{"id":"60642d26-4ae4-4014-b31a-09e0f862b399","arxiv_id":"2607.02899","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.5,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Interfacial thermal conductance at Au/Bi1-xSbx topological insulator junctions is non-monotonic in temperature and reversibly tunable by bias current via topological interface states, confirmed by trivial and insulating controls.","lead":"Experiments show heat flow across gold–topological-insulator contacts can be turned up and down with electric current by moving carriers between protected interface states and bulk bands. This offers a solid-state way to actively manage heat without moving parts, useful for dense high-power electronics.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified beyond the reader's already-flagged eDMM hierarchy and ARPES-under-Au assumption.","rationale":"The manuscript is a careful experimental study whose central claim—that TIS mediate and electrically modulate interfacial thermal conductance—is supported by the non-monotonic T dependence and bias-current peaks that appear only in the topological samples and vanish in both the trivial-semimetal and electronically-decoupled controls. Bulk μ, κ, and junction resistance are shown to be current-independent, ruling out bulk or Joule-heating artifacts. The quantitative eDMM estimates and WKB tunneling argument contain free parameters and idealizations, so absolute magnitudes remain less certain; this is precisely the medium correctness risk already identified by the reader and is the reason for CONDITIONAL rather than unconditional ACCEPT. No additional load-bearing flaw (circularity, missing control, or internal contradiction) is required to explain residual uncertainty. The proposed ARPES-under-Au check would directly test the single most important modeling assumption without altering the experimental pattern that already justifies the conditional verdict.","tokens_in":19845,"tokens_out":538,"duration_ms":5941,"concrete_test":"Deposit a sub-monolayer Au coverage on Bi89Sb11 and re-measure ARPES of the S1/S2 dispersions (or equivalent interface-sensitive spectroscopy); if the Dirac-like features and ΔE ≈ 24 meV survive with only minor hybridization shifts, the eDMM input parameters remain valid and the hierarchy claim is strengthened. If the surface bands are strongly gapped or quenched, the TIS-dominance interpretation of the G[001] anomalies would need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest_assumption already isolates the central soft spot: that G[001] anomalies are dominated by electron–electron transmission through TIS (Gtot ≈ Gee2) while residual phonon–phonon and bulk channels remain negligible for the observed non-monotonic T and bias responses. This rests on the hierarchy Gee2 ≫ Gee1, Gbulk ≫ Gee2, and on the persistence of bare-surface ARPES dispersions under Au contact (invoked via Bi2Se3 literature). The paper itself notes residual phonon–phonon cannot be rigorously excluded, and absolute Gee estimates use free parameters (W = 2.5 nm, ARPES masses). No deeper internal inconsistency or unaddressed experimental confound is present; the topology-specific controls (trivial Bi97Sb3 and Al2O3 interlayer) and bulk-property invariance under current already constrain alternative explanations tightly.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports frequency-domain thermoreflectance measurements of interfacial thermal conductance G[001] at Au/Bi1-xSbx junctions. For topological compositions (x = 11% and 13%), G[001] is non-monotonic in temperature (rise, drop near 120 K, rise above 220 K) and shows a reversible, polarity-symmetric peak versus bias current density below 120 K. Both features are absent in a trivial-semimetal control (Au/Bi97Sb3) and an electronically decoupled control (Au/Al2O3/Bi87Sb13). The authors attribute the responses to carrier redistribution between topological interface states (TIS) and bulk bands, driven thermally by Fermi–Dirac broadening and electrically by quasi-Fermi-level shifts plus WKB tunneling into L-band states. ARPES on Bi89Sb11 confirms the expected fivefold surface-state crossings; an electronic diffusive-mismatch model (eDMM) estimates that parabolic TIS pockets dominate Gee over bulk L-band channels. Bulk mobility, thermal conductivity, and junction resistance are shown to be current-independent, supporting an interfacial origin.","tokens_in":20073,"tokens_out":1139,"duration_ms":10463,"significance":"If the interpretation holds, the work supplies direct experimental evidence that topological interface states can dominate heat flow across a metal–TI junction and that this channel can be electrically reconfigured. That combination is rare: most active thermal-control schemes act on bulk conductivity or on structural/chemical interface engineering, whereas here the control variable is interface electronic structure. The experimental design is strong—highly repeatable FDTR phase spectra, relative G precision of ~1%, and three independent controls that all suppress the anomalies—and the ARPES characterization anchors the surface-state picture. The result therefore opens a concrete materials route (larger-gap TIs, gated interfaces, multilayer Au/TI stacks) for solid-state thermal switches compatible with dense electronics.","major_comments":[{"comment":"Results (eDMM hierarchy and Eqs. 1–4) and Materials (eDMM derivation): the central claim that the observed anomalies are TIS-mediated rests on Gtot ≈ Gee2 with Gee2 ≫ Gee1 and Gbulk ≫ Gee2. The absolute Gee estimates (0.26 vs 1.3–14 MW m−2 K−1 at 80 K) use bare-surface ARPES parameters and a literature Bi2Se3 analogy for Au-contact persistence; residual phonon–phonon coupling is acknowledged but not bounded. A quantitative upper