Statistical accuracy in fluorescence correlation spectroscopy

Abstract

Fluorescence correlation spectroscopy is a promising technique for the study of chemical kinetics and diffusion processes in small, well-defined sample volumes. Laser-induced fluorescence is used as a specific highly sensitive probe of concentration, permitting the analysis of the characteristic time dependence of spontaneous concentration fluctuations. We present here a quantitative analysis of the statistical and systematic errors inherent in such an experiment. Theoretical expressions are derived for the signal-to-noise ratio for various operational definitions of the photocount correlation signal. It is shown that the critical parameter governing the statistical accuracy of the experiment is not the total number of photocounts per correlation time, but (when background fluorescence is negligible) the photocounts per correlation time, per fluorescent molecule, a parameter independent of the number of fluorescent molecules in the sample volume. Comparisons are made with parallel results for laser-light scattering experiments derived by other authors. Finally, we consider the effects of background fluorescence, photolysis, and other deviations from ideality.