Introduction

Vinylboronic esters are highly valuable organic intermediates and are intensively used in various transformations including C–C bond formations1,2,3, electrophilic or radical additions, and hydrogenation reactions4,5,6,7,8,9. Among these, the most prominent reaction is Pd-catalyzed Suzuki coupling, which could supply important substituted olefins with aryl, alkenyl, alkynyl, and alkyl halides2,7,10. Among the many ways to synthesize multi-substituted olefins (Fig. 1a)11,12,13,14, functional groups are needed to induce the vinyl group of boronates via precedent process or complicated conditions. Among them, the hydroboration of alkynes has gained much attention owing to efficiently access to arylvinylboronates via employing transition metal such as copper15,16, silver17, ruthenium18, etc. as catalyst (Fig. 1b-a). In addition, metal-photocatalyzed borylation reaction of vinyl halides has also been developed in recent years (Fig. 1b-b)19,20. Apparently, the direct modification of C–H on vinyl group is the most attractive way due to the efficiency. However, there is a big challenge for this strategy since the coupling reactions of aryl-electrophile with vinylic C–H bonds are normally catalyzed by Pd-catalyst (Heck-type reaction) (Fig. 1b-c)21,22, under which reaction conditions, boronate groups are generally not tolerant and thus such transformation is hardly realized23,24,25,26. Herein we report a wet base-promoted reaction of vinyl pinacolboronates and diaryliodonium salts (Ar1I+Ar2OTf) to afford the corresponding trans-arylvinylboronates with high yields and selectivity. Our new findings disclose the radical arylation of vinyl pinacolborate 2 can be realized with diaryliodonium salts (Ar1I+Ar2OTf) 1 promoted by wet base (such as carbonate, typically), so it is characterized by the simplicity and the possibility of further functionalization (Fig. 1c). Consequently, a new pathway for efficient employment of both aromatic moieties of (Ar1I+Ar2OTf) is realized, engaged in two types of C–C bond forming reactions in the iterative synthesis of olefins27,28,29.

Fig. 1: Arylvinylboronates and olefins.
figure 1

a Multi-substituted olefins. b Synthesis of trans-arylvinylboronates. c This work for a novel iterative synthesis of olefins.

Recently, diaryliodonium salts, Ar2I+X, have received considerable attention due to their powerful arylation for various nucleophiles to synthesize valuable aromatic compounds. A big challenge for these arylation reactions is how to efficiently use both aryl groups of the diaryliodonium salts since only one aryl group was utilized and the other one was deposited in most cases29. On the basis of our group’s previous research on diaryliodonium salts30,31,32,33,34, here we report a tandom process in which an aryl iodide generated in situ is captured in a second step by a Suzuki reaction, yielding aryl olefins. This atom-economical use of diaryliodonium salts may offer a useful approach to the iterative synthesis of aryl olefins using alkenyl boronic esters as intermediates (Fig. 1c). Of note, it will be a novel approach for the iterative synthesis of aryl olefins using alkenyl boronic esters as intermediate.

Results and discussion

Investigation of reaction conditions

To achieve this goal, we initially examined the reaction of di(4-tolyl) iodonium triflate 1a and pinacol vinylboronate 2a serving as model substrates. As shown in Table 1, no product or low yield was observed when the reaction was performed at 80 °C in DCE with CuCl as catalyst and 1.0 equivalents of DIPEA or potassium carbonates as base. The desired product 3a could be detected in 3% yield (determined by GC with n-dodecane as internal standard) in the presence of additive of tetra-butylammonium fluoride (TBAF). It was pleasingly found that the yield could be improved when the 40 equivalents of water was employed as the additive, affording 3a in 62% yield (in a mixture of Z− and E− isomers, Table 1, entry 4). As a comparison, Pd-catalyzed systems tended to give the Suzuki-type coupling product 4a (Table 1, entries 5 and 6). In addition, it was surprised to find that the reaction could also work without CuCl catalyst to give the only E-isomers in 64% yield (entry 7), indicating that the method was a novel approach for constructing C(sp2)–C(sp2) bonds. The element analysis showed that the content of transition metal was below 5 ppm. Subsequently, various solvents including DCM, PhMe, THF, CH3OH, DMF (entries 8-9, 12-14) were screened. As a result, DCM was the best choice, affording 3a in 81% yield (entry 8). The yield was increased when elevating the reaction temperature, eventually, affording the isolated yield in 88% at 100 °C in a sealed tube (entry 11). Further base screening of inorganic base such as K3PO4, NaHCO3, and Li2CO3 proceeded compatibly (Supplementary Table 1), and Li2CO3 also gave the isolated yield of 82%. Of note, treatment of the reaction using insoluble Ag2CO3 as base did not give the desired product 3a (entry 17), and the control experiment revealed that no reaction occurred in the absence of base (entry 15). Moreover, increasing or decreasing the equivalent of water significantly resulted in reduced yields. These results proved that water could dramatically influence the reaction. In addition, the counter anion of Ar1I+Ar2OTf 1 including OTs and OAc could give comparable yield of 3a (Supplementary Table 1).

