Attosecond delay metrology beyond the photon coherence time with spectrally resolved Hong-Ou-Mandel interferometry
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.
Kai: We're looking at the paper: "Attosecond delay metrology beyond the photon coherence time with spectrally resolved Hong-Ou-Mandel interferometry".
Mira: Here's the gist of the paper. Hong–Ou–Mandel interferometry enables delay estimation at the quantum precision limit but is traditionally constrained to path differences within the coherence time of the interfering photons Here, we demonstrate single-measurement path-delay sensing at the measurement Cram´er–Rao bound using spectrally resolved HOM interference, thereby removing the conventional dynamicrange limitation imposed by the photon coherence window, with no scanning required for calibration.
Kai: Core Methodology The paper introduces a method for single-measurement path-delay sensing at the Cram´er–Rao bound using spectrally resolved Hong–Ou–Mandel interferometry of spectrally entangled photons.
Mira: Instead of inferring the delay from the depth of the HOM dip, it extracts it from the periodicity of the joint spectral interference fringes.
Kai: This approach allows for delay measurements over a substantially extended dynamic range compared to conventional HOM-based sensing.
Paper summary: Kai: So we're looking at this paper now called "Attosecond delay metrology beyond the photon coherence time with spectrally resolved Hong-Ou-Mandel interferometry". What's the main idea here?
Mira: Basically, they show a way to measure path delays at the quantum limit without being stuck by how long those photons stay coherent. It claims they can do this using spectrally resolved Hong-Ou-Mandel interference of spectrally entangled photons, removing that usual range limitation imposed by the photon coherence window.
Lev: If I'm looking at that idea, it sounds like they are trying to get past the standard limitations of delay estimation in HOM setups where you usually need the delay to be within the time scale of those photons interfering. What's their specific technique for this?
Kai: They aren't inferring the delay just from how deep that HOM dip is anymore. Instead, they extract it from a different thing: the periodicity of these joint spectral interference fringes. It’s about looking at the structure in the spectrum itself rather than just looking at a single coincidence count.
Mira: That periodicity is key because it lets them measure over a much wider dynamic range than what you get with just integrated coincidence counts alone <ref:2607.16849#pg2>. It’s about using the spectral structure to encode that time-delay information over a broader window.
Lev: But how do they guarantee this is still quantum-limited? That's the big question for me, because we always worry about measurement noise leaking in when you try to push precision this far.
Kai: They determine the Fisher information using two specific things: first, the bandwidth of each photon, and second, the nondegeneracy between their signal and idler photons <ref:2607.16849#pg2>. The first part gives you a four sigma squared contribution from each photon's bandwidth, while the second part is the omega squared term from that nondegeneracy.
Mira: And they say in an ideal case where V equals one and gamma equals one, the total Fisher information ends up being four sigma squared plus omega squared. Crucially, they state this result is independent of delta t, which means it keeps the same quantum-limited sensitivity even when the time delay is outside that conventional HOM coherence window.
Lev: That independence from delta t is what makes it tempting for real hardware implementation, but I always worry about the assumptions needed to get there. What are those specific assumptions they need to hold true for that four sigma squared plus omega squared result?
Kai: Those assumptions are fairly specific—they need V equals one and gamma equals one for that perfect scaling <ref:2607.16849#pg3>. They set up a conceptual setup involving a fifty:fifty beamsplitter where the photons go through an object that introduces the delay, and then they use custom spectrometers based on time-tagging event cameras to get those spectral details <ref:2607.16849#pg3>.
Paper summary: Mira: And they define V as the effective HOM interference visibility as η squared times epsilon, where eta is the two-photon indistinguishability parameter and epsilon accounts for the finite spectral resolution of the spectrometer <ref:2607.16849#pg3>. That means any imperfections in those components directly eat into that ideal four sigma squared plus omega squared performance.
Lev: I see what you're saying about V being a variable, which is always true in practice. If the visibility drops because of polarization mismatch or spatial mode overlap, how much of that four sigma squared contribution is actually lost?
Kai: The effective visibility V equals eta squared times epsilon, and that whole expression depends on things like polarization mismatch and spatial mode overlap <ref:2607.16849#pg3>. They are showing that this approach is robust because it relies on fringe periodicity rather than just absolute coincidence rates, which makes it less sensitive to photon losses or variations in the visibility itself.
