Photoinduced Brook-Type Rearrangement of Acylcyclopolysilanes.

Previously unknown 1,1,4-tris(trimethylsilyl)-4-acyldodecamethylcyclohexasilanes (Me3Si)2Si6Me12(Me3Si)COR (16a, R = tert-butyl; 16b, R = 1-adamantyl) have been synthesized by the reaction of the potassium silanides (Me3Si)2Si6Me12(Me3Si)K with acid chlorides ClCOR, and their photochemical rearrangement reactions have been studied. The molecular structures of 16a,b as determined by single-crystal X-ray diffraction analysis exhibit an unusual twist-boat conformation of the cyclohexasilane ring. When 16a,b were photolyzed with λ >300 nm radiation, they underwent Brook type 1,3-Si → O migration reactions to generate the cyclohexasilanes 17a,b with exocyclic Si=C bonds along with smaller amounts of the ring-enlarged species 19a,b with endocyclic Si=C double bonds. While 17a,b were stable enough to allow characterization by NMR and UV absorption spectroscopy, the less stable products 19a,b could only be observed in the form of their methanol adducts.

Introduction

Silenes with various structures have been isolated and characterized since Brook and subsequently Wiberg about 30 years ago reported on the first stable species which contain Si=C double bonds (compare Chart 1).[1] Brook utilized a photochemical 1,3-Si → O shift of a SiMe3 group of the acylpolysilane (Me3Si)3SiCOAd (Ad = 1-adamantyl) to generate the silene 1,[2] while Wiberg obtained the donor-free species 2 by a route involving intramolecular lithium fluoride elimination.[3] Further advances in Si=C double-bond chemistry include the synthesis of the stable silene 3 by Apeloig and co-workers via a sila-Peterson type reaction[4] and the synthesis of the endocyclic silene 4, obtained by rearrangement of a silylene.[5] In addition, stable 1-silaallenes 5 and 6 have been reported by the groups of West and Pietschnig,[6] while Okazaki and Tokitoh published a series of papers on stable silaaromatic compounds such as the 2-silanaphthalene 7.[7] Only a few years ago, Kira et al. isolated and structurally characterized the 4-silatriafulvene 8.[8] Even more recently, the first example of the stable metal-substituted silene 9 was synthesized by Bravo-Zhivotovskii and Apeloig et al.,[9] while Scheschkewitz and Sekiguchi et al. reported on the formation of the cyclic Brook-type silene 10 from the disilenide Tip2Si=SiTip2Li (Tip = 2,4,6-iPr3C6H2) and AdCOCl.[10] Charged silenes have been described by Sekiguchi et al., who synthesized the silyl anion substituted silene 11,[11] and by Ottosson et al., who succeeded in the isolation of the first stable 2-silenolate 12.[12] In the most recent paper, finally, Iwamoto et al. reported on the synthesis of the exocyclic silene 13 showing a distinct intramolecular charge transfer transition from the π orbital of the Si=C double bond to the π* orbital of the anthryl moiety.[13]

Chart 1

Selected Stable Silenesa

Abbreviations: Tip, 2,4,6-iPr3C6H2; Tbt, 2,4,6-bis(trimethylsilyl)phenyl; Dmp, 2,6-dimesitylphenyl; Ad, 1-adamantyl); Ant, 9-anthryl.

Data on substituted silenes clearly demonstrate that the stability of silenes is strongly influenced by the choice of the substituents attached to the Si=C moiety.[14] As already has been realized by Brook, the steric bulk of the substituents on the carbon atom is certainly a key factor in the kinetic stabilization of silenes with regard to dimerization. Additional stability is gained in silenes influenced by reversed (Siδ−=Cδ+) bond polarization effected by π-donor substituents on the carbon atom.[15] Reverse-polarized silenes such as the Brook-type silenes 1, the silatriafulvene 8, or the transient 2-amino-2-siloxysilenes recently investigated by Ottosson et al.[16] are less reactive toward moisture and alcohols than naturally polarized silenes (Siδ+=Cδ−), and addition of alcohol often proceeds by C–O instead of Si–O bond formation. In contrast to the naturally polarized silenes, which give [2 + 2] and ene adducts as well, the reverse-polarized silenes react selectively with dienes to yield only [4 + 2] adducts.

Most silenes are acyclic molecules. To the best of our knowledge, in addition to the silaaromatics 1- and 2-silanaphthalene, 9-silaanthracene, and 9-silaphenantrene, 4 and 10 are the only stable silenes with the unsaturated silicon atom incorporated into cyclic structures which have been isolated and structurally fully characterized so far. According to DFT calculations on small model compounds the slight pyramidalization of the tricoordinate Si atom in 10 is due to reverse (Siδ−=Cδ+) bond polarization rather than steric congestion around the Si=C double bond. In line with reverse polarization 10 turned out to be remarkably stable. In contrast to Brook’s silene 1 it reacts only slowly with air and moisture and does not react with MeOH at any appreciable rate.[10]

Larger cyclopolysilanes containing either endo- or exocyclic Si=C double bonds have not been described in the literature before. Because we are interested in substituent effects on polysilane frameworks in general,[17] we now want to report on the outcome of our attempts to synthesize the previously unknown Brook-type cyclic silenes 17a,b by the photolysis of the acylcyclohexasilanes 16a,b, which we prepared successfully for the first time employing standard procedures for cyclopolysilane synthesis (Scheme 1).[18]

Scheme 1

Synthetic Approach toward Exocyclic Silenes

Results and Discussion

Synthesis of Acylcyclohexasilanes

According to Scheme 1, the potassium silanide 15 cleanly reacts with equimolar amounts of acid chlorides ClCOR (R = tBu, Ad) in diethyl ether solution at −80 °C to give the air-stable and crystalline acylcyclohexasilanes 16a,b in yields of >60%. Analytical data obtained for 16a,b (see the Experimental Section) are consistent with the proposed structures. Substitution of one SiMe3 group in 14 produces four magnetically nonequivalent endocyclic silicon atoms. Thus, the 29Si NMR spectra of 16a,b exhibit two resonance lines near −38 ppm for the endocyclic SiMe2 groups, one signal for the Si atom bearing the acyl group near −70 ppm, and one signal for the tertiary Si atom around −130 ppm.[19] The methyl and SiMe3 substituents, furthermore, can be attached either cis or trans relative to the acyl group, which leads to four nonequivalent methyl and three nonequivalent SiMe3 groups. The resulting number of resonance lines actually appears in the experimental NMR spectra, although some 1H signals are too close to each other to be completely resolved.

