Synthesis of Stable Dianionic Cyclic Silenolates and Germenolates

In this contribution a convenient synthetic method to obtain the previously unknown dianionic cyclic silenolates and germenolates is described. These dianions 2a,b and 4a,b are easily accessible via a one-pot synthetic protocol in high yields. Their structural properties were analyzed by a combination of NMR, single-crystal X-ray crystallography, and DFT quantum mechanical calculations. Moreover, the reactivity of 2a,b and 4a,b with selected examples of electrophiles was investigated. 2a and 4a were reacted with ClSiiPr3 to give new examples of polysilanes and polygermanes with exocyclic double bonds. The reaction of 2b with ClSiMe2SiMe2Cl led to the formation of the acyl bicyclo[2.2.2]octasilane 6. Moreover, the reaction of 2a,b and 4a,b with MeI, as an example of a carbon-centered electrophile, led to selective alkylation reactions at the negatively charged silicon and germanium atoms. The corresponding methylated structures 9a,b and 10a,b were formed in nearly quantitative yields. The competitive reactivity of the silyl and silenolate anion toward 1 equiv of ClSiMe3 showed that the outcome of the reaction was strongly influenced by the substituent at the carbonyl moiety. 2a reacted with 1 equiv of ClSiMe3 to give the corresponding cyclic silenolate S1a, which demonstrated that the silyl anion is more nucleophilic than the silenolate with attached aromatic groups. 2b, on the other hand, reacted with 1 equiv of ClSiMe3 to give the bicyclic compound 11via an intramolecular sila-Peterson alkenation reaction. These findings clearly showed that the alkyl-substituted silenolate is more nucleophilic than the silyl anion. This paper demonstrates that 2a,b and 4a,b have the potential to be used as unique building blocks for complex polysilane and polygermane frameworks.


■ INTRODUCTION
The synthesis of defined polysilanes in which more than five silicon atoms are connected is challenging. The standard approaches for such polysilanes are Wurtz-type coupling 1−3 or Lewis acid catalyzed rearrangement reactions. 4 These two methods generally give rise to structurally simple polysilanes with a low set of functionalities for further derivatization, which prevent the construction of molecules of even moderate complexity.
A potent strategy for the construction of structurally more challenging silicon frameworks is the use of di-or multifunctionalized starting materials such as α,ω-dianions. Gilman 5,6 and Hengge 7,8 were pioneers in this area and developed the cleavage of strained cyclosilanes to obtain dianions. Sekiguchi, 9 Tokitoh, 10 Kira, 11 and Apeloig 12 also contributed with their groups to this research field and prepared some previously unknown 1,1-, 1,2-, and 1,4-dilithiooligosilanes. Marschner and Baumgartner, who introduced KOtBu into the field of polysilane chemistry, achieved a milestone in polysilane synthesis. Consequently, the construction of relative complex polysilanes could be accomplished in a straightforward way. 13−16 Recently, Klausen and co-workers established new phenyl-substituted dianions. 17 These dianions were used as building blocks for the formation of defined polysilanes as well as for the synthesis of heteroelement substituted polysilanes. 18−20 Scheschkewitz et al. treated their hexasilabenzene with lithium naphthalenide and obtained a novel dianionic silicon cluster. This dianion turned out to be a valuable synthon for the generation of unprecedented molecular heterosiliconoids with boron and phosphorus directly incorporated into the cluster scaffold. 21 In addition, we just published a paper about the synthesis of a mixed substituted dianion, which allows straightforward access to a hitherto unknown tricyclic polysilane (see Chart 1). 22 Nevertheless, the synthesis of mixed functionalized disilanides has not been reported so far, although these substances would represent ideal building blocks for highly complex silicon frameworks.
As we have reported earlier, it is possible to synthesize and characterize cyclic silenolates as well as cyclic germenolates and convert them with suitable electrophiles in order to gain a new set of differently substituted acylsilanes as well as acylgermanes (see Chart 2).
Furthermore, we could show that the reaction of these enolates with chlorosilanes ClSiR 3 allowed straightforward access to silenes and germenes with exocyclic structures. 23,24 Due to the straightforward accessibility of these cyclic enolates, we saw the potential to investigate their chemical behavior in greater depth. In this context, we have established a novel synthetic strategy for the synthesis of previously unknown dianionic cyclic silenolates and germenolates. The aim of this work is to investigate the spectroscopic properties and the reactivity of these new types of dianions with selected examples of electrophiles.

