{"id":"40a39d6e-7696-4497-b7f6-e4913ee6b96e","arxiv_id":"2505.02454","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Monolayer MoS2 on an SBN relaxor ferroelectric substrate shows reversible photoluminescence enhancement and electron density reduction as the substrate crosses its ferroelectric-to-paraelectric transition.","lead":"Researchers placed a single layer of MoS2 onto a relaxor ferroelectric crystal (SBN) and found that heating through the crystal's phase transition reversibly brightens the MoS2's light emission and lowers its electron density. The effect offers a temperature-based, electric-field-free way to tune electronic properties of 2D materials, with built-in memory from thermal hysteresis.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported 14 to 8 microcoulomb/cm² decrease is not established because Eq. (1) is inverted on a PL spectrum averaged over oppositely polarized domains, making the extracted value a model-dependent and possibly spurious net rather than the actual doping change.","rationale":"The paper provides a clean empirical observation: 1L-MoS2 on SBN shows a temperature-dependent exciton/trion conversion that correlates with the SHG-derived transition, and the SiO2 control supports the attribution to the substrate. The hysteresis and reversibility are also real features of the data. However, the quantitative central claim, 14 to 8 microcoulomb/cm², requires converting a spatially averaged PL ratio into an electron density. The paper itself acknowledges the averaging over anti-parallel domains and the opposite signs of the expected doping changes, but it does not model how the average of a nonlinear mass-action response should be inverted. Applying the single-domain formula to the averaged spectrum is therefore not justified without additional information about domain area fractions and per-domain radiative yields. The reader's conditional verdict is appropriate: the concern is substantial but addressable with a quantitative two-domain model or spatially resolved measurements. I do not see a basis for outright rejection because the qualitative trend and its correlation with the phase transition are plausible and partially supported by control and SHG data.","tokens_in":13813,"tokens_out":12722,"duration_ms":163633,"concrete_test":"Perform a two-domain nonlinear inversion of the data: model the measured ratio as R(T) = [f I_-^+(n_+) + (1-f) I_-^-(n_-)] / [f I_+^+(n_+) + (1-f) I_+^-(n_-)] with n_+ = n_0^+ - delta(T) and n_- = n_0^- + delta(T), using the same mass-action and eta_r values. Fit f, n_0^+, n_0^-, and delta(T) to the experimental R(T) and the hysteresis loop. If no solution exists with 0.3 < f < 0.7 and a single physical delta(T) that vanishes above TC, the net decrease is an averaging artifact. As a complementary experiment, resolve the PL (or Kelvin probe force microscopy) on individual domains with sub-500 nm spatial resolution across 30-90 degree Celsius to verify that the trion/exciton ratio changes in opposite directions on the two domain orientations.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing weakness is the conversion of the averaged PL ratio into a net electron density. The paper states that the SBN substrate contains anti-parallel domains whose electron-density changes have opposite signs across the phase transition, and that the ~2 micrometer PL spot averages over ~500 nm domains. Equation (1), based on mass action, is a nonlinear relation between I_X-/I_X and n_e. Inverting it on an averaged spectrum therefore does not yield the spatial average of n_e; it yields a mixed functional of the two domain populations and their radiative yields. The observed drop from 14 to 8 microcoulomb/cm² is thus not directly readable from the data. To obtain a net decrease from a model with two opposite-sign Delta n_e values, one must specify either unequal domain area fractions inside the spot or a particular nonlinearity in the intensity response; neither is provided. In addition, the constant eta_r = 20/3 and fixed E_b = 20 meV are assumed over the full 30-90 degree Celsius range without independent verification, so the entire quantitative scale and sign of Delta sigma depend on unmeasured parameters. The plateau above TC and the hysteresis in Figure 4 are empirical, but their designation as electron-density modulation relies on this insecure inversion.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports temperature-dependent photoluminescence (PL) measurements of monolayer MoS₂ transferred onto a relaxor ferroelectric Sr₀.₆₁Ba₀.