Cu-Catalyzed Site-Selective Benzylic Chlorination Enabling Net C─H Coupling with Oxidatively Sensitive Nucleophiles

Abstract

Site-selective chlorination of benzylic C─H bonds is achieved using a Cu^ICl/bis(oxazoline) catalyst with N-fluorobenzenesulfonimide as the oxidant and KCl as a chloride source. This method exhibits higher benzylic selectivity relative to established chlorination protocols and is compatible with diverse alkyl arenes. Sequential benzylic C─H chlorination/nucleophilic substitution affords C─O, C─S, and C─N coupling products with oxidatively sensitive coupling partners.

Graphical Abstract

Benzylic chlorides are versatile synthetic intermediates that react readily with heteroatom- and carbon-based nucleophiles in conventional substitution¹ and catalytic cross-coupling reactions.^2-6 These compounds may be accessed via radical-chain chlorination,^7,8 leveraging the comparatively low bond dissociation energy and higher reactivity of benzylic C─H bonds (Figure 1A). Radical-chain chlorination methods that involve chlorine radical (Cl•) as the hydrogen-atom transfer (HAT) reagent, however, often exhibit poor C─H site selectivity. Consequently, benzylic chlorides are commonly prepared by the reaction of benzylic alcohols with SOCl₂ or via other functional-group interconversion methods.^2,3,6,9,10 C─H chlorination methods that employ species other than Cl• for the HAT step can show improved selectivity.^11-22 In this context, Cu catalysts in combination with N-fluorobenzenesulfonimide as the oxidant (Cu/NFSI) promote diverse C─H functionalization and oxidative cross-coupling reactions (e.g., Figure 1B)^23-35 that exhibit high benzylic site selectivity. These methods involve a radical-relay mechanism, in which HAT generates a diffusible benzylic radical that undergoes subsequent functionalization by Cu^II and a nucleophilic coupling partner. Here, we report the development of a Cu/NFSI method for benzylic chlorination that uses KCl as the source of chloride (Figure 1C). A comparison with other chlorination reactions highlights the unique benzylic selectivity of Cu/NFSI reactivity. The method uses the C─H substrate as the limiting reagent, and it may be paired with a subsequent nucleophilic substitution step to achieve net C─H coupling with nucleophiles that are not compatible with direct oxidative C─H functionalization.

Figure 1.

Chlorine radical as an unselective HAT agent (A), representative Cu/NFSI benzylic functionalization methods (B), and reaction sequence illustrating Cu/NFSI benzylic chlorination (C).

Several mechanistic considerations contributed to initial efforts in this study. A possible C─H chlorination mechanism, shown in Figure 2, reflects insights from other Cu/NFSI C─H functionalization reactions.^30,32 Copper(I) reacts with NFSI to generate a bis(sulfonyl)imidyl radical (•NSI) that can promote HAT. The resulting benzylic radical could undergo Cu^II-mediated functionalization via one of several possible mechanisms (inner-sphere^24,26,29 or outer-sphere^30,31 coupling with a coordinated ligand, or radical-polar crossover involving a benzylic cation intermediate^30,31,34,35). Cu/NFSI C─H functionalization reactions often require a sacrificial reductant or "redox buffer," such as a dialkylphosphite^30,34,35) to support catalytic turnover. The reductant provides a means to regenerate Cu^I if all of the Cu^I catalyst is oxidized to Cu^II via undesired reaction of •NSI (see red/blue pathways in Figure 2).

Figure 2.

