Source-linked AI summary
Entanglement-based wavelength multiplexed quantum communication network
Sören Wengerowsky, Siddarth Koduru Joshi, Fabian Steinlechner, Hannes Hübel, Rupert Ursin
TL;DR
Most quantum-key-distribution implementations and protocols remain limited to two communicating parties, while fully connected networks must minimize client resources. This paper presents a passive, wavelength-multiplexed architecture using one entanglement source and demonstrates a scalable network with communication speed limited by source brightness and detector quality rather than active switching.
Problem
Most quantum-key-distribution implementations and protocols are limited to two communicating parties, making minimized resource requirements with full connectivity a key requirement.
Method
The architecture combines a single bipartite-entanglement source with auxiliary frequency correlations, one detection module and one single-mode fiber per user, and wavelength multiplexing for independent communication links.
Results
The proof-of-principle network is fully connected, linearly scalable in client resources, adaptable to additional users without changing client hardware, and limited in speed by source brightness and detector quality rather than active-component duty cycles.
Takeaways & Limitations
Telecommunication-wavelength operation, scalability, and ease of upgrading make the architecture a candidate for commercial quantum communication networks.
Takeaways & Limitations
Multiplexing many quantum channels onto a single detector increases accidental counts and introduces noise that reduces entanglement fidelity.
Abstract
from arXiv · showhide
Quantum networks scale the advantages of quantum communication protocols to more than just two distant users. Here we present a fully connected quantum network architecture in which a single entangled photon source distributes quantum states to a multitude of users. Our network architecture thus minimizes the resources required of each user without sacrificing security or functionality. As no adaptations of the source are required to add users, the network can readily be scaled to a large number of clients, whereby no trust in the provider of the quantum source is required. Unlike previous attempts at multi-user networks, which have been based on active components, and thus limited to some duty cycle, our implementation is fully passive and thus provides the potential for unprecedented quantum communication speeds. We experimentally demonstrate the feasibility of our approach using a single source of bi-partite polarization entanglement which is multiplexed into 12 wavelength channels to distribute 6 states between 4 users in a fully connected graph using only 1 fiber and polarization analysis module per user.
I. QUANTUM KEY DISTRIBUTION NETWORKS
Quantum networks address QKD’s limitation to two communicating parties, but existing multi-user approaches involve practical, security, or resource constraints. This paper presents a fully connected, passive architecture using one entangled source and wavelength multiplexing.
- QKD’s practical applicability is curtailed because most implementations and protocols support only two communicating parties.
- Quantum repeater networks require further technological advancement in quantum memories before becoming practical.
- High-dimensional or multipartite-entanglement networks require source changes when users are added or removed because system dimensionality changes.
- Trusted-node networks extend bipartite communication but weaken quantum-cryptographic security and duplicate sender and receiver hardware.
- Point-to-multipoint networks allow cross-set communication and have used passive beam splitters, active switches, or frequency multiplexing.
- The presented architecture connects four users to one polarization-entangled source through one fiber each, using frequency correlations and WDM to distribute entanglement between every pair.
- All user pairs can generate private keys from a single source, while the passive design avoids active-switch duty-cycle limits and supports resource and scalability benefits.
II. NETWORK ARCHITECTURE
The architecture uses a fully connected logical network implemented with one fiber and one detection module per user, while wavelength channels distribute pairwise entanglement from a shared source. Users can be added without changing source, client hardware, or software, and the topology can be adapted to sub-graphs.
- A fully connected graph lets every user pair communicate and can be adapted to other network topologies.
- A single bipartite-entanglement source and auxiliary correlations realize the network, with one single-mode fiber and one detection module per user.
- The network’s layers represent physical connections, shared entangled states, and classical communication, post-processing, and secret keys.
- The four-user implementation distributes six bipartite entangled states through three wavelength channels per user over single-mode fibers.
- Adding a user requires the provider to multiplex more channels into each fiber, without changing the source, quantum state, user hardware, or classical post-processing.
III. SETUP
The experimental setup separates photon pairs from a polarization-entangled source into wavelength channels, then passively reroutes selected channel combinations to users through single fibers.
- Photon pairs from a polarization-entangled source are separated into wavelength channels and multiplexed into single fibers for passive rerouting to users.
A. Entangled photon source
The network uses a novel frequency-correlated polarization-entangled photon source at telecommunications wavelengths, based on a MgO:ppLN crystal in a Sagnac configuration. The source ultimately distributes six entangled photon-pair connections among four users.
- A novel source generates frequency-correlated polarization-entangled photon pairs at telecommunications wavelengths for Dense WDM networks.
- The source uses type-0 spontaneous parametric down-conversion in a 4-cm MgO:ppLN crystal with a 19.2 µm poling period.A 775.075 nm continuous-wave pump is converted into co-polarized signal and idler photons in the telecom C-band.
- Bidirectional pumping inside a Sagnac-type setup creates polarization entanglement across two wavelength channels.
- The spectrum is centered at 1550.15 nm and split using symmetric 100 GHz ITU band-pass filters into channels 27–32 and 36–41.
- Six pairs of polarization-entangled photons are distributed among four users so that every user pair shares one photon pair.
B. De-multiplexing and Multiplexing
Energy correlations identify paired signal and idler wavelength channels, which are passively multiplexed into four user fibers. This arrangement gives each user three channels and an entangled connection with every other user.
