Introduction

Transition metal-catalyzed coupling reactions of organic electrophiles with organometallic reagents has been well established to meet the demand for the construction of chiral centers in organic synthesis (Fig. 1a)1,2,3. On the other hand, the allyl-allyl cross-coupling reaction is of great importance since it affords synthetically versatile 1,5-diene products4. However, there is an issue of regioselectivity forming the linear and the branched products (Fig. 1b). Due to the steric effect, usually linear products are formed as observed by Trost5, Stille6, and their coworkers. Interestingly, the branched 1,5-alkadienes are very common in natural products and the two C=C bonds would also make such products a class of platform molecules for efficient syntheses of natural products and biologically active molecules, for example, (-)-preclamol7, (-)-protrifenbute8, (-)-paroxetine9, rottnestol10, herboxidiene11, and α-cuparenone12 (Fig. 1e)13,14,15. In 2010, the first asymmetric enantioselective coupling between allylic t-butyl carbonates and allylBpin, forming such branched products with a tertiary chiral center, was reported by Morken et al. with ee of 74–91%. The regioselectivity for 3-aryl substituted allylic t-butyl carbonates is >20:1 while that of 3-alkyl ones is ~10:116. Feringa et al. observed that in the Cu-catalyzed reaction of 2-alkenyl bromide with allyl Grignard reagent17 in 58–94% ee and 37:63–91:9 branched over linear selectivity as is observed also by Ohmiya/Sawamura in the Cu-catalyzed reaction of 2-alkenylic phosphonates with allylboronates18 and by Wang/Ding in the Pd-catalyzed reaction of 1-arylallylic acetates with allylboronates19. Only in the Ir-catalysis, a high selectivity for both enantio- and branched selectivity has been observed by Carreira and Yang independently in the Ir-catalyzed coupling of allylic alcohols with allyl silanes20 or boronates21. The allyl-allyl coupling for the enantioselective establishment of quaternary carbon stereocenter is even more challenging. The only report is also from Morken and his coworkers: the reaction of racemic tertiary 1,1-disubstituted allylic carbonates with allylboronates afforded the products with a quaternary chiral center in 52–92% ee, and due to the greatly increased steric hinderance, the coupling reactivity is lower, thus, resulting in β-elimination as the side reaction22. Since the boracurpation of allenes would provide allylic metallic reagents23,24,25,26,27,28,29,30,31,32, it would be interesting to develop the three-component reaction of differently substituted allenes, B2Pin2, and allylic electrophiles for such a purpose with an attractive diversity for organic synthesis. Hoveyda and coworkers reported such a three-component reaction of OTBS-substituted 1,2-pentadiene and OTBDPS-substituted 1,2-hexadiene with 3-substituted allylic phosphonates, and the reaction demonstrated a linear selectivity referring to allenes33, as is also observed by Xiong et al34. Fañanás-Mastral et al. observed the first branched selectivity referring to the reaction between mono-substituted allenes and allyl t-butyl carbonate, albeit with ~40% ee35. One solution could be to identify a finely tuned chiral catalyst with a much higher reactivity, avoiding β-elimination while dictating the branched vs linear selectivity (Fig. 1c).

Fig. 1: Catalytic asymmetric C-C and allyl-allyl couplings as well as selected examples of related natural products.
figure 1

a Transition metal-catalyzed coupling reactions for the construction of chiral centers. b Challenges in allyl-allyl cross-coupling reactions. c Strategy: ligand-controlled regio- and enantioselectivity of allyl-allyl coupling. d This work: Synergistic bimetallic catalysis for asymmetric allylboration of allenes. e Selected natural products containing chiral 1,5-diene unit and bioactive molecules.

Full size image

Here, we wish to disclose our recent observation on the synergistic Cu/Pd bimetallic catalyzed three-component highly regio- and enantioselective allylboration of monoaryl- as well as monoalkyl-substituted allenes with B2Pin2 and allylic phosphonates generating chiral 1,5-dien-2-yl boronates with tertiary chiral centers in excellent branched selectivity towards to allenes with an enantioselectivity of up to 97% ee (Fig. 1d). (-)-Protrifenbute and its two derivatives are synthesized efficiently by using this protocol as the key step. Furthermore, it is exciting to observe that such a strategy could also be applied to the construction of much more challenging all-carbon quaternary stereocenters as well when applying 1,1-disubstituted allenes. As is known, the construction of all-carbon quaternary stereocenters is still a challenge due to the requirement of steric bias of the four substituents.

