Stellar Intensity Interferometry

Light from thermal black body radiators such as stars exhibits photon bunching behaviour at sufficiently short time-scales [1]. However, with available detector bandwidths, this bunching signal g(2) is difficult observe directly. We present an experimental technique to increase the photon bunching signal in blackbody radiation via spectral filtering of the light source. Our measurements reveal strong temporal photon bunching from blackbody radiation, including the Sun.

This technique allows for an absolute measurement of the photon bunching signature and thereby a direct statement on the statistical nature of a light source. Such filtering techniques may help revive the interest in intensity interferometry as a tool in astronomy which was first demonstrated by Hanbury-Brown and Twiss [2].

Top | Detector Resolution | Thermal Spectrum | Experimental Setup | Filtered Transmission | Photon Bunching | References

Photon Detector Resolution

Two-photon coincidence measurements of a pair source generated via spontaneous parametric down conversion (SPDC) reveals the timing jitter introduced by the avalanches photon detectors (APDs).
Timing resolution	of two APDs
Typical APD timing resolution is in the GHz regime as shown above: a FWHM of 1.2 ns for the passively quenched APD (CS30902S) and an effective timing resolution of 40 ps for the actively quenched APD (PDF) which constrains the smallest useful time bin for timing correlation and thus g(2) measurements.

Spectrum of Thermal Light Sources

The coherence time scale of thermal photon bunching is approximately inversely proportional to the spectral width of the associated light source.

Spectra of different light sources

For a discrete line spectrum such as that of a Mercury (Hg) discharge lamp, the coherence time of its photon bunching behaviour is approximately 0.5 ns. But for continuous blackbody spectrum such as that of the Sun or for an Arc lamp with temperatures that peak at 6000 K, the associated coherence times is in the regime of 5 fs, which is about four orders of magnitude too small to be directly resolved by the APDs. This makes it difficult for the g(2)(\tau;) of blackbodies such as stars to be directly observed.

Experimental Setup

The coherence time of the light source is increased via spectral filtering through the scheme illustrated below:

Image of experimental setup

We first spatially filter the light via a single mode fiber (SMF) into a single Gaussian transverse mode. The light passes through a two-stage spectral filter. The Glan-Taylor polarizer (PBS) then transmits only linearly polarized light and in conjunction with the half-wave plate, allows us to balance the detector (APD) photon count rates for optimising the duration of data collection. Finally, two-photon coincidences between the two APDs are then correlated as a function of their mutual time delay.

Spectral Filter

The spectral filter is two-stage: firstly, a reflective diffraction grating functions as a grating monochromator and selects a transmission profile with FWHM of 0.12 nm as shown in the top half of the diagram.

spectral resolution of filter

The light is then further filtered down by a solid etalon into a narrowband transmission of 2 GHz bandwidth corresponding to approximately 0.5 ns coherence time. As demonstrated by the lower half of the diagram, the etalon is able to partially resolve the hyperfine structure of a Mercury (Hg) discharge lamp with various isotopes.

Thermal Photon Bunching

Various types of light sources are passed through the filtering scheme, after which temporal g(2) measurements are performed on them as shown in the plots below:

Intensity correlation for different light sources

We thus successfully demonstrate the ability to directly observe temporal photon bunching signature from blackbodies, such as g(2) = 1.37 for the Sun with a coherence time of 0.32 ns. This opens up the possibility into further investigation of intensity interferometry and its applications for stellar measurements. Further details regarding this work can be found here [3].


[1] R.J. Glauber, Physical Review 130, 2529 (1963).
[2] R. Hanbury-Brown & R.Q. Twiss, Nature 178, 1046-1048 (1956).
[3] P.K.Tan, Astrophysical Journal Letters, 789, L10 (2014).