{"id":"78e41419-32f8-4928-ad16-bf5f670b15c6","arxiv_id":"2602.20907","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In dense phonon spectra, isotope substitution rotates vibrational eigenvectors within coupled mode manifolds, and overlap-based tracking along a mass-interpolation path resolves how vibrational identity is preserved or redistributed.","lead":"This paper shows that swapping hydrogen for deuterium in the metal–organic framework ZIF-8 does more than shift vibrational frequencies: in crowded spectral regions it also rotates the shape of the vibrations, so matching modes between the two isotopes requires tracking eigenvectors, not just peak positions. It offers a framework for doing that tracking and argues neutron spectra can flag where such reorganization will happen.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Experimental validation rests on Γ-point modes with no q-grid or dispersion check; if dispersion is non-negligible, the measured TOSCA peaks are not the same objects as the tracked eigenmodes.","rationale":"The reader's weakest assumption is the same concern I identified: the entire experimental comparison is built on Γ-point normal modes without any q-grid or dispersion analysis. This is the most load-bearing issue because the paper's empirical claims—that INS validates the force constants and that spectral intensity indicates eigenvector reorganization—depend on the measured peaks being the same objects as the computed modes. If dispersion is non-negligible, then the overlap statistics and transport analysis, while internally consistent, are not connected to the experimental spectra in the way claimed. The paper is otherwise methodical: the perturbation derivation in Appendix A is coherent, the grid-convergence study of the adiabatic continuation is a genuine check, and the qualitative examples are plausible. The 'quantitative predictor' from INS intensity is asserted rather than quantitatively tested, but that is secondary to the missing q-grid validation. Since the reader already conditioned on this issue, the verdict should remain CONDITIONAL; a concrete dispersion check would settle whether the concern actually invalidates the experimental connection.","tokens_in":11853,"tokens_out":5733,"duration_ms":60608,"concrete_test":"Recompute the simulated INS spectrum using a finite q-grid (e.g., 2×2×2 and 4×4×4 Monkhorst-Pack, or at least a high-symmetry path) with the same DFPT force constants and AbINS, and compare to the Γ-only spectrum and to TOSCA. Specifically, extract the phonon bandwidth of the tracked modes around 416, 639, 734 cm^-1 and in the 450–1200 cm^-1 fingerprint region. If any mode's dispersion exceeds roughly 10 cm^-1 or shifts the simulated peak positions by more than the isotope shifts being discussed, the Γ-point-only validation is quantitatively unsafe and the tracked eigenvectors cannot be claimed to explain the measured spectra.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central empirical support is Fig. 1, which compares TOSCA powder INS spectra to simulated neutron-weighted DFT spectra. Appendix A.2 states explicitly: 'All modes considered in this work correspond to Γ-point normal modes.' No q-grid, supercell, or dispersion bandwidth is reported anywhere in Methods or Results. TOSCA is a powder spectrometer: the measured one-phonon intensity is a Brillouin-zone average, weighted by q-dependent eigenvectors and the neutron weighting in Eq. (5). If the 200–1200 cm^-1 modes of ZIF-8 have bandwidths comparable to the isotope shifts used for assignment (e.g., the 416→374 cm^-1 pair), then a feature appearing at a measured energy may be dominated by zone-boundary phonons whose eigenvectors differ substantially from the Γ-point eigenvectors tracked in Figs. 2–9. The paper asserts that 'the calculations reproduce peak positions and relative intensities across the vibrational fingerprint region, confirming that the DFT-LD force constants accurately describe the vibrational eigenproblem relevant to neutron scattering,' but this is exactly the step that requires a q-sampling validation. Without it, the endpoint overlaps, the transport amplitudes, and the 'experimental indicator' arguments are benchmarked against an unverified reference. This is a missing validation rather than an internal inconsistency, but it is load-bearing because the entire connection to experiment depends on it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the vibrational response of protiated and deuterated ZIF-8 using inelastic neutron scattering (INS) and DFT lattice dynamics. It argues that mass substitution acts as a continuous Hermitian deformation of the mass-weighted dynamical matrix, and that in spectrally dense regions the associated eigenvectors rotate substantially within coupled subspaces, so mode correspondence cannot be inferred from frequency shifts alone. The authors introduce an adiabatic eigenvector-continuation framework with overlap-based tracking and an overlap-gap stability diagnostic, apply it to the H→D interpolation, and report grid-convergence tests showing that apparent mode crossings mostly disappear with increasing interpolation resolution. They further interpret transport amplitudes and propose that INS peak intensities, being hydrogen-weighted, provide an experimental predictor of whether a mode will shift predictably or reorganize.","tokens_in":12088,"tokens_out":2444,"duration_ms":28710,"significance":"If the central claim holds, the paper offers a practical framework