{"id":"c560609e-be6c-4142-a92a-01d648ad0758","arxiv_id":"2607.21045","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"All 15 Raman-active phonons of beta-Ga2O3 are separated, including previously unresolved B_g(3) and B_g(4) modes, and their relative Raman tensor elements are determined from birefringence-corrected polarization-angle-resolved Raman spectra.","lead":"This paper reports the phonon energies and relative Raman tensor elements of all 15 Raman-active modes of the semiconductor beta-Ga2O3 from polarization-angle-resolved Raman measurements on four crystal planes. The single-stage hyperspectral fitting procedure used here may also help separate overlapping Raman modes in other birefringent anisotropic crystals.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The newly reported B_g(3)/B_g(4) tensor elements are extracted only by imposing theoretical PARRS profiles built on a privately communicated dielectric tensor; no identifiability or sensitivity check is provided, so the quantitative separation claim is conditional on that model.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the single-stage fit's ability to separate overlapping modes and extract tensor elements depends on the theoretical PARRS profiles and the dielectric tensor from private communication. The paper gives no identifiability analysis, no uncertainties, and no independent validation of the new B_g values. This is a genuine weakness because the central claim is quantitative and the most novel outputs are precisely the previously unresolved B_g(3)/B_g(4) elements. The absence of raw data and the non-public alpha version of peak-o-mat2 further limit external checking. However, the concern does not by itself invalidate the result: the angular profiles for A_g and B_g modes are largely orthogonal (e.g., sin^2(2φ) vs cos^2(2φ) in crossed geometry), so some separation robustness is likely; the issue is that this has not been demonstrated for the reported precision. Therefore the appropriate verdict remains CONDITIONAL, not a rejection. The proposed synthetic-data and dielectric-perturbation test would concretely decide whether the condition is satisfied.","tokens_in":14353,"tokens_out":5429,"duration_ms":59910,"concrete_test":"Perform a synthetic-data identifiability and sensitivity test: generate hyperspectral datasets from the best-fit model with the reported noise level, then refit with (i) 100 random initializations and bootstrap resampling of residuals, and (ii) dielectric-tensor perturbations within literature uncertainty (e.g., ±0.05 on εxx, εyy, εzz and ±0.5° on the principal-axis angle α). Track recovery of the injected A_g(7)/B_g(4) and B_g(3)/B_g(4) tensor elements and report the shift in Table I values. If any newly reported B_g element shifts by more than ~10% or the pair assignment swaps, the separation is not robust to model uncertainty and the central claim should be softened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is full spectral separation and quantitative determination of all 15 Raman tensor elements, including the previously unresolved A_g(5)/B_g(3), A_g(7)/B_g(4), and B_g(5)/A_g(9) pairs. The separation is not a direct measurement: the single-stage fit in Sec. III uses the Kranert birefringence model (Eqs. 4-8) with numerical coefficients in Table S2 that depend on the dielectric tensor of Eq. 2, which is taken from private communication (Ref. [65]). For the A_g(7)/B_g(4) pair, separated by only 1.0 cm^-1, the fit can trade intensity between the overlapping modes if the dielectric tensor or the model's Fresnel/solid-angle corrections (Eqs. 6-7) are slightly off. The paper provides no identifiability analysis, no parameter uncertainties, and no independent validation of the newly determined B_g(3) and B_g(4) values; the stated agreement with theory [42] is not quantified. The sign of the c tensor element is also explicitly left unresolved, qualifying the 'complete' determination. Thus the quantitative output is load-bearing on model assumptions that are not stress-tested.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a single-stage hyperspectral fitting procedure for polarization angle-resolved Raman spectra (PARRS) of birefringent β-Ga2O3. The model explicitly accounts for birefringence using the Kranert approach (Eqs. 4-8) with a dielectric tensor taken from private communication (Eq. 2). Measurements on (100), (010), (001), and (2̄01) planes are combined to extract the energies and relative Raman tensor elements of all 15 Raman-active modes, including previously unresolved overlapping pairs A_g(5)/B_g(3), A_g(7)/B_g(4), and B_g(5)/A_g(9). The central claim is full spectral separation and quantitative determination of all tensor elements, with relative signs within the (a,b,d) and (e,f) groups, though the sign of c is not determined.","tokens_in":14718,"tokens_out":4179,"duration_ms":45613,"significance":"If the method is robust, it provides a significant advance in quantitative Raman spectroscopy of optically