{"id":"83854a8a-be95-4a2f-b977-13e2e21c6c9f","arxiv_id":"2506.03985","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Jackiw-Rebbi interface states are realized in stacked van der Waals WS2 gratings and used to directionally enhance exciton emission from an embedded WSe2 monolayer.","lead":"This paper demonstrates topologically protected Jackiw-Rebbi interface states in photonic gratings made from layered WS2, with light confined to the interface between two gratings and emitted directionally. It also couples light emission from a WSe2 monolayer to these states, boosting directional photoluminescence by up to 22 times.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Near-field localization evidence depends on per-side median leveling; without raw s-SNOM data the interface peak could be a normalization artifact, and this is the main load-bearing weakness.","rationale":"The paper has multiple independent pillars: far-field angle-resolved reflectance shows a mid-gap interface state absent from either grating alone, FDTD simulations localize the field at the interface, and an additional sample reproduces the far-field result. The topological interpretation via band inversion and BIC flipping is standard and consistent with the Dirac-model analysis. The weakest link is the experimental demonstration of spatial confinement, which relies on s-SNOM images processed with per-side median leveling. The paper itself flags this normalization in Supplementary Note 5 but does not show raw data, so the possibility of an artificial contrast step at the boundary cannot be excluded. The wavelength scan is helpful but not definitive because the material-response contrast is wavelength-dependent. This is the same concern the reader identified as the weakest assumption, and it is genuinely load-bearing for the abstract's claim of strong spatial confinement. Other issues, such as the absence of a direct robustness test and the reference-region choices in the PL enhancement, are secondary; they support a conditional verdict rather than a rejection. A straightforward raw-data comparison and a trivial-interface control would settle whether the near-field peak is a true JR state signature.","tokens_in":22740,"tokens_out":12760,"duration_ms":134173,"concrete_test":"Provide the raw (un-normalized) s-SNOM amplitude maps or line profiles across the interface for at least 730, 736, and 742 nm. Recompute the normalized profiles without per-side median leveling, e.g., by subtracting a single global baseline fitted over the whole scan, and check whether a peak at x=0 persists only at 736 nm. As a control, measure a structure with two identical gratings (trivial interface) under identical illumination and processing; a true JR state should produce no mid-gap peak in either raw or processed data, while a normalization artifact would produce a step at the boundary.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section V and Supplementary Note 5 describe how the s-SNOM amplitude images are 'median line leveled separately either side of the grating interface' to suppress the different intrinsic responses of the high- and low-FF gratings. This per-side normalization subtracts a different baseline from each half of each row. If the raw scattering baseline changes smoothly across the interface, the normalization itself creates a step/edge feature at x=0, and a genuine interface-localized enhancement would be superimposed on that artificial step. The wavelength scan in Fig. 5(e), showing a peak only near 736 nm, mitigates but does not eliminate the concern: the dielectric contrast between the two gratings is itself wavelength-dependent, so a wavelength-dependent baseline artifact cannot be excluded. No raw, un-normalized near-field maps are shown, and the explanation in Supplementary Note 5 explicitly acknowledges that the total signal combines intrinsic material properties with local fields. Because the abstract lists real-space confinement as a headline result, this is the most load-bearing part of the experimental evidence.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the design, fabrication, and characterization of one-dimensional topological photonic gratings made from multilayer WS2 on gold. The authors use an effective non-Hermitian Dirac Hamiltonian to argue that two grating regions with different filling factors realize opposite topological phases through band inversion of a symmetry-protected BIC, and that their interface hosts a Jackiw-Rebbi mid-gap state. Far-field angle-resolved reflectance contrast shows a mid-gap mode at the interface with a 10 meV linewidth and an 8.0° angular bandwidth. Near-field s-SNOM at 736 nm shows an interface-localized scattering peak that disappears at other wavelengths, in agreement with FDTD simulations. Finally, in a five-layer heterostructure with an embedded hBN-encapsulated WSe2 monolayer, angle-resolved PL measurements show directional enhancement up to 22× relative to an unpatterned reference region.","tokens_in":22823,"tokens_out":4982,"duration_ms":50872,"significance":"If the claims hold, this is a significant advance: it brings topological interface states to compact (<100 nm) van der Waals photonic devices, demonstrates active emitter coupling to a JR state, and provides a transfer-stack fabrication route. The theoretical framework is standard and the simulations use independently measured optical constants, so the JR state energy is not an ad hoc fitted parameter. The far-field band-inversion evidence is supported by