β-Amino- and Alkoxy-Substituted Disilanides

Our recent study on formal halide adducts of disilenes led to the investigation of the synthesis and properties of β-fluoro- and chlorodisilanides. The reaction of the functionalized neopentasilanes (Me3Si)3SiSiPh2NEt2 and (Me3Si)3SiSiMe2OMe with KOtBu in the presence of 18-crown-6 provided access to structurally related β-alkoxy- and amino-substituted disilanides. The obtained Et2NPh2Si(Me3Si)2SiK·18-crown-6 was converted to a magnesium silanide and further on to Et2NPh2Si(Me3Si)2Si-substituted ziroconocene and hafnocene chlorides. In addition, an example of a silanide containing both Et2NPh2Si and FPh2Si groups was prepared with moderate selectivity. Also, the analogous germanide Et2NPh2Si(Me3Si)2GeK·18-crown-6 could be obtained.


Introduction
The development of silanide chemistry [1][2][3][4][5] has advanced the progress of organosilicon chemistry over the last decades. Being the silicon analogs of carbanions, these compounds are invaluable for the construction of Si-X bonds. Especially, the formation of Si-Si bonds, which has long been restricted to Wurtz-type coupling of halosilanes, has been facilitated by the use of silanides, which allow much easier access to inherently asymmetric di-and oligosilanes [6]. In addition, these compounds also allow a much more flexible access to transition-metal silyl complexes and silylated organic compounds.
Typically, silanides are highly reactive species which are easily hydrolyzed to hydrosilanes. They need to be prepared, stored, and handled under strict exclusion of moisture. Initially, the chemistry of silanides was mainly restricted to alkyl-, aryl-, and silyl-substituted compounds. The introduction of functional groups to silanides represented a logical advance to allow further transformations of the formed di-and oligosilanes. Examples of hydrogen-substituted silanides date back to Gilman's cleavage of HPh 2 SiSiPh 2 H with lithium [7], and related compounds have been used also later [8,9]. The attachment of Lewis basic heteroatoms to silanides was pioneered by Kawachi and Tamao, who introduced α-aminosilanides in the early 1990s [10,11] and could demonstrate their use as hydroxy anion equivalent in organic synthesis [11]. Later on, α-alkoxy- [12][13][14] and fluorosilanides [15][16][17] were introduced, and these compounds (silylenoids) turned out to be ambiphilic with a tendency to self-condensation. This was not observed for aminosilanides, although by NMR spectroscopy they clearly showed a relationship to those more reactive compounds [18].

Results and Discussion
After investigating the chemistry of α-aminooligosilanides some time ago [20], we now wish to extend our studies to β-aminooligodisilanides, which can be considered the formal condensation products of α-aminooligosilanides. As the latter do not undergo facile dimerization, a straightforward way to obtain β-aminooligodisilanides is the reaction of aminosilylchlorides with oligosilanides followed by further silanide formation via a reaction with potassium tert-butoxide. The synthesis of (Me 3 Si) 3 SiSiPh 2 NEt 2 (1) was thus achieved by reaction of (Me 3 Si) 3 SiK [21,22] with Et 2 NPh 2 SiCl [23,24]. Single-crystal XRD analysis of the obtained compound ( Figure 1) showed that it crystallized in the orthorhombic space group P2(1)2(1)2(1) (Table S1). An observed Si-N distance of 1.717(3) Å reflects the diminished steric demand of the diphenylsilylene unit compared to the respective Si(SiMe 3 ) 2 element of [Et 2 N(SiMe 3 ) 2 SiSiMe 2 ] 2 [20], where the Si-N bond length is increased to 1.741(7) Å. A Si-Si-bond length of 2.3744(14) Å for the Et 2 NPh 2 Si-Si bond and other Si-Si lenghts between 2.3540(15) and 2.3706(14) Å (Table 1) as well as Si-Si-Si bond angles close to the