Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Statistics of doubly stochastic light : An experimental investigation of photon arrival statistics

Martin Hilbers, Josef Malsher, Ask Olaussen

Published Jan 1, 2026• 0 citations
OpenAlex

Abstract

If a beam of light is coherent, then the photons constituting the beam arrive at random moments in time. Consequently, the photon counting statistics in a given time interval below the coherence time, should in theory follow a Poissonian distribution. When the measurement interval exceeds the coherence time, the distribution should broaden and become a super-Poissonian distribution. The aim of this study is to experimentally demonstrate the Poissonian nature of light, as well as quantifying the transition to a super-Poissonian distribution. In the experiment, a low-intensity laser beam was measured with single-photon detectors, able to capture individual photons. The arrival time of the photons was detected with an oscilloscope, and the data further analysed with our own Python code. The probability distribution of the arrival time of photons was constructed and fitted with a Poissonian distribution for different time intervals. The results showed a clear Poissonian distribution for all intervals from 0.2 μs up to 2000 μs. We did however, observe tendencies of randomness for the longest measured intervals. A Poissonian distribution is defined by the variance being equal to the mean, whilst for a super-Poissonian distribution the variance is higher. The biggest measured ratio between the variance and the mean in our experiment was 1.06 for interval time 2000 μs, which is too small to infer super-Poissonian. Due to limited data memory, our dataset was restricted in length and resolution. Therefore, longer intervals could not be studied without statistical uncertainty. To cross-check the result we also calculated the probability distribution for the arrival time itself. The result perfectly fitted an exponential distribution which was in accordance with the theory of coherent light. Hence, within the given measurement accuracy and statistical uncertainty, it is safe to conclude that the laser beam is coherent for all tested time intervals. However, this does not allow us to conclude that it is a single-mode laser.

Related papers

Photon counting statistics with imperfect detectors

Aleksei Reutov, Сыч Денис Васильевич

Abstract Measurement of photon statistics is an important tool for the verification of quantum properties of light. Due to the various imperfections of real single photon detectors, the observed statistics of photon counts deviates from the underlying statistics of photons. Here we analyze statistical properties of coherent states, and investigate a connection between Poissonian distribution of photons and sub-Poissonian distribution of photon counts due to the detector dead-time corrections. We derive a functional dependence between the mean number of photons and the mean number of photon counts, as well as connection between higher-order statistical moments, for the pulsed or continuous wave coherent light sources, and confirm the results by numerical simulations.

Journal of Physics Conference Series•Dec 1, 2021•0 citations

Time development of photon statistics in fluorescence from single atoms

Theodore Hodapp, M. A. Finn, G. W. Greenlees

Photons spontaneously emitted from single atoms are necessarily antibunched in time. The statistics of the photon time record, however, can show sub-Poissonian or even super-Poissonian statistics and depend strongly on the time interval used for analysis. In this work we have analyzed photon arrival times from resonantly excited barium atoms in an atomic beam observed under single-atom conditions. The time-interval dependence of photon statistics is presented, showing the connection between antibunching and sub-Poissonian statistics (at times comparable to the spontaneous-decay lifetime of the atom) as well as the evolution of the statistics in the long-time limit.

Physical Review A•Oct 1, 1992•2 citations

Sub- and super-Poissonian photon statistics of single-molecule fluorescence blinking

Игорь Сергеевич Осадько

An analysis of intermittent fluorescence is presented for a single molecule driven by a continuous-wave laser field. The interruptions of fluorescence are caused by transition of the molecule to a triplet state. A method previously developed to calculate photon distribution for continuous-wave fluorescence is applied to analyze photon statistics of fluorescence blinking. The probability w N (T) of counting N photons over a time interval T is derived for intermittent fluorescence. The photons counted over relatively short intervals are found to have a sub-Poissonian (narrower than Poisson) distribution. The photon distribution over intervals longer than the mean off time has a complicated form with two maxima; i.e., a super-Poissonian (wider than Poisson) distribution is obtained.

Journal of Experimental and Theoretical Physics•Jun 1, 2007•9 citations

Experimental study of sub-Poissonian statistics in the transfer of momentum from light to atoms

Bruce G. Oldaker, Peter J. Martin, P. L. Gould +2 more

We have studied the statistics of the absorption of photons by atoms by measuring the momentum distribution of an atomic beam that is deflected by a laser beam. We observe the transition from a Poissonian to a sub-Poissonian distribution, as we vary the detuning of the laser. The largest observed fractional deviation of the variance from Poissonian statistics is Q=-0.51\ifmmode\pm\else\textpm\fi{}0.05, in agreement with theoretical predictions for our experimental conditions.

