{"id":"b56572d1-c002-4024-abca-daea2e75c226","arxiv_id":"2506.20889","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Photometric binaries in 35 open clusters show binary fraction and mass-ratio rising with cluster dynamical age, evidence that dynamical encounters reshape binary populations.","lead":"Using three sky surveys, the authors counted binary stars in 35 open star clusters and found that older, dynamically evolved clusters contain more binaries and more equal-mass binaries. The result suggests that gravitational encounters inside clusters reshape binary systems over time, which may explain why field stars often have similar companions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Distance-dependent 2MASS completeness can mimic the mass-ratio trends: dynamically old clusters are systematically more distant, and low-q binaries are hardest to recover at 2-4 kpc, so the reported q-shape and median-q correlations may be selection artifacts.","rationale":"The reader's weakest assumption is that the mass-limited main-sequence sample is equally complete across clusters spanning 500-4000 pc and a factor of about 100 in age. My analysis agrees and sharpens this concern: the mechanism is the differential availability of 2MASS photometry, which the paper itself shows is the largest contributor to low-q binary recovery. Table 1 reveals a strong distance-age collinearity in the sample, with old/dynamically old clusters concentrated at distances where 2MASS is incomplete for the quoted mass range. The q-shape and median-q trends are therefore the most fragile part of the central claim. The binary-fraction correlation is less threatened by this particular selection effect because missing low-q binaries would bias old clusters toward lower measured binary fractions, opposite to the observed sign, so I would not reject the paper on that basis. The paper contains genuine independent support: it reproduces known cluster parameters, validates metallicities against APOGEE, and includes a direct test of survey blending in Section 3.2. What is missing is a quantitative completeness correction or a distance-limited reanalysis. The conditional verdict is appropriate; the proposed injection-recovery test would settle whether the mass-ratio evolution is real or an artifact.","tokens_in":27767,"tokens_out":5218,"duration_ms":59669,"concrete_test":"Run an injection-recovery test with the same BASE-9 pipeline. For each of the 31 primary-sample clusters, take the observed member stars in the ML-MS range, inject synthetic binaries with known q in steps of ~0.05 from 0.5 to 1.0 and primary masses over 0.7-1.1 Msun, place them at the cluster's distance, reddening, and filter set, add realistic photometric scatter and magnitude limits for Gaia DR3, Pan-STARRS1, and 2MASS, and measure recovery probability as a function of distance, q, and primary mass. Then do either of the following: (a) use the completeness matrix to correct fb,q,M, the core binary fraction, and the q distributions, and recompute the correlations in Figures 6 and 8 and the young-versus-old K-S test in Figure 7; or (b) restrict the sample to clusters with distance < 1.5 kpc and redo the same correlations.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim bundles two correlations: binary fraction rising with dynamical age, and mass-ratio distributions shifting toward q=1 with dynamical age. The first correlation is not easily produced by photometric incompleteness, because missing faint low-q binaries would tend to lower the measured binary fraction in old clusters, opposite to the observed trend. The mass-ratio part of the claim is much more vulnerable.\n\nSection 3.2 shows that adding 2MASS photometry increases binary recovery, and that the gain is largest for low-q binaries. Yet 2MASS is far shallower than Gaia or Pan-STARRS. The sample spans 500-4000 pc (Section 2.1), and Table 1 shows that the dynamically old clusters (Age/trh > 15) are mostly at 1.5-4 kpc, while the dynamically young clusters (Age/trh < 2) are mostly at 0.6-1.2 kpc. At 3-4 kpc, a 0.7-1.1 Msun primary with q near 0.5 has a secondary near or below the 2MASS limit, so BASE-9 is deprived of the very bands that most help it recognize low-q binaries. The sample cuts in Section 3.3 (q > 0.5, M1 = 0.7-1.1 Msun, r < 3 rc) enforce a common analysis range but do not establish equal completeness across distance. No injection-recovery test or completeness map is presented anywhere in the paper.\n\nIf low-q binaries are selectively missed in old, more distant clusters, the observed q distributions will shift toward q=1 and the median q will increase with dynamical age, exactly as reported in Figures 7 and 8. The K-S and Anderson-Darling tests comparing young versus old q distributions would then partly or wholly reflect distance-dependent completeness. The authors explicitly note the 2MASS dependence in Section 3.2, and Section 4 acknowledges other caveats, but no quantitative check is made for this differential completeness.