₃₉Nb₂O₆ (SBN) substrate. As the SBN undergoes its ferro-to-paraelectric phase transition between 30 and 90 °C, the authors observe a reversible PL enhancement and a trion-to-exciton conversion, correlated with the substrate's transition temperature independently determined by second-harmonic generation. Using a mass-action model (Eq. 1) with literature parameters, they convert the exciton/trion intensity ratio into an absolute electron density, reporting a decrease from about 14 µC/cm² at room temperature to about 8 µC/cm² above the transition, with thermal hysteresis between heating and cooling cycles. The modulation is attributed to the vanishing spontaneous polarization of the SBN substrate, which changes the screening-charge balance at the MoS₂/SBN interface. The authors compare with a SiO₂ substrate to separate intrinsic MoS₂ thermal effects from the phase-transition effect.","tokens_in":14020,"tokens_out":8206,"duration_ms":103344,"significance":"If the quantitative interpretation is accepted, this work introduces relaxor ferroelectrics as a reconfigurable platform for gradually tuning the electronic and optical properties of monolayer TMDs, with potential applications in temperature-responsive optoelectronics and memories. The qualitative phenomenology—PL enhancement and trion-to-exciton conversion linked to the independently measured transition temperature—is well supported by the SiO₂ comparison. The paper also includes a thoughtful discussion of the polydomain averaging effect, which is a strength. The potential significance is genuine but rests on the reliability of the absolute electron-density extraction.","major_comments":[{"comment":"The quantitative extraction of n_e from a PL spectrum that averages over anti-parallel ferroelectric domains is not justified. The paper states that the ~2 μm PL spot averages over ~500 nm domains of opposite polarization, and that the recorded spectrum is an average of the emissions from both domain types. Since Eq. (1) is linear in n_e for each individual domain but is applied to the intensity-weighted average of the two populations, the extracted n_e is a weighted mean whose weights (the exciton PL intensities) depend on the doping level of each domain. Consequently, the reported decrease from about 14 to about 8 µC/cm² cannot be directly read as a net electron-density change; it could be substantially influenced by the temperature-dependent redistribution of the weights. The paragraph acknowledging the polydomain structure states that the extracted net variation is smaller than the single-domain polarization change, but this is only a qualitative statement and does not quantify the weighting effect. To support the quantitative claim, the authors should either perform a domain-resolved measurement or construct a explicit two-domain model with area fractions and intensity weights, and show that the extracted Δσ is robust to reasonable assumptions about these parameters.","section":"Eq. (1) and Fig. 3a"},{"comment":"The assumption that the relative quantum yield η_r = 20/3 and the trion binding energy E_b = 20 meV are constant over the full 30–90 °C range is unjustified. The extracted n_e is proportional to η_r and depends exponentially on E_b/k_BT through C(T). If either parameter varies with temperature, the inferred n_e and hence the reported Δσ will be in error. The paper states that η_r is 'kept fixed' but provides no sensitivity analysis or literature support for the temperature independence of these quantities. At minimum, the authors should discuss the possible temperature dependence of η_r and E_b and estimate the resulting uncertainty in the reported electron-density values.","section":"Eq. (1) and the paragraph following it"},{"comment":"The sentence 'Similar ∆σ values have been obtained for a different 1L-MoS2 with higher initial electron doping' is presented as a reproducibility check, but no data for this second sample are shown in any figure or table. This claim is currently not verifiable; it should either be documented with a figure or removed from the text.","section":"Results, reproducibility statement"}],"minor_comments":[{"comment":"Equation (1) is typeset incorrectly: the ratio I_{X^-}/I_X is missing the division slash, making the equation appear as a product. It should read I_{X^-}/I_X = n_e/(η_r C(T)).","section":"Eq. (1)"},{"comment":"The word 'poli-domain' in the paragraph beginning 'A point that should be addressed...' is a typo; it should be 'polydomain'.","section":"Polydomain paragraph"},{"comment":"The caption contains the typo 'paralectric' for 'paraelectric'.","section":"Figure 3b caption"},{"comment":"Figure 3b caption states 'The net electron doping of 1L-MoS2 is not represented', yet Figure 3a presents a curve labeled as electron doping. The relationship between the plotted quantity in Fig. 3a and the statement in Fig. 3b should be clarified; if the plotted n_e is a domain-averaged value, this should be stated explicitly for both figures.","section":"Figure 3a vs. Figure 3b"},{"comment":"The sign convention for Δσ in Figure 4 is not defined. The text says 'the gradual decrease in Δσ is observed until it becomes zero', implying Δσ is a positive quantity that vanishes above TC; please define Δσ and its sign clearly in the figure caption or text.","section":"Figure 4"},{"comment":"Reference 19 appears to have an incorrect volume number: 'Nano Lett. 2, 959 (2021)' should probably be 'Nano Lett. 21, 959 (2021)'.