MSc Thesis: Formation and Reactions of C=Si and C=Ge Double Bonds, Technion - Israel Institute of Technology, Haifa, Israel, 1994.

Abstract

Compounds with multiple bonds to silicon and germanium have been prepared as stable compounds only a decade ago and they are now of great contemporary interest. The growing attention to these compounds stems not only from the interest in their fundamental properties and chemistry, but also from their potential for producing new materials for various applications (novel polymers, as precursors for high-value ceramic materials). Yet, the arsenal of methods available for synthesizing silenes and germenes is very limited. The few known synthetic pathway provide only a limited range of compounds, and each of the known methods suffers from various disadvantages. One of the major goals of this work was to develop new methods for this synthesis.

The Peterson olefination reaction provides a useful method for the preparation of a variety of alkenes from alfa-silyl carbanions and carbonyl compounds (equation 1).

We wanted to study if a Peterson-type reaction using a silyl-substituted silyl anion instead of a carbanion may be used to synthesize silenes (equation 2).

Previous work in our group has shown very recently that in non-polar media the reaction of (Me3Si)3SiLi.3THF (1) with sveral ketones indeed proceeds via a Peterson-type reaction, providing a new route for generating silenes. For example, reaction of 2-adamantanone with an equimolar amount of (1) in hexane or benzene gave 85% yield of 1,2-disilacyclobutane 2 (equation 3). The observed product was interpreted in terms of the intermediacy of the silene 4. The silene 4 being kinetically unstable dimerizes in a head-to-head fashion to 2. Further evidence for the intermediacy of 4 was obtained by trapping experiments, where the corresponding Diels-Alder product 5 of the reaction of 4 with butadienes was formed (equation 3).

The enolizable ketone 4-tert-butylcyclohexanone gave at -78 oC the noncyclic dimer 6, that points to the intermediacy of silene 7 which gives 6 via a ene-reaction (equation 4).

These and other reactions have been studied in detail in this work. Our first goal was to explore the generality and the limitations of this new synthetic method for generating silenes and to explore if it can be extended also to germenes. We concentrated our attention to the following questions:

(1) what is the influence of the structure of the carbonyl compounds on the Peterson-type elimination?

(2) what is the effect of the substituents on the silyl anion on the reaction?

(3) what is the effect of the solvent polarity?

A second goal was to study the reactivity and the selectivity of the resulting silenes and germenes, which as stated above have different substitution patterns than those of the previously known silenes and germenes.

A. The preparation of a germene.

Reaction of tris(trimethylsilyl)germyl lithium (1') with 2-adamantanone in hexane leads to the formation of the thermally unstable 1,2-digermacyclobutane 8 (which structure was assigned on the basis of a X-ray analysis of a single crystal) via the the germene 9 (equation 5). Further evidence for the intermediacy of the germene 9 was provided by trapping experiments. Reaction of 1' with 2-adamantanone in the presence of 1,3-butadiene leads to the expected Diels-Alder product 10.

B. The role of the structure of the carbonyl compound.

In order to explore the generality of the Peterson reaction of tris(trimethylsilyl)-silyllithium (1) we have carried out reactions of 1 with various ketones and aldehydes. Reaction of 1 with ketone 11 in hexane gave 1,2-disilacyclobutane 12 in an 90% yield. The best explanation for the obtained product is the intermediacy of the silene 14 (equation 6).

Silenes 4 and 14 are not stable enough kinetically to allow their isolation. It was therefore of interest to increase the steric hindrance at either the carbon or the silicon end of the Si=C double bond in order to prepare kinetically stable and isolable silenes. Reaction of 1 with the sterically crowded di-tert-butyl ketone (15), 2,2,6,6-tetramethylcyclohexanone (16), fenchone (17), 2,2-dimethylcamphor (18), diadamantylketone (19) and 1-methyl-2-adamantanone (20) in hexane or benzene was studied. In all cases yellow-green solutions were obtained after addition of 1.

Surprisingly, hydrolysis produced the starting ketones and (Me3Si)3SiH. We conclude that with these ketones addition does not occur (too crowded) and that the coloured solutions contain some complexes probably of type 21 (equation 7).

The reaction of 1 with 4-tert-butylcyclohexanone was studied in some more detail. After aqueous work-up no traces of Me3SiOSiMe3 or of the dimer 6 were detected in the reaction mixture. This shows that spontaneous elimination did not occur in contrast to the analogous reaction with adamantanone (equation 3). The major product of the reaction carbinol 23 is stable at room temperature in hexane or benzene solution. Under heating or evacuation, 23 readily gives dimer 6. At the same time, lithium alkoxide 22, the obvious precursor of 23, does not produce silene 5 upon heating or removal of hexane (equation 8).

The very different stability of the alkoxides and carbinols towards elimination are discussed in terms of the steric repulsion and of aggregation.

The reactions of 1 with several aromatic carbonyl compounds, both enolizable and non-enolizable, were studied. Reaction of 1 with benzaldehyde in hexane at -78 ºC leads after hydrolysis to the carbinol 25 and also to products resulting from the corresponding anion-radical (e.g., diphenylethanediol). However, in this case heating in the presence of 1,3-butadiene gave in a low yield silacyclohexene 27, suggesting the intermediacy of the corresponding transient silene 26 (equation 9).

