{"id":"cfcf23e2-ee48-491e-aeaa-38a714c604ef","arxiv_id":"2412.01239","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"The authors observe path-dependent angular asymmetry in the dielectric response of antiferromagnetic FePS3 at 2.0 eV and attribute it to light-induced octupolar electronic polarization.","lead":"This paper reports that below its magnetic ordering temperature, the antiferromagnet FePS3 shows an angle-sweep asymmetry in its optical dielectric response at photon energy 2.0 eV, which the authors interpret as light-induced hysteresis of electronic polarization. The work is potentially relevant to optical control of multiferroics, but the claim rests on a physically problematic extraction of a so-called octupolar component.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The gap between the 0–90° and 90–180° branches is static mirror asymmetry, not hysteresis; no history dependence is shown, so the central claim of light-induced memory is unsupported.","rationale":"The reader's weakest_assumption identifies the cos(6θ) octupolar component as physically implausible in a linear rank-2 dielectric response, and that is a strong technical objection to the proposed origin. However, the more fundamental problem is the claim of 'hysteresis' itself: the reported gap between the 0–90° and 90–180° branches is simply a static angular asymmetry, not a memory effect. Without demonstrated history dependence, the abstract's central claim—light-induced hysteresis of electronic polarization—is unsupported regardless of whether an octupolar component exists. The reader's rationale also mentions this hysteresis overclaim, but their weakest_assumption focuses on the octupole. I partially agree: the octupole concern is valid, but the hysteresis definition is logically prior. My proposed bidirectional sweep would settle the question directly. Since the reader already reached REJECT, and my concern reinforces that conclusion, the verdict remains unchanged.","tokens_in":15722,"tokens_out":3515,"duration_ms":32728,"concrete_test":"Perform a bidirectional angular sweep at 2.0 eV and 10 K: measure Re[ε](θ) for θ increasing from 0° to 180°, then immediately for θ decreasing from 180° to 0° at the same spot. If the two curves coincide at every angle, the 'hysteresis gap' is a static mirror-asymmetric response, not a memory effect. Alternatively, hold the polarization at θ = 45° and compare Re[ε] after first illuminating at θ = 0° for 10 minutes versus after θ = 90°; identical values indicate no light-induced memory.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is hysteresis of electronic polarization, but the evidence is a difference between Re[ε](θ) measured on two adjacent angle intervals in a single static angular sweep. This gap is equivalent to Re[ε](θ) ≠ Re[ε](180°−θ), i.e., mirror symmetry breaking. It does not require any dependence on how the state was reached. True hysteresis requires that the response at a given θ depends on the history (e.g., previous angles, illumination time, sweep direction). The paper never shows such memory; all reproducibility and temperature checks are consistent with a static angular anisotropy that appears below TN. Therefore, 'light-induced hysteresis' is an overclaim. Even if the cos(6θ) octupolar component were real, it produces a static 6-fold angular modulation, not a hysteresis loop. The experimental design (transmission at each angle) cannot distinguish memory from static asymmetry unless the angle is swept bidirectionally or the illumination history is varied. This concern is load-bearing because removing 'hysteresis' removes the central novelty claimed in the abstract and conclusion.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports polarization-resolved transmission spectroscopy of the antiferromagnetic semiconductor FePS3 below its Néel temperature (TN ~117 K). From angle-dependent transmission spectra, the authors extract the real and imaginary parts of the dielectric constant via multi-oscillator fits (six Gaussians plus a Cauchy term) and observe linear dichroism below TN. They further plot Re[ε](θ) − min(Re[ε](θ)) over 0°–90° and 90°–180°, finding that the two angular intervals ('Path I' and 'Path II') overlap at 1.6 eV but split at 2.0 eV. This splitting is described as a 'hysteresis gap' and is attributed to a light-induced electronic octupolar polarization combined with dipolar polarization, associated with a broken mirror symmetry of the light-excited state. The paper claims light-induced hysteresis of electronic polarization and proposes a mechanism for light-induced multiferroicity.","tokens_in":97,"tokens_out":5016,"duration_ms":99450,"significance":"If