{"id":"bda4af53-e54c-4ff8-b774-f7f282985967","arxiv_id":"2509.03469","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Transport measurements show that the beta phase of Bi4I4 conducts along its side surfaces, that this surface conduction switches off in the alpha phase, and that resistance noise near the transition is consistent with fluctuating topological domains.","lead":"Using conductive atomic force microscopy and resistance noise spectroscopy, this experiment reports that Bi4I4 conducts electricity along its side surfaces in the high-temperature beta phase and loses that surface conduction when it transforms to the low-temperature alpha phase. A smart generalist might read it because it is a candidate material whose topological surface transport can be switched on and off by a room-temperature structural transition.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No trivial-material or top-surface-step control for c-AFM edge-current contrast; the beta-phase surface-state transport claim rests on an artifact-prone measurement.","rationale":"I reviewed the full text and SI. The strongest claim—surface-state-mediated transport in beta-Bi4I4 that vanishes in alpha—rests on two experimental legs: c-AFM current maps (Sec. 3) and telegraphic noise (Sec. 4). The most load-bearing assumption is that the c-AFM edge-current contrast in Fig. 2a is intrinsic surface-state conduction. The alpha-phase null is not a sufficient control because the transition changes bulk conductivity and may alter contact. The noise-based inference of alpha hinge modes is plausible but also depends on the c-AFM claim for the beta phase and on a parameter-free-domain-size assumption that is partially circular (SI Sec. S10 inverts Eq. 19 to extract l from the observed fc). However, the direct observation of edge currents is the linchpin. This concern is exactly the reader's weakest_assumption. Since the concern is addressable by a control experiment and does not by itself falsify the reported data, the CONDITIONAL verdict remains appropriate. I recommend UNCHANGED.","tokens_in":22657,"tokens_out":6042,"duration_ms":67653,"concrete_test":"Exfoliate a topologically trivial control flake (e.g., MoS2 or hBN) with a stepped topography similar to the beta-Bi4I4 flake in Fig. 2 and perform identical c-AFM scans (same tip, 1 V bias, scan rate, and temperature). If the current map shows comparable edge enhancement on the trivial control, the beta-Bi4I4 edge contrast is a topographic/contact artifact; if the trivial flake shows no edge enhancement, the surface-state interpretation remains viable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central new experimental evidence that beta-Bi4I4 transports through gapless side-surface states is the c-AFM current map in Fig. 2a,b, where line scans across a ~400-nm flake show ~400 pA at the side-surface steps under 1 V DC bias while the top surface carries near-zero current. The same flake in alpha phase (Fig. 2c,d) shows no edge current, which is a partial control, but the beta-to-alpha transition also changes the bulk resistivity and may alter tip-sample contact, so this is not a clean null for topographic/contact artifacts. No c-AFM measurement is reported on a topologically trivial flake with comparable stepped topography, no top-surface step on the same flake is used as a facet-specific control, and no quantitative comparison of the measured current to the expected surface-state conductance (or to a metallic reference) is provided. Because c-AFM tips are well known to show current enhancement at sharp topographic steps due to field concentration, contact-area changes, and local pressure effects, the possibility remains that the edge contrast in Fig. 2a is a measurement artifact rather than surface-state conduction. This premise enters in Section 3 and underpins the headline claim; if it fails, the paper retains the structural/transition characterization and the noise data, but loses its main direct transport evidence for the topological surface states.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports experimental and theoretical work on the room-temperature β→α transition in the quasi-one-dimensional topological material Bi4I4. Using XRD and resistivity, the authors identify a first-order displacive transition near 300 K. Conductive-AFM current maps on exfoliated flakes show current contrast at side-surface step edges in the β phase, its absence in the α phase, and its reappearance upon reheating. Low-frequency resistance noise measurements near the transition show telegraphic switching and a Lorentzian component in the power spectral density. A φ^3 theory of the displacive transition is coupled to a one-dimensional fermion model, yielding topological edge modes