Synthesis and Photochemistry of Tris(trimethoxysilyl)acyl-silanes and 1,4-Tetrakis(silyl)-1,4-bisacylsilanes

In this contribution, we present the synthesis of two groups of novel acylsilanes 1–6. Compounds 1 and 2 represent tris(trimethoxysilyl)acylsilanes, and compounds 3–6 are 1,4-tetrakis(silyl)-1,4-bisacylsilanes. All isolated compounds were characterized by infrared (IR) and nuclear magnetic resonance (NMR) spectroscopy and X-ray crystallography. Additionally, these compounds were further analyzed by ultraviolet/visible (UV/vis) spectroscopy and their longest wavelength absorption bands were assigned by density functional theory (DFT) calculations. On the basis of the well-known Brook rearrangement of acylsilanes, we irradiated 1–6 in benzene solutions at 405 nm (λ) for several hours. Photolysis of compounds 1 and 2 afforded the same silene rearrangement products as found in previous investigations of structurally related acylsilanes. In addition, trapping experiments with MeOH further support our proposed mechanism for silene formation. The photolysis of tetrakis(trimethylsilyl)bisacylsilane 3 gave rise to the formation of a monosilene intermediate 10; upon prolonged irradiation, the subsequently formed bissilene undergoes a fast dimerization to bicyclic product 11. Interestingly, unlike the expected head-to-head dimerization of Brook-type silenes, this bissilene undergoes a selective head-to-tail dimerization. In contrast, tetrakis(trimethylsilyl)bisacylsilane 4 undergoes a selective and completely stereoselective double CH activation to air stable bicyclic system 12. The mechanism of this rearrangement is fully described by DTF calculations. Unfortunately, tetrakis(trimethoxysilyl)bisacylsilanes 5 and 6 underwent unselective photochemical rearrangements.


