Photoinduced Formation of Hybrid Aryl Pd-Radical Species Capable of 1,5-HAT: Selective Catalytic Oxidation of Silyl Ethers into Silyl Enol Ethers

Abstract

A direct visible light-induced generation of a hybrid aryl Pd-radical species from aryl iodide and Pd(0) is reported to enable an unprecedented (for hybrid Pd-radical species) hydrogen atom-transfer event. This approach allowed for efficient desaturation of readily available silyl ethers into synthetically valuable silyl enols. Moreover, this oxidation reaction proceeds at room temperature without the aid of exogenous photosensitizers or oxidants.

Aryl halides are vital starting materials for a variety of important Pd(0)-catalyzed C–C and C–heteroatom bond-forming reactions, which usually involve Pd(II) intermediates A, formed via a two-electron oxidative addition process (Scheme 1a).¹ Alternatively, if a hybrid aryl Pd-radical species B, possessing both traditional Pd and radical nature, can be realized, it may exhibit novel reactivity.² Thus, if conditions for a direct conversion of easily available aryl halides into B are found, it could trigger the development of an array of novel transformations. The work described herein features: (1) the first direct visible light-induced generation of a hybrid aryl Pd-radical species (i) from aryl iodide and Pd(0); (2) capability of the formed hybrid aryl Pd-radical species to undergo a 1,5-hydrogen atom-transfer (HAT) process (i → ii); (3) a new method for general, efficient, and mild desaturation of silyl ethers (1) into silyl enol ethers (2) (Scheme 1b).

Scheme 1.

Traditional and Proposed Novel Reactivity of Aryl Halides in Pd(0)-Catalyzed Reactions

In his pioneering work, Curran reported a remote C–H functionalization of alcohols via a 1,5-HAT process utilizing a halo-aryl silane tether (1′) (eq 1).^3a Standard radical initiation of 1′ led to the aryl radical iii, which underwent 1,5-HAT to produce alkyl radical iv capable of undergoing further useful free radical reactions.³ We thought of developing an oxidative hybrid Pd-radical version of Curran’s free radical chemistry as a potentially useful new method. Indeed, if a hybrid aryl Pd-radical complex i, capable of HAT and a subsequent β-hydride elimination, could be generated (Scheme 1b), it would allow for a direct oxidation of silyl ethers (1) into silyl enols (2).^4,5 Clearly, the success of this process hinges on the efficient generation of the hybrid aryl Pd-radical species (i) from aryl halides and capability of the latter to undergo HAT, both of which, to the best of our knowledge, are unprecedented.^6,7 We envisioned that putative species i could possibly be generated either via thermal^6,8 or photochemical^9,10 activation of an aryl iodide in the presence of Pd(0) complex.

(1)

To this end, we first tested silyl-tethered alcohol 1a under various thermal Pd-catalyzed conditions that have been developed for hybrid Pd-radical reactions of alkyl halides (Table 1).^2,6,11 However, employment of Pd(PPh₃)₄ or PEPPSI-IPr complexes resulted in hydro-dehalogenation¹² of 1a (entries 1, 2). The combination of Pd(OAc)₂ and bidentate ferrocene ligands was also incompetent (entries 3, 4). Evidently, other means of activation were needed to stimulate this process. We were intrigued by recent reports of beneficial effect of visible light on promoting radical-type transition-metal-catalyzed transformations.⁹ Thus, we examined the reaction under irradiation of visible light (blue LED). Excitingly, product 2a was obtained in 72% using Pd(PPh₃)₄ catalyst at room temperature (Table 1, entry 5). Remarkably, the reaction proceeded quite efficiently without the use of any exogenous photosensitizers.¹³ Next, bidentate phosphine ligand dppf with Pd(OAc)₂ was tried, although it was less efficient (entry 6). Delightfully, 1-diphenylphosphino-1′-(di-tert-butylphosphino)ferrocene (L) was found to be the best ligand, producing desaturated product 2a in 79% yield (entry 7). Notably, performing the reaction with a 23W CFL led to lower efficiency (entry 8). A control experiment indicated that the Pd-catalyst is crucial for this reaction (entry 9). It was also found that both dimethyl- and diisopropyl-substituted substrates were less efficient.¹⁴ Obviously, this novel desaturation protocol requires no external oxidants, which makes it milder compared to other desaturation methods.^4b,c,15

Table 1.

Optimization of the Reaction Conditions^a

entry	catalyst	ligand	conditions	yield, %b
1	Pd(PPh3)4	–	120 °C	0c
2	PEPPSI-IPr	–	120 °C	0c,d
3	Pd(OAc)2	dppf	120 °C	0c
4	Pd(OAc)2	L	120 °C	0c
5	Pd(PPh3)4	–	rt, blue LED	72
6	Pd(OAc)2	dppf	rt, blue LED	67
7	Pd(OAc)2	L	rt, blue LED	79 (79)e
8	Pd(OAc)2	L	rt, 23W CFL	49f
9	–	L	rt, blue LED	NR

^aStandard conditions: 1a 0.05 mmol scale, PhH 0.1 M, 120 °C or 34W blue LED.

^bNMR Yield.

^c14–35% of hydro-dehalogenation of 1a was obtained.

^d5 mol % of catalyst was used.

^eIsolated yield.

^f51% conversion of 1a.

