{"id":"3137a74e-88dd-461f-98ad-73cf2111bdf7","arxiv_id":"2505.00542","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"With existing or near-term microwave-to-optics transducers, optical links can deliver on-demand entanglement between separated quantum processors with fidelities above 90 percent, and above 99 percent after distillation, enabling distributed quantum computing.","lead":"This paper argues that optical links between separate quantum processor modules can already, or with near-term hardware, deliver on-demand entanglement with high fidelity, even though microwave-to-optics converters are inefficient and noisy. It lays out protocols, transducer parameters, and architectures for scaling quantum computers into modular, distributed systems.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"On-demand >99% Bell-state fidelity hinges on non-demolition qubit-to-transducer transfer (η_MW=0.95), which the paper itself says has not been demonstrated; without it, Example 3's fidelity collapses.","rationale":"After reading the full manuscript, I find the reader's assessment well-calibrated. The paper is a perspective with a clear quantitative framework, and its tables encode a plausible roadmap. The most load-bearing assumption is indeed η_MW, for three reasons. First, the manuscript itself singles it out as 'the only not yet accomplished parameter' in Transducer 1, which is otherwise assembled from demonstrated values. Second, η_MW appears multiplicatively in the protocol infidelity of the one-photon TMS scheme and in the heralding probability of both protocols, so it controls both fidelity and on-demand delivery time. Third, unlike detector efficiency or thermal noise, which have partial demonstrations in separate devices, non-demolition qubit-to-transducer transfer has not been demonstrated at all; reference [67] is explicitly destructive. I considered whether the 16-pair distillation assumption (infidelity reduction by an order of magnitude) was a stronger concern, but the cited literature [34] and standard recurrence results make a 10x reduction plausible for depolarizing errors; the bigger unknown is the input fidelity itself, which is governed by η_MW. I also considered the optical-memory requirement in Example 2; however, the memory affects rate more than asymptotic fidelity, and Example 3 does not require an optical memory. Therefore I agree with the reader's weakest-assumption identification and would keep the verdict conditional: the quantitative claims are internally consistent but their near-term realization is gated by a single unperformed experimental step. The proposed concrete test would settle whether that step is genuinely required by mapping the fidelity threshold in η_MW.","tokens_in":31141,"tokens_out":12784,"duration_ms":122202,"concrete_test":"Recompute Example 3 (and Example 2) fidelities from the Table III formulas as a function of η_MW from 0.5 to 0.95, with all other Table V parameters fixed, including the timeout contribution (t_del=15/400 µs, T_2=200/2500 µs, 20/1 parallel links) and the stated distillation gain (infidelity ×0.1). Determine the minimum η_MW needed to keep the distilled on-demand fidelity ≥0.99. Compare that threshold with the highest non-demolition qubit–transducer transfer efficiency reported in the literature; if no experiment reaches the threshold, the headline claim is a projection rather than a statement about existing or near-term performance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III D admits: 'The only not yet accomplished parameter is non-demolition state transfer between a qubit and a transducer with high efficiency, η_MW.' Yet the headline claim of >99% fidelity rests on Examples 2 and 3 (Table VI), both assuming η_MW=0.95. In the one-photon TMS protocol (Box 3), a microwave-optical pair is generated in the transducer; if the microwave photon is not absorbed by the qubit (probability 1−η_MW) while the optical photon is detected, a herald is registered without qubit excitation, contributing (1−η_MW) directly to the protocol infidelity (Table III). Dropping η_MW from 0.95 to 0.5 increases this term from 0.05 to 0.5, reducing Example 3's heralded fidelity from ~0.91 to ~0.5; with 20 parallel links and t_del=15 µs the on-demand fidelity would fall well below the 0.99 threshold even before distillation. η_MW also enters p_her through η_tot, increasing the timeout contribution. Reference [67] demonstrated only destructive qubit-to-optical transduction, not the required non-demolition transfer. Thus the quantitative case for near-term >99% links depends on an experimental capability that has not been shown at any efficiency, let alone at 0.95 with n_th≤0.01 simultaneously.