High-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity

arXiv:2507.09873 · quant-ph · Submitted 2025-07-14 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: I'm Kai, and with me are Mira and Lev, guest researcher.

Mira: Today's paper: "High-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity".

Kai: The gist: A broadband quantum channel capacity derived for microwave-optical transducers provides a practical metric to assess transducer performance by combining throughput and added noise.

Mira: First, who's behind it and why it matters.

Title and authors: Kai: So we’re looking at this paper now, "High-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity." Basically, they’ve derived a way to measure how good these kinds of microwave and optical transducers actually are.

Mira: Right. They use this broadband quantum channel capacity as a metric that combines throughput and added noise in one place, which should give us something more concrete than just looking at efficiency or noise separately.

Lev: From an error correction standpoint, it’s useful because it connects the physical performance of the transducer to how much entanglement you can actually send over a noisy channel. If we know this capacity, we can predict how reliable those quantum links are in practice.

Kai: That’s right, and they show that this capacity depends on things like efficiency and bandwidth, which is a very practical way to characterize the total rate of transduction.

Mira: They define throughput as the product of efficiency, bandwidth, and duty cycle—that’s —and they find that the broadband capacity C depends on both that throughput and the added noise.

Lev: When we look at how this works for real hardware, like a membrane-based transducer with a throughput of seven kHz and an input-referred added noise of three photons, the numbers are quite telling.

Kai: And they show that in downconversion specifically, at those few-photon noise levels, the performance improvement over previous work is nearly four orders of magnitude.

Mira: They quantify this as an apparent efficiency eta Aet which includes that gain factor and gives you the fraction of the incident signal that actually exits the device under ideal conditions.

Lev: In upconversion, they report an input-referred added noise of two point six, while in downconversion it’s three point six photons; these are comparable numbers across both modes for their respective measurements.

Kai: They give us an approximation for the broadband capacity C that is roughly linear in the throughput when efficiencies are small, which is helpful for initial analysis.

Mira: That approximation looks like C about two pi l n p squared q over one - N add - N add p N addq +(one - eta p1 - N addq) over two(<ref:2507.09873#pg1>...) which is quite complex, but it’s accurate to within seventeen percent for efficiencies below zero point five.

Lev: For running this on real hardware, we need to consider the noise integration over the bandwidth, because in downconversion, wideband noise extends beyond the transduction mode due to occupation of a microwave resonator.

Kai: That’s a key point; that added noise N downadd is three point six photons at omega m q, which is comparable to the upconversion noise at two point six, but they find an underlying directional asymmetry there.

Mira: They suggest that this difference between upconversion and downconversion spectra points to some underlying directional asymmetry in the noise baths involved, which they plan to explore later in the paper.

Title and authors: Lev: They even suggest tuning parameters like o and e(o) to balance different noise terms against each other, which means you have to tailor the pump strengths based on whether those noise components differ between upconversion and downconversion.

Kai: So, when we look at the numbers, the contour plots of C give us a good intuition for how important throughput is versus added noise in this quantum-enabled regime where N add is less than one <ref:2507.09873#pg1>.

Mira: The paper states that improving a channel’s throughput is always beneficial because C scales linearly with bandwidth and duty cycle, and at least linearly with efficiency.

Lev: And in downconversion, they show a performance gain factor of five thousand four hundred when you increase the throughput compared to previous work, which is significant scaling data for high-speed conversion systems.

Kai: So, what does this mean for the hardware guys? They’re pointing toward specific design changes that could boost performance further.

Mira: They mention recovering mechanical coherence measured in previous generations of the device and improving the electromechanical coupling rate to increase throughput in both modes.

Lev: It sounds like they are targeting reducing stiff modes in the mechanical spectrum and addressing intrinsic loss rates they measured, which points toward refining the engineered mechanical isolation of their pinned membrane.

Kai: This whole paper, "High-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity," gives us a practical metric for benchmarking transducers by combining throughput and added noise.

Mira: It also provides analytical expressions for directional asymmetry in noise performance between upconversion and downconversion, which helps designers understand the trade-offs involved in network architecture.

