A Silicon Analogue of Vinyllithium: Structural Characterization of a Disilenide^†

Abstract

The reduction of Tip₂SiCl₂ by finely dispersed lithium powder directly affords the disilenide 1 (see picture), the structure of which could be characterized by X‐ray crystallography. The reaction of 1 with Me₃SiCl leads to the corresponding asymmetrical disilene. The ready availability of 1 is expected to open new possibilities for the synthesis of disilenes in general. Tip=2,4,6‐iPr₃C₆H₂

Organosilyl anions are highly useful reagents in organic and inorganic synthesis. Although the structure and reactivity of lithiated derivatives have been studied in detail,¹ information concerning unsaturated anionic silicon compounds is still scarce. In view of the rapid development of the chemistry of disilenes² the availability of anionic derivatives would be highly desirable. In particular, the recently growing interest in conjugated homonuclear π systems of silicon justifies the need for new synthetic methods.

In this context cyclotetrasilenide 1, a silicon analogue of the allyl anion, has been recently described.³ In 1997 Weidenbruch et al. proposed that reduction of disilene 2 by lithium metal leads to disilenide 3, which reasonably explained the formation of tetrasilabutadiene 4 upon treatment of 3 with mesitylbromide (Scheme 1).⁴ However, 3 could not be isolated nor characterized.⁵ The present work reports the isolation by means of a simplified synthetic protocol and the structural characterization of disilenide 3.

Scheme 1

R=Si(Me)tBu₂; Tip=2,4,6‐iPr₃C₆H₂; Mes=2,4,6‐Me₃C₆H₂.

Several approaches to the preparation of disilene 2 were reported,⁶ two of which employed reductive conditions. The original procedure by Watanabe et al. made use of the reductive dimerization of Tip₂SiCl₂ with a stoichiometric amount of lithium naphthalide.^6a In light of the observations made by Weidenbruch et al.,⁴ the assumption that the disilenide 3 could be accessible by direct reaction of Tip₂SiCl₂ with the required amount of reduction equivalents seemed warranted.

Indeed the reduction of Tip₂SiCl₂⁷ with an excess of finely dispersed lithium powder in 1,2‐dimethoxy ethane (DME) at 25 °C and subsequent crystallization from hexane leads to a uniform orange product in acceptable yield (51 %). The spectroscopic data of the thermally surprisingly stable compound (m.p. 121 °C) agree with the constitution of the disilenide 3 (Scheme 2).

Scheme 2

The UV/Vis absorption of 3 in hexane at 417 nm is bathochromically shifted compared with that of 2 at 266 nm. Such a shift is typical for disilenes that bear electropositive substituents, for example, the highest wavelength absorptions of tetra(silyl)disilenes are found around 400 nm.⁸

The ²⁹Si NMR spectrum shows two resonances at δ=100.5 and 94.5 ppm, both shifted downfield significantly with respect to that of 2 (δ=53.4 ppm)^6a Vinyllithium exhibits a similar deshielding effect only for the α carbon atom (Δδ(C_α)=+60.8, Δδ(C_β)=+9.9 ppm).⁹ The unusually strong effect of asymmetrical substitution on the shifts of the ²⁹Si NMR spectrum of disilenes has been discussed recently in a quantum‐chemical study.¹⁰ The broadening of the signal at 100.5 ppm, which shows only a cross signal to one of the arene protons in the ²⁹Si–¹H COSY spectrum of 3, might arise from coupling of the Si atom to the coordinated ⁷Li nucleus. However, even at −80 °C in toluene no quartet splitting of the signal was observed.

The structure of the disilenide 3 in the solid state as determined by X‐ray diffraction analysis¹¹ (Figure 1) shows the presence of a close‐contact ion pair. The Si1–Li bond length of 285.3(3) pm, however, is long in comparison with those usually observed in silyllithium compounds, for example, 267 pm in Ph₃SiLi⋅3 THF.¹²

