{"id":"97870057-b630-483c-87fc-a3c8b927de9a","arxiv_id":"2412.16099","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Tantalum resonators on silicon with a niobium seed layer reach internal quality factors up to 3.6 million at high power and kinetic inductance up to 0.6 pH per square, with thinner films giving more inductance but lower quality.","lead":"Researchers built superconducting microwave resonators from thin tantalum films on silicon and mapped how quality factor and kinetic inductance change with film thickness. The results offer a practical thickness tradeoff for compact, high-impedance quantum circuits made with CMOS-compatible fabrication.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim depends on unverified α-Ta phase attribution: no XRD/TEM or resistivity data confirm the Nb seed nucleates bcc Ta, so the quoted Qi and LK may describe β-Ta or a Ta/Nb bilayer, not α-Ta.","rationale":"The reader identified the same weakest assumption: the lack of structural verification that the Nb seed layer actually nucleates α-Ta. My stress test agrees and sharpens the technical consequence. The kinetic inductance estimate in Eq. (2) is derived from the normal-state sheet resistance and Tc, both of which are properties of the full Ta/Nb stack, not of α-Ta alone. Without phase identification and without de-embedding the Nb seed contribution, the quoted LK values and the 'α-Ta' label in the central claim are not independently supported. This does not overturn the paper's practical engineering message—thin Ta/Nb resonators on silicon with Qi > 10^6 and LK up to 0.6 pH/sq are useful—but it does mean the material-mechanism claim is conditional on structural characterization. Other issues noted by the reader (typos, missing error bars, no repository) are secondary and do not affect the core correctness risk as strongly. Therefore the appropriate verdict remains conditional, with the condition being the addition of phase and bilayer characterization.","tokens_in":14243,"tokens_out":6557,"duration_ms":60862,"concrete_test":"Prepare companion Ta/Nb films from the same sputter run at 40, 80, and 100 nm Ta thicknesses and perform grazing-incidence X-ray diffraction (2θ/ω) on each. If the α-Ta (110) reflection near 2θ ≈ 38.5° is observed with no β-Ta peaks (e.g., β-Ta (002) or (412)), the phase attribution is confirmed; if β-Ta or mixed phase is found, the central claim must be revised. As a secondary check, measure the room-temperature sheet resistance and compare with the known α-Ta (~15–60 μΩ·cm) versus β-Ta (~150–200 μΩ·cm) resistivity ranges, and de-embed the Nb seed contribution by measuring a Nb-only control film.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that Nb-seeded films on unheated silicon are α-Ta and that the measured high Qi and kinetic inductance are properties of that phase. The only evidence offered is the statement in the fabrication paragraph: 'Before Ta deposition, a 5 nm Nb seed layer was sputtered to promote the growth of the Ta α-phase.' No XRD, TEM, electron diffraction, or resistivity-ratio measurement is presented. This matters because α-Ta (bcc) and β-Ta (tetragonal) have very different normal-state resistivities and superconducting properties; β-Ta has substantially higher sheet resistance, which would inflate the LK value estimated from Eq. (2), LK ≈ ħRs/(πΔ0), even if the microwave loss remained low. In addition, Eq. (2) uses the total normal-state sheet resistance of the Ta/Nb bilayer, so the 5 nm Nb seed layer's own kinetic-inductance contribution is automatically folded into the quoted 0.6 pH/sq value for the 40 nm film. If the film is actually β-Ta or a mixed phase, the thickness-dependent Q-LK tradeoff is still a valid engineering result for the fabricated bilayer stack, but the material attribution to α-Ta and the mechanistic narrative in the abstract and conclusion are unsupported. This is the load-bearing premise because the novelty claim—'α-Ta on unheated Si with a Nb seed layer'—collapses if the phase is wrong, while the numerical resonator data alone do not establish the phase.