Lev: And from an error correction view, it tells us that optimizing the device means tuning pump strengths based on those specific noise differences to achieve better overall quantum channel capacity.

Kai: So to wrap up this paper, they show how you can use a broadband capacity metric to guide design improvements for both upconversion and downconversion systems.

Mira: The practical implication is that we have a clearer way to see which factors—throughput or added noise—are dominating the performance in different parts of the transducer operation.

Lev: It’s a useful tool for predicting how reliably entanglement can be distributed across these kinds of physical quantum channels when they are operating under realistic, finite-bandwidth conditions.

Kai: So, we have this paper here detailing the high-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity.

Mira: It’s a solid piece of work because it links theoretical capacity bounds to measurable quantities like efficiency and noise in both upconversion and downconversion modes.

Lev: We’ll keep an eye on those suggested design targets, especially the ones about mechanical isolation and coupling rates, as they might lead to the next generation of high-performance hardware.

The paper's summary: Kai: So we’re looking at this paper now, "High-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity." Basically, they've derived a way to measure how good these kinds of microwave and optical transducers actually are.

Kai: That’s right, and they show that this capacity depends on things like efficiency and bandwidth, which is a very practical way to characterize the total rate of transduction.

Mira: They define throughput as the product of efficiency, bandwidth, and duty cycle—that’s —and they find that the broadband capacity C depends on both that throughput and the added noise.

Kai: And they show that in downconversion specifically, at those few-photon noise levels, the performance improvement over previous work is nearly four orders of magnitude.

Mira: They quantify this as an apparent efficiency eta Aet which includes that gain factor and gives you the fraction of the incident signal that actually exits the device under ideal conditions.

Kai: They give us an approximation for the broadband capacity C that is roughly linear in the throughput when efficiencies are small, which is helpful for initial analysis.

Mira: That approximation looks like C about two pi l n p squared q over one - N add - N add p N addq +(one - eta p1 - N addq) over two(...) which is quite complex, but it’s accurate to within seventeen percent for efficiencies below zero point five.

Lev: They even suggest tuning parameters like o and e(o) to balance different noise terms against each other, which means you have to tailor the pump strengths based on whether those noise components differ between upconversion and downconversion.

Mira: The paper states that improving a channel’s throughput is always beneficial because C scales linearly with bandwidth and duty cycle, and at least linearly with efficiency.

Kai: So what this means for the hardware guys is that we have a clear way to benchmark these devices by combining throughput and noise into one practical number.

Mira: It also gives us analytical expressions for directional asymmetry in noise performance between upconversion and downconversion, which helps designers understand the trade-offs involved in network architecture.

Kai: So to wrap up this paper, they show how you can use a broadband capacity metric to guide design improvements for both upconversion and downconversion systems.

Mira: The practical implication is that we have a clearer way to see which factors—throughput or added noise—are dominating the performance in different parts of the transducer operation.

Kai: So we've seen how they use this capacity to guide design improvements for both upconversion and downconversion systems.

Mira: We also saw that the paper points toward specific mechanical fixes, like better isolation of the membrane and improving coupling rates, as ways to push performance even higher.

Lev: That’s exactly right; they suggest recovering mechanical coherence and addressing intrinsic loss rates in the device design to get better results in both modes.

Kai: So this capacity metric gives us a roadmap for making these transducers more robust and faster for real quantum networks.

The paper's improvements: Kai: So we’ve seen how they use this capacity to guide design improvements for both upconversion and downconversion systems. Now, what are the actual suggestions from the authors on how to make these transducers even better?

Mira: The paper points toward a few concrete things, first, they say you need to recover mechanical coherence in the device because that’s what was lost in previous generations.

Lev: That makes sense for hardware; if you lose coherence in the mechanical part of the system, it messes up your ability to keep the signal stable and coherent over time.

Kai: Yeah, so they're suggesting better mechanical isolation or maybe tuning how they pin that membrane to suppress those stiff modes we saw in their noise spectra.

Mira: They also mention improving the electromechanical coupling rate; if you can increase how strongly the optical part couples to the mechanical part, it should boost that throughput number significantly.

Lev: If they improve those coupling rates, that directly feeds back into increasing the overall throughput of the system, which is what we want when designing something for quantum networking.