Results

To avoid the issue of β-elimination, at the beginning, p-methoxyphenylpropadiene 1a, B2pin2, and allyl diethyl phosphate 2a were chosen as the model substrates to develop the asymmetric three-component reaction with C2-symmetric chiral ligand (Table 1). Firstly, we used CuCl and [Pd(allyl)Cl]2 as metallic catalyst precursors36,37,38,39,40,41,42,43,44,45,46,47 and tBuOLi as the base to screen typical chiral ligand skeletons in THF at room temperature: When mono-dentated phosphoramidite ligand (R)-L148, monophosphine ligand (S)-L2, biphenyl-based bisphosphine ligands (R)-L3, or Trost ligand (R,R)-L4 was applied, no branched product 3aa was generated (Table 1, entries 1–4); the linear coupling products (Z)-4aa and (E)-4aa as well as hydroboration products 5aa and 5aa’ were detected in presence of (R)-L1 or (S)-L2 as the ligand as expected (Table 1, entries 1 and 2). Interestingly, with BINAP (R)-L5 as the ligand, the targeted branched coupling product 3aa was afforded albeit in 5% NMR yield (Table 1, entry 5); with SYNPHOS (R)-L6 and BPE-PHOS (R,R)-L7, the desired product 3aa was obtained in 33% yield and -24% ee, 54% yield and 48% ee, respectively (Table 1, entries 6 and 7); interestingly, Quinox P* (R,R)-L8 improved the ee to 84%, albeit still merely in 28% yield (Table 1, entry 8). To our delight, when non-C2-symmetric phosphine-oxazoline chiral ligands L9-L16 were applied (Table 1, entries 9–16), (R)-L16 was identified as the optimal chiral ligand affording the product 3aa smoothly with 83% NMR yield and 89% ee (Table 1, entry 16).

Table 1 Selected representative results of ligand screening
Full size table

However, the branched/linear selectivity is still a problem (Table 1, entries 6-16). Thus, (R)-L16 was chosen as the ligand to optimize other reaction parameters (Table 2). Firstly, using [Pd(allyl)Cl]2 as the palladium catalyst precursor, several copper salts were screened (Table 2, entries 1–5) and CuTc was identified to be the best (Table 2, entry 4). Then with CuTc, PdCl2, Pd2(dba)3, and Pd(OAc)2 were tested, however, no better results were obtained (Table 2, entries 6–8). Since the base may play a critical role, different bases were applied (Table 2, entries 9–12): When MeOLi was used, the target compound 3aa was obtained with 74% yield and 92% ee. However, more hydroboration products 5aa and 5aa’ were formed (Table 2, entry 12). Subsequently, with MeOLi as the base, the solvent effect was explored and high yields and selectivity were achieved in tetrahydrofuran (Table 2, entries 12–16). When the reaction was conducted at 0 oC, the chemo-, regio-, and enantioselectivity were improved (Table 2, entry 17). Better results were obtained at −10 oC with the concentration of 0.5 M (Table 2, entry 18). A lower catalyst loading led to 47% NMR yield (Table 2, entry 19). The reaction at a concentration of 0.5 M was too viscous (entry 18), which may have a negative effect on the reaction yield. We reasoned that the trace amount of the water in the ambient environment may result in the formation of the hydroboration products 5aa and 5aa’, therefore, all the starting materials 1a, 2a and B2pin2 were dried over 5 Å molecular sieves before use and the branched product 3aa was obtained with 96% ee, 70% isolated yield, and a remarkable branched/linear selectivity of 95/5 at a concentration of 0.25 M, which has been defined as the optimized conditions (Table 2, entry 21).

