{"id":"5153d32c-9c8d-4023-96fd-d07680221ca7","arxiv_id":"2507.23014","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Sb thin films show thickness-dependent band shifts in ARPES and transport, supporting the predicted quantum-confinement-driven topological phase transition.","lead":"Antimony films five to thirteen nanometers thick show electronic changes as they get thinner: the conduction band moves up, consistent with an insulating gap opening. The study combines transport and photoemission to test a predicted transition from a topological semimetal to a topological insulator.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The ARPES evidence for the central band-shift claim is an uncontrolled absence comparison: one 15 nm film shows an M-point feature, one 6 nm film does not, with no quantitative shift, buffer-only control, or thickness series.","rationale":"I read the paper in good faith: the authors have grown Sb films and combined transport and ARPES, which is a reasonable approach. The most direct evidence for the central claim is the M-point ARPES comparison in Section III.C. That comparison is the least secure step because it infers an energy shift from the presence/absence of a feature in only two samples, without quantifying the shift or ruling out alternative causes. I do not claim the conclusion is false; I claim the current data are insufficient to establish it. The reader's verdict (CONDITIONAL) already captures this, so I recommend no change in verdict. The proposed ARPES thickness-series experiment with EDC fits and a buffer-only reference would settle whether the concern lands. I also note the transport carrier densities are 2D sheet densities and should be thickness-normalized if used as independent evidence.","tokens_in":12639,"tokens_out":4904,"duration_ms":63002,"concrete_test":"Repeat ARPES on a thickness series (e.g., 6, 9, 12, 15 nm) grown on identical (Bi,Sb)2Te3/sapphire buffers, including a buffer-only control sample. At the M point, collect high-statistics EDCs and fit the conduction-band peak position relative to a measured Fermi-level reference (e.g., a gold foil in electrical contact). If the M-point peak shifts monotonically upward by more than the energy resolution (6 meV) with decreasing thickness, and the buffer-only control shows no M-point intensity, the claim is supported. If the feature merely weakens without shifting, or the buffer control contributes M-point intensity, the claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that reducing Sb thickness shifts the conduction band upward and the valence band downward—rests primarily on ARPES Section III.C. The only ARPES evidence is that an M-point feature assigned to the bulk conduction band is visible in the 15 nm film and absent in the 6 nm film (Figs. 4(c)-(d) vs 4(g)-(h)). Absence is not a measured shift. No energy-distribution curves, peak binding energies, or Fermi-level reference are provided, and the two films differ only at one intermediate thickness on each side of the predicted transition. The disappearance could instead be caused by thickness-dependent Fermi-level pinning, reduced photoemission cross-section, a matrix-element change, degraded crystallinity in the 6 nm film, or a contribution from the (Bi,Sb)2Te3 buffer that is thickness-dependent. Fig. 5's valence-band downward shift is a schematic, not data. The transport samples on GaSb(111) with a 20 nm GaSb buffer and seed layer are a different heterostructure, so they do not independently confirm the ARPES assignment; moreover, the Hall 'carrier concentrations' are 2D sheet densities from Eq. S1 and are not normalized by film thickness, making the observed decrease partly geometric. Thus the current data do not establish the central band-shift claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an experimental study of Sb thin films grown by molecular beam epitaxy on GaSb(111)B and on (Bi,Sb)2Te3/sapphire substrates, combining electrical transport measurements (Hall effect, magnetoresistance, weak antilocalization) with angle-resolved photoemission spectroscopy (ARPES). The central claim, stated in the Conclusions, is that reducing the Sb film thickness shifts the conduction band upward and the valence band downward, consistent with a quantum-confinement-driven approach to a topological insulator state. The evidence includes ARPES comparison of 15 nm and 6 nm films, where an M-point feature attributed to the bulk conduction band in the thicker film is absent in the thinner film, and transport data showing decreasing bulk carrier concentrations with thickness, metallic-to-insulating crossover in R(T), and thickness-dependent weak antilocalization.","tokens_in":12943,"tokens_out":3178,"duration_ms":39560,"significance":"If the central claim is correct, the work would provide a notable experimental step towards realizing the predicted topological transition in elemental Sb thin films, a material of interest for spintronics and hybrid devices. The manuscript has clear strengths: systematic MBE growth over a thickness range, structural characterization by XRD and AFM, transport measurements across multiple samples, and a direct ARPES comparison of two thicknesses. The WAL measurements and analysis of the phase coherence length are also competently executed. However, the significance is currently limited by the qualitative nature of the ARPES evidence and by the lack of quantitative