Heterotrimetallic Carbon Dioxide Copolymerization and Switchable Catalysts: Sodium is the Key to High Activity and Unusual Selectivity

Abstract A challenge in polymer synthesis using CO2 is to precisely control CO2 placement in the backbone and chain end groups. Here, a new catalyst class delivers unusual selectivity and is self‐switched between different polymerization cycles to construct specific sequences and desirable chain‐end chemistries. The best catalyst is a trinuclear dizinc(II)sodium(I) complex and it functions without additives or co‐catalysts. It shows excellent rates across different ring‐opening (co)polymerization catalytic cycles and allows precise control of CO2 incorporation within polyesters and polyethers, thereby allowing access to new polymer chemistries without requiring esoteric monomers, multi‐reactor processes or complex post‐polymerization procedures. The structures, kinetics and mechanisms of the catalysts are investigated, providing evidence for intermediate speciation and uncovering the factors governing structure and composition and thereby guiding future catalyst design.

Section S1: Methods Materials: Solvents and reagents were obtained from commercial sources and used as received unless stated otherwise. If "dried solvents" were used these were obtained from an SPS system, degassed by several freeze-pump-thaw cycles and further dried with 3 Å molecular sieves and stored under nitrogen. Cyclohexene oxide was dried over calcium hydride overnight and purified via fractional distillation prior to use and stored in an inert atmosphere. Phthalic anhydride was extracted with dry benzene, recrystallised from dry chloroform and sublimed at 100 °C and 0.05 bar. Commercial 1,2cyclohexenediol was recrystallized from dry chloroform and dried under dynamic vacuum before use. 99.8% carbon dioxide supplied by BOC Ltd was dried by passing it through two drying columns (VICI Metronics carbon dioxide purifier) prior to use. Compound A and H2L were obtained according to a procedure published by Akine et al. 1 The spectroscopic data was in agreement with the literature.
NMR spectra were recorded on a Bruker Advance 400 QNP or Bruker Avance 500 MHz cryo spectrometer. All spectra were recorded with external standards. Unambiguous assignments of NMR resonances were made on the basis of 2D NMR experiments.
UV-visible spectra were collected on a Varian Cary 50 UV spectrometer.
Gel permeation chromatography analysis was carried out on a a Shimadzu LC-20AD instrument equipped with two mixed bed PSS SDV linear S columns in series, THF as the eluent at a flow rate of 1mL/min and at 40 °C. Polymer molecular mass (Mn) was determined by comparison against narrow molecular mass polystyrene standards which were used to calibrate the instrument. Each polymer sample was dissolved in HPLC-grade THF (10 mg/mL) and filtered through a 0.20 μm porous filter frit prior to analysis.
High-resolution ESI mass spectra were obtained using a Thermofisher LTQ Orbitrap XL, by the EPSRC UK National Mass Spectrometry Facility at Swansea University.
Elemental analysis was obtained using a Perkin Elmer 240 Elemental Analyser, by Dr Nigel Howard from the elemental analysis lab at the University of Cambridge.

