Download How To Find Newborn Black Holes Kazumi Kashiyama (UCB)

How To Find Newborn Black Holes
Kazumi Kashiyama (UCB)
Eliot Quataert and Rodrigo Fernandez (UCB)
1
2
3
4
IBIS
IBIS/SPI/JEM-X
IBIS/SPI/JEM-X
IBIS/JEM-X
IBIS/SPI/JEM-X
(date yy/mm/dd)
02/11/25–02/12/15
02/12/09–02/12/11
03/06/07–03/06/11
03/03/24–04/09/10
04/11/22
(ks)
Type
d.o.f. 810
583
582
581
staring, 5 × 5a , hexb584
365/365/31
5×5
Notes.
The assumed5stellar
292/296/275
× 5 temperature is Teff = 31,000 K. Ropt is the stellar radius derived from the models. Rdist is the stellar radius derive
c
measured
assumed
temperature. Vrot sin i = 95 ± 6 km s−1 is the projected rotational velocity of the O-star derived from the model
269/35parallax and
GPS
assumes
orbit and synchronous rotation. Model B assumes a circular orbit and nonsynchronous rotation. Model C assumes an eccentri
8/8/6a circularcalibration
pseudosynchronous rotation. Model D assumes an eccentric orbit and nonsynchronous rotation. The adopted parameters are the weighted aver
values for Model D derived for each temperature in the range of 30,000 K ! Teff ! 32,000 K.
Stellar-Mass Black Holes
Notes:
a
5×5 dither pattern around the nominal target location.
b
Hexagonal pattern around the nominal target location.
c
Individual exposures separated by 6◦ along the scan path, shifted by 27.5◦ in galactic longitude.
(Observations indicated on the first line also used, together with epochs 1–4, for Figs. 2 and 3.)
U
X-ray
(1.5-12 keV)
B
tion of the Binary Mass Function
Eric Addison
Fig. 1. RXTE/ASM daily average (1.5–12
keV)Bel+06
light curve
Cadolle
of Cygnus X-1 from 2002 November to 2004 November
(MJD = JD − 2 400 000.5) with the periods of our INTEGRAL
observations (see text and Tables 1 and 2 for epoch definitions).
were active. For the IBIS/ISGRI spectral extraction however,
we implemented the most recent module (prepared for the
OSA 5.0 delivery) which is based on the least squares fit
done on background and efficiency corrected data, using coded
source zones only. This option minimizes spurious features
in the extracted spectra, which appear in particular when the
Orosz+11
sources are weak, partially coded and the background
poorly
corrected (A. Gros, private communication). For the PICsIT
spectral extraction, we took the flux and error values in the
mosaic image at the best-fit position found for the source. We
used the response matrices officially released with OSA 4.2,
rebinned to the 8 energy channels of the imaging output.
The SPI data were preprocessed with OSA 4.2 using the
standard energy calibration gain coefficients per orbit and excluding bad quality pointings which have anomalous exposure
and dead time values (or with a high final χ2 during imaging). The spiros 9.2 release (Skinner & Connell 2003) was
used to extract the spectra of Cygnus X-1, Cygnus X-3 and
EXO 2030+375, with a background model proportional to the
saturating event count rates in the Ge detectors. Concerning the
instrumental response, version 15 of the IRF (Image Response
Files) and version 2 of the RMF (Redistribution Matrix Files)
were used for epoch 1 and 2, e.g., prior to detector losses, while
Credit: Chandra X-­Ray Observatory, NASA
V
The Astrophysical Journal, 785:28 (6pp), 2014 April 10
Fig. 2. The 20–200 keV IBIS/ISGRI light curve of Cygnus X-1 from
2002 November 25 until 2004 November 22 and corresponding HR
between the 40–100 and the 20–30 keV energy bands (see text and
Tables 1 and 2 for epoch definitions).
versions 17 and 4 respectively were taken for epoch 4, e.g.,
after the failure of two detectors.