error analysis for the transport fits.","major_comments":[{"comment":"The central claim that reducing the Sb thickness lifts the conduction band at the M-point is based on an uncontrolled absence comparison: one 15 nm film shows an M-point feature and one 6 nm film does not. No energy-distribution curves, peak binding energies, or Fermi-level reference are provided, so no quantitative band shift is extracted. The absence of the feature could equally result from thickness-dependent Fermi-level pinning, a matrix-element change, reduced photoemission cross-section, or degraded crystallinity in the 6 nm film. A buffer-only control, a thickness series with at least three thicknesses, or a measured binding-energy shift would be needed to support the claimed upward shift.","section":"Section III.C, Figs. 4(c)-(h)"},{"comment":"The transport samples (on GaSb(111)B with a 20 nm GaSb buffer and 2 nm seed layer) and the ARPES samples (on sapphire with a 2 nm (Bi,Sb)2Te3 buffer) are different heterostructures, so the transport trends do not independently corroborate the ARPES band-shift assignment. The thinnest transport film is 5.1 nm and the thickest is 13.2 nm, while the ARPES films are 6 nm and 15 nm, so the thickness ranges barely overlap and the buffer layers differ. The paper should either measure ARPES on the same heterostructure used for transport or explicitly justify the assumption that the band shifts are independent of the buffer and substrate.","section":"Section II and Table 1"},{"comment":"The three-band Hall model has six free parameters (three carrier densities and three mobilities), is fitted only in a low-field regime, and no error bars, fitting residuals, or uniqueness analysis are provided. As a result, the reported decrease of the bulk carrier concentrations with thickness is not established quantitatively. Additionally, the carrier concentrations extracted from Eq. (S1) are two-dimensional sheet densities, not normalized by film thickness; since the sheet density naturally scales with thickness for a fixed volumetric density, the observed decrease is partly geometric. The paper should show thickness-normalized densities or otherwise correct for this trivial thickness dependence.","section":"Section III.B and Supplementary S2, Eq. (S1)"},{"comment":"The WAL fits use a 'suitable magnetic field range' chosen per sample without an objective criterion, which can bias the extracted values of alpha and l_phi. This matters because the paper uses alpha ≈ -0.5 to infer a single coherent conducting channel and the l_phi(T) exponent beta to infer a change in the phase decoherence mechanism. The manuscript should state the fitting range for each sample or use a fixed field range, and report the confidence intervals for alpha and l_phi.","section":"Section III.D, Fig. 6"}],"minor_comments":[{"comment":"The power-law exponents gamma for the MR curvature are listed without uncertainties; given that the central transport trend is a decrease from 1.38 to 0.57, error bars or at least a fit-quality statement would help assess whether the change is significant.","section":"Section III.B"},{"comment":"Figure 5 is described as a 'schematic' reconstruction of the band dispersion, but the caption and text could more explicitly warn readers that this is not measured data; the current phrasing risks being read as an ARPES result.","section":"Section III.C, Fig. 5"},{"comment":"The temperature-dependence exponent is written inconsistently as both 'β' and 'ß' (e.g., 'lØ~𝑇−𝛽' in the text and 'ß' in Fig. 6(d) and Fig. S4); the manuscript should use a single symbol.","section":"Section III.D"},{"comment":"The text refers to 'GaSb(111)A' when the main text uses 'GaSb(111)B'; also, the XRR fit in Fig. S1 gives a thickness of 13.1 nm while the main text and Table 1 quote 13.2 nm. These inconsistencies should be reconciled.","section":"Supplementary S1"},{"comment":"The statement in the abstract and main text that the MR curvature transitions 'from quadratic to linear' is imprecise because the reported gamma values range from 1.38 to 0.57, i.e., from super-linear to sub-linear; the wording should match the data.","section":"Section III.B"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concerns are well placed and align with my own reading. The central band-shift claim rests on a qualitative ARPES absence comparison and on transport data from a different heterostructure with under-constrained fits. These are fixable in a revision—for instance, by adding quantitative ARPES analysis, a buffer control, or a thickness series—so I would not reject the paper outright. The topic is within the scope of the journal, and the experimental growth effort is valuable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on the Sb thin-film paper.\n\nThe genuinely new part is the systematic MBE growth and characterization of elemental Sb films between 5.1 and 13.2 nm, with transport and ARPES on the same material system. That is useful. The transport data show a coherent trend: sheet resistance becomes more insulating, Hall data indicate multiple carriers and lower overall carrier density, WAL appears with alpha near -0.5, and l_phi drops with thickness. If the band-structure interpretation holds, this is a nice confirmation of Zhang et al.'s 2012 prediction and a practical platform for Bi1-xSbx.