{"id":"5d06ad8d-9c09-4520-ac1d-0515d3fb43d3","arxiv_id":"2412.11178","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"In 59 type 1 AGNs, the soft X-ray excess profile correlates with its strength: stronger excess is more powerlaw-like, and only a hybrid reflection plus warm corona model reproduces the trend.","lead":"The authors analyzed X-ray and UV spectra of 59 bright active galaxies and found that stronger soft X-ray excess emission tends to have a more extended, powerlaw-like spectral shape, while weaker excess is more blackbody-like. They show that this shape-strength link cannot be produced by ionized disk reflection alone, but a mix of reflection plus a warm corona can reproduce it, suggesting a hybrid origin for the soft excess.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Hybrid-model validation is circular: §5.3 fits relxilllp+compTT to the same 59 spectra before 'reproducing' the observed correlation, so the ubiquitous hybrid-nature claim is not independently tested.","rationale":"The reader's weakest_assumption is residual ionized absorption/A Appendix A limitations. That is a legitimate systematic concern, and the paper's Appendix A does flag the limitation of using only the first RGS-detected absorber with fixed covering factor. However, the most load-bearing weakness is the circular validation of the hybrid model in §5.3. The central claim, as stated in the abstract, pairs a quantitative correlation with the statement that the correlation 'cannot be explained by ionized disk reflection alone' and that a toy hybrid scenario 'can successfully reproduce' it, 'highlighting the ubiquitous hybrid nature.' The simulation is not an independent test: the model parameters are fit to the same data used to define the correlation, so the 'reproduction' is expected from the fitting procedure. The paper's own limitation statement in §5.3 ('properly decomposing the two components ... deferred to a future work') underscores that the decomposition is degenerate. The log q–log Ecut degeneracy test in §5.1 is also flawed as a null test, but the paper has mitigating evidence: the chi-square-difference metric is not affected by the same parameter degeneracy, and the unfolded spectra in Fig. 5 visually show the trend. Thus the empirical correlation is likely robust, but the physical interpretation is not independently supported. The reader's verdict of CONDITIONAL is appropriate; my concern reinforces it rather than changing it, hence UNCHANGED. Agreement is partial because the reader's rationale lists the hybrid model issue among the reasons for conditional, though the formal weakest_assumption field identifies absorption instead.","tokens_in":43545,"tokens_out":11099,"duration_ms":112467,"concrete_test":"Hold out a random subset of ~20 sources. Fit relxilllp+compTT to the remaining 39 sources, and use the resulting distribution of best-fit parameters (or a physically motivated prior, e.g., the REXCOR fX/hf parameter space) to generate simulated spectra for the held-out sources without fitting their data. Fit those simulated spectra with model 3 and compare the predicted log q–log Ecut and log q–Δχν² relations to the observed ones. If the out-of-sample prediction reproduces the correlation within bootstrap errors, the hybrid interpretation gains real support; if not, the §5.3 simulation is circular and the ubiquitous-hybrid claim remains unsubstantiated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's physical conclusion—that the soft X-ray excess is ubiquitously hybrid—rests on §5.3, where relxilllp+compTT is claimed to reproduce the observed log q–log Ecut correlation. But the 59 input parameter sets are obtained by fitting the same 59 core spectra with relxilllp+compTT (with kT fixed at 1 keV and reflection parameters fixed at defaults). The fake spectra are therefore drawn from a model already tuned to each source, and the recovered distribution of model-3 parameters is expected to resemble the data whether or not the hybrid model is physically correct. This is an in-sample consistency check, not a falsifiable prediction. The paper itself acknowledges in §5.3 that 'properly decomposing the two components through spectral fitting could be challenging, and is deferred to a future work,' and that the toy setup is adopted precisely because of the strong degeneracy between reflection and warm corona. As a result, the per-source decomposition is not unique, and the correlation between fitted q and Ecut is largely inherited from the fitting process rather than independently predicted. The contrast with relxilllp alone in §5.2 is similarly weakened: those fits are also made to the same data, and their fit quality is not reported, so the simulation does not cleanly represent the reflection-only hypothesis. The empirical correlation itself is supported by the unfolded spectra in Fig. 5 and by the chi-square-difference metric, so the central correlation is likely real; what is not established is the stronger claim that a hybrid model explains it in a meaningful, predictive sense.