{"id":"e79118de-18e4-4045-b01a-0c8304a34217","arxiv_id":"2505.03894","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In charge-neutral Sb-doped MnBi6Te10, time-resolved ARPES shows multiple surface photovoltage shifts from adjacent one-unit-cell terraces, revealing intra-unit-cell charge redistribution that prevents QAHE via chemical doping.","lead":"Ultrafast photoemission reveals two different electric potential shifts on the same crystal surface of Sb-doped MnBi6Te10, exposing an internal redistribution of charge that ordinary doping cannot fix. The result explains why chemical doping alone has failed to turn this magnetic topological insulator into a quantum anomalous Hall system.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Dual-SPV interpretation depends on an unverified one-unit-cell vertical-offset geometry; trARPES cannot spatially resolve the two domains.","rationale":"The reader's weakest assumption correctly identifies the unverified domain geometry as the load-bearing step in the argument. My independent reading of the full text confirms this: the paper explicitly concedes that trARPES cannot determine the vertical offset, and the retreat to cleaving statistics is not a measurement. The 24×25 µm2 probe size and the static-ARPES linewidth analysis demonstrate that multiple domains contribute to the spectra, so lateral inhomogeneity is a genuine alternative. The dual-SPV effect is a real and interesting observation, and the paper's supporting evidence (SPV saturation, consistent termination-dependent electron doping, and the triple-SPV observation) makes the observation credible; however, the interpretation as intra-unit-cell charge redistribution relies on a single geometric scenario that is not directly tested. This does not warrant rejection, because the proposed model is plausible and testable, and the experimental observation stands regardless. A conditional verdict is appropriate: the paper should be accepted only if the domain geometry can be corroborated by spatial characterization. I did not find a more serious internal inconsistency, and I do not see evidence of fraud or overreach beyond the acknowledged inference. The concrete test proposed above would settle whether the concern actually lands.","tokens_in":9745,"tokens_out":6392,"duration_ms":71685,"concrete_test":"Map the step-edge topography of the exact cleaved surface with STM or AFM, and repeat trARPES with a probe small enough to fit on a single flat 1-BT terrace (e.g., <5 µm). The decisive check: if the two-replica splitting at -1 ps disappears when the probe is centered on a terrace with no step edge, and reappears only when a ~3.3 nm step is within the probe area, the intra-unit-cell interpretation is supported. If the splitting persists on a single flat terrace, it arises from lateral inhomogeneity or an experimental artifact, and the central conclusion does not follow.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central inference—that the two SPV replicas on the 1-BT termination prove intra-unit-cell charge redistribution—requires that the two bands originate from adjacent 1-BT domains separated by exactly one MnBi6Te10 unit cell (~3.3 nm). The paper itself states that 'the trARPES experiment alone cannot determine the vertical offset,' and it falls back on cleaving statistics: most step edges are one MBT (1.3 nm) or one BT (1 nm) layer, a one-unit-cell step is 'already rare,' and bunched steps are 'extremely difficult to obtain.' This is an unmeasured inference, not a constraint from the data. The trARPES probe covers a 24×25 µm2 area containing many terraces, so the observed splitting is a spatial average. Static ARPES MDC fits already indicate 'superposition of multiple domains with slight angular misalignments' (static linewidth 0.028 Å-1 versus 0.023/0.026 Å-1 for the split peaks), so multiple domains are certainly present. Different domains could have different local Sb concentrations, anti-site defect densities, or surface band bending even with the same termination and zero vertical offset; such lateral inhomogeneity would also produce two different SPV shifts under a pump beam whose intensity profile is not perfectly flat. The model further assumes the two contributing regions are electrically connected through the topological surface state, but no step-height or spatial measurement establishes this geometry. The triple-SPV observation on the MBT termination (Fig. S5) is interpreted with the same unverified domain-offset assumption. Therefore, the jump from 'multiple SPV shifts' to 'spontaneous electrical polarization within a unit cell' is underdetermined: a static potential difference between domains separated by one unit cell is a sufficient explanation, but not the only one consistent with the data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports static ARPES and trARPES measurements on 18% Sb-doped MnBi6Te10, a magnetic topological insulator that Hall measurements show to be near global charge neutrality. Static ARPES finds that the MBT and 1-BT terminations remain electron-doped, while the unobserved 