In 1963 Glauber /Phys. Rev. 130, 2529 and 131, 2766/ published a basic theoretical study that revealed the broad variety of possible photon states. Later on a quantum-statistical analysis showed that states with classical analog (this light Is termed classical light) as well as states without classical analog (so-called nonclassical light) can occur. Following the definition given by Teich, Saleh, and Perina /JOSAB2 (1985) 275) nonclassical light exhibits one or more of the attributes squeezed, antibunched, sub-Poissonian, what is connected with a negative value of the field-fluctuation excess, the bunching excess, or the Poisson excess. These negative values mean quantum fluctuations below the classical standard quantum limits of ideal laser light; thus, nonclassical light may yield an essential improvement of the optical measurement accuracy, in particular of the signal-to-noise ratio. There exists a common property of the three attributes concerning the Glauber-Sudarshan representation P(c ), which is the real weight function belonging to the incoherent mixture (density operators) of Glauber states with different complex amplitudes. In the case of classical light P(oç ) is a positive definite function, whereas nonclassical light requires a nonpositive definite function P(o.), which cannot be interpreted as a proper probability distribution - as it is possible in the case of light with a classical analog /M. Schubert, Ann. d. Phys. 4k (1987) 53/. In addition to its possible important applications nonclassical light yields a deep insight into the quantum nature of the radiation field.
In 1963 Glauber /Phys. Rev. 130, 2529 and 131, 2766/ published a basic theoretical study that revealed the broad variety of possible photon states. Later on a quantum-statistical analysis showed that states with classical analog (this light Is termed classical light) as well as states without classical analog (so-called nonclassical light) can occur. Following the definition given by Teich, Saleh, and Perina /JOSAB2 (1985) 275) nonclassical light exhibits one or more of the attributes squeezed, antibunched, sub-Poissonian, what is connected with a negative value of the field-fluctuation excess, the bunching excess, or the Poisson excess. These negative values mean quantum fluctuations below the classical standard quantum limits of ideal laser light; thus, nonclassical light may yield an essential improvement of the optical measurement accuracy, in particular of the signal-to-noise ratio. There exists a common property of the three attributes concerning the Glauber-Sudarshan representation P(c ), which is the real weight function belonging to the incoherent mixture (density operators) of Glauber states with different complex amplitudes. In the case of classical light P(oç ) is a positive definite function, whereas nonclassical light requires a nonpositive definite function P(o.), which cannot be interpreted as a proper probability distribution - as it is possible in the case of light with a classical analog /M. Schubert, Ann. d. Phys. 4k (1987) 53/. In addition to its possible important applications nonclassical light yields a deep insight into the quantum nature of the radiation field.
Abstract In the present paper the attributes of nonclassical light with respect to the theoretical basis, the experimental verification, and the applications are discussed and compared. If one ore more of the corresponding statistical measures (fluctuation excess, bunching excess, Poisson excess) attain negative values, surprising new properties of photon states occur. The analysis shows that under certain temporal and interaction conditions squeezing, antibunching, and sub‐Poisson behavior can accompany one or more of the other attributes. Also the connection of the attributes with the Glauber‐Sudarshan P‐representation of the corresponding states is studied and yields important assertions.
There is a common understanding in quantum optics that nonclassical states of light are states that do not have a positive semidefinite and sufficiently regular Glauber-Sudarshan P function. Almost all known nonclassical states have P functions that are highly irregular, which makes working with them difficult and direct experimental reconstruction impossible. Here we introduce classes of nonclassical states with regular, non-positive-definite P functions. They are constructed by ``puncturing'' regular smooth positive P functions with negative Dirac-$\ensuremath{\delta}$ peaks or other sufficiently narrow smooth negative functions. We determine the parameter ranges for which such punctures are possible without losing the positivity of the state, the regimes yielding antibunching of light, and the expressions of the Wigner functions for all investigated punctured states. Finally, we propose some possible experimental realizations of such states.