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that microwave-to-optical transducers with currently demonstrated or near-term parameters can enable on-demand, high-fidelity entanglement links between superconducting quantum processors, contrary to the common view that deterministic transducers are required. The authors introduce figures of merit for transducers (efficiency, bandwidth, noise, integration), review state-of-the-art devices, propose two idealized transducer parameter sets (Tables II and V), and analyze four entanglement protocols (one- and two-photon upconversion and two-mode squeezing). Three example link configurations are computed, with claimed on-demand fidelities of 0.73, 0.91, and 0.91 (the last two boosted to 99% by distillation). The paper also discusses architectures such as lattice surgery, sparse links, remote graph states, and optical switching, concluding that sparse links and graph-state generation are feasible in the near term, while full lattice surgery is not.","tokens_in":31521,"tokens_out":5682,"duration_ms":56753,"significance":"The central claim—if correct—is significant: it would reframe distributed quantum computing over optical links as a near-term engineering target rather than a distant ideal. The paper draws on established entanglement protocols (DLCZ, Barrett-Kok) and clearly states its hardware assumptions, which helps the community calibrate feasibility. Its strength is the concrete parameter-based analysis linking transducer metrics to protocol performance, and its explicit admission of which parameters remain undemonstrated. However, the quantitative examples contain arithmetic inconsistencies and rely on a microwave transfer efficiency (η_MW) that the authors themselves state has not been achieved, so the numerical headline claims are not yet reproducible.","major_comments":[{"comment":"The stated fidelity of 0.73 for Example 1 is not consistent with the stated infidelities. From Table III, the one-photon TMS protocol infidelity is ηMW·pMO + (1−ηMW) = 0.8×0.01 + 0.2 = 0.208 (≈0.21), and the thermal infidelity is 2·nth·ηMW² = 2×0.1×0.64 = 0.128 (≈0.13). Combining these as additive errors gives fidelity ≈ 1−0.21−0.13 = 0.66; as independent multiplicative errors it gives (1−0.21)(1−0.13) ≈ 0.69. Neither equals 0.73. Please show the exact expression used or correct the number, because this is the first quantitative demonstration that on-demand links exceed the classical bound.","section":"§IV.F, Example 1 (Table VI)"},{"comment":"All three examples depend critically on η_MW = 0.8 or 0.95, yet §III.D explicitly states: 'The only not yet accomplished parameter is non-demolition state transfer between a qubit and a transducer with high efficiency, η_MW.' For the one-photon TMS protocol used in Examples 1 and 3, Table III shows that (1−η_MW) enters the protocol infidelity directly, and η_MW also affects p_her through η_tot. If η_MW were 0.5 instead of 0.95, Example 3's protocol infidelity would rise from ≈0.07 to ≈0.52, collapsing the claimed 91% fidelity to roughly 50%. The abstract's phrase 'existing or near-term performance levels' therefore overstates the current evidence; the possibility of non-demolition transfer at any efficiency is not yet demonstrated. Please qualify the central claim accordingly and treat this parameter as a projection, not a near-term given.","section":"§III.D and §IV.F, Tables V and VI"},{"comment":"The formulas for pher, protocol infidelity, and thermal noise infidelity in Table III are presented without derivation or explicit source. The protocols are attributed to prior work (Krastanov et al. [84], Zeuthen et al. [37], Barrett and Kok [85]), but the specific expressions are not linked to equations in those references. Since the entire quantitative analysis in §IV.F rests on these entries, please either provide a self-contained derivation in an appendix or state exactly which equation in which reference yields each formula (including the one-photon upconversion thermal term 6nth/ηMW and the two-photon TMS terms).","section":"§IV.B and Table III"},{"comment":"The claimed heralding probability of 3×10⁻² and