{"id":"dfd3d4ba-d84b-4d67-b016-e036745e8093","arxiv_id":"1908.03247","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Under biaxial strain, the interlayer exciton in bilayer MoS2 shifts more than the A exciton (gauge factors -48 vs -41 meV/%), attributed to strain-tuned interlayer spacing.","lead":"Researchers stretched thin sheets of bilayer MoS2 by heating a plastic base, and watched its light emission features shift. The interlayer exciton, a particle whose electron and hole sit in different layers, shifted more than the ordinary A exciton, providing a new way to tune 2D materials.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"IL-vs-A gauge factor difference rests on a ~0.09 meV/°C slope residual after a common thermal-shift subtraction; the IL thermal shift is not separately reported.","rationale":"The reader's verdict is thoughtful and largely justified: the biaxial strain method, the reversibility checks, the spatial mapping, and the multi-flake statistics are genuine strengths, and the Poisson-effect mechanism is consistent with prior DFT. I diverge from the reader on where the weakest link lies. The reader flags the identification of the intermediate reflectance peak as the interlayer exciton. I agree that identity is an assumption, but it is an assumption carried over from strong prior work (Gerber et al. PRB 99, 035443; Slobodeniuk et al. 2D Mater. 6, 025026; Deilmann and Thygesen Nano Lett. 18, 2984). Assigning a well-known spectral feature based on such literature is standard practice and, by itself, is not the most fragile step. The more fragile step is internal: the central comparison survives only if the intrinsic thermal shift of the IL peak equals that of the A peak to within about 0.05 meV/°C. The measured total slope difference on PP is only about 0.09 meV/°C, and the manuscript subtracts a common alpha = -0.4 meV/°C without reporting the IL-specific value or its uncertainty. This is a systematic, not statistical, risk: repeating the measurement on six flakes cannot average it out. The concrete re-analysis of Fig. S15 would settle this. I also note that SI Section S6, cited for the finite-element strain-transfer estimate, contains duplicated slippage text instead of the promised description, leaving that figure unsupported; this is a presentation defect that does not directly bear on the central claim. Given the gap, I recommend CONDITIONAL rather than ACCEPT, because the paper should either report the measured alpha_IL with uncertainty or demonstrate the robustness of DeltaGF to a +/-0.1 meV/C variation in alpha_IL - alpha_A.","tokens_in":10928,"tokens_out":15432,"duration_ms":157140,"concrete_test":"Re-fit the bilayer MoS2-on-SiO2/Si temperature series (Supporting Information, Fig. S15) and extract the thermal shift of the IL peak, α_IL, with a proper uncertainty that includes fit covariance and any sample-to-sample variability. If |α_IL - α_A| ≥ 0.05 meV/°C, recompute the IL-A gauge factor difference using the individually measured α values. The headline claim is unsupported if the sign of ΔGF flips or the recomputed difference becomes comparable to its combined uncertainty. The same check should be applied to the trilayer on SiO2/Si if the IL peak is resolvable there.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Central claim: the interlayer (IL) exciton gauge factor is systematically larger than the A exciton's, ΔGF = (-48+41) = -7 meV/%. Strain is applied thermally on PP; for each exciton, GF = (s_X - α_X)/β, with s_X the measured slope on PP, α_X the intrinsic thermal slope on SiO2/Si, and β = 128e-6 per °C. Thus ΔGF = (s_IL - s_A - (α_IL - α_A))/β. The reported ΔGF implies s_IL - s_A ≈ -0.09 meV/°C. The manuscript states that on SiO2/Si 'all the excitons shift by -0.4 meV/°C' but never reports α_IL separately, nor its uncertainty. If α_IL is only 0.09 meV/°C less negative than α_A (e.g., -0.31 vs -0.40), the strain-induced difference vanishes; if more negative, it is amplified. Because the same α is subtracted for all six flakes, the sample statistics in Fig. S11/S12 cannot remove this systematic bias. The load-bearing assumption is therefore that α_IL = α_A to within ~0.05 meV/°C, a precision not documented here.