High-performance source of indistinguishable polarization-entangled photons with a local oscillator reference for quantum networking

TL;DR

The paper tackles the need for a versatile, high-performance source of indistinguishable polarization-entangled photons at telecom wavelengths that includes a local oscillator reference for quantum networking. It presents a free-space, compact platform that co-designs SPDC in an apodized ppKTP crystal, pump/LO generation via a customized CPA-enhanced chain, and a beam-displacer Mach-Zehnder interferometer to switch between polarization entanglement and heralded single photons, all integrated with high-efficiency SNSPDs. Key results include polarization entanglement visibilities near 99% in multiple bases, a Schmidt number $K\approx1.009$ indicating near-spectral separability, a heralded efficiency of $68.0\%$, a successive-photon HOM visibility of $96.3\%$, and an LO-HOM visibility of $88.6\%$, demonstrating strong indistinguishability and LO compatibility. The platform offers a flexible, scalable route toward multiprotocol telecom quantum networks, with potential for entanglement swapping, path-entangled measurements, and network-wide phase locking of lasers.

Abstract

Optical quantum networking protocols impose stringent requirements on the states produced by sources of entanglement. We demonstrate a free-space, compact, source of indistinguishable pairs of polarization entangled photons, with an integrated local oscillator reference as a significant step towards this goal. This source achieves $(99.11 \pm 0.01)\%$ polarization entanglement visibility, $(96.3 \pm 0.6)\%$ successive-photon Hong-Ou-Mandel interference visibility, $(68.0 \pm 0.1)\%$ heralded efficiency as detected, and $(88.6 \pm 0.2)\%$ interference visibility with a local oscillator. This simultaneous achievement of state-of-the-art metrics demonstrates an adaptable platform for quantum networking.

Figures (3)

• Figure 1: a) A simplified depiction of our SPDC source. We use a 980 nm pump laser to excite Er:Yb doped glass in a 100 MHz cavity to produce a broad spectrum seed. This acts as both a stable seed for the down-conversion and the LO reference. We filter and amplify this seed using Martinez-style compressors and an erbium doped fiber amplifier (EDFA). This allows us to match the seed spectrum to the apodized down-conversion crystal and produce a high-power pulse. We then up-convert this 1550 nm light to 775 nm using a commercial magnesium-oxide-doped, periodically poled lithium niobate (MgO:ppLN) crystal. We down-convert this light using apodized periodically poled potassium titanyl phosphate (ppKTP) in a beam displacer Mach-Zehnder interferometer (MZI). This produces indistinguishable polarization-entangled photons that we fiber couple with high efficiency and detect with superconducting nanowire single-photon detectors (SNSPDs). b) An image of the monolithic aluminum mount used to passively align all polarization optics in our down-conversion system. The beam displacers are held in by orange 3D printed tabs. The waveplates are mounted using compact rotation mounts. Finally, the crystal is mounted in a black crystal oven.
• Figure 2: a) Measured joint spectral intensity (JSI) of the photon pairs, obtained using time of flight spectroscopy. The JSI depicts the correlations in emission wavelengths in the signal and idler of generated pairs. We tune the crystal temperature and angle for degenerate photon pairs and low Schmidt number of $1.0089 \pm 0.0002$, calculated using Eq. \ref{['eq:K']}. This implies spectrally separable photons. b) Measured coincidences when we rotate the basis of the second polarization analyzer. The first polarization analyzer is set to the A,D,V,H basis. The change in absolute intensity for different bases is due to the polarization sensitivity of the detectors.
• Figure 3: The results of our successive-photon and single-photon-LO HOM interference measurements. a) The setup we use to interfere consecutive heralded single photons is shown above. The source emits a pair of photons probabilistically, one of which is measured by two herald detectors. The heralded photons then interfere on a beam splitter. b) A plot of the observed HOM dip as a function of delay for a power of $(48.3 \pm 0.3)$ mW. The coincidences are minimized when the two photons arrive simultaneously at the beam splitter. c) Power dependence of the successive-photon HOM. This allows us to infer the purity of our photons if we had ideal number-resolving detectors. The visibility in this case is limited by multi-photon effects. The four-fold coincidences were integrated for a range of approximately $1$ to $6$ hours as the pump power was reduced. d) A depiction of the heralded single photon and LO interference. e) Measured coincidence rate plotted as a function of time delay between single photon and LO. Similar to the successive-photon case, the LO and the single photon interfere, leading to a bunching effect at the output port. In this case the maximum visibility is limited by multi-photon effects in both the heralded single photon and the LO. f) Measured visibility between single photon and LO plotted as a function of pump power, with LO power held constant. This experiment has a much higher rate due to it being a 3-fold detection event instead of four fold and only corresponded to 1 hour for each power.