{"id":"03c99507-d47f-454f-b32f-018a19eef382","arxiv_id":"2607.20788","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Interpolating phonons over a localized transition force yields converged Huang-Rhys sidebands for defects without giant supercells, validated on the NV- center in diamond.","lead":"Defect optical spectra are smeared into a phonon sideband by lattice vibrations; this paper computes the sideband from a small, localized 'transition force' plus interpolated vibrations on grids equivalent to ~17 million atoms. On the NV- center in diamond it reproduces the measured emission spectrum, including the 63 meV band and fine phonon structure.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Hypercell phonons are those of a periodic defect array; the paper asserts but never tests that the 4×4×4 force-constant kernel is free of defect-image contamination, so the isolated-defect equivalence is unproven.","rationale":"The reader's weakest assumption is precisely this: the equivalence between the phonons of a periodic defect supercell array and the vibrational continuum of an isolated defect requires that the force-constant perturbation be converged with respect to defect-image separation. The paper states this condition in Sec. 5 but never directly measures the range of the force-constant perturbation, and no supercell larger than 4×4×4 is tested. My reading of the derivations finds no internal algebraic error: Eq. (S_{q,ν}) follows correctly from D Δx = f, and the 1/N_q normalization is right. The missing piece is empirical validation of the key physical assumption. This does not invalidate the method, but it means the central claim is conditionally supported. The experimental agreement for emission is encouraging, but it is a single system and a broadened lineshape comparison; it does not by itself certify that the periodic-array phonons have converged to the isolated-defect limit. The proposed ΔΦ range test would settle whether the 4×4×4 kernel is large enough; if it is, the method is on much firmer footing. Therefore the reader's CONDITIONAL verdict remains appropriate, and I do not recommend changing it.","tokens_in":23959,"tokens_out":11373,"duration_ms":98637,"concrete_test":"Compute the real-space force-constant perturbation ΔΦ(R) = Φ_NV(R) − Φ_diamond(R) in the 4×4×4 cell (the paper already has both force-constant sets). If |ΔΦ| for atom pairs at the supercell boundary (R ≳ 4 Å) is larger than ~1% of the on-site/maximum value, the kernel is image-contaminated and a 5×5×5 (or embedding [2]) calculation is needed. Alternatively, rerun the entire workflow with a 5×5×5 supercell and compare Stot and α to the 4×4×4 extrapolation; a shift in Stot larger than ~0.1 or in α larger than ~10% would confirm that the 4×4×4 kernel is not converged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The core interpolation step (Sec. 5) builds hypercell dynamical matrices by Fourier-transforming the force-constant kernel of the defect-containing supercell. The resulting phonons are exactly those of a periodic array of defects, not of an isolated defect. The paper's stated escape clause is that the local static Green function of the hypercell becomes indistinguishable from that of an isolated defect 'if the defect-containing supercell is sufficiently large that the defect-induced perturbation of the force-constant kernel is converged with respect to the defect-image separation' (Sec. 5, final paragraph). This condition is never checked. The presented evidence — localization of the transition force (Fig. 2, Table 1) and convergence of S(ℏω) with hypercell size (Fig. 3) — does not certify it. Force localization constrains the source f^tr, not the kernel Φ; and increasing hypercell size only refines q-point sampling within the same supercell Φ, so it cannot detect errors in those force constants. The linear acoustic scaling S(ℏω)=αℏω follows from Σ f^tr=0 and is independent of force-constant convergence. Hence periodic-image contamination of Φ (or of the reconstructed f^tr) could silently shift the acoustic slope, the 63 meV band shape, and the van Hove features without being visible in any reported convergence test.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a phonon-interpolation method for computing Huang–Rhys spectral densities and optical lineshapes of point defects. The central idea is to reconstruct a localized transition-induced force from DFT supercell calculations via the harmonic relation F̃tr = Φ̃f ΔR̃, then couple that force source to Fourier-interpolated dynamical matrices of a much larger hypercell. The partial Huang–Rhys factor is recast as S_{q,ν} = |Σ_a f^tr_a·e*_{ν,a}(q)e^{-iq·R_a}|²/(2N_qℏω³). The method is benchmarked on the NV⁻ centre in diamond: the transition force is shown to be localized in a 4×4×4 supercell, hypercells up to 32×32×32 are used to resolve the vibrational continuum, and the resulting emission lineshape is compared with experiment, recovering a dominant 63 meV band, linear low-energy acoustic scaling, and van Hove-related high-energy structure.","tokens_in":24138,"tokens_out":5489,"duration_ms":53841,"significance":"If the central equivalence between hypercell phonons and the isolated-defect vibrational continuum