Figure 3. Schematic diagram of Cygnus X-1, shown as it wo
the sky plane. The offsets are in milliarcseconds (mas), assum
of 1.86 kpc. The orbital phase is φ = 0.5, which corresponds
conjunction of the O-star. The orbit of the black hole is indicated
where the major and minor axes have been drawn in for clarity
dashed line, respectively). The direction of the orbital motion
as determined by the radio observations (Reid et al. 2011). T
represents the local effective temperature. The star is much c
inner Lagrangian point because of the well-known effect of “grav
(Orosz & Hauschildt 2000). The temperatures referred to in
Figure 1 and Table 1 refer to intensity-weighted average values.
(A color version of this figure is available in the online journal.)
, f (m1 ) is a piece of information constructed from observables of a
Resultsorbital
of the analysis period P
Assuming we have values for3.the
orb and the maximum
As shown in Fig. 1, during the epoch 2 INTEGRAL observa3
tions, the
1.5–12 keV ASM
average
countquantity
rate of Cygnus X-1
rs, say V2 , we can compute the
value
for
the
m
sin
i through
1
(∼1.3 Crab) was larger than during epoch 1 (∼290 mCrab)
by
a factor
of 4.5.
derived
IBIS/ISGRI
20–200 keV
light
Figure
2. Top:
the The
optical
light curves
and best-fitting
models
assuming an
eccentric
with e Ratio
= 0.018
D, leftX-1
panels)
and a recircular orbit
curves
andorbit
Hardness
(HR)(Model
of Cygnus
are shown
(Model A, in
right
Noteoverview
how muchofbetter
thePV-Phase
unequal maxima
of the
spectively
Fig.panels).
2 (general
part of
and
of the recent work considers circular orbits and s
light curves are accommodated by the model that includes eccentricity. Bottom:
epochs 1 to 4) and Fig. 3 (specific zooms on part of PV-Phase,
the radial velocity measurements binned into 50 bins (filled circles) and the
rotation (e.g., Caballero-Nieves et al. 2009). While
epochs
1 and
2). From
1 toorbit
epoch
2, while
thecircular
ASM orbit
aver-(right).
best-fitting
model
for theepoch
eccentric
(left)
and the
tricity we find is small (e = 0.018 ± 0.002 for
age count rate increased, the 20–200 keV IBIS/ISGRI one deTeff = 31,000 K), it is highly significant. Allowi
creased from ∼910 to ∼670 mCrab as shown in Fig. 3 (where,
Ozel+12
−1
in the 20–200 keV range, 1 Crab = 205 cts s ). This proba5
bly indicates a state transition between epochs 1 and 2, as also
0
2
4
6
8
10
suggested by the decrease in the IBIS HR (Fig. 3). Similar
transition, with a change in the ASM light curves and an
evolving IBIS HR, occurred again during GPS data (epoch 3).
Figu
Figure
(200
Figure 2 (bottom) shows the IBIS HR behaviour over the
whole1. Observed masses of neutron stars (filled triangles) and black holes
(filled squares) from Özel et al. (2010, 2012). The thick solid line at 2.25 M⊙
rapi
2002–2004 period indicated in Table 1.
Porb V22
m1 sin3 i
=
f (m1 ) ≡
2
(1 + q)
2πG
(1)
rst component of the binary, i is the inclination angle of the orbit,
of the system, and G is the gravitational constant.
Which progenitors produce BHs not NSs?
Supernova shock is stalled or not?
How much material fallback on protoNS?
The key will be inner density structure within
r ⇠ 1000 km, M r ⇠ 2-3 M
Red supergiant
O’Connor & Ott 2011; Ugliano et al. 2012;
K.
Kashiyama,
R.
Fernandez,
and
E.