anisotropic materials, addressing a problem that has been considered 'pointless' by some earlier work. The explicit inclusion of Fresnel and solid-angle corrections in a global fit to hyperspectral data, rather than fitting each spectrum individually, is a valuable contribution. The residual plots suggest good agreement between model and data. However, the central quantitative output—the tensor elements in Table I—is reported without uncertainties, and no identifiability or sensitivity analysis is provided. The reliance on a privately communicated dielectric tensor and the unresolved sign of c further temper the claim of a 'complete' determination.","major_comments":[{"comment":"The central output—the relative Raman tensor elements a, b, c, d, e, f for all 15 modes in Table I—is reported without any uncertainties. This is particularly concerning for the overlapping pairs A_g(5)/B_g(3), A_g(7)/B_g(4), and B_g(5)/A_g(9), where the fit must separate modes separated by as little as 1.0 cm^-1. No confidence intervals, bootstrap results, or parameter-correlation analysis are provided. Without these, the reader cannot judge whether the reported values are statistically meaningful or whether the fit can trade intensity between overlapping modes. Please provide uncertainties (e.g., from the covariance matrix or a bootstrap) and discuss their implications.","section":"Sec. III, Table I"},{"comment":"The quantitative separation of overlapping modes rests on the theoretical PARRS profiles from the Kranert model, which depend on the dielectric tensor of Eq. (2). This tensor is from private communication (Ref. [65]) and is given without stated uncertainties. The Fresnel and solid-angle corrections in Eqs. (6)-(7) also depend on this tensor. No sensitivity analysis is shown: it is not demonstrated how the fitted tensor elements change when ε is varied within a plausible range or when model approximations are modified. This is load-bearing because a small model error could trade intensity between overlapping A_g and B_g modes. Please add a sensitivity analysis, or fit the dielectric tensor components as free parameters if the data constrain them.","section":"Sec. III, Eqs. (2), (6)-(7)"},{"comment":"The title and abstract claim 'complete Raman tensor determination,' yet the sign of the c tensor element is not determined (Sec. III: 'Solely the element c enters quadratically'). The paper suggests a (110) plane would link c to the other elements, but such a measurement is not reported. Thus the claim of completeness is overstated. Either modify the claim to 'magnitudes and relative signs within the (a,b,d) and (e,f) groups' or include a measurement that resolves the sign of c.","section":"Abstract/Table I and Sec. III"},{"comment":"The statement that the newly determined B_g(3) and B_g(4) tensor elements are 'in good agreement with the values predicted by theory [42]' is not quantified. Please provide a numerical comparison with the theoretical values and define what 'good agreement' means (e.g., within a certain percentage). This is important because these modes are the main new experimental result and the only validation offered for the separation method.","section":"Sec. III"}],"minor_comments":[{"comment":"The statement 'The error of the phonon energies can be considered well below the spectral resolution of 0.5 cm^-1' lacks justification from the fit residuals or a covariance analysis. Please provide a quantitative error estimate.","section":"Sec. III"},{"comment":"The analysis used a non-public alpha version of peak-o-mat2; for reproducibility, consider providing the fit model definition, representative data, and a version-controlled script, or at least a detailed pseudocode of the single-stage fitting procedure.","section":"Sec. II and Data Availability"},{"comment":"The contour plots are informative, but the color scale uses the square root of intensity while the residual plots use a linear scale, which can obscure systematic deviations in low-intensity regions. Consider presenting residuals with the same transformation for direct comparability.","section":"Fig. 3 and SI"}],"recommendation":"major_revision","confidential_remarks":"The manuscript presents a plausible and potentially useful fitting approach, and the residual plots are encouraging. However, the missing uncertainty quantification and sensitivity analysis are serious gaps, particularly because the separation of overlapping modes is model-dependent. The reliance on a private communication for a key input parameter (the dielectric tensor) without any test of its influence on the results is a concern that should be addressed before publication. The title's 'complete' claim is also too strong given the unresolved sign of c. I recommend major revision rather than rejection because the core method appears sound and the missing pieces can be supplied within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth your time. The genuinely new thing here is the fitting procedure: instead of fitting each spectrum slice and then fitting the angle profiles in a second step, they fit the whole hyperspectral dataset in one pass, with Lorentzian peaks multiplied by the predicted angle-dependent PARRS profiles. That is a real improvement for strongly overlapping modes, and it pays off by giving the first experimental values for B_g(3) and B_g(4), which earlier work couldn't separate. They also report all 15 modes' energies and tensor elements, and the residuals in the supplemental material look good.