Table 2 Optimization of reaction conditions
Full size table

The optimized reaction conditions (Table 2, entry 21) have then been applied to evaluate the scope of the allenes and allylic phosphates. This synergistic Cu/Pd enantioselective catalysis protocol tolerates a range of monosubstituted and 1,1-disubstituted allenes, generating a series of boron contained chiral 1,5-dienes 3 or 7 with moderate to good yields and excellent regio- and enantioselectivity (Fig. 2). The substituent location in the phenyl group has no obvious effect since both the o- or m-MeOC6H4 substituted allenes 1b and 1c gave corresponding products (S)-3ba and (S)-3ca with 94 and 96% ee, respectively. Even the o-, m-, and p-bromophenyl substituted allenes 1d, 1e, and 1f afforded (S)-3da, (S)-3ea, and (S)-3fa without touching the reactive and versatile C-Br bond. The gram-scale preparation of chiral 1,5-diene (S)-3ca could be achieved readily in 75% yield and 96% ee. A variety of synthetically useful functional groups, such as F, Cl, Me, CO2Me, and CN, were accommodated, affording (S)-3ha - (S)-3la successfully with ee ranging from 91 to 97%. As expected, when (S)-L16 was introduced instead of (R)-L16, the corresponding enantiomer (R)-3ia was produced on a 10 mmol scale with 97% ee and 77% isolated yield with only 1.2 equiv of 2a. Besides the monoaryl-substituted substrates, monoalkyl-substituted allenes also worked to afford (S)-3ma, (R)-3na, and (S)-3oa with moderate yields and a decent enantioselectivity of 84 to 86% ee. The reaction also tolerated heteroaryl groups, such as pyrrole (1p), indole (1q), and pyridine (1r), affording the corresponding products (S)-3pa, (S)-3qa, and (S)-3ra with 88% to 90% ees in 49 to 73% yields. Employing cyclohexyl substituted allene 1s as the starting material, which had the similar steric hindrance to the phenyl substituted allene 1g, target product (R)-3sa was obtained in 66% NMR yield with 83% ee, which was similar to those of other alkyl-substituted allenes [(R)-3na - (S)-3oa (84-86% ees)]. These results ruled out that the steric hindrance led to the difference for aryl- and alkyl-substituted allenes, and we reasoned that the electronic effect may be the key point.

Fig. 2: Substrate scope.
figure 2

a Conditions A: 1 (1.0 mmol, 1.0 equiv.), 2 (1.5 equiv.), B2Pin2 (1.5 equiv.), MeOLi (1.5 equiv.), CuTc (5 mol%)/(R)-L16 (6 mol%), [Pd(allyl)Cl]2 (2.5 mol%)/(R)-L16 (6 mol%), THF (4.0 mL, 0.25 M), −10 oC. b Conditions B: 1 or 6 (1.0 mmol, 1.0 equiv.), 2 (2.0 equiv.), B2Pin2 (2.0 equiv.), MeOLi (2.0 equiv.), CuTc (5 mol%)/(R)-L16 (6 mol%), [Pd(allyl)Cl]2 (2.5 mol%)/(R)-L16 (6 mol%), THF (4.0 mL, 0.25 M), −10 oC. c determined by HPLC analysis after oxidation of 3 or 7 using NaBO3•4H2O. d 1i (10.0 mmol, 1.0 equiv.), 2a (1.2 equiv.), (S)-L16 was used instead of (R)-L16. e MeOLi (4.0 equiv.). f MeOLi (3.0 equiv.). g [Pd(2-methylallyl)Cl]2 instead of [Pd(allyl)Cl]2. h determined by HPLC analysis after being converted to (R)-10sa, for details, see Supplementary Information. ND not detected.

Full size image

Furthermore, even all-carbon quaternary stereocenters may be constructed by applying the 1,1-disubstituted allenes forming 1,5-pentadiene 7 with excellent levels of regio- and enantioselectivity: the reaction of 3-phenylbuta-1,2-diene 6a afforded the desired branched product (R)-7aa with 72% isolated yield and 94% ee. It worth to mention that the linear products and hydroboration products were not detected, and only a trace amount of β-H elimination product (<4% NMR yield, if any) was detected by 1H NMR analysis of crude products (see Supplementary Information). Synthetically versatile F, Cl, Br, Me, and methoxy group were all tolerated on the phenyl ring of the 1,1-disubstituted allenes affording (R)-7ba - (R)-7ha with moderate yields and excellent ee values (Fig. 2). The gram-scale synthesis of (R)-7ea has been achieved easily (1.55 g, 82% isolated yield and 94% ee). Furthermore, the absolute configuration of the all-carbon quaternary stereocenters was determined by the X-ray single crystal diffraction analysis of (R)-7ea. Methyl or methoxy-substituent within the phenyl ring also worked, requiring a higher loading of MeOLi to ensure a high yield (Fig. 2, (R)-7fa and (R)-7ga). Besides, 3,4-diphenylbuta-1,2-diene 6i may also afford the target product (S)-7ia with 71% yield and 94% ee. The methyl group may be replaced with Et, nPr, nBu, nPent, or nHex group (6a, 6j-n), affording (R)-7ja, (R)-7ka, (R)-7la, (R)-7ma, or (R)-7na with an excellent ee. However, when 1,1-diaryl-substituted allene 6o was employed, the corresponding product (R)-7oa was not detected. Diethyl 2-methylallyl phosphate 2b and diethyl 2-phenylallyl phosphate 2c may also be applied to afford the desired products (R)-7ab and (R)-7ac in 93% and 97% ee, respectively.

