{"id":"cb693f7d-0af0-4237-ab99-e21d117fd604","arxiv_id":"2412.16045","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A neon-FIB-defined Nb nanoSQUID-embedded resonator achieves flux tuning and high Q at 16 mK, with phase noise dominated by dielectric two-level systems.","lead":"Researchers built a niobium microwave resonator with an embedded SQUID tuned by magnetic flux and showed it stays tuneable and low-loss at 16 millikelvin, colder than prior niobium devices. The device is aimed at quantum memories that store information in the spins of silicon atoms.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Noise attribution is underdetermined: without the phase-to-flux transfer coefficient, flux-independent phase noise does not establish that SQUID noise is below resonator TLS noise.","rationale":"I read the paper in good faith. The core demonstration, a monolithic Nb nanoSQUID-embedded resonator tuned by 300 kHz at 16 mK with Qi ~ 1.4e5 and no Q penalty versus the control, is directly measured and internally consistent. The reader's CONDITIONAL verdict is appropriate. The load-bearing gap is the quantitative relation between the observed phase noise and an equivalent flux noise. The flux-independence of S_theta is suggestive but not sufficient, because S_theta = (d_theta/d_Phi)^2 * S_Phi and d_theta/d_Phi is small here. The authors themselves flag the missing transfer coefficient, which is exactly the point at which the applicability claim rests. My concrete test would use the existing tuning-curve fit to set an upper bound on S_Phi, thereby settling whether the 'dominated by TLS' conclusion is quantitatively justified. I do not see a fatal flaw; the device demonstration stands. Field resilience at the 100 mT scale relevant to Si:Bi clock transitions is not demonstrated, but the strongest claim as quoted does not hinge on it, and the reader's concern is the more direct logical gap.","tokens_in":10161,"tokens_out":8399,"duration_ms":77772,"concrete_test":"From the Fig. 2(c) fit, compute the maximum phase-to-flux coefficient d_theta/d_Phi = (2Qi/omega0)(d_omega/d_Phi) at the most flux-sensitive bias (near Phi_ext/Phi0 ~ 0.25-0.3, where df0/dPhi is largest). Using the Fig. 3(b) phase-noise floor S_theta at that bias, form the upper bound S_Phi_upper = S_theta / (d_theta/d_Phi)^2. If S_Phi_upper is below 0.5 micro-Phi0/Hz^0.5, the TLS attribution is supported; if S_Phi_upper exceeds it, the device's flux sensitivity is too low to certify that SQUID noise is negligible. This is a one-line calculation from already-reported data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's strongest claim includes 'intrinsic phase noise comparable to a bare CPW resonator' and that the measured noise is dominated by dielectric TLS. The zero-field phase noise comparison is direct, but the TLS attribution relies on the flux independence of S_theta (Fig. 4) together with the statement that 'intrinsic SQUID noise is significantly lower than the measured resonator noise.' The relation is S_theta = (d_theta/d_Phi)^2 * S_Phi. The authors explicitly note that SQUID flux noise and resonator phase noise can be related through an experimentally obtained transfer coefficient, but they do not measure it. Because the frequency tuning range is only 300 kHz (Fig. 2c), d_theta/d_Phi is modest; a constant SQUID flux noise would produce a flux-independent contribution that could be hidden below the TLS floor. The paper's own forward estimate, that 10-20 MHz tuneability is needed to reach 0.5 micro-Phi0/Hz^0.5, indicates that at the current 300 kHz tuning the same assumed S_Phi would not be negligible. Without computing S_Phi_upper from the measured S_theta and the known tuning curve, the conclusion that SQUID noise is not the limiting noise source is not quantitatively established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a monolithic Nb CPW resonator terminated by an rf nanoSQUID whose weak link is fabricated with a neon focused-ion-beam, and it characterizes the device at 16 mK. It presents the zero-field resonance with Qi = 1.41e5, internal-Q versus photon-number behavior comparable to a same-chip control resonator, flux tuning over a 300 kHz range with an extracted screening