Reaction of 1 with acetophenone in hexane at -78 ºC gave after aqueous work-up the starting ketone due to the formation of the corresponding enolate, a process which could not be suppressed. Reaction of 1 with benzophenone in hexane at -78 ºC produced an intensive blue solution which results from the presence of the well known anion-radical [Ph2C=O]., as was shown by ESR spectroscopy. Reaction of 1 with tert-butyl-phenyl ketone and 1-adamantyl-tolyl ketone in hexane at -78 ºC gave complex reaction mixtures containing products of the reduction of these ketones by a SET-reaction and products resulting from nucleophilic addition of 1 to the carbonyl group.

C. The role of the solvent.

The reaction of 1 with 4-tert-butylcyclohexanone was studied in hexane, toluene and THF. The differences found in the reaction course between these solvents are extraordinary, revealing in each solvent completely different products.

In THF at -78 ºC, this reaction results (after aqueous work-up) in ca. 75% yield of the silyl ether 28. On the other hand, in hexane at -78 ºC the alcoholate 22 is obtained, which after hydrolysis gives the carbinol 23. At the same time, the lithium alcoholate 22, when isolated from hexane, does not provide 28 upon reflux in THF either in the presence of 15-crown-5 or without it, providing after aqueous work-up only carbinol 23. On the other hand, when carbinol 23 is treated with a catalytic amount of potassium hydride in hexane, it undergoes a facile rearrangement to give the silyl ether 28 (equation 10).

In contrast to the other solvents when the reaction between 1 and 4-tert-butyl-cyclohexanone was carried out in toluene, a mixture of the rearrangement product 28 and of the carbinol 23 in ca. 1 : 1 ratio was obtained.

We propose that the formation of 28 in THF in the reaction of 1 with 4-tert-butylcyclohexanone might proceed via a catalytic mechanism, where the acidic a-protons of the precursor ketone undertake the role of the proton donor. In order to confirm this suggestion, different experiments were carried out.

We propose that the stability of alcoholate 22 in hexane towards base-catalyzed rearrangement as well as towards the Peterson-type elimination might be explaned in terms of strong aggregation, which prevents further reactions.

D. The role of the substituents in the silyl anions.

We have also studied the effect of the substituents in the silyl anions on their reaction with 2-adamantanone and on the behavior of the alcohol products. We have synthesized the series of anions Rn(Me3Si)3-nSiLi (29' - 33') with combinations of one to three silyl and alkyl (aryl) substituents, following known in the literature procedures, and have studied their reactions with 2-adamantanone in hexane at room temperature.

Me(Me3Si)2SiLi (29') tBu(Me3Si)2SiLi (30') Ph(Me3Si)2SiLi (31') 2,4,6-iPr3C6H2(Me3Si)2SiLi (32') MePh(Me3Si)SiLi (33')

The reactions resulted in all cases in the addition producing after hydrolysis the corresponding carbinols 29 - 33 in good yields (e.g., 70-80%). No traces of the corresponding silene dimers or of Me3SiOSiMe3 were detected in the reaction mixtures.

The carbinols 29 - 33 could be isolated even in the neat. This contrasts with the behavior of 23, which was stable only in hexane solution and which spontaneously eliminates Me3SiOLi upon evaporation of the solvent.

Thus, replacing one Me3Si-group in 23 by an alkyl or an aryl substituent leads to the production of an isolable carbinol, which is stable towards spontaneous elimination. We conclude that electronic factors must be responsible for this dramatic effect of substitution of one Me3Si group by an alkyl or aryl substituent on the thermal stability of the carbinol.

E. Reactivity of the obtained silene.

Thermolysis of 2 in the presence of different reagents, which are known to react with silenes leads to trapping products of the silene 4. The trapping reactions which have been carried out with 4 are summarized in Scheme 1. The regiochemistry obtained in the reactions with 1-methoxy-1,3-butadiene, styrene and phenylacetylene was analized using FMO theory analysis.

In order to gain some information on the reactivity and selectivity of silene 4 we have carried out competition reactions of silene 4 with various reagents. The relative rates (or the competition ratios) were obtained. The reactivity and selectivity of silene 4 was compared with the other data known for silenes, and it was found to be less selective than e. g., (Me3Si)2C=SiAlk2.

Scheme 1

F. Reaction of (Me3Si)3SiLi . 3THF with tetramethyl-1,3-cyclobutanedione.

In contrast to the other ketones, the reaction of tetramethyl-1,3-cyclobutanedione (40) with one molar equivalent of 1 did not lead to the corresponding alcoholate or to the silene. Instead the ring opened product 41a was obtained and its X-ray structure was determined. Aqueous work-up of the reaction mixture gave the first known diacylsilane 42. Reaction of 41a with Me3SiCl leads, as expected, to the silyl enol ether 43a (Scheme 2). In order to check the generality of the ring opening reaction we have also carried out reactions of 40 with methyllithium, triethylsilyl- and triethylgermyl lithium (Scheme 2).

Scheme 2

The lithium enolates 41a, 41b and 41c exhibit two new interesting absorption maxima (compared to known acylsilanes), at ca. 330 nm and 410 nm. These new absorptions were explained using ab initio MO calculations for the model compounds.