Type IIb Supernova Progenitors in 3D: Variability and Episodic Mass Loss revealed by Radiation-Hydrodynamics Simulations

Jared A. Goldberg; Lars Bildsten; Matteo Cantiello; Yan-Fei Jiang

REVIEW 3 major objections 6 minor 1 cited by

Type IIb Supernova Progenitors in 3D: Variability and Episodic Mass Loss revealed by Radiation-Hydrodynamics Simulations

T0 review · 3 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read 3D radiation-hydrodynamics simulations show that the outer envelopes of Type IIb supernova progenitors are violently time-dependent, with order-of-magnitude luminosity swings on tens-of-day timescales and episodic mass loss that shapes a cl

desk verdict First 3D RHD simulations of stripped-envelope YSGs show real new physics, but partial-sphere domains make the headline luminosity and mass-loss numbers softer than they look. read the letter →

arxiv 2508.12486 v1 pith:SBDC3GKX submitted 2025-08-17 astro-ph.SR astro-ph.HE

Jared A. Goldberg , Yan-Fei Jiang , Lars Bildsten , Matteo Cantiello This is my paper

classification astro-ph.SRastro-ph.HE

keywords typeIIbsupernovaeyellowsupergiantsradiationhydrodynamicsstellarpulsationsepisodicmasslosscircumstellarmattershockbreakout3Dsimulations

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that the outer envelopes of Type IIb supernova progenitors—yellow supergiants that retain a thin hydrogen shell—are intrinsically three-dimensional and time-dependent, not smooth spherical shells. Its radiation-hydrodynamics simulations show that surface convection, driven by the hydrogen and helium opacity bumps, excites large-scale radial pulsations that swing the star's luminosity by roughly an order of magnitude on timescales of tens of days. The same pulsations launch both successful and failed mass ejections, yielding episodic mass loss of about $10^{-6}$ to $10^{-5}\,M_\odot\,\mathrm{yr}^{-1}$ and a surrounding halo whose density varies by orders of magnitude in space and time. If this is what real progenitors do, early shock-breakout and shock-cooling emission depend on exactly when during the pulsation cycle the star explodes, so spherical models cannot uniquely decode the explosion's outer structure. A sympathetic reader would care because this provides a self-consistent origin for dense circumstellar material around stripped supernovae and predicts observable variability before explosion.

What carries the argument

The engine is a self-excited, opacity-driven pulsation. In these stripped envelopes the luminosity-to-mass ratio is large, so convection is vigorous and organized into cells comparable to the pressure scale height; because the temperature near the surface crosses the hydrogen and first helium opacity peaks, the radiation field and the radial motion couple, sustaining low-azimuthal-order p-modes. The simulations evolve the coupled radiation and gas equations over a spherical wedge of the star, with the inner boundary continuously supplying the stellar luminosity, allowing convection, pulsation, and mass ejection to arise from the same dynamics rather than being prescribed.

What would settle it

Continuously monitor a sample of candidate stripped yellow supergiants at daily cadence for two years. Predicted: order-of-magnitude photometric swings with periods of 10–35 days and occasional outbursts; observed photometric quiet would rule out the pulsation mechanism in this regime. Alternatively, if many Type IIb supernovae show early light curves that are all well fit by a smooth spherical density profile, the predicted chaotic, time-varying halo is not universal.

Watch

Extended reading notes

Core claim

The simulations model two partially stripped yellow supergiant envelopes in convective steady state. In the lower-luminosity model the photospheric radius oscillates between about 135 and 185 $R_\odot$, effective temperature between roughly 4800 and 9000 K, and luminosity between $\log(L/L_\odot)\approx4$ and 5.4, with persistent modes at 19.4 and 9.5 days; the higher-luminosity model swings between 120 and 250 $R_\odot$, 5000 and 10,000 K, and $\log(L/L_\odot)\approx4.5$ to 5.6, with a 35-day mode. Material in the outer layers becomes supersonic, producing dense plumes that reach several times the photospheric radius and either escape or fall back. Time-averaged unbound mass-loss rates are

Load-bearing premise

The conclusions rest on the assumption that roughly one to three years of simulated post-relaxation behavior is a fair sample of what a real yellow supergiant does over the thousands of years before explosion, rather than an artifact of the initial setup or of the inner and outer boundaries.

