Cobalt-catalysed alkene hydrogenation: a metallacycle can explain the hydroxyl activating effect and the diastereoselectivity

A new non-redox metallacycle mechanism explains the substrate preference, the diastereoselectivity, and the hydroxyl activating effect in cobalt-catalyzed alkene hydrogenation.


ALTERNATE REDOX MECHANISM 1: HYDRIDE TRANSFER TO THE NON-METHYLATED CARBON
. Computed free energies (kcal/mol) for hydrogenation of terpinen-4-ol to the cis-product via a mechanism in which the hydride transfer occurs to the non-substituted sp 2 carbon atom of the substrate (TS 2-3R_2 ) resulting in cobalt coordination to the methylated carbon atom. This is followed by proton transfer to the methylated carbon (TS 4-5R_2 ). The overall barrier for this mechanism (24.4 kcal/mol) exceeds the redox pathway described in Figure 3 of the main text (barrier of 23.6 kcal/mol).  Figure S3. Computed free energies (kcal/mol) for hydrogenation of terpinen-4-ol to the cis-product via a mechanism in which hydride transfer occurs to the methyl-substituted sp 2 carbon atom of the substrate (TS 2-3R ) followed by β-hydride elimination from the carbon atom adjacent to the OR substituted carbon,  Figure S4. Computed free energies (kcal/mol) for hydrogenation of terpinen-4-ol to the cis-product via a non-redox σ-bond metathesis mechanism, in which the hydride transfer occurs to the methylsubstituted sp 2 carbon atom of the substrate (TS 2-3R ), followed by H 2 coordination (TS 3-4R_4 ) and proton transfer to the substrate (TS 4-5R_4 ). The overall barrier for this mechanism (28.5 kcal/mol) exceeds the redox pathway described in Figure 3 of the main text (barrier of 23.6 kcal/mol). For optimised coordinates, see XYZ file.

GENERATION OF THE ACTIVE CATALYST
We predict that generation of the active species 1 M occurs through σ-bond metathesis from the precatalyst C1 (Fig. S5). The starting trimethylsilyl (TMS) complex can undergo ligand substitutions by   either splitting H 2 or by activating the O-H bond of a substrate molecule, each liberating one TMS group. H 2 splitting (barrier of 29.3 kcal/mol) followed by splitting of an OH group (barrier 7.7 kcal/mol) leads to formation of 1 M , with a relative energy of -32.3 kcal/mol. Formation of the dihydride complex has a higher barrier for the second step (12.0 kcal/mol) and results in a species of higher energy (-23.6 kcal/mol). This indicates that in presence of OH-bearing substrates, formation of 1 M is preferred over formation of the previously proposed dihydride species. Formation of a bis-alkoxide species (-28.9 kcal/mol) has a higher barrier (32.5 kcal/mol) than initial H 2 splitting (29.3 kcal/mol).  Figure S5. Energetic flow chart (kcal/mol) comparing transformation of the precatalyst C1 to either a dihydride, the mono-alkoxide 1 M or a bis-alkoxide.   Figure S9. Comparison of computed free energies (kcal/mol) for hydrogenation of terpinen-4-ol to the cisor trans-product, respectively, via a mechanism in which the hydride transfer occurs to the nonsubstituted sp 2 carbon atom of the substrate (TS 2-3R_2 ). This is followed by proton transfer to the methylated carbon (TS 4-5R_2 ). Here black represents the cis-product, while blue represents the trans product. The schematic drawings reflect the cis-geometries. For optimised coordinates, see XYZ file.  Figure S10. Comparison of computed free energies (kcal/mol) for hydrogenation of terpinen-4-ol to the cisor trans-product, respectively, via a mechanism in which the hydride transfer occurs to the methylsubstituted sp 2 carbon atom of the substrate (TS 2-3R ) followed by β-H elimination from the carbon atom adjacent to the OR substituted carbon, TS 3-4R_3 . Black represents the cis-product, blue the trans-product.

S9
Note that this is the most favorable pathway computed for formation of the trans-product (barrier 24.2 kcal/mol relative to 1 R ). The drawings reflect the cis-geometries. For optimised coordinates, see XYZ file.    SiMe 4 using chemical shifts of the solvent as a secondary standard.
Gas chromatography for the alkane products was performed on a Shimadzu GC-2010 gas chromatograph. GC analyses were performed using a Restek 15 m x 0.25 mm RTX-5 5% diphenyl/95% dimethyl polysiloxane column with a film thickness of 0.25 μm.
Single crystals suitable for X-ray diffraction were coated with polyisobutylene oil in a glovebox, transferred to a nylon loop and then quickly transferred to the goniometer head of a Bruker D8 APEX3 Venture diffractometer equipped with a molybdenum X-ray tube (λ = 0.71073 Å) and a Cu X-ray tube (λ = 1.54178 Å). Preliminary data revealed the crystal system. The data collection strategy was optimised for completeness and redundancy using the Bruker COSMO software suite. The space group was identified, and the data were processed using the Bruker SAINT+ program and corrected for absorption using SADABS. The structures were solved using direct methods (SHELXS) completed by subsequent Fourier synthesis and refined by full-matrix least-squares procedures.

