Palladium-Catalyzed γ-Arylation of α,β-Unsaturated Esters from Silyl Ketene Acetals

The palladium-catalyzed coupling of enolates with aryl and vinyl electrophiles has become a useful method for the construction of carbon–carbon bonds.^[1] Although many palladium complexes have been identified that catalyze the coupling of enolates of ketones, esters, amides, and aldehydes with haloarenes, the vinylogous coupling of dienolates of α,β-unsaturated carbonyl compounds with haloarenes (γ-arylation) has been less studied. γ-Aryl α,β-unsaturated carbonyl compounds are useful building blocks because they can be further functionalized at the carbonyl or olefinic positions.^[2]

To develop a catalyst for the γ-arylation of α,β-unsaturated carbonyl compounds, one must address several issues, including the regioselectivity for forming products from α-, β-, or γ-arylation, the selectivity for monoarylation versus diarylation, and the potential condensation of the reactive products. The selectivity for formation of products from α or γ-arylation versus β-arylation depends on whether an enolate complex forms by transmetalation with the arylpalladium halide intermediate, as occurs during the α-arylation of carbonyl compounds, or whether it forms by insertion of the α,β-unsaturated carbonyl compound into the metal–aryl bond, as occurs during a Heck reaction. The selectivity for products from α- versus γ-arylation likely depends on the stability and reactivity of the series of isomeric arylpalladium dienolate intermediates that result from transmetalation.

Despite these obstacles, a few γ-arylation reactions of α,β-unsaturated ketones and aldehydes have been reported. Terao et al. reported the palladium-catalyzed γ-arylation of unfuntionalized α,β-unsaturated aldehydes and ketones.^[3] Martin and Buchwald reported the γ-arylation of similar α-substituted α,β-unsaturated aldehydes.^[4] Varseev and Maier later reported a sequential γ-arylation and dehydrogenation for the synthesis of substituted tetralones.^[5] Finally, Hyde and Buchwald reported the γ-arylation of γ-substituted α,β-or β,γ-unsaturated ketones^[6] and lactones^[7] with aryl chlorides and bromides in moderate to good yield.

Despite these advances, the coupling of enolates of α,β-unsaturated esters is underdeveloped. The γ-arylation of acyclic esters has been limited to reactions of tin enolates in low to moderate yields.^[8] Moreover, the previous couplings of α,β-unsaturated ketones and aldehydes have required elevated temperatures and substrates that contain substituents at either the α or γ position of the carbonyl compound for high yield.

Herein, we report the γ-arylation and γ-vinylation of α,β-unsaturated esters in high yield with broad scope by coupling aryl, heteroaryl, and vinyl halides with the corresponding silicon dienolates. These reactions occur in high yield with a variety of esters, including those lacking substituents at the α and γ positions, and with aromatic and vinylic electrophiles that contain potentially reactive functional groups. Moreover, these reactions occur without typical fluoride additives to activate silicon enolates.

Our studies of the coupling of haloarenes with the dienolates of α,β-unsaturated esters began with alkali metal dienolates of α,β-unsaturated esters. However, the reaction of bromobenzene with the alkali metal dienolate of methyl trans-2-hexenoate did not form the γ-aryl α,β-unsaturated esters in significant yield. Thus, we investigated reactions of silicon enolates of α,β-unsaturated esters. These studies were based on our prior work on the coupling of aryl halides with silyl enol ethers and silyl ketene acetals.^[9–11]

Our initial studies on the coupling of silyl ketene acetals derived from α,β-unsaturated esters focused on the model palladium-catalyzed coupling of bromobenzene with the trimethylsilyl ketene acetal of methyl-2-hexenoate (1a). Most previous α-arylations of silyl ketene acetals or silyl enol ethers were conducted with fluoride additives, such as ZnF₂, CsF, MgF₂, Bu₃SnF, or CuF₂. In contrast to these previous studies, the coupling of silyl ketene acetal 1a occurred without a fluoride additive. Instead, the reaction formed the γ-aryl E-product 2 in high yield in the presence of inexpensive zinc chloride. Previously reported coupling reactions of aryl halides with silicon enolates without a fluoride additive have been limited to the α-arylation and vinylation of silicon enolates in the presence of thallium acetate^[11] or to intramolecular α-vinylations.^[9]

