Cross-Coupling Reaction of Allylic Ethers with Aryl Grignard Reagents Catalyzed by a Nickel Pincer Complex

A cross-coupling reaction of allylic aryl ethers with arylmagnesium reagents was investigated using β-aminoketonato- and β-diketiminato-based pincer-type nickel(II) complexes as catalysts. An β-aminoketonato nickel(II) complex bearing a diphenylphosphino group as a third donor effectively catalyzed the reaction to afford the target cross-coupled products, allylbenzene derivatives, in high yield. The regioselective reaction of a variety of substituted cinnamyl ethers proceeded to give the corresponding linear products. In contrast, α- and γ-alkyl substituted allylic ethers afforded a mixture of the linear and branched products. These results indicated that the coupling reaction proceeded via a π-allyl nickel intermediate.


Introduction
Transition metal-catalyzed cross-coupling reactions are efficient and widely utilized synthetic protocols for constructing carbon-carbon bonds [1]. The reactions of allylic electrophiles with aryl metal reagents provide promising methodologies for the formation of C(sp 3 )-C(sp 2 ) bonds. Allylic electrophiles with efficient leaving groups, such as allylic halides, acetates, and phosphates, are commonly used in these reactions [2][3][4][5][6]. In contrast to these allylic electrophiles, allylic ethers have been considered less-reactive electrophiles in cross-coupling reactions, whereas these electrophiles are easily accessible, cheap, and easy to handle. Therefore, the development of effective metal catalysts for coupling reactions utilizing allylic ethers as electrophiles in a highly regioselective manner has attracted considerable attention. The choice of the metal catalyst is critical for linear and branched selectivity. Although precious metals such as Pd and Ir have been extensively explored in selective allylic substitution reactions [2][3][4], recent efforts have focused on first-row late transition-metals such as Cu, Co, Fe, and Ni, whose complexes act as highly active catalysts. For example, Feringa reported the asymmetric allylic alkylation of acyclic allylic ethers with organolithium reagents using a Cu/phosphoramidite system [7]. Oshima and Yorimitsu reported a cobalt-catalyzed cross-coupling reaction of allylic ethers with Grignard reagents in a linear manner [8]. Iron catalysts also enable the allylic arylation reaction to be performed [9][10][11][12]. Recently, Ni pincer complexes have been reported as effective catalysts for the coupling reaction of allylic ethers with aryl zinc reagents [13].
In recent years, tridentate pincer-type complexes have generated significant interest because the pincer-type ligand stabilizes the metal complex and its properties can be tuned to achieve optimal reactivity [14][15][16][17][18][19][20][21][22][23][24][25]. We have recently reported the synthesis of a range of pincer-type nickel(II) complexes (1a-d) utilizing a combination of β-aminoketonato or β-diketiminato frameworks with a third donor such as an amino or phosphino group [26,27]. Our systematic study on these nickel(II) complexes revealed that modifying the ligand framework has a significant influence on the catalytic performance in the cross-coupling reaction of aryl halides with arylmagnesium reagents. In particular, nickel(II) complexes 1b and 1d bearing a diphenylphosphino group as the third donor enable the utilization of aryl fluorides as electrophiles in the reaction [27]. In order to elucidate the reactivity of the nickel(II) complex toward the relatively robust C-O bond in an allylic aryl ether, we herein describe the cross-coupling reaction of allylic ethers with arylmagnesium reagents using nickel(II) complexes (1a-d) (Scheme 1).
Molecules 2019, 24 FOR PEER REVIEW 2 nickel(II) complexes (1a-d) utilizing a combination of β-aminoketonato or β-diketiminato frameworks with a third donor such as an amino or phosphino group [26,27]. Our systematic study on these nickel(II) complexes revealed that modifying the ligand framework has a significant influence on the catalytic performance in the cross-coupling reaction of aryl halides with arylmagnesium reagents. In particular, nickel(II) complexes 1b and 1d bearing a diphenylphosphino group as the third donor enable the utilization of aryl fluorides as electrophiles in the reaction [27]. In order to elucidate the reactivity of the nickel(II) complex toward the relatively robust C-O bond in an allylic aryl ether, we herein describe the cross-coupling reaction of allylic ethers with arylmagnesium reagents using nickel(II) complexes (1a-d) (Scheme 1).

