{"id":"56ea77b0-c2a8-42f9-a784-791725e30e30","arxiv_id":"2608.07101","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"In disordered NbN films, the superconductor-insulator transition is driven by quantum phase fluctuations: the superfluid stiffness drops sharply near critical disorder while the pairing temperature stays finite, and the finite-temperature transition remains of BKT type.","lead":"This paper measures how the superconducting transition changes as disorder increases in thin niobium nitride films, right up to the point where superconductivity disappears. It finds that the loss of superconductivity is driven by quantum fluctuations of the phase of the superconducting order, not by the destruction of pairing, and that the finite-temperature transition stays of the Berezinskii-Kosterlitz-Thouless type even near the transition.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Near the SIT, the inflection-point Tc0 is the linchpin of the finite-Tc0 claim; the acknowledged loss of the independent cross-check leaves this unvalidated.","rationale":"The paper is strong where it does not depend on the disputed estimator: the internal consistency between TBKT from R(T) and from Js(T), the Halperin-Nelson fits over 3-4 decades, and the low-disorder behavior all support the fermionic suppression scenario and the BKT description. However, the central new claim is precisely the separation of Tc0 from Js(0) and TBKT near the SIT. The authors themselves flag in Supplementary Sec. I.E.a that the only independent cross-check is unavailable in exactly that regime. This is not a disagreement with the BKT interpretation or the data quality; it is an unvalidated estimator at the parameter range where the conclusion is drawn. The proposed synthetic-data test and the constraint-free amplitude-fluctuation refit would settle whether the inflection point tracks Tc0 near critical disorder. A conditional verdict is the right outcome, and no adjustment to the reader's verdict is needed.","tokens_in":19586,"tokens_out":6036,"duration_ms":56898,"concrete_test":"Use the existing R(T) data for RN >= 12 kOhm: fit the high-temperature branch above the inflection point to the amplitude-fluctuation conductivity expression of Ref. [21] (Aslamazov-Larkin and Maki-Thompson plus weak-localization corrections) with Tc0 as a free parameter and no constraint from the inflection point, and compare with the reported inflection-point values. Simultaneously, generate synthetic R(T) curves with known input Tc0 and tc0 = 0.65 and 0.8, using Eq. 1 below Tc0 and the same fluctuation expression above, add realistic noise, and run both estimators. If the amplitude-fluctuation fit cannot constrain Tc0 to within +/-10% for noisy synthetic data at tc0 >= 0.65, or if the fitted Tc0 values for real data deviate systematically from inflection-point values in the direction of the Js(0) deficit, then the finite-Tc0 line in Fig. 3f is not established by the current analysis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The decisive experimental input for the paper's central claim is the mean-field pairing temperature Tc0 in the high-disorder regime. All of the contrast in Fig. 3f between a finite Tc0 (light blue diamonds) and a sharply suppressed Js(0) (black circles) depends on identifying Tc0 with the center of the linear segment/inflection point of R(T) (Supplementary Sec. I.E.a, Fig. S7). The authors explicitly state that for strongly disordered samples the previous independent sanity check—quasiparticle suppression of Js(T) consistent with Eq. 3—is no longer possible because Js(T) is almost rectangular (Sec. I.E.a, Fig. 2a). The only near-SIT validation offered is a fit of the amplitude-fluctuation conductivity (red line in Fig. S7), but that fit is not independent: it uses the same R(T) trace and the same fluctuation theory that motivates the inflection-point convention. At tc0 = (Tc0 - TBKT)/Tc0 ~ 0.65-0.8, the vortex-dominated phase-fluctuation regime occupies most of the transition, so the curvature crossover defining the inflection point is itself shaped by vortex unbinding. If the true pairing scale actually tracks Js(0) and TBKT near Wc, then the apparent 'drop' of Js(0) below the mean-field prediction