Mechanistic basis for high stereoselectivity and broad substrate scope in the (salen)Co(III)-catalyzed hydrolytic kinetic resolution

Abstract

In the (salen)Co(III)-catalyzed hydrolytic kinetic resolution (HKR) of terminal epoxides, the rate- and stereoselectivity-determining epoxide ring-opening step occurs by a cooperative bimetallic mechanism with one Co(III) complex acting as a Lewis acid and another serving to deliver the hydroxide nucleophile. In this paper, we analyze the basis for the extraordinarily high stereoselectivity and broad substrate scope observed in the HKR. We demonstrate that the stereochemistry of each of the two (salen)Co(III) complexes in the rate-determining transition structure is important for productive catalysis: a measurable rate of hydrolysis occurs only if the absolute stereochemistry of each of these (salen)Co(III) complexes is the same. Experimental and computational studies provide strong evidence that stereochemical communication in the HKR is mediated by the stepped conformation of the salen ligand, and not the shape of the chiral diamine backbone of the ligand. A detailed computational analysis reveals that the epoxide binds the Lewis acidic Co(III) complex in a well-defined geometry imposed by stereoelectronic rather than steric effects. This insight serves as the basis of a complete stereochemical and transition structure model that sheds light on the reasons for the broad substrate generality of the HKR.

Figure 1

Limiting models for stereoinduction in the bimetallic epoxide ring-opening step include: A) The stereochemistry of the Lewis acidic complex determines stereoselectivity, with the stereochemistry of the nucleophiledelivery agent (salen)Co–OH being inconsequential; B) The stereochemistry of the nucleophilic (salen)Co–OH complex controls stereoselectivity, with that of the Lewis acidic complex being unimportant; C) High stereoselectivity is contingent on a matched relationship between the stereochemistry of both catalysts.

Figure 2

The eight possible stereochemically distinct pathways in a (salen)Co(III)-catalyzed hydrolysis of a terminal epoxide. In each case, the reaction component that is “mismatched” with respect to the other two components is shown in red.

Figure 3

Dependence of the loading of matched catalyst (S,S)-1b and mismatched catalyst (R,R)-1b on the rate of hydrolysis of (R)-1,2-epoxyhexane ([epoxide]_i = 6.6 M) in 1,2-hexanediol at 25 °C. The reaction rate is plotted as a function of conversion, with water ([H₂O]_i = 2.8 M) as the limiting reagent. To generate the (salen)Co–OH complex 1b quantitatively, the (salen)Co–Cl complex (S,S)-1c and/or (R,R)-1c (0.1–0.5 mol%) was added to the mixture of epoxide and diol and was aged for 60 min prior to addition of water (ref. 6).

Figure 4

Chiral, stepped conformation of a cationic (salen)Co(III) bis(aziridine) complex. Figure generated from data retrieved from the Cambridge Structural Database, submission number CCDC 185815 and ref. . The counterions (a 1:1 mixture of chloride and acetate) and solvent (methylene chloride) are omitted for clarity.

Figure 5

The salen step is an element of chirality in metal salen complexes. The structure shown is derived from single crystal X-ray diffraction data from ref. . The structure on the right is the unit cell, containing two complexes of each enantiomeric conformer.

Figure 6

Optimized structures of neutral (salen)Co(III) complexes calculated as closed-shell singlets at the B3LYP/6-31G(d) level.

Figure 7

Summary of possible pathways for 1b to engage in cooperative catalysis with a Co(III) complex of an achiral salen ligand. If the salen step mediates stereochemical communication, each (salen)Co(III) complex would only be able to undergo cooperative catalysis with another identical complex or with a different (salen)Co(III) complex of the same absolute step stereochemistry. L = H₂O or epoxide.

Figure 8

Comparison of rate of epoxide hydrolysis catalyzed by 3b (0.7 mol %) and (P,S,S)-1b (0.35 mol %). The rates of hydrolysis of (R)-1,2-epoxyhexane ([epoxide]_i = 6.0 M) in 1,2-hexanediol at 25 °C as a function of conversion of water ([H₂O]_i = 3.4 M). In each experiment, 3c or (R,R)-1c was added to the mixture of epoxide and diol and aged for 60 min, followed by water to generate 3b or 1b, respectively, in situ.

