Eruptive mass loss less than a year before the explosion of superluminous supernovae. II. A systematic search for pre-explosion eruptions with VLT/X-shooter

1. Problem Statement

The evolution and fate of massive stars are intimately linked to their mass-loss history. While steady stellar winds are well-understood, episodic, violent eruptive mass-loss events shortly before core collapse remain poorly characterized.

• Context: Superluminous supernovae (SLSNe-I) are rare, extremely luminous explosions thought to originate from massive, hydrogen-poor progenitors.
• The Phenomenon: Previous serendipitous discoveries (e.g., iPTF16eh, SN 2018ibb) revealed broad Mg II absorption features in SLSN-I spectra. These features are blueshifted by thousands of km/s relative to the host galaxy, interpreted as resonance scattering by a rapidly expanding shell of circumstellar material (CSM) ejected shortly before the explosion.
• The Gap: It was unclear whether these pre-explosion eruptions are a common feature of the SLSN-I population or a rare occurrence. Previous detections were limited to a few objects, and systematic searches were hindered by the need for high-quality near-ultraviolet (UV) spectroscopy to resolve these specific features.

2. Methodology

The authors conducted a systematic search for CSM shells in a sample of SLSNe-I using the VLT/X-shooter spectrograph.

• Sample Selection:
  • A total of 21 hydrogen-poor SLSNe-I were analyzed, spanning redshifts $z = 0.13 - 0.95$.
  • The sample included 13 objects from a dedicated triggered X-shooter program and 8 objects from the literature/ESO archive with existing high-quality spectra.
  • Selection criteria required redshifts $z > 0.11$ (to place Mg II resonance lines within the UVB arm) and sufficient signal-to-noise (S/N).
• Data Reduction & Analysis:
  • Raw data were re-reduced using the X-shooter pipeline (v3.6.3) and custom scripts to ensure uniformity.
  • Spectra were corrected for Milky Way extinction and flux-calibrated.
  • Focus Region: The analysis centered on the 2800 Å region (rest-frame) where the Mg II doublet resides.
• Modeling Approach:
  • Monte Carlo Scattering Code: A radiative transfer code (based on Fransson et al. 2014) was used to model the interaction between the SN continuum and a homologously expanding CSM shell.
  • Parameter Space: A grid of $\sim 200,000$ models was generated, varying:
    • Maximum shell velocity ($v_{max}$): 500–5500 km/s.
    • Inner/Outer radii ($R_{in}, R_{out}$) relative to the photosphere.
    • Shell thickness ($\Delta r$).
    • Optical depth ($\tau$).
  • Statistical Framework: A Bayesian likelihood approach was employed to compare "continuum-only" models against "CSM-shell" models. A detection was claimed if the CSM model was preferred at > 5$\sigma$ significance.
  • Constraints: For non-detections, the authors calculated the column density limits above which a shell would have been detectable given the specific S/N of each spectrum.

3. Key Contributions

• Systematic Survey: This is the first systematic search for pre-explosion CSM shells in a large, homogeneous sample of SLSNe-I using high-resolution UV spectroscopy.
• New Detections: The study identified two previously undiscovered SLSNe-I with CSM shells (DES15S2nr and DES16C3ggu), bringing the total number of confirmed cases to six (including SN 2018ibb, SN 2020xga, SN 2022xgc, and iPTF16eh).
• Diagnostic Tool: The authors developed and validated a robust statistical diagnostic to distinguish between real CSM features and noise, allowing them to place strict upper limits on the existence of shells in non-detection cases.
• Physical Characterization: Detailed modeling of the new detections provided precise constraints on shell location, velocity, and ejection timescales.

4. Key Results

A. Detections and Shell Properties

Five objects in the sample showed robust CSM detections (> 5$\sigma$):

1. SN 2018ibb, SN 2020xga, SN 2022xgc: Confirmed previous findings.
2. DES15S2nr:
   • Shell location: $\sim 3.4 \times 10^{15}$ cm.
   • Velocity: $\sim 4800$ km/s.
   • Ejection time: $\sim 2$ months before explosion.
3. DES16C3ggu:
   • Shell location: $\sim 4.8 \times 10^{15}$ cm.
   • Velocity: $\sim 4700$ km/s.
   • Ejection time: $\sim 3$ months before explosion.

B. Common Characteristics

• Velocity Range: All detected shells share a narrow velocity range of 3000–5000 km/s. This suggests a common physical mechanism driving the mass ejection, distinct from typical stellar winds.
• Geometry: The shells are relatively thin ($\Delta r \sim 0.05 - 0.2 R_{ph}$) and located close to the photosphere ($R_{in} \sim 1.8 - 2.9 R_{ph}$).
• Ejection Timescale: The shells were expelled within months to less than a year prior to core collapse, consistent with late-stage eruptive mass loss.

C. Non-Detections and Limits

• Excluded Parameter Space: For the 16 non-detections, the authors ruled out shells that are both optically thick and geometrically thick.
• Selection Bias: The non-detections are unlikely to be due to selection effects (e.g., low S/N). The analysis shows that shells with properties similar to the detected ones would have been visible in spectra with S/N > 5.
• Implication: The absence of detections in the majority of the sample suggests that only a subclass of SLSN-I progenitors undergoes these specific late-stage eruptive mass-loss episodes.

D. Correlations with SN Properties

• No Strong Correlations: No significant correlations were found between CSM properties and SN peak luminosity, rise time, or color.
• Marginal Correlation: A tentative negative correlation was found between CSM shell velocity and light curve decline timescale ($\rho = -0.90, p=0.04$). This suggests that faster-declining SLSNe might host faster-moving shells, potentially linking the explosion mechanism to the pre-explosion mass ejection. However, the small sample size prevents definitive conclusions.

5. Significance and Implications

• Progenitor Physics: The detection of shells ejected only months before explosion challenges standard steady-wind models. The velocities ($\sim 5000$ km/s) and timescales are consistent with Pulsational Pair-Instability (PPI) eruptions or LBV-like giant eruptions.
• Mass Estimates: While direct mass estimates are uncertain due to ionization and geometry assumptions, the inferred properties (especially if PPI-driven) suggest progenitor He-core masses $> 45 M_{\odot}$.
• Population Diversity: The fact that only $\sim 30\%$ of the sample shows these features implies a physical diversity among SLSN-I progenitors. It is not a universal trait of all SLSN-I, but rather a specific channel of massive star evolution.
• Future Outlook: The study highlights the necessity of high-quality UV spectroscopy for understanding massive star death. Future surveys (e.g., LSST) combined with targeted UV follow-up will be crucial for expanding this sample and distinguishing between intrinsic physical diversity and observational limitations.

In conclusion, this paper establishes that eruptive mass loss shortly before explosion is a real, albeit not universal, phenomenon in SLSNe-I, characterized by fast-moving, dense shells that provide a unique window into the final stages of massive stellar evolution.