is an artifact of an overestimated Tc0, and the quantum-phase-fluctuation conclusion loses its experimental foundation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a systematic study of the disorder-driven superconductor-insulator transition (SIT) in ultrathin NbN films, measuring both sheet resistance R□(T) and superfluid stiffness Js(T) on the same meander devices for fifteen superconducting and three insulating samples with normal-state resistance up to the critical range 16-18 kΩ. The finite-temperature transition is found to remain of Berezinskii-Kosterlitz-Thouless (BKT) type across the whole superconducting range: Halperin-Nelson fits describe R□(T) over 3-4 decades, and the vortex-unbinding temperature TBKT from resistance agrees with that from the universal stiffness jump Js(TBKT) = 2TBKT/π. The central claim is that close to the SIT the zero-temperature stiffness Js(0) and TBKT drop sharply below the mean-field expectation, quantified through a one-parameter Finkel'stein fit for Tc0 and the Ginzburg-Levanyuk formula, while the pairing temperature Tc0, assigned to the inflection point of R(T), remains finite and follows the Finkel'stein curve. This is interpreted as evidence that the SIT is a second-order quantum phase transition driven by quantum phase fluctuations, with pairing surviving into the insulator, in contrast to the first-order scenario reported for InOx [26]. The paper also reports the vortex-core energy µ and the activation energy E* extracted from the low-temperature stiffness, and comparisons with self-consistent BdG simulations of the pairing amplitude distribution.","tokens_in":19903,"tokens_out":21271,"duration_ms":168655,"significance":"If correct, the main payoff is a clean discrimination between fermionic (pairing-destruction) and bosonic (phase-fluctuation) mechanisms for the SIT in a canonical disordered two-dimensional superconductor, together with the demonstration that the BKT character of the finite-T transition survives up to critical disorder. The manuscript has genuine strengths: simultaneous R□(T) and Js(T) measurements on identical devices; two fully independent determinations of TBKT (Halperin-Nelson resistance fits and the universal stiffness jump) that agree across the entire disorder range, which solidly grounds the BKT-character claim; a cross-check of Js from nonlinear IV characteristics near the SIT (Extended Data Fig. 2); a one-parameter Finkel'stein baseline for Tc0; and falsifiable predictions (power-law suppression of Js(0) and TBKT with z ≃ 1) that connect to quantum-critical scaling. The persistence of the tunneling gap Eg across the SIT (Fig. 3g) gives partial independent support for surviving pairing at strong disorder.","major_comments":[{"comment":"The central contrast of the paper—a finite Tc0 (light blue diamonds in Fig. 3f) coexisting with a sharply suppressed Js(0) (black circles) in regime III—rests on identifying Tc0 with the center of the linear segment / inflection point of R(T) at disorder levels where tc0 = (Tc0 - TBKT)/Tc0 reaches 0.65-0.8, i.e., where the phase-fluctuation window is about 80% of Tc0. Supplementary Sec. I.E.a states explicitly that the previous independent sanity check (quasiparticle suppression of Js(T) consistent with Eq. 3) is no longer possible for strongly disordered samples because Js(T) is almost rectangular (Fig. 2a). The near-SIT validation shown in Fig. S7 (red line) is a fit of the same amplitude-fluctuation theory to the same R(T) data that motivates the inflection-point convention, so it does not constitute an independent determination. In the regime tc0 = 0.65-0.8, the curvature crossover that defines the inflection point is itself shaped by the vortex-unbinding regime that spans most of the transition; if the true pairing temperature actually tracks Js(0) and TBKT near Wc, the apparent drop of Js(0) below the mean-field expectation would be an artifact of an overestimated Tc0, and the quantum-phase-fluctuation conclusion would lose its experimental foundation. I request (i) a sensitivity analysis of the Fig. 3f comparison under alternative Tc0 assignments (e.g., extrapolating the low-disorder amplitude-fluctuation fits without anchoring to the inflection point, or