ideal tetrahedral angle, characterized compound 1 as a rather typical neopentasilane. The same conclusion could be drawn from the 29 Si NMR chemical shifts of 1 (Table 2). Resonances at −10.1 and −133.4 ppm for the trimethylsilyl groups and the central silicon are close to that of tetrakis(trimethylsilyl)silane. The shift of +2.0 ppm found for the Et 2 NPh 2 Si group is more unusual. Comparable compounds such as (Et 2 N)Ph 2 Si-SiMe 3 (−15.3 ppm), (Et 2 N)Ph 2 Si-SiMe 2 -SiPh 2 (NEt 2 ) (−15.7 ppm) [25], and (Et 2 N)Ph 2 Si-Si t Bu 2 H (−8.7 ppm) [26] featured the respective resonances more up-field. The reaction of compound 1 with potassium tert-butoxide in the presence of 18-crown-6 [22] led to the formation of silanide 1a (Scheme 1). Crystal structure analysis of 1a (Figure 2), which crystallizes in the triclinic space group P-1, showed a silanide with shortened Si-Si bond lengths close to 2.33 Å (Table 1). A sum of Si-Si-Si bond angles of 319.85 deg showed a fairly pyramidalized silanide atom. The potassium ion appeared coordinated by the crown ether, and the Si-K distance of 3.5064 (19) Å was within the range usually observed for this type of compounds. No intramolecular N . . . K interaction could be observed, but a weak interaction of the NEt 2 group with a neighboring molecule might be possible. The Si-N distance 1.743(2) of 1a is clearly large compared to that of 1.717(3) Å for 1, which likely reflects a repulsive interaction between the lone pairs on nitrogen and silicon.    When comparing the structural parameters of 1a to those of FPh 2 Si(Me 3 Si) 2 SiK·18-crown-6 [19], we found that the contraction of the central Si-Si bond was not so pronounced. This can be due to the fact that the disilene adduct character of 1a is smaller than that of FPh 2 Si(Me 3 Si) 2 SiK·18-crown-6. Further comparison with the recently reported Ph 3 Si(Me 3 Si) 2 SiK·18-crown-6 [27] showed a high degree of similarity, which was further emphasized by the 29 Si-NMR characterization of 1a (Table 2), featuring the silanide resonance at −186.6 ppm and the signal for SiMe 3 at −6.3 ppm (δ for Ph 3 Si(Me 3 Si) 2 SiK: −189.0 and −5.8 ppm [27]). The expected down-field shift for the SiPh 2 NEt 2 caused the respective signal to appear at +16.3 ppm.
Frequently, in reactions with redox labile metal salts, potassium silanides turn out to be too reducing. To moderate the reactivity of these compounds, we found it convenient to transmetallate them to magnesium by metathesis reaction with MgBr 2 ·Et 2 O [28,29]. The application of this method on 1a caused the formation of 1b (Scheme 1). Usually, we prepare silyl magnesium compounds in situ from the respective potassium compounds. Assuming quantitative conversion both for silanide formation and for transmetallation reactions, the stoichiometry of the magnesium silanide is determined by the amount of oligosilane starting material. Typically, we use an excess of MgBr 2 ·Et 2 O for reasons of convenience. While more MgBr 2 has no effect on reactivity, single-crystal XRD analysis of 1b provided us with an example where an additional equivalent of MgBr 2 co-crystallized with two molecules of Et 2 NPh 2 Si(Me 3 Si) 2 SiMgBr ( Figure 3).
The slightly less anionic character of 1b, compared to 1a, is nicely reflected by its crystal structure. The central Si-Si bond distance of 2.338(3) Å is slightly longer than in 1a, conversely, the Si-N distance of 1.730 (7) is somewhat shortened, and the sum of bond angles (328.6(1) deg) is larger, indicating diminished pyramidalization. This is consistent with 29 Si-NMR shifts of −165.5 ppm for the silanide atom and of +8.7 ppm for the SiPhNEt 2 group (Table 2). Again, this is caused by the increased covalent character of the Si-Mg bond compared the to the Si-K interaction of 1a.