the central claim were valid, the observation of light-induced, history-dependent electronic polarization in an antiferromagnetic semiconductor would be a substantial new result with implications for optical control of multiferroic order. The paper has some strengths: it includes temperature-dependent data, reproducibility across three samples of different thickness, a stability check after 9 months of air exposure, and a BCS-type fit for the temperature dependence of the claimed gap. However, the manuscript's central claims rely on two problematic assumptions: (i) that the observed difference between two angular intervals constitutes true hysteresis (a history-dependent memory effect), and (ii) that a cos(6θ) component can appear in the linear dielectric constant of a homogeneous crystal. Both assumptions are unsupported, and the second is physically inconsistent for a rank-2 tensor. Because these issues undermine the main conclusions, the paper in its current form is unlikely to make a sound contribution.","major_comments":[{"comment":"The central claim of 'hysteresis' is not supported by the experimental evidence. The data are obtained from a static angular sweep: Re[ε](θ) is measured over 0°–90° and then over 90°–180° in a single pass, and the difference between these two intervals is a static mirror asymmetry of the function Re[ε](θ), i.e., Re[ε](θ) ≠ Re[ε](180°−θ). True hysteresis requires the response to depend on the history of the external parameter (here, the polarization angle or illumination history), typically shown by bidirectional sweeps yielding different branches. The manuscript provides no test of history dependence or memory. Thus the abstract's claim of 'light-induced hysteresis of electronic polarization' is an overclaim; what is observed is a static angular anisotropy with broken mirror symmetry below TN. This concern is load-bearing because the novelty claimed in the title and abstract rests entirely on the word 'hysteresis'.","section":"§2 (Figure 4a,b)"},{"comment":"The identification of a cos(6θ) 'octupolar' component in the linear dielectric constant is physically inconsistent. For a homogeneous medium, the linear permittivity is a rank-2 tensor, and the measured quantity Re[ε](θ) = e_i(θ) Re[ε_ij] e_j(θ) is bilinear in the direction cosines; its angular dependence can contain only a constant and cos(2θ)/sin(2θ) terms (equivalently l = 0 and l = 2). A cos(6θ) component cannot arise in the linear response of a bulk crystal, independent of any static octupolar order parameter, unless a nonlinear optical process or a different observable (e.g., intensity) is involved. The claimed octupolar component in Fig. 4d therefore most likely arises from the multi-oscillator fitting procedure (six Gaussians plus a Cauchy term) rather than from a physical electronic octupole. Removing this component removes the proposed origin of the hysteresis gap, so this is a load-bearing error.","section":"§2 (Figure 4d; Fourier analysis of Re[ε](θ))"},{"comment":"The 'explanation' of the hysteresis gap by octupolar polarization is circular. Both the hysteresis gap |δmax| and the octupolar amplitude are extracted from the same fitted Re[ε](θ) spectra, so their agreement in Fig. 4e does not provide independent confirmation. Furthermore, the linear relation |δmax| = α × ΔRe[εb] + β (α = 0.13, β = 0.015) introduces two additional fitted parameters, and the BCS-type fit |δmax|(T) = A tanh(2.3√(TN/T − 1)) introduces A as a free parameter. The paper does not discuss the number of free parameters used to generate the key quantities, nor does it provide an error analysis that would allow the reader to judge whether the apparent correlations are statistically meaningful. Without an independent test of the octupolar hypothesis, the claim that octupolar polarization 'drives' the hysteresis is not established.","section":"§2 (Figure 4e and correlation |δmax| = α ΔRe[εb] + β)"}],"minor_comments":[{"comment":"The term 'hysteresis' is used even when the two paths overlap (e.g., 'hysteresis without a gap' in the caption of Fig. S6). This loose terminology obscures the distinction between a two-path plot constructed from two angular intervals and a genuine history-dependent loop. A more neutral descriptor, such as 'angular asymmetry' or 'mirror-symmetry breaking,' would be more accurate.","section":"Abstract and Fig. 4a,b"},{"comment":"The BCS-type model for |δmax|(T) is introduced without justification or reference. The constant 2.3 in the exponent and the choice of tanh functional form need to be motivated, or the fit should be presented as phenomenological with suitable