Single crystals suitable for X-ray structure analysis could be grown from compounds 16a,b. The obtained molecular structures are depicted in Figures 1 and 2 together with selected bond distances, bond angles, and dihedral angles. 16a,b crystallize in the monoclinic space group P21/n and in the triclinic space group P1̅, respectively. In both structures the cyclohexasilane ring adopts a twist-boat conformation, which is rather unusual, because most cyclohexasilanes studied so far feature a chair conformation of the cyclopolysilane cycle.[20] The geometry around the endocyclic silicon atoms is approximately tetrahedral with Si–Si–Si bond angles close to the respective angles found in other cyclohexasilane structures,[18,20,21] although Si(1) exhibits some distortion due to the steric bulk of the attached t-Bu or Ad group.

Figure 1

ORTEP diagram for compound 16a. Thermal ellipsoids are depicted at the 50% probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and bond and torsion angles (deg) with estimated standard deviations: Si–Si(mean) 2.356, C(1)–O(1) 1.213(2), Si(1)–C(1) 1.970(1), Si–Cmethyl(mean) 1.880; Si(6)–Si(1)–Si(2) 111.73(2), Si(3)–Si(2)–Si(1) 111.88(2), Si(4)–Si(3)–Si(2) 113.50(3), Si(5)–Si(4)–Si(3) 112.03(2), Si(4)–Si(5)–Si(6) 113.04(2), Si(5)–Si(6)–Si(1) 114.46(2), O(1)–C(1)–C(2) 120.2(1), O(1)–C(1)–Si(1) 113.9(1), C(2)–C(1)–Si(1) 125.9(1), C(1)–Si(1)–Si(7) 120.8(5), C(1)–Si(1)–Si(6) 99.1(5), C(1)–Si(1)–Si(2) 101.2(5), C–Si–C(mean) 107.6; O(1)–C(1)–C(2)–Si(1) 177.2(2).

Figure 2

ORTEP diagram for compound 16b. Thermal ellipsoids are depicted at the 50% probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and bond and torsion angles (deg) with estimated standard deviations: Si–Si(mean) 2.358, C(1)–O(1) 1.226(3), Si(1)–C(1) 1.962(2), Si–Cmethyl(mean) 1.883; Si(6)–Si(1)–Si(2) 111.73(3), Si(3)–Si(2)–Si(1) 115.54(3), Si(4)–Si(3)–Si(2) 113.61(3), Si(5)–Si(4)–Si(3) 111.67(3), Si(4)–Si(5)–Si(6) 114.40(3), Si(5)–Si(6)–Si(1) 112.18(3), O(1)–C(1)–C(2) 119.9(2), O(1)–C(1)–Si(1) 113.2(2), C(2)–C(1)–Si(1) 126.9(2), C(1)–Si(1)–Si(9) 121.78(7), C(1)–Si(1)–Si(6) 102.07(7), C(1)–Si(1)–Si(2) 99.12(7), C–Si–C(mean) 107.6; O(1)–C(1)–C(2)–Si(1) −176.9(3).

Si–Si bond lengths between 2.34 and 2.37 Å were observed; the average Si–Si bond distance of 2.36 Å is typical for Si–Si single bonds in cyclopolysilanes[22] and agrees well with the Si–Si covalent bond length of 2.34 Å. The sum of the bond angles around the carbonyl C atom in 16a,b is close to 360° and reflects the trigonal-planar geometry within the SiRC=O moiety. Unexceptional carbonyl C=O bond lengths of 1.21 and 1.23 Å were measured,[23] while the silicon carbonyl group bond distances at 1.97 and 1.96 Å are considerably elongated relative to the length of an average Si–C(sp3) bond,[24] as observed earlier for other acyl silanes.[25] In order to minimize steric congestion, the bulky t-Bu or Ad groups are oriented toward the outside of the molecules, which brings the carbonyl oxygen atoms in 16a,b in rather close contact to the endocyclic Si–Si bond system, with nonbonding distances to the plane defined by Si(1), Si(2), and Si(6) of only 2.66 and 2.65 Å, respectively. This structural feature will gain importance because it can be used to rationalize the unprecedented course of the photolysis experiments described in the next section of this paper.

Photolysis of Acylcyclohexasilanes

When the acylsilanes 16a,b in d6-benzene solution (c ≈ 0.07 M) in the absence of any other reactant were photolyzed with λ >300 nm radiation, yellow solutions were obtained. NMR analysis performed after an irradiation time of 4 h showed the formation of the exocyclic silenes 17a,b along with unreacted starting material. Figure 3 presents the 29Si NMR spectrum of the photolysis product derived from 16b, including the assignment of the observed resonance lines. The remaining spectra can be found in the Supporting Information. 29Si and 13C NMR chemical shift data of 17a,b are summarized in Table 1. 13C and 29Si NMR signals characteristic of Si=C were observed near 215 and 37 ppm, respectively, while the 29Si signal near 13 ppm is easily assigned to the OSiMe3 moiety formed by the photochemical 1,3-trimethylsilyl shift. These values compare reasonably well with those measured for the acyclic Brook-type silenes (Me3Si)2Si=C(OSiMe3)R (R = t-Bu, Ad).[2] The 1H NMR spectrum of the photolysis product of 16a contains, in addition to poorly resolved signals between δ 0.2 and 0.6 ppm for the SiMe3 and SiMe2 groups, absorptions at δ 1.26 and 0.96 ppm for the C(CH3)3 substituents in 16a and 17a, respectively. By integration of these signals a silene/acylsilane ratio of approximately 40/60 could be estimated.