Mira: So what’s the practical result they got when they actually built and measured this? It’s not just theory, right?
Kai: They achieved a time-delay precision of two zero attosecond six nanometer using one million detected photon pairs <ref:2607.16849#pg3>. In a real-time measurement setup, where they run it live, they get a precision of three hundred thirty attosecond one hundred nanometer <ref:2607.16849#pg3>.
Lev: Two zero attosecond six nanometer is good for theory, but for me building this on actual hardware, what’s the practical hurdle? Where does the overhead land when you try to run this real-time measurement architecture?
Kai: The current system's performance is strongly limited by the photon detection efficiency of their TPX3CAM detector, which is only about eight percent at their operating wavelength <ref:2607.16849#pg3>. That efficiency really drags down the practical precision you can achieve right now.
Mira: That eight percent efficiency is a big constraint for any measurement that relies on coincidence counts, so they are pointing toward future detector technologies to fix that bottleneck.
Lev: I agree, and I think that’s where the real scaling potential lies. If we look at emerging detector technologies like superconducting nanowire single-photon detector array sensors, they expect to increase the coincidence rate by more than two orders of magnitude <ref:2607.16849#pg3>. That kind of jump would dramatically improve the measurement record.
Kai: So, even with those future detectors, they are still running into limitations because the precision degrades at large path delays due to the finite spectral resolution and sampling of their two-photon spectrometer <ref:2607.16849#pg3>. The effective HOM interference visibility V equals eta squared times epsilon drops when you get too far from zero delay.
Paper summary: Mira: That is a fair limitation, because if the spectrometer can't resolve the spectrum well enough, that finite spectral resolution reduces that effective visibility V equal to eta squared times epsilon <ref:2607.16849#pg3>. So they aren't just measuring better; they’re still hitting a wall based on the instrument design itself.
Lev: It sounds like the system needs both better photon sources and better detection, and also a spectrometer that can handle the spectral resolution needed to maintain that interference visibility over a wide delay range. That’s a lot of engineering work <ref:2607.16849#pg3>.
Kai: Exactly. But I think the main thing they are demonstrating here is that you don't need to scan the interferometer or recalibrate the dip just to get that precision, which makes this method really attractive for practical sensing applications, even with these current limitations <ref:2607.16849#pg1>.
Mira: So, to wrap up this paper "Attosecond delay metrology beyond the photon coherence time with spectrally resolved Hong-Ou-Mandel interferometry", what's the real implication for how we think about measuring time delays?
Kai: It means we can move toward attosecond delay metrology that goes beyond the coherence time of individual photons by using spectral resolution in a specific way. It shows that by looking at interference fringes in the spectrum, you can extract delay information over a much wider dynamic range than traditional methods allow <ref:2607.16849#pg2>.
Lev: For hardware development, it suggests that the path forward isn't just about making bigger detectors, but about designing spectrometers and measurement architectures that can maintain high effective visibility V equal to eta squared times epsilon across a larger delay range <ref:2607.16849#pg3>.
Mira: And from a theoretical standpoint, it validates the idea that resolving additional photonic degrees of freedom, like frequency or time itself, can recover interference signatures beyond the traditional window <ref:2607.16849#pg2>. It confirms the potential for this type of measurement to exist at the quantum precision limit if you manage those spectral constraints.
Kai: So, while they've shown a way around the coherence time limit for single-measurement sensing at the Cram´er–Rao bound, they also made it clear that current practical performance is heavily tied to detector efficiency and spectrometer resolution <ref:2607.16849#pg3>.
Lev: That's the honest assessment—the paper gives you the potential physics, but you still have to solve the engineering problems of building a system that can handle those spectral constraints and high detection efficiencies <ref:2607.16849#pg3>.
Mira: It’s a strong piece of work because it shows how spectrally resolved HOM interference can be used for single-measurement sensing without needing to scan or recalibrate the delay, which is something I think is very valuable for real-world sensing applications <ref:2607.16849#pg1>.
Conclusion: Kai: So, we're wrapping up this look at this paper about attosecond delay metrology beyond the photon coherence time with spectrally resolved Hong-Ou-Mandel interferometry.
Mira: It really boils down to how they use spectral information in interference fringes instead of just looking at the depth of a standard Hong-Ou-Mandel dip.