Quataert
(RSG)
Horiuchi et al. 2014; Pejcha & Thompson 2015
2
12
13
15
17
20
25
30
35
40
45
50
55
60
70
80
100
Density [g cm ]
105
100
10
0.8
1052
RSG
0
6
10
10
WR
0.6
0.2
-10
10
53
BH?
BSG
0.4
Blue supergiant
(BSG)
-5
Z = Zsun
Z = 0.01 Zsun
1
ξ2.5
10
10
10
7
8
10
9
10
10
10
10
Radius [cm]
11
10
12
Walf-Rayet star
(WR)
10
13
10
14
10
20
30 40 50 60 70 80 100
ZAMS mass [Msun]
Fig. 3.— Compactness of the progenitors. The magenta and
cyan points
to the
cases instar
Figs. can
1 andform
2, respectively.
Allcorrespond
types of
massive
BHs.
lations, the necessary condition for successful SN explosion is ⇠2.5 . 0.2 0.3. In the following sections, we
What is observational signature?
Stellar type (RSG, BSG, or WR)
＊
Rotation profile (disk or not)
＊
Magnetic field (jet or not)
Possible Outcomes in a BH Formation
No. 2, 2008
DO SUPERGIANTS VANISH?
1339
Kochanek+08
Fig. 3.— Possible outcomes in forming a BH. The optical signatures of the ‘‘no explosion’’ scenarios are little explored.
significantly less luminous than before it died. Other known sources
galaxies (axis ratios <0.3). Only 40 galaxies need to be observed
# 5
Possible Outcomes in a BH Formation
No. 2, 2008
DO SUPERGIANTS VANISH?
1339
Gamma-ray bursts?
Hypernovae?
Super-luminous supernovae?
luminous,
but rare
Fig. 3.— Possible outcomes in forming a BH. The optical signatures of the ‘‘no explosion’’ scenarios are little explored.
significantly less luminous than before it died. Other known sources
galaxies (axis ratios <0.3). Only 40 galaxies need to be observed
# 5
Possible Outcomes in a BH Formation
No. 2, 2008
DO SUPERGIANTS VANISH?
1339
“Unnovae”
probably
not rare
Fig. 3.— Possible outcomes in forming a BH. The optical signatures of the ‘‘no explosion’’ scenarios are little explored.
significantly less luminous than before it died. Other known sources
galaxies (axis ratios <0.3). Only 40 galaxies need to be observed
# 5
Very low energy supernovae from neutrino mass loss
Even if the SN shock is stalled,
10.— Shock modeled by CASTRO in RSG15 (TOV = 2.5,
aFig.
weak
shockmapped
can into
be KEPLER.
driven by
full
neutrino
losses)
neutrino mass loss of the PNS.
Fig. 11.— Velocity of the hydrogen envelope at 5×107 s after core
collapse in RSG15, TOV = 2.5, full neutrino loss model, evolved
further in KEPLER.
“Luminous red novae”
then continued to evolve them. Figures 10 and 11 show
the KEPLER
results
RSG15, TOV
= 2.5; 13
Figure 10
Nadezhin
80,for
Lovegrove
& Woosley
5
shows the imported shock, and Figure 11 shows the final
7
velocity of the hydrogen envelope at t = 5 × 10 s.
The shock has decreased significantly in strength by
the time it reaches the base of the hydrogen envelope;
however, this envelope is very tenuously bound in both
RSG15 and RSG25. For each model we tested six choices
of TOV limit (2.0 - 2.5 M⊙ , in 0.1 M⊙ increments) and
evaluated the strength of the shock that reached the hydrogen envelope. Using the full neutrino loss model,
we found in every case tested for RSG15 and in 3 of
6 tested for RSG25 that the shock produced was larger
than 1×1047 ergs, the approximate binding energy of the
envelope (see Table 3). We can therefore realistically expect the envelope to be ejected in these cases. However,
the highest kinetic energy achieved was only of the order of 6×1047 ergs, and most models fell well below that
number. The envelope is therefore ejected with a very
low velocity (50 - 100 km/s). It emits most of its energy
Fig. 12.— KEPLER light curve for a transient from RSG15,
via hydrogen recombination. Optically this transient has
A significant part of the energy
TOV = 2.5. The transient is low luminosity but lasts for around
39
40
a low luminosity ∼ 10 − 10 ergs/s, but maintains this
a year.