Kai: So they’re essentially saying you can push performance higher by fixing the mechanical side and tuning the electrical side to maximize their coupling.

Mira: That brings us back to those directional asymmetries we talked about earlier; they suggest tuning pump strengths based on whether upconversion or downconversion noise terms are dominating in a specific direction.

Lev: So it’s not just about brute-forcing the numbers, it’s about understanding the underlying noise sources and engineering the system to favor one mode over another if that helps your application.

Kai: It sounds like the next step is less about just making things bigger, and more about a very specific kind of precision engineering on both mechanical and electrical levels.

Mira: Exactly; they’re moving from just measuring performance to actually designing the physical device to achieve those higher throughput numbers they predicted with their capacity metric.

Lev: And if you get that better coupling and coherence, the practical quantum capacity we calculated should translate into a much more reliable channel for distributing entanglement.

Kai: So, they’re giving us a clear path: fix the mechanical stiffness, boost the coupling strength, and then use this capacity metric to guide those specific engineering choices.

Conclusion: Kai: So to wrap up this paper, "High-throughput bidirectional electro-optic transduction assessed with a practical quantum capacity," they’ve shown how to use this broadband capacity metric to guide design improvements for both upconversion and downconversion systems.

Mira: It’s a solid piece of work because it links theoretical capacity bounds directly to measurable quantities like efficiency and noise in both modes.

Lev: I think the most important thing is that they give us a practical way to predict how reliable entanglement can be distributed across these kinds of physical quantum channels when they are operating under realistic, finite-bandwidth conditions.

Kai: It’s a clear roadmap for hardware designers because it tells them exactly where to focus their efforts—whether it's on improving mechanical isolation or tweaking the electrical coupling rate.

Mira: They showed that by focusing on those engineering targets, you can get high-throughput quantum performance in both upconversion and downconversion modes.

Lev: I just want to emphasize that this capacity metric is a real tool for error correction researchers because it ties the physical setup directly into the required fidelity of the quantum communication protocol.

Kai: It’s about moving from abstract theory to something tangible that we can actually build and test in a lab setting.

Mira: The implications are big because it gives us a unified way to assess performance across different transduction mechanisms, like upconversion versus downconversion.

Lev: So this work sets a benchmark for how we should evaluate any new opto-electromechanical transducer before we even start the complex error correction calculations on it.

Kai: It’s a useful tool for predicting how reliably entanglement can be distributed across these kinds of physical quantum channels when they are operating under realistic, finite-bandwidth conditions.

Mira: We'll keep an eye out for how this capacity metric influences the design choices for future quantum devices in the next few papers.

Lev: That’s right; it gives us a framework to decide which parts of the physical system need the most immediate refinement to reach higher performance goals.

JILA, National Institute of Standards and Technology · Department of Physics, University of Colorado, Boulder · Department of Physics, Yale University

quant-ph

Submitted: 2025-07-14

Updated: 2026-10-09

Comments: 8 pages + Appendices, 6 figures

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 89/100

The gist: The gist: A broadband quantum channel capacity derived for microwave-optical transducers provides a practical metric to assess transducer performance by combining throughput and added noise.

Key concepts

Broadband Quantum Channel Capacity (Cub): This is a derived metric that bounds the maximum rate at which quantum information can be sent through realistic, finite-bandwidth thermal-loss channels. It combines throughput (efficiency, bandwidth, duty cycle) and added noise to provide a comprehensive measure of the transducer's capability for quantum communication.
Throughput ($\Theta$ or $\eta_{BD}$): Throughput is a key performance indicator defined as the product of efficiency, bandwidth, and duty cycle. It quantifies the total rate at which transduction occurs. The paper found that improving this throughput significantly increases the broadband capacity (Cub) in both upconversion and downconversion regimes.
Added Noise ($N_{add}$): This refers to the noise introduced by the transducer itself, measured in photons. The study found input-referred added noise levels of 3 photons in both upconversion and downconversion. The capacity (Cub) is shown to depend on this added noise, indicating that minimizing this is crucial for high-rate quantum communication.
Upconversion vs. Downconversion Asymmetry: The study found differences in spectral characteristics between upconversion and downconversion modes, suggesting an underlying directional asymmetry. This asymmetry indicates that optimal performance might require tuning pump strengths based on whether the noise baths differ between the two transduction directions.

