Phosphasalen Indium Complexes Showing High Rates and Isoselectivities in rac‐Lactide Polymerizations

Abstract Polylactide (PLA) is the leading bioderived polymer produced commercially by the metal‐catalyzed ring‐opening polymerization of lactide. Control over tacticity to produce stereoblock PLA, from rac‐lactide improves thermal properties but is an outstanding challenge. Here, phosphasalen indium catalysts feature high rates (30±3 m −1 min−1, THF, 298 K), high control, low loadings (0.2 mol %), and isoselectivity (P i=0.92, THF, 258 K). Furthermore, the phosphasalen indium catalysts do not require any chiral additives.


Experimental Section
All solvents and reagents were obtained from commercial sources (Aldrich, VWR and Strem) and used as received unless stated otherwise. For organometallic reactions, tetrahydrofuran (THF), hexane and cyclohexane solvents were refluxed over sodium/benzophenone and stored over activated molecular sieves under nitrogen. Diethylether and toluene used for ligand preparation or as a recrystallisation solvent were taken directly from an MBraun MB-SPS 800 Solvent Purification system and stored over activated molecular sieves. THF-d8, toluene-d8 and benzene-d6 deuterated solvents were distilled over sodium/benzophenone and stored over activated molecular sieves under nitrogen. All dry solvent and reagents were degassed by several freeze-pump-thaw cycles before being stored under nitrogen. Unless stated otherwise, all ligand preparation reactions performed under an inert atmosphere were performed using a double-manifold Schlenck vacuum line under nitrogen. All three-step metallation reactions were performed in a nitrogen-filled glovebox (N2 < 0.1 ppm, O2 < 0.1 ppm). The phosphasalen ligand, L 2 , was synthesised according to literature protocols. [1] Rac-LA was recrystallized twice from anhydrous toluene under a nitrogen atmosphere before twice being sublimed and stored under nitrogen before use. NMR Spectroscopy: 1 H, 13 C, 31 P and 2D NMR (COSY, HSQC, HMBC) spectra were recorded using a Bruker AV 400 MHz spectrometer at 298 K. 1 H{ 1 H} NMR and variable temperature NMR studies were also recorded using a Bruker AV 400 MHz spectrometer. ROESY and DOSY NMR experiments were recorded using a Bruker AV 500 MHz spectrometer. The tacticity of polymers was determined from its homonuclear-decoupled 1 H NMR spectrum, using deconvolution techniques from MestReNova v. 8.0.0. The overall isoselectivity, Pi, was determined by taking an average of tetrad integrals predicted from Bernoullian statistics (equations given in section on 'data fitting'). The following abbreviations are used in the report of spectra: br, broad; s, singlet; d, doublet; dd, doublet of doublets; t, triplet; q, quartet; m, multiplet.

ESI-MS:
ESI-MS spectra were collected by Carl Braybrook and mass spectrometric analyses were performed on a Thermo Scientific Q Exactive mass spectrometer fitted with a HESI-II ion source. Positive and/or negative ion electrospray mass spectra were recorded in an appropriate mass range set for 140,000 mass resolution. The probe was used with 0.3 ml/min flow of solvent. The nitrogen nebulizing/desolvation gas used for vaporization was heated to 350°C in these experiments. The sheath gas flow rate was set to 35 and the auxiliary gas flow rate to 25 (both arbitrary units). The spray voltage was 3.0kV and the capillary temperature was 300°C.

