Space-division multiplexed quantum key distribution exploiting multi-plane light conversion for few-mode fibers

AI summary

Overview

• Research area: Quantum communication — entanglement-based quantum key distribution (QKD), space-division multiplexing (SDM) in optical fibers, and spatial-mode manipulation via multi-plane light conversion (MPLC).
• Technical level: Intermediate. Readers need a working idea of entanglement-based QKD, fiber modes, and detection, but the paper is an experimental demonstration rather than a theory-heavy treatment.
• Scope (1 sentence): The paper experimentally demonstrates a BBM92 QKD link in which polarization-entangled photons are transmitted through two distinct spatial modes of a 10 m few-mode fiber and separated at the receiver by an MPLC module, reporting QBERs of 1.9 ± 0.4% and 6.8 ± 0.8% for the two channels.

What This Paper Is About

Fiber-based QKD has mostly been built on single-mode fiber, where the usable degrees of freedom are limited to polarization, phase, time bin, or wavelength. To scale quantum networks toward multiple users, multiple parallel quantum channels are needed without proportionally increasing loss, footprint, or alignment complexity. The authors' goal is to show that the different guided spatial modes of a single few-mode fiber can act as separate, distinguishable quantum channels — not as a high-dimensional encoding — while still preserving the polarization entanglement that BBM92 needs for security.

Key Contributions

1. A parallel-channel architecture for entanglement-based QKD: Entangled photons are routed into a few-mode fiber through two separate single-mode-fiber inputs, selectively exciting $\mathrm{LP}{01}$ and $\mathrm{LP}{11b}$ as two independent transmission channels.
2. MPLC-based mode demultiplexing at the receiver: A commercial multi-plane light conversion module (Cailabs PROTEUS-S-6-SI-PRY) separates the modes into distinct output ports, serving as a polarization-independent mode sorter so that polarization analysis is not disturbed before detection.
3. Experimental BBM92 operation over both multiplexed channels: The paper reports QBER, secure key rate, and average visibility across three configurations (a non-multiplexed reference plus the two multiplexed links), each run continuously (40 min, 115 min, and 184 min respectively).
4. Diagnosis of the current performance limits: The authors quantify insertion loss, end-to-end transmission, and inter-channel crosstalk, and attribute the QBER and visibility degradation primarily to polarization effects and loss at the FMF input coupling (including the 50:50 non-polarizing beam splitter cube), rather than to modal crosstalk alone.

Main Findings

• QBER stays below the security threshold in all configurations: The 11% security threshold is not exceeded in the non-multiplexed reference, the Input 1–Output 1 channel, or the Input 2–Output 2 channel.
• Channel 1 performs nearly as well as the reference: QBER rises only slightly from 1.5 ± 0.4% (no multiplexing) to 1.9 ± 0.4% (multiplexing with Input 1 and Output 1).
• Channel 2 degrades noticeably: QBER rises to 6.8 ± 0.8% for Input 2–Output 2 — still under threshold, but a substantial increase.
• Stability over many rounds: QBER was averaged over about 500 QKD rounds, and the low standard deviation is presented as evidence of long-term stability of the multiplexed link. Missing QBER values, especially in the Input 2–Output 2 link, are attributed to a mismatch between estimated and real QBER causing failed error correction.
• Secure key rates: Average secure key rates ranged from 550 bit/s (Input 1–Output 1) to 30 bit/s (Input 2–Output 2). The paper explicitly states that the multiplexed scheme currently cannot outperform the non-multiplexed secure key rate and that this was not the scope of the work; a key rate for the non-multiplexed reference is not reported.
• Losses are dominated by input coupling: The home-built multiplexer insertion loss is 3.24 dB for Input 1 and 7.3 dB for Input 2. End-to-end transmission efficiencies from Input 1 to Output 1 and Input 2 to Output 2 were approximately 15% and 5.5%, corresponding to total losses of 8.24 dB and 12.6 dB. 3 dB per channel arises solely from the beam splitter cube at the FMF input.
• Crosstalk figures: −8.53 dB for Input 1→Output 2 leakage relative to Input 1→Output 1, and −4.70 dB for Input 2→Output 1 leakage relative to Input 2→Output 2.
• MPLC insertion loss: 2.0 dB for $\mathrm{LP}{01}$ and 2.3 dB for $\mathrm{LP}{11b}$.
• Visibility degradation is polarization-related, not just loss: For Input 2–Output 2, visibility is mainly reduced for coincidences appearing in H. Damping alone would affect A, D, V, and H equally, so the authors infer a polarization effect, most likely introduced at the FMF input.
• Coincidence imbalance supports this: In the non-multiplexed case coincidence counts are comparable, with only A increased by a factor of about 1.5 relative to D. In the multiplexed cases the A-versus-D difference grows to a factor of about 2 (In 1, Out 1) and about 3.5 (In 2, Out 2), plus a factor of about 2.3 for H versus V in both modes.

Methodology in Plain English

The team generated polarization-entangled photon pairs by spontaneous parametric down-conversion in periodically poled potassium titanyl phosphate (PPKTP) crystals in a beam-displacer configuration, pumped by a 775 nm continuous-wave laser (Toptica) at 85 µW, with the crystals temperature-stabilized at 34 °C. Residual pump light was removed with dichroic mirrors and a coarse wavelength division multiplexer, which also separates the signal (1530 nm) and idler (1570 nm) photons, each with a 20 nm bandwidth.