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript uses the BASE-9 Bayesian code on Gaia DR3, Pan-STARRS, and 2MASS photometry for 35 open clusters to identify photometric binaries, derive cluster parameters, and constrain binary mass ratios. After restricting the sample to main-sequence primaries of 0.7-1.1 solar masses, q > 0.5, and r < 3 core radii, the authors report that the binary fraction increases with cluster dynamical age (r = 0.5, and r = 0.6 for the core), that the mass-ratio distributions of dynamically young and old clusters are statistically distinct, and that the median mass ratio increases with dynamical age (r = 0.5, and r = 0.6 for the core). They interpret these trends as evidence that dynamical encounters, mass segregation, and exchange interactions increase both the binary fraction and the mass ratios of binaries in open clusters, and they connect the old-cluster and field q distributions.","tokens_in":28132,"tokens_out":5136,"duration_ms":53486,"significance":"If the reported trends are real, this would be an important observational result for our understanding of binary evolution in open clusters, providing one of the largest photometric samples to constrain how the mass-ratio distribution changes with dynamical age. The authors should be credited for combining 11 photometric bands, validating their BASE-9 parameters against APOGEE metallicities and literature values, comparing their binary fractions with previous work, and making their data publicly available on Zenodo. However, the central q-shape claim is vulnerable to a distance-dependent completeness artifact: Table 1 shows that dynamically old clusters tend to be more distant, while Section 3.2 shows that 2MASS photometry, which is the shallowest band set used, is especially important for recovering low-q binaries. No injection-recovery test or completeness map is presented. The binary-fraction correlation is less susceptible to this particular artifact, but its significance is also not fully established because uncertainties on the binary fractions and dynamical ages are not propagated into the correlation statistics.","major_comments":[{"comment":"The Pearson correlations in Figure 6 (r = 0.5, t = 3.4; core r = 0.6, t = 3.8) and Figure 8 (median q versus Age/trh, r = 0.5; core r = 0.6) are computed from point estimates without propagating the asymmetric uncertainties listed in Table 1. Binary fractions carry errors as large as about ±0.08 in some clusters, and the dynamical ages often have uncertainties of tens of percent (for example, NGC 2627 has Age/trh = 18.2 with +16.2/-16.2). Consequently, the reported t-statistics do not measure the statistical significance of these correlations; the authors should provide confidence intervals or p-values obtained by bootstrap or Monte Carlo propagation that incorporates the measurement uncertainties.","section":"Section 3.4, Figures 6 and 8"},{"comment":"The central mass-ratio result may be a distance-dependent selection artifact. Section 3.2 demonstrates that adding 2MASS photometry improves binary recovery and that the improvement is largest for low-q binaries, yet 2MASS is the shallowest of the three surveys. Table 1 shows a strong confound: clusters with Age/trh > 15 are mostly at 1.5-4 kpc (e.g., NGC 2506 at 3058 pc, Berkeley 32 at 3333 pc, Berkeley 39 at 4042 pc), while clusters with Age/trh < 2 are mostly at 0.6-1.2 kpc. At 3-4 kpc, a 0.7-1.1 solar mass primary with q near 0.5 has a secondary near or below the 2MASS limit, removing exactly the photometric information that most helps BASE-9 identify low-q binaries. The observed shift toward q near 1 in dynamically old clusters is precisely the signature this selection would produce. The manuscript does not include an injection-recovery test or completeness map, so this alternative explanation remains open and must be addressed before the dynamical interpretation can be accepted.","section":"Section 3.2, Table 1, Figures 7-8"},{"comment":"The K-S and Anderson-Darling tests compare seven clusters with Age/trh < 2 to nine clusters with Age/trh > 15, groups that were selected after inspecting the q distributions shown in Figure 12. Because the grouping is post-hoc and uses the same data used to define the hypothesis, the reported p-values (about 9e-12 and 0.001) are not valid significance levels. In addition, individual stars within a cluster are not independent draws, so tests that pool stars across clusters inflate the apparent significance. A cluster-level permutation test, or a pre-specified split with bootstrap resampling over clusters, is needed to support the claim that the q distributions