Ring-opening polymerization of cyclic esters by 3- and 4-pyridinyl Schiff base Zn(II) and Cu(II) paddlewheel complexes: kinetic, mechanistic and tacticity studies

2.1 Materials and reagents

Schlenk techniques were used for all manipulations using argon as the choice gas. All solvents, tetrahydrofuran (THF) 98%, dichloromethane (DCM) (99%) and hexane (98%) were obtained from Sigma-Aldrich, all dried from a sodium-benzophenone mixture. Metal salts [Cu(OAc)₂·2H₂O and Zn(OAc)₂·2H₂O], 3-pyridinecarboxaldehyde (98%), 4-pyridinecarboxaldehyde (98%), p-toluidine, 4-chloroaniline, 2,6-dimethylaniline (99%), 2,6-diisopropylaniline (97%), aniline (99%), ε-caprolactone (ε-CL) (97%), ʟ-lactide (ʟ-LA) (98 %) and rac-lactide (rac-LA) (98%) were also obtained from Sigma-Aldrich and used without further purification.

2.2 Instrumentation

¹H- and ¹³C NMR spectra were measured at room temperature using a Bruker 400 MHz spectrometer. ¹H- and ¹³C NMR data were recorded in CDCl₃, with a residual internal solvent signal of CDCl₃ as 7.26 and 77.00 ppm, respectively. IR spectra were obtained on a PerkinElmer Universal ATR spectrum 100 FT-IR spectrometer. Mass spectra of the compounds were obtained from a Water synapt GR electrospray positive spectrometer. UV spectra were obtained using UV-3600 Shimadzu UV–VIS-NIR spectrophotometer.

2.3 Synthesis of the Schiff base ligands

Solvent-free grinding was employed for the synthesis of ligands L1 - L10 using the procedure described in literature (Njogu et al., 2017). Pyridinecarboxaldehyde (1 mmol, 0.107 g) mixed with p-toluidine (1 mmol, 0.107 g), 4-chloroaniline (1 mmol, 0.127 g), 2,6-dimethylaniline (1 mmol, 0.121 g), 2,6-diisopropylaniline (1 mmol, 0.177 g) and aniline (1 mmol, 0.093 g) were grinded with a mortar and pestle for 10 – 15 min. The products were dried in vacuo to completely remove the water. (E)-1-(pyridin-4-yl)-N-(p-tolyl)methanimine L1, (E)-N-(2,6-diisopropylphenyl)-1-(pyridin-4-yl)methanimine L2, (E)-N-(4-chlorophenyl)-1-(pyridin-4-yl)methanimine L3, (E)-N-(2,6-dimethylphenyl)-1-(pyridin-4-yl)methanimine L4, (E)-N-phenyl-1-(pyridin-4-yl)methanimine L5, (E)-1-(pyridin-3-yl)-N-(p-tolyl)methanimine L6, (E)-N-(2,6-diisopropylphenyl)-1-(pyridin-3-yl)methanimine L7, (E)-N-(4-chlorophenyl)-1-(pyridin-3-yl)methanimine L8, (E)-N-(2,6-dimethylphenyl)-1-(pyridin-4-yl)methanimine L9 and (E)-N-phenyl-1-(pyridin-4-yl)methanimine L10 shown in Fig. 1 were obtained in excellent yield (97–98%).

Close

Fig. 1

Schiff base ligands L1 – L10 used in this study.

Export to PPT

2.4 Synthesis of metal complexes

The Schiff base metal complexes (1 – 13) were prepared in methanol solution at ambient temperature via the reaction between the metal(II) acetate and the respective Schiff base ligand synthesized using reported literature protocol (Sarma et al., 2009). Generally, the metal carboxylates were obtained by the gradual addition of either ethanolic (ca. 10 mL) Zn(II) acetate (1 mmol, 0.219 g) or Cu(II) acetate (1 mmol, 0.2 g) to a solution of ligands 1–10 (1 mmol) respectively in ethanol (ca. 10 mL) under constant stirring for 6 h. The resulting solids were filtered using a vacuum filter to obtain the crude product, which was recrystallized from dichloromethane (DCM).

