{"id":"5e02c026-2834-4a8c-b187-c978a7ecea48","arxiv_id":"2506.14615","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"BEBOP-3b is a 0.56-Jupiter-mass circumbinary giant planet on a 547-day, moderately eccentric orbit, the first radial-velocity discovery of a previously unknown circumbinary system.","lead":"Astronomers report the discovery of BEBOP-3b, a giant planet about half the mass of Jupiter that orbits a pair of stars once every 547 days. It is the first planet found by the radial velocity method around a binary star system that was not already known to host a planet.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 547-day signal is interpreted as planetary without including a correlated-activity model in the null hypothesis; the H-alpha/Na D non-detections are suggestive but do not formally exclude an activity origin.","rationale":"Independent support is real: 141 SOPHIE RVs over 2197 days, median precision 2.3 m/s, a 5.5-sigma HRCCS detection of the secondary, BF1,0 = 3.4e13, and a FIP periodogram peak below 1% all point to a coherent long-period RV signal. The load-bearing step is attributing that signal to a planet. Section 4.4 shows no significant H-alpha or Na D periodicity at 547 d, but the RV model itself contains no activity or correlated-noise component, so a 547-day activity signal is not part of the null hypothesis being tested. The dependence on only two activity indicators, with the other three described as too noisy, further weakens the exclusion. The candidate 1400-day signal, which is only weakly preferred (BF2,1 = 6), demonstrates the presence of smooth long-period residual power and illustrates why a GP/activity comparison is needed. The requested test is standard and uses existing data; it would either remove the concern or show that the planet term is still strongly favoured. This does not justify rejection, but for a new planet claim the activity-correlated null should be explicitly checked before unconditional acceptance. I would therefore move the verdict to CONDITIONAL. Minor issues (Table A4 Msec typo, the Section 4.3 statement about a ~1000-day activity peak versus the 55.6/84.1-day peaks in Section 4.4, and the abstract's 'eccentricity of 0.25' given the marginal eccentricity) should also be cleaned up but do not drive the verdict.","tokens_in":20619,"tokens_out":11131,"duration_ms":121731,"concrete_test":"Re-run the kima RV analysis on the same SOPHIE data, same outlier flags and priors, replacing the planet Keplerian (or adding to the 0-planet model) with a quasi-periodic Gaussian process kernel (e.g., a simple harmonic oscillator or celerite kernel). Compare the evidence for a model containing the 547-day Keplerian against the GP-only model. If the GP-only model reaches comparable evidence (within a factor of ~150) and absorbs the 547-day power, the planetary interpretation is not uniquely supported; if the Keplerian remains strongly preferred and the GP's inferred period is not near 547 d, the activity objection is settled.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is the detection of BEBOP-3b from an 11.8 m/s, 547-day RV modulation. The paper's main defence against an activity origin is Section 4.4: GLS periodograms of H-alpha and Na D show no significant power near 550 d. This is the least secure link in the argument. The kima model set (Section 3.4) includes only Keplerian signals plus Student-t white noise; there is no quasi-periodic correlated-noise or activity term in the null model. Consequently, the reported Bayes factor (BF1,0 = 3.4e13) and the FIP periodogram measure preference for a Keplerian over pure noise, not over an activity process. A long-lived or quasi-periodic activity signal with a ~547-day timescale could evade detection in H-alpha and Na D if those indices are imperfectly coupled to the RV-producing surface features, and the paper itself notes that other indicators (Ca I, He I, Ca II) were too noisy to use. The candidate 1400-day signal, with BF2,1 = 6, indicates that smooth long-period variability exists in the residuals; whether it is planetary or activity, it is not modelled by the null hypothesis. The planetary attribution therefore rests on the assumption that the two usable activity indicators are sufficient tracers, which is not demonstrated by the present analysis.