Table 1 Optimization of the reaction condition for the arylation of vinylboronic esters.a

Substrate scope

Under the optimized condition, the scope of this novel procedure was sought to be investigated. Initially, functionalized diaryliodonium salts 1 with a broad range of substitutions were examined. As shown in Fig. 2, substitutions with diverse functional groups such as (o, m, p–) methyl (1a,1k,1m), halogen(1b1d,1j,1l), tert-butyl (1g), trifluoromethyl (1i), and methoxy (1h) were all well-tolerated, affording the corresponding products with good stereoselectivity in moderate to good yields. Notably, the substrates 1n and 1o, bearing a steric hindrance substituents, were also furnished in 82% (3n) and 55% (3o) yields. Interestingly, unsymmetrical diaryliodonium salts (ArI+MesOTf) containing 2-chloropyridine (1q), 2-naphthalene (1p), and 4-biphenyl(1r) species also gave corresponding products rather than product 3o, which might account for that steric effect was more obvious than electronic effect in terms of the two competing aryl motifs. Next, a series of substituted vinyl pinacolboronates were reacted with 1a. It was delighted to find that simple alkyl substitutions such as methyl(2t), ethyl(2u), propyl(2v) all went with moderate to good yields, while 2-phenyl substitution only afforded the product 3w in 27% isolated yield. It was worth mentioning that (Z)-formation of 1,2-substuited alky olefin 2x could be productive to exclusively give the product 3x in 41% yield, whereas no product was afforded while (E)-formation of olefins trans-2x served as substrates.

Fig. 2: The scope of diaryliodonium salts and pinacol vinylboronate.
figure 2

Reactions were performed with 1 (0.15 mmol), 2 (2.0 eq.), base (1.0 or 2.0 eq.), and H2O (40 eq.), in DCM (1 mL) at the temperature 100 °C.

After the successful arylation of C–H bond of vinyl pinacolboronates 2 was realized with (Ar1I+Ar2OTf) 1 promoted by wet base, we were keen to explore the further arylation of products 3 for efficient employment of both aromatic moieties of 1 via Suzuki reactions. Thus, a series of palladium catalyst, the temperature and phosphorus ligands were screened (Supplementary Table 2), and the best isolated yield and superior selectivity was obtained with Pd(OAc)2 as the catalyst, NaOH as base, PPh3 as ligand, of which 5a was given in 81% isolated yield (Fig. 3)2,35. Then, this one-pot protocol was extended to other substrates. As desired, symmetrical diaryliodonium salts with a variety of substituents (o,p-methyl, p-tert-butyl, p-trifluoromethyl) were all accomplished smoothly to give products 5a5d. Subsequently, various unsymmetrical substituted diaryliodonium salts and substituted alkenyl borate esters were examined. It was all productive for various diaryliodonium salts when either the methyl (5e), bromo substituents (5f) or a bulky aryl group such as 2,4,6-trimethyl (5h) or naphthalene (5i) moiety. In addition, hetero-aromatic rings including benzothiophene (5j) and pyridine species (5k, 5l) were well-tolerated to offer the respective products. Tri-substituted olefins were also afforded with comparable yields (5g, 5m, 5n), which were ubiquitous building blocks (vide infra).

