{"id":"e2bf8919-16d1-4c98-8288-7579258a51a6","arxiv_id":"2509.05905","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"About ten interstellar objects are likely currently bound in the solar system, concentrated in high-inclination, high-eccentricity orbits, with 25 candidate objects identified.","lead":"This paper uses simulations to show that interstellar objects can be gravitationally captured and stay bound in our solar system, estimating that about ten are bound today and flagging 25 candidate objects. The result matters because captured interstellar material could be studied repeatedly and up close, unlike the one-pass interlopers seen so far.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Survival-law tail beyond the 1.9 Myr integration contributes roughly a third of the predicted steady-state population; the ~10 bound-ISO estimate is not robust to this unvalidated extrapolation.","rationale":"The paper's central quantitative claim is the current bound-ISO population (~10), so the load-bearing quantity is the steady-state integral of capture rate times survival probability over solar-system history. The most fragile input is the survival law: a two-parameter Weibull fit to ejection times from a 1.9 Myr post-capture simulation (Fig. 5, Eq. 15). I quantified the tail contribution: at the end of the simulation t = 1.9 Myr = 10 t0, so S = exp(-10^0.46) = 0.056; because k = 0.46, the upper-incomplete-gamma tail beyond the simulation provides roughly 30% of the total integral. Thus about a third of the predicted ~10 objects is an extrapolation rather than a measurement. The longest simulated orbits (a up to 3500 au, periods ~2e5 yr; ISBOs to 9000 au) have completed fewer than one orbital period, so the tail is dynamically unconverged. This matches the reader's flagged weakest assumption, though I would refine it: the issue is not primarily that the heavy tail may overestimate stability or that early solar-system conditions differed; the steady-state integral is dominated by captures within the last few Myr, so early conditions are largely irrelevant. The issue is that the tail shape beyond 1.9 Myr is unconstrained, and a third of the predicted number depends on it. I considered whether the abstract/full-text capture-interval discrepancy (220 yr vs 48,500 yr) is more load-bearing. It is serious and must be fixed, but the full-text rate follows from a stated simulation setup (Hill-radius shell, Eq. 11), and the factor ~220 is close to (a_J/R_Hill)^2, suggesting the abstract inadvertently used Jupiter's semimajor axis as the cross-section. If so, it is an abstract error rather than a flaw in the full-text estimate. The tail extrapolation, by contrast, is a methodological gap that cannot be resolved by re-reading; it requires new computation. I also note Eq. 15's prefactor (2.79e-5 = 1393/50e6) is not the stated per-year capture rate, a ~35% normalization inconsistency that reinforces the need to re-derive the integral from raw data. The qualitative conclusions (captured ISOs concentrate near TJ=3 at high e and i, providing a phase-space discriminant, and only a few of the 18 candidates are expected to be interstellar) are supported by the simulation distributions and do not depend sensitively on the tail. Therefore the appropriate verdict remains CONDITIONAL: the framework and candidate lists are worth publishing, but the '~10' headline should be stated as provisional until the survival tail is validated and the rate normalization and abstract interval are corrected. Since the reader already reached CONDITIONAL, I do not move the verdict.","tokens_in":16792,"tokens_out":17238,"duration_ms":156397,"concrete_test":"Extend the post-capture REBOUND/REBOUNDx integrations of all 1393 captured single ISOs (at least the 78 survivors at 1.9 Myr) to 50 Myr, record all ejection times, and re-fit the survival law. Recompute the steady-state integral with the prefactor set to the independently derived capture rate from Eq. 11 (expected 1/48,500 yr^-1) and compare with the published prefactor 2.79e-5. If the tail beyond 1.9 Myr contributes less than 10% of the integral and the predicted current number remains 8-12, the headline stands; if the tail contribution or normalization shifts the number by more than a factor of ~2, the ~10 bound-ISO claim must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline number (~10 bound ISOs) is the capture rate (1/48,500 yr) multiplied by the time integral of the survival function fit to 1.9 Myr of post-capture integrations (Eq. 15, Fig. 5). The Weibull tail is not a negligible correction: at t = 1.9 Myr = 10 t0 (t0 = 1.9e5 yr, k = 0.46), the survival fraction is exp(-10^0.46) = 0.056, so 78 of 1393 captured objects are still bound when integration stops. For this Weibull, the upper-incomplete-gamma tail beyond 1.9 Myr contributes roughly 30% of the total integral, meaning about 3 of the predicted ~10 objects rest on extrapolation beyond the simulated time. The simulated interval is only about 9 orbital periods at the 2e5-yr periods shown in Fig. 6c, and less than one period for the a~3500 au (ISBO: 9000 