Figure 3

29Si NMR spectrum after photolysis of 16b in C6D6 solution with a 150 W mercury lamp through Pyrex glass for 4 h at room temperature, including assignment of the resonance lines.

Table 1

13C and 29Si Chemical Shift Data of Silenes 17a,ba

	R = t-Bu (17a)	R = 1-Ad (17b)
13C NMR
Si=C	214.39	215.93
–R	42.70, 30.49	45.04, 42.15, 36.78, 29.03
–SiMe3	3.27, 1.17	3.29, 1.43
–SiMe2–	–0.60, −1.49, −1.66, −2.54	–0.32, −1.43, −1.62, −2.47
29Si NMR
Si=C	37.00	37.45
–OSiMe3	13.16	13.27
–SiMe3	–8.79	–8.69
–SiMe2–	–34.41, −36.81, −37.44, –37.58	–34.68, −36.75, −37.40, –37.48
SiSi4	–132.43	–132.38

In C6D6 solution. All values are vs external TMS, in ppm.

In C6D6 solution. All values are vs external n class="Gene">TMS, in ppm.

The interesting observation that it is possible to generate 17a,b photochemically with two competing chromophores in the acylcyclopolysilane, the cyclohexasilane moiety and the acyl group, can be rationalized on the basis of the UV spectra of 13 and 15. While 14 does not absorb above 300 nm[26]16a,b exhibit weak absorption bands near 370 nm (ε ≈ 200 L mol–1 cm–1), which are easily assigned to the n−π* transition of the C=O group in accordance with the literature.[2] Thus, the acyl group in 16 is selectively excited upon irradiation with 360 nm light, while the endocyclic σ(SiSi) electron system is not affected, which leads to the observed Brook type reaction course.

Further irradiation afforded increasing amounts of unidentified polymeric decomposition products at the expense of 17a,b, illustrated by the appearance of broad signal groups typical for polymers containing (SiMe) (δ 0–0.6 ppm) and t-Bu (Ad) (δ 1.0–2.0 ppm) groups in the 1H NMR spectra of the resulting photolysis solutions (compare Figure 4). Apparently 17a,b are less stable under photolytic conditions than Brook’s open-chain compounds, presumably as a consequence of the presence of the cyclohexasilane cycle. Cyclohexasilanes such as Si6Me12 have been shown earlier to undergo silylene extrusion and ring contraction reactions when photolyzed with 254 nm light.[27] Although experimental evidence is missing,[28] it is not unlikely that 16a,b exhibit similar reactivity upon irradiation at 360 nm because the wavelength for the onset of the photoinduced silylene extrusion might be considerably red-shifted due to conjugation of the exocyclic π(Si=C) chromophore with the endocyclic σ(SiSi) system. Thus, Si6Me12 does not absorb light of wavelengths >270 nm, while the UV absorption spectra of 17a,b show absorption bands at 362 and 302 nm of considerable intensity (Figure 5). It is also interesting to note that there is no evidence for the presence of head-to-head dimers arising from 2 + 2 cycloaddition reactions of 17a,b, which are the decomposition products of most Brook-type silenes such as (Me3Si)2Si=C(OSiMe3)t-Bu.[29] This is conclusive, because the formation of head-to-head dimers of 17a,b would cause severe steric strain due to the cyclic structure of the polysilane backbone, which apparently totally inhibits dimerization. Since no dimers were present in the photolysis mixtures, attempts were made to crystallize 17a,b from the photolysis mixtures obtained after an irradiation time of 4 h. Removal of solvent in vacuo, however, gave viscous yellow oils which contained mainly polymeric material of undefined composition along with unreacted 16a,b.

Figure 4

1H NMR spectrum after photolysis of 16b with a 150 W mercury lamp through Pyrex glass for 0, 4, 6, and 8 h at room temperature.

Figure 5

UV absorption spectra recorded after irradiation of a solution of 16b in C6H12 with a 4 W 366 nm lamp (c = 5 × 10–5 mol L–1; t = 0, 5, 30, 50 min).

UV absorption spectra recorded after irradiation of a solution of n class="Chemical">16b in C6H12 with a 4 W 366 nm lamp (c = 5 × 10–5 mol L–1; t = 0, 5, 30, 50 min).

As mentioned earlier, solutions of the silenes 17a,b are yellow. UV absorption spectra recorded after irradiation of a diluted hydrocarbon solution of 16b with λ 366 nm radiation in a quartz cuvette are presented in Figure 5. The quasi-identical spectra obtained for 16a may be found in the Supporting Information. After an irradiation time of 5 min two absorption bands appear centered at λmax ∼300 and ∼360 nm which were not present prior to photolysis. The latter band tailing into the visible region is easily assigned to a π–π* transition within the Si=C fragment. Quite remarkably, it is shifted to the red by 20 nm with respect to the related acyclic silenes (Me3Si)2Si=C(OSiMe3)R (R = t-Bu, Ad), which have π–π* absorptions of around 340 nm.[2] The bathochromic shift of the π–π* absorption band observed for 17a,b very likely might be explained by enhanced conjugation between the endocyclic σ(Si–Si) bond system with the exocyclic Si=C double bond. The data presented in Figure 5, furthermore, also reflect the instability of 17a,b under photolytic conditions. Upon prolonged irradiation the intensity of the 300 and 360 nm absorption bands decreases until nearly complete photobleaching occurs after about 20 and 50 min for 17a,b, respectively. 17b, therefore, seems to be slightly more stable, which is also evident from the relative intensities of the signals in the NMR spectra of the obtained photolysis solutions.

