{"id":"1549447a-d1f7-4177-8cb8-9f366aab64e2","arxiv_id":"2506.22620","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":7,"one_line_summary":"A scanning high-Q superconducting resonator can couple coherently to transmon qubits and read out their spectra and coherence times without fabricating on-chip readout circuitry.","lead":"The authors built a movable superconducting microwave resonator on a cantilever tip and used it to image capacitance and to measure the quantum states of transmon qubits with no readout circuitry on the chip. If it works as reported, the tool lets researchers probe qubits and candidate quantum materials by scanning a high-quality resonator over them at millikelvin temperatures.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'no tip-induced loss' conclusion rests on an unvalidated simulated coupling g; a wrong g would misattribute tip-induced loss to intrinsic loss, weakening the claim that the scanning resonator is minimally invasive.","rationale":"The paper is a well-executed experimental demonstration of a new scanning resonator tool. The central claim—that a scanning high-Q resonator can be used for dispersive readout, two-tone spectroscopy, and coherence measurements of transmon qubits without on-chip readout circuitry—is supported by direct measurements: the resonator-transmon dispersive shift, the two-tone spectrum with g-e and two-photon g-f transitions, and the T1/T2R/T2E decays. These do not depend on the electromagnetic simulations. The load-bearing weak point is the stronger quantitative conclusion that the tip adds no loss beyond Purcell loss. That conclusion is used to argue the probe is non-invasive and suitable for characterizing quantum devices, and it depends on the simulated coupling g with no independent validation. Because Purcell loss grows as g², even a modest error in g changes the predicted loss substantially, and the absence of error bars on Fig. 4 hides the sensitivity of the fit. This does not overturn the main demonstration, but it does mean the 'minimally invasive' claim is not yet fully established, justifying the reader's conditional verdict. I partially agree with the reader's weakest-assumption identification: both the zF sensitivity and the Purcell analysis rely on simulations, but the Purcell/g dependence is the more central risk for the device claim. The data-availability placeholder and missing error bars on coherence times are additional minor issues, but they are not the most load-bearing concern.","tokens_in":15131,"tokens_out":8891,"duration_ms":102682,"concrete_test":"Extract g from the already-measured dispersive shift in Fig. 3(b). Fit the low-power and high-power resonator responses to obtain the dressed-to-bare shift 2χ at the same 2.5 μm tip position, then invert the transmon dispersive formula χ = g²α/[Δ(Δ+α)] (with α = -E_C, E_C from Table SI) for g. Compare this g with the Ansys/scQubits g/2π used in Supplement Sec. VII at 2.5 μm. Recompute Γ_Purcell = (g/Δ)²κ with the measured g and replot Fig. 4 with error bars. If the distance-dependent Γ1 is still fully accounted for by the recalibrated Purcell term plus a constant, the no-extra-loss claim survives; if a distance-dependent residual remains, that claim must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central demonstration—dispersive readout, two-tone spectroscopy, and T1/T2E measurements—is credible as a first demonstration. The load-bearing quantitative claim is in Sec. IV: 'we do not find additional tip-sample capacitance-dependent contributions to the transmon loss beyond the predicted Purcell loss.' That statement is what makes the probe minimally invasive and is used to argue the tool is suitable for probing quantum coherent devices. It rests entirely on the simulated qubit-resonator coupling g/2π from Ansys Maxwell and scQubits (Supplement Sec. VII), inserted into Γ_Purcell = (g/Δ)²κ for the fit in Fig. 4. No independent measurement of g is reported. Because Γ_Purcell scales as g², a factor-of-2 error in g changes the predicted Purcell rate by 4×. If the true g is larger than simulated, the increasing Γ1 in Fig. 4 would be partly tip-induced loss misattributed to Purcell, and the 'no additional loss' conclusion fails; if g is smaller, the model would overstate Purcell and the fit would hide a distance-dependent discrepancy. The same simulation also fixes the capacitive sensitivity calibration (Supplement Sec. V), but that is less central to the headline device claim. The lack of error bars on the Γ1 points in Fig. 4 makes the agreement harder to assess.