{"id":"7b73c73b-1df9-4a8f-bcc9-29bc8f5a5a10","arxiv_id":"2411.14113","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A sqrt(3) x sqrt(3) reconstruction in tetralayer rhombohedral graphene on MoS2 at 77 K is reported as direct imaging of intervalley coherent order.","lead":"Using a scanning tunneling microscope at 77 K, researchers saw a honeycomb-scale pattern in rhombohedral tetralayer graphene on molybdenum disulfide, which they identify as a predicted electron order called intervalley coherence. The result suggests that a nearby MoS2 layer can stabilize this collective state at much higher temperatures than usual.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The IVC identification hinges on interpreting the √3×√3 STM pattern, but no simulated STM image from the SVL-IVC Hartree-Fock state is shown to match the data.","rationale":"I read the paper as claiming a specific microscopic ground state: spontaneous intervalley coherence (SVL-IVC) in tetralayer rhombohedral graphene on MoS2, stabilized by proximity spin-orbit coupling and screening, observed at 77 K through a √3×√3 STM reconstruction. The data have genuine independent support: clean topographs, homogeneous lattice constant, no observable moiré, a bias- and filling-dependent modulation, and an hBN control sample. These make the observation of an interaction-driven reconstruction plausible. The weakest link is the interpretive step from a periodic LDOS modulation to a specific broken-symmetry many-body state. The Methods symmetry argument shows that an intervalley-coherent state can produce a Kekulé-type pattern, but 'allowed' is not 'the only explanation,' and no simulated STM image from the Hartree-Fock state is shown. The flat-band splitting, while consistent with correlations, does not itself identify IVC. Therefore the only direct evidence for IVC is the √3×√3 periodicity, and that evidence is not quantitatively tied to the SVL-IVC state through a forward model of the STM experiment. The reader's CONDITIONAL verdict already captures this gap; my stress-test reinforces it rather than moving it. I do not see an internal inconsistency or a fatal flaw, only a missing decisive comparison. A single computational check — simulating the STM images from the SVL-IVC Hartree-Fock state and comparing them with Figs. 3 and 4 — would either convert the conditional acceptance into a strong claim or show that the observed pattern is not uniquely attributable to IVC. Hence the verdict remains unchanged at CONDITIONAL.","tokens_in":11558,"tokens_out":5631,"duration_ms":55970,"concrete_test":"Take the converged SVL-IVC Hartree-Fock state described in Methods (96×96 k-grid, ε_r = 5.0, d = 120 nm, and the Ising SOC used in Fig. 5a), compute the top-layer LDOS with a Tersoff-Hamann-type tunneling matrix, and simulate constant-current topographs and dI/dV maps at biases from -500 mV to 500 mV at the experimental ~60% and ~70% fillings. Compare these simulated maps quantitatively with Fig. 3f-h and Fig. 4a-c: if the √3×√3 modulation appears in the same bias window and filling range with comparable normalized FFT intensity, the IVC interpretation is directly supported; if the simulation shows no √3×√3 top-layer LDOS, or the bias/filling dependence mismatches, then the observed pattern cannot be uniquely attributed to SVL-IVC.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central inference is the Results sentence: 'The observed √3 × √3 supercell pattern demonstrates the existence of the IVC order arising from a coherent superposition of electron wavefunctions between the K and K′ valleys, i.e., the Kekulé distortion.' The load-bearing step is identifying a real-space √3×√3 LDOS modulation with a spontaneous SVL-IVC Hartree-Fock ground state. That identification is underdetermined because the manuscript never computes simulated STM topographs or dI/dV maps from the SVL-IVC state and compares them with the data. The Methods section 'The Kekulé distortion in the SVL-IVC state' only proves that SVL-IVC is symmetry-allowed to have a nonzero Kekulé component, while KIVC forbids it; it does not establish that the observed pattern's amplitude, bias window, or filling dependence is uniquely reproduced by SVL-IVC. It also does not quantitatively exclude other √3×√3 mechanisms such as residual intervalley scattering from defects, a substrate-driven reconstruction, or a tip-induced electronic instability. The paper's statement that 'These observations are accurately reproduced by our theoretical calculations' is not supported by any image-level comparison. Since the flat-band splitting is not IVC-specific (it is also observed on hBN), the periodicity is the only evidence tying the data to valley coherence. Without a simulated STM comparison, the central claim remains an interpretation rather than a demonstrated identification.