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the fabrication and cryogenic microwave characterization of coplanar waveguide (CPW) resonators made from tantalum films of 40, 80, and 100 nm thickness deposited on unheated high-resistivity silicon with a 5 nm niobium seed layer. The authors claim that the Nb seed promotes the growth of the body-centred-cubic alpha-Ta phase, and they measure critical temperatures, sheet resistances, and internal quality factors as functions of microwave power and temperature. They report a maximum internal quality factor Qi of about 3.6e6 at high power for the 100 nm film and a kinetic inductance LK of 0.6 pH/sq for the 40 nm film, estimated from the BCS relation using the measured normal-state sheet resistance. The paper fits the power-dependent Qi data with a standard two-level-system (TLS) model and compares the results with earlier Ta resonator work.","tokens_in":14517,"tokens_out":3909,"duration_ms":34207,"significance":"If the results hold, the thickness-dependent Qi-LK tradeoff for Ta films on unheated silicon would be a useful engineering data point for compact, high-impedance circuit-QED devices, particularly because the process is CMOS-compatible and avoids substrate heating. The paper uses established measurement and fitting techniques, including notch-type S21 fitting and the TLS saturation model, and the reported Qi values at high power are competitive with several prior Ta resonator studies. However, the central material claim that the films are alpha-Ta is not directly verified, and the kinetic inductance values are extracted from a Ta/Nb bilayer, so the attribution of the measured performance specifically to alpha-Ta is not yet established.","major_comments":[{"comment":"The claim that the sputtered Ta films are alpha-Ta rests entirely on the sentence 'Before Ta deposition, a 5 nm Nb seed layer was sputtered to promote the growth of the Ta alpha-phase.' No XRD, TEM, electron diffraction, or resistivity-ratio measurement is shown to confirm the phase. This is load-bearing because the abstract and conclusion attribute the high Qi and high LK to alpha-Ta, and because Eq. (2) uses the normal-state sheet resistance Rs, which differs substantially between alpha-Ta (bcc) and beta-Ta (tetragonal). If the films are beta-Ta or a mixed phase, the quoted LK values and the mechanistic narrative change. Please add structural characterization of the films or explicitly reframe the results as properties of the Ta/Nb bilayer stack without the alpha-Ta attribution.","section":"Fabrication (page 3-4) and Conclusion"},{"comment":"The kinetic inductance is estimated as LK ≈ ħRs/(πΔ0) using the measured normal-state sheet resistance of the full film stack. Since the 5 nm Nb seed layer is metallic and superconducting, its own kinetic-inductance contribution is included in the quoted value of 0.6 pH/sq for the 40 nm film. The paper consistently refers to 'Ta samples' and 'Ta films,' but the measured Rs is that of the Ta/Nb bilayer. Please quantify the Nb seed contribution (e.g., by measuring a Nb-only control film or by estimating its sheet resistance separately), or revise the text so the LK values are not presented as intrinsic properties of the Ta layer alone.","section":"Eq. (2) and kinetic inductance values (page 5)"},{"comment":"Table 2 lists the seed layer as '6 nm Nb seed layer' for all three thicknesses, while the fabrication text states '5 nm Nb seed layer.' Additionally, the single-photon Qi range for the 100 nm film is given as '1.65-4.5 × 106', but Table 1 lists the three measured single-photon Qi values for 100 nm as 2 × 10^5, 1.65 × 10^5, and 4.5 × 10^5 — an order of magnitude lower. This internal inconsistency undermines the reliability of the comparison table and must be corrected.","section":"Table 2, 'This work' rows"},{"comment":"No fit uncertainties or measurement uncertainties are reported for the headline quantities: Qi, 1/Q_TLS^0, and LK. Figure 4 shows error bars, but the source of those errors (e.g., fit covariance, repeated measurements, systematic power calibration) is not described. Since the central quantitative claims are the maximum Qi of ~3.6e6 and LK of 0.6 pH/sq, confidence intervals or at least a description of how the errors were estimated are needed to assess whether the thickness-dependent trends are significant.","section":"Tables 1-2 and Figures 4-6"}],"minor_comments":[{"comment":"The abstract states a '5 nm Nb seed layer' while Table 2 lists '6 nm Nb seed layer' for the same devices; please harmonize the reported seed-layer thickness throughout.","section":"Abstract and Table 2"},{"comment":"Eq. (2) is attributed to reference [46], a paper on niobium nitride resonators; please verify the citation is appropriate for the BCS-based