Trapping Experiments

Alcohols are known to be very effective trapping agents for Brook-type silenes.[1] Thus, photolysis of acyltris(trimethylsilyl)silanes in the presence of MeOH to which a trace of weak base such as pyridine or Et3N has been added usually affords the 1,2-addition product across the Si=C double bond with the RO– group attached to silicon and the alcoholic H attached to carbon (Scheme 2). Earlier work on the photolysis of acylsilanes has shown that in the absence of base subsequent acid-catalyzed C–O or Si–O bond cleavage reactions of the initially formed trapping products occur, which lead to the formation of complex mixtures of “solvolysis” products.[30] The acid catalysts appeared to be byproducts of the photolysis reaction. In contrast to the behavior described in Scheme 2, photolysis of the cyclic acyl silanes 16a,b in 2.5/1 C6D6/MeOH gave only about 80% of the adducts 18a,b expected from trapping of the silenes 17a,b which already have been observed in the absence of trapping reagents. Rather surprisingly, about 20% of the cyclosiloxanes 20a,b was also obtained, which are the trapping products of the silenes 19a,b.

Scheme 2

Photolysis of Acyltris(trimethylsilyl)silanes in the Presence of Methanol

These results are presented in Scheme 3 and clearly indicate that photochemical rearrangement of acylcyclohexasilanes to species containing Si=C double bonds can occur in two different ways. In addition to the “normal” Brook-type rearrangement including a 1,3-Si → O trimethylsilyl shift to give silenes 17a,b with exocyclic Si=C bonds (reaction path A), an unprecedented rearrangement takes place, including ring scission and a 1,3-Si → O dimethylsilyl shift with formation of the ring-enlarged silenes 19a,b with endocyclic Si=C double bonds (reaction path B). Likely the formation of 19a,b can be interpreted as arising from the molecular structure of the acyl silanes 16a,b, because reaction path B might become a favorable process, due to last but not least the close proximity of the carbonyl oxygen atom to the adjacent endocyclic Si–Si bonds already mentioned above (compare Figures 1 and 2). Furthermore, it is obvious that 19a,b are not stable under photolytic conditions in the absence of trapping reagents because they were not detected when the photolysis was carried out in pure hydrocarbon solvents.

Scheme 3

Photolysis of Acylcyclohexasilanes 16a,b in the Presence of Methanol/Et3N

Pure 18a and 20a could be isolated from the crude photolysis solution by column chromatography over silica gel with heptane as an eluant followed by crystallization from acetone. Pure 18b was obtained as the less soluble product from the mixture with 20b by repeated crystallization from acetone. Slightly impure crystals of 20b were isolated in low yield from the combined mother liquors after concentration and repeated crystallization from acetone at −30 °C. Nevertheless, the identity of the individual adducts 18a,b and 20a,b was established by NMR and high-resolution mass spectroscopy. Analytical data are summarized in the Experimental Section, and experimental 1H NMR spectra can be found in the Supporting Information. All adducts showed nine 29Si resonances, in line with the presence of nine magnetically inequivalent silicon atoms. For the six-membered cycles 18a,b four signals for the endocyclic SiMe2 groups between −32 and −41 ppm, two SiMe3 resonances near −6 and −9 ppm, and one signal for the OSiMe3 group near +15 ppm were detected. The isomers 20a,b, in contrast, had three SiMe2 signals between −32 and −44 ppm and one signal for the OSiMe2 group at significantly lower field (∼+19 ppm), while three SiMe3 signals appeared near −8.5 (two signals) and −19 ppm. In 13C NMR eight lines appear at the high-field end of the spectra for the magnetically nonequivalent methyl groups attached to the endocyclic silicon atoms.

The molecular structures of 18b and 20a,b in the solid state as determined by single-crystal X-ray diffraction are presented in Figures 6–8 together with selected bond distances, bond angles, and dihedral angles. 18b crystallizes in the monoclinic space group P21/c with the cyclohexasilane ring in a slightly distorted chair conformation and unexceptional bond lengths and angles. 20a,b crystallize in the triclinic space group P1̅. The eight-membered heterocycle adopts a boatlike conformation. The most prominent feature of the molecular structures of 20a,b is the position of the endocyclic oxygen atom, which is oriented toward the center of the ring in order to allow an exocyclic arrangement with minimum steric congestion for the bulky R group at C(1). 20a,b contain two chiral centers. In both cases one molecule of the S,S and one molecule of the R,R enantiomer are present in the unit cell, which are related by an inversion center (compare Figure 9). Obviously the diasteromeric species with R,S and S,R configuration were not formed, because otherwise a second set of NMR signals should have been observed.

Figure 6

ORTEP diagram for compound 18b. Thermal ellipsoids are depicted at the 50% probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and bond angles (deg) with estimated standard deviations: Si–Si(mean) 2.362, C(1)–O(1) 1.434(3), Si(1)–O(2) 1.679(2), Si(2)–O(1) 1.651(2), Si(1)–C(1) 1.930(3), Si–Cmethyl(mean) 1.877; Si(7)–Si(1)–Si(3) 110.14(3), Si(1)–Si(3)–Si(4) 107.90(3), Si(3)-Si4)-Si(5) 115.20(4), Si(4)–Si(5)–Si(6) 112.25(3), Si(5)–Si(6)–Si(7) 117.05(3), Si(6)–Si(7)–Si(1) 107.56(3), Si(1)–O(2)–C(15) 121.7(2), C(1)–O(1)–Si(2) 128.4(2), C–Si–C(mean) 107.8.