■ RESULTS AND DISCUSSION
Synthesis of Dianionic Silenolates. The reaction of the acylcyclohexasilanes 1a,b with 2 equiv of KOtBu led to the formation of compounds 2a,b, whereby two different functionalized anionic silicon atoms were incorporated into one molecule (Scheme 1).
To the best of our knowledge, 2a,b represent the first examples of dianionic polysilanes bearing a silyl anion and a silenolate fragment in one molecule. The dianionic compounds 2a,b were formed in the same fashion as previously described for the corresponding silenolates. Two major differences are worth mentioning. First, the use of an appropriate solvent is highly important. We observed the formation of 2a,b only in DME, Et 2 O, and toluene. In the case of THF, no product was formed, probably due to the reaction of 2a,b with THF leading to degradation. This was also described in the case of α,ω-oligosilyl dianions by Marschner et al., 14 who observed that stable dianionic species were only formed with the use of DME or benzene/toluene with the addition of crown ethers. Second, the reaction is characterized by a two-step reaction sequence. The first 1 equiv of KOtBu is consumed immediately (approximately 10 min), yielding the silenolates S 1 a,b. The second abstraction of the trimethylsilyl group is much slower and takes place within approximately 18 h. For isolation, 2a,b were crystallized from Et 2 O/18-cr-6 at room temperature to give orange crystals of the 1:2 18-cr-6 adducts, which were obtained in isolated yields of >90%. After filtration, the crystals can be stored at −30°C in the absence of air even for prolonged periods. 2a,b afforded crystals of sufficient quality for single-crystal X-ray crystallography. The molecular structures are depicted in Figures 1 and 2; selected bond lengths and the sums of valence angles are summarized in Table 1.
On the basis of the observed structural features, 2a,b are best described as acyl silyl anions (keto form) with Si−C single bonds, CO double bonds, and markedly pyramidal central Si (1) atoms. Interestingly in 2a the K (2) + cation coordinates simultaneously to Si (1) and Si (6) . This is probably caused by a packing phenomenon. Furthermore, this simultaneous coordination is also the reason for 2a to adopt the half-boat conformation, while 2b and 4b (4b is the dianionic germenolate and will be introduced in the next section) Organometallics pubs.acs.org/Organometallics Article adopt chair conformations. Additionally 2a shows short Si (2) − CH contacts which are less than the van der Waals radii of silicon and hydrogen. This can also explain its half-boat coordination. A similar result in terms of Si−CH contacts was obtained by the Klausen group. 17 Synthesis of Dianionic Germenolates. The straightforward synthesis of 2a,b encouraged us to expand our new methodology to other starting materials. As we have reported previously, it is possible to synthesize cyclic acylgermanes 3a,b. 24 The reaction of these cyclic acylgermanes with 2 equiv of KOtBu led to the formation of dianionic germenolates 4a,b (Scheme 2). Again, the reaction is characterized by a two-step reaction sequence with reaction rates similar to those of the corresponding acylsilanes.
For isolation, 4a,b were crystallized from Et 2 O/18-cr-6 at room temperature to give orange crystals of the 1:2 18-cr-6 adducts, which can be stored after filtration at −30°C in the absence of air even for prolonged periods. 4b afforded crystals of sufficient quality for single-crystal X-ray crystallography. The molecular structure is depicted in Figure 3; selected bond lengths and the sums of valence angles are summarized in Table 2.
On the basis of the observed structural features, 4b is best described as an acyl germyl anion (keto form) with a Ge−C single bond, a CO double bond, and markedly pyramidal central Ge (1) and Ge (2) atoms.
