{"id":"51220afe-2775-48d1-a573-bacc54584752","arxiv_id":"2505.08885","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Multi-epoch XMM-Newton data reveal a constant 2-10 keV luminosity and a declining soft X-ray flux in Green Pea galaxy J0822+2241, identifying it as the first confirmed X-ray AGN in a Green Pea galaxy.","lead":"Deep X-ray observations of the Green Pea galaxy J0822+2241 show a constant hard X-ray luminosity while the soft X-ray flux dropped by about 60% over 6.2 years, a pattern best explained by a low-luminosity active galactic nucleus (AGN) whose line-of-sight obscuration increased. This is the first conclusive X-ray identification of an AGN in a Green Pea galaxy, strengthening the idea that faint AGN in compact galaxies could have helped reionize the early universe.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'only plausible scenario' conclusion rests on a tied intrinsic photon index; two epochs do not distinguish NH variability from intrinsic spectral variability.","rationale":"The paper's central claim is conditional on a specific two-epoch spectral decomposition. The strongest evidence for an AGN is genuinely multi-wavelength: constant 2–10 keV luminosity, an SED requiring an AGN component, broad H-alpha stable over 16 years, and a radio-to-X-ray ratio in the Seyfert range. These supports make 'J0822+2241 hosts an AGN' robust. What is not robust is the narrower conclusion that the soft decline is caused by an NH increase rather than by a change in the intrinsic continuum. Section 5.1 imposes Gamma = 1.7 on both epochs; the paper never reports a Bayesian or maximum-likelihood comparison with a variable-continuum model. The independent epoch fits yield observed Gamma values of 1.7 and 0.8, and with only two epochs the degeneracy between NH and Gamma cannot be resolved by the data. The paper's own language is more careful than the title: Section 7 says 'out of all scenarios tested,' and Section 3.1 calls the spectral variability test phenomenological rather than causal. The reader's CONDITIONAL verdict captures exactly this gap, and I would not move it. If the proposed model-comparison test returns Delta ln Z below -5 in favor of variable NH, the conclusion would be substantially strengthened.","tokens_in":37150,"tokens_out":9612,"duration_ms":103226,"concrete_test":"Refit the combined epoch 1 + epoch 2 EPIC-pn spectra with BXA/UltraNest using identical priors and instrument responses for (A) the Section 5.1 model (TBabs x zTBabs x cabs x zpowerlw, Gamma tied at 1.7, NH free per epoch) and (B) the same continuum but with Gamma free per epoch and NH tied to the epoch 1 value. Compare the Bayesian evidence difference Delta ln Z = ln Z_B - ln Z_A. If Delta ln Z is not strongly negative (say below -5), the constant-photon-index assumption is not decisively required and intrinsic spectral variability remains viable; the 'only plausible scenario' claim would need to be softened. If Delta ln Z is below -5, the obscuration-variability interpretation is quantitatively preferred and the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is in Section 5.1: to identify the soft X-ray decline with an NH increase, the simultaneous fit ties the intrinsic photon index to Gamma = 1.7 in both epochs and allows only the absorbing column to vary (Table 2). This is an assumption imposed by the model, not a result of the fit. The independently fitted observed photon indices are 1.7 in epoch 1 and 0.8 in epoch 2 (Section 3.1); a model with constant Gamma necessarily converts that change into NH. The paper does not compare against the plausible alternative in which the intrinsic continuum itself changed, e.g., a harder coronal state in epoch 2 or a soft excess/thermal component present in epoch 1 that faded, nor against a warm or partial-covering absorber. Since epoch 1's soft 0.5–2 keV luminosity is close to the unobscured expectation for Gamma = 1.7, while epoch 2's is suppressed, both the variable-obscuration and variable-continuum interpretations can reproduce the two data points. The paper itself notes in Section 3.1 that the variability detection is purely phenomenological and 'do[es] not necessarily provide a causal link between each epoch.' The causal link is supplied only by the tied-Gamma assumption. With two epochs and no model-selection test between variable-NH and variable-continuum models, 'conclusively identifies' and 'only appropriate possibility' are stronger than the analysis demonstrates.