{"id":"5f63127f-9df0-420b-a429-b74066fbe56f","arxiv_id":"2505.09946","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In the magnetic state of CrPS4, spin splitting of the bands by about 0.5 eV creates fully spin-polarized band edges that quantitatively explain the tunneling spectra and photoluminescence transitions.","lead":"This paper shows that in the magnetic semiconductor CrPS4, the electronic bands split into spin-polarized copies below the magnetic transition, producing new features in tunneling spectra and explaining the photoluminescence lines. It combines scanning tunneling spectroscopy, optical measurements, and density functional theory to map how magnetism reshapes the band structure.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Full spin polarization is inferred from absence of extra PL lines; without quantitative exciton and matrix-element calculations, missing PL cannot rule out unpolarized bands.","rationale":"The reader's weakest assumption identifies the argument from silence: the conclusion that bands are fully spin-polarized relies on the premise that absent PL lines imply forbidden transitions. My concern shares this core but sharpens it with a concrete physical mechanism: exciton binding and dark excitons in 2D semiconductors. The paper's direct equality between STS single-particle energy differences and PL photon energies is not justified without accounting for excitonic effects, and the DFT gap discrepancy (0.6 vs 0.87 eV) further weakens the validation. However, the overall picture—spin-split bands with matching relative energies—remains plausible and well-supported by the STS-PL-DTF agreement. A GW-BSE calculation would settle whether the missing PL lines are truly informative. Since the reader already assigned CONDITIONAL status, this concern reinforces that conditionality rather than changing the verdict; the paper should explicitly address excitonic effects or perform spin-resolved measurements before claiming full spin polarization. Therefore I keep the reader's verdict unchanged.","tokens_in":11745,"tokens_out":7357,"duration_ms":80502,"concrete_test":"Perform GW-BSE calculations (or a comparable excitonic approach) for the magnetic state of monolayer CrPS4 using the same vdW-DF-cx starting point, and list the bright exciton energies and oscillator strengths for all transitions within ±2 eV of the gap. If the computed bright excitons include transitions that are not observed in PL, or if the exciton binding energies shift the STS-PL energy matches by more than ~50 meV, then the argument from silence fails and the full spin-polarization conclusion is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that individual bands are fully spin-polarized rests on the Discussion statement: 'if the bands were not fully spin polarized, optical transition at many more different energies should be detected in the experiments.' This is an argument from silence that assumes every spin-conserving interband transition with the correct energy is bright and observable in photoluminescence. In a 2D semiconductor, many transitions are dark due to exciton binding, symmetry selection rules, momentum mismatch, or small oscillator strength, and PL is dominated by excitons rather than free-carrier interband recombination. The paper compares single-particle STS band-edge differences directly to PL photon energies (e.g., Table I, Fig. 3c), implicitly neglecting exciton binding energies, which in 2D materials can exceed 100 meV. Without a quantitative calculation of optical matrix elements and exciton energies for the proposed band structure, the absence of extra PL lines does not establish full spin polarization. The DFT validation is also not fully quantitative: the calculated bandgap (0.6 eV) differs from the STS gap (0.87 eV) by about 30%, so the DFT spin assignments are only approximately confirmed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports STS, PL, and DFT results on the van der Waals magnetic semiconductor CrPS4. It shows that below the magnetic transition (Tc = 38 K) new features appear in the tunneling spectra, which the authors attribute to spin-split bands with a large exchange energy (about 0.5 eV). The energy differences between the STS features are compared to photoluminescence lines observed in the magnetic state, and a qualitative match is claimed. Spin-polarized DFT calculations are presented that show fully spin-polarized conduction and valence bands, with a calculated gap of 0.6 eV versus the measured 0.87 eV STS gap. Based on the consistency between STS, PL, and DFT, the authors conclude that individual bands are fully spin-polarized over a broad energy interval and that this explains the optoelectronic properties of CrPS4.","tokens_in":11908,"tokens_out":5209,"duration_ms":51279,"significance":"If correct, the paper would provide a comprehensive picture of how magnetism modifies the band structure of a van der Waals semiconductor, with implications for spin-polarized transport and gate-tunable half-metallic devices. The combination of STS, PL, and DFT is strong in principle: the DFT calculation is an independent benchmark, the STM topograph comparison provides a second observable beyond band energies, and the data are deposited in a public repository. The main strength is the multi-technique consistency, but the load-bearing conclusion of full spin polarization is inferred indirectly and needs additional