2.5 General procedure for ring-opening polymerization of ɛ-caprolactone (ɛ-CL) and lactide (LA)

All polymerization reactions were carried out using Schlenk techniques with argon as the gas. The required amount of the respective complex was added into a Schlenk tube containing the monomer ε-CL (1.14 g, 0.01 mol) and LA (1.44 g, 0.01 mol) in bulk and also in toluene (5 mL) immersed in a preheated oil bath at 110 °C to initiate the polymerization reaction. Samples were withdrawn at regular intervals using a syringe, and quenched quickly by dissolving in cooled CDCl₃ in an NMR tube for kinetic experiments. The conversion of polymerization was determined by ¹H NMR spectroscopy and was evaluated using Eqs. (1) and (2). For ε-CL, signal intensities at 4.0 ppm (I_4.0) for OCH₂ protons of PCL and signal intensities at 4.2 ppm (I_4.2) for ε-CL monomer using Eq. (1).

For PLA, the integration values of the methine proton of the LA monomer and that of the PLA polymer were obtained using Eq. (2).

The peak areas of the polymer and monomer were used to deduce the conversion. The observed rate constants were extracted from the slope of the line of best fit from the plot of ln( ${[M]}_o$ / ${[M]}_t$ ) vs t.

2.6 Polymer characterization by size exclusion chromatography (SEC)

Molecular weights and polydispersity indexes (Ð) were determined by size exclusion chromatography (SEC) at the central analytical facilities, Stellenbosch University, South Africa. The samples were dissolved in tetrahydrofuran (THF) stabilized with butylated hydroxytoluene (BHT), giving a sample with a concentration of 2 mg ml⁻¹. Sample solutions were filtered via a syringe through 0.45 mm nylon filters before being subjected to analysis. The SEC instrument consists of a Waters 1515 isocratic HPLC pump, a Waters 717plus auto-sampler, a Waters 600E Paper system controller (run by Breeze Version 3.30 SPA) and a Waters in-line Degasser AF. A Waters 2414 differential refractometer was used at 30 °C in series along with a Waters 2487 dual wavelength absorbance UV/Vis detector operating at variable wavelengths·THF (HPLC grade stabilized with 0.125 % BHT) was used as the eluent at flow rates of 1 mL min⁻¹. The column oven was kept at 30 °C and the injection volume was 100 mL. Two PLgel (Polymer Laboratories) 5 mm Mixed-C (300 × 7.5 mm) columns and a pre-column (PLgel 5 mm Guard, 50 × 7.5 mm) were used. Calibration was done using narrow polystyrene standards ranging from 580 to 2 × 10⁶ g mol⁻¹. All molecular weights were reported as polystyrene equivalents.

2.7 Molecular structure determination by single-crystal X-ray analysis

Crystal evaluation and data collection were done on a Bruker Smart APEXII diffractometer with Mo Kα radiation (I = 0.71073 Å) fitted with an Oxford Cryostream low-temperature apparatus operating at 100 K for 7, 9, 10 and 11. Reflections were collected at different starting angles and the APEXII program suite was used to index the reflections (Bruker, 2009a). Data reduction was performed using the SAINT (Bruker, 2009c) software, and the scaling and absorption corrections were applied using the SADABS (Bruker, 2009b) multi-scan technique. The structures were solved by the direct method using the SHELXS program and refined using the SHELXL program (Sheldrick, 2008). Graphics of the crystal structures were drawn using OLEX2 software (Dolomanov et al. 2009). Non-hydrogen atoms were first refined isotropically and then by anisotropic refinement with the full-matrix least squares method based on F² using SHELXL (Sheldrick, 2008). The crystallographic data and structure refinement parameters for complexes 7, 9, 10 and 11 are given in Table 1.

3.1 Synthesis and spectroscopic studies of Schiff base ligands

The Schiff base ligands used in this work were prepared via mechanochemical procedure in which the aldehydes along with substituted aniline were ground for ten minutes using a mortar and pestle to obtain L1 - L10 in good yields of 97 – 99 % (Scheme 1). The condensation reaction was followed using thin layer chromatography (TLC) and Fourier Transform Infrared (FT-IR) spectroscopy until the N—H and C⚌O bands from the reactants disappeared, confirming the completion of the reaction. The crude product was dried in vacuo and further characterized by FT-IR spectroscopy which gave spectra with characteristic imine (C⚌N) bands around 1620–1640 cm⁻¹. The ¹H and ¹³C NMR analyses of the Schiff base ligands revealed a singlet peak at about 8.20–8.44 ppm and 156 ppm, respectively, which corresponds to the formation of the imine bond (HC⚌N) commonly reported in the literature for Schiff bases (Dong et al. 2016; Munzeiwa et al., 2019; Köppl and Alt, 2000). The ¹H and ¹³C NMR spectra of L1- L10 are available in the supplementary information (Figs. S1 - S20).