Terminology

Entanglement — 1× in this paper · explained in 4 episodes

Summary

The gist: A broadband quantum channel capacity derived for microwave-optical transducers provides a practical metric to assess transducer performance by combining throughput and added noise.

Deriving Practical Capacity

The authors derive a broadband quantum channel capacity that bounds the maximum rate at which quantum information can be sent through realistic finite-bandwidth thermal-loss channels This capacity serves as a comprehensive measure of transducer performance and provides insight into the relative importance of disparate metrics. They find that the broadband capacity depends on the throughput—defined as the product of efficiency, bandwidth, and duty cycle—and on the added noise. The two-way classically assisted quantum capacity c quantifies the number of qubits that can be faithfully communicated, or ebits of entanglement that can be distributed, on average per attempted use of the channel. This capacity is relevant for many applications in sensing and networking, where LOCC is a free resource and the errors are dominated by loss and thermal noise. The broadband capacity Cub is applicable to finite-bandwidth spectrally-uncorrelated thermal-loss channels.

Experimental Platform and Performance Metrics

The study presents measurements of a membrane-based opto-electromechanical transducer with high throughput of 7 kHz and at an input-referred added noise of 3 photons in both upconversion and downconversion. The throughput is quantitatively identified as the efficiency-bandwidth-duty cycle product Θ “ ηBD, as a way to characterize the total rate of transduction. In downconversion, this magnitude at the few-photon noise level marks an improvement of nearly four orders of magnitude over previous work. The apparent efficiency η “ Aηt includes this gain factor and gives the fraction of the incident signal that exits the transducer, where ηt “ ηMp4ΓeΓo(Γ2/Tq) is the ideal transduction efficiency of a device operating with perfect optomechanical and electromechanical sideband resolution.

Upconversion Transducer Performance

In upconversion, the input-referred added noise Nupadd “ 2.6 in upconversion as well as downconversion (Ndownadd “ 3.6) are presented. The broadband capacity Cub is approximately linear in the throughput Θ for small efficiencies, and is given by CubpNadd, Θq « 2π Θ lnp2q / (1 – Nadd) – Nadd ln pNaddq + (1 – ηp1 – Naddq) / log2(…). This approximation is accurate to within 17% for η ď 0.5 and to within 3% for η ď 0.1. The average added noise Nbarupadd “ 2.64p26q over the bandwidth of the transducer is obtained by integrating the total contributed noise.

Downconversion Transducer Performance and Asymmetry

In downconversion, wideband noise extends beyond the bandwidth of the transduction mode due to occupation of the microwave resonator. The input-referred added noise Ndownadd pωmq “ 3.6p9q, comparable to Nupaddpωmq. The difference between the spectra in upconversion and downconversion alludes to an underlying directional asymmetry we explore in Sec. VII. Optimal downconversion performance is achieved by tuning Γo to balance nthγm against ae, and Γe(Γo) to balance nmin,e against be.

Assessing Transducer Performance

Contours of Cub give an intuition for the relative importance of throughput and added noise in the quantum-enabled regime of Nadd ă 1. Improving a channel’s throughput is always beneficial because Cub is linear in B and D, and at least linear in η. In downconversion, the device demonstrates a leap in achieved performance with an increase in throughput of a factor of 5400 compared to previous work.

Directional Asymmetry

The difference between the spectra in upconversion and downconversion suggests that details of the noise baths involved can motivate operating the transducer with diminished efficiency to reduce added noise in one direction at the expense of higher noise in the other. Optimal performance is achieved by tuning pump strengths based on whether nem, nom, and nth differ between upconversion and downconversion. The fact that our membrane-based platform does not exhibit noise terms proportional to Γo is the reason for its exceptional downconversion noise performance.

