{"id":"0357c59d-46c9-46ea-9bec-614f5a3dbebd","arxiv_id":"2504.17464","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Green-Kubo and Lorentz-model infrared dielectric functions of MgO and LiH become consistent when the Lorentz model uses a frequency-dependent phonon self-energy from MD and electronic polarization is included via an ε∞ multiplier or Born effective charges.","lead":"This paper benchmarks two standard methods for predicting infrared optical properties of polar crystals, the Green-Kubo formula and the Lorentz model, using molecular dynamics with rigid-ion and machine-learned potentials. It shows the methods agree only when the Lorentz model is generalized with a frequency-dependent phonon self-energy and when electronic polarization is treated with an ε∞ multiplier or Born effective charges.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Classical MD self-energy restricts the reconciliation to the classical regime; the 8 K MgO comparison in Fig. 4 exposes the quantum/isotope failure, and PIMD is not propagated through the Lorentz self-energy construction.","rationale":"The reader identified the same load-bearing assumption: the classical MD velocity-correlation function, via the Kubo transform of Eq. 9, must yield a phonon self-energy that faithfully represents anharmonic and multi-phonon processes. I agree, and I sharpen the concern by pointing to the paper's own Fig. 4 and Sec. III D: classical MD fails at 8 K for MgO, and the PIMD improvement is not carried through the self-energy/Lorentz branch of the calculation. This does not invalidate the paper's core demonstration for room temperature and above, where the two methods agree and comparison with experiment is good; it does mean the central claim should be scoped to the classical-statistics regime. The paper itself states that atomic dynamics are essentially classical and that classical MD fails at cryogenic temperatures, so the limitation is partially acknowledged, but it is absent from the headline claim. The proposed PIMD-based test would directly separate the classical-statistics limitation from a possible failure of the Lorentz self-energy parameterization itself. The reader's CONDITIONAL verdict seems appropriate; my stress-test does not move it.","tokens_in":18834,"tokens_out":20536,"duration_ms":218666,"concrete_test":"Run the self-energy extraction (Eqs. 9, 14, 15) on PIMD trajectories of MgO at 8 K and 300 K, insert Pi(omega) into Eq. 8, and compare the resulting reflectance with the classical-MD Green-Kubo curve, the PIMD Green-Kubo curve, and experiment in Fig. 4. If the PIMD-based Lorentz reflectance tracks the PIMD-GK curve and experiment at 8 K while the classical Lorentz curve does not, the classical-statistics assumption is the bottleneck and the reconciliation can be extended to the quantum regime. If PIMD-Lorentz fails to reproduce PIMD-GK, the single-self-energy Lorentz parameterization itself is the limiting step rather than the classical approximation alone.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central unification rests on the classical Kubo transform used in Eq. 9 to convert the MD modal-velocity correlation function into a phonon self-energy, and on Eq. 14 to extract Gamma(omega). Because the self-energy is a functional of classical trajectories, it inherits classical statistics: no zero-point motion and no isotope-disorder scattering. The paper's own Fig. 4 shows the classical-MD Green-Kubo reflectance for MgO at 8 K deviates from experiment, while perturbation theory is closer; the same classical self-energy would be fed into the Lorentz model, so the two-method agreement at 8 K is agreement between two equally classical quantities. The PIMD comparison in Sec. III D is performed only at the Green-Kubo level and is not propagated through the self-energy/Lorentz construction. Thus the claimed reconciliation is demonstrated only in the regime where classical nuclear statistics and moderate anharmonicity hold; the introduction's claim that 'Green-Kubo formula reconciles the Lorentz model parameterized with phonon self-energy extracted from MD simulation' lacks an explicit qualification of this regime.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript compares two routes to the infrared dielectric function of polar crystals, MgO and LiH: the Green–Kubo formula evaluated from equilibrium molecular dynamics and the Lorentz model parameterized either by a constant TO linewidth from spectral energy density analysis or by a frequency-dependent phonon self-energy extracted from modal velocity correlation functions. Using both a rigid-ion BKS potential and a NEP machine-learned potential, it shows that the simple Lorentz model misses multi-phonon absorption features, while the Lorentz model with an MD-derived self-energy reproduces the Green–Kubo spectra. It further argues that, with the rigid-ion model, the electronic polarization correction is multiplicative through the factor ε∞ (Eq. 28), whereas with the NEP the electronic contribution is captured by assigning Born effective charges when computing the dipole moment. The paper also compares with experiment and