meta C–H Arylation of Electron-Rich Arenes: Reversing the Conventional Site Selectivity

Abstract

Controlling site selectivity of C–H activation without using a directing group remains a significant challenge. While Pd(II) catalysts modulated by a mutually repulsive pyridine-type ligand have been shown to favor the relatively electron-rich carbon centers of arenes, reversing the selectivity to favor palladation at the relatively electron-deficient positions has not been possible. Herein we report the first catalytic system that effectively performs meta C–H arylation of a variety of alkoxy aromatics including 2,3-dihydrobenzofuran and chromane with exclusive meta site selectivity, thus reversing the conventional site selectivity governed by native electronic effects. The identification of an effective ligand and modified norbornene (NBE-CO₂Me), as well as taking advantage of the statistics, are essential for achieving the exclusive meta selectivity.

Graphical Abstract

1. INTRODUCTION

C–H activation reactions without using a directing group offer synthetic utility that is complementary to the extensively developed chelation-assisted C–H activation reactions. In this context, development of Pd(II)-catalyzed nondirected C–H activation reactions has been hampered by the following two limitations: there is a lack of reactivity when 1.0 equiv of an arene is used;¹ achieving site selectivity based on electronic and steric effects has met with tremendous difficulties.² Although ligand-enabled reactivity with arenes as limiting reagents has been reported recently,² approaches to control the site selectivity of these reactions are lacking.

The lack of site selectivity of Pd(II)-catalyzed nondirected C–H activation is partially due to the ineffective recognition of steric effects by Pd(II) catalysts. For example, nondirected C–H activation reactions of electron-rich arenes with Pd(II) catalysts predominantly afford a mixture of ortho- and para-products.² We therefore proposed to introduce a bulky transient reaction partner to improve the selectivity based on steric effects. Based on our previous research on modified norbornene-assisted directed meta C–H functionalization,^3–5 we envisioned that C–H palladation of electron-rich arenes followed by norbornene insertion could favor the sterically less hindered para-positions. In addition, norbornene insertion at either ortho- or para-positions could relay the final C–H arylation to the same meta-positions, thereby statistically enhancing the site selectivity. Most interestingly, this approach will reverse the conventional site selectivity of electron-rich arenes to afford the meta-selective C–H arylation.

Alkoxy aromatics are ubiquitous in natural products and drug molecules, as well as versatile coupling partners through recent nickel chemistry (Scheme 1).^6,7 Since alkoxy groups are ortho and para directors for electrophilic palladation reactions, reversing this site selectivity to instead favor meta C–H functionalization of alkoxyarenes would enable new strategies for arene synthesis. We have previously reported meta-selective C–H functionalization of phenol using a tailor-designed U-shaped template (Scheme 2a).^8,9 The Larrosa group developed a two-step sequence in one pot, namely, installation of a carboxylic group to direct ortho C–H arylation and subsequent removal of the carboxylic group, to realize a formal meta C–H arylation of phenol (Scheme 2b).¹⁰ However, these two approaches are not compatible with phenyl ethers such as dihydrobenzofurans or chromanes which do not contain the free hydroxyl group. Herein we report the first catalytic system that performs nondirected meta-selective C–H arylation of a variety of electron-rich alkoxy aromatics (Scheme 2c), thus reversing the conventional site selectivity governed by native electronic effects. The use of a modified norbornene (NBE-CO₂Me) to relay the initial ortho- and para-palladation to meta-palladation is crucial for the observed site selectivity.

Scheme 1.

Bioactive Compounds Containing Phenols

Scheme 2.

Methods for meta C−H Activation of Phenols

2. RESULTS AND DISCUSSION

To implement the relay strategy using NBE-CO₂Me, we selected anisole as the model substrate for reaction optimization and began our study by treating the model substrate (anisole, 1.0 equiv) with aryl iodide (methyl 4-iodobenzoate, 2.0 equiv), Pd(OAc)₂ (15 mol %), modified norbornene (NBE-CO₂Me) (1.5 equiv), and AgOAc (3.0 equiv) in hexafluoro-2-propanol (HFIP) (solvent) at 95 °C (Table 1). Encouragingly, the desired meta-arylation products were detected in 15% mono and di combined yield (entry 1). Guided by our recent finding that electron-deficient 2-pyridone ligands can accelerate nondirected C–H activation,^2a,d we wondered whether pyridone would also help this norbornene-mediated arylation. Disappointedly, the optimal 2-pyridone ligand (L₁) improved the total yield to only 25%. The MPAA (monoprotected amino acid) ligand afforded only a slightly increased yield. A variety of pyridine and quinoline ligands were also examined. While electron-rich pyridine-based ligands shut down the reaction (L_2–3), a significantly enhanced yield was obtained when electron-deficient pyridine-based ligands were used (L_5–7). Notably, quinoxaline improved the total yield to 58% (L₆). Extensive screening of pyridine and quinoxaline-based ligands revealed that 6-cyanoquinoxaline was the best ligand for this nondirected arylation reaction, affording the combined mono and di meta-arylation products in 69% yield (L₇). As a control experiment, L₇ alone in the absence of NBE-CO₂Me failed to provide any reactivity, thus highlighting the importance of NBE-CO2Me in this meta-selective arylation reaction (entry 12). The desired products were obtained in 63% yield when using 0.5 equiv of NBE-CO₂Me instead of 1.5 equiv (entry 13).