In a related way, tris(trimethylsilyl)dimethylmethoxysilylsilane (2) [30] was treated in C 6 H 6 with potassium tert-butoxide in the presence of 18-crown-6 (Scheme 1). The fact that the respective methoxylated silanide 2a was formed at all is somewhat surprising, as we previously observed that in the reaction of either chloro-or fluorodimethylsilyltris(trimethylsilyl)silane with potassium tert-butoxide, the attack of the alkoxide occurred exclusively at the halodimethysilyl group. It seems likely that the selectivity in the reaction of 2 does not depend on a stronger steric shielding of the methoxydimethysilyl versus the halodimethysilyl group but rather is due to the fact that the formation of a tert-butoxydimethysilyl group seems thermodynamically not feasible. The same way as FPh 2 SiSi(SiMe 3 ) 2 K is structurally related to F(Me 3 Si) 2 SiSi(SiMe 3 ) 2 K, compound 2a is related to MeO(Me 3 Si) 2 SiSi(SiMe 3 ) 2 K, which we obtained from the self-condensation of MeO(Me 3 Si) 2 SiK [13]. Single-crystal XRD analysis of 2a ( Figure 4) revealed it to be a typical isotetrasilanide with a high degree of pyramidalization (sum of bond angles: 305.46(8) deg), a short Si-SiOMe bond distance of 2.329(2) Å (2.3361(13) Å was found for MeO(Me 3 Si) 2 SiSi(SiMe 3 ) 2 K), and a somewhat longer Si-O distance of 1.678(5) Å. Compared to 1a, the stronger pyramidalization is consistent with a more shielded 29 Si NMR resonance at −201.8 ppm (Table 2). Judging the fact that the silanide resonance of MeO(Me 3 Si) 2 SiSi(SiMe 3 ) 2 K was observed at −170.4 ppm, it can be concluded that MeO(Me 3 Si) 2 SiSi(SiMe 3 ) 2 K can be regarded as a methoxide adduct of a disilene [13], whereas 2a is better described as a β-methoxydisilanide.
Over the last years, we have prepared numerous oligosilanylated zirconocenes and hafnocenes, mostly by reaction of either potassium or magnesium oligosilanides with the respective group 4 metallocene dichlorides [31][32][33][34][35][36]. Reactions of 1b with zirconocene and hafnocene dichlorides proceeded analogously and provided access to complexes 3a and 3b (Scheme 2). In the course of the reaction between 1b (prepared in situ by addition of MgBr 2 ·Et 2 O to a solution of 1a) and Cp 2 ZrCl 2 (Scheme 2), the formed compound 3a reacted by about 20% with the formed MgClBr to yield Cp 2 Zr(Br)Si(SiMe 3 ) 2 SiPh 2 NEt 2 as a side product. The presence of the latter was recognized in the crystal structure of 3a ( Figure 5). Both the fairly covalent interaction between Si and Zr as well as the steric demand of the Cp 2 Zr(Cl) unit caused an elongation of the central Si-Si distance of the Si(SiMe 3 ) 2 SiPh 2 NEt 2 moiety to 2.387(3) Å. The observed Si-Zr bond length of 2.863(2) Å is close to those found previously for Cp 2 Zr(Cl)Si(SiMe 3 ) 2 SiMe 2 Thex (2.853 Å) and Cp 2 Zr[Si(SiMe 3 ) 3 ] 2 (2.878 Å) [32]. However, compared to the structurally related complex Cp 2 Zr(Cl)Si(SiMe 3 ) 2 SiPh 2 F (d Si-Zr = 2.799(1)/2.803(1) Å) [19], the bond is significantly longer. As expected, complex 3b ( Figure 6) was found to be isostructural to 3a, with the Si-Hf bond (2.8339 (8)    An analogous reaction of 1a with Cp 2 TiCl 2 did not give the respective isotetrasilanyl titanocene chloride but the oxidative coupling product 4 (Scheme 2). This is not entirely surprising, as we found out earlier that disilylated titanocenes in the oxidation state +4 tend to undergo reductive elimination of disilanes [34,35]. The crystal structure of 4 ( Figure 7) [41] was reported to exhibit a longer bond. For the tetrasilane unit, a trans-conformation was found, with a torsional angle of 180 deg and diethylamino substituents on different sides of the plane