Editorial extensions

If this is right

If these simulations capture real progenitors, Type IIb supernova precursors should show measurable order-of-magnitude photometric variability with periods of about 10–35 days, giving a warning signal before core collapse.
Early shock-breakout and shock-cooling emission cannot be inverted uniquely for envelope mass and radius without knowing the pulsation phase; the same star at different times yields different inferred density profiles.
Dense circumstellar material around SNe IIb, including the kind invoked for SN 1993J, can be produced by the progenitor itself through episodic, partly failed mass ejection, rather than requiring a separate binary or wind mechanism.
Because the halo mass beyond the photosphere is above roughly $10^{-4}\,M_\odot$ about half the time, a comparable fraction of events should show signatures of confined circumstellar material in their early light curves.
If these stars collapse directly to black holes, the stochastic angular momentum in the outer layers supports the formation of accretion disks and low-luminosity transients, although the low envelope mass limits their energy.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the two runs differ in luminosity, simulation volume, and resolution, the suggested trend toward longer periods at higher luminosity is not a measured scaling; a purpose-built parameter sweep over envelope mass, metallicity, and luminosity would test whether pulsation period and amplitude track $L/M_\mathrm{env}$.
A testable extension is to feed many snapshots from one simulation into a light-curve code and compute the distribution of early Type IIb emission for a single progenitor; the spread would quantify the systematic error in spherical shock-cooling fits.
If the violent variability phase lasts on the order of $10^4$ years, a decade-long deep time-domain survey should catch dozens to hundreds of active progenitors, offering a population-level test of the predicted variability statistics.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Josh, here's my read on arXiv:2508.12486. The genuinely new thing is that the authors put partially stripped YSG envelopes into 3D radiation hydrodynamics for the first time and let the physics run. The result—convection near the opacity peaks driving large-scale pulsations, episodic outflow and fallback, and a circumstellar halo with orders-of-magnitude density fluctuations—is real and worth taking seriously. The density profiles along different lines of sight genuinely differ by orders of magnitude, and that alone is a strong argument that spherical shock-cooling models will miss something. I credit the paper for laying out that case carefully.

The soft spots are real, though. The two simulations are wedges: 35% and 70% of the sphere. The paper renormalizes the wedge-integrated flux and mass flux to 4π. For the luminosity claim—order-of-magnitude total luminosity variability—that renormalization is only valid if the wedge is a fair sample at all times. The authors explicitly attribute the variability to low-ℓ modes (ℓ=2), which are spatially correlated. A wedge can easily see a bright lobe move through it and interpret that as a change in the star's total luminosity. Without a full-sphere control, the factor-of-25 luminosity range and the claimed periods could be partly a projection artifact. The paper's dismissive line about √N statistics doesn't address this. The same problem hits the mass-loss rates: if outflow happens in a few plumes, multiplying the wedge flux by 4π overestimates the total rate whenever the wedge catches a plume. The qualitative story—plumes rise, some fall back, some escape—is directly visible in the snapshots, so that's robust. But the quantitative mass-loss numbers, 10^-6 to 10^-5 Msun/yr, are shakier than the text implies.

Other caveats: only two models, short analyzed durations (1200 and 200 days), no released data or code, and no error bars on the power spectra. These are addressable, and none of them sink the central qualitative conclusion. The paper is also honest about its setup, which I appreciate.

Who should read it: anyone modeling SN IIb light curves or CSM around stripped stars. It deserves a serious referee—the referee should push for a full-sphere or at least a wider-domain run, or a clear quantitative argument for why the wedge renormalization is safe. I'd send it out.