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a scanning superconducting microwave resonator—a lumped-element NbTiN resonator on a cantilever—as a probe for quantum coherent devices. At 10 mK the authors measure an internal quality factor Qi>10^4 in the single-photon regime, demonstrate capacitive imaging with sensitivity of order 3 zF/Hz^{1/2} and about 2 μm spatial resolution, and then couple the tip to transmon qubits. Using power-dependent dispersive shifts, two-tone spectroscopy, T1, Ramsey, and Hahn-echo measurements, they extract qubit spectra and coherence times for multiple transmons without any on-chip readout circuitry. They also report a tip-sample-distance-dependent relaxation rate that they model as Purcell loss plus a distance-independent term, concluding that the tip introduces no additional loss. Finally, they map fge, T1, and T2E for an array of 17 transmons.","tokens_in":15449,"tokens_out":9931,"duration_ms":111674,"significance":"If correct, the results establish a new scanning-probe tool for circuit QED: a high-Q resonator that can be positioned over arbitrary devices and used for dispersive readout, spectroscopy, and coherence measurements without fabricating readout resonators on the sample chip. The central demonstration—coherent coupling to transmons and multi-qubit characterization with no on-chip readout—is supported by direct measurements. Strengths include the direct two-tone and time-domain data, the clear presentation of the setup, and the data-availability statement. The main caveat is that the quantitative 'no additional loss' conclusion depends on an unvalidated simulated coupling strength, and several quantitative results lack error bars.","major_comments":[{"comment":"The conclusion that the tip adds no loss beyond Purcell rests entirely on the simulated qubit-resonator coupling g(d) from the AC Conduction solver and scQubits, inserted into Γ_Purcell = (g/Δ)^2 κ. No independent measurement of g is presented, and the tip-sample distance axis used in Fig. 4 is itself calibrated from the same simulation (Supplement Sec. V). Because the Purcell rate scales as g^2, a factor-of-2 simulation error changes the predicted rate by 4×, and a systematic error in the simulated distance dependence could be partly absorbed into the fit, misattributing distance-dependent tip-induced loss to Purcell or vice versa. To support the 'no additional loss' claim, the authors should either measure g independently (e.g., through an avoided crossing or ac-Stark shift) or provide a sensitivity analysis of the fit to g together with fit residuals. At minimum, the conclusion should be softened to state consistency with Purcell loss within the accuracy of the simulation.","section":"Section IV; Supplement Sec. VII, Eq. (11)"},{"comment":"The coherence times and relaxation rates are reported without error bars or confidence intervals, so the agreement between the Γ1 data and the Purcell model in Fig. 4, and the significance of the T2E increase in Fig. 3(f), cannot be quantitatively assessed. The authors should provide statistical uncertainties (from repeated measurements or fit covariance) for T1, T2R, T2E, and Γ1, and include residuals for the Fig. 4 model.","section":"Figures 3(d-f) and 4"}],"minor_comments":[{"comment":"The abstract and text state 'zeptoFarad sensitivity' but the reported quantity is 3 zF/Hz^{1/2} at a 1 Hz bandwidth; please state the per-root-Hz units consistently in the abstract.","section":"Section III and abstract"},{"comment":"Because the capacitance-to-frequency conversion α and the tip-sample distance are calibrated using simulated Cts(d) with two fitted parameters (d0 and α), the absolute value of the quoted zF sensitivity is simulation-dependent; the main text should state this caveat explicitly and, if possible, provide an independent calibration.","section":"Section III and Supplement Sec. V"},{"comment":"The sentence 'the qubit coherence is affect by pure dephasing processes' contains a typo ('affect' should be 'affected').","section":"Section IV"},{"comment":"The EC and EJ values are extracted using the measured fge together with simulations; the table should include a note that these energies are model-dependent and do not include uncertainties.","section":"Supplement