Previous page     Contents     Next page

Dynamic solid-state properties and phase transitions in organosilicon compounds (X. Helluy, J. Kümmerlen and A. Sebald, in collaboration with W.A. Dollase/Los Angeles and R. Dinnebier/Bayreuth)

Many organosilicon compounds are non-polar, nearly spherically shaped molecular solids, and display concomitant dynamic solid-state properties. The molecular compounds E(SiMe₃)₄, with E = C,Si belong to this category and exist as crystalline solids at ambient temperatures. Previously, only the gas phase structures had been determined for these two compounds and both have molecular T symmetry. For E = Si the crystal structure has now been determined to have space group Fm3m by Rietveld refinement of powder X-ray diffraction data obtained at room temperature. Both compounds E(SiMe₃)₄ undergo structural phase transitions in the temperature range 200 to 220 K. In the respective low-temperature phases the molecular dynamic properties of E(SiMe₃)₄ were characterized by variable-temperature ¹³C and ²⁹Si one- and two-dimensional CP/MAS NMR. Quantitative evaluation of the respective exchange rate constants for molecular reorientation and for internal reorientation of the SiMe₃ groups as a function of temperature reveals striking differences in the dynamic properties of these two chemically closely related methane homologues. For Si(SiMe₃)₄ at low temperatures, internal reorientation of the SiMe₃ groups is observed, and at higher temperatures additional reorientation of the entire molecule around its preferred molecular axes takes place. For C(SiMe₃)₄, the thermally activated modes of reorientation in the solid state are quite different: over the temperature range of the low-temperature phase, no internal reorientation of the SiMe₃ groups takes place, and only reorientation of the entire molecule is observed. A possible explanation for this difference in thermally activated molecular dynamics is offered by force field calculations, based on the gas-phase structures of the two molecules: due to higher intramolecular steric hindrance in C(SiMe₃)₄, whole-molecule reorientation may be the preferred reorientational mode for this compound (Fig. 3.8-10).
 

Fig. 3.8-10: a) schematic illustration of reorientational modes in solid E(SiMe3)4 b) 13C 2D exchange spectra for E = C (left) and E = Si (right) c) force-field calculations of activation barriers for internal SiMe3 reorientation for E = C (left) and E = Si (right).

Numerically exact spectral lineshape simulations for isolated 2- and 3-spin systems under MAS conditions (S. Dusold and A. Sebald)
 

Isolated 2- and 3-spin systems occur (either naturally, or man-made by isotopic labelling) in many different chemical compounds, ranging from biochemicals to inorganic solids. One class of chemicals where such spin systems naturally occur are transition-metal phosphine complexes (the M-³¹P and M(³¹P)₂ fragments, where M represents a spin-1/2 isotope such as ¹¹³Cd, ¹⁹⁵Pt, or ¹⁹⁹Hg). The Hamiltonian appropriate to describe isolated dipolar coupled 2- and 3-spin systems under MAS conditions does not commute with itself at different times. The practical consequences in MAS NMR spectra of such spin systems are observed as complicated spectral lineshapes, strongly dependent on the external magnetic field strength and applied MAS frequency. Interpretation of such MAS NMR spectra "by inspection" is not possible. Their spectral lineshapes not only depend on the magnitudes of NMR interactions but also depend on the mutual orientations of the various interaction tensors. Hence, numerically exact spectral lineshape simulations of MAS NMR spectra of polycrystalline powder samples can be used to yield information about the orientation of e.g. chemical shielding tensors in the principal axes system of a molecule or crystal, or in the determination of internuclear distances. Conventional MAS NMR of polycrystalline powder samples, in conjunction with numerically exact spectral lineshape simulations thus provide information which would otherwise require NMR experiments on oriented single crystals. A complete analysis of the MAS NMR spectra of isolated dipolar coupled 2- and 3-spin systems lifts a severe limitation with regard to sample availability for solid-state NMR, as no single crystals are necessary. With our focus on optimizing efficiency, we have developed computational methods that allow for high-quality iterative fitting of the experimentally observed MAS NMR spectral lineshapes. We illustrate this technique with the example of double- and triple-resonance experimental and calculated ¹¹³Cd CP/MAS NMR spectra of a representative cadmium-phosphine complex (Fig. 3.8-11).
 