support.","major_comments":[{"comment":"The central claim that individual bands are fully spin-polarized over a broad energy interval rests on the argument in the second paragraph: 'if the bands were not fully spin polarized, optical transition at many more different energies should be detected in the experiments.' This is an argument from silence. In a 2D semiconductor, many interband transitions are dark or very weak because of exciton binding, symmetry selection rules, momentum mismatch, or small oscillator strength, and photoluminescence is dominated by excitons rather than free-carrier interband recombination. Without a quantitative calculation of optical matrix elements and exciton energies for the proposed band structure, the absence of additional PL lines cannot rule out the presence of unpolarized or partially polarized bands. The DFT calculation alone is not decisive because it uses a specific exchange-correlation functional whose spin splitting is not experimentally validated. The paper should either soften the claim to 'consistent with full spin polarization' or provide a quantitative estimate of the brightness of all competing transitions.","section":"Discussion and conclusions"},{"comment":"The quantitative comparison of STS band-edge differences to PL photon energies, summarized in Table I and Fig. 3c, implicitly neglects exciton binding energies. In a 2D semiconductor, exciton binding energies can be hundreds of meV, so the energy of a PL line is not simply the difference between two single-particle band edges. This is not a small correction in the present data: in the paramagnetic state the STS gap is reported as 1.25 eV (Table I) while the optical gap is 1.34 eV, meaning the single-particle gap is smaller than the optical gap, which is unphysical for a direct semiconductor with positive exciton binding. The manuscript should address this discrepancy explicitly, for example by discussing whether the STS features are true band edges or whether the optical transition in the paramagnetic state is not the fundamental gap. Without such a discussion, the claimed 'virtually perfect quantitative match' between STS energy differences and PL energies is not reliable evidence for the band-edge identification.","section":"Tunneling spectroscopy and its interpretation"},{"comment":"The statement that the DFT calculated bandgap of 0.6 eV is 'very close' to the measured STS gap of 0.87 eV overstates the agreement; the discrepancy is about 30 percent. Moreover, the colored arrows in Fig. 4 are placed onto the DFT band structure post hoc, connecting extrema of the same spin at the same k-point that match the PL energies. This procedure selects the transitions that reproduce the experimental values, so the agreement does not constitute an independent prediction of the optical spectrum. To support the claim that DFT 'quantitatively predicts the relative positions of all detected bands' and explains which pairs of bands radiate, the authors should compute the optical transition strength (e.g., dipole matrix elements) and, ideally, the exciton spectrum, and show that the predicted bright transitions match the observed PL lines. As written, the validation loop is partly circular.","section":"Comparison with first-principles simulations"}],"minor_comments":[{"comment":"The text says 'Exfoliation exposes a buckled selenium surface', but CrPS4 contains sulfur, not selenium; this should be corrected to 'sulfur'.","section":"Magnetic and semiconducting properties"},{"comment":"The text states 'a bandgap just smaller than 1.5 eV is clearly visible' at 78 K, but Table I reports the PM bandgap as 1.25 eV. Please reconcile these values or state the uncertainty in the gap determination.","section":"Tunneling spectroscopy and its interpretation"},{"comment":"In reference 52 the author list appears garbled ('L. Sun, R. Gibertini, A. Scarfato, ...'); the second author should be 'M. Gibertini' to match the other references.","section":"Data availability"},{"comment":"The statement that electrons in the high-energy spin-split band 'can only relax by flipping their spin or by hopping to adjacent layers' overlooks other relaxation channels such as phonon-assisted scattering within the same spin manifold, which may affect the radiative lifetime argument. This is a minor point because the qualitative conclusion is unaffected.","section":"Discussion and conclusions"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an interesting and potentially important dataset, but the full-spin-polarization conclusion is currently supported more by an argument from silence than by direct evidence. The authors should be encouraged to either temper the claim or supply a concrete calculation of optical matrix elements and exciton effects. The DFT agreement is also overstated given the 30% gap discrepancy. These issues are fixable within the manuscript's scope, so a major revision is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real news here is the first STS mapping of the spin-split band edges in CrPS4, and the unified picture it gives for the previously puzzling PL lines, including the 1.99 eV feature. The data look carefully taken, the comparison across STS, PL, and DFT is systematic, and the STM topography simulations are a nice extra check. The paper earns its keep as an experimental tour de force.