Fig. 3: The synthesis of various di- or tri-substituted aryl olefins.
figure 3

a1.2 eq. of 2.

Mechanistic study

To investigate the mechanism of the arylation on C–H bond of 2 with 1, a few experiments were conducted. First, the effect of K2CO3 amount was investigated in this process. The control experiments (Supplementary Figs. 1 and 2) showed that there was a dramatic rate increase after 1 to 2 h when 1 equivalent of K2CO3 base was used and reached >60% yield; as a comparison, increasing the amounts of K2CO3 led to an evident rate decrease. Due to less solubility, 2 equivalent of Li2CO3 was needed. Above results indicated that the base amount was crucial to this procedure that might be essential for activating alkenyl borates 2 and accelerating dissociation of the OTf group of 136,37,38,39. To elucidate this transformation, 2,2,6,6-tetramethyl-1-piperidinyloxy was introduced to this base-promoted aryl migration process, and the adduct 2,2,6,6-tetramethyl-1-phenylpiperidine 6 was detected40,41, and it could be even obtained in higher yields in the absence of 2a (Fig. 4a). Moreover, the deficient of either base or water could not be capable of getting the desired product 3a and the trapping product 6. The above results were consistent with the EPR experiments (Fig. 4b and Supplementary Fig. 3), which indicated that the base-H2O system could release CO32− and split diaryliodonium salts into a pair of carbonate-stabilized radical 7a (vide infra) and aryl radicals 8a40,42,43,44.

Fig. 4: Preliminary mechanistic study.
figure 4

a Radical trapping experiments. b Radical detected experiments by EPR. c Proposed mechanism for arylation of vinylboronic esters.

On the basis of above results, a plausible mechanistic pathway was proposed in Fig. 4c. Ph2I+OTf 1e was triggered by CO32− to give I, which accelerated radicals 8e and PhI-CO3 formed via homo cleavage. The intermediate II formed from vinylboronate 2a and CO32− enabled the addition of radicals 8e to form the C–C bond45,46,47, generating the “ate” α-boronate adduct radical species III48,49,50,51,52. Species III was capable of occurring SET reaction with radical CO3 to give the intermediate moiety IV and intramolecular dehydrocarbonate give the “ate” intermediate VI, which further eliminated to give the desired product 3e.

To gain further insight of the base-promoted pathway, density functional theory (DFT) calculations were carried out to explore the reaction mechanism. In this system, the role of H2O in the reaction was discussed in Supplementary Figs. 4 and 5 and Supplementary Note 1. Besides, the anion exchange of diaryliodonium salts 3e from OTf to CO32− is easy to generate the intermediate I with quite an exothermic reaction energy release of 39.7 kcal/mol (Supplementary Fig. 6). The complex Ph2I+X (X = K2CO3, KCO3, CO32−, OTf) decompose to Ph-I+X radical and phenyl radical Ph endothermically. The corresponding Gibbs free energy required for the decomposition follows the order: K2CO3 > OTf > KCO3 > CO32− (Supplementary Figs. 7 and 8). The Gibbs free energies for the case X = K2CO3 and OTf are +90.7 and +32.6 kcal/mol, respectively, which are so high that the decomposition can hardly take place under the experimental condition. While on the other hand, the Gibbs free energy of +2.5 kcal/mol for the case X = CO32− coming from the ionization of K2CO3 by water is small enough for the subsequent homo cleavage, the fact that no reaction occurs without water addition to the system. The proposed mechanism suggests that it starts from the combination of vinyl pinacolboronates and carbonate to form the intermediate II, which is calculated to be exothermic by 7.4 kcal/mol. The phenyl radical attacks the =CH2 group of II to give the intermediate III, which is calculated to be exothermic by 12.4 kcal/mol (Supplementary Figs. 9 and 10).