Figure 8

ORTEP diagram for compound 20b. Thermal ellipsoids are depicted at the 50% probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and bond and torsion angles (deg) with estimated standard deviations: Si–Si(mean) 2.361, C(1)–O(1) 1.458(3), Si(1)–O(2) 1.673(2), Si(6)–O(1) 1.654(2), Si(1)–C(1) 1.925(3), Si–Cmethyl(mean) 1.883; Si(1)–Si(2)–Si(3) 123.28(4), Si(2)–Si(3)–Si(4) 120.49(4), Si(3)–Si(4)–Si(5) 111.26(4), Si(4)–Si(5)–Si(6) 114.57(4), Si(1)–O(2)–C(20) 124.6(2), C(1)–O(1)–Si(6) 128.5(2), C–Si–C(mean) 107.6; Si(4)–Si(5)–Si(6)–O(1) 44.2(1), Si(2)–Si(1)–C(1)–O(1) −37.9(2).

Figure 9

Unit cell of compound 20a containing one S,S and R,R enantiomeric pair, which can be interconverted through an inversion center between the molecules.

ORTEP diagram for compound 20a. Thermal ellipsoids are depicted at the 50% probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and bond and torsion angles (deg) with estimated standard deviations: Si–Si(mean) 2.359, C(1)–O(1) 1.448(3), Si(4)–O(2) 1.672(2), Si(5)–O(1) 1.657(2), Si(4)–C(1) 1.929(2), Si–Cmethyl(mean) 1.879; Si(6)–Si(1)–Si(2) 112.35(3), Si(1)–Si(2)–Si(3) 119.26(3), Si(2)–Si(3)–Si(4) 122.98(3), Si(5)–Si(6)–Si(1) 115.00(3), Si(4)–O(2)–C(23) 124.0(2), C(1)–O(1)–Si(5) 128.4(2), C–Si–C(mean) 107.4; Si(1)–Si(6)–Si(5)–O(1) −41.4(1), Si(3)–Si(4)–C(1)–O(1) 38.0(1).

Photolysis of Acylbicyclo[2.2.2]octasilanes

In contrast to the behavior described above for the acylcyclohexasilanes 16a,b the photolysis of the 1-trimethylacyl-4-(trimethylsilyl)dodecamethylbicyclo[2.2.2]octasilane cage 21 in a MeOH/C6D6 mixture gave rise to the formation of the silene–methanol adduct 22 along with polymeric material of unknown structure, with no evidence for the presence of products derived from photochemically induced Me3Si migration (Scheme 4). When it was photolyzed in pure C6D6, 21 reacted very slowly to give unidentified polymeric decomposition products. This finding is not surprising, because 21 does not contain a Me3Si group in a position α to the acyl functionality. A Brook-type 1,3-SiMe3 shift under formation of an exocyclic silene, thus, is simply not possible.

Scheme 4

Photolysis of Acylbicyclo[2.2.2]octasilane 21 in the Presence of Methanol/Et3N

Although pure 22 could not be isolated from the obtained crude product mixture, its structure has been established by NMR and GC-MS analysis. Analytical data are summarized in the Experimental Section. GC-MS showed only one volatile product with a signal corresponding to the molecular ion at m/e 594. Due to the presence of nine inequivalent silicon atoms, nine 29Si resonances were found at −130.5 (SiSi4), −43.7, −43.5, −40.5, −37.4, and −36.7 (SiMe2), −5.6 (SiMe3), +8.6 (SiOMe), and +18.3 ppm ((SiMe2)O). Consistent with the proposed structure, 1H NMR showed a one-proton singlet at δ 3.68 ppm (HCOMe), a three-proton singlet at δ 3.47 ppm (OCH3), a nine-proton singlet at δ 0.99 ppm (C(CH3)3), and a poorly resolved signal group at δ 0.18–0.32 ppm (SiCH3) in addition to some minor impurities and some polymeric background.

Conclusions

In summary, we have demonstrated that acylcyclohexasilanes which contain a Me3Si group in a position α to the acyl substituent upon photolysis undergo Brook-type rearrangement reactions to give products with Si=C double bonds. In contrast to the behavior of branched open-chain substrates as studied extensively by Brook et al. in the past, however, we observed the formation of two products: a cyclohexasilane with an exocyclic Si=C double bond arising from a 1,3-Si → O shift of a SiMe3 group and a ring-enlarged silene with an endocyclic Si=C double bond arising from an unprecedented rearrangement including ring scission and insertion of the Si=C–O fragment into the cyclohexasilane ring. The exocyclic product was stable enough to allow detection by NMR and UV absorption spectroscopy, while the less stable endocyclic product could only be observed in the form of its methanol adduct. Photolysis experiments involving acylbicyclo[2.2.2]octasilane cages, furthermore, indicated that in the absence of α-SiMe3 groups exclusively the endocyclic product is obtained. Further studies with the primary aim to isolate related species with enhanced stability are currently underway.