Figure 9

Rate dependence on amount of 3b and (R,R)-1b catalyst. For each catalyst loading and/or mixture, we plot the rate of hydrolysis of (R)-1,2-epoxyhexane ([epoxide]_i = 6.0 M) in 1,2-hexanediol at 25 °C versus conversion of water ([H₂O]_i = 3.4 M). In each experiment, 3c and/or (R,R)-1c (0.35–0.70 mol%) was added to the mixture of epoxide and diol and aged for 60 min, followed by water to generate 3b or 1b, respectively, in situ.

Figure 10

Rate dependence on amount of 3b and (P,S,S)-1b. Plot of the rates of hydrolysis of (R)-1,2-epoxyhexane ([epoxide]_i = 6.0 M) in 1,2-hexanediol at 25 °C versus conversion of water ([H₂O]_i = 3.4 M) in 1,2-hexanediol. In each experiment, 3c and/or (S,S)-1c (0.35 mol%) was added to the mixture of epoxide and diol and aged for 60 min, followed by water. The dotted black curve represents the rate of hydrolysis expected from the mixture of 3b and (P,S,S)-1b if no cooperative catalysis between these two catalysts occurred.

Figure 11

Plot of relative energy versus O–Co–O–C dihedral angle θ in a neutral (S,S)-(salen)Co(III) complex with bound (R)-1,2-epoxypropane (blue squares), (S)-1,2-epoxypropane (red diamonds) and ethylene oxide (green circles), calculated as closed-shell singlets at the B3LYP/6-31G(d) level. Each point represents the relative uncorrected electronic energy of an optimization performed with θ frozen and all other degrees of freedom permitted to relax. The minimum for each epoxide was set to ΔE = 0 kcal/mol.

Figure 12

Structures of neutral (S,S)-(salen)Co–OH complexes with bound (R)-1,2-epoxypropane and (S)-1,2-epoxypropane, calculated as closed-shell singlets at the B3LYP/6-31G(d) level.

Figure 13

Structures of potential nucleophilic catalysts ³1b and ¹1b•H₂O optimized at the B3LYP/6-31G(d) level of theory.

Figure 14

Relative reactivity of the singlet and triplet nucleophile in the epoxide ring-opening transition structure for (salen)Co–OH at the B3LYP/6-31G(d) level in the gas phase. Structures in the singlet spin state were calculated as closed-shell configurations. Energies reported as the difference in uncorrected electronic energy.

Figure 15

Epoxide ring-opening transition structures optimized at the B3LYP/6-31G(d) level of theory are presented along with the difference in energy between each structure and TS-1•H₂O. The selectivity was also calculated from single-point energies at the M06-L/6-31+G(d,p) level of theory with the B3LYP/6-31G(d) geometry. C–H bonds are omitted for clarity.

Figure 16

While the epoxide ring is held in the same orientation with respect to the catalyst for both enantiomers of epoxide, an incoming nucleophile attacking the less substituted epoxide carbon is expected to experience different steric interactions with the stepped conformation of the catalyst depending on the stereochemistry of the epoxide (red arrows).

Figure 17

The effect of the epoxide substituent on calculated stereoselectivity. Epoxide ring-opening transition structures optimized at the B3LYP/6-31G(d) level of theory are presented along with the difference in energy between the “Matched” and “Mismatched” transition structures. C–H bonds are omitted for clarity.

Figure 18

Computed selectivity for the hypothetical epoxide ring-opening reaction of ethylene oxide. The epoxide methyl substituent in TS-1•H₂O and TS-2•H₂O was replaced with a hydrogen atom and the resulting structure was optimized to a transition structure at the B3LYP/6-31G(d) level. The resulting structures were overlaid with their parent structures by minimizing the RMSD between the six atoms in the Lewis acidic Co(III) center’s coordination sphere. C–H bonds are omitted for clarity.

Scheme 1

Hydrolytic kinetic resolution of terminal epoxides catalyzed by (salen)Co(III) complexes.

Scheme 2

Proposed catalytic mechanism for epoxide hydrolysis by (salen)Co(III) complexes.