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a fresh direction—using relaxor ferroelectrics to modulate TMD properties—and the qualitative PL/trion-exciton correlation with the phase transition is convincing. The main weakness is the absolute electron-density extraction, which the authors partially acknowledge but do not quantitatively resolve. The second-sample reproducibility claim and the fixed-parameter assumption add to the uncertainty. I believe the paper can become publishable after the quantitative claims are properly qualified or supported, hence major_revision rather than reject."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe first thing you should know: this is the first TMD-on-relaxor-ferroelectric study, and the central observation—PL of monolayer MoS2 changes reversibly and with hysteresis as the SBN substrate goes through its diffuse phase transition—is solid. The comparison with SiO2, the SHG-determined Tc, and the absence of anomalies in exciton energy and linewidth all support the idea that the effect is electronic doping from the substrate, not a thermal artifact.\n\nWhat the paper does well: it picks a good control (SiO2), measures the phase transition independently with SHG, shows the trion-to-exciton conversion, and is upfront about the domain-averaging problem. The authors even state that their extracted net charge variation is smaller than the single-domain polarization change. That honesty counts.\n\nThe soft spot is the conversion from PL ratio to absolute electron density. Equation (1) is a mass-action relation that is nonlinear in n_e. The PL spot averages over ~2 µm, containing many ~500 nm anti-parallel domains. Inverting the averaged spectrum through Eq. (1) does not give the spatial average n_e; it gives a mixed functional of the two domain populations. The paper acknowledges the averaging but then still plots the result as \"electron density in 1L-MoS2\" (Figure 3a) and reports 14 to 8 µC/cm². Without a model for the domain area fractions or a single-domain measurement, those numbers are not established. The fixed eta_r = 20/3 and E_b = 20 meV over the entire 30–90°C range add further uncertainty. This is a real flaw in the quantitative claim, but not in the qualitative one.\n\nMinor issues: the second flake with higher initial doping is mentioned without data; the sentence about the sign of doping change on positive/negative domains reads inconsistently, though the intended meaning is clear.\n\nBottom line: the paper deserves a serious referee. The new substrate class is interesting, and the qualitative physics is likely right. The referee should push for a domain-resolved measurement, error bars on n_e, or at least a reanalysis that does not invert Eq. (1) on the averaged spectrum. I would send it out, not desk-reject it.","headline":"First TMD/relaxor-ferroelectric study with a solid qualitative PL observation, but the absolute doping numbers rest on an unjustified inversion of a domain-averaged spectrum.","tokens_in":14592,"tokens_out":3907,"would_cite":true,"duration_ms":47269,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["78.55.-m","77.80.-e","78.67.-n"],"model":"deepseek-v4-flash","headline":"Monolayer MoS2 on a relaxor ferroelectric loses about 6 µC/cm² of electron density as the substrate turns paraelectric, reversibly and with thermal memory.","keywords":["monolayer MoS2","relaxor ferroelectric","strontium barium niobate","photoluminescence","trion-exciton ratio","electron doping modulation","phase transition","thermal hysteresis"],"falsifier":"Perform Hall-effect or capacitance-based carrier-density measurements on the same MoS2/SBN stack while sweeping 30–90 °C and check whether the carrier density drops by about $6\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$ with the same hysteresis; a reversible change of comparable size and sign would support the mechanism, while no change or an opposite sign would falsify it. A second check is to measure the photoluminescence at sub-500 nm resolution on a single ferroelectric domain, where the model predicts larger and opposite-sign doping swings on up versus down domains than the averaged value.","tokens_in":1944,"feed_emoji":"⚡","tokens_out":3675,"duration_ms":138589,"temperature":0.7,"pith_summary":"The paper establishes that a single layer of molybdenum disulfide (MoS$_2$) transferred onto a strontium barium niobate (SBN) relaxor ferroelectric substrate changes its electron density as the substrate passes through its ferroelectric-to-paraelectric phase transition. The measured electron doping falls from roughly $14\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$ at room temperature to