Time-multiplexed measurements of nonclassical light at telecom wavelengths

We report the experimental reconstruction of the statistical properties of an ultrafast pulsed type II parametric down-conversion source in a periodically poled potassium titanyl phosphate waveguide at telecom wavelengths, with almost perfect photon-number correlations. We use a photon-number-resolving time-multiplexed detector based on a fiber-optical setup and a pair of avalanche photodiodes. By resorting to a germane data-pattern tomography, we assess the properties of the nonclassical light states with unprecedented precision.

Figure 1

(a) Probe-state generation. Pulsed light at telecom wavelengths is generated in a Ti:sapphire pumped optical parametric oscillator (OPO). The repetition rate of the reference is lowered by an acousto-optic modulator (AOM). Motorized half-wave plates (HWP) followed by polarizing beam splitters (PBS) are used to vary the attenuation. The light power is measured by power meters (PM), further attenuated by neutral density filters (NDF), and coupled into the single-mode fibers of the time-multiplexing detector (TMD). (b) Nonclassical state generation. The pump light is spectrally tailored by a $4f$ system, coupled into the KTP waveguide (WG) and blocked by a long pass filter. The SPDC photons pass a bandpass filter (BPF), are separated by a PBS and coupled into the TMD. (c) TMD. Two input pulses are distributed into 16 bins to obtain information about the photon number in each input as well as photon-number correlations between the beams.

Figure 2

Average infidelity versus the dimension of the reconstruction space, $n_{\max }$. The number of patterns employed is indicated by symbols: squares (50 patterns), triangles (80 patterns), and circles (150 patterns). Open symbols are used for standard detector tomography and solid ones for data-pattern tomography.

Figure 3

Two-mode reconstruction of the ${\mathrm{SPDC}}_2$ state. Left: Complete photon-number distribution, with $M=50$ probes. Right: Comparison with the theoretical model. Circles with error bars represent the experimental results, whereas theory is indicated by squares.

Figure 4

Single-mode signal reconstructions (dots) of ${\mathrm{SPDC}}_1$ (left) and ${\mathrm{SPDC}}_2$ (right), both with $M=30$ probes. Best fits to Bose-Einstein distributions (squares) are also shown. The reconstruction uncertainties are almost negligible and cannot be appreciated.

Figure 5

Reconstructed single-photon (left) and two-photon (right) heralded idler states generated from ${\mathrm{SPDC}}_1$ (top) and ${\mathrm{SPDC}}_2$ (bottom), with $M=80$ probes. Squares denote again the corresponding theoretical predictions.

Figure 6

Heralded single-photon (left) and double-photon (right) idler states as predicted from the reconstructed two-mode photon-number distributions of ${\mathrm{SPDC}}_3$, with $M=80$ probes.