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports STM/STS measurements of tetralayer rhombohedral graphene (RG) placed on MoS2, showing a √3×√3 supercell modulation in atomic-resolution topographic images and dI/dV maps when the flat band is partially filled at ~60% and ~70%. The authors interpret this pattern as intervalley coherent (IVC) order, specifically the SVL-IVC state, and support this with Hartree–Fock mean-field calculations that find the SVL-IVC state to be a ground state with an allowed Kekulé-type LDOS modulation. They also report the absence of the √3×√3 pattern in hBN-supported RG under the same conditions, which they attribute to the proximity-induced Ising spin–orbit coupling from MoS2.","tokens_in":11760,"tokens_out":5703,"duration_ms":53736,"significance":"If the identification is correct, this would be the first real-space visualization of intervalley coherent order in rhombohedral graphene, and at the unusually high temperature of 77 K. The paper's strengths include careful band-structure fitting to determine γ1, a bias- and filling-dependent dataset, a control experiment on hBN, and Hartree–Fock calculations that show the SVL-IVC state is symmetry-allowed to produce the observed periodicity. However, the central inference that the √3×√3 pattern 'demonstrates the existence of the IVC order' is not backed by simulated STM images or dI/dV maps from the SVL-IVC state, and the theoretical calculations do not quantitatively reproduce the bias or filling dependence of the pattern. The claim is therefore plausible but currently under-supported at the image level.","major_comments":[{"comment":"The central claim that the observed √3×√3 pattern \"demonstrates the existence of the IVC order\" is not supported by any simulated STM image or dI/dV map from the SVL-IVC Hartree–Fock state. The Methods section only proves that the SVL-IVC state is symmetry-allowed to have a nonzero Kekulé component (while KIVC forbids it); it does not compute the LDOS of that state and compare it with the experimental images. Consequently, the statement in the Introduction that \"These observations are accurately reproduced by our theoretical calculations\" is unsubstantiated at the image level. Without such a comparison, alternative √3×√3 mechanisms—such as defect-induced intervalley scattering, substrate-driven reconstruction, or a tip-induced electronic instability—are not quantitatively excluded. This is a load-bearing gap for the paper's principal conclusion.","section":"Results (Fig. 3) and Methods 'The Kekulé distortion in the SVL-IVC state'"},{"comment":"The theoretical support does not reproduce the observed bias and filling dependence. The Hartree–Fock calculation is shown for a single doping (0.9×10^12 cm^-2) and no phase diagram as a function of filling is presented, so the observed appearance of the pattern at ~60% and ~70% filling and its absence at ~50% and ~100% is not compared with theory. The parameters ε_r = 5.0, d = 120 nm, and λ_I = 1 meV are stated without sensitivity analysis. The claim of robust IVC order at 77 K would be substantially strengthened by computing the IVC order parameter or spectral gap as a function of filling, screening, and SOC strength, and by showing that the SVL-IVC state is the stable ground state in the parameter range that matches the experiment.","section":"Fig. 5 and Methods 'The Hartree-Fock mean-field calculation method'"},{"comment":"The paper argues that the bias- and filling-dependence, together with the defect-free and strain-free local topography, \"essentially rule out the scattering mechanism induced by disorder.\" This argument is reasonable but qualitative; it does not exclude all alternative √3×√3 electronic instabilities or substrate-driven reconstructions. Moreover, the paper does not show that the SVL-IVC state is uniquely responsible for the observed pattern as opposed to, for example, the IVC0 state or a charge-density-wave-type instability. The statement that the pattern is \"consistent with\" SVL-IVC is weaker than the claim of accurate reproduction, and the manuscript should either provide simulated STM images that discriminate between candidate orders or temper the conclusion accordingly.","section":"Results (paragraph on disorder exclusion) and Fig. 4"}],"minor_comments":[{"comment":"Typo: \"Pervious theoretical analysis\" should be \"Previous theoretical analysis.