⚛️ quantum physics

Optimal filtering and generation of entangled photons for quantum applications in the presence of noise

This paper investigates optimal filtering strategies for co-propagating time-bin-entangled photons with high-power C-band classical data over long-distance fibers, demonstrating a method to mitigate Raman scattering noise while balancing trade-offs between noise rejection, single-photon purity, and heralding efficiency for scalable quantum networking.

Original authors: Jordan M. Thomas, Andrew R. Cameron, Akil Pathiranage, Si Xie, Raju Valivarthi, Panagiotis Spentzouris, Maria Spiropulu, Cristián Peña, Prem Kumar

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jordan M. Thomas, Andrew R. Cameron, Akil Pathiranage, Si Xie, Raju Valivarthi, Panagiotis Spentzouris, Maria Spiropulu, Cristián Peña, Prem Kumar

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to have a secret, whispered conversation with a friend across a very noisy room. In the world of quantum physics, this "whisper" is a single photon (a particle of light) carrying delicate information, and the "noise" is a flood of other light particles that drown it out.

This paper is about how to make that whisper heard clearly, even when the room is filled with the roar of a jet engine (classical data signals) and the echo of the room itself (background noise).

Here is a breakdown of what the researchers did, using simple analogies:

1. The Problem: The "Noisy Room"

The researchers wanted to send entangled photons (two particles of light that are magically linked, like a pair of dice that always land on the same number) through standard fiber-optic cables. These are the same cables used for your home internet.

However, these cables are also carrying 10-Gbps classical data (like high-speed internet traffic). This data is like a massive, loud crowd shouting in the room. When this loud crowd moves through the fiber, it creates a specific type of "echo" called Spontaneous Raman Scattering (SpRS). This echo creates a sea of random noise photons that look just like the secret whispers the researchers are trying to hear.

If they just turned up the volume on the secret whisper (the quantum signal), it would get lost in the noise. If they turned down the volume on the internet (the classical signal), they wouldn't be able to use the fiber efficiently. They needed a way to filter out the noise without losing the signal.

2. The Solution: The "High-Precision Sieve"

To solve this, the team used three types of filters, acting like a multi-layered security checkpoint:

Frequency Filters (The Color Sieve): They used tunable filters that only let through a very specific "color" (wavelength) of light. Think of this as a sieve that only lets through red marbles and blocks everything else. They made this sieve incredibly fine (narrow bandwidth) to block the noise.
Time Filters (The Stopwatch): They only looked for the secret photons at the exact moment they were supposed to arrive. It's like a bouncer who only lets people in if they knock at the exact second the bell rings.
Polarization Filters (The Orientation Gate): Light waves vibrate in different directions. The researchers set up a gate that only lets through waves vibrating in a specific direction, blocking the rest.

3. The Big Discovery: The "Perfect Match" Trade-off

Here is the tricky part the paper uncovered. When you make your sieve (filter) too fine to block the noise, you accidentally block some of your own secret whispers too.

The researchers found a delicate balance:

The "Heralding Efficiency" Problem: In quantum physics, when you detect one photon of a pair, it "heralds" (announces) that its partner exists. The researchers found that if you filter too narrowly, you lose the ability to know for sure that the partner is there. It's like trying to hear a whisper in a storm; if you put on earplugs to block the wind, you might block the whisper too.
The "Multipair" Trap: Sometimes, the light source accidentally creates two pairs of whispers instead of one. If your filter is too wide, the noise from these extra pairs gets mixed in, ruining the clarity of the message.

4. The Experiment: A 50-Kilometer Test Run

The team successfully sent these entangled photons through 25 kilometers of fiber (for a total of 50 km) while simultaneously sending 10-Gbps internet data through the same cable.

The Result: They managed to keep the internet signal loud (at milliwatt levels, which is much louder than previous experiments allowed) while still hearing the quantum whispers clearly.
The Analogy: Imagine trying to hear a single violinist play a solo while a full orchestra is playing right next to them. Previous attempts required the orchestra to play very quietly. This team managed to let the orchestra play loudly (at a level 10 times stronger than before) and still hear the violinist perfectly, thanks to their ultra-precise "earplugs" (filters).

5. Why This Matters (According to the Paper)

The paper highlights a few key takeaways for building future quantum networks:

Filter Shape Matters: They found that "flat-top" filters (which have a flat, even cut-off) work better than "Gaussian" filters (which have a curved, soft cut-off) for keeping the signal clean.
Pulse Width Matters: The length of the laser pulse used to create the photons changes how well the filters work. Shorter pulses require different filter settings to avoid losing the signal.
The "Noise-Free" Lesson: Even in a perfect, silent room, if your filters aren't designed right, you still lose information. The study shows that "imperfect filtering" hurts the quality of the connection even without any background noise.

Summary

The researchers proved that you can run high-speed internet and delicate quantum networks on the same fiber-optic cable at the same time. They did this by using a combination of color, time, and direction filters to block out the "echo" of the internet traffic. However, they also warned that being too aggressive with filtering can accidentally block the quantum signal itself, so finding the "Goldilocks" zone—where the filter is just right—is the key to making quantum internet a reality.

This work is specifically about fiber-optic networks and quantum communication. It does not discuss medical uses or clinical applications, but rather focuses on the physics of light, noise, and data transmission.

