2307.06192v2

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

MNRAS 000, 1–8 (2015) Preprint 7 September 2023 Compiled using MNRAS LATEX style file v3.

Failed supernova simulations beyond black hole formation


Takami Kuroda1 ∗ and Masaru Shibata1,2
1 Max-Planck-Institut für Gravitationsphysik, Am Mühlenberg 1, D-14476 Potsdam-Golm, Germany
2 Center for Gravitational Physics and Quantum-Information, Yukawa Institute for Theoretical Physics, Kyoto University, Kyoto, 606-8502, Japan

Accepted XXX. Received YYY; in original form ZZZ


arXiv:2307.06192v2 [astro-ph.HE] 6 Sep 2023

ABSTRACT
We present an axisymmetric failed supernova simulation beyond black hole formation, for the first time with numerical relativity
and two-moment multi energy neutrino transport. To ensure stable numerical evolution, we use an excision method for neutrino
radiation-hydrodynamics within the inner part of black hole domain. We demonstrate that our excision method is capable to
stably evolve the radiation-hydrodynamics in dynamical black hole spacetime. As a remarkable signature of the final moment
of PNS, we find the emergence of high energy neutrinos. Those high energy neutrinos are associated with the proto-neutron star
shock surface being swallowed by the central black hole and could be a possible observable of failed supernovae.
Key words: (stars:) supernovae: general – stars: black holes – neutrinos – gravitational waves

1 INTRODUCTION BH is predominantly determined by the compactness of the pro-


genitor star, along with the detailed explosion scenario (but see
Massive stellar collapse is one of the main formation channels
Burrows & Vartanyan (2021) for counterexamples).
of stellar-mass black hole (BH), whose existence was observa-
tionally substantiated through numerous coalescence events (e.g. There are currently numerous multi-dimensional simulations re-
Abbott et al. 2016, 2019). Massive stars heavier than ∼ 8 M⊙ un- porting a successful SN explosion (e.g., Müller & Varma 2020;
dergo a catastrophic gravitational core-collapse (CC) at the end stage Burrows et al. 2020; Stockinger et al. 2020; Bollig et al. 2021;
of their evolution. The subsequent evolutionary path is rich in variety Nakamura et al. 2022; Vartanyan et al. 2022). These studies are pri-
and determines the remnant property. Broadly speaking, less to mod- marily directed towards less massive, or more precisely less compact,
erately massive stars explode as core-collapse supernova (CCSN), progenitor stars, in which the canonical neutrino heating mecha-
whereas more massive stars are prone to fail the explosion, some- nism can trigger the explosion, leaving behind a neutron star (NS).
times completely and sometimes exhibiting only a feeble explosion However, there are several observational evidences of a “failed” su-
(Nomoto et al. 2006; Tanaka et al. 2009). At the same time some of pernova (Kochanek et al. 2008; Smartt 2015; Adams et al. 2017).
more massive stars are known to be accompanied by a very ener- These events report a sudden disappearance of red supergiant, in-
getic explosion termed as hypernova (Iwamoto et al. 1998), whose ferring that the whole progenitor star collapses and becomes a BH
explosion energy is about one order of magnitude larger than those without noticeable explosions. Furthermore exceptionally low en-
of canonical SNe. ergy SNe, e.g., SN 2008ha (Valenti et al. 2009; Foley et al. 2009),
The CCSN explosion scenario and the mass range determining were detected, which could possibly be explained by “fallback” dur-
the fate are yet to be fully understood (for reviews, see Janka et al. ing SN explosion (Kawabata et al. 2010; Fryer et al. 2009). Should
2016; Müller 2016; Burrows & Vartanyan 2021). It is evident, how- these events be a gravitational collapse of massive star, the remnant
ever, that unless the explosion possesses sufficient energy to ex- becomes most likely a BH due to their inferred small ejecta mass.
pel substantial amounts of stellar mantle, the central compact rem- These observations associated possibly with a BH formation
nant will ultimately acquire a mass that surpasses the maximum strongly motivate us to explore the failed and fallback SN sce-
mass limit, above which its internal pressure cannot counteract its narios. There were, however, severe numerical difficulties in per-
own self-gravitational force, thereby leading to the formation of forming SN simulations in BH spacetime. First, multi-dimensional
a black hole. The remnant property is tightly connected with its SN simulations in general relativity (GR), for instance with nu-
progenitor mass (Woosley et al. 2002; Heger et al. 2003). In gen- merical relativity, are still minor, e.g., Müller et al. (2010) (and
eral, the more massive the progenitor is, the higher the probabil- its subsequent works) using the so-called conformal flatness con-
ity of being BH is. Moreover, recent parametric studies, focus- dition (CFC) or Kuroda et al. (2016) with a Baumgarte-Shapiro-
ing on the explodability by the standard neutrino heating mecha- Shibata-Nakamura (BSSN) formalism (Shibata & Nakamura 1995;
nism, have revealed that the compactness (O’Connor & Ott 2011) Baumgarte & Shapiro 1999). Since BHs are fundamentally general
could potentially be a good indicator of BH formation (see also, relativistic objects, the formation process, namely from the onset of
e.g., Ugliano et al. 2012; Sukhbold et al. 2016; Müller et al. 2016; gravitational collapse of massive progenitor to BH formation and be-
Ertl et al. 2016; Ebinger et al. 2019). Like these, the formation of a yond, can be precisely followed only by numerical relativity. Second,
sophisticated neutrino transport is essential for modern SN simu-
lations. However, numerical relativity simulation in BH spacetime
∗E-mail: takami.kuroda@aei.mpg.de combined with sophisticated neutrino transport is currently still chal-