Close

Scheme 1

Synthetic route for the preparation of Schiff base ligands L1 – L10.

Export to PPT

3.2 Synthesis and spectroscopic studies of Zn(II) and Cu(II) carboxylates

Zn(II) and Cu(II) acetate (1:1) were added to methanolic solutions of the Schiff base ligands and stirred at room temperature for 6 h to produce the metal carboxylate complexes in good yields between 75 and 94% (Scheme 2). The characterization of the carboxylate complexes by IR, mass spectroscopy, and elemental analysis points to a dinuclear carboxylate system [MLn(CH₃COO)₂]₂. The Zn(II) complexes were further characterized with ¹H- and ¹³C NMR (Figs. S21 - S26).

Close

Scheme 2

Synthesis of Zn(II) and Cu(II) pyridinyl complexes 1–13.

Export to PPT

The ESI-MS spectra of complexes 1 – 13 gave molecular and fragment ions peaks confirming a metal to ligand to acetate ratio of M₂L₂Ac₂. For instance Zn(II) complex 1 revealed m/z peaks at 765.14, 701.09 and 515.09 which corresponds to (M⁺ + 5H), [M⁺ – C₂H₃O₂] and Zn(L1)(C₂H₃O₂) ions, respectively. The base peak observed at 701.09 matched the [M⁺ – C₂H₃O₂] fragment (Fig. S37). Similarly, Cu(II) complex 10 revealed m/z peaks at 894.9 and 793.1 corresponding to the molecular ion (M⁺) and the [M⁺ – C₄H₆O₄] ion, respectively. The base peak observed at 595.5 matched the Cu(L7)₂ fragment which suggests the formation of the stable mononuclear species in dichloromethane solution (Fig. S38).

Analysis of the metal carboxylates by FTIR spectroscopy revealed bands between 323 and 373 cm⁻¹ corresponding to the M—N_py stretching frequency and between 676 and 714 cm⁻¹ corresponding to M—O stretching frequencies. This is an indication of the coordination of the metal center to the pyridinyl N atom and the acetate O atom. The syn-syn bidentate bands of the coordinated carboxylate were observed between 1421 and 1422 cm⁻¹ and between 1602 and 1626 cm⁻¹, respectively, in all the complexes, which resulted in a Δυ that lie in the range 180 ≥ Δυ ≤ 205 similar to reported paddle wheel geometry (Akpan et al. 2016b; Zhu et al. 2010; Selvakumar et al., 2008; Vagin et al., 2007).

Absorption spectra of 1, 3 and 5 in chloroform (Fig. 2a) and 2, 4, 6–13 in dichloromethane (Fig. 2b) were analyzed in the 200–450 nm wavelength range. Bands observed between 230 and 243 nm and 264 and 265 nm were assigned to π-π* transition and those in the range 322 to 336 nm assigned to n-π* as summarized in Table S1. The absorption spectra of the Zn(II) and Cu(II) complexes were found to be influenced by the nature and position of the ligand substituents. For example, three λ_max values were generally observed with the complexes whose ligand contained both electron-donating (1, 2 and 8) and electron-withdrawing substituents (5 and 6) at the para position whereas two λ_max values were obtained for ortho substituted and unsubstituted systems. The n-π* transition of the ortho substituted complexes (3, 7, 10, and 12) experienced a hypochromic shift whereas the para substituted complexes (1, 5, 6, and 11) experienced a hyperchromic shift.

Close

Fig. 2

Electronic absorption spectra of (a) complexes 1, 3 and 5 in chloroform (b) complexes 2, 4, 6 – 13 in DCM.