\n\nThe soft spots are real but not fatal. First, the relative tensor elements are reported without any uncertainties. That is the main weakness for a paper whose whole point is quantitative values. Second, the separation of the overlapping A_g/B_g pairs is model-dependent: it relies on the Kranert birefringence model and on the dielectric tensor taken from private communication. If that dielectric tensor is slightly off, the fit can trade intensity between overlapping modes and bias the tensor elements. The authors don't do an identifiability or sensitivity analysis, so we don't know how stiff these numbers are. Third, the sign of c is explicitly left unresolved, so 'complete' is a bit strong. Fourth, the fitting code is a non-public alpha, which limits reproducibility. None of these undermine the central claim that the procedure works — the agreement with prior partial data and with theory for other modes supports that — but they do mean the new B_g numbers should be treated as conditional until validated.\n\nWho gets value from this: Raman spectroscopists working on anisotropic or birefringent crystals, the β-Ga2O3 community, and anyone fitting overlapping phonon modes. It's a focused contribution, not a field-reshaping one.\n\nMy recommendation: send it to peer review. A serious referee should push for error bars, a sensitivity check on the dielectric tensor, and a clearer statement that the sign of c remains open. With those additions, the paper would be solid.","headline":"A solid methodological advance: single-stage hyperspectral fitting separates overlapping Raman modes in birefringent β-Ga2O3 and delivers first experimental tensor elements for B_g(3)/B_g(4), but the numbers need uncertainty quantification and a sensitivity check.","tokens_in":15212,"tokens_out":1048,"would_cite":true,"duration_ms":13623,"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 single hyperspectral fit separates all 15 Raman-active modes of β-Ga2O3, including previously overlapped pairs.","keywords":["Raman spectroscopy","polarization angle-resolved Raman","birefringence","β-Ga2O3","monoclinic crystal","Raman tensor","phonon modes","hyperspectral fitting"],"falsifier":"Generate simulated hyperspectral PARRS data for β-Ga2O3 with known tensor elements and the reported spectral resolution and noise, run the same single-stage fit, and check whether the B_g(3)/B_g(4) elements are recovered without bias. Alternatively, measure the three critical mode pairs on an additional crystal plane, such as the (110) plane, or at a second laser wavelength, where the birefringence corrections change; if the extracted tensor elements drift significantly, the result is model-dependent.","tokens_in":14285,"feed_emoji":"🔬","tokens_out":9037,"duration_ms":88497,"temperature":0.7,"pith_summary":"The paper aims to prove that quantitative Raman tensor extraction is feasible in strongly birefringent, low-symmetry crystals if the polarization-angle-resolved data are treated as one coherent block rather than as a series of independent spectra. The authors measure Raman scattering on four crystal planes of β-Ga2O3 while rotating the light polarization, and fit the full hyperspectral dataset in a single stage using a model that carries the birefringence corrections inside the angle-dependent selection rules. They report that this yields the energies and relative Raman tensor elements of all 15 Raman-active modes, including the three pairs—A_g(5)/B_g(3), A_g(7)/B_g(4), and B_g(5)/A_g(9)—whose members are too close in energy for the spectrometer to resolve individually. A reliable complete tensor set matters because β-Ga2O3 is a leading candidate for next-generation power electronics, and its monoclinic anisotropy has previously made quantitative Raman analysis seem impractical.","feed_headline":"All 15 β-Ga2O3 Raman modes fully separated in one fit","feed_subtitle":"Angle-resolved data and a birefringence-aware fit resolve previously hidden A_g/B_g mode pairs.","key_machinery":"The machinery is a single-stage hyperspectral fit function that models, for each scattering geometry, the Raman intensity as a sum over modes of a Lorentzian line shape multiplied by an angle-dependent selection-rule profile, plus a background. Each profile is computed from a birefringence-corrected effective Raman tensor built by transforming Raman and dielectric tensors into the crystal-surface coordinate system and applying Fresnel transmission, internal eigenpolarization mixing, collection-cone solid-angle, and