Initial Optimization of the Nickel(II) Complexes
To evaluate the catalytic activity of nickel(II) complexes 1a-d, we carried out the reaction between cinnamyl phenyl ether 2a and 3 equivalents of p-tolylmagnesium bromide 3a using nickel(II) complexes 1a-d (2.5 mol%) in THF at room temperature (Table 1), referring to our previously study on the biaryl coupling reaction of aryl chlorides with arylmagnesium reagents [26]. Complex 1a gave the linear product 4a in only 7% yield. Complex 1b showed high catalytic activity to give 4a in 86% yield. Nickel(II) complexes 1c and 1d bearing N,N,N-and N,N,P-type ligands were not effective in the reaction and afforded the target product in 2% and 4% yield, respectively. In these cases, a large amount of 2a remained unreacted (see 2.3 Reaction mechanism section). In all cases, the branched product 5a was not detected and 4,4′-bitolyl (6a) derived from 3a was formed as a by-product (9-30%). In the absence of complex 1a-d, the reaction did not give any coupled products. Scheme 1. The coupling reaction of allylic ethers with aryl Grignard reagents using Ni(II) pincer complexes (1a-d).

Initial Optimization of the Nickel(II) Complexes
To evaluate the catalytic activity of nickel(II) complexes 1a-d, we carried out the reaction between cinnamyl phenyl ether 2a and 3 equivalents of p-tolylmagnesium bromide 3a using nickel(II) complexes 1a-d (2.5 mol%) in THF at room temperature (Table 1), referring to our previously study on the biaryl coupling reaction of aryl chlorides with arylmagnesium reagents [26]. Complex 1a gave the linear product 4a in only 7% yield. Complex 1b showed high catalytic activity to give 4a in 86% yield. Nickel(II) complexes 1c and 1d bearing N,N,Nand N,N,P-type ligands were not  We next examined the solvent effect using the reaction between 2a and 3 equivalents of 3a with complex 1b. The results are summarized in Table 2. The choice of the solvent used is essential to produce product 4a in satisfactory yield. When cyclopentyl methyl ether (CPME) was employed as solvent, the coupled product 4a was obtained in 89% yield (entry 2). The yield of 4a decreased in  We next examined the solvent effect using the reaction between 2a and 3 equivalents of 3a with complex 1b. The results are summarized in Table 2. The choice of the solvent used is essential to produce product 4a in satisfactory yield. When cyclopentyl methyl ether (CPME) was employed as solvent, the coupled product 4a was obtained in 89% yield (entry 2). The yield of 4a decreased in  We next examined the solvent effect using the reaction between 2a and 3 equivalents of 3a with complex 1b. The results are summarized in Table 2. The choice of the solvent used is essential to produce product 4a in satisfactory yield. When cyclopentyl methyl ether (CPME) was employed as solvent, the coupled product 4a was obtained in 89% yield (entry 2). The yield of 4a decreased in  We next examined the solvent effect using the reaction between 2a and 3 equivalents of 3a with complex 1b. The results are summarized in Table 2. The choice of the solvent used is essential to produce product 4a in satisfactory yield. When cyclopentyl methyl ether (CPME) was employed as solvent, the coupled product 4a was obtained in 89% yield (entry 2). The yield of 4a decreased in  We next examined the solvent effect using the reaction between 2a and 3 equivalents of 3a with complex 1b. The results are summarized in Table 2. The choice of the solvent used is essential to produce product 4a in satisfactory yield. When cyclopentyl methyl ether (CPME) was employed as solvent, the coupled product 4a was obtained in 89% yield (entry 2). The yield of 4a decreased in 1d 4 90 a The reaction was carried out using 2a (0.5 mmol) and 3a (1.5 mmol) in the presence of 1 (0.0125 mmol, 2.5 mol%) in THF (5 mL) for 24 h at room temperature. b The yield of 4a and recovery of 2a were determined by 1 H NMR analysis using pyrazine as the internal standard. effective in the reaction and afforded the target product in 2% and 4% yield, respectively. In these cases, a large amount of 2a remained unreacted (see 2.3 Reaction mechanism section). In all cases, the branched product 5a was not detected and 4,4 -bitolyl (6a) derived from 3a was formed as a by-product (9-30%). In the absence of complex 1a-d, the reaction did not give any coupled products.

