{"id":"7543832a-76f2-44c3-bc52-cf91c2729316","arxiv_id":"2603.14123","paper_version":4,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.5,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"For fixed-frequency operation of state-of-the-art SPAs, vacuum squeezing shows no significant Kerr dependence and is instead dominated by resonator and microwave-chain loss.","lead":"Experiments on SNAIL parametric amplifiers show that practical single-mode vacuum squeezing is limited by internal and chain losses, not by Kerr nonlinearity. This redirects device and packaging work toward loss reduction for better quantum sensing and qubit readout.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper's strongest claim is well-supported by the flux/power maps, IMD Kerr extraction, and the explicit prediction of K-independence once gain is fixed (Appendix A, Eq. A14). The reader's weakest_assumption correctly identifies the softest modeling points, but those points do not reverse the experimental observation that S is insensitive to the accessible range of K, nor the design implication that further Kerr suppression is secondary to loss reduction. High-gain degradation and IMD anomalies are acknowledged and lie outside the regime used for the main claim. No circularity or internal inconsistency is present. Therefore the ACCEPT verdict and high confidence remain appropriate; no adjustment is required.","tokens_in":23774,"tokens_out":569,"duration_ms":5708,"concrete_test":"Independently extract η_int(Φ_ext) from the VNA S11 fits already shown in Fig. 7 and recompute the predicted S_obs(Δ) curves of Fig. 4 using only those measured η_int values (plus a single fixed η_cold). If the measured S-vs-Δ trend is quantitatively recovered without free efficiency parameters, residual variation is confirmed as loss; if large residuals remain, unmodeled higher-order effects must be re-examined.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest_assumption correctly flags the linearized stiff-pump Gaussian model (Appendix A) and the attribution of residual S-vs-Δ variation to uncalibrated η_int η_cold (0.22–0.42). Those are real modeling gaps: high-gain degradation (Appendix H) and IMD artifacts (Appendix F) sit outside the model, and η_int η_cold is never measured independently. However, they do not undercut the central claim. Experimentally, |K| only varies by ~2\times (main cluster 40–100 kHz) while S shows no clear correlation with |K| (Fig. 3b); theory predicts exact K-independence once gain is fixed and Δ_eff is retuned (Eq. A14, Fig. 5c); and K/κ ~ 10^{-3} is already small enough that Kerr only limits maximum gain via Stark shift, which is readily compensated. The residual S-vs-Δ trend is therefore secondary to the loss-dominated conclusion. The claim that reducing loss, not further Kerr suppression, is the primary route therefore stands on the data and the leading-order theory.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript characterizes single-mode vacuum squeezing from a SNAIL parametric amplifier (SPA) under fixed-frequency conditions typical of sensing and qubit-readout experiments. By sweeping external flux and pump power near the nominal Kerr-free point, the authors extract |K| via IMD (IIP3) and measure squeezing via heterodyne detection, calibrating the hot-chain efficiency η_hot with a Stark-shifted Ramsey protocol on a downstream 3D cavity-qubit. They report ~2 dB of usable squeezing at the qubit plane, with |K| varying by only a factor of ~2 (main cluster 40–100 kHz) and no clear correlation between S and |K|. A linearized input-output theory (Appendix A) shows that once gain is fixed and Δ_eff is retuned, S_obs is independent of K and is set by the product η_int η_cold. The residual S-versus-Δ degradation is attributed to uncalibrated efficiency variation. The central claim is that baseline K/κ ~ 10^{-3} already renders Kerr non-limiting, so loss reduction, not further Kerr suppression, is the primary route to better practical squeezing.","tokens_in":24026,"tokens_out":1136,"duration_ms":9559,"significance":"If correct, the result reorients device and experiment design for microwave squeezed-state applications: further engineering of the SNAIL Kerr-null point is secondary to improving internal Q and reducing package/wiring insertion loss. The work supplies a concrete experimental protocol (fixed fs, gain-targeted flux/power maps, IMD-derived |K|, qubit-plane efficiency calibration) and a transparent leading-order theory (Eqs. A13–A14, Table I) that cleanly separates Kerr Stark-shift effects from loss. These elements are directly usable by groups employing SPAs or related three-wave-mixing amplifiers for axion searches and qubit readout. The data and appendices are sufficiently detailed to allow independent assessment of the loss-dominated conclusion.","major_comments":[{"comment":"The residual S-versus-Δ trend (Fig. 4) is attributed entirely to uncalibrated η_int η_cold variation (0.22–0.42). Appendix A and Fig. 5c correctly show that S is independent of both K and Δ once gain is fixed and Δ_eff = 0, but the manuscript never measures η_int or η_cold independently