{"id":"6bda42db-afda-49a4-9b33-cc2b2f608ad9","arxiv_id":"2511.11803","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A single Fe impurity in Bi2Se3 shows super-Poissonian shot noise that tracks its 16-meV vibrational sidebands, evidence of vibron-mediated electron bunching at the atomic scale.","lead":"This paper reports atomic-scale shot-noise measurements on a single Fe impurity in Bi2Se3 and observes super-Poissonian noise (Fano factor ~1.16) in the same voltage range as vibrational sidebands. It argues this is the first atomic-scale evidence that electron-vibron coupling makes electrons tunnel in bunches, a predicted effect that could later be used to inject paired electrons.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Super-Poissonian Fano factor relies on the assumption that S=2eIF with F constant; the paper only shows near-linear S(I) over a restricted pA–nA range, leaving open a current-proportional artifact.","rationale":"The reader identified the linearity of the shot-noise relation as the weakest assumption; I agree. The paper's own caveat about the restricted current range and the modest F value (≈1.16) make this the single most load-bearing point. A failure of this assumption would invalidate the interpretation of F>1 as bunching. Other weaknesses (quantitative model mismatch, missing error bars, raw data unavailable) are secondary: they affect the strength of the evidence but not the core logical step. Therefore the appropriate verdict remains conditional, pending this direct check.","tokens_in":11947,"tokens_out":5820,"duration_ms":52853,"concrete_test":"At fixed bias in the resonance, measure the noise power S at the Fe1 centre as a function of setpoint current I from 1 pA to 10 nA (at least 3 decades). Plot S vs I and fit S = A I + B. Verify that B equals the amplifier noise floor (measured with the tip retracted) and that A is constant to within, say, 5% across the range; also record the noise spectrum around the LC resonance to check that the excess component is white rather than 1/f. If S/I is not constant, the reported Fano factor does not represent tunnelling statistics.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that F>1 at the Fe1 centre implies vibron-mediated electron bunching depends critically on identifying the measured noise as shot noise, S=2e|I|F. The paper defines F this way (Results) but reports only a 'near-linear' current dependence (SI 2.6) and acknowledges in the Discussion that this may arise from the restricted current range (pA–nA). If part of the excess noise is a current-proportional artefact (e.g., a bias-dependent amplifier gain, a resistance fluctuation ∝ I, or a nonlinear I–V converting voltage noise into a current signal), then the computed F is not a true Fano factor, and F>1 does not establish bunching. The 2Fe1 control and spatial localization partly mitigate this, but they do not exclude a noise source intrinsic to the Fe1 junction. The paper itself flags the unexplained persistence of noise beyond the resonance, further indicating that the shot-noise interpretation is not fully secured. To support the claim, the authors need a direct test of the linearity of S vs I over a wide current range and a demonstration that the excess noise is white in the measured band.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports a combined scanning tunnelling spectroscopy and shot-noise study of an individual subsurface Fe impurity (Fe1) in Bi2Se3. The authors observe periodic fine structure in the normalized differential conductance with spacing 16.0±0.8 meV, identified as vibron sidebands of the A^2_1g surface phonon, and fit the sideband envelope to extract a Franck–Condon coupling λ=2.6±0.5. Simultaneously recorded current noise yields a Fano factor F=S/(2eI) that increases from ~1 to ~1.16 as the bias enters the impurity resonance, is spatially localized at the impurity centre, and is absent in 2Fe1 dimer controls. The paper interprets F>1 as vibron-mediated electron bunching, supporting this with a minimal model F=1+2(Iτ/e)(g^(2)(0+)-1) with τ=1 ps from literature and g^(2)(0+)~13 chosen to reproduce the observed F. The Discussion carefully lists alternative sources (mechanical instabilities, charging, resistance fluctuations, multichannel tunnelling, spin dynamics) and argues they are excluded.","tokens_in":12244,"tokens_out":10623,"duration_ms":90471,"significance":"If the shot-noise interpretation is correct, this is the first atomic-scale observation of vibron-mediated electron bunching, a phenomenon predicted two decades ago and of potential interest for injecting N-electron bundles. The paper's strengths are the simultaneous atomically resolved STS and MHz shot-noise measurements, the clear spatial localization of the excess noise, the 2Fe1 control showing reduced noise, and the explicit, honest discussion of alternative mechanisms and the model's limitations. However, the central claim currently rests on a modest F~1.16 enhancement with no reported error bars in the main text, on a 'near-linear' S(I) relation over