SN 2025kg: Engine-Driven Ic-BL Supernova

• SN 2025kg is a luminous stripped–envelope Type Ic supernova with a double-peaked light curve and early blue thermal emission, marking it as a key event in engine-driven explosions.
• Its multi-wavelength observations—from rapid photometric evolution and soft X-ray emission to detailed spectroscopy—support models including cocoon cooling, magnetar powering, and circumstellar interaction.
• Debates over the early peak origin and compositional complexities underscore SN 2025kg’s pivotal role in bridging GRB-SN dynamics with diverse central-engine scenarios.

Searching arXiv for papers on SN 2025kg / EP250108a to ground the article in current literature. SN 2025kg is a luminous stripped–envelope core-collapse supernova classified as a broad-lined Type Ic supernova (SN Ic-BL) and associated with the fast X-ray transient EP250108a, discovered on 2025 January 8 by the Einstein Probe. It is one of the rare supernovae linked to an X-ray flash or fast X-ray transient, and it has become a reference event for the study of engine-driven stripped-envelope explosions with double-peaked optical light curves, early blue thermal emission, and possible dense circumstellar interaction (Aguilar et al., 28 Jul 2025). Its observational interpretation is not unique: published analyses variously emphasize hydrodynamical magnetar powering, cocoon cooling from a weak or choked jet, and interaction with extended circumstellar material (CSM), but they converge on SN 2025kg as an unusually well-observed Ic-BL event at the interface between low-luminosity GRB-like explosions, X-ray flashes, and fast X-ray transients (Srinivasaragavan et al., 24 Apr 2025).

1. Discovery, identification, and classification

EP250108a was detected by the Einstein Probe as a fast X-ray transient or X-ray flash on 2025 Jan 8, 12:30:28.34 UT, and the detection time is commonly taken as the explosion epoch in subsequent modeling (Aguilar et al., 28 Jul 2025). Follow-up optical observations revealed the transient later designated SN 2025kg. Spectroscopy established a broad-lined Type Ic classification, with hydrogen- and helium-free optical spectra in the canonical classification sense and broad absorption lines indicating high expansion velocities (Aguilar et al., 28 Jul 2025).

A secure redshift of $z = 0.17641 \pm 0.0003$ was measured from host-galaxy emission lines, with a luminosity distance $D_L = 880.6\ {\rm Mpc}$ or $D_L \simeq 881\ \mathrm{Mpc}$ under the cosmological assumptions adopted in the discovery analyses (Eyles-Ferris et al., 11 Apr 2025). At this redshift SN 2025kg became the closest known supernova discovered following an Einstein Probe fast X-ray transient (Rastinejad et al., 11 Apr 2025).

The event occupies a specific phenomenological niche. It is not a classical long GRB, because no prompt gamma-ray counterpart was detected, yet its X-ray and optical properties place it close to low-luminosity GRB and X-ray-flash supernovae such as SN 2006aj/XRF 060218 (Srinivasaragavan et al., 24 Apr 2025). This has made SN 2025kg central to the emerging view that some Ic-BL explosions host weak, low-efficiency, or choked jets whose prompt high-energy signatures appear primarily in soft X-rays rather than in gamma rays (Eyles-Ferris et al., 11 Apr 2025).

2. Photometric evolution and multi-wavelength behavior

The defining photometric property of SN 2025kg is its double-peaked light curve. The first peak occurs during the first $\lesssim 6$ days and is described as an early cooling or interaction bump; the second is the main supernova peak around $\sim 10$–15 rest-frame days (Aguilar et al., 28 Jul 2025). Around the main peak, different bolometric reconstructions agree well, whereas at very early times they diverge, likely because of differing treatments of UV flux and bolometric corrections (Aguilar et al., 28 Jul 2025).

The first peak was very blue and very luminous. One analysis quotes $M_g \approx -19.5$ mag and $M_r \approx -19.1$ mag for the initial bump, with $g-r \approx -0.4$ mag, followed by reddening as the first peak faded (Srinivasaragavan et al., 24 Apr 2025). The main radioactive peak reached $M_r = -19.39 \pm 0.02$ and $M_g = -18.95 \pm 0.06$, brighter than the mean Ic-BL population but still within the observed range (Srinivasaragavan et al., 24 Apr 2025). Another study reports a bolometric peak luminosity $D_L = 880.6\ {\rm Mpc}$0 at $D_L = 880.6\ {\rm Mpc}$1 days after explosion in the rest frame (Zhu et al., 24 Jul 2025).