\n\nThe soft spots are real and they sit exactly where the central claim lives. The ARPES evidence is one 15 nm film and one 6 nm film on (Bi,Sb)2Te3/sapphire. The M-point bulk conduction band feature is present in the thick film and absent in the thin film. That is an absence, not a measured shift. There are no EDCs, no quoted binding energies, no Fermi-level reference, and no buffer-only control. Absence could come from matrix-element changes, Fermi-level pinning, or the buffer, and Fig. 5 is a schematic, not data. So saying 'ARPES confirmed upward shift' overstates what the data show.\n\nThe transport analysis has its own gaps. The 3-band Hall fit has six free parameters with no error bars, no uniqueness check, and the extracted carrier densities are 2D sheet densities not normalized by thickness. Since thickness drops by more than a factor of two across the series, a constant volumetric carrier density would look like a large decrease in sheet density. That weakens the 'decreasing carrier concentration' evidence. The HLN fit range is chosen per sample as 'suitable'; that is common, but makes the alpha values less convincing. The MR curvature change (gamma from 1.38 to 0.57) is real, though the abstract calls it quadratic to linear, which is fair for the thickest versus middle but not for the thinnest.\n\nI don't think there is a load-bearing flaw. The paper is honest about bulk conduction interference, and the conclusion is more strongly worded than the evidence supports. A serious referee should send it out and ask for stronger ARPES (thickness series, quantitative shifts, buffer control) and thickness-normalized Hall densities with error analysis, rather than desk-reject. It is a worthwhile experimental result that needs discipline before it can confirm the predicted transition.","headline":"New MBE transport/ARPES data on Sb films, but the central band-shift claim rests on an uncontrolled absence comparison and needs major strengthening before it can be taken as confirmation.","tokens_in":13521,"tokens_out":2311,"would_cite":false,"duration_ms":27595,"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":"This paper claims that thinning Sb films from 15 nm to 6 nm raises the conduction band and lowers the valence band, so quantum confinement pushes Sb toward a topological insulator.","keywords":["antimony thin films","quantum confinement","topological phase transition","weak antilocalization","ARPES","molecular beam epitaxy","semimetal-to-topological-insulator transition"],"falsifier":"Take the same 6 nm and 15 nm films and measure the M-point conduction band with higher photon flux, different photon energies, and clean surfaces; if the conduction band still appears below the Fermi level in the 6 nm film, the upward-shift claim fails. A complementary test: gate a 6 nm film through the band gap and look for the Hall conductivity to reach a minimum or plateau; absence of any gap-like gate response would also undercut the claim.","tokens_in":12454,"feed_emoji":"⚛️","tokens_out":7150,"duration_ms":77687,"temperature":0.7,"pith_summary":"The paper tries to establish that quantum confinement, rather than alloying or defects, is what changes the band structure as Sb films get thinner. It concludes that reducing the thickness moves the conduction band upward and the valence band downward, opening a band gap and driving Sb toward the predicted semimetal-to-topological-insulator transition. The authors combine two experiments: Hall and magnetoresistance transport show fewer bulk carriers and weak antilocalization as thickness decreases, and ARPES shows a bulk conduction-band feature at the M-point in the 15 nm film that is absent in the 6 nm film. If the claim holds, elemental Sb becomes a practical thin-film platform for topological surface states without the stoichiometry issues of alloys.","feed_headline":"As Sb films thin, conduction and valence bands move apart","feed_subtitle":"ARPES and magnetotransport show quantum confinement lifts the conduction band in the 6 nm film.","key_machinery":"The mechanism is the quantum confinement effect in a semimetal: because electron and hole effective masses have opposite signs, shrinking the film raises the conduction band and lowers the valence band, which increases both the direct gap at Gamma and the indirect overlap gap and eventually turns the semimetal into a three-dimensional topological insulator below about 7.8 nm. Operationally, the paper leans on three probes: three-band Hall fitting to separate surface-state, bulk-electron, and bulk-hole channels; Hikami-Larkin-Nagaoka analysis of weak antilocalization to extract the prefactor alpha and the phase coherence length; and ARPES at the M-point to track the conduction band's position relative to the Fermi level.","core_discovery":"The central claim is that the thickness trend in Sb films is a genuine confinement-driven band rearrangement: the conduction band shifts upward, the valence band shifts downward, and the indirect gap opens as the film approaches a few nanometres, turning the semimetal into a three-dimensional topological insulator with an insulating bulk and spin-momentum-locked surface states. The load-bearing observation is the ARPES comparison between 15 nm and 6 nm films on (Bi,Sb)2Te3/sapphire: the 15 nm film shows a bulk conduction band feature at the M-point below the Fermi level, and that feature is absent in the 6 nm film. Transport is offered as a consistent companion: three-band Hall fits give two electron and one hole bulk/surface