Palladium-Catalyzed Oxidative Arylhalogenation of Alkenes: Synthetic Scope and Mechanistic Insights

Abstract

This article describes the development of a Pd-catalyzed reaction for the arylhalogenation (halogen = Cl or Br) of diverse α-olefins by oxidatively intercepting Mizoroki-Heck intermediates. These transformations afford synthetically useful 1,2- and 1,1-arylhalogenated products in good yields with good to excellent selectivities that can be modulated by changing the nature of the halogenating reagent and/or the reaction conditions. The selectivity of these reactions can be rationally tuned by: (i) controlling the relative rates of oxidative functionalization versus β-hydride elimination from equilibrating Pd^II-alkyl species and (ii) stabilization of organometallic Pd^II intermediates through the formation of π-benzyl adducts. These arylhalogenations exhibit modest to excellent levels of stereospecificity, and the key carbon-halogen bond-forming step proceeds with predominant retention of stereochemistry at carbon.

Introduction

The Mizoroki-Heck reaction involves the Pd-catalyzed coupling of an alkene with an aryl halide or aryl metal species.¹ This transformation is widely used in organic synthesis for the construction of carbon–carbon bonds,² and significant effort has been devoted to both catalyst development^1–3 and mechanistic investigations.^1–4 As shown in eq. 1, the general mechanism of the Heck reaction involves formation of a σ-aryl Pd species followed by olefin insertion to generate σ-alkyl Pd intermediate A and then β-hydride elimination/olefin dissociation to release product B.

(1)

Methods for intercepting Mizoroki-Heck intermediate A via other fundamental organometallic reactions would provide attractive routes for the 1,2-arylfunctionalization of α-olefins (eq. 2). Such transformations would complement the traditional Mizoroki-Heck reaction by providing access to products containing one additional bond and one additional stereocenter as compared to B. Some success has been realized in this area by trapping intermediate A via alkene, alkyne, or CO insertion.^5,6 However, the scope, generality, and range of products accessible from these reactions remains limited, in large part because the relative rate of β-hydride elimination to form B is often competitive in these processes.

(2)

On the basis of our interest in high oxidation state palladium chemistry, we thought that oxidation of A would provide an attractive alternative route for intercepting this intermediate. Indeed, very early studies by Heck demonstrated the feasibility of this transformation. For example, in 1968, he reported the Pd-catalyzed reaction of methylvinylketone with PhHgCl and CuCl₂ to afford 1,2-arylchlorination product 1 in 80% yield (eq. 3).⁷ However, the overall synthetic utility of this transformation was limited by a modest substrate scope, competing β-hydride elimination/olefin dissociation (to form 2), and the requirement for aryl mercury reagents.

(3)

Sporadic subsequent reports have also suggested that Heck intermediate A can be intercepted under oxidative conditions to generate 1,1-difunctionalized products.⁸ For example, Tamaru and coworkers demonstrated the palladium-catalyzed 1,1-phenylchlorination of 1-octene with PhSnBu₃ in the presence of CuCl₂ (eq. 4).^9,10 However, the yield of this transformation was not reported, and 1-octene was the only substrate examined.¹¹ More recently, a similar Pd-catalyzed 1,1-phenylchlorination product 4 was unexpectedly observed in the reaction between 3, PhSnBu₃, and CuCl₂ (eq. 5).¹² Again, this was an isolated example, and the reaction was not explored further.

(4)

(5)

These prior studies suggested that it might be possible to develop oxidative transformations for the predictable and selective conversion of α-olefins into 1,2 or 1,1-arylfunctionalized products of general structures C and D (eq. 6). Recent work from our group and others has shown that σ-alkyl Pd intermediates related to A can be intercepted with a wide variety of oxidants to install diverse functional groups, including C–Cl, C–Br, C–I, C–F, C–N, C–O, and C–C bonds. This has been demonstrated in the context of Pd-catalyzed ligand-directed C–H functionalization,¹³ where sp³ C–H activation has been followed by oxidative halogenation,¹⁴ amination,¹⁵ oxygenation,¹⁶ and arylation.¹⁷ In addition, a variety of sequences involving amino-, oxy- or halopalladation of olefins to generate σ-alkyl Pd intermediates followed by oxidative halogenation,^18,19 amination,²⁰ oxygenation,²¹ and arylation²² have been developed to afford diverse organic products.

(6)

We report herein the implementation of this strategy in the development of general oxidative Mizoroki-Heck reactions for the 1,2- and 1,1-arylhalogenation of alkenes.²³ The full scope of both 1,2-and 1,1-arylchlorination and bromination reactions is described. Additionally, investigations of the mechanism and the stereospecificity of these transformations are discussed in detail.