a second sample (Supp. Note 3), and the PL analysis is documented in detail. The main risk is the near-field normalization procedure, which is load-bearing for the real-space localization claim.","major_comments":[{"comment":"The s-SNOM images in Fig. 5(d–e) are median line-leveled separately on each side of the grating interface to suppress the different intrinsic material responses of the high- and low-FF gratings. This procedure subtracts a different baseline from the two halves of each scan line; if the raw baseline varies smoothly across the interface (e.g., due to the different effective permittivity of the two gratings), the leveling itself creates an artificial step at x=0. Because the intrinsic material response of the two gratings is wavelength-dependent, the wavelength scan in Fig. 5(e) cannot by itself exclude a normalization artifact. The authors should show the raw, un-leveled s-SNOM maps or a control region processed with the same per-side leveling, and quantify the magnitude of the leveling step relative to the claimed interface enhancement. This is needed to support the abstract's claim of spatial confinement.","section":"Section V and Supplementary Note 5"},{"comment":"The 22× directional enhancement factor in Fig. 6(e) is computed using a reference PL signal collected without the variable aperture, and Supplementary Note 6 states that this reference contains weak grating-mode dispersion (Fig. S5(f)). If the no-aperture reference includes any grating-coupled emission within the chosen integration region, the denominator is not a clean uncoupled-monolayer baseline and the enhancement factor is inflated. The authors should either recompute the enhancement using the aperture-based reference (Fig. S5(d)) or quantitatively show that the weak dispersion in the no-aperture reference contributes negligibly to the integrated signal.","section":"Section VI and Supplementary Note 6"},{"comment":"The central identification of the interface state as topologically protected is inferred from the observed band inversion (BIC on the upper vs lower branch) through the effective Hamiltonian mapping of Refs. [11,33], rather than from a direct calculation of the Zak phase or surface impedance for the actual fabricated parameters. The authors should state explicitly whether the topological invariant is computed for the fabricated structure (a1=279 nm, F1=0.81; a2=319 nm, F2=0.41) or only for the idealized Hamiltonian with φ=0/π. If only the latter, this limitation should be acknowledged, or an independent check (e.g., a direct Zak-phase calculation from the RCWA modes) should be provided to support the term 'topologically protected JR state'.","section":"Section IV and Figure 4(a)"}],"minor_comments":[{"comment":"The notation |m(x)·x| is dimensionally inconsistent because m is a scalar mass parameter, not a vector; this should be written as |m(x)| |x| or m0 |x| in the exponential.","section":"Section II, Eq. (5)"},{"comment":"The wavevector normalization for the interface region uses the average of the two grating periods; this should be stated explicitly in the main text rather than only in the figure caption.","section":"Section IV, Fig. 4(b) caption"},{"comment":"The label 'uncoupled monolayer' for the PL reference is misleading, since the reference consists of an hBN-encapsulated monolayer between two unpatterned WS2 slabs; 'unpatterned' or 'ungrated' would be clearer.","section":"Section VI and Supp. Note 6"},{"comment":"The text states that 'complete background removal' was obtained, but the data are still median line-leveled; the distinction between interferometric background removal and the per-side leveling procedure should be clarified to avoid confusion.","section":"Section V, s-SNOM description"},{"comment":"The s-SNOM image lacks a scale bar; adding one would help the reader judge the spatial extent of the interface-localized peak relative to the grating periods.","section":"Fig. 5(d)"}],"recommendation":"major_revision","confidential_remarks":"The main technical uncertainty is whether the near-field localization evidence survives a re-analysis with raw, un-leveled data. The far-field evidence for a mid-gap interface state is solid, and the PL study is carefully documented apart from the reference-baseline concern. I would condition acceptance on addressing the near-field normalization and PL reference issues; the paper is otherwise well within the scope of the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper is a solid, well-executed demonstration of a photonic Jackiw-Rebbi interface state in a sub-100nm WS2 inverted grating platform, with the added bonus of monolayer WSe2 coupling. The main weaknesses are the s-SNOM normalization and the absence of a direct robustness test, but neither guts the central observation.\n\nThe genuinely new thing here is the material platform. Previous JR states in dielectric gratings used SiN, TiO2, or perovskite, with thicknesses in the micron range. The 'inverted grating' idea—adding a top WS2 slab to reduce the effective index contrast so the band gap can close and reopen—is clever and experimentally enabled by vdW stacking. The fabrication looks careful: exfoliation, e-beam patterning, transfer, with two separate samples showing reproducible band inversion and mid-gap state. The simulations use independent optical constants from the literature, and the JR state energy emerges from RCWA/FDTD rather than from fitting the reflectance, so the circularity burden is low.