Export to PPT

The ¹H NMR analysis of the Zn(II) complexes displayed a characteristic singlet peak between 8.24 and 8.51 ppm corresponding to the imine proton (HC⚌N) of the Schiff base ligands. Additionally, the unique singlet peak observed around 2.07 ppm was assigned to the methyl proton of the bridging acetate, which confirmed the formation of the carboxylate complexes (Akpan et al. 2016b). The aromatic protons of the pyridinyl moiety shifted downfield upon coordination to the Zn(II) metal via the nitrogen atom in the pyridinyl pendant (Table S2). For instance, the ¹H NMR spectrum of complex 1 (Fig. S21) having a methyl substituent on the para position experienced the most significant shift (Table S2, entry 1) while complex 5 having a chloro substituent experienced the least shift (Table S2, entry 3). Similarly, ¹³C NMR spectrum of complex 1 (Fig. S24) contained an additional carbon peak of the methyl of acetate observed at 22.76 ppm and the carbonyl peak at 180.26 ppm which established the formation of the complex with similar trends observed for all zinc complexes. The spectroscopic data obtained for the Zn(II) carboxylate complexes are consistent with results reported for similar carboxylate complexes (Akpan et al. 2016b; Kumar et al., 2010). From the spectroscopic analysis, it is quite clear that all complexes coordinated via the pyridinyl N atom, leaving the imine N atom uncoordinated as we intended (Adeleke et al., 2020).

3.3 Molecular structure of 7, 9, 10 and 11

The crystals of complexes 7, 9, 10 and 11 were all acquired by vapor diffusion of hexane into complex dichloromethane solutions. Their molecular structures obtained by single-crystal X-ray diffraction studies and are displayed in Fig. 3a - d with selected bond distances and angles summarized in Table 2. All complexes are neutral, and paddlewheel in the solid-state, and each has only half a molecule in their asymmetric units. The other half in each case is generated through a center of inversion situated at the center of a Cu(II)--- Cu(II) supported interaction located at (½ ½ ½), (0 1 0), (½ 1 0) and (0 ½ ½) for complexes 7, 9, 10 and 11, respectively. These cuprophilic Cu(II)-----Cu(II) interactions have distances equal to 2.6141(1) Å, 2.6280(3) Å, 2.6252(2) Å and 2.6270(3) Å which aligns with similar paddle wheel conformations in literature (Munzeiwa et al., 2018; Mosae Selvakumar et al. 2013; Attandoh et al., 2014). In all five complexes the carboxylate anions bridge the two Cu(II) centers in a bidentate bridging mode. The ligand coordinates to the Cu(II) center through the pyridinyl N atom, with Cu—N distances of 2.170(5), 2.1738(18), 2.2104(12) and 2.1774(13), for 7, 9, 10 and 11, respectively. In doing so, the geometry around the Cu(II) centers is distorted octahedron in which four O atoms occupy the base of the equatorial positions while the pyridinyl N atom occupies the apex of the pyramid. The O—Cu—O bond angles are near orthogonal and so are the N—Cu—O bond angles.

Close

Fig. 3

Molecular structure of complexes (a) 7 and (b) 9, (c) 10 and (d) 11 with thermal ellipsoids drawn at 50% probability level. Hydrogen atoms are omitted for clarity.

Export to PPT

Complexes 9 and 11 are isostructural considering the similarity in the unit-cell parameters even though the unit-cell volume of 9 is slightly higher than that of 11 (Table 1). A quantitative isostructural analysis was obtained using Mercury (CSD version 5.41) which gave a root mean square deviation of 0.0347 Å from an overlay of the two structures (Fig. S27). Further to this, the molecular packing of complexes 9 and 11 (Figs. S28 and S29) as viewed along the crystallographic a-axis also show a level of similarity.

4.1 Kinetics of ROP reactions of ε-CL and LAs

A common observation reported for paddlewheel carboxylates is the slow onset of their polymerization in bulk at elevated temperature (Attandoh et al., 2014; Akpan et al. 2016b; Munzeiwa et al., 2018). This is because of the conformation changes experienced by the catalyst in the formation of the transition state (Motreff et al. 2012). Kinetic studies of samples withdrawn at regular intervals are usually conducted to gain insight into the polymerization mechanism and reaction rate. The polymerization rate of the Cu(II) complexes obtained from the kinetic experiment showed that steric and electronic effects of the ligand substituents slightly influenced catalytic activity though they are situated away from the metal center. Kinetic analysis of the carboxylate complexes showed a pseudo-first order relationship with respect to ε-CL and LAs as shown by the semi-logarithmic plot of ln([M]₀/[M]_t) vs. t in Fig. 4 and S30 – S32. This is in accordance with Eq. (3) where $k_{\mathit{app}}=k$ _p $\left[I\right]$ ^y; $k$ _p = rate of chain growth, I = initiator, and y = the order of reaction.

Close

Fig. 4

Pseudo-first-order logarithmic plot for the bulk polymerization of CL at 110 °C by complexes 1, 3 and 5.

