Sulfamate Esters Guide Selective Radical-Mediated Chlorination of Aliphatic C–H Bonds

Abstract

Masked alcohols are particularly appealing as directing groups because of the ubiquity of hydroxyl groups in organic small molecules. Herein, we disclose a general strategy for aliphatic γ-C(sp³)–H functionalization guided by a masked alcohol. Specifically, we determine that sulfamate ester-derived nitrogen-centered radicals mediate 1,6-hydrogen-atom transfer (HAT) processes to guide γ-C(sp³)–H chlorination. This reaction proceeds through a light-initiated radical chain-propagation process and is capable of installing chlorine atoms at primary, secondary, and tertiary centers.

Graphical abstract

Guided Chlorination: Aliphatic γ-C(sp³)–H chlorination is directed by a sulfamate-ester masked alcohol. This reaction involves a light-initiated N–Cl bond homolysis, followed by an unusual radical-mediated 1,6-hydrogen-atom abstraction with subsequent chlorination enabled by a chain-propagation process. Through this process, chlorine atoms can be selectively installed at primary, secondary, and tertiary centers with predictable selectivity.

Late-stage alkyl C–H functionalization^[1] facilitates investigations into pharmaceutical small molecule leads.^[2] Unfortunately, position-selective reactivity is difficult to achieve at unactivated C(sp³)–H centers, which have high bond dissociation energies (90–105 kcal/mol), are not acidic (pK_a ≥ 50), and do not integrate more polarizable and electronically accessible π-orbitals. Consequently, in most efficient atom-transfer processes, site-selective activation originates at an electron-rich, sterically accessible bond (Scheme 1A),^[1b,d,3] or is controlled kinetically by a directing group (Scheme 1B). A hydroxyl group is attractive as a directing moiety because alcohols are common within readily available small molecules. To our knowledge, there are no general strategies for aliphatic γ-C–H functionalization guided by a masked alcohol.^[4,5] Herein described are the first experiments to demonstrate that sulfamate esters^[6,7] can mediate 1,6-hydrogen-atom transfer (HAT) processes to guide the light-promoted chlorination of unactivatedγ-C(sp³)–H centers.

Scheme 1. C(sp³)–H chlorination reactions.

Our strategy for γ-functionalization is inspired by the modern resurgence of interest^[8,9] in the Hofmann-Löffler-Freytag (HLF)^[10–12] reaction. In traditional HLF processes, position selectivity arises based on formation of an intermediate nitrogen-centered radical that engages in a 1,5-HAT process^[13] through a kinetically favored, six-membered transition state.^[1,4] By contrast, we anticipate sulfamate esters may be capable of mediating 1,6-HAT processes, which are rare^[1,5] as they are expected to proceed through seven-membered transition states, which are often kinetically disfavored.^[16] We hypothesize that sulfamate esters enable 1,6-HAT because their elongated O–S and S–N bonds (∼1.58 Å) and compressed O–S–N bond angles (∼103°)^[1,7] geometrically favor a seven-membered ring transition state for C–H abstraction. The site-selectivity available through these reactions complements that achieved using traditional HLF chlorination transformations,^[18] alternative guided chlorination methods,^[19] or site-selective intermolecular^[20] chlorination processes.

This approach is demonstrated in the course of site-selective chlorine-transfer to convert N-chlorinated sulfamate esters 1 to alkyl chlorides 2 (Scheme 1B). Predictable control has proven more challenging in C(sp³)–H chlorination^[21] than bromination reactions,^[12d,22,23] owing to the promiscuity of chlorine radical promoted hydrogen-atom abstraction. Nevertheless, alkyl chlorides are useful synthetic intermediates, and components of bioactive small molecules, with >2000 known chlorine-containing natural products.^[24] Consequently, new technologies for selective aliphatic C–H chlorination have potential to streamline syntheses of bioactive small molecules.^[25]

To minimize mechanistic uncertainty, we chose to generate nitrogen-centered radicals by using light to homolyze nitrogen–chlorine bonds of N-chlorosulfamate esters 1. N-chlorosulfamate esters 1 are readily available from sulfamate esters 3^[26] upon treatment with trichloroisocyanuric acid or tert-butyl hypochlorite (see supporting information for details).As anticipated, photolysis of pentyl methylchlorosulfamate ester 1a results in selective γ-chlorination at a methylene center (Table 1, entry 1, C–H bond dissociation energy (BDE) ∼ 98 kcal mol^–1).^[27] This reaction does not occur in the absence of light, indicating that photolysis initiates the reaction.