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a multi-epoch X-ray study of the Green Pea galaxy SDSS J082247.66+224144.0 using archival and new XMM-Newton observations separated by 6.2 rest-frame years. The authors find that the 2–10 keV luminosity is consistent between epochs while the 0.5–2 keV flux drops by about 60%, and that the observed photon index hardens from Γ≈1.7 to Γ≈0.8. They then fit a simultaneous absorbed power-law model with a common intrinsic photon index and epoch-dependent line-of-sight column density, reporting an increase in NH from log NH≈20.7 to ≈21.8 cm−2 at >99.7% confidence. Supporting analyses detect a broad Hα component (FWHM≈1360 km/s), perform broadband SED fitting with an AGN component, derive black hole and stellar masses, and compare the source with JWST-selected AGN that lack X-ray detections. The paper concludes that an AGN with transient low-column-density obscuration is the only plausible explanation for the X-ray properties, making J0822+2241 the first conclusively identified X-ray AGN in a Green Pea galaxy.","tokens_in":37291,"tokens_out":6360,"duration_ms":63872,"significance":"If the identification holds, this is an important result: it provides a local template for low-luminosity AGN in compact, low-metallicity star-forming galaxies and supports the idea that such AGN could contribute to reionisation. The study is methodologically careful in several ways: spectral fitting uses Bayesian nested sampling with posterior predictive checks, the NH-ratio significance is checked with a beta-function fit, the epoch-1 NH lower bound is explicitly treated as unconstrained, and the multi-wavelength evidence (broad Hα, SED, radio-to-X-ray ratio) is assembled with appropriate caveats. The main weakness is that the variability mechanism is inferred from a model that assumes the intrinsic continuum is identical between epochs; with only two epochs this assumption is not tested against plausible alternatives, so the 'only plausible scenario' wording is stronger than the data warrant.","major_comments":[{"comment":"The central inference of an obscuration increase is conditional on the assumption that the intrinsic AGN continuum is a power law with a single photon index Γ=1.7 in both epochs. The independent fits in Section 3.1 give Γ=1.7+0.5−0.6 in epoch 1 and Γ=0.8+0.2−0.3 in epoch 2, so the tied-Γ model is forced to map the spectral change into NH. The paper does not fit or compare a model in which the intrinsic continuum changes between epochs (for example, a free photon index in each epoch, a soft excess present in epoch 1 that fades, or a variable warm/partial-covering absorber). Because the two epochs cannot distinguish these physically distinct interpretations, the statement that variable obscuration is the 'only appropriate possibility' is not supported by the analysis. I recommend adding a model-comparison test and softening the conclusion accordingly.","section":"§5.1, Table 2"},{"comment":"The claim that the hard 2–10 keV luminosity is constant rests on weak epoch-1 hard-band data: Table 1 lists a hard-band signal-to-noise ratio of only 1.8 for epoch 1, and the quoted luminosity has an uncertainty of ±0.3 dex. A factor-of-two change in the hard continuum between epochs is therefore not excluded. This weakens the argument that the soft-band decline must be due to absorption rather than to intrinsic continuum variability, and it should be stated explicitly when the scenarios are compared.","section":"Table 1, §3.1"},{"comment":"The abstract and Section 7 claim that the X-ray properties 'conclusively' identify an AGN and that obscuration variability is the only appropriate possibility. Section 3.1 itself notes that the variability tests are 'purely phenomenological' and do not provide a causal link. The AGN identification is supported by multiple independent diagnostics and may well be correct, but the uniqueness of the obscuration-variability scenario is not established. The language should be revised to distinguish a robust AGN identification from a preferred, but not unique, physical interpretation of the epoch-to-epoch spectral change.","section":"Abstract and Section 7"}],"minor_comments":[{"comment":"The caption lists '[Oii]λλ4959, 5507', which appears to be a typo for '[Oiii]λλ4959, 5007'; please check and correct.","section":"Figure 11 caption"},{"comment":"The caption refers to the broad Hα measurement as '(c.f. Section 5.1)', but the measurement is presented in Section 3.2.2; the cross-reference should be corrected.","section":"Figure 7 caption"},{"comment":"The equivalent width of the broad Hα component is reported as EWHα,broad = 90+12−10 Å; please state explicitly whether this is in the rest frame or