\n\nThe far-field reflectance data are the strongest evidence: a clear mid-gap interface state appears only when measuring at the interface, with a 10 meV linewidth and an 8° angular bandwidth in kx. The PL experiment is also a nice step: the WSe2 monolayer couples to the JR state, and the directional enhancement factor of ~22x, while dependent on integration choices, is supported by a careful reference measurement.\n\nNow the soft spots. First, the s-SNOM localization evidence in Fig. 5 is weakened by the per-side median leveling described in Supp. Note 5. Subtracting a different baseline from each half of each row can itself create a step at the interface, and raw, un-normalized maps are not shown. The wavelength scan—peak only at 736 nm—and agreement with the simulated field profile mitigate this, but they do not fully eliminate the concern that at least part of the contrast could be a boundary artifact. This is the most load-bearing caveat for the claim of real-space confinement.\n\nSecond, the 'topologically protected' label is inferred from the observed band inversion and the standard bulk-boundary correspondence. The paper does not directly test robustness to disorder or perturbations, which is the operational meaning of protection. Given the literature on photonic JR states, this is a common approach, but the claim should be stated as inferred.\n\nThird, the key numbers—linewidth, angular bandwidth, and the 22x enhancement—are quoted without experimental uncertainties. The fits look reasonable, but confidence intervals would help.\n\nWho is this for? Researchers in topological photonics and 2D-material nanophotonics. The platform is compact and integrates an active emitter, which makes it potentially useful for on-chip directional light sources. I'd send it to a serious referee; the concerns are addressable in revision.","headline":"Compact WS2 inverted gratings yield a credible photonic JR state with active emitter coupling; near-field normalization and missing robustness test are the main caveats.","tokens_in":23531,"tokens_out":3554,"would_cite":true,"duration_ms":33933,"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":"The paper reports Jackiw-Rebbi topological interface states in stacked WS2 photonic gratings, localised at the boundary, with 10 meV linewidth and directional enhancement of coupled WSe2 emission up to 22 times.","keywords":["Jackiw-Rebbi states","topological photonics","van der Waals heterostructures","WS2 gratings","bound states in the continuum","near-field optical microscopy","photoluminescence enhancement","photonic crystals"],"falsifier":"Re-measure the s-SNOM signal at 736 nm without the separate median line-leveling on the two grating sides; if the interface enhancement disappears in the raw scattering amplitude, the spatial-confinement claim collapses. A complementary check is to fabricate the same double-grating interface between two topologically identical gratings and show that no mid-gap interface peak appears, or to scale the grating period and confirm the interface peak tracks the predicted JR energy.","tokens_in":22459,"feed_emoji":"🔬","tokens_out":7458,"duration_ms":67775,"temperature":0.7,"pith_summary":"The paper claims that a one-dimensional photonic crystal made from stacked layers of quasi-bulk WS2 can host a Jackiw-Rebbi interface state between two gratings with opposite topological phase, working at near-infrared wavelengths around 750 nm. It supports this with far-field angle-resolved reflectance showing a mid-gap mode of 10 meV linewidth and 8.0 degrees angular bandwidth, and with near-field scanning microscopy showing the mode confined to the grating interface. It then embeds a monolayer WSe2 emitter in the stack and reports directional photoluminescence enhancement up to 22 times that of the uncoupled monolayer. If correct, this shows that van der Waals materials can be stacked into compact topological photonic devices that both localise and directionally emit light.","feed_headline":"Stacked WS2 gratings host topological Jackiw-Rebbi light states","feed_subtitle":"A 100-nm-thick van der Waals stack confines light at the boundary and boosts monolayer WSe2 emission by 22 times.","key_machinery":"The mechanism is a non-Hermitian Dirac-like Hamiltonian for two counter-propagating guided modes of a 1D grating, with diffractive coupling $Je^{i\\varphi}$ and radiative loss $\\gamma$. Mirror symmetry restricts $\\varphi$ to $0$ or $\\pi$; changing the filling factor flips this phase, which inverts the band structure and changes the Zak phase from 0 to $\\pi$. After a unitary transform the Hamiltonian becomes a 1D Dirac equation whose mass term $m(x)$ changes sign at the grating interface, and the Jackiw-Rebbi solution localises exactly there. The experimental enabler is the inverted grating, a design where a top WS2 slab reduces the effective index contrast so the gap can actually close and reopen, something a bare high-index WS2 grating cannot do.","core_discovery":"The paper's central claim is that a photonic Jackiw-Rebbi state appears at the interface between two topologically distinct gratings etched into quasi-bulk WS2, and that this state can be seen in far-field reflectance, near-field scattering, and emitter photoluminescence. A top WS2 slab on the grating lowers the effective refractive-index contrast enough that the filling factor closes and reopens the photonic gap, causing band inversion; the boundary between a high-filling-factor and a low-filling-factor grating then carries a mid-gap mode at about 1.68 eV. The mode has a 