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the discovery of BEBOP-3b, a circumbinary giant planet detected with SOPHIE radial velocities. Using 141 spectra spanning 2197 days, the authors fit the binary orbit and planetary signals with the trans-dimensional Bayesian tool kima, finding strong evidence for a 547-day Keplerian signal with K = 11.8 m/s, a planet mass of 0.558 M_Jup, and a moderately eccentric orbit (e = 0.247). They validate the signal against stellar activity using H-alpha and Na D indices, present a marginal 1400-day outer candidate, perform stability simulations for putative inner planets, derive dynamical stellar masses via HRCCS, and compare those with a joint photometric and RV analysis.","tokens_in":20824,"tokens_out":6720,"duration_ms":78743,"significance":"If correct, this is the first radial-velocity detection of a previously unknown circumbinary system and the first RV-discovered circumbinary planet with a significant eccentricity. The paper also strengthens the empirical population of circumbinary planets by adding a long-period, moderately eccentric member outside the usual \"pile-up\" region. The analysis is thorough: the RV dataset is large and homogeneous, the detection is supported by both a Bayes factor and a FIP periodogram, the HRCCS mass measurement is an independent cross-check, and the stability maps address whether the system is dynamically plausible. The authors are appropriately cautious about the eccentricity and the outer candidate. The main limitation, that the RV null model does not include a correlated-activity component, is real but is mitigated by the lack of significant activity-index periodicity near the planet period and is acknowledged in the text; it does not undermine the central detection claim.","major_comments":[],"minor_comments":[{"comment":"The BEBOP-3 b row lists the secondary stellar mass as 0.3615(39) M_sun, which is inconsistent with Table 1 and Table 2, where M_sec = 0.2615 +/- 0.0039 M_sun. One of these values is a typo and should be corrected.","section":"Table A4"},{"comment":"The marginalised time of periastron is reported as T_peri = 59633(+45,-62) BJD, which is inconsistent with Table 3, where T_peri = 9633(+31,-36) BJD. Please verify and correct this entry.","section":"Table A3"},{"comment":"The text states that \"the highest activity signal is found at ~1000 days,\" but Section 4.4 identifies only a 55.57-day H-alpha peak and an 84.07-day Na D peak, with no 1000-day periodicity described. This apparent contradiction should be resolved, as it bears on the discussion of the candidate outer signal.","section":"Section 4.3"},{"comment":"The activity validation would be more quantitative if the authors estimated what amplitude of a 550-day RV signal could be produced by stellar activity given the measured H-alpha and Na D variations. As written, the conclusion that the signal is \"unlikely\" to be activity is reasonable, but an explicit upper limit would strengthen the planetary interpretation.","section":"Section 4.4"},{"comment":"The abstract states an eccentricity of 0.25 without qualification, while Section 5.2 calls the eccentricity detection \"marginal.\" The abstract should reflect this uncertainty, for example by giving the uncertainty or using language such as \"moderately eccentric.\"","section":"Abstract / Section 5.2"}],"recommendation":"accept","confidential_remarks":"The paper is a solid, well-presented detection paper that fits the scope of the journal. The stress-test concern about the lack of a correlated-activity null model is valid in principle, but the enormous Bayes factor, the independent FIP detection, the lack of significant activity-index power near 550 days, and the authors' own cautious wording make the detection credible. The only substantive issues are a few numerical typos and the minor request for a more quantitative activity upper limit, neither of which affects the central result. I recommend acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper you asked about is the BEBOP team's RV discovery of a circumbinary giant planet around the eclipsing binary BD+79230. The central claim holds up: a 547-day, 11.8 m/s Keplerian signal with a Bayes factor of 3.4e13 against the no-planet model, a FIP periodogram peak below 1% at 500-600 days, and 141 RVs over six years. That is a secure detection by any RV standard.\n\nWhat is genuinely new: this is the first RV-detected circumbinary planet around a system not previously known to host one (BEBOP-1c was in a system already known from transits). BEBOP-3b sits far from the pile-up population, at about 4.8 times the Holman-Wiegert stability radius, on an eccentric orbit. The stellar masses from HRCCS agree with the traditional photodynamical fit at 1.7 sigma, a nice cross-check. The stability analysis showing room for an inner planet is a thoughtful addition.\n\nSoft spots, in order of importance. First, the activity argument: the paper relies on H-alpha and Na D periodograms showing no significant power at 550 d, but the kima model only compares Keplerians plus white Student-t noise against each other, not against a correlated-activity process. That is a real limitation, though it is the standard one in this field and the paper's own caveats soften it. The eccentricity is marginal, and the abstract's plain '0.25' overstates it a bit; the text is more careful. The 1400-day candidate has BF 2,1=6, which the authors correctly call weak. There is also an apparent typo in Table A4: the secondary mass is listed as 0.3615 solar masses, inconsistent with 0.2615 in Table 1. Minor, but should be fixed.