Results

Reaction Design

In order to develop general and robust conditions for alkene arylfunctionalization, we needed a working mechanistic hypothesis for the formation of each of the products. As shown in Scheme 1, the Heck product B is known to be generated from σ-alkyl Pd intermediate A via β-hydride elimination/olefin dissociation. The 1,2-arylfunctionalization product C could be formed by direct oxidative functionalization of A. Finally, the isomeric 1,1-product D could derive from A via a sequence involving β-hydride elimination, olefin insertion with opposite regiochemistry, and finally oxidative functionalization of the resulting Pd-benzyl complex (E) (Scheme 1). This mechanistic proposal suggests that the ratio of products B, C, and D should be tunable by modifying the relative rates of four fundamental organometallic transformations: oxidative functionalization, β-hydride elimination, olefin dissociation, and olefin insertion.

Scheme 1.

Working Mechanistic Hypothesis for Formation of B-D

1,2-Arylchlorination Reactions

We reasoned that use of the highly reactive electrophilic chlorinating reagent PhICl₂ should increase the rate of oxidative functionalization of A relative to that of β-hydride elimination, thereby promoting formation of 1,2-product C. Thus, we first studied the Pd-catalyzed reaction between alkene 5, PhICl₂, and PhSnBu₃²⁴ in a variety of common solvents. Unexpectedly, the 1,1-product (6b) was the major isomer observed at room temperature in all solvents examined (Table 1, entries 1–5). The most promising initial lead was in CH₂Cl₂, which showed formation of significant quantities of 6a (1 : 1.3 ratio of 6a : 6b, 46% overall yield, entry 5). Lowering the temperature to −78 °C and increasing the reaction concentration to 0.064 M in CH₂Cl₂ led to a significant enhancement in both selectivity for and yield of 6a (entry 8). The yield was improved further by increasing the amount of PhSnBu₃ from 1.3 to 2.6 equiv (entry 9). Ultimately, optimal conditions were found to be: 10 mol % of PdCl₂(PhCN)₂, 2 equiv of PhICl₂, 2.6 equiv of PhSnBu₃ in CH₂Cl₂ (0.064 M) starting at −78 °C with gradual warming over 5 h to 25 °C. This afforded a quantitative yield of 6 as a 10 : 1 mixture of isomer 6a : 6b (entry 9).

Table 1.

Optimization of the 1,2-Arylchlorination Reaction

Entry	Solvent	Temperature	Concentration	Yielda	6a : 6ba
1	Dioxane	25 °C	0.032 M	77%	1 : 6.7
2	THF	25 °C	0.032 M	53%	1 : 9.6
3	Et2O	25 °C	0.032 M	64%	1 : 9.7
4	C6H6	25 °C	0.032 M	55%	1 : 8.9
5	CH2Cl2	25 °C	0.032 M	46%	1 : 1.3
6	CH2Cl2	0 °C to 25 °C	0.032 M	55%	1.9 : 1
7	CH2Cl2	−78 °C to 25 °C	0.032 M	56%	8 : 1
8	CH2Cl2	−78 °C to 25 °C	0.064 M	82%	10 : 1
9b	CH2Cl2	−78 °C to 25 °C	0.064 M	100%	10 : 1

^aYield and ratio of isomers determined by ¹H NMR spectroscopic analysis of the crude reaction mixture.

^b2.6 equiv of PhSnBu₃.

These conditions proved general for the 1,2-arylchlorination of numerous α-olefins (Table 2). The reactions were tolerant of a wide variety of common functional groups, including esters, aromatic and alkyl halides, benzylic hydrogens, amides, and silyl ethers. Additionally, both electron rich and electron deficient arylstannanes were effective coupling partners. The products were obtained in good to excellent yields and with good selectivity for the 1,2-isomer. In general, the mass balance in these reactions consisted of small amounts of the 2,1-phenylchlorination, dichlorination, and/or β-hydride elimination products.

Table 2.

Substrate Scope for 1,2-Arylchlorination

Entry	Alkene	Stannane	Major Product	Yielda	1,2 : 1,1
1		PhSnBu3		72%	8 : 1
2		PhSnBu3		84%	13 : 1
3		PhSnBu3		96%	9 : 1
4		PhSnBu3		92%	11 : 1
5		p-FC6H4SnBu3		84%	11 : 1
6		PhSnBu3		85%	6: 1
7		p-BrC6H4SnBu3		96%	7 : 1
8		PhSnBu3		86%	8 : 1
9		PhSnBu3		68%	6 : 1
10		PhSnBu3		86%	10 : 1
11		p-CH3C6H4SnBu3		74%	5 : 1
12		PhSnBu3		80%	14: 1

^aIsolated yields and selectivities. Notably, in some cases, the isolated material contained minor impurities (typically 2,1-phenylchlorination or dichlorination products). See Supporting Information for details.

