Reversible On/Off Switching of Lactide Cyclopolymerization with a Redox-Active Formazanate Ligand

Redox-switching of a formazanate zinc catalyst in ring-opening polymerization (ROP) of lactide is described. Using a redox-active ligand bound to an inert metal ion (Zn2+) allows modulation of the catalytic activity by reversible reduction/oxidation chemistry at a purely organic fragment. A combination of kinetic and spectroscopic studies, together with mass spectrometry of the catalysis mixture, provides insight in the nature of the active species and the initiation of lactide ring-opening polymerization. The mechanistic data highlight the key role of the redox-active ligand and provide a rationale for the formation of cyclic polymer.

NMR spectra were recorded on Varian Mercury Plus 400, Varian Inova 500 or Bruker Avance NEO 600 spectrometers. The 1 H and 13 C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm). All electrochemical measurements were performed at ambient temperatures under an inert N2 atmosphere in THF or DCM, containing 0. 1 M [Bu4N][PF6] as the supporting electrolyte. Electrochemical measurements were performed using an Autolab PGSTAT 204 computer-controlled potentiostat and data was recorded with Autolab NOVA software (v.2.1.4). Cyclic voltammetry (CV) was performed using a three-electrode configuration comprising of a Pt wire counter electrode, a Ag wire pseudo-reference electrode and a Pt disk working electrode (CHI102, CH Instruments, diameter = 2 mm). The Pt working electrode was polished before the experiment using an alumina slurry (0.05 μm), rinsed with distilled water and subjected to brief ultrasonication to remove any adhered alumina microparticles. The electrodes were then dried in an oven at 75 °C overnight to remove any residual traces of water. The CV data was referenced by addition of ferrocene to the THF or DCM solution at the end of experiments. UV/Vis spectra were recorded in a DCM solution (≈ 10 -5 M) using an Avantes AvaSpec-2048 UV/Vis spectrophotometer. Electron paramagnetic resonance (EPR) measurements were carried out on a Bruker EMX Nano X-band (9.5 GHz) and performed under nitrogen with a sample concentration of 1 mM in degassed dichloromethane. Polymer samples for GPC and MALDI-TOF measurements were obtained by quenching the reaction with a droplet of MeOH followed by precipitation of the polymer from a DCM solution, by addition to excess hexane (3x) and the precipitated polymer was subsequently dried under vacuum. Molecular weights (Mn) and (Mw) and the dispersity (ĐM) were measured by gel permeation chromatography (GPC) using triple detection, consisting of a Viscotek RALLS detector, Viscotek Viscometer Model H502 and Schambeck RI2012,A Refractive Index detector. The separation was carried out by utilizing two PLgel 5 µm MIXED-C, 300 mm columns from Agilent Technologies at 35 °C. THF 99+%, extra pure, stabilized with BHT was used as the eluent at a flow rate of 1.0 mL/min. The samples were filtered over a 0.2 µm PTFE filter prior to injection. Data acquisition and calculations were performed using Viscotek OmniSec software version 5.0, using a refractive index increment (dn/dc) of 0.042. Molecular weights were determined based on a universal calibration curve generated from narrow dispersity polystyrene standards (Agilent and Polymer Laboratories, Mw from 645 to 3001000 g/mol). MALDI-TOF mass spectra were recorded on an AB Sciex 4800 Plus MALDI-TOF/TOF Analyzer. solutions of DHB as a matrix (20 mg/mL in THF), polymer sample (5 mg/mL in THF) and LiCl (5 mg/mL in THF) were prepared. MALDI-TOF samples were prepared by mixing the before-mentioned solutions in a matrix:sample:salt = 10:2:1 ratio. 1 µL of this solution was applied onto the MALDI target plate. Mass spectrometry by direct injection was performed on a Waters Xevo G2 QTOF spectrometer using ESI, measuring in negative ionization mode. Experimentals 1,5-diphenyl-3-para-tolyl formazanate zinc phenoxide (complex 2) 1,5-diphenyl-3-para-tolyl formazanate zinc methyl (325 mg, 0.825 mmol) was dissolved in 15 mL of toluene. The solution was layered with pentane (10 mL) containing phenol (77.0 mg, 0.818 mmol) and was left overnight to diffuse in a freezer at -30 °C. Crystals had formed overnight which were separated from the solution, washed with toluene (2 times, 2 mL) and pentane (7 times, 2 mL) and dried in vacuo, to afforded 245 mg (0.236 mmol, 57 %) of large crystals with a green metallic shine. The crystals obtained were suitable for X-ray diffraction. 1

