We report measurement of the photon-counting distributions of nanosecond-duration light pulses. Photon counting is performed on light sources with widely different photon statistics, such as (1) a Q-switched Nd:YAG-laserpumped dye laser operating well above threshold and producing single-mode nanosecond pulses, (2) the same laser operating close to threshold, and (3) a rotating ground-glass aberrator illuminated by the laser pulses of case (1). The method is particularly suitable for the investigation of the statistics of light generated by the nonlinear interaction of radiation with matter.We report measurement of the photon-counting distributions of nanosecond-duration light pulses. Photon counting is performed on light sources with widely different photon statistics, such as (1) a Q-switched Nd:YAG-laserpumped dye laser operating well above threshold and producing single-mode nanosecond pulses, (2) the same laser operating close to threshold, and (3) a rotating ground-glass aberrator illuminated by the laser pulses of case (1). The method is particularly suitable for the investigation of the statistics of light generated by the nonlinear interaction of radiation with matter.

Optics LettersNov 1, 198215 citations

Correlations in Light from a Laser at Threshold

Ren Fang Chang, Victor Korenman, C. O. Alley +1 more

In this study, we measured the temporal correlations in the electromagnetic field radiated by a laser in the threshold region of oscillation, from $\frac{1}{10}$ to 10 times threshold intensity. The experimental results are compared with theoretical predictions based on solutions of a Fokker-Planck equation. We stabilized the intensity of a He-Ne cw gas laser by means of a long-time-constant servo system which controlled the cavity length. Using a fast photomultiplier as detector, we recorded the photoelectron count distribution within a short counting time (3 \ensuremath{\mu}sec) while the photomultiplier was exposed to the laser light. From the photoelectron count distribution measurement, we calculated second, third, and fourth normalized cumulants of the intensity probability density function of the light field. The normalized cumulant is a measure of "pure" correlations among photons because the contributions from lower-order correlations are removed. The statistics of the photoelectron count distribution shows that the intensity fluctuations at about $\frac{1}{10}$ threshold are nearly those of a Gaussian field, and continuously approach those of a constant-amplitude field as the intensity is raised to about 10 times threshold. The normalized second, third, and fourth cumulants of the intensity probability density function of the laser light were also measured at 17 and 42% of threshold intensity as the counting time was increased from 3 to 1000 \ensuremath{\mu}sec. The results agree with predictions computed under the assumption that the dependence of the correlation functions on the time variables is the same as for Gaussian light.In this study, we measured the temporal correlations in the electromagnetic field radiated by a laser in the threshold region of oscillation, from $\frac{1}{10}$ to 10 times threshold intensity. The experimental results are compared with theoretical predictions based on solutions of a Fokker-Planck equation. We stabilized the intensity of a He-Ne cw gas laser by means of a long-time-constant servo system which controlled the cavity length. Using a fast photomultiplier as detector, we recorded the photoelectron count distribution within a short counting time (3 \ensuremath{\mu}sec) while the photomultiplier was exposed to the laser light. From the photoelectron count distribution measurement, we calculated second, third, and fourth normalized cumulants of the intensity probability density function of the light field. The normalized cumulant is a measure of "pure" correlations among photons because the contributions from lower-order correlations are removed. The statistics of the photoelectron count distribution shows that the intensity fluctuations at about $\frac{1}{10}$ threshold are nearly those of a Gaussian field, and continuously approach those of a constant-amplitude field as the intensity is raised to about 10 times threshold. The normalized second, third, and fourth cumulants of the intensity probability density function of the laser light were also measured at 17 and 42% of threshold intensity as the counting time was increased from 3 to 1000 \ensuremath{\mu}sec. The results agree with predictions computed under the assumption that the dependence of the correlation functions on the time variables is the same as for Gaussian light.

Physical ReviewFeb 10, 196979 citations

In this study, we measured the temporal correlations in the electromagnetic field radiated by a laser in the threshold region of oscillation, from

\frac{1}{10}

to 10 times threshold intensity. The experimental results are compared with theoretical predictions based on solutions of a Fokker-Planck equation. We stabilized the intensity of a He-Ne cw gas laser by means of a long-time-constant servo system which controlled the cavity length. Using a fast photomultiplier as detector, we recorded the photoelectron count distribution within a short counting time (3 \ensuremath{\mu}sec) while the photomultiplier was exposed to the laser light. From the photoelectron count distribution measurement, we calculated second, third, and fourth normalized cumulants of the intensity probability density function of the light field. The normalized cumulant is a measure of "pure" correlations among photons because the contributions from lower-order correlations are removed. The statistics of the photoelectron count distribution shows that the intensity fluctuations at about

\frac{1}{10}

threshold are nearly those of a Gaussian field, and continuously approach those of a constant-amplitude field as the intensity is raised to about 10 times threshold. The normalized second, third, and fourth cumulants of the intensity probability density function of the laser light were also measured at 17 and 42% of threshold intensity as the counting time was increased from 3 to 1000 \ensuremath{\mu}sec. The results agree with predictions computed under the assumption that the dependence of the correlation functions on the time variables is the same as for Gaussian light.