General Polymerization protocols
Example of low pressure epoxide copolymerization: A mixture of the catalyst (5.2 mg, 5.0 µmol, 1 equiv.), 1,2-cyclohexenediol, (if used, 11.7 mg. 100.0 µmol, 20 equiv.), phthalic anhydride (148.7 mg, 1.0 mmol, 200 equiv.) and cyclohexene oxide (2 mL, 20.0 mmol, 4000 equiv.) was added to a Schlenk tube, under nitrogen. If carbon dioxide was applied, the reaction was subjected to three vacuum/CO2 cycles (pressure regulated at 1 bar CO2 or 1 bar of a 1:1 mixture of CO2 and N2). If carbon dioxide was not used, the Schlenk tube was connected to vacuum/nitrogen line. In some cases, the Schlenk tube was fitted with a DiComp probe for in situ-ATR-IR spectroscopy (REACTIR) and submerged into an oil bath that was pre-heated to the appropriate temperature. At this point the REACTIR instrument was set to begin data-collection (t0). Where a gas atmosphere switch was necessary, it was performed slowly using three vacuum/gas cycles at the given temperature. After the reaction was completed, the mixture was allowed to cool to room temperature and the crude product composition was analysed by NMR spectroscopy of an aliquot to determine the ratio of products (poly(cyclohexene carbonate) PCHC, poly(cyclohexene oxide) PCHO, cyclohexene carbonate c5c). The polymer was isolated by adding the concentrated polymerisation mixture (ca 0.5 mL, achieved by removing excess CHO under a stream of N2) to 100 mL of acidified MeOH (10 µL concentrated HCl, per 100 mL MeOH) resulting in the precipitation of white product.
High pressure epoxide copolymerization: A suspension of catalyst (52.0 mg, 50.0 µmol, 1 equiv.), cyclohexene diol (117.0 mg, 1.0 mmol, 20 equiv.) in CHO (20.0 mL, 0.2 mol, 4000 equiv.) was injected into a 100 mL Parr reactor, under a stream of dry CO2. The reactor was also fitted with a DiComp sentinel probe, attached to an ATR-IR spectrometer, which allowed for continual monitoring of PCHC formation. The reactor was then pressurized with CO2 to the target reaction pressure and allowed to reach the required temperature. Upon reaction completion, the reactor vessel was cooled to room temperature and depressurized; the product composition was analysed by NMR spectroscopy of an aliquot. The polymer was isolated by adding the concentrated polymerisation mixture (ca 5.0 mL, achieved by removing excess CHO under dynamic vacuum) to 500 mL of acidified MeOH (50µL concentrated HCl, per 500 mL MeOH) resulting in precipitation of white product.

Section S2: Synthesis and characterisation of Zn2Na
Scheme S1: Synthesis of Zn2Na from H2L and NMR assignment numbering for Zn2Na.
Alternatively, the addition of just Zn(OAc)2•(H2O)2 (33.3 mg, 151 µmol) to H2L (50.0 mg, 75 µmol) resulted in the precipitation of an orange solid, which was washed with Et2O (100 mL) and dried in vacuo yielding Zn2 (39 mg, 98 µmol, 65%). Zn2 is completely insoluble in all common organic solvents which prevents its characterisation. Yet the addition of CF3C6F4CO2Na (11.2 mg, 98 µmol) in DMSO (5 mL) to a suspension of Zn2, quantitatively yielded Zn2Na as established of 1 H NMR of an aliquot in d6-DMSO.  Reaction Stirring: At epoxide conversions >30%, magnetic stirring becomes inefficient due to viscosity limitations, i.e. the polymerization stirring stops or significantly decreases in terms of speed. To minimize viscosity limitations, rare-earth magnetic stirrer bars (1 cm in length, 0.5 cm in maximal diameter) were applied in a narrow diameter Schlenk tube (12 cm in length, 1.75 cm in max diameter, 1.5-3 mL overall fill volume) at 1400 rpm stirring speed. It is important to note that any deviations from these conditions could change the conversion vs. time data and/or proportion of ether linkages.

Copolymer Determination and characterization procedures:
The following characterisation techniques were used to establish that the different monomer sequences are present the same polymer chain, as opposed to the formation of two separate chains.
(a) Monomodal, narrow dispersity polymer molecular weight distributions were obtained for all copolymers. Also, aliquots taken during polymerizations showed increasing molar mass values with time, but not significant changes to dispersity or modality. The 1 H NMR spectra of these aliquots indicated compositions, at each time point, which correspond to the same composition indicated by in situ IR spectroscopy. (d) Polymer compositions, as determined by NMR spectroscopy, remained identical in crude polymerization samples and in products which were purified by fractionation techniques (i.e. precipitation from DCM solutions using MeOH).