Table 1.

Evaluation of Ligands^a,b

Entry	Ligand (NBE, Y/N)	Yield (%)
		mono	di	total
1	No Ligands (Y)	11	4	15
2	Ac-Gly-OH (Y)	13	6	19
3	L1 (Y)	17	8	25
4	L2 (Y)	-	-	<10
5	L3 (Y)	-	-	<10
6	L4 (Y)	16	8	24
7	L5 (Y)	35	15	50
8	L6 (Y)	40	18	58
9	L7 (Y)	46	23	69
10	L8 (Y)	23	10	33
11	L7 + L8 (Y)	50	26	76
12	L7 (N)	-	-	trace
13	L7 + L8 (0.5 equiv.)	43	20	63

^aConditions: 1a (0.2 mmol, 1.0 equiv), aryl iodide (2.0 equiv), Pd(OAc)₂ (15 mol %), NBE-CO₂Me (1.5 equiv), L (30 mol %), AgOAc (3.0 equiv), HFIP (0.5 mL), 95 °C, under air, 20 h. See SI for workup procedures.

^b1H NMR yields.

Considering the multiple steps involved in this complex catalytic cycle and the optimal ligand for each step may differ, we have also tested a dual ligand system.^2b,c,e Interestingly, the addition of an electron-deficient 3-pyridinesulfonic acid as the second ligand (L₈) improved the combined yield to 76% (entry 11, see SI for detailed ligand screening). Control experiments show that L₈ alone gave much lower reactivity than L₇ (for control experiments with several other substrates, see the SI). 3-Pyridinesulfonic acid (L₈) was previously used to promote sp³ C–H olefination.¹¹ The beneficial role of L₈ is likely due to the formation of a more reactive cationic L₇(L₈)Pd⁺-OAc complex, which could bind the arene substrate more effectively (Scheme 3). The rationale for the design of this type of anion-containing L-type ligand has been previously demonstrated.¹¹

Scheme 3.

Beneficial Effect of Dual Ligand System (L₇, L₈)

With the best conditions established, we next investigated the aryl iodide scope of this reaction (Table 2). Using 3-methylanisole as the model substrate, the corresponding arylation products were obtained in moderate to good yields and high regioselectivity with a series of aryl iodides regardless of their electronic properties (2ba–2bj). Various 2-substituted pyridine-based heteroaryl iodides are also compatible with this reaction (2bk–2bm). Several aryl iodides derived from estrone, borneol, and menthol were also effectively utilized to afford the desired products in 69%, 73%, and 71% yields, respectively (2bn–2bp).

Table 2.

Scope of (Hetero)Aryl Iodides^a,b

^aConditions: 1b (0.2 mmol), aryl iodides (2.0 equiv), Pd(OAc)₂ (15 mol %), NBE-CO₂Me (1.5 equiv), L₇ (30 mol %), L₈ (15 mol %), AgOAc (3.0 equiv), HFIP (0.5 mL), 95 °C, under air, 20 h. See SI for workup procedures.

^bIsolated yields.

^cAryl iodides were used as limiting reagents (1.0 equiv), 2.0 equiv of arenes.

A wide range of arenes were also examined in this meta-selective arylation reaction (Table 3). Various functionalities on the 3-positions of anisole including ester, ether, silyl, alkyl, amide, and phenyl groups were all well tolerated, affording the desired products in 56–78% isolated yields (2c–2l). The reaction remained highly meta-selective when 2-isopropylanisole (2m) and 5-methoxy-1,2,3,4-tetrahydronaphthalene (2n) were used as the substrates. Notably, other readily removable protecting groups of phenols such as isobutyl and tert-butylsilyl groups stayed intact during this arylation reaction (2o,p). Surprisingly, 2-methylanisole (2q) afforded the desired meta-product in only 28% yield, which is much lower than 2-isopropylanisole (2m). Large substituents (2m) at the C-2 position would impede the norbornene-related second C–H activation at C-3, thus selectively affording the C-5 arylation product in good yield. However, the second C–H activation at C-3 could still occur when the C-2 substituent is small (2q), and yet this intermediate is not reactive toward aryl iodide. Instead, reductive elimination led to the formation of the cyclobutane side product (see SI for more information). para-Substituted anisole gave only a trace amount of meta-arylation product due to the steric hindrance of the methyl group (2r). No desired product was observed with substrate containing strong electron-withdrawing groups (see SI for other unsuccessful substrates).

Table 3.

