{"id":"9c90b85f-87fb-4727-bd79-f6615f11b490","arxiv_id":"1908.02659","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Ballistic trajectories from the L2 point of a binary can escape, fall back, hit the star, or self-intersect, and escape occurs even for sub-corotation launches at certain offsets.","lead":"This paper maps the fates of gas particles launched near the L2 point of a binary star system under different launch positions and speeds. It finds that even particles moving slower than the binary's rotation can escape, and that small changes in launch conditions switch outcomes between escape, fallback, stellar collision, and shock.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sub-corotation escape claim rests on an inconsistent inner boundary: Fig. 7 keeps the f=1 Roche lobe for collisions while moving L2 for f<1; the abstract's signature result may be an artifact.","rationale":"The reader and I identify the same weakest point: the f<1 runs keep the corotating f=1 equipotential as collision boundary. I treat it as more consequential than the reader does, because it targets the abstract's new sub-corotation claim rather than just a quantitative overestimate. If the corrected boundary removes the Ef>0 regions in Fig. 7, the headline claim is unsupported; if it merely shrinks them, the claim survives. The authors' own caveat in Section 3.3 makes this possibility concrete. I also note the printed Eq. (2) appears to have a dimensional inconsistency (f x rather than f^2 x), which should be resolved in the same re-run. Other concerns (q=0.792 vs 0.78 discrepancy, lack of released grid data) are real but do not bear on the existence of sub-corotation escape. Because the paper is otherwise careful, with tight tolerances and clearly described methods, a conditional acceptance with one targeted recomputation is proportionate.","tokens_in":10543,"tokens_out":8890,"duration_ms":99296,"concrete_test":"Recompute Fig. 7 for f=0.95, 0.9, 0.8 and q=0.05, 0.5, 0.95 with the self-consistent potential Phi_f = -mu/r1 - (1-mu)/r2 - (f^2/2)(x^2+y^2), an L2_f defined by f^2 x - mu/(x-1+mu)^2 - (1-mu)/(x+mu)^2 = 0, and the collision boundary as the critical equipotential through L2_f. Keep all other integration settings identical. Measure the signed area of initial conditions with Ef>0 outside the collision region. If at least one contiguous escape region survives, the headline claim stands; if none survives, the sub-corotation escape regions in Fig. 7 are artifacts of the f=1 boundary. As a secondary check, rerun the same set using the printed Eq. (2) versus the f^2 form to quantify sensitivity to the Lagrangian-point location.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 3.3's f<1 runs, which carry the abstract's 'even for initial velocities slower than corotation' claim, combine f-dependent launch kinematics with an f=1 collision boundary. The inner boundary used to terminate trajectories is the original corotating equipotential through L2 (Section 2.1), while Eq. (2) puts the f<1 L2 farther out. Launching near the original L2 therefore means starting closer to the star in the modified problem, and the true critical surface is the f<1 Roche lobe, not the f=1 lobe. The authors concede in Section 3.3 that using the modified equipotential 'would increase the number of colliding trajectories.' This moves the concern from caveat to load-bearing: if the reclassified trajectories include the Ef>0 regions in Fig. 7, the central sub-corotation discovery disappears entirely. A compounding issue is that Eq. (2) as printed has f x rather than f^2 x for a frequency ratio f=omega'/omega, so the marked modified L2 positions may themselves be inaccurate. The central claim is therefore not yet verified in the regime where it is new.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies ballistic mass loss from the vicinity of the second Lagrange point L2 in a circular binary, using the restricted three-body problem. For test particles launched with various position offsets and velocities relative to L2, the authors integrate the equations of motion and classify outcomes into unbound outflow (Ef > 0), fallback leading to a decretion disk (Ef < 0), collision with the binary, and self-intersecting/shocked trajectories. They map these outcomes as functions of initial offset, radial velocity, and initial angular frequency ratio f = ω′/ω. The central new claim, stated in the abstract, is that even for initial velocities slower than corotation (f < 1), a set of initial position offsets leads to unbound outflows. The paper also presents a smoothed particle hydrodynamics (SPH) illustration of a self-intersecting stream that produces a shock. The results are intended as a reference for interpreting hydrodynamic simulations and observations of L2 mass loss in merging binaries, common envelope evolution, and related transients.","tokens_in":10732,"tokens_out":5015,"duration_ms":54679,"significance":"If the results hold, this paper provides a useful and previously missing systematic map of ballistic outcomes near L2, going beyond the point-like, corotating injection assumption of earlier work (e.g., Shu et al. 1979). The demonstration that outcomes are sensitive