a restricted current range, and on a model whose key parameter g^(2)(0+) is fixed to reproduce the measured F. These gaps are load-bearing because F>1 only implies bunching if the measured noise is genuine shot noise.","major_comments":[{"comment":"The central definition F=S/(2e|I|) only has the meaning of a Fano factor if the excess noise is shot noise, i.e. S scales linearly with |I| and is white in the measured band. The paper reports only a 'near-linear' current dependence (SI 2.6) and explicitly concedes in the Discussion that the linear behaviour may arise from the restricted pA–nA range. A current-proportional artifact intrinsic to the Fe1 junction (e.g., bias-dependent gain, resistance fluctuations proportional to I, or nonlinear I–V conversion of voltage noise) would produce the same apparent F>1 without electron bunching. The 2Fe1 control and spatial map reduce but do not exclude such an artifact. Please provide a direct S-vs-I measurement over at least a decade at fixed bias and show that the excess noise is white in the measured 20 kHz band.","section":"§2 'Shot-noise spectroscopy'; SI 2.6"},{"comment":"The main text reports F rising to about 1.16, but no error bars or statistical significance are given for this modest enhancement. A 16% excess could be within systematic or statistical uncertainty. Please provide error bars or confidence intervals on F from repeated spectra or a calibrated noise floor, and a quantitative test that F>1 at the relevant biases. Also show representative raw excess-noise spectra to verify that the 20 kHz band around 1.05 MHz is not contaminated and that the noise is white.","section":"Fig. 4c; §2"},{"comment":"The interpretive model is partly circular. λ=2.6 is fit to the same Franck–Condon sidebands used to identify the vibron, and g^(2)(0+) is chosen so that F=1+2(Iτ/e)(g^(2)(0+)-1) reproduces the measured F=1.16 with τ=1 ps from literature. Thus the statement 'we obtain g^(2)(0+)≃13' is a retrodiction, not an independent test of the vibronic-bunching mechanism. Please either derive g^(2)(0+) from a microscopic model using the measured λ, or explicitly present the model as illustrative. The measured F>1 is independent evidence for bunching, but the model should not be cited as independent confirmation.","section":"§3 'Vibron induced noise'; SI 3"},{"comment":"The exclusion of resonant tunnelling via interacting levels (ref. 39) is too quick. A single localized interacting level can itself produce super-Poissonian noise; the DFT result that Fe1 has only one d_z2-derived resonance does not by itself rule out this dynamical mechanism. Please provide a quantitative comparison—e.g., the expected F and its bias/current dependence for the Safonov-type mechanism—and explain why it fails to reproduce the observed spatial localization and the contrast with 2Fe1. This is needed because the paper's case rests on 'in the absence of alternative sources'.","section":"§3 'Origin of super Poissonian noise'"},{"comment":"The paper notes that the noise remains enhanced well beyond the resonance and only returns to Poissonian near the valence band edge, and that no enhancement is observed at the impurity lobes despite sidebands there. These observations are not explained by the minimal vibronic model and are labelled 'puzzling'. For a claim of evidence, the authors should address whether these features are compatible with the vibronic-bunching interpretation or indicate an additional contribution to S. As written, the mismatch between the noise energy dependence and the sideband structure weakens the direct connection between F>1 and the measured vibron.","section":"§4 Outlook; Fig. 4e"}],"minor_comments":[{"comment":"The author list has M. Amato, but the contributions section mentions 'A. Amato'; check spelling and initials.","section":"Author Contributions"},{"comment":"'normalised by (I/V) to remove the setup effect' is unclear; specify what the setup effect is (e.g., the tip–sample distance dependence of the transmission).","section":"Methods"},{"comment":"The notation 'Fe1' and 'Fe 1' is used inconsistently; choose one convention and apply it uniformly.","section":"Throughout"},{"comment":"The equation F=1+2(Iτ/e)(g^(2)(0+)-1) appears without derivation; give a brief derivation or a more explicit reference to Supplementary Information Section 3.","section":"§3 'Vibron induced noise'"},{"comment":"The sentence 'All other noise is linear in current' is confusing because the preceding sentence states that the charging-ring noise is quadratic; clarify which regions are included in 'all other'.","section":"Fig. 4 caption"}],"recommendation":"major_revision","confidential_remarks":"This is a challenging and potentially important experiment, and the 2Fe1 control is a genuine strength. The qualitative observation of enhanced current noise appears real, but the interpretation as vibron-mediated bunching is not yet secure. The requested S(I) linearity and whiteness tests, plus proper error bars, are essential and should be feasible with the existing setup. If the authors