fidelity of 91% are not traceable from the stated parameters. With transducer 2, η_tot = η_MW·p_MO·η_det = 0.95×0.1×0.5 = 0.0475 ≈ 0.05, and Table IV gives the memory-enhanced two-photon upconversion probability as η_tot·η_mem/2 ≈ 0.024 for η_mem=1, not 3×10⁻². The fidelity is described as limited by 'thermal infidelity, link efficiency and decoherence', but the individual contributions are not quantified. Please show the full calculation, including the decoherence term from T2=2500 µs and t_del=400 µs, so the 0.91 result can be reproduced.","section":"§IV.F, Example 2"},{"comment":"The statement that 'By consuming 16 Bell pairs for four rounds of entanglement distillation, the infidelity can be reduced by an order of magnitude ... resulting in a net fidelity of 99% [34]' is not supported by a specific distillation protocol. The fidelity gain of a 16-to-1 distillation schedule depends on the input state structure (e.g., Werner vs. dephased) and the chosen purification circuit. Please specify the protocol and provide a reference or calculation demonstrating the quoted infidelity reduction for the states produced by the examples.","section":"§IV.F, distillation claim"}],"minor_comments":[{"comment":"There are numerous typos, including 'collectivelly', 'archictecture', 'frequenices', 'expored', 'contruct', 'achieavable', and 'succesful'. A careful proofreading pass is needed.","section":"Throughout"},{"comment":"Box 3, step 4 writes the projected state as (|ge> + e^{iφ}|eg>)/√2 + √pMO|ee> without normalization. This is acceptable only in the limit pMO≪1; please state this approximation explicitly.","section":"§IV.A, Fig. 3"},{"comment":"The dashed line at 0.1 is described as a 'threshold for surface code error correction with lattice surgery', but no reference or derivation is given for this threshold. Please cite a source or define the context.","section":"Fig. 7"},{"comment":"The row for '2-photon TMS, catch and release' gives pher with memory = (η_tot·η_MW·η_mem²)/2. It would help to state whether η_mem includes the herald efficiency of the parity check mentioned in the text, since that is not clear from the current definition.","section":"Table IV"},{"comment":"Reference [74] ('Deterministic quantum state and gate teleportation between distant superconducting chips') is missing author names and journal volume/page information; please complete the citation.","section":"Reference [74]"}],"recommendation":"major_revision","confidential_remarks":"The paper's central idea is plausible and the authors are transparent about the key assumption (η_MW). However, the numerical inconsistencies in Examples 1 and 2, the unsubstantiated distillation claim, and the absence of derivations for Table III are load-bearing for the quantitative conclusions. The paper would also benefit from a clearer separation between 'demonstrated', 'projected', and 'assumed' parameters in the abstract and introduction. If the authors can provide reproducible calculations and appropriately qualify the η_MW dependence, a revised version could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is a feasibility study from QphoX arguing that existing or near-term microwave-to-optics transducers can support on-demand entanglement links with >99% fidelity when combined with memories, parallelization, and distillation. That framing is right: it is not a new protocol, but a parameter comparison over known DLCZ/Barrett-Kok/Krastanov schemes. The authors are honest about what is missing—the one ingredient not yet demonstrated is non-demolition high-efficiency qubit-to-transducer state transfer (η_MW=0.95). That admission sits in Section III D, and it is the load-bearing assumption in Examples 2 and 3. Without it, the headline fidelity collapses; the stress-test note is correct that dropping η_MW to 0.5 pushes Example 3 below the classical threshold. So the central quantitative claim is real but conditional.\n\nWhat the paper does well: it is a clear, well-structured compilation of figures of merit (efficiency, bandwidth, noise, integration) and a sensible taxonomy of entanglement protocols with their heralding probabilities and infidelities. The architecture discussion (lattice surgery, sparse links, graph states, optical switching) is correct and usefully distinguishes what near-term links can and cannot do. The paper does not oversell lattice surgery with current transducers, and it flags the η_MW gap explicitly. Citations to prior protocol work are appropriate; self-citation is present but not improper given that the authors are the transducer developers.