{"id":"c8db75a0-1420-448d-9a68-d4a57cf87414","arxiv_id":"2411.12782","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Three SNS bolometers on a single chip are frequency-multiplexed and read out simultaneously with crosstalk between -11.8 and -17.9 dB.","lead":"Researchers built three heat sensors, called bolometers, on one chip and used separate microwave frequencies to trigger and read out each one at the same time with little interference. This is a step toward operating future quantum computers with fewer cables, amplifiers, and other bulky microwave components.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Crosstalk isolation is inferred from a fitted P1dB model, not measured directly, and the printed Table I does not transparently reproduce the stated -11.8 to -17.9 dB range; the ~12 dB safe margin is the load-bearing point that needs independent confirmation.","rationale":"The reader's conditional verdict targets exactly the right load-bearing assumption: the sufficiency of the on-chip filter isolation. My stress-test confirms that this is the weakest step in the central argument. The strongest credited evidence is the direct time-domain demonstration that three bolometers can be individually triggered and read out simultaneously, which is a genuine proof of principle. However, the numerical crosstalk claim, the quantitative basis for 'very low cross talk' and the 12 dB safe margin, is not backed by a direct S-parameter or leakage measurement. It is recovered from fits of the power-dependent response, and the numbers in Table I do not obviously reproduce the quoted -11.8 to -17.9 dB range under the stated P1dB-difference procedure. This is a concrete, checkable inconsistency rather than a matter of disagreement with consensus. Because the reader already marked the paper CONDITIONAL and cited the same underlying gap, my read does not move the verdict; it sharpens the required verification. The proposed direct ratio test would settle whether the inferred isolation is trustworthy.","tokens_in":10943,"tokens_out":12679,"duration_ms":121410,"concrete_test":"Reanalyze the raw data behind Fig. 3(d)-(f) and Table I without the shared phenomenological fit: for each bolometer and each off-resonant heater frequency, take the measured response at a fixed high heater power (e.g., -115 dBm) and divide by the on-resonance response at the same power to obtain a direct isolation ratio in dB. Compare this ratio with the claimed -11.8 to -17.9 dB range and the ~12 dB safe margin. If any channel shows less than 12 dB direct isolation, or if the direct ratio differs by more than 2 dB from the fit-inferred values, the reported crosstalk margin and the scaling conclusion are not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The scaling claim depends on the CPW input filters isolating heater channels with at least ~12 dB attenuation. The paper's evidence is not a direct leakage measurement: it consists of P1dB compression points (Table I) obtained by fitting each bolometer's power-dependent response to a phenomenological model from Ref. [2], with crosstalk inferred from differences between these fitted points. Applying the stated procedure to Table I gives on/off-resonance P1dB differences of roughly 12.0 dB (B3), 19.1-21.2 dB (B2), and 21.9-26.3 dB (B1); the text instead reports a range of -11.8 to -17.9 dB and 'roughly 12 dB' of safe operating range. The printed table therefore does not transparently reproduce the quoted range unless additional fit correlations or model assumptions are invoked. The simultaneous multiplexing runs use low heater power (-135 dBm) and heavy averaging, where leakage can be hidden below noise, so they do not independently validate the margin at powers near saturation or in a larger array. If the true isolation is weaker, or if the P1dB fit systematically shifts the inferred attenuation, the crosstalk-limited scaling claim is the part that fails.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports the design, fabrication, and characterization of three superconductor–normal-conductor–superconductor (SNS) bolometers on a single chip. The heater inputs are isolated by on-chip CPW band-pass filters centered at 7.6, 5.8, and 4.4 GHz, while the three probe tank circuits resonate at 194, 157, and 180 MHz and share a common readout line. The authors characterize the probe resonances and filter responses, measure time-domain responses to individual heater pulses, and fit the power dependence of the detector response to extract one-dB compression points, from which they infer a heater crosstalk of −11.8 dB to −17.9 dB. They then demonstrate simultaneous frequency-multiplexed triggering of all three bolometers in the eight on/off combinations, with per-channel SNR around 7–8, and discuss the implications for calorimetric readout and for scaling to larger arrays.","tokens_in":11136,"tokens_out":12455,"duration_ms":114786,"significance":"If the quantitative results hold, this is a valuable experimental demonstration: it is the first direct showing I am aware of that multiple SNS bolometers on one chip can be individually and simultaneously actuated through frequency-multiplexed heater and probe circuits with little interference at the chosen operating point. The raw time-domain data in Fig. 4 and the signal-to-leakage ratios