An all-fibred, telecom technology compatible, room temperature, single-photon source

TL;DR

This work reports a fully fibre-coupled, room-temperature single-photon source based on GaN color centers, emitting at 1292 nm in the telecom O-band and compatible with CWDM channels. By combining a stable all-fibre excitation/collection scheme with off-the-shelf telecom components, the authors achieve a high single-photon purity of $g^{(2)}(0) = 0.059$ and a practical brightness of 25 kcps/mW, with a saturation around 5 mW and a maximum external brightness near 60 kHz. The system demonstrates robust mechanical stability over hours and confirms the feasibility of transportable quantum cryptography devices, while tomography indicates a partially polarized emission that favors phase-time encoding. The results point toward integrating GaN defect emitters with photonic structures to create compact, QKD-ready sources suitable for real-field deployment in telecom networks.

Abstract

Single photon sources are essential building blocks for fundamental quantum optics but also for quantum information networks. Their widespread is currently hindered by unpractical features, such as operation at cryogenic temperature and emission wavelength lying outside telecom windows. Taking advantage of telecom technology and point defects in GaN crystals, we present, for the first time, the development of a fully-fibred source of single photons operating at room temperature, emitting photons in the telecom O-band and fulfilling the standards of telecom photonics. We characterise an emitter producting single photons at the wavelength of 1292\,nm, a spectral broadening compatible with CWDM channels of 13\,nm, and a brightness of 25 kcps per mW of pump power. The source shows a signal-to-noise ratio of 16.5 and an autocorrelation degree (purity) of 0,059 at room temperature, showing high potential for being integrated transportable quantum cryptography devices.

Figures (6)

• Figure 1: Sketch of fully fibred single photon device operating at telecom wavelength. (a) Scheme of the experimental setup for collecting single photon from GaN samples. We use a CW laser at 976 nm and a standard 980-1310 nm WDM to separate the pump photon (red arrow) from the singles photons (blue arrow). (b) Zoomed view the microlensed fibre pointing at the surface of the GaN sample. (c) Intensity map of the GaN sample in the X and Y directions, showing the locations of single photon emitters (SPEs). BPF : band pass filter, ID281 SNSPD : superconducting nanowire single-photon and TDC : Time to digital converter WDM : wavelength division multiplex, CWDM : Coarse wavelength division multiplex and PC : polarization controller.
• Figure 2: Optical characterisation of an emitter. (a) Spatially-resolved photoluminescence (PL) scan, with the intensity represented along the Z axis and depicted through a color scale. The scan was performed with a pump power of 1 mW and a step size of 100 nm. (b) PL spectrum of a SPS (blue dots) with a 10 s integration time per point, a pump power of 0.5 mW, a step size of 500 pm, and a sampling interval of 900 pm. The red line is the Voigt fit of the SPS PL emission. The green dots represent the device noise (XTA) when the laser is off and the orange dots represent a spectrum measured in a GaN location where there is no SPS, with the pump laser on.
• Figure 3: Classical characterisation of the performance of the experimental set-up. The data have been graphically normalized to intensity (cps). The data are represented by blue dots, the fit by a green line and the 80 % threshold by a red dashed line. (a) PL signal as a function of time when active stabilisation is off. (b) Spatially-resolved PL scan of the SPS. (c) PL signal as a function of the distance between the fibre and the sample surface.
• Figure 4: Single photon emitter purity measurement. (a) Experimental setup of an HBT interferometer to measure the autocorrelation function. SPS : single photon source, BS : beam splitter and PC : polarized controller. (b), (c) & (d) Measurement of intensity autocorrelation with a pump power of 1 mW. The data are shown in blue with their associated error bars, and the fitting curve, following the equation $g^{(2)}(\tau) = 1 - \beta_1 e^{-\tau / \tau_1} + \beta_2 e^{-\tau / \tau_2}$, is shown in red. The measurement lasted for 250 minutes, with a sampling time resolution $\delta t$ of 40 ps and 25,000 photons detected per second per mW. For proper normalization purpose, we show the time delay up 500 ns in (b) and show in (c) & (d) a zoom-in around zero-delay time.
• Figure 5: Comparaison of purity and brightness versus pump power. (a) $g^{(2)}(0)$ as a function of pump power. The measurement was conducted with an integration time of 90 minutes and with the same experimental setup. (b) Normalised counts per second and signal-to-noise ratio as a function of pump power. The green, yellow and purple dots represent normalised emission from the SPE, a SPS-free GaN location and single photon number in the absence of noise (removing the contribution from the GN background), respectively. The blue curve represents the signal-to-noise ratio. The red lines represent the different regressions: linear for SPS-free GaN location and saturation for single photons without noise following the equation $I = I_{sat} \frac{P}{P_{sat} + P}$.