47
48model
comes
frominH
Fig. 6.—
Collapse offor
the
in the
maximum
neutrino
RSG15 model
the recombination.
fully-modeled neuluminosity
ofRSG15
order
a year.
The
color temperature
of Fig. 7.— Collapse of the
mass loss
case,
showing
a
shock
forming
due
to
the
effective
core
trino
mass
loss
case,
showing
a
shock
forming
due
to the such
effective
sh
the transient is very red, of order 3000 K. An example
explosion would leave a black hole. A survey
as that
mass decrement. Positive velocities are outwards, negative are incore mass decrement. Positive velocities are outwards, negative
light curve
can be
Figure
12 for
RSG15,
TOV =are inwards.
proposed
byare
Kochanek
al. (2008),
monitoring
red suwards. Curves
are purple
for seen
early in
times,
shading
to red
for late.
Curves
purple foretearly
times, shading
to red for
2.5.has the TOV limit set to 2.5, resulting in a mass decrepergiants
forthe
anomalous
might
This model
late. This
model has
TOV limittransients
set to 2.5, that
resulting
in asignal
mass the
Ecalculated
of H envelope
of RSG
ment of 0.525
. Bind.
The shock
propagates
out ofare
the helium
core muchdecrement
of 0.523
M⊙ . The
shock
is smaller
in strength
than theThey
TheM⊙transients
here
obviously
birth
of a black
hole,
should
catch
these events.
10 cm). Time shock reaches 1010 cm: 38 s. Time to
casevisible
and reaches
edge of
the simulation
with “star”
a
(r = 3.568×10
fainter
and less energetic than standard core-collapse su-maximum-loss
would be
as a the
sudden
brightening
of the
11 cm: 577 s.
lower velocity. Time shock reaches 1010 cm: 40 s. Time to end of
end of helium
core:
196
s.
Time
to
10
pernova, but they do bear some similarity to a class ofhelium for
of207
order
a year,
bys.a gradual but complete
11 cm: 620
core:
s. Time
to 10followed
recently-observed
transients:
“luminous
novae,”
disappearance.
corresponds
to MGh as given
by Eq. the
4. This
model red
takes
such
as
V838
Mon
(Munari
2002).
Luminous
red
novae
In RSG25 a TOV limit of 2.2 M⊙ or lower resulted
into account the thermal mass loss and ties the cessation
E
⇠ 10 &
erg
T ⇠ 3000 K
No Data No Data
4589
4652
4795
4915
Searching for vanishing supergiants
• Monitoring ~106 RSGs in ~25 Gal.
within ~10 Mpc with ~0.5 yr
cadence for ~5 yrs using the
Large Binocular Telescope
• Examine sources with
(⌫L⌫ )
Figure 14. Select V and Rc band observations for Candidate 1 in NGC
all. The selected observations give a clear picture of the source’s variabi
July 2008 (3 May 2008) and the “Last” observation is on 20 November 2
Kochanek+08, Gerke+15
104 L
• 3 core collapse supernovae
• 1 candidate of vanishing RSG
• Continuous obs. will 14give
J. R. Gerke et al.
meaningful constraints on
failed SN rate.
