Copper-catalyzed silylation of p-quinone methides: new entry to dibenzylic silanes.

An efficient and general copper(I)-catalyzed silylation of p-quinone-methides is described. Non-symmetric dibenzylic silanes are obtained in high yields under mild reaction conditions. These compounds can be used as bench-stable benzylic carbanion precursors.

para-Quinone methides are reactive intermediates formed by a cyclohexadiene moiety in para-conjugation with a carbonyl group and an exo-methylene component. They are neutral entities with a zwitterionic resonance structure that enhances the electrophilic character at the δ-position.[1] Carbanions, aromatic rings, alcohols and amines are typical nucleophiles that quickly react with p-quinone methides to afford a variety of diaryl derivatives (Scheme 1).[2] The use of transition metals could allow the formation of C–C and C–X bonds complementary to those formed by direct addition of a typical nucleophile. Surprisingly, the use of metal-catalyzed transformations to functionalize the exocyclic double bond of p-quinone methides remains largely unexplored.[3]

Scheme 1

p-Quinone methides as diaryl derivatives precursors.

We became interested in probing this approach using a silyl copper(i) complex as a formal silicon nucleophile (Scheme 1). To the best of our knowledge, the addition of nucleophilic silicon species to ortho- or para-quinone methides has not been studied to date. Silicon-containing molecules are valuable synthetic intermediates which can be converted into useful compounds through a number of transformations.[4] Recently, copper-catalyzed silylation reactions have emerged as a powerful tool for C–Si bond formation.[5] We envisioned that insertion of the exocyclic double bond into the Cu–Si bond followed by aromatization would afford non-symmetric benzylic silanes.

The most common way to synthesize benzylic silanes involves the reaction between an in situ generated benzylic carbanion and a silyl chloride.[6] Our method would offer a milder alternative to this classic approach, avoiding the use of stoichiometric amounts of strong bases. Herein, we describe a copper(i)-catalyzed protocol for the silylation–aromatization of p-quinone-methides. The reactions proceed in high yields using only 10% of an inexpensive copper(i) salt and a commercially available silaborane reagent.[7]

We started our study with p-quinone methide 1a, containing removable t-Bu groups at the α-positions (Table 1).[8] A series of ligands were screened (Table 1, entries 1–6) using Cu(CH3CN)4PF6 (10 mol%), Me2PhSiBpin (1.1 equiv.), NaOt-Bu (0.2 equiv.) and MeOH (4 equiv.). We found that NHC ligands (entries 5 and 6) were superior to monodentate or bidentate phosphines (entries 1–4). SIMes gave the best results, affording dibenzylic silane 2a with 86% isolated yield (entry 6, ≥98% conversion. The use of other bases (entries 7–9) or different copper salts (entries 10 and 11) gave poorer results. Lowering the catalyst loading to 5 mol% also resulted in a significantly lowered yield (entry 12). In the absence of MeOH (entry 13) or with only two equivalents (entry 14) compound 2a was obtained in 46% and 57% yield respectively. Finally, to check the role of the NHC–Cu(i) catalyst we carried out the reaction in the absence of copper salt and ligand (entry 15). Under those conditions, a very complex mixture was observed in the 1H NMR spectrum of the crude product. From this mixture, we could identify the product of 1,6-addition of methoxide to 1a as the main compound, unreacted 1a, and a small amount of 2a. The formation of 2a under these conditions could be explained by alkoxide activation of the silaborane in the absence of the copper catalyst.[9]

Table 1

Optimization of the reaction conditions

Entry a	Copper salt	Base	Ligand	2a (%)
1	Cu(CH3CN)4PF6	NaOt-Bu	Ph3P	64 b
2	Cu(CH3CN)4PF6	NaOt-Bu	JohnPhos	64 b
3	Cu(CH3CN)4PF6	NaOt-Bu	Xantphos	50 b
4	Cu(CH3CN)4PF6	NaOt-Bu	(±)-BINAP	45 b
5	Cu(CH3CN)4PF6	NaOt-Bu	IMes	87 b
6	Cu(CH 3 CN) 4 PF 6	NaOt-Bu	SIMes	≥98 b (86) c
7	Cu(CH3CN)4PF6	KOt-Bu	SIMes	60 c
8	Cu(CH3CN)4PF6	LiOt-Bu	SIMes	68 c
9	Cu(CH3CN)4PF6	CsF	SIMes	61 c
10	CuCl	NaOt-Bu	SIMes	55 c
11	Cu2O	NaOt-Bu	SIMes	20 c
12 d	Cu(CH3CN)4PF6	NaOt-Bu	SIMes	35 c
13 e	Cu(CH3CN)4PF6	NaOt-Bu	SIMes	46 c
14 f	Cu(CH3CN)4PF6	NaOt-Bu	SIMes	57 c
15 g	—	NaOt-Bu	—	—

Reaction conditions: 1a (0.2 mmol), Me2PhSiBpin (0.22 mmol), base (20 mol%), Cu(CH3CN)4PF6 (10 mol%), ligand (11 mol%), MeOH (0.8 mmol), THF (0.1 M).

