{"id":"831516fc-84c0-4c70-acba-956da69035a4","arxiv_id":"2506.03456","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"Strong low-frequency parametric drive on a transmon causes ionization above a threshold that matches Floquet predictions, limiting parametric gate speed.","lead":"This paper shows that a superconducting transmon, when driven too hard at low frequency, gets kicked into high-energy states, a breakdown that sets a speed limit for parametric quantum gates. The authors show that a Floquet analysis predicts the threshold, suggesting the same design constraint underlies readout errors and parametric control.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Floquet threshold predictor rests on a hand-defined clustering cutoff (M−2) and visual cluster identification; the central quantitative claim needs a test of whether predicted thresholds are stable under automated variants of this rule.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the Floquet threshold is defined by a hand-chosen k-means criterion with visual identification and an M−2 cutoff. I agree that this is the point on which the paper's quantitative predictive claim is least secure. The core experimental observation of a drive-amplitude threshold is well supported by the combination of measurement, exact time-dependent simulation, and the coherent physics of thresholds lowering for higher initial states and slower ramps. The full simulation including the resonator and higher Josephson harmonics is a genuine independent check, and the agreement in Fig. 1b is striking. The open question is not whether ionization occurs, but whether the Floquet spectrum provides a robust quantitative predictor, which is the basis for the 'unifying framework' and 'fundamental constraints' claims. The proposed invariance test would settle whether the predicted threshold is an intrinsic spectral feature or an artifact of the criterion. Since the reader's verdict is already CONDITIONAL and my concern reinforces rather than overturns that assessment, the appropriate verdict is UNCHANGED.","tokens_in":22267,"tokens_out":6534,"duration_ms":80790,"concrete_test":"Request the code/data underlying Figs. 1–3, then recompute the Floquet threshold over ωd/2π ∈ [1, 2] GHz while varying: (i) number of clusters k=2 vs k=3; (ii) cutoff M−1, M−2, M−3; (iii) Floquet amplitude step 2, 5, 10, 20 MHz; (iv) number of Floquet modes 20, 30, 50; and (v) an automated cluster-identification rule replacing visual inspection, for example the cluster whose mean population is largest at high amplitude. If all variants give threshold curves within about 0.1 GHz of each other and of the red line in Fig. 1, the criterion is robust; if shifts exceed about 0.3 GHz in any frequency range, the quantitative predictive claim is underdetermined and should be re-benchmarked.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the Floquet spectrum quantitatively captures the ionization threshold is only as strong as the criterion used to extract a threshold from the spectrum in Methods C. There, the ng-averaged Floquet-mode populations are grouped into two k-means clusters, one of which is identified as the 'chaotic manifold' by visual inspection, and the threshold is defined as the lowest drive amplitude at which the ground-state mode population exceeds M−2, where M is the mean population of that cluster. The choices k=2, the M−2 offset, the 10 MHz amplitude grid, and the 30-mode truncation are not derived from the underlying dynamics, and the same Floquet data are used both to define and to evaluate the predictor. The validation in Fig. 1 is also against an experimental threshold whose frequency dependence is contaminated by uncalibrated cable impedance, so the comparison mainly constrains the smooth trend. If different reasonable implementations of 'entry into the chaotic manifold' move the predicted threshold by an amount comparable to the experimental uncertainty, the agreement shown would be a property of the criterion rather than evidence of a quantitative Floquet predictor. This does not threaten the robust experimental observation of a drive-amplitude threshold, but it does threaten the stronger quantitative and unifying claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an experimental and numerical study of a transmon qubit driven parametrically at frequencies (1–2 GHz) far below its 0–1 transition. It observes a drive-amplitude threshold beyond which the transmon ionizes into many high-energy states, and it argues that this threshold is quantitatively predicted by the instantaneous Floquet spectrum: specifically, by the onset of strong hybridization and level bunching of Floquet modes, interpreted as a quantum counterpart of the classical chaotic layer of the driven pendulum. The authors support the claim with full time-dependent Schrödinger simulations that include Josephson harmonics, with measurements of the threshold for initial states |0>, |1>, and |2>, and with pulse-shape dependence studies.","tokens_in":22493,"tokens_out":2543,"duration_ms":31847,"significance":"If