to small offsets and velocity deviations is valuable for interpreting hydrodynamic simulations, as the authors connect to MacLeod et al. (2018b). A clear strength is the direct numerical integration of the standard restricted three-body equations with tight tolerances (10^-12), giving parameter-free outcome maps. The SPH shock illustration is a nice complement, though it is explicitly indicative rather than definitive. The main caveat is that the abstract's headline sub-corotation escape claim rests on the f < 1 runs in Fig. 7, whose inner collision boundary is not self-consistent with the modified potential; this needs to be resolved before the claim can be considered established.","major_comments":[{"comment":"The f < 1 runs, which carry the abstract's claim that unbound outflow occurs for initial velocities slower than corotation, use the original f = 1 corotating equipotential as the inner collision boundary while Eq. (2) places the modified L2 point farther out for f < 1. The authors concede in Section 3.3 that 'taking into account the modified equipotential would increase the number of colliding trajectories.' This is a load-bearing inconsistency: trajectories currently classified as escaping (Ef > 0) may instead collide with the binary if the self-consistent f < 1 equipotential is used. Please re-run the f < 1 calculations with the modified equipotential as the inner boundary (or otherwise quantify how many of the Ef > 0 regions survive) before claiming that sub-corotation unbound outflow exists.","section":"Section 3.3, Fig. 7"},{"comment":"Equation (2) as printed reads f x_L2 − μ/(x_L2 − 1 + μ)^2 − (1 − μ)/(x_L2 + μ)^2 = 0. For a particle on a circular orbit with angular frequency ω′ = fω, the radial force balance in units where ω = 1 should have the centrifugal term f^2 x, not f x, because the required centripetal acceleration is ω′^2 x. Consequently, the 'modified position of L2' marked in Fig. 7 is not the correct equilibrium point for the stated frequency ratio. Please correct Eq. (2), recompute the marked L2 positions, and check whether the interpretation of the f < 1 results is affected.","section":"Section 2.1, Eq. (2)"}],"minor_comments":[{"comment":"The unresolved discrepancy with Shu et al. (1979) on the bound/unbound dividing line (q = 0.792 versus 0.78) deserves a brief discussion of possible causes, such as the stopping distance at 200a, the definition of final energy, or the inner boundary treatment.","section":"Section 3.1, footnote"},{"comment":"The inner equipotential crossing is checked only every 0.25/ω time units; please comment on whether this temporal discretization could miss a fast collision and slightly alter the hatched collision regions in Figs. 3, 5, 6, and 7.","section":"Section 2.1, termination condition"},{"comment":"The phrase 'a only a small perturbation' contains a grammatical error; it should be 'only a small perturbation.'","section":"Introduction, first paragraph"},{"comment":"The reference to 'Fig. 5 of MacLeod et al. (2018b)' is vague; please specify what is shown there (e.g., the stream density or velocity field) so the reader can follow the comparison.","section":"Section 4, MacLeod comparison"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid numerical study with clean, reproducible-looking methods, and the outcome maps will be useful to the community. My main concern is that the abstract's headline claim (sub-corotation escape) rests on Fig. 7, which uses an inner boundary that is inconsistent with the modified potential for f < 1; the authors themselves flag this as likely to increase collisions. The Eq. (2) f-versus-f^2 issue compounds the concern. Both are fixable within the scope of the paper, so I recommend major revision rather than rejection. I would also encourage the authors to soften the abstract claim until the f < 1 result is verified with a consistent boundary."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThis is a paper you should know about: Hubova & Pejcha extend the classic Shu et al. (1979) ballistic treatment of L2 mass loss to include positional offsets and non-corotating initial velocities. They map out four outcomes—unbound outflow, fallback into a decretion disk, collision with the binary, and self-intersecting loops—as a function of mass ratio and initial conditions. The f=1 and f>1 parts are new and solid; the collision and loop outcomes are genuinely new classes that were absent from the corotation-only treatment. Figure 3 alone will be a useful reference for anyone interpreting hydrodynamic simulations of binary mass loss.\n\nThe headline claim—that even sub-corotation launch (f<1) can produce unbound outflow—is not yet verified. The f<1 runs in Figure 7 use the original corotating equipotential as the inner collision boundary, while Eq. (2) moves L2 outward. The authors acknowledge this in Section 3.3: the modified equipotential would increase the number of colliding trajectories. That is not a minor caveat. The abstract presents the sub-corotation result without this qualification, and if a significant fraction of the red Ef>0 regions in Figure 7 are reclassified as collisions once the inner boundary is made self-consistent, the central claim disappears entirely. A referee should ask for a recomputation, or at least a quantitative estimate of how many trajectories change outcome.