{"id":"dfe49a2f-ac7d-407d-9469-4da0d8b46a2e","arxiv_id":"2412.05054","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Radiation-pressure-driven dust outflows remove dust too slowly to explain the scarcity of non-accreting transition discs, except in special EUV wind cases with atypically large photon fluxes.","lead":"This paper uses 2D simulations to test whether radiation pressure can blow dust out of protoplanetary discs fast enough to explain why so few observed discs have large dust-free cavities. It finds that in most cases the dust loss is 5 to 10 times too slow, so radiation pressure does not solve the puzzle for typical models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Wind prescription bias is one-sided: corrections would likely lower dust mass-loss, strengthening the negative result","rationale":"The reader correctly identifies the simplified wind prescription as the main numerical limitation, but the direction of the bias is favorable to the paper's central claim. The authors' caveats and the trap's location behind the inner rim mean that a corrected 2D wind could reduce dust mass-loss rates, which would make radiation pressure even less effective. The parameter study already varies the wind model, strength, stellar mass, fragmentation velocity, and turbulence, and only the EUV case with atypically high flux succeeds. The conclusion that radiation pressure cannot solve the relic disc problem in most models is therefore robust to the identified uncertainty. I do not find a load-bearing concern that could reverse the verdict; the CONDITIONAL verdict is appropriate given the quantitative uncertainties, but the qualitative result is well-supported.","tokens_in":33082,"tokens_out":10585,"duration_ms":112982,"concrete_test":"Reproduce the fiducial X-ray O11 evolutionary calculation (Section 5.1) using a full 2D gas velocity field from a hydrodynamical photoevaporation simulation (e.g., Picogna et al. 2019 or Sellek et al. 2024) instead of the prescribed vertical velocity of Eq. 5. Track the dust-trap mass and radial dust mass-flux at 1 Myr. If the trap mass still increases and the mass-loss rate stays below the required ~1e-6 M_Earth/yr, the central claim is robust to the wind treatment.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption is the purely vertical wind prescription, v_g,Z = Sigma_wind / rho_g (Section 3.2.1). If this overestimates vertical drag at the inner rim, dust mass-loss rates would be overestimated, making the central negative result (Sections 5.2 and 8) conservative, not fragile. The authors state two caveats: (1) using hydrostatic gas density in the wind region changes the velocity but not the mass flux, so the drag force on entrained grains is unchanged; (2) the wind is launched radially inward before turning vertical, which is neglected and would push dust inward, further reducing outflow. The dust-trap is situated several au behind the inner rim, where the vertical assumption is more accurate. Thus the identified simplification biases the simulation in favor of dust removal, and a more realistic wind treatment would likely yield even lower mass-loss rates, reinforcing the conclusion that radiation pressure cannot solve the relic disc problem within current photoevaporation models. The only regime where the conclusion could reverse is if a full 2D wind solution produced unexpectedly stronger vertical drag or a smoother, more efficient profile; however, no known physical mechanism supports this, and the authors' parameter study already tests three different wind models with no success except for atypically large EUV fluxes.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates whether radiation-pressure-driven dust outflows can resolve the 'relic disc' problem in photoevaporating protoplanetary discs, in which models predict a population of non-accreting, large-cavity discs that is rarely observed. Using the cuDisc code, the authors couple 2D (R,Z) dust dynamics, coagulation/fragmentation, radiative transfer, and parameterised internal photoevaporation, evolving discs through the primordial phase in 1D and then through outer-disc dispersal in 2D. They measure dust mass-loss rates from the inner dust trap and run evolutionary simulations over ~1 Myr for three wind models (X-ray O11, X-ray P19, EUV), two stellar masses, and several fragmentation velocities and ionising fluxes. The central finding is that, in most of parameter space, radiation-pressure-driven outflows remove dust at rates roughly 5-10 times too small to deplete the dust trap against refilling, so radiation pressure does not solve the relic disc problem within current photoevaporation models; only smooth, EUV-like wind profiles with atypically high ionising fluxes and low dust masses at gap opening clear the trap on the required timescale. The paper also finds that the photosphere height is controlled by levitating micron-sized grains in the outflow rather than by inner-rim shadowing, and it presents synthetic SEDs and scattered-light images.","tokens_in":33350,"tokens_out":19939,"duration_ms":199498,"significance":"If the result holds, it closes a proposed solution to the relic disc problem and redirects attention to other dispersal mechanisms or to processes that reduce the dust mass at gap opening. The study is significant for its scope: it is, to my knowledge, the first global 2D treatment of radiation-pressure dust outflows in photoevaporating discs that includes coagulation/fragmentation and radiative