in a parameter window; the McWhorter generation-recombination model is used to relate the Lorentzian corner frequency fc to the α-domain size. The paper concludes that β-Bi4I4 transports through gapless side-surface states and that α-Bi4I4 hosts hinge/edge modes responsible for the noise signature.","tokens_in":22827,"tokens_out":5771,"duration_ms":61402,"significance":"If the claims hold, the paper would provide a rare transport-level signature of a temperature-driven topological surface-state transition in a quasi-one-dimensional material, complementing existing ARPES/STS studies. The strengths are the reproducible c-AFM contrast on several flakes with reversible thermal cycling, the explicit six-surface DFT surface-state calculation (Fig. S5), and the quantitative fitting of the noise spectra as 1/f plus a Lorentzian. However, the two central interpretive steps—the surface-state origin of the c-AFM edge current and the extraction of α-domain sizes from fc—are not yet established to the standard required for the headline claims. The c-AFM measurement lacks controls for topographic/contact artifacts, and the domain-size estimate is circular as presented.","major_comments":[{"comment":"The c-AFM edge-current contrast is the only direct experimental evidence that β-Bi4I4 transports through side-surface states. The α-phase null is a partial control, but the first-order transition changes bulk resistivity and likely tip-sample contact, so it does not exclude topographic/contact artifacts. The paper reports no measurement on a topologically trivial flake with comparable stepped topography, no top-surface step on the same flake, and no comparison of the measured current (~400 pA at 1 V; several nA at 4–5 V in the SI) to the conductance expected from a gapless surface state or to a metallic reference. Because conductive-AFM tips routinely show enhanced current at sharp steps due to field concentration and contact-area changes, this leaves a plausible artifact path. This premise underpins the headline claim; without such controls, the transport conclusion is not robust.","section":"Section 3, Fig. 2a–f and Figs. S6–S7"},{"comment":"The domain size l is estimated by inverting Eq. (19) from the measured Lorentzian corner frequency fc (Fig. S16), and the same l is then used in the main text to state that 'a typical domain of size l∼380 nm leads to fc∼20 mHz which is consistent with experimentally observed values.' This is circular: fc cannot simultaneously fix l and serve as a confirmation of the model. The exponential sensitivity of Eq. (19) to l/ξ makes even order-of-magnitude agreement weak unless l is determined independently (e.g., by structural imaging or scattering) or unless the extraction is explicitly treated as a fit with uncertainty propagation. As written, the noise interpretation is not falsifiable.","section":"SI Section S10, Eq. (19) and Fig. S16; Section 5, last paragraph"},{"comment":"The abstract asserts that the α-phase hosts 'hinge states' and the paper concludes with 'transport signatures of the topological states in both phases.' The α-phase boundary modes are never directly observed: the c-AFM maps show only the absence of edge current in α, and the noise analysis is an indirect, model-dependent interpretation. The effective one-dimensional Hamiltonian in Eq. (6)/(8) supports zero-energy edge modes of a model chain, but this is not a demonstration of the three-dimensional hinge states of α-Bi4I4. Ordinary two-level fluctuators—domain boundaries, charge traps, contact instabilities—can also produce telegraphic resistance noise near a first-order transition. The claims about α-phase hinge states should be explicitly softened unless direct evidence (e.g., spectroscopy or length-dependent transport on α-phase flakes) is provided.","section":"Abstract and Section 5; Conclusion"}],"minor_comments":[{"comment":"Axis labels in several panels are garbled ('4120.40Current (pA)', '0 856 (nm)', etc.); the figures need re-typesetting.","section":"Fig. 2"},{"comment":"There is a numerical inconsistency: the main text says l≈380 nm gives fc≈20 mHz, while the SI says the same l gives fc≈10 mHz. Please harmonize the statement and check the evaluation of Eq. (19).","section":"Section 5 vs. SI Section S10"},{"comment":"The caption says 'current variation (top left)' for a panel that appears to be bottom left; please correct the panel references.","section":"Fig. S6 caption"},{"comment":"No error bars, number of devices, or cooling-cycle comparison is reported for the fc(T) and ⟨δR²⟩/R² plots. At minimum, a statement on run-to-run and sample-to-sample reproducibility is needed.","section":"Fig. 3c,e"},{"comment":"The terms TLF and RTN are used interchangeably; please define them once and use them