{"id":"795f6fb9-c1bd-49eb-956c-aaf7599a7ff6","arxiv_id":"2501.18346","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"First application of XMM-Newton soft proton response matrices to real flare spectra shows a power-law input spectrum plus a soft excess, and reveals 20-year degradation of MOS cameras relative to PN.","lead":"This paper tests newly built computer models of how XMM-Newton's mirrors focus low-energy protons onto its detectors, using 55 real flare observations. It finds the proton spectrum is a power law with a small extra soft component that likely comes from imperfections in the models, and quantifies how the two MOS cameras have degraded relative to the PN camera over 20 years.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 2–5 keV soft excess is attributed to proton-matrix error without an independent calibration of that band; if the Geant4 matrices are wrong in a different way, the 'pure power law' input claim fails. A thin-vs-medium filter comparison can test this.","rationale":"Reader's CONDITIONAL verdict already captures the main risk; this stress test agrees. The concern is not that the paper is wrong but that its headline inference is underdetermined: a residual is assigned to matrix error on the basis of a Paper I simulation trend rather than an in-band calibration or a real-data test of that trend. The proposed filter split is a direct, low-cost check using the same public data and matrices, and it would discriminate between a matrix artifact and a genuine low-energy component. The paper's own caveats ('cannot be entirely excluded'; laboratory measurements required) support keeping the verdict CONDITIONAL rather than upgrading it. No independent verification beyond public data is required.","tokens_in":19663,"tokens_out":6211,"duration_ms":64599,"concrete_test":"Using the public flare spectra and the proton response matrices, split the MOS1 and MOS2 spectra by optical filter (thin vs medium) within each epoch, matching on hardness ratio or count rate where possible. For each spectrum, fit the 5–11.5 keV power law, freeze it, and measure the 2–5 keV excess fraction (equivalently, the fitted black-body normalization/flux relative to the power-law flux). Paper I predicts that the excess fraction grows with passive material, so medium-filter spectra should show a systematically larger excess than thin-filter spectra; if the excess is a real spectral component and the matrices are correct, the response-corrected excess should be filter-independent. If no filter trend is found, the matrix-artifact interpretation is not supported and the input-power-law claim must be softened; if the predicted trend is present, the interpretation is strengthened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that the proton input spectrum is a power law, with the 2–5 keV excess being a response-matrix artifact. This is an inference from residuals: the 5–11.5 keV power law is extrapolated downward, the excess is fitted by a phenomenological black body, and the excess is then attributed to matrix imprecision. The load-bearing assumption is that the Geant4 proton matrices are correct in the 2–5 keV band, exactly where they are least validated: the matrices are a first release, previously checked only against averaged EXTraS background spectra, and the authors state in Section 6 that a physical origin such as heavier ions 'cannot be entirely excluded' and that laboratory measurements are still required. If the true matrix error differs from the simulated one, the black-body component could be a real soft proton or ion component, and the input would not be a pure power law. The paper cites a Paper I simulation trend that the excess grows with passive material as support, but does not verify that trend on the real thin- versus medium-filter MOS observations analyzed here.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the first application of the Geant4-based soft-proton response matrices of Paper I to real XMM-Newton flare spectra, using 55 simultaneous MOS1/MOS2/PN observations in four epochs spanning solar maximum and minimum. The authors find no epoch-dependent variation in mean flare rates, report that a single power law fits the 5-11.5 keV band but leaves a low-energy excess in 2-5 keV that they model phenomenologically with a black body, and attribute this excess to imprecision in the proton response matrices at softer energies. They also derive inter-detector systematic uncertainties in spectral indices and input fluxes and identify a time-dependent degradation of MOS proton response relative to PN.","tokens_in":19835,"tokens_out":6255,"duration_ms":64530,"significance":"If the conclusions hold, this is a valuable first real-data validation of proton response matrices for XMM-Newton, with direct implications for background modeling and for future missions such as Athena. The matrices themselves are derived from independent Geant4 simulations rather than from the fitted data, so the comparison is a genuine external check rather than circular. The multi-instrument comparison and the use of four epochs provide a useful empirical handle on systematic uncertainties. The main result, however, rests on the interpretation of the low-energy excess as a matrix artifact; that interpretation is plausible but not demonstrated, and the paper itself concedes that a physical origin from heavier ions cannot be entirely excluded.","major_comments":[{"comment":"The central attribution of the 2-5 keV soft excess to matrix imprecision is not demonstrated. The evidence is an extrapolation of the 5-11.5 keV power law plus a phenomenological black body, and the paper explicitly states that 'a more physical origin—that is, a component coming