Photo-Organocatalytic Enantioselective Radical Cascade Reactions of Unactivated Olefins

Cascade processes are powerful strategies for rapidly increasing structural and stereochemical complexity while delivering complex chiral molecules in one step.¹ The high reactivity and selectivity of open-shell intermediates² makes radical chemistry perfectly suited for implementing cascades.³ However, the intrinsic challenge of carrying out reactions with radicals in a stereocontrolled fashion⁴ has greatly limited the development of enantioselective variants. The majority of the reported strategies rely on either chiral substrates or stoichiometric ligands.^{3c, 4, 5} Recently, photoredox catalysis⁶ has provided effective tools to enable single-electron transfer (SET) mediated radical cascade processes under mild reaction conditions.^3a The combination with chiral Lewis acid catalysts has allowed the ensuing radical cascade process to be channeled towards a stereocontrolled pattern.⁷ Although effective, these catalytic asymmetric methods are generally limited to either formal [2+2]^7a or [3+2]^7b,^7c photocycloaddition reactions or cation radical Diels–Alder processes.^{7d, 8}

Recently, our laboratories identified a new photochemical catalytic mode of substrate activation that exploits the excited-state reactivity of chiral iminium ions (I) to generate radicals upon SET oxidation of suitable substrates (II; Figure 1 a).⁹ The resulting chiral intermediate III can also govern the ensuing radical coupling step with high stereoselectivity. We wondered if this new photochemical activation mode could be used to expand the synthetic potential of enantioselective radical cascades beyond the reactivity constraints of formal cycloaddition chemistry.⁸ In an initial attempt, we combined the excited-state iminium ion chemistry with the ground-state reactivity of enamines to access cyclopentanols, from cyclopropanols, with high stereoselectivity (Figure 1 b).¹⁰ However, while the first carbon–carbon bond-forming step in this cascade sequence is a radical process, the subsequent carbon–carbon bond formation follows a two-electron pathway. Herein, we detail how the photochemistry of iminium ions has served to successfully implement an enantioselective catalytic radical cascade reaction that combines two sequential radical-based bond-forming processes without mechanistically relying on a cycloaddition manifold (Figure 1 c). This organocatalytic photochemical process converts readily available enals and unactivated olefins, bearing an oxygen-centered nucleophilic handle, into complex chiral molecules containing butyrolactone^11a or tetrahydrofuran moieties.^11b These two structural elements are present in numerous biologically active and naturally occurring molecules.

Figure 2 details our proposed strategy to develop an enantioselective SET-triggered radical cascade. We envisioned a catalytic cycle wherein a chiral amine catalyst would form the colored iminium ion intermediate I upon condensation with the enal 1. Selective excitation with a violet-light-emitting diode (LED) turns I into a strong oxidant (I*), as implied by its reduction potential, which was estimated at about +2.4 V (E*_red (I*/I^.−) vs. Ag/Ag⁺ in CH₃CN) using electrochemical and spectroscopic methods.^9a Therefore, the photoexcited iminium ion can oxidize an alkene substrate (2)¹² adorned with a suitable oxygen-centred nucleophilic handle (alkenes 2 used in this study have oxidation potentials ranging from +2.03 to +2.19 V vs. Ag/AgCl in MeCN, see section D in the Supporting Information for details). The SET activation of 2 would concomitantly form the chiral 5π-intermediate III and the alkene radical cation IV. The nucleophilic moiety within IV would then trigger a polar-radical-crossover cycloaddition,^{13, 14} capitalizing on the simultaneous polar and radical nature of such radical cation species. This process would proceed with an anti-Markovnikov selectivity to afford the more stable tertiary radical intermediate V.¹⁵ At this juncture, a stereocontrolled radical coupling with III would afford the cascade product 3 with high enantioselectivity. This mechanistic plan finds support in the high reactivity of the distonic intermediate IV, which has been extensively investigated by Nicewicz to develop anti-Markovnikov hydrofunctionalizations of olefins,^{14, 15, 16} and its tendency to initiate cycloaddition sequences leading to heterocyclic products.^{14, 16}

To test the feasibility of our photochemical plan, we selected cinnamaldehyde (1 a) and the gem-difluorinated diarylprolinol silylether catalyst A^9a for the formation of the chiral iminium ion (Table 1). The experiments were conducted in CH₃CN using a single high-power (HP) LED (λ_max=420 nm) with an irradiance at 30 mW cm⁻², as controlled by an external power supply (full details of the illumination set-up are reported in Figure S1 of the Supporting Information). The alkenoic acid 2 a was selected as the model substrate based on its oxidation potential (E_p^ox (2 a^.+/2 a)=+2.09 V vs. Ag/AgCl in MeCN) and its tendency to undergo SET oxidation mediated anti-Markovnikov lactonization.^16a Using 40 % of trifluoroacetic acid (TFA) to assist the formation of I, the desired product 3 a was obtained at ambient temperature in 55 % yield, 88 % ee, and 1:1 d.r. (entry 1). Increasing the amount of TFA to 60 mol % improved the yield to 62 % (entry 2). The chiral amine catalyst B, possessing bulkier perfluoro-isopropyl groups on the arene scaffold, improved the enantioselectivity (91 % ee, entry 3), whereas the chiral imidazolidinone catalyst C afforded the product with low stereocontrol (entry 4). Implementation of a two-phase solvent system, using tetradecafluorohexane in a 1:1 ratio to CH₃CN, further increased the yield of the reaction (72 % yield, entry 5).¹⁷ Control experiments indicated that both the light and the amine catalyst are required for reactivity (entries 7 and 8), while the addition of a radical trapping agent (TEMPO, 1 equiv) completely inhibited the process.