differ with dynamical age.","section":"Section 3.5, Figure 7"},{"comment":"Several sample-definition choices are data-driven: the lower primary-mass cut of 0.7 solar masses is chosen where the observed fb,q(M1)/fb,q ratio is lowest, the upper cut of 1.1 solar masses is set by the turnoff of most clusters, the q > 0.5 cut follows from the claimed completeness limit, and the N >= 100 and r < 3rc cuts are motivated by the same dataset used for the trend analysis. This does not invalidate the trends by itself, but it means the reported Pearson coefficients should be tested for robustness to reasonable alternative cuts (for example, q > 0.6, M1 = 0.8-1.0 solar masses, or r < 2.5rc). Without such robustness checks, part of the observed correlation could reflect the optimization of the sample cuts rather than the underlying physical relation.","section":"Section 3.3"}],"minor_comments":[{"comment":"The phrase 'low-redenning' should be 'low-reddening'.","section":"Section 2.1"},{"comment":"The text refers to 'Coulombs constant' where the Coulomb logarithm is meant; also, the quantities N, M_Tot, and the three-dimensional core radius r_{c,3D} used in Eq. (2) are not fully defined in the text.","section":"Section 3.4, Eq. (2)"},{"comment":"The caption begins with 'T able' and should read 'Table'.","section":"Table 1 caption"},{"comment":"The phrase 'strong correlation' is used without reporting confidence intervals; adding the intervals from the propagated uncertainties would make the strength of the evidence easier to assess.","section":"Abstract and Section 4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and the public data release is commendable. The main risk is the distance-dependent 2MASS completeness confound for the mass-ratio trends; I would ask for injection-recovery tests or a distance-controlled subsample before reconsidering. The authors should also revisit the significance statements for the correlations, since the reported t-statistics and K-S p-values do not account for measurement uncertainties, cluster-level clustering, or post-hoc group selection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the scale and the q-distribution claim: 35 open clusters analyzed uniformly with BASE-9, showing binary fraction rising with dynamical age (r=0.5, t=3.4) and an even stronger core trend (r=0.6, t=3.8), plus a statistical distinction between q distributions of young and old clusters, with old clusters looking like the field. The binary-fraction trend was already in Donada et al. and Cordoni et al., but the q-shape transition is new and, if real, is the kind of anchor that links cluster dynamics to the field's twin excess. The paper is careful, the methods are mostly clear, and the authors are honest that the dynamical interpretation is an interpretation. They also ship the data on Zenodo and an interactive explorer, which should count for something.\n\nNow the soft spots, in proportion. The binary-fraction correlation itself is not easily killed by photometric incompleteness: missing faint low-q binaries in distant clusters would pull the measured fraction down, opposite to the observed rise. That part looks solid. But the q-shape and median-q correlations are much more exposed. The dynamically old clusters in Table 1 skew to 1.5-4 kpc; the young ones are mostly within 1.2 kpc. The paper itself shows that adding 2MASS boosts recovery most for low-q binaries, and 2MASS is shallow. At 3-4 kpc, a q≈0.5 secondary near the bottom of the 0.7-1.1 Msun primary range is right around the 2MASS limit, so the very data that helps you see low-q binaries is differentially missing in the old, distant clusters. That can produce exactly the shift toward q=1 and the rising median q they report. No injection-recovery test or completeness map is presented. That's the load-bearing gap.\n\nThe statistics also have unpolished edges: the reported r and t values do not propagate uncertainties on binary fraction or dynamical age; the sample cuts are data-driven with no stability check; the K-S test compares post-hoc extreme age groups; and the Dip statistic is sensitive to sample size, which is larger for the old, massive clusters. None of these are fatal on their own, but they weaken the quantitative claim.