Abstract

Nonlinear optical coefficients are fundamental 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 A comprehensive set of absolute second-order nonlinear optical coefficients for various materials has recently been given by Roberts.² We believe, however, that many of the values cited should be critically reexamined for the following reasons. First, they were obtained by scaling nonlinear optical coefficients measured at different wavelengths through Miller's Δ. But it has not been clear in what wavelength region and to what ac curacy Miller's Δ is constant. Second, the interference effect arising from multiple reflections that occur in second-harmonic generation (SHG) from (nearly) plane-parallel-plate samples has been neglected in almost all previous measurements. Moreover, a large discrepancy between a parametric fluorescence value³ and an SHG value⁴ on LiIO₃ has not been satisfactorily explained.

© 1994 IEEE