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports differential reflectance measurements on mechanically exfoliated bilayer MoS2 flakes transferred onto polypropylene (PP) substrates, where biaxial strain is generated via the thermal expansion mismatch between PP and MoS2. The energies of the A, B, and interlayer (IL) excitons are extracted from Gaussian fits of spectra acquired between roughly 17 and 95 °C, converted to substrate biaxial strain using a calibrated PP expansion coefficient of 128×10^-6 K^-1, and corrected for intrinsic temperature shifts using a temperature-only control on SiO2/Si. The authors report gauge factors of (-41 ± 2) meV/% for the A exciton, (-45 ± 2) meV/% for the B exciton, and (-48 ± 4) meV/% for the IL exciton, and conclude that the IL exciton is systematically more sensitive to biaxial strain than the A exciton. They attribute this larger response to strain-tunable van der Waals interaction via the Poisson effect, which changes the interlayer distance under in-plane biaxial strain.","tokens_in":11144,"tokens_out":7911,"duration_ms":80933,"significance":"If the central claim is robust, this is a useful quantitative result for strain engineering of naturally stacked bilayer TMDs, establishing an opposite ordering of the IL and A gauge factors relative to the uniaxial-strain case and supporting a mechanism based on interlayer-distance tuning. The experimental work has clear strengths: six bilayer flakes, reversibility cycles, spatial uniformity mapping, an explicitly calibrated substrate expansion, and a finite-element strain-transfer simulation in Section S6. The authors also correctly flag that the reported gauge factors are lower bounds under the complete strain-transfer assumption. However, the headline conclusion that the IL gauge factor is systematically larger than the A gauge factor depends on a small slope difference whose systematic error budget is not documented; this is the main risk to the paper's central claim.","major_comments":[{"comment":"The claim that the IL gauge factor is systematically larger than the A gauge factor rests on a difference of about 7 meV/%, which corresponds to roughly 0.09 meV/°C in the PP-measured slopes after the common intrinsic thermal shift is subtracted. The manuscript states that on SiO2/Si \"all the excitons shift by –0.4 meV/ºC\" but does not report the IL thermal shift separately or its uncertainty. If α_IL were only 0.09 meV/°C less negative than α_A (for example -0.31 instead of -0.40 meV/°C), the strain-induced difference would vanish; if more negative, it would be amplified. Because the same α is subtracted for all six flakes, the sample statistics in Figures S11 and S12 cannot remove this systematic bias. Please report the temperature-only values for A, B, and IL separately with uncertainties (the data appear to be present in Figure S15), and show that α_IL − α_A is small compared with 0.09 meV/°C, or provide an alternative control that does not require this assumption. This is load-bearing for the central claim.","section":null},{"comment":"The text reports a flake-to-flake statistical fluctuation of 0.3-0.4 meV/ºC and a spatial variation of 0.1-0.2 meV/ºC, yet the claimed IL-A difference corresponds to only about 0.09 meV/ºC in slope. The units appear internally inconsistent because gauge factors are defined in meV/% while the quoted fluctuations are in meV/ºC. More importantly, the box-plot comparison in Figure S12 is not a quantitative significance test; the authors state that the differences are \"substantially different from zero\" without reporting per-sample differences, their uncertainties, or a paired test. Please clarify the units and provide a statistical test that explicitly accounts for both sample-to-sample variance and the systematic thermal-shift uncertainty discussed above.","section":null}],"minor_comments":[{"comment":"The same paragraph about ruling out slippage and breakdown appears at the beginning of both Section S5 and Section S6, and the text in Section S6 is cut off mid-sentence; this appears to be a copy-paste error that should be corrected.","section":null},{"comment":"The reversibility test is referenced as \"Figure S15\" in both Sections S5 and S6, but Figure S15 is the SiO2/Si control; the reversibility data are actually shown in Figure S17. Please fix the cross-reference.","section":null},{"comment":"The statement that \"all the excitons shift by –0.4 meV/ºC\" is too coarse for the paper's central claim; please give the individual A, B, and IL thermal shifts with uncertainties, rather than a single rounded value.","section":null},{"comment":"The abstract explicitly labels the A and B gauge factors as lower bounds, but the IL gauge factor is presented without the same caveat in the abstract and conclusions; please state that it is also a lower bound under the same strain-transfer