au) tail, so the ejection statistics for long-lived objects are not converged. A power-law tail, S(t) ~ t^-alpha with alpha <~1, would make the 4.5 Gyr integral much larger or divergent, and the 1.9 Myr data cannot rule it out. The normalization is also internally inconsistent: the prefactor 2.79e-5 in Eq. 15 equals 1393/50e6, the per-particle capture probability, not the stated per-year rate 2.06e-5, a ~35% inflation. The abstract's 220-yr capture interval versus the full-text 48,500-yr interval adds an unresolved order-of-magnitude inconsistency. These issues do not invalidate the qualitative phase-space prediction (TJ<3, high e, high i), but they show the quantitative '~10' is not yet supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper combines a semi-analytic capture model with N-body simulations (REBOUND/REBOUNDx) to estimate the steady-state population of interstellar objects temporarily bound to the solar system. For single ISOs, 5e7 test particles are launched at Jupiter's Hill sphere; 1393 are captured, yielding a quoted rate of 1 bound ISO every 48,500 years. Post-capture survival is fit with a Weibull law and integrated over 4.5 Gyr to predict about 8 Jupiter-captured ISOs today, plus 2 from Saturn, concentrated near the Jupiter Tisserand boundary (TJ = 3) at high eccentricity and high inclination. For binary ISOs, the Hills mechanism is modeled with equal-mass circular binaries; the f = 1/50 case yields an estimated 3 currently bound objects. The predicted phase-space regions are compared with JPL small-body data, giving 18 single-ISO candidates and 7 ISBO candidates, with the caveat that only a few are expected to be genuinely interstellar.","tokens_in":17271,"tokens_out":18077,"duration_ms":166787,"significance":"If the quantitative estimate were fully supported, this would be a valuable prediction: a small, dynamically distinctive population of bound interstellar planetesimals, with candidate lists that can be followed up by surveys. The paper has real strengths: an independent semi-analytic capture check (1011 vs 1393), explicit inclusion of non-gravitational effects and the four giant planets, a public GitHub repository, and a candidate-selection procedure that does not tune the dynamical model to the observed candidates. The qualitative phase-space prediction (TJ < 3, high eccentricity, high inclination) is likely robust to the issues below. However, the headline number of ~10 currently bound objects is not reproducible from the text as written because of an abstract/full-text discrepancy, a normalization inconsistency in Eq. (15), an unvalidated survival-law extrapolation, and a missing derivation of the Monte Carlo rate factor. These issues are fixable with corrected derivations and either longer integrations or a more cautious statement of the prediction.","major_comments":[{"comment":"The abstract reports a mean capture interval of about 220 years, whereas §5.1 states that 1393 captures out of 5e7 simulated particles correspond to 1 bound ISO every 48,500 years. These numbers differ by more than two orders of magnitude, and the 48,500-year value is the one used for all population estimates; the abstract's interval is therefore not supported by the paper. Please correct the abstract and also state explicitly whether the '~10 currently bound ISOs' is the sum of the eight Jupiter-capture objects, the two Saturn objects, and the up-to-three ISBO objects, since these contributions are never summed in the text.","section":"Abstract and §5.1"},{"comment":"The prefactor 2.79e-5 in Eq. (15) is 1393 / 5e7, the fraction of simulated particles that are captured; it is not the per-year capture rate quoted in §5.1 (about 2.06e-5 yr^-1). The sentence 'This is the capture rate multiplied by the rate of ejection' is therefore dimensionally inconsistent, and integrating the stated P_iso(t) over 4.5 Gyr gives a population roughly 35% larger than integration with the quoted per-year rate. Please replace the prefactor with the physical rate, or explain exactly what quantity Eq. (15) is meant to represent, and recompute the inferred current populations (the values 8 and 9 in §5.1).","section":"§5.1, Eq. (15)"},{"comment":"The steady-state population estimate relies on a Weibull survival law fit to 1.9 Myr of post-capture integration. At the end of the simulated interval (t = 1.9e6 yr, about 10 t0), the survival fraction is exp[-(10)^0.46] about 0.056, so 78 of the 1393 captured particles are still bound; the fitted tail beyond 1.9 Myr contributes roughly one-third of the 4.5 Gyr integral. This tail is not validated by data, and the simulation duration is short relative to the orbital periods of the long-lived objects (periods up to about 2e5 yr, semi-major axes up to 3500 au, as reported in §5.1.1). Please extend the integrations for surviving particles, provide a sensitivity analysis to the Weibull tail (for example, versus a power-law tail), and report how much of the predicted ~10 comes from extrapolation beyond the simulated time.","section":"§5.1, Eq. (15) and Fig. 5"},{"comment":"The conversion from 1393 captured particles to