Table 1. Chlorine-transfer robust to N-substitution.

entrya	R		Substrate for chlorine-transfer (1)	Reaction time (min)	Product (2)	Yield (%)c
1	Me	12.65	1a	45	2a	85
2d	tBu	12.74	1b	15	2b	98
3e	tBu		1b	60	2b	83
4	CH2CF3	9.18	1c	15	2c	96
5e	CH2CF3		1c	60	2c	86
6f	CH2CF3		1c	90	2c	70
7		11.05	1d	45	2d	66
generates complex mixture of products

^aGeneral reaction conditions: 1.0 equiv N-chlorosulfamate ester 1, PhH, 22 °C, irradiated with a blue Kessil lamp.

^bAqueous ionization pK_a values have been predicted using SPARC.^[28]

^cIsolated yield

^dTrace amount of δ-chlorinated product 4b (not depicted) detected in crude ¹H NMR. See Supporting Information for details.

^eReaction run in PhCF₃.

^fReaction run in PhCl.

Chlorine-transfer proceeds from N-methyl-, N-tert-butyl-, N-trifluoroethyl- and N-2-(pyridin-2-yl)isopropylchlorosulfamate esters 1a–d (Table 1, entries 1–7). The corresponding series of sulfamate esters 3 are predicted to differ in terms of N–H bond acidity^[28] by more than three pK_a units, indicating that the reaction tolerates broad electronic variations across N-alkyl substituents. In all of these cases, the remaining mass balance can be accounted for principally by reduced 3. Notably, in the reaction of substrate 1b, <2% yield of δ-chlorinated product 4b (not depicted) can be detected in the crude reaction mixture along with desired γ-chlorinated 2b.While N-alkylsulfamate esters facilitate chlorination, some N-arylsulfamate esters may not engage in productive guided chlorine-transfer processes, as photochemical irradiation of arylated 1e generates a complex mixture of products.

In these chlorine-transfer processes, site-selectivity is orthogonal to that available through most atom-transfer processes reliant on innate selectivity. In the productive transformations of N-chlorinated 1a–d, chlorinated alkanes 2a–d are formed with exquisite γ-selectivity, despite the fact that δ-functionalization is expected to be favored under conditions that rely on inherent selectivity (Table 1; Scheme 1A).

Guided selectivity is preserved across a range of O-alkyl sulfamate esters (Table 2). For menthol-derived 1f, C(5)–H and C(8)–H bonds are expected to have very similar electron densities.^[2,9] Nevertheless, unguided tertiary-selective intermolecular oxidation^[29] and amination^[30] processes result in selective functionalization of analogues at C(5), presumably because the C(8)–H bond is sterically encumbered.^[29] By contrast, under the described conditions, chlorination occurs exclusively at C(8)–H, which is poised to interact with the sulfamate ester nitrogen-centered radical (entry 1) via a seven-membered transition state. This position-selectivity is analogous to that displayed in iron- and manganese-nitrene-mediated intramolecular amination reactions of sulfamate esters.^[31,32]

Table 2. Chlorine-transfer robust to O-alkyl substitution.

entrya	substrate	product	time (min)	yield (%)b
1			15	2f, 91
2			15	2g, 85
3			15	2h, 96
4			15	2i, 54
5			240	2j, 82
6			30	2k, 87
7			90	2l, 92
8			120	2m, 19
9			15	2n, 36

^aGeneral reaction conditions: 1.0 equiv N-chlorosulfamate ester 1, PhH, 22 °C, irradiated with a blue Kessil lamp.

^bIsolated yield.

Furthermore, the directed chlorination reaction overcomes the innate site-selectivity that arises from inductive deactivation in unguided oxidation processes. In unguided C–H functionalization reactions, proximity to inductively electron-withdrawing groups deactivates C–H bonds to oxidation. This phenomenon is evident when 3,7-dimethyloctanol is masked with an electron withdrawing group, and engages in undirected fluorination,^[33] oxygenation,^[34] amination,^[35] azidation,^[3,6] or trifluoromethylthiolation^[37] processes that functionalize C(7) in preference to C(3). By contrast, the sulfamate ester-guided chlorination installs chlorine at the electronically deactivated C(3) position (entries 2–3).