© 2015 The Authors


2 T. Kuroda and M. Shibata
lenging. To date, simulations only up to BH formation (Kuroda et al. (Shibata & Nakamura 1995; Baumgarte & Shapiro 1999) with a
2018; Shibagaki et al. 2020; Kuroda et al. 2022) or switching to New- fourth order finite differencing for the spatial derivatives and a
tonian gravity with a large excision region (several times of the four-step Runge-Kutta method. We choose ‘1+log’ slicing condi-
Schwarzschild radius) immediately after BH formation (Chan et al. tion for the lapse and gamma-driver condition for the shift vector
2018; Rahman et al. 2022) are reported. Very recently Sykes et al. (Alcubierre et al. 2003). BH formation is determined by identifying
(2023) reported the first SN simulations solving the full spatial do- the location of apparent horizon (AH) by an AH finder, e.g., Shibata
main above the BH, i.e., without discarding too large computational (1997). After the AH formation, we enforce an excision method for
domain in the vicinity of central BH, based on the CFC metric. radiation-hydrodynamics inside the AH, while we evolve the full
The main obstacle of neutrino transport in BH spacetime, or rather black hole spacetime without excision for geometrical variables.
immediately after BH formation, stems from the rapid change of mat- Here we will briefly explain our excision technique for radiation-
ter field. At the moment of BH formation, the (rest mass) density just hydrodynamics. Once the AH is found, we divide the interior of
above the BH is generally high & 1014 g cm−3 . The density, however, AH into two: inner and outer regions. The interface of these two
quickly decreases to ∼ 1010 g cm−3 within a few ms concomitantly regions is locating at 𝑓 𝑟 AH (𝜃), where 𝑓 ∈ [0, 1] and 𝑟 AH (𝜃) denotes
with the proto-neutron star (PNS) being swallowed by the central the radius of AH at 𝜃-direction with 𝜃 being the angle with respect
BH. This indicates that the region in the vicinity of the BH rapidly to 𝑧-axis. In the outer region, we solve the full neutrino radiation-
shifts from optically thick to thin condition and such extreme con- hydrodynamics in the same way as the outside of AH (i.e. 𝑟 > 𝑟 AH ).
dition makes neutrino transport with full interactions a significantly On the other hand, we excise the inner region and artificially set all
challenging subject. In addition, the matter (and probably also radi- primitive variables, i.e., the rest mass density 𝜌, entropy 𝑠, electron
ation) field inside the BH is typically required to be “excised” for fracion 𝑌𝑒 , spacial components of four-velocity 𝑢 𝑖 , and the zeroth
stable numerical evolution. As of now, however, there is no concrete and first order neutrino radiation moments (𝐸 (𝜈, 𝜀) , 𝐹 𝑖 (𝜈, 𝜀) ), as
method how we should treat the radiation field inside the excised
region and also inside BH for stable numerical evolution.  𝜌   ∼ 0.1𝜌max 
 𝑖
  
In this study, we report our first SN simulation beyond BH forma-  𝑢   0 
   
tion with numerical relativity and multi-energy neutrino transport.  𝑠   ≈ 1.5 𝑘 B baryon−1 
 =  for 𝑟 (𝜃) ≤ 𝑓 𝑟 AH (𝜃). (1)
We use an excision method for both matter and neutrino radiation  𝑌𝑒   ≈ 0.15 
   
fields inside a part of BH domain. Our excision method demonstrates  𝐸 (𝜈, 𝜀)   𝐸 thick (𝜈, 𝜀)

   
 𝐹   𝐹thick (𝜈, 𝜀) 𝑖 
stable evolution immediately after BH formation as well as in the sub-  (𝜈, 𝜀) 𝑖   
sequent BH accretion phase. Furthermore, we find the emergence of
Here 𝜌max represents the maximum rest mass density outside of the
high energy neutrinos associated with the PNS shock surface being
AH, which therefore changes its value with time due to the mass
swallowed by the central BH, which could potentially be a probe
accretion onto BH. Regarding the entropy and electron fraction, we
of the very final moment of PNS. We also show that these high en-
use fixed values taken from typical NS structures. The zeroth and
ergy neutrinos could be detectable by the current and next-generation
first order radiation moments (𝐸 thick (𝜈, 𝜀) , 𝐹thick (𝜈, 𝜀) 𝑖 ) inside the
neutrino detectors if the BH formation happens in our Galaxy.
inner region are enforced to be the moments in the optically thick
This paper is organized as follows. Section 2 starts with a concise
limit (c.f. Eqs. (6.14)–(6.15) in Shibata et al. 2011) assuming the beta
summary of our GR radiation-hydrodynamic scheme with an excision
equilibrium with matter.
scheme and also describe the initial setup of the simulation. The
We shortly touch the appropriate value for 𝑓 . Usually, source terms
main results and detailed analysis of our new findings are presented
for neutrino-matter interactions including gravitational red-shift and
in Section 3. We summarize our results and conclude in Section 4.
Doppler terms are quite stiff. Inside the inner region 𝑟 (𝜃) ≤ 𝑓 𝑟 AH (𝜃),
Throughout the paper, Greek indices run from 0 to 3 and Latin indices
we do not evolve any radiation-matter fields, that is, these stiff source
from 1 to 3, except 𝜈 and 𝜀 which denote neutrino species and energy,
terms are suddenly switched off across the excision boundary. Such
respectively.
artificial treatment inevitably causes spurious behaviours appearing
especially in the radiation fields near the excision boundary. If we
choose the value of 𝑓 to be close to unity, those spurious oscillations
2 METHOD eventually propagate even out to the outside of AH and the simulation
In our full GR radiation-hydrodynamics simulations, we solve the will be crashed. Therefore in this study we set 𝑓 = 0.5 to avoid such
evolution equations of metric, hydrodynamics, and energy-dependent pathological behavior. With these treatments, we found numerically
neutrino radiation. Each of the evolution equations is solved in an stable neutrino radiation-hydrodynamic evolution in BH spacetime.
operator-splitting manner, while the system evolves selfconsistently
as a whole, satisfying the Hamiltonian and momentum constraints
(Kuroda et al. 2016). In Sec. 2.1, we describe our numerical method 2.2 Model
focusing particularly on the excision method applied to the neutrino We use a non-rotating massive star with zero metallicity, whose
radiation-hydrodynamics variables. Sec. 2.2 is devoted to explaining initial mass at its zero-age main sequence is 70 M⊙ (Takahashi et al.
the computed model and numerical setup. 2014). It has a substantially high compactness parameter 𝜉2.5 = 1
(O’Connor & Ott 2011) at the final evolution phase. This progenitor
star was reported to form a BH within a few hundred milliseconds
2.1 Radiation hydrodynamics in BH spacetime
after the first bounce (Kuroda et al. 2018; Shibagaki et al. 2021). We
We solve full GR multi-energy neutrino transport equations in use the DD2 EOS of Typel et al. (2010). The maximum NS mass of
axisymmetric 2 + 1 dimensions (two spatial dimensions and DD2 for cold and non-rotating case is 2.42 M⊙ , which is consistent
one momentum-space dimension). Details of the code are de- with the existence of observationally confirmed massive NSs with ∼
scribed in our previous studies (Kuroda et al. 2016, 2022). The 2 M⊙ (Demorest et al. 2010; Antoniadis et al. 2013; Cromartie et al.
black hole spacetime is evolved using the BSSN formalism 2020).

