{"id":"136f3cb3-4cf7-481a-adf0-476d9a7275de","arxiv_id":"2506.20645","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A compact superconducting reflection-less band-pass filter for quantum circuits achieves low loss, wideband absorption of reflections, and suppresses thermal photons from its termination resistors, verified with a qubit.","lead":"This paper reports a superconducting microwave filter that lets signals pass through with almost no loss while absorbing reflections instead of sending them back. The device also blocks heat noise from its own resistors, which could help protect qubits and amplifiers in quantum computers.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The abstract's 14.5 GHz return-loss bandwidth exceeds the paper's own calibration error ceiling, which limits resolvable 10 dB return loss to 11–13 GHz.","rationale":"The reader's weakest-assumption analysis correctly identifies the return-loss calibration ceiling as the load-bearing issue, and the manuscript text itself confirms it: Section V.A gives two independent upper limits (13 GHz from standards error, 11 GHz from cable tracking/matching error) above which a 10 dB return loss cannot be resolved. Yet the abstract and conclusion claim 14.5 GHz and 14 GHz, respectively. This is not a stylistic discrepancy; the central specification of the device is a measured quantity, so an unverifiable measurement bandwidth makes the headline claim unsupported. The concern is concrete and falsifiable: either the error analysis is too pessimistic, or the abstract overstates the verified range. A corrected abstract or a new calibration measurement would resolve it. I do not see a deeper internal inconsistency in the filter theory or in the thermal-photon suppression experiment; the heating measurement with a qubit device gives independent, if qualitative, support for the pass-band thermal-photon suppression claim, and the normalized S21 data support the low insertion loss. Therefore the appropriate outcome is the reader's conditional verdict, not rejection.","tokens_in":16634,"tokens_out":3945,"duration_ms":51237,"concrete_test":"Recompute the calibrated S11 from the raw switched-path data with an explicit error cutoff: retain only frequency points where the Monte Carlo error locus (Fig. 12) and the cable error-gain expression (Eq. 22) indicate less than 10 dB uncertainty, and report the highest frequency at which the measured return loss is known to exceed 10 dB. If that frequency is below 14.5 GHz, amend the abstract to state the verified bandwidth (e.g., to 11 GHz or 13 GHz) rather than the extrapolated one. To settle the claim affirmatively, repeat the calibration with cryogenic standards and switch paths characterized to 18 GHz, or place a known low-reflection through at the DUT reference plane and verify residual error remains below 10 dB to 14.5 GHz.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is the packaged 10 dB return loss from DC to above 14.5 GHz. Section V.A explicitly undermines this: the corrected-standards error exceeds 0.316 (10 dB) at 13 GHz, and cable tracking/matching error limits observable 10 dB return loss to at or below 11 GHz. The conclusion states 'vector calibrated S11 below 10 dB from 1 to 14 GHz,' which is inconsistent with both the abstract's 14.5 GHz figure and the paper's own error analysis. The headline value is presented as a measured result, not a simulation prediction, so the 14.5 GHz bandwidth is not established. The device may still meet the spec, but the reported measurement cannot resolve it. A second, related caveat is that the measured device used 16-ohm resistors against the 26-ohm design target, so the demonstrated part differs from the designed part; the measured 13.5 dB stop-band return loss still clears 10 dB, but the margin is thinner than the design would imply.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the design, fabrication, and cryogenic characterization of a superconducting reflection-less band-pass filter intended for protecting quantum circuits from return-loss-induced errors and for suppressing thermal photons. The filter is synthesized using Morgan's dual-network theory, fabricated with Al on high-resistivity Si with NiCr resistors, and measured in a connectorized package at 20 mK. The authors report low insertion loss (<1 dB with packaging), wideband return loss, and a differential qubit-heating experiment showing that heating the filter's resistors does not raise the qubit's effective temperature, unlike a conventional attenuator.","tokens_in":16826,"tokens_out":2357,"duration_ms":28393,"significance":"If the headline claims are established, this is a valuable