\n\nThere is also a possible typo in Eq. (2): with f defined as omega'/omega, the centrifugal term should be proportional to f^2, but the printed equation has f. If that formula was used to mark the modified L2 positions in Figure 7, those markers are wrong. This is easy to fix but worth flagging.\n\nThe paper is otherwise honest and well executed. The integration method is standard with tight tolerances, and the description is detailed enough for re-implementation. The q=0.792 vs 0.78 discrepancy with Shu et al. is minor and probably numerical, but it would be nice to see it resolved.\n\nBottom line: this deserves a serious referee. The f>=1 results are publishable as they stand. The sub-corotation claim needs either a self-consistent treatment of the inner boundary or a downgrade to a suggestive, not conclusive, result. I would accept with major revision.","headline":"A useful extension of Shu et al. with a headline sub-corotation result that is not yet self-consistent and needs a closer look.","tokens_in":11306,"tokens_out":8368,"would_cite":true,"duration_ms":85944,"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":"Even streams launched slower than corotation can still escape a binary if they start with the right offset from L2.","keywords":["close binaries","mass loss","L2 point","Roche potential","ballistic trajectories","tidal torques","luminous red novae","decretion disk"],"falsifier":"Compute the f<1 maps with the collision boundary replaced by the self-consistent equipotential through the shifted L2 point; if the positive-final-energy regions shrink or vanish, the sub-corotation escape claim fails as stated. Alternatively, run a hydrodynamic simulation starting a stream at f=0.95 with the specific offset that the ballistic calculation says escapes, and check whether the gas becomes unbound despite being slower than corotation.","tokens_in":10295,"feed_emoji":"💫","tokens_out":9913,"duration_ms":91410,"temperature":0.7,"pith_summary":"The paper sets out to show that the fate of gas leaving a binary star through the outer Lagrange point L2 is controlled by a narrow window of initial conditions, not just by the binary mass ratio. Integrating ballistic test-particle trajectories launched with small spatial offsets and velocity offsets from L2, the authors map four outcomes: unbound outflow, fallback into a decretion disk, collision with the binary surface, and self-intersecting loops that produce shocks. The central new result is that even trajectories starting roughly 20 percent slower than corotation can still become unbound, provided the launch position is offset in the right way. This matters because L2 mass loss is tied to luminous red novae and common-envelope transients, and the resulting maps give hydrodynamic simulations a concrete reference for what final morphologies to expect.","feed_headline":"Slow streams can still break free of a binary at L2","feed_subtitle":"Outcome maps show escape, fallback, and collision depend sharply on where the outflow launches.","key_machinery":"The central machinery is the restricted three-body ballistic trajectory integration in the frame corotating with the binary. The key quantity is the final specific energy $E_f$: positive means unbound outflow, negative means fallback into a decretion disk, with collisions and self-intersections detected geometrically. Initial conditions are parameterized by the offset ($\\Delta x, \\Delta y$) from L2, by the radial speed $v_0$, or by the angular-frequency ratio $f=\\omega'/\\omega$ relative to the binary; for $f\\neq 1$ the paper recomputes the effective L2 position from Eq. (2) using the standard modified Lagrange point for non-corotating orbits. The energy gain $\\Delta E$ is set by tidal torquing, whose sign and strength depend on where the particle sits relative to the binary, and the maps of $E_f$ are the paper's main output.","core_discovery":"On the paper's own terms, the discovery is that the asymptotic outcome of ballistic L2 mass loss is a sensitive function of initial position and velocity. For a binary with mass ratio $q$, launching a particle exactly at L2 in corotation reproduces the classic $q$-dependent division between unbound streams and decretion disks, but any offset in position ($\\Delta x, \\Delta y$) or any deviation of the initial velocity from corotation opens up new outcomes, including collision with the star and self-intersecting trajectories. The most striking new finding is a thin arc of initial offsets toward the side the binary rotates into ($\\Delta y > 0$) where tidal torquing is strong enough to give positive final energy even when the initial angular frequency is only $f=0.8$ of corotation. The paper further shows that the efficiency of tidal energy gain, $\\Delta E$, varies sharply on scales comparable to the stellar size, which explains why small shifts in launch conditions flip the trajectory type; and it demonstrates with smoothed-particle hydrodynamics that the self-intersecting case produces a hot, shocked inner arc near the binary.","pith_inferences":["Because the