for tracking vibrational identity under isotopic substitution in complex materials with dense phonon manifolds, a setting where standard frequency-based assignment is ambiguous. The perturbation derivation in Appendix A is clean, the approach uses no fitted physical parameters for the main results, and the explicit grid-convergence analysis of apparent crossings is a valuable methodological contribution. The potential connection between INS intensity and mode stability is physically appealing and would be useful for interpreting isotope-substitution experiments. However, the experimental validation and several of the derived predictions rely on an unvalidated Γ-point approximation for a powder neutron spectrum, which is a load-bearing gap in the current manuscript.","major_comments":[{"comment":"The paper states that all modes are Γ-point normal modes (Appendix A.2) and uses Fig. 1 to validate the DFT force constants against TOSCA powder INS data. TOSCA measures a Brillouin-zone average; the one-phonon intensity in Eq. (5) and the eigenvector overlaps in Eq. (3) are q-dependent in general. No q-grid, dispersion bandwidth, or zone-boundary check is reported anywhere. If the 200–1200 cm−1 modes of ZIF-8 have appreciable dispersion, the measured peaks are not necessarily the same objects as the tracked Γ-point eigenmodes, and the experimental validation and all mode-correspondence statistics are undermined. This is fixable by adding a q-point convergence study or at least a dispersion estimate in the fingerprint region, but it is load-bearing for the central empirical claim.","section":"Appendix A.2; Section II.B; Fig. 1"},{"comment":"The text reports that at N_s = 10,000 there are still 13 apparent crossings, and that targeted refinement in s∈[0.2,0.35] reduces this number to 3. The authors then state that representative modes are chosen that do not undergo apparent crossings, so mode continuity is ensured for the studied modes. This is a selection, not a demonstration of full correspondence: the residual crossings are located in the most congested regions, which are precisely the regions where the paper claims its framework resolves ambiguity. The claim that the evolution is consistent with avoided-crossing behaviour is therefore not fully established for the entire spectrum. The authors should either resolve all residual crossings, or explicitly state the scope of the tracking claim and explain why the remaining cases do not affect the conclusions.","section":"Section III.E and Figs. 4–6"},{"comment":"Equation (5) is written as I_i(Q,ω) ∝ exp(−2W(Q)) |Q·e_i,H|^2 / ω_i, but the experiment is a powder measurement. For a powder, the intensity involves an orientational average over Q directions, so the simple statement that 'the measured peak intensity provides a direct experimental measure of hydrogen displacement weight' needs a derivation or at least a justification that the powder average preserves the proportionality to the hydrogen displacement amplitude. Without this, the predictive claim connecting INS intensity to the diagonal perturbation in Eq. (4) is not rigorously established, even at the Γ-point.","section":"Eq. (5) and Section III.C"},{"comment":"The 'action at a distance' example is interpreted as a large coupling matrix element acting despite a comparatively large spectral denominator. The text does not provide the numerical values of the relevant N_ji and (ω_i^2−ω_j^2) terms along the interpolation path, so the reader cannot verify that the coupling is indeed structural rather than accidental near-degeneracy at some intermediate s. A short table or trajectory of these quantities would make the interpretation quantitative and falsifiable.","section":"Section III.G and Fig. 7"}],"minor_comments":[{"comment":"There are several typographical issues: 'endpont' (Section III.B), 'sugggest thereofre' (Conclusions), 'Chemsitry Divison' and 'assitance' (Acknowledgments), and 'V ASP' in the DFT methods. These should be corrected.","section":"Throughout"},{"comment":"The notation O_i^(1)(s_k) and O_i^(2)(s_k) is defined verbally but the index j^(1) is not made explicit before use. A compact formal definition of j^(1)(i,s_k) would improve readability.","section":"Eqs. (13)–(14)"},{"comment":"In the two-level effective model, V_ij(s) is introduced but not defined in terms of the original dynamical matrix. It would be helpful to state that it is the off-diagonal matrix element of D(s) in the q_i,q_j basis, or otherwise relate it to the transport amplitudes a_ji.","section":"Section III.F, Eq. (17)"},{"comment":"The overlap map in Fig. 2 would benefit from a color scale (the caption says the color axis, but the axis is not labeled with values). In Fig. 3, the choice of 'representative' modes should be made reproducible by specifying the selection criterion, since the captions state the illustrations are representative examples.","section":"Fig. 2 and Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The reader's stress-test concern about the missing q-grid validation is, in my reading, the correct central issue: the entire experimental anchoring of the eigenvector-tracking analysis rests on the equivalence of Γ-point modes and powder-averaged TOSCA peaks. The perturbation derivation and numerical convergence analysis are solid, and the paper does not fit parameters to the target result, but the missing dispersion check must be addressed before the experimental validation can be trusted. A second, less prominent issue is that the authors deliberately select modes without apparent crossings for the detailed transport analysis, which limits the demonstrated generality of the framework in the very regions where it is most needed. Both are fixable in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the paper is worth reading: it gives a clean, practical way to track how normal modes change under isotope substitution in a dense phonon spectrum, and it shows that in ZIF-8 many apparent mode crossings are just under-sampling artifacts. Second, the experimental anchor is shakier than the authors claim: all the DFT modes are Γ-point, while the TOSCA spectra are powder-averaged over the Brillouin zone, and no q-grid or dispersion check is reported. That gap is load-bearing for the validation, not a cosmetic detail.