can provide them, I would support publication; as it stands, the central claim overreaches the presented evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does something genuinely new: atomically resolved shot-noise maps on a single Fe impurity in Bi2Se3, with simultaneous conductance and Fano-factor imaging, showing a spatially localized super-Poissonian noise correlated with 16-meV Franck-Condon sidebands. The 2Fe1 dimer control behaving oppositely is a nice touch, and the DFT analysis of orbital character supports the single-channel picture. The experimental effort is substantial and the internal consistency is good. Credit is due for taking the prediction of vibron-mediated bunching seriously and building a measurement that can actually test it at the atomic scale.\n\nThe main worry is the identification of the excess noise as shot noise. The paper defines F = S/2e|I| and reports only a \"near-linear\" current dependence in the SI, over a pA–nA range; the Discussion concedes that this linearity may reflect the restricted current range. If part of the excess noise is a current-proportional artifact—say, amplifier gain drift, resistance fluctuations, or voltage noise converted by a nonlinear I–V—then F>1 does not establish bunching. The 2Fe1 control and the spatial localization mitigate but do not exclude a source intrinsic to the Fe1 junction. The F value is modest (≈1.16) with no error bars in the main text, and the theoretical model fixes g^(2)(0+) to reproduce the measured F, so the quantitative agreement is not a parameter-free prediction. The paper itself flags two real anomalies: noise persisting well beyond the resonance and the absence of enhancement at the impurity lobes. Those do not sink the qualitative claim, but they show the picture is incomplete.\n\nNone of this is fatal. The claim is plausible, and the authors are honest about the limitations. But the evidence is not yet a demonstration. A direct test of S vs. I linearity over a much wider current range, plus a demonstration that the excess noise is white in the measured band, would strengthen the case considerably.\n\nThis paper deserves a serious referee. I would send it out rather than desk reject. It is a careful experiment with a well-defined hypothesis, and the weaknesses are identifiable and addressable. For someone working on STM shot noise or vibronic transport, it is worth reading and worth citing; for the broader community, the central claim should be treated as conditional.","headline":"A careful STM shot-noise study that plausibly observes vibron-mediated bunching at a single Fe impurity, but the central Fano factor is modest and the shot-noise identification rests on a near-linear current dependence over a narrow range.","tokens_in":12782,"tokens_out":1781,"would_cite":true,"duration_ms":16576,"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":"Vibron-mediated electron bunching appears in atomic-scale shot noise at a single Fe impurity.","keywords":["shot noise","Fano factor","vibron","electron bunching","Franck-Condon blockade","Bi2Se3","scanning tunneling microscope","single impurity"],"falsifier":"Widen the shot-noise measurement to currents spanning several orders of magnitude (e.g., up to microamperes) at a fixed bias in the resonance. If the noise power departs from linearity in current—or the Fano factor changes systematically with current—the super-Poissonian value may be an artifact of the narrow current window rather than vibronic bunching. Alternatively, tuning the bias so that the 16 meV vibron mode is no longer excited (below the sideband threshold) should return F to 1; a persistent F>1 there would indicate another mechanism.","tokens_in":11840,"feed_emoji":"⚛️","tokens_out":4363,"duration_ms":38331,"temperature":0.7,"pith_summary":"The paper claims that electrons tunnelling through an iron impurity embedded in Bi2Se3 can arrive in bunches because each tunnelling event leaves the impurity's vibrational mode excited. It supports this with two linked measurements: tunnelling spectra showing 16 meV vibronic sidebands with Franck-Condon coupling λ≈2.6, and simultaneous shot-noise maps showing super-Poissonian noise (Fano factor >1) that grows with bias into the resonance and saturates. A control dimer impurity, where Coulomb charging dominates, shows suppressed noise, ruling out generic artifacts. If correct, this is the first atomic-scale demonstration of a theoretical prediction made decades ago and opens a route to injecting N-paired electrons from a single site.","feed_headline":"Electrons tunnel in bunches through a vibrating iron atom","feed_subtitle":"Super-Poissonian shot noise at a single Fe impurity in Bi2Se3 reveals vibron-mediated bunching.","key_machinery":"The central element is a subsurface Fe impurity acting as a nano-electromechanical resonator: its electronic occupation is coupled through Holstein coupling to a 16 meV A1g phonon mode of Bi2Se3, so that each tunnelling event displaces the oscillator and transiently raises the probability of the next event. Shot noise is