\"","section":"Main text (paragraph beginning 'It is worth noting that besides inducing')"},{"comment":"The equation for ρ(q,E) is garbled: the overline indicating complex conjugation and the bracket structure are unclear, and the trace expression appears incomplete. Please rewrite this formula with clear notation.","section":"Methods, 'The Kekulé distortion in the SVL-IVC state'"},{"comment":"Define Δq and the density operator ρ̂(q) explicitly, and state whether q is a wave vector in the reciprocal lattice or a momentum transfer. The current notation makes the derivation hard to follow.","section":"Methods, 'The Kekulé distortion in the SVL-IVC state'"},{"comment":"The normalized IVC strength plots lack error bars and information about the number of independent regions or measurements. Adding these would support the claims of bias- and filling-dependent behavior.","section":"Fig. 3h and Fig. 4c"},{"comment":"The phrase \"2-order IVC\" is not defined; if it refers to second-order diffraction spots in the FFT, please state so explicitly and indicate them in the figure.","section":"Fig. 3f,g,h"},{"comment":"The Hartree–Fock parameters are only partly specified (λ_I = 1 meV in the caption; ε_r = 5.0 and d = 120 nm in Methods). Please list all numerical inputs (grid size, band projection, interaction truncation) and justify the choice λ_I = 1 meV for MoS2-proximitized graphene.","section":"Fig. 5 caption and Methods"},{"comment":"The term \"Kekulé distortion\" is used interchangeably to denote an electronic LDOS modulation and a lattice distortion. Since the STM measures LDOS, please clarify that the observed pattern is an electronic reconstruction, not necessarily a structural distortion.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a potentially important observation, but the central identification of the √3×√3 pattern with SVL-IVC order requires an image-level comparison between simulation and experiment. The authors should be given the opportunity to add such a comparison. If they cannot, the conclusion should be substantially softened to 'consistent with' rather than 'demonstrates.' I also note that the manuscript's novelty statement should be checked against Ref. 49 (Arp et al.) to ensure the claim of lacking direct confirmation is accurate and appropriately worded."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one. It's the first real-space image of a √3×√3 reconstruction in rhombohedral tetralayer graphene that they assign to intervalley coherence, at 77 K, enabled by MoS2 proximity. The experiment is clean: the pattern is bias-dependent, appears only at partial filling, and is absent on hBN under the same conditions. The contrast with hBN is a nice control and makes the SOC-proximity story plausible. The Hartree-Fock calculation is competent and identifies SVL-IVC as the state that can show the Kekulé component, while KIVC cannot.\n\nThe soft spot is the identification itself. The leap from a √3×√3 LDOS modulation to \"IVC order demonstrated\" is underdetermined. The symmetry argument in Methods only shows SVL-IVC is allowed to have the distortion; it does not show that this state uniquely produces the observed pattern's amplitude, bias window, or filling dependence. The paper never computes simulated STM topographs or dI/dV maps from the HF ground state and overlays them on the data. So the phrase 'accurately reproduced by our theoretical calculations' is not supported at the image level. Other mechanisms—defect-induced intervalley scattering, substrate reconstruction, tip effects—are argued against but not quantitatively excluded. The filling-dependent splitting on hBN also weakens the claim that the splitting itself is IVC-specific. Missing error bars on some extracted quantities and no deposited data/code are minor but add to the conditional feel.\n\nStill, the observation is new and likely important. The bias/filling dependence and the hBN control rule out the most boring explanations. If I were refereeing, I would ask for simulated STM images from the SVL-IVC state, a clearer statement of how the IVC intensity is defined, and a more quantitative exclusion of disorder scattering. None of that is fatal. This deserves serious peer review. I'd bring it up in the group and cite it once it passes scrutiny.","headline":"Genuinely new STM observation of a √3×√3 reconstruction in tetralayer RG on MoS2, plausibly IVC but not yet demonstrated without simulated STM comparison.","tokens_in":12458,"tokens_out":4507,"would_cite":true,"duration_ms":31923,"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":"Intervalley coherent order in rhombohedral graphene imaged at 77 K","keywords":["intervalley coherence","rhombohedral graphene","scanning tunneling microscopy","Kekulé distortion","flat bands","spin-orbit proximity","MoS2 substrate","Hartree-Fock mean-field"],"falsifier":"If the $\\sqrt{3}\\times\\sqrt{3}$ reconstruction persisted at full filling or at half filling, or appeared identically on hBN after introducing comparable disorder, the intervalley-coherence assignment would fail; a direct check is to compute simulated STM images from the Hartree-Fock SVL-IVC states and compare the real-space phase and bias dependence with the measured maps.","tokens_in":11262,"feed_emoji":"🔬","tokens_out":8681,"duration_ms":78841,"temperature":0.7,"pith_summary":"This paper claims to have directly imaged intervalley coherent (IVC) order in rhombohedral