Moreover, this guided functionalization process overcomes selectivity that typically arises from differences in bond strength. Pendant silyl ether 1i undergoes efficient and selective γ-chlorination, without oxidation of the weak etherial C–H bonds (BDE ∼ 85 kcal mol⁻¹, entry 4).^[27]

In addition to differentially masked diols, the reaction conditions are compatible with a variety of functional groups. Sulfamate esters derived from N-protected amino alcohols engage in selective atom-transfer, as phthalimidyl 1j provides chlorinated 2j in 82% isolated yield (entry 5). As noted previously, the reaction also allows pendant heteroaromatic moieties (Table 1, entry 7) and fluorides (Table 1, entries 4–6; Table 2, entries 1, 3–5, 9).

This method transforms C(sp³)–H bonds at secondary, tertiary (BDE ∼ 96 kcal mol^–1),^[27] and benzyliccenters (BDE ∼ 90 kcal mol^–1)^[27] in synthetically useful yields (entries 1–8). Generally, chlorination of γ-C(sp³)–H centers occurs in preference to β-C–H centers, even when the γ-C–H bond is primary and significantly stronger (BDE ∼ 101 kcal mol⁻¹)^[27] than a secondary β-C–H bond. For example, chlorination of propyl tert-butylsulfamate ester 1m and (–)-borneol derivative 1n occur at the stronger primary C–H bonds, albeit with diminished efficiency (entries 8–9). While each of these substrates displays a geometrically accessible β-methylene center, chlorination at these positions is not detected. Instead, dehalogenated sulfamate esters 3m and 3n are the primary by products of these transformations. An analogous reactivity trend has been documented with White's manganese-catalyzed intramolecularamination with sulfamate esters.^[32]

In principle, this chlorination reaction could proceed through a closed cycle and/or a radical chain propagation mechanism (Scheme 2). In either pathway, initiation would occur by light-promoted N–Cl bond homolysis, thereby converting N-chlorosulfamate ester 1b to chlorine radical and sulfamyl radical 5b. The resulting nitrogen-centered radical 5b mediates an intramolecular 1,6-HAT to generate a carbon-centered radical 6b with exquisite position selectivity. Subsequent divergence in carbon–chlorine bond forming events then distinguishes between the proposed pathways. In the closed radical cycle mechanism, intermediate carbon-centered radical 6b recombines with the chlorine radical to terminate the reaction and provide 2b (not depicted). Alternatively, carbon-centered radical 6b could engage in chlorine-atom abstraction with another equivalent of N-chlorosulfamate ester 1b (Scheme 2). This sequence would release desired halogenated 2b, along with another equivalent of radicalloid 5b, which could propagate this chain reaction.

Scheme 2. Chlorination proceeds through light-initiated chain-propagation.

These mechanistic hypotheses differ in terms of the number of product equivalents that can be generated per absorbed photon, a relationship that defines quantum yield (Φ). While a closed radical cycle could furnish a maximum of one product molecule per absorbed photon (Φ ≤ 1), a radical chain propagation mechanism could provide multiple equivalents of the product per absorbed photon (Φ > 1).

Quantum yield measurements have been performed to provide insight into the operative reaction mechanism,^[3,8] and provide evidence that the reaction proceeds through a light-initiated chain propagation mechanism. Briefly, standard chemical actinometry using potassium ferrioxalate allowed us to determine the photon flux of a fluorimeter at 313 nm.^[38,39] After 15 minutes of irradiation of N-chlorinated 1b in benzene at 313 nm in the calibrated fluorimeter, 68% conversion to chloroalkane 2b is observed. This yield corresponds to 77 equivalents of product formed per absorbed photon (Φ = 77), indicating that this reaction proceeds through a chain propagation mechanism.

This sulfamate ester-guided HLF reaction is expected to provide a powerful and general platform to complement current HLF technologies. This research is among the first to establish that sulfamate esters can mediate 1,6-HAT such that the generated carbon-centered radicals can be trapped efficiently in guided intermolecular reactions. Furthermore, the method provides efficient access to secondary and tertiary alkyl chlorides, a valuable class of synthetic intermediates, with novel and predictable site-selectivity.