2-BT termination is inferred to be hole-doped. Time-resolved measurements reveal a positive surface photovoltage on the MBT termination and, on the 1-BT termination, a time-dependent splitting of the band structure into two replicas with different SPV shifts and a small momentum offset. The authors interpret this dual-SPV effect as evidence that adjacent 1-BT domains with a one-unit-cell vertical offset experience different photovoltages because of intra-unit-cell charge redistribution, effectively forming an n-p junction within each unit cell. The paper concludes that chemical doping alone cannot place the Fermi level in the magnetic gap needed for the quantum anomalous Hall effect and suggests that the light-induced polarization may be useful for optoelectronics and spintronics.","tokens_in":10154,"tokens_out":4900,"duration_ms":52330,"significance":"If the interpretation holds, the result is significant: it provides a spectroscopic signature of intra-unit-cell charge redistribution in a nominally charge-neutral magnetic topological insulator, directly linking microscopic charge inhomogeneity to the difficulty of realizing the QAHE by doping. The paper has clear strengths: the fluence-saturation behavior of the SPV is a well-controlled check, the static and time-resolved data are presented in detail, the analysis introduces no free fitted parameters, and the observation of multiple time-dependent band replicas is an interesting experimental finding in its own right. The central claim, however, rests on an acknowledged and unverified assumption about the vertical offset of adjacent domains, and the current evidence does not exclude lateral inhomogeneity as an alternative source of the multiple SPV shifts. The result is therefore best viewed as a suggestive observation that needs additional spatial or height-sensitive measurements before the intra-unit-cell interpretation can be considered established.","major_comments":[{"comment":"The central inference that the dual-SPV effect proves intra-unit-cell charge redistribution relies on the assumption that the two 1-BT domains probed by the 24×25 µm2 beam are vertically separated by exactly one unit cell. The authors explicitly state that 'the trARPES experiment alone cannot determine the vertical offset' and then infer the one-unit-cell offset from cleaving statistics, while also noting that a one-unit-cell step is 'already rare.' This is not a measured constraint, and the statistics do not establish that the particular domains contributing to the split spectra have that offset. Under the same pump beam, lateral variations in Sb concentration, anti-site defect density, or surface band bending across domains with the same termination and zero vertical offset would also produce two different SPV shifts. To support the central claim, the authors need direct spatial or step-height information, such as micro-ARPES or STM, or they must explicitly soften the conclusion to 'consistent with' rather than 'evidence of.'","section":"Results, Fig. 3 and accompanying text"},{"comment":"The Discussion states that there is no reason for a global doping asymmetry between the two BT layers and that the asymmetry 'occurs stochastically.' This admission is in tension with the observation of a sharply defined ~90 meV splitting into two discrete replicas. If the intra-unit-cell asymmetry is stochastic, one would expect a distribution of built-in potentials and hence a distribution of SPV shifts, not two well-defined replicas reproducible across the probed area. The authors should either provide a mechanism that selects a preferred orientation of the intra-unit-cell dipole or explain why only two discrete values appear despite the stated stochasticity.","section":"Discussion, paragraph on stochastic doping asymmetry"},{"comment":"The triple-SPV effect on an MBT termination is presented as supporting evidence for the multi-SPV scenario, but the same vertical-offset ambiguity applies. Three replicas with roughly equal energy spacing are consistent with one-unit-cell offsets between adjacent MBT domains, yet the measurement again averages over a large area and provides no spatial resolution. Without information on the lateral positions and step heights of the contributing domains, the triple-SPV could equally arise from three laterally distinct regions with different doping levels. This supporting experiment does not independently strengthen the intra-unit-cell interpretation.","section":"Results, Fig. S5"},{"comment":"There is a quantitative inconsistency in the reported MBT SPV shift: the main text describes a maximal band shift of ~50 meV at time zero, while Supplementary Fig. S4 states that the SPV shift saturates at approximately 100 meV. This discrepancy matters because the argument that the 1-BT termination shows a significantly larger shift (~100 meV) than the MBT termination is part of the basis for the proposed model. The authors should reconcile these numbers; if the MBT shift is actually ~100 meV under the same conditions, the claimed contrast between the two terminations is weakened.","section":"Results, Fig. 