transfer, with a parameter study spanning three wind models and a range of stellar and dust parameters. The public availability of cuDisc, the explicit discussion of approximation caveats, and the synthetic observations tied to a real non-accreting transition disc are concrete strengths. Importantly, the main simplifications (purely vertical wind, removal at the wind base, and use of extinction opacity for radiation pressure) all tend to overestimate dust removal, so the central negative result is conservative rather than fragile.","major_comments":[],"minor_comments":[{"comment":"The purely vertical wind prescription and the use of hydrostatic gas density in the wind region may overestimate vertical drag on dust near the inner rim, and the neglected radially inward launch of the wind would also reduce outward dust transport; because both effects bias the calculation in favour of dust removal, the central negative result is conservative. I recommend stating this one-sidedness explicitly in §8 and in the abstract rather than only in the §4 caveat.","section":"§3.2.1, Eq. (5)"},{"comment":"The radiation-pressure force per unit volume uses the extinction opacity κ_ext; for dust with anisotropic scattering the momentum-transfer efficiency should be κ_abs + (1-g)κ_sca. Using κ_ext overestimates the force, again in the conservative direction for the main conclusion, but the manuscript should either use the proper combination or add a sentence explaining why the difference is unimportant.","section":"§3.2.3, Eq. (6)"},{"comment":"The stellar radii are listed as '1.7 R_*' and '1.1 R_*'; these should be solar radii (R_sun) to avoid ambiguity, and the units in the table should be made consistent.","section":"Table 1"},{"comment":"Please clarify whether the 2D gas density, assumed hydrostatic, is recomputed as the dust distribution and temperature evolve during the 2D phase; if it is fixed at the initial condition, state this and comment on the expected effect on the vertical drag in Eq. (5).","section":"§3.2"},{"comment":"For the successful low-flux EUV case (Φ=10^43 s−1) the disc age at gap opening is 12 Myr, at the high end of observed disc lifetimes; this should be flagged more prominently when presenting this case as satisfying the observational constraints.","section":"§5.2.3"},{"comment":"There is a typo in the description of the solver: 'Godonuv' should be 'Godunov'.","section":"§3.2"}],"recommendation":"minor_revision","confidential_remarks":"The paper is well within the scope of the journal and the central negative result is robust to the main approximations, because those approximations bias the calculation toward dust removal. The remaining issues are local: sharpening the conservative-bias statements, fixing the radiation-pressure opacity convention or justifying it, and a few presentation points. I would not require new simulations for acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nYou should know this paper before the next reading group: it is the most complete attempt yet to test the idea that radiation pressure can disperse the dusty component of photoevaporating transition discs, and the test fails. Across a wide parameter study with three photoevaporation models, two stellar masses, varied fragmentation velocities and opacities, dust mass-loss rates are 5 to 10 times too small to empty the dust trap before it is refilled by radial drift and by the outward motion of the disc's inner rim. The only successful cases are EUV winds with ionising fluxes around 1e43-5e43 photons/s, which are atypical and force disc lifetimes of 6-12 Myr. So radiation pressure likely does not solve the relic disc problem.\n\nThe paper earns its conclusion. It moves the problem from the 1D local calculation of Owen & Kollmeier (2019) to full 2D simulations with dust growth, fragmentation and radiative transfer in cuDisc. That step matters: the 2D treatment reveals that the dust trap is continuously resupplied, something the 1D model could not see. The parameter study is broad, the authors compare three wind models, and they are transparent about where approximations enter. The 'levitating' micron-sized grain enhancement and the finding that small grains in the outflow set the photosphere height are interesting side results with consequences for scattered-light interpretation.\n\nThe soft spots are real but not load-bearing. The wind is treated as a purely vertical flow with the velocity set by mass conservation. This neglects the radially inward launch at the inner rim and, if anything, overestimates vertical drag on dust there. The dust trap is a few au behind the rim, where the vertical assumption is more accurate. So the central negative result is conservative: a full 2D gas flow would likely give even lower dust removal, not higher. The EUV success case is propped up by unusually high photon fluxes and long lifetimes, so it should not be read as a rescue. The authors also discard simulations where the dust-to-gas ratio exceeds unity at the rim, but they give a reasonable argument that those cases would not clear faster.\n\nThis paper deserves a serious referee. It is a solid, reproducible step forward for the photoevaporation and transition disc community. I would send it to review; even if some quantitative details shift with a better wind model, the qualitative verdict against radiation pressure as the dispersal mechanism looks robust.