^bThe 1,1 and 1,2-isomers were generally not separable by column chromatography, but could be isolated in pure form using HPLC.

^cThe variation in 1,2 : 1,1 selectivity as a function of substrate may be due to differing interactions of polar functional groups within the substrate with the palladium catalyst.

1,1 Arylchlorination Reactions

We next sought conditions that would selectively provide 1,1-arylchlorinated products. The solvent screen in Table 1 showed that the reaction of 5 with PhICl₂ significantly favored formation of the 1,1-product 6b in ethereal solvents. We reasoned that the use of a less reactive chlorinating reagent would further slow competing 1,2 functionalization and thus provide higher selectivity for 6b. Gratifyingly, the use of NCS or CuCl₂ under otherwise identical conditions resulted in >20 : 1 selectivity for 6b in a variety of solvents (Table 3). While the yields were modest at room temperature (ranging from 10–58%), they could be significantly improved by conducting the reactions at −78 °C. Under optimal conditions (10 mol % of PdCl₂(PhCN)₂, 4 equiv of CuCl₂, 1.3 equiv of PhSnBu₃ in Et₂O (0.032 M) starting at −78 °C with gradual warming over 5 h to 25 °C), 6b was obtained in 82% yield and >20 : 1 selectivity as determined by ¹H NMR spectroscopy (entry 8).

Table 3.

Optimization of 1,1-Arylchlorination Reaction

Entry	Solvent	Temperature	Yield CuCl2a	Yield NCSa
1	Dioxane	25 °C	13%	34%
2	C6H6	25 °C	28%	48%
3	AcOH	25 °C	10%	46%
4	CH2C12	25 °C	32%	36%
5	THF	25 °C	58%	39%
6	THF	−78 °C to 25 °C	75%	ndb
7	Et2O	25 °C	29%	36%
8	Et2O	−78 °C to 25 °C	82%	44%

^aYield and ratio of isomers determined by ¹H NMR spectroscopic analysis of crude reaction mixture.

^bnd = not determined

The substrate scope of 1,1-arylchlorination with CuCl₂ was similar to that of the PhICl₂ reactions. Aryl stannanes and α-olefins containing diverse functional groups were effective coupling partners, and the 1,1-isomer was consistently obtained in >20 : 1 selectivity (Table 4). In all cases, <10% of alkene products derived from β-hydride elimination/alkene dissociation were observed by ¹H NMR spectroscopic analysis of the crude mixtures.²⁵

Table 4.

Substrate Scope for 1,1-Arylchlorination²⁵

Entry	Alkene	Stannane	Major Product	Yielda	1,2 : 1,1
1		PhSnBu3		75% (53%)	1 : >20
2		PhSnBu3		83% (54%)	1 : >20
3		PhSnBu3		84% (71%)	1 : >20
4		o-CH3C6H4SnBu3		87% (67%)	1 : >20
5		PhSnBu3		86% (66%)	1 : >20
6		p-FC6H4SnBu3		76% (59%)	1 : >20
7		PhSnBu3		73% (71%)	1 : >20
8		PhSnBu3		51% (41%)	1 : >20
9		PhSnBu3		74% (55%)	1 : >20
10		PhSnBu3		73% (71%)	1 : >20
11		p-CH3C6H4SnBu3		78% (72%)	1 : >20
12		PhSnBu3		94% (82%)	1 : >20
13		p-BrC6H4SnBu3		82% (54%)	1 : >20

^aYield determined by ¹H NMR spectroscopic analysis of crude reaction mixture. Isolated yields are in parentheses.

1,2 and 1,1 Arylbromination Reactions

We next pursued analogous palladium-catalyzed arylbromination reactions using CuBr₂ and NBS as electrophilic brominating reagents (PhIBr₂ is not readily available).^26,27 As shown in Table 5, NBS was poorly effective for this transformation. In contrast, CuBr₂ efficiently promoted the formation of both 1,1-arylbromination product 28b and 1,2-arylbromination product 28a. The isomer ratio in these CuBr₂-mediated reactions could be tuned by simply changing the solvent. For example, 1,1-product 28b was formed in high (82%) yield and with >20 : 1 selectivity in Et₂O (entry 6). However, moving from Et₂O to THF under similar conditions resulted in a reversal of selectivity to afford 28a as the major product (28a : 28b = 3 : 1), albeit in modest (49%) yield (Table 5, entry 8). This is in notable contrast to the trend observed for arylchlorination, where the use of CuCl₂ in THF favored 1,1-products. Subsequent optimization revealed that increasing the concentration to 0.128 M resulted in 80% yield of a 10 : 1 mixture of 28a : 28b (entry 10). Under these conditions, the mass balance consisted of 29 (20% yield), which is the product of 1,2-addition of Ph and ring opened THF to the alkene.²⁸

Table 5.