observed frame, for consistency with the other reported quantities.","section":"Section 3.2.2"}],"recommendation":"major_revision","confidential_remarks":"This is a solid and interesting paper. My main concern is overinterpretation: the data are consistent with variable obscuration, but the 'only plausible scenario' claim is not justified by the two-epoch tied-Γ model. I do not see grounds for rejection, but the abstract, title, and Section 7 need to be adjusted and a model-comparison test for a variable continuum should be added. The scope is appropriate for the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is the most credible AGN claim yet in a Green Pea galaxy, and I think the AGN is real. The X-ray luminosity, the broad H-alpha component, the radio-to-X-ray ratio, and the broadband SED all line up. But the paper's own central twist, that we are seeing a column density increase between 2013 and 2020, is not as secure as they say, and the headline 'conclusively identifies' is a step too far given what two epochs can actually do.\n\nWhat is genuinely new: 111 ks of XMM data added to 28 ks archival, giving the deepest X-ray look at a Green Pea. They detect statistically significant soft-band variability while the hard band stays flat. They find a broad H-alpha component in the SDSS spectrum, then check it with fresh Palomar data 16 years later (rest-frame) and show it has not changed, which is a nice supernova test. They fit the SED with Lightning and get a self-consistent AGN plus host picture. The statistical work is careful: BXA/UltraNest, posterior predictive checks via Q-Q plots, explicit treatment of the unconstrained epoch-1 N_H lower bound. They also take alternative explanations seriously, which is more than many papers do.\n\nThe soft spot is the interpretation of the variability. The simultaneous fit that supports the N_H-increase story ties the photon index at Gamma = 1.7 across both epochs and only lets the column change. The standalone fits give observed Gamma ~ 1.7 in epoch 1 and ~ 0.8 in epoch 2. Under the tied-Gamma model, all of that becomes absorption; but a variable intrinsic continuum (a harder coronal state, a fading soft excess, even partial covering) could reproduce the same two data points. The paper notes in Section 3.1 that the spectral-change test is phenomenological and does not by itself give a causal link, then Section 5.1 supplies that link through the model assumption. With two epochs and no model-selection run against a variable-continuum alternative, 'only appropriate possibility' is stronger than what the data support. I do not think this kills the AGN identification, which rests on a pile of independent evidence, but it does mean the obscuration-variability narrative should be presented as preferred, not conclusive.\n\nThe stellar mass and black hole mass are both messy, and the paper says so; the H-alpha-based BH mass is an extrapolation and could be off, but they flag the systematics.\n\nWho should read it: anyone working on low-luminosity AGN, Green Pea galaxies, or the local-analog connection to JWST Little Red Dots. It is a solid, useful data paper. I would send it to a serious referee and expect the authors to soften the 'conclusive' language; ideally they would add a simple test of a variable-Gamma model or an explicit discussion of the degeneracy. That is a revision, not a rejection.","headline":"Strong multi-wavelength case for an AGN in a Green Pea galaxy, though the 'only plausible' variable-obscuration scenario is not uniquely established with two epochs.","tokens_in":38339,"tokens_out":3925,"would_cite":true,"duration_ms":39743,"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":"The paper argues that the X-ray changes seen in the Green Pea galaxy J0822+2241 can only be explained by a low-luminosity active galactic nucleus hidden behind a cloud of gas that thickened tenfold between 2013 and 2020.","keywords":["Green Pea galaxies","active galactic nuclei","X-ray spectroscopy","obscuration variability","XMM-Newton","low-luminosity AGN","reionisation analogues","ultraluminous X-ray sources"],"falsifier":"A third X-ray observation that finds the soft 0.5-2 keV flux recovering toward its 2013 level while the 2-10 keV luminosity remains at $10^{42}$ erg/s would contradict the simple obscuration interpretation, since the column would have to both rise and fall on a timescale shorter than the 6.2-year gap. Conversely, a new measurement showing the hard luminosity changing by more than the quoted uncertainties while the soft flux stays suppressed would also falsify the