about $8\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$ above the transition near 73 °C, and the change reverses on cooling with thermal hysteresis. The claim matters because it provides a contact-free, temperature-driven way to continuously dope a two-dimensional semiconductor using the substrate's own vanishing spontaneous polarization, without electric gates, strain, or chemical treatments. Alongside the doping change, the monolayer's photoluminescence brightens and shifts from trion-dominated to exciton-dominated emission, which the authors attribute to the altered charge balance at the MoS$_2$/SBN interface.","feed_headline":"Heating a crystal substrate shifts MoS2 electron density by 6 µC/cm²","feed_subtitle":"2D MoS2 loses electrons as the crystal turns paraelectric, reversibly and with memory.","key_machinery":"The operative mechanism is the charge balance at the MoS$_2$/SBN interface: in the ferroelectric phase, bound polarization charge on the SBN surface is partially screened by charges in the monolayer, and as the spontaneous polarization $P_S$ vanishes across the relaxor transition, that bound charge disappears and re-weights the electron population in the MoS$_2$. The quantitative tool is a mass-action formula, Eq. (1), that converts the measured trion-to-exciton intensity ratio into an absolute electron density $n_e$ using the trion binding energy ($E_b \\simeq 20$ meV), MoS$_2$ effective masses, and a fixed relative quantum yield $\\eta_r = 20/3$. This conversion is combined with independent second-harmonic-generation tracking of the SBN phase transition and with a SiO$_2$-supported MoS$_2$ reference sample that supplies the temperature baseline, so the reported doping change is attributed to the substrate transition rather than to ordinary thermal effects.","core_discovery":"The central claim is that the relaxor character of SBN, a smeared phase transition instead of a sharp Curie point, turns the substrate into a continuously tunable electrostatic gate for monolayer MoS$_2$. As the substrate is heated from 30 °C through the transition at about 73 °C, its spontaneous polarization gradually vanishes; the screening-charge balance at the interface changes; and the net electron density in the MoS$_2$ monolayer decreases from about $14\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$ ($9\\times10^{13}\\ \\mathrm{e}/\\mathrm{cm}^2$) to about $8\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$, while the exciton-to-trion photoluminescence ratio rises monotonically and then plateaus above the transition. The modulation is reversible on cooling and exhibits hysteresis, which the paper traces to the known thermally hysteretic phase transition of SBN. Because the optical measurement averages over oppositely polarized ferroelectric domains, the authors argue that the observed net change of about $6\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$ is a lower bound on the local polarization-driven doping change, comparing it with spontaneous-polarization values of 15–20 $\\mu\\mathrm{C}/\\mathrm{cm}^2$ reported for single-domain SBN.","pith_inferences":["The same mechanism should transfer to other monolayer transition-metal dichalcogenides whose emission shows resolvable exciton and trion peaks, such as WS$_2$ or WSe$_2$, although their trion binding energies and band alignments will shift the quantitative doping scale.","Because relaxor transition temperatures depend on composition, choosing a different SBN stoichiometry or another relaxor ferroelectric could move the doping-modulation window to higher or lower temperatures, extending the idea beyond the 30–90 °C range reported here.","If the local polarization state can be switched optically or electrically, the bistability seen in the heating-cooling cycle suggests a route to persistent, locally patterned doping in the monolayer; this goes beyond what the paper demonstrates.","A transport measurement across the same temperature range should reveal a comparable reversible change in sheet carrier density or conductivity if the polarization-charge mechanism dominates, giving a non-optical check that the PL-derived densities do not provide."],"forward_implications":["Monolayer MoS$_2$ on SBN can be continuously electron-doped by temperature alone, from about $14$ to about $8\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$, across the 30–90 °C range, with no gate electrode or chemical processing.","The doping modulation is reversible and hysteretic, so the same heterostructure can hold two different electron densities at the same temperature depending on whether it was heated or cooled, a property suited to memory or synaptic-type devices.","Because the phase transition raises exciton emission while leaving trion emission nearly constant, the exciton-to-trion ratio can serve as an all-optical thermometer or a probe of the relaxor transition.","The