from heavier ions—cannot be entirely excluded.' The 2-5 keV band is exactly the range where the Geant4 matrices are least independently validated. The Paper I trend with passive material and the lower PN excess are suggestive but do not establish the artifact interpretation. A concrete test is available in the data: compare the amplitude of the soft excess relative to the power law for thin-filter versus medium-filter observations of the same instrument and epoch. This comparison is not performed. The authors should either perform it or reframe the abstract and conclusions to present the matrix-artifact interpretation as a hypothesis rather than the main result.","section":"Section 6 (and abstract)"},{"comment":"Some PN spectra are excluded from the subsequent analysis after the power-law extrapolation overestimates the 2-5 keV rates. The text says 'we excluded them from the following analysis' for observations 0052140201, 0108061901, 0827241201, 0844210101, 0852190101, 0854590401, and 0862400101. This is a post-hoc selection on the dependent variable: removing cases with negative or zero soft excess biases the sample toward positive excesses and can inflate the reported 21% (MOS) and 5% (PN) excess fractions and the perceived need for a black-body component. The authors should quantify how many spectra are excluded and show that the conclusions are robust when all spectra are included, or adopt a model that can accommodate flattening below 5 keV.","section":"Section 4.1"},{"comment":"The two-step fitting procedure fixes the power-law parameters from the 5-11.5 keV fit before adding the black-body component. As the paper states, allowing all parameters to float in the full 2-11.5 keV range changes the results, and the fixed-parameter approach was adopted to avoid degeneracies. This makes the model comparison unequal: the power-law-plus-black-body model is not compared with alternative models on the same footing, and the reported uncertainties on kT and the excess rates do not propagate the covariance with the fixed power-law parameters. A simultaneous fit with weak priors, or at least a profile-likelihood treatment, is needed to support the claim that the black body is the only model that gives coherent results across detectors.","section":"Section 4.2"},{"comment":"The broken power-law test is not a strong test of spectral steepening within the fitted band. For MOS, the best-fit break energies are 22-27 keV, above the fitted 2-11.5 keV range, so within the analysis band the model is effectively a single power law; the indices below the break are explicitly unconstrained. The statement that a broken power law 'cannot be considered physically reliable' and that there is no evidence for steepening is therefore weaker than presented. The log-parabola model also gives incoherent MOS/PN parameters, but the conclusion that the input spectrum is a pure power law should be based primarily on the simultaneous fits and the missing filter comparison, not on the broken-power-law result.","section":"Section 4.3.1"}],"minor_comments":[{"comment":"There are typographical errors and axis-label inconsistencies: Figure 3 uses 'Dic.' instead of 'Dec.', Section 3 has 'corespondent' for 'corresponding', and Section 6 has 'have have formed'. These should be corrected.","section":"Multiple sections"},{"comment":"The statement that there are no statistically significant variations in the mean rates across epochs is a null result with limited power, given the large rms values and the small number of epochs. It would be more precise to say that no variation is detected rather than that there is no effect.","section":"Section 3 / Table 3"},{"comment":"The NHP threshold of 2.7e-3 is applied per spectrum without multiple-testing correction. With 165 spectra the expected number of false rejections under the null is small (about 0.4), so this is not a serious concern, but the paper should state explicitly that no multiple-testing correction was applied.","section":"Section 4.1"},{"comment":"The caption says the flux ratio is plotted 'as a function of the PN rates in the same band', but the text describes ratios of MOS1/MOS2, MOS1/PN, and MOS2/PN fluxes as functions of MOS2 or PN rates. The caption and axis labels should be aligned with the text for all three panels.","section":"Figure 10 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is a useful first step and the Geant4 matrices are an external input, so the core comparison is not circular. However, the abstract's central claim that the soft excess is a matrix artifact goes beyond what the analysis demonstrates; the authors themselves include the heavier-ion caveat. If the filter-thickness comparison is not possible with the current sample, the manuscript should be revised to present the power-law-plus-artifact interpretation as a preferred hypothesis with quantified systematic uncertainty, rather than as the definitive result."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main thing you should know: this is the first paper to actually fit real XMM-Newton flare spectra with the proton response matrices from Paper I, and it produces two kinds of genuinely useful results. First, the inter-calibration numbers (3% MOS-MOS, ~24% MOS-PN, factor of two in input flux) are what people will quote when using these matrices. Second, the 20-year degradation story for MOS1/MOS2 relative to PN is a real finding, and the rate-ratio analysis that leads to it is careful. The paper also deserves credit for being honest: the black body is labeled phenomenological, and the text explicitly admits a heavier-ion origin cannot be excluded.