A llyl carboxylates are versatile synthetic intermediates that can be applied to a wide spectrum of reliable derivatizations. For example, they can participate in transition-metal (TM)-catalyzed two-electron redox reactions f1 (Figure 1A), which often proceed through TM-π-allyl intermediates that react with nucleophiles 1 or electrophiles, 2 generating a wide array of allylated organic compounds. The olefin moiety of allyl carboxylates can also engage in oneelectron redox reactions with radical species to form alkyl radical intermediates that partake in cross-coupling reactions, furnishing valuable 1,2-difunctionalized products. 3 Despite these advances, catalytic radical 1,3-difunctionalization of allyl carboxylates involving 1,2-radical migration (RaM) remains elusive. 4 Inspired by the Surzur-Tanner rearrangement, in which a β-(acyloxy)alkyl radical undergoes 1,2-RaM via an acyloxy shift to form a more stable radical intermediate, 5 we questioned whether we could achieve the catalytic radical 1,3-difunctionalization of allyl carboxylates via a 1,2-RaM mechanism. If successful, this new reaction manifold could serve as the basis for a broadly useful set of allyl carboxylate derivatization protocols that support the preparation of a wide variety of versatile building blocks for synthesis, medicinal, and material chemistry.

Phosphorus-based catalysis has emerged as a promising strategy in organic synthesis for the formation of carboncarbon and carbon-heteroatom bonds. 6 In this area, the vast majority of the reactions, such as nucleophilic phosphine catalysis and phosphine oxide catalysis, take advantage of the two-electron redox reactivity of phosphorus compounds to generate ionic intermediates. 7 In contrast, phosphine-catalyzed one-electron pathways that generate radical species are underdeveloped. 6b,e,8 One approach to such reactions involves halogen-atom transfer (XAT) reactions 9 in which the highest such as Ph 3 P, n-Bu 3 P, and 2,2′-bis(diphenylphosphino)-1,1′binaphthyl (BINAP) gave lower yields (entries 3-5). While organic photocatalysts such as Eosin Y and a Ru-based photocatalyst were inefficient, the Ir(ppy) 3 photocatalyst afforded the desired product in 72% yield (entries 6-8). We proceeded with the dppm catalyst ($1.375/mmol) because it is more than 1000 times cheaper than Ir(ppy) 3 ($1,784/ mmol). 14 Na 2 CO 3 is critical for the reaction efficiency because no desired product was obtained in its absence (entry 9).

Other inorganic and organic bases such as K 2 CO 3 , KOAc, and 2,6-lutidine could be used, but they afforded lower yields (entries 10-12). 15 Use of acetonitrile as the solvent also diminished the reaction yield (entry 13). Finally, control experiments showed that a temperature of 90 °C, light, and an oxygen-free environment were all essential to the success of the reaction (entries 14-16).

With the optimized conditions in hand, we investigated the t2 scope of allyl carboxylates (Table 2). A series of substituted allyl carboxylates, including linear and cyclic frameworks, were well tolerated, affording the corresponding products 3a-3h in 53-88% yield. Substrates containing cyclic ether, thioether (3i and 3j), or the ethylene ketal group (3k) were also compatible.

However, allyl carboxylates with a secondary alkyl substituent or internal alkene failed under standard reaction conditions.

Next, we examined a wide range of migrating groups with different electron-donating and electron-withdrawing substituents at various positions of the aromatic ring, all of which See the Supporting Information for experimental details. Reaction yields were determined by 1 H NMR using CH 2 Br 2 as an internal standard. Bz, benzoyl; BINAP, 2,2′-bis(diphenylphosphino)-1,1′binaphthyl; LED, light-emitting diode; N.R., no reaction.

