Pentacoordinate silicon compounds anionic species

Although anionic, pentacoordinate silicon species have received attention, the isolation of siliconates 194, 103, 195, 196 (Figure 28) by Perozzi, Martin and colleagues (155) has provided an opportunity to study in detail both structure and stereomutation of a pentacoordinate silicon compound (155, 321). This point will be discussed further in Sect. V-C. The crystal structure of 195 shows that the silicon is pentacoordinated with the fluorine and the two carbon atoms forming the equatorial plane of a somewhat distorted trigonal bipyramid (321). It is essentially the same as that proposed for similar siliconates on the basis of NMR data obtained in the solution phase (155). 

Pentacoordinate Silicon Compounds with SiOs Skeletons 24.2.1 Anionic Species with SiOs Skeletons... 

In 1982, Sakurai [7] described a catalytic version of this reaction (Scheme 13.4). The addition of small quantities of fluoride anions to the allylsilane 1 generates the pentacoordinated silicon species 10, probably in equilibrium with the starting materials 1 and 11. This activated species can react with the carbonyl derivative 6 to yield the alkoxide 12 which is trapped by fluorotrimethylsilane. This last step not only furnishes the silylated compounds 13 but also regenerates the fluoride catalyst 11. Acidic work-up then leads to the desired homoallylic alcohol 7. 

The most intriguing results are those from the reaction promoted by fluoride. The reaction of fluoride with an allylmetal reagent is thought to proceed through either an allyl anion or an allyl fluorosiliconate intermediate [24]. Allyl anions [25] and pentacoordinate silicon species [26] have been proposed as intermediates in the fluoride-induced allylation of a carbonyl compound. The results obtained with the deuterium model rule out the intermediacy of a free allyl anion since the ratio 14/15 is different for the syn compared to the anti Sp/ pathways. 

Hypervalent silicon compounds have found wide utility in organic synthesis. In general, pentacoordinated anionic silicates are more reactive toward nucleophiles than are tetracoordinated silanes. For example, Mes2SiF2 is unreactive toward water, while (the 18-crown-6 potassium salt of) Mes2SiF3 is completely hydrolyzed within minutes. Similarly, the pentacoordinate anion HSi(OEt)4 is an effective reducing agent for aldehydes, ketones, and esters at or below room temperature (Scheme 2) no such reaction occurs with HSi(OEt)3. The difference in relative reactivities of hypervalent and nonhypervalent species is relevant to the intermediates proposed in Section 7.6. 

Reacting la,b with MeLi in HMPA as active solvent and in the presence of MeOH as trapping agent, the attack of a methyl anion at the silicon atom of the silacyclobutane is the first reaction step and gives a pentacoordinated silicon anion (Scheme 2). Such five-coordinated species are discussed as intermediates during the ring opening polymerization of silacyclobutenes, -butanes, and -pentenes [2]. Furthermore, five-coordinated silicon species are well described to be stable compounds [3]. 

Alternatively, unreactive mixtures of organosilicon hydrides and carbonyl compounds react by hydride transfer from the silicon center to the carbon center when certain nucleophilic species with a high affinity for silicon are added to the mixture.76 94 This outcome likely results from the formation of valence-expanded, pentacoordinate hydrosilanide anion reaction intermediates that have stronger hydride-donating capabilities than their tetravalent precursors (Eq. 6).22,95 101... 

The difficulty in interpreting these results, at the time, arose from the lack of reliable points of comparison for the expected Sn chemical shifts of stannylium ions. There was no doubt that values higher than 300 were deshielded to an unprecedented extent, but were these shifts sufficient to demonstrate tricoordination or trivalency Computation in the early 1990s could not provide a reliable answer. Arshadi et al bypassed the calculational problem in 1996 by publishing a remarkable empirical correlation between structurally analogous silicon and tin compounds ( Si chemical shift vs Sn chemical shift). Since reliable calculations were available for the Si chemical shifts of trialkylsilylium ions, this plot could provide at least an indication of the expected Sn chemical shifts of trialkylstannylium ions, which proved to be 5 ca. 1700. The species observed by Birchall, Lambert, and Sakurai thus were very far from the expected chemical shift and hence from the ideal tricoordinate geometry of the stannylium ion. Since the values are deshielded to some extent, pentacoordination could be ruled out. The best description of the structures observed by all these authors, therefore, is the bond-stretched, solvent-coordinated stannyl cation 8. Lambert and Kuhlmann observed high conductivity, so that the neutral anion-coordinated variant 9 could be eliminated. Such structures (8) also apply to those reported in 1992 by Edlund et al as the tetrahedral part of the equilibrium with pentacoordinate species. 

---