Scope of Substrates^a,b

^aConditions: 1c-1zd (0.2 mmol), aryl iodide (2.0 equiv), Pd(OAc)₂ (15 mol %), NBE-CO₂Me (1.5 equiv), L₇ (30 mol %), L₈ (15 mol %), AgOAc (3.0 equiv), HFIP (0.5 mL), 95 °C, under air, 20 h. See SI for workup procedures.

^bIsolated yields.

^cNo byproducts of arenes were observed in most cases.

^d4-Fluoro-1-iodobenzene (Ar^F) was used instead of methyl 4-iodobenzoate for better separation on TLC.

^eConversion is 46% (see SI for information about major byproduct).

^fOther regiosiomers were observed in less than 5%.

^g1H NMR yield, inseparable from aryl iodides and NBE-CO₂Me on TLC.

2,3-Dihydrobenzofurans and chromanes, another class of electron-rich arenes, are ubiquitous in drug molecules and natural products.¹² Although meta-C–H functionalization of 1-indoline and 1,2,3,4-tetrahydroquinolines has been accomplished by attaching a tethered directing template to the nitrogen atom,¹³ this strategy is incompatible with 2,3-dihydrobenzofurans and chromanes. Gratifyingly, our norbornene-mediated nondirected strategy provided a solution to this long-standing problem. Simple 2,3-dihydrobenzofuran was successfully arylated with high site selectivity in 48% isolated yield (2s). Substituents such as alkyl and ester groups on the C-2 position afforded good yields (2t, 2u). Esters and amides containing chromanes were also effective substrates with high site selectivity in 62–76% yields (2v–2y). More structurally complex chromanes with succinimide and pyrrolidone functionalities were successfully arylated (2z, 2za). To showcase the synthetic utility of this method, nondirected meta C–H arylation of the FDA approved drug Tasimelteon was carried out to provide the potentially useful arylated product in 72% yield (2zb). Other oxygen-containing heterocycles such as 1,4-benzodioxane and 2-methoxydiben-zofuran were also compatible substrates (2zc, 2zd). We also tested several unsubstituted phenols with different protecting groups, such as ethyl, benzyl, and triisopropylsilyl ether (TIPS). Ethyl- and benzyl-protected phenols afforded the desired products in a combined yield of 73% with exclusive meta selectivity (2ze and 2zf). Bulky TIPS-protected phenol gave 32% di product but only 10% mono product (2zg), which is probably due to the large steric hindrance of TIPS. Furthermore, meta-arylation of the chromane derivative (1w) was carried out on a 5.0 mmol scale, affording the desired product in 58% yield (Scheme 4).

Scheme 4.

Scale-up Reaction

Based on our design, the observed meta C–H arylation proceeds through the initial palladation at the ortho-position, which is then intercepted by the norbornene insertion followed by subsequent Catellani C–H arylation (Scheme 5a). To obtain evidence for this proposed mechanism, we performed deuterium incorporation experiments with ethox-ybenzene under the standard conditions in the presence of HFIP-ol-D (see SI for more details). The presence of 70% deuterium incorporation at exclusively the C-2 position in the monoarylated product suggests that the monoarylation product was derived from ortho C–H palladation and the catalytic cycle is terminated by the C-2 protonation (deuterium incorporation step) of the arylpalladium intermediate E (Scheme 5a). Interestingly, 27% D-incorporation at the ortho-position and 55% D-incorporation at the para-position are observed in the diarylated product. The para-deuteration originates from initial C–H para-palladation and subsequent consecutive diarylation as expected from the Catellani reaction. Conceivably, the monoarylated product could also undergo further ortho-palladation and proceeds to give the diarylation product, which could lead to the incorporation of the ortho-deuterium. This hypothesis is experimentally supported by the D-labeling in the monoarylation of 1b (Scheme 5b).

Scheme 5.

Deuteration Experiments and Proposed Mechanism

3. CONCLUSION

In conclusion, we have developed the first example of nondirected meta C–H arylation of electron-rich arenes. This reaction features broad substrate scope and good functional group compatibility, exemplified by successful C–H arylation of a range of complex molecules and FDA approved drug molecule Tasimelteon. Importantly, the ortho and para C–H palladation intermediates are not arylated prior to being intercepted by NBE-CO₂Me, thus affording exclusive meta-arylation products and reversing the conventional site selectivity of electron-rich arenes. Efforts to develop other transformations guided by similar design are currently underway in our laboratory.

4. EXPERIMENTAL SECTION

General Procedure for Nondirected meta C–H Arylation.

Alkoxybenzene (0.2 mmol), aryl iodide (2.0 equiv), Pd(OAc)₂ (15 mol %), modified norbornene (1.5 equiv), L₇ (30 mol %), L₈ (15 mol %), AgOAc (3.0 equiv), and HFIP (0.5 mL) were added to a reaction vial (10 mL). The vial was capped and closed tightly. Then the reaction mixture was stirred at 95 °C for 20 h. After cooling to room temperature, the mixture was passed through a pad of Celite with ethyl acetate as the eluent to remove the insoluble precipitate. The resulting solution was concentrated and purified by preparative thin-layer chromatography to afford the desired product.