parameter beta_L = 1.51, homodyne phase-noise spectra, a P^-0.5 power dependence for both devices, and phase noise that is independent of applied flux. The authors conclude that the device is flux-tuneable at millikelvin temperatures without a quality-factor penalty and that its noise is dominated by dielectric TLS rather than by the SQUID, and they discuss design improvements toward single-spin ESR sensitivity.","tokens_in":10343,"tokens_out":7326,"duration_ms":70143,"significance":"If fully supported, this is a useful experimental advance for hybrid superconductor-spin systems: it demonstrates a monolithic Nb SQUID-embedded resonator operating at 16 mK with a quality factor and phase noise comparable to a control resonator on the same chip. The same-chip comparison, the direct Q versus photon-number data, and the explicit modeling of the flux-tuning curve are clear strengths. The principal shortcoming is that the phase-to-flux transfer coefficient needed to convert the measured phase noise into a calibrated flux sensitivity is not measured, so the quantitative conclusion that SQUID noise is negligible is not established by the presented data.","major_comments":[{"comment":"The statement that 'the intrinsic SQUID noise is significantly lower than the measured resonator noise' is not quantitatively established. The relation S_theta = (d_theta/d_Phi)^2 S_Phi requires an experimentally determined transfer coefficient, which the manuscript notes but does not provide. The observed flux independence of S_theta, combined with the 300 kHz tuning range of Fig. 2(c), places only a very weak upper bound on S_Phi: the paper's own forward estimate that 10–20 MHz tuning is needed to reach 0.5 uPhi0/Hz^0.5 implies that at 300 kHz an S_Phi orders of magnitude larger could be masked by the dielectric noise floor. Please measure d_theta/d_Phi (for example, by applying a known flux modulation and detecting the phase response, or by using the slope of the tuning curve together with the resonator transfer function) and report the resulting S_Phi upper bound, or restrict the conclusion to flux-independent phase noise.","section":"Noise characterisation (Fig. 4 and p. 5)"},{"comment":"The abstract states that the authors 'characterise the flux sensitivity ... and find that the noise is dominated by dielectric noise', but the measurements reported are phase noise versus drive power and versus applied flux, without a calibration to flux units. Without the transfer coefficient, the device's flux sensitivity in units of Phi0/Hz^0.5 is not actually characterized. Please either supply the calibration or rephrase these claims so that they describe uncalibrated phase-noise measurements and a qualitative noise-source attribution.","section":"Abstract and p. 5"},{"comment":"The attribution of the noise to dielectric TLS rests partly on the f^-0.5 spectral region of Fig. 3(b), but the text itself states that this functional form is likely an artifact of insufficient sampling time. This does not invalidate the comparison between the two devices made at the same offset frequency, nor the P^-0.5 power dependence measured at 1 kHz, but the paper should make clear that the TLS conclusion is based on the power dependence and the control-resonator comparison rather than on the spectral exponent, which is not physically meaningful in the measured band.","section":"Noise characterisation (Fig. 3 and p. 4)"}],"minor_comments":[{"comment":"In Eq. (1), it should be stated explicitly whether Pin is the nominal room-temperature power or the power at the sample after accounting for the -60 dB attenuation, since the inferred photon number depends on this choice.","section":"Methods, Eq. (1)"},{"comment":"The fit to the flux-tuning data is shown as a solid line, but the model parameters L1 and L2 that enter Eq. (4) are not given, and no uncertainty is reported for the extracted beta_L = 1.51 and I0 = 320 uA; please provide these values or a reference for the inductance partition.","section":"Fig. 2(c) and accompanying text"},{"comment":"The text describes an f^-1 dependence below 100 Hz, but no fit line is shown in the figure; adding guide lines for the f^-1 and f^-0.5 regions would make the spectral description easier