consistently throughout.","section":"General notation"}],"recommendation":"major_revision","confidential_remarks":"To the editor: I am sympathetic to the system and the experimental effort, but the paper's main new claim rests on a single artifact-prone c-AFM measurement and a circular domain-size estimate. If the authors can supply the missing controls and/or reframe the claims to match the direct evidence, a revised version could be appropriate. The paper likely fits a specialist transport or topological-materials journal; in its present form it is not yet suitable for acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a serious experimental report with a genuinely new transport angle on Bi4I4 — c-AFM current maps showing side-surface conduction in the beta phase, disappearing in the alpha phase, plus telegraphic resistance noise near the 300 K transition. The prior literature (ARPES, STS, theory) is cited honestly; the transport evidence is new.\n\nWhat's good: multiple exfoliated flakes show the side-surface edge current reproducibly, and the same flake loses the edge current in the alpha phase and recovers on reheating. The first-principles surface calculations support the idea that only side surfaces carry gapless states. The effective phi3-plus-fermion theory is a nice toy model connecting a first-order displacive transition to a topological change, though schematic.\n\nWhere it's soft: the c-AFM edge current is the load-bearing evidence, and there is no control for the known topographic field-enhancement artifact. c-AFM tips routinely show enhanced current at sharp steps simply from field concentration and tip contact, and the beta-to-alpha comparison is not a clean null because the bulk resistivity and tip-sample contact change across the transition. No trivial flake, no top-surface step on the same flake, and no comparison of measured current to expected surface-state conductance (or a metallic reference) is provided. At 1–5 V bias, contact effects can dominate. So the central transport claim is plausible but not airtight.\n\nThe noise analysis is partially circular: Eq. 19 is inverted in the SI to extract alpha-domain size from the measured fc, so the 'consistent with observed values' line in the main text is a parameter consistency check with one free parameter, not an independent prediction. That's not fatal, but it should be labeled as such. The alpha-phase hinge states asserted in the abstract are not directly imaged; the telegraphic noise is consistent but not conclusive. And the theory's topological regime is stated in dimensionless units; the mapping to the physical |lambda phi| ~ 15 meV is plausible but not fully developed.\n\nBottom line: this is a paper a serious referee should see. The experiment is timely, the authors clearly did a lot of work, and the concern is a missing control, not an obvious internal contradiction. A competent referee can ask for the controls and for clearer language on prediction vs fit. I would not cite it as established transport evidence yet, but I would bring it to reading group to debate how much weight the current maps should carry.\n\nRecommendation: send to peer review, conditional on requests for artifact controls and explicit acknowledgment of the extraction step.","headline":"Promising transport evidence for switchable surface states in Bi4I4, but the c-AFM edge current needs a real artifact control before the headline claim can be trusted.","tokens_in":23550,"tokens_out":3524,"would_cite":false,"duration_ms":37137,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["72.70.+m","73.20.At","71.70.Ej"],"model":"deepseek-v4-flash","headline":"Bi4I4 crosses a displacive topological transition near 300 K: the β phase's gapless surface states carry current, the α phase's surfaces gap out, and telegraphic noise near the switch traces tunneling through α-domain edge modes.","keywords":["Bi4I4","topological phase transition","displacive structural transition","weak topological insulator","surface states","conductive atomic force microscopy","resistance noise spectroscopy","telegraphic noise"],"falsifier":"Cool the same exfoliated flake into the α phase and scan a region with intentionally steep, freshly cleaved steps under identical bias: if the edge-current contrast persists at comparable magnitude, the β-phase contrast is a topographic tip artifact rather than surface-state transport. Alternatively, a two-terminal measurement of a single side surface in the β phase should show a finite, roughly temperature-independent conductance of order e²/h per surface that drops abruptly and reproducibly near 300 K; if the conductance change tracks bulk resistivity instead, the edge current is not