2 and Supplementary Fig. S4"}],"minor_comments":[{"comment":"The authors write that the static linewidth (0.028 Å-1) being broader than the split peaks (0.023 and 0.026 Å-1) indicates 'a superposition of multiple domains with slight angular misalignments.' This is reasonable, but it also concedes that multiple domains contribute to the static spectrum; the text should clarify that this linewidth analysis does not by itself constrain the vertical offset between those domains.","section":"Results, static linewidth analysis"},{"comment":"The caption contains the label 'MBST1610' without defining it. Please explain the sample or measurement identifier in the caption or methods.","section":"Fig. S5 caption"},{"comment":"The statement that the 2-BT termination 'must be p-doped' is an inference from overall charge neutrality plus the observed electron doping of the other terminations. Since the 2-BT termination was not directly observed, this should be framed explicitly as an inference rather than a measured result.","section":"Discussion, 2-BT inference"},{"comment":"The paper alternates between '1-BT termination' and 'BT termination' when referring to the same surface, and Fig. 4 introduces 'electron-doped BT' and 'hole-doped BT' sublayers in a way that can be confused with the 1-BT and 2-BT surface terminations. Please define the notation for surface terminations versus intra-unit-cell layers consistently.","section":"Introduction, terminology"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal, and the experimental observation of multiple SPV replicas is interesting. The main concern is that the load-bearing assumption about the one-unit-cell vertical offset between adjacent domains is unverified and arguably in tension with the authors' own statistics; a spatially resolved or step-height-sensitive measurement, or a substantially weakened claim, would be needed to make the central conclusion convincing. I do not see a circularity problem: the paper does not fit parameters or derive its conclusion from the assumption. The discrepancy between the 50 meV and 100 meV MBT shifts should be fixed in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe real contribution here is the experimental observation: on the 1-BT termination of 18% Sb-doped MnBi6Te10, trARPES resolves two replicas of the band structure that shift differently in time, and on an MBT termination three replicas appear. That multi-SPV splitting on a single nominal termination is new and not in the cited prior work. The static ARPES showing that MBT and 1-BT surfaces remain electron-doped despite near-zero Hall carrier density is useful, and it backs the practical takeaway: chemical doping alone won't put the Fermi level in the magnetic gap of this material. The fluence-saturation check for SPV is clean.\n\nThe soft spot is the one the stress-test flags, and it's real. The dual-SPV model assumes the two probed 1-BT regions are adjacent domains with a one-unit-cell vertical offset. trARPES can't measure that, and the cleaving-statistics fallback is a weak prior. The probe area is 24×25 µm2, so lateral inhomogeneity—different local Sb concentration, anti-site defects, band bending—could produce two different SPV shifts without any vertical offset. The static MDC linewidth already indicates a superposition of slightly misaligned domains, so inhomogeneity is present. The momentum splitting could be a lateral field, again not uniquely tied to a one-unit-cell step.\n\nTo the authors' credit, they state the limitation explicitly: 'the trARPES experiment alone cannot determine the vertical offset.' So it's an acknowledged inference, not a hidden assumption. But that inference supports the paper's headline claim of intra-unit-cell charge redistribution and spontaneous polarization. The claim is consistent with the termination-dependent doping seen in micro-ARPES, but it's underdetermined by the multi-SPV data alone.\n\nMinor issues: one baseline carrier density comes from private communication, and raw data aren't deposited. Not deal-breakers.\n\nWho this is for: people working on Mn-Bi-Te materials and on SPV in topological insulators. It deserves a serious referee; the new observation is publishable even if the interpretation needs work. I'd send it out and ask for either spatial-resolved measurements (STM or nano-ARPES) or a quantitative treatment of lateral-inhomogeneity alternatives. My own verdict would be conditional, not reject.","headline":"A genuinely new trARPES observation—multiple SPV replicas on one termination—but the intra-unit-cell polarization story leans on an unverified one-unit-cell step geometry.","tokens_in":10606,"tokens_out":2945,"would_cite":true,"duration_ms":28814,"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":"Sb-doped MnBi6Te10 is globally charge-neutral yet separates charge inside each unit cell, so chemical doping alone cannot realize the quantum anomalous Hall effect.","keywords":["MnBi6Te10","magnetic topological insulator","trARPES","surface photovoltage","intra-unit-cell charge redistribution","quantum anomalous Hall