Physical Review Letters•Sep 24, 1990•16 citations

Atoms used to detect non-classical light

Richard Claude Thompson

It is a well known fact that photons behave like particles in many ways. One consequence of this is that any measurements of the intensity of a light beam made by counting photons (e.g. with a photomultiplier tube) in fixed time intervals will show statistical fluctuations. In most cases these fluctuations are characterised by Poissonian statistics, which means that if the mean count is n, the fluctuations are of order n½.

Physics World•Dec 1, 1990•0 citations

Photon antibunching

A. M. Fox

Abstract In the previous chapter we studied how light beams can be classified according to their photon statistics. We saw that the observation of Poissonian and super Poissonian statistics could be explained by classical wave theory, but not sub Poissonian statistics. Hence sub-Poissonian statistics is a clear signature of the photon nature of light.

Oxford University Press eBooks•Apr 27, 2006•2 citations
If a beam of light is coherent, then the photons constituting the beam arrive at random moments in time. Consequently, the photon counting statistics in a given time interval below the coherence time, should in theory follow a Poissonian distribution. When the measurement interval exceeds the coherence time, the distribution should broaden and become a super-Poissonian distribution. The aim of this study is to experimentally demonstrate the Poissonian nature of light, as well as quantifying the transition to a super-Poissonian distribution. In the experiment, a low-intensity laser beam was measured with single-photon detectors, able to capture individual photons. The arrival time of the photons was detected with an oscilloscope, and the data further analysed with our own Python code. The probability distribution of the arrival time of photons was constructed and fitted with a Poissonian distribution for different time intervals. The results showed a clear Poissonian distribution for all intervals from 0.2 μs up to 2000 μs. We did however, observe tendencies of randomness for the longest measured intervals. A Poissonian distribution is defined by the variance being equal to the mean, whilst for a super-Poissonian distribution the variance is higher. The biggest measured ratio between the variance and the mean in our experiment was 1.06 for interval time 2000 μs, which is too small to infer super-Poissonian. Due to limited data memory, our dataset was restricted in length and resolution. Therefore, longer intervals could not be studied without statistical uncertainty. To cross-check the result we also calculated the probability distribution for the arrival time itself. The result perfectly fitted an exponential distribution which was in accordance with the theory of coherent light. Hence, within the given measurement accuracy and statistical uncertainty, it is safe to conclude that the laser beam is coherent for all tested time intervals. However, this does not allow us to conclude that it is a single-mode laser.
Abstract Measurement of photon statistics is an important tool for the verification of quantum properties of light. Due to the various imperfections of real single photon detectors, the observed statistics of photon counts deviates from the underlying statistics of photons. Here we analyze statistical properties of coherent states, and investigate a connection between Poissonian distribution of photons and sub-Poissonian distribution of photon counts due to the detector dead-time corrections. We derive a functional dependence between the mean number of photons and the mean number of photon counts, as well as connection between higher-order statistical moments, for the pulsed or continuous wave coherent light sources, and confirm the results by numerical simulations.
Photons spontaneously emitted from single atoms are necessarily antibunched in time. The statistics of the photon time record, however, can show sub-Poissonian or even super-Poissonian statistics and depend strongly on the time interval used for analysis. In this work we have analyzed photon arrival times from resonantly excited barium atoms in an atomic beam observed under single-atom conditions. The time-interval dependence of photon statistics is presented, showing the connection between antibunching and sub-Poissonian statistics (at times comparable to the spontaneous-decay lifetime of the atom) as well as the evolution of the statistics in the long-time limit.
An analysis of intermittent fluorescence is presented for a single molecule driven by a continuous-wave laser field. The interruptions of fluorescence are caused by transition of the molecule to a triplet state. A method previously developed to calculate photon distribution for continuous-wave fluorescence is applied to analyze photon statistics of fluorescence blinking. The probability w N (T) of counting N photons over a time interval T is derived for intermittent fluorescence. The photons counted over relatively short intervals are found to have a sub-Poissonian (narrower than Poisson) distribution. The photon distribution over intervals longer than the mean off time has a complicated form with two maxima; i.e., a super-Poissonian (wider than Poisson) distribution is obtained.