Dual gold and photoredox catalysis: visible light-mediated intermolecular atom transfer thiosulfonylation of alkenes

An unprecedented intermolecular atom transfer thiosulfonylation reaction of alkenes was achieved by combining Au catalysis and visible-light photoredox catalysis. A SCF3 group and other functionalized thio groups together with sulfonyl group were regioselectively introduced into alkenes.


Introduction
Regioselective difunctionalization of alkenes has the advantage of introducing diverse functional groups into vicinal carbons of common alkene moieties in a single operation and has attracted extensive attention from synthetic chemists. 1 Of the many catalytic processes that have been developed, a majority requires stoichiometric amounts of an external strong oxidant, such as PhI(OAc) 2 , Selectuor (eqn (1)), 2 and are typically initiated by a transition metal-catalyzed intramolecular addition. Intermolecular difunctionalization of alkenes is more challenging because of the regiochemical issue (eqn (2)). Recently, Stephenson developed an elegant visible light-mediated atom transfer radical addition reaction converting haloalkanes and a-halocarbonyl compounds into alkenes. 3 These neutral redox reactions are very attractive because they are atom-economical and require no additional oxidants. (1) The triuoromethylthio (SCF 3 ) group is a key structural unit in many pharmaceutical and agrochemical products such as tiorex, toltrazuril, and vaniliprole. 4 It is well known that SCF 3 groups in a molecule induce even higher lipophilicity than tri-uoromethyl substituents, and consequently, the incorporation of SCF 3 group into pharmaceuticals could greatly improve their ability to cross lipid membranes. 5 Because of this, the introduction of a SCF 3 group into small molecules has attracted signicant attention in synthetic chemistry. 6,7 Current methods for the construction of C-SCF 3 bonds involve electrophilic tri-uoromethylthio reagents 6 or the nucleophilic AgSCF 3 reagent. 7 Sulfonyl groups are similar to carboxyl or phosphate groups in terms of molecular size and charge distribution, and the sulfonyl group has been introduced into bioactive molecules to improve their activity. 8 A sulfonyl group has two receptors for hydrogen bonds and this can enhance the binding affinities of drug molecules with target proteins. Sulfones can be easily transformed into other functional groups, such as alkenes, via Julia olenation. 9 We investigated that whether both SCF 3 and sulfonyl groups could be simultaneously introduced into organic compounds in a single step. Moreover, this transformation has not been described to date. This report describes a dual gold and photoredox catalytic approach to the intermolecular atom transfer thiosulfonylation of alkenes.
A combination of visible light-mediated photoredox catalysis 10 and transition-metal-catalysis is possible to bring two distinctive catalytic systems together and achieve unprecedented new reactions. 11 Recently, the use of gold(I) complexes in photoredox catalysis has gained considerable attention. 12,13 This photoredox catalytic cycle triggers the conversion of Au(I) into Au(III) under mild conditions, a conversion which previously had only been achieved with stoichiometric quantities of strong oxidants. 14 In 2013, Glorius reported a rst dual gold and photoredox-catalyzed reaction, which achieved intramolecular oxy-and amino-arylation of alkenes with aryldiazonium salts (Scheme 1A). 12b Toste et al. took advantage of the visible lightmediated Au(I)/Au(III) cycle to produce arylative ring expansion reactions and carbon-phosphorus cross-coupling reactions (Scheme 1B). 12d,e This visible light-mediated single electron oxidative reaction has been utilized to access gold(III) complexes from gold(I) species. 12k Recently, Hashmi reported an aryldiazonium salts mediated Au(I) to Au(III) transformation upon irradiation with blue LED in the absence of a photosensitizer. 12q In all these reactions, 12 the same aryldiazonium salts were used. The development of new gold/photoredox catalysis mode is highly desirable. To achieve the proposed triuoromethyl-thiosulfonylation reaction in the most atom-economical manner, a difunctionalization reagent, such as PhSO 2 SCF 3 (2a), is required. This reagent can be easily prepared from PhSO 2 Na and AgSCF 3 (for details, see the ESI †). We envisioned that the reaction of PhSO 2 SCF 3 with a cationic gold catalyst in the presence of a photocatalyst would generate an LAuSCF 3 species and a benzenesulfonyl radical, which can then add to the alkene forming a new alkyl radical. 3a This radical may oxidize LAuSCF 3 into a Au(II) intermediate, which is further oxidized to a Au(III) derivative by the Ru III catalyst. Subsequent reductive elimination forms the target difunctionalized product, regenerating the Au(I) catalyst (Scheme 1C). This reaction provided a new approach to introduce triuoromethylthio (SCF 3 ) group at the a position of styrenes.

