{"id":"b88e7f99-beb4-4f27-a145-6f5854453288","arxiv_id":"2506.03539","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"New germanium Compton scattering data at 180 eV/c to 25 keV/c momentum transfer favor the Geant4 Livermore model, reject the Monash model at up to 7.7 sigma, and support an MCDF-based scattering function that raises predicted dark matter backgrounds by 10-50%.","lead":"This paper measures how gamma rays scatter off germanium atoms at very small angles and compares the data to Monte Carlo models used in dark matter detectors. The measurements favor one model, rule out another, and suggest that a new atomic calculation changes the expected low-energy background for sub-GeV dark matter searches.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The measured scattering-function points do not directly validate the MCDF-RIA versus HF-RIA difference in the momentum-transfer region driving the 10–50% background claim: the decisive low-angle points depend on a simulated 'soft wall' angle shift (1.5° to 2.02°) that is not independently validated.","rationale":"The paper contains two distinct claims: (1) Geant4 Livermore and Penelope Compton models are consistent with the measured DDCS while Monash is rejected, and (2) the MCDF-RIA scattering function predicts a Compton background 10–50% higher than HF-RIA at low gamma energies. Claim (1) is well supported by the chi-square statistics in Table IV and by the direct spectral comparisons, and I do not see a model-independent flaw there. Claim (2), which is the DM-relevant headline, is the less secure part. The experimental SF extraction is normalized to theory at 12°, so the comparison is only about relative shape; the lowest-X points are displaced by a simulated soft-wall correction of roughly 0.5°, much larger than the stated angular precision. No independent validation of that simulated angle distribution is presented, and the measured points stop at X = 0.94 Å^-1, which is the same region relevant to the 239-keV sub-keV background enhancement. If the effective-angle model is incorrect, the apparent preference for MCDF-RIA could disappear. This is essentially the reader's weakest assumption, so the reader's CONDITIONAL verdict is appropriate. A further minor issue is that the abstract's 2.3% precision figure is never defined or derived in the body; the authors should state what it refers to, alongside providing the effective-angle validation and the data/code underlying the analysis.","tokens_in":19068,"tokens_out":9902,"duration_ms":120170,"concrete_test":"Reconstruct the per-event scattering angle from the coincident HPGe recoil energy and NaI scattered-photon energy using Compton kinematics, θ = arccos[1 − m_e c^2 (1/E_γ' − 1/E_γ)], and compare the resulting angular distribution at nominal 1.5° with the Geant4 distribution that produces the effective angle 2.02°. If the reconstructed distribution has mean consistent with 2.02° within about 0.05°, the soft-wall correction is confirmed and the low-angle SF points stand; if it is consistent with the nominal 1.5°, the SF comparison must be redone and the apparent MCDF-RIA preference may vanish.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The comparison that supports MCDF-RIA over HF-RIA is not a direct measurement of the scattering function in the region where the two theories differ; it is a relative shape extracted using simulations that contain the model under test. In Eq. (8), S_exp is defined as (dσ/dΩ_exp / dσ/dΩ_sim) × S_sim, and the 12° point is forced onto the theoretical curve by the normalization described in Section V A. Thus only the shape from 12° down to the low-angle points carries information. For those decisive low-angle points, Table IV gives effective angles of 2.02° and 2.34° for nominal angles of 1.5° and 2°, i.e., shifts of 0.52° and 0.34°. These are far larger than the quoted 0.03° angular precision, and they come entirely from the Geant4 model of source dispersion, collimator geometry, and the sub-keV PSD efficiency. If the simulated angle distribution is wrong at the 0.3–0.5° level, the X-coordinate of the 1.5° point moves by an amount comparable to the separation between the MCDF-RIA and HF-RIA curves, changing the 'favors MCDF-RIA' conclusion. Moreover, no measured point lies below X = 0.94 Å^-1, while the 10–50% background enhancement at 239 keV is dominated by sub-keV recoils at X values at or below the lowest measured point. The experimental SF data therefore do not directly test the background claim that motivates the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a coincidence measurement of Compton doubly differential cross sections (DDCS) from a 10-g HPGe detector irradiated by a collimated 662 keV 137Cs source, at nominal scattering angles of 1.5°, 2°, 3°, 4°, 5°, and 12°, with scattered photons detected by a NaI[Tl] detector. The measured spectra are compared with Geant4 implementations of three low-energy Compton models (Livermore, Monash, Penelope), and the Monash model is reported to be rejected at 4.4σ–7.7σ while Livermore/Penelope