MALDI-TOF MS: MALDI ToF analysis was carried out by Carl Braybrook with spectra run on a Bruker
Autoflex III MALDI TOF/TOF mass spectrometer using positive ion in reflectron mode. Laser power was optimized and set to 40. The matrix employed was DCTB at a concentration of 10 mg ml -1 in CHCl3/EtOH. 10 µL of this solution was co-applied with the 1 μL KTFA and 1 ul of the analyte (1 mg mL -1 ; CHCl3/EtOH) to a steel target plate for analysis. Mn, Mw and Đ values were calculated using the Bruker PolyTools Synthetic Polymer Analysis software package version 1.18. X-ray diffraction: X-ray diffraction data was collected using a Bruker APEX II CCD diffractometer for 1 and an Agilent Xcalibur 3 E diffractometer for 2. Further crystallographic data can be found in Figs. S16-S18 and Tables S1-S2.
Elemental analysis: Elemental analysis (C,H,N) for all novel ligands and complexes were were carried out by Mr Stephen Boyer at London Metropolitan University except ligand L 1 , which was analysed by Mr Bob McAllister from the Campbell Microanalytical Laboratory at the University of Otago, with a standard error of ±0.3%.

Gel permeation chromatography:
For polymers obtained by ROP of rac-LA initiated by compound 1, Four Agilent PL-Gel columns (3 x PL-Gel Mixed C (5 μm) and 1 x PL-Gel Mixed E (3 μm) columns) were used in series, with HPLC grade chloroform (amylene-stabilised) as the eluent, at a flow rate of 1 mL min -1 , on a Waters Alliance system equipped with an Alliance 2695 Separation Module at 30 o C. Polymer number-average molecular weight (Mn) and polydispersity index (Mw/Mn; PDI) were calibrated against low dispersity polystyrene standards using a 3 rd order polynomial fit, linear across molar mass ranges. As molar masses were reported as polystyrene equivalents, a Mark-Houwink correction factor of 0.58 was used. [2] For polymers obtained by ROP of rac-LA initiated by compound 2, PLA molecular weight information (Mn and PDI) were determined by gel permeation chromatography, equipped with multi-angle laser light scattering (GPC-MALLS). Two Mixed Bed PSS SDV linear S columns were used in series, with THF as the eluent, at a flow rate of 1 mL min -1 , on a Shimadzu LC-20AD instrument at 40 o C. The light scattering detector was a triple-angle detector (Dawn 8, Wyatt Technology), and the data was analysed using Astra Version 6.1. The refractive angle increment for polylactide (dn/dc) in THF was 0.040 mL g -1 . [3] All polymers were filtered prior to analysis. DSC: Differential scanning calorimetry was performed on a DSC 821 e using 100 μl aluminium crucibles (Mettler Toledo) under nitrogen. For each sample, approximately 5 -6 mg of material was heated at a rate of 5 K min -1 . The glass transition temperature, Tg, and the melting temperature, Tm, were determined using the software pack Star e (Mettler Toledo, version 14.0). The Tg was determined by the peak onset and the Tm, the peak maximum was used.

Synthetic protocols
Compound S1 At 273 K, N-bromosuccinimide (5.4 g, 30.3 mmol) was added slowly to a solution of 2,4-di-cumylphenol (10.0 g, 30.3 mmol) in acetonitrile (150 mL) under continuous stirring. The reaction mixture was allowed to warm to 298 K and left stirring overnight for 16 h. To the afforded orange solution, a saturated aqueous solution of sodium sulphite (~10 mL) was then added, inducing the formation of a white precipitate. The mixture was then filtered, the white solid separated and the filtrate extracted with petroleum ether (2 x 50 mL). The organic layer was then dried (MgSO4) and the solvent removed in vacuo. The product was isolated as a yellow oil (7.5 g, 61 %).

Compound S2
Under a dry atmosphere of N2, a solution of 2-bromo-3,5-di-tert-cumylphenol (4.0 g, 9.8 mmol) in Et2O (~100 mL) was cooled to 195 K before subsequently, a solution of n-butyl lithium (1.5 M in hexanes, 13.0 mL, 19.6 mmol) was added affording a white suspension. The reaction mixture under continuous stirring was allowed to warm to 298 K and left for 3 h. The resultant turbid white suspension was cooled back down to 195 K and chlorodiphenylphosphine (1.75 mL, 9.8