The synthetic potentials of these chiral 1,5-pentadienes have been demonstrated (Fig. 3): the gram-scale reactions of 1c and 6e with allyl diethyl phosphate 2a under the standard conditions afforded (S)-3ca and (R)-7ea smoothly, which readily underwent Suzuki coupling49 selectively with the C-I bond in 1,3-dibromo-5-iodobenzene and the C-Br bond in methyl 4-bromobenzoate affording the arylation products (S)-10ca and (S)-11ea, respectively; oxidation of the alkenyl boronate unit in (R)-7ea with NaBO3•4H2O results in the synthesis of α-chiral homoallyl ketone (S)-12ea with 91% yield and 94% ee; oxidation of (S)-3ca followed by condensation with (2,4-dinitrophenyl)hydrazine50 led to the formation of optically active solid product (R,E)-13ca in 73% yield and 95% ee, and its X-ray single crystal diffraction analysis further confirmed the stereoselectivity of this reaction. Catalytic cross metathesis reaction of (R)-7ea occurred highly selectively with mono-substituted terminal C=C bond, furnishing functionalized chiral 2,5-dienoate (R,E)-14ea with 87% isolated yield and 93% ee51; the reaction of (S)-3ca with nBuLi followed by the addition of ICH2Cl and hydrolysis afforded 1,5-dienol (S)-15ca in 52% yield and 96% ee52.

Fig. 3: Gram-scale synthesis and transformations of (S)-3ca and (R)-7ea.
figure 3

Conditions A and B: see Fig. 2. a determined by HPLC analysis after oxidation using NaBO3•4H2O. C: (S)-3ca (1.0 equiv.), 1,3-dibromo-5-iodobenzene (3.0 equiv.), Pd(PPh3)4 (10 mol%), Cs2CO3 (3.0 equiv.), THF (2.0 mL), 80 oC, 24 h. D: (R)-7ea (1.0 equiv.), methyl 4-bromobenzoate (3.0 equiv.), Pd(PPh3)4 (10 mol%), Cs2CO3 (3.0 equiv.), THF (2.0 mL), 80 oC, 59 h. E: (R)-7ea (1.0 equiv.), NaBO3•4H2O (5.0 equiv.), THF (5.0 mL), H2O (5.0 mL), r.t., 5 h. F: (R)-7ea (1.0 equiv.), methyl acrylate (3.0 equiv.), second generation Grubbs catalyst (5 mol%), CuI (7.5 mol%), Et2O (2.0 mL), 35 oC, 7 h. G: (S)-3ca (1.0 equiv.), NaBO3•4H2O (5.0 equiv.), THF (5.0 mL), H2O (5.0 mL), r.t., 3 h, then (2,4-dinitrophenyl)hydrazine (1.0 equiv.), BF3•Et2O (1.0 equiv.), MeOH (2.0 mL), −78 oC-r.t., 6 h. b after recrystallization in DCM/PE. H: (S)-3ca (1.0 equiv.), ICH2Cl (1.5 equiv.), nBuLi (1.1 equiv.), THF (1.0 mL), −78  oC-r.t., 16.5 h, then, NaOH (2.5 M, aq., 0.5 mL), H2O2 (30%, 0.5 mL), 0 oC, 12 h.

Full size image

To further demonstrate the synthetic potential of this synergistic Cu/Pd catalysis strategy, the synthesis of (-)-protrifenbute8 has been executed (Fig. 4). Firstly, (R)-3ia was prepared smoothly by following this protocol (for details, see Fig. 2); upon selective catalytic olefin cross metathesis reaction with 4-fluoro-3-phenoxystryene, chiral 1,5-diene (R,E)-16ia was obtained51. Subsequently, (R)-17ia was produced with a 60% of isolated yield and 96% ee after selective hydrogenation using Pd/C and (MeO)2MeSiH in EtOH. The C-B bond in (R)-17ia was transformed to the C-H bond upon its treatment with AgF in the mixture of THF/MeOH/H2O, affording (R)-18ia53. The synthesis of (-)-protrifenbute (R)-21ia has been realized upon cyclopropanation20. Due to the versatile nature of the C-B bond in (R)-17ia, the derivatives of protrifenbute, (R)-22ia and (S)-23ia, were prepared with decent yields and excellent ee after coupling49,54 and cyclopropanation.