Optimization of Arylbromination Reaction

Entry	Solvent	Temperature	Concentration	Yield CuBr2a	28a : 28ba	Yield NBSa	28a : 28ba
1	Dioxane	25 °C	0.032 M	29%	1 : >20	19%	1 : >20
2	C6H6	25 °C	0.032 M	24%	1 : 13	8%	1 : >20
3	AcOH	25 °C	0.032 M	16%	1 : >20	0%	ndb
4	CH2Cl2	25 °C	0.032 M	27%	1 : >20	6%	1 : >20
5	Et2O	25 °C	0.032 M	46%	1 : >20	8%	1 : >20
6	Et2O	−78 °C to 25 °C	0.032 M	82%	1 : >20	ndb	ndb
7	THF	25 °C	0.032 M	25%	1 : >20	11%	1 : >20
8c	THF	−78 °C to 25 °C	0.032 M	49% (17%)	3 : 1	ndb	ndb
9c	THF	−78 °C to 25 °C	0.064 M	50% (5%)	10 : 1	ndb	ndb
10c	THF	−78 °C to 25 °C	0.128 M	80% (20%)	10 : 1	ndb	ndb

^aYield and ratio of isomers determined by ¹H NMR spectroscopic analysis of crude reaction mixture. The yield of 29 is shown in parentheses where applicable.

^bnd = not determined.

^c1.9 equiv of PhSnBu₃.

As shown in Table 6 and Table 7, the optimal conditions for formation of the 1,1- and 1,2-arylbrominated products were general for numerous α-olefins. These transformations exhibited a scope and functional group tolerance similar to the arylchlorination reactions. Notably, THF-addition products analogous to 29 were formed as significant by-products (5–20% yield as determined by ¹H NMR spectroscopic analysis of the crude reaction mixtures) in the 1,2-arylbrominations.²⁹

Table 6.

Substrate Scope for 1,2-Arylbromination²⁹

Entry	Alkene	Major Productb	Yielda	1,2 : 1,1a
1			73% (54%)	23 : 1(15 : 1)
2			86% (64%)	17: 1 (15: 1)
3			72% (56%)	14 : 1 (11 : 1)
4			84% (60%)	13 : 1 (15: 1)
5			87% (63%)	14 : 1 (20 : l)c
6			87% (63%)	14 : 1 (22 : 1)
7			65% (52%)	12 : 1 (10: 1)

^aYield and ratio of isomers determined by ¹H NMR spectroscopic analysis of crude reaction mixture. Isolated yields and selectivities are in parentheses.

^bBetween 5 and 20% of THF-addition products analogous to 29 was observed by crude ¹H NMR spectroscopic analysis with all substrates.

^cIsolated ratio reflects 1,2 product to a mixture of 1,1 and 1,4 products.

^dThe 1,1 and 1,2-isomers were generally not separable by column chromatography, but could be isolated in pure form using HPLC.

^eThe variation in 1,2 : 1,1 selectivity as a function of substrate may be due to differing interactions of polar functional groups within the substrate with the palladium catalyst.

Table 7.

Substrate Scope for 1,1-Arylbromination²⁹

Entry	Alkene	Major Product	Yielda	1,2 : 1,1a
1			66% (56%)	1 : >20
2			85% (70%)	1 : 15
3			70% (67%)	1 : >20
4			88% (60%)	1 : >20
5			79% (69%)	1 : >20
6			50% (41%)b	1 : >20
7			80% (68%)	1 : >20
8			64% (45%)	1:>20

^aYield and selectivity determined by ¹H NMR spectroscopic analysis of crude reaction mixture. Isolated yields are in parentheses.

^bApproximately 10% of the Heck byproduct was observed by crude NMR.

Mechanistic Investigations: General Goals

We next turned our efforts to investigating the mechanism of these transformations. In particular, we sought to: (i) interrogate the proposed reaction pathways for formation of the 1,2- and the 1,1-products, (ii) understand the role of oxidant and solvent in imparting 1,2- versus 1,1 selectivity, (iii) gain mechanistic insights into the high selectivity for benzylic functionalization with CuCl₂ and CuBr₂ in Et₂O, and (iv) probe the stereochemical course of the key carbon-halogen bond-forming step of these transformations. The results of these mechanistic studies are detailed below, and their implications are discussed later in the manuscript.