\n\nSoft spots: the arithmetic in Example 1 does not check out—stated infidelities 0.21 and 0.13 combine to at most 0.69 fidelity, not 0.73. Table III formulas are presented without derivation, so the numbers are not independently traceable. And the 99% claim requires 16 Bell pairs for distillation per link, a resource overhead the paper glosses over in the summary. These are fixable in revision, but they matter.\n\nThe paper is worth a serious referee—not a breakthrough, but a useful, clearly-written roadmap with specific numbers the community can test. The referee should ask for derivations of Table III, corrected arithmetic for Example 1, and a sensitivity analysis around η_MW. I would bring it to a reading group.\n\nRecommendation: send to peer review, with those requests.","headline":"A useful, honest roadmap for optical links between superconducting processors, but the >99% fidelity claim hinges on an undemonstrated parameter (η_MW=0.95) and Example 1's arithmetic is off.","tokens_in":31997,"tokens_out":3824,"would_cite":false,"duration_ms":36542,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Hk","03.67.Lx","03.67.Bg"],"model":"deepseek-v4-flash","headline":"Near-term microwave-to-optics transducers, despite low efficiency and added noise, can generate on-demand entanglement between separated quantum processors with greater than 99% Bell-state fidelity after distillation, turning distributed…","keywords":["microwave-to-optics transduction","distributed quantum computing","on-demand entanglement","Bell-state fidelity","superconducting qubits","entanglement distillation","quantum memories","optical switching"],"falsifier":"Assemble a single integrated transducer with the Table V Transducer 1 parameters and measure the qubit state after a microwave round trip through the transducer; if the nondestructive transfer efficiency $\\eta_{\\mathrm{MW}}$ falls below 0.8 at an input-referred added noise of 0.1, the 88 microsecond on-demand link fidelity of Example 1 cannot be reached.","tokens_in":30977,"feed_emoji":"🔗","tokens_out":13603,"duration_ms":131299,"temperature":0.7,"pith_summary":"Microwave-to-optics converters are widely assumed to need near-unity efficiency and negligible noise before they can connect quantum processors. This paper challenges that assumption: using loss-tolerant, heralded entanglement protocols with on-demand delivery, converters with already demonstrated or near-term performance can deliver interprocessor entanglement above the classical threshold today and about 99% fidelity after modest improvements and distillation. If the argument is right, distributed quantum computing—several cryogenic modules linked by optical fibers—becomes a near-term engineering target rather than a distant one, and optical links can outperform a single cryostat processor even while the converters remain imperfect. The paper ties this to concrete protocol choices, memory and parallelization trade-offs, and architectures where the links pay off first.","feed_headline":"Near-term transducers can deliver 99% Bell-state links","feed_subtitle":"On-demand heralded entanglement could let today's converters stitch cryogenic modules into one distributed quantum computer","key_machinery":"The load-bearing element is the on-demand entanglement link: each processor module has a communication qubit coupled to a transducer plus a storage qubit, and the link repeatedly runs a heralded entanglement attempt until an optical Bell-state measurement (a beamsplitter and photon detectors that project the qubits into an entangled state) reports success, or until a timeout $t_{\\mathrm{del}}$; the storage qubit then delivers an entangled state on schedule, whether or not a herald arrived. The link is loss-resilient because the protocols herald success with photon detection instead of requiring the transduced photon to survive with high probability. Two protocol families carry the examples: one-photon two-mode squeezing, in which the transducer emits a microwave-optical pair with small probability $p_{\\mathrm{MO}}$ and the microwave photon is absorbed by the qubit, and two-photon upconversion, in which time-bin photons emitted by the qubit are transduced to optics and both early and late herald photons must be detected. The