Referee Report

3 major / 6 minor

Summary. This paper presents the first 3D radiation-hydrodynamics simulations (Athena++) of partially-stripped yellow supergiant envelopes as Type IIb supernova progenitors. Two models, YSG1L4.7 and YSG2L5.1, are evolved in spherical-polar wedges from hydrostatic MESA-inspired initial conditions. The authors report that surface convection coupled with opacity changes sustains large-amplitude, low-azimuthal-order radial pulsations, producing order-of-magnitude luminosity variability on tens-of-days timescales. The simulations also show supersonic flows, episodic mass loss at roughly 1e-6 to 1e-5 Msun/yr, and a strongly time-variable, clumpy circumstellar halo. The authors argue that these effects make spherically symmetric, single-epoch predictions of SN-IIb shock breakout and shock cooling unreliable, and that the simulations provide a self-consistent explanation for dense circumstellar material around pulsating evolved stars.

Significance. If the quantitative results hold, this is a significant step: it provides a self-consistent 3D picture of Type IIb progenitors in which variability and episodic mass loss emerge from the radiation-hydrodynamics rather than being inserted by hand. The simulations are expensive, the numerical method is well established, and the qualitative scenario is compelling. The authors also show appropriate caution in places, for example explicitly stating that the two models should not be interpreted as a period-luminosity relation. However, the headline quantitative claims—factor-of-25 luminosity variability, specific pulsation periods, and mass-loss rates—depend on renormalizing partial-sphere wedges to 4π and on relatively short post-relaxation intervals. As written, the paper robustly supports the qualitative picture but not the specific amplitudes and rates.

major comments (3)

[§2 and §5] The light curves are constructed by integrating the radiation flux over the simulation wedge and normalizing to the wedge solid angle (§5). YSG1L4.7 covers only θ∈[π/4,3π/4] and φ∈[0,π], i.e. ~35% of 4π, while YSG2L5.1 covers ~70% of 4π. The paper simultaneously claims that the pulsations are 'low-azimuthal-order' (ℓ=2). For an ℓ=2 pattern, positive and negative flux lobes cancel only when integrated over the full sphere; a wedge-averaged flux can therefore show much larger temporal variations than the true monopole luminosity. The same issue applies to the mass-loss rates in §6, which are computed as 4πr²⟨ρv_r⟩ over outward-moving zones and then implicitly treated as global. A full-sphere control run, or at least a spherical-harmonic decomposition of F_r and v_r, is needed to establish that the order-of-magnitude variability and the quoted Ṁ values are not dominated by the wedge geometr
[§2 and §6] The analyzed post-relaxation intervals are short, especially for YSG2L5.1: the model reaches the stated steady-state criterion after ≈150 days and is analyzed from 180 days onward, giving only ~200 days, or about 5.7 cycles of the claimed 35-day mode. The r²⟨F_tot⟩≈const criterion is a necessary condition for energy balance but not sufficient to demonstrate that the pulsation amplitude, the mass-loss rate, and the halo mass statistics have converged. The initial relaxation from a hydrostatic envelope, the density floor, and the outer boundary at 6436 R_sun could all influence these statistics. A convergence test, a longer YSG2L5.1 run, or an explicit systematic uncertainty estimate is needed before the quoted mass-loss rates and halo masses can be taken as quasi-steady progenitor properties.
[§5] The periodogram peaks at 19.4 d, 9.5 d, and 35 d are identified as low-order p-modes (ℓ=2) sustained by opacity feedback, but no mode identification is shown. There is no spherical-harmonic decomposition of the velocity or flux perturbations, no comparison with linear nonadiabatic pulsation periods, and no discussion of how the θ and φ wedge boundaries might shift mode frequencies or select certain azimuthal orders. Given that the light curve itself is obtained from a partial sphere, the dominant periods could be set by the domain geometry rather than by the stellar envelope. The text's caveat that the differences between the two models 'should not be interpreted as a true Period-Luminosity relation' is helpful, but it does not address the amplitude or period bias within each model.