The remaining calculated pathway for the reaction is shown in Fig. 5 and Supplementary Data 1, which shows that the intermediate III reacts with PhI-CO3 radical to produce the intermediate IV. Due to the weak interaction of PhI with CO3•− radical anion, PhI will directly leave the reaction system and CO3•− is bonding to intermediate III simultaneously. The next step of the reaction is the rate-determining one with a barrier of 25.5 kcal/mol overcoming the TS1, corresponding to proton abstraction by the CO3 group and giving the intermediate V. The O–H bond length 1.06 Å of the intermediate V is longer than that of bicarbonate anion (0.97 Å), indicating that the benzylic proton is not abstracted completely, as long as the basicity of carbonate is not large enough. Then, V could convert into VI through the TS2. In this process, with the CO3 group bonding to the α carbon left, the charge transfer occurs from the benzylic carbanion to the leaving CO3 group, leading to the formation of C=C bond by overcoming a very small barrier of 6.2 kcal/mol. In the TS2, the distance between the leaving CO3 group and α carbon is 2.43 Å, and the formed C=C bond length is 1.39 Å, which is approximately equal to that of C2H4 (1.33 Å). Subsequently the product 3e is obtained from VI via releasing bicarbonate, bearing successive barriers of 5.0 kcal/mol. The DFT pathway shows that the relative location Ha and Hb of intermediate IV can be attributed to the stereo-configuration of the final product, because the carbonate bonded to the boron atom abstracts the Ha atom, while the Hb atom remains. The Hb and Hc atoms are in the opposite direction along the C–C bond. Besides, we have also considered the situation that reaction starts from the binding of 2a and KCO3 (the detail in Supplementary Fig. 11), which shows a less preferred reaction mechanism comparing with that of CO32−.

Fig. 5: The calculated Gibbs free energy profile.
figure 5

The reaction pathway of the arylation of vinyl pinacolboronates.

The skeletons of aryl olefins widely occur in many biologically active compounds. The potential utility of this method was also assessed, as shown in Fig. 6, some illustrative cases were accomplished. The product 5l could be precisely transformed into [18F]AV-45, an effective PET agent for targeting Aβ plaques in human cerebrovascular, under the standard conditions of the Fig. 353. In addition, Chlorotrianisene 10 was furnished from compound 5n with 98% yield in one step of chlorination reaction54. Finally, the one-pot process of constructing tri-substituted olefins was applied to the synthesis of (Z)-tamoxifen precursor 5m with good selectivity, and then a series of downstream reactions were manipulated to afford the (Z)-tamoxifen in 68% yield55,56.

Fig. 6: Chemical derivatives.
figure 6

The synthetic efforts toward: a [18F] AV-45; b Chlorotrianisene; c tamoxifen.

In summary, we have developed an approach for selective arylation of C–H bond of vinyl pinacolboronates utilizing diaryliodonium salts and water-base as additive. This new strategy was exemplified of two-component arylation of diaryliodonium salts accessing aryl olefins via radical-type and Suzuki-type cross-coupling reaction in one-pot, which has been demonstrated as an iterative synthesis of multi-substituted olefins. The mechanistic experiments and DFT theoretical studies revealed the multi-function of carbonate and a novel radical-pair pathway of diaryliodonium salts promoted by wet carbonate.

Methods

General considerations

Unless specified, all substrates were obtained commercially from various chemical companies and their purity has been checked before use. Unless otherwise stated, all commercial reagents were used as received without purification. The synthesis of 3: mixture of diaryliodonium salt (0.15 mmol, 1.0 eq.) and base [condition a: K2CO3 (1 eq.) condition b: Li2CO3 (2 eq.)] was added into a schlenk tube and then evacuated and recharged with N2 for three times. After that, 1.0 ml DCM was added in, followed by vinyl pinacol boronic esters (0.30 mmol, 51 μl) and pure water (6.0 mmol, 100 μl). The tube and mixture were stirred at 100 °C for 12 h. After completion, the tube was cooled to room temperature, then NaCl aq. (10 ml) was added and the mixture was extracted with EtOAc (10 ml × 3), and then dried by anhydrous Na2SO4. The mixture was evaporated then purified on silica gel (petroleum ether/EtOAc = 50:1) provided the corresponding product. Full experimental details can be found in the Supplementary Methods. NMR spectra can be found in Supplementary Figs. 1250.