Abstract

We propose a new approach to the analysis of stochastic processes with non-Gaussian statistics: the expansion of stochastic processes in terms of first and higher powers of Gaussian components (like a basis set). This approach is applied to treat the non-Gaussian correlation effects observed in coherent anti-Stokes Raman scattering (CARS) measurements performed on the O(18) transition of nitrogen using frequency-doubled pump-laser fields.¹ A proper account of these effects is critical to the accuracy of CARS measurements on species concentrations. By comparison of our calculated results with experimental data at 500 Torr, the field statistics of the pump lasers are first determined and then used to make further predictions at 4500 Torr, where the Raman relaxation time is comparable with the time scale of the pump–laser fluctuations. In both cases, calculations based on this approach give much better agreement with experiments than previous theories.^1,2

© 1988 Optical Society of America