minor comments (6)

[§2] The second run is introduced as 'YSG1L5.1' but is called YSG2L5.1 everywhere else; please correct the typo.
[Fig. 4 caption] The caption labels the light-curve panel as 'YSG1L4.9'; the model is YSG1L4.7.
[§4.3] The text refers to 'YSG2L4.9' when computing r_isco; this should be YSG2L5.1.
[§2] The boundary conditions at the θ and φ boundaries are not described. Please state whether they are periodic, reflecting, or outflow, and justify that the wedge boundaries do not artificially constrain the low-order modes that are central to the paper.
[Abstract and §2] The abstract quotes M_env~0.1–1 M_sun, but YSG1L4.7 has only ~0.05 M_sun exterior to the inner boundary. Please clarify whether the quoted range includes only the simulation domain or the full envelope exterior to the core.
[Fig. 3] Several axis labels in the lower panels appear as '10□1M⊙/yr' with missing superscripts; these should be rendered as 10^{-1}–10^{-6} M⊙/yr.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported variability and mass-loss rates are emergent simulation outputs, not quantities encoded in the setup.

full rationale

The core claims—order-of-magnitude luminosity variability, pulsation periods of 19.4 d/9.5 d/35 d, and episodic mass loss ~1e-6 to 1e-5 Msun/yr—are extracted from the Athena++ RHD evolution after a steady-state criterion (r^2<Ftot> constant within 5%) is satisfied, rather than being imposed or fitted. The inner boundary uses a fixed total luminosity from a MESA model as an input; this sets the mean luminosity but not the time-dependent fluctuations, which emerge from convection and opacity feedback. The citations to Goldberg et al. (2022a) and Schultz et al. (2022) are methodological precedents (domain setup, Lomb-Scargle/SLF cleaning), not results assumed in order to obtain the paper's conclusions. The partial-sphere domain (35-70% of 4pi) and short steady-state durations are legitimate robustness concerns about whether the amplitudes and rates are converged or biased, but that is an external validity/correctness matter, not circularity: nothing in the paper defines the predicted variability in terms of the boundary conditions or renames a fitted parameter as a prediction. The paper even cautions 'should not be interpreted as a true Period-Luminosity relation given the differences in the geometry of the simulation domain,' acknowledging geometry rather than hiding it. No self-citation is load-bearing in the derivation chain, so the paper is self-contained against external benchmarks and receives score 0.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claims rest on the chosen simulation setup: grid resolution, boundary conditions, opacity tables, stripping history, and initial conditions. These are reasonable choices but not externally constrained, so the quantitative predictions inherit their uncertainty.

free parameters (5)

Envelope composition = X=0.425, Y=0.555, Z=0.02 with Asplund et al. 2009 abundances
Chosen to match the motivating MESA models; opacity from this mixture drives the pulsations, so the result depends on this input.
Inner boundary luminosity = Fixed total luminosity at the inner boundary (about 30% radiative, rest convective)
The 'Fixed L' boundary condition injects the full stellar luminosity as radiation flux, driving the convection; the exact thermal structure near the boundary may affect the pulsation amplitude.
Density floor = 6e-16 g/cm3
A numerical floor imposed to avoid vacuum; it can affect the mass and density of the outer halo and the inferred mass-loss rate.
MESA mixing length alpha = 3
Used to construct the initial 1D models; a standard but arbitrary choice that influences the envelope structure imported into 3D.
Stripping parameters = Enhanced Dutch wind scaling after core He depletion to yield Menv 0.1-1 Msun
The degree and timing of stripping set the envelope mass and Teff; only two realizations are considered, so the results may not represent all SN-IIb progenitors.