Early UV–optical spectral energy distributions are well described by a rapidly expanding cooling blackbody. Representative fits give $D_L = 880.6\ {\rm Mpc}$2 K and $D_L = 880.6\ {\rm Mpc}$3 cm at 0.97 days, with $D_L = 880.6\ {\rm Mpc}$4, while by 3.31 days the temperature had cooled to $D_L = 880.6\ {\rm Mpc}$5 K and the radius had expanded to $D_L = 880.6\ {\rm Mpc}$6 cm (Eyles-Ferris et al., 11 Apr 2025). A time-dependent fit implies an average expansion speed over the first 0.5 days of $D_L = 880.6\ {\rm Mpc}$7, suggesting mildly relativistic ejecta during the earliest phase (Eyles-Ferris et al., 11 Apr 2025).

The associated X-ray transient was soft and long-lived compared with classical GRBs. Reported values include a duration of $D_L = 880.6\ {\rm Mpc}$8 s or $D_L = 880.6\ {\rm Mpc}$9 s, peak 0.5–4 keV luminosity of order $D_L \simeq 881\ \mathrm{Mpc}$0, and total radiated X-ray energy of order $D_L \simeq 881\ \mathrm{Mpc}$1 (Eyles-Ferris et al., 11 Apr 2025). Radio follow-up yielded only upper limits, including MeerKAT non-detections at 3.06 GHz and VLA limits at 10 GHz, excluding standard bright on-axis GRB-like afterglows but remaining consistent with low-energy jets, choked jets, or low-density environments (Srinivasaragavan et al., 24 Apr 2025).

This combination—soft X-ray prompt emission, weak or absent gamma rays, blue early optical cooling, and a GRB-SN-like main peak—has motivated comparisons to both XRF-associated supernovae and engine-driven Ic-BL events more generally (Rastinejad et al., 11 Apr 2025).

3. Spectroscopy, velocities, helium, and hydrogen signatures

Optical spectroscopy shows the transition from a blue continuum-dominated transient to a typical Ic-BL spectral sequence (Srinivasaragavan et al., 24 Apr 2025). Broad absorption features attributed to Fe II and Si II appear prominently, and template matching links SN 2025kg closely to SN 2002ap, SN 2006aj, and SN 1998bw at different epochs (Rastinejad et al., 11 Apr 2025). The Fe II $D_L \simeq 881\ \mathrm{Mpc}$2 velocity evolution, commonly used as a photospheric proxy in stripped-envelope supernovae, declines from $D_L \simeq 881\ \mathrm{Mpc}$3 at $D_L \simeq 881\ \mathrm{Mpc}$4 days to $D_L \simeq 881\ \mathrm{Mpc}$5 at $D_L \simeq 881\ \mathrm{Mpc}$6 days (Srinivasaragavan et al., 24 Apr 2025).

Near maximum light, JWST/NIRSpec prism spectroscopy from 0.5 to 5 $D_L \simeq 881\ \mathrm{Mpc}$7 revealed two unusual features for an Ic-BL event: weak He I absorption near $D_L \simeq 881\ \mathrm{Mpc}$8 and $D_L \simeq 881\ \mathrm{Mpc}$9, and a broad unidentified emission structure around 4–4.5 $\lesssim 6$0 (Rastinejad et al., 11 Apr 2025). Multi-component fitting disfavors interpreting the $\lesssim 6$1 absorption as only C I and Mg II, because that would imply velocities $\lesssim 6$2 km s$\lesssim 6$3, substantially larger than the optical Fe/Si velocities; a model including He I yields more consistent velocities of $\lesssim 6$4 km s$\lesssim 6$5 (Rastinejad et al., 11 Apr 2025). The reported conclusion is that He I is almost certainly present, though likely at low to moderate strength, with a conservative estimate $\lesssim 6$6 (Rastinejad et al., 11 Apr 2025).

The 4–4.5 $\lesssim 6$7 emission remains unidentified. Thermal dust or r-process-powered emission is rejected because a blackbody explanation would require $\lesssim 6$8 and $\lesssim 6$9 by $\sim 10$0 rest-frame days, which is inconsistent with the measured ejecta kinematics (Rastinejad et al., 11 Apr 2025). A plausible suggestion in the published analysis is that it may arise from an Fe-group nebular blend, but no definitive identification is given (Rastinejad et al., 11 Apr 2025).

