Synthesis and conformational properties of 1,3-dimethyl-3-phenyl-1,3-azasilinane. Low temperature dynamic NMR and computational study

1,3-Dimethyl-3-phenyl-1,3-azasilinane was synthesized and its conformational behavior was studied by the low temperature NMR spectroscopy and quantum chemical calculations. The compound was shown to exist as an equilibrium mixture of the Ph ax Me eq and Ph eq Me ax chair conformers with the N-methyl substituent in equatorial position. The barrier to ring inversion was also determined.


Introduction
Azasilinanes, and in particular 1,4-azasilinanes, are the most studied heterocyclic organosilicon compounds. 14][5][6][7][8] Recently, a renewal of interest in 1,3-azasilinanes has occurred as potential biologically active compounds. 9,10A distinctive feature of bioactive 1,3-and 1,4-azasilinanes is the presence of aryl at silicon and that the biological activity proved to be dependent on the axial or equatorial orientation of the aryl group.The stereochemistry of the carbon analogs, geminally disubstituted 3-Me-3-Ph-piperidines has been studied by 1 H and 13 C NMR spectroscopy, although only at room temperature. 11yclic organosilicon compounds having a Ph-Si bond are of general interest from both synthetic and theoretical point of view.Cleavage of this bond with electrophilic agents provides the possibility for preparing the corresponding sila-functionalized heterocycles such as silacyclohexanes, 1,12 1,3-thiasilinanes 12 and e.g.1,3-azasilinanes. 10We also prepared 3-X-3-Me-1,3-thiasilinanes (X = H, F) from 3-Me-3-Ph-1,3-thiasilinanes 12 in order to study the conformational behavior of the latter thiasilinanes 13,14 as well as of 3-Me-3-Ph-1,3-thiasilinane itself. 14t is a general rule that monosubstituted cyclohexanes prefer the equatorial conformation because of unfavorable 1,3-diaxial interactions in the axial conformation. 15In the certain case of geminally disubstituted cyclohexanes, the substituent of the larger conformational energy (A value) shows a greater preference for the equatorial position.However, even the simplest representative of this class of compounds, 1-methyl-1-phenylcyclohexane, exhibits an exception from this rule: 16,17 the MeeqPhax conformer is 0.32 kcal mol -1 more stable than the MeaxPheq conformer in spite of the larger conformational energy of the phenyl group (2.87 kcal mol -1 ) 15 as compared to the methyl group (1.80 kcal mol -1 ) 15 and the stability of the MeeqPhax and MeaxPheq conformers is reversed again in favor of the latter when the two substituents are attached to silicon, rather than to carbon, as we have exemplified for 1-methyl-1-phenyl-1-silacyclohexane and 3-methyl-3-phenyl-3-thiasilinane. 14n continuation of our systematic studies of the effect of heteroatoms on the conformational preferences of the geminally Si-Me,Ph substituted six-membered heterocycles we report in this paper the synthesis of 1,3-dimethyl-3-phenyl-1,3-azasilinane 1 and the elucidation of its conformational behavior by low-temperature NMR spectroscopy and quantum chemical calculations.