Lev: So, they're taking this delay estimation and pushing it past the time scale set by how long those photons stay coherent in a typical setup.
Kai: Exactly. They use the periodicity of those joint spectral interference fringes to extract the delay measurement, which lets them measure over a much wider range than what you usually see with standard HOM setups.
Mira: The core idea is that this method can maintain that quantum limit sensitivity even when the time delay is way outside your conventional coherence window, provided you have the right resources.
Lev: But we gotta remember the constraints they put on it. They needed specific conditions, like perfect two photon indistinguishability and certain visibility settings to keep that four sigma squared plus omega squared scaling true.
Kai: Right, so the real world is messy because of those assumptions, especially when you get into practical measurements.
Mira: And they showed that while the theoretical potential is high, the actual performance right now is really tied to how good your detector efficiency and your spectrometer resolution are.
Lev: That’s where I get nervous for hardware deployment. You need a really efficient detector to make those measurements practical, and you still have that spectral resolution limit causing precision degradation at larger delays.
Kai: So the implication here is that we can move toward attosecond delay measurements without having to constantly scan the interferometer or recalibrate every time we want a new measurement point.
Mira: It suggests that resolving additional degrees of freedom in the photon, like their frequency, is a legitimate way to recover interference signatures you normally lose.
Lev: But that means for anyone building this, it’s not just about getting better detectors; it’s about designing measurement architectures that can keep the effective visibility high over those larger delay ranges.
Kai: So we've seen how they tackle the theory and the numbers, but now we have to figure out how to build a system that actually delivers on that promise.
Nexus for Quantum Technologies University of Ottawa National Research Council of Canada Joint Centre for Extreme Photonics University of Portsmouth Institute of Cosmology and Gravitation Institute for Quantum Studies Chapman University
quant-ph, physics.optics
Submitted: 2026-07-18
Updated: 2026-11-04
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 89/100
The gist: Hong–Ou–Mandel interferometry enables delay estimation at the quantum precision limit but is traditionally constrained to path differences within the coherence time of the interfering photons…
Key concepts
- Hong–Ou–Mandel (HOM) interferometry
- A technique using two photons interfering at a beam splitter to create bunching or antibunching effects. Traditionally, this method limits delay measurements to distances shorter than the photon's coherence time, which is what this paper aims to overcome.
- Spectrally resolved HOM interference
- Instead of looking at a single interference dip, researchers analyze the pattern of fringes created when using spectrally entangled photons. This allows them to extract delay information from the periodicity of these fringes rather than relying on the coherence window alone.
- Cramér–Rao bound
- This is a theoretical limit defining the best possible precision achievable for estimating an unknown parameter, like time delay, given a set of measurements. The paper shows their method achieves this fundamental quantum-limited sensitivity.
- Fisher information
- A measure of how much information can be extracted from a measurement. The paper calculates that using frequency-resolved measurements provides Fisher information independent of the time delay ($Δt), ensuring the same quantum precision regardless of how large the delay is.
Terminology
Summary
Hong–Ou–Mandel interferometry enables delay estimation at the quantum precision limit but is traditionally constrained to path differences within the coherence time of the interfering photons Here, we demonstrate single-measurement path-delay sensing at the measurement Cram´er–Rao bound using spectrally resolved HOM interference, thereby removing the conventional dynamicrange limitation imposed by the photon coherence window, with no scanning required for calibration. p. 0 Attosecond delay metrology beyond the photon coherence time with spectrally
resolved Hong-Ou-Mandel interferometry
Yingwen Zhang,1, 2, 3, ∗ Kyle Jordan,1, 2, 3 Duncan England,2, 3 Vincenzo
Tamma,4, 5 Ebrahim Karimi,1, 2, 3, 6 and Benjamin Sussman2, 1, 3
1Nexus for Quantum Technologies, University of Ottawa, Ottawa ON, Canada, K1N6N5
2National Research Council of Canada, 100 Sussex Drive, Ottawa ON, Canada, K1A0R6
3
Joint Centre for Extreme Photonics, National Research Council and University of Ottawa, Ottawa, Ontario, Canada
4School of Mathematics and Physics, University of Portsmouth, Portsmouth PO1 3QL, United Kingdom
5
Institute of Cosmology and Gravitation, University of Portsmouth, Portsmouth PO1 3FX, United Kingdom
6
Institute for Quantum Studies, Chapman University, Orange, California 92866, USA
Hong–Ou–Mandel (HOM) interferometry enables delay estimation at the quantum precision limit
but is traditionally constrained to path differences within the coherence time of the interfering photons. Here, we demonstrate single-measurement path-delay sensing at the measurement Cram´er–Rao
bound using spectrally resolved HOM interference, thereby removing the conventional dynamicrange limitation imposed by the photon coherence window, with no scanning required for calibration.Where the paper says it
Core Methodology
