{"id":"6aa9c676-7fb4-4a96-90e1-83d8a4fd6173","arxiv_id":"2412.00975","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Stellar collisions in galactic nuclei can deflect stars into the supermassive black hole's tidal radius or eject them at speeds up to the hypervelocity regime, producing a small but observable population of TDEs and runaway stars.","lead":"A simulation of stars orbiting the Milky Way's supermassive black hole shows that direct stellar collisions can kick stars onto paths that end in tidal disruption or in ejection from the cluster. The work offers a new way to explain unusual tidal disruption events and hypervelocity stars, though the predicted rates are uncertain.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative rates hinge on the untested deflection law Eq. 5; a hydro calibration that changes the deflection or dissipation by order unity could shift collision-TDE and ejection rates by orders of magnitude, although the qualitative plausibility claim likely survives.","rationale":"I agree with the reader's weakest assumption: Eq. 5 is the least secure element of the argument. The paper is transparent about its proof-of-concept nature and explicitly acknowledges the need for hydro calibration, which strengthens credibility. The qualitative mechanism—collisions change orbital energy and angular momentum—is physically sound and follows from momentum conservation (for mergers) and scattering (for high-speed collisions). Even a weaker deflection would still generate some TDEs and ejections, so the central 'plausible outcomes' claim is likely robust. However, the Section 5 rates and the hypervelocity tail are quantitative statements that depend on a hand-built interpolation and a dissipation model with no impact-parameter dependence. The paper's own 10%/50% speed scalings test only one axis of uncertainty; the other axis, the deflection angle, is untested. The small event counts add to the uncertainty but are secondary to the missing collision-physics calibration. Therefore the reader's CONDITIONAL verdict is appropriate: the model is a useful proof of concept, but the quantitative rates should be considered provisional until Eq. 5 is calibrated against hydrodynamics.","tokens_in":22438,"tokens_out":3949,"duration_ms":36304,"concrete_test":"Run smoothed-particle-hydrodynamics (SPH) simulations of 1 Msun solar-type collisions across a grid of relative speeds (e.g., 100–3000 km/s) and impact parameters b/rc = 0, 0.2, 0.4, 0.6, 0.8, 1.0. For each run, measure the post-collision velocities and masses of the two interaction products, extract an effective deflection angle and speed change as functions of b and v_rel, and replace Eq. 5 and the uniform dissipation scalings in the semi-analytic model with these calibrated functions. Recompute the collision-TDE and ejection rates in the fiducial Milky Way cluster. If the rates stay within a factor of a few of the quoted values, the quantitative claims are supported; if they change by an order of magnitude or more, the rates in Section 5 are not yet supported and the paper's conclusions remain qualitative.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.2.2 introduces Eq. 5, theta_coll = 2 (b/rc) arctan(b90/b), as an interpolation between hyperbolic deflection at grazing incidence and zero deflection at b=0. The paper states this is a proof-of-concept choice that 'can be tuned by hydro simulations.' All high-speed collision orbital changes—TDEs placed on radial orbits and stars ejected to unbound orbits—are generated by this deflection combined with an impact-parameter-independent speed reduction (0%, 10%, or 50%). The quantitative results in Section 5 (collision-TDE rate 10^-8 to 10^-7 per galaxy per year, unbound fractions ~0.3–1%) assume this specific functional form. If real stellar collisions deflect more weakly at intermediate b, or dissipate energy in an impact-parameter-dependent way (e.g., stronger dissipation for head-on hits), the number of stars placed on TDE orbits and the hypervelocity tail could drop dramatically. The paper's own sensitivity test only varies a uniform multiplicative speed scale; it does not vary the deflection law itself. With only 0–3 TDEs per 10,000-star run, Poisson noise is large, but the dominant unknown is the collision physics, not the counting statistics. The ejection mechanism may still operate because even a small deflection near periapsis can unbind a star, so the qualitative conclusion