Fig. 1. Comparison of the second-order correlation of thermal light and ASPDC. (a) Thermal light correlation. The thermal light ET is mixed with vacuum field EV at a beam splitter. Therefore, the light fields E1 and E2 are from the same source even though G1,2(2) is measured at different space–time points. G1,2(2) can be expressed as G1,2(2)=GCor(2)+G1(1)G2(1), where the correlated term GCor(2) comes from the same frequency components in the two split beams, and the background term G1(1)G2(1) comes from different frequency components in the two split beams. (b) ASPDC correlation. The temporal correlation can be measured directly between the signal and idler beams, where the correlated term GCor(2) comes from signal frequency mode ωs and its inherently correlated counterpart in idler mode ωi=ωp−ωs, while the background term G1(1)G2(1) comes from the random combination with ω′s and ω′i≠ωp−ω′s.

Fig. 2. Simulated visibility as a function of gain parameter. The visibility drops from one to about 48.3%, as the parametric process transforms from SPDC to ASPDC. The two vertical bars mark the range of the gain level achieved in our experiment, where the visibility can be considered as constant.

Fig. 3. Schematic of experiment setup. BS, beam splitter; DM, dichromatic mirror; IF, interference filter; FFPC, fiber Fabry–Perot cavity; FPC, fiber polarization controller; FBG, fiber Bragg grating; Ai and Bi, fiber connectors; VOA, variable optical attenuator; HPF, 600 MHz high-pass RF filter (removable).

Fig. 4. ASPDC energy as a function of pump pulse energy.

Fig. 5. (a) Correlation of twin beams as a function of relative time delay, where the red curve shows the background given by the pulse profile of the twin beams. (b) Normalized correlation with exponential decay fit. (c) Visibility as a function of pump pulse energy and transmission of variable optical attenuator (VOA), respectively. (d) Correlation measurement with 600 MHz high-pass RF filters in the photon current.