Fig. 1. Experimental setup for generating polarization-entangled state by using a four-stage uneven NLI. DSF, dispersion shifted fiber; SMF, single-mode fiber; PBS, polarization beam splitter; QWP, quarter-wave plate; HWP, half-wave plate; FPC, fiber polarization controller; DBF, dual-band filter; PA, polarization analyzer; SPD, single-photon detector.

Fig. 2. (a) Theoretically calculated and (b) experimentally measured joint spectral intensity of the entangled photon pairs.

Fig. 3. (a) Measured count rate of idler photons R_i versus average pump power P_ave. Inset: measured true coincidence C_t versus P_ave. The solid line represents the fitting polynomial R_i = s₁P_ave + s₂P_ave², where s₁ and s₂ are the fitting parameters. The dotted line and dashed line are the linear and quadratic terms, respectively. (b) Measured intensity correlation function of the idler field g⁽²⁾ versus proportion of photons from FWM, R = s₂P_ave²/R_i. Inset: measured g⁽²⁾ versus pump power P_ave.

Fig. 4. Measured (a) single-channel counts and (b) coincidence counts versus the relative angle θ₁–θ₂ of the two polarization analyzers. The solid curve in (b) is the fitting cosine function, which corresponds to a TPI visibility of 92%.

Fig. 5. Spectra of idler photons generated from spontaneous FWM in uneven NLIs. (a) and (b) are the results of three-stage NLI with interstage efficiencies of 40% and 78%, respectively. (c) and (d) are the results of four-stage NLI. The points are measured data, while the curves are fitting results. The visibility of interference fringe V is obtained from the leftmost island in each plot.