250:1:1	90	10	Bulk	>99	0.81	2.83	1.64	1485
24	250:1:1	-	10	Hexane	Trace
25	250:1:1	-	10	THF	Trace
26	250:1:1	-	10	CH2Cl2	9	0.40	0.27	2.03	135
27 e	250:1:1	90	30	Bulk	trace
28 e	250:1:1	60	30	Hexane	83	1.85	1.90
29 e	250:1:1	60	30	THF	42	0.80	1.20
30 e	250:1:1	35	30	CH2Cl2	29	0.59	1.16

^a Conditions: 20 μmol [Al]; 2 mL solution. ^b Determined by ¹H NMR spectroscopy. ^c GPC data were recorded in THF vs. polystyrene standards, using a correcting factor of 0.56 [40]. ^d M_n (calcd) = (monomer/initiator) × (conversion) × 114 + 108 (molar mass of BnOH). ^e under reflux.

). Below 40 °C, there was no monomer consumption after 10 min. However, on raising the temperature from 60 to 90 °C, the conversion gradually increased from 47% at 60 °C to a maximum of >99% at 90 °C (runs 4, 5, Table 4). Further increasing the temperature up to 110 °C led to a steady decrease in the conversion. These findings contrast with observations seen for M (Chart 1) in which only 20% conversion was observed at 80 °C and 95% at 90 °C [37].

With the ε-CL:Al:BnOH ratio fixed at 250:1:1 and the run temperature at 90 °C, the influence of reaction time on the polymerization was explored (runs 5, 8–10, Table 4). Monitoring the polymerization by ¹H NMR spectroscopy at intervals of 3, 5, 7 and 10 min saw the conversion rise from 43% to approaching 100% while the molecular weight of polymer increased from 1.04 × 10⁴ to 1.89 × 10⁴ g/mol. By contrast, the TOF decreased with time implying gradual deactivation of the active species over longer run times. To investigate the importance of BnOH in the activation of C1, the polymerization runs were performed with and without benzyl alcohol over different run times but with the temperature maintained at 90 °C; the results are collected in Table 4 (runs 11–17). Without benzyl alcohol, C1 showed no activity after 20 min while after 30 min 15% conversion was observed, highlighting the induction period needed when using C1 alone. This result contrasts with the virtually 100% conversion observed after 10 min using an equimolar ratio of C1 to BnOH (run 5, Table 4). On increasing the BnOH/Al ratio from 1 to 10, the conversion gradually decreased from 99 to 69% (runs 5, 14–17, Table 4) while the M_n of the PCL lowered from 1.89 × 10⁴ to 0.28 × 10⁴ g/mol, suggesting the possible role of BnOH as a chain transfer reagent. An alternative explanation could be due to the likely intermediate ‘L1Al(OCH₂Ph)₂’ undergoing disproportionation and dimerization to form an inactive species in a manner similar to that previously reported [41].

In order to gain some information on the structural features of the PCLs, all the samples obtained by employing different amounts of BnOH were analyzed by ¹H NMR spectroscopy and by MALDI-TOF mass spectrometry. In the case where no BnOH was employed, the MALDI-TOF spectrum indicated that the polymer was comprised of linear polymers capped by two different types of end-group (α and β in Figure 3). In α (the major species) the capping group is CH₃ while in β it is L1. These findings are supported by the ¹H NMR data that reveal aromatic signals characteristic of L1 in β and a distinct singlet peak at δ 2.14 that can be assigned to the acetyl end group (CH₃C=O) in α (Figure 4). Therefore, it would seem probable that this polymerization proceeded by a coordination-insertion mechanism in which the Al-Me and Al-L1 independently initiated the ring opening polymerization by pathways I and II (Scheme 2) [42,43,44,45].

On the other hand, when one equivalent or more of BnOH was employed in combination with C1, the ¹H NMR spectra of the PCLs showed only signals typical of BnO end groups (e.g., δ 5.0 (OCH₂Ph) in Figure 5). This observation was corroborated by the MALDI-TOF spectrum that showed solely peaks corresponding to a linear chain polymer capped by a BnO group (Figure 6). Similarly, the polymers obtained by using larger amounts of BnOH possessed linear structures with BnO end groups, which were again confirmed by ¹H NMR spectroscopy and MALDI-TOF mass spectrometry (Figures S1–S4). As an additional feature of the spectra, it was evident that the more BnOH employed the lower the molecular weight of the polymer formed (runs 5, 14–17, Table 4).