\n\nNone of these threaten the planet's existence. The detection is robust, and the authors disclose their data and outliers. I would send this to peer review without hesitation, and I'd be surprised if the referee found a fatal flaw.\n\nRecommendation: accept after minor revision.","headline":"BEBOP-3b is a convincing, well-documented RV discovery of a circumbinary giant planet; the main weakness is the usual activity-null-model caveat, not the detection itself.","tokens_in":21574,"tokens_out":3469,"would_cite":true,"duration_ms":34076,"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 547-day radial-velocity wobble of the binary BEBOP-3 is caused by a circumbinary giant planet of about half a Jupiter mass, making this the first radial-velocity detection of a previously unknown circumbinary system.","keywords":["circumbinary planets","radial velocity method","eclipsing binaries","giant planet","orbital eccentricity","dynamical masses","orbital stability","high-resolution cross-correlation spectroscopy"],"falsifier":"If a stellar activity tracer with a period near 547 days is found in the same spectra—e.g. in the Ca II H&K S-index, line bisectors, or the CCF full-width-at-half-maximum—the planetary interpretation fails. Alternatively, if a further three to five years of radial velocities do not preserve a coherent Keplerian phase at 547 days, with the period locked to better than a few days, the signal is not a planet.","tokens_in":20367,"feed_emoji":"🪐","tokens_out":10694,"duration_ms":93171,"temperature":0.7,"pith_summary":"The paper claims that the 547-day periodic wobble in the radial velocities of the eclipsing binary BEBOP-3 is caused by a circumbinary giant planet, BEBOP-3b, with a mass of about $0.56\\ M_{\\rm Jup}$ on a moderately eccentric orbit ($e \\approx 0.25$). If correct, this is the first radial-velocity detection of a previously unknown circumbinary system, and a rare example of a circumbinary planet orbiting well beyond the stability pile-up region where most known cases sit. The claim matters because circumbinary planets are direct tests of how planets form and migrate in the disc around a binary, and the sample with measured masses is still small. The paper also derives model-independent dynamical masses for both stars in the binary and shows that an undetected inner planet near the stability limit could be hiding below the current detection threshold.","feed_headline":"547-day wobble reveals a giant planet orbiting a binary star","feed_subtitle":"Six years of radial-velocity data reveal a half-Jupiter planet around a pair of stars, a first for the method.","key_machinery":"The central mechanism is Doppler reflex motion: the planet pulls the binary's center of mass, producing an 11.8 m/s sinusoidal variation in the radial velocities of the primary star. The detection rests on the BINARIES model in the kima nested-sampling code, which fits the binary's Keplerian orbit (including apsidal precession and general relativity) and simultaneously treats the number of additional Keplerian signals as a free parameter, reporting Bayes factors between models. The False Inclusion Probability periodogram and the activity indices H$\\alpha$ and Na D serve as the filters that separate a planetary signal from stellar noise. High-resolution cross-correlation spectroscopy supplies the secondary's velocity semi-amplitude, converting the single-lined binary into an effectively double-lined one to give model-independent masses.","core_discovery":"From 141 high-resolution spectra obtained over six years, the authors model the binary orbit and search the residuals for additional Keplerian signals with a nested-sampling algorithm that lets the number of planets float. The evidence for one planet over none is overwhelming (Bayes factor $3.4 \\times 10^{13}$), with a signal at $P = 547^{+6.2}_{-7.6}$ days, semi-amplitude $K = 11.8 \\pm 1.1$ m/s, eccentricity $e = 0.247^{+0.077}_{-0.089}$, and a mass of $0.558^{+0.051}_{-0.048}\\ M_{\\rm Jup}$ if the orbit is coplanar with the binary. Because the chromospheric activity indices H$\\alpha$ and Na D show no significant periodicity near 547 days, they interpret the wobble as planetary. A second signal near 1400 days is reported only as a candidate. By applying high-resolution cross-correlation spectroscopy to the same spectra, they recover the secondary star's absorption at $5.5\\sigma$, obtaining dynamical masses of $1.083 \\pm 0.026\\ M_\\odot$ and $0.2615 \\pm 0.0039\\ M_\\odot$ for the primary and secondary. Stability integrations show stable orbits exist for a hypothetical inner planet near the boundary where the Kepler circumbinary planets are found.","pith_inferences":["If wide-orbit eccentric circumbinary giants like BEBOP-3b are common, the apparent pile-up near the stability limit is likely inflated by the transit method's bias toward short periods; RV surveys with long baselines should be able to test this by measuring the orbital-separation distribution of a larger sample.","The 11.8 m/s amplitude is only about twice the 6.7 m/s jitter, so