Export to PPT

The kinetic analysis of the Zn(II) complexes showed that the polymerization rate of complex 5 (Table 3, entry 5) with a chloro substituent is higher than complexes 1 and 3 (Table 3, entries 1 and 3) with electron-donating substituents. Generally, the activity decreased with a decrease in electron withdrawing effects because of reduced metal nucleophilicity, which retarded monomer biding for activation. Related studies have also shown varying correlations between the substituents effects and activity using different ligands (García-Valle et al. 2015; Hormnirun et al. 2006; Nomura et al. 2002). The catalytic activity of the Cu(II) complexes followed the order: 7 > 2 > 9 > 12 > 4 > 13 > 10 > 8 > 11 > 6 showing that the polymerization rate of complexes 8 and 13 (Table 3, entries 8 and 13) with no substituent are higher than complexes 6 and 11 (Table 3, entries 6 and 11) with electron-withdrawing substituent but lower than complexes 2 and 9 (Table 3, entries 2 and 9) with electron-donating substituents. The polymerization rate of complexes 7 and 9 with less bulky ortho substituent are higher than complexes 4 and 10 with bulky ⁱPr substituent. Among the Cu(II) complexes, the polymerization rate of complex 7 with the ortho-substituted electron-donating methyl group had the highest. The steric hindrance in the ligand framework of a metal complex either decrease the catalytic activity by reducing the monomer coordination or increase it by stabilizing the active species (Qin et al. 2020; Wei et al. 2019). In this case, since the substituent are far from the metal center little crowding is expected. This can be observed from the slight difference in reaction rates of complexes 9 and 10.

Complex 5 being the most active was selected for the polymerization of rac-LA and _L-LA at 110 °C in toluene using a [monomer]/[catalyst] ratio of 200:1. A maximum conversion of 98 % was obtained between 5 and 9 h with an apparent rate constant (k_app) of 0.2513 h⁻¹ and 0.1518 h⁻¹ for rac-LA and _L-LA polymerization, respectively (Fig. S30). The polymerization rate of rac-LA is higher than that of _L-LA, a trend that has been reported (Pietrangelo et al. 2010). This has been linked to the ROP mechanism which favors racemic enchainment which is encoded by the monomer and/or the polymer stereocenters resulting in alternate _D- and _L-LA monomer insertion.

A plot of ln k_app vs. ln [5] using various catalyst concentrations at a fixed monomer concentration (Table 4) was employed to unravel the order of reaction with regards to complex 5. The reaction order extracted from the slope of the line of best fit was found to be 0.75 (Fig. 5). Fast, polymerization kinetics have been reported mainly, using some zinc alkoxide (Stasiw et al. 2017) and in situ generated copper alkoxide catalyst (Kwon et al. 2015).

Close

Fig. 5

Order of reaction determination from the plot of lnk_app against ln[5].

Export to PPT

4.2 Effect of solvent and temperature on ε-CL polymerization

Complexes 5 and 7 (Zn(II) and Cu(II)) having superior activities were chosen as the model systems to understand better the structure (metal–ligand) effects on activity. Complex 5 was employed for the polymerization of ε-CL at different temperatures and complex 7 was used to investigate the effect of different alcohol co-initiators. The activity of complex 7 was further examined in the presence of different alcohol initiators at 110 °C using a [M]:[I]:[ROH] ratio of 200:1:1 (Fig. 6).

Close

Fig. 6

Kinetic plots of ln[CL]₀/[CL]_t vs. t for complex 7 in the polymerization of ε-CL using alcohol initiators.

Export to PPT

¹H NMR analysis of samples withdrawn at regular intervals confirmed the formation of alkoxide species (Fig. S33) which improved the catalytic activity obtained with benzyl alcohol, ethanol and isopropyl alcohol (Table 4 entries 7 – 9) relative to the bulk polymerization (Table 3 entry 9). The co-initiation trend followed the order EtOH > BnOH > t-BuOH, and this is consistent with reported trend using Sn(OR)₂ (Sattayanon et al. 2013; Sattayanon et al. 2015). From theoretical and available experimental data, they observed that the polymerization rate depends on three factors:hydrocarbon chain length, number and size of branched chains and number of coordinated alkoxide groups. An increase in hydrocarbon chain length, number and size of branched chains in the alkoxide results in greater steric hindrance thereby retarding monomer coordination. In addition, an increase in the number of coordinated -OR groups and the alkyl chain length decreases the positive charge of the nucleophilic center, hence making it less susceptible to monomer coordination resulting in reduced polymerization rates.