MNRAS 000, 1–8 (2015)


Failed SN simulations beyond BH 3
ties show a decresing trend, while heavy-lepton neutrinos show a
rapid increase in both its luminosity and mean energy. These features
were previously identified in 1D full-GR simulations with Boltzmann
neutrino transport Liebendörfer et al. (2004) and are commonly ob-
served in the literature, due to rapid contraction of the PNS to the
forming BH (see also, Sumiyoshi et al. (2007); Fischer et al. (2009);
Hempel et al. (2012); Gullin et al. (2022) as well as 3D models by
Kuroda et al. (2018); Shibagaki et al. (2021)). The overall features
before the BH formation are in a good agreement with our former
model 𝑧70 reported in Kuroda et al. (2022), in which the DD2-based
nuclear EOS taking into account a first-order quantum chromody-
namics (QCD) phase transition was used. Taking into account the
fact that the QCD phase transition occurs after the PNS starts col-
lapsing (Kuroda et al. 2022), the agreement between the current and
previous models is quite reasonable.
We also compare BH formation time with previous related stud-
ies. O’Connor & Ott (2011) presented a nice correlation between BH
formation time, obtained from various 1D GR models, and compact-
ness parameter of progenitor star. According to their Fig. 6, mas-
sive stars having 𝜉2.5 = 1, which is the case for the current model,
are forming BH at 𝑡pb ∼ 250 − 750 ms, where the time variation
reflects the different nuclear EOS. Powell et al. (2021) performed
faint SN simulations in 3D using a zero-metallicity progenitor star
Figure 1. Overall evolution feature. Panel (a): the maximum rest-mass den- with 85 M⊙ , whose compactness parameter is 𝜉2.5 = 0.86. They
sity 𝜌max (black), central lapse function 𝛼c (red), and baryon mass of PNS witnessed shock revival prior to BH formation, which to some ex-
𝑀PNS (blue); (b): neutrino luminosity 𝐿𝜈,51 in units of 1051 erg s −1 ; and (c)
tent suppresses subsequent mass accretions onto the PNS and may
neutrino mean energy h 𝜀𝜈 i. Neutrino profiles are evaluated at 𝑟 = 400 km.
delay the BH formation. Their models exhibited BH formation occur-
In panels (b) and (c), the color represents neutrino species: electron type neu-
trino (black), electron type antineutrino (red), and heavy lepton type neutrino ring at 𝑡pb ∼ 290 − 590 ms. Using similar massive progenitor stars,
(blue). Rahman et al. (2022) also demonstrated faint SN scenarios with BH
formation occuring at 𝑡pb ∼ 350 − 400 ms. In addition, a recent study
of Sykes et al. (2023) presented BH formation at 𝑡pb ∼ 220 ms for
The 2D axially symmetric computational domain extends to 1.5 × the same progenitor model used in Powell et al. (2021). Considering
104 km from the center. In the cylindrical computational domain, that our numerical formalism is totally independent from these pre-
2:1 ratio nested boxes with 11 refinement levels are embedded, and vious studies and also that we use a different progenitor model, some
each nested box contains 64 × 64 cells so that the finest resolution time variations in BH formation time are expected to emerge. At the
at the center becomes ≈230 m. In this work, we assume the plane same time, comparing to less massive stars, e.g., with 𝜉2.5 ∼ 0.25,
symmetry with respect to the equatorial plane. The neutrino energy which are predicted to form BH at 𝑡pb &2 s (O’Connor & Ott 2011),
space 𝜀 logarithmically covers from 3 to 400 MeV with 14 energy unless successful shock revival does not occur, all previous studies
bins. In this study, we use the up-to-date neutrino rates of Kotake et al. including this study are presenting consistent BH formation time, i.e.,
(2018), which are used also in our recent studies (Kuroda et al. 2022; substantially quicker than 𝑡pb &2 s expected in less massive stars.
Kuroda & Shibata 2023). Next we discuss the neutrino radiation-hydrodynamics evolution
after the BH formation, focusing mainly on how effectively our exci-
sion method manage to prevent propagation of spurious behaviours
3 RESULTS often appeared at the excision boundary. Fig. 2 displays spherically
averaged spatial profiles of the rest mass density (top-left), elec-
We first describe the picture of post-bounce evolution till the forma- tron fraction (top-right), entropy (middle-left), radial component of
tion of BH. Fig. 1 shows: (a) the maximum rest-mass density 𝜌max,15 the three velocity (middle-right), electron type neutrino luminosity
in units of 1015 g cm−3 (black), baryon mass of PNS 𝑀PNS (blue), (bottom-left), and anti-electron type (solid-line) and heavy-lepton
and central lapse function 𝛼c (red); (b) neutrino luminosity 𝐿 𝜈,51 type (dash-dotted line) neutrino luminosities (bottom-right), at sev-
in units of 1051 erg s−1 for neutrino species; and (c) neutrino mean eral time slices. In the middle-left panel, we supplementary plot a
energy h𝜀 𝜈 i. The PNS surface is defined by the location for which the temperature profile, but only at the formation of BH (red dash-dotted
rest mass density drops below 1010 g cm−3 . 𝐿 𝜈 and h𝜀 𝜈 i are evalu- line), which is used in the later discussion with Fig. 3. Each color
ated from the emergent neutrino spectra measured at 𝑟 = 400 km. In represents the post BH formation time 𝑡BH , denoted in the top-left
panel (a), we also plot the maximum mass of DD2 EOS for cold and panel. Once the AH is formed, we plot structures only aoutside the
non-rotating stars by the horizontal dash-dotted line of 2.42 M⊙ . AH.
Panel (a) exhibits that the 𝑀PNS exceeds the maximum allowed Slightly before AH formation at 𝑡BH = −0.1 ms, the central density
mass of current EOS at 𝑡pb ∼ 100 ms. However, because of an addi- exceeds 1015 g cm−3 and the velocity profile inside the PNS shows
tional contribution from thermal pressure, the PNS does not immedi- the infalling structure. For 𝑡BH ≥ 0 ms, for which we apply an ex-
ately collapse to a black hole. From the maximum density evolution, cision method described in the previous section, we see essentially
we see a sharp increase at 𝑡pb ∼ 177 ms, at the same time 𝛼𝑐 decreases no numerical instabilities at the interface of the AH. All the neu-
to ∼ 0. This signals the BH formation. Prior to the BH formation at trino radiation fields and hydrodynamical variables exhibit smooth
𝑡pb & 160 ms, electron and anti-electron type neutrino luminosi- structures across the AH and subsequently swallowed into its inside.