the central quantitative claim holds, the paper establishes a fundamental speed limit for parametric control of superconducting circuits and provides a computationally cheap predictor (Floquet spectrum) for the breakdown threshold. The experimental phenomenology is genuinely new in the parametric-drive context and is backed by full time-dependent simulations whose parameters (EC, EJ,m, g, resonator frequency) are fitted to independent spectroscopy data, not to the ionization threshold itself. The drive-amplitude calibration via low-amplitude Rabi fringes is also independent of the threshold. The paper further connects the quantum bunching to classical Poincaré sections, giving a physically appealing semiclassical picture. The main weakness is that the quantitative Floquet predictor relies on a hand-defined clustering criterion (Methods C), so the strength of the 'quantitative' claim is currently not fully supported.","major_comments":[{"comment":"The threshold extraction is defined by a particular implementation: k-means with k=2, visual identification of the 'chaotic' cluster, the M−2 cutoff, a 10 MHz amplitude grid, and 30-mode truncation. None of these choices is derived from the dynamics, and the same Floquet data are used both to define and to evaluate the predictor. The manuscript should report robustness tests: for example, how the predicted threshold changes when the cutoff is varied (M−1, M−3), when the grid spacing is halved or doubled, when the number of Floquet modes is changed, and when an automated cluster-identification rule replaces the visual step. If the resulting threshold moves by an amount comparable to the experimental uncertainty, the agreement in Figs. 1 and 3 is a property of the criterion, not of the Floquet framework.","section":"Methods C (Floquet analysis)"},{"comment":"The experimental threshold frequency dependence is explicitly contaminated by uncalibrated, frequency-dependent impedance mismatch in the drive line (stated in the Results). This means the comparison between the Floquet red line and the experimental ionization boundary mainly tests the smooth trend, not quantitative per-frequency prediction. The authors should quantify this uncertainty, either by calibrating the impedance and re-deriving the experimental threshold, or by showing that the residual scatter between the Floquet prediction and the experimental boundary is within the impedance-induced uncertainty. As it stands, the claim of 'excellent agreement' over frequency is not quantitatively supported.","section":"Results, Fig. 1a; Methods B"},{"comment":"The Floquet predictor neglects the readout resonator and dissipation, and the authors themselves note in Fig. 3 that the threshold prediction breaks down near strong low-amplitude resonances due to hybridization with the |0> Floquet mode. This is an acknowledged limitation, but the manuscript does not delineate the domain of validity of the predictor. The paper should state clearly the conditions under which the Floquet threshold is expected to agree with the full dynamics (e.g., away from strong (n:n) multiphoton resonances, for drive frequencies in the window studied, and for ramp times in the quasi-adiabatic regime), or provide a systematic comparison of Floquet-predicted thresholds against full simulation thresholds across the entire frequency–amplitude plane, including the resonance regions.","section":"Methods C and Fig. 3"}],"minor_comments":[{"comment":"The caption for Fig. 2c states 'Additional distributions for drive amplitudes differing by 1 MHz' but the text says 0.1% and 0.2% changes; please make the amplitude offsets explicit and consistent.","section":"Fig. 2 caption and main text"},{"comment":"The caption refers to panels 'c-d' but the figure appears to contain panels (a), (b), and (c)-(d); correct the panel labels and referencing.","section":"Supplemental Fig. S7 caption"},{"comment":"There is a typo 'A veraging' in the paragraph on pulse-shape dependence (should be 'Averaging').","section":"Main text, Results"},{"comment":"The tanh-box pulse envelope is intricate; please define the parameter k clearly, since it appears to be computed from tramp but the relation is only given implicitly.","section":"Eq. (3) and definitions"},{"comment":"The fit of EJ,m to spectroscopy is described, but the errors on the fitted parameters are not reported; adding these would help assess the sensitivity of the Floquet threshold to model parameters.","section":"Supplemental Sec. III"},{"comment":"The statement that the 10 MHz grid 'skips most avoided crossings of size ≲ 1 MHz' is not quantitatively justified; a sentence explaining the relationship between grid spacing and adiabaticity during the ramp would improve clarity.","section":"Methods C"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong experimental and numerical study, and the central observation of a parametric-drive ionization threshold is likely correct. My major concern is that the quantitative predictive claim of the Floquet framework rests on a hand-defined clustering criterion, and the experimental validation is partly masked by impedance mismatch. These issues are addressable with additional robustness tests and uncertainty quantification. I would not reject, but the current version overclaims the predictive power. The reference list is appropriate and the authors engage closely with prior work on transmon ionization and chaos."