of plausible outcomes is likely robust; however, the rate estimates and the claim that collision-ejected stars can reach v_inf > 1000 km/s are conditional on a deflection law that has not been calibrated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proof-of-concept paper presents a semi-analytic Monte Carlo model of 1 Msun stars in a fixed Milky Way nuclear cluster, following direct stellar collisions and two-body relaxation over 10 Gyr. Collision locations and relative velocities are sampled from the cluster properties; post-collision orbits are computed either with a sticky-sphere merger or with the collision deflection law of Eq. (5) plus three dissipation prescriptions (0%, 10%, and 50% speed reduction). The authors find that collisions place stars on nearly radial orbits, producing TDEs of stripped or merged stars, and that high-speed collisions near periapsis can eject stars with speeds up to >1000 km/s. They estimate a collision-TDE rate of 10^-8 to 10^-7 per galaxy per year and argue that collision-affected TDEs may be a factor of 3 more common, though still subdominant to the overall TDE rate.","tokens_in":22763,"tokens_out":6952,"duration_ms":62046,"significance":"If the channel is real, the paper identifies a physically motivated way to produce TDEs of unusual stars (stripped or merged) and a periapsis-collision mechanism for hypervelocity stars, both of which are testable with LSST and Galactic-center surveys. The strengths of the manuscript are its transparent forward-modeling setup, explicit stopping conditions (β = 0.5 and E≥0), systematic variation of mass-loss prescriptions, cusp slope, dissipation, and eccentricity distributions, and its honest statement that rates are uncertain and future SPH calibration is needed. The qualitative plausibility of collision-induced orbital changes is well supported by the simple mechanics of the model; the quantitative rates, however, rest on an uncalibrated deflection law and on small event counts, so the numerical rate estimates should be treated as illustrative until the collision physics is better constrained.","major_comments":[{"comment":"The deflection law θ_coll = 2(b/r_c) arctan(b_90/b) is an ad hoc interpolation between the hyperbolic-grazing limit and zero head-on deflection, as the authors state. All quantitative results in Section 5 (collision-TDE rate, unbound fractions, and the >1000 km/s tail) depend on this specific functional form, which is not calibrated against hydrodynamic simulations. The sensitivity tests vary only a uniform multiplicative speed reduction (0%, 10%, 50%), not the deflection law itself. If real collisions deflect more weakly at intermediate impact parameters, or dissipate energy in an impact-parameter-dependent way, the collision-TDE rate and the hypervelocity ejection fraction could change by orders of magnitude. I ask the authors to either calibrate Eq. (5) to existing SPH results, test robust alternative deflection prescriptions, or explicitly demote the Section 5 rates to illustrative order-of-magnitude estimates that are not central conclusions.","section":"Section 3.2.2, Eq. (5)"},{"comment":"The quantitative rate estimates are built from very small event counts: the twenty fiducial 10,000-star runs yield between 0 and 3 TDEs each (10 TDEs total in Section 4.1), and the factor-of-3 enhancement for collision-affected TDEs in Section 5 is derived from a single 4,000-star run with no reported confidence interval. The paper should propagate Poisson uncertainties into the rate quoted as 10^-8 to 10^-7 per galaxy per year and clearly state the statistical error on the factor-of-3 claim, or label both as order-of-magnitude estimates rather than rates. This is needed because the claimed rate is already a derived quantity built from scaling small sample counts to the full cluster.","section":"Section 4.1 and Section 5"}],"minor_comments":[{"comment":"The line styles for α=1.25 and α=1.75 are described inconsistently: Section 3.1 calls α=1.25 dashed and α=1.75 solid, while the Figure 1 caption reverses these assignments.","section":"Section 3.1 vs. Figure 1 caption"},{"comment":"The text 'orbital orbital parameters' contains a duplicated word; similar typos elsewhere include 'the the' in Section 4.1, 'negligable' in Section 3.2.2, 'sample of sample stars' in Section 6, and 'conveivably' in Figure 2's caption.","section":"Section 3.3"},{"comment":"The sentence 'The maximum speeds at infinity suggest that high