Proposed mechanistic cycle for benzylic chlorination leveraging redox-buffering

Initial screening studies were initiated with isobutyl benzene (1a) as the substrate and reaction conditions incorporating a Cu salt, ancillary ligand, NFSI, and a chloride source (Table 1). Trimethylsilyl (TMS)-substituted nucleophiles have been used in previous Cu/NFSI reactions, including those with TMSCN,²⁴ TMSN₃,³¹ TMSNCO.³⁴ These reagents are expected to transfer anionic ligands to Cu^II, promoted by Si─F bond formation; however, TMSCl proved rather ineffective. No reactivity was observed in the absence of a phosphite additive (Table 1, entry 1), and only 22% yield of the benzylic chloride was observed in the presence of phosphite (entry 2). Chlorination of the phosphite additive was observed under these conditions (see Figure S3 in the Supporting Information). Use of alkali metal chlorides improved the reaction outcome (entries 3-7), with use of KCl contributing to the best yield (entry 7). Soluble chloride sources, such as NBu₄Cl, inhibit the reaction (see Table S8 in the Supporting Information). The identity of the ancillary ligand influenced the reaction outcome (entries 7-10; see Table S3 in the Supporting Information for other ligands). The best yields were obtained with bathophenantroline L1 and the bis(oxazoline) (BisOx) ligand L4 (entries 7 and 10); however, the latter generally led to higher yields for different substrates (see Figure S2 in Supporting Information).³⁶ Thus, L4 was selected for subsequent studies, even though no enantioselectivity was observed from the reactions. The latter observation is rationalized by a radical-polar crossover mechanism in which the intermediate benzylic radical is oxidized by Cu^II to a benzylic cation, which then reacts with the chloride nucleophile. The latter mechanism is consistent with the seminal early studies of Kochi,³⁷ which showed that Cu^II salts promote oxidation of organic radicals to the corresponding cations. This mechanism resembles other recently reported Cu/NFSI C─H functionalization reactions.^30,31,34,35

Table 1.

Cu/NFSI Benzylic Chlorination Reaction Optimization.

entry	ligand (mol%)	M─Cl	Conv. (%)	yield 1-B:1-T (%)b
1	L1	TMSCl	0	0:0c
2	L1	TMSCl	22	17:1
3	L1	LiCl	29	14:2
4	L1	NaCl	39	28:2
5	L1	RbCl	53	39:4
6	L1	CsCl	72	57:5
7	L1	KCl	75	59:7
8	L2	KCl	68	56:4
9	L3	KCl	65	58:6
10	L4	KCl	80	66:5

^a0.2 mmol scale.

^b1H NMR yields, ext. std. = mesitylene.

^cReaction conducted without phosphite reductant

Once optimal reaction conditions were identified, the Cu/NFSI-promoted chlorination reaction was then compared with other established C─H chlorination methods (Figure 3). Specifically, isobutyl and isopentyl benzene (1a and 2a) were used to compare the reaction yield and selectivity for 2° benzylic versus 3° aliphatic C─H chlorination. The different methods included the following: (a) Cu/NFSI (see Table 1), (b) Ag(phen)/^tBuOCl (phen = 1,10-phenanthroline),¹⁵ in which ^tBuO• or a Ag^IIO^tBu species is proposed to initiate HAT, (c) N-chlorosuccinimide (NCS)/dibenzoylperoxide (DBPO),³⁸ with radical-chain chlorination propagated by HAT from a succinimidyl radical, (d) Cu(OAc)₂/N-hydroxyphthalimide (NHPI)/trichloroisocynanuric acid (TCA),¹³ proposed to involve HAT by oxygen- and nitrogen-centered radicals, (e) light-promoted benzophenone (Ph₂CO)/NCS,¹² with HAT promoted by the excited state ketone, (f and g) radical-chain chlorination SO₂Cl₂/DBPO,⁸ propagated by Cl•-mediated HAT, and (h) Mn(tetraphenylporphyrin)Cl (Mn(TPP)Cl)/NaOCl,¹¹ which involves HAT by a Mn^v=O species.

Figure 3.

Applications of varied C─H chlorination methods on isobutylbenzene and isopentylbenzene. Yields determined by ¹H NMR spectroscopy (external std. = mesitylene).

Bond dissociation energies for the benzylic and tertiary C─H sites (~89 and ~93 kcal/mol, respectively³⁹) suggest that reactivity should be strongly favored at the benzylic position. The data in Figure 3, however, reveal a wide range of outcomes from the different chlorination methods. Cu/NFSI conditions support benzylic chlorination in good yield and selectivity, with benzylic:tertiary (B:T) ratios of 10:1 and 20:1 for 1a and 2a. Ag(phen)/^tBuOCl affords the highest yields and selectivities among the other seven conditions, with benzylic chlorination yields of 37% and 53% for 1a and 2a, respectively, and B:T ratios of ~2:1, relative to Cu/NFSI. Conventional radical-chain chlorination conditions (methods c, f, and g) afford lower yields and B:T selectivities, and even slightly favoring tertiary over benzylic chlorination in some cases (see methods f with 1a and method g with 1a and 2a). All reaction conditions generate side products corresponding to dichlorination and/or oxygenation of the substrate (in gray).