113 underwent 1,3-bromodifluoroalkylation in 51-81% yield (3l-114 3q). Allyl carboxylates with ortho-, meta-, and para-substituted 115 benzoate substituent migrating groups reacted with similar 116 efficiency. 2-Naphthoates and aliphatic esters also migrated 117 successfully, affording the desired products 3r-3v in moderate to good yields. Furthermore, we demonstrated that alternative primary, secondary, and tertiary ester substituents on the bromodifluoroacetates worked well and delivered 3w-3y in 63-69% yield. An array of bromodifluoroacetamides such as N-diethyl, piperidinyl, morpholinyl, N-adamantyl, and N-Table 2. Visible-Light-Induced Phosphine-Catalyzed 1,3-Bromodifluoroalkylation of Allyl Carboxylates a a See the Supporting Information for experimental details; 1 (1.00 equiv), 2 (3.00 equiv), dppm (10.0 mol %), Na 2 CO 3 (2.50 equiv), DCE (0.040 M), 100 W blue LED, 90 °C, 24 h. b 5.00 equiv of BrC 6 F 13 or BrC 8 F 17 was used and reacted for 48 h. conditions, forming the desired products 3aj-3an in 48-84% yield.

To showcase the scalability of this process, a gram-scale s1 reaction was performed by using substrates 1a and 2a (Scheme s1 1A). A satisfactory 82% isolated yield of 3a was obtained under the irradiation of two 100 W blue LEDs for 72 h. We also performed a series of transformations to demonstrate the synthetic utility of the 1,3-bromodifluoroalkylated products in forming useful molecular scaffolds (Scheme 1B). The bromine and ester functionality in the product could be used for subsequent synthetic manipulations. For example, selective HBr elimination was achieved in the presence of the benzoate leaving group, generating the useful terminal alkene 4. The synthetic utility of the ester was exemplified by selective reduction, hydrolysis, and amidation of the ethyl ester without affecting the alkyl benzoate, furnishing the corresponding alcohol 5, acid 6, and amide 7 in 76%, 62%, and 94% yields, respectively.

To gain a better understanding of the reaction mechanism, we carried out a series of experimental and computational f2 studies (Figure 2). The addition of 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO) as a radical scavenger significantly inhibited the reaction (Figure 2A). The TEMPO-trapped difluoroacetate adduct (8) was isolated and characterized by NMR and HRMS spectrometry. A radical clock experiment using substrate 1ao formed the ring-opening product 3ao in 42% yield. These experiments suggest a radical mechanism. We then subjected 1,2-bromodifluoroalkylated compound 9 to the standard conditions (Figure 2B). We did not obtain the desired 1,3-product, indicating that the reaction does not proceed through 1,2-bromodifluoroalkylation followed by ionic translocation of the benzoate and bromide groups. Crossover experiments using substrates 1s and 1e furnished only the noncrossover products 3s and 3e, respectively, implying that the acyloxy shift probably proceeds through concerted mechanisms (Figure 2C). 18 The O 18 -labeling experiment using 1-O 18 See the Supporting Information for experimental details; 1 (1.00 equiv), 2 (3.00 equiv), dppm (10.0 mol %), Na 2 CO 3 (2.50 equiv), DCE (0.040 M), 100 W blue LED, 90 °C, 24 h. Scheme 1. Large-Scale Reaction and Post-Functionalization a a See the Supporting Information for experimental details. (i) DBU (1.50 equiv), DMSO (0.200 M), 60 °C, 6 h; (ii) NaBH 4 (2.00 equiv), MeOH (0.200 M), 0 °C to rt, 3 h; (iii) K 2 CO 3 (5.00 equiv), MeOH/ H 2 O (1:1, 0.100 M), rt, 12 h; (iv) NH 3 (10.0 equiv), MeOH (0.100 M), rt, 16 h. D 178 with 95% O 18 incorporation afforded the products 3a-O 18 -1 179 and 3a-O 18 -2 in a 1.4:1 ratio, indicating that both [1,2]-and 180 [2,3]-acyloxy group shift pathways may operate simultaneously 181 (Figure 2D). 19 Light on/off experiments and quantum yield measurement (ϕ = 0.0029) suggested that an extended radical chain mechanism is unlikely (Figures S4 andS7).

To elucidate the role of dppm in this reaction, we performed 19 F NMR titration experiments, which showed an upfield shift