In order to examine the differences between aromatic and saturated substituents at the carbonyl moiety, the mesityl-and adamantyl-substituted derivatives 2a,b and 4a,b were investigated. All UV−vis calculations were performed on the geometry-optimized X-ray crystal structures via TDDFT-PCM in toluene at the CAM-B3LYP/6-31+G(d,p) level of theory. 26 Noteworthy, CAM-B3LYP achieved a better consistency for dianions 2a,b and 4a,b in calculated vertical excitations in comparison to B3LYP, which was previously applied to UV−vis calculations on silenolates S 1 a,b and germenolates S 2 a,b. 24,27 The silenolates 2a,b exhibit intense absorption maxima in the range between 433 and 450 nm, which are red-shifted in the order 2b → 2a. The same bathochromic trend 4b → 4a also applies to the germenolates 4a,b with absorption maxima between 420 and 447 nm. The acyl substituent (aryl vs alkyl) significantly affects the HOMO orbital density and hence its shape, which ultimately leads to different reaction centers in conversion with electrophiles (see section below). The HOMO-1 and HOMO of 2a ( Figure 4) correspond to the p z orbital of the silenolate with a significant part of the corresponding silanide mixed in, respectively. This contribution makes the silanide equally nucleophilic regarding reactions of cyclic silenolates with aromatic acyl substituents. In contrast, the HOMO of 2b only exhibits the p z character of the silenolate, whereas the HOMO-1 of 2b shows the silanide orbital alone, allowing a site-specific functionalization. In addition, the energy difference between the HOMO-1 and the HOMO is in 2a significantly larger than in 2b (0.28 eV vs 0.09 eV). Similar observations were made with the dianionic germenolates 4a,b. Upon excitation, electron density is displaced into the π* orbital of the corresponding carbonyl orbitals. In the corresponding LUMOs of the aryl-substituted species 2a and 4a, our calculations additionally showed considerable conjugation of the carbonyl group and the aromatic π systems, which is not possible for the alkylsubstituted silenolate 2b and germanolate 4b. As a consequence of this, the empty orbitals are energetically stabilized in the order 2b → 2a. This stabilization results in smaller excitation energies and in the observed bathochromic shifts of the corresponding absorption bands. The obtained experimental and computational data are summarized in Table  4 and show reasonable agreement.

Reactivity of 2a,b versus Selected Examples of
Chlorosilanes. The reactivity of 2a,b versus chlorosilanes parallels the observed reactivities for silenolates and silyl anions. The same reactivity was found by Ohshita and Ishikawa, by Marschner et al., and by our group. [13][14][15]23,28,29 Thus, 2b with an alkyl group attached to the carbonyl moiety reacted with an equimolar amount of tetramethyldichlorodisilane (ClSiMe 2 SiMe 2 Cl) at 0°C in THF with formation of the acyl bicyclo[2.2.2]octasilane 6. 6 was obtained in nearly quantitative yield (95% yield). The asymmetrically substituted acylsilane 6 exhibits two 29 Si resonance lines for the SiMe 2 Table 3. Selected 13 C and 29 Si NMR Chemical Shifts for the Silenolates 2a,b and 4a,b a 2 (ppm) 4 (ppm)   . NMR spectral data of 5 (see the Experimental Section) are also typical for a Brook-type silene. 13 C and 29 Si signals characteristic for SiC were observed at δ 29 Si 32.8 ppm and δ 13 C 198.9 ppm, respectively. 5 was obtained in a good yield (64% yield).
Reactivity of 4a,b versus Selected Examples of Chlorosilanes. Furthermore, the reactivity of 4a,b versus chlorosilanes was investigated and parallels that previously observed for germenolates and germyl anions. 24,31 The reaction of 4a with 2 equiv of ClSiiPr 3 afforded the formation of the O-silylated germene 7 in excellent yields (compare Scheme 4). NMR spectral data of 7 (see the Experimental Section) are again typical for a Brook-type germene. A 13 C signal characteristic for GeC was observed at δ 13 C 210.17 ppm. Interestingly, the reaction of 4b with an equimolar amount of ClSiMe 2 SiMe 2 Cl did not lead to the formation of the expected product 8; instead, an undefined polymeric material was formed.
The unsuccessful derivatization of 4b with ClSiMe 2 SiMe 2 Cl encouraged us to investigate the reactivity of 4b with 2 equiv of ClSiiPr 3 . Again, no expected product formation was observed (Scheme 5). The same experiment was further repeated with 2b and gave also rise to an undefined polymer. Therefore, we reasoned that ClSiiPr 3 is too sterically demanding to allow M− Si (M = Si for 2b and M = Ge for 4b) bond formation in the presence of an adamantoyl group.