channels whose bulk densities fall with thickness, the resistance turns from metallic to insulating at 5.1 nm, and weak antilocalization with an alpha near -0.5 persists below 16 K. The paper's own summary is that reducing the Sb thickness leads to upward and downward shifts of conduction and valence bands, respectively, establishing a foundation for realizing the predicted topological phase transition.","pith_inferences":["A quantitative test the authors did not perform: extracting the conduction-band binding energy at the M-point from ARPES as a function of thickness would separate a true upward shift from Fermi-level pinning or reduced photoemission intensity.","The transport and ARPES films sit on different substrates (GaSb versus sapphire with a (Bi,Sb)2Te3 buffer); comparing same-thickness films on both substrates would test whether the observed gap opening is intrinsic to Sb or partly substrate-driven.","If the confinement picture is right, the topological index should be preserved across the transition, so thin Sb should show surface states without a bulk gap-closing-reopening; this distinguishes the mechanism from a Chern-like transition.","A direct spin-resolved ARPES or spin-torque experiment on gated thin films would extend the paper's claim from band shifts to the helical spin texture that the topological phase requires."],"forward_implications":["If the band shift is real, ultrathin Sb films should become bulk-insulating with only surface conduction when the Fermi level sits in the gap.","The predicted transition thickness is inside the measured range, so thickness becomes a control knob for switching Sb between topological-semimetal and topological-insulator behaviour.","Electrostatic gating of a thin Sb film should suppress the bulk Hall channels and reveal the topological surface states that the weak-antilocalization signal points to.","Because Sb is elemental, this route avoids the compositional disorder of Bi1-xSbx and gives a cleaner test bed for confinement-driven topology.","The same thickness-tuning logic can be applied to Bi1-xSbx, where composition and thickness could be adjusted independently."],"supporting_citations":[{"why":"Supplies the theoretical prediction that quantum confinement in Sb(111) nanofilms shifts the bands, opens a gap, and drives a transition from topological semimetal to topological insulator below about 7.8 nm.","marker":"[7]"},{"why":"Establishes that in thick Sb the conduction band extends below the Fermi level at the M-point, so the 15 nm ARPES feature is identified as the bulk conduction band.","marker":"[30]"},{"why":"Identify the two V-shaped surface states of Sb in ARPES, letting the authors separate surface from bulk features.","marker":"[28,29]"},{"why":"Provides the prior transport study of ultrathin Sb that saw surface-state transport but could not confirm the topological transition, defining the gap this paper aims to fill.","marker":"[22]"},{"why":"Gives the thickness-scaling framework for weak-antilocalization phase coherence length, used to interpret the sublinear lØ versus thickness as coexistence of bulk bands and topological surface states.","marker":"[37]"},{"why":"Supplies the three-band Hall model used to extract carrier concentrations and mobilities for the surface-state, conduction-band, and valence-band channels.","marker":"[27]"},{"why":"Supplies the Hikami-Larkin-Nagaoka formula used to fit the weak-antilocalization magnetoconductance and extract alpha and the phase coherence length.","marker":"[34,35]"}],"fun_headline_variants":["Thinning Sb films opens a gap, hinting at topological insulator","Quantum confinement lifts Sb conduction band as films thin","Sb thin films turn insulating as gap opens from confinement","Thickness-driven band gap in Sb films hints at topological phase"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the 6 nm and 15 nm films differ only in Sb thickness, so the vanishing of the M-point conduction-band feature means the band moved up, not that the signal faded, the Fermi level pinned, or the buffer contaminated the sample, and that the GaSb transport films follow the same physics as the sapphire ARPES films.","fun_headline_variants_meta":{"raw":{"variants":["Thinning Sb films opens a gap, hinting at topological insulator","Quantum confinement lifts Sb conduction band as films thin","Sb thin films turn insulating as gap opens from confinement","Thickness-driven band gap in Sb films hints at topological phase"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000922,"raw_usage":{"total_tokens":3992,"prompt_tokens":1025,"completion_tokens":2967,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":641,"completion_tokens_details":{"reasoning_tokens":2900}},"tokens_in":641,"tokens_out":2967,"duration_ms":24381,"temperature":1.0,"reasoning_tokens":2900,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T11:07:50.904088+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same 6 nm and 15 nm films and measure the M-point conduction band with higher photon flux, different photon energies, and clean surfaces; if the conduction band still appears below the Fermi level in the 6 nm film, the upward-shift claim fails. A complementary test: gate a 6 nm film through the band gap and look for the Hall conductivity to reach a minimum or plateau; absence of any gap-like gate response would also undercut the claim.","supporting_citations":[],"review_version":1}