\n\nBottom line: this is a solid, useful paper that deserves a serious referee, but the referee should send it back for a completeness test (simulated binaries injected into the actual photometry, as a function of distance and q) and for uncertainty propagation on the correlations. I would bring it to reading group and would probably cite the binary-fraction sample, but I would not hang the field-twin link on the q-shape until the completeness question is answered.","headline":"A careful, data-rich extension to 35 clusters whose binary-fraction trend looks robust, but whose q-shape correlation with dynamical age is exposed to a distance-dependent completeness confound that needs a real test before the dynamical interpretation is accepted.","tokens_in":28744,"tokens_out":1535,"would_cite":true,"duration_ms":18451,"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":"Binary fractions and mass ratios in open clusters rise with dynamical age.","keywords":["open clusters","binary stars","mass ratio","dynamical age","relaxation time","photometric binaries","Bayesian inference","stellar dynamics"],"falsifier":"Inject synthetic binary populations with a fixed input binary fraction and mass-ratio distribution into the observed color-magnitude diagrams of these clusters across the full distance and age range, then run the same BASE-9 recovery; if the recovered binary fraction and mass-ratio distribution still rise with dynamical age, the trends are real, while if the trends vanish or shrink to the injected population, they are selection effects. A complementary check is a radial-velocity survey of the same clusters, which detects binaries independently of photometric mass ratio.","tokens_in":27524,"feed_emoji":"🔭","tokens_out":7483,"duration_ms":81043,"temperature":0.7,"pith_summary":"The paper aims to show that stellar dynamics, not just conditions at birth, reshape the binary populations of open clusters. Using the BASE-9 Bayesian fitting code on Gaia DR3, Pan-STARRS, and 2MASS photometry of 35 clusters, the authors identify unresolved binary stars and measure their mass ratios. They report that the binary fraction rises with cluster dynamical age—the number of half-mass relaxation times a cluster has lived—and that the trend is stronger in cluster cores. The mass-ratio distribution also changes: dynamically young clusters show multi-modal distributions, while dynamically old clusters show uniform distributions with a peak near equal masses, resembling field binaries. The authors interpret this as dynamical encounters and exchanges that destroy low-mass-ratio binaries and build up high-mass-ratio ones, with the caveat that the sample is small.","feed_headline":"Older open clusters hold more binaries, and more equal-mass ones","feed_subtitle":"If right, stellar encounters—not just birth conditions—set the binaries that later populate the galaxy.","key_machinery":"The load-bearing tool is BASE-9, a Bayesian code that fits PARSEC isochrones through all 11 photometric bands simultaneously and returns posterior probabilities for each star's membership, primary mass, secondary mass, and binarity. The comparison sample is then defined carefully: main-sequence stars with primary mass 0.7–1.1 solar masses (slightly lowered for the two oldest clusters), binaries with $q > 0.5$, and members within three King-model core radii, with at least 100 such stars per cluster (31 of 35 clusters). Dynamical age is the cluster age in units of the half-mass relaxation time, $t_{\\rm rh} = 0.346 N r_{c,3D}^{3/2}/\\sqrt{G M_{\\rm tot} \\ln \\Lambda}$, which quantifies how many relaxation timescales the cluster has experienced. The $q$-distribution shapes are diagnosed with the Hartigan Dip test for unimodality and compared with Kolmogorov–Smirnov and Anderson–Darling tests.","core_discovery":"On the authors' own terms, the central discovery is a statistical connection between how dynamically evolved an open cluster is and what its binary population looks like. Across 35 clusters, both the global binary fraction and the core binary fraction correlate with dynamical age (Pearson r = 0.5 and 0.6, respectively), and the median mass ratio of main-sequence binaries also rises with dynamical age (r = 0.5, or 0.6 in the core). The mass-ratio distributions of dynamically young clusters (less than about 2 half-mass relaxation times) and old clusters (greater than about 15) are statistically distinct: young clusters show multi-modal $q$ distributions rising toward $q = 1$ and toward the survey limit $q = 0.5$, while old clusters show a uniform distribution with a peak near $q = 1$ that matches field solar-type binaries. The paper interprets these patterns as evidence that dynamical encounters preferentially disrupt or evaporate low-$q$ binaries and that exchange encounters favor similar-mass companions, pushing binary populations toward higher mass ratios as clusters age.","pith_inferences":["A sharper test of the dynamical-processing interpretation would compare clusters of similar dynamical age but very different central density: if encounters drive the effect, denser clusters should show more extreme shifts in the mass-ratio distribution.","The photometric method cannot distinguish a true $q \\approx 1$ binary from a triple whose combined light mimics one; if hidden tertiaries are common, part of the twin peak in old clusters could be unidentified triples, a possibility