Scheme 1
In the room temperature 13 C NMR spectrum each carbon atom is represented by a single line.Due to the presence of the chiral silicon center in the molecule, protons of methylene groups are diastereotopic: at 273 K, the 4-CH A and 4-CH B resonate separately at 0.85 and 1.00 ppm as multiplets of 8 lines each, the 5-CH2 group exhibiting a very complex multiplet at about 2.00 ppm, the 2-CH2 group proves to be an AB system at 2.10 and 2.15 ppm, respectively, and the 6-CH A and 6-CH B protons also resonate separately at 2.4 and 2.5 ppm as multiplets, the former overlapped with the NMe singlet.With lowering the temperature, all proton signals are broadened and split finally indicating that compound 1 exists as an equilibrium mixture of two conformers shown in Scheme 2. The dynamic 1 H NMR study is given in Figure 1, the resolved 1 H NMR spectrum with assignment of the signals is shown in Figure 2, the 1 H and 13 C signals of the aliphatic part are given in Table 1; aromatic 1 H and 13 C parts of the spectra are congruent.The assignment of signals in the frozen 1 H NMR spectrum (Figure 2) was made taking into account the splitting [axial protons with two (at C4 and C-6) or three large couplings (at C5), 3 J(gem), 3 J(ax,ax)] and the integral intensities of the signals, and using the HMQC and HMBC 2D NMR spectra at 143 K.The ratio of the conformers is ca.2:1.The assignment of the proton and carbon signals to the two preferred conformers 1a and 1b (Figure 2, Table 1) is in agreement with criteria established for 1-methyl-1-phenylcyclohexane, 16 3,5-and 3,6-dimethyl substituted 3-arylpiperidines: 11 (i) equatorial H-2 and H-4 are deshielded in the Me eq Ph ax conformer 1a as compared to those in the MeaxPheq conformer 1b; (ii) axial H-2 and H-4 are shielded in 1a as compared to 1b; (iii) axial 3-Me protons are deshielded by the nitrogen lone pair in 1b as compared to 1a; (iv) the axial SiMe carbon atom in 1b must resonate at higher field than the equatorial analog in 1a.Applying these criteria, the frozen spectra in Figure 2 allow us to conclude that 1a (3-Meeq-3-Phax) is the minor conformer and 1b (3-Meax-3-Pheq) is the major conformer.This assignment was unequivocally proved by the 2D HMBC NMR spectrum: the methyl carbon atom of the major conformer exhibits cross peaks to C(2,4)−H(ax) while the phenyl ipso carbon atom of the minor conformers exhibit the corresponding cross peaks to C(2,4)−H(ax) only; cross peaks between SiMe carbon and Cipso to H(2,4)eq protons, respectively, 6-eq(1b) 6-eq(1a) 2-eq(1a) 2-eq(1b) NMe(1a) NMe(1b) 5-eq(1b) 6-ax(1b) 4-eq(1b) 4-eq(1a) 2-ax(1a) 5-ax(1a) 6-ax(1a) 5-eq(1a)  are not present due to the dihedral angle dependence of the vicinal 3 J(C,H) coupling as origin for the cross peaks.
A lower stability of the 3-Meeq-3-Phax conformer 1a is consistent with the same order of stability earlier found by us for 1-methyl-1-phenyl-1-silacyclohexane and 3-methyl-3-phenyl-1,3-thiasilinane 14 and with the larger conformational energy of the phenyl group.This order of stability is different from that for 1-methyl-1-phenylcyclohexane; 17 the reasons were analyzed in detail previously. 14Possible explanation of the lower stability of 1a vs. 1b comes from consideration of the optimized structure of the molecule (Figure 3).In conformer 1b the plane of the equatorial phenyl group is perpendicular to the mean plane N1-C2-C4-C6 of the 1,3-azasilinane ring system, whereas in conformer 1a the plane of the axial phenyl group is turned about the Si-Cipso bond in such a way that one of the ortho-protons is oriented directly to the lone pair of the nitrogen atom.This results in a reduced contact N•••H ortho of 2.49 Å, which is 0.21 Å shorter than the sum of the van der Waals radii (2.7 Å) and provides an additional steric destabilization of 1a as confirmed by some shortening of the interacting C-Hortho bond as compared to the corresponding noninteracting C-Hortho bond.
The presence of only one NMe signal for each conformer is indicative of the fact that the nitrogen inversion is still fast on the NMR time scale at the temperature of our measurements.
The ratio of the conformers 1a : 1b of 1 : 2 in the frozen equilibrium corresponds to a free energy difference G o = 0.24 kcal mol -1 .
The barrier to ring inversion ΔG ≠ of 9.0 kcal mol -1 was calculated from decoalescence of the SiMe protons, SiMe carbons and C-2, C-4 and C-5 signals, as shown in Table 2.The mean value of the barrier is practically the same as that for the recently studied close analog of 1, 1,3,3-trimethyl-1,3-azasilinane (9.1 kcal mol -1 ). 18The measured barrier is lower than those in N-methylpiperidine (14.4 kcal mol -1 ) 19 or 1,3,3-trimethylpiperidine (10.8 kcal mol -1 ), 20 but markedly higher than that in 1-methyl-1-phenylsilacyclohexane (5.5 -6.0 kcal mol -1 ). 14n order to verify the conclusions about structure and participation of the predominant conformers in the conformational equilibrium made on the basis of the NMR criteria (vide supra), we have performed quantum chemical calculations of conformers 1a and 1b as well as their NMeax counterparts at the MP2/6-311G(d,p) level of theory.The MeeqPhax conformer 1a turned out to be 1.42 kcal mol -1 more stable than 1b.The gas phase free energy difference ΔG o (298 K) = ΔG o (1a) -ΔG o (1b) obtained from vibrational calculations at the same level of theory is twice as low as that and equal to 0.74 kcal mol -1 .This result is opposite to the unequivocal experimental result obtained in freon mixture, which may be attributed to additional stabilization of 1b in solution due to more accessible lone pair of the nitrogen atom in it (Fig. 4).Inclusion of PCM has changed slightly the calculated values but has not affected the trend discussed.Still, both the calculated and experimental values of ΔG o are small, which is consistent with comparable amounts of the two conformers.
The corresponding NMeax conformers are 5.23 and 4.97 kcal mol -1 higher in energy than 1a and 1b, respectively, approving the presence of only one conformer for the given configuration of the substituents at silicon.