assumptions (5)

standard math Radiation-hydrodynamics equations as implemented in Athena++ (Jiang 2021)
The simulation results depend on the numerical solution of the RHD equations; no independent code or analytic solution is used to cross-check.
domain assumption Spherical polar domain with partial solid-angle coverage
YSG1L4.7 covers phi in [0,pi] and the theta range excludes the poles; angle-averaged quantities are normalized to the domain, which could bias the global luminosity and mass-loss rates.
domain assumption Non-rotating, non-magnetic initial models
Rotation and magnetic fields are neglected; they could modify convection, pulsation, and mass loss in real stripped stars.
domain assumption No dust formation in the outer layers
The authors note dust would further levitate outbound material and raise the mass-loss rate; ignoring it gives a lower bound.
ad hoc to paper Initial conditions constructed from a relaxed hydrostatic envelope rather than evolved self-consistently from an earlier phase
The 3D runs begin from a constructed radiative envelope that is then allowed to relax; the relaxation process may imprint on the early dynamics and the measured mass-loss rates.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Type IIb Supernova Progenitors in 3D: Variability and Episodic Mass Loss revealed by Radiation-Hydrodynamics Simulations." pith.science (2026). https://pith.science/paper/SBDC3GKX

@misc{pith2026250812486,
  author       = {Pith},
  title        = {Pith review of: Type IIb Supernova Progenitors in 3D: Variability and Episodic Mass Loss revealed by Radiation-Hydrodynamics Simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SBDC3GKX}},
  note         = {Machine review of arXiv:2508.12486}
}

abstract

We present the first 3D Radiation-Hydrodynamics simulations of partially-stripped ($M_\mathrm{core}\sim10M_\odot$, $M_\mathrm{env}\sim0.1-1M_\odot$) Yellow Supergiant ($L\sim10^5$, $T_\mathrm{eff}\approx5000-8000$K) envelopes, constructed with Athena++. These envelope models represent the progenitors of Type IIb supernovae (SNe-IIb), which have lost a substantial fraction of their H-rich envelope before undergoing core-collapse. The luminosity-to-mass ratio is high in these extended envelopes, and convection is strongly driven by Hydrogen and Helium opacity peaks. This surface convection, coupled with changes in the opacity, sustains large-amplitude low-azimuthal-order radial pulsations, creating order-of-magnitude variability in the stellar luminosity on a timescale of tens of days. If persistent prior to a SN-IIb, these variations could herald the upcoming explosion. Supersonic fluid motions across the outer layers of the star lead to both successful and failed mass ejection events, which shape the circumstellar environment and drive episodic mass loss ($\sim10^{-6}-10^{-5}M_\odot/$yr, in outbursts). The resulting 3D gas distribution in the outer atmosphere, responsible for early-time supernova shock-breakout and shock-cooling emission, shows orders-of-magnitude fluctuations in both space and time at any given radial location. This intrinsically complex halo of bound and unbound material complicates predictions for early SN-IIb lightcurves relative to spherically-symmetric models. However, it does provide a natural, self-consistent explanation for the presence and diversity of dense circumstellar material observed or inferred around pulsating evolved stars.

Figures

Figures reproduced from arXiv: 2508.12486 by the authors.

**Figure 1.** History of the angle-averaged radial profiles for the YSG1L4.7 (left) and YSG2L5.1 (right) models, zoomed in on a subset of times when the envelope is in convective steady-state. Panels show (top to bottom): log10(density), log10(gas temperature), mass-averaged radial velocity, and magnitude of the mass-averaged tangential velocity. boundary is supplied with the larger total luminosity as radiation flux, via the “Fi… view at source ↗

**Figure 2.** Upper Panels: Gas density (purple-orange colors, left panels) and radial velocity (red-blue colors, right panels) slices through a representative snapshot of the YSG1L4.7 model at day 1845.5. The central panels show equatorial slices, indicating fixed radii at r = 100R⊙ (grey dashed circle), r = Rph = 167R⊙ (black dashed circle), and r = 250R⊙ (grey dotted circle). These radii, respectively, correspond to the upper,… view at source ↗