REF
RMS
First
52
Figure 13. The B, V and Rc band differential light curves for
Candidate 1 in NGC 6946. The open circle in the V band light
curve was an observation that fell just outside our quality criteria
No Data
band
that was later added as a check on the measurements. The V
vertical axis is in units of 104 L⊙ (νLν ) and has been normalized to
20080503 20080705 20081125
20090325
20100318 20110923
20120612
20130610difference
20141120 between
Last
the first
observation
so that the
luminosity
the
first and last observations can be easily seen. A change in lumiNo Data No Data
R band
nosity by 104 L⊙ in either direction, as indicated by the horizontal
lines, would lead to the source being selected as a candidate. The
4589
4652
4795
4915
5273
6090
bottom
two panels
are5827
the 3.6µm
and6453
4.5µm 6981
SST archival light
curves normalized to the first epoch and on a different y-axis
scale.
F
H
se
ob
1
fo
m
14
in
ob
sh
ca
an
a
th
fr
de
un
w
If
va
Possible Outcomes in a BH Formation
No. 2, 2008
DO SUPERGIANTS VANISH?
1339
BH + super Eddington disk
maybe
not rare
Fig. 3.— Possible outcomes in forming a BH. The optical signatures of the ‘‘no explosion’’ scenarios are little explored.
significantly less luminous than before it died. Other known sources
galaxies (axis ratios <0.3). Only 40 galaxies need to be observed
# 5
Fall back disk may be ubiquitous
Woosley & Heger 12, Perna+14
The Astrophysical Journal, 752:32 (11pp), 2012 June 10
Woosley & Heger
BSG
WR
in binary
Figure
1. Distribution Journal,
of angular752:32
momentum
with
respect
angular momentum
The Astrophysical
(11pp),
2012
June to
10 mass in the pre-supernova models for V24 and V36. The smooth curves show the Woosley
& Heger
required to form a stable disk at the last stable orbit of a Schwarzschild black hole (lower curve) or Kerr black hole (upper curve) including the given mass. The
intermediate curve that follows the Schwarzschild curve until far out in the star uses the integrated angular momentum in the model to determine the last stable orbit.
Where the irregular dark line showing the actual angular momentum on the star intersects this line a disk can form. The outer 9 M⊙ of Model V24 and the outer 10 M⊙
of Model V36 will form a disk. The edge of the helium cores of the two models is apparent in the sharp inflection in the angular momentum at 8.3 M⊙ and 15.0 M⊙ .
Mass loss was included in the calculation, but due to the low metallicity, only 0.05 M⊙ and 0.15 M⊙ was lost in V24 and V36, respectively. Note that, if all the surface
material accreted here, the black hole would rotate at nearly its maximum allowed value (i.e., the red line intersects the green one at the end).
for massive main-sequence stars depends upon some fractional
power of the metallicity (Kudritzki 2002). More important and
less certain is the dependence of mass lost as a red or blue
supergiant on metallicity. It is thought that mass loss from cool
giants is more dependent upon pulsations and grain formation
(Reimers 1977; Smith et al. 2011). One might expect therefore
a rapid falloff in mass loss below some value necessary for
significant grain production. It has been estimated that red giant
mass loss will be significantly less below 0.1 Z⊙ (Bowen &
Willson 1991; Zijlstra 2004).
If stars do not lose mass then they conserve angular momentum. Since the natural course of evolution leads to the contraction and spin up of the inner star, shear instabilities and
magnetic torques will concentrate an increasing amount of angular momentum in the outer regions of the star. While the
Outer layers of up to ~ a few
M
angular momentum distributions as shown in Figure 1. While
the lack of sufficient angular momentum within the helium core
precludes making a disk around a black hole within that mass,
there is ample rotation in the outer part of the hydrogen envelope
to do so. This could be a very common occurrence. From the
survey of A. Heger & S. E. Woosley (2012, in preparation), the
total angular momentum of a massive star at birth that has an
equatorial rotation speed of about 20% Keplerian on the main
sequence is
!
"
M 1.8
52
Jtot = I ω ≈ 2 × 10
erg s.
(1)
M⊙
can “naturally” have sufficient
j
If the star does not lose mass, most of this angular momentum becomes concentrated, in the pre-supernova star, in a nearly rigidly
13
v
r0 ≈
fr × for
∼ 3&
×ionized
10
cm helium,
. (1)
min
(MRI).