Technical Summary: Optimal Filtering and Generation of Entangled Photons for Quantum Applications in the Presence of Noise

Problem Statement
The deployment of quantum optical technologies, such as quantum teleportation and distributed quantum computation, faces significant hurdles in real-world environments due to background noise. A primary challenge is the coexistence of quantum and classical networks within existing fiber infrastructure. When classical communication signals propagate in the same fibers as entangled photons, spontaneous Raman scattering (SpRS) generates broadband noise photons that obscure single-photon detection. While frequency, time, polarization, and spatial filtering are standard methods to reject noise, filtering the joint spectral amplitude (JSA) of photon pairs can degrade the "filter heralding efficiency" (FHE). This reduction in FHE lowers the probability of detecting true coincidence events, even in lossless environments. Furthermore, imperfect FHE exacerbates the negative impact of multiphoton pair emission, potentially hindering the fidelity of multiphoton quantum applications more than systems with ideal FHE. There is a critical need to understand the trade-offs between noise rejection, single-mode purity, and coincidence rates, particularly for C-band quantum networking where low-loss transmission is essential but Raman gain is high.

Methodology
The authors experimentally investigated these trade-offs by distributing time-bin entangled photon pairs (degenerate at 1536.5 nm) over 50 km of standard single-mode fiber (two 25-km spools). In each spool, the quantum signal co-propagated with a 10-Gbps C-band classical data signal (1547.72 nm). The photon pairs were generated via type-II spontaneous parametric down-conversion (SPDC) in a periodically poled lithium niobate waveguide (PPLN-WG), pumped by 200 MHz, ~50 ps pulses.

The experimental setup employed a multi-stage filtering strategy:

Spectral Filtering: Tunable-bandwidth optical filters (32 pm to 600 pm FWHM) were used to reject SpRS noise.
Temporal Filtering: Low-jitter superconducting nanowire single-photon detectors (SNSPDs, ~45 ps jitter) and time-to-digital converters allowed for precise coincidence windows (down to 200 ps).
Polarization Filtering: Fiber polarizing beam splitters (FPBS) and active electronic fiber polarization controllers (E-FPCs) were used to reject unpolarized Raman noise and correct for time-dependent polarization rotations in the fiber.

The study compared different filtering configurations (e.g., varying spectral bandwidth $\Delta\lambda$ versus temporal window $\Delta T$ ) to analyze their impact on the Coincidence-to-Accidental Ratio (CAR), heralding efficiency, and entanglement visibility. Theoretical models were developed to account for imperfect FHE, multipair emission, and noise, simulating scenarios with Gaussian and flat-top filter shapes, varying pump pulse widths, and different noise levels.

Key Contributions and Results

Demonstration of C-Band Coexistence: The authors successfully distributed time-bin entangled photons alongside 10-Gbps classical data over 50 km of fiber. By utilizing narrow spectrotemporal and polarization filtering, the system tolerated milliwatt-level C-band classical power, exceeding comparable C-band/C-band entangled photon experiments by roughly an order of magnitude.
Impact of Imperfect FHE: The study quantified how filtering the JSA reduces FHE. Experiments showed that for the same single-detector signal-to-noise ratio (SNR), a system with lower FHE (achieved via narrower spectral filtering) exhibited a significantly lower CAR compared to a system with higher FHE. For instance, 50-pm spectral filtering yielded a maximum CAR of ~40, whereas 300-pm filtering yielded ~155 under similar noise conditions.
Trade-off Between Noise Rejection and Purity: The results highlight a fundamental trade-off: while narrowing filters reduces noise and can improve single-mode purity (beneficial for Hong-Ou-Mandel interference), it simultaneously lowers the FHE. This reduction in FHE makes the system more susceptible to degradation from multipair emission and accidental coincidences, limiting the maximum achievable fidelity.
Filter Shape and Pump Pulse Optimization: Simulations revealed that flat-top filters achieve higher CAR and noise tolerance compared to Gaussian filters for a given spectral purity. Additionally, the study found that pump pulse width significantly affects performance; shorter pulses broaden the unfiltered JSA, causing FHE to drop more rapidly as filter bandwidth increases. This suggests that for high-repetition-rate sources requiring short pulses, optimizing the relationship between pump width and filter bandwidth is critical.
Quantitative Limits: The authors derived an expression for the maximum achievable CAR as a function of the pair-symmetric heralding efficiency (PSHE). They demonstrated that an imperfect PSHE (e.g., 0.2) linearly reduces the maximum CAR and coincidence rates compared to an ideal system, limiting the visibility of entanglement verification.

Significance and Claims
The paper claims that its findings provide essential insights for designing sources and receivers for multiphoton quantum applications, extending beyond simple quantum-classical integration. The authors assert that their work:

Enables Scalable Networking: By demonstrating robust coexistence with mW-level classical power, the results support the feasibility of scalable, low-loss entanglement-based networks using existing fiber infrastructure and C-band wavelengths compatible with erbium-ion quantum memories.
Clarifies Filtering Physics: The study elucidates that filtering-induced degradation of single-photon purity and rates occurs even in noise-free environments due to the physics of imperfect FHE and multipair emission. This challenges the assumption that simply increasing pump power can compensate for filtering losses.
Guides Future Design: The results offer a framework for optimizing the trade-off between noise rejection and heralding efficiency. The authors suggest that future optimal sources should aim for factorable JSAs that allow for narrow filtering without sacrificing FHE, or utilize coherent filtering techniques (e.g., mode-selective detection) to maximize both purity and noise rejection simultaneously.
Broad Applicability: While focused on fiber networks, the authors note that these principles apply to noise in free-space links and various quantum devices (sources, frequency converters, switches), providing a general methodology for evaluating the impact of filtering on quantum fidelity.

The work concludes that while narrow filtering is necessary for noise suppression in coexistence scenarios, the design of quantum sources and filters must carefully balance the reduction of noise against the preservation of heralding efficiency to ensure high-fidelity multiphoton operations.