\n\nThe genuinely new content is the adiabatic eigenvector-continuation framework: interpolating the mass matrix, propagating eigenvectors by maximum overlap, and using the overlap gap ΔO as a stability diagnostic. The grid-convergence study (Figs. 4 and 5) is a good idea and well executed; it makes a real point that coarse interpolation can manufacture crossings. The perturbation derivation in Appendix A is clean and standard. The observation that INS intensity and the isotope shift both depend on the hydrogen-displacement projection is physically correct and useful.\n\nThe soft spots, in rough order of severity. First, the Γ-point issue. The paper explicitly says \"All modes considered in this work correspond to Γ-point normal modes.\" TOSCA measures a BZ average. If ZIF-8's 400–1200 cm⁻¹ modes have appreciable dispersion, the measured peaks are not the same objects as the tracked modes, and the Fig. 1 comparison doesn't validate the eigenvectors used for the overlap and transport analysis. This needs a q-sampling check or at least a frank discussion. Second, the \"quantitative predictor\" is never actually tested: the authors assert that isolated-peak intensity predicts the isotope shift, but they don't compare predicted shifts to measured deuterated intensities across a set of modes. Third, there is an internal inconsistency about the crossing count: the text says crossings tend asymptotically to zero, but their own fit in Appendix B has a constant C=6.8, so the asymptote is about 7 crossings, not zero. The targeted refinement reduces them to 3, but that is still not zero. Fourth, the detailed transport-amplitude analysis deliberately selects modes that do not undergo apparent crossings, so the \"collective mixing\" examples are chosen to be clean, which is fine but limits how representative they are.\n\nNo serious circularity: the equations are standard perturbation identities, and nothing is fitted to the target result. There is no data or code deposit, so reproducing the ZIF-8 calculation and spectra would take real effort.\n\nWho will get value? People working on phonon assignments in MOFs and molecular crystals, and anyone using isotope substitution to interpret INS. The framework deserves a serious referee, but the revision should either add a q-grid validation or soften the experimental claims.","headline":"A useful eigenvector-tracking framework for isotope-substitution phonon problems, with a clean derivation and a nice convergence check, but the experimental validation rests on Γ-point modes without a q-grid check, and the quantitative predictor is never actually tested.","tokens_in":12607,"tokens_out":2755,"would_cite":true,"duration_ms":27129,"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":"Isotope substitution in dense phonon systems rotates vibrational eigenvectors within coupled subspaces, so mode correspondence between isotopic endpoints cannot be inferred from frequency shifts alone.","keywords":["isotope substitution","phonon eigenvectors","eigenvector continuation","ZIF-8","inelastic neutron scattering","mode mixing","avoided crossings","metal-organic frameworks"],"falsifier":"Compute the phonon dispersion of ZIF-8 on a dense q-point mesh and simulate the powder INS spectrum including q-dependence. If the simulated fingerprint-region spectrum changes substantially relative to the zone-centre-only calculation, or if the H/D eigenvector overlap statistics differ once finite-q modes are included, then the experimental validation and the tracked mode identities are undermined. A cleaner experiment: use oriented or single-crystal INS to resolve individual modes and check directly whether the 734 cm^-1 protiated mode's displacement character is actually transferred to the","tokens_in":11681,"feed_emoji":"🔀","tokens_out":6075,"duration_ms":54972,"temperature":0.7,"pith_summary":"The paper argues that replacing hydrogen with deuterium in a complex solid does more than rescale vibrational frequencies: in spectrally dense regions it rotates the shapes (eigenvectors) of the vibrations, so a mode at the starting isotope cannot be identified by matching frequencies at the endpoint. Using ZIF-8 as a test case, the authors show that inelastic neutron scattering intensity—weighted by hydrogen displacement—can flag which modes will shift predictably and which will undergo collective reorganisation. They introduce an adiabatic eigenvector-continuation scheme that tracks each mode through a continuous mass change, with overlap-based assignment and stability diagnostics, and find that apparent mode crossings on coarse grids are undersampling artefacts. The result reframes vibrational identity in framework materials as a continuous trajectory in eigenvector