1,2- and 1,1-Migratory Insertion Reactions of Silylated Germylene Adducts

The reactions of the PMe3 adduct of the silylated germylene [(Me3Si)3Si]2Ge: with GeCl2·dioxane were found to yield 1,1-migratory insertion products of GeCl2 into one or two Ge–Si bonds. In a related reaction, a germylene was inserted with tris(trimethylsilyl)silyl and vinyl substituents into a Ge–Cl bond of GeCl2. This was followed by intramolecular trimethylsilyl chloride elimination to another cyclic germylene PMe3 adduct. The reaction of the GeCl2 mono-insertion product (Me3Si)3SiGe:GeCl2Si(SiMe3)3 with Me3SiC≡CH gave a mixture of alkyne mono- and diinsertion products. While the reaction of a divinylgermylene with GeCl2·dioxane only results in the exchange of the dioxane of GeCl2 against the divinylgermylene as base, the reaction of [(Me3Si)3Si]2Ge: with one GeCl2·dioxane and three phenylacetylenes gives a trivinylated germane with a chlorogermylene attached to one of the vinyl units.


Introduction
Low valent main group compounds such as carbenes and their higher congeners have attracted much attention due to their transition metal-like behavior [1][2][3][4]. While most of this similarity refers to the oxidative addition and reductive elimination processes, other elementary reaction steps typical for transition metals are also possible.
The 1,2-migratory insertion is not restricted to phenylacetylene. The same reaction could be also carried out with trimethylsilylactylene and 1-hexyne (Scheme 1) [13]. When phosphane adducts of cyclic disilylated germylenes were subjected to reactions with alkynes spirocyclic, vinylgermanes were obtained [14]. Similar alkyne insertion reactivity was later observed for a diborylstannylene [15]. A related ethylene insertion was observed previously in the Si-Sn bond of a stannylated silylene adduct [16], and later in the Si-Si bond of the stable acyclic silylene dipp(Me 3 Si)N-Si-Si(SiMe 3 ) 3 (dipp = 2,6-iPr 2 C 6 H 3 ) [17]. More recently, Sasamori, Tokitoh and co-workers reported the spectacular example of a digermyne acting as a main group catalyst for the cyclotrimerization of arylacetylenes.
The key steps of the catalytic cycle are 1,2-migratory insertion reactions of arylacetylenes into Ge-C bonds of vinylgermylenes [18]. In the current account, we present the chemistry of 1 and related compounds with GeCl 2 ·dioxane.
In the 29 Si NMR spectrum of 3, the Ge(PMe 3 )Si(SiMe 3 ) 3 part (δ = −8.7 (d, 3 J Si-P = 11.4 Hz, GeSi(SiMe 3 ) 3 ) and −112.6 (d, 2 J Si-P = 16.1 Hz, Si(SiMe 3 ) 3 ) very much resembles the spectrum of 1, while the GeCl 2 unit attached to the second Si(SiMe 3 ) 3 group caused a deshielding effect of the chemical shift for the central silicon atom of the Si(SiMe 3 ) 3 unit to −76.7 ppm. The same was observed for the signal at −73.0 ppm in the 29 Si NMR spectrum of 4. With the phenylacetylene insertion product 2 readily available [5], we were interested in whether the 1,1-migratory insertion of GeCl 2 into the remaining Ge-Si bond of 2 would occur with similar ease. The reaction of vinylgermylene PMe 3 adduct 2 with GeCl 2 ·dioxane indeed proceeded quantitatively, however, the obtained compound 5 ( Figure 3, Scheme 3) was not the expected insertion product of the Ge-Si bond, but rather a cyclic germylene PMe 3 adduct.  It is known that germylene phosphane adducts are in equilibrium with the free germylene [13]. For the formation of 5 we assumed that, in the first step, the respective free germylene of 2 inserted into a Ge-Cl bond of GeCl 2 , forming a germylchlorogermylene PMe 3 adduct intermediate (6). The intramolecular cyclization of 6 can occur via the metathesis of a Ge-Cl with a Si-SiMe 3 bond, releasing Me 3 SiCl with simultaneous Si-Ge bond formation (Scheme 3). In this mechanism, the eventual germylene involves not the original germanium atom but the one introduced by the addition of GeCl 2 .