With the optimized conditions in hand, the generality of this reaction was examined. Thus, cyclohexanol derivatives (1b–d) reacted smoothly to produce silyl enol ethers 2b–d in good yields. Heterocycles, possessing 4-hydroxyltetrahydropyran (1e), 4-chromanol (1f), and 4- hydroxylpiperidine (1g) motifs, afforded products 2e–g in good yields (Table 2). Cyclopentanol derivatives 1h–j were found to produce 2h–j in reasonable yields. Moreover, cyclobutanol (1k) and medium-sized cyclicalcohol derivatives (1l–m) were found to be competent substrates as well, producing the corresponding products 2k–m in moderate to good yields. Next, we tested this methodology on desaturation of acyclic ethers, which are known to react inefficiently in dehydrogenation reactions.¹⁶ Hence, desaturation of 2-propanol derivative 1n produced 2n efficiently. Likewise, substrates possessing bulky groups at the β-position, such as -t-Bu (1o) and -TMS (1p), provided 2o and 2p in 95% and 97% yields, respectively. The practicality of this method was supported by a scale-up experiment with 1.2 g of 1o, where the oxidation product 2o was produced in nearly quantitative yield. Desaturation of 4-substituted benzylic alcohol derivatives 1q–w containing different functionalities, including aryl chloride (1r), cyanide (1s), and unprotected 2° alcohol (1w), proceeded well, producing moderate to good yields of the corresponding silyl enol ethers. Moreover, a double-fold desaturation of diol derivative 1x led to the dienol ether 2x in good yield. This methodology also has proven to be efficient on challenging linear substrates, which are known to be unselective in HAT protocols.^3a,d,17 Thus, 4-heptanol 1y and 2-hexanol 1z, selectively generated their respective α–β desaturated products (2y,z) in good yields, although as a mixture of stereoisomers.¹⁸ Then, we investigated our technology toward functionalization of more complex molecules. Thus, terpenes 1aa and 1ab reacted well, though the former produced an unstable silyl enol (2aa) that resulted in 40% isolated yield along with 55% yield of desilylated ketone product (camphor). Oxidation of steroid androstanone 1ac, bearing multiple sites of desaturation, proceeded smoothly, producing 2ac in 75% yield as a sole regioisomer.

Table 2.

Photoinduced Pd-Cat. Oxidation of Silyl Ethers^a

^aStandard Conditions: 1 0.2 mmol, Pd(OAc)₂ 0.02 mmol, L 0.04 mmol, Cs₂CO₃ 0.4 mmol, PhH 0.1 M, 34W blue LED, rt. Isolated yields.

^b10:1 ratio, major isomer shown.

^c3 mmol scale = 99% isolated yield.

^dYield based on NMR ratios.

^eZ:E = 3:1.

^fZ:E:terminal =7:3:1.

^gYield based on isolation of desilylated product, camphor (55% isolated yield).

We foresee two distinct mechanistic scenarios for this novel oxidation reaction, a radical pathway (Path A) and a concerted metalation deprotonation (CMD) protocol (Path B) (Scheme 2). First, under visible light (blue LED), a direct SET^2,6,10,11 from excited Pd(0) complex to aryl iodide 1a produces a hybrid Pd-radical intermediate 3 (Path A). In an alternative, though, less likely scenario,¹⁹ the latter can form through a conventional oxidative addition of Pd(0) complex to 1a to form the Pd(II) intermediate 4, followed by its excitation into 5²⁰ and a subsequent homolysis.²¹ Then, 3 undergoes a 1,5-HAT of H_α to generate the hybrid alkyl Pd-radical complex 6, which subsequently produces silyl enol 2a either via recombination and successive β-hydride elimination of H_β (Path A1)¹¹ or via a direct H_β-atom elimination²² with Pd(I)I (Path A2). Potentially, 6 can also be converted to 2a either via oxidation by Pd(I)I into a Scheme 2. Possible Mechanisms cationic intermediate 8, followed by its deprotonation (Path A3), or via an atom-transfer/HI-elimination sequence (Path A4).^23,24 In the alternative pathway (Path B), intermediate 4 undergoes a CMD process^16a,25 of H_α (4 → 10) to produce palladacycle 11. The following β-hydride (H_β) elimination of the latter occurs to generate the aryl Pd(II) silyl enol intermediate 12. A subsequent reductive elimination yields 2a′ and regenerates the Pd(0) catalyst. However, the performed deuterium-labeling experiments disproved this mechanism (Path B).¹⁴

Scheme 2.

Possible Mechanisms

The intermediacy of radical species (Path A) was supported by the use of radical scavengers¹¹ and radical clock experiments.^14,26 Hence, it was found that employment of radical traps, such as galvinyloxy and TEMPO, completely suppressed the reaction.¹⁴ It was also found that cyclopropyl-containing substrate 13 under standard reaction conditions underwent smooth ring opening of the cyclopropyl ring producing 16 as the major product (Scheme 3). Notably, formation of product 14 with preserved cyclopropyl unit and the β–C elimination product²⁷ 15 were not detected.

Scheme 3.

Radical Clock Experiment

In conclusion, we demonstrated the first direct generation of a hybrid aryl Pd-radical species from aryl halides and Pd(0) and its capability of promoting 1,5-HAT process, which allowed for a general mild and efficient conversion of easily available silyl ethers into valuable silyl enol ethers. Notably, this photoinduced Pd-catalyzed reaction proceeds at room temperature without any exogenous photosensitizers or oxidants. We envision that this methodology can be applied for a late-stage desaturation of complex molecules.