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a sample of 59 unobscured type 1 AGNs with simultaneous XMM-Newton EPIC-pn and OM UVW1 observations. It fits the 0.5-10 keV spectra with blackbody, powerlaw, and cut-off powerlaw soft excess models (plus pexrav and Fe Kalpha lines), classifies sources as bb-like or po-like, and defines the soft excess strength log q as the 0.5-2 keV luminosity ratio of the soft excess to the primary continuum. The main empirical claim is a correlation between the SE profile (Ecut or reduced chi-square difference) and log q, with Spearman rho about 0.56-0.57. The paper further argues via relxilllp and relxilllp+compTT simulations that ionized reflection alone cannot produce this correlation while a hybrid model can, concluding that the soft X-ray excess is ubiquitously hybrid.","tokens_in":43848,"tokens_out":8892,"duration_ms":81499,"significance":"The empirical correlation, if robust, is a valuable observational constraint on soft excess models. The paper has real strengths: a homogeneous XMM-Newton sample with simultaneous UV data, a cut-off powerlaw that uniformly describes both SE profile families, two independent profile metrics including a chi-square-difference metric that is not degenerate with the model-3 quantity log q, direct unfolded SE spectra, and a careful treatment of censored Ecut values with ASURV. The model simulations are less persuasive: the hybrid-model 'reproduction' is an in-sample consistency check rather than a falsifiable prediction, and the reflection-only comparison lacks reported fit quality. The central observational result and the toy-model illustration can be separated; the physical conclusion should be presented with correspondingly weaker epistemic weight.","major_comments":[{"comment":"The fakeit test in Section 5.1 does not test whether the log q - log Ecut degeneracy can produce an artificial correlation. Because each fake spectrum is generated from the observed best-fit model 3 parameters, the input sample already contains the observed correlation; the mean recovered rho = 0.63 only shows that re-fitting preserves the input correlation. A valid null experiment would decouple the two parameters (for example, shuffle Ecut across sources before generating fake spectra) and then measure the correlation induced purely by the fitting degeneracy. As presented, the test cannot support the statement that the correlation 'cannot be attributed to parameter degeneracy.' In addition, since 32/42 po-like Ecut values are lower limits, the reported rho = 0.56 for log Ecut depends on the ASURV censoring treatment; please also report a detection-only analysis and a shuffled null distribution.","section":"Section 5.1 and Fig. 7"},{"comment":"The relxilllp+compTT simulation is calibrated by fitting the same 59 core spectra with relxilllp+compTT, so the simulation is an in-sample retrodiction rather than an independent confirmation. It shows that the hybrid model family is flexible enough to accommodate the observed correlation, but it does not establish the abstract's conclusion of a 'ubiquitous hybrid nature.' The manuscript itself notes that the two components are strongly degenerate and that proper decomposition is deferred to future work. The claim should be reframed as consistency with a hybrid scenario, or supported by out-of-sample predictions or by a model-selection statistic (for example, Delta chi-square or AIC between relxilllp-only and relxilllp+compTT fits on the observed spectra).","section":"Section 5.3 and Fig. 9"},{"comment":"The fit quality of the relxilllp-only and relxilllp+compTT fits to the core sample is not reported. Without reduced chi-square values or other goodness-of-fit statistics, the reader cannot tell whether the reflection-only model is an acceptable description of the data; if it is statistically rejected, the simulated parameter distribution does not fairly represent the reflection-only hypothesis. Please report the fit statistics and parameter distributions for these fits, and consider a likelihood-based model comparison.","section":"Sections 5.2 and 5.3"},{"comment":"The residual ionized absorption check in Appendix A only includes the first and most significant RGS-detected absorber, imposes its column density as an upper limit, and fixes the covering factor to unity. Weaker or multiple ionized absorbers could remain unmodeled