■ INTRODUCTION
−11 The pioneer of this research field was A. G. Brook, who discovered and elaborated the photochemistry of acylsilanes. 12He showed that these compounds undergo a thermally induced or a photoinduced 1,2-shift of the silyl group forming siloxycarbenes as labile intermediates.The accepted mechanism for the formation of siloxycarbenes in the singlet state is shown in Scheme 1. Upon irradiation, the acylsilane system is promoted to excited state 1*.On the basis of the small energy gap between the nonbonding and antibonding orbitals of carbonyl compounds, the excited singlet state of acylsilane 2S* is formed by relaxation.An intersystem crossing (ISC) enables the formation of an excited triplet state of acylsilane 2T*, which is followed by 1,2-Brook rearrangement, leading to the corresponding triplet siloxycarbene intermediate 3T.Subsequently, 3T undergoes another ISC to form the singlet state of siloxycarbene 3S. 13 These carbenes have shown the ability to react as reagents in a variety of organic reactions.The nucleophilic addition of siloxycarbenes to various electrophiles (E + ) has been used by many groups (Scheme 1, path A).In addition, such carbenes can insert into a large variety of C−H, B−H, or C−B bonds (Scheme 1, path B).Their addition to alkenes and alkynes has also been published (Scheme 1, path C). 14,15 In the presence of Si−Si bonds in the backbone of the molecule, a photochemically induced 1,3-trimethlysilyl shift is possible.This "unusual" Brook rearrangement was exploited by Brook et al. and led to the isolation of the first stable silene (see Scheme 2). 16This was a milestone in organosilicon chemistry and tremendously boosted the number of papers on silene chemistry.
While Brook used a variety of different substituents at the carbonyl moiety, he did not change the substituents at the silyl groups.The implementation of KO t Bu as a standard reagent for the synthesis of polysilanes opened the door for new molecules. 17By a slight modification of the substituent pattern, we found the formation of exocylic 18 as well as endocyclic silenes. 19,20Recently, we introduced a pathway toward tris(trimethoxysilyl)silanides as a new building block for polysilane chemistry. 19,21Moreover, we could also show that the nucleophilicity of the central Si atoms is significantly higher than that of structurally related silanides.The reason for this is the negative hyperconjugation induced by the three trimethoxysilyl groups.We therefore asked ourselves whether these groups can also influence the photochemistry of acylsilanes.Consequently, we synthesized novel tris-(trimethoxy)acylsilanes and investigated their photochemistry.The second part of this work is devoted to α,ω-bis(acyl)polysilanes.Here, we report on their synthesis and their photochemistry.
■ RESULTS AND DISCUSSION Synthesis of Precursor Molecules.Tris(trimethoxysilyl)acylsilanes.Dodecamethoxyneopentasilane, which was synthesized according to the literature, 22 was treated with equimolar amounts of potassium tert-butoxide (KO t Bu) in dry tetrahydrofuran (THF) to generate the silanide.Afterward, the anion is reacted with 1 equiv of the corresponding acid chloride (see Scheme 3).The reaction results in the clean formation of acylsilanes 1 and 2 in good yields.Analytical data are consistent with the proposed structures, exhibiting one resonance line in the 29 Si NMR spectra for the three trimethoxysilyl groups near −42 ppm and one signal for the Si atom bearing the acyl group near −107 ppm for 1 and −98 ppm for 2. In the 13 C NMR spectrum, the signal at approximately δ = 242 ppm is characteristic of a carbonyl group bound to silicon.
Tetrakis(silyl)-1,4-bisacylsilanes. Hexakis(trimethylsilyl)tetramethyltetrasilane and hexakis(trimethoxysilyl)tetramethyltetrasilane, which were synthesized according to the literature, 23,24 were treated with 2.10 equiv of KO t Bu in dry DME and THF, respectively, to generate the dianionic species.Afterward, the dianions are reacted with 2.10 equiv of the corresponding acid chlorides (R = Ad and Mes).After recrystallization, products 3−6 can be isolated in high yields (see Scheme 4).Analytical data for 3 and 4 are consistent with the proposed structures, exhibiting one resonance line in the 29 Si NMR spectra for the four trimethylsilyl groups near δ = −11 ppm, one signal for the dimethylsilyl groups near δ = −30 ppm, and one signal for the Si atom bearing the acyl group at approximately δ = −70 ppm.For compounds 5 and 6, the 29 Si NMR spectra show one signal for the four trimethoxysilyl groups near δ = −41 ppm, one signal for the dimethylsilyl groups near δ = −31 ppm, and one signal for the Si atom bearing the acyl group at δ = −93 ppm for compound 5 and at δ = −84 ppm for compound 6.Again, in the 13 C NMR spectrum, peaks for the carbonyl groups are observed at approximately δ = 245 ppm.
We were able to grow crystals suitable for single-crystal Xray structure analysis by cooling a solution of 3 and 4 in acetone to −30 °C (Figures 1 and 2).Compound 3 crystallizes in tetragonal space group P4 3 2 1 2 with four molecules per unit cell.Compound 4 crystallizes in monoclinic space group P2 1 /c with two molecules per unit cell.Spectroscopy.To accurately determine the transition for the following photochemical experiments, UV/vis spectra of 1−6 were recorded.The UV/vis spectra of 1−6 reveal characteristic bands centered between 350 and 410 nm for the