A summary of the effect of the silyl group and additive on the coupling of bromobenzene with the silyl ketene acetals of methyl hexenoate is shown in Table 1. The γ-arylation product was formed in the highest yields with the triethylsilyl (TES) ketene acetal (1b; Table 1, entry 3). The coupling between 1b and bromobenzene in the presence of ZnF₂ occurred in low yield (Table 1, entry 4), perhaps owing to the poor solubility of ZnF₂ in tetrahydrofuran. Reactions of 1b with bromobenzene in the presence of ZnBr₂ and ZnI₂ (Table 1, entries 5 and 6) formed the γ-arylation product 2 in lower yield than the same reaction in the presence of ZnCl₂. The coupling between 1b and bromobenzene with catalytic amounts of ZnCl₂ occurred in yields similar to those of reactions with stoichiometric ZnCl₂, although the reaction was slower (Table 1, entry 7). Ester 2 was not formed from the reaction of 1b with bromobenzene in the absence of ZnCl₂ (Table 1, entry 1).

Table 1.

Effects of the silyl group and zinc halides on the γ-arylation of silyl ketene acetals with bromobenzene.^[a]

Entry	Acetal	Additive	t[h]	Yield [%][b]
1	1b	none	24	<5
2	1a	ZnCl2	2	82
3	1b	ZnCl2	4	87
4	1b	ZnF2	24	11
5	1b	ZnBr2	4	81
6	1b	ZnI2	25	<5
7 [c]	1b	ZnCl2	12	83
8	1c	ZnCl2	12	64

^[a]Reaction conditions: acetal (0.26 mmol), PhBr (0.20 mmol), additive (0.30 mmol), [Pd(dba)₂] (0.010 mmol), PtBu₃ (0.01 mmol), THF (0.8 mL).

^[b]Yields were determined by GC analysis with dodecane as an internal standard (average of two runs).

^[c]0.2 equiv ZnCl₂

Several palladium catalyst systems for the coupling of silyl ketone acetal 1b with bromobenzene were also explored; ester 2 did not form in a detectable amounts from the coupling of 1b with bromobenzene in the presence of catalyst systems containing X-phos, dppe, dppf, Xantphos, rac-BINAP, PCy₃, or PPh₃ ligands. Ester 2 formed in the presence of [Pd(dba)₂] and Qphos, but the yield was lower than that from the reaction conducted with a catalyst containing PtBu₃.

The coupling of silyl ketene acetal 1b with chlorobenzene and iodobenzene was also attempted. The coupling product 2 was not detected from the reaction with chlorobenzene, and the γ-aryl α,β-unsaturated ester 2 was formed in only 29% yield from the coupling of 1b with iodobenzene after 24 hours.

The scope of the γ-arylation with a variety of aryl bromides is shown in Table 2. γ-Aryl α,β-unsaturated esters were obtained in high yields from reactions with both electron-rich and electron-poor aryl bromides. The reaction proceeded in high yield with aryl bromides containing a single ortho substituent (Table 2, entries 6 and 7); higher catalyst loadings were required for more sterically hindered aryl bromides (Table 2, entry 8), but the reaction of 2,6-dimethyl-bromobenzene occurred in high yield. The reaction tolerated a variety of functional groups that could be used for further elaboration, such as a chloro, nitro, cyano, acyl, and a phenacyl group (Table 2, entries 21–25). The coupling also proceeded with aryl bromides that contained acidic N–H groups (Table 2, entries 16, 17, and 30).

Table 2.

