{"id":"981711d3-ced5-4780-a716-6d7885e8ad44","arxiv_id":"2505.02708","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An eROSITA search for alternating up-down-up-down X-ray light curves yields 16 SMBH binary candidates, with eRASSt J0530-4125 showing a ~1 year quasi-period.","lead":"This paper presents 16 X-ray selected candidate supermassive black hole binaries found by looking for alternating bright and faint states in the eROSITA all-sky survey light curves. The best candidate shows about a one-year quasi-period in X-rays, and the authors estimate that at most a few percent of galaxies host such systems.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Selection's red-noise false-positive rate is untested; the candidate list and the ~0.05 upper limit therefore rest on an explicitly deferred calibration.","rationale":"The reader identified the same load-bearing assumption: the up-down-up-down eROSITA pattern is not yet tested against a red-noise null, and the paper explicitly defers the quantitative false-positive analysis to Paper II. This stress-test confirms that this is the central weakness, and it is precisely the condition on which the candidate list, the ~1-year quasi-periodicity claim for eRASSt J0530-4125, and the ~0.05 upper limit rest. The paper has independent support where it matters: 15/16 sources are spectroscopically confirmed extragalactic, X-ray positions and luminosities support nuclear accretion, stacked X-ray spectra show AGN-like power laws, and multiwavelength follow-up of the best candidate spans several instruments and bands. None of that, however, distinguishes periodic modulation from red-noise stochastic variability in a few cycles. The conditional verdict is therefore appropriate: the paper is a well-executed candidate search with an explicitly stated limitation, but the central physical interpretation should not be accepted as quantitative until the red-noise false-positive rate is computed. No internal inconsistency or unsupported claim beyond the deferred calibration was found, so the reader's verdict need not change.","tokens_in":46466,"tokens_out":3543,"duration_ms":42739,"concrete_test":"Simulate the full selection pipeline: generate millions of red-noise light curves with PSD slopes 1-2 and normalization matched to the observed eROSITA variability-amplitude distribution; resample them at the actual eRASS1-eRASS5 epochs with Poisson errors and 3-sigma upper limits for non-detections; apply the exact Sect. 2.2.1 criteria (alternating >3-sigma changes, F>5, S>3, bright-state flux threshold). Count the expected number of selected sources among the ~40,000 parent galaxies. In parallel, compute the false-alarm probability of the ~1-year peak of eRASSt J0530-4125 by fitting a red-noise PSD model to the combined eROSITA+Swift+NICER light curve using a maximum-likelihood periodogram method.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the 16 (15 extragalactic) up-down-up-down eROSITA sources are plausible SMBHB candidates, and the derived optimistic fraction of ~0.05 SMBHBs per X-ray-emitting galaxy in Sect. 6.3, depends on the assumption that the alternating 4-5 epoch pattern is not produced by red-noise stochastic variability of single AGN. Section 6.1 explicitly concedes: 'We cannot rule out the presence of red-noise variability in our sample until further follow-up campaigns provide strong evidence of either periodic or stochastic variability.' Section 5.2 likewise admits that the Lomb-Scargle periodogram peaks are evaluated only against a white-noise null, not against the red-noise PSD that characterizes AGN. The false-positive rate is deferred entirely to Paper II, so neither the candidate status of the 16 sources nor the ~0.05 upper limit is currently calibrated. Since red noise with PSD slope 1-2 is known to produce few-cycle alternating patterns (Krishnan et al. 2021; Witt et al. 2022), and the adopted thresholds F>5, S>3 over only four 6-month epochs are not obviously rare under a red-noise null, this missing control is the load-bearing weakness. The paper is otherwise internally consistent and appropriately cautious; the issue is an explicit, quantified missing control rather than a contradiction or overclaim beyond what the authors themselves flag.