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nQuick take: this paper convincingly shows that a transmon driven far below its transition frequency ionizes above a drive-amplitude threshold, and that a Floquet-based criterion predicts that threshold across frequency, initial state, and ramp time. It extends the ionization phenomenology from readout to parametric control, which is the genuinely new piece. The experiment and the full time-dependent simulations match strikingly well, and the model parameters are fitted to independent spectroscopy, with the drive amplitude calibrated through Rabi fringes, so the threshold is a prediction rather than a fit to the target data. The Floquet predictor is compared against both experiment and the independent TDSE simulation, so the core quantitative claim is not circular.\n\nThe main soft spot is the threshold extraction itself. Methods C defines \"entry into the chaotic manifold\" via k-means clustering of ng-averaged Floquet-mode populations, with a hand-chosen M−2 cutoff and visual identification of the chaotic cluster. That is arbitrary, and the stress-test worry is legitimate: if reasonable automated variants move the predicted threshold by an amount comparable to the experimental uncertainty, the agreement would be partly a property of the criterion. That said, the predictor tracks the TDSE ionization boundary across frequency, initial state, ramp time, and pulse duration, which suggests the criterion is not fine-tuned to a single curve. Still, the paper would be stronger with a robustness check, e.g., showing how the predicted threshold varies with the clustering cutoff or with an automated cluster-identification rule, and with code/data to let others test it.\n\nTwo smaller reservations. First, the Floquet model neglects dissipation, yet Fig. 4 attributes the pulse-duration dependence in the experiment to spontaneous emission; that attribution is plausible but unmodeled, so the quantitative reach of the framework is unclear. Second, the experimental frequency dependence is contaminated by uncalibrated cable impedance, which they acknowledge; the comparison mainly constrains the smooth trend. The broader \"unifying framework\" claim is a reasonable hypothesis but goes beyond the single-device demonstration, and the paper itself flags that the threshold prediction breaks down near strong low-amplitude resonances (Fig. 3 caption).\n\nOverall, this is a solid, honest experimental paper with a new result and a useful predictor. The central claim holds up. The ad hoc elements are worth fixing or at least defending with a sensitivity analysis, but they do not undermine the observation or the qualitative framework.\n\nRecommendation: yes, send this to peer review. A serious referee should focus on the robustness of the Floquet threshold criterion and on quantifying how dissipation and cable impedance affect the comparison. I'd bring it to a reading group and would cite it if I were working on parametric gates or transmon ionization.","headline":"A solid experimental demonstration that parametric drive breakdown in a transmon is predicted by Floquet analysis, with an ad hoc threshold criterion that deserves a robustness check.","tokens_in":23142,"tokens_out":2096,"would_cite":true,"duration_ms":21720,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A transmon driven far below its transition frequency ionizes above a drive-amplitude threshold, and the instantaneous Floquet spectrum quantitatively predicts that threshold across drive frequency and initial state.","keywords":["transmon ionization","parametric control","Floquet spectrum","drive-induced chaos","superconducting circuits","subharmonic drive","Josephson harmonics","strong-drive threshold"],"falsifier":"Replace the k-means M-2 threshold rule with an objective criterion, for instance the drive amplitude at which the ground-state Floquet mode population reaches half the chaotic-cluster level or the point where the population gap between clusters closes, and recompute the predicted threshold curve shown in Fig. 1; if the curve moves by more than the observed width of the ionization transition, the hand-defined rule is load-bearing and the predictive claim fails. A device-level test would compare predicted and measured thresholds for a transmon with a substantially different anharmonicity or Josephson-harmonic content.","tokens_in":1907,"feed_emoji":"⚛️","tokens_out":5581,"duration_ms":123240,"temperature":0.7,"pith_summary":"Superconducting quantum processors often use strong off-resonant drives to activate parametric gates, and increasing drive amplitude is the natural way to make those