Visible light-promoted ring-opening functionalization of unstrained cycloalkanols via inert C-C bond scission.

Described herein is a novel, useful, visible light-promoted ring-opening functionalization of unstrained cycloalkanols. Upon scission of an inert cyclic C-C σ-bond, a set of medium- and large-sized rings are readily brominated under mild reaction conditions to afford the corresponding distal bromo-substituted alkyl ketones that are hard to synthesize otherwise. The products are versatile building blocks, which are easily converted to other valuable molecules in one-step operation. This protocol is also applicable to the unprecedented ring-opening cyanation and alkynylation of unstrained cycloalkanols.

Introduction

C–C bond activation is always a challenging issue in synthetic chemistry. Over the past few decades, this area has received great attention relying on the cleavage of cyclic C–C bonds.1 Strained cycloalkanols such as cyclopropanols and cyclobutanols have emerged as privileged precursors for the preparation of β- and γ-substituted ketones through radical-mediated ring-opening functionalization.2,3 Recently, we have consecutively reported the ring-opening functionalization of cyclobutanols to construct various chemical bonds, e.g., C–F, C–Cl, C–N, C–S, C–C, etc., based on silver or manganese catalysis.4 However, the opening of unstrained rings (in particular, 5–7 and 12–15 membered rings) is always confronted with a formidable challenge. It is mainly attributed to the significantly decreased ring-strain energy (Scheme 1A).5 During our research, it was also found that the intramolecular dehydration took place in competition with a ring-opening pathway to suppress the reaction outcome. Therefore, an efficient catalytic protocol should be sought for the ring opening of unstrained cycloalkanols.

Scheme 1

Ring-opening functionalization of unstrained cycloalkanols.

Photoredox catalysis has proven to be a powerful tool for the mild generation of an alkyloxy radical that triggers the subsequent ring-opening reactions or other transformations.6 Recently, Knowles et al. applied intramolecular proton-coupled electron transfer (PCET) to the photocatalytic ring opening of unstrained cycloalkanols.7 In this remarkable process, however, electron-rich tertiary alcohols were required to generate the arene radical cation, a key intermediate for the intramolecular PCET process. This somewhat limited the practicality of the protocol (Scheme 1B). Later, Zuo et al. disclosed an elegant photocatalytic ring-opening amination of unstrained cycloalkanols using a cerium chloride complex (Scheme 1C).8 Beyond this, other types of ring-opening functionalization of unstrained cycloalkanols are still anticipated.

Alkyl bromides are versatile building blocks in synthetic chemistry, which can be easily converted to other valuable molecules through nucleophilic substitution or cross coupling reactions. We considered that the ring-opening bromination of unstrained cycloalkanols would give rise to distally bromo-substituted alkyl ketones which are synthetically valuable but hard to synthesize otherwise. Herein, we provide support for this hypothesis (Scheme 1D). A set of medium- and large-sized rings are readily brominated through the mild cleavage of an inert cyclic C–C σ-bond with the assistance of visible-light irradiation. The newly formed C–Br bonds can function as privileged precursors for many other useful chemical bonds. Moreover, this protocol is also applicable to the unprecedented ring-opening cyanation and alkynylation of unstrained cycloalkanols.

Results and discussion

At the outset, we implemented the reaction parameter survey with the less strained 1-phenylcyclopentanol 1a as a model substrate and NBS as a bromine source (Table 1). It was found that the use of biphasic solvents was crucial to the reaction. In the presence of a photoredox catalyst (PC) and a hypervalent iodine reagent, the ring-opening bromination readily proceeded under visible-light irradiation, affording the desired δ-bromo alkyl ketone 2a. A brief evaluation of solvents indicated that CCl4/H2O delivered the best yield (entry 4), while the use of PhCF3/H2O also resulted in a comparable yield (entry 9). The yield was slightly improved to 76% by using PC 2 instead of PC 1 (entry 13). Other than PIDA, hypervalent iodine reagents such as BI-OH, IBX, and DMP were also suitable for the reaction, but led to lower yields (entries 16–18). The amount of H2O was important to the reaction outcome. Increasing the volume of H2O in biphasic solvents decreased the yield (entry 19), whereas using anhydrous CCl4 only led to trace amounts of product 2a (entry 20). Both PC and PIDA were not indispensable; removing either one still afforded the product in modest yields (entries 21 and 22). It might be suggested that the overall ring-opening process was the outcome of two different pathways. The reaction did not proceed in the absence of both PC and PIDA or in the dark (entries 23 and 24). Finally, the yield of 2a was further improved to 78% with the use of PhCF3/H2O as the mixed solvent (entry 25).