The molecular weight of PCL obtained with t-butanol co-initiator was higher than the other alcohol initiators even though it had the lowest activity recorded (Table 4 entry 10). This observation is linked to the steric hindrance present in the molecule. The molecular weight of PCL at 110 °C is much higher than that at lower temperatures 80 and 90 °C, at similar conversion. The lower Mw at lower temperature was a result of the long reaction times required which induced transesterification reactions, while at 110 °C the polymerization is quantitative after a shorter period (Stridsberg and Albertsson, 1999).

The polymerization of ε-CL was examined at different temperatures (80 °C to 110 °C) using the most active zinc catalyst complex 5. A marked reduction in the rate constant from 0.4043 h⁻¹ at 110 °C to 0.0242 h⁻¹ at 80 °C confirmed the temperature dependence of the reaction. Longer induction periods observed at 80 °C and 90 °C resulted in lower k_app of 0.0242 h⁻¹ (Table 4 entry 1) and 0.0878 h⁻¹ (Table 4 entry 2), respectively relative to the k_app of 0.1186 h⁻¹ (Table 4 entry 3) and 0.4043 h⁻¹ (Table 4 entry 4) obtained at 100 °C and 110 °C, respectively. An overall activation energy barrier (E_a) of 98.34 kJ mol⁻¹ was obtained from the slope of the Arrhenius plot shown in Fig. 7. A large energy barrier of this magnitude is due to the strong M—O bond of the bridging acetate co-ligand (Munzeiwa et al., 2018). The enthalpy (ΔH) and entropy (ΔS) of activation were obtained from the slope of the Eyring plot as 95.28 kJ mol⁻¹ and −7.47 JK⁻¹ mol⁻¹, respectively, both of which conforms to an ordered transition (Fig. 7).

Close

Fig. 7

Arrhenius plot of lnk vs T⁻¹ and Eyring plot of ln(k_app/T) against T⁻¹ for the bulk polymerization of ε-CL initiated by 5.

Export to PPT

4.3 Molecular weight and molecular weight distribution of polymers.

The number molecular weight (M_n) and polydispersity index (Ð) of PCL and PLA obtained from gel permeation chromatography (GPC) with adequate corrections as prescribed by Mark Houwink were compared to those obtained from ¹H NMR analysis and summarized in Table 5. Low M_n polymers of 0.83–3.48 kg mol⁻¹ and 1.28–2.39 kg mol⁻¹ were obtained for PCL and PLA, respectively, with broad Ð of 1.7–2.1 indicating a level of trans-esterification during the ROP reactions (Dubois et al. 1991). (Ikpo et al. 2012) The M_n of the PCL obtained from GPC and NMR analysis increased as the [M]₀:[I]₀ increased from 100:1 to 400:1. This behavior can be alluded to the fact that there are few polymer chains propagating per initiating species at low catalyst loading, hence they grow much longer (Table 5 entries 1–4). The expected molecular weights slightly deviate from the experimental values obtained from GPC and NMR. This is as a result of self-termination intermolecular transesterification chain transfer reactions (vide infra) which kill the growing chains (Dubois et al. 1991; Ikpo et al. 2012). Also, molecular weights from GPC are comparable to those deduced from NMR and the inconsistency, hinges on separation mode in GPC which is based on polymer hydrodynamic volume rather than actual molecular weight (Balke and Cheng, 1991).

