Palladium-Catalyzed Atom-Transfer Radical Cyclization at Remote Unactivated C(sp³)–H Sites: Hydrogen-Atom Transfer of Hybrid Vinyl Palladium Radical Intermediates

Abstract

A novel mild, visible-light-induced palladium-catalyzed hydrogen atom translocation/atom-transfer radical cyclization (HAT/ATRC) cascade has been developed. This protocol involves a 1,5-HAT process of previously unknown hybrid vinyl palladium radical intermediates, thus leading to iodomethyl carbo- and heterocyclic structures.

A rad transfer

A novel mild, visible-light-induced palladium-catalyzed hydrogen atom translocation/atom-transfer radical cyclization (HAT/ATRC) cascade has been developed. This protocol involves a 1,5-HAT process of previously unknown hybrid vinyl palladium radical intermediates, thus leading to iodomethyl carbo- and heterocyclic structures.

Transformations involving translocation of radical intermediates by hydrogen-atom transfer (HAT) have received much attention over the past few years as a highly regioselective and mild approach toward functionalization of remote C(sp³)–H bonds.^[1] Recently, our group reported a HAT process triggered by a hybrid aryl palladium radical intermediate (Scheme 1a).^[2] Induced by visible light, these aryl palladium radical species underwent a 1,5-HAT step followed by the β-H elimination to generate silyl enol ethers in an efficient manner. Expanding this strategy by utilizing hybrid alkyl palladium radical species^[3] led to the development of a method for remote desaturation of alcohols at unactivated C(sp³)–H sites (Scheme 1b).^[4] Based on the ability of these hybrid palladium radical species to activate C–H bonds by a HAT process, we were eager to investigate the possibility of achieving a HAT process with vinyl hybrid palladium radical intermediates as a platform for development of new transformations. To the best of our knowledge, there have been no reports on the generation and reactivity of vinyl hybrid palladium radical intermediates.^[5] Herein, we report a mild, visible-light-induced exogenous photosensitizer free^[6,7] formation of novel hybrid vinyl palladium radical species which trigger a HAT process at unactivated C(sp³)–H sites^[8]/non-chain atom-transfer radical cyclization sequence to produce valuable iodoalkyl carbo- and heterocycles (Scheme 1c).

Scheme 1.

Hydrogen atom translocation of hybrid palladium radical intermediates.

Curran and Shen disclosed translocation of vinyl radicals, which, under standard tin hydride conditions, enabled reductive formation of methylcyclopentane cores by a sequential 1,5-HAT/5-exo-trig cyclization.^[9] Inspired by this work, we aimed at developing an oxidative version of this transformation toward methylenecyclopentane fragments [Eq. (1)]. The proposed sequence presumed formation of a hybrid vinyl palladium intermediate (1a→i), followed by its transposition (i→ii), cyclization (ii→iii), and a β-H elimination (iii→3). However, reaction of the benchmark substrate 1a under our previously employed conditions^[2] did not result in formation of the expected Heck type-product 3, instead, the iodomethyl dihydrobenzofuran 2a was produced selectively. Apparently, the latter is a product of an unprecedented HAT/ATRC-type cascade reaction involving activation of the C(sp³)–H bond.^[10,11]

(1)

Inspired by the discovery of a new HAT/ATRC reaction [Eq. (1)], and motivated by the importance of iodomethyl cyclopentane derivatives,^[12] we commenced investigation of this interesting transformation. First, optimization of the reaction conditions, employing 1a, was conducted (Table 1). It was found that this reaction proceeded more efficiently at higher dilution, thus producing 2 a in 51% yield (entry 3). Employment of sterically demanding tertiary amines (Cy₂NMe) was equally efficient when compared to Cs₂CO₃ (entry 5). Switching to more bulky amines or other inorganic bases, however, played a detrimental role (see Table S2 in the Supporting Information). Different palladium sources exhibited minor effects on the outcome of this reaction, thus producing the products in nearly equal amounts (see Table S3). Extensive ligand screening proved DPEphos as the ligand of choice (Table 1, entry 6). Surprisingly, except for dtbdppf, neither the parent dppf (entry 7), nor any of its derivatives (see Table S4) produced appreciable amounts of the product.^[13] Commonly for radical chemistry,^[1] benzene proved to be the best among variety of solvents tested (see Table S5). Employment of degassed benzene (Table 1, entry 8), as well as lowering the amount of the base (entry 9), led to further improvement of the yield. Performing the reaction under standard free-radical conditions^[9] delivered no target product (see Table S6). Finally, control experiments indicated that both light and the catalyst are vital for this transformation (Table 1, entries 10–12).

Table 1.

Optimization of reaction conditions.^[a]

Entry	[M]	Base	Ligand	Yield [%][b]
1	0.02	Cs2CO3	dtbdppf	35
2	0.05	Cs2CO3	dtbdppf	5
3	0.01	Cs2CO3	dtbdppf	51
4	0.01	iPr2NEt	dtbdppf	42
5	0.01	Cy2NMe	dtbdppf	53
6	0.01	Cy2NMe	DPEphos	55
7	0.01	Cy2NMe	dppf	nr
8	0.01	Cy2NMe	DPEphos	64[c]
9	0.01	Cy2NMe	DPEphos	71,[c,d] (69[e])
10	0.01	Cy2NMe	DPEphos	0[f]
11	0.01	Cy2NMe	DPEphos	7[g]
12	0.01	Cy2NMe	DPEphos	decomp[h]

^[a]All reactions were performed on a 0.05 mmol scale.