Organometallics
adamantoyl-substituted silanes and between 388 and 425 nm for the mesitoyl-substituted silanes (compare Figures 3 and 4).On the basis of time-dependent density functional theory (TDDFT) calculations, all of these transitions can be straightforwardly attributed to nσ → π* transitions.Consequently, we performed irradiations with light-emitting diodes (LEDs) having emission maxima centered at λ = 405 nm (≈295 kJ mol −1 ) for all compounds to selectively address these transitions.
Photochemistry of 1−6.Tris(trimethoxysilyl)acylsilanes.First, we investigated the photochemistry of tris-(trimethoxysilyl)acylsilanes 1 and 2. On the basis of our spectroscopic investigation, we performed the irradiation experiments at λ = 405 nm in benzene and in the absence of air and moisture.Photolysis of compounds 1 and 2 afforded the same silene rearrangement products as found in previous investigations of structurally related acylsilanes. 25,26In the case of 1, we were able to observe the formation of silene 7a and the corresponding dimerization product 7b by performing NMR spectroscopy after irradiation for only 5 min.Upon prolonged irradiation, the staring material was completely consumed after 120 min and silene 7a and dimer 7b were obtained in a ratio of ∼1:1 with small amounts of unidentified compounds (see Scheme 5).
Further irradiation does not change the ratio of the products but afforded increased amounts of unidentified polymeric decomposition products at the expense of 7a and 7b (see Figure 23).Figure 5 shows the 29 Si NMR spectrum after irradiation for 120 min, including the assignment of the observed resonance lines (all other obtained spectroscopic data that strongly support the structural assignment are shown in the Experimental Section, together with experimental details).Unfortunately, all attempts to separate both compounds failed, but crystals of sufficient quality for X-ray analysis for dimer 7b were obtained by cooling a concentrated solution of the mixture in n-pentane to −30 °C (see Figure 6).As for all other structurally related Brook-type silenes, head-to-head dimerization was observed.
As outlined in Figure 5, the 29 Si NMR spectrum depicts several well-separated shifts, which can be easily assigned to the two corresponding products.For silene 7a, the characteristic signal for the Si�C fragment was observed at δ = 34.3ppm, which is shifted significantly upfield to the corresponding trimethylsilyl-substituted silene isolated by Brook et al. 26    Furthermore, the signals at δ = −36.2ppm and δ = −51.70ppm can be assigned to the Si(OMe) 3 groups.Finally, the OSi(OMe) 3 moiety appeared at δ = −56.2ppm, which is shifted strongly downfield to other structurally related OSi(OMe) 3 groups (see compound 7b).We assume the ylidic resonance structure is responsible for this shift.For dimer 7b, the Si(OMe) 3 signals were assigned at δ = −41.9ppm and δ = −42.0ppm.The signal for the OSi(OMe) 3 appeared at δ = −90.8ppm.Previous experience with structurally related compounds allowed the full assignment of the 29 Si NMR spectrum. 21The 13 C NMR spectrum showed the characteristic shift for the Si�C fragment at δ = 172.9ppm, which is also shifted strongly upfield to the corresponding trimethylsilyl-substituted silene isolated by Brook et al. 26 Again, we assume that the ylidic resonance structure is responsible for this shift.
Again, the photolysis of mesityl-substituted acylsilane 2 was performed at λ = 405 nm in benzene and in the absence of air moisture.After irradiation for 25 min, the starting material was completely consumed.The performed NMR analysis at this stage showed the formation of three products.Again, the resonance lines characteristic for silene 8a and its dimer, 8b, were determined.In contrast to 1, the ratio of this equilibrium is strongly shifted to the dimeric form (∼1:0.05).In addition, silene 8a is photochemically unstable and underwent an intramolecular C−H bond addition reaction of one o-CH 3 groups at the aromatic ring under formation of spirocyclic compound 8c (see Scheme 6).This C−H bond addition reaction is a well-known reactivity of mesityl-substituted Brook-type silenes. 20,25igure 7 shows the 29 Si NMR spectrum after the irradiation for 25 min, including the assignment of the observed resonance lines (all other obtained spectroscopic data that strongly support the structural assignment are shown in the Experimental Section, together with experimental details).
Again, for silene 8a, the characteristic signal for the Si�C fragment was observed at δ = 30.3ppm, and similar shifts for structurally related compounds are found in the same region. 18s all other shifts for the silene and the corresponding dimer are found in the same region as those for compound 7a and 7b, we summarize them in Table 1.
Upon prolonged irradiation, the signals for 8a and 8b slowly vanished, leading to the formation of C−H activation product 8c as the main product.Unfortunately, all attempts to separate 8c from small amounts of uncharacterizable side products failed.However, all other obtained spectroscopic data that strongly support the structural assignment are given in the Experimental Section, together with experimental details.
Our proposed mechanism for silene formation was supported by trapping experiments with MeOH in the presence of Et 3 N.For both acylsilanes, the expected 1,2-MeOH addition products of the silenes were obtained nearly quantitatively (see Scheme 7).Analytical and spectroscopic data that strongly support the structural assignment are summarized in the Experimental Section, together with experimental details.
Tetrakis(trimethylsilyl)-1,4-bisacylsilanes.On the basis of our spectroscopic investigation, we performed the irradiation experiments at λ = 405 nm with compounds 3 and 4 in