speed collisions may represent another mechanism to launch hypervelocity stars' is a reasonable qualitative statement, but the histogram in Figure 4 would be more informative if it included the effect of the stellar-cluster potential on v_inf, as the paper itself notes that only the SMBH potential is used.","section":"Section 4.3"},{"comment":"The statement that 'even if only 1% of those events get detailed follow-up, we may still observe one to a few of these collision-induced TDEs with LSST' conflates detection with spectroscopic classification; the fraction of LSST TDEs with the follow-up needed to identify a stripped or merged stellar progenitor will likely be much smaller than 1%, so the optimistic number should be presented as an upper bound.","section":"Section 5"}],"recommendation":"major_revision","confidential_remarks":"The paper is a legitimate proof-of-concept and the qualitative mechanism is likely correct, but the quantitative rate claims in Section 5 need either calibration of Eq. (5), robustness tests against alternative deflection laws, or an explicit downgrade to illustrative estimates. The self-citations to the authors' own prior code papers are appropriate given the method lineage. The manuscript is within scope for an astrophysics journal; I would not reject it, but it is not ready for acceptance in its current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a transparent, well-scoped proof-of-concept that stellar collisions can both drop stars onto TDE orbits and eject stars at hypervelocity speeds. The qualitative claim holds up. The quantitative rates are conditional on a deflection law that has not been calibrated, so treat them as order-of-magnitude until hydro simulations weigh in.\n\nWhat is actually new: prior work (MacLeod et al. 2012; Yu & Tremaine 2003) touched collisions in TDE contexts and HVS ejection, but not the orbital evolution of collision products under both merger and stripping prescriptions. The authors sample collision location and collider velocity along each star's orbit, update the SMBH orbit via momentum conservation for mergers or via a deflection angle plus uniform speed scaling for high-speed collisions, and test two mass-loss recipes and three dissipation scalings. That is a sensible semi-analytic framework. The paper is also honest about its ad hoc pieces.\n\nThe findings that collision-TDEs are mostly high-speed, low-mass-loss events, and that merger TDEs require anti-aligned angular momenta, are interesting. The connection to stripped stars explaining the Miller et al. (2023) N/C ratio is speculative but plausible.\n\nThe soft spots are real but not fatal. Equation (5), the deflection law, is a hand-built interpolation between zero and hyperbolic deflection. The authors note it can be tuned by hydro simulations, but they do not vary it in the sensitivity tests. Dissipation is applied as a uniform 10% or 50% speed reduction, not an impact-parameter-dependent prescription. With 0-3 TDEs per 10,000-star run, Poisson noise is substantial. No code or data are released, which limits reproducibility. That said, the ejection mechanism is robust: a small deflection near periapsis can raise a star's energy past zero, so the qualitative conclusion likely survives calibration. The 50% dissipation run, where stars sink inward and get disrupted at higher rates, is a nice sanity check.\n\nWho this is for: anyone working on TDE rates, stellar collisions in galactic nuclei, or hypervelocity star production. It is a proof-of-concept, so the final rates are not the deliverable; the channel is now clearly defined and the missing input (hydrocalibrated deflection and dissipation) is well identified. I would send it to peer review. The referees should ask for a sensitivity run that varies Eq. (5) itself, and ideally a public version of the code.","headline":"A clear proof-of-concept that collisions can place stars on TDE orbits and eject hypervelocity stars, with rates that are order-of-magnitude until the deflection law is calibrated.","tokens_in":23295,"tokens_out":2820,"would_cite":true,"duration_ms":26040,"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":"Stellar collisions in galactic nuclei can place stars on nearly radial orbits that end in tidal disruption, and high-speed collisions near periapsis can eject stars from the cluster at hypervelocity speeds.","keywords":["stellar collisions","tidal disruption events","hypervelocity stars","nuclear