effect","Sb doping"],"falsifier":"Measure the topography of the exact region used in trARPES with a technique such as STM or atomic-force microscopy on the same cleaved crystal: if the two band replicas appear on a single flat 1-BT terrace with no one-unit-cell step, or if the step height between the two emitting domains is not an integer multiple of the unit-cell height, the intra-unit-cell charge-redistribution explanation fails. A complementary test is to repeat trARPES with a probe spot smaller than the typical domain size; if both SPV replicas persist when the probe is centered far from any step edge, the dual-domain interpretation is falsified.","tokens_in":9555,"feed_emoji":"⚡","tokens_out":7819,"duration_ms":69617,"temperature":0.7,"pith_summary":"The paper aims to establish that in 18% Sb-doped MnBi6Te10, a magnetic topological insulator that Hall measurements place near global charge neutrality, charge is not actually uniform: electrons and holes separate across the layers within each unit cell, producing a built-in electric dipole. The evidence comes from time- and angle-resolved photoemission: after an infrared pump pulse, the band structure on the single-Bi2Te3 termination splits into two replicas shifting by different amounts, and on a MnBi2Te4 termination it splits into three. The authors read these multiple surface photovoltages as different local band bendings on adjacent terraces whose vertical offset is one unit cell. If correct, this means chemical doping alone cannot shift the Fermi level into the magnetic gap needed for the quantum anomalous Hall effect; the material also needs uniform carrier doping. It also suggests a light-induced polarization within each unit cell that could be exploited optoelectronically.","feed_headline":"Charge-neutral Sb-doped MnBi6Te10 hides charge split inside unit cells","feed_subtitle":"Two photovoltage replicas in trARPES reveal a per-unit-cell electric dipole that dopant chemistry cannot erase.","key_machinery":"The central mechanism is surface photovoltage (SPV): when an infrared pump pulse generates electron-hole pairs, the photoexcited carriers screen the band bending at the surface, transiently shifting the whole band structure in energy. In this material the shift is not a single rigid motion: the band splits into two replicas with different SPV shifts and a momentum offset of about $0.01$ Å$^{-1}$, which the paper interprets as two adjacent 1-BT terraces, vertically offset by one unit cell, sitting in different local electric fields. The unit cell itself is the load-bearing object: it is modeled as a built-in p-n junction whose net charge is zero but whose internal dipole produces the different photovoltages, and the momentum splitting is attributed to lateral electric fields near the domain boundary acting on the emitted photoelectrons.","core_discovery":"The central claim is that a nominally charge-neutral magnetic topological insulator, 18% Sb-doped MnBi6Te10, nevertheless exhibits intra-unit-cell charge redistribution: within one unit cell, the MnBi2Te4 layer and one Bi2Te3 layer are electron-doped while the other Bi2Te3 layer is hole-doped, forming an internal n-p junction. The evidence is the transient splitting of the band structure into two, or in another spot three, replicas during surface photovoltage. Because the replicas shift by different energies and slightly different momenta, the authors argue that adjacent 1-BT surface domains with a one-unit-cell vertical offset sit at different photovoltages, which can only happen if each unit cell carries a spontaneous electric polarization. This reconciles the contradiction between global charge neutrality (Hall carrier density about $6.5\\times10^{18}$ cm$^{-3}$) and the electron-doped surfaces seen in static ARPES. The consequence the paper draws is that realizing the quantum anomalous Hall effect in this family requires both uniform carrier doping and a ferromagnetic ground state, not just chemical doping.","pith_inferences":["If the one-unit-cell-step interpretation is right, the roughly 50 meV spacing between replicas quantifies the built-in dipole potential per unit cell; a testable extension would be to vary the Sb concentration and anti-site defect density and look for a proportional change in the replica splitting.","Because the paper notes the top and bottom BT layers can be swapped by flipping the crystal, the sign of the intra-unit-cell polarization should be stochastic from domain to domain; spatially resolved pump-probe measurements could map domains with opposite polarization.","The same dual-SPV probe might apply to other MnBi$_{2n}$Te$_{3n+1}$ members such as MnBi$_2$Te$_4$ and MnBi$_4$Te$_7$: if the effect scales with unit-cell thickness, it would confirm that the charge redistribution is a generic property of the heterostructure family rather than a peculiarity of MnBi$_6$Te$_{10}$.","If anti-site defects are the driver, then defect engineering rather than doping chemistry is the natural next lever; samples with fewer Mn-Bi/Sb anti-site defects should show smaller replica splitting and more uniform