Fig. 4: Synthesis of (-)-protrifenbute and its derivatives.
figure 4

Conditions A: see Fig. 2. I: (R,E)-16ia (1.0 equiv.), 5% Pd/C (5 mol%), (MeO)2MeSiH (3.0 equiv.), EtOH, air, r.t., 4.5 h. J: (R)-17ia (1.0 equiv.), AgF (2.1 equiv.), THF/MeOH/H2O (10/9/1), Ar, 50 oC, 6 h. K: (R)-17ia (1.0 equiv.), (1) MeLi (1.1 equiv.), Et2O, Ar, −78–0 oC, 1.5 h; (2) I2 (1.0 equiv.), MeOH, Ar, −78 oC- r.t., 3 h; (3) NaOH (3M, aq), 0.5 h. L: (R)-17ia (1.0 equiv.), methyl 4-bromobenzoate (3.0 equiv.), Pd(PPh3)4 (10 mol%), Cs2CO3 (3.0 equiv.), THF (0.1 M), 80 oC, 50 h. M: (R)-18ia, (R)-19ia, or (S)-20ia (1.0 equiv.), Et2O (5.0 equiv.), CF3COOH (5.0 equiv.), CH2I2 (5.0 equiv.), DCM (0.1 M), 0 oC-r.t., 3 h.

Full size image

To elucidate the mechanism, several control experiments have been performed (Table 3): no products were detected, and the starting material allene 1a was recovered in 92% yield in the absence of CuTc/(R)-L16 (Table 3, compare entry 2 with entry 1). When only chiral copper catalyst was used, the reaction yielded racemic product 3aa in 22% yield, with 1a being recovered with 39% (Table 3, entry 3). In the absence of allyl phosphonate 2a, the reaction with [Pd(allyl)Cl]2 (0.75 equiv.)/(R)-L16 (1.5 equiv.) also afforded the target compound 3aa with 34% yield and 95% ee, indicating that the in situ formed allyl palladium species must have provided the allyl group (Table 3, entry 4). We also carried out the reactions with the different ligands for copper and palladium. When CuTc/(R)-iPr-PHOX in combination with Pd(PPh3)4 were employed as the bimetallic catalytic system, the target product (S)-3aa was obtained in only 3% NMR yield (Table 3, entry 5). When the combination was modified to CuTc/(R)-iPr-PHOX and [Pd(allyl)Cl]2/DPPF, (S)-3aa was obtained in 21% NMR yield with 81% ee (Table 3, entry 6). When PPh3 was employed as the ligand for CuTc, (S)-3aa was not detected (Table 3, entry 7). When DPPF was applied for CuTc, (S)-3aa was obtained in 37% NMR yield with 75% ee (Table 3, entry 8). Thus, the optimal chiral ligand (R)-iPr-PHOX was necessary for both copper and palladium catalysts. In order to further demonstrate the nature of allyl metallic intermediate in this synergistic Cu/Pd catalytic cycles, isotopic labeling experiments were designed and conducted (Fig. 5a): with CuTc/(R)-L16 as the catalyst precursors, only racemic d2-3oa was produced with 21% isolated yield, which is consistent with the above results (Fig. 5a, entry 1 and Table 3, entry 3), and the absence of d2-3oa’ as judged by 1H NMR analysis of isolated products indicated that copper catalytic cycle must have proceeded through SN2-type process with d2-2a to produce the racemic product. However, when synergistic Cu/Pd catalysts were used, the desired products d2-3oa and d2-3oa’ were obtained in 1:1 ratio and a combined isolated yield of 82% and enantioselectivity of 86% ee, which provides clear evidence for the involvement of ƞ3-allyl palladium species in the catalytic cycles and reveals the necessity of the Pd catalysis for this asymmetric allylboration reaction.

