Rack-integrated quantum dot-based source of single and entangled photons at telecom C-band

AI summary

Overview

Research area: Quantum photonics — semiconductor quantum dot (QD) sources of single and polarisation-entangled photons, engineered for the telecom C-band and packaged into a transportable, rack-based unit for deployed quantum communication.

Technical level: Intermediate. The paper is written for quantum-optics and photonics researchers; the experimental narrative is readable, but the optical engineering (volume Bragg gratings, etalons, polarisation compensation, quantum state tomography) assumes familiarity with photonic instrumentation.

Scope: The paper describes the design, optical characterisation, and field-deployment of a compact rack-integrated InAs/GaAs quantum-dot source emitting single photons and polarisation-entangled photon pairs in the telecom C-band, including a deployed 35.8 km metropolitan fiber link test and a simulated secret-key-rate analysis for entanglement-based quantum key distribution.

What This Paper Is About

Quantum light sources based on epitaxial quantum dots deliver excellent single-photon and entangled-photon performance, but they normally live on optical tables in carefully controlled laboratories. The authors build a compact, transportable, remotely controllable rack-based system that contains everything needed to run a telecom C-band quantum dot source — cryostat, pulsed excitation laser, filtering optics, and fiber coupling — and show that it retains state-of-the-art performance outside the lab.

Key Contributions

1. A fully integrated, transportable source unit: the optical stack, cryostat, and all active optomechanics are enclosed in a box of 600 × 1649 × 1003 mm (W × H × D) on four heavy-duty wheels of 125 mm radius; when stacked with a separate two-rack electronics box (two 19" racks, 10 standard rack units total), the system is 1980 mm tall and fits standard transport facilities.

2. A three-floor optical architecture (excitation, cryostat, collection) mounted on independent breadboards above the cryostat vacuum chamber and connected via vibration-damping posts, with three combinable excitation sources (triggered white-light, a 640 nm CW diode, and a tunable pulsed Er-doped fiber laser with EDFA covering 1530–1560 nm) so that above-band, LA-phonon-assisted, and two-photon excitation (TPE) are all available in the deployed unit.

3. Demonstration of both operating modes at record or near-record performance: single-photon emission from a QD at 1543 nm (QD-LA) with g⁽²⁾(0) = 0.056 ± 0.007 at saturation, and polarisation-entangled pair emission from a second QD (QD-TPE) with a fiber-coupled coincidence rate of 201 ± 13 kcps and entanglement fidelity of 0.964 for a 5 ps coincidence window.

4. A field test over a deployed 35.8 km metropolitan fiber loop between the University of Stuttgart Vaihingen campus and the Deutsche Telekom AG site in Stuttgart-Feuerbach, plus a numerical BBM92 secret-key-rate estimate showing 4.2 cps at 175 km.

Main Findings

• Single-photon performance (QD-LA): Under LA-phonon-assisted excitation at a wavelength of 1543 nm (laser blue-detuned by 1.5 nm) and a 76 MHz repetition rate, the source reached a saturation count rate of 1.3 Mcps in the fiber-coupled main collection port, corresponding to an overall end-to-end system efficiency of approximately 2%. The zero-delay autocorrelation at saturation was g⁽²⁾(0) = 0.056 ± 0.007; at 75% of the saturation power it improved to g⁽²⁾(0) = 0.031 ± 0.002, with the count rate reduced to 86% of its saturation value. The emitter was optically active for 60.9% of the time.

• Entangled-pair performance (QD-TPE): Under two-photon excitation, exciton and biexciton lines were centred at λ_X = 1532.1 nm and λ_XX = 1537.1 nm with linewidths of 6.93(12) GHz and 4.27(3) GHz. Corrected fiber-coupled count rates were 5.04 ± 0.16 Mcps (biexciton) and 1.97 ± 0.06 Mcps (exciton) at 76 MHz, corresponding to end-to-end efficiencies of approximately 6.6% and 2.6%. Autocorrelation values were g⁽²⁾(0) = 0.009 ± 0.001 (biexciton) and 0.015 ± 0.001 (exciton), and the emitter was active for 81% of the time.

• Entanglement quality: The entangled-photon coincidence rate reached 201 ± 13 kcps, a fine-structure splitting of only 3.2 μeV was measured, and a maximum entanglement fidelity of 0.964 was obtained for a 5 ps coincidence time window; the reconstructed two-photon density matrix is shown for a 10 ps window.

• Transmission efficiency: The compact setup achieves above 50% transmission for both exciton and biexciton photons, and the authors report a system transmission exceeding 65% after correcting for detector efficiency, compared with roughly 30% for previously reported highly efficient telecom C-band table-top setups.

• Deployed fiber operation: Single photons from QD-LA were transmitted through a 35.8 km deployed metropolitan fiber loop with an overall transmission loss of 18 dB (14.4 dB propagation loss in the loop, with most of the overall loss attributed to fiber connectors). After correcting for detector efficiency, 20 kcps was detected at the link output while the purity was preserved, with g⁽²⁾(0) = 0.036 ± 0.006.

• Projected QKD performance: Numerical BBM92 simulations using the measured time-resolved density matrices and coincidence rates (assuming detector efficiency 0.8 and fiber attenuation 0.21 dB/km, with the coincidence window optimised per distance) give a maximum secret key rate of 4.2 cps at 175 km, corresponding to 36.75 dB of channel loss — reported as exceeding all previously reported deployed entanglement-based QKD values using QD entangled-photon sources.