Table SI"},{"comment":"The calibration of the π-pulse amplitude used in the T1 measurement is not described; a brief statement on pulse calibration would aid reproducibility.","section":"Section IV"}],"recommendation":"major_revision","confidential_remarks":"To the editor: This is a solid experimental demonstration with a clear novelty in combining a scanning high-Q resonator with single-photon-level circuit QED readout. The main concern is the load-bearing nature of the simulated coupling g for the 'no additional loss' claim; requesting an independent g measurement or a clear caveat should be part of the revision. The paper is well within the scope of a specialized quantum-engineering journal; I see no citation or novelty-disclosure issues. The manuscript would also benefit from a more careful treatment of uncertainties throughout."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The central demonstration is real and worth knowing about: the authors mount a high-Q superconducting resonator on a cantilever, scan it over transmon qubits, and use it as both a dispersive readout port and a drive port, extracting the transmon spectrum, T1, Ramsey and echo times, and mapping 17 qubits with no on-chip readout circuitry. Prior scanning-resonator work (refs 30, 31) stayed in the dielectric-imaging regime, so coherent coupling to a qubit is the genuinely new step. The raw measurements that support it—power-dependent dispersive shift, two-tone spectroscopy showing the one- and two-photon transitions, and the time-domain decays—look credible. Qi > 10^4 at single-photon powers and zF/Hz^1/2 capacitive sensitivity at 500 nm are solid benchmarks.\n\nThe soft spot is the quantitative loss claim. In Sec. IV the authors conclude the tip adds no loss beyond Purcell, but that conclusion rests entirely on the simulated qubit-resonator coupling g from Ansys Maxwell/scQubits, with no independent measurement. Since the Purcell rate scales as g^2, a factor-of-two error in g changes it by 4x. If the true g is larger, the rising Gamma1 with decreasing distance could be partly tip-induced loss misattributed to Purcell; if smaller, the model overstates Purcell and the fit would mask a distance-dependent discrepancy. The 'minimally invasive' claim is therefore not as firm as the text implies. The lack of error bars on the Gamma1 points in Fig. 4 makes the agreement hard to assess. This is not fatal—the coherent-coupling demonstration stands independently—but the authors should either validate g or soften the claim.\n\nMinor issues: the coherence times are quoted without error bars, and the data availability statement is still a placeholder. Both are easy fixes.\n\nWho this is for: experimentalists in circuit QED, scanning microwave microscopy, and quantum device characterization. It deserves a serious referee—send it out. I would accept after revision.","headline":"First credible demonstration of coherent transmon readout with a scanning resonator; the 'no extra tip loss' claim is softer than the text admits because it rests on an unvalidated simulated coupling.","tokens_in":15950,"tokens_out":2971,"would_cite":true,"duration_ms":29469,"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 high-quality superconducting resonator mounted on a scanning cantilever can be positioned over a transmon qubit and serve as its dispersive readout and control port, allowing the qubit's energy spectrum and coherence times to be…","keywords":["scanning resonator","circuit quantum electrodynamics","transmon qubit","dispersive readout","capacitive imaging","superconducting resonator","qubit coherence","scanning probe microscopy"],"falsifier":"Measure the qubit-resonator coupling directly from an avoided-level crossing (vacuum Rabi splitting) at several tip-sample distances, and compare the Purcell-limited relaxation rate predicted by $(g/\\Delta)^2\\kappa$ with the measured $T_1$; agreement within uncertainty would confirm the calibration, while a systematic offset would show that the simulated coupling or capacitance is wrong.","tokens_in":14931,"feed_emoji":"🔬","tokens_out":4531,"duration_ms":47162,"temperature":0.7,"pith_summary":"The paper aims to establish that a high-quality superconducting resonator on a scanning cantilever can act as a movable, reconfigurable readout port for quantum coherent devices. The authors show that by bringing a small tip terminated in a lumped-element