bounding Tc0 by the tunneling-gap scale), and (ii) an explicit statement of how much the regime-III Tc0 values would have to change for the claimed drop of Js(0) and TBKT relative to the Finkel'stein baseline to disappear. The zero-temperature tunneling gap Eg persisting into the insulator (Fig. 3g) supports the survival of pairing in the ground state, but it does not by itself pin the temperature scale Tc0, so it does not remove this concern.","section":"Supplementary Sec. I.E.a and Fig. S7; Fig. 3f; Fig. 2a"},{"comment":"The mean-field baselines against which the regime-III drop is measured are not independent of the disputed Tc0 convention. The black line in Fig. 3f is obtained by inserting the Finkel'stein fit for Tc0(RN)—a one-parameter fit (γc = 5.312, Extended Data Fig. 4) to the very same inflection-point-based Tc0 data—into Eq. 4, and the red line by inserting that fit into Eq. 5. Any systematic error in the inflection-point assignment in regime III therefore enters both the data points and the baselines, and the manuscript does not quantify how robust the regime-III deviation is to this shared input. Please (i) report the residuals and confidence interval of the one-parameter Finkel'stein fit, and (ii) re-examine the deviation of Js(0) and TBKT from the baselines when the Finkel'stein fit is performed on regime I-II data only (RN ≲ 10-12 kΩ) and extrapolated into regime III, where the inflection-point assignment is least constrained. This test would separate the claimed drop from the calibration of the baseline.","section":"Methods (Finkel'stein fit, Eq. 6); Fig. 3f; Extended Data Fig. 4"},{"comment":"The quantum-critical part of the conclusion (second-order transition, z ≃ 1) leans on the power-law analysis in the inset of Fig. 3f, where Js(0) ∝ (δg)^ν and TBKT ∝ (δg)^(zν) are stated to be 'consistent with' ν = 0.67. No fitting procedure, goodness-of-fit, or uncertainty is given, and the number of superconducting samples in regime III is small (of order four). Moreover δg = (Rc - RN)/Rc² depends on Rc, which is only bracketed as 16 kΩ < Rc < 18 kΩ; the resulting spread in δg can be large for near-critical points, and the implied uncertainty in (ν, z) is not propagated. Please provide the fit range, confidence intervals for ν and z, the number of points, and the sensitivity of the exponents to the choice of Rc within the stated bracket. If the exponent is too weakly constrained by the present data, the claim should be softened accordingly, since the first-order-versus-second-order distinction drawn from the InOx comparison in the Summary partly rests on this scaling analysis.","section":"Fig. 3f inset"}],"minor_comments":[{"comment":"The closing sentence of the abstract invokes the experimental discovery by Haviland, Liu, and Goldman, but the original paper (Phys. Rev. Lett. 62, 2180 (1989)) is not among the references; it should be cited there and where the SIT is introduced in the opening paragraph.","section":"Abstract"},{"comment":"The definition of the Finkel'stein parameter is inconsistent between the main text (γc = ln[ħ/(kBTc00τ)] in Eq. 6) and the Supplement (γc = 1/log(Tc00τ) in Eq. S.11), and the supplementary form is dimensionally inconsistent as written. Since γc = 5.312 is the single free parameter of the central baseline fit, the two definitions must be reconciled.","section":"Eq. (6) vs. Supplementary Eq. S.11"},{"comment":"The universal BKT relation appears with an inverted factor in two places: 'TBKT = 2Js(TBKT)/π' and 'Js(TBKT) = πTBKT/2' should both read Js(TBKT) = 2TBKT/π (equivalently TBKT = (π/2)Js(TBKT)); elsewhere the paper uses the correct form (Fig. 2 caption), so this is presumably a typographical slip, but it should be corrected because the 2/π factor is invoked in the inset of Fig. 3f.","section":"Main text, 'Similarly, Eq. 5 ...' and 'Close to the SIT ...' paragraphs"},{"comment":"Several typographical errors should be cleaned up: 'One film film with lower disorder' (Methods), 'previous tudies' (Supplementary Sec. I.D.c), 'the the measured Tc0' (Methods, analysis of Tc0), and 'Rychaudhuri' (Extended Data Fig. 4 caption, for Raychaudhuri).","section":"Methods; Supplementary Sec. I.D.c; Extended Data Fig. 4 caption"},{"comment":"The