and shift both the bb-like/po-like classification and Ecut. Please quantify how much residual absorption would be required to remove the observed log q - profile correlation, or fit the full set of RGS-detected absorption zones for at least a subset of sources.","section":"Appendix A and Section 3.2"}],"minor_comments":[{"comment":"The abstract says all relxilllp parameters are fixed except ionization, but Section 5.3 leaves the relxilllp luminosity and photon index free. Please harmonize the parameter description.","section":"Abstract vs. Section 5.3"},{"comment":"The definition of log q is missing a parenthesis; it should read log q = log(L_SE,0.5-2,cpl / L_PC,0.5-2,cpl).","section":"Equation (2)"},{"comment":"The phrase 'see Appendix 3.2' appears to be a wrong cross-reference; it should be Appendix A or Section 3.2.","section":"Section 3.1, item 3"},{"comment":"The sentence 'with a median chi2_bb - chi2_cpl of 5.98 for chi2_bb - chi2_cpl' is redundant and should be rewritten.","section":"Section 3.1"},{"comment":"The labels 'Median = 0.14 keV (KM)' and 'Median = 1.40 keV' should explicitly identify which subgroup (bb-like and po-like) each median refers to.","section":"Figure 6, upper panel"},{"comment":"The p-values are not corrected for multiple comparisons; with five physical quantities tested, the 2% p-value for LUV/LPC should be treated as marginal evidence only.","section":"Table 1"},{"comment":"The name 'Zogbhi' in the Cackett et al. 2013 reference entry should be 'Zoghbi'.","section":"Bibliography"},{"comment":"The 'unfolded' soft excess spectra are still model-dependent because they are obtained by subtracting the assumed primary continuum and Fe Kalpha model; please state this caveat explicitly.","section":"Section 4.2"}],"recommendation":"major_revision","confidential_remarks":"The empirical correlation is the strongest part of the paper and is likely publishable after a major revision that repositions the hybrid-model section as an existence proof rather than a physical uniqueness claim. Releasing the spectral fitting and simulation scripts would substantially increase confidence. The 'first time' claim should also be checked carefully against earlier qualitative trends in Boissay et al. (2016) and related work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: the empirical correlation is the real result. The paper reports that in 59 type 1 AGNs, a stronger soft X-ray excess tends to have a more powerlaw-like profile (higher Ecut), with Spearman ρ ~ 0.56–0.57. The correlation is supported by two independent profile metrics, including a blackbody-vs-powerlaw reduced chi-square difference that is not degenerate with the soft excess strength log q, and by direct unfolded soft excess spectra. That is a clean quantitative advance over the qualitative stacking in Boissay et al. 2016, and the uniform cutoff-powerlaw description (median Ecut 1.40 keV for po-like, 0.14 keV for bb-like) is a handy phenomenological tool.\n\nThe soft spots are in the interpretation, not the correlation. The relxilllp+compTT simulation of §5.3 is not an independent test: the input parameters come from fitting the same 59 spectra with that hybrid model, so the recovered distribution of q and Ecut is expected to resemble the data. The paper honestly calls it a toy and notes the strong degeneracy, but the abstract's \"ubiquitous hybrid nature\" claim overreaches what the simulation establishes. A proper test would fix the hybrid model from independent constraints or make a forward prediction on a new sample. The Appendix A absorption check is also limited—only the first RGS-detected absorber, column as an upper limit, covering factor fixed to unity—so weak residual absorption could still bias the bb/po classification, though probably not the overall correlation.\n\nCredit where due: the sample selection is careful (pile-up handling, simultaneous UV, explicit exclusion of strong warm absorbers), and the fits and tables are well documented. The contrast with reflection-only relxilllp simulation is useful even if not definitive.