Results and discussion
To validate this concept, styrene (1a) and PhSO 2 SCF 3 (2a) were selected as model substrates to test the feasibility of this hypothesis. Aer a detailed optimization of reaction conditions, the proposed alkene triuoromethylthio-sulfonylation product (3a) was achieved in 94% yield under the standard conditions: a mixture of IPrAuCl (10 mol%), AgSbF 6 (15 mol%), Ru(bpy) 3 Cl 2 (2.5 mol%) in DCE (1 mL) was stirred under irradiation for 1-3 h with 100 W blue LED in a N 2 atmosphere (Table 1, entry 1). The dual catalytic nature of this reaction was investigated by control experiments, which conrmed that the gold catalyst, the Ru(II) photosensitizer, and visible light irradiation are all necessary for the reaction (Table S2, † entries 2-5). In the absence of silver salt, the reaction led to a dramatic decrease in the yields (NMR yields <5%), showing that the formation of a cationic gold species is highly important (Table S2, † entry 4). Other gold catalysts, such as AuCl, Ph 3 PAuCl, and Gagosz catalyst (Ph 3 PAuNTf 2 ), are all less effective than IPrAuCl (entries 2-5). A silver free system using IPrAuSbF 6 and Ru(bpy) 3 (PF6) 2 led to a slightly lower yield (88%, entry 6); thus, silver is not necessary in this reaction. Reactions in other solvents, such as acetonitrile and methanol, led to only traces of products and no solvent addition products, indicating that a carbocation mechanism is not involved (entries 7 and 8). Other iridium photocatalysts including, Ir[dF(CF 3 )ppy] 2 (dtbbpy)PF 6 and fac-Ir(ppy) 3 , were also tested and no product was observed in the reaction system (entries [10][11]. The attempt to lower the catalysts loading also led to a lower reaction yield (entry 12).
With these optimized conditions established, the scope of the dual gold and visible light-mediated alkene difunctionalization reaction was explored. As summarized in Table 2, a wide range of styrene-type alkenes were applicable and moderate to good isolated yields of the product were achieved (3a-3l). Substrates bearing different electron-withdrawing or electrondonating groups at different positions in the aromatic ring were all compatible with this reaction. A series of functional groups, such as ester, cyano, triuoromethyl, and halogen, were all welltolerated under the reaction conditions. In particular, the alkene 1m, containing an alkyne moiety, was also applicable to this reaction, giving the corresponding alkene difunctionalized product (3m) in moderate yield. Note that dienes, such as isoprene, were also amenable to this transformation, generating the 1,4-addition product (3n) as the major product in 55% yield, and its structure was conrmed by single crystal X-ray analysis.
Internal alkenes, including acyclic and cyclic alkenes, were also suitable substrates. Most of the reactions exhibited excellent diastereoselectivities. Interestingly, trans-alkenes and cisalkenes afforded the same major products (4o) in similar yields with similar diastereoselectivities, which is a very important feature and also an advantage of this reaction. The relative conguration of the seven-membered product 3s was unambiguously characterized by single X-ray crystallography. However, the reactions of aliphatic alkenes were unsuccessful under the same conditions.
Organosulfur compounds are ubiquitous in the pharmaceutical industry, materials science, and food chemistry. 15 The construction of C-S bonds is important but challenging 16 because a sulfur atom could coordinate with a metal catalyst, such as Au, leading to catalyst inactivation. When we applied the abovementioned dual catalysis system to the general thiosulfonylation reaction of alkenes using PhSO 2 SC 4 H 9 (À1.64 V, vs. SCE) as the reagent, which might be more challenging because of its lower oxidative potential compared to that of PhSO 2 SCF 3 (À1.11 V, vs. SCE, Fig. S5, ESI †), to our delight, the reactions were very successful (Table 3). Various alkenes, including styrenes, internal alkenes, and dienes, were all compatible with this thiosulfonylation reaction, giving the corresponding products in good yields (4a-4r). A large variety of alkyl and aromatic thio groups can be easily introduced into a styrene molecule, generating difunctional products generally in very good yields (5a-5h).