Experimental Section

General Considerations

All experiments were performed under a nitrogen atmosphere using standard Schlenk techniques. Solvents were dried using a column solvent purification system.[31] Commercial KO-t-Bu (97%), ClCO-t-Bu (99%), and ClCOAd (98%) were used as purchased, Et3N (99%) was dried by distillation from solid KOH, and commercial anhydrous MeOH was dried with 3 Å molecular sieves. 14 and 15 were synthesized as previously reported.[18]1H (299.95 MHz), 13C (75.43 MHz), and 29Si NMR spectra (59.59 MHz) were recorded on a Varian INOVA 300 spectrometer in C6D6 or CDCl3 solution and referenced versus TMS using the internal 2H-lock signal of the solvent. Mass spectra were run either on an HP 5971/A/5890-II GC/MS coupling (HP 1 capillary column, length 25 m, diameter 0.2 mm, 0.33 μm polydimethylsiloxane) or on a Kratos Profile mass spectrometer equipped with a solid probe inlet. Infrared spectra were obtained on a Bruker Alpha-P Diamond ATR spectrometer from the solid sample. Melting points were determined using a Büchi 535 apparatus and are uncorrected. Elemental analyses were carried out on a Hanau Vario Elementar EL apparatus. Photolyses were performed by using a 150 W medium-pressure mercury lamp (Heraeus TQ 150). Sample solutions were photolyzed under an atmosphere of nitrogen in Pyrex Schlenk or NMR tubes immersed in cold water to ensure ambient sample temperature and to prevent irradiation with light of wavelengths λ <300 nm. UV absorption spectra of 17a,b were recorded on a PerkinElmer Lambda 5 spectrometer after irradiation of diluted cyclohexane solutions of 16a,b (c = 5 × 10–5 mol L–1) with a 4 W 366 nm lamp at room temperature in quartz cuvettes.

Synthesis of 1-Trimethylacyloctamethyl-1,4,4-tris(trimethylsilyl)cyclohexasilane (16a)

A solution of 15 in 20 mL of DME was freshly prepared from 1.74 g (3.0 mmol) of 14 and 0.37 g (3.3 mmol) of KO-t-Bu and slowly added to a solution of 0.40 g (3.3 mmol) of ClC-Ot-Bu in 50 mL of diethyl ether at −80 °C. Subsequently the mixture was stirred for another 30 min at −80 °C, warmed to room temperature, and finally stirred for an additional 60 min. After aqueous workup with 100 mL of 3% sulfuric acid the organic layer was separated and dried over Na2SO4 and the solvents were stripped off with a rotary evaporator. Drying in vacuo (0.02 mbar) and crystallization from acetone solution by slow evaporation of the solvent at room temperature afforded 1.37 g (77%) of analytically pure 16a as colorless crystals.

Mp: 159–161 °C. Anal. Found: C, 44.15; H, 9.82. Calcd for C22H60OSi9: C, 45.52; H, 10.19. 29Si NMR (C6D6, TMS, ppm): −7.77, −8.62, −11.55 (SiMe3); −37.04, −38.02 (SiMe2); −71.86 (SiCOtBu); −131.46 (Si(SiMe3)2). 13C NMR (C6D6, TMS, ppm): 245.41 (SiC=O); 48.77 (C(CH3)3); 24.75 (C(CH3)3); 3.75, 3.70, 2.55 (Si(CH3)3); −0.74, −0.87, −2.20, −2.26 (Si(CH3)2). 1H NMR (C6D6, TMS, ppm, relative intensity): 0.96 (9H, s, C(CH3)3); 0.50, 0.36, 0.35, (6H each, s, Si(CH3)2); 0.32 (15H, s, Si(CH3)2 + Si(CH3)3); 0.29, 0.28 (9H each, s, Si(CH3)3). IR (neat): ν(C=O) 1623 (m) cm–1. UV absorption (hexane solution): λ1 367 nm (ε1 = 190 mol–1 cm–1), absorption shoulder at 255 nm (ε2 = 11000 mol–1 cm–1). HRMS: calcd for [C22H60OSi9]•+ (M+) 592.2568, found 592.2603.

Synthesis of 1-Adamantylcarbonyloctamethyl-1,4,4-tris(trimethylsilyl)cyclohexasilane (16b)

The procedure followed was that used for 16a with 2.91 g (5.0 mmol) of 14, 0.62 g (5.5 mmol) of KO-t-Bu, and 0.99 g (5.0 mmol) of ClCOAd. Yield: 2.02 g (60%) of analytically pure 16b as colorless crystals.

Mp: 148–151 °C. Anal. Found: C, 49.91; H, 9.72. Calcd for C28H66OSi9: C, 50.07; H, 9.91. 29Si NMR (CDCl3, TMS, ppm): −7.60, −8.30, −10.90 (SiMe3); −36.80, −37.96 (SiMe2); −72.06 (SiCOAd); −131.32 (Si(SiMe3)2). 13C NMR (CDCl3, TMS, ppm): 248.22 (SiC=O); 51.50 (AdCCO); 36.99, 36.69 (Ad-CH2); 28.03 (Ad-CH); 3.86, 3.83, 2.83 (Si(CH3)3); −0.69, −0.92, −1.68, −2.01 (Si(CH3)2). 1H NMR (CDCl3, TMS, ppm, relative intensity): 2.06 (3H, b, Ad–CH); 1.72, 1.67 (6H each, b, Ad–CH2), 0.35, 0.30 (6H each, s, Si(CH3)2); 0.28, 0.25 (15H each, s, Si(CH3)3 + Si(CH3)2); 0.23 (9H, s, Si(CH3)3). IR (neat): ν(C=O) 1621 (m) cm–1. UV absorption (hexane solution): λ1 371 nm (ε1 = 200 mol–1 cm–1), absorption shoulder at 250 nm (ε2 = 12000 mol–1 cm–1). HRMS: calcd for [C28H66OSi9]•+ (M+) 670.3037, found 670.3065.

Photolysis of 16a,b in Hydrocarbon Solution

A solution of 0.05 mmol of the acylsilane 16a or 16b in 0.7 mL of d6-benzene in an NMR tube was photolyzed with a 150 W mercury lamp at 25 °C for 4 h. At this time 1H and 29Si NMR analysis showed the formation of the silene 17a or 17b, respectively, along with unreacted starting material. For 16a/17a a silene/acylsilane ratio of approximately 40/60 in the resulting yellow solution was estimated by integration of the C(CH3)3 signals in the 1H NMR spectra. Further irradiation afforded increasing amounts of polymeric decomposition products at the expense of 17a,b. Attempted removal of the solvent in vacuo after an irradiation time of 4 h also gave rise to extensive product decomposition.