perturbation theory and includes a path-integral MD analysis of nuclear quantum effects.","tokens_in":19210,"tokens_out":8422,"duration_ms":85842,"significance":"If the claims hold, the paper supplies a useful practical recipe: MD-derived phonon self-energies can be injected into a Lorentz-form dielectric function to capture anharmonic and multi-phonon infrared absorption, and the ε∞ correction for rigid-ion MD clarifies an often approximate treatment. Strengths include the two-material/two-potential benchmark, the comparison against experimental reflectance and perturbation theory in Fig. 4, the force-error sensitivity test in Appendix C, and openly available NEP training data. The principal caveat is that the Green–Kubo and Lorentz-plus-self-energy curves are derived from the same MD trajectories, so their agreement is a consistency check between two reduction formulas rather than independent cross-validation; the manuscript should state this explicitly and qualify the regime of validity of the claimed unification.","major_comments":[{"comment":"The Lorentz model with Π(ω) obtained from Eqs. (9)–(15) and the Green–Kubo susceptibility of Eq. (1) are both computed from the same equilibrium MD trajectories. The close agreement in Figs. 3(a)–3(f) therefore demonstrates internal consistency of two spectral reductions, not an independent cross-validation of the underlying dynamics. The Introduction's statement that 'the Green-Kubo formula reconciles the Lorentz model parameterized with phonon self-energy extracted from MD simulation' and the Abstract's 'cross-validation' wording should be tempered: the independent tests are the experimental reflectance and the perturbation-theory comparison, while the GK–Lorentz agreement tests the single-mode parametrization. I recommend adding an explicit statement of this distinction and, if feasible, an analytical demonstration of how Eq. (8) with Π from Eq. (14) reduces to the same spectral function as Eq. (1) for a single infrared-active mode.","section":"Section III C and Fig. 3"},{"comment":"The extraction of Γ(ω) uses the classical Kubo transform, so the phonon self-energy inherits classical statistics: there is no zero-point motion and no isotope-disorder scattering. The authors' own Fig. 4 shows that the MD-NEP reflectance at 8 K deviates from experiment while perturbation theory is much closer, and the PIMD result is propagated only at the Green–Kubo level, not through the self-energy/Lorentz construction. Consequently, the reconciliation is demonstrated only in the regime of classical nuclear statistics and moderate anharmonicity. This regime restriction should appear near the Introduction's central claim and in the Abstract, not only in the Conclusion.","section":"Section II B, Eq. (9), and Section III D, Fig. 4"},{"comment":"For the NEP calculations, fixed DFPT Born effective charges are assigned only when computing the dipole moment, while the underlying potential has no explicit charge or polarization degrees of freedom and was trained on neutral DFT cells. The paper correctly notes that the mechanism remains to be fully elucidated, but this approximation is load-bearing for the MLP branch of the unification claim. A quantitative test would strengthen the argument, such as comparing the ε∞εion result obtained from the RIM charges with the Z*-based NEP result, or checking against density-functional perturbation theory mode effective charges. At minimum, the statement that MLP 'automatically' includes the electronic contribution should be softened to an effective description in which the fitted forces absorb polarization effects while the dipole is assigned via fixed Born effective charges.","section":"Section II D 2"}],"minor_comments":[{"comment":"The caption contains a duplicate panel label '(d)' for the MgO reflectance panel; the MgO panels should be labeled (a)–(c) and the LiH panels (d)–(f).","section":"Fig. 3 caption"},{"comment":"The summation notation '3,n' in the SED expression is unclear; the index ranges should be written explicitly, for example as a sum over the 3 Cartesian components and the n basis atoms.","section":"Eq. (6)"},{"comment":"The sentence contains the typo 'simulation detals'; also, the statement that the 10×10×10 supercell 'has been tested to be sufficiently capture the long-range interaction well' should be rephrased.","section":"Section II D"},{"comment":"A brief statement on statistical uncertainty or convergence of the MD spectra with respect to trajectory length and supercell size would help the reader assess how robust the agreement in Fig. 3 is.","section":"Section III C"},{"comment":"The text says PIMD brings the 8 K MgO reflectance closer to experiment, but it should state explicitly at which temperature the PIMD curve in Fig. 4 is computed and whether PIMD data at 295 K and 950 K are also shown or only the 8 K case.","section":"Section III D"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope and the derivations are mostly sound, but the central claim is currently framed as a reconciliation/cross-validation when the GK–Lorentz agreement is largely a same-trajectory consistency check. The