Organometallics
benzene and in the absence of air and moisture.In the case of 3, we were able to observe the formation of a monosilene intermediate 10 and small amounts of dimerization product 11 by performing 29 Si NMR spectroscopy after irradiation for 4 h (see Figure 10).These deeply yellow intermediate showed the characteristic Si�C shift at δ = 41.6 ppm, whereas the OSiMe 3 signal can be found at δ = 13.2 ppm.Upon prolonged irradiation, the staring material and the silene intermediate were completely consumed and dimer 11 was obtained alongside with minor amounts of an uncharacterizable polymer (see Scheme 8).The end product showed a high photochemical stability with unchanged NMR spectra over a prolonged period of time.Compound 11 was isolated by crystallization from n-pentane at −30 °C.Analytical and spectroscopic data that strongly support the structural assignment are summarized in the Experimental Section, together with experimental details.
Interestingly, in contrast to the expected head-to-head dimerization of Brook-type silenes, this bissilene undergoes a selective head-to-tail dimerization.
Moreover, we were able to grow crystals suitable for X-ray analysis, which were obtained by slowly evaporating a concentrated solution of 11 in acetone at room temperature.Compound 11 crystallized in triclinic space group P1, and the unit cell contains four molecules (see Figure 8).
Mesityl-substituted bisacylsilane 4 was irradiated for 4 h at λ = 405 nm in benzene and in the absence of air and moisture.The progress was again monitored by 29 Si NMR spectroscopy.
Here we found that the starting material was nearly consumed, and in addition to the selective formation of one product, signals for the Si�C group at δ = 35.6 ppm and for the OSiMe 3 group at δ = 15.61 ppm were observed, indicating a monosilane formation.After irradiation for an additional 1 h, the starting material was completely consumed (see Figure 11).Final product 12 is air stable and was isolated by crystallization from acetone at room temperature in an excellent yield of 64.7% (see Scheme 9).
Moreover, we were able to grow crystals suitable for X-ray analysis, which were obtained by slowly evaporating a concentrated solution of 12 in acetone at room temperature.Compound 12 crystallized in monoclinic space group P2 1/c , and the unit cell contains four molecules.Interestingly, for compound 4, we would assume a C−H activation at both double bonds, as known in the literature, 18,20,25 generating a four-membered ring system at both former silene moieties.In contrast to this assumption, compound 4 undergoes an intramolecular ring closure leading to bicyclic derivative 12, which is confirmed by NMR spectroscopy and X-ray crystallography.Photochemical product 12 contains five chiral centers at Si2, Si5, C1, C4, and C5, with alternating S and R enantiomers.X-ray crystallography could not clarify if the SRSRS or the RSRSR configuration crystallized (see Figure 9).
The lack of important intermediates after the silene formation toward the ring closure made quantum mechanistic calculations necessary to shed light on the mechanism.The mechanism of this photolysis reaction is proposed in Figures  12 and 13.Due to the high energies during the irradiation, isolation and detection of silene formation on both sides are not possible.The Si backbone shows a 108°conformation (A), which favors the following steps.The first reaction step after silene formation is the transfer of one hydrogen of the o-CH 3 group to the double-bonded silicon (TS1).The reaction pathway according to the literature, where first the hydrogen is transferred to the carbon, is unfavored.B exhibits a structure with an acyclic double bond.This double bond interferes with the aromatic ring system leading to a nonplanar formation.Another hydrogen is transferred from the now -CH 2 group to the carbon of the former double bond, leading to a saturated Si−C bond (TS2).C shows the minimum of the carbene with the saturated Si−C bond.next transition, TS3, is the approach and rotation of the carbene to the remaining Si�C bond.Due to the geometry of this intermediate, the partly negatively charged carbon (-CH) interacts (D) with the second double bond, generating the photochemical and air stable bicyclic endproduct 12, containing an eight-and threemembered ring system.
We also performed the trapping experiments in MeOH in the presence of Et 3 N.However, for both compounds the main products were again bicyclic derivatives 11 and 12.We assume that the intramolecular rearrangement also occurs in the presence of MeOH, wehich is the dominant process.
Tetrakis(trimethoxysilyl)-1,4-bisacylsilanes. Compounds 5 and 6 were photolyzed at λ = 405 nm in C 6 D 6 for several hours.In contrast to the other derivatives, silene formation was not detected by NMR analysis.However, we observed different peaks in the O(SiOMe) 3 region, which speaks to the Brook rearrangement step with unstable silene formation and a low activation barrier for the next step.However, no selective products after prolonged photolysis were formed according to NMR spectroscopy.Consequently, on the basis of the experience gained with compounds 1 and 2, we stopped our investigations here.