Aryl and vinyl bromide scope of the γ-arylation of 1b.^[a]

Entry		[Pd] cat.[b]	Yield [%][c]
1	PhBr	A	82
2[d]	PhBr	B	77
3[e]	PhBr	C	77
4	4-tBu-C6H4Br	A	82
5	3,5-Me2-C6H3Br	A	78
6	2-MeC6H4Br	A	83
7	2-PhC6H4Br	A	78
8[f]	2,6-Me2C6H3Br	A	96
9[g]	2,6-Me2C6H3Br	B	83
10	2,6-Me2C6H3Br	C	95
11	1-bromonaphthalene	A	82
12	2-bromonaphthalene	A	90
13	2-OMe-C6H4Br	A	83
14	3-OMe-C6H4Br	A	81
15	4-OMe-C6H4Br	A	83
16[f][h]	4-NH2-C6H4Br	A	74
17[h]	4-MeN(H)-C6H4Br	A	80
18	4-NMe2-C6H4Br	A	83
19	4-CF3-C6H4Br	A	77
20	4-F-C6H4Br	A	91
21	4-Cl-C6H4Br	A	91
22	4-CN-C6H4Br	A	81
23	4-NO2-C6H4Br	A	82
24	4-C(O)Me-C6H4Br	A	58
25	4-C(O)Et-C6H4Br	A	84
26	4-CO2Me-C6H4Br	A	89
27	3-bromothiophene	A	76
28	3-bromopyridine•BEt3	A	58[i]
29	6-bromoquinoline	A	69[j]
30	5-bromoindole	A	64
31	2-bromoindene	A	51
32	1-Br-2-Me-1-propene	A	34

^[a]Reaction conditions: 1b (1.3 mmol), aryl or vinyl halide (1.0 mmol), ZnCl₂ (1.5 mmol), [Pd] cat. (0.010 mmol), THF (2 mL).

^[b][Pd] cat.: A: [Pd(dba)₂], PtBu₃ (1:1); B: [Pd(PtBu₃)₂]; C: [{P(tBu₃)Pd(m-Br)}₂].

^[c]Yield of isolated product (average oftwo runs).

^[d]0.4% [Pd] cat.

^[e]0.1 % [Pd] cat.

^[f]5% [Pd] cat.

^[g]2% [Pd] cat.

^[h]2.6equiv 1b, 3.0equiv ZnCl₂.

^[i]Isolated as uncoordinated pyridine.

^[j]Product was isolated in 85% purity.

A variety of heteroaryl bromides also successfully coupled with silyl ketene acetal 1b, including the coupling of 1b with bromothiophene, bromopyridine, and bromoquinoline (Table 2, entries 27–29), and the coupling of 1b with 5-bromo indole containing an unprotected N–H bond (Table 2, entry 30). The coupling of 1b with the BEt₃ adduct of 3-bromopyridine occurred faster and in comparable yields to the reactions of the free pyridine. The use of a borane to inhibit coordination of a pyridine to the catalyst and to promote carbon–halogen bond cleavage and subsequent reductive elimination has been previously reported for the amidation of aryl halides.^[12]

The coupling also occurred with vinyl bromides; γ-vinyl-α,β-unsaturated esters were isolated in moderate yields from the reactions of 2-bromoindene and 1-bromo-2-methyl-1-propene (Table 2, entries 31 and 32).

The isolated complexes [Pd(PtBu₃)₂] and [{(PtBu₃)Pd(μ-Br)}₂] were particularly active for the coupling of silyl ketene acetal 1b with aryl bromides. For example, the coupling of 1b with bromobenzene proceeded in high yield in the presence of just 0.4 % [Pd(PtBu₃)₂] (Table 2, entry 2) or 0.2 % [{(PtBu₃)Pd(μ-Br)}₂] (Table 2, entry 3). The coupling between silyl ketene acetal 1b and a larger 10 mmol scale of bromobenzene in the presence of 0.2 % [{(PtBu₃)Pd(μ-Br)}₂] gave the coupled product ester 2 in 69 % yield.

The coupling of the more sterically demanding substrates also proceeded with lower catalyst loadings of [Pd(PtBu₃)₂] or [{(PtBu₃)Pd(μ-Br)}₂] than of the combination of [Pd(dba)₂] and PtBu₃ (Table 2, entries 9 and 10). Thus, [Pd(PtBu₃)₂] or [{(PtBu₃)Pd(μ-Br)}₂] could used in place of [Pd(dba)₂] and PtBu₃ in most cases when a single-component catalyst is preferred.

