Multinuclear Zinc–Magnesium Hydroxide Carboxylates: A Predesigned Model System for Copolymerization of CO2 with Epoxides

Among numerous catalysts in the ring-opening copolymerization of epoxides with carbon dioxide (CO2), zinc dicarboxylate complexes are the most common type, and in the family of metal-based homogeneous catalysts, zinc and magnesium complexes have attracted widespread attention. We report on the synthesis and structural characterization of a zinc–magnesium benzoate framework templated by the central hydroxide anion with μ3-κ2:κ2:κ2 coordination mode, [ZnMg2(μ3-OH)(O2CPh)5]n (n = 1 or 2). The resulting heterometallic system forms stable Lewis acid–base adducts with tetrahydrofuran (THF) and cyclohexene oxide (CHO), which crystallize as the hexanuclear zinc–magnesium hydroxide carboxylate cluster [ZnMg2(μ3-OH)(O2CPh)5(L)2]2 (L = THF or CHO). Their X-ray crystal structure analysis revealed that the Zn center prefers 4-fold coordination and the Mg centers demonstrated the ability to accommodate higher coordination numbers, and as a result, the heterocyclic molecules are exclusively bonded to 6-fold Mg atoms. The heteronuclear carboxylate aggregates appeared active in the copolymerization reaction at elevated temperatures to produce an alternating poly(cyclohexene carbonate).

General Considerations.Unless otherwise stated, all manipulations involving air and moisturesensitive organometallic compounds were conducted under a dry, oxygen-free argon atmosphere either using standard Schlenk techniques or in a glovebox (MBraun UniLab Plus; < 0.1 ppm O2, < 0.1 ppm H2O).All glassware was stored in a 150°C oven overnight before use.All reagents were purchased from commercial vendors: benzoic acid (ABCR), diethyl zinc (ABCR), n-butyl magnesium (1M solution in heptane) (sigma) and used as received.Solvents were purified by passage through activated aluminium oxide (MBraun SPS) and stored over 3Å molecular sieves.The deuterated solvents were dried over Na/K, distilled under an argon atmosphere before use, and stored over molecular sieves.Cyclohexene oxide (CHO) was obtained from ABCR and prior to its use it was distilled from CaH2, thoroughly degassed and stored under argon.Research grade carbon dioxide was used for copolymerization reactions.

Synthetic Procedures
Initial reaction to obtain heteronuclear Zn/Mg oxide/hydroxide carboxylates.Et2Zn (2.0 M in hexane, 1 mL, 2 mmol) and di-n-butylmagnesium (1.0 M in heptane, 2 mL, 2 mmol) was added dropwise to a solution of benzoic acid (0.73 g, 6 mmol) in THF (30 mL) at -78 °C.The mixture was warmed to room temperature, and stirred for 8 h.Afterwards, to the vigorously stirred reaction degassed water (18 µL, 1 mmol) was added and the reaction mixture was stirred for another 24 h.A mixture of the product was isolated as colourless crystals after filtration and crystallization in THFhexane at 0-5 °C.However, it was not possible to separate the different compounds; but we were fortunate enough to get the crystal structure of two different compounds from the batch of crystals.
General procedure for the ROCOP.Cyclohexene oxide (2 mL, 20 mmol) and the catalyst (0.1 mol%) were added to a Schlenk tube in the glovebox.This Schlenk tube was then subjected to five rapid vacuum/CO2 (pressure regulated to 1 bar) cycles, before it was left stirring under 1 atm CO2, at 80 °C in an oil bath.Aliquots were taken under a positive pressure of CO2.Reaction was quenched by cooling the sample and exposing it to air. 1 H NMR spectra were taken, in air and in CDCl3, before the crude product was obtained through removal of volatile CHO under vacuum.GPC analysis was carried out on the crude sample which was re-dissolved in dichloromethane for analysis.