component for superconducting quantum microwave chains: a compact (0.6 mm^2) filter that absorbs out-of-band reflections rather than reflecting them, with low pass-band loss and a built-in thermal-photon suppression mechanism. The qubit-heating differential measurement is a strong, independent validation of the thermal-photon suppression claim, and the design methodology using S-parameter cascades and partial full-wave simulation is practical and transferable. However, the central return-loss bandwidth claim is not supported by the paper's own error analysis, so the measured performance envelope needs to be restated or the measurement capability improved.","major_comments":[{"comment":"The abstract claims '10 dB packaged return loss from DC to above 14.5 GHz' and the conclusion claims 'vector calibrated S11 below 10 dB from 1 to 14 GHz', but Section V.A states that the standards correction error exceeds 0.316 (i.e., 10 dB return loss) at 13 GHz, and further that cable tracking/matching errors limit the observable 10 dB return loss to at or below 11 GHz. These statements are mutually inconsistent: the highest frequency at which the measurement can resolve a 10 dB return loss is 11 GHz, not 14.5 GHz. Since the wideband return-loss specification is a primary advertised feature, the authors must either (a) present a calibrated measurement with a demonstrably adequate error budget up to 14.5 GHz, or (b) explicitly revise all headline claims (abstract, conclusion, and Table I) to the supported frequency range, with a clear statement that higher-frequency return loss is not established by the presented data.","section":"V.A"},{"comment":"The measured device used 16-ohm resistors against a 26-ohm design target, and the paper notes this limits the stop-band return loss to 13.5 dB. The abstract and Table I do not mention this deviation, so a reader could reasonably infer that the reported performance is for the designed circuit. The paper should separate the as-designed simulation from the as-measured device, state the resistor-value discrepancy prominently in the results section and abstract, and discuss how the reduced resistor value affects the confidence in the claimed 10 dB return-loss bandwidth. The margin above 10 dB is thinner than the design would imply, particularly in the stop band, which matters for the headline claim.","section":"VI"}],"minor_comments":[{"comment":"In the text before Eq. (5), 'band bass version' should be 'band-pass version'.","section":"II.A"},{"comment":"The text says 'Monty Carlo' twice; it should be 'Monte Carlo'.","section":"V.A"},{"comment":"The phrase 'over a 80% fractional bandwidth' is grammatically incorrect; it should be 'over an 80% fractional bandwidth'.","section":"Abstract"},{"comment":"The sentence 'The standards’ error exceeds 0.316 (10 dB) at 13 GHz' should clarify that 0.316 is the voltage reflection coefficient magnitude corresponding to 10 dB return loss, to avoid ambiguity.","section":"V.A"},{"comment":"Reference [15] is incomplete: the URL contains '[insert date]' and should be replaced with the actual access date or a stable citation.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is technically interesting and the thermal-photon suppression experiment is convincing, but the abstract and conclusion overstate the measured return-loss bandwidth relative to the paper's own error analysis. The authors should either improve the calibration to actually verify 10 dB return loss to 14.5 GHz or revise all headline numbers. The 'DRAFT – NOT SUBMITTED' header and incomplete reference also suggest the manuscript is not finalized; these should be cleaned up before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about this paper. First, the thermal-photon suppression measurement is the real gem: a qubit-based differential test showing that heating the filter's resistors does not raise the qubit's effective temperature, while an attenuator does. That's a clean, useful result. Second, the main advertised spec — 10 dB packaged return loss from DC to above 14.5 GHz — is not supported by the paper's own calibration error analysis. Section V.A says the standards error exceeds 10 dB at 13 GHz and that the observable 10 dB return loss is limited to at or below 11 GHz. The conclusion even says '1 to 14 GHz,' which disagrees with both the abstract and the error budget. So the headline bandwidth is unestablished.\n\nWhat's new and good: the cryogenic superconducting implementation with NiCr resistors on Al-on-Si, the explicit treatment of symmetry-plane delay and mutual coupling via Neumann integrals and tee networks, and the qubit heating experiment. The design methodology is detailed and honest about non-ideal effects. The fabricated device was off-target — resistors came out 16 ohms against the 26-ohm design — but the measured stop-band return loss still clears 10 dB, so this is a caveat, not a fatal flaw.