to verify.","section":"Fig. 3(b)"},{"comment":"The statement that flux sensitivity is characterized 'as a function of microwave drive power and externally applied magnetic field' overstates the parameter coverage: Fig. 3(c) is at zero flux and Fig. 4 is at a single power; this should be clarified.","section":"Abstract and Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The experimental demonstration is solid and likely of interest to the journal's readership. The missing phase-to-flux calibration is the load-bearing gap: it directly affects the validity of the noise-attribution claim in the abstract and conclusion. If the authors can add a direct transfer-coefficient measurement or a conservative upper bound on S_Phi, I would be willing to support acceptance; without it, the central conclusion about SQUID noise being negligible is stronger than the data support."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: the central device result holds up. A monolithic Nb nanoSQUID-embedded resonator, made with a Ne FIB 3D nanobridge, tunes with flux at 16 mK, with Q on par with a bare CPW on the same chip (1.4e5 at -80 dBm, 7e3 at single photon). That is a real step past the 300 mK Nb constriction work, and the fabrication story is clean: the 3D bridge profile comes naturally from beam spread rather than multi-layer deposition, with honest discussion of FIB damage and ~50% yield. The model for flux tuning with beta_L=1.51 fits the data, and parameter extraction from tuning is standard. Credit where due: this is useful and well-executed incremental device physics.\n\nThe soft spot is exactly where the reader and stress-test point. The paper claims the measured phase noise is dominated by dielectric TLS and that intrinsic SQUID noise is significantly lower. The evidence is that S_theta is independent of flux. But S_theta = (dtheta/dPhi)^2 S_Phi, and dtheta/dPhi is never measured—the 300 kHz tuning range is the only handle, and it is small. A constant SQUID flux noise would also give flux-independent phase noise, hidden under the TLS floor. The authors even state the transfer coefficient is needed and then don't provide it. Their own forward estimate—10–20 MHz tuneability needed to resolve 0.5 uPhi0/Hz^0.5—says plainly that at the present 300 kHz, the assumed SQUID noise would not be negligible. So the noise-attribution conclusion is not quantitatively established. This is a missing measurement, not a fatal flaw. Also in proportion: the f^-0.5 noise spectra are acknowledged as sampling-limited, single-shot, no error bars; that weakens the quantitative spectral shape but not the zero-field comparison to the control.\n\nWho this is for: experimentalists working on hybrid superconductor–spin memories, especially Si:Bi clock transitions. They will value the 16 mK operation and the control comparison. The paper deserves a serious referee. I would send it out with a request that the authors either measure dtheta/dPhi and report flux-equivalent noise, or soften the SQUID-noise conclusion to a statement about what they can bound. With that one addition it becomes a solid, citable device paper.","headline":"Solid incremental Nb SQUID-resonator demo at 16 mK, but the noise-attribution claim needs the transfer coefficient the authors didn't measure.","tokens_in":10987,"tokens_out":2282,"would_cite":true,"duration_ms":20847,"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 niobium SQUID resonator is flux-tuneable at 16 mK with no loss in quality factor.","keywords":["nanoSQUID","niobium resonator","neon focused ion beam","millikelvin operation","flux-tuneable resonator","Josephson nanobridge","phase noise","hybrid quantum memory"],"falsifier":"A direct test would be to calibrate the transfer coefficient by applying a known small alternating magnetic flux to the SQUID at its most sensitive bias point and measuring the induced phase modulation; using that calibration to convert the measured phase noise into an equivalent flux noise would either confirm or refute the claim that the SQUID contributes negligibly to the measured noise.","tokens_in":9917,"feed_emoji":"❄️","tokens_out":8405,"duration_ms":64797,"temperature":0.7,"pith_summary":"The paper reports a niobium microwave