MNRAS 000, 1–8 (2015)


4 T. Kuroda and M. Shibata

Figure 3. Post BH formation evolution of: (a) the irreducible mass 𝑀irr and
2-norm of Hamiltonian constraint violation | | 𝐻 | | 2, (b) neutrino luminosities,
and (c) mean neutrino energies, as a function of 𝑡BH . In panels (b) and (c),
the color represents neutrino species: electron type neutrino (black), electron
type antineutrino (red), and heavy lepton type neutrino (blue).

Figure 2. Spherically averaged radial profiles of the rest mass density 𝜌


(top-left), electron fraction 𝑌𝑒 (top-right), entropy per baryon 𝑠 (middle- (see blue and magenta lines in the bottom-left panel in Fig. 2), while
left), radial component of the three-velocity 𝑣 𝑟 ≡ 𝑢𝑟 /𝑢𝑡 (middle-right), the rest of neutrino species has essentially no production channel
neutrino luminosity for 𝜈𝑒 (bottom-left), 𝜈¯𝑒 (solid, bottom-right), and 𝜈𝑥 and their neutrino luminosities quickly subside. Sykes et al. (2023)
(solid-dashed, bottom-right) at different times denoted in the top-left panel. In reported a BH excision scheme with neutrino transport. According
the middle-left panel, we also plot a temperature profile, but only at 𝑡BH = 0 ms to their long time failed CCSN simulation in 1D spherical symmetry,
(red dash-dotted line). qualitatively similar spatial profiles of neutrino luminosities, namely
relatively strong 𝜈𝑒 emission continuing even after BH formation,
was also reported.
From the density structural evolution, the maximum density drops Fig. 3 displays: (a) the irreducible mass 𝑀irr and 2-norm of
by four orders of magnitude, from ∼ 1014 g cm−3 to ∼ 1010 g cm−3 , Hamiltonian constraint violation ||𝐻|| 2 , (b) neutrino luminosities,
within a few ms, presenting a clear transition from optically thick and (c) mean neutrino energies, as a function of 𝑡BH . Here, p 𝑀irr
to thin conditions. This feature makes SN simulations in dynamical is defined by the area of apparent horizon 𝐴 as 𝑀irr = 𝐴/16𝜋
BH spacetime one of numerically challenging subjects. We found (cf. Baumgarte et al. 1996; Shibata 1997) and ||𝐻|| 2 measures the
that, if we suddenly switch off the neutrino-matter interactions inside constraint violation only for numerical cells outside the AH. From
the AH, it causes spurious behaviors, which eventually leak out to panel (a), the irreducible mass shows an increasing trend from
the outside and lead to a code crash. Therefore we believe that it is 𝑀irr ∼ 2.88 M⊙ to ∼ 3.06 M⊙ during the first 40 ms. At the moment
essential to ensure a buffer zone between the AH and the excised of the AH formation, the measured value of the protoneutron star
region, especially when the neutrino radiation fields are taken into mass, 𝑀PNS , is ∼ 2.76 M⊙ , which rapidly decreases to . 0.001 M⊙
account. During the first few ms after AH formation, low-𝑌𝑒 and high (the total mass outside of the AH and where 𝜌 ≥ 1010 g cm−3 is met)
entropy material, which represent typical PNS shocked material, are within a few ms. It means that the estimated 𝑀irr is slightly larger
still present outside the AH. They are, however, immediately swal- than 𝑀PNS at 𝑡BH = 0 ms. Furthermore, from panel (a), 𝑀irr ini-
lowed by the BH and for 𝑡BH & 3 ms the BH accretion enters a nearly tially shows a slightly odd behavior, a nearly constant evolution until
steady state, exhibiting high-𝑌𝑒 (∼ 0.49) and relatively low entropy 𝑡BH ∼ 8 ms, and it increases afterward. From these, we naively sus-
(∼ 5 kB baryon−1 ) flows (see magenta lines). pect that the current numerical resolution at the center Δ𝑥 ∼ 230 m
Next we focus on how the neutrino signals in association with might not be high enough1 to accurately resolve the location of ap-
the BH formation are radiated away. Bottom two panels indicate parent horizon and may tend to overestimate the initial BH mass
that all neutrino species have an outgoing flux for 𝑟 & 30 km at the approximately by ∼ 0.1 M⊙ , i.e., ∼ 3 % error in the evaluation for
time of the BH formation. In the vicinity of AH, on the other hand, the total BH mass or the AH radius. However, once the system re-
neutrino radiation fields experience a strong drag by infalling high laxes to a quasi-steady state for 𝑡BH & 10 ms, 𝑀irr increases with a
density component (& 1012 g cm−3 ) and have an inward flux. After reasonable growth rate of 𝑀¤ irr ≈ 4.66 M⊙ s−1 , which agrees approx-
the mass accretion becomes a nearly steady state flow for 𝑡BH & 3 ms, imately with that of the PNS mass, 𝑀¤ PNS ≈ 4.73 M⊙ s−1 , before the
the dominant neutrino-matter interaction is the electron capture due BH formation (see panel (a) in Fig. 1). The 2-norm of Hamiltonian
to continuous replenishment of high-𝑌𝑒 materials (∼ 0.49, see top-
right panel) from stellar mantle. It results in a sustained neutrino
emission even after the BH formation for electron type neutrinos 1 The BH is resolved by ∼ 13 − 14 grid points at its formation.

MNRAS 000, 1–8 (2015)