the paper itself notes.","The same approach applied to globular clusters or to clusters closer than 500 pc could reveal whether the shift toward high $q$ saturates once clusters are dynamically old or continues to strengthen.","One could test whether the binary-fraction trend with dynamical age is driven by evaporation of low-mass single stars rather than by binary creation; measuring the mass functions of the same clusters should show that low-mass single stars are depleted in dynamically old systems."],"forward_implications":["If the trend is real, the binary fraction of an open cluster is not set at birth; it grows over relaxation timescales, most rapidly in the core.","The mass-ratio distribution of dynamically old clusters matching the field suggests many solar-type field binaries were dynamically processed inside clusters before dispersal.","Cluster mass matters: lower-mass clusters retain multi-modal $q$ distributions, while higher-mass clusters show unimodal distributions, meaning total mass acts as a proxy for how much dynamical processing has occurred.","The $q \\approx 1$ 'twin' peak seen in old clusters and in the field may be partly built by stellar exchanges rather than being purely primordial."],"supporting_citations":[{"why":"Supplies the starting catalog of open clusters and their members from which the 35-cluster sample is drawn.","marker":"(Hunt & Reffert 2023)"},{"why":"Paper I; establishes the BASE-9 membership, parameter, and binarity procedure that this work extends to 35 clusters.","marker":"(Childs et al. 2024)"},{"why":"Shows that BASE-9 recovers photometric binaries reliably for q > 0.5, motivating the mass-ratio cut used here.","marker":"(Cohen et al. 2020)"},{"why":"Introduces the Bayesian BASE-9 code that fits stellar evolution models to photometry and classifies stars as single or binary.","marker":"(von Hippel et al. 2006)"},{"why":"Provides the field solar-type binary mass-ratio distribution with a uniform shape and a q near 1 peak that old clusters are compared against.","marker":"(Raghavan et al. 2010)"},{"why":"Gives the hard-soft boundary and the binary disruption and exchange physics used to interpret the dynamical-age trends.","marker":"(Heggie 1975)"},{"why":"Argues that stellar exchanges in clusters tend to increase binary mass ratios over time.","marker":"(Marks et al. 2011)"},{"why":"Shows that encounters and exchange dynamics in clusters tend to raise binary mass ratios, supporting the paper's interpretation.","marker":"(Geller et al. 2012)"}],"fun_headline_variants":["Stellar encounters skew open cluster binaries to equal masses","Dynamically old clusters hold more equal-mass binaries","Cluster dynamics push binary pairs toward equal mass","Open cluster binaries shift with stellar dynamics","Age of cluster shapes its binary mass ratios"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the mass-limited main-sequence sample (primary masses 0.7–1.1 solar masses, mass ratios above 0.5, within three core radii) is equally complete in all 35 clusters; if older or more distant clusters systematically miss faint, low-mass-ratio binaries, the correlations with dynamical age could be selection artifacts rather than dynamical processing.","fun_headline_variants_meta":{"raw":{"variants":["Stellar encounters skew open cluster binaries to equal masses","Dynamically old clusters hold more equal-mass binaries","Cluster dynamics push binary pairs toward equal mass","Open cluster binaries shift with stellar dynamics","Age of cluster shapes its binary mass ratios"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000547,"raw_usage":{"total_tokens":2661,"prompt_tokens":1039,"completion_tokens":1622,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":655,"completion_tokens_details":{"reasoning_tokens":1553}},"tokens_in":655,"tokens_out":1622,"duration_ms":14909,"temperature":1.0,"reasoning_tokens":1553,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:40:03.518650+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Inject synthetic binary populations with a fixed input binary fraction and mass-ratio distribution into the observed color-magnitude diagrams of these clusters across the full distance and age range, then run the same BASE-9 recovery; if the recovered binary fraction and mass-ratio distribution still rise with dynamical age, the trends are real, while if the trends vanish or shrink to the injected population, they are selection effects. A complementary check is a radial-velocity survey of the same clusters, which detects binaries independently of photometric mass ratio.","supporting_citations":[{"cited_title":"M., Hurley, J","cited_arxiv_id":null,"evidence_quote":"Shows that encounters and exchange dynamics in clusters tend to raise binary mass ratios, supporting the paper's interpretation."}],"review_version":1}