\n\nYours,\n\n[Name]","headline":"The definitive 2D test shows radiation pressure does not solve the relic disc problem, and the negative result is robust and likely conservative.","tokens_in":33889,"tokens_out":4584,"would_cite":true,"duration_ms":45691,"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":"Radiation pressure cannot solve the relic disc problem: dust outflows remove material 5–10 times too slowly.","keywords":["protoplanetary discs","photoevaporation","radiation pressure","dust growth and fragmentation","transition discs","relic discs","radiative transfer","dust mass loss"],"falsifier":"Run the same dust-growth and radiative-transfer calculation with a full 2D gas velocity solution for the photoevaporative wind; if micron-sized grains at the inner rim then cross the wind base at rates above roughly $10^{-6}\\,M_\\oplus$ yr$^{-1}$, the paper's central negative result is overturned. Alternatively, a survey finding non-accreting transition discs with X-ray winds whose mid-IR excess fades within a few $10^5$ yr would contradict the predicted clearing timescales.","tokens_in":32891,"feed_emoji":"🌪️","tokens_out":9264,"duration_ms":81889,"temperature":0.7,"pith_summary":"Photoevaporation models predict that many discs should linger as non-accreting transition discs with large dusty cavities, but such 'relic discs' are rarely observed. This paper asks whether radiation pressure from the central star can disperse the remaining dust quickly enough to resolve that mismatch, using global 2D simulations that track dust growth, fragmentation, radiative transfer, and photoevaporative winds. The simulations find that dust mass-loss rates from the inner rim of the outer disc are typically 5–10 times too small to keep pace with refilling of the dust trap, because grains larger than about a micron are launched outward but fall back into the disc. Only discs with smooth, EUV-like wind profiles, gas mass-loss rates above about $5\\times10^{-9}\\,M_\\odot\\,\\mathrm{yr}^{-1}$, and less than about $1\\,M_\\oplus$ of dust at gap opening clear within the observed timescales. The paper concludes that radiation pressure is not the missing dispersal mechanism in current photoevaporation models.","feed_headline":"Radiation pressure disperses disc dust 5–10 times too slowly","feed_subtitle":"A full 2D simulation shows launched grains fall back and the trap refills; only extreme EUV winds match observed clearing.","key_machinery":"The load-bearing element is the balance of forces on a dust grain just above the disc photosphere: radiation pressure acts radially outward, gravity acts radially inward, and the grain's centrifugal force alone cannot lift its trajectory above a flared disc surface. Escape therefore requires vertical drag from the photoevaporative wind, which the simulations prescribe through a vertical gas velocity $v_{g,Z}=\\dot{\\Sigma}_{\\rm wind}/\\rho_g$. The code tracks a dust size distribution from $0.1\\,\\mu$m to $50$ cm, computes the radiation force per grain size via wavelength-dependent opacities and a 2D optical depth to the star, and removes any dust that crosses the wind base, defined by a neutral hydrogen column of $10^{21}\\,\\mathrm{cm}^{-2}$. The crucial diagnostic is the angle of the grain trajectory relative to the photosphere; micron-sized grains have trajectories that point back into the disc interior, so the outflow is geometrically unable to carry them away. A related mechanism is the pile-up of 'levitating' micron-sized grains at heights where wind drag balances settling, which sets the photosphere height and controls how much dust is exposed to direct stellar radiation.","core_discovery":"The paper's central claim is that radiation-pressure-driven dust outflows remove dust from photoevaporating transition discs too slowly, by factors of 5–10, to prevent the inner edge of the outer disc from remaining optically thick to mid-infrared radiation for longer than observed. The simulations show that sub-micron grains are accelerated along trajectories that leave the disc, but micron-sized grains—which carry much of the outflow mass—are pushed radially outward and then re-enter the disc because their trajectories lie below the photosphere; the radial mass flux consequently falls by at least an order of magnitude between the dust trap and twice its radius. The trap is replenished by inward radial drift and by the outward sweeping of dust as photoevaporation pushes the gas inner rim outward, so in the fiducial models the dust-trap mass increases over 1 Myr after gap opening. The exception is an EUV wind with a smooth mass-loss profile and an unusually high photon flux, which leads to low dust masses at gap opening and a low photosphere that lets larger grains escape; in those runs the near-to-mid-IR excess fades within about 0.25–0.8 Myr. The authors conclude that the relic disc problem is not solved by radiation pressure, and that solutions must instead come from reducing the dust mass at gap opening or from revising the photoevaporation models themselves.","pith_inferences":["Because dust mass-loss scales sub-linearly with dust surface density, discs that have already lost mass to planetesimal formation should show faster mid-IR fading after gap opening; this is testable with surveys that compare transition-disc lifetimes to initial dust masses.","The result that photosphere height is set by levitating outflow grains rather than shadowing implies that scattered-light scale-height measurements may overestimate vertical mixing; comparing scattered-light scale heights with gas scale heights from CO observations