Failed SN simulations beyond BH 5
constraint ||𝐻|| 2 stays around ∼ 10−4 without any secular increase 𝑡BH ∼ 3 ms. Neutrinos at higher energie bins (𝜀 = 176 and 265 MeV)
after BH formation. also show a sudden increase with slight time delays of ∼ 0.5 ms
Regarding the neutrino signals, the neutrino luminosity for all from the peak time for 𝑓 𝜀=117 . These time delays are mostly due to
species show a rapid distinction and eventually migrate to a quasi that higher energy neutrinos require a longer time for escaping from
steady state for 𝑡BH & 5 ms. From panel (b), 𝐿 𝜈𝑒 stays around ∼ collapsing stellar mantle. On the other hand, regarding 𝜈¯𝑒 (as well as
2 × 1049 erg s−1 till the end of our calculation, which features a long 𝜈𝑒 ), the less population of high energy neutrinos (𝜀 & 50 MeV) prior
term steady state mass accretion onto the BH. Nearly constant 𝐿 𝜈𝑒 of to the BH formation than that of 𝜈 𝑥 (compare two thin lines in panel
the order of O (1049 ) erg s−1 is also reported in Sykes et al. (2023). (a)) leads simply to a less noticeable increase at 𝑡BH ∼ 3 − 4 ms. Ad-
The neutrino mean energy h𝜀 𝜈 i may reveal the final moment of ditionally, the presence of charged current reactions tend to suppress
devastating PNS collapse. As can be clearly seen, h𝜀 𝜈 i for all neutrino their increase. In fact, 𝑓 𝜀≥117 for 𝜈¯𝑒 shows approximately an order
species show a drastic increase at 𝑡BH ∼ 3 ms. This is particularly the of magnitude smaller values than that for 𝜈 𝑥 . These features result
case for heavy lepton type neutrinos, which show a remarkably high in the observed high energy neutrinos pronounced for heavy lepton
mean energy of h𝜀 𝜈𝑥 i ∼ 90 MeV. These values are even higher than type ones (Fig.3). Although our moment formalism cannot capture
those from the QCD CCSN models (Fischer et al. 2018; Kuroda et al. the particle acceleration mechanisms at the shock front, non-thermal
2022), which are also known to emit high energy neutrinos h𝜀 𝜈𝑥 i ∼ shock acceleration (Kazanas & Ellison 1981; Giovanoni et al. 1989;
40 MeV due to strong shock heating in association with the quark Nagakura & Hotokezaka 2021) is also reported to excite high energy
core bounce. We will now shortly discuss their possible excitation neutrinos from CCSNe.
mechanism. First, since we measure the emergent neutrino signals As a comparison with previous studies, Gullin et al. (2022) has
at 𝑟 = 400 km, these high energy neutrinos are produced at 𝑡BH ∼ perofrmed a GR Monte Carlo neutrino transport and reported high
1 − 2 ms. From Fig. 2, this time corresponds exactly to the time when energy neutrinos with h𝜀 𝜈𝑥 i ∼ 50 MeV in association with BH for-
huge amounts of hot PNS envelope together with a shock surface mation. Since their calculations are performed on the fixed spacetime
infall with a relativistic speed of ∼ 0.3𝑐. The highest temperature of and matter fields after BH formation, quantitative differences in h𝜀 𝜈 i
collapsing PNS material (middle-left panel in Fig. 2) for the regions from ours are inevitable. We, however, believe that the emission of
of 𝑟 & 30 km, where 𝐹𝜈𝑥 has a positive sign (bottom-right panel) high energy neutrinos just after the BH formation seem to be a com-
and can contribute to the emergent neutrino spectrum, is merely mon feature and might be used as a smoking gun of infall of PNS
𝑇 ∼ 10 MeV. It indicates that heavy lepton type neutrinos, whose surface. Rahman et al. (2022) performed CCSN simulations with
energy are h𝜀 𝜈𝑥 i ∼ 30 MeV, could be barely explained via such as BH formation. However, since they excise the innermost 400 km
pair production channel, although it is not likely for much higher once they find the AH and also their models present a successful
neutrino energy of ∼ 90 MeV. shock expansion, i.e., corresponding to the fallback SN model, the
To further discuss their origin, we examine their spectral features. emergence of high energy neutrinos similar to ours was not reported.
Fig. 4 depicts: (a) the distribution function 𝑓 𝜀 2 for 𝜈¯𝑒 (black lines) Finally, we discuss observable multi messenger signals for a cur-
and 𝜈 𝑥 (red lines) at three different time slices: 𝑡BH = 0 ms, 3 ms rent failed CCSN model. Fig. 5 displays from top: (a) the neu-
(corresponding to the time when high energy neutrinos are observed), trino detection rate Γ of Hyper-Kamiokande (HK) (Abe et al. 2011;
and 7 ms, (b) time evolution of distribution function 𝑓 𝜀 for all energy Hyper-Kamiokande Proto-Collaboration et al. 2018); (b) Γ of Ice-
bins higher than 𝜀 ≥ 52 MeV (this time, 52, 78, 117, 176, and Cube (IC) (Abbasi et al. 2011; Salathe et al. 2012); (c) matter origin
265 MeV) (solid lines) and mean energy h𝜀i (dashed line) for 𝜈¯𝑒 , and gravitational waves (GWs) 𝐷ℎ+ ; and (d) spectrogram of ℎ+ obtained
(c) same as the panel (b) but for 𝜈 𝑥 . All these values are measured at by a short-time Fourier transform. We assume a source distance of
𝑟 = 400 km. 𝐷 = 10 kpc. ℎ+ is the gravitational wave strain, which is calcu-
From panel (a), the energy spectrum at 𝑡BH = 3 ms for 𝜈 𝑥 exhibits lated from a standard quadrupole formula, and we show only the
a flatter profile with relatively more populations for neutrinos with non-vanishing component in axisymmetric profile observed along
& 50 MeV. Such feature cannot be seen in other two time snapshots. the equatorial plane. The neutrino detection rate Γ is evaluated
We attribute the flatter profile to a consequence of more effective in the same way as Kuroda et al. (2022) assuming a Fermi-Dirac
isoenergy scatterings taking place in the upstream to the relativisti- distribution for the neutrino energy spectrum (Lund et al. 2010;
cally infalling shock surface. Because of the rapid infall of the PNS Takiwaki & Kotake 2018). Note that in the evaluation for Γ, we
shock surface (see 𝑣𝑟 -profiles from 𝑡BH = −0.1 ms to 1 ms in Fig. 2), consider two extreme cases: all 𝜈¯𝑒 emitted from the source reach
the outgoing comoving neutrino flux ahead of the shock becomes the detectors without neutrino flavor conversion and cause the signal
relatively larger. Consequently the effect of isoenergy neutrino scat- at the detectors (black lines in the figure); all 𝜈¯ 𝑥 (identical to 𝜈 𝑥 in
terings becomes more prominent compared to the case with a station- this study) emitted from the source are completely swapped by 𝜈¯𝑒
ary shock surface. Furthermore, that impact is more visible for high and cause the signals (red lines). In inset of the upper two panels,
energy neutrinos as the cross section of the isoenergy scatterings is we show a magnified view of Γ relative to BH formation time 𝑡BH to
proportional to the square of the incoming neutrino energy. Indeed, feature detection of high energy neutrinos.
from panel (c), the distribution function for heavy lepton type neu- Regarding the neutrino detection rate Γ, both of the two extreme
trinos shows an increase(decreasing) trend for 𝜀 ≥ 117(≤ 78) MeV cases, i.e., with and without neutrino flavor conversion, essentially
at 𝑡BH . 3 ms. Particularly at the energy bin 𝜀 = 117 MeV ( 𝑓 𝜀=117 : show a quantitatively similar monotonic increase until the BH forma-
red line), its increase is noteworthy with its maximum appearing at tion. This feature can be seen for both detectors. This indicates that
the possible range of neutrino oscillation effects (see Mirizzi et al.
2 2016, for a review), i.e. the region bounded by two lines in panels
We reconstruct the distribution function 𝑓 𝜀 simply by 𝑓 𝜀 = 𝐽 𝜀 /4 𝜋 𝜀 3 ,
where 𝐽 𝜀 denotes the zeroth order neutrino radiation moment measured in (a,b), is quite small, compared to previous studies using less mas-
the comoving frame at the energy bin 𝜀. With an appropriate closure relation, sive progenitor stars (e.g. Tamborra et al. 2012; Kuroda et al. 2022).
𝜇
𝐽 𝜀 is determined from the zeroth and first order radiation momenta (𝐸 𝜀 , 𝐹𝜀 ), For instance, Γ𝜈¯𝑒 →𝜈¯𝑒 becomes ∼ 1.5 times higher than Γ𝜈¯ 𝑥 →𝜈¯𝑒 for
which are measured in the Eulerian frame and are the basic variables evolved 𝑡pb & 100 ms for CCSN models with less massive progenitor stars,
in our M1 neutrino transport. while the current one with a more massive progenitor star presents