a physical per-year rate is not shown explicitly. In Eq. (11), the geometric factor is written as 4 pi^2 r_max^2 n_iso times a Monte Carlo average over sampled velocities and directions. If the velocity directions are sampled uniformly over the inward hemisphere, the unbiased estimator for the flux integral in Eq. (10) requires an additional factor of 2 pi from the uniform hemisphere measure, because the cos(theta) factor is part of the integrand; as written, the rate would be low by a factor of two unless the sampling is cosine-weighted. Please provide the explicit unbiased estimator, including all factors of 2 pi and the sampling measure, or otherwise demonstrate that 1393 / 5e7 corresponds to 2.06e-5 yr^-1. This is load-bearing because the 48,500-year interval is a direct input to the final population estimate.","section":"§4.1, Eqs. (10)-(11)"}],"minor_comments":[{"comment":"Please use standard scientific notation (10^6, 3e6, etc.) in Table 1; as typeset, '106' and '3·106' are ambiguous and likely to be misread.","section":"Table 1"},{"comment":"The name '96P/Macholz' in Table 3 is spelled '96P/Machholz' in the text and in the cited literature; please make the spelling consistent.","section":"Table 3"},{"comment":"The sentence 'Many of these objects have not been studied well and represent excellent candidates for future surveys' would be more informative if accompanied by a quantitative statement of how many of the 18 candidates have measured physical properties (colors, activity, or albedo) and by a machine-readable table in the GitHub repository.","section":"§5.1.2"},{"comment":"The scaling discussion for the ISBO capture rate (C proportional to f^4 for high f) is terse; a short derivation connecting Eq. (12) to the collision and encounter probability would help the reader verify the claimed scaling.","section":"§5.2"},{"comment":"The maximal-dot-product assumption in Eq. (8) and the equal-mass circular-binary assumption are stated clearly, but the paper would benefit from a brief quantitative estimate of how much these assumptions could bias the ISBO capture rate in either direction, beyond the qualitative over- and underestimate discussion in §5.2.","section":"§3 and §4.2"}],"recommendation":"major_revision","confidential_remarks":"The abstract's 220-year interval and the normalization issues in Eqs. (11) and (15) should have been caught before submission, but they are fixable. I do not see grounds for rejection: the qualitative phase-space prediction and the candidate lists are valuable, and the survival-tail concern can be addressed either by more simulation or by softening the headline claim. The manuscript is within the scope of the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the qualitative prediction—bound ISOs should cluster near the Jupiter Tisserand boundary (TJ=3) at high eccentricity and modest-to-high inclination—is solid and genuinely useful. The Hills-mechanism treatment for binary ISOs is new, and the candidate lists come with appropriately cautious odds (~3 of 18 single, ~1 of 7 binary expected to be real). The paper deserves a serious referee, but the headline \"~10 bound ISOs\" needs work before it can be taken quantitatively.\n\nWhat's actually new: the long-term survival analysis that reshapes the captured phase-space distribution, the first quantitative look at interstellar binary capture via the Hills mechanism, and the inclusion of non-gravitational effects (radiation pressure, Poynting-Robertson drag, Yarkovsky, GR, outgassing) with code and parameters in the appendix. The N-body capture rate (1 per 48,500 yr) is in line with prior analytic estimates, and the semi-analytic cross-check (1011 vs 1393) is a nice sanity check.\n\nSoft spots, in order of severity:\n\n1. Internal inconsistency: the abstract says a mean capture interval of ~220 yr; the full text says 1 per 48,500 yr. That is a factor of ~220 and cannot stand as-is.\n\n2. Mis-normalized survival law: Eq. (15) uses 2.79e-5 as if it were the per-year capture rate, but that is actually 1393/5e7, the per-particle capture probability. The true per-year rate is 2.06e-5, so the integrated current population is inflated by ~35%. The paper's \"eight bound ISOs\" should be more like six before any other corrections.\n\n3. The Weibull tail is load-bearing: only 1.9 Myr of post-capture integration goes into the fit, yet about 3 of the ~8 predicted objects (after the normalization fix) come from the extrapolated tail beyond 1.9 Myr. The data cannot rule out a power-law tail with alpha <= 1, which would change the 4.5-Gyr integral substantially. The authors should either extend the integrations far enough to sample the tail, or present the steady-state number as a range with explicit sensitivity to the tail index.\n\n4. Minor: the ISO number density is uncertain by factors of 2-3 and enters linearly; the authors acknowledge this. The early solar system's different configuration could also change capture and survival, but that is an honest assumption, not a hidden flaw.