Diffusivity measurement
The DOSY spectra were acquired on Bruker AVANCE II 300 MHz spectrometer at 298 K. Pulsed field gradient double stimulated echo convection-compensated (PFGSTE) sequence with total of 16 diffusion encoding bipolar gradients (ranging from 3 to 48 G/cm, smoothed-square shaped, equal steps in gradient squared) was used and the total width of the gradient pulse was optimized to achieve attenuation of about 90% of the initial intensity of the signals.Overall, the key acquisition parameters were as follows: total length of gradient encoding pulses gradient -2ms, diffusion delay -150ms, gradient recovery delay -0.1ms, relaxation delay -2.8s.Steady-state scans in number of 4 were performed prior to acquisition of the data.Raw data was processed with powerful DOSY Toolbox which is extensively described in its author's paper.S1 Samples were dissolved in dry and degassed THF-d8 at concentration ca.15mM.The molecular masses of analyzed compounds were estimated utilizing an external calibration curve (ECC) approach with normalized diffusion coefficients, with 9-methylanthracene (9-MeA; MW = 192) as an internal reference.S2, S3 Moreover, we have applied the van-der-Waals radii-based correction in order to account for the underestimation of MW due to the presence of heavy atoms, according to the literature methods.

S-7 X-ray Diffraction Studies
The single crystal X-ray data for complexes 1-THF, 1-CHO and 2 were collected at 100(2)K on a SuperNova Agilent diffractometer using graphite monochromated MoK radiation ( = 0.71073 Å) for 1-THF, 2 and CuK radiation ( = 1.54184Å) for 1-CHO.The crystals of all complexes were selected under Paratone-N oil, mounted on the nylon loops, and positioned in the cold stream on the diffractometer.The data were processed with CrysAlisPro.S5 The structures 1-THF, 1-CHO and 2 were solved by direct methods using the SHELXT program and were refined by full matrix least-squares on F 2 using the program SHELXL.S6 S14).e TON = molCHO converted x (molcat) -1 .f TOF = TON per hour.
All non-hydrogen atoms were refined with anisotropic displacement parameters.Hydrogen atoms were added to the structure model at geometrically idealized coordinates and refined as riding atoms.Crystallographic data (excluding structure factors) for the structure reported in this paper have been deposited with the Cambridge Crystallographic Data Centre as a supplementary publication.Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax: (+44)1223-336-033; e-mail: deposit@ccdc.cam.ac.uk).CCDC: 2239055 (1-THF), 2239056 (1-CHO), 2240656 (2).Powder XRD data were collected on a PANalytical Empyrean diffractometer.Measurements employed Ni-filtered Cu Ka radiation of a copper sealed tube charged with 40 kV voltage and 40 mA current in a Bragg-Brentano geometry with a beam divergence of 1 deg. in the scattering plane.The sample was spread over the surface of a porous glass plate fixed to the sample holder.Diffraction patterns were measured in the scattering angle range of 3-50 degrees by step scanning in steps of 0.02 degree.

Figure S6 .
Figure S6.The molecular structure of 1-THF with thermal ellipsoids is set at 30% probability.Hydrogen atoms and lattice THF molecules have been omitted for clarity.

Figure S10 .
Figure S10.The distance between the metal centers in 1-THF.

Figure S11 .
Figure S11.Powder X-ray diffraction pattern for 1-THF compared with the simulated PXRD pattern obtained from single crystal X-ray diffraction analysis. .

Figure S12 .
Figure S12.Powder X-ray diffraction pattern for 1-CHO compared with the simulated PXRD pattern obtained from single crystal X-ray diffraction analysis.

Figure S13 .
Figure S13.Powder X-ray diffraction pattern for 2 compared with the simulated PXRD pattern obtained from single crystal X-ray diffraction analysis.

Figure S19 .
Figure S19.A linear plot of lnkobs versus ln[cat]0 showing a first-order dependence on the catalyst concentration.

Figure S20 .
Figure S20.Conversion of CHO vs CO2 pressure in the presence of 0.1 mol % catalyst at 80 °C for 24 h reaction.

Table S5 .
ROCOP of CHO and CO2 initiated by 1-THF a .Reactions were run using different catalyst concentrations (0.05-0.15 mol%) in 20 mmol of CHO (8 M in diethylcarbonate), 80 C, 1 bar pressure of CO2.b Expressed as percentage CHO conversion, determined from the 1 H NMR spectroscopy (FigureS14).cExpressedas a percentage of CO2 uptake versus the theoretical maximum (100%), determined from the 1 H NMR spectroscopy by comparing normalized integrals for carbonate (4.64 ppm) and ether (3.45 ppm) resonances in the polymer backbone (FigureS14).d Expressed as a percentage of polymer formation versus the theoretical maximum (100%), determined from the 1 H NMR spectroscopy by comparing normalized integrals for polymer (4.65 ppm), cis-cyclic carbonate (4.68 ppm) and trans-cyclic carbonate (δ 4.01 ppm) (Figure a