Kosterlitz 1974 paper is cited twice with different page numbers (J. Phys. C 7, 1060 in [10] and 1046 in [56]); reference [12] also contains a garbled author string ('V. Ambegaokar, D. R. N., B. I. Halperin & Siggia, E. D.'). These should be normalized.","section":"References [10], [12], [56]"},{"comment":"The claim that broadening in two-coil measurements 'occurs for the two-coil method only' and is caused by roughly 10% edge-to-center variations of Js(0) and RN is supported only by an 'in preparation' citation; since this claim is used to reinterpret previously published BKT broadening results [27], it should either be presented with its supporting data or clearly flagged as preliminary.","section":"Methods, two-coil comparison"},{"comment":"Given the sensitivity of the Tc0 assignment discussed in the major comments, the phrases 'unambiguous evidence for quantum phase fluctuations' (Abstract) and 'strongly suggest' (Summary) overstate the present support; I recommend qualifying these statements until the near-SIT Tc0 determination is independently corroborated.","section":"Abstract and Summary"}],"recommendation":"major_revision","confidential_remarks":"For the editor's attention: (1) Several claims rely on unpublished sources—Refs. [29] (Delta-distribution crossover), [30] (tunneling Eg data of Fig. 3g), [42], and the 'in preparation' Weitzel-Dutta-Pfaffinger work cited in Methods for the two-coil comparison; the authors should be asked to either include the necessary data or explicitly mark the affected claims as preliminary. (2) The Haviland-Liu-Goldman 1989 paper is invoked in the abstract but missing from the references. (3) The authors' own Supplementary Sec. I.E.a concedes the missing independent cross-check at high disorder; this is the point referees will converge on, and the authors would be well served to address it head-on in a revision. (4) The manuscript's strong experimental dataset and one contentious identification fit a broad physics journal, but the promotional language ('unambiguous evidence', 'for the first time') may invite scrutiny given the outstanding Tc0 question."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the systematic comparison of Js(0), TBKT, and Tc0 across disorder up to the SIT, with transport and kinetic inductance measured on the same NbN devices. The BKT analysis is convincing: Halperin-Nelson fits span 3–4 decades, and TBKT from resistance agrees with the universal jump in stiffness. The sharp drop of Js(0) below the BCS/Finkel'stein baseline near 12 kOhm is a real trend, not an artifact of a single point.\n\nThe soft spot is exactly where the stress-test note lands. The central contrast—finite Tc0 while Js(0) collapses—relies on identifying Tc0 with the inflection point of R(T) in the high-disorder regime, where the previous cross-check (quasiparticle suppression of Js(T)) is no longer available because Js(T) is almost rectangular. The supplementary fit to amplitude fluctuations is called independent, but it uses the same R(T) trace and the same fluctuation theory, so it is not a separate test. If the inflection point overestimates Tc0 near the SIT—e.g., because the curvature crossover is shaped by vortex unbinding—then the apparent drop of Js(0) is partly an artifact.\n\nI don't think this is fatal. The Finkel'stein fit describes Tc0 across the whole range, the tunneling gap persists into the insulating side, and the authors are explicit about the loss of the cross-check. The abstract's 'unambiguous' is too strong; 'consistent with' would be fairer. Missing error bars on the central energy-scale plots and reliance on unpublished manuscripts for two key inputs (the Delta-distribution crossover and the tunneling data) add friction for a referee but don't undermine the core measurement.\n\nThis deserves a serious referee. The data set is valuable and the BKT analysis is rock solid. The main fix is a robustness analysis of Tc0 near the SIT—how much does the conclusion shift under alternative definitions, or with a paraconductivity theory that includes phase fluctuations. No red flags in the citation pattern; they engage the competing InOx first-order scenario directly.