Data for 17a are as follows. 29Si NMR (C6D6, TMS, ppm): 37.00 (Si=C), 13.16 (OSiMe3), −8.79 (Si(SiMe3)2), −34.41, −36.81, −37.44, −37.58 (SiMe2), −132.43 (Si(SiMe3)2). 13C NMR (C6D6, TMS, ppm): 214.39 (Si=C), 42.70 (C(CH3)3), 30.49 (C(CH3)3), 3.27 (Si(Si(CH3)3)2), 1.17 (OSi(CH3)3), −0.60, −1.49, −1.66, −2.54 (Si(CH3)2). 1H NMR (C6D6, TMS, ppm, relative intensity): 1.26 (s, C(CH3)3); 0.55–0.28 (several overlapping signals, Si(CH3)).

Data for 17b are as follows. 29Si NMR (C6D6, TMS, ppm): 37.45 (Si=C), 13.27 (OSiMe3), −8.69 (Si(SiMe3)2), −34.68, −36.75, −37.40, −37.48 (SiMe2), −132.38 (Si(SiMe3)2). 13C NMR (C6D6, TMS, ppm): 215.93 (Si=C), 45.04, 42.15, 36.78, 29.03 (Ad-C), 3.29 (Si(Si(CH3)3)3), 1.43 (OSi(CH3)2), −0.32, −1.43, −1.62, −2.47 (Si(CH3)2). 1H NMR (C6D6, TMS, ppm, relative intensity): 1.98 (b, Ad–CH); 1.89, 1.65 (b, Ad–CH2); 0.57–0.29 (several overlapping signals, Si(CH3)).

Photolysis of 16a/Methanol

A solution of 0.50 g (0.84 mmol) of 16a and 3 drops of anhydrous Et3N in 5 mL of benzene and 2 mL of methanol was photolyzed with a 150 W mercury lamp at 25 °C over 10 h. At this time NMR analysis showed that the starting material had been completely consumed. After removal of the volatile components on a rotary evaporator, 20 mL of pentane was added and the resulting solution was filtered over silica gel. Evaporation of pentane afforded 0.46 g of a semisolid residue containing approximately 80% of 18a and 20% of 20a. Pure and colorless crystals of 18a and 20a could be isolated from the crude product by column chromatography (heptane, silica gel) followed by crystallization of the individual components from acetone at −30 °C.

Data for 18a are as follows. Yield: 0.1 g (20%). Mp: 124–127 °C. 29Si NMR (C6D6, TMS, ppm): 16.13, 14.95 (SiOMe, OSiMe3); −6.49, −9.23 (Si(SiMe3)2); −33.06, −33.23, −39.57, −41.02 (SiMe2); −132.22 (Si(SiMe3)2). 13C NMR (C6D6, TMS, ppm): 80.12 (CH(t-Bu)OSi); 53.40 (OCH3); 35.07 (CCH3)3); 28.33 (C(CH3)3); 3.96, 3.92, 0.85 (Si(CH3)3); −0.25, −0.54, −1.34, −1.54, −3.02, −3.15, −3.23, −3.81 (Si(CH3)2). 1H NMR (C6D6, TMS, ppm, relative intensity): 3.55 (1H, s, CH(t-Bu)OSi)); 3.28 (3H, s, OCH3); 0.93 (s, 9H, C(CH3)3); 0.55, 0.47, 0.41, 0.405, 0.37, 0.31, 0.30, 0.28 (3H each, s, Si(CH3)2); 0.35, 0.345, 0.252 (9H each, s, Si(CH3)3). HRMS: calcd for [C23H64O2Si9]•+ (M+) 624.2830, found 624.2814.

Data for 20a are as follows. Yield: 0.05 g (10%). Mp: 136–138 °C. 29Si NMR (C6D6, TMS, ppm): 19.61, 16.78 (OSiMe2, SiOMe); −8.41, −8.45 (Si(SiMe3)2); −19.67 (MeOSiSiMe3); −32.25, −41.54, −43.11 (SiMe2); −130.64 (Si(SiMe3)2). 13C NMR (C6D6, TMS, ppm): 80.96 (OCHSi-t-Bu); 53.39 (OCH3); 34.61 (CCH3)3); 28.92 (CCH3)3); 3.72, 3.54, 1.48 (Si(CH3)3); 3.22, −0.10, −1.105, −1.11, −1.50, −2.10, −2.78, −4.11 (Si(CH3)2). 1H NMR (C6D6, TMS, ppm, relative intensity): 3.55 (1H, s, OCHSi-t-Bu); 3.28 (3H, s, OCH3); 0.93 (s, 9H, C(CH3)3); 0.55, 0.47, 0.41, 0.405, 0.37, 0.31, 0.30, 0.28 (3H each, s, Si(CH3)2); 0.35, 0.345, 0.25 (9H each, s, Si(CH3)3). HRMS: calcd for [C23H64O2Si9]•+ (M+) 624.2830, found 624.2855.

Photolysis of 16b/Methanol

A solution of 0.40 g (0.60 mmol) of 16b and 3 drops of anhydrous Et3N in 5 mL of benzene and 2 mL of methanol was photolyzed at 25 °C with a 150 W mercury lamp over 10 h. At this time NMR analysis showed that the starting material had been completely consumed. After removal of the volatile components on a rotary evaporator 20 mL of pentane was added and the resulting solution was filtered over silica gel. Evaporation of pentane afforded 0.38 g of a semisolid residue containing approximately 80% of 18b and 20% of 20b. Small amounts of pure and colorless crystals of 18b could be isolated as the less soluble product after repeated recrystallization from acetone solution. White crystals of slightly impure 20b were isolated in low yield from the combined mother liquors after concentration and repeated crystallization from acetone at −30 °C.