resonator close to a transmon qubit, the resonator couples through the tip-sample capacitance and performs dispersive readout and qubit control without any readout circuitry fabricated on the sample chip. They report a resonator internal quality factor above 10,000 at single-photon powers, capacitive imaging with zeptoFarad sensitivity and micron spatial resolution at milliKelvin temperatures, and full characterization of the energy spectrum and coherence times of a 17-qubit array. If the approach holds, it would give experimentalists a way to probe materials and qubit platforms that cannot host high-quality on-chip resonators.","feed_headline":"A scanning resonator reads qubits with no on-chip wiring","feed_subtitle":"High-Q superconducting tip maps 17 qubit spectra and coherence with zeptoFarad sensitivity at millikelvin temperatures.","key_machinery":"The central object is a lumped-element superconducting resonator, a 7.955 GHz NbTiN device on a SiN/Si cantilever, terminated in a 2-micron tip. The tip-sample capacitance $C_{ts}$ is the coupling port: changes in $C_{ts}$ shift the resonator frequency (the basis of capacitive imaging), and when the tip is over a transmon the same capacitance produces a qubit-resonator coupling $g$ that enables dispersive readout. The paper calibrates tip-sample distance by fitting measured frequency shifts to simulated $C_{ts}$ curves, and computes the induced qubit relaxation from the Purcell formula $\\Gamma_{\\text{Purcell}} = (g/\\Delta)^2\\kappa$, where $\\Delta$ is the qubit-resonator detuning and $\\kappa$ the resonator linewidth, to separate tip-induced loss from intrinsic qubit loss.","core_discovery":"The central claim is that a single high-Q superconducting resonator on a scanning tip can be coherently coupled to a transmon qubit through the tip-sample capacitance and serve as both the readout and control port, so that the qubit's energy spectrum, $T_1$, Ramsey dephasing, and echo coherence can be measured with no on-chip readout circuitry. The authors demonstrate this by positioning the tip over individual transmons, observing the expected dispersive, power-dependent resonator response, resolving the $|g\\rangle\\to|e\\rangle$ transition and a two-photon $|g\\rangle\\to|f\\rangle$ transition, and measuring coherence times. The measured relaxation rate as the tip approaches the qubit matches the Purcell rate predicted from the simulated coupling, which the authors take as evidence that the scanning tip introduces no loss beyond the Purcell channel.","pith_inferences":["The same readout mechanism should extend to other quantum coherent systems whose energy scales overlap the resonator frequency, such as spin ensembles or magnons, where on-chip high-Q resonators are hard to fabricate.","The sensitivity floor is set by mechanical vibration; a vibration-isolated or floated scanning stage could push the capacitance noise below the reported zF/Hz$^{1/2}$ and allow smaller tip-sample distances.","The demonstration suggests a path toward spatially resolved noise spectroscopy: by rastering the tip, local two-level-system defects and quasiparticle traps could be located and correlated with qubit decoherence.","A direct test of the method's generality would be to compare extracted coherence times for the same qubit measured both with the scanning resonator and with a conventional on-chip readout."],"forward_implications":["A single scanning resonator can characterize many qubits on one chip without fabricating readout resonators, which speeds up testing of qubit arrays and of chips where readout circuitry is impractical.","Materials that cannot host high-quality on-chip resonators, such as those with large microwave loss, can still be probed at the single-photon level by bringing the resonator tip near them.","Because the tip is positionable, the same resonator can map the spatial variation of qubit properties and local loss sources across a device.","Adding a Purcell filter to the resonator tip should suppress the distance-dependent relaxation channel and lengthen qubit $T_1$.","The demonstrated zF/Hz$^{1/2}$ capacitive sensitivity at single-photon powers extends scanning microwave microscopy into a regime suitable for probing quantum coherent systems."],"supporting_citations":[{"why":"Establishes circuit QED as the framework for resonator-mediated qubit readout and control.","marker":"[1]"},{"why":"Supplies the