Application of linear prediction least-squares fitting to time-resolved optical spectroscopy

Abstract

Linear prediction least-squares fitting is shown to be a powerful tool for the quantitative analysis of pump-probe-type experiments. When used correctly, it is capable of extracting multiple exponential or exponentially damped sinusoidal components from noisy data. Furthermore, it can be used to analyze data that are not simply the response of the system being studied, but rather the convolution of a nonnegligible instrument function with the desired response. In pump-probe experiments, it can be used to deduce time constants that are considerably shorter than the laser pulse itself.