revision should reframe the abstract and introduction, and explicitly delimit the classical-statistics regime; the 8 K MgO failure and the GK-only PIMD comparison should be connected to that limitation. With those changes the paper is likely acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about this paper. First, its real contribution is practical: it shows that a Lorentz model fed with an MD-derived frequency-dependent phonon self-energy reproduces the Green-Kubo dielectric function, including the multiphonon absorption that a constant linewidth misses. Second, the ε∞ multiplicative correction for rigid-ion MD (Eq. 28, ε = ε∞ ε_ion) is a small but genuine fix for a systematic error in earlier work that simply added ε∞ − 1. Anyone computing IR spectra from RIM-based MD should read that section.\n\nWhat is solid: the Born-Huang chain derivation in Sec. II C is clean, the self-energy extraction via Kubo transform and Dyson's equation is standard machinery, and the MD self-energy is compared against experimental fits in Appendix E, which is a real external anchor. The benchmarks are also careful: identical potentials for both routes, MgO and LiH chosen for different anharmonicity, and RIM and MLP both tested. Citation practice is fair; the earlier GK-MD papers are cited and the difference with them is concrete. The trained potentials and data are on GitHub. The paper is also honest about its failure modes: Fig. 4 at 8 K shows classical MD missing the experiment, and the conclusion states that classical MD fails at cryogenic temperatures because of isotope disorder and nuclear quantum effects.\n\nWhere it is soft. The two sides of the claimed reconciliation are not independent; both are functionals of the same classical MD trajectories. The agreement in Fig. 3 is therefore a self-consistency check of the extraction procedure and the Lorentz parameterization, not a validation of the potential or the classical statistics. The text frames this as 'excellent consistency' without flagging the shared origin of the data; one sentence noting that the agreement is expected and that the real tests are the experimental comparisons would sharpen the claim. Relatedly, the introduction says Green-Kubo 'reconciles' the Lorentz model, which is too broad: the demonstration holds in the classical, moderate-anharmonicity regime. The stress-test point about PIMD is fair; the PIMD results are only shown at the Green-Kubo level, so there is no quantum version of the self-energy route yet. No error bars anywhere, which is minor for this kind of benchmark. The RIM results deviate from experiment for both materials, and the authors attribute this to two-body potential limitations, which is plausible.\n\nWho it is for: people computing temperature-dependent IR dielectric functions or thermal radiative properties from MD, RIM users who need the ε∞ correction, and anyone who wants a compact Lorentz-parameterized form that still carries multiphonon physics. For peer review: yes, it deserves a serious referee. The claims are substantive, the work is reproducible, and the main limitations are acknowledged inside the paper. A referee should push on the scoping of the reconciliation claim and on error estimation, but neither is a load-bearing flaw.","headline":"A useful, honest benchmark paper showing that a Lorentz model with an MD-derived frequency-dependent phonon self-energy reproduces the Green-Kubo dielectric function, plus a real ε∞ correction for rigid-ion models; the main caveat is that both sides come from the same classical trajectories.","tokens_in":19611,"tokens_out":4863,"would_cite":true,"duration_ms":46253,"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 Green-Kubo formula and the Lorentz model give unified predictions of infrared dielectric functions when the Lorentz model is parameterized with a frequency-dependent phonon self-energy from molecular dynamics and electronic…","keywords":["infrared dielectric function","Green-Kubo formula","Lorentz model","phonon self-energy","multi-phonon absorption","molecular dynamics","polar materials","Born effective charge"],"falsifier":"Repeat the two-route comparison on a third strongly anharmonic polar crystal at a temperature where classical statistics should hold—say 300 K for a heavier-ion oxide—using the same machine-learned potential; if the Lorentz model fed with the MD-derived self-energy does not match the Green-Kubo dielectric function wherever multi-phonon features appear, the claimed reconciliation is not general.","tokens_in":18655,"feed_emoji":"🔬","tokens_out":8947,"duration_ms":79201,"temperature":0.7,"pith_summary":"This paper argues that two standard routes to the infrared dielectric function of polar crystals—the Green-Kubo formula computed from equilibrium molecular dynamics and the phenomenological Lorentz model—can be made quantitatively consistent. It shows that the usual Lorentz model with a constant linewidth misses the multi-phonon absorption that the Green-Kubo correlation function contains, and that replacing the linewidth by a frequency-dependent phonon self-energy removes the discrepancy. It also identifies how