Research ArticlePHYSICS

High-efficiency single-photon generation via large-scale active time multiplexing

See allHide authors and affiliations

Science Advances  04 Oct 2019:
Vol. 5, no. 10, eaaw8586
DOI: 10.1126/sciadv.aaw8586
  • Fig. 1 Multiplexed HSPSs and our implementation.

    (A) Simplified diagram of a general multiplexed HSPS. Multimode SPDC source(s) probabilistically generate photon pairs in which a signal-photon state is correlated to its twin idler-photon state. According to a mode analysis of an idler photon, an adaptive N × 1 optical switch converts a signal-photon state to a predetermined output mode, e.g., time bin. (B) Timing diagram of our time-multiplexed scheme. Our HSPS pumped with a period τ (probabilistically) generates photons in N different time bins. An adjustable delay line can delay signal photons for an arbitrary integer multiple of τ so that any initial time bin state of a heralded photon is converted to a fixed output time bin. (C) Schematic diagram of our experimental setup. See Materials and Methods for experimental details. SHG, second harmonic generation; PPKTP, periodically poled potassium titanyl phosphate; DM, dichroic mirror; FC, fiber coupler; SMF, single-mode fiber; FPGA, field-programmable gate array.

  • Fig. 2 Experimental results.

    (A) heralding signal probability PH, (B) single-photon probability P1, and (C) second-order autocorrelation function g(2)(t = 0) versus the number of multiplexed time bins N. PH and P1 are significantly enhanced as N, while g(2)(0) is approximately unchanged for all different mean photon numbers μ. (D to F) Observed Hong-Ou-Mandel interference for synchronized photons with N = 40. Empty circles, squares, and diamonds in (D) to (F) show data points after subtracting accidental coincidences. Solid and dashed lines are the best-fit theoretical curves (23) for raw coincidences and those after subtracting accidental coincidences, respectively. V denotes the interference visibility without (with) subtracting accidental coincidences. Error bars are estimated by Poissonian photon counting statistics.

  • Fig. 3 Predicted M-photon coincidence production rate CM for different single-photon sources.

    While our time-multiplexed source is less efficient for lower M due to its lower repetition rate, orders of magnitudes better success rates are expected for M > 15.

  • Table 1 Comparison of performances of single-photon sources.

    MUX, multiplexed HSPS; E, single-photon enhancement factor in MUX; I, indistinguishability; CM = P1MR, predicted M-fold coincidence generation rate, assuming that M-independent sources can be prepared and synchronously operated. Note that P1 is the probability of preparing a single photon that is coupled into a single-mode fiber. For sources reported for different experimental parameters, results with conditions demonstrating the highest P1 are shown.

    ReferencesType of sourceR (MHz)PHP1Eg(2)(t = 0)I
    This work (MUX)μ =
    μ =
    μ = 0.0040.50.080.051(2)27.9(20)0.007(7)0.91(4)
    Possible improvementμ = 0.150.990.7512.50.050.99

Supplementary Materials

Stay Connected to Science Advances

Navigate This Article