bound on Gpp (or a control that isolates it) and a clearer statement of how much of the ~2% T anomaly and the ΔG[001] ≈ 0.12 MW m−2 K−1 bias peak can be carried by residual channels would make the hierarchy load-bearing rather than assumed.","section":null},{"comment":"Results (bias section, WKB estimate): the high-j suppression is ascribed to L-band activation via WKB tunneling with an assumed barrier width W = 2.5 nm taken from Ref. 46. Because P is exponentially sensitive to W, the claimed 21% accessibility at the 80 K peak (and the analogy to the 120 K thermal threshold) is only semi-quantitative. Either a measured or constrained W, or an explicit sensitivity analysis showing that the peak position remains consistent over a plausible W range, is needed before the electrical and thermal activation routes can be presented as quantitatively convergent.","section":null}],"minor_comments":[{"comment":"Fig. 4B caption states G[001] is “~2% larger at T < 120 K than in the 120–220 K range”; the main text should quote the absolute G scale (or the absolute ΔG) so readers can compare with the eDMM estimates of several MW m−2 K−1.","section":null},{"comment":"Materials and Methods: the multilayer thermal-diffusion model parameters (Au κ, C; Bi1-xSbx heat capacity and anisotropy) are fixed from literature; a short table of the numerical values used would improve reproducibility.","section":null},{"comment":"Fig. 6: error bars on the individual G[001](j) points are not shown; given the stated ~1% relative precision, they would help the reader judge the significance of the peak and the subsequent drop.","section":null},{"comment":"Typographical: “fom the dual electron and hole TIS channels” (Results, bias paragraph) should be “from”; “storng sensitivity” in Fig. S5 caption should be “strong”.","section":null},{"comment":"The dedication to J.P. Heremans is appropriate and moving; ensure the corresponding-author list and acknowledgements remain consistent with journal policy on posthumous authorship.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The experimental controls and measurement precision are unusually clean for an interfacial-thermal-transport paper; the two major comments are about tightening the quantitative interpretation rather than about the existence of the effect. I would not require new experiments if the authors can supply a reasoned bound on Gpp and a W-sensitivity plot. Scope is a good fit for a high-impact condensed-matter or applied-physics journal."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The real news here is the first clear experimental isolation of topological interface states as the dominant, electrically reconfigurable channel for cross-plane thermal conductance at a metal/TI junction. Au on Bi89Sb11 and Bi87Sb13 shows non-monotonic G(T) with a drop near 120 K and a sharp rise above 220 K, plus reversible bias-current peaks below 120 K that reverse with polarity. Both vanish in the trivial Bi97Sb3 control and when an Al2O3 spacer is inserted. Bulk mobility, kappa, and junction resistance stay flat under current, so the anomalies are not bulk artifacts. ARPES confirms the fivefold surface crossings and supplies the velocities and masses used later. FDTR phase spectra are rock-solid (sub-0.5 % RSD), relative G changes resolve at ~1 %, and the three independent controls all suppress the effect. That pattern is new and hard to dismiss.\n\nWhat they do well is the experimental design and the honest framing. They treat the eDMM estimates as order-of-magnitude checks rather than fits, note that residual phonon-phonon cannot be rigorously excluded, and keep the absolute G numbers secondary to the relative trends. Citations are standard and appropriate; no padding or missing priors that I can see.\n\nThe soft spot is exactly the one the reader flagged: the hierarchy Gee2 ≫ Gee1 and the assumption that bare-surface ARPES dispersions survive Au contact (borrowed from Bi2Se3 literature). Barrier width W = 2.5 nm and the pocket masses are free parameters, so the absolute Gee numbers are illustrative. That weakens the quantitative story but does not undercut the qualitative claim, because the topology-specific controls already rule out the obvious alternatives. No circularity, no internal contradiction with their own equations.\n\nThis is for anyone working on interfacial thermal transport, topological materials, or active heat management in dense electronics. The data are solid enough that a serious editor should send it to referees rather than desk-reject; expect requests for more interface characterization or a tighter bound on the phonon residual, but the core result deserves the airtime. I would engage with it.","headline":"Clean FDTR data isolate TIS as the electrically tunable channel for interfacial heat flow; modeling assumptions are the softest part but the topology-specific controls carry the claim.","tokens_in":20721,"tokens_out":537,"would_cite":true,"duration_ms":15633,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Current injection reversibly tunes heat flow across gold–topological-insulator junctions by shifting carriers between interface and bulk electronic states.","keywords":["topological interface states","interfacial thermal conductance","Bi1-xSbx","frequency-domain thermoreflectance","active thermal management","electronic structure engineering","Wiedemann–Franz","metal–topological insulator junction"],"falsifier":"Repeat the FDTR temperature and bias sweeps on the same Au/Bi1-xSbx junctions after a surface treatment or overlayer that demonstrably destroys or buries the topological surface dispersion (verified by ARPES); if the non-monotonic G(T) and the peaked G(j) both disappear while bulk transport remains unchanged, the TIS assignment is confirmed; if