electronic polarization must enter each approach: a multiplicative $\\varepsilon_\\infty$ factor for rigid-ion models, and Born effective charges for machine-learned potentials. Demonstrations on MgO and LiH support the claim, with classical molecular dynamics accurate at high temperature and perturbation theory better at cryogenic temperature. If the reconciliation is right, molecular dynamics becomes a dependable predictor of temperature-dependent infrared spectra, including multiphonon features.","feed_headline":"Phonon self-energy unifies MD and Lorentz infrared spectra","feed_subtitle":"Frequency-dependent damping reproduces multiphonon absorption, so the two standard methods agree.","key_machinery":"The central object is the frequency-dependent phonon self-energy $\\Pi(\\omega)$ of the infrared-active transverse-optical phonon, which generalizes the constant linewidth of the textbook Lorentz model. It is extracted from equilibrium molecular dynamics by projecting atomic velocities onto phonon modes, constructing the retarded single-phonon Green's function through the Kubo transform, and inverting Dyson's equation; the imaginary part gives the frequency-dependent damping and the real part the anharmonic frequency shift. Inserting $\\Pi(\\omega)$ into the generalized Lorentz model (Eq. 8) is what makes the dielectric function carry multi-phonon absorption. The second piece of machinery is the electronic-polarization correction: a derivation from the standard oscillator equations shows that rigid-ion models must multiply the ionic dielectric function by $\\varepsilon_\\infty$ rather than add $\\varepsilon_\\infty-1$, while machine-learned potentials supply the electronic response implicitly through Born effective charges when the dipole moment is evaluated.","core_discovery":"The central claim is that the Green-Kubo formula and the Lorentz model are the same physics expressed twice, once through dipole-moment fluctuations and once through damped oscillator response, and that they agree quantitatively when the Lorentz model is fed the full frequency-dependent phonon self-energy $\\Pi(\\omega)=\\Delta(\\omega)-i\\Gamma(\\omega)$ extracted from MD velocity-correlation functions via the Kubo transform and Dyson's equation. With this generalized damping, the Lorentz model reproduces the multi-phonon absorption structure of the Green-Kubo dielectric function for MgO and LiH, which a constant linewidth $\\tau^{-1}$ cannot do. The paper further claims that electronic polarization is not a simple additive background: for a rigid-ion model the correct correction is multiplicative, $\\varepsilon=\\varepsilon_\\infty\\varepsilon_{\\mathrm{ion}}$, derived from the standard ionic-polarization equations, while a machine-learned potential already encodes the electron-polarization response in its atomic dynamics and needs only Born effective charges to compute the dipole moment. Classical molecular dynamics with a machine-learned potential achieves quantitative agreement with experiment at 295 K and 950 K for MgO but fails at 8 K, where perturbation theory and path-integral methods perform better.","pith_inferences":["The same self-energy route should extend to other infrared-active crystals and to anisotropic or two-dimensional polar materials, where multi-phonon features and frequency-dependent damping are even more prominent; a test on such a material would sharpen the generality of the reconciliation.","The multiplicative $\\varepsilon_\\infty$ correction implies that the static dielectric constant predicted by rigid-ion molecular dynamics should be multiplied by $\\varepsilon_\\infty$, a prediction that can be checked directly against measured $\\varepsilon(0)$ across a series of alkali halides.","If the equivalence is exact in the classical regime, the phonon self-energy from molecular dynamics can be used as a temperature-dependent input to mesoscale radiative-heat-transfer models, giving them an atomistic grounding without running MD at every design point.","The observed failure of the additive correction suggests prior rigid-ion molecular-dynamics studies of near-field radiative heat transfer may have systematically underestimated infrared responses; recomputing those quantities with the multiplicative correction is a concrete test of the paper's mechanism."],"forward_implications":["Molecular dynamics and the Lorentz model can now be cross-validated: either route yields the same infrared dielectric function when the Lorentz model uses an MD-derived frequency-dependent self-energy, so results from one method can be checked against the other.","The conventional rigid-ion correction $\\varepsilon_{\\mathrm{GK}}+\\varepsilon_\\infty-1$ underestimates the ionic infrared response; the physical correction is $\\varepsilon_{\\mathrm{GK}}$ multiplied by $\\varepsilon_\\infty$, which changes the predicted longitudinal-optical frequency and the reflectance outside the main reflection band.","Machine-learned potentials trained on DFT data, combined with Born effective charges, reproduce the measured infrared reflectance of MgO at room and elevated temperatures, including multi-phonon