\n\nThe main soft spot is the inconsistency between the measured data, the error analysis, and the abstract/conclusion. The authors need to either present corrected data with error bars showing the actual resolvable bandwidth, or soften the claim to what the measurement can support (e.g., 10 dB up to ~11 GHz, or state the spec as simulated). The 14.5 GHz number looks like it comes from design simulation, not measurement. That's a load-bearing issue for the abstract but fixable in revision.\n\nThe paper deserves peer review: the device concept is useful, the thermal suppression result is a meaningful verification, and the fabrication process is relevant for integration with qubits and TWPAs. I'd send it with a request to clarify the measurement limits and correct the claims. I wouldn't cite the 14.5 GHz spec as measured until that's resolved.","headline":"The device work is real and the thermal-photon suppression measurement is the strongest part, but the paper's own error budget undercuts the headline 14.5 GHz return-loss claim; that needs fixing.","tokens_in":17327,"tokens_out":2580,"would_cite":false,"duration_ms":26470,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A 0.6 mm² superconducting filter absorbs reflections and blocks thermal photons in one package, with insertion loss below 1 dB over an 80% bandwidth centered at 8 GHz.","keywords":["reflection-less filters","superconducting circuits","cryogenic microwave filters","thermal photon suppression","traveling-wave parametric amplifiers","return loss","dual network synthesis","quantum computing hardware"],"falsifier":"Recalibrate the packaged filter at 20 mK using a low-loss cryogenic switch or a direct on-wafer two-port measurement with a vector network analyzer whose tracking error is below 1 dB up to 15 GHz, then check whether $S_{11}$ remains below -10 dB continuously from DC through 14.5 GHz; if the return loss crosses -10 dB at any frequency below 14.5 GHz, the abstract's headline bandwidth is not established.","tokens_in":2056,"feed_emoji":"❄️","tokens_out":3953,"duration_ms":67675,"temperature":0.7,"pith_summary":"This paper claims that a monolithic superconducting reflection-less filter can combine three functions that quantum microwave chains usually need separate components for: low-loss band-pass filtering, wideband absorption of reflections, and suppression of thermal photons emitted by its own terminating resistors. The device, fabricated with aluminum on silicon plus nichrome resistors, is small enough to co-fabricate with qubits and traveling-wave parametric amplifiers. If the claim holds, a single compact component could replace or supplement circulators and attenuators, improving amplifier gain flatness and qubit coherence by preserving a 50-ohm impedance environment. The paper also identifies symmetry-plane delay and mutual inductance as the main physical limits to the ideal reflection-less condition, and presents a design method that incorporates these effects without full-wave simulation of the entire filter.","feed_headline":"Tiny superconducting filter stops reflections and stray heat","feed_subtitle":"A 0.6 mm² dual-network device keeps pass-band loss under 1 dB while absorbing out-of-band signals for qubit lines.","key_machinery":"The central mechanism is the dual-network reflection-less filter: a symmetric two-port network designed so that its even-mode and odd-mode reflection coefficients are negatives of each other, which forces $S_{11}=S_{22}=0$ while allowing arbitrary reciprocal transmission. The band-pass version is synthesized from a lumped-element filter whose reactances are replaced by their reciprocal impedances (inductors become capacitors and vice versa), with resistors equal to the characteristic impedance. The paper identifies three non-ideal effects that spoil the perfect reflection-less condition — component tolerance, delay across the symmetry plane, and mutual coupling between spiral inductors — and handles the last two explicitly: delay is modeled analytically, and mutual inductance is computed from a Neumann-integral approximation and included in circuit simulation by converting coupled inductors into cascaded tee networks. This S-parameter cascade method lets the designers optimize inductor winding directions without full-wave simulation of the entire structure.","core_discovery":"The paper's central claim is that a reflection-less band-pass filter can be realized in a