Ultratunable Quantum Frequency Conversion in Photonic Crystal Fiber

K. A. G. Bonsma-Fisher, P. J. Bustard, C. Parry, T. A. Wright, D. G. England, B. J. Sussman, and P. J. Mosley

Phys. Rev. Lett. 129, 203603 – Published 10 November 2022

Abstract

Quantum frequency conversion of single photons between wavelength bands is a key enabler to realizing widespread quantum networks. We demonstrate the quantum frequency conversion of a heralded 1551 nm photon to any wavelength within an ultrabroad (1226–1408 nm) range in a group-velocity-symmetric photonic crystal fiber, covering over 150 independent frequency bins. The target wavelength is controlled by tuning only a single pump laser wavelength. We find internal, and total, conversion efficiencies of 12(1)% and 1.4(2)%, respectively. For the case of converting 1551 to 1300 nm we measure a heralded $g^{(2)} (0) = 0.25 (6)$ for converted light from an input with $g^{(2)} (0) = 0.034 (8)$ . We expect that this photonic crystal fiber can be used for myriad quantum networking tasks.

Received 27 July 2022
Accepted 21 October 2022

DOI:https://doi.org/10.1103/PhysRevLett.129.203603

Physics Subject Headings (PhySH)

Photonic crystal waveguide Quantum networks Quantum optics

Atomic, Molecular & OpticalQuantum Information, Science & Technology

Authors & Affiliations

K. A. G. Bonsma-Fisher ¹, P. J. Bustard ^1,*, C. Parry², T. A. Wright ², D. G. England¹, B. J. Sussman^1,3, and P. J. Mosley^2,†

¹National Research Council of Canada, 100 Sussex Drive, Ottawa, Ontario K1A 0R6, Canada
²Centre for Photonics and Photonic Materials, Department of Physics, University of Bath, Bath BA2 7AY, United Kingdom
³Department of Physics, University of Ottawa, Advanced Research Complex, 25 Templeton Street, Ottawa, Ontario K1N 6N5, Canada

^*philip.bustard@nrc-cnrc.gc.ca
^†p.mosley@bath.ac.uk

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 129, Iss. 20 — 11 November 2022

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a) Four-wave mixing process leading to frequency conversion from an input signal at $ω_{s}$ to a target at $ω_{t}$ . (b) The PCF group velocity remains symmetric about the zero dispersion frequency ( $ω_{0}$ ) across an ultrabroad range. Signal and pump fields are group velocity matched on opposite sides of the zero dispersion point; $ω_{p}$ is matched to $ω_{s}$ , and $ω_{q}$ to $ω_{t}$ . By changing $ω_{q}$ (blue shading), the target frequency $ω_{t}$ is tuned across an ultrabroad range (orange shading). (c) Experimental setup. Pump fields $p$ and $q$ , as well as input photon $s$ (see main text for details), are coupled into the PCF. Upon exiting the PCF, converted photon $t$ is collected into SM fiber and sent to SNSPDs, or first to a scanning monochromator (MC). Half-(quarter-)wave plate (HWP, QWP); interference filter (IF); polarizing beam splitter (PBS); dichroic mirror (DM). (d) Scanning electron micrograph of the PCF with a pitch (interhole spacing) of $Λ = 2.11 μ m$ and hole diameter $d = 0.71 μ m$ .
Reuse & Permissions
Figure 2
(a) Coincidence-to-accidental ratio $R_{CA}$ between $t$ and $h$ photons (filled points, left axis) is $≫ 1$ across an ultrabroad range of frequency conversion, including the entire telecoms o- and e-bands (blue and orange shaded regions, respectively). The expected second-order coherence function, $g_{t, t}^{(2)} (0)$ , across the range is plotted (right axis, dashed lines) with the measured data point for $λ_{t} = 1300 nm$ (hollow circle). Uncertainties are derived from Poissonian error in photon detection. (b) Fraction of 1551 nm source light depleted by pumps. Solid lines in (b) show theoretical conversion efficiencies rescaled for comparison with the data.
Reuse & Permissions
Figure 3
Monochromator (a), (b), and pump power (c) scans when $λ_{p} = 787.0$ , $λ_{q} = 872.7 nm$ . In (a)–(c), coincidence counts are between herald ( $h$ ) and converted ( $t$ ) photons (left axis, green squares), noise photons (left axis, red triangles), input ( $s$ ) photons with(out) pumps (right axis, solid (hollow) orange circles). Error bars are derived from Poissonian error in photon detection. Solid and dashed lines show Gaussian fits to data. (a) Converted spectrum at 1300 nm. (b) Input photons, with no pumps present (hollow orange circles) and with pumps present (filled orange circles). (c) Scan of $p$ -pump pulse energy, with $q$ pulse energy fixed at 5.5 nJ. Noise counts are scaled by a factor of 10 for clarity. Solid line is a linear fit to noise data. Noise counts have been subtracted from the depleted signal (filled orange circles, right axis) for direct comparison with the input.
Reuse & Permissions
Figure 4
Detection times of 1551 nm photons (orange), 1300 nm target (green) and noise (red) photons relative to the Pockels cell trigger. The detector jitter time is $\sim 100 ps$ . The 1 ns coincidence window is shown for reference (shaded gray). Inset: Measured spectrum showing broadband noise. Input and target spectra are shown for reference. Here, the input (1551 nm) is an attenuated laser pulse with 1 photon on average.
Reuse & Permissions

Physical Review Letters