An unexpected synthesis of azepinone derivatives through a metal-free photochemical cascade reaction

Azepinone derivatives are privileged in organic synthesis and pharmaceuticals. Synthetic approaches to these frameworks are limited to complex substrates, strong bases, high power UV light or noble metal catalysis. We herein report a mild synthesis of azepinone derivatives by a photochemical generation of 2-aryloxyaryl nitrene, [2 + 1] annulation, ring expansion/water addition cascade reaction without using any metal catalyst. Among the different nitrene precursors tested, 2-aryloxyaryl azides performed best under blue light irradiation and Brønsted acid catalysis. The reaction scope is broad and the obtained products underwent divergent transformations to afford other related compounds. A computational study suggests a pathway involving a step-wise aziridine formation, followed by a ring-expansion to the seven-membered heterocycle. Finally, water is added in a regio-selective manner, this is accelerated by the added TsOH.

CH2Cl2 (3 × 30 mL). The combined organic layer was dried over anhydrous Na2SO4, filtered and concentrated. The residue was purified by silica gel column chromatography using a mixture of petroleum ether and EtOAc as eluent to give 2-aryloxy anilines (70-98% yield).
To a mixture of substituted 2-aryloxy aniline (5.00 mmol, 1 eq.) in EA (9.2 mL) and H2O (1.2 mL) was added conc. HCl (2.7 mL). The resulting miture was stirred for 10 min at room temperature and then it was cooled to 0 °C and a solution of NaNO2 (583 mg, 8.45 mmol, 1.69 eq.) in H2O (2.0 mL) was added slowly. Upon completion of the addition, stirring continued for 30 min at 0 °C followed by a slow addition of NaN3 (553 mg in 2 mL H2O, 8.50 mmol, 1.7 eq.) aqueous solution. After 30 min at 0 °C, the reaction mixture was diluted with 15 mL water, extracted with EtOAc (3 × 20 mL). The combined organic layer was washed with 5% NaOH solution (1 × 50 mL), then with water (1 × 50 mL), dried over Na2SO4, filtered and concentrated to afford azide 3. Under nitrogen atmosphere, to a solution of substrate 3 (0.2 mmol) in 2.0 mL THF were added TsOH·H 2O (19 mg, 0.1 mmol, 0.5 eq.) and H2O (36 mg, 2.0 mmol, 10 eq.). The resulting mixture was irradiated at room temperature with 29 W blue LEDs for 48 h. The solvent was removed under reduced pressure and the residue was purified by silica gel column chromatography using a mixture of petroleum ether and EtOAc as eluent to provide the desired product 5. To a solution of 5a (48 mg, 0.2 mmol, 1.0 eq.) in 2.0 mL acetone were added K2CO3 (42 mg, 0.3 mmol, 1.5 eq.) and CH3I (43 mg, 0.3 mmol, 1.5 eq.). The resulting mixture was heated to reflux for 3 h. Then the reaction mixture was cooled, diluted with water (15 mL) and extracted with EA (3 × 15 mL). The combined organic layer was dried over Na2SO4, filtered and concentrated. The residue was purified by silica gel column chromatography using a mixture of petroleum ether and EtOAc as eluent to provide the desired product 6 (46 mg, 92%) To a solution of 5a (47 mg, 0.2 mmol, 1.0 eq.) in 2.0 mL acetone was added K2CO3 (110 mg, 0.8 mmol, 4.0 eq.) at room temperature. The resulting mixture was stirred at 70 °C for 0.5 h followed by the dropwise addition of 3-bromoprop-1-yne (48 mg, 0.4 mmol, 2.0 eq.) and then stirring continued for 8 h at 70 °C. Upon completion, the reaction mixture was filtered through a pad of silica gel. The filtrate was concentrated under reduced pressure and the residue was purified by flash silica gel column chromatography to give product 7 (47 mg, 87%). To a solution of 5b (140 mg, 0.7 mmol, 1.0 eq.) and pyridine (0.12 mL, 1.4 mmol, 2.0 eq.) in anhydrous CH2Cl2 (5 mL) was added dropwise a solution of Tf2O (0.14 mL, 0.84 mmol, 1.2 eq.) in 5 mL anhydrous CH2Cl2 at 0 °C. The resulting mixture was stirred at room temperature for 1.5 h and then quenched by the addition of 1 mL H2O. The resulting mixture was transferred into a separating funnel followed by the addition of 10% NaHCO3 aqueous solution (20 mL) and 15 mL CH2Cl2. The organic layer was collected and the aqueous phase was extracted with CH2Cl2 (2 × 20 mL). The combined organic layer was dried over Na2SO4, filtered and concentrated. The residue was purified by column chromatography on silica gel to provide compound 8 (163 mg, 70%). A mixture of Pd(PPh3)4 (5.0%, 12 mg, 0.01 mmol), AuCl(PPh3) (5.0%, 5 mg, 0.01 mmol) and compound 8 (67 mg, 0.2 mmol, 1.0 eq.) in an oven-dried flask was degassed and backfilled with nitrogen. A solution of ethynylbenzene (32 mg, 0.3 mmol, 1.5 eq.) and Et3N (62 mg, 0.6 mmol, 3.0 eq.) in degassed DMF (1.0 mL) was added via syringe. The resulting mixture was stirred at 90 °C for 7 h, cooled to room temperature and then filtered. The filtrate was evaporated and the residue was purified by silica gel column chromatography to afford product 9 (54 mg, 95%). To a solution of 5 g (118 mg, 0.38 mmol, 1.0 eq.) and pyridine (60 mg, 0.76 mmol, 2.0 eq.) in anhydrous CH2Cl2 (3 mL) was added dropwise a solution of Tf2O (0.08 mL, 0.46 mmol, 1.2 eq.) in 3 mL anhydrous CH2Cl2 at 0 °C. The resulting mixture was stirred at room temperature for 1.5 h and then quenched by the addition of 1 mL H2O. The resulting mixture was transferred into a separating funnel followed by the addition of 10% NaHCO3 aqueous solution (20 mL) and 15 mL CH2Cl2. The organic layer was collected and the aqueous phase was extracted with CH2Cl2 (2 × 20 mL). The combined organic layer was dried over Na2SO4, filtered and concentrated. The residue was purified by column chromatography on silica gel to provide compound 11 (156 mg, 93%). A mixture of Pd(PPh3)4 (3.0%, 6.9 mg, 0.006 mmol), arylboronic acid (0.95 mmol, 3.0 eq.), compound 11 (140 mg, 0.32 mmol, 1.0 eq.) or 8 (67 mg, 0.2 mmol, 1.0 eq.), toluene (4.0 mL), ethanol (1.0 mL), H2O (1.0 mL) and K2CO3 (150 mg, 1.1 mmol, 3.4 eq.) was degasses and stirred at 100 °C for 6 h under nitrogen. Upon completion, the reaction was neutralized by 5% aqueous HCl, then the aqueous phase was separated and further extracted with EtOAc (2 × 5 mL). The combined organic layers were dried over Na2SO4, filtered and concentrated. The residue was purified by silica gel column chromatography to afford product 12 or 10.

