Abstract

The second-order nonlinear optical coefficients are essential material parameters that determine the performance of nonlinear optical devices. Unfortunately, however, there have been considerable discrepancies among their absolute values reported to date [1]. This confusion, we believe, arises from the following factors. First, quite different values have been reported between parametric fluorescence (PF) and second-harmonic generation (SHG) measurements; e.g., d₃₁(LiIO₃) measured by PF method is 70 % larger than by SHG method [2. 3]. Second, wavelength scaling of nonlinear optical coefficients has been made based on the constancy of Miller’s Δ, although it has not been experimentally confirmed in what wavelength region and to what accuracy Miller’s Δ is constant. Moreover, the interference effect caused by multiple reflection effect in (nearly) plane-parallel samples has been overlooked in almost all the previous measurements, which can be the source of considerable error.

© 1996 Optical Society of America