are consistent with data. Using the ratio of measured to simulated spectra, the authors extract experimental scattering functions S(X) and compare them with two theoretical scattering functions: the HF-RIA one adopted in Geant4 and their own MCDF-RIA calculation, concluding that the data favor MCDF-RIA at low momentum transfer. The paper then uses Geant4 simulations to estimate the impact of the MCDF-RIA versus HF-RIA scattering functions, detector mass, and source position on Compton backgrounds relevant to electronic-recoil dark matter searches, and presents a combined background analysis for several dark matter models. The abstract additionally claims a \"2.3% measurement\" of atomic Compton scattering in the momentum-transfer range 180 eV/c to 25 keV/c, but no 2.3% figure appears in the body of the manuscript.","tokens_in":19390,"tokens_out":3279,"duration_ms":36523,"significance":"If the central claims hold, the paper would provide a useful experimental test of low-energy Compton models in Geant4, a new ab initio MCDF-RIA calculation of the germanium scattering function, and a concrete quantification of how the choice of scattering function affects sub-keV dark matter backgrounds. The DDCS comparison that rejects the Monash model is substantially independent of the scattering-function normalization and is valuable for validating Monte Carlo transport codes in the low-energy regime. The machine-readable style of the analysis, with explicit tables of efficiencies, systematic errors, and chi-square statistics, is a strength. However, the scattering-function extraction is not a direct measurement of S(X) in the region where MCDF-RIA and HF-RIA differ most, and the headline precision claim is not substantiated in the text; these issues materially weaken the parts of the paper that motivate the dark matter implications.","major_comments":[{"comment":"The abstract claims a \"2.3% measurement of atomic Compton scattering\", but no 2.3% uncertainty appears anywhere in Sections IV E, V, or the tables. Table III lists total systematic errors of 0.46%–0.77%, and the normalization procedure in Section V A fixes the overall scale at 12 degrees by construction. The abstract should either be revised to quote the actual systematic errors (which are the ones quantified in the paper) or a derivation of the 2.3% number must be supplied; as written, the headline claim is unsupported and internally inconsistent with the body text.","section":"Abstract and Section V"},{"comment":"The measured scattering function is defined as S_exp = (dσ/dΩ_exp / dσ/dΩ_sim) × S_sim, and the 12 degree point is pinned to the theoretical scattering function by the normalization described in Section V A. The only information that can discriminate MCDF-RIA from HF-RIA therefore comes from the low-angle points, whose effective angles are 2.02° and 2.34° for nominal 1.5° and 2°, i.e., shifts of 0.52° and 0.34° relative to the nominal angles. These shifts far exceed the 0.03° angular calibration precision and are derived entirely from the Geant4 simulation of source dispersion, geometry, and the sub-keV PSD efficiency. If that simulated angle distribution is wrong at the 0.3–0.5° level, the X-coordinates of the decisive points shift by an amount comparable to the separation between the MCDF-RIA and HF-RIA curves, changing the conclusion that the data favor MCDF-RIA. The paper should quantify this sensitivity explicitly, for example by varying the simulated source divergence or PSD efficiency within their uncertainties and recomputing the effective angles and the resulting χ² values.","section":"Section V B, Eq. (8), Table IV"},{"comment":"The 10–50% background enhancement at 239 keV is stated to arise in the sub-keV region, which corresponds to momentum transfers at or below the lowest measured point in Fig. 10 (X ≈ 0.94 Å^-1 at the effective angle of 2.02°). No measured scattering-function point lies in this region, so the experimental data do not directly test the MCDF-RIA versus HF-RIA difference that drives the background claim. The conclusion in Section VI A rests on the MCDF-RIA calculation itself plus the fragile low-angle extraction discussed in the previous comment. The manuscript should explicitly state that the background comparison is a prediction from the ab initio calculation, not a tested consequence of the measured scattering function, and should provide a quantitative estimate of how the systematic uncertainty in the effective-angle shifts propagates to the 10–50% claim.","section":"Section VI A, Fig. 10, Fig. 11"}],"minor_comments":[{"comment":"There are frequent typographical and grammatical errors (e.g., \"defination\", \"discrypancy\", \"bule\", \"corsses\", \"diviations\", \"choosen\", \"background\" for background) that should be corrected in a thorough language edit.","section":"Throughout"},{"comment":"Reference [43] reads \"And still have three models in version (11.03)\" and is not a proper bibliographic entry; it should be replaced with a complete citation or removed.","section":"Reference list"},{"comment":"The