assumption.","section":null}],"recommendation":"major_revision","confidential_remarks":"The load-bearing issue is the undocumented assumption that the intrinsic thermal shift of the IL exciton equals that of the A exciton to within about 0.05 meV/°C. This is fixable by re-analyzing the existing SiO2/Si control data and reporting the per-exciton thermal shifts with uncertainties, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look: this is the first biaxial strain study of the interlayer exciton in bilayer MoS2, and the data are carefully taken. The paper reports gauge factors for A, B, and IL excitons, with the IL response (-48±4 meV/%) larger than A (-41±2 meV/%). That ordering is the claim that makes the paper publishable, since it contrasts with uniaxial strain. The measurements are supported by six bilayer flakes, reversibility checks, spatial maps, and a temperature-only control on SiO2/Si. The authors also honestly label the gauge factors as lower bounds due to unknown strain transfer.\n\nThe soft spot is the thermal-shift subtraction. Strain is applied by heating the PP substrate, so the authors measure the intrinsic shift on SiO2/Si and subtract it. They state that all excitons shift by -0.4 meV/°C on SiO2/Si, but they never report the IL value separately or its uncertainty. The entire IL-vs-A ordering reduces to a slope difference of about 0.09 meV/°C after subtraction. If the IL thermal shift is just 0.09 meV/°C less negative than the A shift, the ordering disappears; if more negative, it is amplified. Because the same thermal shift is subtracted for all six flakes, the sample statistics cannot fix a systematic error in this difference. This is a real gap. It may be that the authors did measure it and found no difference, but the paper does not give us the number.\n\nThe rest holds up. The A and B gauge factors are consistent across samples, the Poisson-effect explanation is plausible and linked to prior DFT, and the citation pattern looks normal. The paper is not sloppy; it just has one under-documented assumption in the central claim.\n\nMy recommendation: engage with it, but read the SI carefully. If the authors can show α_IL ≈ α_A with uncertainty below ~0.05 meV/°C, the ordering stands. As it stands, the claim is suggestive rather than proven. A serious referee would have asked for that number.","headline":"Careful biaxial strain study with a plausible new ordering for the interlayer exciton, but the paper never isolates the IL thermal shift and the ordering could be an artifact of that subtraction.","tokens_in":11683,"tokens_out":4090,"would_cite":true,"duration_ms":40272,"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":"In bilayer MoS2, the interlayer exciton responds more strongly to biaxial strain than the A exciton does, with a gauge factor of (-48 ± 4) meV/%.","keywords":["biaxial strain","interlayer exciton","bilayer MoS2","gauge factor","Poisson effect","van der Waals interaction","differential reflectance spectroscopy","transition metal dichalcogenides"],"falsifier":"Directly measure the interlayer spacing of a biaxially strained bilayer MoS2 flake (via X-ray diffraction or electron diffraction) while tracking the intermediate reflectance peak; if the peak shifts by about -48 meV/% without any roughly 0.2%-per-1% compression of the interlayer spacing, or if the peak survives in a sample where interlayer excitons are suppressed, the central claim fails.","tokens_in":10730,"feed_emoji":"🔬","tokens_out":7172,"duration_ms":67932,"temperature":0.7,"pith_summary":"This paper reports that biaxial strain—stretching or compressing a material equally in two in-plane directions—shifts the optical resonances of bilayer molybdenum disulfide (MoS2), and that the interlayer exciton, a bound electron-hole pair whose partners live in different layers, shifts more than the ordinary intralayer A exciton. From differential reflectance spectra taken while a polypropylene substrate thermally expands or contracts, the authors extract gauge factors (energy shift per percent strain) of (-41 ± 2) meV/% for the A exciton and (-48 ± 4) meV/% for the interlayer exciton, with the latter systematically larger across several flakes. They explain the extra sensitivity through the Poisson effect: in-plane expansion shrinks the interlayer distance, modulating the van der Waals interaction that sets the interlayer exciton energy. If correct, this gives a simple, reversible knob for tuning interlayer excitons in naturally stacked bilayers without changing the material or