they survive, it is falsified.","tokens_in":20730,"feed_emoji":"🔥","tokens_out":990,"duration_ms":8818,"temperature":0.7,"pith_summary":"Modern electronics pack so much power into small volumes that heat often bottlenecks performance, and the solid-solid interface is usually the worst thermal bottleneck. This paper shows that when gold is placed on Bi1-xSbx topological insulators, the special electronic states that form at that interface carry a measurable fraction of the heat. Those states can be filled or emptied by temperature (through ordinary Fermi–Dirac broadening) or by a small bias current (through a quasi-Fermi-level shift and tunneling into nearby bulk bands). The result is a non-monotonic temperature dependence of interfacial thermal conductance and a reversible, polarity-symmetric modulation with current density—both of which vanish in control samples that lack topological interface states or that are electronically decoupled by an oxide barrier. If the mechanism is general, interface electronic-structure engineering becomes a new, mechanically passive route to active thermal management inside dense solid-state devices.","feed_headline":"Current tunes heat flow at gold–topological-insulator interfaces","feed_subtitle":"Topological interface states let bias current reversibly open and close a solid-state heat channel","key_machinery":"Topological interface states (TIS) treated within the electronic diffusive-mismatch model (eDMM): heat crosses the junction primarily by electron–electron coupling into the high-density TIS pockets (Gee2), whose occupation is thermally broadened by the Fermi–Dirac distribution and electrically shifted by a quasi-Fermi level of order qV together with WKB tunneling into nearby bulk L-band states.","core_discovery":"The interfacial thermal conductance of Au/Bi89Sb11 and Au/Bi87Sb13 junctions exhibits a non-monotonic temperature dependence and a reversible, polarity-symmetric modulation under bias current; both signatures are absent in trivial-semimetal and Al2O3-decoupled controls and are explained by carrier redistribution between topological interface states and bulk bands.","pith_inferences":["If the eDMM hierarchy holds, similar electrically tunable G should appear at other metal/TI contacts whose surface Dirac or pocket states survive weak hybridization.","The polarity symmetry of G(j) implies that bipolar TIS (electron and hole pockets) are advantageous; monopolar surface states would produce an asymmetric response that could itself be a diagnostic.","Device-scale thermal transistors or diodes could be built by placing the Au/TI junction in series with a fixed-conductance path, converting the G modulation into a binary heat-routing element.","Because the effect is even in current and vanishes at high temperature, it is naturally compatible with pulsed-current thermal management schemes that avoid continuous power dissipation."],"forward_implications":["Interface electronic structure, rather than geometry or chemistry alone, becomes a design variable for active thermal interfaces.","Larger-gap topological insulators should widen the quasi-Fermi-level window before bulk bands activate, increasing the usable modulation range of G.","Electrostatic gating can be combined with current injection to set the zero-bias Fermi level independently of the bias-induced shift.","Multilayer stacks of repeated Au/TI interfaces could amplify the even-in-field conductance response for practical thermal switches.","The same TIS-mediated pathway offers a mechanically robust alternative to phase-change or strain-based thermal control inside dense high-power electronics."],"fun_headline_variants":["Bias current reversibly tunes heat flow across Au–BiSb topological interfaces","Electronic structure engineering electrically controls Au/TI interfacial heat","Topological interface states enable current-driven modulation of thermal conductance","Carrier redistribution lets bias open and close heat channels at Au/BiSb junctions","Fermi-level shifts electrically gate solid-state heat conduction at gold–TI contacts"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The measured conductance changes are assumed to be dominated by electron transmission through the topological interface states, with phonon and bulk channels remaining negligible for the observed anomalies.","fun_headline_variants_meta":{"raw":{"variants":["Bias current reversibly tunes heat flow across Au–BiSb topological interfaces","Electronic structure engineering electrically controls Au/TI interfacial heat","Topological interface states enable current-driven modulation of thermal conductance","Carrier redistribution lets bias open and close heat channels at Au/BiSb junctions","Fermi-level shifts electrically gate solid-state heat conduction at gold–TI contacts"]},"model":"grok-4.5","effort":"low","cost_usd":0.004072,"raw_usage":{"total_tokens":1210,"prompt_tokens":749,"num_sources_used":0,"completion_tokens":96,"cost_in_usd_ticks":40720000,"prompt_tokens_details":{"text_tokens":749,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":365,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":749,"tokens_out":96,"duration_ms":4581,"temperature":1.0,"reasoning_tokens":365,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-12T06:16:04.081504+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Repeat the FDTR temperature and bias sweeps on the same Au/Bi1-xSbx junctions after a surface treatment or overlayer that demonstrably destroys or buries the topological surface dispersion (verified by ARPES); if the non-monotonic G(T) and the peaked G(j) both disappear while bulk transport remains unchanged, the TIS assignment is confirmed; if they survive, it is falsified.","supporting_citations":[],"review_version":1}