In the solid state structure of 5, the planar ring features of Ge-Ge (2.5052(5) Å) and Ge-P (2.3610(9) Å) bond distances, which are somewhat shorter than that of 3 and 4, were the bond of the germylene atom to the Si(SiMe 3 ) 2 unit (2.4409(9) Å), which is longer than in 3 but still short of the 2.4816(8) Å observed for 2. The Si-Ge-Ge angle is very acute (84.97(3) • ), with the Ge-PMe 3 interaction located almost orthogonal to the ring plane. The 29 Si NMR spectrum features the signal for the germylene attached Si(SiMe 3 ) 2 unit at −50.1 ppm (d, 2 J Si-P = 15 Hz) and that for the CSi(SiMe 3 ) 3 at −100.7 (d, 3 J Si-P = 2 Hz).
Both NMR spectroscopic results and XRD data showed selective formation of only one isomer of 5. It is reasonable to assume that the other possible isomer where PMe 3 and Cl are on different sides of the ring plane is less stable. Since we have previously shown that the dissociation of PMe 3 from silylated germylenes is a facile process, re-association of PMe 3 would occur in a selective way to form the more thermodynamically stable product [13].
Compound 3 offered the opportunity to also study the reaction sequence of 1,1-insertion of GeCl 2 followed by 1,2-insertion of an alkyne. In particular, the regioselectivity of the alkyne insertion reaction was of interest. As the germylene atom of adduct 3 has two weak bonds, the question arose whether the alkyne would insert into the Ge-Si bond or into the, presumably, weaker Ge-Ge bond. Much to our surprise, we observed two products. One was the double regioselective insertion of Me 3 SiC≡CH (7), not into the expected Ge-Si and Ge-Ge bonds adjacent to the germylene atom but into the two different Ge-Si bonds presented in 3 (Scheme 4). While the structure of compound 7 was unequivocally established by single crystal XRD analysis, the structure of the other compound 8a was assigned by NMR spectroscopy as the product of a single alkyne insertion into the germylene-Si bond. The outcome of the experiment was rationalized by assuming that the first step in the reaction occurred analogously to what is shown in Scheme 1. The 1,2-insertion of the alkyne into the (Me 3 Si) 3 Si-Ge bond gave the germylene adduct 8a (Scheme 5). The initial formation of 8a was consistent with NMR spectroscopic analysis of the reaction course. Dissociation of PMe 3 can lead to the free germylene of 8a, which can rearrange to 1,2-dichlorodigermene 8b and, in a further rearrangement step, lead to isomer 8c, where formally the two germanium atoms of 8a and 8c exchanged valency (Scheme 5). The insertion of another equivalent of Me 3 SiC≡CH into the (Me 3 Si) 3 Si-Ge bond of isomer 8c led to 7.  When an N-heterocyclic carbene (NHC) adduct of the double phenylacetylene insertion product 9 [13] was subjected to reaction with GeCl 2 ·dioxane, the germylene lone pair served as a Lewis base to replace the dioxane of GeCl 2 (10) (Scheme 6 top reaction, Figure 5). A similar reaction of Roesky's germylene with GeCl 2 ·dioxane has been reported [25].