Compound L 1
At 195 K, bromine (149 µL, 2.91 mmol) was added to a solution of S2 (1.50 g, 2.91 mmol) in dichloromethane (50 mL). The solution was then allowed to warm to 298 K and left to stir for 2 h. The solution was then cooled again to 195 K, after which, tributylamine (694 µL, 2.91 mmol) was added, followed by ethylenediamine (97 µL, 1.46 mmol). The solution was then allowed to warm to 298 K and left stirring for 16 h overnight affording a pale yellow turbid suspension The solvent was then removed in vacuo leaving behind a yellow gelatinous product. To the crude mixture, tetrahydrofuran (15 mL) was added followed by petroleum ether (2 mL

Compound 1
Sodium hydride (30.8 mg, 1.28 mmol) was added to a slurry of ligand L 1 (200 mg, 0.16 mmol) in THF (10 mL) under continuous stirring at 298 K. After 16 h, the afforded turbid white suspension was checked by 31 P{ 1 H} NMR spectroscopy, which showed clean formation of the deprotonated species with a singlet at 19.1 ppm. The solution was then filtered to remove insoluble sodium salts before subsequent addition of indium trichloride (35.5 mg, 0.16 mmol) and the reaction stirred for another 16 h at 298 K. The 31 P{ 1 H} NMR spectrum of the afforded cloudy solution was then checked for clean formation of an indium chloride complex confirmed by a singlet at 40.8 ppm. Potassium ethoxide (13.5 mg, 0.16 mmol) was then added to the reaction mixture and stirring continued for a further 12 h at 298 K. The 31 P{ 1 H} NMR spectrum was again checked and indicated clean formation of the indium alkoxide species with a singlet at 39.5 ppm. The cloudy solution was then filtered and the solvent removed in vacuo. Finally, the residue was dissolved in cyclohexane (5 mL) and left to crystallise, affording the product as colourless crystals (63 mg, 31 %).

Compound 2
Sodium hydride (28.8 mg, 1.20 mmol) was added to a slurry of ligand L 2 (200 mg, 0.20 mmol) in THF (10 mL) under continuous stirring at 298 K. After 16 h, the afforded turbid white suspension was checked by 31 P{ 1 H} NMR spectrum, which showed clean formation of the deprotonated species by a singlet at 25.4 ppm. The insoluble sodium salts were then removed from the solution by centrifugation before subsequent addition of indium trichloride (44.3 mg, 0.20 mmol) and the reaction stirred for another 16 h at 298 K. The 31 P{ 1 H} NMR spectrum of the afforded cloudy solution was then checked for clean formation of an indium chloride complex confirmed by a singlet at 41.9 ppm. Potassium tertbutoxide (22.5 mg, 0.20 mmol) was then added to the reaction mixture and stirring continued for a further 12 h at 298 K. The 31 P{ 1 H} NMR spectrum was again checked and indicated clean formation of the indium alkoxide species with a singlet at 40.3 ppm. The cloudy solution was then centrifuged to remove all remaining insoluble salts and the solvent removed in vacuo. Finally, the residue was washed with hexane (5 mL) inducing precipitation of a white solid which was then isolated and dried (90 mg, 44 %).  ,7.19;N,2.74. Found: C,67.87;H,7.32;N,2.82.

Typical polymerisation procedure (ambient):
In a nitrogen-filled glovebox, a silanised vial was charged with rac-LA (232 mg, 1.6 mmol) and dissolved in THF (1.3 mL). A stock solution of the initiator (0.3 mL, 0.011 M) was injected into the solution of monomer such that the overall concentration of lactide was 1 M and the initiator, 2 mM. Aliquots were taken inside the glovebox at specific time intervals and precipitated in hexane (~ 1 mL). The aliquots were then removed from the glovebox and left to evaporate. The crude product was analysed by 1 H NMR (and 1 H{ 1 H} NMR where applicable) spectroscopy and GPC. Polymer conversion yields were determined by relative integration of the methine proton resonances for the polymer and monomer. The tacticity (Pi or Ps) was determined by integration of the tetrad resonances in the 1 H{ 1 H} decoupled NMR of the polymer. In the cases of high Pi values (Pi ≥ 0.87), the polymers were purified and their thermal properties analysed by DSC.
Typical polymerisation procedure (low temperature): In a nitrogen-filled glovebox, a silanised vial was charged with rac-LA (288 mg, 2.0 mmol) and dissolved in THF (1.66 mL). A stock solution of the initiator (1.0 mL, 0.004 M) was injected into the solution of monomer such that the overall concentration of lactide was 0.75 M and the initiator 1.5 mM. The vial was then sealed, removed from the glovebox placed into a (-15 o C) freezer equipped with a stirring functionality. After the allotted reaction time, the vial was removed from the freezer and quenched into hexane (5 mL) and the solvent allowed to evaporate. The sample was then analysed used the methods described above.