caveats.","section":"§2 (Fig. 4c)"},{"comment":"The dielectric constants are obtained from fits with a large number of free parameters (six Gaussians plus Cauchy terms for each spectrum), but the manuscript gives no uncertainties on the fitted parameters or on the resulting Re[ε](θ) values. The reader cannot assess whether the small differences that produce |δmax| are within the fitting error.","section":"Methods (Modeled transfer function)"},{"comment":"The Fourier decomposition labels components as 'dipolar' and 'octupolar.' Since the linear dielectric tensor is rank-2, such labels are misleading; the angular components in the linear response should be described by their l = 0,2 (and possibly higher if nonlinear) origins, not by static multipole moments.","section":"Fig. 4d"}],"recommendation":"reject","confidential_remarks":"The manuscript's central claims rest on two fundamental issues that cannot be fixed locally: the mischaracterization of static angular asymmetry as hysteresis, and the assertion of a cos(6θ) component in the linear dielectric response. The second point is a fatal inconsistency with basic tensor properties, and the first removes the claimed novelty. Even if the experimental data are reproducible, the interpretation and title would need to be substantially reworked, and it is unclear whether a supportable claim of light-induced polarization memory can be extracted from the present measurements. I recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the data quality is good: three samples, temperature sweeps, sum-rule checks, and a stability check after nine months. The linear dichroism below TN is reproduced and consistent with earlier work. That part of the paper is solid.\n\nThe problems are with the interpretation. What they call the hysteresis gap is the difference between Re[ε](θ) on the 0–90° and 90–180° branches of a single static angular sweep. That is equivalent to Re[ε](θ) ≠ Re[ε](180°−θ), i.e., mirror asymmetry of the dielectric tensor. No history dependence is shown: no bidirectional sweeps, no variation of illumination time, no demonstration that the state depends on how you got there. Calling this 'hysteresis' and invoking a 'light-induced memory effect' is an overclaim.\n\nThe second issue is the octupolar polarization. The linear dielectric function is a rank-2 tensor, so its angular dependence in a fixed plane is limited to a constant, cos(2θ), and sin(2θ). A cos(6θ) component cannot arise from any linear response of a crystal. The Fourier analysis that extracts both cos(2θ) and cos(6θ) is fitting a functional form that violates the symmetry of the linear response. The octupole term is almost certainly an artifact of the multi-oscillator model or of the data reduction. The correlation between |δmax| and the octupole amplitude is circular because both come from the same fitted Re[ε](θ), and the correlation uses two fitted parameters.\n\nThere is also a reporting gap: the dielectric functions come from fits with roughly 23 free parameters per spectrum, and no uncertainties are propagated into the derived quantities such as |δmax| or the octupole amplitude. This makes it hard to evaluate whether the residual after subtracting the biaxial model is statistically meaningful.\n\nThe physical picture—charge transfer between zigzag and armchair directions leading to polarization along a at 1.6 eV and b at 2.0 eV—is plausible as a qualitative interpretation of the LD shift, and the spectral weight transfer is nicely demonstrated. But the central new claim, light-induced hysteresis of electronic polarization, is not established.\n\nI think this paper deserves a serious referee. The experimental work is careful enough that a revised version, with 'hysteresis' replaced by 'mirror symmetry breaking' and the octupole analysis removed, could be a publishable study of optical anisotropy in FePS3. As it stands, the main conclusion is not supported. If I were the editor, I would send it to review and ask the referee to insist on the rank-2 tensor point and on a proper test of history dependence, e.g., bidirectional angle sweeps.","headline":"Careful data, but the 'hysteresis' is a static mirror asymmetry and the octupolar cos(6θ) component is not allowed in a linear dielectric tensor; the central claim does not survive.","tokens_in":16576,"tokens_out":5394,"would_cite":false,"duration_ms":48731,"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":"The paper claims that light-induced electronic polarization in FePS3 is hysteretic at 2.0 eV but not at 1.6 eV, and that an octupolar component drives the memory effect.","keywords":["light-induced electronic