absorption.","Classical molecular dynamics is reliable for infrared spectra at elevated temperatures but not at cryogenic temperatures, where quantum nuclear effects and isotope scattering dominate; perturbation theory handles that regime better, and path-integral molecular dynamics partially restores the agreement.","The generalized Lorentz model with a frequency-dependent self-energy, not a constant linewidth, is the correct phenomenological target for atomistic predictions of temperature-dependent infrared optical properties."],"supporting_citations":[{"why":"Supplies the Green-Kubo formula connecting dipole fluctuations in polar crystals to infrared absorption, the foundation of the molecular-dynamics route.","marker":"[19]"},{"why":"Showed that rigid-ion Green-Kubo results give $\\varepsilon(\\omega)/\\varepsilon_\\infty$, motivating the multiplicative electronic-polarization correction analyzed in this paper.","marker":"[21]"},{"why":"Represents the earlier additive-correction practice of adding $\\varepsilon_\\infty-1$, which the paper argues underestimates the ionic infrared response.","marker":"[20]"},{"why":"Provides the generalized Lorentz model with a frequency-dependent phonon self-energy used to capture multi-phonon absorption.","marker":"[44]"},{"why":"Provides the Kubo transform linking classical velocity-correlation functions to the retarded phonon Green's function, the step that extracts the self-energy from MD.","marker":"[50]"},{"why":"Supplies the spectral-energy-density method used to extract phonon frequencies and linewidths for the conventional Lorentz parameterization.","marker":"[38]"},{"why":"Provides experimental phonon self-energy spectra used to validate the MD-derived self-energy of MgO.","marker":"[32]"},{"why":"Provides the perturbation-theory reflectance curves compared against molecular dynamics at 8 K, 295 K, and 950 K.","marker":"[34]"}],"fun_headline_variants":["Phonon self-energy reconciles two infrared methods","One dielectric theory: Green-Kubo meets Lorentz","Self-energy fixes Lorentz model's multiphonon gap","MD and Lorentz unified by frequency-dependent damping","Infrared dielectric function: two methods, one answer"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument holds only if the thermal motion of atoms in a classical molecular dynamics simulation reproduces the same anharmonic vibrations that cause real infrared absorption, so that the damping pulled out of simulated velocity correlations faithfully represents the crystal's actual multiphonon processes.","fun_headline_variants_meta":{"raw":{"variants":["Phonon self-energy reconciles two infrared methods","One dielectric theory: Green-Kubo meets Lorentz","Self-energy fixes Lorentz model's multiphonon gap","MD and Lorentz unified by frequency-dependent damping","Infrared dielectric function: two methods, one answer"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000711,"raw_usage":{"total_tokens":3205,"prompt_tokens":956,"completion_tokens":2249,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":572,"completion_tokens_details":{"reasoning_tokens":2186}},"tokens_in":572,"tokens_out":2249,"duration_ms":14182,"temperature":1.0,"reasoning_tokens":2186,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:39:49.406684+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the two-route comparison on a third strongly anharmonic polar crystal at a temperature where classical statistics should hold—say 300 K for a heavier-ion oxide—using the same machine-learned potential; if the Lorentz model fed with the MD-derived self-energy does not match the Green-Kubo dielectric function wherever multi-phonon features appear, the claimed reconciliation is not general.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Green-Kubo formula connecting dipole fluctuations in polar crystals to infrared absorption, the foundation of the molecular-dynamics route."},{"cited_title":"Simoncelli, N","cited_arxiv_id":null,"evidence_quote":"Showed that rigid-ion Green-Kubo results give $\\varepsilon(\\omega)/\\varepsilon_\\infty$, motivating the multiplicative electronic-polarization correction analyzed in this paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Represents the earlier additive-correction practice of adding $\\varepsilon_\\infty-1$, which the paper argues underestimates the ionic infrared response."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Kubo transform linking classical velocity-correlation functions to the retarded phonon Green's function, the step that extracts the self-energy from MD."},{"cited_title":"Fugallo, B","cited_arxiv_id":null,"evidence_quote":"Supplies the spectral-energy-density method used to extract phonon frequencies and linewidths for the conventional Lorentz parameterization."},{"cited_title":"Shanker and S","cited_arxiv_id":null,"evidence_quote":"Provides experimental phonon self-energy spectra used to validate the MD-derived self-energy of MgO."},{"cited_title":"Carati and A","cited_arxiv_id":null,"evidence_quote":"Provides the perturbation-theory reflectance curves compared against molecular dynamics at 8 K, 295 K, and 950 K."}],"review_version":1}