gates faster. This paper demonstrates experimentally and numerically that a transmon serving as a nonlinear coupler ionizes once the drive amplitude crosses a threshold, with its population spreading from the computational states into roughly thirty excited levels, many above the cosine-potential well. The ionization is not caused by a single multiphoton resonance; it occurs when the ground-state Floquet mode joins a strongly hybridized layer of modes that grows from the top of the potential well downward. A predictor built from the instantaneous Floquet spectrum reproduces the measured threshold over the full 1–2 GHz drive-frequency window and for initial states |0t>, |1t>, and |2t>. If correct, this places a fundamental speed limit on parametric control of superconducting circuits and gives a concrete design criterion for avoiding drive-induced decoherence.","feed_headline":"Far-off-resonance drive ionizes transmons above a sharp threshold","feed_subtitle":"Parametric gates on superconducting circuits hit a speed limit set by drive-induced chaotic Floquet modes.","key_machinery":"The central object is the instantaneous Floquet spectrum of the periodically driven transmon: the quasienergies and Floquet modes obtained by diagonalizing the unitary propagator over one drive period, with the Hamiltonian including four Josephson harmonics and offset charge. Its job is to expose avoided crossings that signal multiphoton resonances and to reveal, through each mode's average transmon population, a band of strongly hybridized modes that grows as the drive amplitude increases. The threshold predictor itself is an operational rule: after averaging over offset charge, the mode populations are separated into two clusters by k-means, one cluster is identified as the chaotic layer, and the ionization threshold is the lowest amplitude at which the ground-state Floquet mode's population exceeds the cluster mean minus two (M-2). A finite 10 MHz step in drive amplitude is used to skip weak avoided crossings that would be crossed diabatically during fast ramps.","core_discovery":"The paper's central claim is that ionization of a strongly and far-detuned driven transmon is a general phenomenon, governed by drive-induced chaos-like hybridization, and that the instantaneous Floquet spectrum provides a quantitative predictor of the onset. Using the full transmon Hamiltonian with four Josephson harmonics and offset charge, the authors diagonalize the one-period propagator U(T,0) and track quasienergies and Floquet modes as the drive amplitude grows. Avoided crossings mark multiphoton resonances, and the offset-charge-averaged populations of the Floquet modes show a bunching layer of strongly hybridized high-energy modes that expands toward the computational states as the amplitude increases; ionization occurs when the computational Floquet mode joins this layer. The threshold is operationally defined by k-means clustering of the averaged populations with an M-2 cutoff, and it agrees with both exact Schrödinger simulations that include the readout resonator and with the measured two-dimensional frequency-amplitude maps. The paper also shows that initializing in |1t> or |2t> lowers the threshold, slower ramps lower it through adiabatic Landau-Zener passage, and longer pulses lower it experimentally through dissipation not included in the closed-system simulation.","pith_inferences":["Editorial inference: If the Floquet-cluster criterion is as robust as the two-dimensional agreement suggests, the same one-period diagonalization could be applied to other nonlinear elements such as fluxonium or SQUID-based couplers to predict their parametric thresholds before fabrication; the paper stops at hypothesizing generality.","Editorial inference: The observed 0.1%-amplitude sensitivity above threshold implies that in a multi-qubit device a slightly mistuned drive could push one coupler over threshold and, through coupling, disturb neighboring modes; the paper raises this as a concern but does not demonstrate cascades.","Editorial inference: The pulse-duration dependence seen experimentally but not in the closed-system simulation hints that a Lindblad master-equation extension of the Floquet predictor, including decay into the readout resonator, would be needed for quantitative threshold curves in lossy devices.","Editorial inference: The fact that the threshold persists across all drive frequencies and initial states suggests that the mechanism is generic to Josephson-junction circuits with low impedance, and that drive-amplitude limits should be included in the error budget of any parametric gate protocol."],"forward_implications":["Fast parametric gates on transmon-based couplers are bounded: above the threshold the coupler leaves the computational subspace, so increasing drive amplitude stops buying gate speed and instead ionizes the device.","The Floquet predictor yields a frequency-dependent maximum drive curve that can be computed from spectroscopy data before fabrication, providing a design rule for avoiding the