MNRAS 000, 1–8 (2015)


6 T. Kuroda and M. Shibata

Figure 4. From left: (a) the distribution function 𝑓 𝜀 for 𝜈¯𝑒 (black lines) and 𝜈𝑥 (red lines) at three different time slices: 𝑡BH = 0 ms, 3 ms (corresponding to the
time when high energy neutrinos are observed), and 7 ms, (b) time evolution of distribution function 𝑓 𝜀 for all energy bins higher than 𝜀 ≥ 52 MeV (this time,
52, 78, 117, 176, and 265 MeV) (solid lines) and mean energy h 𝜀i (dashed line) for 𝜈¯𝑒 , and (c) same as the panel (b) but for 𝜈𝑥 . All these values are measured
at 𝑟 = 400 km.

∼ 100 cm are emitted presenting a broad band emission. Once the


BH is formed and BH accretion settles into a quasi steady state
for 𝑡BH & 3 ms, we observe essentially no GWs for the current
non-rotating model. As a comparison to a previous 2D GR study
(Rahman et al. 2022), which performed faint SN simulations using
an 80 M⊙ progenitor star, the current GWs are showing consistent
behaviors in the initial convection phase (𝑡pb . 50 ms). During this
phase, the amplitude and typical frequency reach 𝐷ℎ ∼ 30 − 40 cm
and 𝐹 ∼ 100 Hz, respectively, in their non-rotating model. These
values are quite consistent with our findings. Although a direct com-
parison in the subsequent phase (𝑡pb & 50 ms till BH formation)
may not be so meaningful, as their models are faint SN, i.e., ex-
hibiting shock revival before BH formation, high frequency GWs
(𝐹 ∼ 1000 Hz) are also observed prior to BH formation, which could
potentially be another common feature.

4 SUMMARY
Figure 5. From top: (a) the neutrino detection rate Γ of Hyper-Kamiokande
We have presented a results of 2D axisymmetric CCSN simulation
(HK); (b) Γ of IceCube (IC); (c) matter origin GWs 𝐷ℎ+ ; and (d) spectrogram
for a massive star with 70 M⊙ . Our core-collapse supernova model is
of ℎ+ obtained by a short-time Fourier transform. We assume a source distance
of 𝐷 = 10 kpc. based on numerical relativity, which solves the GR neutrino-radiation
hydrodynamics equations together with the two-moment (M1) neu-
trino transport equations of Kuroda et al. (2016). We used up-to-
roughly comparable values. Another remarkable feature is rapid in- date neutrino opacities following Kotake et al. (2018) and employed
crease of Γ𝜈¯ 𝑥 →𝜈¯𝑒 (red lines) as the PNS approaches BH formation the DD2 EOS of Typel et al. (2010). In this framework, we follow
(𝑡pb & 150 ms). It is a clear signature of the increasing behavior of for the first time “beyond BH formation”. To ensure stable numer-
both 𝐿 𝜈𝑥 and h𝜀 𝜈𝑥 i shown in Fig. 1. We also discuss if the high en- ical evolution, we use an excision method for neutrino radiation-
ergy heavy lepton type neutrinos, as a possible signature of the shock hydrodynamics, while we evolve the geometrical variables for entire
surface being swallowed by BH, could be observed. From insets, we computational domain.
can marginally observe a slight increase for Γ𝜈¯ 𝑥 →𝜈¯𝑒 (red lines) at Our results showed consistent PNS evolution and multi-messenger
𝑡BH ∼ 3 ms, which is more visible for IC. This time is consistent with signals during the PNS contraction phase with previous studies, for
the emission time of high energy neutrinos (see panel (c) in Fig. 3). which the same progenitor model was used (Kuroda et al. 2018;
If we could observe such a tentative increase of neutrino detection Shibagaki et al. 2021; Kuroda et al. 2022). The current non-rotating
during the exponential decay, it could be a possible signature of the PNS model exceeds the maximum NS mass for DD2 EOS at ∼ 100 ms
aforementioned final moment of the PNS shock surface. after bounce. Subsequently, it initiates the second gravitational col-
Bottom two panels show the emitted GWs. We see essentially the lapse, resulting in BH formation at 𝑡pb ∼ 177 ms. After we identify
same features as have been discussed for model 𝑧70 in Kuroda et al. the AH, our excision technique demonstrates its capability to stably
(2022). During the first ∼ 50 ms after bounce, relatively large and evolve the radiation-hydrodynamics in dynamical BH spacetime. We
low frequency GWs originated from postbounce convective motions solve the full neutrino-matter interactions taking into account the
are observed, whose amplitudes and frequencies reach ∼ 50 cm and gravitational redshift and Doppler terms from the AH down to the
∼ 100 Hz, respectively. Afterward the gravitational waveform shows excision domain, so that spurious oscillations often appearing around
a considerable subsidence, which is then disrupted at 𝑡pb & 120 ms. the excision surface do not leak outside the AH. We also mention
At the moment of BH formation, burst like GWs of the order of that our current numerical method satisfies the Hamiltonian con-

MNRAS 000, 1–8 (2015)