Table 1. Optimization studies.^[a]

Entry	Catalyst	Reaction conditions	Yield [%][b]	ee [%][c]
1	A	40 % TFA	55	88
2	A	60 % TFA	62	88
3	B	60 % TFA	63	91
4	C	60 % TFA	60	33
5	B	CH3CN/HexF (1:1)	72 (58)	91
7	A	no light	0	–
8	–	no catalyst	0	–

• [a] Reactions performed on a 0.1 mmol scale at 30 °C for 16 h using 0.2 mL of solvent under illumination by a single high-power (HP) LED (λ_max=420 nm) with an irradiance of 30 mW cm⁻². [b] Yield of 3 a determined by ¹H NMR analysis of the crude reaction mixture using trichloroethylene as the internal standard. Yields of the isolated 3 a are reported within parentheses. [c] Enantiomeric excess determined by HPLC analysis on a chiral stationary phase. Hex_F=tetradecafluorohexane, TDS=tert-hexyldimethylsilyl, TFA=trifluoroacetic acid.

Using the optimized reaction conditions described in entry 5 of Table 1, we then turned our attention to the scope of the photocatalytic asymmetric radical cascade process (Figure 3). We first evaluated the possibility of using differently substituted olefins. The presence of geminal methyl groups at the α-position of 2 significantly improved the yield and enantioselectivity (adduct 3 b, 75 % yield and 94 % ee). An ethyl fragment could also be installed (3 c), while the inclusion of a cyclohexyl group afforded the spirocyclic adduct 3 d in moderate yield and high enantioselectivity. The absence of α-substituents in 2 did not significantly affect the overall efficiency (product 3 e). This observation indicates that the Thorpe–Ingold effect is not critical to the cascade process. Concerning the substitution of the double bond, introducing a cyclohexyl in place of the dimethyl moiety slightly increased the diastereoselectivity to 1.3:1 (3 f) while maintaining a high chemical yield and enantioselectivity. Importantly, tetrahydrofuran derivatives 4 a and 4 b could also be synthesized when decorating the alkene substrate with an alcohol acting as the nucleophilic moiety.

Crystals from the compound 3 a were suitable for X-ray crystallographic analysis,¹⁸ which established the stereochemical course of the radical cascade process.

Regarding the scope of the α,β-unsaturated aldehydes 1, moderately electron-withdrawing and electron-donating groups on the aromatic ring are tolerated without diminishing either the yield or enantioselectivity (Figure 4). The substituent can also be moved around the ring without impeding reactivity (3 g–l). Notably, synthetically useful groups, which can undergo further functionalization, were tolerated. These included halogens (3 g–j), trimethylsilyl (3 m), and pinacolborane (3 n). One limitation of this reaction is that aliphatic enals, including tert-butylacroleine and (E)-2-octenal, remained unreacted under the reaction conditions.

We then focused on the design of a three-component radical cascade process to assemble complex chiral molecules from simple substrates in a single step. Specifically, we surmized that the excited iminium ion I* could oxidize amylene (5), an unfunctionalized alkene hydrocarbon (E_p^ox (5^.+/5)=+2.03 V vs. Ag/AgCl in MeCN; Figure 5).¹⁹ The anti-Markovnikov intermolecular addition of a carboxylic acid²⁰ to the less substituted position of the resulting VI would then provide the key radical intermediate VII, which would eventually engage in a stereoselective radical coupling governed by the chiral 5π-intermediate III (structure shown in Figure 2). This catalytic asymmetric cascade process, which combines two intermolecular radical steps, would provide a one-step access to complex aldehydes (7).

The use of α,α-difluorinated acids (6) enabled the successful realization of this enantioselective three-component cascade. The aldehydes 7 were converted into enoates 8 a–c, which could be isolated in moderate yields and enantioselectivities (Figure 5). The acids play a dual role in this transformation, since they assist the iminium ion formation while acting as nucleophilic substrates in the three-component cascade.

In summary, the photochemical activity of a chiral iminium ion has been exploited to activate unfunctionalized olefins and trigger an enantioselective radical cascade process. We expect that this photochemical organocatalytic activation mode could be useful for designing other radical cascade processes and for stereoselectively synthesizing more complex chiral scaffolds.