• Stability under two-colour illumination: Additional white-light illumination stabilised the charge-carrier environment, suppressed blinking, increased the detected count rate, and improved the fidelity of coherent state preparation, as evidenced by measured Rabi oscillations.

Methodology in Plain English

The team grew InAs/GaAs quantum dots on a metamorphic buffer layer, which strain-engineers the emission into the telecom C-band, and embedded them in a planar λ-cavity with a bottom AlAs/GaAs distributed Bragg reflector (23 pairs) and a top SiO₂/TiO₂ dielectric DBR (4 pairs) to boost brightness. The sample sits in a compact closed-cycle cryostat (attoDRY800xs) driven by an air-cooled helium compressor, holding 4 K without needing cooling water.

Excitation light is prepared on one floor of the optical stack: a white-light source for imaging and charge-carrier stabilisation, a 640 nm continuous-wave diode combined with it on a 50:50 beamsplitter, and a tunable pulsed Er-doped fiber laser (10 ps pulses, 400 pm FWHM, repetition rate switchable from 76 MHz to 1 GHz) that drives LA-phonon-assisted or two-photon excitation. A 99:1 non-polarising beamsplitter sends 1% of the light to the sample and the rest to a power meter for active stabilisation, and a 3 mm lens inside the cryostat focuses the light and collects the emission.

On the collection floor, two 1.2 nm-bandwidth volume Bragg gratings suppress the laser by 10⁶ per filter, two further 0.55 nm VBGs split exciton from biexciton photons into separate paths, and angle-tunable etalons (291.6 GHz free spectral range, finesse 18.1) provide extra spectral selectivity. A sequence of motorised quarter- and half-wave plates allows arbitrary polarisation transformations for tomography and for compensating the deployed fiber's polarisation drift, characterised with horizontal and diagonal input states from calibrated linear polarisers. Photons are then either fiber-coupled or, via a flip mirror, routed through a Hanbury-Brown and Twiss setup, polarisation-analysis modules, or a spectrometer using a polarisation-independent MEMS fiber switch. Low-group-delay-dispersion mirrors throughout the collection path preserve the phase relationship needed for entanglement fidelity.

Finally, the source was connected to a real metropolitan fiber loop and characterised after transmission; entanglement-based QKD performance at longer distances was estimated by simulation rather than measured.

Why This Matters

Impact on research: The work shows that quantum dot sources in the telecom C-band can leave the optical table without losing state-of-the-art performance, moving quantum light sources from laboratory demonstrations toward deployable infrastructure. It also sets a coincidence-rate record for entangled photon emission in the telecom C-band within a rack-based unit.

Real-world applications:

• Entanglement-based quantum key distribution (BBM92) over installed metropolitan and long-haul fiber.
• Single-photon sources for prepare-and-measure QKD and other quantum cryptography protocols requiring low multi-photon probability.
• Entanglement swapping and photonic quantum repeaters, where photon indistinguishability and fidelity are crucial.
• Integration with existing fiber networks operated by telecom providers, as demonstrated with the Deutsche Telekom AG link.

Industry relevance: The system requires no dedicated infrastructure such as cooling water, is transportable through standard facilities at 1980 mm stacked height, fits two 19" racks, and is fully remotely controllable — all prerequisites for a telecom operator or a quantum-technology vendor to deploy and operate a quantum light source at a remote site.

Future Directions

• Improving the multi-photon suppression further by optimising spectral detuning and excitation power, building on the g⁽²⁾(0) = 0.031 ± 0.002 already obtained at reduced power.
• Reducing connector losses in deployed links, since the majority of the 18 dB loss on the 35.8 km loop came from connectors rather than fiber attenuation.
• Experimental realisation of the simulated BBM92 secret key rate over long distances, moving from the numerical 4.2 cps at 175 km toward a measured field deployment.
• Addressing emitter blinking and charge-carrier instability (QD-LA was active only 60.9% of the time), for example through the two-colour illumination scheme that already improved QD-TPE stability.

Target Audience

Researchers and engineers in quantum photonics, quantum communication, and integrated quantum optics; experimentalists working with epitaxial quantum dots and single-photon sources; telecom and network engineers evaluating hardware for quantum key distribution deployment; and graduate students seeking a concrete example of how laboratory quantum optics is engineered into a transportable, field-deployable system.

Authors’ abstract

For quantum light sources in everyday telecommunication networks, quantum science needs to be fully transformed into quantum technology. The first necessary step to move outside a well-controlled lab environment requires the use of quantum light sources operating in the technologically relevant telecom O- and C-band. This can be provided using epitaxial quantum dots as deterministic sources of quantum light. Particularly intriguing is that emitters operating at telecom wavelengths are rapidly catching up with their short wavelength counterparts in terms of performances. Here, we make a decisive step forward in the development of quantum communication networks: a state-of-the-art source of quantum light, a semiconductor quantum dot (QD), with record coincidence rate for entangled photon emission in the telecom C-band, is operated inside an optimized rack-based setup. This setup includes a tunable pulsed laser for the QD excitation (from quasi- to fully resonant excitation), all optics for the excitation filtering, and QD signal coupling into single-mode fibers. Overall, the setup allows for above $50\%$ transmission for both exciton and biexciton photons. These results show that quantum dots-based telecom light sources can now be transported and integrated into existing fiber infrastructures, an important step to demonstrate the feasibility of the upcoming quantum internet.