Typical NMR scale polymerizations
In a glove box, rac-lactide (250 µmol), 1,3,5-trimethoxybenzene (25 µmol) and catalyst 2 (2.5 µmol) were added to a vial and dissolved in 0.4 mL CD2Cl2 and subsequently transferred to a screw-cap NMR tube, fitted with a PTFE/silicone septum. A solution of Cp2Co (5.0 µmol, 1 eq. per formazanate, in 100 µL CD2Cl2) was prepared and kept in a 100 µL micro-syringe fitted with a rubber stopper. The sample tube was removed from the glove box and a 1 H NMR spectrum was measured on a Varian Inova 500 spectrometer prior to 'activating' the catalyst. The Cp2Co solution (5.0 µmol, 1 eq. per formazanate, in 100 µL CD2Cl2) was added through the septum using a micro-syringe, which initiated the polymerization. Data was collected automatically using an arrayed experiment with a pre-acquisition delay ("pad"). Conversion was determined on the basis of the 1 H NMR integrations of methine peak of LA and PLA versus the integration of the aromatic signal of the internal standard TMB.

Procedure for NMR scale polymerizations with different equivalents of Cp2Co
In a glove box, rac-lactide (250 µmol), 1,3,5-trimethoxybenzene (25 µmol) and catalyst 2 (2.5 µmol) were added to a vial and dissolved in 0.4 mL CD2Cl2 and subsequently transferred to a screw-cap NMR tube, fitted with a PTFE/silicone septum. Solutions with different concentration of Cp2Co were prepared and kept in a 100 µL micro-syringe fitted with a rubber stopper.* The sample tube was removed from the glove box and a 1 H NMR spectrum was measured on a Varian Inova 500 prior to 'activating' the catalyst. The Cp2Co solution was added through the septum using a micro-syringe, which initiated the polymerization. *A larger amount of CD2Cl2 was needed to dissolve the 4 equivalents of Cp2Co, this resulted in the use of 250 µL micro-syringe. The amount of CD2Cl2 used to dissolve the rac-lactide, 1,3,5trimethoxybenzene and catalyst was scaled down, so that the total volume after addition amounted to 500 µL. Due to the smaller amount of solvent, no spectrum was recorded prior to activation of the catalyst.

Procedure for NMR scale switching studies
In a glove box, rac-lactide (250 µmol), 1,3,5-trimethoxybenzene (25 µmol) and catalyst 1 (2.5 µmol) were added to a vial and dissolved in 0.4 mL CD2Cl2 and subsequently transferred to a screw-cap NMR tube, fitted with a PTFE/silicone septum. Solutions of Cp2Co and FcPF6 were prepared and kept in a 100 µL micro-syringe fitted with a rubber stopper. The sample tube was removed from the glove box and a 1 H NMR spectrum was measured on a Varian Inova 500 prior to 'activating' the catalyst. The Cp2Co solution (5.0 µmol in 100 µL CD2Cl2, [Co]:[Zn] 1.0) was added through the septum using a microsyringe, which initiated the polymerization. Data was collected automatically every 2 minutes for a total of 60 minutes (using the array "pad" command). The sample tube was removed from the instrument, FcPF6 solution (5.2 µmol in 100 µL CD2Cl2, [Fe]:[Zn] ∼ 1.05) was added through the septum and the tube was given a good shake to mix everything thoroughly. The sample was placed back in the instrument and measurements were continued. After another 60 minutes, the same procedure was repeated with a Cp2Co solution (5.2 µmol in 100 µL CD2Cl2, [Co]:[Zn] ∼ 1.05) to reactivate the catalyst. The polymerization was monitored to approximately 70% conversion. Conversion was determined on the basis of the 1 H NMR integrations of methine peak of LA and PLA versus the integration of the aromatic signal of the internal standard TMB.

Procedure for the NMR scale stability of the OFF-state
In a glove box, rac-lactide (250 µmol) and catalyst 2 (2.5 µmol) were added to a vial and dissolved in 0.4 mL CD2Cl2 and subsequently transferred to a J. Young NMR tube. A Cp2Co solution (5.0 µmol in 100 µL CD2Cl2, [Co]:[Zn] 1.0) was added using a micro-syringe, which initiated the polymerization, after which the NMR tube was taken out of the glove box and placed in the NMR (Varian Inova 500) and spectra were collected automatically every 5 minutes up to 60 minutes after initiation (using the array "pad" command). The sample tube was removed from the instrument, taken into the glove box, FcPF6 solution (5.2 µmol in 100 µL CD2Cl2, [Fe]:[Zn] ∼ 1.05) was added and the tube was given a good shake to mix everything thoroughly, after which another spectrum was measured. The sample was stored in the glove box overnight, and the next morning a spectrum was measured (still in the OFF-state) after a total of 18 hours. After 17 hours in the OFF-state, Cp2Co solution (5.2 µmol in 100 µL CD2Cl2, [Co]:[Zn] ∼ 1.05) was added to the tube (in glove box) to reactivate the catalyst. The polymerization was monitored to approximately 95% conversion.