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents the first in a series of searches for supermassive black hole binary (SMBHB) candidates in SRG/eROSITA data. The authors select extragalactic X-ray sources showing alternating 'up-down-up-down' flux patterns across the four (or five) eROSITA all-sky surveys, requiring adjacent epoch flux differences >3σ, a bright-to-faint flux factor F>5, and a significance S>3 for the primary sample of 16 sources. Optical spectroscopy confirms 15/16 as extragalactic nuclear sources; one is classified as an off-nuclear ULX-like source. X-ray follow-up with Swift, NICER, and XMM-Newton is presented, with eRASSt J0530-4125 showing tentative ~1-year quasi-periodic variability. Stacked X-ray spectra are mostly power laws with Γ~1.8–2.8. Using the selected candidates and an assumed eROSITA sensitivity window, the authors estimate an optimistic upper limit of ~0.05 SMBHB candidates per X-ray-emitting galaxy. The paper repeatedly emphasizes that red-noise AGN variability can mimic the selection pattern and defers a quantitative false-positive analysis to Paper II.","tokens_in":46758,"tokens_out":7230,"duration_ms":77789,"significance":"If the selection is robust, this is the first systematic X-ray-selected SMBHB candidate sample from an all-sky survey, providing a valuable target list for multi-wavelength follow-up and a demographic constraint in a regime where most searches are optical. The strengths of the paper are the careful optical spectroscopic confirmation, the multi-wavelength characterization of the candidates, the transparent treatment of alternative scenarios (TDEs, QPEs, HLXs), and the explicit acknowledgment of the red-noise caveat. The ~0.05 fraction, once properly calibrated against a red-noise null, would be an interesting observational constraint. However, the main quantitative claims currently rest on an unquantified selection completeness and false-positive rate, so the significance is conditional.","major_comments":[{"comment":"The selection's false-positive rate under red-noise AGN variability is not quantified. The paper states in §6.1 that 'We cannot rule out the presence of red-noise variability in our sample until further follow-up campaigns provide strong evidence of either periodic or stochastic variability,' and §5.2 notes that the Lomb-Scargle periodogram is evaluated only against a white-noise null. Since the candidate list of 16 sources and the fraction estimate in §6.3 are derived from the F>5, S>3 selection over only 4–5 epochs spaced six months apart, a red-noise false-positive control is load-bearing. Red-noise PSDs with slopes 1–2 are documented to produce few-cycle alternating patterns (Krishnan et al. 2021; Witt et al. 2022), so the authors should provide at least a preliminary simulation-based false-positive rate, or explicitly reframe the candidate list and the fraction as raw selection rates pending Paper II.","section":"§6.1 and §5.2"},{"comment":"The conversion from the raw selection rate to the ~0.05 fraction relies on two ad hoc inputs: a 'sensitive time window' of ~3 months and a uniform period distribution over 30 years. Neither is derived from a simulation of the selection function, and the fraction is directly proportional to both. Furthermore, the numerator is not corrected for red-noise false positives. The conclusion's phrasing '~0.05 SMBHB per X-ray emitting galaxy' overstates the support, since the calculation is an upper limit on the number of selected candidates, not on the confirmed SMBHB population. The authors should either recalibrate these inputs or restrict the claims to selection rates only.","section":"§6.3 and §7"}],"minor_comments":[{"comment":"The significance criterion is written as F_i - F_{i+1} over the error sum, without absolute value; as written it cannot select 'down-up-down-up' profiles. Presumably the absolute value is intended; please clarify.","section":"§2.2.1"},{"comment":"The repeated 'Obs dates [years]' ticks on the sub-panels are visually noisy; the axis label should be placed once per column to improve readability.","section":"Figures 7 and 8"},{"comment":"The sentence 'Most of these candidates are have timescales in the regime...' contains a typo ('are have'); please correct.","section":"Introduction"},{"comment":"The phrase 'We do not discard the possibility that the variability pattern is the result of stochastic processes' is awkward; 'We do not rule out' would be clearer and consistent with the rest of the paper's language.","section":"§5.2.