A further distinctive feature is transient broadened H$\sim 10$1 at $\sim 10$2 days. A broad H$\sim 10$3-like component was detected in one epoch but not in neighboring spectra (Zhu et al., 24 Jul 2025). Its line width implies $\sim 10$4, comparable to the Si II velocity at the same phase, which argues that the hydrogen resides in inner, low-velocity ejecta rather than in a distant high-velocity interaction zone (Zhu et al., 24 Jul 2025). Other analyses interpret the same phenomenon as evidence for interaction with hydrogen-rich material at $\sim 10$5 cm, perhaps in a thin shell or clumps (Rastinejad et al., 11 Apr 2025). These interpretations are not mutually exclusive, but they locate the hydrogen in a geometrically or temporally localized component rather than a canonical extended hydrogen envelope.

4. Models for the early peak: cooling emission, cocoon emission, and circumstellar interaction

The early optical bump has been modeled in several ways. One line of work describes the first $\sim 10$6 days as a rapidly expanding cooling blackbody and argues that the observed X-ray and radio properties are consistent with a collapsar-powered jet that is low energy $\sim 10$7 and/or fails to break out of dense surrounding material (Eyles-Ferris et al., 11 Apr 2025). In that interpretation, the early optical emission is favored to arise from a shocked cocoon rather than from ordinary supernova ejecta alone (Eyles-Ferris et al., 11 Apr 2025).

Analytic cocoon fits to the early light curve yield a cocoon mass $\sim 10$8, cocoon kinetic energy $\sim 10$9, opening angle $M_g \approx -19.5$0, and shock breakout radius $M_g \approx -19.5$1 (Eyles-Ferris et al., 11 Apr 2025). These values are consistent with shock breakout from a compact Wolf–Rayet-like progenitor and with the requirement that only $M_g \approx -19.5$2 of material at $M_g \approx -19.5$3 is needed to reproduce the prompt X-ray light curve in jet-driven models (Eyles-Ferris et al., 11 Apr 2025).

A second line of work jointly fits the optical light curve with radioactive powering plus an early shock-cooling or cocoon component and compares three scenarios: SN ejecta interacting with extended CSM, a cocoon from a jet choked in the stellar envelope, and a cocoon from a jet choked in extended CSM (Srinivasaragavan et al., 24 Apr 2025). In the extended CSM interaction model, the inferred envelope parameters are $M_g \approx -19.5$4, $M_g \approx -19.5$5, and $M_g \approx -19.5$6 (Srinivasaragavan et al., 24 Apr 2025). In the jet-choked-in-CSM model, the corresponding values are $M_g \approx -19.5$7, $M_g \approx -19.5$8, and $M_g \approx -19.5$9 (Srinivasaragavan et al., 24 Apr 2025).

Both of those CSM-based scenarios naturally reproduce a soft X-ray shock breakout with $M_r \approx -19.1$0, duration $M_r \approx -19.1$1 s, and characteristic observed temperature near 1.2 keV, close to the observed EP250108a properties (Srinivasaragavan et al., 24 Apr 2025). By contrast, a cocoon confined to the stellar envelope in the Nakar–Piran formalism does not by itself naturally yield such a soft, long-lived prompt X-ray signal (Srinivasaragavan et al., 24 Apr 2025).

Hydrodynamical modeling offers a related but numerically distinct picture. In a preferred “Mag+CSM” solution, the early cooling phase is reproduced by a dense wind-like CSM with $M_r \approx -19.1$2, $M_r \approx -19.1$3, and $M_r \approx -19.1$4, attached to a compact progenitor of radius $M_r \approx -19.1$5 (Aguilar et al., 28 Jul 2025). This larger inferred CSM radius reflects differences in modeling assumptions and parameterization rather than a settled consensus.

These variations underscore a central point: the early bump is widely attributed to shock-heated extended material, but whether that material is best described as a cocoon, a dense CSM shell, a wind-like CSM, or some hybrid remains unresolved.

5. Power source of the main peak: radioactive nickel versus magnetar injection

The main peak of SN 2025kg poses a luminosity problem for purely radioactive models. In 1D radiation–hydrodynamical calculations, a Ni-only model reproduces the main peak and intermediate-time photospheric velocities with $M_r \approx -19.1$6, $M_r \approx -19.1$7, and $M_r \approx -19.1$8 (Aguilar et al., 28 Jul 2025). The implied ratio $M_r \approx -19.1$9 is explicitly identified as problematic, because it requires nearly half of the ejecta to be radioactive nickel (Aguilar et al., 28 Jul 2025).

