Abstract : It is generally taken for granted that the spectrum of light does not change on propagation. This belief is implicit in all spectroscopy. In the last few years it was discovered that while such spectral invariance holds for radiation from commonly used sources it does not hold, in general, because the spectrum of emitted radiation depends not only on the source spectrum but also on the coherence properties of the source. It was found that in some cases the induced changes may be appreciable, possibly resulting in frequency shifts of spectral lines. It has also been demonstrated that similar spectral changes may be produced by scattering on random media whose dielectric susceptibility fluctuations are appropriately correlated. Very recently it was shown that under certain circumstances the frequency shifts generated by dynamic scattering may imitate the Doppler effect, even though the source, the scatterer and the observer are all at rest relative to each other. (emk)
Abstract : It is generally taken for granted that the spectrum of light does not change on propagation. This belief is implicit in all spectroscopy. In the last few years it was discovered that while such spectral invariance holds for radiation from commonly used sources it does not hold, in general, because the spectrum of emitted radiation depends not only on the source spectrum but also on the coherence properties of the source. It was found that in some cases the induced changes may be appreciable, possibly resulting in frequency shifts of spectral lines. It has also been demonstrated that similar spectral changes may be produced by scattering on random media whose dielectric susceptibility fluctuations are appropriately correlated. Very recently it was shown that under certain circumstances the frequency shifts generated by dynamic scattering may imitate the Doppler effect, even though the source, the scatterer and the observer are all at rest relative to each other. (emk)
It is generally taken for granted that the spectrum of light does not change on propagation. This belief is implicit in all spectroscopy. In the last few years it was discovered that while such spectral invariance holds for radiation from commonly used sources it does not hold, in general, because the spectrum of emitted radiation depends not only on the source spectrum but also on the coherence properties of the source. It was found that in some cases the induced changes may be appreciable, possibly resulting in frequency shifts of spectral lines. It has also been demonstrated that similar spectral changes may be produced by scattering on media whose dielectric susceptibility is appropriately correlated.
It was recently shown that, when light is scattered from a random medium whose response function (e.g., the dielectric susceptibility) has appropriate correlation properties, the spectrum of the scattered light may be frequency shifted with respect to the spectrum of the incident light.1,2 In general the relative frequency shifts produced by this mechanism will be different for different spectral lines. We show that under certain circumstances the relative shifts generated by scattering from appropriately correlated space-time fluctuations may become frequency independent and may be arbitrarily large. This scattering mechanism then imitates the Doppler effect, even though the source, the scattering medium, and the observer are at relative rest with respect to each other.
This paper presents a review of research, both theoretical and experimental, concerning the influence of coherence properties of fluctuating light sources and of correlation properties of scattering media on the spectra of radiated and scattered fields. Much of this research followed a discovery made in 1986, that the spectrum of light may change on propagation, even in free space. More than 100 papers on this topic have been published to date and many of them are reviewed, or at least mentioned, in this article. After an introduction and a summary of some of the main mathematical results relating to second-order coherence theory of statistically stationary optical fields, spectral changes that may take place on superposing fields produced by two partially correlated sources are discussed. Spectral effects in fields produced by two-dimensional secondary sources and by three-dimensional primary sources are then considered. The section which follows describes spectral changes that may arise when polychromatic light is scattered on media whose physical properties vary randomly either in space and/or in time. A review is also presented of recent research, which has revealed that under certain circumstances the changes in the spectrum of light scattered on random media may imitate the Doppler effect, even though the source, the medium and the observer are all at rest with respect to one another. In the final section a brief review is given of a new emerging technique sometimes called spatial-coherence spectroscopy. It is based on the discovery that it is possible, under certain circumstances, to determine field correlations from spectral measurements.
The effect of the statistical properties of fluctuating light sources on the spectrum of the emitted radiation has been the subject of intense research over the last few years. It was first shown that due to spatial correlations, the spectrum of light was not, in general, invariant on propagation, even in free space.1 It was also shown that these effects can lead to shifts of spectral lines, a result which is of interest particularly for astrophysics.2 In view of the intimate connection between radiation from localized sources and scattering from media of finite extent, similar spectral changes may also be produced by scattering. In this paper a model scatterer is introduced, whose dielectric response function is a random function of space and time and whose space-time correlation function is a Gaussian function in all its variables. Using the first-order Born approximation, such scattering media are shown to generate frequency shifts of spectral lines that imitate the Doppler effect in its main features. The shifts may be arbitrarily large. Some numerical simulations of such shifts are presented and compared with Doppler shifts.
It has recently been demonstrated both theoretically1-3 and experimentally4 that contrary to general belief, the spectrum of light emitted by a steady-state source depends, in general, not only on the source spectrum but also on the degree of correlation between the source fluctuations. When the source spectrum consists of a single line centered at frequency wo , the source correlation may cause the spectrum of the emitted light to be a line centered on a lower frequency < wo or at a higher frequency w'o > wA, depending on the particular form of the degree of correla?ion. Thus source correlations provide a new physical mechanism for redshifts or blueshifts of spectral lines. This newly discovered phenomenon is obviously-of particular interest for astronomy, where line shifts observed in the spectra of radiation from stellar objects have up to now always been assumed to be caused either by the motion of the stellar object relative to the observer (the Doppler effect) or as being due to gravitation. Examples of redshifted lines due to source correlations are shown in Fig. 1.
There is an implicit assumption in all of spectroscopy that the relative energy distribution in the spectrum of radiation which propagates in free space is independent of the location of the observer, provided that the observer and the source are at rest relative to each other. That this assumption is not valid, in general, was suggested by the results of an investigation of Mandel111, carried out almost thirty years ago. Mandel showed that when light beams from two small correlated sources which have the same normalized spectra are superposed, the normalized spectrum in the region of superposition will differ, in general, from the normalized spectrum of each source. More recently it was predicted theoretically2 and confirmed experimentally soon therwards3 that the normalized spectrum of light produced by an extended source, will, in general, differ from the normalized source spectrum (assumed to be the same at every source point); and that it will depend on the location of the point of observation. The spectral changes were shown to depend on the correlation properties of the source.