defined by the main chain. The 29 Si-NMR resonance (−115.4 ppm, Table 2) of the central silicon atoms of 4 reflects its extended branched octasilane framework [42]. Since compound 1a appeared somewhat different from FPh 2 Si(Me 3 Si) 2 SiK [19], we wondered if it was possible to prepare a silanide containing both the Et 2 NPh 2 Si and the FPh 2 Si groups. For this reason, we reacted 1a with Ph 2 SiF 2 , obtaining neopentasilane 5 in good yield (Scheme 3). The reaction of the latter with KO t Bu in the presence of 18-crown-6 gave access to the respective silanide 5a (Scheme 3) at low temperature (−30 • C). Unfortunately, single-crystal XRD analysis of 5a was not possible, but its NMR spectroscopic properties, in particular the 1 J SiF coupling constant and the SiK and SiPh 2 F chemical shifts, were very similar to those of FPh 2 Si(Me 3 Si) 2 SiK [19]. As tris(trimethylsilyl)germyl potassium is a readily available compound [43], we decided to extend the chemistry of β-aminooligosilanides to that of aminosilyl-substituted germanides. For this reason, we reacted tris(trimethylsilyl)germyl potassium with Et 2 NPh 2 SiCl to obtain diethylaminodiphenylsilyltris(trimethylsilyl)germane (6), which in a subsequent step, was converted to potassium diethylaminodiphenylsilylbis(trimethylsilyl)germanide 6a by reaction with KO t Bu (Scheme 4). The compound can be regarded as the diethylamide adduct of a silagermene. As could be expected, 1 H and 13 C-NMR spectroscopic properties of 6 and 6a were fairly close to those of 1 and 1a. The 29 Si-NMR spectra of 6 and 6a displayed the typical behavior of silylated germanes, where the silyl resonances were shifted a few ppm towards a lower field [43].

General Remarks
All reactions involving air-sensitive compounds were carried out under an atmosphere of dry nitrogen or argon using either Schlenk techniques or a glove box. All solvents were dried using a column-based solvent purification system [44]. Chemicals were obtained from different suppliers and used without further purification. 1  To compensate for the low isotopic abundance of 29 Si, the INEPT pulse sequence was used for the amplification of the signal [45,46]. Frequently, this did not allow observing diphenylsilyl Si signals; therefore, a simple inverse-gated single pulse experiment was used for those cases. Elementary analyses were carried out using a Heraeus VARIO ELEMENTAR instrument. As potassium and magnesium silanides usually give poor analysis data, the purity of these compounds was confirmed by 1 H, 13 C and 29 Si-NMR spectra.

X-Ray Structure Determination
For X-ray structure analyses, the crystals were mounted onto the tip of a glass fiber, and data collection was performed with a BRUKER-AXS SMART APEX CCD diffractometer using graphite-monochromated Mo Kα radiation (0.71073 Å). The data were reduced to F 2 o and corrected for absorption effects with SAINT [47] and SADABS [48,49], respectively. The structures were solved by direct methods and refined by a full-matrix least-squares method (SHELXL97) [50]. If not noted otherwise, all non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were located in calculated positions to correspond to standard bond lengths and angles. All diagrams were drawn with 30% probability thermal ellipsoids, and all hydrogen atoms were omitted for clarity. Crystallographic data (excluding structure factors) for the structures of compounds  1, 2a, 1a, 1b, 3a, 3b, and 4 (4). Copies of data can be obtained free of charge at: http://www.ccdc.cam.ac.uk/products/csd/request/. The figures of solid-state molecular structures were generated using Ortep-3 as implemented in WINGX [51] and rendered using POV-Ray 3.6 [52].