Table 1

Reaction parameter survey

Entry	Solvent (v/v 20/1)	Photoredox catalyst	Hypervalent iodine	Yield (%) b
1	DCM/H2O	PC 1	PIDA	22
2	DCE/H2O	PC 1	PIDA	21
3	CHCl3/H2O	PC 1	PIDA	15
4	CCl4/H2O	PC 1	PIDA	74
5	DMF/H2O	PC 1	PIDA	0
6	DMSO/H2O	PC 1	PIDA	0
7	Acetone/H2O	PC 1	PIDA	<5
8	Toluene/H2O	PC 1	PIDA	<5
9	PhCF3/H2O	PC 1	PIDA	68
10	CH3CN/H2O	PC 1	PIDA	35
11	THF/H2O	PC 1	PIDA	0
12	MeOH/H2O	PC 1	PIDA	0
13	CCl4/H2O	PC 2	PIDA	76
14	CCl4/H2O	PC 3	PIDA	59
15	CCl4/H2O	PC 4	PIDA	24
16	CCl4/H2O	PC 2	BI-OH	70
17	CCl4/H2O	PC 2	IBX	55
18	CCl4/H2O	PC 2	DMP	65
19 c	CCl4/H2O	PC 2	PIDA	70
20	CCl4	PC 2	PIDA	<5
21	CCl4/H2O	PC 2	—	53
22	CCl4/H2O	—	PIDA	45
23	CCl4/H2O	—	—	0
24 d	CCl4/H2O	PC 2	PIDA	0
25	PhCF3/H2O	PC 2	PIDA	78

1a (0.2 mmol), NBS (0.3 mmol, 1.5 equiv.), hypervalent iodine (0.4 mmol, 2.0 equiv.), and PC (0.006 mmol, 3 mol%) in mixed solvents (2.0 mL/0.1 mL) at rt, 14 W blue LED irradiation.

Yields of isolated products.

CCl4/H2O (2.0 mL/0.2 mL).

In the dark.

With the optimized reaction conditions in hand, we set out to investigate the generality of the protocol (Scheme 2). First, a variety of cyclopentanols were tested. Generally, electron-deficient aryl substituents were preferred for the reaction, leading to δ-bromo alkyl ketones in synthetically useful yields. Halides (2b–2d), in particular bromides (2d), were well tolerated, providing a platform for the late-stage product manipulation. Positional change (para-, meta-, and ortho-) of substituents did not impede the transformation (2g–2i). Although highly electron-rich aryl compounds such as anisole were not suitable due to the electrophilic bromination of the aryl group, it could be overcome by mounting an additional electron-withdrawing group (2h). This method also allowed for the preparation of secondary bromides (2j). The ring opening of unsymmetric cycloalkanols proceeded in a regioselective manner. However, the synthesis of tertiary bromides failed under the current reaction conditions. The scope of cyclohexanols was investigated more extensively than that of cyclopentanols. Besides the functional groups already examined, other groups such as cyano (2q and 2t) and ester (2z) were also compatible to the reaction conditions. In addition to aryl-substituted cyclohexanols, alkyl-substituted cyclohexanols were apt to furnish the elusive distal-bromo dialkyl ketones (2w and 2x). Various substituents could be incorporated into the alkyl chain by using the functionalized cycloalkanol precursors (2y–2ab). In the presence of excess NBS, three-fold bromination took place on the toluene-substituted cycloalkanol to give product 2ac which was brominated at both benzylic and distal positions. The reaction with the cyclohexanol derived from 1,4-cyclohexandione afforded the unsaturated 1,4-diketone via dehydrobromination (2ad). Likewise, the ring-opening bromination of cycloheptanols bearing various functional groups also readily proceeded, affording the corresponding distal-bromo alkyl ketones in useful yields (2ae–2ak). Unfortunately, the ring opening of eight- and ten-membered cycloalkanols failed, leading to complex mixtures. The reason is unclear so far. Remarkably, the opening of large-sized rings such as cyclododecanols and cyclopentadecanols occurred efficiently, yielding a variety of distally brominated alkyl ketones that are hard to synthesize otherwise (2al–2ar).