The scope of the process with different silyl ketene acetals is shown in Table 3. The reaction tolerates substitution at the α, β, γ, or δ positions of the acetal. Products from reactions of silyl ketene acetals containing substituents at the α or β positions were isolated as a mixture of olefin geometries (Table 3, entries 1 and 2). The coupling of a silyl ketene acetal containing two substituents at the γ position required increased catalyst loading, but occurred in 75% yield (Table 3, entry 3). Finally a substrate substituted in the δ position reacted in similar yields. Under our reaction conditions, no product was observed from the γ-arylation of the silyl ketene acetals derived from α,β-unsaturated lactones.

Table 3.

γ-Arylation of substituted silyl ketene acetals.^[a]

Entry	R1	R2	R3	R4	Yield [%][b]
1[c][d]	Me	H	H	Me	62 (E)
					10 (Z)
2	Me	H	Me	H	57 (E)
					5 (Z)
3[d]	Me	Me	H	H	75
4	iPr	H	H	H	68

^[a]Reaction conditions: acetal (1.5mmol), PhBr (1.0mmol), ZnCl₂ (1.5 mmol), [Pd(dba)₂] (0.010 mmol), PtBu₃ (0.01 mmol), THF (2 mL).

^[b]Yield of isolated product (average of two runs).

^[c]2.2 equiv silyl ketene acetal, 3.0 equiv ZnCl₂.

^[d]2% [Pd(dba)₂],2% PtBu₃.

A potential mechanism for the palladium-catalyzed coupling of 1b with aryl bromides is shown in Scheme 1. This catalytic cycle is analogous to that proposed for the palladium-catalyzed α-arylation of carbonyl compounds. In this cycle, the zinc chloride would facilitate the formation of the palladium dienolate complex (4 and 5), and reductive elimination from a palladium dienolate complex would generate the desired product. Reductive elimination from an η³-palladium dienolate complex (4) has been reported by Kurosawa and co-workers,^[13] but the zinc chloride induced transmetalation of a silyl ketene acetal or silyl enol ether with the palladium aryl halide complex has not been studied.

Scheme 1.

Potential mechanism for the coupling of 1b with PhBr. TES = triethylsilyl.

To understand the promoting effect of zinc chloride, we studied reactions containing additives that possess Lewis acidic properties or chloride ions. To determine if the Lewis acidity of the zinc halide facilitates transmetalation of the enolate from silicon to palladium, we conducted the coupling of enolate 1b with bromobenzene in the presence of BF₃·THF. The reaction with BF₃·THF in place of ZnCl₂ proceeded to full conversion, albeit over a longer reaction time and in diminished yield (Table 4, entry 3). To assess whether the zinc chloride served as a source of halide,^[14] the coupling of enolate 1b with bromobenzene was performed in the presence of zinc triflate. This reaction occurred in a yield that was comparable to that from the reaction in the presence of zinc chloride (Table 4. entry 2). Because the yield and reaction times for the coupling in the presence of zinc triflate were comparable to those of reactions in the presence of zinc chloride, we conclude that the ZnCl₂ is not simply a source of halide; instead the Lewis acidic property likely helps labilize the halide on palladium.^[15]

Table 4.

Effect of different additives on the coupling of 1b with bromobenzene.^[a]

Entry	Additive	t[h]	Yield [%][b]
1	ZnCl2	4	87
2	Zn(OTf)2	4	95
3	BF3•THF	48	39
4	N(octyl)3MeCl	8	83
5[c]	N(octyl)3MeCl	24	45

^[a]Reaction conditions: 11 (0.030 mmol), 1b (0.13 mmol), additive (0.15 mmol), THF (0.8 mL).

^[b]Yields were determined by GC analysis with dodecane as an internal standard (average of two runs).

^[c]Reaction performed in the presence of 4 equiv 4-ClC₆H₄Br.

At the same time, a genuine source of halide without Lewis acid did promote the coupling process. The reaction of 1b with bromobenzene catalyzed by [Pd(dba)₂] and PtBu₃ in the presence of methyltrioctylammonium chloride occurred in high yield (Table 4, entry 4), although the rate was slower than that for reactions conducted with zinc chloride. Because ammonium halides are not Lewis acidic, it appears that the γ-arylation process with the silyl enolates can be accelerated by either a Lewis acid or a Lewis base. The palladium-catalyzed coupling of 1b with bromobenzene in the presence of methyltrioctylammonium chloride and zinc chloride together required long times and did not occur in high yield (Table 4, entry 5); thus, we conclude that the ammonium chloride and zinc chloride promote the coupling of silyl ketene acetal and bromobenzene through different mechanisms.

