Visible-Light-Driven Stereoselective Annulation of Alkyl Anilines and Dibenzoylethylenes via Electron Donor–Acceptor Complexes

A catalyst-free, stereoselective visible-light-driven annulation reaction between alkenes and N,N-substituted dialkyl anilines for the synthesis of substituted tetrahydroquinolines is presented. The reaction is driven by the photoexcitation of an electron donor–acceptor (EDA) complex, and the resulting products are obtained in good to high yields with complete diastereoselectivity. Mechanistic rationale and photochemical characterization of the EDA-complex are provided.


Emission spectrum of irradiation source
Emission spectrum of the irradiation source ( Figure S2) used for the light mediated annulation reaction (20 W 1200 lm Osram white CFL bulbs) was recorded using an AvaSpec-2048 spectrometer. Characteristic peaks are 435 nm, 541 nm, 545 nm and 610 nm.

Determination of association constant of EDA complex
The equilibrium constant of the EDA complex formed between 1a and 2a was determined by using the Benesi-Hildebrand method (Eq. 1). 1 A series of solutions of aniline (1a) and DBE (2a) in 1,4-dioxane were prepared in which the concentration of 2a was held constant (0.06 M) but with varying concentration of 1a. Absorption spectra were recorded of the samples and the 1/ΔA at 500 nm versus 1/[4'-N,N-trimethylaniline] was plotted. Association constant KEDA was calculated from linear regression ( Figure S3). Using the Benesi-Hildebrand relationship (Eq. 1) KEDA was determined to be 0.42 Where AEDA is the absorbance of the EDA complex, KEDA is the association constant for the EDA complex and εEDA is the extinction coefficient for the EDA complex. Based on the linear fitting in Figure S3, the extinction coefficient is determined from the intersection and then the equilibrium constant is calculated from the extinction coefficient and the slope of the fitted line. To verify the formation of hydrogen peroxide in the reaction mixture a series of experiments were conducted ( Figure S4). First, a mixture of potassium iodide and starch was prepared ( Figure S4A). To verify its efficiency in detecting hydrogen peroxide, a drop of 30% hydrogen peroxide in water was added and a color change to dark purple was observed. Next, the reaction mixture was prepared according to the general procedure ( Figure S4C) and a portion was mixed with the indicator solution, no color change to violet was observed indicating that no hydrogen peroxide was present before irradiation ( Figure S4D). The reaction mixture was then subjected to the light source four hours ( Figure  S4E), and a portion was mixed with the indicator solution giving a strong color change to dark purple indicating presence of hydrogen peroxide ( Figure S4F).

Reaction monitored over time
The formation of 3a as a result of the irradiation of a mixture of 1a (0.7 mmol) and 2a (0.1 mmol) in 1,4-dioxane (3 mL) was monitored over time using gas chromatography with durene as internal standard ( Figure S5). Highest concentration is reached after ca four hours. Figure S5. Concentration of product 3a over the course of the irradiation using gas chromatography and durene as internal standard.

Light on-off cycles
The formation of 3a in the model reaction (see section 1.9 for conditions) as the irradiation source was switched on and off in cycles was monitored by gas chromatography using durene as internal standard ( Figure S5). The Gibbs free energy associated with the photoinduced electron transfer process (ΔGPET) in the EDA complex between 1a and 2a was estimated using eq. 1 2-4 where NA is the Avogadro constant, e is the elementary charge, Eox(D .+ /D) and Ered(A/A .-) are the oxidation and reduction potentials of the donor and acceptor, ΔE0,0 is the energy associated to the S0 à S1 transition of the excited partner, and w describes the electrostatic attractions between donor and acceptor. The terms w can be assumed to be small in polar solvents and be omitted. Literature value of the oxidation potential of 1a is 0.72 (V vs. SCE), 5 and reduction potential of 2a is -1.52 (V vs. SCE). 6 Excited state energy of 1a can be estimated by the emission wavelength (350 nm). 4,7 Given these assumptions and values, the Gibbs free energy can be calculated: This suggests that the photoinduced electron transfer within the excited EDA complex between 1a and 2a is thermodynamically favorable.

Determination of quantum yield
Following a published procedure 1 the quantum yield of the light mediated annulation reaction was determined using a standard ferrioxalate actinometer. All solutions and samples were prepared protected from light or handled under red light: Solution A: Potassium ferrioxalate (295 mg) and 1 mL sulfuric acid was diluted to 100 mL with distilled water. The solution was carefully handled only under red light.
Solution C: 4.94 g sodium acetate and 1 mL sulfuric acid were dissolved in 100 mL distilled water.
For the measurement of the photon flux of the spectrophotometer, 1x1 cm cuvettes containing 2 mL of solution A was irradiated (450 nm excitation) for different times (1, 10, 20 and 40 minutes). After each irradiation period, the irradiated solution was transferred to a 10 mL volumetric flask containing 0.5 mL of solution B and 2 mL solution C. Water was then added to make up a total of 10 mL solution.
After being kept for 1 hour in the dark to promote full coordination of phenanthroline, the absorption of the solutions was measured at 510 nm. The amount of iron(II) formed during irradiation was determined by employing Beer-Lambert's law, molar extinction coefficient of the iron phenanthroline complex (11100 L M -1 cm -1 ), and the difference in absorbance at 510 nm between the irradiated solutions and a blank that had been kept in dark.
( ) = 1 * 3 * ∆ 9!7 .; 10 < * 2 * * 9!7 .; Where V1 is the volume of the irradiated sample, V2 is the volume of the irradiated sample taken for the determination of amount of Fe(II), V3 is the final volume of the sample measured, l is the path length of light (1 cm), ∆ 9!7 .; is the difference in absorption between the irradiated samples and the reference sample kept in dark, and 9!7 .; is the molar extinction coefficient of the iron(II) phenanthroline complex.
The amount of iron(II) was then plotted as a function of irradiation time ( Figure S7).
where dx/dt is the rate of formation of iron(II), (450 ) is the quantum yield for the ferrioxalate actinometer at 450 nm excitation (0.9) and A(450 nm) is the absorbance of the ferrioxalate actinometer at the irradiation wavelength (0.214). Consequently, the photon flux was determined to be 1.6 *10 -10 einstein s -1 . Finally, the model reaction was prepared according to the general procedure and 3 mL of the prepared solution was placed in 1x1 cm cuvettes and irradiated for 30 and 60 minutes, respectively. Solvent was then removed, and the amount of product was determined by 1 H NMR using durene as internal standard to be 1.1 µmol and 2.2 µmol for 30 and 60 minutes irradiation time, respectively. Based on the previously determined photon flux, the irradiation time t, and the absorption of light at the irradiation wavelength A(reaction mixture), the quantum yield of the reaction was determined according to eq. 3.
Consequently, a quantum yield of 4.5 was determined.
S9 9 X-ray Compound 3q Figure S8. Thermal ellipsoid plot (50% ellipsoid probability) of 3q showing the labelling of the asymmetric unit. Hydrogen atoms drawn as fixed-size spheres with a of radius 0.20 Å.
Crystals of 3q for single-crystal X-ray diffraction were grown using the layering technique from 3q in dichloromethane and fresh hexane. The colorless-yellow prism shaped crystals appeared after 4 days.

Crystal data for compound 3q:
C24H20BrNO2