tetralayer graphene, a long-predicted broken-symmetry ground state that had previously been inferred only from transport. In four-layer ABC-stacked graphene placed on MoS2, the authors see a $\\sqrt{3}\\times\\sqrt{3}$ supercell in scanning tunnelling microscopy at 77 K when the flat band is partially filled (~60% and ~70%), and they attribute this pattern to a Kekulé distortion arising from coherent superposition of K and K' valley wavefunctions. The pattern appears only at biases near the flat band and is absent in hBN-supported samples under the same conditions, which the authors take as evidence that MoS2's spin-orbit proximity and screening promote the order. If correct, the result makes a predicted correlated phase routinely observable at liquid-nitrogen temperature rather than millikelvin temperatures and points to substrate proximity as a practical knob for collective states in graphene multilayers.","feed_headline":"Intervalley order in rhombohedral graphene imaged at 77 K","feed_subtitle":"A √3×√3 reconstruction at partial filling reveals the substrate-boosted Kekulé state, which vanishes on hBN.","key_machinery":"The load-bearing object is the $\\sqrt{3}\\times\\sqrt{3}$ supercell in the local density of states, which the paper reads as the Kekulé distortion of an intervalley coherent state. The essential theoretical object is the SVL-IVC order parameter $(\\tau_x s_x,\\tau_z s_z)$, where $\\tau$ and $s$ are Pauli matrices in valley and spin space: the $\\tau_x s_x$ term is a coherent intervalley component that survives large Ising spin-orbit coupling, and the $\\tau_z s_z$ term encodes spin-valley locking. Because this state's effective time-reversal symmetry transforms the intervalley density operator $\\hat{\\rho}(\\mathbf{q}=K-K'+\\Delta\\mathbf{q})$ without a sign flip, the Kekulé distortion is allowed in real space, in contrast to the Kramers IVC state where it cancels. The supporting calculation is a Hartree-Fock mean-field treatment with a dual-gate screened interaction and a 96x96 momentum grid, using Ising SOC from the MoS2 substrate.","core_discovery":"The central claim is that the $\\sqrt{3}\\times\\sqrt{3}$ atomic reconstruction observed in partially filled tetralayer rhombohedral graphene on MoS2 is the intervalley coherent state, i.e. the Kekulé distortion. The state appears at 77 K, forms at partial fillings of the flat band where the spectroscopic peak splits, and is not seen in hBN-supported tetralayer graphene measured the same way. The paper identifies the microscopic mechanism as a spin-valley-locked IVC (SVL-IVC) state with order parameter $(\\tau_x s_x,\\tau_z s_z)$: the spin-valley locking breaks enough symmetry that the usual cancellation of the Kekulé density modulation in a Kramers IVC state no longer occurs, and the effective time-reversal symmetry $\\mathcal{T}=\\tau_y s_y \\mathcal{K}$ allows a real-space $\\sqrt{3}\\times\\sqrt{3}$ modulation. Hartree-Fock mean-field calculations with screened Coulomb interaction and Ising spin-orbit coupling reproduce this state, and the calculations connect its stability to the combined effect of MoS2-induced spin-orbit coupling and enhanced screening.","pith_inferences":["A gate-tunable device should show the $\\sqrt{3}\\times\\sqrt{3}$ pattern appearing and disappearing as the filling passes through the ~60-70% window; watching this switch in one sample would test the filling dependence without relying on different local doping regions.","The same mechanism predicts that replacing MoS2 with other dichalcogenide substrates of different spin-orbit strength will shift the filling range or temperature ceiling of the IVC order; this is a testable extension the paper does not carry out.","Comparing the measured STM images with simulated LDOS maps computed from the Hartree-Fock SVL-IVC wavefunctions would independently confirm the identification; the paper does not report such a comparison.","If the order is truly spontaneous rather than pinned by disorder, cooling below 77 K should reveal domain walls or phase slips of the $\\sqrt{3}\\times\\sqrt{3}$ modulation where the coherent intervalley phase changes."],"forward_implications":["Rhombohedral graphene inherits an experimentally accessible IVC phase at 77 K, so the predicted valley-coherent state can be studied with real-space probes instead of only transport.","Substrate proximity becomes a control parameter: the same graphene on hBN shows no reconstruction, so choosing a transition-metal dichalcogenide substrate can switch the correlated ground state on and off.","The bias and filling dependence of the $\\sqrt{3}\\times\\sqrt{3}$ intensity gives a spectroscopic fingerprint for IVC order that can be searched for in other flat-band graphene systems.","The observed flat-band splitting and the IVC wavevector constrain the ground-state flavor ordering in tetralayer rhombohedral graphene, favoring