polarization","hysteresis","antiferromagnet","FePS3","octupolar polarization","linear dichroism","van der Waals semiconductor","multiferroicity"],"falsifier":"Take the same FePS3 flake below TN and extract Re[ε](θ) at 2.0 eV from raw angle-resolved transmission data without the six-Gaussian transfer-function model; if no cos(6θ) component survives a direct Fourier transform of the absorbance, the octupolar origin of the hysteresis is not supported.","tokens_in":15352,"feed_emoji":"🧲","tokens_out":11867,"duration_ms":95350,"temperature":0.7,"pith_summary":"The paper reports that in the antiferromagnetic semiconductor FePS3, visible light can create an electronic polarization whose direction depends on the photon energy. Below the Néel temperature, 1.6 eV light produces polarization along the a-axis, whereas 2.0 eV light produces net polarization along the b-axis. The central observation is a hysteresis in the angle-dependent dielectric response at 2.0 eV—the response on the 0° to 90° sweep does not retrace the 90° to 180° sweep—while no such gap appears at 1.6 eV. The paper attributes the 2.0 eV hysteresis to a cos(6θ) octupolar component of the electronic polarization that, combined with the antiferromagnetic spin lattice, breaks mirror symmetry. If correct, this is a light-controlled, energy-selective electronic polarization memory in a magnetic semiconductor, a step toward multiferroic devices.","feed_headline":"Light writes a polarization memory into an antiferromagnet","feed_subtitle":"At 2.0 eV the state remembers the sweep direction; at 1.6 eV it does not—a light-controlled electronic memory.","key_machinery":"The load-bearing object is the angle-resolved real dielectric function Re[ε](θ), measured by confocal transmission spectroscopy as the electric field direction θ is swept from 0° to 180° and then over the mirror range 90° to 180°. Its Fourier decomposition into dipolar cos(2θ) and octupolar cos(6θ) components is what lets the paper separate Pa from the octupole and assign the hysteresis. The multipole analysis, together with a BCS-type temperature fit of |δmax| (TN = 117 K) and the proportionality |δmax| = 0.13·ΔRe[εb] + 0.015, carries the quantitative argument that the octupolar polarization is the driver of the hysteresis.","core_discovery":"The central claim is that light-induced electronic polarization in FePS3 is hysteretic at 2.0 eV and not at 1.6 eV, and that the hysteresis comes from an octupolar component. The authors identify dipolar (cos(2θ)) and octupolar (cos(6θ)) terms in the Fourier decomposition of Re[ε](θ); the dipolar term alone explains the 1.6 eV behavior and linear dichroism, while the octupolar term appears only near 2.0 eV and produces the hysteresis gap |δmax|. They connect the gap to the b-axis polarization Pb and ultimately to a mirror-symmetry breaking of the combined spin-lattice-plus-light state below TN = 117 K. The proposal is a new mechanism for multiferroicity: light-induced electronic multipoles, not lattice or magnetic order alone, break the symmetry needed for a switchable polarization.","pith_inferences":["Because a strictly linear rank-2 dielectric tensor can only produce a cos(2θ) angular variation, an independent probe of the proposed octupole—for example, optical second-harmonic generation or a nonlinear susceptibility measurement—would test whether the cos(6θ) term is a genuine electronic multipole.","A natural extension is to search for the same energy-selective hysteresis in other Ising-type honeycomb antiferromagnets; if the octupole mechanism is generic, the hysteresis gap should appear at the high-energy d-d transition in each material and vanish above its Néel temperature.","Time-resolved pump-probe measurements could distinguish an electronic octupole memory from a slow structural rearrangement: if the hysteresis forms within the d-d excitation lifetime, it is electronic; if it requires sustained illumination, a longer-lived lattice or spin reorganization is involved."],"forward_implications":["At 2.0 eV, sweeping the light polarization angle forward and backward in FePS3 below TN gives two distinct response paths, so the final electronic state depends on the history of the illumination angle—an optical memory element in a magnetic semiconductor.","At 1.6 eV the response is non-hysteretic, so photon energy selects between a reversible polarization channel and a memory channel in the same material.","The hysteresis gap follows the antiferromagnetic order parameter, turning on sharply below 117 K, so the effect is controlled by the magnetic state rather than by the lattice.","The magnitude of the gap is proportional to the b-axis change