Experimental Procedures:
The catalyst dppeCo(CH 2 SiMe) 3 (C1) was synthesised using a previously reported method. vii Catalytic hydrogenation reactions were carried out in thick-walled glass vessels under 1 or 4 atm H 2 using the previously reported procedures. vii

Synthesis of deuterium-labeled terpinen-4-ol:
In a nitrogen-filled glovebox, a 20 mL scintillation vial was charged with 1.00 g (6.48 mmol) of terpinen-4-ol, 10 mL of THF and a stir bar. The solution was cooled at 77 K. n BuLi solution (7.13 mmol, 2.8 mL, 2.5 M in hexane) was diluted using 10 mL hexanes and cooled at 77 K. Into the thawing solution of terpinen-4-ol, the n BuLi solution was added dropwise. The mixture was stirred at room temperature for 1 hour, and 1 mL of CD 3 OD (99.8% atom %D) was added dropwise to the mixture. The resulting solution was stirred at r.t. for 15 minutes, and the volatiles were removed en vacuo. The residue was reconstituted in toluene, filtered through celite followed by passage through a thin pad of alumina. The volatiles were removed en vacuo to afford (0.94 g, 93% yield) of deuterium-labeled terpinen-4-ol (C 10 H 8 -OD Hydrogenation of deuterium-labeled alcohol: (96% D-incorporation) Figure S12. Hydrogenation of deuterium-labeled terpinen-4-ol. 1 H-NMR integration suggests 1,2-H 2 alkane product. No observable deuterium-attached carbons were found in 13 C-NMR.

Evidence for protonolysis of cobalt dialkyl by terpinen-4-ol and formation of cobalt alkoxide:
In a nitrogen-filled glovebox, a 20 mL scintillation vial was charged with 0.015 g of dppeCo(CH 2 SiMe 3 ) 2 , 0.5 mL of benzene-d 6 . The resulting solution was transferred into a J-Young tube and cooled at 77 K. Into another 20 mL scintillation vial was charged with 0.018 g of terpinen-4-ol and 0.5 mL of benzene -d 6 . The resulting solution was cooled at 77 K, and upon thawing, transferred into the J-Young tube. The tube was quickly sealed, brought outside the box, and let stand at room temperature for 12 hours. Vacuum distillation of the volatiles of the reaction content (benzene-d 6 , SiMe 4 , unreacted terpinen-4-ol) under high vacuum into another J-Young tube was performed, and 1 H-NMR of the volatiles identifies both SiMe 4 product and unreacted excess terpinen-4-ol. The non-volatile components were brought into the glovebox, transferred into a 20 mL-scintillation vial in THF. The volatiles were removed, affording 0.016 g of residue. Into the residue was added 5 mL THF and 0.095 g (2.0 equiv, assuming mono-alkoxy product based on 1 H-NMR integration of volatiles) of TMSI. The mixture was stirred at r.t. for 12 h, and the volatiles (THF, unreacted TMSI) was removed en vacuo. The residue (dppeCo(CH 2 SiMe 3 ) x (I) n , TMSterpinen-4-ol) was dissolved in benzene-d 6 into a J-Young tube, and the contents were distilled on high vacuum line into another J-Young tube. 13 C-NMR of the volatiles confirmed the identity of TMS-terpinen-4-ol.  Figure S13. Reaction between dppeCo(CH 2 SiMe 3 ) 2 and substrate occurs at room temperature to generate SiMe 4 and a proposed cobalt alkoxide species S13 Catalyst deactivation occurs from reacting with H 2 in the absence of substrate: In a nitrogen-filled glovebox, a thick-walled glass vessel was charged with 0.100 g of dppeCoCH 2 SiMe 3 , 5 mL Et 2 O, a stir bar and sealed. The vessel was bought outside the box and attached to a high vacuum line. The contents of the vessel were frozen in liquid nitrogen, and the head space of N 2 was removed. Then 4 atm H 2 was added to the vessel and sealed. The mixture was stirred for 15 minutes at room temperature and the contents were frozen in liquid nitrogen again. The headspace of H 2 was removed and the vessel was brought back into the glovebox. The mixture was concentrated en vacuo, affording a red solid in 0.086 g yield and was sampled for 1 H-NMR analysis. The solid was tested for catalytic activity, and hydrogenation of terpinen-4-ol using the sample solid at 2.5 mol% loading afforded no desired alkane product.
Crystallization from a saturated pentane solution of the red solid at -35 °C overnight afforded crystals suitable for x-ray diffraction.  Figure S14. Treating dppeCo(dialkyl) with H 2 in the absence of substrate leads to rapid formation of catalytically inactive bis(ligand)-cobalt species.