dipole-scattering prefactor corrections. Because A_g and B_g profiles have different functional dependences on the polarization angle—sin^4/cos^4 combinations versus sin^2(2φ)/cos","core_discovery":"The central claim is that the full Raman response of monoclinic β-Ga2O3 can be inverted from polarization angle-resolved Raman spectra once birefringence is built into the model: all ten A_g and five B_g modes are individually resolved and assigned energies, linewidths, and relative Raman tensor elements in a self-consistent fit across the (100), (010), (001), and (2̄01) planes. The decisive advantage over the standard two-step workflow—fit each spectrum, then fit the extracted peak intensities—is that overlapping peaks beyond the spectral resolution are separable because each symmetry class is forced to follow a distinct, theoretically predicted angle profile. As a byproduct, the multi-plan","pith_inferences":["The strongest model dependence lies in the three unresolved pairs: if the dielectric tensor or the birefringence corrections are slightly off, the fit can trade intensity between the overlapping A_g and B_g peaks, so the newly reported B_g(3) and B_g(4) tensor values should be treated as model-dependent until independently validated.","A natural validation is a synthetic-data recovery test: simulate PARRS hyperspectral data with known tensor elements and the same noise level, run the single-stage fit, and measure the bias in the overlapped modes; this would quantify identifiability.","Repeating the measurements at a second excitation wavelength, where the dielectric tensor and birefringence change, would alter the angular profiles and provide an independent check that the separated modes are genuine rather than fitting artifacts.","The same framework could be extended to extract strain or temperature dependence of Raman tensors, since it already resolves overlapping peaks that per-spectrum fitting cannot."],"forward_implications":["The energies and relative Raman tensor elements of all 15 Raman-active phonon modes of β-Ga2O3 are now determined experimentally, including B_g(3) and B_g(4), which previously had no measured values.","The single-stage hyperspectral fitting strategy is offered as a route for other birefringent or low-symmetry crystals where peak overlap has prevented quantitative Raman tensor analysis.","For the four planes studied, the relative signs of the tensor elements in the (a,b,d) and (e,f) groups are fixed by the data rather than by convention; only the sign of c remains undetermined.","The resulting tensor set provides a complete benchmark for density-functional-theory calculations of the Raman response of β-Ga2O3.","The reported agreement with earlier experimental and theoretical values for isolated modes indicates that the method reproduces known results while correcting those affected by overlap."],"fun_headline_variants":["Birefringence-aware single fit separates all 15 β-Ga2O3 Raman modes","Complete Raman tensor of β-Ga2O3 from one angle-resolved fit","Full Raman tensor for β-Ga2O3 via birefringence-corrected single fit","All 15 β-Ga2O3 Raman modes resolved by single-stage analysis","One fit captures all 15 β-Ga2O3 Raman modes with birefringence"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The separation of the overlapped A_g/B_g mode pairs rests on the assumption that the polarization-angle profiles computed from the birefringence model—using dielectric-tensor values provided to the authors privately—are accurate enough that the fit can correctly divide intensity between two peaks the spectrometer cannot resolve on its own.","fun_headline_variants_meta":{"raw":{"variants":["Birefringence-aware single fit separates all 15 β-Ga2O3 Raman modes","Complete Raman tensor of β-Ga2O3 from one angle-resolved fit","Full Raman tensor for β-Ga2O3 via birefringence-corrected single fit","All 15 β-Ga2O3 Raman modes resolved by single-stage analysis","One fit captures all 15 β-Ga2O3 Raman modes with birefringence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000816,"raw_usage":{"total_tokens":3399,"prompt_tokens":722,"completion_tokens":2677,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":466,"completion_tokens_details":{"reasoning_tokens":2558}},"tokens_in":466,"tokens_out":2677,"duration_ms":19584,"temperature":1.0,"reasoning_tokens":2558,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T08:36:35.342653+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Generate simulated hyperspectral PARRS data for β-Ga2O3 with known tensor elements and the reported spectral resolution and noise, run the same single-stage fit, and check whether the B_g(3)/B_g(4) elements are recovered without bias. Alternatively, measure the three critical mode pairs on an additional crystal plane, such as the (110) plane, or at a second laser wavelength, where the birefringence corrections change; if the extracted tensor elements drift significantly, the result is model-dependent.","supporting_citations":[],"review_version":1}