Reactivity of 2a,b and 4a,b versus Carbon-Centered Electrophiles. We selected MeI as a carbon-centered electrophile, because it represents a benchmark reagent with numerous examples found in the literature. 27,29,32 In the reaction of 2a,b and 4a,b with MeI, the same reactivities in terms of reaction sites were observed. In all cases, alkylation of the negatively charged silicon as well as germanium atoms were found in nearly quantitative yields (Scheme 6). Again, the same tendency was reported in the case of acyclic silenolates and silanides by Ohshita, Ottosson, and Marschner earlier. 14,15,28,29,32 The methylated silicon derivatives 9a,b and germanium derivatives 10a,b were obtained as cis/trans mixtures. The silicon atoms of 9a,b undergo a significant low-field shift from −70 to 45 ppm (in the case of the acylsubstituted silicon atom) and from −131 to −84 ppm (for the silyl-substituted silicon atom). This is caused by the lower shielding of the methyl group in comparison to the trimethylsilyl group (see the Experimental Section).
Competitive Reactivity of the Silyl Anion and the Silenolate. Finally, we investigated which silanide is more nucleophilic, the silyl anion or the silenolates. Therefore, we  examined the competitive reactivity of the silyl anion and the silenolate toward 1 equiv of ClSiMe 3 . The outcome of the reaction was strongly influenced by the substituent at the carbonyl moiety, which was in alignment with our computational analysis. 2a reacted with 1 equiv of ClSiMe 3 to give the corresponding cyclic silenolate S 1 a and demonstrated that the silyl anion is more nucleophilic than the silenolate. In contrast to that, 2b reacted with 1 equiv of ClSiMe 3 to give the bicyclic compound 11 via an intramolecular sila-Peterson alkenation reaction. This observation clearly showed that the alkylsubstituted silenolate is more nucleophilic than the silyl anion. Further studies to probe the scope of these new dianions are currently in progress.

■ EXPERIMENTAL SECTION
All experiments were performed under a nitrogen atmosphere using standard Schlenk techniques. Solvents were dried using a column solvent purification system. 33 ClSiMe 3 (95%), KOtBu (>98%), ClCOMes (99%), ClCOAd (98%) and 18-cr-6 (99%), were used without any further purification. 1 H, 13 C, and 29 Si NMR spectra were recorded on a Varian INOVA 300 spectrometer in C 6 D 6 , THF-d 8 , or CDCl 3 solutions and were referenced versus TMS using the internal 2 H-lock signal of the solvent. HRMS spectra were obtained on a Kratos Profile mass spectrometer. Infrared spectra were obtained on a Bruker Alpha-P Diamond ATR spectrometer from the solid sample. Melting points were determined using a Stuart SMP50 apparatus and are uncorrected. Elemental analyses were carried out on a Hanau Vario Elementar EL apparatus. UV absorption spectra were recorded on a PerkinElmer Lambda 5 spectrometer.   29 Si NMR (THF-d 8  Synthesis of 5. A 500 mg portion (0.76 mmol) of 1a was dissolved in 20 mL DME and cooled to −30°C, and 180 mg (1.60 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 2a was completely formed. Subsequently, the reaction mixture was cooled to −30°C and 309 mg (1.60 mmol) of ClSiiPr 3 was added dropwise. The red solution immediately turned yellow. After removal of the volatile components under vacuum, the remaining yellow solid was dissolved in heptane, the solution was filtered through dry Celite, and the solvent was Synthesis of 6. A 500 mg portion (0.74 mmol) of 1b was dissolved in 20 mL of DME and cooled to −30°C, and 175 mg (1.60 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 2b was completely formed. Subsequently, the reaction mixture was cooled to −30°C and 146 mg (0.74 mmol) of ClSiMe 2 SiMe 2 Cl was added dropwise. The red solution immediately turned colorless. After aqueous workup with 10 mL of 3% sulfuric acid, the organic layer was separated and dried over Na 2 SO 4  Synthesis of 7. A 500 mg portion (0.67 mmol) of 3a was dissolved in 20 mL of DME and cooled to −30°C, and 158 mg (1.41 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 4a was completely formed. Subsequently, the reaction mixture was cooled to −30°C and 272 mg (1.41 mmol) of ClSiiPr 3 was added dropwise. The red solution immediately turned yellow. After removal of the volatile components under vacuum, the remaining yellow solid was dissolved in heptane, the solution was filtered through dry Celite, and the solvent was stripped off again. Synthesis of 8. A 500 mg portion (0.66 mmol) of 3b was dissolved in 20 mL of DME and cooled to −30°C, and 155 mg (1.38 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 4b was completely formed. Subsequently, the reaction mixture was cooled to −30°C and 123 mg (0.66 mmol) of ClSiMe 2 SiMe 2 Cl was added dropwise. The red solution immediately turned colorless. After aqueous workup with 10 mL of 3% sulfuric acid, the organic layer was separated and dried over Na 2 SO 4 and the solvent was stripped off with a rotary evaporator. Drying under vacuum and subsequent NMR measurement showed complete degradation to an uncharacterizable polymer.