kinetic-inductance relation as used here.","section":"Page 5, Eq. (2) reference"},{"comment":"The caption states 'VNA power = 0 dBm' but the text refers to 'high power'; please clarify whether 0 dBm is the power at the VNA port or at the resonator after accounting for the 80 dB of total attenuation.","section":"Page 6, Figure 3 caption"},{"comment":"The TLS model in Eq. (4) uses n_ph, while the text and figures use <n_ph>; please use consistent notation for the average photon number.","section":"Eq. (4) and Figures 4-5"},{"comment":"The text labels both the quasi-particle loss in Eq. (6) and the Mattis-Bardeen expression in Eq. (7) as δ_qp; please state explicitly how these two expressions are related and which one is used for the fits shown in Fig. 6.","section":"Page 8, Eq. (6) and Eq. (7)"},{"comment":"The expression Δf_qp = -(1/2) α f_r (ΔL_k/L_k) uses α, which is defined earlier as the kinetic-inductance fraction; please ensure the same symbol is not confused with the fit parameter α in Eq. (1).","section":"Page 12, Eq. (9)"}],"recommendation":"major_revision","confidential_remarks":"The paper's strength is the direct engineering comparison across three Ta thicknesses on unheated silicon, which is of practical interest. The main risk is the unverified alpha-Ta phase attribution: if the authors cannot supply XRD, TEM, or resistivity-ratio evidence, the central novelty claim is substantially weakened and the paper should be reframed as a study of the Ta/Nb bilayer stack. The internal inconsistency in Table 2 (seed-layer thickness and the order-of-magnitude error in the 100 nm single-photon Qi range) should be fixed before resubmission. The manuscript is within the scope of the journal, but the quantitative claims need uncertainty reporting."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thickness scan is the worthwhile part: 40/80/100 nm Nb-seeded Ta on unheated high-resistivity Si, with Qi and LK for each thickness. The reported numbers — 0.6 pH/sq at 40 nm, Qi ~3.6e6 at high power for 100 nm — are plausible and come from standard notch fitting and the BCS relation. That gives engineers a useful tradeoff curve for compact, high-inductance circuits. The measurement and analysis are mostly standard, and the TLS/quasiparticle modeling is descriptive rather than overreaching.\n\nThe soft spots are real. The biggest is that the paper attributes the performance to alpha-Ta without a single piece of structural evidence. No XRD, no TEM, no resistivity ratio. The fabrication paragraph simply asserts that the 5 nm Nb seed promotes alpha-phase growth. This matters because the LK estimate uses the normal-state sheet resistance in Eq. (2), and beta-Ta has a much higher resistivity; the phase attribution changes how the data should be interpreted relative to other alpha-Ta work. The data set still stands as a characterization of this particular Ta/Nb bilayer stack, but the central claim of 'alpha-Ta on unheated Si' is unsupported as written. That is a load-bearing gap, not a cosmetic one.\n\nThere are also smaller issues: no error bars on any Qi or LK value, no raw data or repository, the Nb seed thickness is 5 nm in the text and 6 nm in Table 2, and Table 2 lists the 100 nm single-photon Qi as 1.65-4.5e6 while Table 1 gives 1.65-4.5e5. The last one is a clear typo, but it needs fixing.\n\nI disagree with any reading that calls the measurements themselves suspect. The notch fits look standard, the photon-number calibration is standard, and the temperature dependence behaves as expected. The paper is incremental engineering, not a new capability, but incremental data with a clean thickness scan is still citable.\n\nWho should read it: groups working on Ta resonator fabrication or high-kinetic-inductance circuits who want a quick thickness comparison. It deserves a serious referee, but the referee should ask for structural characterization (XRD or TEM) or a rewritten claim that does not depend on alpha-Ta. Without that, the abstract overstates what is actually demonstrated.