“The word ‘classical’ means only one thing in science: it's wrong!” We have previously emphasized the fact that all states of light are quantum mechanical and are thus nonclassical, deriving some quantum features from the discreteness of the photons. Of course, in practice, the nonclassical features of light are difficult to observe. (We shall use “quantum mechanical” and “nonclassical” more or less interchangeably here.) Already we have discussed what must certainly be the most nonclassical of all nonclassical states of light – the single-photon state. Yet, as we shall see, it is possible to have nonclassical states involving a very large number of photons. But we need a criterion for nonclassicality. Recall that in Chapter 5 we discussed such a criterion in terms of the quasi-probability distribution known as the P function, P (α). States for which P (α) is positive everywhere or no more singular than a delta function, are classical whereas those for which P (α) is negative or more singular than a delta function are nonclassical. We have shown, in fact, that P (α) for a coherent state is a delta function, and Hillery has shown that all other pure states of the field will have functions P (α) that are negative in some regions of phase space and are more singular than a delta. It is evident that the variety of possible nonclassical states of the field is quite large.
Nonclassical effects such as squeezing, antibunching, and sub-Poissonian statistics of photons have been attracting attention in quantum optics over the last decade. Up to now most theoretical and experimental investigations have been carried out exclusively in the time domain while neglecting the spatial aspects by considering only one spatial mode of the electromagnetic field. In many situations such an approximation is well justified. There are, however, problems that do not allow in principle a single-mode consideration. This is the case when one wants to investigate the quantum fluctuations of light at different spatial points in the plane perpendicular to the direction of propagation of the light beam. Such an investigation requires a complete description of quantum fluctuations of light in both time and space and cannot be done within a single-mode theory. This space-time description brings about a natural generalization into the spatial domain of such notions as the standard quantum limit, squeezing, antibunching, etc. It predicts, for example, the possibility of generating a light beam with sub-Poissonian statistics of photons not only in time but also in the beam's transverse plane. Of particular relevance to the applications is a situation in which the cross section of the light beam contains several nonoverlapping areas with sub-Poissonian statistics of photons in each. Photodetection of such a beam produces several sub-shot-noise photocurrents depending on the number of independent areas with sub-Poissonian statistics. This is in marked contrast to the case of a single-mode sub-Poissonian light beam in which any attempt to collect light from only a part of the beam deteriorates the degree of shot-noise reduction. This property of multimode squeezed light opens a range of interesting new applications in optical imaging, optical parallel processing of information, parallel computing, and many other areas in which it is desirable to have a light beam with regular photon statistics across its transverse area. The aim of this review is to describe the recent development in this branch of quantum optics.
Although light quanta were known since the works by Planck and Einstein, for several decades after these works, there was no need to treat light quantum mechanically, and all experiments could be well explained using the quantum description of matter and classical description of light (the semiclassical approach). The situation changed with the development of sources of so-called nonclassical light, i.e., light whose properties cannot be understood in the framework of the semiclassical approach. Two simple examples can be given here. The first one is the suppression of shot noise in optical detectors. Shot noise, i.e., fluctuations of photocurrent scaling as the square root of the photocurrent mean value, can be, in principle, attributed to the discrete nature of the charge generated by the detector. At the same time, the effect of shot-noise suppression, which is observed with certain light sources (squeezed light) is clearly related to the statistics of light, and quantum mechanics is completely necessary for its description. Another example, which brings us closer to the subject of the present paper, is two-photon light. This type of light contains only pairs of photons (often called biphotons), and hence, a pair of properly aligned photon-counting detectors registering two-photon light, in an ideal case, will produce only simultaneous photocounts. This behavior, again, can only be explained in terms of a quantum-mechanical description for the light.
The definition of nonclassical states in quantum optics by the nonpositivity of their Glauber–Sudarshan quasiprobability is investigated and it is shown that it hides some serious problems. It leads to a subdivision of squeezed thermal states into classical and nonclassical states which is difficult to interpret physically by some qualitatively different behaviour of the states. Nonclassical states are found in arbitrarily small neighbourhoods of every classical state that is illustrated by a very artificial modified thermal state. The observability of the criterion in comparison to that for nonclassicality of states determined by the nearest Hilbert–Schmidt distance to a class of reference states is discussed. The behaviour of the nonclassicality of states in models of phase-insensitive processes of damping and amplification is investigated and it is found that every nonclassical state eventually makes a transition to a classical state. However, this is not specific for the negativities or singularities of the Glauber–Sudarshan quasiprobability and is found in similar form for other quasiprobabilities, for example, for the Wigner quasiprobability. We discuss in quite general form some defects of the Glauber–Sudarshan quasiprobability if compared with classical distribution functions over the phase space, in particular the failure of an earlier advertised superposition formula.