**Figure 3.** Left Panels: Top to bottom show the line-of-sight luminosity 4πr2Fr,01, ratio of kinetic energy to internal energy, radially-integrated column density y, and gas density ρ for a representative snapshot of the YSG1L4.7 model. Orange colors indicate the volume-weighted probability of finding a fluid element with a given y-axis value at each radial coordinate. Black curves of varying thickness show four arbitrary indiv… view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: shows the lightcurve and power spectra for the YSG1L4.7 model, which shows periodic and stochastic luminosity fluctuations ranging from log(L/L⊙) ≈ 4 − 5.4. Due to uneven time-sampling of our simulation output, the power spectra were calculated with a Lomb-Scargle pe…

**Figure 5.** Figure 5: Eruptive Mass loss and the resulting circumstellar halo for the YSG1L4.7 model. Upper left panel: Mass exterior to select radii (xm) as a function of time. The black line at r = 180R⊙ approximates the mass external to the photosphere. Upper right panel: X-axis shows th…

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Radio and X-ray Observations of the Transitional Supernova 2019yvr: Insights into the Progenitor Mass-Loss History
astro-ph.HE 2026-07 accept novelty 6.0 of 10

Radio SSA modeling of SN 2019yvr yields a CSM density profile ρ ∝ r^{-1.65} and mass-loss rates ~1–3 imes10^{-5} M⊙ yr^{-1} with no dramatic density jump at the optical Ib o IIn transition.

Reference graph

Works this paper leans on

2 extracted references · 1 linked inside Pith · cited by 1 Pith paper

[1]

Aerts, C., Puls, J., Godart, M., & Dupret, M. A. 2009, A&A, 508, 409, doi: 10.1051/0004-6361/200810471 Aldering, G., Humphreys, R. M., & Richmond, M. 1994, AJ, 107, 662, doi: 10.1086/116886 Anders, E. H., Lecoanet, D., Cantiello, M., et al. 2023, Nature Astronomy, 7, 1228, doi: 10.1038/s41550-023-02040-7 Antoni, A., & Quataert, E. 2022, MNRAS, 511, 176, d...

arXiv 2009
[224]

M., Davidson, K., & Smith, N

https://arxiv.org/abs/astro-ph/9705097 Humphreys, R. M., Davidson, K., & Smith, N. 2002, AJ, 124, 1026, doi: 10.1086/341380 Humphreys, R. M., Jones, T. J., & Martin, J. C. 2023, AJ, 166, 50, doi: 10.3847/1538-3881/acdd6c Irani, I., Morag, J., Gal-Yam, A., et al. 2024, ApJ, 970, 96, doi: 10.3847/1538-4357/ad3de8 Irwin, C. M., & Hotokezaka, K. 2025, MNRAS, ...

arXiv 2002

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Aerts, C., Puls, J., Godart, M., & Dupret, M. A. 2009, A&A, 508, 409, doi: 10.1051/0004-6361/200810471 Aldering, G., Humphreys, R. M., & Richmond, M. 1994, AJ, 107, 662, doi: 10.1086/116886 Anders, E. H., Lecoanet, D., Cantiello, M., et al. 2023, Nature Astronomy, 7, 1228, doi: 10.1038/s41550-023-02040-7 Antoni, A., & Quataert, E. 2022, MNRAS, 511, 176, d...

arXiv 2009

[2] [224]

M., Davidson, K., & Smith, N

https://arxiv.org/abs/astro-ph/9705097 Humphreys, R. M., Davidson, K., & Smith, N. 2002, AJ, 124, 1026, doi: 10.1086/341380 Humphreys, R. M., Jones, T. J., & Martin, J. C. 2023, AJ, 166, 50, doi: 10.3847/1538-3881/acdd6c Irani, I., Morag, J., Gal-Yam, A., et al. 2024, ApJ, 970, 96, doi: 10.3847/1538-4357/ad3de8 Irwin, C. M., & Hotokezaka, K. 2025, MNRAS, ...

arXiv 2002