“…Equation (15) has been proposed for moderately ionic compounds and its failure, when applied to highly ionic crystals, can be discussed in terms of the non-applicability of the harmonic oscillator approximation in describing the motion of q eff in these compounds (Levine 1973a;Kucharczyk 1995). A different way of obtaining the effective bond charge of LiF, NaF, NaCl, NaBr, KF, KCl and KBr, as derived from experimental third-order nonlinear optical susceptibility, has been presented in (Kucharczyk 1995), where q eff has been found to be closely related with the Szigeti effective charge and the Phillips-Van Vechten ionicity f i . The correlations have been next extended to other alkali halides, thus allowing us to estimate q eff for those crystals for which the third-order optical susceptibility has not been measured yet (Kucharczyk et al 2000).…”