Investigation of the Effect of the Solvent and Nickel(II) Complexes
We next examined the solvent effect using the reaction between 2a and 3 equivalents of 3a with complex 1b. The results are summarized in Table 2. The choice of the solvent used is essential to produce product 4a in satisfactory yield. When cyclopentyl methyl ether (CPME) was employed as solvent, the coupled product 4a was obtained in 89% yield (entry 2). The yield of 4a decreased in 1,4-dioxane (71%, entry 3). t-Butyl methyl ether (MTBE) and toluene were also effective solvents for the reaction and produced 4a in 82% and 83%, respectively (entries 4 and 5). CH 2 Cl 2 and N,N-dimethylformamide (DMF) were not suitable for the reaction (entries 6 and 7). Et 2 O provided the best result, affording product 4a in 99% yield, from which 4a was isolated in 94% yield (entry 8). The other nickel(II) complexes 1a, 1c, and 1d exhibited relatively low catalytic activity, even in Et 2 O (entries 9-11). When the amount of Grignard reagent 3a used was reduced to 1.5 equivalents, complex 1b acted as an effective catalyst to give 4a in an isolated yield of 91% (entry 12).  In order to investigate the influence of the leaving group (LG) on the allylic substrate, we carried out the reaction between various cinnamyl electrophiles (2a-d, 7, and 8) with 1.5 equivalents of 3a in the presence of 1b (2.5 mol%) in Et2O (Table 3). Cinnamyl phenyl ether 2a was converted into 4a in 91% isolated yield. In the case of cinnamyl p-methoxyphenyl ether 2b, the cross-coupled product 4a was isolated in excellent yield (96%). When the amount of nickel(II) complex 1b used was reduced to 1 mol%, 2a was converted into 4a in 60% yield, whereas in the case of 2b, product 4a was isolated in 95% yield. The reaction of cinnamyl ether 2c bearing a p-trifluoromethylphenyl group in the leaving group afforded 4a in 62% yield. p-Methoxyphenoxide, which has an electron-donating group on the aromatic ring, acted as an effective leaving group. Methyl ether 2d was also applicable in the reaction using 2.5 mol% of 1b to give product 4a in 90% yield. When 1 mol% of 1b was employed, the product was obtained in 73% yield. Cinnamyl alcohol 7 was also able to be used as an electrophile, however, the yield of 4a was low (40%) and 7 was recovered in 54%. In the case of cinnamyl chloride 8, compound 4a was formed in low yield (38%). In order to investigate the influence of the leaving group (LG) on the allylic substrate, we carried out the reaction between various cinnamyl electrophiles (2a-d, 7, and 8) with 1.5 equivalents of 3a in the presence of 1b (2.5 mol%) in Et 2 O (Table 3). Cinnamyl phenyl ether 2a was converted into 4a in 91% isolated yield. In the case of cinnamyl p-methoxyphenyl ether 2b, the cross-coupled product 4a was isolated in excellent yield (96%). When the amount of nickel(II) complex 1b used was reduced to 1 mol%, 2a was converted into 4a in 60% yield, whereas in the case of 2b, product 4a was isolated in 95% yield. The reaction of cinnamyl ether 2c bearing a p-trifluoromethylphenyl group in the leaving group afforded 4a in 62% yield. p-Methoxyphenoxide, which has an electron-donating group on the aromatic ring, acted as an effective leaving group. Methyl ether 2d was also applicable in the reaction using 2.5 mol% of 1b to give product 4a in 90% yield. When 1 mol% of 1b was employed, the product was obtained in 73% yield. Cinnamyl alcohol 7 was also able to be used as an electrophile, however, the yield of 4a was low (40%) and 7 was recovered in 54%. In the case of cinnamyl chloride 8, compound 4a was formed in low yield (38%).

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b' (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected

Substrate Scope
With the optimized conditions in hand, we next examined the substrate scope of the reaction using allylic ethers 2 bearing a p-methoxyphenoxide group as the leaving group and a range of arylmagnesium reagents (3) in the presence of 1 mol% of 1b as the catalyst ( Table 4). The reaction of cinnamyl ether 2b with phenylmagnesium bromide 3b gave the coupled product 4b in excellent yield (94%) with linear selectivity. In the case of 4-methoxyphenylmagnesium bromide 3c, which has an electron-donating group on the aromatic ring, the coupled product 4c was isolated in 94% yield. The reaction with 4-N,N-dimethylaminophenylmagnesium bromide 3d gave product 4d in 95% yield. We recently reported that complexes 1b and 1d act as an effective catalyst for the cross-coupling reaction of aryl fluorides with arylmagnesium reagents [27]. In the reaction with 4-fluorophenylmagnesium bromide 3e, product 4e was obtained in 96% yield. In the present reaction conditions, the reaction of 4e as an electrophile with Grignard reagent 3e did not occur. In cases of sterically congested arylmagnesium reagents, such as 1-naphthylmagnesium bromide 3f, o-tolylmagnesium bromide 3g and 2,4,6-trimethylphenylmagnesium bromide 3h, the corresponding products 4f, 4g, and 4h were obtained in 85%, 79%, and 80%, respectively, in the presence of 2.5 mol% of 1b under refluxing conditions. The reaction of p-substituted cinnamyl ether 2i, which has an electron-donating group (OMe) on the aromatic ring, with p-tolylmagnesium bromide 3a gave product 4i in 78% yield. In the case of electrophile 2j bearing an electron-withdrawing group (F), the coupled product 4j was obtained in 89% yield. In the coupling reaction, alkyl-and benzylmagnesium reagents were not suitable as nucleophiles. In addition, the reaction using a cis/trans mixture of 2b (cis:trans = 86:14) with Grignard reagent 3a afforded the trans product (4a) in 88% yield as a sole product ( Table 5, entry 1). In contrast to cinnamyl ethers, γ-alkyl substituted allylic ether 2g afforded a mixture of the linear and branched isomers (4k and 5k) in 96% yield. The ratio of 4k:5k was determined to be 54:46 using 1 H NMR spectroscopy (entry 2). These results indicate that the reaction proceeds via a π-allyl nickel intermediate (vide infra). Therefore, we examined the reaction using secondary allylic ethers as electrophiles. The reaction of an allylic ether bearing a phenyl group at the α-position (2h) proceeded with linear selectivity to afford product 4a as a single regioisomer in high yield (94%). In this reaction, branched isomer 5a was not detected (entry 3). In contrast, an allylic ether bearing an alkyl group at the α-position (2i) was converted into a mixture of regioisomers (4k and 5k) in 93% yield. The ratio of 4k:5k was determined to be 57:43 using 1 H NMR spectroscopy (entry 4). The reaction of a mixture of 2j and 2k (2j:2k = 67:33) gave the linear product 4l in 90% yield with high regioselectivity. The branched isomer (5l) was not detected (entry 5). The regioselectivity in this reaction probably depends on both maximizing the resonance stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product. Table 4. The cross-coupling reaction of allylic ethers 2 with arylmagnesium reagents 3. a linear product 4l in 90% yield with high regioselectivity. The branched isomer (5l) was not detected (entry 5). The regioselectivity in this reaction probably depends on both maximizing the resonance stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product. linear product 4l in 90% yield with high regioselectivity. The branched isomer (5l) was not detected (entry 5). The regioselectivity in this reaction probably depends on both maximizing the resonance stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product. stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product. linear product 4l in 90% yield with high regioselectivity. The branched isomer (5l) was not detected (entry 5). The regioselectivity in this reaction probably depends on both maximizing the resonance stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product. stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product. substituted allylic ether, the steric influence should operate predominantly for the formation of the product. stabilization energy on the coupling product and minimizing steric hindrance in the coupling of the Grignard reagent with the putative π-allyl nickel species [13]. In the reaction of the γ-alkyl substituted allylic ether, the steric influence should operate predominantly for the formation of the product.  product.

Reaction Mechanism
We examined the stoichiometric reaction to gain insight into the reaction mechanism (Table 6). In the reaction of nickel(II) complex 1b with an equimolar amount of allylic ether 2b, oxidative addition of the C-O bond of 2b to 1b did not occur and 2b was recovered in 96%. In contrast, when 1b was treated with an excess amount of phenylmagnesium reagent 3b, 1b was converted into another nickel species. In the 31 P{ 1 H} NMR spectrum, the signal derived from 1b (37.5 ppm) disappeared [26] and a new signal was observed at 49.1 ppm, whose resonance was due to the formation of a new nickel species. However, characterization of this nickel complex was not successful. Subsequently, we examined the reaction with various amounts of 3a in the presence of 1b and 2b. The obtained results are summarized in Table 6. In the reaction using a mixture of 1b and 2b with 1 equivalent of 3a, the coupled product 4a was not detected with 2b being recovered in 83%. In the case of the reaction using 2 equivalents of 3a, the coupled product (4a) was formed in 23% yield. Using 3 equivalents of 3a, the reaction proceeded smoothly and 2b was completely consumed to give the coupled product (4a) in 83% yield, along with the formation of 4,4 -bitolyl (6a, 0.19 mmol). The reaction using 4 equivalents of 3a also gave the product (4a) in 83% yield. Therefore, we assumed that two or more equivalents of the arylmagnesium reagent are required to for the reaction to proceed via the generation of an anionic nickel(0) species, which is the active species for the activation of the allylic ether (vide infra).