To determine which portion of the catalytic cycle is affected by the zinc and ammonium additives, reactions of enolate 1b were performed with an arylpalladium bromide complex that is closely related to the complexes that would be intermediates in the coupling process. Arylpalladium halide complexes of P(1-Ad)tBu₂ (1-Ad = 1-adamantyl) are more easily isolated than those containing PtBu₃, and the combination of [Pd(dba)₂] and P(1-Ad)tBu₂ catalyzed the formation of 2 in 86 % yield. Thus, we studied complex 6 containing P(1-Ad)tBu₂. The reaction of silyl ketene acetal 1b with complex 6 was faster in the presence of zinc chloride than in the absence of any additive (Table 5 entries 1 and 2). When the reaction between 6 and 1b was performed in the presence of an excess of an aryl halide (with an aryl moiety that is different from that bound to palladium) to trap the palladium(0) product, the γ-arylation product 2 was formed in yields that were comparable to those of the catalytic reaction (Table 5, entry 3). Thus, the zinc halide accelerates the transmetalation between silicon and palladium. Chloride sources also triggered the stoichiometric reaction of 6 with 1b, but we have not yet identified the role of tetraalkylammonium halides in the catalytic process.

Table 5.

Stoichiometric reactions between arylpalladium halide specides with 1b.^[a]

Entry	Additive	t	Yield 2 [%][b]
1	none	30 h	27
2	ZnCl2	10 min	65
3[c]	ZnCl2	10 min	82
4	N(octyl)4Cl	30 min	69

^[a]Reaction conditions: 6 (0.030 mmol), 1b (0.13 mmol), additive (0.15 mmol), THF (0.8 mL).

^[b]Yields were determined by GC analysis with dodecane as an internal standard (average of two runs).

^[c]Reaction performed in the presence of 4 equiv 4-ClC₆H₄Br.

In summary, we have developed the first palladium-catalyzed coupling of aryl halides with acyclic α,β-unsaturated esters to form products from coupling at the γ position. These reactions are conducted via silyl ketene acetals, and, unlike many couplings of silicon enolates, does not require a fluoride additive. These reactions occur at room temperature with low catalyst loadings, in high yield with aryl bromides that contain a wide array of functional groups and with silyl ketene acetals that contain substitution at the α, β, γ, and δ positions. The product is formed by reductive elimination from a palladium dienolate complex. Further studies are required to determine the precise mechanism for activation of silyl ketene acetal, but the formation of this dienolate complex is promoted by the mildly acidic zinc chloride. Further studies on reactions of dienolates and efforts to develop stereoselective transformations of these species are in progress.

Experimental Section

General procedure for the γ-arylation of silyl ketene acetals catalyzed by PtBu₃ and [Pd(dba)₂] (Table 2): PtBu₃ (2.0 mg, 0.010 mmol; in a 1 dram vial) was placed inside a drybox under either an argon or nitrogen atmosphere. Bromobenzene (157 mg, 1.00 mmol) was added, followed by triethyl(1-methoxyhexa-1,3-dienyloxy)silane (346 mg, 1.30 mmol). Zinc chloride (204 mg, 1.50 mmol) was then added, followed by [Pd(dba)₂] (5.8 mg, 0.010 mmol). A Teflon-coated stirrer bar was then added, followed by 2 mL of THF. The vial was closed with a cap that was lined with a PTFE/silicone septum and removed from the drybox. The reaction mixture was stirred at room temperature until GC analysis showed that full consumption of the aryl bromide had occurred. Upon completion, the reaction mixture was diluted with 30 mL of ethyl acetate (EtOAc) and washed three times with saturated NaHCO₃. An additional 10 mL of EtOAc was added to the combined aqueous washes for back-extraction. The combined organic solutions were washed once with brine and dried with anhydrous MgSO₄. The suspension was filtered through a plug of Celite. Volatile materials were removed by rotary evaporation. The crude mixture was then purified by column chromatography on silica gel (acetone/pentane) to give the γ-aryl α,β-unsaturated ester.