Figure 1

Structure of 3⋅2 DME in the solid state at 100 K. Protons are omitted for clarity. Thermal ellipsoids are set at 70 % probability. Selected distances [pm] and angles [°]: Si1‐Si2 219.2(1), Si1‐Li 285.3(3), Si1‐C1 193.1(1), Si2‐C16 191.5(1), Si2‐C31 190.5(1), Si2‐Si1‐Li 131.7(1) , C1‐Si1‐Li 119.0(1), C1‐Si1‐Si2 107.6(1), C16‐Si2‐Si1 140.3(1), C31‐Si2‐Si1 111.0(1), C31‐Si2‐C16 108.2(1), C31‐Si2‐Si1 111.0(1), C31‐Si2‐C16 108.2(1), C1‐Si1‐Si2‐C31 174.9(1), Li‐Si1‐Si2‐C31 10.4(1), Si1‐Si2‐C16‐C17 −140.5(1), Si1‐Si2‐C31‐C32 −109.5(1), Si2‐Si1‐C1‐C2 70.1(1).

The core of the disilenide 3 is visibly distorted from that of the planar uncharged disilene 2. The Si1Si2 bond of 219.2(1) pm is significantly longer (2: 214.4 pm),^6a despite the obvious relieve of steric strain. Ab initio calculations by Apeloig and Karni predicted a slight elongation of the SiSi bond in the unsubstituted disilenide (3 u) compared to H₂SiSiH₂.^13a While both silicon atoms reside nearly in one plane with the ipso‐carbon atoms C1, C16, and C31 (largest deviation: 6.6(1) pm for Si2), the distance of the lithium cation to this plane amounts to 54.7(3) pm. The angle formed by C1‐Si1‐Si2 (107.6(1)°) is noticeably smaller than the corresponding angles in 2 (120.8°, 121.6°). Conversely the Si1‐Si2‐C16 angle is strongly widened (140.3(1)°). Density functional theory (DFT) calculations by Schäfer III and co‐workers on 3 u without the countercation even found an Si‐Si‐H angle of 90° at the negatively charged silicon atom, thus indicating the avoidance of hybridization at the central Si atom.^13b

The structure of the solvent separated ion pair of 3 is therefore of considerable interest. In fact, addition of [2.2.1]cryptand to a THF solution of 3 and subsequent layering with hexane afforded red crystals, which, unlike those of 3, were totally insoluble in hexane. The X‐ray diffraction analysis left no doubt about the constitution as a solvent‐separated ion pair of 3. However, because of heavy disorder in the disilenide moiety the structure could not be refined satisfactorily. Attempts to improve the quality of crystals are underway.

The availability of pure 3 should be very useful for the synthesis of low‐valent silicon compounds, for example, asymmetrically substituted disilenes, the preparation of which has so far been limited to only a couple of examples.¹⁴ The reaction of 3 with Me₃SiCl in toluene indeed affords the asymmetric disilene 5 (Scheme 2), whose structure is derived from spectroscopic data: the ²⁹Si NMR spectrum shows three resonances at δ=97.7, 50.9, and −8.3 ppm, which were unambiguously assigned on the grounds of the very well‐resolved ²⁹Si satellite signals.¹⁵ The signal at δ=50.9 ppm exhibits two sets of satellites and is therefore attributed to the central sp² Si atom. The coupling constant of the signal at δ=97.7 ppm (¹J(Si,Si)=116.3 Hz) is more than 1.5 times larger than that of the signal at −8.3 ppm (¹J(Si,Si)=75.1 Hz). ¹J(SiSi) coupling constants for tetra(aryl)disilenes were found to be in the range of 155 to 158 Hz and have been interpreted in terms of higher s character of the SiSi bond compared to single bonds.^16b Hence, the s character of the double bond in 5 is considerably lowered by the electropositive silyl substituent.

Reactions of the disilenide 3 with a variety of other electrophiles are currently under investigation.