\n\nThe qualitative phase-space diagram (Figure 8) is supported by the simulations and is the paper's strongest contribution. The candidate lists are useful even if most of the listed objects are probably not ISOs.\n\nVerdict: worth a serious referee, but the quantitative claims need correction or heavy caveating. I would suggest major revision rather than desk rejection.","headline":"Qualitatively interesting and worth refereeing, but the '~10 bound ISOs' headline is not yet supported: the abstract and full text disagree by a factor of ~220, the survival-law prefactor is mis-normalized, and the heavy-tailed Weibull reaches beyond what the 1.9 Myr integrations can justify.","tokens_in":17771,"tokens_out":5637,"would_cite":true,"duration_ms":46232,"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 roughly ten interstellar planetesimals are now gravitationally bound to the solar system, most captured by Jupiter's gravitational braking, and that 25 observed objects are candidate members of that population.","keywords":["interstellar objects","gravitational capture","Jupiter gravitational braking","Hills mechanism","Tisserand parameter","N-body simulations","long-term orbital evolution","planetesimal formation"],"falsifier":"Re-run the post-capture N-body integrations for at least 100 million years instead of 1.9 million years; if the ejection rate no longer follows the heavy-tailed Weibull law with $k\\approx0.46$ and nearly all captured objects are ejected within a few million years, the predicted $\\sim10$ bound ISOs would collapse. Observational check: a complete survey of the 25% highest-density contour near $T_J=3$ should find roughly two bound ISOs; finding zero after full sky coverage would falsify the model.","tokens_in":16594,"feed_emoji":"☄️","tokens_out":8601,"duration_ms":72413,"temperature":0.7,"pith_summary":"This paper tries to establish that interstellar objects (ISOs) are not only transient passers-through but can be gravitationally captured, and that a small population of them is bound to the solar system today. The authors build a semi-analytic model of the incoming ISO flux near Jupiter, pair it with N-body simulations of capture and long-term ejection, and extend the same logic to binary interstellar objects captured through the Sun's Hills mechanism. Their central estimate is that about ten ISOs are currently bound, most captured by Jupiter's gravitational braking, with a steady-state orbital distribution concentrated near the Jupiter Tisserand boundary $T_J=3$ at high eccentricity and high inclination. From that predicted phase-space density they flag 18 observed objects as candidate Jupiter-captured ISOs and seven as candidate Hills-captured binaries. The result matters because any confirmed bound ISO would be a repeatedly observable sample of planetesimal material from another planetary system.","feed_headline":"Around 10 interstellar objects may be orbiting the Sun today","feed_subtitle":"Jupiter's gravity is the main capture agent; 25 known objects are put forward as candidates.","key_machinery":"The load-bearing object is the gravitational slingshot at Jupiter, characterized by a two-body deflection angle $\\delta=2\\tan^{-1}(GM_J/(b v^2))$ that determines whether an incoming ISO leaves the encounter with negative heliocentric energy. The organizing variable for the resulting population is the Tisserand parameter with respect to Jupiter, $T_J\\approx a_J/a + 2\\sqrt{(a/a_J)(1-e^2)}\\cos i$, whose $T_J=3$ boundary marks the transition from planet-dominated to Sun-dominated dynamics; captured ISOs pile up along it. The conversion from capture events to a present-day census is carried by a Weibull survival fit $P(t)\\propto\\exp(-(t/t_0)^k)$ with $t_0=1.9\\times10^5\\,\\mathrm{yr}$ and $k=0.46$, applied to N-body integrations that include the other giant planets, radiation pressure, Poynting-Robertson drag, Yarkovsky drift, general-relativistic precession, and outgassing.","core_discovery":"On the paper's own terms, the discovery is that gravitational braking by Jupiter is a small but steady capture channel for interstellar planetesimals: the simulations produce one bound ISO roughly every $4.85\\times10^4$ years, and a Weibull survival law with $t_0=1.9\\times10^5$ years and $k=0.46$ describes how long those captures stay bound before ejection. Integrating this capture-and-survival balance over 4.5 billion years yields about eight bound ISOs from Jupiter and two from Saturn, with the Hills mechanism contributing fewer than about three, for a total near ten. The captured population clusters at $T_J\\approx3$ with semimajor axes beyond Jupiter's, eccentricities mostly $0.9$--$1$, and inclinations biased prograde at capture but reshaped toward high values by long-term survival. Comparing that phase-space region with the catalog of known solar system objects yields 18 single-ISO candidates and 7 binary-capture candidates, of which the paper expects roughly three and one, respectively, to be genuinely interstellar.","pith_inferences":["A testable extension the paper does not pursue: because the predicted bound population scales linearly with the assumed interstellar number density, counting bound ISOs