Synthesis of 1c. A 500 mg portion (0.74 mmol) of 1b was dissolved in 20 mL of DME and cooled to −30°C, and 175 mg (1.60 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR Organometallics pubs.acs.org/Organometallics Article showed that the dianionic species 2b was completely formed. Subsequently, the reaction mixture was cooled to −30°C and 287 mg (1.49 mmol) of ClSiiPr 3 was added dropwise. The red solution immediately turned colorless. After aqueous workup with 10 mL of 3% sulfuric acid, the organic layer was separated and dried over Na 2 SO 4 and the solvent was stripped off with a rotary evaporator. Drying under vacuum and subsequent NMR measurement showed complete degradation to an uncharacterizable polymer. Synthesis of 4c. A 500 mg portion (0.66 mmol) of 3b was dissolved in 20 mL of DME and cooled to −30°C, and 155 mg (1.38 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 4b was completely formed. Subsequently, the reaction mixture was cooled to −30°C and 267 mg (1.38 mmol) of ClSiiPr 3 was added dropwise. The red solution immediately turned colorless. After aqueous workup with 10 mL of 3% sulfuric acid, the organic layer was separated and dried over Na 2 SO 4 and the solvent was stripped off with a rotary evaporator. Drying under vacuum and subsequent NMR measurement showed complete degradation to an uncharacterizable polymer.
Synthesis of 9a. A 500 mg portion (0.76 mmol) of 1a was dissolved in 20 mL of DME and cooled to −30°C, and 180 mg (1.60 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 2a was completely formed. Subsequently, the reaction mixture was cooled to −30°C and an excess of MeI was added dropwise. The red solution immediately turned yellow. After aqueous workup with 10 mL of 3% sulfuric acid, the organic layer was separated and dried over Na 2 SO 4 and the solvent was stripped off with a rotary evaporator. Drying under vacuum afforded 366 mg (89%) of the analytically pure cyclic acylsilane 9a as a cis/trans mixture. The obtained product was recrystallized from acetone, giving yellow crystals of one isomer. Yield: 144 mg (35%) of analytically pure 9a (isomer 1).
Data for 9a (isomer 1) are as follows. Mp: 124−126°C. Anal. Calcd for C 23  Data for 9a (isomer 2) are as follows. 29 Si NMR (C 6 D 6 , TMS, ppm): Synthesis of 9b. A 500 mg portion (0.74 mmol) of 1b was dissolved in 20 mL of DME and cooled to −30°C, and 175 mg (1.60 mmol) of KOtBu was added. After it was stirred for an additional 30 min, the mixture was warmed to room temperature and finally stirred for an additional 14 h. At this time, reaction control by 29 Si NMR showed that the dianionic species 2b was completely formed. Subsequently, the reaction mixture was cooled to −30°C and an excess of MeI was added dropwise. The red solution immediately turned colorless. After aqueous workup with 10 mL of 3% sulfuric acid, the organic layer was separated and dried over Na 2 SO 4 and the solvent was stripped off with a rotary evaporator. Drying under vacuum afforded 335 mg (81%) of the analytically pure cyclic acylsilane 9b as a cis/trans mixture. Finally, this mixture of isomers was chromatographed on a precoated TLC SIL G-200 UV 254 plate, with toluene/heptane (1/5) as eluent, to separate both isomers. Yield: 145 mg (35%) of analytically pure 9b (isomer 1). Yield: 103 mg (25%) of analytically pure 9b (isomer 2).