Multi-Epoch X-Ray Detection of SLSN-I 2018bsz: Constraints on the Powering Mechanism and Ejecta Structure

Technical Summary: Multi-Epoch X-Ray Detection of SLSN-I 2018bsz

Problem and Motivation
Superluminous supernovae (SLSNe) are rare stellar explosions with energies approximately 100 times greater than ordinary core-collapse supernovae. The mechanisms powering these events remain debated, with primary candidates including strong interaction between the ejecta and circumstellar medium (CSM), or energy injection from a rapidly rotating, highly magnetized neutron star (magnetar). While optical modeling has successfully reproduced light curves for many SLSNe-I using magnetar models, distinguishing between power sources is challenging as different mechanisms can produce similar optical signatures. X-ray observations offer a critical probe to break this degeneracy; however, detections are rare, particularly at late epochs when ejecta absorption decreases and magnetar emission might become visible. Prior to this study, only one SLSN-I (SCP 06F6) had been definitively detected in X-rays.

Methodology
The authors present a multi-epoch X-ray spectroscopic study of SN 2018bsz, the closest known stripped SLSN-I (distance ~111 Mpc). The dataset comprises four Chandra observations (ACIS-S/S3) taken between 87 and 304 days post-explosion, and one XMM-Newton observation (EPIC-pn/MOS) taken at 1253 days post-explosion.

Data reduction followed standard procedures using CIAO (v4.13) for Chandra and SAS (v22.1.0) for XMM. Spectral analysis was performed using XSPEC (v12.15.0c) with C-statistics due to low count rates.

• Source Identification: Chandra images revealed the SN source and three fainter, nearby X-ray sources (S1–S3) within the XMM point-spread function.
• Spectral Modeling: Spectra were fitted with Galactic absorption ($N_H = 2.02 \times 10^{21}$ cm$^{-2}$). Models included power laws (representing magnetar spin-down or non-thermal interaction) and thermal plasma models (mekal, representing CSM shock interaction).
• Contamination Handling: For the late-time XMM observation, the authors modeled the contribution of the nearby sources (S1–S3) using a "contamination annulus" derived from Chandra data. They fitted the XMM spectrum with a composite model: tbabs(powCont.A + pow) or tbabs(powCont.A + mekal), where the contamination component was fixed based on Chandra results.
• Power Source Constraints: The authors tested two scenarios:
  1. Magnetar: A Markov Chain Monte Carlo (MCMC) fit was performed to constrain the initial spin period ($P_0$) and magnetic field ($B$) assuming a dipole spin-down model. They compared these results against optical light-curve fits from Gomez et al. (2024) and considered the effects of ejecta absorption and ionization breakout.
  2. CSM Interaction: Mass-loss rates ($\dot{M}$) were estimated using the luminosity-mass loss relation from Chevalier & Fransson (2017).

Key Results

1. Detection: SN 2018bsz was detected at >3$\sigma$ significance in all four Chandra epochs. A tentative detection was made in the XMM observation at 1253 days, though this is sensitive to the level of contamination from nearby sources. This makes SN 2018bsz the second X-ray detected SLSN-I and the third X-ray detected SLSN overall.
2. Spectral Evolution:
   • Early Epochs (Chandra): The spectra are well-described by hard power laws with photon indices $\Gamma \approx 0.9–1.9$ (best fit $\Gamma = 1.32$ for simultaneous fit).
   • Late Epoch (XMM): The spectrum softens significantly. At 1253 days, the data are consistent with a soft power law ($\Gamma < 4.0$) or a thermal plasma model with $kT \approx 0.20$ keV.
3. Luminosity Evolution: The X-ray luminosity remains relatively flat over time ($L_{0.5-8 \text{ keV}} \approx 1.1 \times 10^{40}$ erg s$^{-1}$), showing little decline between 87 and 304 days.
4. Magnetar Constraints:
   • MCMC fits to the X-ray data (assuming negligible ejecta absorption) yield a weak magnetic field ($\log B \approx 13.1$) and a long initial spin period ($P_0 \approx 16$ ms).
   • These parameters are inconsistent with the "best-fit" magnetar parameters derived from optical light curves by Gomez et al. (2024) ($P_0 \approx 0.9$ ms, $\log B \approx 14.4$).
   • Using the optical-derived parameters ($B_{opt}, P_{opt}$), the magnetar model underpredicts the observed X-ray luminosities at all epochs except the earliest and fails to reproduce the flat light curve.
   • Accounting for ejecta absorption increases the discrepancy. Ionization breakout models suggest that even with favorable parameters, the magnetar emission would not escape the ejecta until $\sim$2–72 years post-explosion, well beyond the observational window.
5. CSM Interaction Constraints:
   • The spectral softening and flat luminosity evolution are consistent with an initially strong ejecta-CSM interaction that weakens over time.
   • Derived mass-loss rates are $\sim 5.4 \times 10^{-3}$ to $1.4 \times 10^{-2}$ $M_{\odot}$ yr$^{-1}$. These rates exceed those of typical stripped SNe and most SLSN-I, aligning more closely with SLSN-IIs or stripped SNe that evolve into interacting systems via eruptive mass loss.
   • The estimated time of the progenitor mass-loss episode is 5–18 years before explosion, consistent with a shell-like or disk-like CSM structure proposed in previous optical studies.

Significance and Conclusions
The paper concludes that the X-ray emission from SN 2018bsz is unlikely to be powered by a standard millisecond magnetar central engine, as the required parameters to fit the X-ray data contradict those derived from optical modeling, and the predicted ionization breakout timescales are too long.

Instead, the observations favor a scenario where the X-ray emission is dominated by the interaction between the ejecta and a dense, asymmetric (disk-like) CSM. This interpretation is supported by:

• The flat temporal evolution of the X-ray luminosity.
• The spectral evolution from hard to soft.
• High inferred mass-loss rates consistent with eruptive progenitor activity.
• Previous optical results indicating hydrogen emission and interaction with a disk-like CSM.

The authors posit that SN 2018bsz represents a distinct subclass of SLSN-I where CSM interaction is crucial for the strong emission, and where a magnetar, if present, does not dominate the observed high-energy output. This study highlights the necessity of multi-wavelength observations to distinguish between powering mechanisms in extreme supernovae.