{"id":"b26472ab-854a-4f7e-b383-c57833010e3e","arxiv_id":"2607.19937","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In twisted MoS2 bilayers, the second-harmonic generation polar lobes rotate by half the structural twist angle, a frequency-independent optical fingerprint.","lead":"This paper derives and computes the second-harmonic generation tensor for several stackings of MoS2 bilayers, finding that in twisted bilayers the polar pattern rotates by exactly half the twist angle, independent of light frequency. If the rule holds generally, it offers a simple optical ruler for measuring twist angles in layered materials without diffraction.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"D3 symmetry of the ideal twisted bilayer would forbid χ_yyy, so the C3-supercell result φ0=θ_tw/2 is not the universal design rule claimed.","rationale":"The paper's symmetry tables for D3h, D3d, C3v, and C3 are individually correct, and the DFT validation of the tensor structures for those groups is useful. The load-bearing issue is the universality of θ_tw=2φ0. The authors' own caveat (Sec. IV A) states that the ideal twisted homobilayer has C2′ axes absent in the 42-atom supercell. Applying the C2′ operation to the in-plane components shows that D3 forbids χ_yyy in the C2′-aligned frame, so the lab-frame SHG pattern rotates by β−30°, not β, for β=θ_tw/2. The computed φ0=θ_tw/2 is therefore a consequence of the artificial C3 symmetry, not a property of the ideal D3 system. This is a correctness risk, not a disagreement with external consensus: the point group of an ideal twisted homobilayer is a well-established group-theoretic fact. The independent-particle approximation and the lack of multi-angle testing are secondary concerns; even within the independent-particle framework, the D3 symmetry would produce a different in-plane tensor and a different lobe shift. The central claim as stated overreaches, and the paper should be conditional on either re-scoping the claim to C3-type supercells or providing a D3-symmetric calculation that confirms the relation.","tokens_in":18329,"tokens_out":34404,"duration_ms":317381,"concrete_test":"Construct a D3-symmetric commensurate supercell for the same 21.8° twisted bilayer by placing the twist center at a high-symmetry point so that a C2′ axis is an exact symmetry (e.g., layers at ±θ_tw/2). Recompute χ(2) with the same VASP/Wannier90/Postw90 pipeline and check whether the in-plane component χ_yyy in the C2′-aligned frame vanishes and whether the co-polarized polar maximum sits at θ_lobe=θ_tw/2−30° rather than +10.9°. A complementary check: compute φ0(ω) for a second commensurate angle (e.g., 13.2°) using the same C3-type supercell; if φ0≠θ_tw/2, the claimed universal lock also fails for C3 systems.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim φ0=θ_tw/2 rests on a single 21.8° C3 supercell calculation. The authors acknowledge in Sec. IV A that the idealized macroscopic twisted homobilayer has C2′ axes absent in this supercell, but they only say these axes enforce χ_zzz=χ_zxx=χ_xxz=0. They miss that C2′ also constrains the in-plane tensor. For point group D3 (C3 + C2′), the 180° rotation about an in-plane axis, e.g., x→x, y→−y, z→−z, forces any component with an odd number of y indices to vanish, in particular χ_yyy=0 and χ_yxx=χ_xxy=χ_xyx=0. The surviving in-plane component is χ_xxx with the C3 relations χ_xxx=−χ_xyy=−χ_yxy=−χ_yyx. In the lab frame where the C2′ axis is at β=θ_tw/2 from the bottom-layer zigzag, this gives χ_xxx=A cos3β and χ_yyy=−A sin3β, so Eq. (20) becomes I∥∝cos²(3(θ−β)). The lobe maximum is at θ=β; relative to the D3h monolayer maximum at 30°, the shift is β−30°, i.e., φ0=θ_tw/2−30°. For θ_tw=21.8° this is −19.1°, not +10.9°. The calculated +10.9° arises because the C3 supercell artificially allows χ_yyy through the fixed-layer translation; it is not representative of the ideal D3 system. Hence Eq. (30) is not universally valid and the claim needs to be re-scoped.