4.4 Stereochemistry of synthesized PLA

Homo-nuclear decoupled ¹H and ¹³C NMR spectroscopy is an established technique for the structural analysis of PLA (Spassky et al. 1996). The tacticity of the PLA is obtained by inspection of the tetrad sequence of the methine region of the ¹H NMR in conjunction with the methine and carbonyl carbon region of the ¹³C NMR (Fig. S36). Highly isotactic poly(_L-LA) is produced by the ROP of _L-LA if no side reactions take place during polymerization. However, zinc based ROP of rac-LA resulted in isotactic PLAs which were dependent on the electronic effects of the para substituent (Stasiw et al. 2017). The Homo-nuclear decoupled ¹H NMR spectra of the synthesized PLA obtained from _L-LA by complex 5 (Fig. 8) revealed pure isotactic sequence (intense iii tetrad peak at 5.12 ppm) and a satellite peak at 5.16 ppm (sii/iis syndioatactic diads) pointing to metal induced racemization (Gadomska-Gajadhur and Ruśkowski, 2020). Similarly, the methine and carbonyl regions of the ¹³C NMR were characterized by two signals observed at about 69.04 and 169.60 ppm, respectively (Fig. S36). Thus, we postulated that Zn(II) metal center in Shiff base complexes could induce racemization reactions by introducing the _D-LA monomer in the polymer chain thereby resulting in slight syndiotactic diads (Gadomska-Gajadhur and Ruśkowski, 2020). With the rac-LA monomer the peaks observed in the range 5.15 – 5.17 ppm and 5.21 – 5.22 ppm in the decoupled ¹H NMR spectrum were assigned to isi, iii, iis/sii and sii, iis tetrads respectively (Fig. 8) (Chabot et al. 1983; Chisholm et al. 1999), which point to heterotactic PLAs with P_r = 0.68. The probability of heterotactic (P_r) enrichment was computed from intensities from homodecoupled ¹H NMR spectra, using the formula, P_r = 2I₁/[I₁ + I₂], where I₁ =(sis + sii) and I₂ = (iis + iii + isi) (Cho et al., 2016; Honrado et al. 2015). The methine and carbonyl carbon peaks were observed at 69.02 and 169.19 ppm in the ¹³C NMR spectrum (Fig. S36), confirming heteroatactic stereo-block PLA.

Close

Fig. 8

Homo-nuclear decoupling of poly(LA) initiated by complex 5 at room temperature in CDCl₃ (400 MHz). Reaction conditions: [CL]₀:[I]₀ = 200:1, bulk, T = 110 °C.

Export to PPT

4.5 End group analysis and mechanistic study

The end group analysis and mechanism of ε-CL and LA polymerization obtained from ¹H NMR and ESI-MS analysis is majorly described by the activated monomer mechanism (AMM) and the coordination insertion mechanism (CIM) (Liu et al. 2009; Ajellal et al. 2010; Marlier et al. 2016). The AMM is characterized by the inclusion of an exogenous alcohol as a co-initiator, which activates the monomer upon coordination, whereas the CIM is characterized by the direct coordination of the cyclic ester monomer to the active site of the catalyst (Munzeiwa et al., 2019).

A CIM mechanism is proposed initially seeing the monomer becomes activated when the catalytic center generates a monomer-catalyst activated transition state that is preceded by insertion of monomer into the M—O bond; giving the first generation alkoxide macro initiator (Scheme 3). There are two possible initiation pathways for the paddlewheel complexes: the mononuclear (A) and binuclear (B). In the presence of an alcohol, path (A) is projected, where the two acetates would have been substituted by the alcohol to generate the alkoxide in situ. In a dinuclear indium system by Fang et al (Fang et al. 2013), a binuclear type preference was found to be the most probable pathway and they alluded the behavior to high energy barrier of dissociation to generate the active species. In order to understand, the preferred pathway, theoretical Density-functional theory (DFT) calculations are underway, and the details of which are not yet conclusive at this stage.

Close

Scheme 3

Proposed mechanism pathways for ROP using paddlewheel acetate complexes.

Export to PPT

The ¹H NMR spectra of PCL and PLA synthesized using complex 5 (Figs. 9 and S34) revealed the OCH₂ peak at 3.65 and 4.87 ppm, respectively. The additional peak observed around 2.10 ppm was assigned to the methyl proton of the acetate capped to the end of the polymer. This observation points to a coordination insertion mechanism. Similarly, the ESI-MS analysis of the synthesized PLA obtained using complex 5 revealed fragment ions with successive peaks differing by a mass of 0.072.07 kg mol⁻¹ corresponding to a half lactide unit (Fig. S35). For instance, a mass of 0.734 kg mol⁻¹ contains a total of ten (n = 10) half lactide moiety in the polymeric chain (Fig. 9). Characteristic m/z peaks for other PCL and PLA in the presence and absence of alcohol initiators are provided in Figs. S39 – S42. However, the acetate group cannot be accounted in the ESI-MS, though it utilize soft ionization technique, collision-induced dissociation reaction may occur, and the anhydride can dissociate leading to the elimination of the acetate (Demarque et al. 2016).

Close

Fig. 9

The ¹H NMR spectrum of poly(ε-CL) initiated by complex 5 at room temperature in CDCl₃ (400 MHz). Reaction conditions: [CL] ₀ :[I]₀ = 200:1, bulk, T = 110 °C.

Export to PPT