Reaction Mechanism
We examined the stoichiometric reaction to gain insight into the reaction mechanism (Table 6). In the reaction of nickel(II) complex 1b with an equimolar amount of allylic ether 2b, oxidative addition of the C-O bond of 2b to 1b did not occur and 2b was recovered in 96%. In contrast, when 1b was treated with an excess amount of phenylmagnesium reagent 3b, 1b was converted into another nickel species. In the 31 P{ 1 H} NMR spectrum, the signal derived from 1b (37.5 ppm) disappeared [26] and a new signal was observed at 49.1 ppm, whose resonance was due to the formation of a new nickel species. However, characterization of this nickel complex was not successful. Subsequently, we examined the reaction with various amounts of 3a in the presence of 1b and 2b. The obtained results are summarized in Table 6. In the reaction using a mixture of 1b and 2b with 1 equivalent of 3a, the coupled product 4a was not detected with 2b being recovered in 83%. In the case of the reaction using 2 equivalents of 3a, the coupled product (4a) was formed in 23% yield. Using 3 equivalents of 3a, the reaction proceeded smoothly and 2b was completely consumed to give the coupled product (4a) in 83% yield, along with the formation of 4,4′-bitolyl (6a, 0.19 mmol). The reaction using 4 equivalents of 3a also gave the product (4a) in 83% yield. Therefore, we assumed that two or more equivalents of the arylmagnesium reagent are required to for the reaction to proceed via the generation of an anionic nickel(0) species, which is the active species for the activation of the allylic ether (vide infra).  Although the detailed reaction mechanism for the nickel catalyzed allyl-aryl coupling reaction was not clear, a plausible mechanism is proposed in Scheme 3 [13,28]. Initially, nickel(II) complex 1b is converted into anionic nickel(0) species I through its reaction with at least 2 equivalents of Although the detailed reaction mechanism for the nickel catalyzed allyl-aryl coupling reaction was not clear, a plausible mechanism is proposed in Scheme 3 [13,28]. Initially, nickel(II) complex 1b is converted into anionic nickel(0) species I through its reaction with at least 2 equivalents of arylmagnesium reagent accompanied by the formation of a biaryl. Then, the low valent nickel(0) species (I) reacts with the allylic electrophile at the nickel center to afford π-allyl nickel species II (as shown in Table 5): For example, γ-alkyl substituted allylic ether 2g was converted into a mixture of linear and branched isomers. We have recently reported that the steric and electronic properties of the nickel(II) complexes 1 [26,27]. The DFT calculations of complexes 1 reveal that the highest occupied molecular orbital (HOMO) levels of these complexes increase in the order 1a ≈ 1b < 1c ≈ 1d. We expected that complexes 1c and 1d worked as effective catalysts for this coupling reaction. However, the bulky substituent (mesityl group) on the ligand framework inhibits to form π-allyl nickel species derived from 1c and 1d. Therefore, complexes 1c and 1d show poor catalytic activities. In contrast, complex 1b acts as an effective catalyst for the coupling reaction, because of constructing an appropriate steric and electronic environment around nickel center by the O,N,P-pincer ligand. After the formation of π-allyl nickel species II, its subsequent transmetalation with the arylmagnesium reagent leads to the formation of a σ-allyl nickel species bearing aryl ligand III and/or IV. We assumed that the formation of intermediate III is more favorable than IV due to steric hindrance. Therefore, the selectivity in the reaction was influenced by the substituent on the allylic substrate. Finally, the desired products 4 and/or 5 are released via reductive elimination from the Ni center to regenerate the nickel(0) species (I) and thus, the catalytic cycle is completed. One of the possible reaction pathways involved in an allylic substitution reaction is the attack of the nucleophile on π-allyl nickel species I; this mechanism cannot be excluded.
Molecules 2019, 24 FOR PEER REVIEW 9 with the arylmagnesium reagent leads to the formation of a σ-allyl nickel species bearing aryl ligand III and/or IV. We assumed that the formation of intermediate III is more favorable than IV due to steric hindrance. Therefore, the selectivity in the reaction was influenced by the substituent on the allylic substrate. Finally, the desired products 4 and/or 5 are released via reductive elimination from the Ni center to regenerate the nickel(0) species (I) and thus, the catalytic cycle is completed. One of the possible reaction pathways involved in an allylic substitution reaction is the attack of the nucleophile on π-allyl nickel species I; this mechanism cannot be excluded.