\n\nWhere it gets soft is the central claim that the proton input is a pure power law. That claim depends on the 2-5 keV excess being a response-matrix artifact. The evidence is circumstantial: the excess is a residual after extrapolating the 5-11.5 keV power law, the black body is the only simple component that makes all three detectors agree, and Paper I reportedly shows a trend with passive material. But the matrices are least validated in exactly that 2-5 keV band, and the thin- vs medium-filter MOS observations in this very sample could have tested the passive-material trend directly. The paper doesn't do that. As the stress-test note says, if the simulated matrix error is wrong in a different way, the black body could be a real soft component and the input spectrum would not be a pure power law. The two-step fitting procedure (fix the power law, then add the black body) and the post-hoc exclusion of some PN spectra add to the sense that the conclusion is somewhat engineered.\n\nNone of this is fatal. The data work is solid, the conclusions are mostly hedged in the text, and the abstract is only mildly overassertive. The paper is genuinely useful for XMM background analysts, instrument teams, and anyone building proton response for Athena.\n\nWho should read it: people who use XMM-Newton EPIC background or care about soft proton contamination. It deserves a serious referee. My recommendation: send it to review, and ask the authors to either do the thin/medium filter comparison within the same epoch or soften the abstract's wording about the power law being the physical model. If they can't strengthen the evidence, the claim should be stated as what it is: a model that works well, with an unresolved low-energy component.","headline":"First real-data test of the XMM soft-proton response matrices; the systematics are useful, but the pure power-law claim rests on an interpretation of the soft excess that the paper doesn't fully nail down.","tokens_in":20446,"tokens_out":1512,"would_cite":true,"duration_ms":16937,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that the input spectrum of XMM-Newton soft proton flares is a power law, and that the soft excess below 5 keV is an artifact of the proton response matrices rather than a real proton component.","keywords":["soft proton flares","XMM-Newton","proton response matrices","power-law spectrum","spectral deconvolution","MOS and PN detectors","magnetospheric protons","solar activity"],"falsifier":"A laboratory measurement of proton transmission through the actual MOS optical filters and electrode structures, compared with the simulated response, would settle whether the 2–5 keV excess is a matrix artifact; if the simulated transmission is correct, the black-body component should disappear once the matrices are corrected, while a residual excess would support a real low-energy component such as heavier ions.","tokens_in":19442,"feed_emoji":"🛰️","tokens_out":7905,"duration_ms":69827,"temperature":0.7,"pith_summary":"XMM-Newton's mirrors focus not only X-rays but also low-energy protons, which arrive as sudden soft proton flares that contaminate observations. This paper tests, for the first time on real data, the response matrices built to convert those proton detections back into the incoming proton spectrum. The central claim is that the input proton spectrum is a power law; the extra soft component needed to fit the 2–5 keV band is a symptom of imperfect matrices, not a real proton population. The analysis of 55 simultaneous MOS and PN flare spectra also quantifies how much the two detector types disagree, giving systematic uncertainties of about 3% between the two MOS cameras and about 24% between MOS and PN. If right, the result clears the way for using flare spectra to monitor the Earth's magnetospheric proton environment across the mission's 20-year lifetime.","feed_headline":"XMM proton flares are a pure power law at the source","feed_subtitle":"First real-data test of the proton response matrices shows the low-energy excess is a simulation artifact.","key_machinery":"The central object is the proton response matrix: a Monte Carlo–generated transfer function that maps an incoming proton energy to the charge deposited in the MOS or PN CCD, folded into the ancillary response and redistribution files used by standard spectral fitting software. It is what lets the authors turn a measured flare spectrum into an inferred input spectrum. The second ingredient is a phenomenological black-body component added below 5 keV to absorb the discrepancy between the extrapolated power law and the data; the paper's argument is that this component tracks matrix error rather than real emission, because it is strongest where passive material in front of the detectors is largest.","core_discovery":"Using 55 simultaneous MOS and PN flare observations taken at solar maximum (2001–2002) and solar minimum (2019–2020), the authors deconvolved background-subtracted spectra in the 2–11.5 keV band with the proton response matrices. A single power law fails for 72% of the spectra over the full band, but fits 96% of them in the 5–11.5 keV range; extrapolating that power law downward leaves an excess that is 21% of the MOS count rate and 5% of the PN count rate in the 2–5 keV band. Adding a phenomenological black-body component with a temperature near 1 keV recovers acceptable fits for 83% of the spectra, and the authors argue this component is an artifact of the matrices: the excess is larger for the front-illuminated MOS, which have electrode structure in front of the detector, than for the back-illuminated PN, and its size grows with passive material in the filters. They conclude that the physical input spectrum of soft proton flares is a power law, that the soft excess reflects imprecise modeling of proton