N,N-Diethylaminodiphenylsilyltris(trimethylsilyl)silane (1)
A solution of tris(trimethylsilyl)silyl potassium (freshly prepared from tetrakis(trimethylsilyl)silane (1.50 g, 4.68 mmol) and KO t Bu (0.54 g, 4.81 mmol) in THF (10 mL)) was added dropwise to a vigorously stirred solution of N,N-diethylaminochlorodiphenylsilane (1.36 g, 4.71 mmol) in THF (10 mL). After the addition, the yellow reaction mixture was stirred for 1 d at ambient temperature, after which all volatiles were evaporated under reduced pressure. The slightly yellow residue was extracted with pentane, and the concentrated extracts were stored at −50 • C to yield colorless crystals of 1 ( Bu (117 mg, 1.05 mmol), and 18-crown-6 (276 mg, 1.05 mmol) were dissolved in a minimum amount of benzene. The color of the mixture immediately turned to orange. After a few minutes, the reaction was finished as could be determined by 29 Si-NMR spectroscopy of an aliquot sample. The solvent was removed in vacuum, and pentane was added to the residue. Product 1a (171 mg, 32%) was crystallized as deep-orange crystals from pentane. NMR

N,N-Diethylaminodiphenylsilylbis(trimethylsilyl)silyl Zirconocene Chloride (3a)
A cold solution of 1b (449 mg, 1.00 mmol) (stored at −35 • C prior to the reaction) in THF (2 mL) was added dropwise to a cold solution of Cp 2 ZrCl 2 (350 mg, 1.10 mmol) in THF (5 mL) under vigorous stirring. After the addition, the orange reaction mixture was kept at −35 • C for another 1 h, followed by evaporation of all volatiles under reduced pressure. The orange residue was extracted with benzene/pentane (1:2 ratio), the solutions were combined and evaporated to dryness, and the solid residue was washed with pentane (3 × 3 mL). The remaining solid was taken up in benzene (3 mL),w and the solution was layered with pentane, which afforded red crystalline 3a (520 mg, 73%). NMR  (4) To a solution of 1 (400 mg, 0.80 mmol) in THF (3 mL), KO t Bu (94 mg, 0.84 mmol) was added. The color immediately turned to dark reddish. The reaction mixture was stirred for 15 h, whereupon the conversion was quantitative according to NMR measurements, and the solution was cooled to −78 • C. A solution of Cp 2 TiCl 2 in THF (1 mL) was added dropwise to the reaction mixture, and stirring was continued at rt for 24 h. The solvent was removed in vacuo, and the residue was dissolved in pentane. The precipitated salts were removed by decantation, and the product was crystallized from pentane by slow evaporation, yielding crystalline colorless 4 (274 mg, 80%  29 Si (inverse-gated): 36.0 (d, 1 J Si-F = 359 Hz, SiPh 2 F), 14.9 (SiPh 2 NEt 2 ), −6.6 (d, 3 J Si-F = 5 Hz, SiMe 3 ), −198.9 (SiK).
3.2.11. N,N-Diethylaminodiphenylsilyltris(trimethylsilyl)germane (6) A solution of tris(trimethylsilyl)germyl potassium (freshly prepared from tetrakis(trimethylsilyl)germane (1.10 g, 3.00 mmol) and KO t Bu (354 mg, 3.15 mmol) in THF (6 mL)) was added dropwise to a vigorously stirred solution of chloro(diethylamino)diphenylsilane in THF (20 mL). The yellowish reaction mixture was stirred for another 0.5 h after the addition was finished. Then, all volatiles were evaporated to dryness, and the yellow residue was extracted with pentane. The yellow extracts were concentrated to a volume of about 4 mL and stored at −35 • C for crystallization, yielding 6 as colorless crystals (230 mg
For fluoro-and alkoxysilanides, we found a tendency to self-condensation, yielding β-fluoroand alkoxydisilanides. Unusual spectroscopic and chemical properties suggested that these compounds should be regarded as base adducts of symmetrical disilenes. Recently, we reported the synthesis of some β-halodisilanides with different substituents at the 1-and 2-positions of the disilane unit. NMR spectroscopic, structural, and chemical characterization clearly showed that the disilene adduct character of the novel compounds was much diminished.
The preparation and characterization of β-N,N-diethylaminodisilanides, outlined in the current study, clearly showed that also these compounds should be regarded as oligosilanides and not as amide disilene adducts. The study includes the conversion of the initially obtained potassium silanides to the respective derivatives of magnesium, zirconium, and hafnium. In addition, we extended this chemistry to a β-N,N-diethylaminodiphenylsilylbis(trimethylsilyl)germanide.