With the amount of benzyl alcohol maintained at one molar equivalent, the effect of changing the amount of ε-CL on the performance of C1 was then examined. On increasing the molar ratio of ε-CL:Al from 100 to 300, all the conversions were greater than 90% with the TOF showing an upward trend (runs 5, 18, 19, Table 4). However, on further increasing the ε-CL:Al ratio between 400 and 750, a dramatic decrease of both the conversion and TOF was evident (runs 20–22, Table 4). On the other hand, the molecular weight of the PCL was found to gradually increase from 0.75 × 10⁴ to 2.51 × 10⁴ g/mol as the ε-CL:Al molar ratio was raised from 100 to 750 (runs 5, 18–22, Table 4), which can be explained by the faster rate of coordination and rate of propagation at higher monomer concentration.

The effect of solvent on the polymerizations using C1 was also studied. As a bulk polymerization was conducted at 90 °C, almost 100% conversion was observed after 10 min. The molecular weight was, however, much lower than that obtained in toluene (run 23, Table 4). When using n-hexane, dichloromethane or THF, the activity dramatically decreased and only trace amounts of polymer could be obtained (runs 24–26, Table 4). It is worthy to note that the above findings are significantly different to that found using analogue M (Chart 1). In that case, no activity was observed in bulk polymerization and moderate conversion was observed in other solvents [37]. One reason to account for these differences may be due to the improved solubility of C1 leading to more facile reactivity.

Based on the optimal conditions established for C1, the remaining five complexes, C2–C6, were also investigated for ROP; the results are collected in Table 4 alongside the data for C1. In terms of the steric properties of the aluminum-R³ group, replacing Me with an i-Bu group slightly decreased the conversion from 99 to 91% and the molecular weight from 1.89 × 10⁴ to 1.21 × 10⁴ g/mol (runs 1–3,

Table 5. The ROP of ε-CL promoted by C1–C6/BnOH ^a.

Run	Pro-Initiator (R1, R2, R3)	Conv. (%) b	Mn (×104) c	Mn (calc.) (×104) d	PDI c	TOF (h−1)
1	C1 (H, H, Me)	>99	1.89	2.83	1.61	1485
2	C2 (H, H, Et)	95	1.35	2.72	1.32	1425
3	C3 (H, H, i-Bu)	91	1.21	2.60	1.25	1365
4	C4 (Cl, H, Me)	83	1.54	2.38	1.55	1245
5	C5 (H, Me, Me)	97	0.73	2.78	1.30	1455
6	C6 (Cl, Me, Me)	82	1.24	2.35	1.48	1230

^a Conditions: 20 μmol [Al], [ε-CL]:[Al]:[BnOH] = 250:1:1, 2 mL toluene solution, 90 °C, 10 min. ^b Determined by ¹H NMR spectroscopy. ^c GPC data were recorded in THF vs. polystyrene standards, using a correcting factor of 0.56 [40]. ^d M_n (calcd) = (monomer/initiator) × (conversion) × 114 + 108 (molar mass of BnOH).

); the increased reactivity of an Al-Me over an Al-i-Bu towards BnOH may be a contributing factor. With a chloride present at the 2-position of the ligand, the conversion decreased rapidly as exemplified by C1 > C4 and C5 > C6. The reason for this observation is likely due to the electron withdrawing properties of a chloride resulting in decreased electron density at aluminum and in-turn slower reaction; steric factors could no doubt also be influential. In addition, the molecular weight of the PCLs generated by C5 and C6 (R² = Me) are slightly lower than that seen with C1–C4 (R² = H), suggesting the detrimental effect of the Cy-CMeO methyl group on chain propagation. Similar to C1, the MALDI-TOF mass spectra of the PCL generated by C2–C5 showed linear polymers capped with BnO groups as the unique polymer class. Likewise, the ¹H NMR spectra of PCLs generated by C2–C5 also clearly showed signals characteristic of a BnO group (Figures S5–S9). Indeed, these results are consistent with the polymerizations proceeding by ‘coordination-insertion’ route with ‘LAl-OCH₂Ph’ presumed as the active species. Unexpectedly, the structural analysis of the PCL obtained by C6/BnOH indicated that two polymer families were present, the major one being a linear structure capped with a PhCH₂O group and the other a linear one capped with a methoxy group. These results suggest that the different R¹ and R² substituents have little effect on the polymer structure but have large effect on their reactivity.

3.3. Mechanistic Analysis

On account of the differences in polymerization behavior between C1–C6 and its unsubstituted counterpart M (Chart 1), the mode of activation was subject to a new ¹H and ²⁷Al NMR spectroscopic investigation. In particular, toluene-d₈ solutions of C1, C1/BnOH and C1/BnOH in the presence of 10 molar equivalents of ε-CL, were examined at various temperatures.