table columns labeled \"(a) (b) (c) (d)\" and \"(a) (b) (c)\" are not self-explanatory; the captions should explicitly map each column letter to the corresponding systematic-error item described in the text.","section":"Section IV E, Table III"},{"comment":"The sentence \"The equivalent significance of the discrepancy between the data and Monash model ranges from 5.49σ to 7.67σ\" is immediately followed by an exception for the 1.5° point (4.4σ); the text should be rephrased to present the 4.4σ value together with the range, since the current wording creates an apparent contradiction.","section":"Section V A"},{"comment":"The terms \"effective scattering angle\" and \"soft wall\" are introduced in Section V B without a formal definition at first use; a short definition or a reference to Table IV would help the reader follow the subsequent argument.","section":"Section III B and Section V B"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a valuable experimental dataset and a clear model-discrimination result for Geant4 Compton models, which is likely publishable after revision. The main concern is that the abstract's precision claim is not present in the body, and the scattering-function/background conclusions rely on simulation-derived effective-angle shifts that are not adequately validated. I would ask the authors to either strengthen the validation (e.g., by comparing the simulated effective-angle distribution against an independent experimental check) or soften the claims accordingly. The 10–50% dark matter background statement should be clearly labeled as a theoretical prediction rather than a measurement. No concerns about citation patterns beyond the malformed reference [43]."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, it is a genuinely careful measurement: six scattering angles from 1.5 to 12 degrees, with meticulous attention to angle calibration, efficiency corrections, and background subtraction. The DDCS comparison is the strong part. The data reject the Monash model at 4.4–7.7 sigma across all angles, and the Livermore/Penelope models are consistent. That conclusion is robust because it compares full spectral shapes, not a normalized point, and it does not depend on the scattering-function extraction. Second, the more ambitious claims—the new MCDF-RIA scattering function and the 10–50% higher Compton background for sub-GeV dark matter searches—are softer than the abstract suggests. The 2.3% measurement precision quoted in the abstract never appears in the body text; what is actually quantified are systematic errors of 0.5–0.8% per angle plus a normalization uncertainty. The scattering function measurement is defined through Eq. (8) as the experimental spectrum divided by the simulated spectrum times the simulated S(X), and it is pinned to theory at 12 degrees. So only the shape from 12 degrees down to 1.5 degrees carries information, and even that shape depends on the simulated effective angle distribution—the \"soft wall\" that shifts 1.5 degrees to 2.02 degrees. That shift is larger than the quoted 0.03-degree angular precision and is not independently validated. The stress-test note is right: the measured points only go down to X = 0.94 Å⁻¹, while the dark-matter background claim is driven by sub-keV events at lower X. The experiment therefore does not directly test the MCDF-RIA versus HF-RIA difference in the region that matters for the background enhancement. That does not make the background calculation wrong, but it means it is an extrapolation, not a measurement. The paper would benefit from an explicit statement of the measured X-range limitation and a sensitivity study of the normalization choice. Who is this for? Dark matter experimentalists using germanium detectors, and Geant4 developers maintaining low-energy Compton models. The Livermore-versus-Monash result alone is worth refereeing. Yes, send it to peer review, but ask for the 2.3% claim to be substantiated or removed, for the scattering-function extraction to be shown with and without the 12-degree pinning, and for the data and simulation configuration to be released. I would cite the DDCS result, but not the background-enhancement figure without checking the normalization sensitivity first.","headline":"A careful Compton-scattering measurement that convincingly kills the Geant4 Monash model, but the scattering-function and dark-matter background claims rest on simulation-dependent normalization and extrapolation beyond the measured range.","tokens_in":20010,"tokens_out":1368,"would_cite":true,"duration_ms":16932,"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 2.3% measurement of atomic Compton scattering in germanium distinguishes between Geant4's Compton models and raises the predicted low-energy dark-matter background by 10–50%.","keywords":["Compton scattering","doubly differential cross section","incoherent scattering function","relativistic impulse approximation","multi-configuration Dirac-Fock","germanium detector","dark matter background","Geant4 simulation"],"falsifier":"Repeat