transmission at the focal plane, and that the remaining cross-instrument differences set the systematic errors: about 3% on spectral indices between the two MOS cameras, about 24% between MOS and PN, and a factor of about two on the inferred input flux.","pith_inferences":["If the soft excess is a matrix artifact, previous XMM background studies that treated a soft component as real may need to revisit their flare-contamination modeling.","The same deconvolution method, applied to the full archive of flares once MOS degradation is modeled, could produce a 20-year map of the magnetospheric proton environment at different orbital distances.","A direct laboratory measurement of proton transmission through the filters, which the paper calls for, would either confirm the artifact explanation or revive the heavier-ion alternative.","For future grazing-incidence X-ray missions, the lesson is that proton response matrices must be validated on real flares before being used to predict focal-plane proton backgrounds."],"forward_implications":["The soft proton input spectrum can be treated as a power law, so flare observations can be converted into physical proton fluxes at the telescope entrance.","The 2–5 keV excess should be absorbed as a systematic (21% for MOS, 5% for PN) rather than modeled as a separate proton population.","No seasonal or solar-cycle dependence of mean flare rates is found across the four epochs.","MOS proton response has degraded over 20 years (about 30% for MOS1 on top of CCD loss, and more for MOS2), so flare-rate comparisons need epoch-dependent corrections.","Cross-instrument systematic uncertainties are quantified: about 3% on spectral indices between MOS1 and MOS2, about 24% between MOS and PN, and a factor of about two on input flux."],"supporting_citations":[{"why":"Builds the proton response matrices for MOS and PN from Monte Carlo simulations; these matrices are the objects being tested on real data.","marker":"Fioretti et al. 2024 (Paper I)"},{"why":"Derives power-law proton environment models above 50 keV that the paper adopts as the expected input spectrum.","marker":"Lotti et al. 2018"},{"why":"Predicts power-law distributions for different magnetosphere regions, supporting the power-law input model.","marker":"Fioretti et al. 2018"},{"why":"Documents the contamination layers on the MOS CCDs used to explain the 20-year degradation in proton response.","marker":"Plucinsky et al. 2017"},{"why":"Describes the RGS grating shading, the basis for halving the MOS effective area in the proton matrices.","marker":"den Herder et al. 2004"},{"why":"Provides simulations of soft proton focusing by XMM and Chandra optics that support the RGS attenuation assumption.","marker":"Nartallo 2002"}],"fun_headline_variants":["XMM proton flares: power law at source, excess from matrices","Real-data test: soft proton flare source is a pure power law","Proton flares deconvolved: power law plus simulation artifact","Soft excess in XMM flares traced to matrix imprecision","Power law confirmed for proton flares; soft bump is artifact"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes the simulated proton response matrices accurately describe how protons of each energy are transmitted and deposit charge in the real MOS and PN detectors; if the simulated effective area is wrong in the 2–5 keV band, what looks like a matrix artifact could be a real spectral component.","fun_headline_variants_meta":{"raw":{"variants":["XMM proton flares: power law at source, excess from matrices","Real-data test: soft proton flare source is a pure power law","Proton flares deconvolved: power law plus simulation artifact","Soft excess in XMM flares traced to matrix imprecision","Power law confirmed for proton flares; soft bump is artifact"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000195,"raw_usage":{"total_tokens":1403,"prompt_tokens":1039,"completion_tokens":364,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":655,"completion_tokens_details":{"reasoning_tokens":278}},"tokens_in":655,"tokens_out":364,"duration_ms":4327,"temperature":1.0,"reasoning_tokens":278,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T23:50:01.334719+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A laboratory measurement of proton transmission through the actual MOS optical filters and electrode structures, compared with the simulated response, would settle whether the 2–5 keV excess is a matrix artifact; if the simulated transmission is correct, the black-body component should disappear once the matrices are corrected, while a residual excess would support a real low-energy component such as heavier ions.","supporting_citations":[{"cited_title":"2024, A&A, same vol ume","cited_arxiv_id":null,"evidence_quote":"Builds the proton response matrices for MOS and PN from Monte Carlo simulations; these matrices are the objects being tested on real data."},{"cited_title":"2018, Experimental Astronomy, 45, 411","cited_arxiv_id":null,"evidence_quote":"Derives power-law proton environment models above 50 keV that the paper adopts as the expected input spectrum."},{"cited_title":"2018, ApJ, 867, 9","cited_arxiv_id":null,"evidence_quote":"Predicts power-law distributions for different magnetosphere regions, supporting the power-law input model."},{"cited_title":"P ., Beardmore, A","cited_arxiv_id":null,"evidence_quote":"Documents the contamination layers on the MOS CCDs used to explain the 20-year degradation in proton response."},{"cited_title":"2002, Esa /estec/tos-ema/02-067/RN Technical Note","cited_arxiv_id":null,"evidence_quote":"Provides simulations of soft proton focusing by XMM and Chandra optics that support the RGS attenuation assumption."}],"review_version":1}