Typical polymerisation procedure (high temperature-melt):
In a nitrogen-filled glovebox, a silanised vial was loaded with rac-LA (191 mg, 1.3 mmol) and the initiator (3.5 mg, 0.0026 mmol). The vial was then sealed, removed from the glovebox and placed into an oil bath set at 130 o C. The reaction was stirred at 130 o C for 10 mins before the vial was exposed to air and a few drops of CDCl3 added to destroy the active catalyst. The mixture was then analysed using the methods described above.
Silanisation procedure: Hot vials heated in an oven at 105 o C were rinsed several times in a solution of dichlorodimethylsilane in dichloromethane (0.8 M, 10 mL) before being left to dry in an oven prior to use.

Polymer purification (for DSC):
The crude polymer mixture was dissolved in THF (~2 mL) and injected into a solution of methanol (~5-10 mL) to precipitate the polymer. The mixture was then centrifuged and the filtrate discarded. This was repeated twice more and the 1 H NMR subsequently checked to verify the absence of monomer signals. Finally, the polymer was dissolved in dichloromethane (~3 mL) and rapidly filtered through a small pad of silica. The solvent was removed in vacuo and dried both under vacuum and by oven, set to 40 o C.

Stereocontrol determination in polymers:
The tacticity of a polymer was determined from its homonuclear-decoupled 1 H NMR spectrum, which displayed 5 peaks, each corresponding to a tetrad resonance and reflects the relative stereochemistry along the polymer chain, see Figs. S50-S53 and S55-S58 for 1 H{ 1 H} spectra. Peak deconvolution (with residuals minimised), was carried out using MestReNova v. 8.0.0, and was used to determine tetrad integrals. The Pi values were determined from each of the tetrad integrals using the probabilities predicted from Bernoullian statistics. [4] The equations used are given below: Pi values were then determined for each tetrad and were highly reproducible, thus the Pi reported is an average of multiple runs and are reported with an approximate standard error in the mean of ±3 %.
i. ii.
Reciprocal space analysis of the data set for the structure of 2 clearly showed the crystal to be twinned. Despite numerous attempts, modelling this twinning at the data processing stage gave unsatisfactory results, and so the data was processed without any twin modelling.
The included solvent was found to be highly disordered, and the best approach to handling this diffuse electron density was found to be the SQUEEZE routine of PLATON. [8] This suggested a total of 905 electrons per unit cell, equivalent to 226.25 electrons per asymmetric unit. Before the use of SQUEEZE the solvent clearly resembled thf (C4H8O, 40 electrons), and 5.5 thf molecules corresponds to 220 electrons, so this was used as the solvent present. As a result, the atom list for the asymmetric unit is low by 5.5(C4H8O) = C22H44O5.5 (and that for the unit cell low by C88H176O22) compared to what is actually presumed to be present.
The C23-and C51-based t-butyl groups were found to be disordered, and in each case two orientations were identified, of ca. 79:21 and 85:15% occupancy respectively. The geometries of each pair of orientations were optimised, the thermal parameters of adjacent atoms were restrained to be similar, and only the non-hydrogen atoms of the major occupancy orientations were refined anisotropically (those of the minor occupancy orientations were refined isotropically). Table S1. Selected bond lengths (Å) and angles ( o ) for compound 1.