Failed SN simulations beyond BH 7
straint well and its violation after BH formation is free from secular in the context of collapsar scenario, the impact of magnetic fields
growth. threading the central BH is undoubtedly worth to be explored as a
After the BH formation, the PNS envelope was simply swallowed possible origin of relativistic jets generated via, e.g., the Blandford-
by the BH and the system transitions to a nearly steady BH-accretion Znajek mechanism (Blandford & Znajek 1977). It has been recently
phase within a few ms. Afterward the BH mass, i.e. the area of demonstrated by Christie et al. (2019); Hayashi et al. (2022) that the
AH, gradually increases because of the continuous mass inflow. The Blandford-Znajek mechanism is a promising mechanism for launch-
accretion flow is composed of high-𝑌𝑒 (∼ 0.5) material, reflecting ing a jet, but only in the framework of compact mergers. We will
the component of progenitor core (i.e. iron). explore this fascinating topic in our future CCSN studies.
On the contrary to the simple collapse dynamics of PNS, its impact
on the emergent neutrino signals was not so trivial. Our findings are:
(1) neutrinos with significantly high energies, especially for heavy
ACKNOWLEDGEMENTS
lepton type neutrinos whose mean energy reaches ∼ 90 MeV, are
observed during the infall phase of PNS envelope and (2) a steady We acknowledge K. Kiuchi, S. Fujibayashi, and A. Betranhandy for
state neutrino emission of electron type neutrinos in the BH accre- fruitful discussions. This work was in part supported by Grant-in-Aid
tion phase. Possible observations of high energy neutrinos from BH for Scientific Research (Nos. 20H00158 and 23H04900) of Japanese
formation are also reported in a previous similar (but spherical sym- MEXT/JSPS. Numerical computations were carried out on Sakura
metric) study by Gullin et al. (2022). We attribute the first feature to and Raven clusters at Max Planck Computing and Data Facility.
more efficient isoenergy scatterings between neutrinos, which strive
to emerge from the shock surface, and infalling stellar mantle ahead
of the shock, which is mainly composed of heavy nuclei. Using time
evolution of neutrino spectral property, we showed that propagation DATA AVAILABILITY
of high energy neutrinos is indeed hindered, when the PNS shock The data underlying this article will be shared on reasonable request
surface drastically collapses (i.e. 1 ms. 𝑡BH . 2 ms). Once the shock to the corresponding author.
surface is engulfed by the BH, those neutrinos are radiated away, with
some time delays for higher energy neutrinos. In the BH accretion
phase, the main component of accretion flow is high-𝑌𝑒 stellar mantle,
whose temperature is at the highest a few MeV. Therefore the main REFERENCES
neutrino emission channel is the electron capture on heavy nuclei Abbasi R., et al., 2011, A&A, 535, A109
occurring in the vicinity of AH. It results in a nearly constant elec- Abbott B. P., et al., 2016, ApJ, 818, L22
tron type neutrino luminosity as also reported in Sykes et al. (2023). Abbott T. M. C., et al., 2019, ApJ, 872, L30
We would like to emphasize that these neutrino properties could be Abe K., et al., 2011, arXiv e-prints, p. arXiv:1109.3262
revealed only by full neutrino radiation-hydrodynamic simulations Adams S. M., Kochanek C. S., Gerke J. R., Stanek K. Z., Dai X., 2017,
with numerical relativity without excising the relevant region outside MNRAS, 468, 4968
the AH, i.e., by fully solving the region outside the BH. Alcubierre M., Brügmann B., Diener P., Koppitz M., Pollney D., Seidel E.,
In this study we employed only one non-rotating progenitor model. Takahashi R., 2003, Phys. Rev. D, 67, 084023
Antoniadis J., et al., 2013, Science, 340, 448
In our future works, we are interested in exploring various CCSN
Baumgarte T. W., Shapiro S. L., 1999, Phys. Rev. D, 59, 024007
models accompanied by BH formation. For instance, a fallback
Baumgarte T. W., Cook G. B., Scheel M. A., Shapiro S. L., Teukolsky S. A.,
scenario is one of the interesting topics. The current progenitor 1996, Phys. Rev. D, 54, 4849
model has a significantly high compactness 𝜉2.5 = 1.0 at precollapse Blandford R. D., Znajek R. L., 1977, MNRAS, 179, 433
stage (O’Connor & Ott (2011) and see also Table 1 in Kuroda et al. Bollig R., Yadav N., Kresse D., Janka H.-T., Müller B., Heger A., 2021, ApJ,
(2022)), which leads to strong mass accretions during the PNS con- 915, 28
traction phase. Therefore it induces the PNS core-collapse without Burrows A., Vartanyan D., 2021, Nature, 589, 29
affording an opportunity for shock revival. However, if one considers Burrows A., Radice D., Vartanyan D., Nagakura H., Skinner M. A., Dolence
less compact stars (Chan et al. 2018; Powell et al. 2021) or rotat- J. C., 2020, MNRAS, 491, 2715
ing stars (Rahman et al. 2022), the shock revival aided by neutrino Chan C., Müller B., Heger A., Pakmor R., Springel V., 2018, ApJ, 852, L19
heating could happen before BH formation. Such systems could be Christie I. M., Lalakos A., Tchekhovskoy A., Fernández R., Foucart F.,
Quataert E., Kasen D., 2019, MNRAS, 490, 4811
observed as a faint supernova (Kochanek et al. 2008; Adams et al.
Cromartie H. T., et al., 2020, Nature Astronomy, 4, 72
2017) and should be distinguished from the current failed SN (or Demorest P. B., Pennucci T., Ransom S. M., Roberts M. S. E., Hessels J. W. T.,
direct BH formation) model with no shock revival. Progenitor model 2010, Nature, 467, 1081
dependency should definitely be explored in the future study to ex- Ebinger K., Curtis S., Fröhlich C., Hempel M., Perego A., Liebendörfer M.,
plain various observations. Thielemann F.-K., 2019, ApJ, 870, 1
Another interesting topic to be explored is the collapsar scenario Ertl T., Janka H. T., Woosley S. E., Sukhbold T., Ugliano M., 2016, ApJ,
(MacFadyen & Woosley 1999) as a possible route to long gamma-ray 818, 124
bursts and hypernovae. In the collapsar scenario, a BH surrounded Fernández R., Metzger B. D., 2013, MNRAS, 435, 502
by a massive disk is formed, i.e., highly non spherical system is Fischer T., Whitehouse S. C., Mezzacappa A., Thielemann F.-K., Liebendör-
fer M., 2009, A&A, 499, 1
formed. Such systems can be followed only in numerical relativity
Fischer T., et al., 2018, Nature Astronomy, 2, 980
with no approximation like CFC approximation. For instance, after
Foley R. J., et al., 2009, AJ, 138, 376
the formation of a massive disk, viscous effects significantly heat Fryer C. L., et al., 2009, ApJ, 707, 193
the disk, leading eventually to the launch of energetic outflows (in Fujibayashi S., Shibata M., Wanajo S., Kiuchi K., Kyutoku K., Sekiguchi Y.,
the context of both NS mergers and massive stellar collapse, see, 2020a, Phys. Rev. D, 101, 083029
e.g., Fernández & Metzger 2013; Just et al. 2015; Fujibayashi et al. Fujibayashi S., Shibata M., Wanajo S., Kiuchi K., Kyutoku K., Sekiguchi Y.,
2020a,b, 2023). As another intriguing and also a challenging topic 2020b, Phys. Rev. D, 102, 123014

