Asymmetrical thermonuclear supernovae triggered by the tidal disruption of white dwarfs

Abstract: In a dense star cluster core, a tidal disruption event (TDE) of a white dwarf (WD) can occur if the WD passes within the tidal radius of an intermediate-mass black hole (IMBH). Very close encounters cause extreme tidal compression in the WD, raising temperatures enough to induce runaway fusion and produce a thermonuclear supernova (SN). Using the hydrodynamics code AREPO augmented with a 55-isotope nuclear reaction network, we performed high-resolution simulations of the TDE of a $0.6$ Msun C/O WD by a $500$ Msun IMBH for different values of the scaled impact parameter $b$ (i.e., the ratio of periapsis distance to tidal radius). Closer encounters produce combined TDE+SN events, with a partial burning of $<sup>{12}$C and $<sup>{16}$O into heavier isotopes -- the $<sup>{56}$Ni fractions of the disrupted WD material vary from 1% at $b = 0.19$ to 82% at $b = 0.10$, while wider ones ($b \gtrsim 0.20$) lead to standard TDEs. In all cases, the material away from the denser regions remains unburnt, spanning a wide range of radial velocities. Such WD TDEs also exhibit a central cavity, wherein little material is found below a radial velocity of several $1000 \,\mathrm{km s}<sup>{-1}$. We also performed 1D and 2D radiative-transfer calculations for these WD-TDEs using the codes CMFGEN and LONGPOL, respectively, covering epochs from a few days to one hundred days. We recover the typical rise times and peak luminosities of SNe Ia, but with an extremely strong viewing-angle dependence of both light curves and spectra. At nebular times, isolated strong emission lines like [Ca ii] λλ 7291, 7323 may appear both displaced and skewed by many $1000 \,\mathrm{km s}<sup>{-1}$ -- such extreme offsets are harder to identify at earlier times due to optical depth effects and line overlap. WD TDEs may produce a diverse set of transients with extreme asymmetry and peculiar composition.

• The paper presents a systematic study of how the outcome of a white-dwarf-IMPBHD encounter transitions from a standard TDE to a thermonuclear supernova with a high-speed velocity dependent on the heat of the encounter utilizing a real time nuclear reaction network.
• The authors found that the father the IMBH disruption distance, the larger the output nickel production was, turning the tail end of the White Dwarf in to peculiarly intermediate mass radius elements.
• The study shows viewing-distance dependent observations of the white dwarf's ejecta producing new viewing-angle dependent features depending on whether viewing from the pole or the side of the transient.
• The paper notes that the transition to ignition source solution was dependent on resolution issues of the hydrocel compute code. Higher resolutions produced better thermodynamic approximatons while some lower resolution required subsequent improvements in Nuclear reaction acccuracies.
• The ejecta contains Nickel and other intermediate mass elements that could be useful in nuclear chemistry-related studies

Tidal disruption events (TDEs) of white dwarfs (WDs) by intermediate-mass black holes (IMBHs) occupy a theoretically attractive niche: they are among the few transients that could unambiguously reveal IMBHs, since WDs are swallowed whole by supermassive black holes without observable signatures. The paper by Vynatheya, Dessart, Ryu, and Pakmar presents a systematic numerical study of how the outcome of a WD–IMBH encounter transitions from a standard TDE to a thermonuclear supernova as a function of the scaled impact parameter $b = r_p/r_t$, and how such events would appear observationally. The work combines high-resolution moving-mesh hydrodynamics with an on-the-fly nuclear reaction network and both 1D and 2D radiative transfer, extending previous treatments to nebular epochs and to a much richer compositional description.

Hydrodynamic setup

The simulations use AREPO, an arbitrary Lagrangian–Eulerian code on a Voronoi mesh, augmented with the Helmholtz equation of state and a 55-isotope nuclear reaction network drawn from JINA REACLIB. This is a substantial improvement over earlier work: MacLeod et al.'s 3D Monte Carlo radiative-transfer study used only seven species, which limits the accuracy of both energy release and ejecta composition — quantities that directly control the radiative-transfer predictions.