governing quantity is the link capacity $\\eta_{\\mathrm{link}}=T_{\\mathrm{coh}}p_{\\mathrm{her}}/t_{\\mathrm{rep}}$: storage coherence must outlast the mean time to a herald. Parallelization multiplies the heralding rate by the number of channels, and entanglement distillation consumes extra raw pairs to compress infidelity, while the transducer parameter sets in Table V (microwave loading $\\eta_{\\mathrm{MW}}$, pair probability $p_{\\mathrm{MO}}$, detector efficiency $\\eta_{\\mathrm{det}}$, added noise $n_{\\mathrm{th}}$, repetition time $t_{\\mathrm{rep}}$) supply the numerical input for the examples.","core_discovery":"On the paper's own terms, the central discovery is that deterministic conversion is the wrong benchmark. The authors show that loss-resilient entanglement protocols in which a Bell-state measurement at an optical beamsplitter herald success, combined with storage qubits that hold the entangled state until a fixed delivery time, make transducer efficiency unimportant for link fidelity: a conservative parameter set (total efficiency $4\\times10^{-3}$, added noise 0.1, microwave loading efficiency 0.8) already gives an on-demand link fidelity of 0.73 in 88 microseconds, above the classical threshold of 0.55. With a projected parameter set (added noise 0.01, efficiency 0.1, microwave loading 0.95) plus a qubit-cavity optical memory and a 2.5 ms microwave memory, the memory-enhanced two-photon upconversion protocol delivers 91% fidelity in 400 microseconds, and a 20-link parallel one-photon two-mode-squeezing scheme delivers 91% in 15 microseconds. Consuming 16 raw pairs in four rounds of entanglement distillation reduces the infidelity by an order of magnitude, giving net fidelities near 99%. The authors conclude that on-demand optical links can surpass the performance of individual cryogenic modules before transducers reach unity efficiency, with near-term gains in sparse-link, graph-state, and switching-based architectures rather than full lattice surgery.","pith_inferences":["If these links mature as described, the practical bottleneck may shift from the converter to the supporting qubit infrastructure: the 99% example consumes 16 raw Bell pairs per distilled link, and clock-rate delivery would need hundreds of qubits and transducers per processor per link, so communication-qubit count and memory lifetime become the scarce resources.","The protocol toolbox generalizes beyond superconducting qubits: because the argument rests on heralding and storage rather than on any particular qubit platform, visible-wavelength systems such as trapped ions or neutral atoms could adopt the same on-demand link design with a different frequency converter at the front end.","A concrete near-term milestone implied by the paper is to demonstrate one on-demand link with greater than 90% fidelity in the 15- to 400-microsecond delivery window using a single integrated transducer; success would validate the projected parameter set before attempting the hundreds-of-channels scaling needed for architectural demonstrations."],"forward_implications":["With transducer parameters already demonstrated piecewise, a one-photon two-mode-squeezing link delivers entangled states on demand in 88 microseconds with fidelity 0.73, above the classical threshold of 0.55.","With modest projected improvements plus a qubit-cavity optical memory and longer-lived microwave storage, a memory-enhanced two-photon upconversion link reaches 91% fidelity in 400 microseconds, and 20 parallel one-photon links reach 91% in 15 microseconds.","Four rounds of entanglement distillation over 16 raw Bell pairs reduce the infidelity by an order of magnitude, giving about 99% net fidelity, at the cost of more transducers and communication qubits per link.","Full clock-cycle lattice surgery would require a 10- to 100-fold increase in link speed or efficiency, so near-term benefits come from sparse links, remote graph-state generation, and optical switching rather than full edge connectivity.","Scaling estimates place the requirement at 100 to 10,000 transducer channels per cryostat, a number compatible with already demonstrated integrated, low-power transducer designs, making dozens of on-demand links per processor a realistic target."],"supporting_citations":[{"why":"It supplies the one-photon two-mode-squeezing (DLCZ-type) protocol that the paper adapts to microwave-to-optics transducers.","marker":"[83]"},{"why":"It