the signal would be vulnerable to a slowly varying stellar activity cycle; computing line-bisector or full CCF shape indicators from the same spectra would provide an immediate, low-cost check.","The period ratio between the 1400-day candidate and BEBOP-3b is close to 5:2; if confirmed, resonant dynamics would link this system to the period-ratio structure seen among giant planets around single stars, and would favor migration rather than in-situ formation.","Eccentric circumbinary orbits beyond the stability limit are a natural outcome if the planet was scattered by a companion or if the circumbinary disc was eccentric; comparing the eccentricity--$a_{\\rm sc}$ distribution across all known circumbinary planets could distinguish these pathways."],"forward_implications":["BEBOP-3b becomes the first previously unknown circumbinary system found by radial velocities, showing that the method can discover planets in configurations transits would miss: long-period and eccentric.","The system's scaled semi-major axis $a_{\\rm sc} \\approx 4.85$ places it far beyond the pile-up region occupied by almost all other circumbinary planets, joining Kepler-1647 as an outlier that formation and migration models will need to reproduce.","The stability analysis implies that an inner planet of roughly ten Earth masses near the stability limit would be stable and would have escaped detection in the present data, so the apparent absence of such planets in RV surveys may be a sensitivity limit rather than a real scarcity.","The HRCCS result demonstrates that dynamical masses for both binary components can be extracted from a ground-based échelle spectrograph with 141 spectra, independent of stellar models.","If the ~1400-day candidate signal is confirmed with further data, BEBOP-3 would contain a second, lower-mass planet ($M \\sin i \\approx 0.2\\ M_{\\rm Jup}$), making it a multi-planet circumbinary system with widely separated orbits."],"supporting_citations":[{"why":"Supplies the stability-radius criterion used to compute BEBOP-3b's scaled semi-major axis and to interpret the stability map.","marker":"Holman & Wiegert (1999)"},{"why":"Established the kima-based radial-velocity detection and detection-limit methodology for circumbinary planets that this work extends.","marker":"Standing et al. (2023)"},{"why":"Provides the False Inclusion Probability periodogram that identifies the 547-day signal and the candidate 1400-day signal.","marker":"Hara et al. (2022)"},{"why":"Adapts HRCCS to high-contrast binaries and provides the recipe whose measured K2 yields the dynamical masses.","marker":"Sebastian et al. (2024b)"},{"why":"Supplies the kima BINARIES model with apsidal precession and general relativity used to fit the binary and search for planets.","marker":"Baycroft et al. (2023a)"},{"why":"Defines the eccentricity prior applied to the planetary Keplerian; the paper flags that this prior was designed for single-star planets.","marker":"Kipping (2013)"},{"why":"Documents the SOPHIE radial-velocity corrections (template/color, moonlight, drift) that give the median precision underpinning the detection.","marker":"Heidari et al. (2024)"}],"fun_headline_variants":["First RV-detected circumbinary planet: BEBOP-3b, a 0.56-Jupiter mass","Eccentric giant planet found orbiting a binary star: BEBOP-3b","Half-Jupiter-mass planet discovered around twin stars in 547 days","BEBOP-3b: a 547-day giant planet with an eccentric orbit around a binary"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The planet interpretation assumes that the 547-day wobble is not produced by a stellar activity signal that leaves no detectable periodicity in the H-alpha and Na D indices, which are the only activity tracers precise enough to use.","fun_headline_variants_meta":{"raw":{"variants":["First RV-detected circumbinary planet: BEBOP-3b, a 0.56-Jupiter mass","Eccentric giant planet found orbiting a binary star: BEBOP-3b","Half-Jupiter-mass planet discovered around twin stars in 547 days","BEBOP-3b: a 547-day giant planet with an eccentric orbit around a binary"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000667,"raw_usage":{"total_tokens":3129,"prompt_tokens":1115,"completion_tokens":2014,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":731,"completion_tokens_details":{"reasoning_tokens":1917}},"tokens_in":731,"tokens_out":2014,"duration_ms":15253,"temperature":1.0,"reasoning_tokens":1917,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:50:22.456010+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If a stellar activity tracer with a period near 547 days is found in the same spectra—e.g. in the Ca II H&K S-index, line bisectors, or the CCF full-width-at-half-maximum—the planetary interpretation fails. Alternatively, if a further three to five years of radial velocities do not preserve a coherent Keplerian phase at 547 days, with the period locked to better than a few days, the signal is not a planet.","supporting_citations":[],"review_version":2}