The progenitor is a $0.6\,M_\odot$ C/O WD (50% $^{12}$C, 50% $^{16}$O), chosen because the observed WD mass distribution peaks near this value, resolved with $\sim 6\times10^6$ cells at $10^{-7}\,M_\odot$/cell after convergence testing. The IMBH is a non-rotating $500\,M_\odot$ point mass treated with Newtonian gravity; for periapsis distances of $\sim 5.5$–$11\,r_g$, relativistic corrections are expected to be mild. Eight parabolic encounters were run with $b$ from 0.20 down to 0.10, each evolved 200 dynamical timescales ($0.6\,M_\odot$0 s at $0.6\,M_\odot$1), i.e., roughly 500 s past periapsis.

Transition from TDE to thermonuclear explosion

The central result is a sharp dependence of nucleosynthesis on $0.6\,M_\odot$2. Wider encounters ($0.6\,M_\odot$3) produce standard full TDEs with negligible burning; closer ones ignite runaway fusion during peak tidal compression at periapsis, producing combined TDE+SN events. The $0.6\,M_\odot$4Ni mass fraction rises steeply across a narrow range of impact parameters:

Quantity	$0.6\,M_\odot$5	$0.6\,M_\odot$6	$0.6\,M_\odot$7	$0.6\,M_\odot$8	$0.6\,M_\odot$9	$^{12}$0
Unbound mass [$^{12}$1]	0.34	0.36	0.46	0.47	0.50	0.51
Kinetic energy [$^{12}$2 erg]	0.98	0.95	1.10	1.08	1.32	1.60
$^{12}$3 [$^{12}$4]	0.00	0.01	0.14	0.35	0.47	0.49

The largest relative increase in nickel production occurs between $^{12}$5 and $^{12}$6, where the $^{12}$7Ni fraction grows by a factor of 40. At $^{12}$8, roughly 82% of the ejecta is iron-group material and nearly half the WD becomes unbound, versus ~56% in the standard-TDE limit. Intermediate-mass elements (Si, S, Ar, Ca) peak at $^{12}$9–$^{16}$0 before being consumed at deeper encounters. A controlled comparison at $^{16}$1 with the network disabled confirms that nuclear energy injection, not tidal dynamics alone, drives the fanning out of the debris.

The resulting abundance patterns span an unusually wide range relative to canonical SNe Ia: the $^{16}$2Ni fraction extends beyond the ~61–73% typical of Chandrasekhar- or sub-Chandrasekhar-mass delayed-detonation models, covering compositions resembling stripped-envelope Type Ic events (C/O-rich, $^{16}$3), faint 91bg-like SNe Ia, and IGE-dominated 91T-like events. Notably, the authors find that the same $^{16}$4 configuration produced a supernova in Rosswog et al.'s SPH simulations but not here; they attribute the discrepancy to resolution, since lower-resolution runs overproduce fusion due to larger ignition volumes. This is a pointed claim about the reliability of prior low-resolution nucleosynthetic results in this regime.

Ejecta geometry

The debris is strongly asymmetric in all cases: most unbound material lies within a cone of half-opening angle ~30° about the direction opposite the incoming orbit, expanding ballistically at up to $^{16}$5. The ejecta exhibit an onion-like stratification in physical space — $^{16}$6Ni-dominated interiors, intermediate-mass elements in the middle, unburnt C/O outside — but because the plume centers move away from the black hole rather than sitting at rest, this stratification does not map onto radial velocity space. All velocities contain mixed composition, a fundamental structural difference from standard SNe Ia.