The solid state structure of 10 ( Figure 5) formally includes two divalent germanium atoms. Ge1 acts as a Lewis acid to accept the NHC as a base (Ge1-C1 = 2.006(4) Å). In addition Ge1 donates via its lone pair to the GeCl 2 unit (Ge(1)-Ge(2) 2.5628(8) Å). Compared to the GeCl 2 units in compounds 3, 4 and 7, the Ge-Cl distances of 10 were elongated to 2.30 Å, consistent with the divalent character of the germylene atom. As expected, the NMR spectra of 10 were fairly similar to that of 9. The most pronounced difference was found for the 13 C signal of the NHC carbene carbon atom, while for 9 a value of 179.9 ppm was observed, the respective signal for 10 can be detected at 150.6 ppm. Similar up-field shifts of the NHC carbene signal were observed previously for cases where the lone pair of a tetrylene NHC adduct acted as a base to a Lewis acid [26]. Scheme 6. Top reaction: Reaction of divinyl germylene NHC adduct 9 with GeCl 2 ·dioxane resulting in the formation of 10, where the NHC adduct 9 acts as a Lewis base to GeCl 2 . Bottom reaction: Reaction of germylene 1 with three equivalents phenylacetylene and GeCl 2 ·dioxane to give a trivinylated germane with a germylene PMe 3 adduct moiety attached to one of the vinyl groups. In a related reaction, we reacted germylene adduct 1 with three equivalents phenylacetylene and one equivalent GeCl 2 ·dioxane. Unfortunately, the reaction was not very selective but it was possible to select a crystal of PMe 3 germylene adduct 11 under the microscope (Scheme 6, Figure 6). It seems reasonable that 11 was formed via the initial formation of the previously described divinylgermylene [5,13]. In a similar way, as was proposed for the formation of 5, insertion of the divinylgermylene into a Ge-Cl bond of GeCl 2 ·dioxane was supposed to give a chlorogermylgermylene. Phenylacetylene can now insert into the Ge-Ge bond of this new germylene, which eventually is isolated as its PMe 3 adduct 11. The germanium atom of the initial germylene then attached to three vinyl substituents. The third vinyl group, however, contained the phenyl group in the 2-position as its regioselectivity was then determined by the newly introduced germylene atom. The crystal structure of 11 ( Figure 6) features a tetravalent germanium atom with the three Ge-C distances covering a range between 1.94 and 1.96 Å and a Ge-Cl distance of 2.1662(9) Å. Again, the Ge-C bond to the divalent germanium atom Ge2 is longer (2.031(3) Å) and the same is true for the Ge-Cl bond (2.296(1) Å).
The 1 H (300 MHz), 13 C (75.4 MHz), 31 P (121.4 MHz), and 29 Si (59.3 MHz) NMR spectra were recorded on a Varian INOVA 300 spectrometer and are referenced to tetramethylsilane (TMS). If not otherwise noted, the solvent used was C 6 D 6 and the samples were measured at rt. In the case of the reaction samples, a D 2 O capillary was used to provide an external lock frequency signal. To compensate for the low isotopic abundance of 29 Si, the INEPT pulse sequence was used for the amplification of the signal [30,31].
Elementary analysis was carried out using a Heraeus VARIO ELEMENTAR. For a number of compounds, the obtained elemental analysis showed carbon values that were too low, which is a typical problem for these compounds likely caused by silicon carbide formation during the combustion process.
X-ray Structure Determination: For X-ray structure analyses, the crystals were mounted onto the tip of glass fibers. 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 [32] and SADABS [33,34], respectively. The structures were solved by direct methods and refined by full-matrix least-squares method (SHELXL97) [35]. If not noted otherwise all non-hydrogen atoms were refined with anisotropic displacement parameters and all hydrogen atoms were located in calculated positions to correspond to standard bond lengths and angles. Crystallographic data (excluding structure factors) for the structures of compounds 3, 4, 5, 7, 10, and 11 reported in this paper were deposited with the Cambridge Crystallographic Data Center as supplementary publication number CCDC-969562 (3), 969565 (4), 1918810 (5), 1918811 (7), 1918808 (10), and 1918809 (11) (see Supplementary Materials). The 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 [36] and rendered using POV-Ray 3.6 [37].