\n\nI'd cite it, and I'd bring it to reading group with the explicit task of scrutinizing the Tc0 extraction.","headline":"Careful NbN stiffness and transport study; the BKT analysis is strong, but the key contrast between finite Tc0 and suppressed Js(0) near the SIT depends on an inflection-point Tc0 that lacks an independent cross-check.","tokens_in":20503,"tokens_out":3636,"would_cite":true,"duration_ms":32604,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["74.25.Dw","74.40.-n","74.62.En"],"model":"deepseek-v4-flash","headline":"Phase fluctuations of the superconducting order parameter, not the destruction of pairing, drive the superconductor-insulator transition in disordered two-dimensional NbN films.","keywords":["superconductor-insulator transition","quantum phase fluctuations","BKT transition","superfluid stiffness","disordered NbN films","two-dimensional superconductivity","kinetic inductance","pairing temperature"],"falsifier":"A direct falsifier would be tunneling spectroscopy on the same NbN films for $R_N$ between roughly 12 and 16 kΩ: if the single-particle gap $E_g$ closes as $J_s(0)$ vanishes near the critical resistance, the fermionic (pairing-destruction) scenario is correct; if the gap stays open on the insulating side while the stiffness approaches zero, the phase-fluctuation claim survives. A sharper version is to measure $E_g$ and $J_s(0)$ on the same device and check whether $J_s(0)$ reaches zero at a disorder strength where $E_g$ is still a finite fraction of its clean value.","tokens_in":19339,"feed_emoji":"❄️","tokens_out":7491,"duration_ms":62457,"temperature":0.7,"pith_summary":"This paper reports measurements of resistance and superfluid stiffness on fifteen superconducting and three insulating disordered NbN films spanning a wide range of disorder. It establishes that the finite-temperature transition from superconductor to resistive state remains of Berezinskii-Kosterlitz-Thouless (BKT) type all the way to the superconductor-insulator transition (SIT), with Halperin-Nelson fits holding over three to four decades of resistance. The central finding is that very close to the SIT the zero-temperature superfluid stiffness and the BKT temperature drop sharply below the mean-field expectation while the pairing temperature stays finite and continues to follow the fermionic suppression law. The authors conclude that the SIT in these films is a second-order quantum phase transition driven by quantum fluctuations of the superconducting phase, not by disorder destroying the pairing amplitude.","feed_headline":"Phase fluctuations, not broken pairing, drive the SIT in 2D films","feed_subtitle":"Superfluid stiffness collapses while pairing stays intact, keeping BKT physics all the way to critical disorder.","key_machinery":"The argument is carried by three objects: the superfluid phase stiffness $J_s(T)$, extracted from the sheet kinetic inductance in a resonant circuit; the mean-field pairing temperature $T_{c0}$, assigned to the inflection point of $R(T)$ where phase- and amplitude-fluctuation regimes meet; and the BKT universal line $J_s(T_{BKT}) = 2T_{BKT}/\\pi$. Resistance in the phase-fluctuation window is fitted to the Halperin-Nelson formula $R(T) = C \\exp(-b/\\sqrt{T/T_{BKT}-1})$, and $T_{c0}$ is compared with the fermionic suppression formula and with the Ginzburg-Levanyuk correction. Vortex core energy $\\mu$ extracted from numerical solution of the Kosterlitz renormalization-group equations links the thermally activated depletion of $J_s$ to vortex-antivortex pairs, and a mapping between normal-state resistance and Anderson disorder strength $W$ ties the experiment to mean-field Bogoliubov-de Gennes simulations of the pairing landscape.","core_discovery":"On the authors' own terms, the paper establishes that the superconductor-insulator transition in homogeneously disordered two-dimensional NbN films is driven by quantum phase fluctuations. The evidence: as disorder increases toward the critical value, the mean-field pairing temperature $T_{c0}$ extracted from the inflection point of $R(T)$ remains finite and follows the fermionic suppression law, while the zero-temperature superfluid stiffness $J_s(0)$ and the BKT temperature $T_{BKT}$ drop sharply below their mean-field and Ginzburg-Levanyuk