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest and well-executed, but the central demographic number is a raw candidate rate dressed as an upper limit on the SMBHB fraction. I recommend requiring the authors to either provide a preliminary red-noise false-positive estimate or clearly downgrade the demographic claim to a selection rate before publication. The promised Paper II may take time, so the current manuscript should be self-contained for the claims it makes. No concerns about citation integrity or novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know two things about this paper before reading it. First, it is the first systematic use of eROSITA's four all-sky scans to select X-ray SMBHB candidates via an 'up-down-up-down' flux pattern, and it ships a genuinely new sample of 16 sources with follow-up X-ray and optical data. Second, the selection's false-positive rate is explicitly deferred to Paper II, so the candidate list and the ~0.05 upper limit are uncalibrated against red noise. That is the whole story, and the authors know it.\n\nThe new stuff is real. The eROSITA-based sample, the ~1-year quasi-period reported for eRASSt J0530-4125, and the X-ray selection-function upper limit are not in the cited literature. The spectroscopic confirmation, spectral stacking, and multi-wavelength presentation are careful. I particularly credit the authors for not overclaiming: they label sources as candidates, state plainly that red noise can mimic the signal, and outline exactly what Paper II will do. The treatment of eRASSt J1003-2607 as an off-nuclear ULX rather than a binary is a good example of the paper's honesty.\n\nThe soft spots are the ones the paper itself flags. The 4-5 eROSITA epochs are too few to distinguish quasi-periodicity from red noise, and the Lomb-Scargle peaks are tested only against white noise. Section 6.1 says the red-noise null cannot be ruled out, and Section 6.3's ~0.05 fraction compounds that uncertainty with an assumed BLR size, a uniform period distribution, and a hand-chosen sensitivity window. These are not hidden flaws; they are stated limitations. My only real criticism is structural: the central product of this paper — a list of candidates — depends on a calibration that does not exist yet. If Paper II finds a high false-positive rate, most of these sources will fall out, and the upper limit will drop. That does not make the present paper wrong; it makes it preliminary.\n\nWho benefits? Observers planning SMBHB monitoring campaigns, anyone working on PTA/LISA follow-up target lists, and modelers who need an X-ray-selected upper limit. The paper deserves a serious referee. I would send it to review with the clear instruction that the red-noise calibration is the essential next step.\n\nBest,\n---","headline":"An honest, well-scoped candidate paper whose selection is explicitly uncalibrated against red noise; worth refereeing now, with the false-positive analysis being the load-bearing next step.","tokens_in":47399,"tokens_out":601,"would_cite":true,"duration_ms":9450,"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":"Four eROSITA X-ray scans alternating bright and faint identify 16 plausible supermassive black hole binaries, one with an apparent one-year period.","keywords":["supermassive black hole binaries","eROSITA all-sky surveys","quasi-periodic X-ray variability","AGN red-noise variability","X-ray monitoring","SMBHB candidates","soft X-ray selection","galactic nuclei"],"falsifier":"Simulate millions of AGN red-noise light curves with power-spectral slopes 1–2, sample each at the eROSITA cadence (one point every six months, four or five epochs), and count how many pass the up-down-up-down, $F>3$, $S>3$ criteria; if the false-positive rate predicts a sample comparable to 16, the candidate list and the ~0.05 upper limit lose their force. On the observational side, additional cycles of NICER/Swift monitoring that show eRASSt J0530-4125's ~1-year phase and amplitude drifting or disappearing would falsify the periodic interpretation.","tokens_in":46302,"feed_emoji":"🕳️","tokens_out":10249,"duration_ms":99807,"temperature":0.7,"pith_summary":"This paper starts the first systematic X-ray search for supermassive black hole binaries (SMBHBs) using the repeated all-sky scans of eROSITA, which sampled the sky roughly every six months over two and a half years. The authors argue that galaxies whose soft X-ray flux alternates between bright and faint states on consecutive scans ('up-down-up-down') are plausible SMBHB candidates, and they present 16 such objects with optical confirmation