Radiative transfer: 1D equivalents

Because AREPO runs stop at ~500 s, the cone-averaged debris was first evolved to homologous expansion (~1 d) with the V1D radiation-hydrodynamics code, including gravity from the central BH and radioactive decay but assuming a dormant BH. Two dynamical effects emerge that have no analogue in ordinary SNe Ia. First, gravitational fallback removes about a tenth of a solar mass within hours, carving a central cavity devoid of material below several thousand km/s — a signature tied directly to the BH mass. Second, layer velocities drop by up to a factor of two while climbing out of the potential well, delaying homologous expansion by hours and slightly lengthening diffusion times.

CMFGEN calculations from 1 d onward yield rise times of ~14–23 d and bolometric peaks of $^{16}$7 to $^{16}$8 erg s$^{16}$9. Spectra follow the expected progression with nickel mass: Fe II/Co II dominance at high $\sim 6\times10^6$0, IME lines (O I, Si II, Ca II NIR triplet) at low $\sim 6\times10^6$1, with [Co II] and [Fe II] dominating nebular spectra in the former case and [O II]/[Ca II] in the latter. Because of enforced spherical symmetry, these 1D models carry ejecta masses 10–20 times the true values (e.g., $\sim 6\times10^6$2 and $\sim 6\times10^6$3 of Ni for $\sim 6\times10^6$4), so their absolute luminosities are inflated; they serve mainly to set the stage for the 2D treatment.

Radiative transfer: viewing-angle effects in 2D

The 2D LONG calculations encode the asymmetry through a latitudinal scaling of opacities and emissivities, reducing ejecta volumes by a factor of ~20 and bringing masses back to the 3D values. Peak $\sim 6\times10^6$5-band magnitudes then fall in the SN Ia range, with rise times clustering near 20 d — slightly longer than comparable standard models, plausibly because BH gravity slows expansion.

The distinctive prediction is extreme viewing-angle dependence. Light curves differ by 2–3 mag between pole-on and side-on views for the same event, persisting even into the nebular phase. At 100 d, strong isolated emission lines ([Ca II], [Fe II], the Ca NIR triplet) are displaced and skewed by 5000–10000 km s$\sim 6\times10^6$6 depending on viewing angle, an effect amplified by the inner velocity cavity. The authors note that analogous blueshifted, displaced emission has been observed in CSS 161010, interpreted there as a stellar TDE — suggesting such line asymmetries may be a generic diagnostic of off-center ejecta around compact objects. At photospheric epochs, optical depth effects and line overlap make these offsets much harder to identify. The 2D models also predict polarization at the ~1% level near maximum light, far above what is measured for standard SNe Ia.

Limitations and open questions

The paper is explicit about its approximations. The 2D axisymmetric treatment ignores the fact that, once asymmetry is allowed, deep ejecta layers are no longer enshrouded by overlying material; using 1D emissivities therefore overestimates temperature, ionization, and brightness for viewing angles away from the cone axis, though the extreme density contrast (inner regions $\sim 6\times10^6$7 times denser than outer layers) mitigates the error. Radioactive decay power deposition is also handled with 1D assumptions, affecting the viewing-angle-dependent luminosity. Enforced homogeneity within radial shells alters coolant balance and suppresses [O II] emission in the lowest-nickel model. More fundamentally, all results derive from a single WD mass ($\sim 6\times10^6$8) and single IMBH mass; outcomes for He WDs, O/Ne WDs, or different IMBH masses remain unexplored, as do relativistic effects at smaller periapsis distances. No direct comparison to the observed SN Ia sample is attempted, partly because no firm SN-Ia-like WD TDE has yet been confirmed.

Conclusion

This work quantifies, with converged hydrodynamics and a realistic network, how a narrow window in encounter geometry converts a WD TDE into a thermonuclear transient whose nickel yield spans essentially the full diversity of hydrogen-deficient supernovae. Its most robust observational predictions — a central velocity cavity tied to BH mass, percent-level polarization, and kilometer-per-second-scale displacement of nebular emission lines under strong viewing-angle dependence — provide concrete targets for identifying WD TDEs in transient surveys, and hence for confirming IMBHs in dense star clusters.