Dichloro[tris(trimethylsilyl)silyl]germyl[tris(trimethylsilyl)silyl]germylene·PMe 3 Adduct (3)
Method A: A solution of tris(trimethylsilyl)silyl potassium (0.31 mmol) in THF (2 mL) was slowly added dropwise to a vigorously stirred solution of GeCl 2 ·dioxane (72 mg, 0.31 mmol) and trimethylphosphane (11 mg, 0.16 mmol) in THF (40 mL) at −30 • C. Stirring of the reaction mixture was continued for 1 h at rt. After the solvent was removed, the residue was extracted with pentane (3 × 5 mL). The orange extract was concentrated to 3 mL and stored at −35 • C. Yellow crystalline 3 (54 mg, 43%) was obtained. was determined by 29 Si NMR after 24 h. The solvent was removed and the residue was extracted with toluene (5 × 5 mL). The solvent was removed to yield a brown solid (229 mg). Crystallization from toluene gave pale yellow crystals. 29 Si NMR analysis indicated formation of at least four different compounds of which 11 is likely the major component. NMR: δ in ppm: 29 Si{ 1 H}: −11.5 (SiMe 3 ), −84.9 (Si q ). The selection of a well-shaped crystal from the batch under the microscope allowed the single crystal XRD analysis of 11.

Conclusions
Here, we reported a few examples of a main group of analogs of what would be called 1,2-migratory insertions in transition metal chemistry. Mono-substituted alkynes were found to insert into the Si-Ge bonds of disilylated germylene PMe 3 adducts [5,13,14]. The current account extends this chemistry to main group analogs of 1,1-migratory insertions. Reactions of one or two equivalents of GeCl 2 ·dioxane with the PMe 3 adduct of [(Me 3 Si) 3 Si] 2 Ge: gave either the mono-or diinsertion product of GeCl 2 insertion into the Ge-Si bonds. Only a few examples of insertion reactions of GeCl 2 into weak E-E bonds of group 14 oligomers are known [19][20][21][22], and the current chemistry seems to provide an interesting pathway to precursors for unsaturated group 14 cluster molecules.
How subtle changes in the starting material can change the reaction course profoundly became evident when we subjected another silylated germylene to reaction with GeCl 2 ·dioxane. This time the alkyne mono-insertion product (Me 3 Si) 3 SiGe:(Ph)C=CHSi(SiMe 3 ) 3 served as the precursor and, instead of the expected GeCl 2 insertion into the Si-Ge bond, we observed the substrate acting as germylene inserting into a Ge-Cl bond. The resulting chlorogermylchlorogermylene reacted further by trimethylsilyl chloride elimination to a cyclic germylene adduct.
In an attempt to reverse the order of 1,1 and 1,2-insertion, we next reacted the GeCl 2 mono-insertion product (Me 3 Si) 3 SiGe:GeCl 2 Si(SiMe 3 ) 3 with Me 3 SiC≡CH. An unexpected mixture of alkyne monoand diinsertion products was obtained, where the first alkyne insertion proceeded into the Si-Ge bond of the germylene but the second insertion did not occur into the Ge-Ge bond as expected but into the Ge-Si bond of the second Si(SiMe 3 ) 3 unit. For this to happen, the involved germanium atoms needed to swap oxidation states, meaning that the two Cl atoms of the GeCl 2 would migrate to the divalent Ge-atom. We assumed that this rearrangement proceeds via known 1,2-shifts of chlorides involving germylene-digermene-germylene interconversion.
As expected, in the eventual reaction of a divinylgermylene with GeCl 2 ·dioxane no germylene insertion chemistry was observed. In this case only the exchange of the dioxane ligand of GeCl 2 against the divinylgermylene base could be detected. A final one-pot reaction of [(Me 3 Si) 3 Si] 2 Ge:·PMe 3 with one equivalent GeCl 2 ·dioxane and three equivalents of phenylacetylene gave a number of products with an interesting trivinylated chlorogermane, where a chlorogermylene is attached to one of the vinyl units. We suppose that this compound forms by initial divinylgermylene formation, followed by insertion of this germylene into a Ge-Cl bond of GeCl 2 . The thus formed germylene with a germyl and a chloride substituent can insert a third alkyne into the Ge-Ge bond. Since the newly introduced Ge atom now acts as the germylene the regiochemistry of the last inserted vinyl unit is reversed with respect to the central germanium atom.
We think that the demonstrated chemistry bears the potential become a useful method in the chemistry of heavy group 14 compounds. The facile introduction of GeCl 2 into molecules that already contain a certain degree of unsaturation opens an entry into precursor chemistry of interesting oligogermanium cluster chemistry.