of their extragalactic, nuclear origins. The best candidate, eRASSt J0530-4125, shows a quasi-periodic bright-faint cycle of about one year in the observed frame, which, if it is the binary orbital period, would place the two black holes about 1.2 milli-parsecs apart. From the rarity of the selected pattern, they estimate an optimistic upper limit of about 0.05 SMBHB candidates per X-ray-emitting galaxy, consistent with theoretical predictions. The paper stresses that single-black-hole red-noise variability can mimic a few cycles of periodicity, so the candidate status depends on longer monitoring and a false-positive analysis deferred to a companion paper.","feed_headline":"A one-year X-ray heartbeat points to a black hole binary","feed_subtitle":"Four eROSITA scans spaced six months apart flag 16 candidates, with the best repeating every ~12 months.","key_machinery":"The load-bearing object is the eROSITA multi-epoch light curve: four all-sky scans spaced six months apart (plus a partial fifth) create a repeating sample of the sky at half-year intervals. The selection machinery is the 'up-down-up-down' pattern: a source must alternate between bright and faint flux levels from scan to scan, with adjacent epochs differing by more than $3\\sigma$, an average bright-to-faint factor $F>3$, and a normalized significance $S>3$; the monitored primary sample tightens this to $F>5$ and a bright-state flux above roughly $5\\times10^{-13}\\,\\mathrm{erg\\,s^{-1}\\,cm^{-2}}$. The pattern is the observable proxy for the predicted periodic feeding of mini-disks in a binary surrounded by a circumbinary disk. The rest of the machinery—Gaia astrometry to remove stars, optical spectroscopy to confirm nuclei and measure black-hole masses, and Swift/NICER/XMM follow-up to fill the six-month eROSITA gaps—exists to test whether the alternation survives denser sampling and to exclude non-binary interlopers such as flares or off-nuclear ultraluminous sources.","core_discovery":"The discovery claim is that the eROSITA 'up-down-up-down' pattern is a workable X-ray selection for sub-parsec supermassive black hole binaries. Selecting sources whose consecutive 0.2–2.3 keV flux measurements differ by more than $3\\sigma$, with a bright-to-faint flux ratio above 3 and a normalized bright-faint separation above 3, yields 16 monitored candidates; optical spectroscopy confirms 15 of them as nuclear extragalactic sources, with single-epoch black hole masses near $10^7\\,M_\\odot$. The most promising object, eRASSt J0530-4125, displays a tentative period of $\\sim$1 year in its combined eROSITA, Swift-XRT, and NICER light curves, with no change in spectral shape, and, under the binary interpretation, would have an orbital separation of 1.2 milli-pc and an orbital velocity of roughly $8000\\,\\mathrm{km\\,s^{-1}}$. Stacked spectra show power-law photon indices $\\Gamma\\sim1.8$–$2.8$ for 14 of 15 sources, consistent with AGN-like emission. One of the 16, eRASSt J1003-2607, was reclassified as a likely off-nuclear ultra-luminous X-ray source and removed. The paper's headline demographic result is an optimistic upper limit of $\\sim0.05$ SMBHB candidates per X-ray-emitting galaxy, and it explicitly leaves confirmation to future multi-cycle monitoring and to the red-noise false-positive analysis of the companion paper.","pith_inferences":["If the eROSITA pattern survives red-noise testing, the ~0.05 per-galaxy upper limit would be a floor for X-ray-selected SMBHBs, since eROSITA is only sensitive to periods near 4 and 12 months; longer-period binaries would be invisible to this search and would raise the true fraction.","The quick disappearance of eRASSt J1141+0635's apparent 120-day period under denser NICER/Swift cadence suggests that future surveys should treat low-cadence 'up-down' patterns as triage filters whose candidates need immediate higher-cadence observations before any demographic claim.","The one reclassified off-nuclear source, likely an ultra-luminous X-ray source, shows that X-ray-selected SMBHB samples require positional and spectroscopic vetting; similar contamination could affect purely photometric searches in other bands.","If the one-year period of eRASSt J0530-4125 is confirmed, it becomes one of the few SMBHB candidates where X-ray monitoring and high-resolution Fe Kα spectroscopy could be combined to test the binary interpretation directly."],"forward_implications":["eROSITA-style all-sky monitoring becomes a discovery