resonator whose resonance frequency can be tuned with a magnetic field while operating at 16 millikelvin, a temperature low enough to be useful for coupling to solid-state spin quantum memories. The central claim is that a neon focused-ion-beam-made nanobridge acts as a Josephson weak link without degrading the resonator's quality factor or adding measurable phase noise. The authors show that the resonator's noise is the same dielectric two-level-system noise seen in a bare resonator, implying the SQUID itself is not the noise bottleneck. If correct, this removes a materials obstacle to hybrid superconductor-spin circuits, since niobium — unlike aluminium — is magnetically resilient and thermally compatible with silicon spin qubits.","feed_headline":"Niobium resonator tunes at 16 mK, keeps its quality factor","feed_subtitle":"Flux-tuneable SQUID resonator shows no added phase noise, clearing a path to silicon spin quantum memories.","key_machinery":"The key object is an rf-SQUID embedded at the shorted end of a $\\lambda/4$ coplanar-waveguide resonator: a superconducting loop interrupted by a single nanobridge weak link, whose Josephson inductance makes the resonance frequency flux-tuneable. The nanobridge is cut with a neon focused ion beam; because the beam intensity falls off towards its edges, the bridge is thinner at the top than at the base, confining the nonlinearity to a short effective length below $3.5$ times the Ginzburg–Landau coherence length at 16 mK. The argument is carried by a transmission-line model in which the SQUID appears as a flux-dependent load impedance, combined with the flux-quantization relation $\\Phi_{\\mathrm{tot}}/\\Phi_0 = \\Phi_{\\mathrm{ext}}/\\Phi_0 - (\\beta_L/2\\pi)\\sin(2\\pi\\Phi_{\\mathrm{tot}}/\\Phi_0)$; fitting this model to the measured frequency tuning yields the SQUID parameters.","core_discovery":"The paper establishes that a monolithic niobium SQUID-embedded coplanar-waveguide resonator, with the SQUID weak link formed by neon focused-ion-beam milling, is flux-tuneable at $T = 16$ mK while retaining an internal quality factor and phase-noise spectrum essentially identical to a bare control resonator on the same chip. The tuneability is achieved through the SQUID's flux-dependent Josephson inductance, which shifts the resonance by up to 300 kHz before a discontinuity at $0.545\\,\\Phi_0$; fitting the tuning curve gives an inductive screening parameter $\\beta_L = 1.51$ and a nanobridge critical current of about $320\\ \\mu\\mathrm{A}$. The phase-noise measurements, which follow the TLS power dependence $S_\\theta \\propto P^{-0.5}$ and are independent of applied flux, support the conclusion that the noise is dominated by dielectric two-level systems in the resonator rather than by the SQUID. The authors argue this combination — low temperature, field resilience, high $Q$, and tuneability — satisfies the requirements for strong coupling to clock-transition spin ensembles such as bismuth donors in silicon.","pith_inferences":["If the flux-independence of the phase noise holds at higher magnetic fields near the clock transition, the SQUID's noise contribution could remain negligible under realistic operating conditions, but the paper does not report such a measurement.","The authors' proposed recipe for increasing tuneability—reducing resonator inductance and increasing the non-junction arm inductance—could be simulated in advance to predict the achievable tuning range and its effect on $\\beta_L$, offering a design target before fabrication.","A testable extension would be to measure the phase noise as a function of drive power at the flux-frustration point where the SQUID's flux-to-frequency transfer is maximal; the current data show no flux dependence even at the most sensitive bias, which indirectly constrains the SQUID's flux noise only if the transfer coefficient is independently calibrated."],"forward_implications":["The device meets the operating conditions ($T < 50$ mK, magnetic-field resilience, high $Q$, flux tuneability) needed for coupling a resonator to clock-transition spin ensembles such as bismuth donors in silicon.","Since the SQUID adds no measurable loss or phase noise, the same fabrication route can be used