quantified by the Fano factor F=S/(2e|I|). The authors connect F to the vibron dynamics through the second-order correlation function g(2)(t), yielding F=1+2(Iτ/e)(g(2)(0+)−1), where τ is the vibron lifetime and g(2)(0+)>1 encodes vibron-enhanced tunnelling. Franck-Condon factors set the weights of multi-vibron sidebands and give the fitted λ, and the rate hierarchy Γ≫γ,1/τ","core_discovery":"Tunnelling through a single Fe impurity that replaces a Bi atom beneath the Bi2Se3 surface is accompanied by excitation of a 16 meV surface vibron mode. The differential conductance shows oscillatory sidebands with a spacing of 16.0±0.8 meV matching the A1g surface phonon, fitted with Franck-Condon factors giving an electron-vibron coupling of λ=2.6±0.5. Simultaneous current-noise measurements at the centre of the impurity give a Fano factor F=1.16 that rises with voltage as the resonance and higher vibron sidebands are entered, saturates where the sidebands fade, and returns to 1 at the valence band edge. The authors argue that after a careful elimination of magnetic, interference, and char","pith_inferences":["The model's prediction that F increases with vibron lifetime suggests that molecules with longer-lived vibrational modes should show avalanche-like bunching; looking for giant Fano factors in such systems would be a direct test of this paper's picture.","The sharp spatial contrast—noise at the impurity centre but not at the lobes—implies that the orbital character of the tunnelling path matters for vibronic feedback; a theory incorporating impurity wavefunctions rather than point coupling would be needed to capture it.","If the linear current dependence is confirmed over a wider range, the inferred inverse-current phonon lifetime would indicate that inelastic tunnelling itself damps the vibron; measuring F as a function of temperature could separate this from lifetime-limited mechanisms.","The same shot-noise technique applied to other dopants with known phonon modes could map which vibrations mediate bunching and estimate coupling strengths from a single noise map."],"forward_implications":["Super-Poissonian shot noise (F>1) at a single impurity is a direct, atomic-scale signature of electron bunching, a regime previously seen only in engineered mesoscopic devices.","The Fano factor rises with the number of accessible vibron excitations and saturates when the sidebands disappear, tying the noise enhancement to the vibron spectrum.","The absence of noise enhancement on a neighbouring dimer impurity, where Coulomb charging dominates, indicates that vibronic coupling rather than generic disorder drives the bunching.","Because the enhanced noise is spatially localized at the impurity centre and insensitive to magnetic fields, it can be used as a local probe of electron-phonon correlations in other doped topological insulators.","If electron coherence can be added, the same mechanism could serve as an on-demand injector of N-paired electrons into surrounding quantum matter."],"fun_headline_variants":["Vibrating iron atom bunches electrons in shot noise","Shot noise reveals vibron-driven electron bunching","Electron bunching from a single vibrating atom","Vibron coupling bunches electrons at a single iron site","Super-Poissonian noise shows electron bunches via vibron"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the measured current noise at the Fe1 centre is genuine shot noise with a power that scales linearly with the DC current; the authors report only a near-linear dependence over a narrow pA–nA range, so if a current-proportional mechanical or electronic artifact contributes, the elevated Fano factor would not prove electron bunching.","fun_headline_variants_meta":{"raw":{"variants":["Vibrating iron atom bunches electrons in shot noise","Shot noise reveals vibron-driven electron bunching","Electron bunching from a single vibrating atom","Vibron coupling bunches electrons at a single iron site","Super-Poissonian noise shows electron bunches via vibron"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000208,"raw_usage":{"total_tokens":1224,"prompt_tokens":710,"completion_tokens":514,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":454,"completion_tokens_details":{"reasoning_tokens":434}},"tokens_in":454,"tokens_out":514,"duration_ms":4089,"temperature":1.0,"reasoning_tokens":434,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T22:09:10.846273+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Widen the shot-noise measurement to currents spanning several orders of magnitude (e.g., up to microamperes) at a fixed bias in the resonance. If the noise power departs from linearity in current—or the Fano factor changes systematically with current—the super-Poissonian value may be an artifact of the narrow current window rather than vibronic bunching. Alternatively, tuning the bias so that the 16 meV vibron mode is no longer excited (below the sideband threshold) should return F to 1; a persistent F>1 there would indicate another mechanism.","supporting_citations":[],"review_version":1}