stacking.","feed_headline":"Interlayer excitons redshift 48 meV per 1% biaxial strain","feed_subtitle":"Interlayer excitons respond to biaxial strain more strongly than A excitons, revealing strain-tunable interlayer spacing.","key_machinery":"The load-bearing object is the interlayer (IL) exciton peak in the differential reflectance spectrum of bilayer MoS2, the intermediate peak between the A and B excitons. The mechanism proposed is the Poisson effect: biaxial in-plane strain changes the out-of-plane interlayer spacing (about 0.2% per 1% in-plane strain, using an out-of-plane Poisson ratio of about 0.2), which tunes the interlayer van der Waals interaction and hence the IL exciton energy. The experimental machinery is a thermal-expansion straining setup with a polypropylene substrate, Gaussian-peak fitting of reflectance spectra, and control measurements on SiO2/Si to separate temperature effects from strain effects.","core_discovery":"The paper's central claim is that in bilayer MoS2 the interlayer (IL) exciton—an exciton whose electron and hole sit in different layers—has a biaxial-strain gauge factor of (-48 ± 4) meV/%, which is systematically larger than the A-exciton gauge factor of (-41 ± 2) meV/% (and the B-exciton's (-45 ± 2) meV/%). This contrasts with uniaxial strain, where the IL exciton was reported to shift slightly less than the A exciton. The authors attribute the difference to the Poisson effect: biaxial tension compresses the interlayer spacing, tuning the van der Waals interaction and therefore the IL exciton energy, while uniaxial strain on a polymer substrate partially counteracts this through in-plane perpendicular compression. The measurements use the thermal expansion mismatch between MoS2 and a polypropylene substrate to apply controlled biaxial strain, with the intrinsic temperature shift subtracted using control samples on SiO2/Si.","pith_inferences":["If the interlayer-exciton assignment holds, the Poisson-effect mechanism predicts that the interlayer gauge factor should scale with the out-of-plane Poisson ratio and interlayer stiffness; comparing different transition-metal dichalcogenides would test this scaling.","A spatially patterned strain field, created for example by local heating of the substrate, could write gradients in interlayer exciton energy and act as an exciton funnel in naturally stacked bilayers without heterostructure fabrication.","Since the reported values are lower bounds, a direct measurement of flake strain—using Raman phonon shifts or X-ray diffraction alongside reflectance—would sharpen the gauge factors and likely increase them.","Extending the measurement to twisted bilayers could separate the interlayer-distance contribution from the stacking-dependent contribution, because twist angle changes the interlayer registry."],"forward_implications":["Biaxial strain provides a continuous, reversible way to tune interlayer exciton energies in naturally stacked bilayer MoS2, with the interlayer exciton responding more strongly than intralayer excitons.","Because the reported gauge factors are lower bounds, the true strain sensitivity of these excitons could be even larger if strain transfer from the polypropylene substrate is imperfect.","The larger interlayer-exciton gauge factor means biaxial strain changes the energy separation between interlayer and intralayer excitons, potentially controlling their coupling or relaxation pathways.","The same thermal-expansion straining method should extend to other layered materials whose interlayer excitons are sensitive to interlayer distance, generalizing the result beyond MoS2.","The reversibility of the temperature cycling indicates that slippage or stacking changes do not occur, making the effect usable in repeated strain cycles."],"supporting_citations":[{"why":"Established from temperature-dependent spectroscopy and theory that the middle reflectance peak in bilayer MoS2 is an interlayer exciton; this is the assignment the central comparison relies on.","marker":"14"},{"why":"Supported the interlayer-exciton interpretation of the K-point fine structure in multilayers, reinforcing the peak assignment.","marker":"15"},{"why":"DFT calculation showing the interlayer exciton energy depends strongly on interlayer distance; used to argue the Poisson-effect change in spacing explains the larger gauge factor.","marker":"24"},{"why":"Uniaxial strain study of bilayer MoS2 reporting an interlayer-exciton gauge factor slightly lower than the A exciton; provides the contrast that highlights the biaxial