constant-intrinsic-continuum assumption.","tokens_in":36856,"feed_emoji":"🔭","tokens_out":6192,"duration_ms":59210,"temperature":0.7,"pith_summary":"This paper reports the deepest X-ray campaign of a Green Pea galaxy to date, targeting SDSS J082247.66+224144.0. Across two XMM-Newton epochs separated by 6.2 years in the galaxy's rest frame, the hard 2-10 keV luminosity stayed constant near $10^{42}$ erg/s while the soft 0.5-2 keV flux dropped by about 60%. After testing transient supernovae, unresolved X-ray binaries, ultraluminous X-ray sources, and unobscured AGN variability, the authors conclude that only an AGN whose line-of-sight gas column increased by roughly a factor of ten can explain the data. If correct, this is the first conclusive X-ray identification of an AGN in a Green Pea galaxy, supporting the idea that low-luminosity black hole accretion in compact star-forming galaxies could have contributed to cosmic reionisation.","feed_headline":"First conclusive X-ray AGN found in a Green Pea galaxy","feed_subtitle":"Hard X-rays stay steady while soft X-rays fade, a signature of a black hole slipping behind gas in a nearby compact galaxy.","key_machinery":"The central machinery is a two-epoch X-ray spectral fit with an absorbed powerlaw model (zTBabs*cabs*zpowerlw), in which the intrinsic photon index is tied to Gamma = 1.7 across both epochs while the line-of-sight column density NH is allowed to vary independently in each epoch. Bayesian posterior sampling with nested sampling propagates the full covariance between the two column density measurements into a single posterior on the ratio NH,2/NH,1, and integrating the probability mass above unity gives the >99.7% confidence in an obscuration increase. Quantile-Quantile difference plots, which compare cumulative detected counts to model-predicted counts without binning, quantify the significance of the soft-band spectral change and confirm that neither epoch's model can reproduce the other epoch's data below 2 keV.","core_discovery":"Out of all the scenarios tested, an AGN displaying obscuration variability is the only appropriate possibility to explain the X-ray properties of J0822+2241 observed by XMM-Newton. The source's rest-frame 2-10 keV luminosity was constant within uncertainties between 2013 and 2020, while its 0.5-2 keV flux decreased significantly, hardening the observed spectrum from a photon index near 1.7 to about 0.8. A simultaneous spectral fit that ties the intrinsic photon index at Gamma = 1.7 and lets only the line-of-sight column density vary between epochs finds log(NH,2/NH,1) = 1.08+0.66-0.72, with the increase in obscuration required at more than 99.7% confidence. The paper also identifies a broad component to H-alpha with FWHM 1360 km/s in the archival SDSS spectrum, argues it is stable over a 16-year rest-frame baseline, and uses the radio-to-X-ray ratio from VLA data to further support an AGN origin. A fundamental requirement is that the AGN dominates the observed X-ray spectrum below 10 keV, a condition that is rare among highly star-forming galaxies.","pith_inferences":["If a larger sample of Green Pea galaxies shows the same two-epoch pattern of a steady hard X-ray band and a fading soft band, obscuration variability would become a generic test for low-luminosity AGN in compact star-forming galaxies rather than a single-object oddity.","The ~6 year rest-frame timescale of the column density change constrains the obscuring gas to be fairly close to the black hole; dense clumps or an outflow on sub-parsec scales could reproduce the observed change, and monitoring over additional epochs would directly test whether the column continues to rise, stays high, or returns.","Because the photon index is tied rather than free, a natural stress test is to let Gamma vary independently between epochs in the same Bayesian framework; the posterior on NH,2/NH,1 would then quantify how much of the inferred obscuration increase is tied to the constant-slope assumption.","The results suggest a concrete search strategy: a shallow X-ray snapshot of a Green Pea galaxy that shows a hard spectrum (Gamma below about 1.0) and a bright 2-10 keV luminosity could be a sign of ongoing or recent obscuration, making such objects high-value targets for follow-up spectroscopy above 10 keV."],"forward_implications":["J0822+2241 becomes the first Green Pea galaxy with a conclusive X-ray identification of an AGN, giving astronomers a local, well-studied analogue for interpreting X-ray-faint AGN candidates at high redshift.","The constant hard X-ray luminosity implies a stable, low-luminosity AGN