breakdown regime.","Initializing a coupler in a higher-energy state lowers the threshold by nearly a factor of two for |2t> compared with |0t>, so excited-state operations and reset protocols are especially constrained.","Pulse shaping matters: slower ramps ionize at lower amplitudes because the system adiabatically follows avoided crossings, while longer pulses ionize earlier in experiment through dissipation absent from the closed-system simulation.","Including higher-order Josephson harmonics shifts the predicted threshold by up to 2 GHz at low drive frequencies, so the multi-harmonic transmon model is essential for quantitative predictions."],"supporting_citations":[{"why":"Provides the Floquet-spectrum method for predicting transmon ionization during dispersive readout that this paper extends to far-detuned parametric drives.","marker":"[10]"},{"why":"Shows that driven transmon ionization tracks the chaotic layer of the corresponding classical rotor, the interpretation used for the bunching layer.","marker":"[9]"},{"why":"Supplies the Floquet Landau-Zener framework for adiabatic microwave ionization from which the instantaneous-spectrum analysis descends.","marker":"[14]"},{"why":"Develops the exact time-dynamics simulation of transmon ionization that the paper adapts, including the readout resonator and pulse ramps.","marker":"[8]"},{"why":"Characterizes measurement-induced state transitions in the rotating-wave approximation, the readout counterpart this paper contrasts with large-detuning parametric driving.","marker":"[7]"},{"why":"Introduces subharmonic-drive transmon control, the fast-gate application whose speed is bounded by the predicted amplitude threshold.","marker":"[27]"},{"why":"Supplies the offset-charge dependence of transmon transitions used for averaging the Floquet populations over gate charge.","marker":"[31]"},{"why":"Documents why weak avoided crossings are traversed diabatically during fast ramps, justifying the 10 MHz spectral resolution of the Floquet calculation.","marker":"[33]"},{"why":"Reports reduced fidelity in high-on-off-ratio parametric operations, the practical effect the paper suggests the threshold may explain.","marker":"[30]"},{"why":"Underpins the association between chaotic classical dynamics and parameter sensitivity of quantum motion used for the 0.1%-amplitude dependence.","marker":"[38]"}],"fun_headline_variants":["Chaotic Floquet modes explain transmon drive breakdown","Floquet spectrum pinpoints drive-induced transmon ionization","Strong parametric drives ionize transmons via chaotic modes","Transmon ionization threshold set by chaotic Floquet spectrum"],"cache_read_input_tokens":25088,"weakest_assumption_plain":"The entire threshold prediction rests on the hand-chosen rule that groups the offset-charge-averaged Floquet mode populations into a chaotic cluster and declares the threshold where the computational mode's population exceeds the cluster mean minus two; if that rule does not track the physical ionization boundary, the predictor's apparent agreement may be coincidental.","fun_headline_variants_meta":{"raw":{"variants":["Chaotic Floquet modes explain transmon drive breakdown","Floquet spectrum pinpoints drive-induced transmon ionization","Strong parametric drives ionize transmons via chaotic modes","Transmon ionization threshold set by chaotic Floquet spectrum"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001012,"raw_usage":{"total_tokens":4286,"prompt_tokens":965,"completion_tokens":3321,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":581,"completion_tokens_details":{"reasoning_tokens":3255}},"tokens_in":581,"tokens_out":3321,"duration_ms":26623,"temperature":1.0,"reasoning_tokens":3255,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:02:04.835187+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Replace the k-means M-2 threshold rule with an objective criterion, for instance the drive amplitude at which the ground-state Floquet mode population reaches half the chaotic-cluster level or the point where the population gap between clusters closes, and recompute the predicted threshold curve shown in Fig. 1; if the curve moves by more than the observed width of the ionization transition, the hand-defined rule is load-bearing and the predictive claim fails. A device-level test would compare predicted and measured thresholds for a transmon with a substantially different anharmonicity or Josephson-harmonic content.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Develops the exact time-dynamics simulation of transmon ionization that the paper adapts, including the readout resonator and pulse ramps."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Characterizes measurement-induced state transitions in the rotating-wave approximation, the readout counterpart this paper contrasts with large-detuning parametric driving."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the offset-charge dependence of transmon transitions used for averaging the Floquet populations over gate charge."}],"review_version":1}