Abstract
The dependence of Si-O(br) and Si-O(nbr) bond lengths on Si-O-Si angle and average electronegativity, x, of non-tetrahedral cations has been examined for four C2/m amphiboles and a number of other silicates with both bridging, O(br), and non-bridging, O(nbr), oxygens. When both x and Si-O-Si angle are large, the mean Si-O(nbr) bond lengths, <Si-O(nbr)>, are observed to be longer than the mean Si-O(br) bond lengths, <Si-O(br)>. However, when x is small, <Si-O(nbr)> are usually shorter than or equal to <Si-O(br)>, irrespective of angle. A correlation of <T-O> bond length (T ≡ Si,Al,P) with O … O non-bonded distance in O—T—O linkages for compounds with SiO4 as well as AlO4 and PO4 tetrahedra is consistent with Cruickshank's (1961) π-bonding hypothesis, with McDonald and Cruickshank's (1967) assumption that O(nbr) have larger negative charges than O(br) and with Gillespie's (1963) VSEPR Theory.
The distributions of tetrahedral Si and Al, B, Be or Mg in a number of framework silicates are examined in terms of Cruickshank's hypothesis and the difference in predicted Si-O, Al-O, B-O, Be-O and Mg-O bond orders for TO4n− tetrahedral ions and are found to be consistent with the proposal that Si should prefer those tetrahedra involved in the widest average T-O-T angles. Although the distribution of tetrahedral cations in the nonframework silicates cannot in general be explained in terms of this proposal, other factors involving d-pπ-bonding in TO4n− ions are helpful in rationalizing the observed tetrahedral distribution of cations in the pyroxenes, hornblendes and micas. The role of OH and F in this ordering process is also briefly discussed.