A related semi-analytic treatment reaches a similar conclusion: matching $g-r \approx -0.4$0 and $g-r \approx -0.4$1 with radioactivity alone implies $g-r \approx -0.4$2 and $g-r \approx -0.4$3, above the typical range for SNe Ic-BL and even above the 0.025–0.3 range quoted from recent collapsar simulations (Zhu et al., 24 Jul 2025). This is the main reason published studies have explored central-engine power.

The preferred hydrodynamical alternative is a magnetar + CSM model with $g-r \approx -0.4$4, $g-r \approx -0.4$5, $g-r \approx -0.4$6, magnetar spin period $g-r \approx -0.4$7, and dipole magnetic field $g-r \approx -0.4$8 (Aguilar et al., 28 Jul 2025). This model reproduces the main bolometric peak, early decline, and Fe II velocity evolution while requiring a much more ordinary nickel mass (Aguilar et al., 28 Jul 2025). The close similarity of these parameters to those inferred for the GRB-SN 2023pel is explicitly cited as support for a magnetar scenario (Aguilar et al., 28 Jul 2025).

A separate magnetar-based study, combining magnetar injection, radioactive decay, and outer-cocoon cooling, derives tighter semi-analytic parameters from MCMC fitting: $g-r \approx -0.4$9, $M_r = -19.39 \pm 0.02$0, $M_r = -19.39 \pm 0.02$1, $M_r = -19.39 \pm 0.02$2, and $M_r = -19.39 \pm 0.02$3 (Zhu et al., 24 Jul 2025). In that framework the magnetar initial rotational energy is $M_r = -19.39 \pm 0.02$4, and $M_r = -19.39 \pm 0.02$5 of that energy is converted into ejecta kinetic energy, yielding a total kinetic energy $M_r = -19.39 \pm 0.02$6 (Zhu et al., 24 Jul 2025).

By contrast, purely radioactive fits in the observational papers give more modest and model-dependent results. An Arnett-like fit to the second peak yields $M_r = -19.39 \pm 0.02$7, $M_r = -19.39 \pm 0.02$8, and $M_r = -19.39 \pm 0.02$9 (Srinivasaragavan et al., 24 Apr 2025), while another study reports a conservative nickel mass range $M_g = -18.95 \pm 0.06$0 depending on whether one uses a one-zone Arnett model or a multi-zone Ni-mixing model (Rastinejad et al., 11 Apr 2025).

The technical disagreement is therefore not about whether radioactivity contributes—it certainly does—but about whether it can plausibly dominate the peak. Hydrodynamical and engine-based studies argue that the required nickel fraction is too high, favoring magnetar input; more phenomenological light-curve fits can reproduce the data with radioactive models, but only within broad parameter uncertainties.

6. Progenitor system, central engine, and relation to other events

SN 2025kg has been linked to several progenitor scenarios, all involving a stripped massive star. Hydrodynamical models use compact H-free CO-core progenitors with $M_g = -18.95 \pm 0.06$1, favoring a progenitor with $M_g = -18.95 \pm 0.06$2 in the preferred magnetar+CSM solution (Aguilar et al., 28 Jul 2025). Observational and semi-analytic studies instead emphasize a low-mass helium star with an extended helium envelope and pre-supernova mass $M_g = -18.95 \pm 0.06$3, consistent with the He I detections and the inferred radius $M_g = -18.95 \pm 0.06$4 (Zhu et al., 24 Jul 2025).

The binary-evolution interpretation is especially explicit in one Letter, which argues that near-solar host metallicity disfavors quasi-chemically homogeneous evolution and instead points to a close helium-star + main-sequence binary formed through isolated binary evolution (Zhu et al., 24 Jul 2025). In that picture, tidal torques in an orbit with $M_g = -18.95 \pm 0.06$5 days can spin up the helium star sufficiently to form a millisecond magnetar, while a main-sequence companion can also supply hydrogen-rich material that later produces broad H$M_g = -18.95 \pm 0.06$6 (Zhu et al., 24 Jul 2025). A plausible implication is that SN 2025kg probes not only central-engine physics but also the role of binary angular-momentum transfer in metal-rich collapsars.

The event is repeatedly compared with SN 2006aj, SN 2020bvc, SN 2023pel, and EP240414a/SN 2024gsa. Hydrodynamical comparison finds that SN 2025kg and SN 2023pel have strikingly similar bolometric light curves and Fe II velocities, with almost identical best-fit parameters in a magnetar scenario (Aguilar et al., 28 Jul 2025). By contrast, SN 2006aj and SN 2020bvc can be modeled with Ni+CSM solutions alone, whereas SN 2025kg is too luminous for its velocities and ejecta mass if powered only by Ni (Aguilar et al., 28 Jul 2025).

