A PROPOSED "NOISE" LEVEL FOR THE DISCRIMINATION BETWEEN INTRA- AND INTERMOLECULAR GEOMETRIC EFFECTS IN PT(II)-COMPLEXES. Åke Oskarsson, Karin Lövqvist, Lars-Ivar Elding, Inorganic Chemistry 1, Chemical Center, University of Lund, P. O. Box 124, S-22100 Lund, Sweden

Crystallographic data are often directly used as a basis for conclusions of chemical significance. For example, observed differences in Pt-Cl bond lengths in different structures may be used to evaluate cis- and trans-influence in square-planar complexes. However, crystallographic bond lengths are not determined by intramolecular effects alone but also packing effects. To establish a reasonable level of significance for an evaluation of an intramolecular "chemical" effect, it is necessary to consider the combined outcome of the experimental errors and packing effects. We have determined each of the crystal structures of [PtCl(dms)₃](PF₆) and [PtCl(dms)₃](CF₃SO₃) from two independent data sets using different crystals. Comparisons of the two data sets for one compound using half-normal probability plots revealed only small systematic errors and the e.s.d.'s are not under-estimated more than 15-20 %. Cross-comparisons of the geometry in the Pt(II)-complex in the different compounds gave "noise" levels for differences in Pt-Cl and Pt-S distances of 0.014 and 0.024 Å for 95 and 99% confidence interval, respectively. Therefore we propose that observed differences of Pt-Cl or Pt-S distances should be at least 0.02 Å to be claimed to be the result of intramolecular effects.