net measured charge swing being smaller than the spontaneous polarization of single-domain SBN implies that on a single ferroelectric domain the local doping change in MoS$_2$ is expected to be larger than the observed $6\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$."],"supporting_citations":[{"why":"Establishes the prior ferroelectric photodoping behavior of monolayer MoS2 that motivates the SBN experiment and the choice of excitation power.","marker":"[20]"},{"why":"Provides pyroelectric-gating context and the MoS2 effective masses used in the temperature function C(T) of Eq. (2).","marker":"[22]"},{"why":"Documents thermal hysteresis across the SBN phase transition, which the paper invokes to explain the observed hysteresis in MoS2 doping.","marker":"[28]"},{"why":"Reports the low Curie temperature near 70 °C for near-identical SBN composition, used to identify the transition range.","marker":"[29]"},{"why":"Supplies the mass-action formula and the relative quantum yield eta_r = 20/3 used to convert trion/exciton ratios into electron density.","marker":"[43]"},{"why":"Provides the trion binding energy Eb about 20 meV required by Eq. (2).","marker":"[46]"},{"why":"Supports the claim that exciton PL is strongly doping-sensitive while trion PL is not, which is central to the domain-averaging interpretation.","marker":"[47]"},{"why":"Gives spontaneous-polarization values of 15-20 microcoulombs per square centimeter for close SBN compositions, used as the comparison for the observed net charge change.","marker":"[48]"}],"fun_headline_variants":["Relaxor ferroelectric substrate gives continuous MoS2 doping control","MoS2 on relaxor ferroelectric: continuous electron tuning with thermal memory","Heating sweeps MoS2 doping from 14 to 8 µC/cm² continuously","Relaxor ferroelectric turns heating into a smooth MoS2 doping knob","MoS2 electron doping tunes smoothly across a relaxor ferroelectric's transition"],"cache_read_input_tokens":16768,"weakest_assumption_plain":"The load-bearing premise is that the trion-to-exciton brightness ratio follows a fixed mass-action conversion with a temperature-independent relative quantum yield of 20/3, and that the collected photoluminescence is a fair average over the up- and down-polarized ferroelectric domains; if either of those fails, the absolute electron densities and the sign and size of the reported modulation are not reliable.","fun_headline_variants_meta":{"raw":{"variants":["Relaxor ferroelectric substrate gives continuous MoS2 doping control","MoS2 on relaxor ferroelectric: continuous electron tuning with thermal memory","Heating sweeps MoS2 doping from 14 to 8 µC/cm² continuously","Relaxor ferroelectric turns heating into a smooth MoS2 doping knob","MoS2 electron doping tunes smoothly across a relaxor ferroelectric's transition"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000963,"raw_usage":{"total_tokens":4151,"prompt_tokens":1048,"completion_tokens":3103,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":664,"completion_tokens_details":{"reasoning_tokens":2996}},"tokens_in":664,"tokens_out":3103,"duration_ms":24377,"temperature":1.0,"reasoning_tokens":2996,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:51:31.752902+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform Hall-effect or capacitance-based carrier-density measurements on the same MoS2/SBN stack while sweeping 30–90 °C and check whether the carrier density drops by about $6\\ \\mu\\mathrm{C}/\\mathrm{cm}^2$ with the same hysteresis; a reversible change of comparable size and sign would support the mechanism, while no change or an opposite sign would falsify it. A second check is to measure the photoluminescence at sub-500 nm resolution on a single ferroelectric domain, where the model predicts larger and opposite-sign doping swings on up versus down domains than the averaged value.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the prior ferroelectric photodoping behavior of monolayer MoS2 that motivates the SBN experiment and the choice of excitation power."},{"cited_title":"Pyroelectric doping reversal of MoS2 p-n junctions on ferroelectric domain walls probed by photoluminescence","cited_arxiv_id":"2504.04886","evidence_quote":"Provides pyroelectric-gating context and the MoS2 effective masses used in the temperature function C(T) of Eq. (2)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the low Curie temperature near 70 °C for near-identical SBN composition, used to identify the transition range."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the mass-action formula and the relative quantum yield eta_r = 20/3 used to convert trion/exciton ratios into electron density."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the claim that exciton PL is strongly doping-sensitive while trion PL is not, which is central to the domain-averaging interpretation."}],"review_version":1}