expectations. Near the transition $J_s(0)$ and $T_{BKT}$ vanish together, consistent with the universal BKT relation and with $J_s(0) \\propto (\\delta g)^{\\nu}$ and $T_{BKT} \\propto (\\delta g)^{z\\nu}$, where $\\delta g = (R_N - R_c)/R_c^2$, $\\nu \\simeq 0.67$, and $z \\simeq 1$. The finite-temperature transition remains of BKT type throughout, so the zero-temperature SIT is a continuous quantum phase transition of the phase variable.","pith_inferences":["If pairing indeed survives into the insulating phase, one expects a regime of preformed or localized Cooper pairs on the insulating side; a testable prediction is that diamagnetic or kinetic-inductance signatures should continue tracking the gap rather than the stiffness.","Because the inflection-point $T_{c0}$ is the only anchor for the pairing scale at high disorder, the paper's central contrast would be strengthened by a concurrent tunneling measurement of the single-particle gap on the same films used for stiffness data; the authors note this cross-check is currently missing.","The $\\mu \\simeq (\\pi^2/2)J_s(0)$ result and the near-power-law stiffness depletion reported in other disordered superconductors suggest that a single vortex-core-energy scale may govern low-temperature electrodynamics of high-kinetic-inductance films; if so, the design rules for microwave circuits should account for vortex excitations rather than only quasiparticles.","A direct extension of the comparison to other materials with the same combined resistance-plus-stiffness protocol could test whether the order of the transition (continuous versus first-order) is controlled by screening strength and dimensionality, as the authors conjecture."],"forward_implications":["The zero-temperature superconductor-insulator transition in disordered NbN films is a continuous (second-order) quantum phase transition driven by phase fluctuations, not by destruction of the pairing amplitude.","The finite-temperature transition remains of BKT type up to the critical disorder; the zero-resistance state is always bounded by the BKT line $J_s = 2T/\\pi$.","Very close to the SIT, $J_s(0) \\propto (\\delta g)^{\\nu}$ and $T_{BKT} \\propto (\\delta g)^{z\\nu}$ with $\\nu \\simeq 0.67$ and $z \\simeq 1$, following the bosonic quantum-phase-fluctuation scenario.","Thermal depletion of superfluid stiffness crosses over from quasiparticle-dominated to vortex-like excitations near the transition, which matters for quantum circuits built from disordered superconductors.","The SIT in NbN is distinct from the first-order, glassy transition reported in InO$_x$; the difference is attributed to material parameters such as thickness, disorder strength, and screening."],"supporting_citations":[{"why":"Supplies the fermionic suppression law for $T_{c0}$ that the paper fits to the pairing-temperature data.","marker":"[3]"},{"why":"Predicts that quantum phase fluctuations drive the superfluid-insulator transition and supplies the scaling form for $J_s(0)$.","marker":"[4]"},{"why":"Predicts power-law vanishing of $J_s(0)$ and $T_{BKT}$ near the quantum critical point, which the paper compares with its exponent $\\nu \\simeq 0.67$.","marker":"[5]"},{"why":"Provides the bosonic quantum-diffusion picture and universal resistance at the transition, framing the interpretation of the critical regime.","marker":"[6]"},{"why":"Gives the Halperin-Nelson formula (Eq. 1) used to fit resistance data and extract $T_{BKT}$ over three to four decades.","marker":"[11]"},{"why":"Provides the BKT relation for disordered films and the Ginzburg-Levanyuk number used to compute the expected $T_{BKT}(R_N)$.","marker":"[19]"},{"why":"Previous work by the same group supplying the extraction procedures for $T_{c0}$, $J_s(T)$, and the BKT parameters.","marker":"[21]"},{"why":"Reports a first-order superconductor-insulator transition in amorphous InO$_x$, serving as the comparison case for the NbN result.","marker":"[26]"},{"why":"Provides the renormalization-group procedure for broadening of the BKT transition by inhomogeneity, used for the stiffness fits.","marker":"[27]"},{"why":"Gives a mean-field theory of first-order quantum superconductor-insulator transitions, cited to contrast with the continuous