Conclusion

The broadband capacity presented here is a practical metric in both the experimental relevance of the model, and in its dependence on the experimentally accessible quantities Nadd and Θ. Reasonable improvements to our device design would build upon the high upconversion throughput and unprecedented downconversion throughput we show here, bringing bidirectional, kHz-scale broadband quantum capacities within reach. The refinement of the engineered mechanical isolation of our pinned membrane would suppress the stiff modes in the mechanical spectrum and address the elevated mechanical intrinsic loss rate measured in this device. Recovering the mechanical coherence we have measured in previous generations of this device, combined with parallel improvement to the electromechanical design to increase the electromechanical coupling rate would yield high-throughput quantum-enabled performance in both upconversion and downconversion.

Acknowledgments

We thank Terry Brown, Michael Vissers, and John Teufel for technical assistance, and Georg Arnold, Srujan Meesala, Ravid Shaniv, Benjamin Brubaker, Jonathan Kindem, and Maxwell Olberding for helpful discussions.

Data Availability

The experimental data and analysis code used to generate the figures for this work are available at https://doi.org/10.5281/zenodo.15872216.

Appendix A

The right hand side of Eq. 23 of Ref. [29] gives cubpη, n¯q, which is an upper bound to the two-way quantum capacity cpη, n¯q of a bosonic thermal-loss channel, modeled as a beamsplitter of transmissivity η with a thermal state of ¯n mean photons. We can then integrate cub over a frequency dependent ηpωq and Naddpωq. Taking the limit as η Ñ 1 gives CubpNadd, Θq « 4πΘ ln p2q / (1 – Nadd).

Appendix B

The additional optical filtering, combined with improved electrical filtering of RF signals used in beam preparation, reduces the phase noise of the optical pump incident on the transducer at a detuning of ωm to-155 dBc/Hz. We attribute the significant performance improvement to two design changes.

Appendix C

Due to two simultaneous changes to our transducer design, the PNC was less effective at suppressing stiff modes. The degraded performance of the PnC combined with the shifted mode frequency led to the optomechanically stiff modes seen in the noise spectra shown in Figs. 2 and 3.

Appendix D

In downconversion, we have Ndownadd pωmq “ nthγm nlockγlock nomΓo nemΓe Aoϵ κo,ext κo Γo / (n¯eκe,extΓ2T AηMκeΓeΓo ´ 2¯neΓ2T Aoϵ κo,ext κo ΓopAeΓe AoΓoq.

Appendix F

The microwave readout efficiency can thus be expressed as ξe “ ξoϵCL Ndet,epωsq Ndet,opωsq κeκe,ext κo,ext κo Γo Γe Ao Ae<ref:2507.

Improvements for AI systems

textbfQuantifying Transducer Performance via Quantum Capacity Metric: Improved Benchmarking Tools for Quantum Hardware Design and Selection. The paper derives a broadband quantum channel capacity as a comprehensive measure of transducer performance, which allows researchers to compare disparate metrics like noise and throughput directly against the decay rates of superconducting qubits, enabling better selection of transducers for specific quantum network applications.
textbfBidirectional Performance Assessment: Enabling Robust Quantum Network Architectures. The work provides analytical expressions for directional asymmetry in noise performance between upconversion and downconversion, allowing designers to understand how diminished efficiency to reduce added noise in one direction at the expense of higher noise in the other impacts overall network utility.
textbfModel-Based Noise Characterization: Predicting Performance Under Realistic Channel Conditions. By deriving an expression for the broadband capacity Cub, researchers can predict performance for realistic finite-bandwidth thermal-loss channels, providing a theoretical justification for using the efficiency-bandwidth-duty cycle product as a practical metric.
textbfMaterial and Design Optimization Guidance: Accelerating Transducer Development Through Targeted Engineering. The analysis shows that performance improvements are achievable by specific design changes, such as pinning the membrane to increase coupling rates or annealing the Si3N4 film to reduce noise, offering concrete engineering targets for next-generation transducer fabrication.
textbfEnhanced Downconversion Throughput Analysis: Identifying Scaling Laws for Advanced Transduction Techniques. The paper demonstrates a leap in achieved performance in downconversion, showing an increase in throughput of a factor of 5400 compared to previous work, which provides crucial scaling data for developing high-speed optical-to-microwave conversion systems.