Light Induced Cobalt(III) Carbene Radical Formation from Dimethyl Malonate As Carbene Precursor

Radical-type carbene transfer catalysis is an efficient method for the direct functionalization of C–H and C=C bonds. However, carbene radical complexes are currently formed via high-energy carbene precursors, such as diazo compounds or iodonium ylides. Many of these carbene precursors require additional synthetic steps, have an explosive nature, or generate halogenated waste. Consequently, the utilization of carbene radical catalysis is limited by specific carbene precursors that access the carbene radical intermediate. In this study, we generate a cobalt(III) carbene radical complex from dimethyl malonate, which is commercially available and bench-stable. EPR and NMR spectroscopy were used to identify the intermediates and showed that the cobalt(III) carbene radical complex is formed upon light irradiation. In the presence of styrene, carbene transfer occurred, forming cyclopropane as the product. With this photochemical method, we demonstrate that dimethyl malonate can be used as an alternative carbene precursor in the formation of a cobalt(III) carbene radical complex.


DFT calculations
DFT geometry optimizations were performed without simplifications on full atomic models using TURBOMOLE 7.5.1, 2 coupled to the PQS Baker optimizer, 3,4 via the BOpt package. 5All calculations were performed in the gas phase with convergence criteria (scfconv = 7) on a m4 grid and Grimme's version 3 zero-damping dispersion corrections to compensate for the underestimation of metal-ligand interactions from uncorrected DFT calculations. 6For a description of the functional and basis sets used, see below.All minima, without imaginary frequencies, were characterized by calculating the analytical Hessian matrix.Energy output generated in Hartree units was converted to kcal/mol by multiplication with 627.51.Graphical representations of orbitals are obtained using IboView 7 and visualization of spin densities using IQMol. 8lculation of the bond dissociation energies (BDEs) 1) All input structures were generated as protein database files (.pdb).
3) Depending on the bond, the BDE was calculated via the following calculations: a. R-X bond: R-H bond: The DFT calculated BDEs were calculated using a semiempirical Hess cycle, as shown in Figure S1. 11,12 ure S1.Hess cycle used to calculated the R-H BDE in a semi-empirical manner.

Synthesis of [Co III (TPP)Cl]
[Co III (TPP)Cl] was synthesized according to a literature procedure. 14[Co II (TPP)] (300 mg, 0.44 mmol, 1 eq) was suspended in MeOH (300 mL) and HCl (3 mL, 12 M) was added dropwise.The purple suspension was stirred for 3 hours without a stopper on the flask.The red solution was filtered and the filtrate was concentrated in vacuo until a green precipitate was formed.It was filtered and washed with H2O (100 mL) and MeOH:H2O 1:1 (30 mL).The purple powder was dried over P2O5 overnight and stored in the glovebox immediately after.Purple powder (92%).
Note: it is important to store the obtained [Co III (TPP)Cl] under dry conditions immediately after synthesis to avoid hydration and thus deactivation (see section 3.3).

Oxidation experiments:
Table S2.The cyclopropanation reaction between dimethyl malonate and styrene with oxidants as additives (under aerobic conditions).a

Moisture affecting [Co III (TPP)Cl]:
During investigation of the cyclopropanation reaction, we occasionally encountered inactivity when the reaction was performed according to the conditions of entry 1 in Table S1.We hypothesized that moisture affects [Co III (TPP)Cl] over time, leading to inactivity during the cyclopropanation reaction.UV-Vis and HRMS spectroscopy was used to determine changes in the structure of [Co III (TPP)Cl] (Figure S12).When the UV-Vis spectrum was measured immediately after [Co III (TPP)Cl] was synthesized, a Soret band at 405 nm and a Q-band at 543 nm was observed (Figure 12a, black line).After 126 days, the UV-Vis spectrum was measured again (blue line).A small change around 440 nm can be observed when comparing the two spectra.Upon addition of an excess of H2O, a shoulder was observed at the same wavelength (440 nm) (Figure S12b).HRMS measurements confirmed our hypothesis of moisture affecting the cobalt complex (Figure S12c).In the presence of an excess of H2O, a signal at 688.1713 m/z was observed that corresponds to [Co III (TPP)OH].The same signal, although significantly less intense, was observed when no H2O was added to the sample.This result demonstrates that a small amount of water changes the composition of the cobalt metal complex.When newly synthesized [Co III (TPP)Cl] was dried over P2O5 and the reaction was prepared in the glovebox, cyclopropane 3 could be observed again in 10% yield (based on 1 H NMR analysis).
Figure S16.X-Band EPR spectrum obtained upon irradiation (370 nm) of in situ formed complex   15 using the cwEPR plugin. 16Hyperfine couplings in MHz.Parameters based on previously reported calculations and used for the spectral simulations. 17