is certified, the method would be a practically valuable, systematically improvable route to converged defect spectral functions without prohibitively large direct supercell calculations. The algebraic reformulation in Eq. (22) is clean and appears exact within the stated harmonic, parallel-mode, equal-frequency approximation; the 1/N_q normalization is dimensionally consistent. The paper is also honest about the limitations of the parallel-mode model for absorption. A public implementation is provided, which strengthens reproducibility.","major_comments":[{"comment":"The central equivalence between the hypercell phonons (a periodic array of defect supercells) and the isolated-defect vibrational continuum is asserted but not tested. The condition that the defect-induced perturbation of the force-constant kernel is converged with respect to defect-image separation is never checked. The locality of the transition force (Fig. 2, Table 1) constrains only the source f^tr, not the kernel Φ. Hypercell-size convergence (Fig. 3) only refines q-point sampling on the same supercell Φ and cannot detect errors in those force constants. The authors should (i) compute the real-space difference ΔΦ = Φ_defect − Φ_pristine in the 4×4×4 cell and show it decays well before the cell boundary, and (ii) compare S(ℏω) obtained from 4×4×4 and at least one larger (e.g. 5×5×5) defect supercell at the same hypercell q-point density. Without such a test, periodic-image contaminat","section":"§6.2, Fig. 4; Eq. (22)"},{"comment":"The linear acoustic scaling S(ℏω)=αℏω is presented as a consistency check, but it is a necessary consequence of Σ_a f^tr_a=0 for a charge-conserving transition and of the ω→0 density of states; it does not certify convergence of the acoustic modes. The coefficient α≈387.7 eV⁻² depends on the first moment of the localized force and on the long-wavelength sound velocities of the hypercell, which are those of a periodic array of defects at finite concentration, not an isolated defect. The authors should test whether α and the low-energy S(ω) are stable when the defect supercell size is increased at fixed hypercell volume, or when the force-constant kernel tail is truncated or compared with the pristine kernel. As it stands, the reported linear scaling is a vanishing-net-force identity, not a convergence test.","section":"§6.2, Fig. 4; Eq. (22)"},{"comment":"The reconstructed transition force is obtained by applying the supercell force-constant matrix to the relaxed displacement field ΔR̃. Because ΔR̃ is the equilibrium displacement of a periodic array of defects, it contains the elastic response of that array. The reconstruction removes the long-range displacement only to the extent that Φ̃f is the exact harmonic kernel of the same array. The paper does not quantify the difference between the reconstructed f^tr and a force directly evaluated from DFT at the initial-state geometry, nor the anharmonic error in using the harmonic relation. A direct comparison in the 4×4×4 cell would test both the reconstruction and the locality assumption on which the method rests.","section":"§5; Eq. (17)"}],"minor_comments":[{"comment":"The caption and panels are difficult to parse because both the supercell size, hypercell size, and Gaussian smearing width σ vary simultaneously. The apparent convergence may partly reflect the decreasing σ. Please label each panel clearly with all three parameters and, ideally, show fixed-σ comparisons.","section":"Fig. 3"},{"comment":"The power-law exponent β differs substantially between the 2×2×2 (1.838) and the 3×3×3/4×4×4 cells (≈1.1–1.2). This undercuts the statement that the power-law behaviour is universal. Please discuss the sensitivity of the extrapolated S_tot to the fitting form and to the number of hypercell points included.","section":"Table 2"},{"comment":"The claim of 'excellent agreement' with experiment is qualitative. Please provide a quantitative discrepancy measure (e.g. integrated absolute difference over the sideband) and state the experimental uncertainty or the sensitivity to the 1 meV broadening choices.","section":"§6.2, Fig. 6"},{"comment":"The manuscript says the hypercell dynamical matrices are 'diagonalized' for a 32×32×32 hypercell corresponding to ~17 million atoms. In practice one diagonalizes the q-dependent dynamical matrix for each q. Please clarify the wording to avoid implying a direct diagonalization of a 17-million-atom matrix.","section":"§5, text near Eq. (23)"}],"recommendation":"major_revision","confidential_remarks":"The algebraic core is credible and the code availability is a plus. The main issue is the unverified isolated-defect equivalence of the hypercell phonons, which is load-bearing for the method. This is fixable with additional calculations and analyses, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this: the paper is a solid, useful method paper for defect phonon sidebands, with central algebra that checks out, but it rests on an unproven convergence assumption about periodic-image effects in the force-constant kernel.