The paper introduces a method for single-measurement path-delay sensing at the Cram´er–Rao bound using spectrally resolved Hong–Ou–Mandel interferometry of spectrally entangled photons. Instead of inferring the delay from the depth of the HOM dip, it extracts it from the periodicity of the joint spectral interference fringes. p. 1 In this work, we demonstrate single-measurement path-delay sensing at the Cram´er–Rao bound (CRB) using
spectrally resolved HOM interferometry of spectrally entangled photons implemented with a custom-built two-photon
spectrometer based on a time-tagging event camera [42]. Instead of inferring the delay from the depth of the HOM
dip, we extract it from the periodicity of the joint spectral interference fringes. Instead of inferring the delay from the depth of the HOM
dip, we extract it from the periodicity of the joint spectral interference fringes. These fringes persist well beyond the
photon coherence length (the width of the HOM dip), with an oscillation frequency that is directly proportional to
the path delay, thereby enabling delay measurements over a substantially extended dynamic range compared with
conventional HOM-based sensing. The photon pairs are then recombined at a
50:50 BS to undergo HOM interference. The probability distributions for spectrally entangled photon pairs exiting
the HOM interferometer in the bunched (same exit port) and antibunched (different exit ports) configurations, are
given by [45, 46]
PB(∆t, ω1, ω2) = 1
2
|Φ(ω1, ω2)|
2
1 + V cos[(ω1 − ω2)∆t]
bunching
PA(∆t, ω1, ω2) = 1
2
|Φ(ω1, ω2)|
2
1 − V cos[(ω1 − ω2)∆t]
antibunching.Where the paper says it
Where the paper says it
Where the paper says it
Quantum Precision Limit
The Fisher information associated with frequency-resolved measurements is determined by two distinct spectral resources: the bandwidth of each photon and the nondegeneracy between signal and idler photons. The first contribution, 4σ2, arises from the bandwidth of each photon, while the second contribution, omega2, arises from nondegeneracy between the signal and idler photons. p. 4 The first contribution, 4σ
2
, arises from the bandwidth of each photon and reflects the fact that broader spectra
produce more rapidly varying spectral interference fringes for a given delay, thereby increasing the available timing
information. The second contribution, Ω2
, arises from nondegeneracy between the signal and idler photons. In the present experiment,
degenerate photon pairs are used, so Ω = 0 and the Fisher information is determined solely by the SPDC bandwidth. Importantly, F0 is independent of ∆t, indicating that spectrally resolved HOM measurements can, in principle,
maintain the same quantum-limited delay sensitivity over a range far exceeding the conventional HOM coherence
window.Where the paper says it
Where the paper says it
Experimental Performance and Robustness
The authors achieved a time-delay precision of 20 attosecond (6 nm) using one million detected photon pairs, while real-time operation yields 330 attosecond (100 nm) precision. The estimator relies on fringe periodicity rather than absolute coincidence rates, making the method intrinsically robust to photon losses and variations in interference visibility, eliminating the need for recalibration. p. 0 Using one million detected photon pairs, we
achieve a time-delay precision of 20 attosecond (6 nm), while real-time operation (at 1 Hz) yields
330 attosecond (100 nm) precision. Because the estimator relies on fringe periodicity rather than
absolute coincidence rates, the method is intrinsically robust to photon losses and variations in interference visibility, eliminating the need for recalibration. Because the estimator relies on fringe periodicity rather than
absolute coincidence rates, the method is intrinsically robust to photon losses and variations in interference visibility, eliminating the need for recalibration. As a practical demonstration, we measure
the thickness of a 300 µm transmissive target with nanometer-scale precision.Where the paper says it
Where the paper says it
System Limitations and Future Directions
The precision degradation at large path delays arises from the finite spectral resolution and sampling of the two-photon spectrometer, which reduces the effective HOM interference visibility Veff(∆z). The current system's performance is strongly limited by the photon detection efficiency of the TPX3CAM, which is only ∼ 8% at the operating wavelength. Emerging detector technologies, such as superconducting nanowire single-photon detector array sensors, are expected to increase the coincidence rate by more than two orders of magnitude. p. 8 Emerging
detector technologies, such as superconducting nanowire single-photon detector array sensors [59–61], offer quantum
efficiencies exceeding 80% together with substantially lower timing jitter. Replacing the current camera with such
detectors would be expected to increase the coincidence rate by more than two orders of magnitude. Further improvements in precision can be achieved by using broadband
non-degenerate SPDC photon pairs, which introduces a non-zero Ω in Eq. More broadly, the ability to perform high-precision delay estimation without scanning the interferometer or recalibrating the HOM dip makes spectrally resolved HOM interferometry particularly attractive for real-world sensing
applications.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved real-time displacement sensing for quantum metrology systems by extracting delay information from spectral interference fringes rather than relying on absolute coincidence rates, as stated in
we extract it from the periodicity of the joint spectral interference fringes.