NMR studies General procedure NMR studies
To a flame-dried 10 mL Schlenk flask, complex 1 (Figure S21, 0.00137 mmol, 1 eq.), benzophenone (Figure S22, 0.002 mmol, 1.5 eq.) and benzene (2 µL) as internal standard were dissolved in CD2Cl2 (0.5 mL).The dark red solution was freeze-pump-thawed and transferred to a J-young NMR tube after filtration.The tube was kept under dark conditions until the NMR measurements were performed.All NMR samples were irradiated with a 390-500 nm light source via a light probe attached to the bottom of the NMR machine.To determine the concentration of complex 1, signals at Figure S22 shows the conversion of complex 1 in presence of benzophenone over time.Initially, an increase in rate was expected when benzophenone was present in the NMR sample.However, the sample was irradiated with a 390-500 nm light source, a region in which benzophenone does not absorb.The effect of benzophenone on the rate of the conversion of complex 1 can therefore not be concluded based on this NMR experiment.We do observe an increase in yield when benzophenone is present (Table S1, entry 1 and entry 4), proving that benzophenone has a positive effect on the formation of cyclopropane in the cyclopropanation reaction.

General procedure UV-Vis studies
The UV-Vis spectrum of complex 1 was also measured over time (Figure S23).Complex 1 (0.690 mg) was dissolved in DCM (10 mL) inside a N2-filled glovebox.The red solution was diluted to obtain a concentration of 21.5 µM and it was transferred to a J-young UV-Vis cuvette.The UV-Vis cuvette was kept in the dark before the measurements started.In between the measurements, the cuvette was irradiated with a green light source (525 nm, 43 W) for a defined period of time.
Under dark conditions, the UV-Vis spectrum of complex 1 remains constant (Figure S23a).Next, we irradiated the sample with 525 nm (43 W) (Figure S23b), because this wavelength is the lowest energy at which complex 1 still absorbs light.After 5 minutes, the shoulder at 375 nm disappeared and a blue shift of the Soret-band (initially 413 nm) was observed with a higher intensity.After 2.5 hours, the Soret-band lowered in intensity again and showed a stronger blue shift to 407 nm.The Q-band moved from 525 nm to 545 nm.Complex 1 clearly converts to other species upon irradiation of green light.While the nature of these species cannot be derived from the UV-Vis spectra, it is most likely that a mixture of species is formed, including [Co II (TPP)] and the cobalt(III) carbene radical complex.

Figure S11 .
Figure S11.General crude 1 H NMR spectrum of the cyclopropanation reaction under conditions according to entry 1 in TableS1.1,3,5-tri-tertbutylbenzene was added as an external standard.

Figure S21. a )
Figure S21.a) Conversion of complex 1 under dark conditions (blue, 2.75 mM) and upon 390-500 nm light irradiation (purple, 2.68 mM) in DCM-d2.b) Conversion of complex 1 and formation of [Co II (TPP)] upon 390-500 nm light irradiation in DCM-d2.c) 1/[Complex 1] plotted against time.Outliers in red do not contribute to the fitting.

Figure
Figure S23.a) UV-Vis spectrum of complex 1 (21.5 µM) in DCM under dark conditions.b) UV-Vis spectrum of complex 1 (21.5 µM) in DCM upon 525 nm (43 W) light irradiation measured over time.All spectra were normalized at 725 nm.
a   a