Conclusions
1,3-Dimethyl-3-phenyl-1,3-azasilinane 1 exists as an equilibrium mixture of the two conformers with equatorial N-Me substitution and differing by the configuration of the phenyl and methyl substituents at the silicon atom, the PheqMeax conformer 1b being preferable.The ratio of the conformers Ph ax Me eq : Ph eq Me ax is 1 : 2, and the barrier to the silapiperidine ring inversion calculated from decoalescence of five pairs of the 1 H and 13 C signals is 9.0 kcal mol -1 , which practically coincides with the barrier in its close analog 1,3,3-trimethyl-1,3-azasilinane (9.1 kcal mol -1 ).

Experimental Section
General.Solvents were freshly distilled from the appropriate drying agents immediately before use (diethyl ether, sodium benzophenone ketyl; benzene, CaH2).The 1 H, 13 C, and 29 Si NMR spectra were recorded on a Brucker DPX 400 spectrometer ( 1 H, 400.1 MHz; 13 C, 100.6 MHz; 29 Si, 79.5 MHz).Chemical shifts (ppm) were determined relative to residual CHCl3 ( 1 H,  7.27, CDCl3), internal CDCl3 ( 13 C,  77.0, CDCl3), and external TMS ( 29 Si,  0.00, CDCl3).The assignment of the 1 H and 13 C signals was made from the j-mod and 2D{ 1 Н- 13 С} NMR spectra at 298K and 173K.Low-temperature 1 H and 13 C NMR spectra were recorded on a Bruker AV-600 (at 600 and 150 MHz, respectively).Chemical shifts were determined relative to residual internal CD2Cl2 ( 13 C, δ 53.73) and are given in ppm downfield to TMS.Analysis and assignment of the 1 H NMR data were supported by homonuclear (COSY) and heteronuclear (HSQC 13 C-1 H, HMBC 13 C-1 H) 2D correlation experiments.A solvent mixture of CD2Cl2, CHFCl2, and CHF2Cl in a ratio of 1:1:3 was used for the low temperature measurements.The probe temperature was calibrated by means of a thermocouple PT 100 inserted into a dummy tube.The low temperature measurements were estimated to be accurate to ±2 K.The chemical shifts difference c, Hz was determined by extrapolation from the lowest temperatures available to Tc used to calculate kc and the ring inversion barriers by the Eyring equation at Tc.The HRMS ESI spectra were recorded using a Micromass Q-TOFmicro mass spectrometer in positive electrospray mode.All samples were injected (10L/min) with a Harvard syringe pump.The capillary voltage was set to 3.7 kV, with a cone voltage between 20-25 V. Elemental compositions were determined by accurate mass measurement with standard deviation < 5ppm.H3PO4 was used as reference compound.

Theoretical calculations
Optimization of geometry and vibrational calculations were performed at the MP2 level of theory with the 6-311G(d,p) basis set.Unscaled ZPE corrections were used.All calculations were performed with the Gaussian 09 computational program. 23No restrictions on the variation of geometric parameters were imposed during the optimization procedure.For 1a and 1b the calculated MP2/6-311G(d,p) energies are -811.0102473au and -811.0079776au, respectively; ΔE is 1.42 kcal mol -1 and ΔG°(298 K) proves to be 0.74 kcal mol -1 (from vibrational calculations) in the gas phase.The solvent in PCM calculations was CHCl 3 (maximally close to the NMR experiment).PCM calculations were performed with full geometry optimization and changed ΔG°(298 K) only negligibly to 0.79 kcal mol -1 compared with the gas phase value.