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a group-theoretic framework for the second-order susceptibility tensor of MoS2 in four point groups (D3h, D3d, C3v, C3), and validates the symmetry-allowed components with independent-particle DFT calculations and Wannier interpolation. For the C3 twisted bilayer at θ_tw = 21.8°, the authors find that the activation of χ_xxx rotates the SHG polar-lobe pattern by φ0 ≈ 10.9°, and they claim the universal relation θ_tw = 2φ0 that is frequency-independent across the optical spectrum, proposing SHG as a wavelength-independent twist-angle probe. The Neumann-principle reductions per point group are internally consistent, but the universal and frequency-independent claims are not adequately supported.","tokens_in":18760,"tokens_out":24253,"duration_ms":275655,"significance":"If the central relation θ_tw = 2φ0 were established, it would provide a simple, non-destructive optical route for twist-angle metrology in vdW heterostructures, with broad applicability to TMDs. The paper also offers a useful systematic classification of χ^(2) tensor structures across stacking symmetries and verifies these selection rules numerically, including the complete suppression in centrosymmetric 2H stacking and the activation of out-of-plane components when σ_h is broken. The data are shared on Zenodo, and the computational workflow is standard. However, the headline universal law rests on a single C3 supercell and ignores the in-plane consequences of the C2′ axes that the authors themselves attribute to the idealized twisted homobilayer. These issues undermine the generality of Eq. (30), although the underlying tensor tables and symmetry analysis remain valuable.","major_comments":[{"comment":"The authors acknowledge that an idealized macroscopic twisted homobilayer has in-plane C2′ axes, but they apply the constraint only to out-of-plane components. For point group D3, a C2′ rotation about an in-plane x axis (R=diag(1,-1,-1)) also constrains the in-plane tensor: components with an odd number of y indices vanish, so the C3 relations in Eq. (10) reduce to χ_yyy = χ_yxx = χ_xxy = χ_xyx = 0 and χ_xxx = -χ_xyy = -χ_yxy = -χ_yyx ≠ 0. In a lab frame with the C2′ axis at β = θ_tw/2 from the bottom zigzag axis, this gives χ_xxx/χ_yyy = -cot(3β), and Eq. (22) yields φ0 = β - 30°, which for θ_tw = 21.8° is -19.1°, not +10.9°. The 42-atom supercell is explicitly C3, so the computed +10.9° and activated out-of-plane components are consequences of the reduced cell symmetry, not of the ideal twisted bilayer. Eq. (30) cannot be asserted as universal without a D3-preserving calculation or a c","section":"Sec. IV A, Eq. (10), Eq. (30)"},{"comment":"The compact phase-shifted form Eq. (21) assumes that χ_xxx/χ_yyy is real. For complex tensor components, I_∥(θ) = |χ_yyy sin3θ - χ_xxx cos3θ|² contains the cross term -2 Re(χ_xxx χ_yyy*) sin3θ cos3θ and does not factor into (|χ_yyy|²+|χ_xxx|²) sin²(3(θ-φ0)) with a real φ0; Eq. (22) is then undefined. The manuscript neither states this reality condition nor provides the relative phase of the computed components. Since the six-lobed nodal pattern and the interpretation of φ0 as a rigid lobe shift depend on this factorization, this missing step is load-bearing.","section":"Eqs. (19)–(22)"},{"comment":"The universal relation θ_tw = 2φ0 is validated at a single commensurate angle, 21.8°, from one C3 supercell. No second twist angle, no incommensurate cell, and no error bar or convergence estimate are reported. Deviations in Fig. 4(d) are described only qualitatively. The statement that the ratio is “solely dictated by the spatial projection of the atomic coordinates” is an assertion; without an analytical derivation or multiple test angles, it cannot support a material-agnostic law.","section":"Sec. IV C, Fig. 4(d), Eq. (30)"},{"comment":"The frequency-independence claim covers only the independent-particle DFT window shown in Figs. 2 and 4 (roughly 0.6–1.6 eV), not “the entire optical spectrum” as stated in the Abstract. Many-body effects are neglected; excitonic and self-energy corrections can renormalize the relative spectral weights of χ_xxx and χ_yyy and could make φ0 frequency-dependent. The claims should be re-scoped or supported by calculations beyond the independent-particle approximation.","section":"Sec. III and Sec. IV C"}],"minor_comments":[{"comment":"Typo: “although the become accessible” should read “although they become accessible.”","section":"Sec. II C"},{"comment":"The text says C3 leaves “13 unique independent tensor components”; the later C3 tensor structure in Eq. (10) contains fewer independent parameters after permutation and C3 relations. Please clarify the counting convention.","section":"Appendix A 2"},{"comment":"The vertical axis label “Peak shift ¢peak” is unclear; it should be φ0(ω) or a related notation. Also, the color-coding by normalized |χ_yyy| should be explained in the caption.","section":"Fig. 4(d)"}],"recommendation":"major_revision","confidential_remarks":"The decisive issue is the D3/C2′ objection. If the authors can show that realistic twisted bilayers always break C2′ (with a commensurate-cell argument or experiment), or if they present D3-preserving calculations and revise Eq. (30) accordingly, the paper could become publishable. Otherwise the headline twist-angle metrology