superconducting, qubit-compatible process and packaged without losing its key specifications. The filter is built from a symmetrical two-port dual network in which the even-mode and odd-mode reflection coefficients are made equal and opposite, so the two-port reflection coefficients vanish while transmission remains frequency-selective. The fabricated device occupies 0.6 mm² and, including its connectorized package at 20 mK, achieves insertion loss below 1 dB over an 80% fractional bandwidth centered at 8 GHz, with return loss of 10 dB or better from DC to above 14.5 GHz as stated in the abstract. The paper further claims that the dual topology makes the transfer function from the termination resistors to the microwave ports a notch filter in the pass band, so thermal photons generated by the resistors are suppressed by orders of magnitude; this is supported by a qubit-based heating experiment in which DC heating of the filter resistors produced no measurable increase in the qubit's effective temperature, unlike a cryogenic attenuator in the same configuration.","pith_inferences":["The paper's abstract claim of '10 dB packaged return loss from DC to above 14.5 GHz' is not actually supported by its own measurement-error analysis, which states that the observable 10 dB return loss is limited to at or below 11 GHz because of cable and switch tracking and matching errors; a reader should treat the higher-frequency portion of the headline specification as an extrapolation rather ","The dual-network thermal-photon suppression property is likely general: any absorptive component built as a matched dual network should show a notch between its terminating resistors and its pass band, so the same idea could be applied to other cryogenic absorptive elements such as matched attenuators or isolator replacements.","The measured stop-band return loss of 13.5 dB was limited by the realized nichrome sheet resistance being 16 ohms instead of the 26-ohm target; improving resistor process control should directly improve the out-of-band reflection-less performance without changing the topology.","The qubit heating experiment demonstrates a practical test protocol for comparing any microwave component's thermal-photon emission: heat its internal resistors with DC power and monitor qubit effective temperature, which is a more direct metric for quantum hardware than raw noise-power measurements."],"forward_implications":["If the filter performs as claimed, a single compact device can provide band-pass filtering, wideband impedance matching, and thermal-photon suppression on the same qubit-compatible chip.","Placing the filter at the output of a traveling-wave parametric amplifier should flatten amplifier gain ripple caused by out-of-band reflections from circulators and isolators, as the paper's simulations show.","The filter's notch transfer function from its resistors to the RF ports means that even when pump or control signals dissipate power in the filter's stop band, the qubit pass band remains largely shielded from the resulting thermal photons.","The S-parameter cascade design methodology, including the Neumann-integral mutual-inductance search, makes it practical to optimize reflection-less filters monolithically without time-consuming full-wave simulation of every layout iteration.","Because the process is compatible with standard aluminum-on-silicon superconducting circuits, the filter can be added to existing qubit and readout fabrication flows without a separate packaging step."],"supporting_citations":[{"why":"Morgan and Boyd's synthesis procedure for reflection-less filters supplies the core even/odd-mode dual-network method that this paper adapts to a superconducting band-pass version.","marker":"[12]"},{"why":"Kurokawa's scattering-matrix power-wave formalism provides the passivity condition that permits a frequency-selective two-port network with zero reflection.","marker":"[17]"},{"why":"Reed and Wheeler's even/odd-mode analysis of symmetrical networks is the basis for the $S_{11}=0$ condition that makes the filter reflection-less.","marker":"[20]"},{"why":"Yeh et al.'s superconducting microwave attenuators define the target nichrome resistor process and the low-temperature heating measurement methodology used here.","marker":"[25]"},{"why":"Yan et al.'s flux-qubit coherence model provides the method for extracting qubit effective temperature from measured T1 and T2e, which underpins the thermal-photon suppression claim.","marker":"[8]"},{"why":"Wang et al.'s cryogenic single-port calibration approach is the starting point for the switched-standards S11 calibration system used at 20 mK.","marker":"[29]"},{"why":"Glasser's