Preferential geminal bis-silylation of 3,4-benzothiophane is caused by the dominance of electron withdrawal by R3Si over steric shielding effects.

Benzylic C-H lithiation of 3,4-benzothiophane and subsequent treatment with triphenyl- or trimethylchlorosilane under a variety of conditions leads to α,α- rather than α,α'-bis-silylation products as a consequence of anion stabilization by R3Si and very fast deprotonation of the intermediate monosilylated product, even with a sterically bulky base such as lithium diisopropylamide.

The origin of the research reported herein was the question of whether chiral cyclic sulfides of general structure 1 (or enantiomers) might be useful for catalytic enantioselective methylene transfer to C=C or C=O from methylene precursors such as Et2Zn/ICH2Cl or Zn/CH2I2, following a process in which 1 serves as a carrier to provide chiral ylide 2.[1] Initial studies were directed toward the synthesis of α,α′-bis-silanes such as 1, R = Me3Si, Ph3Si, and t-BuMe2Si, from the readily available starting material 3 (Figure 1).[2] It seemed reasonable that rapid access to the series of silyl ligands 1 might be secured simply by sequential, one-flask silylation of monoanions derived from 3 or by one-step bis-silylation of the α,α′-dianion of 3, a 12-π-electron (non-Hückel stabilized) system. Surprisingly, however, access to the desired 2,5-bis-silylated 3,4-benzothiophane structure proved to be difficult because the behavior of anions desired from 3 was strongly influenced by the electronic and steric effects of silyl substituents.

Figure 1

Some 3,4-benzothiophanes.

Reaction of 3 in THF at −78 °C[3] with 1.0 equiv of n-BuLi (2.5 M in hexanes) for 3 h followed by slow addition of Ph3SiCl (in THF at −78 °C) afforded as major product (along with unreacted 3) the crystalline gem-bis-silane 4 in 31% yield.[3] The structure of 4 was established by X-ray crystallographic analysis (Figure 2).[4]

Figure 2

X-ray crystallographic structure of 4.

The formation of 4 likely occurs via the C(2)-monosilylation product 5 by further deprotonation at C(2) (either by monolithiated 3 or residual BuLi) followed by a second silylation (Scheme 1). The putative intermediate 5 could be obtained by the gradual addition of a cold (−78 °C) solution of lithium diisopropyl amide (LDA) in THF to a mixture of 3 and 1.0 equiv of Ph3SiCl in THF at −78 °C, which gave 5 in 42% yield along with recovered 3 (40%).

Scheme 1

Facile Geminal Silylation

Consistent with these observations was the finding that the addition of 1.0 equiv of LDA in THF (at −78 °C) to a solution of 5 and Ph3SiCl in THF at −78 °C afforded 4 (76%) as major product, without the formation of a detectable amount of the α, α′-silylated product 1, R = Ph (Scheme 2). Evidently even with the use of the very bulky base LDA, the acidifying effect of an C(2) Ph3Si group is sufficiently great to override its steric shielding of the geminal C(2)-H proton. An α-triphenylsilyl group has been found to increase the acidity of the C(9)-proton of fluorene by about 104 in previous work.[5,6] Despite extensive experimentation, no conditions could be found to convert 3 to the α,α′-bis-silylated derivative 1, R = Ph.

Scheme 2

Preferential Geminal Triphenylsilylation

The same preference for the formation of gem-bis-silylated product was demonstrated using Me3SiCl as the silylation reagent. Thus, the addition of 3 to a mixture of 2.0 equiv of LDA and 6.0 equiv of TMSCl at −78 °C produced the gem-bis-silylation product 7 in 64% yield with no measurable amount of the isomeric α,α′-bis-TMS compound (Figure 3).

Figure 3

Possible position isomers for disilyation.

At this point, it should be mentioned that all of the above silylation reactions must be carried out at low temperatures because the intermediate 6 is unstable above −60 °C. There is considerable precedent for decomposition via the cycloelimination pathway shown in Scheme 3.[7]

Scheme 3

Possible Pathway for Anion Decomposition

1H, 6Li, and 13C NMR studies were carried out at −78 °C to ascertain the nature of the lithiated intermediate (6) generated from 3 under various lithium–hydrogen exchange conditions. Reaction of 3 with n-butyllithium in THF at −78 °C occurred with a half-life (t1/2) of greater than 1 h and gave only the monolithiated form, 6 (Scheme 1).[4] The 1H, 6Li, and 13C NMR spectra on samples generated using [6Li] n-BuLi showed asymmetry attesting to monolithiation, and the appearance of the benzylic CH2 moiety as an A–B quartet showed that the 6Li counterion differentiated the two faces. Although we observed no evidence to indicate aggregation, it cannot be rigorously excluded either. No other lithiated species could be observed even using more than 2 equiv of n-BuLi at −78 °C for 12 h, and quenching after 12 h with MeOD led to only monodeuterated 3.[4]

Although the reaction of monolithium intermediate 6 with triphenylchlorosilane occurs readily at −78 °C to form 5, NMR studies showed that proton transfer from 5 to 6 is even faster to form the silyl stabilized lithio-anion 9. Further, NMR measurements also revealed that the reaction of 9 with Ph3SiCl (Scheme 4) is slower than expected and requires several hours at −78 °C. The slower rate of the silylation 9 to 4 appears to be a consequence of the anion-stabilizing effect (electron withdrawal) of Ph3Si combined with the steric shielding by Ph3Si in both 9 and Ph3SiCl.

Scheme 4

NMR Rate Studies

The research described above has revealed significant nuances in the behavior of lithiation products derived from 3,4-benzothophane involving surprising rates of C–H lithiation and C–Li silylation.