claim should be withdrawn or significantly weakened. The remaining symmetry tables and DFT validation are useful but not sufficient for acceptance in their current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth a serious referee, but the central rule is overclaimed. The group-theory part is clean and correct: the tensor tables for D3h, D3d, C3v, and C3 are standard, and the DFT spectra confirm the symmetry-allowed components. That confirmation is useful as a reference, even if not surprising. The genuinely new observation is that in their 21.8° C3 supercell, φ0(ω) stays locked at θ_tw/2 across the whole computed spectrum. If that holds at other angles, it would be a genuinely handy calibration rule for SHG-based twist metrology. The oblique-incidence discussion for accessing out-of-plane components is also clear and useful.\n\nThe soft spots are the universality claim. The φ0 = θ_tw/2 relation rests on a single twist angle, a single stacking, and the independent-particle approximation. There are no error bars, no second-angle test, and no analytical derivation. More importantly, the paper itself notes that the ideal macroscopic twisted bilayer would have C2′ axes that are absent in the C3 supercell. The stress-test note pushes this correctly: those C2′ axes would force χ_yyy = 0 in D3 symmetry, which changes the in-plane tensor and the lobe shift entirely. The paper never addresses why the C3 supercell is the representative case for real twisted bilayers. That does not invalidate the DFT result for the specific C3 cell they studied, but it does mean the headline — a wavelength-independent optical route for twist-angle determination — is not established as a universal design rule.\n\nThe symmetry tables alone are textbook-neat, and the paper is clearly written. I would cite it for the C3 tensor structure and the frequency-flat φ0 observation, but I would not build anything on the 2φ0 relation without more angles and an explicit treatment of the C2′ problem. A serious referee should see this; the correct outcome is major revision, not rejection.","headline":"Solid symmetry analysis and one clean numerical observation, but the θ_tw = 2φ0 rule is single-angle empiricism wearing a universal-design-rule costume.","tokens_in":19242,"tokens_out":26181,"would_cite":true,"duration_ms":244338,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The twist angle of a MoS2 bilayer can be read directly from the rotation of its second-harmonic generation lobes.","keywords":["second-harmonic generation","MoS2 bilayer","twist angle","point-group symmetry","Neumann's principle","χ(2) tensor","transition metal dichalcogenides"],"falsifier":"Measure the co-polarized SHG polar pattern of a MoS2 twisted bilayer with an independently known twist angle (e.g., by electron diffraction) at two or more fundamental wavelengths, and check whether the node/lobe rotation is exactly θ_tw/2 and identical at both wavelengths. Alternatively, repeat the first-principles calculation in a larger or symmetrized supercell that includes the C2′ axes: if the lobe shift changes or becomes frequency-dependent, the central claim is refuted.","tokens_in":18214,"feed_emoji":"🌀","tokens_out":6057,"duration_ms":54569,"temperature":0.7,"pith_summary":"This paper argues that the second-harmonic generation (SHG) response of a bilayer MoS2 is fully controlled by the stacking point group, and that twisting the layers activates a new in-plane susceptibility component that was silent in all untwisted stackings. That activation rigidly rotates the six-lobed SHG polar pattern by an angle equal to half the twist angle, and the rotation does not change with photon energy across the computed optical spectrum. If correct, SHG becomes a direct, wavelength-independent optical ruler for twist angles, alongside a set of design rules that connect stacking configuration to allowed and forbidden nonlinear tensor components. The argument combines group-theoretic tensor analysis with first-principles spectra for five configurations spanning four point groups.","feed_headline":"Twist angle is twice the SHG lobe rotation in MoS2 bilayers","feed_subtitle":"Six-lobed SHG pattern rotates by half the twist angle at every photon energy, giving a direct optical route to measure twist.","key_machinery":"The central object is the second-order susceptibility tensor χ^(2) — the coefficient relating an induced polarization at twice the optical frequency to the product of two fundamental fields — constrained by Neumann's principle: every symmetry operation of the crystal point group must leave the tensor invariant. The decisive symmetry element is the vertical mirror plane σv; in