Ground-based and JWST Observations of SN 2022pul: II. Evidence from Nebular Spectroscopy for a Violent Merger in a Peculiar Type Ia Supernova

Nebular-phase spectroscopy (≳100 days past maximum) reveals the composition, geometry, kinematics, and ionization of SN ejecta via forbidden emission lines powered by 56Co decay. Type Ia supernovae (SNe Ia) include a variety of peculiar subtypes that deviate from the Phillips relation, offering clearer diagnostics of progenitor and explosion physics than normal SNe Ia. Among these, “03fg-like” (super-Chandrasekhar) SNe Ia show high luminosities, broad light curves, early-time C II, relatively low ejecta velocities, and low-ionization nebular spectra, potentially requiring total masses above the Chandrasekhar limit or alternative power sources. Proposed progenitors include: (i) differentially rotating near-MCh WDs; (ii) core-degenerate scenarios (near-MCh WD exploding within an AGB-star envelope); and (iii) violent mergers of two sub-MCh WDs. In the violent merger, the primary detonates first; ≈1 s later, the secondary burns at lower density, placing centrally the O/Ne/Mg and IMEs and yielding lower IGE production, asymmetries, and low ionization at late times. Nebular NIR/MIR spectroscopy with JWST provides relatively isolated IME/IGE lines, mitigating optical blending and enabling detailed diagnostics. Prior JWST nebular observations of normal SN Ia 2021aefx indicated a near-MCh delayed detonation. Here we analyze the dust-continuum-subtracted optical–MIR spectrum of the 03fg-like SN 2022pul at 338 d to identify lines, fit profiles, and compare with violent-merger model predictions, concluding that SN 2022pul likely resulted from a violent WD–WD merger.

Peculiar SNe Ia have been reviewed by Taubenberger (2017), Jha et al. (2019), and Liu et al. (2023). 03fg-like events (Howell 2006; Hicken 2007; Scalzo 2010; Silverman 2011; Taubenberger 2013; Hsiao 2020; Ashall 2021) challenge standard models, prompting scenarios including core-degenerate explosions and violent WD mergers (Dimitriadis 2022; Srivastav 2023; Siebert 2023). Violent mergers (Pakmor 2011, 2012; Kromer 2013; Pakmor 2017) predict central O/Ne and IMEs, reduced IGEs, asymmetry, and low ionization. JWST nebular spectra enable robust line identification beyond the ground-based NIR (Kwok 2023; DerKacy 2023). Blondin et al. (2023) computed nebular spectra at ~270 d for multiple explosion models; only the violent merger exhibited strong, centrally peaked [Ne II] 12.81 µm. Previous observations of 03fg-like SNe sometimes show [O I] λλ6300,6364 at late times, rare in SNe Ia (e.g., SN 2012dn, SN 2021zny).

Data: Combined optical+NIR+MIR nebular spectrum at 338 rest-frame days post-explosion (explosion date MJD 59785.3). Optical: MMT/Binospec at 332 d. Ground-based NIR: Keck/NIRES at 316 d (higher resolution, used for 1.0–1.83 µm). JWST: NIRSpec+MIRI beyond 1.83 µm at 338 d. A thermal dust continuum resembling a ~500 K blackbody is present; for line analysis, a T=500 K blackbody was subtracted from 2.5–14 µm, and residual MIR continuum was removed. No CO/SiO emission was identified. Spectral properties assessed include ionization state, line strengths, and profile asymmetries. Line identification and fitting: The observed spectrum was modeled as a superposition of emission from ions Fe I–III, Co II–III, Ni II–IV, S IV, Ca II/IV, Ar II/III, O I, and Ne II, following nebular fitting approaches (Maguire 2018; Flörs 2018, 2020; Kwok 2023). Constraints: all lines of the same ion share profile-shape parameters and kinematic offset; relative line strengths within an ion fixed to those from the Blondin et al. (2023) nebular spectrum of the Pakmor et al. (2012) violent merger (MERGER) model at 338 d, with inter-ion amplitudes free. Non-Gaussian profiles (double Gaussians, asymmetric flat-topped shells with Gaussian wings, skewed Gaussians) were used to capture asymmetries and shell-like geometries. Resolution uncertainties in JWST prism/LRS data were propagated (e.g., ~1200 km s−1 near 3 µm, ~2000 km s−1 near 6 µm, ~500 km s−1 near 12 µm), and fitting uncertainties estimated by bootstrap resampling; width uncertainties combined in quadrature with resolution. Model comparisons: Four explosion models (MCh delayed detonation DDT, MCh pulsationally-assisted GCD, sub-MCh double detonation DBLEDET, and sub-MCh violent merger MERGER) were recomputed to 338 d for comparison (Blondin et al. 2023). A clumped MERGER variant (volume-filling factor f=0.5) was computed to reduce ionization via enhanced recombination. To reproduce the narrow [O I] λλ6300,6364, an additional central mass (~0.1 M⊙ below ~2000 km s−1) was added by a Gaussian density enhancement centered at v≈580 km s−1 (σ≈600 km s−1). 56Ni mass estimate: Integrated 0.4–14 µm flux converted to luminosity using d=16±2 Mpc, then corrected by the gamma/positron energy deposition fraction from the MERGER model (Eabsorbed/Etotal=0.055) to infer 56Ni mass using standard 56Co decay power relations (Nadyozhin 1994; Branch & Wheeler 2017).