SVL-IVC over competing states in the MoS2-supported case.","Because IVC order has been linked theoretically to superconductivity in rhombohedral graphene, a stable 77 K IVC state narrows the set of pairing mechanisms that need to be considered."],"supporting_citations":[{"why":"supplies the theoretical prediction that IVC order is a ground state of rhombohedral graphene and that it produces a Kekulé distortion.","marker":"[27]"},{"why":"supplies the companion theory linking intervalley coherence and Kohn-Luttinger superconductivity in rhombohedral trilayer graphene.","marker":"[28]"},{"why":"proposes switchable isospin-induced IVC order through proximity SOC in rhombohedral multilayers, the starting point for the SVL-IVC state.","marker":"[46]"},{"why":"gives the correlated-phase theory for spin-orbit-coupled rhombohedral trilayer graphene, including the SVL-IVC state.","marker":"[47]"},{"why":"provides the flavor-symmetry-breaking analysis in spin-orbit-coupled graphene that underlies the order parameter $(\\tau_x s_x,\\tau_z s_z)$.","marker":"[48]"},{"why":"documents flat-band splitting in tetralayer rhombohedral graphene on hBN at 77 K, the main experimental comparison showing the MoS2 specificity.","marker":"[39]"},{"why":"demonstrates STM imaging of intervalley coherent order in magic-angle twisted trilayer graphene, the experimental template for reading the reconstruction.","marker":"[11]"},{"why":"explains why a Kramers IVC state has no Kekulé distortion in STM, which motivates the need for the spin-valley-locked variant.","marker":"[56]"}],"fun_headline_variants":["Kekulé order imaged in tetralayer graphene on MoS2 at 77 K","Direct STM image of intervalley coherent state at 77 K","√3×√3 Kekulé pattern imaged on MoS2 graphene at 77 K","Kekulé order visible at 77 K only on MoS2-supported graphene","Intervalley coherence imaged in real space: Kekulé on MoS2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's conclusion turns on the assumption that the $\\sqrt{3}\\times\\sqrt{3}$ pattern is a spontaneous electronic order rather than a static reconstruction, defect-induced scattering, or tip artifact; the authors argue against disorder and substrate strain but do not directly measure valley coherence.","fun_headline_variants_meta":{"raw":{"variants":["Kekulé order imaged in tetralayer graphene on MoS2 at 77 K","Direct STM image of intervalley coherent state at 77 K","√3×√3 Kekulé pattern imaged on MoS2 graphene at 77 K","Kekulé order visible at 77 K only on MoS2-supported graphene","Intervalley coherence imaged in real space: Kekulé on MoS2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000904,"raw_usage":{"total_tokens":3938,"prompt_tokens":1043,"completion_tokens":2895,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":659,"completion_tokens_details":{"reasoning_tokens":2781}},"tokens_in":659,"tokens_out":2895,"duration_ms":20495,"temperature":1.0,"reasoning_tokens":2781,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:31:23.018727+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If the $\\sqrt{3}\\times\\sqrt{3}$ reconstruction persisted at full filling or at half filling, or appeared identically on hBN after introducing comparable disorder, the intervalley-coherence assignment would fail; a direct check is to compute simulated STM images from the Hartree-Fock SVL-IVC states and compare the real-space phase and bias dependence with the measured maps.","supporting_citations":[{"cited_title":"& Zaletel, M","cited_arxiv_id":null,"evidence_quote":"supplies the theoretical prediction that IVC order is a ground state of rhombohedral graphene and that it produces a Kekulé distortion."},{"cited_title":"& Vishwanath, A","cited_arxiv_id":null,"evidence_quote":"supplies the companion theory linking intervalley coherence and Kohn-Luttinger superconductivity in rhombohedral trilayer graphene."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"proposes switchable isospin-induced IVC order through proximity SOC in rhombohedral multilayers, the starting point for the SVL-IVC state."},{"cited_title":"M., Alicea, J","cited_arxiv_id":null,"evidence_quote":"gives the correlated-phase theory for spin-orbit-coupled rhombohedral trilayer graphene, including the SVL-IVC state."},{"cited_title":"& Das Sarma, S","cited_arxiv_id":null,"evidence_quote":"provides the flavor-symmetry-breaking analysis in spin-orbit-coupled graphene that underlies the order parameter $(\\tau_x s_x,\\tau_z s_z)$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"demonstrates STM imaging of intervalley coherent order in magic-angle twisted trilayer graphene, the experimental template for reading the reconstruction."},{"cited_title":"P., Soejima, T","cited_arxiv_id":null,"evidence_quote":"explains why a Kramers IVC state has no Kekulé distortion in STM, which motivates the need for the spin-valley-locked variant."}],"review_version":1}