\n\nMy recommendation: send it to review, conditional on major revision.","headline":"Useful thickness-dependent data for Ta-on-Si resonators, but the alpha-Ta claim is not backed by any structural characterization; send to review with a request for XRD/TEM or a softened claim.","tokens_in":15151,"tokens_out":4385,"would_cite":true,"duration_ms":35439,"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":"Alpha-tantalum resonators on silicon reach Qi above 3 million","keywords":["alpha-tantalum","kinetic inductance","coplanar waveguide resonator","internal quality factor","niobium seed layer","two-level system loss","superconducting quantum circuits","microwave resonators"],"falsifier":"An X-ray diffraction or transmission electron microscopy measurement of the 40 nm film that shows $\\beta$-Ta or a mixed phase rather than bcc $\\alpha$-Ta would falsify the material attribution; alternatively, a direct kinetic inductance measurement disagreeing strongly with the $0.6\\ \\mathrm{pH/sq}$ estimate would falsify the electrical interpretation.","tokens_in":83,"feed_emoji":"⚛️","tokens_out":4854,"duration_ms":57390,"temperature":0.7,"pith_summary":"The paper reports fabrication and measurement of tantalum coplanar waveguide resonators on high-resistivity silicon, using a thin niobium seed layer so the tantalum grows in its low-loss body-centered cubic phase. It shows that internal quality factor and kinetic inductance can be tuned by film thickness: a 100 nm film reaches an internal quality factor around $3.6\\times 10^6$ at high microwave power, while a 40 nm film gives an estimated kinetic inductance of $0.6\\ \\mathrm{pH/sq}$. The authors argue this combination is valuable for compact, high-impedance superconducting quantum circuits where strong coupling and small footprints matter. They fit the power and temperature dependence of the loss to the standard two-level-system and quasiparticle model.","feed_headline":"Alpha-tantalum circuits on silicon reach Qi above 3 million","feed_subtitle":"A 100 nm film gives low loss, while a 40 nm film maximizes kinetic inductance for compact quantum circuits.","key_machinery":"The load-bearing fabrication step is a 5 nm niobium seed layer sputtered before tantalum; the paper states this seed promotes the $\\alpha$ (body-centered cubic) phase on unheated silicon, a phase with low microwave loss. The tantalum film thickness then controls kinetic inductance through the standard relation $L_K \\approx \\hbar R_s/(\\pi \\Delta_0)$, estimated from measured normal-state sheet resistance and critical temperature. Performance is evaluated with a notch-type complex $S_{21}$ model to extract quality factors, and the TLS saturation model is used to separate two-level-system loss from quasiparticle loss.","core_discovery":"The central claim is that room-temperature sputtered $\\alpha$-Ta films on silicon, seeded by 5 nm niobium, can simultaneously deliver high internal quality factor and high kinetic inductance, and that the two quantities trade off with film thickness. Across 40, 80, and 100 nm films, the maximum internal quality factor at high power rises to about $3.6\\times 10^6$ for the 100 nm film, while the estimated kinetic inductance falls from $0.6$ to $0.2\\ \\mathrm{pH/sq}$ as the film thickens. This thickness-dependent dataset is presented as an engineering curve: choose thin tantalum when kinetic inductance and compact high-impedance circuits matter, and thicker tantalum when low loss matters.","pith_inferences":["A testable extension is to measure the actual crystallographic phase of the films with X-ray diffraction or transmission electron microscopy; if the seed layer produces mixed-phase or beta-Ta, the thickness-dependent story would need reinterpretation.","The kinetic inductance values are inferred from normal-state sheet resistance and the BCS gap, not from a direct microwave kinetic inductance extraction, so a direct measurement would pin down the numbers more firmly.","If the tradeoff holds, compact high-impedance resonators could couple more strongly to qubits and enable smaller or more densely integrated quantum processors, though a qubit demonstration remains beyond this paper."],"forward_implications":["A 100 nm alpha-Ta resonator on silicon reaches $Q_i \\sim 3.6\\times 10^6$ at high power, confirming Nb-seeded room-temperature tantalum as a viable low-loss material on CMOS-compatible substrates.","The 40 nm film's estimated $0.6\\ \\mathrm{pH/sq}$ kinetic inductance is competitive with other high-kinetic-inductance platforms and enables higher characteristic impedance at fixed geometry.","Because kinetic inductance rises as thickness falls, designers can choose a thickness to set the resonator's impedance and frequency without changing lithographic dimensions.","The power and temperature dependence of $Q_i$ follows the standard TLS plus quasiparticle model, so further gains should come from reducing interface TLS and quasiparticle generation."],"supporting_citations":[{"why":"Demonstrates tantalum resonators on silicon with a niobium buffer layer, providing the closest prior room-temperature seed-layer approach.","marker":"[31]"},{"why":"Provides an earlier tantalum-on-silicon seed-layer resonator baseline with high quality factors.","marker":"[33]"},{"why":"Reports high-Q alpha-tantalum resonators on heated silicon without a seed layer, the main alternative growth route this work compares against.","marker":"[35]"},{"why":"Supplies the formula used to estimate kinetic inductance from sheet resistance and critical temperature.","marker":"[46]"},{"why":"Gives the notch-type complex scattering model used to extract internal and coupling quality factors.","marker":"[53]"},{"why":"Provides the TLS saturation model used to fit power and temperature dependence of the loss.","marker":"[63]"},{"why":"Shows high-Q alpha-tantalum on sapphire, a benchmark for low-loss tantalum grown on heated substrates.","marker":"[30]"},{"why":"Supplies the effective penetration depth relation and a buffer-layer sapphire comparison for kinetic inductance behavior.","marker":"[38]"}],"fun_headline_variants":["Thin Ta boosts inductance; thick Ta boosts Q for compact qubits","Ta thickness tunes between low loss and high kinetic inductance","Room-temp alpha-Ta films give both high Q and high kinetic inductance","Thickness dial: 40 nm gives inductance, 100 nm gives Qi=3.6M","Alpha-Ta resonators hit Qi 3.6M and thickness-tuned inductance"],"cache_read_input_tokens":17152,"weakest_assumption_plain":"The whole interpretation rests on the 5 nm niobium seed layer actually growing the alpha (body-centered cubic) phase of tantalum, but the paper does not include direct structural characterization such as X-ray diffraction or electron microscopy.","fun_headline_variants_meta":{"raw":{"variants":["Thin Ta boosts inductance; thick Ta boosts Q for compact qubits","Ta thickness tunes between low loss and high kinetic inductance","Room-temp alpha-Ta films give both high Q and high kinetic inductance","Thickness dial: 40 nm gives inductance, 100 nm gives Qi=3.6M","Alpha-Ta resonators hit Qi 3.6M and thickness-tuned inductance"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0013,"raw_usage":{"total_tokens":5305,"prompt_tokens":946,"completion_tokens":4359,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":4257}},"tokens_in":562,"tokens_out":4359,"duration_ms":29664,"temperature":1.0,"reasoning_tokens":4257,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T10:46:57.602408+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An X-ray diffraction or transmission electron microscopy measurement of the 40 nm film that shows $\\beta$-Ta or a mixed phase rather than bcc $\\alpha$-Ta would falsify the material attribution; alternatively, a direct kinetic inductance measurement disagreeing strongly with the $0.6\\ \\mathrm{pH/sq}$ estimate would falsify the electrical interpretation.","supporting_citations":[{"cited_title":"Gao, The physics of superconducting microwave resonators, California Institute of Technology, 2008","cited_arxiv_id":null,"evidence_quote":"Provides the TLS saturation model used to fit power and temperature dependence of the loss."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates tantalum resonators on silicon with a niobium buffer layer, providing the closest prior room-temperature seed-layer approach."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports high-Q alpha-tantalum resonators on heated silicon without a seed layer, the main alternative growth route this work compares against."},{"cited_title":"Jones, N","cited_arxiv_id":null,"evidence_quote":"Supplies the formula used to estimate kinetic inductance from sheet resistance and critical temperature."},{"cited_title":"Moskalenko, I.y.S","cited_arxiv_id":null,"evidence_quote":"Gives the notch-type complex scattering model used to extract internal and coupling quality factors."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows high-Q alpha-tantalum on sapphire, a benchmark for low-loss tantalum grown on heated substrates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the effective penetration depth relation and a buffer-layer sapphire comparison for kinetic inductance behavior."}],"review_version":1}