Deuterium Labeling

Deuterium labeled 1-octene-(1,1-d₂) was utilized as a substrate for both the 1,2- and 1,1-phenylhalogenations. Reaction with PhICl₂ or with CuBr₂/THF afforded 40a–Cl/Br as the sole 1,2-phenylhalogenated products (eq. 7). Less than 5% deuterium incorporation was observed at any other site along the alkyl chain.

(7)

The reaction of 1-octene-(1,1-d₂) was also examined with CuCl₂ and CuBr₂/Et₂O. In both cases, 40b–Cl/Br was the only 1,1-phenylhalogenated product formed (eq. 8). Again, less than 5% of other isomers were observed, indicating clean migration of the deuterium from the 1- to the 2-position.⁸

(8)

4-(4-Chlorophenyl)-1-butene

The reactivity of 4-(4-chlorophenyl)-1-butene (33) was examined under both 1,2- and 1,1-arylhalogenation conditions. With CuCl₂ and CuBr₂/Et₂O, mixtures of two isomeric products were formed (eq. 9). In both cases, the major isomer was the expected 1,1-product 41-Cl/Br. However, a minor amount of a second benzylic halogenation product 42-Cl/Br was also observed. The ratio of 41-Cl : 42-Cl was 4 : 1 in the chlorination reaction and the ratio of 41-Br : 42-Br was 7 : 1 under the bromination conditions as determined by ¹H NMR spectroscopic analysis of the crude reaction mixtures.

(9)

The use of PhICl₂ or CuBr₂/THF as the oxidant produced 1,2-arylhalogenated products 43-Cl/Br in 62/81% crude yield (eq. 10). The analogous 1,1-products were also formed (17/4% crude yield) along with traces (2/2%) of the 1,4-arylfunctionalized compounds 42-Cl/Br (eq. 10). Notably, isomers derived from chlorination/bromination at other sites in the middle of the alkyl chain were not detected in these transformations.

(10)

Styrene Substrates

The Pd-catalyzed reaction of 4-fluorostyrene with PhSnBu₃ under 1,2-arylhalogenation conditions afforded the expected 1,2-isomer (44-Cl/Br) with >20 : 1 selectivity (Table 8, entries 1 and 3). Intriguingly, the 1,2-product 44-Cl/Br was also favored in this transformation with CuCl₂ or CuBr₂/Et₂O as the oxidant (selectivity = 1.4 : 1 in the chlorination and 9 : 1 in the bromination reaction) (Table 8, entries 2 and 4). This experiment was repeated with styrene as the alkene substrate and (p-FC₆H₄)SnBu₃, and, again, 1,2-arylhalogenated products predominated in all four cases (Table 8, entries 5–8).

Table 8.

Reactions of Styrene Substrates

Entry	Oxidant/Solvent	X	Y	Z	Yield	Ratio 1,2 : 1,1
1	PhICl2/CH2Cl2	Cl	F	H	50%a	>20 : 1a
2	CuCl2/Et2O	Cl	F	H	48%b	1.4 : 1b
3c	CuBr2/THF	Br	F	H	45%b	>20 : 1b
4d	CuBr2/Et2O	Br	F	H	69%b	9 : 1b
5	PhICl2/CH2Cl2	Cl	H	F	79%a	> 20 : 1a
6	CuCl2/Et2O	Cl	H	F	50%b	2.2 : 1b
7c	CuBr2/THF	Br	H	F	30%b	> 20 : 1b
8d	CuBr2/Et2O	Br	H	F	50%b	24 : 1b

^aYield and selectivity determined based on isolated material.

^bYield and selectivity determined by ¹H NMR spectroscopic analysis of crude reaction mixture.

^cReaction conditions were optimized to −78 °C to 25 °C with 2 equiv of organostannane to increase the yield – there was no difference in selectivity compared to the reaction at 0 °C.

^dPd(acac)₂ used as [Pd].

Vinylnaphthalene Substrate

The Pd-catalyzed reaction between vinylnaphthalene and (p-FC₆H₄)SnBu₃ was also explored, and the 1,2-arylhalogenation product 46-Cl/Br was the major isomer observed under all four sets of conditions (Table 9).³⁰ The regioselectivity with CuCl₂ and CuBr₂/Et₂O increased significantly in comparison to that observed in the corresponding reactions of styrene. For example, upon changing from styrene to vinylnaphthalene, the 1,2 : 1,1 selectivity changed from 2.2 : 1 to >50 : 1 for chlorination with CuCl₂ and from 24 : 1 to >50 : 1 for bromination with CuBr₂/Et₂O.