Table 2 Substrate scope of alkene trifluoromethylthiosulfonylation reactions a isolated yield and 8 in 32% yield (Scheme 2c). This benzenesulfonyl radical could also be generated by PhSO 2 Cl (11) in the presence of visible light, 3a and the reaction between 8, 11, and 2a afforded the target product (3a) in 98% yield (Scheme 2d). In contrast, without this gold intermediate, the direct atom transfer radical addition adduct 12 was obtained in 81% yield (Scheme 2e). This experiment demonstrated that IPrAuSCF 3 is the possible reaction intermediate. 17 Stern-Volmer uorescence quenching experiments were performed to gain an insight into the photoredox catalytic cycle (for details, see the ESI ( Fig. S1 and S2, ESI †)). The photoluminescence of Ru(bpy) 3 2+ was quenched by IPrAuSbF 6 with a rate constant of 8.85 Â 10 2 L mol À1 . In contrast, PhSO 2 SCF 3 (À1.11 V vs. SCE) and styrene cannot serve as emission quenchers. The cyclic voltammogram of IPrAuSbF 6 contains a reversible reduction peak at À0.11 V vs. SCE (Fig. S5, ESI †), indicating that this cationic gold catalyst is easily reduced by the excited state of the photocatalyst Ru(bpy) 3 2+ (E III/*II 1/2 ¼ À0.81 V vs. SCE). 10b To further characterize this electron transfer reaction, a ash-photolysis study was carried out (Fig. S3 †). Upon laser excitation by 355 nm light, the ground state absorption at $450 nm was obviously bleached and a characteristic absorption band at $360 nm was detected. This is ascribed to the reductive state of bipyridine in Ru(bpy) 3 (SbF 6 ) 2 ( Fig. S4-a, ESI †). 18 When PhSO 2 SCF 3 was introduced into a solution of Ru(bpy) 3 (SbF 6 ) 2 , the transient absorption spectra did not show any difference ( Fig. S4-b, ESI †). The lifetime of the excited state of Ru(bpy) 3 (SbF 6 ) 2 slightly decreased from 493 to 473 ns, as observed from the kinetics probed at 450 nm. However, when IPrAuSbF 6 was added to a solution of Ru(bpy) 3 (SbF 6 ) 2 , a new absorption peak appeared at $530 nm. This is characteristic of gold nanoparticles. At the same time, the lifetime of the excited state of Ru(bpy) 3 (SbF 6 ) 2 decreased from 493 to 394 ns ( Fig. S4-c, ESI †). All these results suggested that the electron transfer between IPrAuSbF 6 and the excited 3 MLCT state of Ru(bpy) 3 (SbF 6 ) 2 occurred, generating the active IPrAu(0) species that might aggregate to form gold nanoparticles. 19 The reductive IPrAu(0) catalyst formed in situ is highly reactive 20 and might react with PhSO 2 SCF 3 , providing IPrAuSCF 3 (8) and a benzenesulfonyl radical.
On the basis of the abovementioned results and previous reports, 12 a tentative proposed mechanism is shown in Scheme 3. Irradiation of Ru (bpy)  catalyst. This alkyl radical directly reacting with another equivalent of PhSO 2 SCF 3 could also afford the target product via a radical chain mechanism. In our reaction, when the internal alkenes were employed as reaction substrates, excellent diastereoselectivities were observed and both trans-and cisalkenes yielded the same diastereomer, which is not common in a pure radical reaction. The IPrAu(I)SCF 3 (8) approached the in situ formed alkyl radical from the less sterically hindered side, which might contribute to the excellent diastereoselectivity of this reaction. The proposed dual gold photoredox catalytic cycle may be the major pathway although the radical chain reaction couldn't be excluded. This gold catalytic cycle from Au (0) to Au(III) is similar to that reported for nickel participating in photoredox catalytic cycles.

Conclusions
In summary, we have reported an intermolecular atom transfer thiosulfonylation reaction of alkenes. Both triuoromethylthio group and other functionalized thio groups can be introduced into alkenes with excellent regioselectivity and diastereoselectivity. These reactions are promoted by a synergistic combination of gold catalysis and visible light photoredox catalysis. Detailed control experiments and reaction mechanism studies indicate that the gold catalyst experiences four different valencies from Au(0) to Au(III), which is unprecedented in previously reported Au-catalyzed transformations. This new dual gold and photocatalysis mode holds potential for applications to other important transformations.