carryin","tokens_in":22402,"feed_emoji":"⚛️","tokens_out":12139,"duration_ms":117914,"temperature":0.7,"pith_summary":"Bi4I4 is a quasi-one-dimensional crystal that changes its electron topology along with its lattice near room temperature: in the high-temperature β phase, current flows through gapless surface states on the crystal's side faces, while in the low-temperature α phase those surfaces go gapped and conduction shifts to localized hinge channels. The paper claims direct transport evidence for this switch. Conductive atomic force microscopy maps show bright current along side-surface step edges in the β phase, vanishing when the sample cools into the α phase and reappearing when it is reheated. Resistance-fluctuation spectroscopy near the transition reveals two-level telegraphic noise with a characteristic corner frequency around 20 mHz, which the authors attribute to electrons tunneling between β domains through the edge modes of intervening α domains. A minimal theory coupling a phonon order parameter to the topological edge electrons reproduces both the first-order displacive character of the transition and the noise frequency scale, marking Bi4I4 as a candidate where a classical structural transition drives a topological one.","feed_headline":"At 300 K, Bi4I4 switches from surface to hinge conduction","feed_subtitle":"Current maps and noise spectra tie the switch to a reversible displacive topological transition at room temperature.","key_machinery":"The load-bearing object is the scalar order parameter φ of a φ³ free energy, L = (a/2)φ² − (1/3)φ³ + (1/4)φ⁴ with a ∝ (T − TS), which describes the first-order displacive lattice transition between the β phase (⟨φ⟩ = 0) and the α phase (⟨φ⟩ ≠ 0). The low-energy electrons are treated as quantum spin Hall edge fermions with Hamiltonian H = ħvF ky σx ⊗ τz + λφ(1 ⊗ τz + 1 ⊗ τx): the φ coupling gaps the surface Dirac cone while preserving a topological regime for −2 < λφ < 0, characterized by a quantized spin winding number. The two-level fluctuations are carried by a generation-recombination trap mechanism: edge modes of α domains act as trap centers, and their energy splitting ΔE_edge = W exp(−","core_discovery":"The central claim is that the room-temperature structural change in Bi4I4 is a displacive topological phase transition, not just a lattice change: below roughly 300 K the α phase is gapped on its surfaces and hosts localized hinge (edge) modes, while above it the β phase supports gapless Dirac surface states that actually conduct. The evidence is twofold. First, c-AFM current maps of exfoliated flakes show strong conduction at side-surface step edges in the β phase (about 304 K), which disappears in the α phase (about 292 K) and returns on reheating, with the same hysteresis seen in resistivity. Second, normalized resistance-noise spectra near the transition acquire a Lorentzian component on","pith_inferences":["If the surface-state reading of the c-AFM maps is correct, a suitably contacted β-phase flake should show a nearly quantized two-terminal conductance (of order e²/h per surface) that drops abruptly at the transition; that measurement would turn local imaging into a transport number.","A decisive control would be the same c-AFM scan on a topologically trivial flake with comparable step morphology, or on the same flake in the α phase with deliberately steep edges: persistent edge contrast would point to a topographic or contact artifact rather than surface-state transport. The paper reports no such control.","The noise signature offers a practical probe of phase coexistence: monitoring fc across the hysteresis loop could map α-domain sizes as a function of sweep rate and thermal history, using edge modes as a local thermometer of the displacive transition.","The same φ-coupled-edge-fermion mechanism should apply to isostructural bismuth halides such as Bi4Br4 and Bi4Br2I2, so the predicted telegraphic-noise peak and its fc-versus-temperature curve are testable predictions for the broader material family."],"forward_implications":["The β phase of Bi4I4 transports electricity on its side (100)-type surfaces through gapless topological states; the same flakes show no such edge current in the α phase, so the current map is a direct readout of the surface band topology.","The β↔α transition is first-order and displacive, with roughly 10 K of resistivity hysteresis; the surface conduction reappears after repeated thermal cycling, so the effect is intrinsic and reversible.","Near the transition, resistance fluctuations carry a Lorentzian two-level component whose corner frequency encodes the domain size; the observed fc of about 20 mHz implies α domains of order 380 nm interleaved with β domains.","Because the surface gap in the α phase is only about 25 meV while the bulk gaps