- Signal channels 27–32 pair with idler channels 36–41 at equal spectral distances from the center wavelength.Examples include channel pairs 27–41 and 28–40.
- Twelve wavelength channels are combined into four fibers using two band-pass filters per fiber.
- Each of four users receives three channels through one fiber, creating a shared entangled-photon pair for every user pair.
- Source fidelity was characterized before multiplexing to confirm high-quality entanglement across the available channel pairs.
- The source spectrum and its band-pass filters distinguish signal and idler channels, with colors marking the entangled photon pairs.
B. Fidelities with Multiplexing
The multiplexed network was tested with four simultaneously active users and six entangled links. Measurements demonstrated entanglement across every user pair, while detector timing and multi-channel detection affected fidelity and key rates.
- Four simultaneously active users received three channels each through the multiplexed network.
- All 12 channels were compensated in two mutually unbiased bases, and temporal cross-correlations identified entangled pairs across the four detectors.
- Six entangled links were measured in the HV and DA bases, with coincidence rates ranging from 10 to 65 Hz.
- A 1 ns coincidence window caused false pair identifications that reduced measured fidelity; corrected values were reported separately.
- Raw key rates of 10–34 Hz were estimated to yield secure key rates of 3–15 Hz, and entangled states were shared between every user pair.
- Figure 6 compares corrected and uncorrected fidelities for WDM channel pairs as wavelength separation varies.
V. CONCLUSIONS
The paper demonstrates a proof-of-principle quantum communication network whose passive, wavelength-multiplexed architecture supports scalability, flexible topology, and compatibility with telecommunications infrastructure.
- The authors successfully realized a proof-of-principle quantum communication network that can be adapted to other topologies and expanded with minimal modifications.
- The architecture supports practical usage scenarios similar to existing everyday networks.
- Telecommunication-wavelength operation makes the network compatible with existing infrastructure, while scalability and upgradeability support commercial-network applications.
- The passive design avoids communication-rate limits imposed by active switching duty cycles; source brightness and detector quality instead determine speed.
- Wavelength multiplexing improves the fully connected design over probabilistic 1:N beam-splitter distribution, which reduces signal-to-noise ratio as users are added.
- Client resources scale linearly, and additional users can be added without changing existing client hardware.
B. Outlook
The outlook focuses on preserving performance as users and wavelength channels increase, while identifying detector noise and accidental coincidences as central constraints.
- Measured fidelities indicate that the architecture remains sound despite increased noise when one detector handles multiple frequency channels.
- Detecting multiple channels on one detector triples the singles count rate while leaving the coincidence rate per link unchanged, increasing noise relative to two-party communication.
- Publicly announcing time-tags and correlation functions lets users discard unrelated counts and improve the signal-to-noise ratio.
- For a network with N nodes, link and system efficiency η, and dark count rate D, the per-user count rate follows Si = D + (N −1)P.
- Accidental coincidences establish the minimum observable coincidence level and reduce contrast.
- Reducing the coincidence window with faster detectors can improve fidelity; a 100 ps window is evaluated in Fig. 8.
2. Scalability
The network is designed for straightforward expansion, but scalability is bounded by source brightness, source bandwidth, and accidental coincidences that grow with user number.
- A 100 ps detector timing jitter supports higher losses and more users than a 1 ns jitter in the calculated Fig. 8 performance.
- The three main scalability limitations are source brightness, the source’s limited wavelength bandwidth, and accidental coincidences that increase with user number.
- An active channel-selecting switch can create custom sub-graphs, while detecting only selected channels limits accidental coincidences.
3. Pulsed network scheme
A pulsed implementation addresses multiplexing-induced accidental counts by synchronizing gated detection with channel delays, while introducing gating and delay-compensation requirements.
- The pulsed scheme can overcome increased accidental counts from multiplexing many quantum channels onto one detector.
- Each user’s gated detector opens n−1 times per laser pulse, with each opening matched to a coincidence-peak delay between users.
- With ideal detectors, pulsed performance is equivalent to n −1 separate communication setups using the same detectors and comparable per-link count rates.
- For fixed users, gating can improve throughput by reducing accidental counts by a factor equal to the gating duty cycle.
- The pulsed scheme requires an additional gating signal and may be unsuitable for mobile nodes because users must compensate delays to all other nodes.
4. Multiplexing and types of entanglement
The architecture is compatible with different entanglement and multiplexing choices, while WDM offers operational advantages over TDM. The authors select polarization entanglement to simplify user hardware and discuss detector requirements for LAN and inter-city deployments.
- Multiplexing: WDM avoids TDM’s active-switching vulnerabilities, including mechanical breakdown and the need for synchronous operation in complex networks.
- Multiplexing: Adding a WDM channel does not reduce existing users’ coincidence rates, whereas adding a TDM channel reduces coincidence rates for all users.
- Types of entanglement: Polarization entanglement was chosen because it simplifies user hardware compared with time-bin entanglement, which requires matched and stabilized interferometers.
- LAN deployment: LAN deployments can use inexpensive SPADs despite low detection efficiency and large timing jitter, tolerating more than 30 dB loss with up to 12 users.
- Inter-city deployment: Inter-city networks can use high-efficiency, low-timing-jitter nanowire detectors, tolerating more than 43 dB loss with up to 25 users over distances exceeding 200 km.