\n\nWhat's new is the formulation: instead of using the relaxed displacement field of a supercell (which includes the periodic array response), they reconstruct the transition-induced force, show it is localized (~6.7 participating atoms, RMS radius ~2.1 Å), and couple that localized source to Fourier-interpolated dynamical matrices on hypercells up to 32×32×32 via a δ_{L,0} source with 1/N_q normalization. That is a legitimate workaround for the dense-phonon-sampling bottleneck. The derivation of Sν from the force source is clean, and the convergence of S(ℏω) with hypercell size is convincing. The emission lineshape matches experiment well, including the 63 meV band and van Hove features. Credit where it's due: the paper is clearly written, the algebra is correct, and the limitations of the parallel-mode approximation are openly discussed.\n\nThe soft spots, in rough order of importance:\n\n1. The load-bearing claim is that phonons of a periodic array of defect supercells, in the source region, match an isolated defect. That is asserted rather than tested. The locality of the transition force is good evidence for the source, but not for the force-constant kernel. Enlarging the hypercell only refines q-point sampling inside the same supercell's force constants; it cannot detect defect-image contamination in those force constants. The linear acoustic scaling is a consequence of a vanishing net force and doesn't certify the acoustic modes. I'd want a direct measure of force-constant range, or a 5×5×5 supercell test, before trusting the method blindly. This is not a fatal issue—diamond force constants are short-ranged—but it's a real hole.\n\n2. The parallel-mode equal-frequency approximation is known to be shaky for absorption; they flag this honestly and show partial improvement with excited-state phonons. Fine for a benchmark, but it limits the claim.\n\n3. Reproducibility: code is on GitHub, but there's no pinned version or data archive. For a methods paper, that should be tightened.\n\nOverall, I'd send this to a serious referee. The method is novel, the execution is careful, and the NV− demonstration is strong enough to merit a hearing. The referee should be asked to focus on the periodic-image equivalence and to demand the missing convergence test.","headline":"A clever and mostly sound method for defect phonon sidebands that needs one missing convergence test before I'd fully trust it.","tokens_in":24824,"tokens_out":5336,"would_cite":true,"duration_ms":46292,"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":"A localized transition-force extracted from a 4×4×4 supercell, coupled to phonons interpolated on hypercells of up to 32×32×32, yields converged Huang-Rhys spectral densities and an emission lineshape in agreement with experiment for the ni","keywords":["phonon interpolation","Huang-Rhys spectral density","transition-induced force","nitrogen-vacancy centre","electron-phonon coupling","optical lineshape","hypercell","diamond defects"],"falsifier":"Compute the reconstructed transition force and the interpolated Huang-Rhys density with a 5×5×5 or larger first-principles supercell and compare with the 4×4×4 result. If the participation number, RMS localization radius, the position or weight of the 63 meV band, or the acoustic coefficient α change measurably, the claim that the 4×4×4 supercell represents the isolated-defect force source is refuted. A direct check is also to measure the real-space range of the force-constant perturbation itself, which the paper does not isolate.","tokens_in":23645,"feed_emoji":"💎","tokens_out":7646,"duration_ms":64710,"temperature":0.7,"pith_summary":"The paper establishes that for charge-conserving optical transitions between localized defect states in a gapped host, the force driving the lattice relaxation is concentrated in a small region around the defect: about seven effective ions and a root-mean-square radius near 2.1 Å in a 4×4×4 supercell. It then shows that this transition-induced force can be reconstructed from a first-principles supercell and used as a localized source that couples the defect to phonons obtained by Fourier-interpolating dynamical matrices on hypercells of up to 32×32×32 (about 17 million atoms). Applied to the nitrogen-vacancy centre in diamond, the procedure recovers a smooth, continuous Huang-Rhys spectral density with a dominant band near 63 meV and a linear low-energy acoustic contribution, and an emission lineshape that agrees with experiment over the full sideband. The approach is presented as a general route to converged defect spectral functions without direct vibrational calculations in prohibitively large cells, with the absorption case showing that final-state vibronic mode mixing remains a limitation of the parallel-mode harmonic treatment.","feed_headline":"A localized force from a 512-atom cell reproduces the NV- sideband","feed_subtitle":"Interpolated phonons on million-atom hypercells resolve the 63 meV band and the acoustic tail.","key_machinery":"The central object is the reconstructed transition-force source F_tr = Φ_f ΔR (mass-weighted f_tr = F_tr/√M), obtained from the relaxed initial- and final-state geometries in a finite supercell. The paper inserts this source into the partial Huang-Rhys factor S_{q,ν} = |Σ_a f^tr_a · e*_{ν,a}(q) e^{-iq·R_a}|² / (2 N_q ℏ ω_ν(q)^3), where e_ν(q) are eigenvectors of the interpolated dynamical matrix. A