the small-angle measurement with a well-collimated source and a lower detector threshold so the soft-wall smearing is reduced; if the extracted scattering-function values move toward the HF-RIA curve rather than the MCDF-RIA curve as the soft-wall is removed, the claimed preference would be directly contradicted. Alternatively, a band-structure calculation of germanium's valence Compton profile could be compared with the sub-keV region of the measured spectra to test whether the mild overestimation seen for the Livermore model comes from the isolated-atom assumption.","tokens_in":18879,"feed_emoji":"⚛️","tokens_out":6173,"duration_ms":66259,"temperature":0.7,"pith_summary":"The paper measures the doubly differential cross section of atomic Compton scattering in germanium at six small angles, reaching momentum transfers as low as 180 eV/c, and uses the data to judge which low-energy Compton model built into the Geant4 simulation code is correct. The measurement, with a total systematic error around 0.5–0.8%, is accurate enough to reject the Monash model at 4.4 to 7.7 sigma while keeping the Livermore and Penelope models. The same data, converted into an experimental incoherent scattering function, favor a fully relativistic ab initio multi-configuration Dirac-Fock treatment over the older Hartree-Fock function. Why this matters: the Compton background in germanium dark-matter detectors is predicted to be 10–50% higher at low gamma energies when the MCDF scattering function is used, which changes how sub-GeV dark-matter limits are drawn.","feed_headline":"Ge Compton data reject Monash model at up to 7.7σ","feed_subtitle":"Small-angle data also favor a relativistic ab initio scattering function, lifting predicted dark-matter backgrounds 10–50%.","key_machinery":"The central object is the incoherent scattering function $S(X)$ in the relativistic impulse approximation, which factorizes the Compton doubly differential cross section into a free-electron term times $S(X)$. The paper computes $S(X)$ from first principles using multi-configuration Dirac-Fock wavefunctions (MCDF-RIA) and compares it with the Hartree-Fock-based function (HF-RIA) used by Geant4. A second load-bearing mechanism is the soft-wall effect: the experimental scattering angle is not a delta function but a simulated distribution, so the effective scattering angle exceeds the nominal one (1.5 degrees maps to 2.02 degrees), which limits how low in momentum transfer the measurement can reach.","core_discovery":"The paper establishes that the Livermore and Penelope Compton models in Geant4 reproduce the measured energy spectra of 662 keV photons Compton-scattered in a 10 g germanium crystal at angles from 1.5 to 12 degrees, whereas the Monash model fails decisively, with discrepancies equivalent to 4.4–7.7 sigma. It also establishes, from the same spectra, an experimental incoherent scattering function that at low momentum transfer tracks the MCDF-RIA calculation more closely than the HF-RIA function adopted in Geant4. Because the scattering function multiplies the free-electron cross section, this changes the predicted low-energy Compton background in germanium: the MCDF-RIA function gives steps in the spectrum about 10–50% higher than HF-RIA at 239 keV, with the largest difference below the L-shell ionization edge.","pith_inferences":["The paper's isolated-atom caveat suggests that a band-structure treatment of germanium's valence electrons could reduce the mild overestimation seen for the Livermore model below 500 eV; if so, the sub-keV Compton background in dark-matter detectors would be slightly lower than the MCDF-RIA isolated-atom prediction.","The same measurement strategy can be extended to other detector materials: applying MCDF-RIA calculations to silicon or germanium at different photon energies would give a direct, model-independent way to calibrate low-energy Compton backgrounds before designing next-generation ionization detectors.","Because the MCDF-RIA difference grows as gamma energy decreases, the largest background-correction impact will be for low-energy environmental lines around 200–350 keV, which are common in underground laboratories; this could matter for experiments searching for dark matter with masses near the sub-GeV range."],"forward_implications":["The Livermore and Penelope models can be used with confidence for angle-dependent Compton simulations in germanium, while the Monash model should not be used for such small-angle work.","Compton backgrounds in germanium detectors computed with the MCDF-RIA scattering function are 10–50% higher than the default Geant4 values at low gamma energies such as 239 keV, with the difference concentrated below the L-shell ionization edge.","Detector mass strongly affects the shape of the Compton background spectrum, with small detectors showing non-flat structures from electron escape and larger detectors approaching a flatter spectrum, independent of gamma source position.","For electronic-recoil dark-matter searches, the step-like structures in the expected