in Re[ε] with a fitted coefficient (0.13), giving a quantitative relation that can be checked by independent optical measurements.","The proposed octupolar mechanism implies that higher-order electronic multipoles, not just dipoles, can drive multiferroic-type symmetry breaking under light."],"supporting_citations":[{"why":"Supplies the transfer-matrix and dielectric-function modeling that converts measured transmission into Re[ε](θ), the central quantity of the hysteresis analysis.","marker":"[39]"},{"why":"Provides the Gaussian oscillator dielectric function model used in the six-Gaussian fit from which Re[ε] and Im[ε] are extracted.","marker":"[38]"},{"why":"Establishes from Raman data that no structural phase transition accompanies the magnetic ordering, supporting the claim that the anisotropy is electronic and spin-driven.","marker":"[37]"},{"why":"Assigns the ~1.7 eV absorption to trigonally allowed d-d transitions in FePS3, identifying the excitation channel that carries the light-induced polarization.","marker":"[31]"},{"why":"Supplies the d6 to d5+d7 orbital configuration and the Coulomb (U) and Hund's (J) energy scales used to split the transitions into U−J (~1.6 eV) and U+J (~2.0 eV).","marker":"[42]"},{"why":"Provides the closest prior example of light-induced electronic polarization in an antiferromagnet (Cr2O3), the effect this work extends by adding hysteresis.","marker":"[22]"},{"why":"Reports the low-temperature optical anisotropy and joint-density-of-states changes in FePS3 that the paper's LD and ΔIm[ε] data build on.","marker":"[33]"}],"fun_headline_variants":["Light-induced polarization hysteresis in FePS3","Antiferromagnet FePS3 remembers light at 2 eV","Light-induced bistable polarization in a magnetic semiconductor","2 eV light writes a polarization memory in FePS3"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the sixfold (cos(6θ)) angular variation in the measured dielectric response is a genuine electronic octupole and not a by-product of the multi-oscillator fitting or of an unmodeled nonlinear optical process; if that component is not real, the proposed origin of the hysteresis gap collapses.","fun_headline_variants_meta":{"raw":{"variants":["Light-induced polarization hysteresis in FePS3","Antiferromagnet FePS3 remembers light at 2 eV","Light-induced bistable polarization in a magnetic semiconductor","2 eV light writes a polarization memory in FePS3"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001289,"raw_usage":{"total_tokens":5260,"prompt_tokens":935,"completion_tokens":4325,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":551,"completion_tokens_details":{"reasoning_tokens":4261}},"tokens_in":551,"tokens_out":4325,"duration_ms":25901,"temperature":1.0,"reasoning_tokens":4261,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T04:35:42.217762+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same FePS3 flake below TN and extract Re[ε](θ) at 2.0 eV from raw angle-resolved transmission data without the six-Gaussian transfer-function model; if no cos(6θ) component survives a direct Fourier transform of the absorbance, the octupolar origin of the hysteresis is not supported.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the transfer-matrix and dielectric-function modeling that converts measured transmission into Re[ε](θ), the central quantity of the hysteresis analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Gaussian oscillator dielectric function model used in the six-Gaussian fit from which Re[ε] and Im[ε] are extracted."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes from Raman data that no structural phase transition accompanies the magnetic ordering, supporting the claim that the anisotropy is electronic and spin-driven."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Assigns the ~1.7 eV absorption to trigonally allowed d-d transitions in FePS3, identifying the excitation channel that carries the light-induced polarization."},{"cited_title":"Koitzsch, T","cited_arxiv_id":null,"evidence_quote":"Supplies the d6 to d5+d7 orbital configuration and the Coulomb (U) and Hund's (J) energy scales used to split the transitions into U−J (~1.6 eV) and U+J (~2.0 eV)."},{"cited_title":"Zhang, T","cited_arxiv_id":null,"evidence_quote":"Provides the closest prior example of light-induced electronic polarization in an antiferromagnet (Cr2O3), the effect this work extends by adding hysteresis."},{"cited_title":"Zhang, K","cited_arxiv_id":null,"evidence_quote":"Reports the low-temperature optical anisotropy and joint-density-of-states changes in FePS3 that the paper's LD and ΔIm[ε] data build on."}],"review_version":1}