The results in Figure 3 highlight the unique benzylic selectivity of •NSI-mediated HAT and provide a starting point for testing of Cu/NSFI C─H chlorination with a series of other substrates (Figure 4). Para-substituted alkyl benzenes and 2-ethyl naphthalene, 3-5, 9, 10, undergo chlorination in good-to-excellent yields (69-97%), with the exception of the 4-ethyl anisole 5 (24%), which generates the benzylic bis(sulfonyl)imide as a significant byproduct (~50%). Propyl arenes with various substituents at the terminal position lead to benzylic chlorination products in good yield (75-84%, 6-8), providing versatile bifunctional electrophilic building blocks. 3-Phenylpropanoic acid methyl ester and alkyl trifluoroacetamides are chlorinated solely at the benzylic position in good yields (11-14, 70-74%).⁴⁰ The lack of functionalization next to the carbonyl or trifluoroacetamide groups is consistent with electronic deactivation of these sites toward reaction with the electrophilic •NSI species. A bromo-substituted fluorene derivative reacts effectively (15, 79%). Common pharmacophores 6-bromochroman 16 and alpha-tetralone 17 afford chlorination products in good yield (75% and 60%, respectively), while pyridine-containing substrates exhibit lower yields (20, 34%; 21, 38%). Benzylic chlorination of more complexes structures also proved effective (cf. 18, 19, 22-26), including successful reaction of Celestolide (18, >99%), Ibuprofen methyl ester (19, 52%), the tetrahydroquinoline precursor to a GnRH antagonist (22, 40%), the xanthine oxidase inhibitor, benzbromarone (23, 68%), and a dronedarone derivative (24, 89%). The data in Figure 4 reflect isolated yields, with product volatility accounting for ¹H NMR yields in a few cases (1, 3-5). In addition, separation challenges resulted in several products being isolated as a mixture with unreacted C─H substrate (2, 6, 7, 12, 13, 20, 21). This issue is not considered problematic because the substrate will be inert in the subsequent reactions of the benzylic chloride and is readily separated at that stage (see below).

Figure 4.

Scope of alkylarene benzylic chlorination. Reactions were run on a 0.2, 0.5, or 1.2 mmol scale. ^aReaction run at 35 °C. ^bReaction run at RT. ^cReaction run at 70 °C. ^c10 mol% TMSCl added. ^dYields determined by ¹H NMR spectroscopy (ext. std. = mesitylene). ^eIsolated as the alcohol.

The data in Figure 4 complement previously reported Cu/NFSI methods for C─H functionalization (cf. Figure 1B), but they also introduce new synthetic opportunities. Cu/NFSI oxidative coupling methods are not compatible with oxidatively sensitive nucleophiles, such as phenols, thiophenols, and alkylamines, due to undesirable side reactivity between these nucleophiles and NFSI. The reactivity here provides a means to bypass this limitation via sequential C─H chlorination/diversification in which the benzylic chlorides are used in subsequent coupling reactions. This concept is illustrated in Figure 5, using benzylic chlorides derived from the trifluoroacetyl-protected aminopropylbenzene (12) and the dronedarone derivative (26). Coupling of 12 with a series of phenols and thiophenols affords analogues of the antidepressant drug, fluoxetine (i.e., Prozac) (27-30). The observed C─O coupling products 27 and 28 are noteworthy because a recently reported³³ C─H fluorination/substitution sequence leads to C─C coupling (i.e., Friedel-Crafts alkylation) in reactions with phenol nucleophiles. The different outcome reflects the different reaction conditions employed to support phenol displacement of the halide leaving group: basic conditions (chloride) versus acidic (fluoride). A complementary set of nucleophilic substitution reactions were carried out with 26, using thiophenol and cyclic secondary amines coupling partners.

Figure 5.

Benzylic chloride displacement. Reactions were run on a 0.2 mmol scale. ^aReaction run with DMF. ^bReaction run with MeCN.

Overall, the reactions presented herein establish a new Cu/NFSI-based method for benzylic C─H chlorination, with KCl as the source of chloride. The •NSI HAT species contributes to very high benzylic site selectivity, relative to a broad cross section of other C─H chlorination methods. The ability to pair C─H chlorination in sequence with nucleophilic substitution or cross-coupling steps provide a strategy to achieve rapid diversification of core structures bearing benzylic C─H bonds.