^[b]GC/MS yields.

^[c]Degassed benzene was used.

^[d]1 equivalent of Cy₂NMe was used.

^[e]Yield of isolated product.

^[f]In dark.

^[g]No Pd(OAc)₂ was used.

^[h]90°C, no light.

With the optimized reaction conditions in hand, the scope of this transformation was explored (Table 2). Reaction of electronically diverse iodovinylbenzyl derivatives (1a–c) provided the corresponding iodomethyl dehydrobenzofurans (2a–c) in good yield. Scaling up this reaction to 1 mmol provided comparable yield of 2a. Replacing the isopropyl group with a cyclohexyl group led to the spirocyclic derivative 2d in reasonable yield. Next, we were eager to probe this methodology on linear aliphatic systems toward assembly of an important cyclopentyl core. Gratifyingly, acyclic precursors proved to be competent reactants leading to the formation of a variety of iodomethyl cyclopentyl derivatives. Substrates, possessing various ester substituents at the carbon tether were nearly as equally efficient (2e–h). Notably, the substrate 1i, bearing a pendant chlorine atom, smoothly underwent reaction to produce the dihalogenated 2i in 57% yield, albeit as the mixture of diastereomers. Markedly, substrates possessing cycloalkyl groups reacted smoothly, thus producing 5/5- (2j) and 5/6- (2k) spirocyclic molecules in 64 and 41% yield, respectively. Importantly, employment of substrates, possessing a heterocyclic substituent [tetrahydrofuryl- (1l) or tetrahydropyranyl- (1m)] with a heteroatom adjacent to the C–H abstraction site reacted diastereoselectively, thus producing the iodomethyl-containing spiroheterocycles 2l,m in good yields. Reaction of substrates, possessing 4-susbtituted tetrahydropyran and piperidine led to efficient formation of the spiroheterocyclic molecules 2 n,o. Finally, a cascade reaction of the γ-butyrolactone-containing substrate led to a valuable spirolactone core^[14] (2p) in 47% yield.

Table 2.

Scope of HAT/ATRC cascade reaction.^[a]

^[a]Reaction conditions: 1 0.1 mmol, Pd(OAc)₂ 0.01 mmol, DPEphos 0.02 mmol, Cy₂NMe 0.1 mmol, benzene 0.01 M, 34 W blue LED.

^[b]1 mmol scale reaction.

^[c]Yield determined by NMR spectroscopy.

^[d]1.1:1 d.r.

^[e]5.3 (trans):1 d.r.

^[f]Single diastereomer (trans).

^[g]1.2:1 d.r.

Apparently, the presence of an iodomethyl moiety in the products of the HAT/ATRC cascade provides a convenient handle for further modification. Indeed, nucleophilic substitution of the iodide in 2o with cyanide and azide groups resulted in 4 and 5, respectively, in good yields (Scheme 2). Alternatively, base-mediated elimination of HI provided the exo-methylene-containing spiroheterocycle 6 in 71% yield.

Scheme 2.

Further transformations of HAT/ATRC cascade products. Boc = tert-butoxycarbonyl, DBU = 1,8-diazabicyclo[5-4-0]undec-7-ene, DMF = N,N-dimethylformamide.

We propose the following mechanism for this novel HAT/ATRC cascade transformation (Scheme 3). The active photo-excited^[15] Pd⁰ complex undergoes a single-electron transfer (SET) with 1 to produce, upon homolysis of the C–I bond, the hybrid vinyl palladium radical intermediate 7. The latter, upon 1,5-HAT generates the tertiary alkyl hybrid palladium radical 8, which upon 5-exo-trig cyclization forms the primary alkyl radical species 9. A subsequent iodine-atom transfer from the putative Pd^II species to 9 furnishes the product 2 and regenerates the Pd⁰ catalyst.^[16] The radical nature of this transformation was confirmed by radical trap and deuterium-labeling experiments.^[17] The UV-vis analysis indicated that the Pd⁰ complex is the photoabsorbing species. Alternatively, a radical-chain mechanism could be operative,^[18] where the palladium catalyst and light would only be needed for the initiation step of the reaction (1→9). Then, an atom-transfer chain reaction (AT-I) between 9 and 1 (9 + 1 → 2 + 7) would propagate the process. However, because of the unfavorable bond dissociation energies (BDEs) of the C–I bonds in vinyl versus alkyl iodides,^[19] this scenario is unlikely.

Scheme 3.

Proposed mechanism for HAT/ATRC cascade.

In summary, we have shown that exposure of vinyl iodides to a palladium catalyst and visible light in the absence of exogenous photosensitizers at room temperature leads to formation of a novel vinyl hybrid palladium radical species. This intermediate triggers a novel 1,5-HAT/non-chain ATRC cascade reaction to form valuable iodomethyl-containing cyclopentanes, as well as spiroannulated cyclopentanes with carbo- and heterocycles. The iodomethyl functionality in the formed reaction products can easily be further functionalized. It is believed that the discovery of novel reactivity of vinyl halides under visible-light/palladium-catalyzed conditions would trigger the elaboration of new methods, and that the developed methods would find applications in synthesis.