■ CONCLUSION
To conclude, we were able to synthesize and fully characterize novel mono-and bis(acyl)polysilanes. On the basis of TDDFT calculations, their longest wavelength absorptions can be straightforwardly attributed to nσ → π* transitions.Furthermore, we performed photochemistry with all isolated derivatives at a preparative scale.Compounds 1 and 2 undergo a selective Brook-type rearrangement and form an equilibrium with the corresponding dimers.Moreover, mesityl-substituted silene 8a undergoes a CH activation and forms spirocyclic compound 8c.On the basis of the lack of crystallization, the instability on silica gel made the isolation of the formed photoproducts impossible.However, the proposed mechanism for the silene formation was supported by trapping experiments with MeOH in the presence of Et 3 N.For tetrakis-(trimethylsilyl)-1,4-bisacylsilanes 3 and 4, the same photochemical experiments were performed.Instead of the expected bissilenes, both compounds undergo highly selective hitherto unknown rearrangements.In case of adamantyl derivative 3, we observed an unprecedented head-to-tail intramolecular ring closure of the bissilene as the sole product.In contrast, mesityl derivative 4 undergoes a selective and completely stereoselective double C−H activation to the air stable bicyclic system, which is fully characterized and validated by X-ray crystallography.The mechanism of this rearrangement is fully described by DTF calculations.Finally, tetrakis-(trimethoxysilyl)bisacylsilanes 5 and 6 underwent unselective photochemical rearrangements.

Organometallics
the measurement of air-sensitive samples, benzene-d 6 was additionally dried above a sodium/potassium alloy during a 12 h reflux.Melting points were determined using the Stuart SMP50 apparatus and are uncorrected.Elemental analyses were carried out on a Hanau Vario Elementar EL apparatus.
Irradiation Experiments.Photochemical experiments were performed on a self-made photoreactor with arrays of various LEDs having an emission spectrum centered at 365, 405, 550, and 590 nm.The photoreactor setup comprises 24 LEDs per defined wavelength.Experiments were performed with a total electrical output power of 25 W. Additionally, an integrated cooling system maintained a constant reaction temperature of 23 °C.
UV/Vis Spectroscopy.UV/vis spectra were acquired using a PerkinElmer Lambda 5 spectrometer.
Fourier Transform Infrared (FT-IR) Spectroscopy.The steady-state and time-resolved FT-IR spectra in solution were recorded on a Bruker Alpha spectrometer running OPUS 7.5 software in transmission mode.
NMR Spectroscopy. 1 H, 13 C, and 29 Si NMR spectra were recorded on a Varian INOVA 400, a Varian INOVA 300, a 200 MHz Bruker AVANCE DPX, or a Bruker Avance 300 MHz spectrometer in benzene-d 6 and referenced versus TMS using the internal 2 H lock signal of the solvent.
DFT Calculations.All structures of the investigated compounds have been optimized and verified by vibrational frequency calculations
Synthesis of 3. A solution of (Me 3 Si) 2 KSi 2 Me 4 (SiMe 3 ) 2 K in 20 mL of DME was freshly prepared from 3.00 g (4.91 mmol, 1.00 equiv) of (Me 3 Si) 3 Si 2 Me 4 (SiMe 3 ) 3 and 1.16 g (10.3 mmol, 2.10 equiv) of KO t Bu and slowly added to a solution of 2.05 g (10.3 mmol, 2.10 equiv) of 1-adamantoyl chloride in 50 mL of diethyl ether at −78 °C.Subsequently, the mixture was stirred for an additional 30 min at −78 °C, allowed to warm 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 Na 2 SO 4 and the   . 29Si NMR spectra before and after irradiation at λ = 405 nm for 4 and 16 h, including the assignment of the observed resonance lines for 10 and 11.
Figure 11. 29Si NMR spectra before and after irradiation at λ = 405 nm for 5 h and isolation via crystallization in acetone, including the assignment of the observed resonance lines for 12.
−78 °C.Subsequently, the mixture was stirred for an additional 30 min at −78 °C, allowed to warm to room temperature, and finally stirred for an additional 60 min.Afterward, the volatile compounds were removed under reduced pressure.Again, toluene was added to filter the polymeric residue.Drying under vacuo (0.02 mbar) and recrystallization from a n-pentane solution at −70 °C afforded 2.30 g of 5 (70%) as a white solid.Mp: 181−182 °C.Anal.Calcd (%) for C 38 H 78 O 14 Si 8 : C, 46.40; H, 7.99.Found: C, 46.56; H, 8.20. 29 Photolysis of 1 in C 6 D 6 .First, 57 mg (0.10 mmol) of 1 in C 6 D 6 (0.6 mL) in an NMR tube was photolyzed with the photoreactor at λ = 405 nm for 2 h.At this time, NMR analysis of the resulting clear solution showed that the starting material had been consumed completely.A mixture of silene 7a and dimer 7b could be obtained Scheme 9. Photolysis of 4 Figure 12.Computed reaction profile of the whole process of all intermediates and TSs using the B3LYP/GD3 functional with the 6-31G** basis set.and characterized by 29 Si NMR spectroscopy.Complete conversion to the dimer or separation was not successful.
For 7a. 29  For 7b. 29  Photolysis of 2 in C 6 D 6 .First, 56 mg (0.10 mmol) of 2 in C 6 D 6 (0.6 mL) in an NMR tube was photolyzed with the photoreactor at λ = 405 nm for 25 min.At this time, NMR analysis of the resulting solution showed that the starting material had been consumed completely.At this stage, the signals for three different products (silene 8a, the corresponding dimer 8b, and C−H bond activation product 8c) were observed.Upon prolonged irradiation, silene 8a is completely converted into 8c.After removal of the solvents, 8c could be obtained in nearly quantitative yield as a colorless oil with small amounts of uncharacterizable side products.
Photolysis of 4 in C 6 H 6 /MeOH.First, 150 mg (0.2 mmol) of 4 and 3 drops of anhydrous Et 3 N in 2.0 mL of C 6 H 6 and 1 mL of methanol were photolyzed at λ = 405 nm for 6 h.After removal of the solvents, 12 was obtained as the only characterizable compound.
NMR spectra and UV/vis spectra (PDF) X-ray crystallographic information (XYZ)