in photoevaporating discs would test this.","The vertical-drag prescription is likely generous to dust escape, so a full 2D gas velocity solution that includes the inward radial component of the wind at the inner rim would probably strengthen the negative conclusion rather than reverse it.","For discs evolving under MHD winds or external photoevaporation, where the gas outer radius shrinks and inward dust flux increases, radiation-pressure clearing should be even less efficient than in the purely viscous discs modelled here."],"forward_implications":["In the X-ray O11 and P19 photoevaporation models, radiation pressure does not reduce the dust-trap mass; the trap mass typically increases over 1 Myr after gap opening, so the discs remain mid-IR bright for too long.","The only successful cases are EUV winds with smooth mass-loss profiles and unusually high photon fluxes, where dust disperses within about 0.25–0.8 Myr; these require gas mass-loss rates above about $5\\times10^{-9}\\,M_\\odot$ yr$^{-1}$ and dust masses below about $1\\,M_\\oplus$ at gap opening.","Radiation pressure competes with wind entrainment: it pushes dust outward and reduces the mass flux crossing the wind base at the inner rim, so including radiation pressure can lower, not raise, the dust lost to the wind.","The photosphere in photoevaporating discs is shaped by outflow grains rather than by shadowing from a hot inner rim, so the disc appears flared in scattered light even when the gas scale height is lower.","If current X-ray wind models overestimate gas mass-loss rates, as suggested by more recent hydrodynamic models, the computed dust mass-loss rates are upper limits and the negative conclusion becomes stronger."],"supporting_citations":[{"why":"Proposed radiation-pressure-driven outflows as a solution to the relic disc problem; this paper extends that calculation from a local 1D model to global 2D simulations with dust growth and fragmentation.","marker":"Owen & Kollmeier (2019)"},{"why":"Supplies the X-ray photoevaporation wind profiles and the relic disc population that motivates the study; used as the X-ray O11 model.","marker":"Owen et al. (2011b)"},{"why":"Supplies the alternative X-ray wind model (P19) with variable column density used for comparison.","marker":"Picogna et al. (2019)"},{"why":"Provides the EUV photoevaporation wind model used in the simulations that meet observational constraints.","marker":"Font et al. (2004)"},{"why":"Provides the EUV profile implementation and shows the build-up of dust-to-gas ratio at the inner rim, a feature the paper reproduces.","marker":"Alexander & Armitage (2007)"},{"why":"Supplies the two-population dust evolution model and the fragmentation/drift limits used for the 1D primordial phase and initial grain-size distributions.","marker":"Birnstiel et al. (2012)"},{"why":"Establishes the vertical advection of dust by winds and the result that radiation pressure can reduce dust entrainment, which the paper confirms.","marker":"Booth & Clarke (2021)"},{"why":"Justifies treating dust that reaches the wind base as lost, since grains delivered there are generally removed from the system.","marker":"Hutchison & Clarke (2021)"},{"why":"Describes cuDisc, the code used to evolve dust dynamics, growth/fragmentation and radiative transfer in 2D.","marker":"Robinson et al. (2024)"},{"why":"Recent X-ray photoevaporation models with additional cooling imply lower gas mass-loss rates, used to argue the computed dust loss rates are likely upper limits.","marker":"Sellek et al. (2024)"}],"fun_headline_variants":["Radiation pressure fails to clear disc dust fast enough","Disc dust dispersal 5-10x too slow via radiation pressure","Radiation pressure can't solve relic disc conundrum","Only extreme EUV winds match disc clearing observations","Dust outflows too weak to explain fast disc clearing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the photoevaporative wind can be treated as a purely vertical gas flow with the vertical velocity set by mass conservation, which exaggerates the upward drag on dust at the inner rim; if the real wind launches dust inward instead, the dust mass-loss rates computed here are upper limits.","fun_headline_variants_meta":{"raw":{"variants":["Radiation pressure fails to clear disc dust fast enough","Disc dust dispersal 5-10x too slow via radiation pressure","Radiation pressure can't solve relic disc conundrum","Only extreme EUV winds match disc clearing observations","Dust outflows too weak to explain fast disc clearing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000277,"raw_usage":{"total_tokens":1748,"prompt_tokens":1144,"completion_tokens":604,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":760,"completion_tokens_details":{"reasoning_tokens":534}},"tokens_in":760,"tokens_out":604,"duration_ms":6252,"temperature":1.0,"reasoning_tokens":534,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T20:56:03.385455+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same dust-growth and radiative-transfer calculation with a full 2D gas velocity solution for the photoevaporative wind; if micron-sized grains at the inner rim then cross the wind base at rates above roughly $10^{-6}\\,M_\\oplus$ yr$^{-1}$, the paper's central negative result is overturned. Alternatively, a survey finding non-accreting transition discs with X-ray winds whose mid-IR excess fades within a few $10^5$ yr would contradict the predicted clearing timescales.","supporting_citations":[],"review_version":1}