space rather than a frequency-ordered list.","feed_headline":"Isotope swaps remix vibrations in crowded spectra","feed_subtitle":"Neutron intensity predicts which modes shift cleanly and which rewire; eigenvector tracking settles isotope mode maps.","key_machinery":"The machinery is the one-parameter family of mass-weighted dynamical matrices D(s) = M(s)^(-1/2) Φ M(s)^(-1/2), built by linearly interpolating the mass matrix between protiated and deuterated forms while keeping the force-constant matrix fixed. On a discrete grid in s, each eigenvector is continued by the maximum-overlap assignment rule, and the gap between the largest and second-largest overlap quantifies tracking stability. First-order perturbation theory supplies the transport amplitudes a_ji = (q_j^T δD q_i)/(ω_i^2 - ω_j^2), with the two amplification channels—small frequency denominators and large coupling matrix elements—explaining collective rotation and 'action at a distance' mixing","core_discovery":"The core claim is that mass substitution is a parametric Hermitian deformation of the mass-weighted dynamical matrix, and as the mass parameter s runs from hydrogen to deuterium, eigenvectors in spectrally congested regions rotate substantially within coupled subspaces; hence continuity of frequency ordering does not imply continuity of vibrational identity. In ZIF-8 the paper demonstrates a progression from near-ideal behaviour (416 to 374 cm^-1, preserved displacement pattern) to quantitative redistribution (639 to 546 cm^-1, altered amplitudes) to qualitative mixing (734 to 553 cm^-1, new methyl-rocking participation). Because one-phonon INS intensity is proportional to hydrogen displacem","pith_inferences":["The same continuation machinery should carry over to other parametric perturbations of the dynamical matrix—pressure, strain, or chemical substitution—though there the potential surface also changes, making the mass-only interpolation an approximation rather than exact.","Because the paper works with zone-centre modes while INS averages over a powder, a q-resolved version of the tracking would show whether the observed mixing and long-range coupling persist away from the zone centre; this is a direct extension.","The identification of structural similarity (shared isotope-sensitive coordinates) as a strong coupling channel suggests selection rules for long-range eigenvector transport that could predict mixing without running the full interpolation.","The intensity-based predictor could be developed into a routine experimental check of force-field quality in framework materials, using a single protiated spectrum to anticipate the full deuteration response."],"forward_implications":["In spectrally sparse regions, an isolated peak's INS intensity directly predicts the size of its deuteration frequency shift, because both derive from the same hydrogen-displacement projection.","In congested regions, comparing protiated and deuterated spectra reveals whether neighbouring modes share hydrogen displacement weight, thus whether collective eigenvector reorganisation will occur.","Mode correspondence between isotopic endpoints should be built from overlap-based eigenvector continuation, not frequency ordering; coarse-grid apparent crossings are undersampling artefacts and disappear on refinement.","Deuteration, including selective deuteration of specific groups, becomes a practical tool to deliberately shift frequencies or redesign vibrational displacement patterns in framework materials."],"fun_headline_variants":["Isotope swaps scramble phonon eigenvectors in crowded spectra","Deuteration rewires vibrations where modes crowd together","Neutron intensity predicts isotope-driven mode makeovers","Vibrational identity flows with isotope mass"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper assumes that the zone-centre normal modes computed by density functional theory are the same vibrations the powder neutron experiment measures—i.e., that phonon dispersion and the wavevector dependence of eigenvectors are negligible across the measured fingerprint region—yet no q-grid or dispersion check is reported.","fun_headline_variants_meta":{"raw":{"variants":["Isotope swaps scramble phonon eigenvectors in crowded spectra","Deuteration rewires vibrations where modes crowd together","Neutron intensity predicts isotope-driven mode makeovers","Vibrational identity flows with isotope mass"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000251,"raw_usage":{"total_tokens":1391,"prompt_tokens":741,"completion_tokens":650,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":485,"completion_tokens_details":{"reasoning_tokens":588}},"tokens_in":485,"tokens_out":650,"duration_ms":7844,"temperature":1.0,"reasoning_tokens":588,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T21:11:56.540274+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the phonon dispersion of ZIF-8 on a dense q-point mesh and simulate the powder INS spectrum including q-dependence. If the simulated fingerprint-region spectrum changes substantially relative to the zone-centre-only calculation, or if the H/D eigenvector overlap statistics differ once finite-q modes are included, then the experimental validation and the tracked mode identities are undermined. A cleaner experiment: use oriented or single-crystal INS to resolve individual modes and check directly whether the 734 cm^-1 protiated mode's displacement character is actually transferred to the","supporting_citations":[],"review_version":1}