makes the optically heralded entanglement protocol concrete for superconducting systems, providing the transducer-specific version used in the example boxes.","marker":"[84]"},{"why":"It provides the two-photon Barrett-Kok protocol that underlies the time-bin upconversion entanglement scheme.","marker":"[85]"},{"why":"It defines the figures of merit and protocol variants for quantum transducers that the paper uses to compare upconversion and two-mode-squeezing schemes.","marker":"[37]"},{"why":"It supplies the microwave-optical entanglement protocol details, including the parity-check and loss constraints that shape the two-mode-squeezing analysis.","marker":"[39]"},{"why":"It demonstrates deterministic delivery of remote entanglement by repeated attempts and storage, which is the core on-demand mechanism the paper relies on.","marker":"[87]"},{"why":"It shows an on-demand optical link already running deterministic remote gates in a distributed computation, supporting the claim that such links are a usable computing resource.","marker":"[90]"},{"why":"It provides the distributed-architecture analysis and the distillation thresholds used to reach about 99% fidelity and to assess sparse-link benefits.","marker":"[34]"},{"why":"It establishes that noisy interprocessor links have a roughly tenfold relaxed fault-tolerance threshold, justifying the 90% fidelity target in the examples.","marker":"[117]"}],"fun_headline_variants":["Loss-resilient links beat cryogenic bottlenecks","Heralded quantum links work with lossy transducers","Distillation lifts optical link fidelity to 99%","No deterministic conversion needed for quantum links"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a qubit can hand its quantum state to the converter and receive it back nondestructively at 80 to 95 percent efficiency while all the other claimed noise and efficiency numbers hold simultaneously in one device, a combined capability that has not yet been demonstrated.","fun_headline_variants_meta":{"raw":{"variants":["Loss-resilient links beat cryogenic bottlenecks","Heralded quantum links work with lossy transducers","Distillation lifts optical link fidelity to 99%","No deterministic conversion needed for quantum links"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000874,"raw_usage":{"total_tokens":3826,"prompt_tokens":1036,"completion_tokens":2790,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":652,"completion_tokens_details":{"reasoning_tokens":2730}},"tokens_in":652,"tokens_out":2790,"duration_ms":22138,"temperature":1.0,"reasoning_tokens":2730,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:40:13.955770+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Assemble a single integrated transducer with the Table V Transducer 1 parameters and measure the qubit state after a microwave round trip through the transducer; if the nondestructive transfer efficiency $\\eta_{\\mathrm{MW}}$ falls below 0.8 at an input-referred added noise of 0.1, the 88 microsecond on-demand link fidelity of Example 1 cannot be reached.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the one-photon two-mode-squeezing (DLCZ-type) protocol that the paper adapts to microwave-to-optics transducers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It makes the optically heralded entanglement protocol concrete for superconducting systems, providing the transducer-specific version used in the example boxes."},{"cited_title":"Distributed Quantum Computing: Applications and Challenges","cited_arxiv_id":"2410.00609","evidence_quote":"It supplies the microwave-optical entanglement protocol details, including the parity-check and loss constraints that shape the two-mode-squeezing analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It demonstrates deterministic delivery of remote entanglement by repeated attempts and storage, which is the core on-demand mechanism the paper relies on."},{"cited_title":"Ilves, S","cited_arxiv_id":null,"evidence_quote":"It shows an on-demand optical link already running deterministic remote gates in a distributed computation, supporting the claim that such links are a usable computing resource."},{"cited_title":"Krutyanskiy, M","cited_arxiv_id":null,"evidence_quote":"It establishes that noisy interprocessor links have a roughly tenfold relaxed fault-tolerance threshold, justifying the 90% fidelity target in the examples."}],"review_version":1}