Table 9.

Reactions of Vinylnaphthalene

Entry	Oxidant/Solvent	X	Yielda	Ratio 1,2 : 1,1a
1	PhICl2/CH2Cl2	Cl	60%b	> 50 :1
2	CuCl2/Et2O	Cl	76%	> 50 : 1
3c	CuBr2/THF	Br	12%d	> 50 : 1
4c	CuBr2/Et2O	Br	48%d	> 50 : 1

^aYield and selectivity determined by ¹H NMR spectroscopic analysis of crude reaction mixture.

^bMass balance in 1,2-chlorination is accounted for by formation of 25% of the 1,2-dichloro product.

^cPd(acac)₂ used as [Pd].

^dMass balance in arylbromination reactions is predominantly accounted for by incomplete conversion (15% conversion of vinylnaphthalene in entry 3 and 70% conversion of vinylnaphthylene in entry 4).

Solvent Effects in Arylbromination

To explore the role of solvent in the 1,1- versus 1,2-arylbrominations with CuBr₂/THF, the reactions were examined in 2-methyltetrahydrofuran (Me-THF) and 2,5-dimethyltetrahydrofuran (Me₂-THF). As shown in Table 10, changing the solvent resulted in a dramatic decrease in selectivity for the 1,2-product. With Me-THF, a 1.5 : 1 ratio of 28a : 28b was obtained, with similar results (28a : 28b = 1.8 : 1) in Me₂-THF.

Table 10.

Effect of Substituted THF-Derivatives on 1,2- versus 1,1-Phenylbromination with CuBr₂

Entry	Solvent	Yielda	Ratio 1,2 : 1,1a
1	THF	80%	15 : 1
2	MeTHFb	37%	1.5 : 1
3	Me2-THFc	57%	1.8 : 1

^aYield and selectivity determined by ¹H NMR spectroscopic analysis of crude reaction mixture.

^bMe-THF = 2-methyltetrahydrofuran.

^cMe₂-THF = 2,5-dimethyltetrahydrofuran.

These reactions were also examined in a series of aprotic solvents with dielectric constants (ε) ranging from 7.58 to 2.25. As shown in Table 11, there is a rough correlation between product ratio and ε, with more of 1,2-product 28a generally formed in higher dielectric solvents.

Table 11.

Effect of Aprotic Solvents on 1,2 versus 1,1 Arylbromination with CuBr₂^a

Entry	Solvent	ε	Yieldb	Ratio 1,2 : 1,1b
1	THF	7.58	49%c	3.4 : 1
2	1,2 Dimethoxyethane	7.2	14%	> 20:1
3	EtOAc	6.02	63%	1 : 8
4	CHCl3	4.81	45%	1: > 20
5	Et2O	4.33	82%	1 : 24
6	Anisole	4.33	23%d	1: > 20
7	iPr2O	3.9	76%	1 : 5.9
8	MTBE	2.6	56%	1: > 20
9	Dioxane	2.25	30%	1 : 20

^aThese reactions were performed at 0.032 M to maximize solubility and improve reproducibility across this broad range of solvents.

^bDetermined by ¹H NMR spectroscopy.

^c1.9 equiv PhSnBu₃ used.

^d0 °C → 25 °C.

Stereochemistry

Finally, we examined the stereochemical outcome of these transformations with alkenes cis-48 and trans-48. With cis-48, all of the reactions produced arylfunctionalization product 49-Cl/Br as the major diastereomer with modest to excellent levels of selectivity (Table 12, entries 1–3).³¹ Compound 49-Cl/Br is the result of net syn addition of Ph and X across the alkene.^31,32 With trans 48, the stereoselectivity was significantly diminished, particularly with PhICl₂ (Table 12, entries 4–6).³² However, in all cases, the major product was diastereomer 50-Cl/Br, which is again the product of syn Ph/X addition.

Table 12.

Arylhalogenation of cis-48 and trans-48^a

Entry	Substrate	Oxidant	Solvent	Yieldb,c	Ratio 49 : 50b
1	cis-48	PhICl2	CH2Cl2	40%	12 : 1
2	cis-48	CuCl2	Et2O	51%	>30 : 1
3	cis-48	CuBr2	Et2O	41%	5 : 1
4	trans-48	PhICl2	CH2Cl2	21%	1 : 1.6
5	trans-48	CuCl2	Et2O	45%	1 : 8
6	trans-48	CuBr2	Et2O	9%	1 : 8

^aConditions: 2 equiv PhSnBu₃, 10 mol % PdCl₂(PhCN)₂, 0.032 M, −78 → 23 °C, 36 h.