Table 3 Control experiments
Full size table
Fig. 5: Isotopic labeling experiments and proposed reaction mechanism.
figure 5

a Isotopic labeling experiments. Yields and the recoveries were determined by 1H NMR analysis using CH3NO2 as the internal standard. a determined by HPLC analysis after oxidation of d2-3oa/3oa’ using NaBO3•4H2O. b The reaction was conducted on a 0.5 mmol scale at 0.25 M. c isolated yield. d determined by 1H NMR analysis of isolated products. ND not detected. b Proposed reaction mechanism. c Proposed rationale for the regio- and enantioselectivity.

Full size image

Based on control experiments, catalytic cycles of synergistic Cu/Pd regio- and enantioselective catalysis have been proposed (Fig. 5b): Firstly, chiral Bpin-Cu[(R)-L16] species B was generated via the transmetalation between B2pin2 and (R)-L16 coordinated copper complex A. The ƞ3-allylic copper species C was generated via the borylcupration of allenes, which may react with the in situ generated chiral π-allylpalladium species E via the oxidative addition of allyl diethyl phosphate d2-2a with Pd0[(R)-L16] complex to afford bis(allyl) palladium complex F. Due to the nature of the Pd2+ and the 18 e electron rule, it is believed that most likely the ƞ3-allylic unit was transferred from the Cu atom to the Pd atom to afford ƞ1-allylic species F1 and F255,56,57. Finally, target products d2-3 or d2-7 with deuterium atom exclusively at either end of the three carbon atoms from the deuterated allyl phosphonate d2-2a were obtained upon reductive elimination of these two palladium intermediates F1 and F2 due to steric effect for pushing more sterically hindered terminal away from this non-C2-symmetric P, N-ligand (R)-L16, resulting in an excellent regio- and enantioselectivity, which may be rationalized by the bigger steric hindrance in TS_(R)-3aa and TS_4aa, thus, favoring the formation of (S)-3aa by the reductive elimination via TS_(S)-3aa (Fig. 5c).

In summary, synergistic Cu/Pd catalyzed three-component branched regioselective asymmetric allylboration of allenes with B2pin2 and allylic phosphates has been developed, affording a series of chiral 1,5-dienes containing a sp2 C-B bond with moderate to good yields. Both mono- or 1,1-disubstituted terminal allenes substituted with functional groups are ideal partners. Due to the versatile nature of the products, a set of chiral molecules may be formed efficiently. Synthesis of (-)-protrifenbute and the preparation of its derivatives have been realized upon using this protocol as the key step. Control experiments reveal that the synergistic Cu/Pd catalysis with non-C2-symmetric phosphine-oxazoline chiral ligand L16 is responsible for the regulation of the regio- and enantioselectivity in the current bifunctionalization reaction of allenes.

Methods

General procedure for the synthesis of allylboration products 3 or 7

Preparation of the Cu pre-catalyst in a nitrogen-filled glovebox: to an oven-dried 4 mL vial were added a stirring bar, CuTc (9.5 mg, 0.05 mmol), (R)-L16 (22.4 mg, 0.06 mmol), and THF (1.0 mL). The resulting mixture was stirred at room temperature for 10 min to afford the Cu pre-catalyst, which was transferred via a 1.0 mL syringe.

Preparation of the Pd pre-catalyst in a nitrogen-filled glovebox: to an oven-dried 4 mL vial were added a stirring bar, [Pd(allyl)Cl]2 (9.2 mg, 0.025 mmol), (R)-L16 (22.4 mg, 0.06 mmol), and THF (1.0 mL). The resulting mixture was stirred at room temperature for 10 min to afford the Pd pre-catalyst, which was transferred via a 1.0 mL syringe.

In a nitrogen-filled glovebox, to an oven-dried 25 mL tube with screw cap were added a stirring bar, B2Pin2, MeOLi, allene 1 or 6, allylic phosphates 2, and THF (2.0 mL). The tube with a screw cap and Cu/Pd pre-catalysts in syringes were then taken out of the glovebox together. The sealed tube was connected to the Schlenk line and degassed/refilled with argon three times. Then, the sealed tube was immersed in the EtOH bath controlled at −10 oC. Under argon atmosphere, the screw cap was removed, and the Pd pre-catalyst and Cu pre-catalyst were added by syringes sequentially and the tube was screwed tightly. The resulting mixture was vigorously stirred at −10 oC as monitored by TLC and filtered over a plug of silica gel eluted with ethyl acetate (50 mL). The crude product was purified by column chromatography on silica gel to afford the product 3 or 7.