 Fluorescence fluctuation spectroscopy

Fluorescence fluctuation spectroscopy infers molecular properties from variations in detected fluorescence, rather than only from its time-averaged intensity. Diffusion of fluorescent molecules through a small illuminated region changes their number in that region; binding or changes of fluorescent state can change their brightness. Fluorescence correlation spectroscopy studies the time dependence of these fluctuations through an autocorrelation function.^[ref]

For a stationary recorded intensity $I(t)$, put $\delta I(t)=I(t)-⟨I⟩$. A common normalized convention isFor an ideal uniformly weighted observation volume with independent equally bright particles, a Poisson distribution for the particle number gives $G(0)=1/\overline{N}$, excluding the detector's instantaneous shot-noise contribution. The same mean signal could therefore arise from many dim particles or fewer bright ones, while their relative fluctuations differ. Real confocal profiles introduce an effective observation volume and model-dependent shape factors.

A confocal implementation uses a focused laser, spatial filtering of collected fluorescence and photon-sensitive detection to record an intensity trace. Faster diffusion usually shortens the correlation time. Binding kinetics, background, bleaching and detector effects can also modify the trace, so a fitted timescale needs a physical model.^[ref]

The 1972 experiment by Douglas Magde and colleagues measured equilibrium fluctuations during binding of ethidium bromide to DNA. The measured correlations revealed both diffusion coefficients and chemical rate constants. It showed how apparently unwanted fluctuations can expose molecular dynamics without imposing a macroscopic concentration jump.^[ref]