Detectors & Imaging

Photon pairs produce spectral-radiance data

Aug. 1, 1997

FIGURE 1. Light is produced by the process of parametric downconversion in which photons input into a nonlinear crystal, in effect, decay into pairs of photons under the constraints of energy and momentum conservation. In this particular case, a laser beam at 351 nm was sent into a KDP crystal. The central spot is a small amount of 351-nm light leaking around a beam stop. For the case in which the input photons split frequency evenly, two photons at 702 nm (red) are produced; the process also produced other pairs consisting of photons redder and bluer than 702 nm.

Using optical-parametric-downconversion techniques to produce correlated photon pairs, physicists at the National Institute of Standards and Technology (NIST; Gaithersburg, MD) have developed a method to measure absolute infrared (IR) spectral radiance to an initial accuracy of better than 3%. This is the only method directly sensitive to radiance, in contrast to conventional techniques, which use radiant power and aperture geometry to evaluate spectral radiance indirectly. The two-photon method also allows measurement of IR sources using high-performance visible radiometric detectors.

In two-photon downconversion, a nonlinear crystal pumped by a laser beam converts each photon of the pump-laser beam into a pair of lower-frequency photons (the signal and idler beams). The downconverted photons are created simultaneously, so the detection of one photon implies the existence of the other. The photon pairs obey energy- and momentum-conservation laws—energy and momentum information for one photon can be used to determine the energy and momentum of the other. The frequencies of the signal- and idler-beam photons sum to the frequency of the pump-beam photons (wp = ws + wi). A variety of nonlinear crystals can be used to create this effect. By adjusting the crystal optical axis tilt and choosing a particular output direction, the desired measurement wavelength can be selected from a wide spectral range (see Fig. 1).

Spectral-radiance measurements

Two-photon generation can be thought of as a spontaneous decay process, with a single pump photon spontaneously creating the correlated photon pair. If an additional beam with the same frequency and direction as one of the output beams is trained on the optically pumped nonlinear crystal, the beam will enhance the incidence of downconversion, essentially causing stimulated decay of pump photons into correlated photon pairs. The ratio of the output for the stimulated versus spontaneous output can be used to determine the spectral radiance of the stimulation source. This ratio is essentially the radiance of the source in units of photons/mode, which may be thought of as natural absolute units of radiance.

In the NIST system, a 300-mW beam from an argon-ion laser emitting at 457.9 nm pumps a 15 × 15 × 9.7-mm lithium iodate (LiIO₃) crystal (see Fig. 2). Alan Migdall and collaborators configured the system to emit visible and IR correlated photon pairs (0.5288 µm/3.415 µm, 0.5065 µm/4.772 µm) with the visible beam exiting the crystal at approximately 4 from the system optical axis and the IR beam exiting at 25 to 45 from the axis.

FIGURE 2. Absolute infrared spectral radiance can be measured using parametric downconversion in which a nonlinear crystal converts photons of a visible pump-laser beam into correlated visible and infrared photon pairs (spontaneous decay). By properly irradiating the crystal with a beam from the test source, one can trigger stimulated decay of pump photons into correlated photon pairs and capture the output with a high-performance visible wavelength detector. The values for the stimulated versus spontaneous correlated photon pair signals are used to calculate absolute spectral radiance, so no absolute calibration of the detector is required; only detector linearity is needed.

A 0.26-mm-diameter thermoelectrically cooled silicon avalanche photodiode (APD) detects the visible downconverted photons. A focusing lens concentrates the beam, and a pinhole restricts the collection angle of the detector to 0.71 mrad. Narrowband visible interference filters placed in front of the detector determine the IR spectral bandwidth of the radiance measurement.

The NIST group used the system to calibrate an argon arc-discharge source previously measured by conventional techniques. By focusing the arc-discharge source output beam onto the pump region in the LiIO3 crystal and angling the beam to overlap the beam of IR downconverted photons, Migdall was able to trigger stimulated decay of pump photons into downconverted IR/visible photon pairs. The APD detected the visible downconverted photons, which correlate directly with the infrared photons.

The actual spectral radiance can be calculated from a ratio based on the correlated visible photon signal produced with and without the input from the arc-discharge source beam. Initial measurements agreed with the conventional measurement system to within 3%. While some of this difference stems from the new method, a significant portion is caused by uncertainties in the conventional measurement itself. Migdall expects to reduce the error to less than 1% as the technique is optimized.