across the flux/power map. A short additional measurement (e.g., flux-dependent κ_int from S11 fits already shown in Fig. 7, or a cold-chain efficiency estimate) would make this attribution quantitative rather than post-hoc and would strengthen the claim that loss, not unmodeled higher-order effects, dominates the observed variation.","section":null},{"comment":"Appendix H shows clear degradation of S above ~25 dB gain (S > 0), which the linearized stiff-pump Gaussian model does not capture. The main text correctly notes that this lies outside the model, yet the abstract and conclusion state that baseline Kerr is already too small to impose a practical limitation without quantifying the gain range over which that statement holds. Clarifying that the loss-dominated, Kerr-insensitive regime is limited to the moderate-gain window used in the main experiment (G ≤ 17.5 dB) would prevent over-generalization.","section":null}],"minor_comments":[{"comment":"Fig. 3(b) and Fig. 4 use the same color coding for gain settings; a single legend shared across both panels (or an explicit statement in the captions) would improve readability.","section":null},{"comment":"Eq. (8) and Appendix G introduce GIL versus the experimental G; a one-sentence reminder in the main text of the numerical range of G/GIL (0 to −3.5 dB) would help readers who skip the appendices.","section":null},{"comment":"Appendix F dismisses TLS as the origin of IMD artifacts on three phenomenological grounds; a brief quantitative bound on any residual TLS contribution to the low-power IIP3 used for |K| extraction would make the argument tighter.","section":null},{"comment":"Typographical consistency: “P ARAMETRIC” and “DEGENERA TE” in section headings appear to contain stray spaces; likewise “SNAIL P arametric” in the title of Sec. II.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The central experimental claim (no significant S–K correlation under fixed-fs operation) is solid and does not rely on the theory. The modeling gaps flagged by the reader (stiff-pump assumption, unmeasured η_int η_cold) are real but secondary; they do not overturn the loss-dominated conclusion. Minor revision is appropriate; the paper is a useful, practical contribution for the SPA/squeezing community."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The punchline is simple and useful: for a state-of-the-art SPA run at fixed squeezing frequency (the constraint that actually matters for readout and sensing), Kerr only varies by about a factor of two across the accessible flux/power map, and measured single-mode vacuum squeezing shows no clear dependence on it. Loss (internal + chain) dominates. That reorders priorities for people building these devices.\n\nWhat is new is the systematic fixed-frequency experimental map of S vs flux, pump power, and IMD-extracted |K|, plus the explicit conclusion that baseline K/κ ~ 10^{-3} is already small enough that further Kerr hunting is secondary. Prior theory (Boutin, Frattini/Sivak) already flagged both Kerr and loss; this paper supplies the practical operating-point data that were missing. The measurement chain is careful: heterodyne histograms interleaved with vacuum, η_hot from Stark-shifted Ramsey on a downstream cavity-qubit, gain maps at fixed fs, and IMD for |K|. Appendix A cleanly shows that once gain is fixed and Δ_eff retuned, S_obs is independent of K to leading order (Eq. A14). That matches the main cluster in Fig. 3b.\n\nSoft spots exist but are proportionate. They never independently measure η_int η_cold; they back-fit a 0.22–0.42 range to explain the residual S-vs-Δ trend. High-gain degradation (Appendix H) and IMD artifacts (Appendix F) sit outside the stiff-pump Gaussian model. Those are real modeling gaps, not fatal ones: the central claim rests on the absence of S–K correlation in the data and on K/κ already being small, not on perfect efficiency calibration. The theory is used to interpret, not to force, the result.\n\nMath and citations look solid; the SPA design lineage is properly referenced. This is for device physicists and anyone packaging squeezers for qubits or axion searches. It deserves a serious referee. I would accept it for peer review and would cite the design takeaway.","headline":"Solid device paper: under fixed-frequency practical constraints, SPA squeezing is already loss-limited, not Kerr-limited; the data and leading-order theory support the design takeaway.","tokens_in":24663,"tokens_out":529,"would_cite":true,"duration_ms":6138,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"In practical SNAIL parametric amplifiers, vacuum squeezing is limited by loss, not residual Kerr, so cutting resonator and chain loss is the main path to better performance.","keywords":["SNAIL parametric amplifier","vacuum squeezing","Kerr nonlinearity","microwave quantum optics","Josephson parametric amplifier","qubit readout","quantum sensing"],"falsifier":"Operate the same SPA (or an otherwise identical device) at fixed fs while independently lowering K/κ by an order of magnitude without changing κ_int/κ or chain efficiency; if measured squeezing still fails to improve once gain