\n\nBottom line: the correlation stands and will be a useful constraint on soft excess models. The hybrid-nature interpretation should be treated as suggestive, not demonstrated. The paper deserves a serious referee; I would send it to review, expecting the hybrid claim to be softened or tested.","headline":"The empirical correlation between soft excess strength and profile is solid; the hybrid-nature interpretation is a plausible but in-sample consistency check.","tokens_in":44423,"tokens_out":3249,"would_cite":true,"duration_ms":26965,"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 reports a quantitative correlation between the soft X-ray excess's spectral profile and its strength in 59 type 1 AGNs, and shows that only a hybrid warm-corona-plus-reflection model reproduces it.","keywords":["soft X-ray excess","type 1 AGN","warm corona","ionized disk reflection","cutoff power law","XMM-Newton","relxilllp","spectral profile"],"falsifier":"Fit the 59 core-sample spectra with high-resolution RGS-informed multi-zone ionized absorption models that allow several absorbers with free covering fractions, then compare the resulting bb-like/po-like classifications and $E_{\\rm cut}$ values with the paper's. If residual ionized absorption systematically shifts the weaker, blackbody-like sources toward power-law-like shapes (or vice versa) and erases the $\\log q$-$E_{\\rm cut}$ correlation, the central claim would be undermined.","tokens_in":43324,"feed_emoji":"🔭","tokens_out":8426,"duration_ms":63746,"temperature":0.7,"pith_summary":"Using a core sample of 59 unobscured type 1 active galactic nuclei observed simultaneously in X-rays and UV with XMM-Newton, the paper tries to establish how the shape of the soft X-ray excess depends on its strength. It finds that roughly 71% of sources show a power-law-like soft excess and 29% a blackbody-like one, and that a single cutoff power law can describe both, with median cutoff energies of 1.40 keV and 0.14 keV respectively. The central new result is a quantitative correlation between the soft-excess profile and its strength: stronger soft excesses are systematically more power-law-like (higher cutoff energy), with Spearman $\\rho$ around 0.56-0.57. The paper argues that ionized disk reflection alone cannot produce this relation, whereas a toy hybrid model combining reflection with a warm corona reproduces it, implying that both components contribute to the soft excess.","feed_headline":"Soft X-ray excess shape tracks its strength in 59 AGNs","feed_subtitle":"Stronger soft excess is power-law-like; weaker is blackbody-like, pointing to a hybrid warm-corona plus reflection origin.","key_machinery":"The load-bearing machinery is the cutoff power law as a unified phenomenological description of the soft excess, with its cutoff energy $E_{\\rm cut}$ acting as a shape parameter: $E_{\\rm cut}\\sim 0.1$ keV mimics a blackbody while $E_{\\rm cut}\\gtrsim 1$ keV mimics a power law. This single family lets the paper place all 59 sources on one profile axis and correlate that axis with the soft-excess strength $\\log q$. The other half of the machinery is the simulation comparison: spectra generated with the ionized reflection model relxilllp alone, versus spectra generated with a toy hybrid model (relxilllp plus the Comptonization model compTT with temperature fixed at 1 keV), are fitted with the same phenomenological models; only the hybrid reproduces the observed $\\log q$-$E_{\\rm cut}$ relation, which is the quantitative argument for a two-component origin.","core_discovery":"The paper's central claim is that the spectral profile of the soft X-ray excess in type 1 AGNs is tied to its luminosity relative to the primary power law: the stronger the soft excess (larger $\\log q$, the 0.5-2 keV soft-excess luminosity divided by the primary continuum luminosity), the more power-law-like its shape, equivalently the higher its cutoff energy $E_{\\rm cut}$. This is quantified by Spearman correlations of $\\rho = 0.56$ between $\\log q$ and $\\log E_{\\rm cut}$ and $\\rho = 0.57$ between $\\log q$ and the reduced chi-square difference $\\chi^2_{\\nu,\\rm bb}-\\chi^2_{\\nu,\\rm po}$, with the correlation checked against parameter degeneracy via simulations. Simulations with the relxilllp ionized reflection model yield mostly blackbody-like soft excesses with $E_{\\rm cut}\\sim 0.1$ keV and fail to reproduce the observed correlation, while a toy hybrid input of relxilllp plus a 1 keV warm corona (compTT) reproduces both the distribution and the correlation. The paper therefore concludes that the soft excess is ubiquitously hybrid, with the relative contribution of ionized reflection and warm corona determining the observed profile.","pith_inferences":["If the hybrid interpretation is correct, the scatter around the $\\log q$-$E_{\\rm cut}$ relation should correlate with independent reflection indicators such as the Fe K$\\alpha$ line strength or soft-lag amplitude; this is testable with the same sample.","The toy model fixes the warm-corona temperature at 1 keV; allowing it to vary while fitting the reflection fraction could either strengthen the hybrid scenario or reveal that part of the correlation is driven by temperature changes rather than by the reflection/warm-corona ratio.","Applying the same analysis to the intermediate 123-observation sample, instead of one