Studies of complex formation from permittivity measurements: a statistical reexamination

Abstract

Permittivity measurements on solutions of two compounds present in a variable ratio may reveal complex formation: both the equilibrium constant and the dipole moment of the complex can be in principle estimated. However, an incorrect statistical treatment involving transformation of variables is used quite often, for example, in the classical method of Few and Smith. In this case the estimates obtained are usually biased, or loaded with such a great uncertainty that they are practically worthless. The results of numerous papers, even some recent ones, are thus completely at variance with facts. This is shown in this paper by recalculating equilibrium constants and dipole moments of the complex in 17 examples from the literature. Simultaneously, a correct and simple statistical procedure is suggested. The danger of an incorrect and (or) incomplete statistical treatment may be encountered, for the same reason, even in other, quite different, areas.