general quantum-sensing context for using high-Q resonators to detect weak signals.","marker":"[2]"},{"why":"Demonstrates a high-Q scanning resonator operating at low photon number for capacitive imaging, the direct predecessor of this setup.","marker":"[30]"},{"why":"Shows that SiN/Si substrates support resonators with internal quality factors above 10^4 at low power, justifying the materials stack.","marker":"[33]"},{"why":"Provides a recent example of power- and temperature-dependent TLS loss in superconducting resonators on similar substrates.","marker":"[35]"},{"why":"Introduces the transmon qubit geometry used as the target device.","marker":"[38]"},{"why":"Derives the dispersive state-dependent frequency shift used for qubit readout.","marker":"[39]"},{"why":"Establishes two-tone spectroscopy through a resonator-qubit coupling, the method used to measure the transmon spectrum.","marker":"[42]"},{"why":"Supplies the Purcell formula used to compute the expected relaxation rate from qubit-resonator coupling.","marker":"[43]"},{"why":"Shows how Purcell filters suppress resonator-induced qubit relaxation, motivating a proposed improvement.","marker":"[44]"}],"fun_headline_variants":["Scanning tip reads qubits without on-chip readout","Mobile resonator probes qubit spectra and coherence","No on-chip wiring: scanning resonator measures qubits","Tunable scanning tip couples to qubits via capacitance","High-Q tip sensor maps qubit coherence at millikelvin"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported zeptoFarad sensitivity and the conclusion that the tip adds no loss beyond Purcell loss depend on simulated values of tip-sample capacitance and qubit-resonator coupling; if those simulations are wrong in absolute scale, both quantitative claims would need revision.","fun_headline_variants_meta":{"raw":{"variants":["Scanning tip reads qubits without on-chip readout","Mobile resonator probes qubit spectra and coherence","No on-chip wiring: scanning resonator measures qubits","Tunable scanning tip couples to qubits via capacitance","High-Q tip sensor maps qubit coherence at millikelvin"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000306,"raw_usage":{"total_tokens":1723,"prompt_tokens":882,"completion_tokens":841,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":498,"completion_tokens_details":{"reasoning_tokens":763}},"tokens_in":498,"tokens_out":841,"duration_ms":8886,"temperature":1.0,"reasoning_tokens":763,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:01:44.127922+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the qubit-resonator coupling directly from an avoided-level crossing (vacuum Rabi splitting) at several tip-sample distances, and compare the Purcell-limited relaxation rate predicted by $(g/\\Delta)^2\\kappa$ with the measured $T_1$; agreement within uncertainty would confirm the calibration, while a systematic offset would show that the simulated coupling or capacitance is wrong.","supporting_citations":[{"cited_title":"Blais, A","cited_arxiv_id":null,"evidence_quote":"Establishes circuit QED as the framework for resonator-mediated qubit readout and control."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the general quantum-sensing context for using high-Q resonators to detect weak signals."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that SiN/Si substrates support resonators with internal quality factors above 10^4 at low power, justifying the materials stack."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides a recent example of power- and temperature-dependent TLS loss in superconducting resonators on similar substrates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the transmon qubit geometry used as the target device."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives the dispersive state-dependent frequency shift used for qubit readout."},{"cited_title":"Boissonneault, J","cited_arxiv_id":null,"evidence_quote":"Establishes two-tone spectroscopy through a resonator-qubit coupling, the method used to measure the transmon spectrum."},{"cited_title":"Blais, J","cited_arxiv_id":null,"evidence_quote":"Supplies the Purcell formula used to compute the expected relaxation rate from qubit-resonator coupling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how Purcell filters suppress resonator-induced qubit relaxation, motivating a proposed improvement."}],"review_version":1}