channel for sub-parsec SMBHBs: 16 candidates from the first four scans, with one showing a one-year X-ray cycle.","The one-year periodicity of eRASSt J0530-4125, if binary in origin, fixes the orbital separation at about 1.2 milli-pc and an orbital speed near 8000 km/s, giving a predicted Fe Kα energy shift of 0.17 keV that XMM-Newton can barely resolve.","The pattern's rarity translates into an optimistic upper limit of ~0.05 SMBHB candidates per X-ray-emitting galaxy, consistent with theoretical expectations below 1–10%.","Most candidates show AGN-like X-ray spectra (photon index 1.8–2.8) and low black-hole masses near $10^7$ solar masses, making them plausible targets for high-resolution Fe Kα spectroscopy with future missions.","The candidate list is explicitly provisional: confirmation requires the periodic signal to persist across several cycles in denser monitoring, and red-noise false positives are deferred to the companion paper."],"supporting_citations":[{"why":"Describes the eROSITA instrument and its all-sky scanning pattern that yields four to five epochs six months apart.","marker":"Predehl et al. (2021)"},{"why":"Provides the stacked eRASS:4 catalog with epoch-resolved fluxes used for the light curves.","marker":"Merloni et al. (2024)"},{"why":"Supplies the 3-sigma upper flux limits that allow 'down' states to include non-detections.","marker":"Tubín-Arenas et al. (2024)"},{"why":"Gives parallax and proper motion measurements used to remove stellar contaminants.","marker":"Gaia Collaboration et al. (2021)"},{"why":"Provides the Hβ scaling relation used for single-epoch SMBH masses.","marker":"Vestergaard & Peterson (2006)"},{"why":"Provides the Hα scaling relation that supplements or replaces Hβ mass estimates.","marker":"Mejía-Restrepo et al. (2022)"},{"why":"Documents how red-noise AGN variability can mimic periodicities, motivating the deferred false-positive analysis.","marker":"Krishnan et al. (2021)"},{"why":"Quantifies how period and amplitude relative to noise govern detectability, shaping the monitoring strategy.","marker":"Witt et al. (2022)"},{"why":"Predicts an ~1% SMBHB fraction at z<0.6, the theoretical benchmark for the observed upper limit.","marker":"Kelley et al. (2019)"},{"why":"Predicts sub-parsec binary quasar surface densities used to check the observed candidate fraction.","marker":"Volonteri et al. (2009)"}],"fun_headline_variants":["eROSITA's up-down X-ray pattern uncovers 16 black hole binary candidates","X-ray flicker search finds 16 supermassive black hole binary candidates","Quasi-periodic X-ray flicker exposes hidden supermassive black hole binaries","Supermassive black hole binaries: eROSITA finds 16 new candidates","One-year X-ray heartbeat points to 16 black hole binary candidates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a four-epoch alternating X-ray pattern signals genuine quasi-periodicity, rather than the random flickering ('red noise') that a single accreting black hole can produce.","fun_headline_variants_meta":{"raw":{"variants":["eROSITA's up-down X-ray pattern uncovers 16 black hole binary candidates","X-ray flicker search finds 16 supermassive black hole binary candidates","Quasi-periodic X-ray flicker exposes hidden supermassive black hole binaries","Supermassive black hole binaries: eROSITA finds 16 new candidates","One-year X-ray heartbeat points to 16 black hole binary candidates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001297,"raw_usage":{"total_tokens":5462,"prompt_tokens":1281,"completion_tokens":4181,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":897,"completion_tokens_details":{"reasoning_tokens":4078}},"tokens_in":897,"tokens_out":4181,"duration_ms":31557,"temperature":1.0,"reasoning_tokens":4078,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:43:12.119500+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate millions of AGN red-noise light curves with power-spectral slopes 1–2, sample each at the eROSITA cadence (one point every six months, four or five epochs), and count how many pass the up-down-up-down, $F>3$, $S>3$ criteria; if the false-positive rate predicts a sample comparable to 16, the candidate list and the ~0.05 upper limit lose their force. On the observational side, additional cycles of NICER/Swift monitoring that show eRASSt J0530-4125's ~1-year phase and amplitude drifting or disappearing would falsify the periodic interpretation.","supporting_citations":[],"review_version":1}