in Table II are concrete evidence for low crosstalk at the operating power, and the paper contains useful engineering details such as measured filter distortions, thermal time constants, and operation-point selection. The main limitation is that the headline isolation range reported in Section IV.A is not transparently reproduced by Table I, and the crosstalk is inferred from fitted compression points rather than measured directly; these issues affect the strength of the crosstalk claim and the extrapolation to larger arrays.","major_comments":[{"comment":"The quoted crosstalk range of −11.8 dB to −17.9 dB is not reproducible from the stated procedure and the data in Table I. Taking the difference between on-resonance and off-resonance P1dB values as the paper suggests yields 12.0 dB for B3 (4.4 GHz versus both 5.8 and 7.6 GHz), 19.1–21.2 dB for B2 (5.8 GHz versus 7.6 and 4.4 GHz), and 21.9–26.3 dB for B1 (7.6 GHz versus 4.4 and 5.8 GHz). Neither 11.8 dB nor 17.9 dB appears in this calculation, and the discrepancy is far larger than the quoted fit uncertainties. Please state explicitly how the P1dB differences are converted to crosstalk, including any use of fit correlations, correct the numerical range, or replace it with a direct leakage measurement. The statement that the safe range is limited by f1 is also unclear because the smallest on/off-resonance P1dB difference in Table I occurs for B3 at f3 = 4.4 GHz.","section":"Section IV.A, Table I"},{"comment":"The crosstalk estimate is inferred from a phenomenological P1dB fit rather than from a direct measurement of the power leaking through the input filters, and the raw time traces at a heater power of −135 dBm can only establish that leakage is below the detection threshold of those traces; they do not quantify the isolation. For example, B3 has an on-resonance P1dB of −132 dBm, so a −135 dBm pulse is only 3 dB below the fitted compression point, whereas a leakage signal suppressed by roughly 12 dB would arrive at about −147 dBm and might be hidden by the noise floor of the time-domain measurement. The simultaneous multiplexing demonstration therefore validates crosstalk at the specific low operating power but not the general isolation claim near saturation or in a larger array. Please provide a direct leakage measurement, such as the power transmitted to a nonaddressed bolometer's absorber, or quantify the smallest leakage detectable in the time-domain traces.","section":"Section IV.A, Figs. 3 and 4, Table II"}],"minor_comments":[{"comment":"There are several duplicated-word and typographical errors, including “are are” in the abstract and Section I, “Thi” in Section I, and “for for” and “the the” in Section IV.B.","section":"Abstract, Section I, Section IV.B"},{"comment":"The caption for panels (d)–(f) refers to them as “(a) 157 MHz, (b) 180 MHz, and (c) 194 MHz”; it should refer to panels (d), (e), and (f), respectively.","section":"Fig. 3 caption"},{"comment":"The text states the probe power is −144 dBm at the device input, while the Fig. 2 caption says approximately −140 dBm; these values should be reconciled or the reference point clarified.","section":"Section III.A and Fig. 2 caption"},{"comment":"The phrase “approximately −125 dB” should read “−125 dBm” to be dimensionally consistent.","section":"Section III.A"},{"comment":"The estimate of 180 multiplexed bolometers is presented as if it followed from probe-frequency crowding alone; it does not include heater-side filter isolation, the measured filter line-shape distortions, or crosstalk-induced saturation, so it should be presented as an upper bound from probe-frequency considerations only or removed.","section":"Section V"},{"comment":"The statement that leakage is “lower by more than an order of magnitude than the actual signal” is imprecise; Table II reports ratios of averaged signal to pre-pulse standard deviation, not calibrated leakage in dB.","section":"Section IV.B, Table II"}],"recommendation":"major_revision","confidential_remarks":"The experimental demonstration appears sound in its raw time-domain form, and the work is within scope for the journal. The main reason for major revision is the mismatch between the quoted crosstalk range and what Table I yields under the stated procedure, plus the indirect nature of the crosstalk extraction. Both issues are fixable without new physics, but they are load-bearing for the headline claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is the first chip with three SNS bolometers selectively triggered through frequency-multiplexed heaters and read out through a shared probe line. The time traces in Fig. 4 and the SNR numbers in Table II show the central claim is real: each bolometer responds only when its own heater is on, even with two or three heaters running at once. That is a genuine step for bolometric qubit readout and deserves refereeing.\n\nWhat's new: previous SNS work had single devices, and the multiplexing scheme was suggested in Ref. [2]; here it is actually built and operated. The on-chip CPW filters, shared probe line, and tank circuits tuned to 157/180/194 MHz are practical engineering, and the probe-signal crosstalk between channels is small. The measured time constants (4–13 µs) are also useful for future calorimetric operation.