Figure 3 .
Figure 3. (a) UV/vis spectra of 1 and 2. Experimental spectra in nhexane at 1 × 10 −3 mol L −1 .(b) Orbitals involved in the first transition for the global minimum of compound 1 (with a contour value of 0.02 au).(c) Orbitals involved in the first transition for compound 2 (with a contour value of 0.02 au).

Figure 4 .
Figure 4. (a) UV/vis spectra of 3−6.Experimental spectra in nhexane at 1 × 10 −3 mol L −1 .(b) Orbitals involved in the first transition for compound 3 (with a contour value of 0.02 au).(c) Orbitals involved in the first transition for compound 6 (with a contour value of 0.02 au).Orbital pictures for the other compounds can be found in Figure S42.

Figure 5 .
Figure5.29 Si NMR spectrum after irradiation at λ = 405 nm for 120 min, including the assignment of the observed resonance lines for 7a and 7b.
Figure5.29 Si NMR spectrum after irradiation at λ = 405 nm for 120 min, including the assignment of the observed resonance lines for 7a and 7b.

Figure 7 .
Figure7.29 Si NMR spectrum after the irradiation at λ = 405 nm for 25 min, including the assignment of the observed resonance lines for 8a−c.

Figure 10
Figure10.29 Si NMR spectra before and after irradiation at λ = 405 nm for 4 and 16 h, including the assignment of the observed resonance lines for 10 and 11.

Figure 13 .
Figure 13.DFT-calculated mechanism with important intermediates and transition states.

Table 1 .
29mmarized29Si NMR Shifts for 7a, 7b, 8a, and 8b Table S1 contains crystallographic data and details of measurements and refinement for all compounds.Crystallographic data (excluding structure factors) have been deposited with the Cambridge Crystallographic Data Centre (CCDC): 2314683 for 3, 2314681 for 4, 2314684 for 7b, 2314680 for 11, and 2314682 for 12.