^bYield and selectivity determined by ¹H NMR spectroscopic analysis of crude reaction mixture.

^cThe mass balance in these reactions is accounted for by recovered starting material as well as the corresponding Mizoroki-Heck product. See Supporting Information for full details.

Discussion

1,2-Arylhalogenation Reactions

All of the mechanistic experiments presented above are consistent with the mechanism proposed in Scheme 1. In this scenario, 1,2-arylhalogenation products are generated via oxidative halogenation of Mizoroki-Heck intermediate A. The deuterium labeling studies in eq. 7 are in line with this proposal, as they show clean formation of 40-Cl/Br without accompanying migration of the deuterium label.

The 1,2-arylbrominations are particularly intriguing because simply moving from THF to Et₂O reverses product selectivity. We were thus very interested to understand the role of solvent in these transformations. As shown in Table 10, changing the solvent from THF to Me-THF or Me₂-THF resulted in a steep drop in selectivity for the 1,2-product 28a. Me-THF and Me₂-THF are expected to have very similar polarity but significantly poorer coordinating capabilities than THF.³³ Thus, this result suggests that THF coordination (either to Pd³⁴ or Cu³⁵) may play a key role in switching the selectivity in this system. The results in Table 11 suggest that solvent polarity may also be an important factor for selectivity, with more polar solvents favoring the 1,2 product. Polar coordinating solvents are likely to enhance the solubility of CuBr₂, which should increase the rate of oxidative bromination relative to that of β-hydride elimination, thereby providing improved selectivity for 1,2-product 28a.

1,1-Arylhalogenation Reactions

These mechanistic studies provide evidence that the 1,1-products are derived from the β-hydride elimination/reinsertion/oxidation mechanism shown in Scheme 1 and in Scheme 2, path 1. The migration of deuterium from the 1 to the 2 position to generate 40b-Cl/Br (eq. 8) is fully consistent with this mechanism. In addition, the formation of 1,4-addition product 42-Cl/Br from the reactions of 4-(4-chlorophenyl)-1-butene with CuX₂/ether (eq. 9) provides strong evidence for equilibrating β-hydride elimination/reinsertion steps prior to oxidative cleavage.

Scheme 2.

Two Possible Mechanisms for Formation of 1,1-Products

A possible alternative pathway involving a standard Mizoroki-Heck mechanism followed by hydrohalogenation of the resulting alkene (Scheme 2, path 2) can be ruled out based on the data in Table 8. In path 2, the reaction of styrene/(p-FC₆H₄)SnBu₃ should produce the same alkene product as that of 4-fluorostyrene/PhSnBu₃. Accordingly, the mechanism in path 2 should afford identical mixtures of 44 to 45 for both reactions. However, as shown in Table 8, the ratio of 44 : 45 changed depending on the alkene/stannane used (entries 2 versus 6 and 4 versus 8).

Selective 1,1-arylfunctionalization with CuX₂/ether appears to derive from a preference for oxidative functionalization of Pd^II–benzyl intermediates (like E) versus Pd^II–alkyl intermediates (like A) (Scheme 2). This preference is reflected in the sole formation of 42 and 41 (as opposed to isomers resulting from halogenation at other sites along the alkyl chain) in reactions of 4-(4-chlorophenyl)-1-butene with CuX₂/ether (eq. 9). Additionally, the observation that 1,2-products predominate in the functionalization of styrene derivatives with CuX₂/ether (Table 8) is consistent with fast oxidative functionalization of an initially formed Pd^II-benzyl species in these systems.

We hypothesize that the selectivity for Pd-benzyl functionalization under equilibrating conditions derives from a stabilizing π-benzyl interaction (F-π, eq. 11). This proposal is supported by the experiments in Table 8 and Table 9, which compare CuX₂/ether reactions of styrene to those of vinylnaphthalene. With both CuCl₂ and CuBr₂, the selectivity for 1,2-functionalization increases significantly upon moving from styrene to vinylnaphthalene (from 2.2 : 1 to >50 : 1 with CuCl₂, and from 24 : 1 to >50 : 1 with CuBr₂). In both cases, initial alkene insertion should generate a Pd^II-benzyl intermediate F-σ/π (eq. 11) or Pd^II-naphthyl intermediate G-σ/π (eq. 12). Thus, the increase in 1,2-product with vinylnaphthalene suggests that k_Ox/k_β-H (where k_Ox = rate constant for direct oxidative functionalization and k_β-H = rate constant for β-hydride elimination) is significantly larger for intermediate G than for F. Importantly, this is consistent with literature precedent, which has shown that π-naphthyl complexes are more kinetically reactive towards nucleophilic functionalization than their π-benzyl counterparts.³⁶

