A Late-time Radio Survey of Type Ia-CSM Supernovae with the Very Large Array

TL;DR

This work performs the largest late-time (>$1$ yr) radio survey of Type Ia-CSM SNe with the VLA at 6 GHz to constrain circumstellar densities via wind-like CSM models. By applying a TC17-based radio light-curve framework that includes a transition from ejecta-dominated to Sedov-Taylor evolution and exploring a range of microphysical parameters, the authors derive per-object limits on $\dot{M}/v_{wind}$ and compare them with optical mass-loss inferences. The majority of SNe Ia-CSM in the sample are radio non-detections, implying mass-loss rates in the range $\sim 10^{-4}$–$10^{-2}$ M$_{\odot}$ yr$^{-1}$ (for $v_{wind}=100$ km s$^{-1}$), though uncertainties in $\epsilon_B$, $\epsilon_e$, $T_{wind}$, and CSM geometry can alleviate tension with optical Hα results. The single detected case, SN 2022esa, is likely not Ia-CSM, while SN 2020eyj's radio data suggest modest mass-loss rates; overall, the results favor either non-spherical or shell-like CSM or significantly smaller microphysical efficiencies, highlighting the need for coordinated optical-radio campaigns with future facilities like Rubin, SKA, and ngVLA.

Abstract

Type Ia-CSM supernovae (SNe) are a rare and peculiar subclass of thermonuclear SNe characterized by emission lines of hydrogen or helium, indicative of a high-density circumstellar medium (CSM). Their implied mass-loss rates of $\sim 10^{-4}-10^{-1}$ M$_{\odot}$ yr$^{-1}$ (assuming $\mathrm{ \sim 100 \ km\ s^{-1}}$ winds) from optical observations are generally in excess of values observed in realistic SN Ia progenitors. In this paper, we present an independent study of CSM densities around a sample of 29 archival Ia-CSM SNe using radio observations with the Very Large Array at 6 GHz. Motivated by the late ($\sim$2 yr) radio detection of the Ia-CSM SN 2020eyj, we observed old ($>$1 yr) SNe where we are more likely to see the emergent synchrotron emission that may have been suppressed earlier by free-free absorption by the CSM. We do not detect radio emission down to 3$σ$ limits of $\sim$35 $μ$Jy in our sample. The only radio-detected candidate in our sample, SN 2022esa, was likely mis-classified as a Ia-CSM with early spectra, and appears more consistent with a peculiar Ic based on later-epochs. Assuming a wind-like CSM with temperatures between $2 \times 10^4$ K and $10^5$ K, and magnetic field-to-shock energy fraction ($ε_B$) = $0.01-0.1$, the radio upper limits rule out mass-loss rates between $\sim 10^{-4}-10^{-2}$ M$_{\odot}$ yr$^{-1}$ (100 km s$^{-1}$)$^{-1}$. This is somewhat in tension with the estimates from optical observations, and may indicate that more complex CSM geometries and/or lower values of $ε_B$ may be present.

Figures (7)

• Figure 1: The time evolution of the forward shock radius (left panel) and velocity (right panel) from the TC17 model (Section \ref{['sec:randv']}) for the case of $E_{51}= 1.0$, $M_{ej}= 1.44$ M$_{\odot}$, and $\dot{M}/v_{wind}$ = 0.1 M$_{\odot}$ yr$^{-1}$ (100 km s$^{-1}$)$^{-1}$. The vertical blue shaded section represents the time range at which our radio observations were obtained. Dashed lines represent the asymptotic solutions for the ejecta-dominated ($n=10$ density profile) and Sedov-Taylor cases.
• Figure 2: Predicted 6 GHz radio light curves for SNe Ia-CSM, shown alongside radio observations in this paper and from the literature. Solid curves correspond to the case of $T_{{wind}} = 2.0 \times$10$^{4}$ K, and dashed curves for $T_{{wind}} =$10$^{5}$ K. Left panel shows the case of $\epsilon_B = 0.1$, and the right panel shows $\epsilon_B = 0.01$ (both panels assume $\epsilon_e = 0.1$). Light curves span a range of $\dot{M}/v_{wind}$ from $3 \times 10^{-5}$ M$_{\odot}$ yr$^{-1}$ (dark blue lines) to $0.1$ M$_{\odot}$ yr$^{-1}$ (yellow lines), assuming $v_{wind} = 100$ km s$^{-1}$. The shaded gray region represents SNe Ia $<$ 1 yr, which are excluded in this study. SNe with 3$\sigma$ upper limits are plotted as inverted black triangles. The radio detections of SN 2022esa from 2023 and 2024 are shown as red circles connected by a dashed line. Literature measurements of SN 2020eyj from Kool23+ and yang23 (scaled to 6 GHz) are shown as orange points.
• Figure 3: Illustration of how we use radio upper limits to constrain the range of ruled-out mass-loss rates $\dot{M}/v_{wind}$ (green shade). Each light curve (green solid) shows the predicted luminosity for different mass-loss rates at a given SN age (30 days, 1 year, 10 years). The range of mass-loss rates where the predicted luminosity exceeds the measured upper limit (dashed line) are ruled out. Due to the light curve shape for wind-like CSM, later-time observations are more effective at ruling out a greater (and higher) range of mass-loss rates.
• Figure 4: The range of mass-loss rates ruled out by the radio upper limits for each SN. For each SN, we show ranges for the three models in Table \ref{['tab:params']} for the parameters $\epsilon_B$ and $T_{wind}$. See Section \ref{['sec:results']} for details.
• Figure 5: 6 GHz light curves produced by interaction with shell-like CSM, compared with our radio observations of SNe Ia-CSM. The light curves accounting for absorption are solid, whereas the ones with no absorption are dashed. The left panel shows models for $\epsilon_B = 0.1$, and the right panel for $\epsilon_B = 0.01$ (both panels use $\epsilon_e = 0.1$). Each light curve's color is specific to the density and mass of the shell, with the dark purple color representing $\rho_{csm} = 10^{-20}$ g cm$^{-3}$ and $M_{shell} = 0.147$$M_{\odot}$, teal corresponding to $\rho_{csm} = 10^{-19}$ g cm$^{-3}$ and $M_{shell} = 1.47$$M_{\odot}$, and yellow depicting $\rho_{csm} = 10^{-18}$ g cm$^{-3}$ and $M_{shell} = 14.7$$M_{\odot}$. We assume $R_{in} = 10^{17}$ cm and $f_{R} = 1$ for this model (see Section \ref{['sec:discussion']}).