\n\nWhat is genuinely new: above Tc you see a clean gap; below Tc you see extra features that line up with PL transitions if you read them as spin-split band edges. The match between the STS energy differences and the PL lines is impressive, and the DFT calculation, while not perfect, reproduces the relative band positions and the wavefunction maps. Data are deposited, which is good.\n\nNow the soft spots. The claim that individual bands are fully spin-polarized over a broad energy interval is not measured directly. It rests on the statement in the Discussion that if bands were not fully spin-polarized, many more optical transitions should be seen. That is an argument from silence, and in a 2D semiconductor it is fragile: dark excitons, momentum mismatch, and small oscillator strengths can easily hide transitions. The paper also compares single-particle STS band-edge differences directly to PL photon energies, ignoring exciton binding energies, which in 2D can be hundreds of meV. The DFT gap is 0.6 eV versus 0.87 eV measured—about 30% off—so the DFT support is qualitative for the gap size, even if the same-spin transition energies match better. The colored arrows on the DFT bands are placed post hoc, which is fine for illustration but not independent confirmation.\n\nNone of this kills the central picture. The evidence that exchange splitting rearranges the band edges and drives the optical response is strong. But the strong conclusion about full spin polarization should be flagged as an inference, not a measurement, and the authors should either tone it down or point to a quantitative exciton/matrix-element calculation. A spin-resolved STS experiment would settle it, and the paper itself notes none has been done.\n\nWho is this for? Anyone working on 2D magnetic semiconductors, especially CrPS4. It deserves a serious referee and, after revision, publication. My recommendation: send it to review, but ask the authors to address the argument-from-silence issue and the exciton-binding-energy point explicitly.","headline":"Solid STS/PL/DFT study mapping spin-split bands in CrPS4, but the full spin-polarization claim rests on an argument from silence that needs softening or spin-resolved confirmation.","tokens_in":706,"tokens_out":834,"would_cite":true,"duration_ms":20240,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"When CrPS4 enters its magnetic state, its conduction and valence bands split by about 0.5 eV into fully spin-polarized copies, and the energy differences between these band edges account for every photoluminescence line observed.","keywords":["CrPS4","van der Waals magnetic semiconductor","spin splitting","exchange energy","scanning tunneling spectroscopy","photoluminescence","band structure","two-dimensional semiconductor"],"falsifier":"A decisive test would be a direct measurement of the spin of the bands, for example spin-resolved photoemission or spin-polarized tunneling on the same CrPS4 surface: full spin polarization at the band edges would confirm the claim, and a detectable minority-spin signal would refute it. Alternatively, a broad search for weak, dark or momentum-indirect emission lines in the magnetic state would check the assumption that all spin-allowed transitions would be visible in photoluminescence.","tokens_in":11518,"feed_emoji":"🧲","tokens_out":6014,"duration_ms":58675,"temperature":0.7,"pith_summary":"The paper sets out to show that the semiconducting and magnetic properties of the van der Waals semiconductor CrPS4 are coupled through a single mechanism: when the material enters its magnetic state below 38 K, the electronic bands are split by a large exchange energy of roughly 0.5 eV, leaving the low-energy conduction and valence bands fully spin-polarized. The authors combine scanning tunneling spectroscopy, photoluminescence, and density functional theory to argue that every observed optical transition in the magnetic state is the energy difference between two band edges of the same spin, while the spin-opposite pairs remain dark. What makes this worth caring about is that it explains not only the ordinary optical gap but also a mysterious high-energy photoluminescence line deep in the band continuum, and it identifies CrPS4 as a platform where gate-tunable half-metallic conduction should be achievable. The central claim is that all basic optoelectronic processes in this material are governed by the spin-split band structure that develops on entering the magnetic state.","feed_headline":"Magnetic order splits CrPS4 bands by 0.5 eV","feed_subtitle":"Below 38 K, every conduction and valence band edge becomes fully spin-polarized, and gap differences match each photoluminescence line.","key_machinery":"The central object is the exchange-split band structure of a single CrPS4 layer in its A-type antiferromagnetic state, in which each layer is ferromagnetically ordered and neighboring layers point oppositely. Bands that are paired in the paramagnetic phase split into two spin copies separated by roughly 0.5 eV, so the lowest conduction band and highest valence band have opposite spins. The argument is carried by matching energy differences between STS-detected band edges to photoluminescence transition energies, then confirming the assignment with spin-resolved density functional theory, which also reproduces the measured spatial pattern of the wavefunctions. Whether a transition is bright or dark is decided by whether the two bands used in the transition have the same spin and share a common point in momentum space.","core_discovery":"In