are 60–130 meV, ambient-temperature transport in Bi4I4 is dominated by surface and hinge electrons, making the material a workable platform for room-temperature topological surface conduction.","The same theory predicts that the α phase hosts topological hinge modes (spin winding number 1) even though its surfaces are gapped, so one displacive transition connects a weak topological insulator (β) to a higher-order topological insulator (α)."],"supporting_citations":[{"why":"Prior prediction of a room-temperature topological phase transition in Bi4I4 and of the ~35 meV surface gap of the α phase; defines the phenomenon under test.","marker":"[23]"},{"why":"Earlier identification of β-Bi4I4 as a weak topological insulator whose side surfaces host gapless states; grounds the expectation that β edge conduction should be observable.","marker":"[25]"},{"why":"Original classification of β-Bi4I4 as a quasi-one-dimensional topological insulator and the crystal-growth recipe the samples follow.","marker":"[29]"},{"why":"Characterization of the β↔α transition and of the α phase's higher-order (hinge) topology; basis for treating α-domain edge modes as the fluctuating channels.","marker":"[28]"},{"why":"The φ³ free-energy theory of first-order displacive structural transitions used to model the order parameter across the transition.","marker":"[31]"},{"why":"The generation-recombination noise model that converts two-level trap fluctuations into a Lorentzian spectrum; supplies the mechanism linking edge modes to the observed noise.","marker":"[51]"},{"why":"The Green's-function surface-state calculation tool used to compute the gapless β (100) and gapped α (¯201) surface spectra.","marker":"[32]"}],"fun_headline_variants":["Bi4I4 flips from surface to hinge conduction at 300 K","Room-temp topological flip: Bi4I4 surface states become hinge states","Bi4I4 surface conduction gives way to hinge modes at 300 K","Bi4I4's 300 K switch: surface states yield to hinge states"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The bright current the c-AFM tip records at side-surface step edges in the β phase is assumed to be genuine topological surface-state conduction, although sharp steps are known to enhance tip current for purely topographic and contact reasons, and no control on a topologically trivial sample or quantitative conductance check is reported.","fun_headline_variants_meta":{"raw":{"variants":["Bi4I4 flips from surface to hinge conduction at 300 K","Room-temp topological flip: Bi4I4 surface states become hinge states","Bi4I4 surface conduction gives way to hinge modes at 300 K","Bi4I4's 300 K switch: surface states yield to hinge states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000937,"raw_usage":{"total_tokens":3854,"prompt_tokens":764,"completion_tokens":3090,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":508,"completion_tokens_details":{"reasoning_tokens":3015}},"tokens_in":508,"tokens_out":3090,"duration_ms":21609,"temperature":1.0,"reasoning_tokens":3015,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T10:54:27.824129+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cool the same exfoliated flake into the α phase and scan a region with intentionally steep, freshly cleaved steps under identical bias: if the edge-current contrast persists at comparable magnitude, the β-phase contrast is a topographic tip artifact rather than surface-state transport. Alternatively, a two-terminal measurement of a single side surface in the β phase should show a finite, roughly temperature-independent conductance of order e²/h per surface that drops abruptly and reproducibly near 300 K; if the conductance change tracks bulk resistivity instead, the edge current is not carryin","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier identification of β-Bi4I4 as a weak topological insulator whose side surfaces host gapless states; grounds the expectation that β edge conduction should be observable."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Original classification of β-Bi4I4 as a quasi-one-dimensional topological insulator and the crystal-growth recipe the samples follow."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Characterization of the β↔α transition and of the α phase's higher-order (hinge) topology; basis for treating α-domain edge modes as the fluctuating channels."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The φ³ free-energy theory of first-order displacive structural transitions used to model the order parameter across the transition."},{"cited_title":"& Soluyanov, A","cited_arxiv_id":null,"evidence_quote":"The Green's-function surface-state calculation tool used to compute the gapless β (100) and gapped α (¯201) surface spectra."}],"review_version":1}