Firstly, the variable temperature ¹H NMR experiment was performed on C1 alone; the stacked spectra are shown in Figure 7. The data reveal the signals for the ligand show some modest shift to lower field with increasing temperature. On the other hand, two separate singlet peaks (δ −0.11, −0.24) for the Al-Me’s at room temperature significantly shifted to lower field and became progressively broader on increasing the temperature. When the temperature reached 100 °C, two peaks effectively collapsed at ca. δ 0, which may be due to the onset of an exchange process at high temperature [46]. In the ²⁷Al NMR spectrum of C1, one broad peak around δ 176.83 (a) was evident at room temperature. However, on raising the temperature from 25 to 100 °C, a second more upfield peak (b) (δ 127.9) gradually grew in intensity (Figure 8). Based on these findings, we presume that C1 exists mainly as a dimeric species at low temperature which is partly transformed into a monomeric species at higher temperature in a manner similar to that previously reported [37].

We also monitored the variable temperature ¹H and ²⁷Al NMR spectra of a 1:1 mixture of C1 and BnOH; the corresponding spectra recorded between 25 and 100 °C are collected in Figure 9 and Figure 10. The ¹H NMR spectrum of C1/BnOH shows multiple upfield peaks for the Al-Me resonances at room temperature which converged into two sharp peaks on raising the temperature. By contrast, there were no major shifts of the downfield H_aryl protons over the temperature range. At room temperature, the coordinated benzyl alcohol methylene protons are inequivalent leading two mutually coupled doublets, this splitting pattern is maintained up to 70 °C above which the signals start to merge and broaden. Unexpectedly, the ratio of the PhCH₂O:Al-CH₃ protons is close to 2:6 at each temperature, which would appear to rule out a species of the type L1Al(OCH₂Ph)(CH₃). In the ²⁷Al NMR spectra of C1/BnOH, two broad peaks were observed in the temperature range 25 to 80 °C. Further increasing the temperature to 90 °C led to apparent coalescence and then at 100 °C the formation of a single broad peak. By analogy with the NMR findings for C1 alone, we propose an equilibrating mixture of dimeric and monomeric species each coordinated by intact BnOH ligands that becomes time-averaged at high temperature.

In order to gain information about the catalyst when in the presence of monomer, the ¹H and ²⁷Al NMR spectra of C1/BnOH in the presence of 10 equivalents of ε-CL were also conducted over the 25 to 100 °C temperature range; the spectra are shown in Figure 11. In all the VT ¹H NMR spectra the PhCH₂O resonance appeared as a singlet (around δ 5.0), which is indicative of a PhCH₂O end group in PCL. Hence, it is apparent that the polymerization occurred rapidly when the monomer was mixed with C1/BnOH even at room temperature. As the temperature was raised, the peaks for PCL (δ 3.94, 1.52) gradually increased as the resonance for ε-CL (δ 3.58) reduced, in accord with the higher polymerization rate at higher temperature. On the other hand, the VT-²⁷Al NMR spectra /BnOH/10 ε-CL gave a single broad resonance at δ 69.0 across the temperature range that we tentatively assign to the active species; unfortunately, this chemical shift also coincides with the probe signal. Nevertheless, related aluminum complexes containing bound α-alkoxy esters, that have been considered as the active species in the ROP, have also shown resonances around δ 70 in their ²⁷Al NMR spectra [13].

4. Conclusions

A series of soluble aluminum 5,6-dihydro-7,7-dimethylquinolin-8-olates, C1–C6, have been successfully prepared and fully characterized. The molecular structures of C1 and C4 indicate that they adopt dimeric forms in the solid state. In the absence of BnOH, C1 showed only low catalytic efficiency for the ROP of ε-CL with the spectroscopic and spectrometric data for the resulting polymer in agreement with a linear structure capped with either a methyl or a L1 group. By contrast, in the presence of BnOH, C1 exhibited excellent efficiency for the ROP of ε-CL with essentially 100% conversion in only 10 min at 90 °C. The chloro-substituted complexes C4 and C6 showed lower activity than that seen with C1–C3 and C5, highlighting the importance of both electronic and steric factors on initiator performance. All polymers displayed linear structures that were capped with PhCH₂O groups, which was confirmed by ¹H NMR spectroscopy and by MALDI-TOF mass spectrometry. The solution state properties of C1 and C1/BnOH over various temperatures have been investigated by multinuclear NMR spectroscopy and highlight the importance of both monomeric and dimeric species. In the presence of 10 molar equivalents of monomer, the active initiator generated from C1/BnOH, could be monitored by ²⁷Al NMR spectroscopy while the chain propagation by ¹H NMR spectroscopy.