The design of efficient metal catalysts for the selective coupling of epoxides and carbon dioxide to afford completely alternating copolymers has made significant gains over the past decade. Hence, it is becoming increasingly clear that this "greener" route to polycarbonates has the potential to supplement or supplant current processes for the production of these important thermoplastics, which involve the condensation polymerization of diols and phosgene or organic carbonates. On the basis of the experiences in our laboratory, this Account summarizes our efforts at optimizing (salen)CrIIIX catalysts for the selective formation of polycarbonates from alicyclic and aliphatic epoxides with CO2. An iterative catalyst design process is employed in which the salen ligand, initiator, cocatalyst, and reaction conditions are systematically varied, with the reaction rates and product selectivity being monitored by in situ infrared spectroscopy.

A detailed mechanistic study into the copolymerization of CO2 and cyclohexene oxide utilizing CrIII(salen)X complexes and N-methylimidazole, where H2salen = N,N'-bis(3,5-di-tert-butylsalicylidene)-1,2-ethylenediimine and other salen derivatives and X = Cl or N3, has been conducted. By studying salen ligands with various groups on the diimine backbone, we have observed that bulky groups oriented perpendicular to the salen plane reduce the activity of the catalyst significantly, while such groups oriented parallel to the salen plane do not retard copolymer formation. This is not surprising in that the mechanism for asymmetric ring opening of epoxides was found to occur in a bimetallic fashion, whereas these perpendicularly oriented groups along with the tert-butyl groups on the phenolate rings produce considerable steric requirements for the two metal centers to communicate and thus initiate the copolymerization process. It was also observed that altering the substituents on the phenolate rings of the salen ligand had a 2-fold effect, controlling both catalyst solubility as well as electron density around the metal center, producing significant effects on the rate of copolymer formation. This and other data discussed herein have led us to propose a more detailed mechanistic delineation, wherein the rate of copolymerization is dictated by two separate equilibria. The first equilibrium involves the initial second-order epoxide ring opening and is inhibited by excess amounts of cocatalyst. The second equilibrium involves the propagation step and is enhanced by excess cocatalyst. This gives the [cocatalyst] both a positive and negative effect on the overall rate of copolymerization.

A series of complexes of the form (salen)AlZ, where H2salen = N,N'-bis(salicylidene)-1,2-phenylenediimine and various other salen derivatives and Z = Et or Cl, have been synthesized. Several of these complexes have been characterized by X-ray crystallography. An investigation of the utilization of these aluminum derivatives along with both ionic and neutral bases as cocatalysts for the copolymerization of carbon dioxide and cyclohexene oxide has been conducted. By studying the reactivity of these complexes for this process as substituents on the diimine backbone and phenolate rings are altered, we have observed that aluminum prefers electron-withdrawing groups on the salen ligands, thereby producing an electrophilic metal center to be most active toward production of polycarbonates from CO2 and cyclohexene oxide. For example, the complex derived from H2salen = N,N'-bis(3,5-di-tert-butylsalicylidene)-1,2-ethylenediimine is essentially inactive when compared to the analogous derivative containing nitro substituents in the 3-positions of the phenolate groups. This is to be contrasted with the catalytic activity observed for the (salen)CrX systems, where electron-donating salen ligands greatly enhanced the reactivity of these complexes for the coupling of CO2 and epoxides. While (salen)AlZ complexes are capable of producing poly(cyclohexene oxide) carbonate with low amounts of polyether linkage along with small quantities of cyclic carbonate byproducts, their reactivities, covering a turnover frequency range of 5.2-35.4 mol of epoxide consumed/(mol of Al x h), are greatly reduced when compared to their (salen)CrX analogues under identical reaction conditions.

The rate of the copolymerization reaction of cyclohexene oxide and carbon dioxide in the presence of (salen)CrIIIN3 and various cocatalysts has been determined as a function of CO2 pressure. Carbon dioxide insertion into the (salen)Cr-alkoxide intermediates, afforded following epoxide ring-opening, was shown to be rate-limiting at pressures below 35 bar. Higher pressures of carbon dioxide resulted in catalyst/substrate dilution with a concomitant decrease in the rate of copolymer formation. On the other hand, cyclic carbonate formation was inhibited as the CO2 pressure was increased. The most active (salen)CrN3 catalyst (H2salen = N,N‘-bis(3-tert-butyl-5-methoxysalicylidene)-(1R,2R)-cyclohexenediimine), along with a [PPN][N3] cocatalyst, exhibited a TOF of 1153 mol epoxide consumed/mol chromium·h at 80 °C and a CO2 pressure of 34.5 bar.

Aluminum and tin salen complexes have been shown to effectively catalyze the ring-opening polymerization (ROP) of trimethylene carbonate (TMC) to polycarbonate. The most active salen derivative in each instance contained a phenylene backbone with chloro substituents in the 3,5-positions of the phenolate rings, with the aluminum derivatives being significantly more active than their tin(IV) counterparts. Importantly, the resultant polycarbonate was shown by 1 H NMR to be void of ether linkages. The reaction was demonstrated to proceed via a mechanism first order in both [catalyst] and [monomer] and to involve TMC ring-opening by way of acyl oxygen bond cleavage. Consistent with a reaction pathway involving an insertion of the monomer into the metal-nucleophile bond (e.g., Al-Cl or Sn-Cl), the activation parameters were determined to be ∆H q ) 51 kJ/mol and ∆S q ) -141 J/(mol deg).