is matched, the claim that Kerr is already irrelevant holds; any clear improvement would falsify it.","tokens_in":24675,"feed_emoji":"🔋","tokens_out":952,"duration_ms":8068,"temperature":0.7,"pith_summary":"Squeezed vacuum from microwave parametric amplifiers is a resource for quantum sensing and qubit readout, but usable squeezing is often only a few decibels. Prior work suggested that residual Kerr nonlinearity warps the state and sets a hard ceiling. This paper tests that idea on a SNAIL parametric amplifier under realistic constraints: the squeezing frequency is fixed by the downstream experiment, and flux and pump power are varied near the nominal Kerr-free point. Across those points Kerr changes by only about a factor of two and measured squeezing shows no clear dependence on Kerr. Linearized theory confirms that once gain is held fixed, the observed squeezing is independent of Kerr for the small K/κ values already typical of these devices. The dominant degradations are internal resonator loss and insertion loss in the microwave chain. The practical message is therefore that device and packaging loss, not further Kerr suppression, is the bottleneck for usable single-mode squeezing.","feed_headline":"Loss, not Kerr, caps squeezing in practical SPAs","feed_subtitle":"Baseline Kerr is already too small; cut resonator and chain loss to raise usable vacuum squeezing.","key_machinery":"The linearized, stiff-pump input-output model of the pump-dressed SPA (Appendix A). Once the amplifier is tuned to fixed gain with Δ_eff ≈ 0, the observed squeezed-quadrature variance S_obs depends on the product of efficiencies η η_int and is independent of Kerr to leading order; Kerr enters only through a Stark shift that can be retuned away.","core_discovery":"When a SNAIL parametric amplifier is operated at fixed squeezing frequency under conditions typical of sensing or qubit readout, residual Kerr varies only modestly and does not set the achievable vacuum squeezing; the observed squeezing is instead limited by internal resonator loss and insertion loss, so reducing those losses is the primary route to improved performance.","pith_inferences":["If packaging and wiring loss can be driven low enough that η η_int approaches unity, residual higher-order nonlinearities or soft-pump effects may reappear as the next ceiling, reopening a role for Kerr engineering.","Because the paper finds S nearly constant from ~10 dB to 25 dB of gain before degradation, many practical experiments can safely operate at moderate gain without sacrificing squeezing.","The same loss accounting implies that two-mode or multi-mode squeezed sources will face analogous efficiency floors unless the entire interconnect chain is redesigned."],"forward_implications":["Further flux or pump optimization aimed solely at nulling Kerr will not materially raise usable single-mode squeezing in present-generation SPAs when fs is fixed by a downstream cavity.","Design effort should prioritize higher internal quality factors and lower package, circulator, and wiring loss between squeezer and sensor or qubit.","The same loss-limited conclusion is expected to apply to other flux-tunable three-wave-mixing parametric amplifiers used as squeezers.","Reported 2 dB of delivered squeezing already reflects realistic chain loss; device-level squeezing is higher once efficiency is calibrated out."],"fun_headline_variants":["Loss not Kerr caps SPA vacuum squeezing","Resonator loss sets practical SPA squeeze limit","SPA squeeze limited by loss not residual Kerr","Reduce losses for better SPA vacuum squeezing","Baseline Kerr already too small for SPA squeeze"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The claim rests on a linearized stiff-pump Gaussian model that makes squeezing independent of Kerr once gain is fixed, and on attributing the remaining detuning dependence entirely to uncalibrated efficiency changes rather than higher-order or soft-pump effects.","fun_headline_variants_meta":{"raw":{"variants":["Loss not Kerr caps SPA vacuum squeezing","Resonator loss sets practical SPA squeeze limit","SPA squeeze limited by loss not residual Kerr","Reduce losses for better SPA vacuum squeezing","Baseline Kerr already too small for SPA squeeze"]},"model":"grok-4.5","effort":"low","cost_usd":0.002596,"raw_usage":{"total_tokens":922,"prompt_tokens":686,"num_sources_used":0,"completion_tokens":48,"cost_in_usd_ticks":25960000,"prompt_tokens_details":{"text_tokens":686,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":188,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":686,"tokens_out":48,"duration_ms":2639,"temperature":1.0,"reasoning_tokens":188,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T21:28:59.495922+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Operate the same SPA (or an otherwise identical device) at fixed fs while independently lowering K/κ by an order of magnitude without changing κ_int/κ or chain efficiency; if measured squeezing still fails to improve once gain is matched, the claim that Kerr is already irrelevant holds; any clear improvement would falsify it.","supporting_citations":[],"review_version":3}