ALP, dark-photon, and Migdal signals remain distinguishable from the step-like Compton and photon-coherent-scattering background, despite sharing the same sub-shell ionization energies.","The soft-wall effect sets a practical floor on how low in momentum transfer the present generation of small-angle scattering experiments can reach with a 662 keV source and a 10 g germanium detector."],"supporting_citations":[{"why":"Supplies the Hartree-Fock scattering function and binding-energy database that Geant4's default Compton background uses and against which MCDF-RIA is compared.","marker":"[33]"},{"why":"Defines the Monash Compton model, the one the measured spectra reject at 4.4–7.7 sigma.","marker":"[23]"},{"why":"Defines the Livermore Compton model whose low-energy behavior the measurements validate.","marker":"[41]"},{"why":"Defines the Penelope model, found consistent with the Livermore model and with the data.","marker":"[42]"},{"why":"The simulation code whose three low-energy Compton models and scattering-function implementation are being tested.","marker":"[27]"},{"why":"Earlier measurement of the same type that showed inconsistency with Livermore; the present work refines data processing and background removal to resolve it.","marker":"[31]"},{"why":"Introduces the multi-configuration Dirac-Fock method used for the ab initio wavefunctions and scattering function.","marker":"[36]"},{"why":"Supplies Hartree-Fock Compton profiles used as the comparison for the MCDF relativistic Compton-profile calculation.","marker":"[40]"}],"fun_headline_variants":["Ge Compton data nix Monash model at up to 7.7σ","Ge Compton scattering favors MCDF-RIA over HF-RIA","Monash Compton model fails 4.4–7.7σ in germanium","Ge data: Monash model off by up to 7.7σ","MCDF-RIA Compton background 10–50% higher in Ge"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extraction of the scattering function assumes that the simulated effective-angle distribution, including the soft-wall shift, correctly models the source dispersion and geometry; if that simulation is wrong, the X-coordinates of the measured scattering-function points shift and the apparent preference for MCDF-RIA over HF-RIA could change.","fun_headline_variants_meta":{"raw":{"variants":["Ge Compton data nix Monash model at up to 7.7σ","Ge Compton scattering favors MCDF-RIA over HF-RIA","Monash Compton model fails 4.4–7.7σ in germanium","Ge data: Monash model off by up to 7.7σ","MCDF-RIA Compton background 10–50% higher in Ge"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000552,"raw_usage":{"total_tokens":2664,"prompt_tokens":1006,"completion_tokens":1658,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":622,"completion_tokens_details":{"reasoning_tokens":1558}},"tokens_in":622,"tokens_out":1658,"duration_ms":12795,"temperature":1.0,"reasoning_tokens":1558,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:00:36.930269+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the small-angle measurement with a well-collimated source and a lower detector threshold so the soft-wall smearing is reduced; if the extracted scattering-function values move toward the HF-RIA curve rather than the MCDF-RIA curve as the soft-wall is removed, the claimed preference would be directly contradicted. Alternatively, a band-structure calculation of germanium's valence Compton profile could be compared with the sub-keV region of the measured spectra to test whether the mild overestimation seen for the Livermore model comes from the isolated-atom assumption.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hartree-Fock scattering function and binding-energy database that Geant4's default Compton background uses and against which MCDF-RIA is compared."},{"cited_title":"Brown, M","cited_arxiv_id":null,"evidence_quote":"Defines the Monash Compton model, the one the measured spectra reject at 4.4–7.7 sigma."},{"cited_title":"Apostolakis, S","cited_arxiv_id":null,"evidence_quote":"Defines the Livermore Compton model whose low-energy behavior the measurements validate."},{"cited_title":"Salvat Gavald` a, J","cited_arxiv_id":null,"evidence_quote":"Defines the Penelope model, found consistent with the Livermore model and with the data."},{"cited_title":"Agostinelli, J","cited_arxiv_id":null,"evidence_quote":"The simulation code whose three low-energy Compton models and scattering-function implementation are being tested."},{"cited_title":"Jia, S.-T","cited_arxiv_id":null,"evidence_quote":"Earlier measurement of the same type that showed inconsistency with Livermore; the present work refines data processing and background removal to resolve it."},{"cited_title":"Desclaux, Computer Physics Communications9, 31 (1975)","cited_arxiv_id":null,"evidence_quote":"Introduces the multi-configuration Dirac-Fock method used for the ab initio wavefunctions and scattering function."},{"cited_title":"Biggs, L","cited_arxiv_id":null,"evidence_quote":"Supplies Hartree-Fock Compton profiles used as the comparison for the MCDF relativistic Compton-profile calculation."}],"review_version":1}