(11)

(12)

Stereochemistry

The stereochemical course of oxidatively-induced sp³-carbon–halogen bond-formation is of interest from a fundamental mechanistic perspective. Additionally, the future development of asymmetric arylfunctionalization reactions hinges upon the ability to achieve clean retention or inversion of stereochemistry at carbon during C–X coupling. Without high stereochemical fidelity in this step, the new stereocenter will be eroded over the course of the reaction.

As summarized in Table 12, the arylfunctionalization of cis-48 and trans-48 favors products of syn addition. The insertion of alkenes into Pd–Ph bonds is well-known to occur with high syn-selectivity.³⁷ Thus, this result indicates that C–X bond-formation occurs primarily with retention of configuration at carbon (eq. 13).

(13)

Carbon–halogen formation most likely occurs via initial 1 or 2 e⁻oxidation of Pd^II–alkyl species A with PhICl₂ or CuX₂ to form a Pd^IV or Pd^III intermediate.^38,39,40 The observed retention of configuration is consistent with direct C–X bond-forming reductive elimination from this high oxidation state Pd center (eq. 14, path 1)⁴¹ or with a phenyl group assistance mechanism involving double inversion (eq, 14, path 2).^41b Both of these pathways have been proposed in the literature for related transformations.⁴¹

(14)

While syn addition predominated in every case, modest levels of stereospecificity were observed in some reactions, particularly with trans-48 and PhICl₂ (Table 12, entry 4). There are several potential explanations for these observations. First, selectivity would be eroded if oxidative functionalization occurred with competing retention and inversion of configuration at carbon. Literature studies of related oxidatively-induced carbon-heteroatom coupling reactions at Pd suggest that both retention⁴¹ and inversion⁴² are possible. Further, small modifications of the reaction conditions can sometimes lead to large changes in the stereochemical outcome.⁴³ However, such competing pathways would most likely affect reactions of cis-48 and trans-48 to similar degrees and thereby provide similar levels of stereochemical erosion with both the cis and trans substrates. Thus, this explanation does not readily account for the dramatically different results for cis-48/PhICl₂ (12 : 1 isomer ratio) versus trans-48/PhICl₂ (1 : 1.6 isomer ratio).

A second possible explanation for formation of mixtures of diastereomers would be competing isomerization of the alkene starting material under the reaction conditions. This scenario would lead to an “apparent” erosion of stereospecificity even if C–X coupling proceeded with clean retention. Olefin isomerization is likely to be particularly problematic with trans-48, since it undergoes arylfunctionalization at a slower rate than cis-48.⁴⁴ Thus, the observation of lower diastereoselectivities in reactions of trans-48 is consistent with a contribution from this pathway. In addition, ¹H NMR spectroscopic analysis of the reaction between trans-48/PhSnBu₃, and PhICl₂ at 50% conversion revealed the presence of detectable quantities (2%) of cis-48. Furthermore, reaction of cis-48 at 23 °C with 2 equiv of PhICl₂ and 2 equiv of PhSnBu₃ resulted in 18% yield and an 8 : 1 ratio of 49-Cl : 50-Cl) along with 17% recovery of trans-48. These results clearly indicate that alkene isomerization can occur to a significant extent under the PhICl₂/CH₂Cl₂ reaction conditions. While this side reaction undoubtedly needs to be addressed, the observation of good levels of stereospecificity with the cis substrates provides impetus for moving forward in the development of asymmetric versions of these transformations.

Summary

In summary, we have developed Pd-catalyzed reactions for the arylchlorination and arylbromination of α-olefins by oxidatively intercepting Mizoroki-Heck intermediate A (Scheme 1). The selectivity of these reactions can be rationally tuned by controlling the relative rates of oxidative functionalization versus β-hydride elimination from equilibrating Pd^II-alkyl species and by π-benzyl stabilization of Pd^II intermediates. This work provides a mechanistic basis for the future development of synthetically useful aryloxygenation, arylamination, arylhalogenation, and diarylation reactions using a variety of oxidants and transmetalating reagents. Additionally, the insights gained from these studies will be valuable for the design of enantioselective versions of these transformations. Work in all of these areas is ongoing in our laboratory and will be reported in due course.