SN 2025kg also serves as a comparison standard for later XRF-SN discoveries. In the study of XRF 241001A/SN 2024aiiq, SN 2025kg is treated as a benchmark NIR spectral analogue: the JWST spectrum of SN 2024aiiq shows excellent agreement with SN 2025kg in continuum shape and broad features, especially from 1.5 to 5 $M_g = -18.95 \pm 0.06$7, and both objects display a broad $M_g = -18.95 \pm 0.06$8 feature and a shallower $M_g = -18.95 \pm 0.06$9 absorption that may indicate a small amount of helium in the ejecta (Schneider et al., 22 Apr 2026). This suggests that “SN 2025kg-like” events may define a subclass of soft X-ray transient–associated Ic-BL supernovae with GRB-SN-like optical properties but unusually informative NIR signatures.

Population arguments remain provisional, but early Einstein Probe discoveries imply that FXT-SNe may be more common than successful GRB jets. One observational paper states that the sample of EP FXT SNe supports rate estimates that low-luminosity jets seen through FXTs are more common than successful GRB jets and that similar FXT-like signatures are likely present in at least a few percent of the brightest Ic-BL SNe (Rastinejad et al., 11 Apr 2025). This suggests that SN 2025kg may be representative not of an isolated peculiarity but of a broader, previously under-sampled channel of engine-driven stellar collapse.

7. Open questions and interpretive tensions

Several aspects of SN 2025kg remain unsettled. The first concerns the main power source of the supernova peak. Hydrodynamical and engine-based models regard a Ni-only solution as formally possible but physically suspect because of the extreme required nickel fraction (Aguilar et al., 28 Jul 2025), whereas more empirical light-curve models recover radioactive parameters within the broad Ic-BL/GRB-SN range (Srinivasaragavan et al., 24 Apr 2025). The absence of very late-time photometry and nebular spectroscopy is repeatedly identified as limiting tighter constraints on $D_L = 880.6\ {\rm Mpc}$00 (Zhu et al., 24 Jul 2025).

The second concerns the geometry and composition of the extended material. Some models favor a wind-like CSM with $D_L = 880.6\ {\rm Mpc}$01 extending to $D_L = 880.6\ {\rm Mpc}$02 (Aguilar et al., 28 Jul 2025); others infer $D_L = 880.6\ {\rm Mpc}$03 at radii of a few $D_L = 880.6\ {\rm Mpc}$04 cm (Srinivasaragavan et al., 24 Apr 2025); cocoon fits prefer much smaller shocked masses but still invoke angularly confined mildly relativistic ejecta (Eyles-Ferris et al., 11 Apr 2025). These discrepancies reflect different assumptions about whether the early optical peak is dominated by shocked CSM, a cocoon, or both.

A third tension lies in the classification versus composition. SN 2025kg is spectroscopically Ic-BL, yet JWST spectroscopy provides evidence for weak He I and late-time H$D_L = 880.6\ {\rm Mpc}$05 (Rastinejad et al., 11 Apr 2025). This does not invalidate the Ic-BL classification, which depends primarily on the absence of strong optical H and He in the classical sequence, but it does suggest a more compositionally complex progenitor than the label alone implies.

Finally, the nature of EP250108a remains debated. One model interprets it as off-axis cooling emission from an inner cocoon viewed at $D_L = 880.6\ {\rm Mpc}$06 (Zhu et al., 24 Jul 2025), whereas other studies prefer soft X-ray shock breakout in extended CSM (Srinivasaragavan et al., 24 Apr 2025) or a trapped or low-energy jet whose shocked cocoon powers both the X-rays and the early blue optical light (Eyles-Ferris et al., 11 Apr 2025). All of these frameworks are engine-driven; the controversy is about breakout conditions, viewing geometry, and the radial structure of the surrounding material rather than about whether a central engine was present.

Taken together, the literature establishes SN 2025kg as a pivotal event for understanding broad-lined Type Ic supernovae associated with soft high-energy transients. It combines GRB-SN-like optical luminosity and kinematics with an X-ray-flash-like prompt counterpart, unusually rich infrared spectroscopy, and clear evidence that the outer stellar environment played an essential role in shaping the observed transient (Rastinejad et al., 11 Apr 2025).