The time-resolved intensity correlation function for an incident pulse of linewidth σ has a universal form in terms of the square of the time-independent field correlation function and the time-varying degree of intensity correlation, κσ(t). The probability distribution of intensity is determined exclusively by κσ(t).
Introduction In this chapter we will put the concept of coherence on a more mathematical basis. This will involve the formal definition of a number of functions that are used to describe the coherence properties of optical fields. These include the analytic signal, various correlation functions, and the degree of coherence for describing both temporal and spatial coherence. We shall see that the degree of temporal coherence is quantitatively related to the lineshape function and that the degree of spatial coherence between two points is determined by the size, intensity distribution, and location of an illuminating source. We will use the wave equation, and special solutions to the wave equation called Green's functions, to show how spatial coherence varies from point to point. The chapter will conclude with a discussion of how the intensity fluctuations of a source depend on its coherence properties, and we will examine a specific scheme, namely the Hanbury Brown–Twiss experiment, by means of which this relationship is studied. This discussion will involve a discussion of “photon statistics,” namely the time variation of the “detection” of photons from a source. In a quantum-mechanical context, square-law detectors respond to these quantized excitations of the optical field, which we call photons. In classical coherence theory it is advantageous to represent the real electromagnetic field by a complex quantity, both for its mathematical simplicity and because it serves to emphasize that coherence theory deals with phenomena that are sensitive to the “envelope” or average intensity of the field.
In recent years the utilization of "optical mixing" spectroscopic techniques in laser light scattering experiments has proved to be a successful method for the investigation of the dynamical properties of physical systems. By illuminating a photosensitive detector with the scattered light and measuring the spectrum or autocorrelation function of the resulting photocurrent, one can, in general, obtain the ensemble-average time dependence of specific collective excitations within the scattering medium. We present here a detailed quantitative analysis of the statistical errors inherent in such measurements due to the stochastic nature of both the scattering and photoemission processes. We determine the statistical errors on the optical-intensity correlation function as measured by two photocounting digital correlator models and on the intensity spectrum as measured by a "self-beat" optical mixing spectrometer. From these errors and a generalized least-mean-squares fitting procedure we calculate the uncertainty on the measured correlation time (linewidth) for the case of a Gaussian optical field with an exponential intensity correlation function (a Lorentzian spectrum). Scaling relationships are given which permit our numerical results to be applied to an arbitrary set of experimental parameters.
The signal-to-noise ratio (S/N) of the Brown–Twiss stellar intensity interferometer, in which the resolving time (τr) of the photodetectors is much longer than the coherence time of incident radiation (τc), is derived, especially for incident radiation with a very high effective temperature. The statistical properties of the wave amplitude of the incident radiation is assumed to be described by a gaussian distribution function. In order to evaluate the mean square error in the intensity-correlation measurement, the eighth-order moment of complex wave amplitudes is obtained by using the moment theorem for a gaussian random process. The mean square error of the high-intensity correlation experiment during a finite observation time (τ0) is determined chiefly by the random correlation of noise currents due to intensity fluctuations of the incident radiation. The S/N in the case of extremely high-intensity radiation, thus, is given by (τ0/τc)12∣γ12∣2, where γ12 is the complex degree of coherence between two observation points independent of the intensity of incident radiation. In the case of low-intensity radiation the mean square error of an intensity-correlation measurement is determined by the random correlation of shot noise in photoelectric currents. The S/N for low-intensity radiation, as originally discussed by Hanbury Brown and Twiss, is given by (τ0/τr)12∣γ12∣2ηnτc where n is the number of incident photons per second and η is the quantum efficiency of a photodetector. The S/N for the intermediate case is also discussed.