Additional calculation results and computational details
Benchmark calculations. The transformation given in Supplementary Fig. 15 was experimentally reported by Ellison et al. to have a ΔEST value of 18 ± 2 kcal/mol in the gas phase 7 . Consistent with this experimental finding, this ΔEST is calculated to be 19.4 kcal/mol at the CCSD(T)/def2-TZVP//M06-2X/6-31G(d) level of theory. To find out which DFT method gives more accurate result, the single point calculations were carried out at the different levels of theory using the Method/def2-TZVP//M06-2X/6-31G(d) calculations. As can be seen from Supplementary Fig. 15a, compared to other methods, the ΔEST calculated by the CAM-B3LYP functional is in closer agreement with the CCSD(T) reference and the experimental value. To investigate whether this is also true for a different nitrene molecule, we calculated the ΔEST for the transformation given in Supplementary  Fig. 15b at different levels of theory, SDD/Method/def2-TZVP//SDD/M06-2X/6-31G(d), in THF and again found that the CAM-B3LYP functional gives a ΔEST closer to that obtained by the CCSD(T) reference.
Although our calculations indicated that M06-2X is not a very accurate method for estimating the ΔEST of nitrenes, it is benchmarked as a reliable method for studying pericyclic reactions ( Supplementary Fig. 15c). We found that compared to other functionals, the M06-2X functional show better agreement with the CCSD(T) reference; for this benchmark evaluation, the calculations were carried out at the SDD/Method/def2-TZVP//SDD/M06-2X/6-31G(d) level of theory in THF.
As can be seen from Fig. 7a, transformation 3a  A T + N2 is calculated to be exergonic. This exergonicity is supported by an additional calculation at the SMD/CCSD(T)/def2-TZVP//SMD/M06-2X/6-31G(d) level of theory in THF for N2 release from phenyl azide with ΔG = -11.1 kcal/mol ( Supplementary Fig. 15d). This result explicitly indicates that the liberation of N2 from aryl azides to give a triplet nitrene is most likely a downhill process.

Computational details
We used TD-DFT calculations using the Coulomb-attenuating method functional by Handy and coworkers 8 (CAM-B3LYP) to optimize structures and investigate the mechanism of photoactivation using Q-Chem 5.4. 9 For this part of the calculations, solvent effects were considered using the closely-related conductor-like PCM (C-PCM) model 10 with tetrahydrofuran as the solvent.
Gaussian 16 11 was used to fully optimize all the structures reported in this paper at the M06-2X level of theory. 12 It is well-documented that the M06-2X functional gives more accurate activation barrier and reaction energies than other DFT methods for an organic transformation. 13 For this part of the calculations, solvent effects were considered using the SMD solvation model 14 with tetrahydrofuran as the solvent.
The 6-31G(d) basis set 15 was used for optimization. Frequency calculations were carried out at the same level of theory as those for structural optimization. Transition structures were located using the Berny algorithm. Intrinsic reaction coordinate (IRC) calculations were used to confirm the connectivity between transition structures and minima. 16 To further refine the energies obtained from the 6-31G(d) calculations, we carried out single-point energy calculations using the def2-TZVP basis set. 17 Tight convergence criterion and ultrafine integral grid were also employed to increase the accuracy of all calculations. In this work, the free energy for each species in solution was calculated using the following formula: G = E(def2-TZVP) + G(6-31G(D)) -E(6-31G(D)) + ∆G 1atm→1M (5) where ∆G 1atm→1M = 1.89 kcal/mol is the free-energy change for compression of 1 mol of an ideal gas from 1 atm to the 1 M solution phase standard state. In simple word, the ∆G 1atm→1M term results in a correction of -1.89 (or +1.89) kcal/mol for a 2 to 1 (or a 1 to 2) transformation.
Minimum energy crossing points (MECPs) between closed-shell singlet and triplet states were located using both Q-Chem for the photoactivation part and the code developed by Harvey et al 18