\n\nSoft spots: the headline crosstalk number (-11.8 to -17.9 dB) is not directly measured. It is inferred from P1dB compression points fitted to a phenomenological model, and Table I as printed does not reproduce that range if you just take differences of the P1dB values. I get 12 dB for B3's off-resonance channels but 19–26 dB for the other two, so either there is an additional step in the analysis or the text is describing something else. Readers should go back to the fits and SI before trusting that number. The claimed \"roughly 12 dB safe margin\" is load-bearing for scaling; if the real isolation is closer to 12 dB, that is fine for three channels at -135 dBm and 10^4 averages, but not obviously enough for a larger array or higher power. Also, \"low-noise\" in the abstract is not supported by a measured NEP here; it leans on earlier work. That is a minor wording issue.\n\nNone of this kills the paper. The simultaneous triggering is shown directly, not just through fits. The crosstalk quantification is the soft part, not the central demonstration. I would send it to peer review and ask for a direct leakage measurement or a clear derivation of the crosstalk from the fits.","headline":"First real frequency-multiplexed SNS bolometer demonstration, with solid time-domain evidence, but the crosstalk number is more model-dependent than the text admits.","tokens_in":11771,"tokens_out":2508,"would_cite":true,"duration_ms":24457,"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":"This paper shows that three superconductor–normal-conductor–superconductor (SNS) bolometers on a single chip can be individually and simultaneously triggered through frequency-multiplexed heater and probe circuits, with crosstalk between…","keywords":["SNS bolometer","frequency multiplexing","calorimetric readout","superconducting qubit readout","crosstalk","coplanar waveguide filter","thermal detector","multiplexed readout"],"falsifier":"A direct test would be to operate a larger array with the same filter design and the same per-channel heater power: if the response of a non-addressed bolometer to a heater pulse at the resonance of an adjacent filter rises above its noise floor when two or more neighbouring heaters are active, the isolation margin is insufficient. Concretely, for the current device, a heater pulse at f1 = 7.6 GHz with power near the -106 dBm 1-dB compression point of B1 should not produce a detectable response in B2 or B3; measuring a response above the pre-pulse noise level would contradict the claimed crosstalk range.","tokens_in":10693,"feed_emoji":"⚛️","tokens_out":5453,"duration_ms":47874,"temperature":0.7,"pith_summary":"The paper sets out to show that ultrasensitive superconductor–normal-conductor–superconductor (SNS) thermal detectors can be frequency-multiplexed on a single chip, so that several bolometers share one probe line and one heater input line while still being individually addressable. The authors design and fabricate three bolometers with probe resonances near 150–200 MHz and input filters at 4.4, 5.8, and 7.6 GHz, then demonstrate that each bolometer responds only to its own heater tone. They measure crosstalk between -11.8 dB and -17.9 dB and show real-time simultaneous excitation of any subset of the three detectors. If this approach scales, it would reduce the number of readout lines and microwave isolators needed for large superconducting quantum processors, since one sensor per qubit could be read out through a shared line.","feed_headline":"Three bolometers on one chip share a single readout line","feed_subtitle":"Frequency-multiplexed SNS sensors could cut readout lines and isolators in large-scale superconducting quantum computers.","key_machinery":"The device is built around the SNS bolometer itself: a thin gold-palladium nanowire acts as both the microwave absorber and the thermometer, with a chain of superconductor–normal-conductor–superconductor junctions whose impedance depends on the electron temperature. Each bolometer is embedded in an LC tank circuit whose resonance frequency shifts with absorbed power, and is read out in reflection at a distinct probe frequency around 150–200 MHz. On the input side, each absorber is coupled to a shared heater line through a half-wavelength coplanar-waveguide band-pass filter, with center frequencies at 4.4, 5.8, and 7.6 GHz. Frequency multiplexing works because the heater filters isolate the input channels from one another and the probe resonances are separated by far more than their linewidths, so a single probe tone containing all three carrier frequencies can be demodulated digitally into three independent time traces.","core_discovery":"The central claim is that frequency multiplexing of SNS bolometers is experimentally practical. Three bolometers on one chip, each with a half-wavelength coplanar-waveguide band-pass filter at its absorber input and an LC tank circuit with a distinct probe resonance, can be actuated and read out simultaneously through shared heater and probe lines. The measured crosstalk between heater channels is between -11.8 dB and -17.9 dB, giving roughly 12 dB of safe operating range, and the probe signals show no measurable interference because the tank resonance frequencies are far apart relative to their linewidths. The authors demonstrate all seven on/off combinations of the three heater