- Spectral coverage 0.4–14 µm (nebular, 338 d) reveals: lower mean ionization than normal SN Ia 2021aefx; asymmetric line profiles; strong IME emission (Ar, Ca), relatively weak higher-ionization IGEs; and first unambiguous [Ne II] 12.81 µm detection in a SN Ia. - IMEs: [Ar II] 6.98 µm dominates the MIR spectrum, blueshifted peak by ~−2000 km s−1; [Ar III] 8.99 µm is strong with boxy, slanted flat-top profile (Gaussian-wing FWHM ~10,300±1100 km s−1; inner-shell vmin ~5400±900 km s−1; overall offset ~+700±300 km s−1; shell-thickness offset ~−1000±200 km s−1). [S IV] 10.51 µm is firmly detected; the 10.5 µm feature requires [S IV] in addition to [Co II] 10.52 µm. [Ca II] λλ7291, 7324 is remarkably strong; [Ca IV] 3.21 µm shows an asymmetric shell-like profile (FWHM ~7500±1300 km s−1; vmin ~5900±1100 km s−1; offset ~−2400±200 km s−1). Lower ionization IMEs (Ar II, Ca II) require an extra blueshifted Gaussian component (e.g., Ar II Gaussian FWHM ~9300±1700 km s−1, offset ~−2300±100 km s−1; Ca II Gaussian FWHM ~4700±200 km s−1, offset ~−1500±100 km s−1), indicating ionization stratification. - IGEs: Relative weakening of high-ionization IGE lines compared to SN 2021aefx; [Co III] 11.89 µm is isolated and asymmetric with a redshifted peak (~+3500 km s−1), well fit by a sum of two Gaussians (FWHM ~5400±1900 km s−1 at offset ~+4600±2300 km s−1 and FWHM ~8000±2200 km s−1 at offset ~+1400±1800 km s−1). [Fe II] NIR lines share a similar redshifted asymmetry (double-Gaussian with FWHM ~4700±1200 km s−1 at offset ~+3800±500 km s−1 and FWHM ~7400±2100 km s−1 at offset ~−600±1300 km s−1). Ni II stronger than Ni III/IV (low ionization); overall IGE profiles consistently redshifted, suggesting a common emitting region distribution without strong ionization stratification. - Oxygen and neon: Narrow [O I] λλ6300, 6364 (FWHM ~2500±100 km s−1), centrally located with slight redshift (~+260±10 km s−1), and skewed (sawtooth) asymmetry; [O II]/[O III] not detected (blending with low-ionization IGEs). Strong, broad, centrally peaked [Ne II] 12.81 µm with unique asymmetric profile: narrow central peak near 0 km s−1 plus broad red Gaussian (FWHM ~10,000±500 km s−1) and blue asymmetric flat-top (FWHM ~7200±600 km s−1; vmin ~7000±500 km s−1; overall offset ~+1400±200 km s−1; shell-thickness offset ~+800±100 km s−1), implying distinct Ne distribution relative to IMEs/IGEs. - Geometry: Summed velocity-space profiles for element groups show distinct distributions: IMEs peak blueshifted (~−2000 km s−1) with extended red tails; IGEs peak redshifted (~+3500 km s−1) with extended blue tails; O+Ne centrally concentrated with additional extended Ne component. This indicates large-scale composition and velocity asymmetries consistent with violent merger ejecta. - 56Ni mass: Lbol,dep(338 d) ≈ (4.6 ± 1.2)×10^41 erg s−1 (distance-dominated uncertainty), implying M(56Ni) ≈ 0.66 ± 0.17 M⊙. - Model comparison: Of four models, only the violent MERGER reproduces key features: lower ionization (matching strong NIR [Fe II]), centrally peaked [Ne II] 12.81 µm, and rounded IME profiles implying central IMEs. Clumping (f=0.5) improves the [Fe III] 0.5 µm complex and boosts [Ar II] while reducing [Ni III], but over-weakens [Ca IV] and [Ar III]; Ni lines still over-predicted, suggesting lower stable Ni abundance and higher Ar/Ca abundance in SN 2022pul. Adding ~0.1 M⊙ within ≲2000 km s−1 reproduces the narrow [O I] doublet without degrading other fits, consistent with a higher-mass or less-disrupted secondary WD. - Pure deflagration (e.g., N100def) underpredicts luminosity, produces narrower lines, and weaker IMEs; disfavored.