Conversion determined by 1H NMR analysis of the crude mixture.

Yield of isolated 2a.

Reaction conditions: 1a (0.2 mmol), Me2PhSiBpin (0.22 mmol), NaOt-Bu (20 mol%), Cu(CH3CN)4PF6 (5 mol%), ligand (11 mol%), MeOH (0.8 mmol), THF (0.1 M).

The reaction was carried out in the absence of MeOH.

0.4 mmol of MeOH were used.

Reaction conditions: 1a (0.2 mmol), Me2PhSiBpin (0.22 mmol), NaOt-Bu (20 mol%), MeOH (0.8 mmol), THF (0.1 M).

With these optimal conditions in hand, we proceeded to study the scope of the silylation–aromatization process (Table 2). We first modified the stereoelectronic properties of the exomethylene substituent (R3). Dibenzylic silanes with electron donating groups (compounds 2b–2c), heterocycles (compound 2d), and a larger naphthyl group (compound 2e) were prepared in high yields. The conditions also worked for p-quinone methides with electron withdrawing groups in para (compounds 2f, 2i, 2j), ortho (compound 2g) and meta (compound 2h) positions. It should be pointed out that our method allows for the synthesis of compounds with halogen substituents (2f, 2g) and an ester group (2i), which would be difficult to obtain by the reaction of a dibenzylic carbanion and a silyl chloride. Interestingly, monobenzylic silane 2k, in which R3 is an alkyl group, was also obtained using the optimized conditions.

Table 2

Copper(i)-catalyzed silylation of p-quinone methides ,

Reaction conditions: 1a (0.2 mmol), Me2PhSiBpin (0.22 mmol), NaOt-Bu (20 mol%), Cu(CH3CN)4PF6 (10 mol%), SIMes (11 mol%), MeOH (0.8 mmol), THF (0.1 M).

Yield of isolated 2.

Additionally, we modified the R1 and R2 substituents. Compounds 2l and 2m, with two methyl groups, and compound 2n, with two isopropyl groups, were obtained in good yields. It is also possible to introduce two different alkyl groups in the α-position (compound 2o) starting from a non-symmetrical p-quinone methide. Finally, the structure of compound 2g was confirmed by single crystal X-ray crystallography (Fig. 1).

Fig. 1

X-ray structure of compound 2g.

One interesting feature of benzylic silanes is their ability to be used as bench-stable benzylic anion equivalents under mild reaction conditions.[10] However, most known examples of these transformations have been performed with monobenzylic trimethylsilane derivatives. Therefore, our method provided an opportunity to check if dibenzylic dimethylphenyl silanes such as 2 could be also used as carbanion precursors. To the best of our knowledge, the generation of dibenzylic carbanions from silanes has not previously been reported. Gratifyingly, treatment of silane 2a with cesium fluoride in DMF, followed by addition of p-chloro benzaldehyde, provided the desired compound 3 as a 1 : 1 mixture of diastereomers (Scheme 2). Oxidation followed by removal of the t-butyl groups[8] using AlCl3 afforded α,α-diaryl ketone 4 in a good overall yield.

Scheme 2

Functionalization of the C–Si bond and de-tert-butylation.

A possible mechanism for the silylation–aromatization reaction of p-quinone methides is shown in Scheme 3. First, a silyl-Cu(i)–NHC complex B is formed by reaction of a copper alkoxide A and the silaborane reagent. Insertion of the exocyclic double bond of the p-quinone methide into the Cu–Si bond affords a π-allyl-copper intermediate (C) that could isomerize to copper phenoxide E. At this point two pathways are possible. Protonolysis in the presence of MeOH would provide silane 2 with release of NHC–CuOMe to restart the catalytic cycle. On the other hand, copper phenoxide E could react directly with the silaborane to provide 2 and silyl-copper complex B.

Scheme 3

Plausible mechanism for the silylation–aromatization.

In conclusion, we have found that copper(i) salts can catalyze the silylation–aromatization process of p-quinone methides. This study represents the first silicon addition to a quinone methide and provides new insight for the development of novel metal-catalyzed transformations. Mono- and dibenzylic silanes can be prepared in high yields under mild reaction conditions. We have also demonstrated that dibenzylic silanes can be used as stable dibenzylic carbanion equivalents. The development of asymmetric versions of this and related transformations is underway.

We thank the European Research Council (ERC-337776) and MINECO (CTQ2012-35957) for financial support. M. T. and A. P. thank MICINN for RyC and JdC contracts. We acknowledge Dr Josefina Perles for X-ray structure analysis.