transition observed in NbN.","marker":"[45]"}],"fun_headline_variants":["Quantum phase fluctuations, not broken pairing, drive 2D film SIT","Pairing stays, phase order falls: the secret to 2D SIT","SIT in 2D films: phase fluctuations, not pairing destruction","Disorder kills phase order, not pairing, in 2D SIT","The 2D SIT: a quantum phase transition, not a pairing collapse"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the inflection point of the resistance curve continues to mark the mean-field pairing temperature $T_{c0}$ in the high-disorder regime, where the independent quasiparticle cross-check used at lower disorder is no longer available; if the inflection point stops tracking $T_{c0}$ near the SIT, the contrast between a finite $T_{c0}$ and a sharply suppressed $J_s(0)$ collapses.","fun_headline_variants_meta":{"raw":{"variants":["Quantum phase fluctuations, not broken pairing, drive 2D film SIT","Pairing stays, phase order falls: the secret to 2D SIT","SIT in 2D films: phase fluctuations, not pairing destruction","Disorder kills phase order, not pairing, in 2D SIT","The 2D SIT: a quantum phase transition, not a pairing collapse"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000882,"raw_usage":{"total_tokens":3883,"prompt_tokens":1091,"completion_tokens":2792,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":707,"completion_tokens_details":{"reasoning_tokens":2690}},"tokens_in":707,"tokens_out":2792,"duration_ms":19369,"temperature":1.0,"reasoning_tokens":2690,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T14:52:28.505157+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct falsifier would be tunneling spectroscopy on the same NbN films for $R_N$ between roughly 12 and 16 kΩ: if the single-particle gap $E_g$ closes as $J_s(0)$ vanishes near the critical resistance, the fermionic (pairing-destruction) scenario is correct; if the gap stays open on the insulating side while the stiffness approaches zero, the phase-fluctuation claim survives. A sharper version is to measure $E_g$ and $J_s(0)$ on the same device and check whether $J_s(0)$ reaches zero at a disorder strength where $E_g$ is still a finite fraction of its clean value.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the fermionic suppression law for $T_{c0}$ that the paper fits to the pairing-temperature data."},{"cited_title":"P., Weichman, P","cited_arxiv_id":null,"evidence_quote":"Predicts that quantum phase fluctuations drive the superfluid-insulator transition and supplies the scaling form for $J_s(0)$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Predicts power-law vanishing of $J_s(0)$ and $T_{BKT}$ near the quantum critical point, which the paper compares with its exponent $\\nu \\simeq 0.67$."},{"cited_title":"P., Grinstein, G","cited_arxiv_id":null,"evidence_quote":"Provides the bosonic quantum-diffusion picture and universal resistance at the transition, framing the interpretation of the critical regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Halperin-Nelson formula (Eq. 1) used to fit resistance data and extract $T_{BKT}$ over three to four decades."},{"cited_title":"J.et al.Berezinskii-kosterlitz-thouless transi- tion in homogeneously disordered superconducting films","cited_arxiv_id":null,"evidence_quote":"Provides the BKT relation for disordered films and the Ginzburg-Levanyuk number used to compute the expected $T_{BKT}(R_N)$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous work by the same group supplying the extraction procedures for $T_{c0}$, $J_s(T)$, and the BKT parameters."},{"cited_title":"Phys.21, 104–109 (2025)","cited_arxiv_id":null,"evidence_quote":"Reports a first-order superconductor-insulator transition in amorphous InO$_x$, serving as the comparison case for the NbN result."},{"cited_title":"& Giamarchi, T","cited_arxiv_id":null,"evidence_quote":"Provides the renormalization-group procedure for broadening of the BKT transition by inhomogeneity, used for the stiffness fits."},{"cited_title":"& Feigel’man, M","cited_arxiv_id":null,"evidence_quote":"Gives a mean-field theory of first-order quantum superconductor-insulator transitions, cited to contrast with the continuous transition observed in NbN."}],"review_version":1}