This allows forreal-time operation (at 1 Hz) yields 330 attosecond (100 nm) precision,
enabling immediate feedback and control loops in quantum sensing. -
Enhanced robustness against experimental imperfections, as the method is
intrinsically robust to photon losses and variations in interference visibility, eliminating the need for recalibration.
This means AI-driven systems can operate reliably over long durations or across different physical samples without requiring manual recalibration of the interferometer path delay. -
High-precision thickness metrology using quantum displacement sensing, where
we measure the thickness of a 300 µm transmissive target with nanometer-scale precision in a single shot.
This capability allows AI systems to perform quality control or precise material characterization in real-world industrial settings using quantum measurements. -
Optimized measurement strategy based on spectral information, leveraging the Fisher information derived from the joint spectral amplitude, which is determined by
two distinct spectral resources,
specificallythe bandwidth of each photon and nondegeneracy between the signal and idler photons.
This guides AI algorithms to select optimal entangled photon sources or operating conditions to maximize time delay precision. -
Adaptive error mitigation for large path delays, where the modeling accounts for finite spectrometer resolution by using a weighted sum of spectra based on
centroiding errors,
as described inthe final modeled spectrum is obtained from a weighted sum of these spectra, Smodel(x, ∆z) = X p W(p)Sp(x, ∆z).
This allows AI to estimate the true delay even when spectral resolution limits are approached at large path differences.
Abstract
Hong-Ou-Mandel (HOM) interferometry enables delay estimation at the quantum precision limit but is traditionally constrained to path differences within the coherence time of the interfering photons. Here, we demonstrate single-shot path-delay sensing using spectrally resolved HOM interference, achieving a precision near the measurement Cramér-Rao bound over an operating range of approximately plus or minus 300 um, removing the conventional dynamic-range limitation imposed by the photon coherence window. By extracting delay information from the spectral interference fringes of spectrally entangled photon pairs, we retain near-optimal sensitivity over an operational range exceeding the photon coherence time by over two orders of magnitude. Using one million detected photon pairs, we achieve a time-delay precision of 13 attosecond (4 nm), while real-time operation (at 1 Hz) with 2000 detected photons yields 330 attosecond (100 nm) precision. Because the estimator relies on fringe periodicity rather than absolute coincidence rates, the method is intrinsically robust to variations in interference visibility, eliminating the need for recalibration. As a practical demonstration, we measure the thickness of a nominally 300 um transmissive target with nanometer-scale precision. These results mark a significant step towards deploying quantum-enhanced HOM interferometry for high-precision displacement sensing in real-world applications.
Sources
- Quantum-limited estimation of the difference between photonic momenta via spatially resolved two-photon interference
- Simultaneous amplitude and phase spectroscopy using two-photon interference
Related papers
- Computational regimes in matrix-product-state-based quantum trajectory simulations
- Toward chemically accurate quantum simulations on quantum hardware: A quantum-HPC hybrid approach
- Transversal surface-code game with reconfigurable qubits
- Fulqrum: a generalized framework for quantum subspace diagonalization
- Flux-Activated Resonant Control of a Bosonic Quantum Memory
- Machine learning the effects of many quantum measurements