One photon of each pair was sent through the few-mode fiber and the other kept as a reference arm. On the transmitter side the signal was split into two paths and coupled into the fiber through two single-mode fiber inputs, using two orthogonally aligned fiber couplers and a non-polarizing beam splitter cube. By adjusting the input coupling position and angle, different guided modes were selectively excited. The fiber was a Prysmian SI-FMF 4LP with a 15 µm core diameter, NA = 0.1, and 10 m length, supporting 4 LP modes in the C-band; the two channels used were $\mathrm{LP}{01}$ and $\mathrm{LP}{11b}$.

At the receiver an MPLC module performed the mode separation. Before running the quantum measurements, the authors calibrated with a conventional laser at the signal wavelength, optimizing coupling and MPLC alignment to maximize power at the target output port — necessary because degenerate or near-degenerate modes have nearly identical propagation constants and can couple during propagation. Both Alice and Bob had polarization analysis modules with a 50:50 beam splitter cube for random basis choice, quarter- and half-wave plates, and polarizing beam splitters, measuring in the H/V and D/A bases. Photons were detected with superconducting nanowire single-photon detectors (IDQuantique, about 85% detection efficiency, dark counts about 10 cts/s) and time-tagged with a qutools module (jitter below 20 ps), using a 100 ps coincidence window after cross-correlation alignment. QBER was obtained by transmitting part of the key material from Bob to Alice for direct comparison, after which that material was discarded.

Why This Matters

• A route to parallel quantum channels in a single fiber: Instead of physically separate cores (multicore fiber), this work shows that distinct guided modes in one core can carry independent QKD channels while preserving entanglement — a compact alternative for scaling quantum-secured links.
• Distinguishes two uses of spatial modes: Unlike high-dimensional QKD relying on coherent modal superpositions and mutually unbiased mode-basis measurements, this approach treats modes as parallel, distinguishable channels, which sidesteps the need for full high-dimensional state tomography and analysis.
• Quantifies the real bottleneck: The results isolate input-coupling loss and polarization imperfections as the dominant limits, giving a concrete engineering target rather than a vague efficiency problem.
• Real-world applications:
  • Multi-user entanglement distribution, where several receivers share one fiber infrastructure.
  • Metropolitan and backbone quantum-secured networks that need higher aggregate key throughput.
  • Quantum internet architectures based on entanglement distribution and swapping, which require parallel quantum links.
  • Classical/quantum coexistence in mode-division multiplexed fiber systems, since MPLC can separate signals in the mode domain.
• Industry relevance: The demonstration uses commercially available components (a Cailabs PROTEUS MPLC module and a standard few-mode fiber), making it directly relevant to photonics vendors, fiber manufacturers, and network operators planning SDM-based quantum-safe infrastructure.

Future Directions

1. Reduce input-coupling loss: The authors identify the FMF input coupling — particularly the 50% loss and uncontrolled polarization transformations of the non-polarizing beam splitter cube — as the main limitation on key rate and QBER, and plan to address these losses in the near future.
2. Improve robustness to modal crosstalk over time: Since the scheme relies on a static MPLC and stable fiber transmission, the authors propose weakly coupled or ultra-low-crosstalk few-mode fibers, or adaptive optics, wavefront-shaping feedback, and dynamically reconfigurable MPLC to compensate time-varying crosstalk.
3. Scaling channel count and distance: Extending beyond the two-mode, linear Alice–Bob link to longer connections and to routers with more complex connectivity, using new fiber types and intelligent optoelectronics to add more spatial channels.
4. AI-assisted multiplexing: Applying artificial intelligence, in particular physics-informed deep learning, to enhance multiplexing and demultiplexing — for instance to compensate polarization and modal impairments more effectively.

Target Audience

Quantum communication experimentalists and photonics engineers working on entanglement distribution, space-division multiplexing, or mode-division multiplexing; network architects evaluating how to scale QKD beyond single-mode fiber; and graduate students or researchers entering fiber-based quantum communication who want a concrete, component-level account of how a few-mode-fiber plus MPLC quantum link is built and characterized. Readers seeking high-dimensional QKD theory or long-distance field-trial results will find this is a proof-of-concept laboratory demonstration rather than a deployment study.

Authors’ abstract

As quantum key distribution (QKD) progresses from laboratory demonstrations toward practical deployment, quantum communi- cation networks increasingly require higher key rates and the ability to distribute independent secret keys among multiple users and network nodes. In this paper we present a promsing approach with multi-plane light conversion (MPLC) for demultiplexing entangled photons, transmitted through a few-mode fiber (FMF). We experimentally demonstrated a spatially multiplexed BBM92 QKD scheme. The modes are selectively excited through separate single-mode-fibers and subsequently separated by MPLC into distinct output ports. Unlike high-dimensional QKD based on coherent modal superpositions, our approach exploits distinguishable guided modes as parallel channels while preserving the entanglement required for QKD. For the multiplexed links, we obtain quantum bit error rates of $1.9 \pm 0.4\%$ for the channel 1 and $6.8 \pm 0.8\%$ for the channel 2. These results are relevant for scalable quantum-secured networks, space-division-multiplexed QKD systems, multi-user entanglement distribution, and future quantum internet architectures, where parallel quantum channels must be implemented without compromising the quantum correlations required for security.