Figure S 11:
Plot showing conversion of CHO to polymer (PCHC and PCHO) vs. time. Data are obtained using in situ ATR-IR spectroscopic monitoring of polymerizations conducted using 1 equiv. Zn2Na, 20 equiv. cyclohexane diol (CHD), 4000 equiv. cyclohexene oxide (CHO) and 1 bar carbon dioxide pressure at 120 °C (Table S1, run #5). The conversion data was obtained by monitoring the changes to absorptions at 1744 cm -1 (PCHC) and 1089 cm -1 (PCHO). The initial rate and turn-overfrequency (TOF) were obtained from 0-15% polymer conversion. Data are obtained using in situ ATR-IR spectroscopic monitoring of polymerizations conducted using 1 equiv. Zn2Na, 20 equiv. cyclohexane diol (CHD), 4000 equiv. cyclohexene oxide (CHO) and 20 bar carbon dioxide pressure at 120 °C (Table S1, run #6). The conversion data was obtained by monitoring the changes to absorptions at 1744 cm -1 (PCHC). The initial rate and turn-over-frequency (TOF) were obtained from 5-20% polymer conversion. Note that the non-linearity from 0-5% conversion arises due to the high-pressure reactor not reaching the target polymerization temperature until 5% conversion    120 °C, upper). Note the increase in the resonance assigned to PCHO (~ 3.35 ppm) arises from the increase in temperature.

Figure S 18:
GPC traces of aliquots removed during the polymerization described above (Fig. S15, 16). The red trace is from the aliquot taken from the reaction at 80 °C and the purple trace taken after the reaction temperature was increased to 120 °C.  Table S1 . The plot shows that the carbonate and ether resonances diffuse at the same rate and are, therefore, attached to one another. Data are obtained using in situ ATR-IR spectroscopic monitoring of polymerizations conducted using 1 equiv. Zn2Na, 20 equiv. cyclohexane diol (CHD), 4000 equiv. cyclohexene oxide (CHO), 1 bar nitrogen pressure, at 120 °C. The conversion data was obtained by monitoring the changes to absorptions at 1089 cm -1 (PCHO). The initial rate and turn-over-frequency (TOF) were obtained from 3-14% polymer conversion Note that during CHO ROP, using Zn2Na, complete CHO conversion was not achieved. The maximum CHO conversion was also observed to decrease with increasing temperature. This phenomenon does not arise due to catalyst decomposition as the reaction can be switched to CO2/CHO ROCOP (after CHO ROP conversion has apparently stopped). These data appear to indicate that CHO ROP may be equilibrium limited but such an explanation is not consistent with the fact that epoxide ROP is heavily exergonic. We infer that Zn2Na catalyses the depolymerisation of PCHO, through a different mechanism to CHO ROP, and hence seemingly pseudo-equilibrates the ROP of CHO. This effect will be the subject to further studies and lies outside the scope of this contribution.

Section S5: Synthesis and characterisation of Zn2NaBArF
In order to confirm that CHO ROP catalysed by Zn2Na features an anionic rather than a cationic propagating chain end,a derivative without anionic initiator was prepared: Zn2NaBArF and its performane in CHO ROP compared. As presented in this section, polymerization results in significantly broader disperity values and (near instantaneous) much faster CHO ROP for Zn2NaBArF than for Zn2Na. These findings are proposed to be a consequence of cationic rather than anionic propagation mechanisms. The findings suggest that CHO ROP using Zn2Na occurs by anionic (coordination insertion) propagation mechanisms.
Scheme S2: Synthesis of Zn2NaBArF from H2L and NMR assignment numbering for Zn2NaBArF.

Synthesis of Zn2NaBArF:
A solution of Zn(OAc)2(H2O)2 (33.3 mg, 151 µmol) and tetrakis(3,5bis(trifluoromethyl)phenyl)borate NaBArF (67.1 mg, 75 µmol) in MeOH (5 mL) was added to a solution of H2L (50.0 mg, 75 µmol) in DCM (5 mL). The resulting solution was left unperturbed for 5 min. Afterwards all volatiles were removed in vacuum yielding a semi-solid. In order to remove residual AcOH by-product, the crude material was suspended in toluene (20 mL), which was afterwards removed undedr vacuum. This process was repeated yielding Zn2NaBArF·4H2O as an orange powder (123 mg, 73 µmol, 98%).         Table S1 #1. Note that heating increasing the temperature to 120°C after PCHO block formation (in order to investigate potential PCHC depolymerisation) left polymer composition unchanged.  Table S1 #2.  Table S1 #2.  (Figure S 34). The black trace corresponds to the polymer produced under a carbon dioxide atmosphere (PCHC), the red trace to the copolymer produced when the gas was changed to nitrogen and the blue trace for the copolymer produced after the gas was changed back to carbon dioxide.        To solve this relation, requires knowledge of [CO2], and therefore the solubility of CO2 in neat CHO over the temperature range 80 -120 °C. The solubility data was approximated using an exponential fit (in line with Henry's law) to the published solubility of CO2 in diethyl carbonate (which shows similar polymerization performance to reactions conducted in neat CHO).