near 10^42 erg/s, meaning its observed soft-band variability is dominated by gas crossing the line of sight rather than by intrinsic changes in accretion power.","The Eddington ratio estimate of roughly 1.4% places the source near the effective Eddington limit on dusty gas, a regime associated with radiation-pressure-driven outflows and changing obscuration.","The broad H-alpha component that persists across a ~16 year rest-frame baseline supports an AGN broad line region rather than a single supernova event.","J0822+2241's broad H-alpha and X-ray luminosities sit about 1-2 dex below local AGN scaling relations, matching the X-ray deficit seen in JWST-detected AGN candidates and suggesting that dedicated X-ray campaigns on Green Pea galaxies could clarify that population's nature."],"supporting_citations":[{"why":"Supplied the initial X-ray detection, photon index, and luminosity for J0822+2241 that motivated this deeper multi-epoch campaign.","marker":"Svoboda et al. (2019)"},{"why":"Showed that X-ray binary emission could contribute at most about 20% of the observed luminosity, ruling out a purely stellar-power explanation.","marker":"Adamcová et al. (2024)"},{"why":"Provided VLA 6 and 10 GHz radio data used to derive a radio-to-X-ray ratio consistent with local AGN and inconsistent with X-ray binaries.","marker":"Borkar et al. (2024)"},{"why":"Established the NuSTAR non-detection and WISE colour constraints that any Compton-thick or AGN-dominated interpretation must satisfy.","marker":"Kawamuro et al. (2019)"},{"why":"Provided the typical AGN photon index distribution and obscured AGN fraction used to benchmark the fitted intrinsic Gamma = 1.7.","marker":"Ricci et al. (2017a)"},{"why":"Supplied the JWST-detected AGN sample with mostly undetected X-ray counterparts to which J0822+2241's broad H-alpha and X-ray luminosities are directly compared.","marker":"Maiolino et al. (2025)"},{"why":"Gave the black hole versus stellar mass scaling relation used to interpret the broad-H-alpha black hole mass and the Lightning SED results.","marker":"Reines & Volonteri (2015)"},{"why":"Provided the multi-mission catalogue of ULX peak luminosities that rules out an individual off-nuclear source as the origin of the observed X-ray emission.","marker":"Walton et al. (2022)"}],"fun_headline_variants":["First X-ray AGN confirmed in a Green Pea galaxy","XMM-Newton finds AGN in Green Pea: soft X-rays fade","Green Pea galaxy hosts AGN with changing obscuration","Hard X-rays steady, soft fade: AGN found in Green Pea"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model assumes the AGN's intrinsic X-ray spectrum stayed exactly the same shape from 2013 to 2020 while only the amount of absorbing gas on the line of sight changed, and with just two epochs of data, an intrinsic spectral change could mimic the claimed obscuration increase.","fun_headline_variants_meta":{"raw":{"variants":["First X-ray AGN confirmed in a Green Pea galaxy","XMM-Newton finds AGN in Green Pea: soft X-rays fade","Green Pea galaxy hosts AGN with changing obscuration","Hard X-rays steady, soft fade: AGN found in Green Pea"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000191,"raw_usage":{"total_tokens":1467,"prompt_tokens":1196,"completion_tokens":271,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":812,"completion_tokens_details":{"reasoning_tokens":194}},"tokens_in":812,"tokens_out":271,"duration_ms":2918,"temperature":1.0,"reasoning_tokens":194,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:46:12.945227+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A third X-ray observation that finds the soft 0.5-2 keV flux recovering toward its 2013 level while the 2-10 keV luminosity remains at $10^{42}$ erg/s would contradict the simple obscuration interpretation, since the column would have to both rise and fall on a timescale shorter than the 6.2-year gap. Conversely, a new measurement showing the hard luminosity changing by more than the quoted uncertainties while the soft flux stays suppressed would also falsify the constant-intrinsic-continuum assumption.","supporting_citations":[{"cited_title":"2019, , 880, 144, 10.3847/1538-4357/ab2b39","cited_arxiv_id":null,"evidence_quote":"Supplied the initial X-ray detection, photon index, and luminosity for J0822+2241 that motivated this deeper multi-epoch campaign."},{"cited_title":"2019, , 881, 48, 10.3847/1538-4357/ab2bf6","cited_arxiv_id":null,"evidence_quote":"Established the NuSTAR non-detection and WISE colour constraints that any Compton-thick or AGN-dominated interpretation must satisfy."}],"review_version":1}