pulses and show that each bolometer's response tracks only its own heater state, with leakage more than an order of magnitude below the signal.","pith_inferences":["If the same isolation can be maintained as channel counts grow, the scheme could replace per-qubit cabling with two shared lines per module, but the worst-case -11.8 dB isolation would need to improve before moving to much larger arrays, since leakage accumulates with the number of neighbours.","The distorted filter lineshapes suggest that lumped-element or other filter designs could raise isolation and tighten the frequency plan, potentially increasing the number of multiplexed bolometers beyond the current 100-MHz-spaced channels.","Because the probe circuit is in the lumped-element regime (wavelength > 1 m), the same multiplexing approach may transfer to other thermal detectors such as NIS or graphene bolometers, provided their input filters can be engineered at the desired heater frequencies."],"forward_implications":["A single frequency-multiplexed probe line can read out multiple bolometers simultaneously, cutting the number of readout lines per channel.","SNS bolometer arrays can be scaled to larger numbers by assigning each sensor a distinct heater and probe frequency; the authors estimate up to 180 bolometers in the 100 MHz–1 GHz probe range.","The demonstrated 10 µs pulses and 4–13 µs time constants suggest faster calorimetric readout than earlier devices, supporting single-shot qubit readout extensions.","The crosstalk level of -11.8 dB or better defines a safe operating window of roughly 12 dB, within which simultaneous excitation does not false-trigger non-addressed bolometers."],"supporting_citations":[{"why":"Motivates the one-sensor-per-qubit frequency-multiplexing scheme and provides the phenomenological power-dependence model used to extract crosstalk.","marker":"[2]"},{"why":"Provides the original nanobolometer design, including the AuPd nanowire absorber and impedance matching to 50 Ω.","marker":"[18]"},{"why":"Supplies the electrothermal-feedback measurement technique and the characterization of the detector nonlinear response.","marker":"[22]"},{"why":"Basis for the half-wavelength coplanar-waveguide band-pass filters used to isolate the heater channels.","marker":"[29]"},{"why":"Demonstrates probe frequencies above 950 MHz for similar nanowire devices, used to estimate the scaling limit of 180 multiplexed bolometers.","marker":"[35]"}],"fun_headline_variants":["Three SNS bolometers share one readout line on a chip","Frequency multiplexing cuts bolometer readout lines","SNS bolometers multiplexed on a single chip","Low-crosstalk multiplexed bolometer readout demonstrated","Three bolometers, one chip, one readout line"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim rests on the assumption that roughly 12 dB of worst-case heater isolation between channels is enough to prevent false triggering, despite filter lineshapes that deviate from the ideal Lorentzian response.","fun_headline_variants_meta":{"raw":{"variants":["Three SNS bolometers share one readout line on a chip","Frequency multiplexing cuts bolometer readout lines","SNS bolometers multiplexed on a single chip","Low-crosstalk multiplexed bolometer readout demonstrated","Three bolometers, one chip, one readout line"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00052,"raw_usage":{"total_tokens":2494,"prompt_tokens":899,"completion_tokens":1595,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":515,"completion_tokens_details":{"reasoning_tokens":1515}},"tokens_in":515,"tokens_out":1595,"duration_ms":12985,"temperature":1.0,"reasoning_tokens":1515,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:36:28.032421+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test would be to operate a larger array with the same filter design and the same per-channel heater power: if the response of a non-addressed bolometer to a heater pulse at the resonance of an adjacent filter rises above its noise floor when two or more neighbouring heaters are active, the isolation margin is insufficient. Concretely, for the current device, a heater pulse at f1 = 7.6 GHz with power near the -106 dBm 1-dB compression point of B1 should not produce a detectable response in B2 or B3; measuring a response above the pre-pulse noise level would contradict the claimed crosstalk range.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Motivates the one-sensor-per-qubit frequency-multiplexing scheme and provides the phenomenological power-dependence model used to extract crosstalk."},{"cited_title":"Kokkoniemi, J","cited_arxiv_id":null,"evidence_quote":"Provides the original nanobolometer design, including the AuPd nanowire absorber and impedance matching to 50 Ω."},{"cited_title":"Govenius, R","cited_arxiv_id":null,"evidence_quote":"Supplies the electrothermal-feedback measurement technique and the characterization of the detector nonlinear response."},{"cited_title":"G¨ oppl, A","cited_arxiv_id":null,"evidence_quote":"Basis for the half-wavelength coplanar-waveguide band-pass filters used to isolate the heater channels."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates probe frequencies above 950 MHz for similar nanowire devices, used to estimate the scaling limit of 180 multiplexed bolometers."}],"review_version":1}