for other niobium quantum circuits without compromising coherence.","The measured TLS-dominated noise implies that reducing resonator dielectric noise should directly improve the sensitivity of the SQUID-embedded resonator.","With a modest increase in tuning range (to 10–20 MHz), the device should reach a flux sensitivity of about $0.5\\ \\mu\\Phi_0/\\mathrm{Hz}^{0.5}$, suitable for single-spin ESR detection.","The 3d nanobridge geometry produced by the finite beam spread is a single-step route to small effective-length weak links, potentially simplifying fabrication of Nb SQUID devices."],"supporting_citations":[{"why":"The prior Nb constriction nanoSQUID resonator, whose 300 mK operating floor this work lowers to 16 mK.","marker":"[37]"},{"why":"Monolithic three-dimensional nanobridge Nb microwave circuits, the approach extended here with a single neon-FIB step.","marker":"[26]"},{"why":"The weak-link criterion $l \\lesssim 3.5\\,\\xi_{GL}$ used to set the nanobridge dimensions.","marker":"[38]"},{"why":"The complex scattering-parameter fit used to extract resonant frequency and quality factor.","marker":"[16]"},{"why":"The homodyne phase-noise measurement method used to compare the SQUID and control resonators.","marker":"[51]"},{"why":"The interacting TLS model whose $P^{-0.5}$ power dependence matches the measured phase noise.","marker":"[57]"},{"why":"Defines the clock-transition frequency and field range that motivate the resonator's operating requirements.","marker":"[8]"}],"fun_headline_variants":["Niobium SQUID resonator tunes at 16 mK, no noise penalty","Flux-tuneable niobium resonator runs at 16 mK for spin coupling","Millikelvin niobium nanoSQUID resonator: tuneable and quiet","Neon FIB-built nanoSQUID resonator tunes at 16 mK","Tuneable niobium resonator at 16 mK keeps quality factor"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's noise conclusion hinges on the unmeasured phase-to-flux transfer coefficient; without calibrating it, the observed flux-independent phase noise does not by itself prove the SQUID is quieter than the resonator.","fun_headline_variants_meta":{"raw":{"variants":["Niobium SQUID resonator tunes at 16 mK, no noise penalty","Flux-tuneable niobium resonator runs at 16 mK for spin coupling","Millikelvin niobium nanoSQUID resonator: tuneable and quiet","Neon FIB-built nanoSQUID resonator tunes at 16 mK","Tuneable niobium resonator at 16 mK keeps quality factor"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00068,"raw_usage":{"total_tokens":3112,"prompt_tokens":988,"completion_tokens":2124,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":604,"completion_tokens_details":{"reasoning_tokens":2021}},"tokens_in":604,"tokens_out":2124,"duration_ms":14463,"temperature":1.0,"reasoning_tokens":2021,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T10:51:22.980731+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test would be to calibrate the transfer coefficient by applying a known small alternating magnetic flux to the SQUID at its most sensitive bias point and measuring the induced phase modulation; using that calibration to convert the measured phase noise into an equivalent flux noise would either confirm or refute the claim that the SQUID contributes negligibly to the measured noise.","supporting_citations":[{"cited_title":"Rodrigo , author M","cited_arxiv_id":null,"evidence_quote":"The prior Nb constriction nanoSQUID resonator, whose 300 mK operating floor this work lowers to 16 mK."},{"cited_title":"Polychroniou , author J","cited_arxiv_id":null,"evidence_quote":"The weak-link criterion $l \\lesssim 3.5\\,\\xi_{GL}$ used to set the nanobridge dimensions."},{"cited_title":"o nigl-Decrinis , author R. Shaikhaidarov , author S. E. \\ Kubatkin , author T. Lindstr \\","cited_arxiv_id":null,"evidence_quote":"The homodyne phase-noise measurement method used to compare the SQUID and control resonators."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The interacting TLS model whose $P^{-0.5}$ power dependence matches the measured phase noise."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the clock-transition frequency and field range that motivate the resonator's operating requirements."}],"review_version":1}