outcome boundaries in the maps are only a few hundredths of the orbital separation wide, real stellar surface activity such as convection or magnetic spots could push a stream across a boundary on short timescales, making the observed outflow type effectively stochastic even at fixed binary parameters.","The authors note self-intersections and possible deterministic chaos near other Lagrange points; a finite-width stream would therefore mix trajectory types, so real outflows may look like superpositions of the clean map categories rather than a single one.","A concrete test: repeat the $f<1$ maps with the collision boundary replaced by the self-consistent modified equipotential; the authors themselves expect more collisions, so the true sub-corotation escape regions are probably smaller than plotted.","Extending the ballistic model with gas pressure and radiative cooling would likely blur the sharp outcome boundaries, so hydrodynamic simulations at finite resolution may miss the thin escape arc even where it exists in the ballistic limit."],"forward_implications":["Small shifts in the launch point near L2 can switch a stream between colliding with the binary, forming a decretion disk, and escaping, so simulations of binary mass loss must treat initial conditions as a range rather than a single corotating point.","Even streams launched slower than corotation (down to about $f=0.8$) can still become unbound for some offsets, so a sub-corotation launch velocity does not by itself imply fallback.","The newly identified self-intersecting trajectories deposit low-angular-momentum, shock-heated gas close to the binary, which resembles decretion-disk formation but with hotter material nearer the stars.","For the pre-merger binary V1309 Sco, the maps imply that decreasing $f$ toward merger raises the fraction of initial conditions that collide with the binary, naturally explaining the rising initial temperature of the L2 stream inferred from observations.","The outcome maps provide a reference for interpreting hydrodynamic simulations: a simulated stream that remains bound may be reflecting its particular initial velocity rather than a universal rule for L2 mass loss."],"supporting_citations":[{"why":"Provides the baseline corotating-launch result that for mass ratios q between 0.064 and 0.78 tidal torqueing unbinds the L2 stream, which this paper extends by adding offsets and velocities.","marker":"Shu et al. (1979)"},{"why":"The hydrodynamic simulation that found a sub-corotation L2 stream remains bound and forms a decretion disk, motivating the need for velocity-dependent outcome maps.","marker":"MacLeod et al. (2018b)"},{"why":"Supplies the V1309 Sco inference that the L2 stream temperature rises and f approaches 0.95 near merger, used to interpret the f<1 results.","marker":"Pejcha et al. (2017)"},{"why":"Gives the modified Lagrange point locations for non-corotating orbits, used to define the effective L2 position for f different from 1.","marker":"Sepinsky et al. (2007)"},{"why":"Establishes that tidal torques transfer energy from the binary orbit to the L2 outflow, the mechanism behind the energy gain Delta E.","marker":"Kuiper (1941)"},{"why":"Provides earlier hydrodynamic outcome classifications that the ballistic trajectory types are compared with, and the basis of the smoothed-particle-hydrodynamics calculation used for the shock case.","marker":"Pejcha et al. (2016a,b)"}],"fun_headline_variants":["Tiny offsets at L2 decide binary mass loss fate","L2 launch offsets unlock escape at sub-corotation speeds","Where you leave L2 matters: escape or collide","Even slow L2 streams can escape if launched with offset"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that a particle collides with the star when it crosses the equipotential surface of the corotating binary, even for cases where the launch point is based on a slower rotation rate; if that surface is wrong for those cases, some trajectories counted as escaping would actually hit the star.","fun_headline_variants_meta":{"raw":{"variants":["Tiny offsets at L2 decide binary mass loss fate","L2 launch offsets unlock escape at sub-corotation speeds","Where you leave L2 matters: escape or collide","Even slow L2 streams can escape if launched with offset"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000182,"raw_usage":{"total_tokens":1312,"prompt_tokens":950,"completion_tokens":362,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":566,"completion_tokens_details":{"reasoning_tokens":296}},"tokens_in":566,"tokens_out":362,"duration_ms":3994,"temperature":1.0,"reasoning_tokens":296,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:38:05.075075+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the f<1 maps with the collision boundary replaced by the self-consistent equipotential through the shifted L2 point; if the positive-final-energy regions shrink or vanish, the sub-corotation escape claim fails as stated. Alternatively, run a hydrodynamic simulation starting a stream at f=0.95 with the specific offset that the ballistic calculation says escapes, and check whether the gas becomes unbound despite being slower than corotation.","supporting_citations":[],"review_version":1}