The observations directly address the progenitor question for 03fg-like SNe Ia. The centrally peaked [Ne II] 12.81 µm and narrow central [O I] are strong signatures of violent WD–WD mergers, where the lower-density secondary burns later and its ejecta expand into the center, placing O/Ne/IMEs at low velocities. The pronounced asymmetries and opposite bulk velocity offsets of IMEs (blueshifted) and IGEs (redshifted) indicate distinct ejecta components consistent with merger geometry. Lower ionization across species further supports higher densities in central layers expected from mergers. Model comparisons show only the violent merger can reproduce these qualitative features; with modest adjustments (clumping to reduce ionization; reduced stable Ni; increased Ar/Ca; and enhanced central density), the model matches most diagnostics, including the narrow [O I]. Pure deflagration models cannot simultaneously match luminosity, line widths, and strong IMEs. The data also suggest diversity within 03fg-like events is likely driven by merger mass ratios, WD masses, burning conditions, and viewing angles. The presence of a dust continuum and possible C/O-rich CSM implied by early light-curve behavior may be linked to pre- or post-merger mass loss or disk formation, though reconciling substantial CSM masses with central O/Ne remains challenging. High-resolution, multi-wavelength nebular spectroscopy combined with spectropolarimetry and angle-dependent merger simulations are needed to refine ejecta geometry, clumping, and abundances and to constrain the CSM connection.

SN 2022pul, a 03fg-like SN Ia, shows a continuous 0.4–14 µm nebular spectrum with low ionization, strong IMEs, weaker high-ionization IGEs, highly asymmetric line profiles, narrow central [O I], and the first unambiguous [Ne II] 12.81 µm detection in a SN Ia. These features, especially central O/Ne and opposite bulk velocity offsets of IMEs vs. IGEs, are naturally explained by a violent merger of two sub-MCh WDs. Among several models, only the violent merger reproduces the key observables; further improvements arise from introducing clumping and an added ~0.1 M⊙ central mass, and from adjusting Ni/Ar/Ca abundances. A pure deflagration model is disfavored. Future work should explore a broader parameter space of violent mergers (mass ratios, metallicity, burning modes, and viewing angles), incorporate non-uniform clumping, refine stable Ni and IME abundances, and model CSM production and interaction consistent with early light-curve and dust constraints. Additional high-resolution JWST spectroscopy and spectropolarimetry of 03fg-like SNe will test the merger scenario and map ejecta asymmetries and dust formation.

- Authoritative mapping of individual authors to affiliations is not provided in the text fragment. Scientifically, the main limitations are: (i) JWST prism/LRS spectral resolution limits width and shape parameter precision, especially at 5–10 µm; (ii) assumptions fixing intra-ion relative line strengths to a specific model (MERGER) introduce model dependence; (iii) use of non-spherical phenomena represented by simplified profile components (Gaussian and asymmetric shells) cannot recover full 3D geometry; (iv) dust subtraction assumes an external 500 K blackbody and removal of residual MIR continuum; (v) distance uncertainty (16±2 Mpc) dominates the 56Ni mass error; (vi) gamma-ray/positron deposition fraction taken from the model; (vii) clumping implemented with small-scale, uniform filling factor affects ionization globally but cannot test large-scale clumps; (viii) abundance adjustments are suggested qualitatively (stable Ni lower, Ar/Ca higher) without a full parameter exploration; (ix) potential small-scale structure in [Ne II] may be affected by noise at long wavelengths.