the magnetic state of CrPS4, the paper claims, the conduction and valence bands are each split into two spin copies separated by about 0.5 eV of exchange energy, and the band edges that determine transport and optics are fully spin-polarized over a broad energy interval. This spin splitting reduces the tunneling bandgap to about 0.87 eV, makes the optical gap of about 1.34 eV arise from same-spin transitions, and accounts for a Fano-shaped line at 1.37–1.40 eV and a sharp high-energy photoluminescence line at 1.99 eV. Density functional theory calculations reproduce the relative positions of all detected band edges and the spatial structure of the wavefunctions, and the paper concludes that every basic optoelectronic process in CrPS4 follows from the evolution of the spin-polarized band structure when magnetic order appears.","pith_inferences":["A testable extension of the paper's picture: the high-energy line's energy should track the exchange splitting as a function of magnetic field or uniaxial strain, since it is set by the same-spin separation of a deep valence band and a conduction band.","If the bands are truly fully spin-polarized over a broad interval, optical excitation at photon energies well above the gap should generate spin-polarized carriers without any spin-polarized contacts, a property useful for spin injection.","The same STS-to-PL comparison could be applied to other layered magnetic semiconductors to see whether the exchange-split band picture is general or whether excitonic and defect effects mask it in other materials."],"forward_implications":["The tunneling bandgap of about 0.87 eV is invisible in photoluminescence because it connects opposite-spin band edges, so direct radiative recombination is forbidden.","The 1.99 eV photoluminescence line is emitted from a fully spin-polarized conduction band whose same-spin valence counterpart lies deep in the continuum; the electron cannot relax without a rare spin flip, so it survives to recombine radiatively.","The spin-polarized band structure over a broad energy interval makes CrPS4 a candidate platform for gate-tunable half-metallic conductors based on van der Waals magnetic semiconductors.","The same exchange-splitting mechanism explains the exponential gate-voltage dependence of magnetoconductance reported for CrPS4 transistors, because aligning spins shifts the conduction band edge downward."],"supporting_citations":[{"why":"Reports the high-energy photoluminescence line near 1.99 eV and the Fano-shaped line that the paper's band-structure picture is designed to explain.","marker":"[38]"},{"why":"Provides the optical-gap and photoluminescence measurements that establish the 1.34 eV comparison in the paramagnetic state.","marker":"[34]"},{"why":"Documents the temperature dependence of Fano resonances in CrPS4, supporting assignment of the 1.37 eV line.","marker":"[37]"},{"why":"Shows magnetoconductance in CrPS4 transistors caused by magnetic alignment shifting the conduction band edge, connecting the band-edge picture to transport.","marker":"[41]"},{"why":"Determines the magnetic structure and confirms the A-type antiferromagnetic order and transition temperature used to interpret the low-temperature state.","marker":"[39]"},{"why":"Establishes the magnetic structure and exchange interactions of CrPS4, providing the layer-resolved spin picture needed for the single-layer interpretation.","marker":"[40]"},{"why":"Provides the density functional theory code used for the band-structure and topography calculations.","marker":"[46]"}],"fun_headline_variants":["CrPS4 magnetism splits bands by 0.5 eV","Spin-polarized bands define all CrPS4 optics","Magnetic order sets CrPS4 band edges apart","0.5 eV exchange splitting reshapes CrPS4 bands","How CrPS4 magnetism explains every optical line"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that the bands are fully spin-polarized rests on the assumption that every allowed optical transition between partially polarized bands would produce a detectable photoluminescence line; if dark excitons, defect states, or momentum mismatch hide such transitions, the absence of extra lines would not prove full spin polarization.","fun_headline_variants_meta":{"raw":{"variants":["CrPS4 magnetism splits bands by 0.5 eV","Spin-polarized bands define all CrPS4 optics","Magnetic order sets CrPS4 band edges apart","0.5 eV exchange splitting reshapes CrPS4 bands","How CrPS4 magnetism explains every optical line"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00088,"raw_usage":{"total_tokens":3802,"prompt_tokens":944,"completion_tokens":2858,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":560,"completion_tokens_details":{"reasoning_tokens":2779}},"tokens_in":560,"tokens_out":2858,"duration_ms":19474,"temperature":1.0,"reasoning_tokens":2779,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:20:13.577103+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be a direct measurement of the spin of the bands, for example spin-resolved photoemission or spin-polarized tunneling on the same CrPS4 surface: full spin polarization at the band edges would confirm the claim, and a detectable minority-spin signal would refute it. Alternatively, a broad search for weak, dark or momentum-indirect emission lines in the magnetic state would check the assumption that all spin-allowed transitions would be visible in photoluminescence.","supporting_citations":[],"review_version":1}