observation per source, could test whether the correlation is sensitive to which epoch is chosen for multi-observation sources.","A direct extension would connect the profile dichotomy to UV/optical variability: if the warm-corona component is tied to disk seed photons, the same power-law-like versus blackbody-like distinction should appear in broadband SED or variability behaviour."],"forward_implications":["Single-component ionized reflection models would be incomplete for the majority of type 1 AGNs, so spectral fits should include both a reflection component and a warm corona.","Measured warm-corona temperatures from simple spectral fits would be biased unless the ionized reflection contribution is subtracted first, because the apparent cutoff energy can be set by the reflection-to-warm-corona ratio rather than by temperature alone.","Sources with blackbody-like soft excess and low $E_{\\rm cut}$ are the best candidates for X-ray reverberation mapping, since their soft excess is likely reflection-dominated, consistent with the soft-lag detections among such sources.","The correlation gives a new observational constraint that any self-consistent hybrid model of the soft excess must reproduce, including the REXCOR-type scenario mentioned in the paper."],"supporting_citations":[{"why":"Supplies the CAIXA catalog from which many candidate type 1 AGNs are drawn, defining the initial sample pool.","marker":"Bianchi et al. 2009"},{"why":"Supplies the Swift-BAT 105-month catalog, the other principal source of candidate AGNs for the sample.","marker":"Koss et al. 2017"},{"why":"Provides the relxilllp lamppost reflection model used for the ionized-reflection simulations and for the reflection component in the toy hybrid model.","marker":"García et al. 2014"},{"why":"Predicts from the REXCOR hybrid model that the soft-excess profile varies with the relative warm-corona and reflection contributions, motivating the double-component scenario.","marker":"Xiang et al. 2022"},{"why":"Earlier work found that soft-excess shape varies with strength in stacked spectra, which the paper turns into a quantitative per-source correlation.","marker":"Boissay et al. 2016"},{"why":"Shows that a pure warm corona predicts strong soft-band absorption lines, supporting the need for a hybrid scenario and used to interpret individual sources.","marker":"García et al. 2019"},{"why":"Provides the ASURV Kaplan-Meier and generalized Spearman statistics used for censored Ecut values and correlation estimates.","marker":"Feigelson & Nelson 1985"},{"why":"Supplies the soft-lag detections and non-detections matched to the core sample to test whether reflection-dominated sources are preferentially blackbody-like.","marker":"De Marco et al. 2013"}],"fun_headline_variants":["Soft excess shape tracks strength in 59 AGNs","Stronger soft excess: power-law-like, weaker: blackbody-like","AGN soft excess: strength drives spectral profile","Soft X-ray excess: strength predicts shape in AGNs","Hybrid origin explains soft excess shape-strength link"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that any ionized gas absorption left unmodelled after the zxipcf-based exclusion is too weak to systematically distort the measured soft-excess shape; if weaker or multiple ionized absorbers remain, the blackbody-versus-power-law classification and the reported correlation with strength could be biased.","fun_headline_variants_meta":{"raw":{"variants":["Soft excess shape tracks strength in 59 AGNs","Stronger soft excess: power-law-like, weaker: blackbody-like","AGN soft excess: strength drives spectral profile","Soft X-ray excess: strength predicts shape in AGNs","Hybrid origin explains soft excess shape-strength link"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000808,"raw_usage":{"total_tokens":3648,"prompt_tokens":1150,"completion_tokens":2498,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":766,"completion_tokens_details":{"reasoning_tokens":2418}},"tokens_in":766,"tokens_out":2498,"duration_ms":17008,"temperature":1.0,"reasoning_tokens":2418,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:13:15.729318+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fit the 59 core-sample spectra with high-resolution RGS-informed multi-zone ionized absorption models that allow several absorbers with free covering fractions, then compare the resulting bb-like/po-like classifications and $E_{\\rm cut}$ values with the paper's. If residual ionized absorption systematically shifts the weaker, blackbody-like sources toward power-law-like shapes (or vice versa) and erases the $\\log q$-$E_{\\rm cut}$ correlation, the central claim would be undermined.","supporting_citations":[],"review_version":1}