Scheme 2

Scope of unstrained cycloalkanols. Reaction conditions: 1 (0.2 mmol), NBS (0.3 mmol, 1.5 equiv.), PIDA (0.4 mmol, 2.0 equiv.), and PC (0.006 mmol, 3 mol%) in mixed solvents (2.0 mL/0.1 mL) at rt, 14 W blue LED irradiation. Yields of isolated products are given. With condition b. [Ir(dF(CF3)ppy)2(dtbbpy)]PF6 as PC. NBS (0.8 mmol, 4.0 equiv.).

To manifest the utility of the protocol, the products were converted to other valuable molecules via nucleophilic substitution or cross-coupling in one or two steps (Scheme 3). First, the bromide in 2a was easily replaced by azide and iodide, forming new C–N3 and C–I bonds in good yields (3 and 4).9 Then, the ketone in 2a was reduced to alcohol that intramolecularly attacked the alkyl bromide, leading to tetrahydropyran 5 in two steps with high yield.10 Alternatively, the C–Br bond was readily converted to C–B and C–C bonds via transition-metal catalyzed cross-coupling reactions, affording synthetically useful building blocks or molecules (6 and 7).11

Scheme 3

Transformation of products to other useful molecules.

Haloperidol is a marketed antipsychotic drug extensively used to treat schizophrenia and Tourette's syndrome. After gaining a portfolio of distal-bromo alkyl bromides as precursors, we accomplished the preparation of haloperidol analogues in high yields (8a–8d, Scheme 4).12 Testing of their biological activities is ongoing.

Scheme 4

Production of haloperidol analogues.

According to the experimental results, the plausible mechanism is depicted (Scheme 5, top). Although the detailed process for the generation of cycloalkoxy radical I is not fully understood, the radical I might be formed via two pathways. First, Stern–Volmer studies disclose that the excited state of IrIII catalyst PC 2 could be oxidatively quenched by NBS to generate the IrIV complex (see the ESI†). However, the oxidation potential of this IrIV complex (E1/2IV/III = 1.21 V in MeCN vs. SCE) is insufficient to oxidize cycloalkanol 1 (Ep/2 > 2.0 V in MeCN vs. SCE) to alkoxy radical Ivia the SET process (for the cyclic voltammetry studies, see the ESI†). Thus, we postulate an intermolecular proton-coupled electron transfer (PCET) process for the generation of alkoxy radical I in the presence of an IrIV complex and weak base such as a succinimide anion (path a).13 Alternatively, homolysis of the O–I bond in situ formed by the reaction of cycloalkanol 1 with PIDA might also lead to the alkoxy radical I (path b).14 In either way, the visible-light irradiation is indispensable. The subsequent β-C scission of I leads to the ring-opened alkyl radical II, which is intercepted by NBS to give the final product 2. Notably, CBr4 is also a competent bromine source in lieu of NBS to afford the same brominated product, suggesting the formation of intermediate II during the reaction (Scheme 5, bottom).

Scheme 5

Plausible pathways of ring-opening bromination.

This protocol is also applicable to the challenging ring-opening cyanation and alkynylation of unstrained cycloalkanols.15 Under the similar reaction conditions, 1-phenyl cyclododecanol was readily converted to the distally cyano- or alkynyl-substituted alkyl ketones in synthetically useful yields (9 and 10), providing a non-trivial approach to construct remote C–C bonds (Scheme 6).

Scheme 6

Ring-opening cyanation and alkynylation of unstrained cycloalkanols.

Conclusions

We have described a novel and efficient visible light-enabled ring-opening functionalization of unstrained cycloalkanols via the mild cleavage of cyclic C–C σ-bonds. Ring opening of a variety of cyclopentanols, cyclohexanols, cycloheptanols, cyclododecanols, and cyclopentadecanols readily proceeded to furnish the distally brominated alkyl ketones that often function as precursors of complex molecules in organic and medicinal synthesis. As shown above, this method provides important feedstocks for the production of haloperidol analogues. The newly formed C–Br bonds in products are easily transformed into other valuable chemical bonds, thus illustrating the utility of this method. This protocol is also applicable to the elusive ring-opening cyanation and alkynylation of unstrained cycloalkanols, providing an unusual strategy for the construction of remote C–C bonds.

Conflicts of interest

There are no conflicts to declare.

Click here for additional data file.