SPV."],"forward_implications":["Global charge neutrality measured by Hall transport does not imply local charge neutrality: the MBT and 1-BT terminations remain electron-doped, so transport and surface-sensitive probes are answering different questions.","Because total charge is conserved, the unobserved 2-BT termination must be hole-doped, completing an intrinsic n-p junction inside each unit cell rather than a uniformly doped bulk.","Sb doping alone cannot place the Fermi level in the magnetic gap of MnBi6Te10; realizing the quantum anomalous Hall effect requires simultaneous control of uniform carrier density and the ferromagnetic ground state.","The multi-SPV effect (dual on 1-BT, triple on MBT) appears on both terminations, indicating it is intrinsic to the unit-cell stacking of the material rather than a single-termination artifact.","Light-induced intra-unit-cell polarization in ferromagnetic Mn(Bi$_{0.82}$Sb$_{0.18}$)$_6$Te$_{10}$ could be used in optoelectronic and spintronic devices, including magnetic tunnel junctions with charge-to-spin conversion and electric-field control."],"supporting_citations":[{"why":"Establishes the quantum anomalous Hall effect in the intrinsic magnetic topological insulator MnBi2Te4, the target phase that the paper argues chemical doping alone cannot reach.","marker":"[1]"},{"why":"Prior ARPES study of pristine MnBi6Te10 that identifies the three terminations, their different dopings, and the delicate ferromagnetism stabilized by anti-site defects.","marker":"[5]"},{"why":"Shows that low-concentration Sb doping can stabilize ferromagnetism in the Mn-Bi-Te family, the doping route whose uniformity the paper questions.","marker":"[14]"},{"why":"Documents surface photovoltage in bulk-insulating topological insulators Bi2Se3 and Bi2Te3, providing the baseline SPV interpretation used for the band shifts.","marker":"[17]"},{"why":"Explains the negative-delay SPV energy shifts in time-resolved photoemission, which the paper relies on to identify photovoltage rather than simple charging.","marker":"[25]"},{"why":"Demonstrates ultrafast photoexcitation of a persistent surface-state population in a topological insulator, supporting the SPV picture for surface band bending.","marker":"[26]"},{"why":"Provides cleaving statistics and step-height information that the paper uses to argue the adjacent 1-BT domains are vertically offset by one unit cell.","marker":"[31]"},{"why":"Identifies how Sb doping promotes anti-site defects in MnBi2Te4, the microscopic mechanism the paper invokes to explain the intra-unit-cell charge redistribution.","marker":"[32]"}],"fun_headline_variants":["Sb-doped MnBi6Te10 splits charge inside unit cells despite neutrality","Internal n-p junction found in charge-neutral Sb-doped MnBi6Te10","Per-unit-cell dipole emerges in Sb-doped MnBi6Te10, trARPES shows","Charge-neutral MnBi6Te10 still carries hidden electric polarization","Sb-doped MnBi6Te10 reveals light-induced per-unit-cell polarization"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion rests on assuming that the two band replicas in trARPES come from neighboring surface terraces separated by exactly one unit-cell step; if the two signals instead come from unrelated surface patches, mixed terminations, or ordinary space-charge effects, the intra-unit-cell charge-redistribution story does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Sb-doped MnBi6Te10 splits charge inside unit cells despite neutrality","Internal n-p junction found in charge-neutral Sb-doped MnBi6Te10","Per-unit-cell dipole emerges in Sb-doped MnBi6Te10, trARPES shows","Charge-neutral MnBi6Te10 still carries hidden electric polarization","Sb-doped MnBi6Te10 reveals light-induced per-unit-cell polarization"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000202,"raw_usage":{"total_tokens":1459,"prompt_tokens":1098,"completion_tokens":361,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":714,"completion_tokens_details":{"reasoning_tokens":262}},"tokens_in":714,"tokens_out":361,"duration_ms":3589,"temperature":1.0,"reasoning_tokens":262,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:42:09.807046+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the topography of the exact region used in trARPES with a technique such as STM or atomic-force microscopy on the same cleaved crystal: if the two band replicas appear on a single flat 1-BT terrace with no one-unit-cell step, or if the step height between the two emitting domains is not an integer multiple of the unit-cell height, the intra-unit-cell charge-redistribution explanation fails. A complementary test is to repeat trARPES with a probe spot smaller than the typical domain size; if both SPV replicas persist when the probe is centered far from any step edge, the dual-domain interpretation is falsified.","supporting_citations":[{"cited_title":"flat-band condition","cited_arxiv_id":null,"evidence_quote":"Demonstrates ultrafast photoexcitation of a persistent surface-state population in a topological insulator, supporting the SPV picture for surface band bending."}],"review_version":1}