MNRAS 000, 1–8 (2015)


8 T. Kuroda and M. Shibata
Fujibayashi S., Kiuchi K., Wanajo S., Kyutoku K., Sekiguchi Y., Shibata M., Typel S., Röpke G., Klähn T., Blaschke D., Wolter H. H., 2010, Phys. Rev. C,
2023, ApJ, 942, 39 81, 015803
Giovanoni P. M., Ellison D. C., Bruenn S. W., 1989, ApJ, 342, 416 Ugliano M., Janka H.-T., Marek A., Arcones A., 2012, ApJ, 757, 69
Gullin S., O’Connor E. P., Wang J.-S., Tseng J., 2022, ApJ, 926, 212 Valenti S., et al., 2009, Nature, 459, 674
Hayashi K., Fujibayashi S., Kiuchi K., Kyutoku K., Sekiguchi Y., Shibata M., Vartanyan D., Coleman M. S. B., Burrows A., 2022, MNRAS, 510, 4689
2022, Phys. Rev. D, 106, 023008 Woosley S. E., Heger A., Weaver T. A., 2002, Reviews of Modern Physics,
Heger A., Fryer C. L., Woosley S. E., Langer N., Hartmann D. H., 2003, ApJ, 74, 1015
591, 288
Hempel M., Fischer T., Schaffner-Bielich J., Liebendörfer M., 2012, ApJ, This paper has been typeset from a TEX/LATEX file prepared by the author.
748, 70
Hyper-Kamiokande Proto-Collaboration et al., 2018,
Progress of Theoretical and Experimental Physics, 2018, 063C01
Iwamoto K., et al., 1998, Nature, 395, 672
Janka H.-T., Melson T., Summa A., 2016, Annual Review of Nuclear and
Particle Science,
Just O., Bauswein A., Ardevol Pulpillo R., Goriely S., Janka H. T., 2015,
MNRAS, 448, 541
Kawabata K. S., et al., 2010, Nature, 465, 326
Kazanas D., Ellison D. C., 1981, in International Cosmic Ray Conference.
p. 176
Kochanek C. S., Beacom J. F., Kistler M. D., Prieto J. L., Stanek K. Z.,
Thompson T. A., Yüksel H., 2008, ApJ, 684, 1336
Kotake K., Takiwaki T., Fischer T., Nakamura K., Martínez-Pinedo G., 2018,
ApJ, 853, 170
Kuroda T., Shibata M., 2023, Phys. Rev. D, 107, 103025
Kuroda T., Takiwaki T., Kotake K., 2016, ApJS, 222, 20
Kuroda T., Kotake K., Takiwaki T., Thielemann F.-K., 2018, MNRAS,
477, L80
Kuroda T., Fischer T., Takiwaki T., Kotake K., 2022, ApJ, 924, 38
Liebendörfer M., Messer O. E. B., Mezzacappa A., Bruenn S. W., Cardall
C. Y., Thielemann F.-K., 2004, ApJS, 150, 263
Lund T., Marek A., Lunardini C., Janka H.-T., Raffelt G., 2010, Phys. Rev. D,
82, 063007
MacFadyen A. I., Woosley S. E., 1999, ApJ, 524, 262
Mirizzi A., Tamborra I., Janka H. T., Saviano N., Scholberg K., Bollig R.,
Hüdepohl L., Chakraborty S., 2016, Nuovo Cimento Rivista Serie, 39, 1
Müller B., 2016, Publ. Astron. Soc. Australia, 33, e048
Müller B., Varma V., 2020, MNRAS, 498, L109
Müller B., Janka H.-T., Dimmelmeier H., 2010, ApJS, 189, 104
Müller B., Heger A., Liptai D., Cameron J. B., 2016, MNRAS, 460, 742
Nagakura H., Hotokezaka K., 2021, MNRAS, 502, 89
Nakamura K., Takiwaki T., Kotake K., 2022, MNRAS, 514, 3941
Nomoto K., Tominaga N., Umeda H., Kobayashi C., Maeda K., 2006,
Nuclear Phys. A, 777, 424
O’Connor E., Ott C. D., 2011, ApJ, 730, 70
Powell J., Müller B., Heger A., 2021, MNRAS, 503, 2108
Rahman N., Janka H. T., Stockinger G., Woosley S. E., 2022, MNRAS,
512, 4503
Salathe M., Ribordy M., Demirörs L., 2012, Astroparticle Physics, 35, 485
Shibagaki S., Kuroda T., Kotake K., Takiwaki T., 2020, MNRAS, 493, L138
Shibagaki S., Kuroda T., Kotake K., Takiwaki T., 2021, MNRAS, 502, 3066
Shibata M., 1997, Phys. Rev. D, 55, 2002
Shibata M., Nakamura T., 1995, Phys. Rev. D, 52, 5428
Shibata M., Kiuchi K., Sekiguchi Y., Suwa Y., 2011, Progress of Theoretical
Physics, 125, 1255
Smartt S. J., 2015, Publ. Astron. Soc. Australia, 32, e016
Stockinger G., et al., 2020, MNRAS, 496, 2039
Sukhbold T., Ertl T., Woosley S. E., Brown J. M., Janka H. T., 2016, ApJ,
821, 38
Sumiyoshi K., Yamada S., Suzuki H., 2007, ApJ, 667, 382
Sykes B., Mueller B., Cordero-Carrión I., Cerdá-Durán P., Novak J., 2023,
Phys. Rev. D, 107, 103010
Takahashi K., Umeda H., Yoshida T., 2014, ApJ, 794, 40
Takiwaki T., Kotake K., 2018, MNRAS, 475, L91
Tamborra I., Müller B., Hüdepohl L., Janka H.-T., Raffelt G., 2012,
Phys. Rev. D, 86, 125031
Tanaka M., et al., 2009, ApJ, 692, 1131

MNRAS 000, 1–8 (2015)

You might also like