Prediction of Silicon-29 Nuclear Magnetic Resonance Chemical Shifts Using a Group Electronegativity Approach: Applications to Silicate and Aluminosilicate Structures

: Linear relations between group electronegativeity (EN) sums of ligands bonded to tetravalent silicon and silicon-29 nuclear magnetic resonance (NMR) chemical shift (δ_Si) are shown to exist for both type P silicon (all ligands have lone-pair electrons available for (d-p) π-bonding, e.g., in (MeO)₄Si) and type S silicon (all ligands have only σ-bonding electrons available, e.g., in (CH₃)₄Si). For type P silicon having group electronegativity sums greater than 11, a range encompassing all minerals, we have used previously reported EN and δ_Si values (for aryl-, halo-, and alkoxysilanes) to describe the observed silicon-29 NMR chemical shift as δ(Si, P) =-24.336∑EN(P) + 279.27. We then apply this correlation to a wide range of silicates and aluminosilicates (containing insular (Q°) to framework (Q⁴) Si sites) to predict silicon-29 NMR chemical shifts by means of a group fragment electronegativity sum approach, in which all fragments (e.g., OAl, OLi, OCa) attached to Si are assigned, on the basis of experiments on a series of model silicates and the above equation, a characteristic group (or fragment) electronegativity value. OSi group electronegativities are scaled linearly with bridging bond angle. As an example of the use of the method, the electronegativity sum value for the cyclosilicate (Q₂) beryl (Al₂Be₃(SiO₃)₆) is derived as EN(OBe) + EN(OAl) + 2(EN(OSi) (168.2°)) = 15.67, which predicts a silicon-29 chemical shift of-102.1 ppm (from Me₄Si), that compares favorably with the value from experiment, -102.6 ppm. On the basis of a total of 99 sites in 51 different compounds, the mean absolute deviation between theory and experiment is 1.96 ppm (correlation coefficient = 0.979). When all types of silicon are considered (Q°-Q⁴), this empirical approach is the most accurate method of predicting silicon-29 chemical shifts found to date.