Single-degenerate type Ia supernovae without hydrogen contamination

The lack of hydrogen in spectra of type Ia supernovae (SN Ia) is often seen as troublesome for single-degenerate (SD) progenitor models. We argue that, since continued accretion of angular momentum can prevent explosion of the white dwarf, it may be natural for the donor stars in SD progenitors of SN Ia to exhaust their envelopes and shrink rapidly before the explosion. This outcome seems most likely for SD SN Ia progenitors where mass-transfer begins from a giant donor star, and might extend to other SD systems. Not only is the amount of hydrogen left in such a system below the current detection limit, but the donor star is typically orders of magnitude smaller than its Roche lobe by the point when a SD SN Ia occurs, in which case attempts to observe collisions between SN shocks and giant donor stars seem unlikely to succeed. We consider the constraints on this model from the circumstellar structures seen in spectra of SN 2006X and suggest a novel explanation for the origin of this material.

💡 Research Summary

The paper addresses the long‑standing problem of the apparent lack of hydrogen in the spectra of Type Ia supernovae (SNe Ia) that are thought to arise from single‑degenerate (SD) progenitor systems. In the classic SD picture a carbon‑oxygen white dwarf (WD) accretes material from a non‑degenerate companion (main‑sequence, sub‑giant, or red‑giant) until it reaches the Chandrasekhar mass (~1.38 M⊙) and ignites carbon explosively. However, observations place very stringent limits on any hydrogen (≲0.01 M⊙) and have failed to detect the expected signatures of the supernova shock colliding with a large companion star.

The authors propose a two‑fold mechanism that naturally eliminates hydrogen contamination. First, as the WD accretes mass it also gains angular momentum. Rotational support raises the effective mass limit for stability. In the extreme case of differential rotation a WD could remain stable up to ~4 M⊙; even with solid‑body rotation the limit is raised to about 1.5 M⊙. Consequently the mass‑transfer phase can continue well beyond the non‑rotating Chandrasekhar limit, allowing the donor star to lose a substantial fraction of its envelope before the explosion.

Second, when the donor is a red‑giant (RG) the envelope mass (M_env) eventually drops below a critical value (~10⁻² M⊙). Below this threshold the stellar radius no longer follows the flat giant‑branch relation; instead the envelope contracts on its Kelvin‑Helmholtz (thermal) timescale τ_KH,env ≈ GM_star M_env/(R L). For typical RG core masses (0.4–0.5 M⊙) and luminosities (~10² L⊙) this timescale becomes of order years or less once M_env ≲10⁻³ M⊙. The star’s radius shrinks from ≳10 R⊙ to ≲1 R⊙, reducing the geometric cross‑section for interaction with the supernova ejecta to ~10⁻⁶ of the original value. After the envelope has fully contracted the remaining material is tightly bound to the compact core, and any hydrogen left in the system is below the detection threshold.

The explosion itself is delayed until the WD either loses angular momentum or redistributes it internally. Estimates for angular‑momentum loss in differentially rotating models give an upper limit of ~10⁶ yr, but even in the solid‑body case the “simmering” phase of carbon burning lasts ~10³ yr. This delay comfortably exceeds the envelope‑contraction timescale, ensuring that the donor is already shrunken when the thermonuclear runaway finally occurs.

The authors discuss the applicability of this scenario to various SD channels. It is most natural for RG+WD systems, where the donor’s envelope can be exhausted during the prolonged mass‑transfer phase. For supersoft X‑ray sources (main‑sequence or sub‑giant donors) the mechanism can still operate if the donor evolves into a giant while the WD is still below the rotationally‑enhanced mass limit; otherwise the donor will already be a compact object with negligible hydrogen.

Population‑synthesis considerations suggest that, for realistic mass‑transfer efficiencies (∼30 % or less), many RG+WD binaries will indeed reach WD masses of 1.4–1.5 M⊙ before the donor’s envelope becomes too thin, making the proposed pathway common. The paper also revisits the circumstellar Na I D absorption features observed in SN 2006X. While previous work interpreted these as evidence for a SD progenitor with multiple nova shells, the authors argue that such material could arise from prior nova eruptions or wind‑blown shells, not from a still‑present hydrogen‑rich envelope.

In summary, the combination of rotationally‑supported super‑Chandrasekhar white dwarfs and rapid post‑envelope contraction of red‑giant donors provides a self‑consistent explanation for the lack of hydrogen in SN Ia spectra and the absence of detectable shock‑companion interaction. The model predicts that most SD progenitors will explode after the companion has shrunk to a size orders of magnitude smaller than its Roche lobe, rendering hydrogen signatures and collision signatures essentially invisible with current observational capabilities. The authors call for future binary‑population studies that incorporate angular‑momentum effects and envelope contraction to quantify the fraction of SNe Ia that should exhibit any hydrogen contamination.