Experimental Section

3: Thoroughly dried and degassed DME (40 mL) at 25 °C was added to a mixture of Tip₂SiCl₂ (7.50 g, 14.8 mmol) and finely dispersed lithium powder (0.38 g, 55 mmol). The reaction mixture instantly became intensely red and warm (ca. 40 °C). After it had cooled to ambient temperature stirring was maintained for 3 h. Excess lithium and precipitated salts were removed by filtration off and washed with 2×10 mL of DME. The combined filtrates were reduced to dryness under vacuum and hexane (80 mL) was added to the residue. This mixture was then heated to about 60 °C, thus dissolving most of the orange precipitate; the remainder was removed by filtration while hot. The solvent of the resulting red solution was removed under vacuum until about 20 mL remained. The copious amounts of an orange solid that precipitated were redissolved by heating to 60 °C and the solution was left at ambient temperature overnight. Large orange crystals were collected by decantation of the mother liquor and dried under vacuum: 3.25 g (51 %) 3 (mp 120–121 °C). ¹H NMR (300 MHz, [D₈]toluene, 25 °C, TMS): δ=7.09, 7.08, 7.04 (each s, each 2 H, m‐H), 4.76 (sept., 4 H, Me₂CH), 4.21 (br, 2 H, Me₂CH), 2.99 (s, 12 H, DME), 2.95 (s, 8 H, DME), 2.86 (m, 3 H, Me₂CH), 1.46, 1.39, 1.35, 1.29, 1.23, 1.19, 1.02 ppm (each d, overall 54 H, Me₂CH); ⁷Li NMR (97 MHz, [D₈]toluene, 25 °C, Li⁺ aq): δ=0.20 ppm; ¹³C NMR (75 MHz, [D₈]toluene, 25 °C, TMS): δ=155.5 (s), 154.1 (s), 149.5 (s), 148.6 (s), 147.9 (s), 145.9 (s), 145.2 (s), 142.6 (s), 121.8 (br d), 121.2 (d), 120.7 (d), 37.2 (d), 36.4 (d), 36.0 (br d), 35.7 (d), 35.5 (d), 35.3 (d), 27.4 (br q), 26.1 (br q), 25.5 (q), 25.4 (q), 25.2 (q), 25.0 ppm (q); ²⁹Si NMR (50 MHz, [D₈]toluene, 25 °C, TMS): δ=100.5 (s, 1Si, SiLi), 94.5 ppm (s, 1Si, SiC₂); UV/Vis (hexane): λ_max(ε)=417.2 (760) nm; IR (neat, 25 °C)=2953 (s), 1456 (m), 1379 (w), 1358 (w), 1306 (w), 1244 (w), 1192 (w), 1162 (w), 1123 (m), 1082 (s), 1029 (m), 937 (w), 869 (s), 839 (m), 752 (m), 646 cm⁻¹ (m).

5: Freshly distilled Me₃SiCl (0.1 mL, 85 mg, 0.78 mmol) was added by syringe at 25 °C to 3 (0.51 g, 0.597 mmol) in dry and degassed toluene (5 mL). After the reaction had been stirred for 30 min, all volatile products were removed under vacuum. Hexane (10 mL) was added to the orange residue. The resulting solution was filtered and reduced to dryness affording 5 (0.42 g, 95 %) as an orange oil. ¹H NMR (300 MHz, C₆D₆, 25 °C): δ=7.15, 7.09, 7.01 (each s, each 2 H, m‐H), 4.25, 4.07, 3.83 (each sept., each 2 H, Me₂CH), 2.76, 2.75, 2.66 (each sept., each 1 H, Me₂CH), 1.32 (d, 12 H, Me₂CH), 1.28 (br., 12 H, Me₂CH), 1.19, 1.18, 1.09 (each d, each 6 H, Me₂CH), 0.96 (d, 12 H, Me₂CH), 0.27 ppm (s, 9 H, SiMe₃); ¹³C NMR (75 MHz, C₆D₆, 25 °C): δ=155.9 (s), 155.1 (s), 154.8 (s), 151.1 (s), 150.5 (s), 150.2 (s), 135.9 (s), 135.5 (s), 131.7 (s), 122.3 (d), 122.1 (d), 121.7 (d), 37.8 (d), 36.8 (d), 34.8 (d), 34.7 (d), 34.4 (d), 25.2 (br), 24.6 (q), 24.2 (q), 24.1 (q), 23.9 (q), 3.0 ppm (q); ²⁹Si NMR (50 MHz, C₆D₆, 25 °C): δ=97.7 (s, 1Si, SiC₂), 50.9 (s,. 1Si, SiSiMe₃), −8.3 ppm (s, 1Si, SiMe₃).

Dedicated to Professor Manfred Weidenbruch