hypercell is built by repeating the defect supercell M₁×M₂×M₃ times; the supercell force-constant kernel is Fourier-interpolated in reciprocal space and diagonalized on a dense q-point grid. Because the transition force is nonzero in only one reference cell, the factor 1/N_q keeps the total Huang-","core_discovery":"The central discovery is that the transition-induced force—the difference between the Born-Oppenheimer forces of the final and initial electronic states evaluated at the same geometry—is a localized source, and that using this force as the source term of the harmonic relation F_tr = Φ_f ΔR removes the need to resolve the long-range displacement field in a large cell. The paper demonstrates the concept on the NV− centre: the force reconstructed from a modest 4×4×4 density-functional supercell has a participation number of about 6.7 and an RMS localization radius of about 2.1 Å, and coupling it to interpolated hypercell phonons yields converged Huang-Rhys spectral densities with a dominant 63","pith_inferences":["Because the locality argument rests on a charge-conserving transition in a gapped, weakly polar host, the same strategy should transfer to other colour centres in diamond, silicon carbide, or hexagonal boron nitride whenever the transition-induced density change is localized and the net charge is unchanged.","For polar hosts or transitions with a net change of charge, the long-range multipolar tail of the transition force could be added analytically through screened Born effective charges, extending the method beyond its current scope.","The emission-absorption asymmetry indicates that applications with strongly vibronically coupled final states will need a full multimode treatment; the interpolation machinery could be paired with such a treatment to test whether the excited-state surface alone explains the residual absorption discrepancy.","The interpolated phonon continuum could also be used for other phonon-sampling-limited quantities, such as non-radiative capture rates or temperature-dependent Debye-Waller factors, since those require the same dense vibrational sampling."],"forward_implications":["A defect transition's full phonon sideband can be computed from an ab initio supercell of a few hundred atoms plus an inexpensive interpolation step, instead of requiring direct vibrational spectra of million-atom cells.","Because both the long-wavelength acoustic tail and van Hove features are resolved, finite-temperature lineshapes and the thermal evolution of the zero-phonon-line region become computable within the same first-principles workflow.","The linear low-energy scaling of the Huang-Rhys density, S(ℏω)=αℏω with α≈387.7 eV⁻², is a consistency check that the force-source formulation avoids an artificial low-frequency divergence.","The total Huang-Rhys factor converges as a power law with hypercell size, enabling systematic extrapolation to the infinite-hypercell limit and yielding a Debye-Waller factor near 2.9% for the NV− transition.","Emission is captured well using ground-state phonons, while absorption requires excited-state phonons; a quantitatively complete absorption description needs to go beyond the parallel-mode equal-frequency harmonic approximation."],"fun_headline_variants":["Localized force unlocks NV- phonon sideband","Tiny defect cell, million-atom phonon grid","Phonon interpolation resolves defect optical spectra","NV- sideband from a single localized force","Small cell, dense phonons: defect spectra solved"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the phonons of the periodic array of defect supercells match, in the region where the transition force acts, the vibrational modes of an isolated defect embedded in the perfect host; if the defect-induced force-constant perturbation reaches the supercell boundary, the interpolated phonons and the resulting sidebands inherit the periodic-image error.","fun_headline_variants_meta":{"raw":{"variants":["Localized force unlocks NV- phonon sideband","Tiny defect cell, million-atom phonon grid","Phonon interpolation resolves defect optical spectra","NV- sideband from a single localized force","Small cell, dense phonons: defect spectra solved"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000235,"raw_usage":{"total_tokens":1397,"prompt_tokens":862,"completion_tokens":535,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":606,"completion_tokens_details":{"reasoning_tokens":471}},"tokens_in":606,"tokens_out":535,"duration_ms":5642,"temperature":1.0,"reasoning_tokens":471,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T09:25:40.539651+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the reconstructed transition force and the interpolated Huang-Rhys density with a 5×5×5 or larger first-principles supercell and compare with the 4×4×4 result. If the participation number, RMS localization radius, the position or weight of the 63 meV band, or the acoustic coefficient α change measurably, the claim that the 4×4×4 supercell represents the isolated-defect force source is refuted. A direct check is also to measure the real-space range of the force-constant perturbation itself, which the paper does not isolate.","supporting_citations":[],"review_version":1}