star clusters","supermassive black holes","stellar dynamics","semi-analytic models","stellar mergers"],"falsifier":"Hydrodynamical simulations of overlapping main-sequence star collisions at relative speeds of hundreds to thousands of km/s that measure the post-collision deflection angle as a function of impact parameter would directly test the deflection law Eq. (5); a measured deflection consistently below $2(b/r_c)\\arctan(b_{90}/b)$ at large impact parameters would lower the expected TDE and ejection rates, and measured dissipation stronger than the 10–50% scalings would shrink the hypervelocity population.","tokens_in":22197,"feed_emoji":"💥","tokens_out":5657,"duration_ms":43167,"temperature":0.7,"pith_summary":"This paper argues that direct collisions between stars in the dense cluster around a supermassive black hole can significantly reshape stellar orbits, sending stars into the black hole's tidal radius or ejecting them from the cluster. Using a semi-analytic model of a Milky Way-like nuclear star cluster, it finds that collisions can produce tidal disruption events involving unusual stars—recent merger products or stars stripped of their outer layers. It also finds that high-speed collisions near periapsis can unbind stars, in some cases to hypervelocity speeds. The authors estimate a collision-driven TDE rate of $10^{-8}$ to $10^{-7}$ per galaxy per year, roughly 100–1000 times below the overall TDE rate, and an ejection fraction near one percent of the inner-parsec population. If correct, collisions form a genuine, though subdominant, channel for TDEs and a new route to hypervelocity stars.","feed_headline":"Colliding stars can feed black holes and launch hypervelocity stars","feed_subtitle":"Semi-analytic models of galactic nuclei show collisions are a genuine, if subdominant, source of flares and ejections.","key_machinery":"The load-bearing object is the collision deflection law $\\theta_{\\rm coll} = 2 (b/r_c) \\arctan(b_{90}/b)$, where $b$ is impact parameter, $r_c$ the sum of stellar radii, and $b_{90} = G(M_\\odot+M_\\star)/v_{\\rm rel}^2$ the impact parameter for a 90-degree deflection. It interpolates between the full hyperbolic deflection at grazing incidence ($b\\approx r_c$) and zero deflection for a head-on hit, and it is the mechanism by which collisions transfer orbital energy and angular momentum to the star about the supermassive black hole. Around it, the model combines a collision probability drawn from the local density and velocity dispersion, mass-loss and merger criteria from Rauch (1999) and Lai et al. (1993), sticky-sphere momentum conservation for mergers, and an optional two-body relaxation kick prescription.","core_discovery":"The central claim is that stellar collisions in galactic nuclei are a plausible direct cause of both tidal disruption events and stellar ejections, including hypervelocity stars. Collisions act in two regimes: low-speed mergers that conserve angular momentum and shrink orbits, and high-speed impacts that deflect stars via a modified hyperbolic encounter, transferring orbital energy between stars. The paper shows that collision-induced TDEs are preferentially produced by high-speed collisions with large impact parameters near apoapsis, which deflect stars onto nearly radial orbits while shedding little mass, and that high-speed collisions near periapsis preferentially unbind stars. Dissipation during collisions reduces but does not eliminate the unbound population.","pith_inferences":["If the deflection law is confirmed by hydrodynamics, collisions would join the Hills mechanism as a physical route to hypervelocity stars; the two channels could be distinguished by the presence or absence of a bound companion and by the stellar properties of the ejected star.","The paper's rate estimate implies that LSST could catch a few collision-induced TDEs among tens of thousands of detections; the most testable signature may be abundance anomalies rather than the overall rate.","In more massive galactic nuclei with shorter collision timescales, the collision channel could rival relaxation-driven TDEs even as the total TDE rate declines, making TDE host-galaxy mass a testable diagnostic.","The model treats collisions as instantaneous impulses; tracking post-collision radius inflation would change the effective tidal radius and could raise the predicted TDE rate."],"forward_implications":["Collision-induced TDEs should occur at a rate of $10^{-8}$ to $10^{-7}$ per galaxy per year, and TDEs of stars that were previously stripped or merged by collisions should be about a factor of 3 more common than that.","A fraction of observed TDEs may involve unusual victims—recent merger products or stripped stars—whose spectra could show anomalous abundances such as high nitrogen-to-carbon ratios.","High-speed collisions near periapsis can eject roughly 0.5–1% of stars in the inner parsec, with speeds at infinity typically 100–600 km/s and occasional stars above 1000 km/s.","Dissipation during collisions cuts the ejected fraction roughly in half for a 10% speed loss, but ejections persist even under a 50% speed loss.","Some collision-produced TDEs have orbital periods short enough to be observed as repeating partial disruptions within about 30 years."],"supporting_citations":[{"why":"Supplies the Rauch99 fitting formulae for mass loss and merger conditions used to set collision outcomes.","marker":"Rauch 1999"},{"why":"Supplies the Lai+93 fitting formulae for mass loss and merger capture radius used as the alternative collision prescription.","marker":"Lai et al. 1993"},{"why":"Provides the hyperbolic deflection formula, Maxwellian velocity sampling, and relaxation timescale used in the orbital update.","marker":"Binney & Tremaine 2008"},{"why":"Establishes the prior link between destructive stellar collisions and TDE rates that this paper extends to orbital deflection.","marker":"MacLeod et al. 2012"},{"why":"Provides the statistical collision treatment, mass and age updates, and merger/stripped-star outcome prescription on which the model builds.","marker":"Rose et al. 2023"},{"why":"Shows that high-speed collisions produce stripped stars and distinguishes imminent versus eventual TDE signatures for collision products.","marker":"Gibson et al. 2024"},{"why":"Defines the canonical hypervelocity star mechanism whose speed range the collision ejection channel is compared against.","marker":"Hills 1988"},{"why":"Treats close encounters as a route to hypervelocity stars and supplies the comparison ejection-rate context.","marker":"Yu & Tremaine 2003"},{"why":"Provides SPH-derived collision outcomes and nuclear cluster collision-rate context used to frame the semi-analytic prescriptions.","marker":"Freitag & Benz 2005"},{"why":"Supplies the eccentricity-dependent functions f1(e) and f2(e) used in the collision timescale formula.","marker":"Rose et al. 2020"}],"fun_headline_variants":["Stellar collisions trigger tidal disruptions and hypervelocity ejections","Colliding stars near black holes spawn flares and runaway stars","Galactic nuclei: collisions produce tidal disruptions and escaped stars","Stellar smashups feed black holes and launch hypervelocity stars","Collision-driven tidal disruptions and ejections in galactic nuclei"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes a hand-built rule for how much two overlapping stars bend each other's paths, and if real collisions bend paths less, or dissipate more energy than the scalings tested, the predicted rates of black-hole plunges and ejections would shrink.","fun_headline_variants_meta":{"raw":{"variants":["Stellar collisions trigger tidal disruptions and hypervelocity ejections","Colliding stars near black holes spawn flares and runaway stars","Galactic nuclei: collisions produce tidal disruptions and escaped stars","Stellar smashups feed black holes and launch hypervelocity stars","Collision-driven tidal disruptions and ejections in galactic nuclei"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000483,"raw_usage":{"total_tokens":2377,"prompt_tokens":927,"completion_tokens":1450,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":1382}},"tokens_in":543,"tokens_out":1450,"duration_ms":9251,"temperature":1.0,"reasoning_tokens":1382,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T04:50:48.010754+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Hydrodynamical simulations of overlapping main-sequence star collisions at relative speeds of hundreds to thousands of km/s that measure the post-collision deflection angle as a function of impact parameter would directly test the deflection law Eq. (5); a measured deflection consistently below $2(b/r_c)\\arctan(b_{90}/b)$ at large impact parameters would lower the expected TDE and ejection rates, and measured dissipation stronger than the 10–50% scalings would shrink the hypervelocity population.","supporting_citations":[],"review_version":1}