Procedure for the Mn versus conversion plot
In a glove box, rac-lactide (2.5 mmol) and catalyst (12.5 µmol) were dissolved in 5 mL DCM. 1.05). Conversion was determined on the basis of the 1 H NMR integrations and the number average molecular weight (Mn) and dispersity were measured by GPC analysis.

Polymerizations with higher LA loading
In a glove box, rac-lactide (0.625 or 1.25 mmol) and catalyst (2.5 µmol) were dissolved in DCM (to yield a 0.5 M solution in LA). The polymerization was initiated by the addition of Cp2Co (5 µmol, [Co]:[Zn] 1.0) and after 14 hours, aliquots were taken to assess the conversion. The reaction was worked up when conversion had reached > 90%. The polymers were purified by precipitated in hexane (3x).

X-ray Crystallography
Suitable crystals of 2 and [LZn(OPh)2][Cp*2Co] (A) were mounted on a cryo-loop in a glove box and transferred, using inert atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. Data collection and reduction was done using the Bruker software suite APEX3. 11 The final unit cell was obtained from the xyz centroids of 9902 (2), 9994 (A), reflections after integration. A multi-scan absorption correction was applied, based on the intensities of symmetryrelated reflections measured at different angular settings (SADABS). 11 The structure was solved by dual space methods using the program SHELXT. 12 Structure refinement was performed with the program package SHELXL. 13 The hydrogen atoms were generated by geometrical considerations and constrained to idealized geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. For complex 2, the toluene solvent molecule was disordered over an inversion center. The symmetry was disregarded in the refinement (negative PART number) and its phenyl ring was constrained to be a regular hexagon (AFIX 66). Three reflections with (Iobs/Icalc)/Sigma(W) > 10 were omitted from the final refinement. For compound A, initial refinement yielded the atom positions of the molecules of interest and one CH2Cl2 solvent molecule. However, an area with high residual electron density was present in the difference Fourier map, which indicated the presence of highly disordered solvent molecules. The contribution from this area was removed by a PLATON/SQUEEZE 14 run which identified a solvent accessible void containing 42 e -(possibly a pentane molecule), subsequent refinement proceeded smoothly. Crystal data and details on data collection and refinement are presented in Table S2.       To evaluate whether 2 would retain its dimeric structure in solution, ccDOSY (convection compensated) was performed on both compound 1 and 2. Compound 1 was set as the benchmark for the approximate size (r) for the monomeric species. The hydrodynamic radius of 2 calculated from the DOSY data is significantly larger than the one found for 1, from which we conclude that that 2 remains dimeric in solution.

Procedure:
Stock solutions of 2 (10.4 mg in 1 mL DCM) and Cp2Co (9.5 mg in 500 µL DCM) were prepared in a glove box. Sample preparation consisted of transferring 250 µL of stock solution 2 to a vial, to which the appropriate amount of stock solution Cp2Co was added, and the total volume was topped up with fresh DCM to 0.5 mL. The samples were transferred to an EPR tube, sealed with a cap, and wrapped with parafilm. The samples were removed from the glove box and measured on a Bruker EMX Nano X-band spectrometer. Double integration of the EPR spectra was performed using the Bruker Xenon software package and corrected for Q-value differences. Figure S20. EPR spectra of (BDI)Zn(OiPr) + Cp2Co (left) and lactide + Cp2Co (right) in DCM.      . Absorption spectra for 2 (black), 2 + 1 eq. Cp2Co (red) and 2 + 1 eq. Cp*2Co (blue).

Procedure:
A polymerization reaction was initiated ([LA]:[Zn] = 10, [Co]:[Zn] = 1) in the glovebox. An aliquot from the reaction mixture was taken using a microsyringe, which was subsequently capped with a rubber stopper. The microsyringe was attached to the mass spectrometer (no column) and the aliquot was directly injected into the ionization chamber. Spectra were measured in negative ion mode and summed over 30 seconds of measuring. The resulting spectrum shows a well-defined envelope, starting at 377.03 Da, with increments of 72 Da. This is in agreement with growing polymer chain attached to a (formazanate)zinc fragment

Characterization of the Resulting Polymers
Gel Permeation Chromatography Figure S28. Representative GPC traces for PLA obtained using 2. Sample measured in THF at 35°C and reported as absolute molecular weight (Table 1).   The polymerization of rac-lactide with the ([BDI]ZnO i Pr)2 catalyst, as described by Coates et al. 15 , was performed as a control experiment. The MALDI-ToF spectrum of the product obtained from polymerization showed a clearly identifiable envelope corresponding to the linear polymer with isopropyl initiator (next to an envelope corresponding to polymer with methanol as end group, produced during the quenching of the polymerization with methanol). This result substantiates that the formation of the cyclic polymer using our system does not come from a measurement error.