10 meV linewidth and an 8.0 degree angular bandwidth perpendicular to the grooves, is spatially localised at the interface in s-SNOM scans, and couples to an embedded hBN-encapsulated WSe2 monolayer to give a directional enhancement of up to 22 times over the uncoupled emitter.","pith_inferences":["Moving the emitting monolayer from the grating layer into the top slab, where the simulated JR field is strongest, could push the directional enhancement above 22; this is a natural next experiment.","The inverted-grating design should transfer to other high-index van der Waals materials and to other wavelengths by rescaling period and thickness, with the same reflectance-contrast signature as a check.","If raw s-SNOM data, before the separate line-leveling, still show the interface peak, the platform becomes a general near-field testbed for topological phase boundaries in layered photonic crystals.","The strong directivity difference between the kx and ky directions suggests the same structure could act as a one-dimensional free-space collimator for integrated van der Waals light sources, a use the paper does not demonstrate."],"forward_implications":["Topological photonic interface states can be made in sub-100 nm van der Waals stacks, removing the need for micrometre-thick gratings.","A monolayer emitter placed at the interface gains a directional emission channel with up to 22-fold enhancement into a narrow angular range.","The same transfer-stamping fabrication can place additional van der Waals layers on pre-etched gratings without lattice matching or chemical bonding.","The JR state's narrow linewidth and mid-gap position make the structure work as a compact directional filter or narrow-linewidth light source.","The measured real-space, k-space, and energy localisation together establish the state as a genuine interface-bound mode accessible to near-field probes."],"supporting_citations":[{"why":"Supplies the original Jackiw-Rebbi solution to the 1D Dirac equation that the interface state is named after.","marker":"[9]"},{"why":"Supplies the Su-Schrieffer-Heeger model whose Zak phase is used to identify the two topological phases.","marker":"[10]"},{"why":"Shows how photonic JR states arise between gratings with opposing Zak phases, providing the design basis.","marker":"[11]"},{"why":"Provides measured anisotropic refractive-index data for WS2 used in the simulations.","marker":"[25]"},{"why":"Supplies the non-Hermitian Dirac-like Hamiltonian for grating modes that the model builds on.","marker":"[34]"},{"why":"Links surface impedance and bulk band geometric phases, supporting the interpretation of the interface state.","marker":"[39]"},{"why":"Supplies the numerical solver used for the grating reflectance simulations.","marker":"[44]"}],"fun_headline_variants":["Topological light states emerge at WS2 grating interface","Jackiw-Rebbi states confine near-IR light in vdW stack","Photonic topological state boosts WSe2 emission 22-fold","Interface between WS2 gratings hosts topological mode","Topological photonic state in stacked WS2 heterostructure"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the near-field scattering peak at the grating interface at 736 nm is the Jackiw-Rebbi mode's local field and not an artificial contrast step created by normalising the signal separately on the two sides of the boundary.","fun_headline_variants_meta":{"raw":{"variants":["Topological light states emerge at WS2 grating interface","Jackiw-Rebbi states confine near-IR light in vdW stack","Photonic topological state boosts WSe2 emission 22-fold","Interface between WS2 gratings hosts topological mode","Topological photonic state in stacked WS2 heterostructure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000204,"raw_usage":{"total_tokens":1444,"prompt_tokens":1055,"completion_tokens":389,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":671,"completion_tokens_details":{"reasoning_tokens":305}},"tokens_in":671,"tokens_out":389,"duration_ms":7049,"temperature":1.0,"reasoning_tokens":305,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:49:59.060349+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-measure the s-SNOM signal at 736 nm without the separate median line-leveling on the two grating sides; if the interface enhancement disappears in the raw scattering amplitude, the spatial-confinement claim collapses. A complementary check is to fabricate the same double-grating interface between two topologically identical gratings and show that no mid-gap interface peak appears, or to scale the grating period and confirm the interface peak tracks the predicted JR energy.","supporting_citations":[{"cited_title":"Arora, T","cited_arxiv_id":null,"evidence_quote":"Supplies the original Jackiw-Rebbi solution to the 1D Dirac equation that the interface state is named after."},{"cited_title":"Orsini, H","cited_arxiv_id":null,"evidence_quote":"Supplies the Su-Schrieffer-Heeger model whose Zak phase is used to identify the two topological phases."},{"cited_title":"Splendiani, L","cited_arxiv_id":null,"evidence_quote":"Provides measured anisotropic refractive-index data for WS2 used in the simulations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the non-Hermitian Dirac-like Hamiltonian for grating modes that the model builds on."},{"cited_title":"Band Inversion Flips the Winding of Bound States in the Continuum","cited_arxiv_id":"2211.09884","evidence_quote":"Supplies the numerical solver used for the grating reflectance simulations."}],"review_version":1}