result.","marker":"16"},{"why":"Introduced the biaxial strain method based on thermal expansion of polypropylene and reported monolayer gauge factors that serve as the comparison baseline.","marker":"30"},{"why":"Supplies the thermal expansion coefficient of MoS2 used to convert temperature changes on the substrate into biaxial strain values.","marker":"29"},{"why":"Provides the out-of-plane Poisson ratio (about 0.2) used to estimate that 1% in-plane biaxial strain changes the interlayer distance by about 0.2%.","marker":"35"},{"why":"Reports strain-tunable van der Waals interactions in few-layer black phosphorus, an analogous system supporting the proposed mechanism.","marker":"36"},{"why":"Describes the micro-reflectance spectroscopy method used to acquire the differential reflectance spectra from which exciton energies are extracted.","marker":"19"}],"fun_headline_variants":["Interlayer excitons feel biaxial strain more than A excitons","Biaxial strain tunes interlayer spacing, boosting exciton shift","Poisson effect boosts interlayer exciton strain response","Interlayer exciton gauge factor beats A exciton under biaxial strain"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The intermediate reflectance peak between the A and B excitons is assumed to be an interlayer exciton on the strength of prior work, not measured directly in this paper; if that peak were an intralayer transition, the claim that the interlayer exciton has a larger gauge factor would not be supported.","fun_headline_variants_meta":{"raw":{"variants":["Interlayer excitons feel biaxial strain more than A excitons","Biaxial strain tunes interlayer spacing, boosting exciton shift","Poisson effect boosts interlayer exciton strain response","Interlayer exciton gauge factor beats A exciton under biaxial strain"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001499,"raw_usage":{"total_tokens":5982,"prompt_tokens":884,"completion_tokens":5098,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":500,"completion_tokens_details":{"reasoning_tokens":5026}},"tokens_in":500,"tokens_out":5098,"duration_ms":40340,"temperature":1.0,"reasoning_tokens":5026,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:19:27.892302+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly measure the interlayer spacing of a biaxially strained bilayer MoS2 flake (via X-ray diffraction or electron diffraction) while tracking the intermediate reflectance peak; if the peak shifts by about -48 meV/% without any roughly 0.2%-per-1% compression of the interlayer spacing, or if the peak survives in a sample where interlayer excitons are suppressed, the central claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Established from temperature-dependent spectroscopy and theory that the middle reflectance peak in bilayer MoS2 is an interlayer exciton; this is the assignment the central comparison relies on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supported the interlayer-exciton interpretation of the K-point fine structure in multilayers, reinforcing the peak assignment."},{"cited_title":"& Thygesen, K","cited_arxiv_id":null,"evidence_quote":"DFT calculation showing the interlayer exciton energy depends strongly on interlayer distance; used to argue the Poisson-effect change in spacing explains the larger gauge factor."},{"cited_title":"& de Vasconcellos, S","cited_arxiv_id":null,"evidence_quote":"Uniaxial strain study of bilayer MoS2 reporting an interlayer-exciton gauge factor slightly lower than the A exciton; provides the contrast that highlights the biaxial result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced the biaxial strain method based on thermal expansion of polypropylene and reported monolayer gauge factors that serve as the comparison baseline."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the thermal expansion coefficient of MoS2 used to convert temperature changes on the substrate into biaxial strain values."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the out-of-plane Poisson ratio (about 0.2) used to estimate that 1% in-plane biaxial strain changes the interlayer distance by about 0.2%."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports strain-tunable van der Waals interactions in few-layer black phosphorus, an analogous system supporting the proposed mechanism."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the micro-reflectance spectroscopy method used to acquire the differential reflectance spectra from which exciton energies are extracted."}],"review_version":1}