all untwisted stackings it enforces χ_xxx = 0, while in the twisted C3 configuration it is absent, so χ_xxx becomes an independent component. The ratio χ_xxx/χ_yyy enters a trigonometric phase φ0 in the SHG polar pattern, and the calculations show this phase equals half the structural twist angle across the whole spectr","core_discovery":"Using Neumann's principle, the paper reduces the 27-component χ^(2) tensor to the allowed independent components for the point groups D3h (monolayer/AA), D3d (2H), C3v (3R), and C3 (twisted). In the twisted C3 bilayer the in-plane components χ_xxx and χ_yyy are both independent and non-zero, which turns the co-polarized SHG intensity into sin²(3(θ−φ0)) with φ0 = (1/3) arctan(χ_xxx/χ_yyy). First-principles results for a 21.8° twisted bilayer give φ0 = 10.9°, exactly half the twist angle, at every photon energy considered; the paper calls this a geometric locking of the lobe shift to the twist.","pith_inferences":["A direct test of the locking relation not reported in the paper: compute or measure φ0 at other commensurate twist angles; if φ0 = θ_tw/2 holds for 9.4°, 13.2°, or 38.2°, the relation is universal, and if it does not, the 21.8° result may be specific to that cell.","The paper's own caveat about the missing C2′ axes suggests the out-of-plane components in twisted bilayers may be an artefact of the small supercell; the in-plane ratio driving φ0 could still survive in a truly D3-symmetric macroscopic layer, but that is an open question the authors do not settle.","Since the independent-particle approximation omits excitons, a natural next step is to check whether many-body corrections preserve the frequency independence of χ_xxx/χ_yyy; excitonic renormalization could alter peak positions without necessarily changing the geometric ratio.","The symmetry logic should extend beyond MoS2 to other 2H TMDs and to local moiré domains, suggesting SHG lobe mapping as a way to image spatial variations of twist angle in heterostructures."],"forward_implications":["In twisted MoS2 bilayers, the SHG polar pattern keeps its six-lobed shape but rotates by half the twist angle, so measuring the lobe orientation gives the twist angle directly.","The rotation is independent of excitation wavelength, so the probe does not require tuning to specific resonances.","Out-of-plane tensor components activated by broken horizontal mirror symmetry can be detected with oblique-incidence p-polarized light, distinguishing 3R and twisted bilayers from monolayers and AA stacks.","AB(2H) stacking has an inversion center and produces identically zero electric-dipole SHG, so SHG can also distinguish stacking configurations.","Because the rules are set by point group alone, they should transfer to other 2H TMD bilayers and related van der Waals crystals."],"fun_headline_variants":["SHG lobes rotate half the twist angle in MoS2 bilayers","Twist MoS2, see SHG rotate half the angle","MoS2 twist angle readout from SHG lobe shift","Frequency-independent twist metrology via SHG rotation"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that a 42-atom supercell with exact C3 symmetry faithfully represents the infinite twisted bilayer — an ideal macroscopic twisted homobilayer would also possess twofold in-plane rotation axes that force out-of-plane tensor components to zero — and that the independent-particle approximation preserves the frequency independence of the in-plane ratio.","fun_headline_variants_meta":{"raw":{"variants":["SHG lobes rotate half the twist angle in MoS2 bilayers","Twist MoS2, see SHG rotate half the angle","MoS2 twist angle readout from SHG lobe shift","Frequency-independent twist metrology via SHG rotation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000151,"raw_usage":{"total_tokens":1034,"prompt_tokens":739,"completion_tokens":295,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":483,"completion_tokens_details":{"reasoning_tokens":224}},"tokens_in":483,"tokens_out":295,"duration_ms":3952,"temperature":1.0,"reasoning_tokens":224,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T11:13:38.536144+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the co-polarized SHG polar pattern of a MoS2 twisted bilayer with an independently known twist angle (e.g., by electron diffraction) at two or more fundamental wavelengths, and check whether the node/lobe rotation is exactly θ_tw/2 and identical at both wavelengths. Alternatively, repeat the first-principles calculation in a larger or symmetrized supercell that includes the C2′ axes: if the lobe shift changes or becomes frequency-dependent, the central claim is refuted.","supporting_citations":[],"review_version":1}