Data for 18b are as follows. Mp: 186–188 °C. 29Si NMR (CDCl3, TMS, ppm): 16.26, 14.82 (SiOMe, OSiMe3); −6.06, −9.08 (Si(SiMe3)2); −32.55, −32.74, −39.23, −40.66 (SiMe2); −131.93 (Si(SiMe3)2). 13C NMR (CDCl3, TMS, ppm): 81.45 (CH(Ad)OSi), 53.50 (OCH3); 40.87, 37.15, 37.10, 28.66 (Ad); 4.10, 4.07, 1.06 (Si(CH3)3); −0.13, −0.57, −1.38, −1.53, −2.82, −2.93, −3.29, −3.94 (Si(CH3)2). 1H NMR (CDCl3, TMS, ppm, relative intensity): 3.74 (1H, s, CH(Ad)OSi); 3.37 (3H, s, OCH3); 1.99 (3H, b, Ad–CH), 1.7–1.5 (12H, b, Ad–CH2); 0.35 (3H, s, Si(CH3)2), 0.30 (6H, b, Si(CH3)2); 0.27, 0.25 (3H each, s, Si(CH3)2); 0.26 (15H, s, Si(CH3)3 + Si(CH3)2); 0.22 (12H, s, Si(CH3)3 + Si(CH3)2); 0.125 (9H, s, Si(CH3)3). HRMS: calcd for [C29H70O2Si9]•+ (M+) 702.3300, found 702.3335.

Data for 20b are as follows. 29Si NMR (CDCl3, TMS, ppm): 19.30, 16.69 (OSiMe2, SiOMe); −8.37, −8.45 (Si(SiMe3)2); −19.74 (MeOSiSiMe3); −32.23, −41.36, −42.93 (SiMe2); −130.52 (Si(SiMe3)2). 13C NMR (CDCl3, TMS, ppm): 81.85 (SiCH(Ad)OSi); 53.66 (OCH3); 41.30, 37.00, 36.72, 28.63 (Ad); 3.83, 3.62, 1.81 (Si(CH3)3); 3.62, −0.02, −1.05, −1.14, −1.38, −2.15, −2.88, −3.94 (Si(CH3)2). 1H NMR (CDCl3, TMS, ppm, relative intensity): 3.47 (1H, s, CH(O)Ad); 3.41 (3H, s, OCH3); 1.97 (3H, b, Ad–CH); 1.7–1.4 (12H, b, Ad–CH2); 0.38, 0.31, 0.27, 0.20, 0.17 (3H each, s, Si(CH3)2); 0.25 (6H, b, Si(CH3)2); 0.28, 0.26 (9H each, s, Si(CH3)3), 0.24 (12H, s, Si(CH3) + Si(CH3)3). HRMS: calcd for [C29H70O2Si9]•+ (M+) 702.3300, found 702.3348.

Photolysis of 1-Trimethylacyl-4-(trimethylsilyl)dodecamethylbicyclo[2.2.2]octasilane (21) in Methanol

A solution of 40 mg (0.071 mmol) of 21 and 3 drops of anhydrous Et3N in 0.3 mL of benzene and 0.3 mL of methanol in an NMR tube was photolyzed with a 150 W mercury lamp at 25 °C for 10 h. At this time NMR and GC-MS analysis showed the formation of methanol adduct 22. Attempted purification by crystallization of the oily crude product obtained after evaporation of the solvents failed to yield a crystalline solid.

29Si NMR (CDCl3, TMS, ppm): 18.35 (OSiMe2), 8.63 (SiOMe), −5.65 (SiMe3); −36.71, −37.38, −40.48, −43.54, −43.75 (SiMe2); −130.53 (SiSi4). 13C NMR (CDCl3, TMS, ppm): 76.31 (CH(O)-t-Bu); 54.53 (OCH3); 35.50 (C(CH3)3); 28.17 (C(CH3)3), 3.82 (Si(CH3)3), 3.75, −0.05, −0.25, −0.36, −0.66, −1.08, −1.48, −1.68, −2.23, −3.32, −3.85, −4.07 (Si(CH3)2). 1H NMR (CDCl3, TMS, ppm, relative intensity): 3.68 (1H, s, CH(O)tBu); 3.47 (3H, s, OCH3); 0.99 (9H, s, C(CH3)3); 0.32, 0.30, 0.29, 0.27, 0.24, 0.18 (3H each, s, Si(CH3)2); 0.31 (6H, s, Si(CH3)3 + Si(CH3)2), 0.28 (9H, s, Si(CH3)3 + Si(CH3)2), 0.26 (12H, s, Si(CH3)3 + Si(CH3)2). MS (m/e (relative intensity)): 594 (0.7%, M+).

X-ray Crystallography

For X-ray structure analysis suitable crystals were mounted onto the tip of glass fibers using mineral oil. Data collection was performed on a Bruker Kappa Apex II CCD diffractometer at 100 K using graphite-monochromated Mo Kα (λ = 0.71073 Å) radiation. Details of the crystal data and structure refinement are provided as Supporting Information. The SHELX version 6.1 program package was used for the structure solution and refinement.[32] Absorption corrections were applied using the SADABS program.[33] All non-hydrogen atoms were refined with anisotropic displacement parameters. Hydrogen atoms were included in the refinement at calculated positions using a riding model as implemented in the SHELXTL program. Crystallographic data (excluding structure factors) have been deposited with the Cambridge Crystallographic Data Centre as supplementary publications CCDC-964365 (16a), CCDC-964366 (16b), CCDC-964367 (18b), CCDC-964368 (20a), and CCDC-964369 (20b). Copies of the data can be obtained free of charge on application to The Director, CCDC, 12 Union Road, Cambridge CB2 1EZ, U.K. (fax (internat.), +44-1223/336-033; e-mail, deposit@ccdc.cam.ac.uk).