Abstract
Bond-length distributions are examined for three configurations of the H(+) ion, 16 configurations of the group 14–16 non-metal ions and seven configurations of the group 17 ions bonded to oxygen, for 223 coordination polyhedra and 452 bond distances for the H(+) ion, 5957 coordination polyhedra and 22 784 bond distances for the group 14–16 non-metal ions, and 248 coordination polyhedra and 1394 bond distances for the group 17 non-metal ions. H⋯O and O—H + H⋯O distances correlate with O⋯O distance (R (2) = 0.94 and 0.96): H⋯O = 1.273 × O⋯O – 1.717 Å; O—H + H⋯O = 1.068 × O⋯O – 0.170 Å. These equations may be used to locate the hydrogen atom more accurately in a structure refined by X-ray diffraction. For non-metal elements that occur with lone-pair electrons, the most observed state between the n versus n+2 oxidation state is that of highest oxidation state for period 3 cations, and lowest oxidation state for period 4 and 5 cations when bonded to O(2−). Observed O—X—O bond angles indicate that the period 3 non-metal ions P(3+), S(4+), Cl(3+) and Cl(5+) are lone-pair seteroactive when bonded to O(2−), even though they do not form secondary bonds. There is no strong correlation between the degree of lone-pair stereoactivity and coordination number when including secondary bonds. There is no correlation between lone-pair stereoactivity and bond-valence sum at the central cation. In synthetic compounds, PO(4) polymerizes via one or two bridging oxygen atoms, but not by three. Partitioning our PO(4) dataset shows that multi-modality in the distribution of bond lengths is caused by the different bond-valence constraints that arise for O(br) = 0, 1 and 2. For strongly bonded cations, i.e. oxyanions, the most probable cause of mean bond length variation is the effect of structure type, i.e. stress induced by the inability of a structure to follow its a priori bond lengths. For ions with stereoactive lone-pair electrons, the most probable cause of variation is bond-length distortion.