Section S8.3: NMR spectroscopy
Scheme S 3: Synthesis of Zn2KO2 13 CO t Bu and NMR assignment numbering.
ZnEt2 (2.0 mg, 16.2 µmol, 2.1 equiv.) was added to a suspension of H2L (5.0 mg, 7.6 µmol, 1.0 equiv.) d8-THF (1 mL) and the resulting mixture was stirred for 10 min, at room temperature. Afterwards KO t Bu (1.5 mg, 13.4 µmol, 1.7 equiv.) was added and the resulting mixture was stirred for 10 min forming a deep brown solution. The reaction was exposed to 1 atm 13 CO2 resulting in the immediate formation of a bright yellow precipitate. The precipitate was dissolved by adding a portion of d6-DMSO (0.3 mL) followed by repeating the pressurisation with 13 CO2. The obtained orange-yellow solution was analysed by NMR spectroscopy. Zn2Na (2.5mg) was heated at 100 °C in CHO (2 mL, 1 equiv. Zn2Na, 4000 equiv. CHO). All volatiles were removed under vacuum and the remaining solid was dissolved in dry d8-THF and analysed by NMR spectroscopy, note that only a broad and undefined spectrum was observed.
Zn2Na (2.5 mg, 1 equiv.) was heated at 100 °C for 30 min, in CHO (2 mL, 4000 equiv.) and with CHD (6 mg, 20 equiv.) inside a J-Youngs NMR tube containing a d8-toluene lock capillary and using (MeOPh)3P as an internal standard. Afterwards the sample was cautiously evacuated, before completely cooling to aid degassing and 13 CO2 was added at 1.5 atm pressure. Note that the 13 CO2 insertion does also occur readily at room temperature, but requires a greater number of de-gassing cycles; the conditions presented here are preferable due to the high cost of 13 CO2. Zn2Na (2.5 mg, 1 equiv.) was heated at 100 °C in CHO (2 mL, 4000 equiv. CHO). The solution was cooled to room temperature and transferred to a N2 filled glovebox where phthalic anhydride (0.37 mg, 1 equiv., from a stock solution in CHO) was added. The resulting solution was stirred for 1 h, at room temperature. All volatiles were removed under vacuum and the remaining solid was dissolved in dry d8-THF and analysed by NMR spectroscopy.  33,164.27,163.26,148.40,140.90,130.67,130.21,128.10,124.88,124.81 (d,J = 272.3 Hz),121.20,120.60,117.71,112.10.           The spectrum on the left shows the product after CHO/CO2 ROCOP; the spectrum on the right shows the product after the gas was switched to N2 and indicates depolymerization occurred (Figure S 78).   Section S13: Synthesis of Zn2La and its Polymerisation Kinetics Synthesis of (CF3C6F4CO2)3La 4-(Trifluoromethyl)benzoic acid (500.0 mg, 2.6 mmol, 1 equiv.) was suspended in deionized water (100 mL). Next, NaOH (105.0 mg, 2.6 mmol, 1 equiv.) was added and the mixture was stirred for 1 h, at room temperature. The mixture was filtered and LaCl3 (150.0 mg, 0.6 mmol, 0.2 equiv.), dissolved in deionized water (10 mL), was added resulting in the immediate formation of a thick white precipitate. The precipitate was isolated by filtration, washed with deionized water (200 mL) and Et2O (200 mL     Section S14: Synthesis of Zn2Y
Conclusive NMR analysis was prevented by the broadness of the spectrum.   PPNCl, 20 equiv. CHD, 4000 equiv. CHO at 100 °C. After 1280 min the gas was switched to N2 and the temperature was increased to 120 °C resulting in depolymerization of PCHC.