error-gain analysis of microwave de-embedding is used to quantify how cable tracking and matching errors limit the observable return-loss bandwidth.","marker":"[32]"},{"why":"Dunsworth et al.'s multi-layer wiring method supplies the trilayer resist and air-bridge fabrication process used to build the filter's capacitors and crossovers.","marker":"[27]"},{"why":"Neumann's formula for mutual inductance is the basis for the numerical integration that estimates coupling between the spiral inductors.","marker":"[24]"}],"fun_headline_variants":["Superconducting filter kills reflections, chills qubits","Reflection-less filter shields qubits from noise and heat","Tiny filter for qubits: no reflections, no stray heat","Wideband filter stops reflections, cools qubit lines","Superconducting filter: zero reflections, low loss for quantum circuits"],"cache_read_input_tokens":19584,"weakest_assumption_plain":"The headline return-loss specification rests on the assumption that the measurement system can actually resolve 10 dB return loss up to 14.5 GHz, but the paper's own error budget says the cables and switches limit observable 10 dB return loss to at or below 11 GHz.","fun_headline_variants_meta":{"raw":{"variants":["Superconducting filter kills reflections, chills qubits","Reflection-less filter shields qubits from noise and heat","Tiny filter for qubits: no reflections, no stray heat","Wideband filter stops reflections, cools qubit lines","Superconducting filter: zero reflections, low loss for quantum circuits"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000161,"raw_usage":{"total_tokens":1240,"prompt_tokens":956,"completion_tokens":284,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":572,"completion_tokens_details":{"reasoning_tokens":199}},"tokens_in":572,"tokens_out":284,"duration_ms":3558,"temperature":1.0,"reasoning_tokens":199,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:43:46.178921+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recalibrate the packaged filter at 20 mK using a low-loss cryogenic switch or a direct on-wafer two-port measurement with a vector network analyzer whose tracking error is below 1 dB up to 15 GHz, then check whether $S_{11}$ remains below -10 dB continuously from DC through 14.5 GHz; if the return loss crosses -10 dB at any frequency below 14.5 GHz, the abstract's headline bandwidth is not established.","supporting_citations":[{"cited_title":"Theoretical and experimental study of a new class of reflectionless filter,","cited_arxiv_id":null,"evidence_quote":"Morgan and Boyd's synthesis procedure for reflection-less filters supplies the core even/odd-mode dual-network method that this paper adapts to a superconducting band-pass version."},{"cited_title":"Power waves and the scattering matrix,","cited_arxiv_id":null,"evidence_quote":"Kurokawa's scattering-matrix power-wave formalism provides the passivity condition that permits a frequency-selective two-port network with zero reflection."},{"cited_title":"A method of analysis of symmetrical four- port networks,","cited_arxiv_id":null,"evidence_quote":"Reed and Wheeler's even/odd-mode analysis of symmetrical networks is the basis for the $S_{11}=0$ condition that makes the filter reflection-less."},{"cited_title":"Microwave attenuators for use with quantum devices below 100 mk,","cited_arxiv_id":null,"evidence_quote":"Yeh et al.'s superconducting microwave attenuators define the target nichrome resistor process and the low-temperature heating measurement methodology used here."},{"cited_title":"The flux qubit revisited to enhance coherence and reproducibility,","cited_arxiv_id":null,"evidence_quote":"Yan et al.'s flux-qubit coherence model provides the method for extracting qubit effective temperature from measured T1 and T2e, which underpins the thermal-photon suppression claim."},{"cited_title":"Cryogenic single-port calibration for superconducting microwave resonator measurements,","cited_arxiv_id":null,"evidence_quote":"Wang et al.'s cryogenic single-port calibration approach is the starting point for the switched-standards S11 calibration system used at 20 mK."},{"cited_title":"An analysis of microwave de-embedding errors (technical notes),","cited_arxiv_id":null,"evidence_quote":"Glasser's error-gain analysis of microwave de-embedding is used to quantify how cable tracking and matching errors limit the observable return-loss bandwidth."},{"cited_title":"Allgemeine gesetze der inducierten elektrischen str ¨ome,","cited_arxiv_id":null,"evidence_quote":"Neumann's formula for mutual inductance is the basis for the numerical integration that estimates coupling between the spiral inductors."}],"review_version":1}