Fig. 3.8-11: Experimental (left) and calculated (right) 22.2 MHz 113Cd MAS NMR spectra of Cd(NO3)2*2PPhMe2;  a) conventional 113Cd CP/MAS NMR spectrum;  b) 113Cd CP/MAS NMR spectrum with additional high-power 31P decoupling; c) 113Cd CP/MAS NMR spectrum with additional low-power 31P recoupling.  The most shielded component of the 31P chemical shielding tensors is typically oriented close to the P-Cd bond direction in the molecular frame.

Unambiguous confirmation of space group Pa3 for the pseudo-cubic phase of SiP₂O₇ by means of ³¹P NMR (J. Kümmerlen and A. Sebald, in collaboration with R. Iuliucci and B.H. Meier/Nijmegen)

The single-crystal X-ray structure of the pseudo-cubic phase of SiP₂O₇ has been refined by simulation and space group Pa3 has been proposed as the correct solution of the structure determination in the literature. This space group requires the presence of eleven crystallographically inequivalent P sites in the asymmetric unit, and for reasons of symmetry, two of the P₂O₇ units must contain linear P-O-P moieties. The validity of the proposed structure solution has been questioned because of the required presence of supposedly unfavorable linear P-O-P arrangements (amongst other, non-linear P-O-P arrangements). In order to resolve this remaining ambiguity concerning the correct space group of the pseudo-cubic phase of SiP₂O₇, ³¹P MAS NMR experiments which only report distance connectivities but are widely insensitive to chemical shifts, must be carried out. Such solid-state rf-driven spin diffusion ³¹P NMR experiments have been carried out, and the distance connectivities (short, intermediate, long) as seen by ³¹P NMR are compared to the connectivities as required by the proposed space group from single-crystal X-ray diffraction. There are 11! = 39916800 assignment permutations for the various ³¹P resonances to the eleven inequivalent P sites in the asymmetric unit. By using formalised "short, intermediate, and long" distance constraints from the ³¹P solid-state NMR experiments, space group Pa3 can be confirmed by NMR, including an unambiguous assignment of the ³¹P resonances to the respective P sites in the asymmetric unit. Figure 3.8-12 shows a ³¹P MAS NMR spectrum of SiP₂O₇, and the resulting model-free assignment of ³¹P resonances (A-K) to the crystallographic P sites (1-11).

Determination of anisotropy of long-range J-coupling by ¹¹⁹Sn OMAS NMR (C. Marichal and A. Sebald)
 

Isotropic J coupling, an interaction transferred through electrons in chemical bonds, is widely exploited in liquid state NMR for purposes of structure elucidation. Like all other NMR interactions, however, J coupling is an anisotropic interaction, described by a second-rank tensor, such that in nonviscous solutions it is motionally averaged to its isotropic value. In principle, anisotropy of J in crystalline solids should be observable. However, J coupling interactions in solids are difficult to observe experimentally as they are usually masked by the simultaneous presence of multiple dipolar couplings much larger in magnitude. So far, only few attempts to determine the anisotropy of J coupling, mostly by means of single-crystal NMR, have been reported in the literature. All of these experimental studies have considered J coupling over a single bond, and to the best of our knowledge no experimental data on the anisotropy of long-range J coupling over more than one bond have been reported. We have
 

Fig. 3.8-12: 31P MAS NMR spectrum of SiP2O7, indicating the final assignment per-mutation of 31P reso-nances to the crystal-lographic P sites in the asymmetric unit.

investigated the anisotropy of ²J(¹¹⁹Sn, ¹¹⁷Sn) in solid (benzyl₃Sn)₂O, using ¹¹⁹Sn off-magic-angle spinning (OMAS) NMR techniques in conjunction with spectral lineshape simulations. Our finding of considerable anisotropy of ²J(¹¹⁹Sn, ¹¹⁷Sn) bears relevance on various other solid-state NMR techniques, designed to determine internuclear distances from NMR experiments: if J anisotropy is present but not taken into account in the analysis of such NMR experiments, the internuclear distances derived will be inaccurate. Figure 3.8-13 shows a comparison of the experimental and the calculated ¹¹⁹Sn OMAS NMR spectra from which we derive an anisotropy of ²J(¹¹⁹Sn, ¹¹⁷Sn) = 850 ± 300 Hz, similar to the magnitude of | ²J_iso(¹¹⁹Sn, ¹¹⁷Sn)| = 955 Hz.
 

Fig. 3.8-13: 119Sn OMAS NMR spectra of (benzyl3Sn)2O; top: calculated spectrum, bottom: experimental spectrum.