in a complete survey would measure $n_\\mathrm{iso}$ independently of single-pass interloper detections.","The same capture physics should operate around massive planets in other planetary systems, so a $T_J\\approx3$ pileup of captured bodies might be imprinted on the orbital distribution of debris or ejected planetesimals around stars with close-in giants.","Orbital position alone cannot certify interstellar origin; if the 25 candidates display the volatile or isotopic anomalies seen in known interstellar interlopers, the dynamical assignments would be strengthened, while native Kuiper-belt-like compositions would argue that most are solar system objects."],"forward_implications":["If the central estimate holds, the solar system currently hosts a handful of bound extrasolar planetesimals, making them repeatable targets for orbit refinement and physical characterization rather than one-pass interlopers.","Bound ISOs should be searched for near $T_J=3$: high eccentricity, semimajor axes larger than Jupiter's, and inclinations biased to high values, a region of orbital space that known solar system populations populate only sparsely.","The model predicts that roughly three of the 18 flagged single-ISO candidates and one of the 7 binary candidates are genuinely interstellar; the strongest single candidate, 2024 XS15, sits in the 10% highest-density contour.","The prograde capture bias combined with high-inclination survival offers a dynamical explanation for part of the small prograde excess among Halley-type comets without invoking a large interstellar component.","Future wide-field surveys that cover the predicted phase-space region should find additional members, and the count would directly test the assumed interstellar number density because the bound population scales linearly with it."],"supporting_citations":[{"why":"Sets the interstellar number density $n_\\mathrm{iso}=0.1\\,\\mathrm{au}^{-3}$ used to convert capture probability into a rate.","marker":"W. Dehnen et al. 2021"},{"why":"Provides the log-normal interstellar velocity distribution used to initialize incoming ISOs and ISBOs.","marker":"T. M. Eubanks et al. 2021"},{"why":"Supplies the N-body integrator used for capture and long-term evolution.","marker":"H. Rein & S.-F. Liu 2012"},{"why":"Supplies the higher-order physical effects (radiation pressure, Poynting-Robertson drag, Yarkovsky, general relativity) added in the survival runs.","marker":"D. Tamayo et al. 2019"},{"why":"Defines the tidal-disruption mechanism by which a binary can leave one component bound to the Sun.","marker":"J. G. Hills 1988"},{"why":"Provides the Kuiper belt binary separation distribution used to model the ISBO population.","marker":"K. S. Noll et al. 2008"},{"why":"Provides the empirical outgassing model applied to all simulated particles.","marker":"T. Kramer & M. Läuter 2019"}],"fun_headline_variants":["Jupiter grabs an interstellar object every 220 years on average","Only about 10 captured interstellar objects still orbit the Sun","Captured interstellar objects linger in high-inclination orbits","25 known Solar System objects may be interstellar captures","Interstellar capture rate: 1 per 220 years, but survival is rare"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the solar system's capture conditions—the interstellar flux and the giant planets' orbits—have stayed roughly constant over 4.5 billion years, so that a survival curve measured from 1.9 million years of post-capture integration can be extrapolated to the present day.","fun_headline_variants_meta":{"raw":{"variants":["Jupiter grabs an interstellar object every 220 years on average","Only about 10 captured interstellar objects still orbit the Sun","Captured interstellar objects linger in high-inclination orbits","25 known Solar System objects may be interstellar captures","Interstellar capture rate: 1 per 220 years, but survival is rare"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000271,"raw_usage":{"total_tokens":1650,"prompt_tokens":989,"completion_tokens":661,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":605,"completion_tokens_details":{"reasoning_tokens":576}},"tokens_in":605,"tokens_out":661,"duration_ms":6232,"temperature":1.0,"reasoning_tokens":576,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:19:36.727069+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the post-capture N-body integrations for at least 100 million years instead of 1.9 million years; if the ejection rate no longer follows the heavy-tailed Weibull law with $k\\approx0.46$ and nearly all captured objects are ejected within a few million years, the predicted $\\sim10$ bound ISOs would collapse. Observational check: a complete survey of the 25% highest-density contour near $T_J=3$ should find roughly two bound ISOs; finding zero after full sky coverage would falsify the model.","supporting_citations":[{"cited_title":"2019, Astronomy & Astrophysics, 630, A4, doi: 10.1051/0004-6361/201935229","cited_arxiv_id":null,"evidence_quote":"Provides the empirical outgassing model applied to all simulated particles."}],"review_version":2}