Scheme 3.
A plausible catalytic cycle for the cross-coupling reaction of an allylic ether and arylmagnesium reagent.

General Information
All manipulations involving air-and moisture-sensitive organometallic compounds were carried out under a nitrogen atmosphere, which was dried with SICAPENT (Merck Co., Inc., Germany, Darmstadt), using standard Schlenk tube or high vacuum techniques. All solvents were distilled over appropriate drying agents prior to use. Nickel complexes 1a-c [26] and 1d [27] were prepared according to literature methods. Allylic aryl ethers 2a-c, 2e-g, and 2l were prepared from the reaction of their corresponding allylic halides and 4-substituted phenols using K2CO3 in acetone.

General Information
All manipulations involving air-and moisture-sensitive organometallic compounds were carried out under a nitrogen atmosphere, which was dried with SICAPENT (Merck Co., Inc., Germany, Darmstadt), using standard Schlenk tube or high vacuum techniques. All solvents were distilled over appropriate drying agents prior to use. Nickel complexes 1a-c [26] and 1d [27] were prepared according to literature methods. Allylic aryl ethers 2a-c, 2e-g, and 2l were prepared from the reaction of their corresponding allylic halides and 4-substituted phenols using K 2 CO 3 in acetone. Allylic methyl ether 2d was prepared via the reduction of the corresponding aldehyde using NaBH 4 followed by methylation using methyl iodide. Allylic methyl ethers 2h and 2i were prepared from addition of the corresponding aldehyde with vinylmagnesium bromide followed by methylation using methyl iodide. A mixture of 2j and 2k was prepared via the reduction of cinnamyl methyl ketone using NaBH 4 followed by a Mitsunobu reaction using 4-methoxyphenol. All other reagents employed in this research were commercially available and used without any further purification.
Proton nuclear magnetic resonance ( 1 H NMR), carbon nuclear magnetic resonance ( 13 C{ 1 H} NMR), and phosphorus nuclear magnetic resonance ( 31 P{ 1 H} NMR) spectra were recorded on BRUKER DRX-500 and JEOL ECA 500 spectrometers at ambient temperature. 1 H NMR chemical shifts were recorded in ppm using tetramethylsilane (TMS) or the solvent resonance peak as an internal standard (TMS: 0.00 ppm, C 6 D 6 : 7.16 ppm). 13 C{ 1 H} NMR chemical shifts were recorded in ppm using the solvent resonance peak as an internal standard (CDCl 3 : 77.0 ppm, C 6 D 6 : 128.0 pm). 31 P{ 1 H} NMR chemical shifts were recorded in ppm using H 3 PO 4 as an external standard (H 3 PO 4 : 0.0 ppm).

General Procedure for the Cross-Coupling Reaction of Allylic Ethers with Arylmagnesium Reagents
A 50 mL Schlenk tube was charged with 1b (1.0-2.5 mol%), the respective allylic electrophile (1.0 mmol), Et 2 O (2.5-5 mL, 0.2 M based on the electrophile), and the aryl Grignard reagent (1.5-3.0 equiv.) at room temperature. The coupling reaction was carried out at room temperature for 24 h. After quenching with 1 M HCl aq. (5 mL), the aqueous layer was extracted with EtOAc (5 × 3 mL). The combined organic layers were washed with brine (10 mL) and dried over anhydrous MgSO 4 . After filtration and removal of all volatiles from the filtrate, the residue was purified by column chromatography on silica gel.    [29].

Conclusions
We have described the cross-coupling reaction of allylic aryl ethers with arylmagnesium reagents catalyzed by nickel(II) pincer complexes 1. Complex 1b bearing a β-aminoketonato unit tethering a diphenylphosphino group showed excellent catalyst performance. The reactions of a range of cinnamyl ether derivatives proceeded with high regioselectively to afford solely the linear coupled products in high yield. The coupling reactions of αand γ-alkyl substituted allylic ethers gave a mixture of linear and branched products. These results indicate that the reaction proceeds via a π-allyl nickel intermediate. Based on stoichiometric reactions, we assumed that an anionic Ni(0) species derived from the reduction of nickel(II) complex 1b with an excess amount of a Grignard reagent was the active intermediate in the reaction. Further investigations on mechanistic aspects and the coupling reactions of various organometallic reagents with organic electrophiles are currently underway in our laboratory.