Proga
Begelman
(2003a,b)
simulated
accre′
a steady
The density
be des
ctronity
scattering
singly
fully
ionized
inner
dis
(MRI).
Proga
accre-structure
ρby≈wind.
ρsimulated
. 2 canmost
(8)
c2 rate, ṀEdd =ity4πGM
10 BH /cκ
10
M
0−2
the
fallback material.
Eddington accretion
∼ ⊙1 &
× Begelman ρ(2003a,b)
"
!
!
t
v
acc
out,min
≈gas
ρ0quali(r/r
,disk
where
ρqualiṀ the
/4πr
, energy
or
0 ≈
3/2
nuclear
energy
released
0 v̄out
ium10tion
and
hydrogen,
respectively.
Theof accretion
rate
nuclear
low
angular
momentum
gas
in
ara-isscenario
−15
−1
−1
tion
angular
momentum
in 0The
a) scenario
accretes
once out
angular
momentum
fṀ
Mof
(κ/0.2
cm2cases
g−1 )in
(M
). Note
that
f
⊙ son marginal
BH /10M
⊙
r
′
−12
−3
We focus
which
the low
circularization
!
"
!
"
We
determine
the
normalization
of
the
density
by
mass
confectively
transported
by
e.g.,
magnetorotational
inim
ρ
∼
4
×
10
g
cm
−3/2
2
−1
dynamically
o below
the
accretion
rate
at
which
there
is
significant
dynamically
importan
0showed
tatively
analogous
toorbit
thatThey
considered
showed
# rmaxhere.2 They
tatively
to0.4that
here.
the
opacity
∼ 0.1,larger
0.2 and
cm
g considered
corresponds
to
f&r Begelman f(2003a,b)
−3
dius
is notκanalogous
much
than
the
innermost
circular
Ṁ
10
0.1
ity
(MRI).
Proga
simulated
servation,
4πr
ρdr ≈
fṀ Md ,the
which yields At t " ta
ρ0 ∼
60
g cm
rmomentum
the
MRI
redistributes
during
min
utrino
(Chen
2007).
this
electron
scattering
for&
singly
ionized
helium,
fullyIn
ionized
At t 0.1
"
tinacc
the
10
that
MRI
redistributes
angular
momentum
during
the
of cooling
the the
BH
(f
−Beloborodov
100);
wethat
show
below
(Fig.
4) case,
that angular
"
!
!
tion
of !
low angular
momentum
gas"
a ,scenario
r ∼ 10
−9/2
"
!
"
!
and
the
outflo
3/2
helium
and
hydrogen,
respectively.
The
accretion
rate
is
circularization,
leading
to
dissipation
which
powers
both
R
M
"
!
"
!
"M
!
∗
d
e cancircularization,
expect
a strong radiation-driven
from
the ′ powers both
tatively
analogous
that
considered
here. essen
They
s
−3/2
−3
fṀoutflow
larger
circularization
radii likely to
leaddissipation
to outflow
fainter more
slowly
frto R
and
the
−12 × M −3
leading
which
M
∗
BH
d
ρ
∼
4
×
10
g
cm
also below the accretion rate at which
there
is
significant
the
outflow
e
0
that
the
MRI
redistributes
angular
momentum
durin
12
andestimated
outflow.
In our scenario,
viscous
×
1the
M⊙
10 0.1cm ejecta 10
k. Such
outflows
haveThe
been
alsoaccretion
confirmed
byannumerievolving
transients.
fallback
rate
can be
as
10
12outflow
the
wid
1
M
10
10
M⊙
accretion
outflow.
In
our
scenario,
the
viscous
circularization,
leading
tocm
dissipation
powers
neutrino
coolingand
(Chen an
& Beloborodov
2007).
In
this
case,
⊙
r/t"which
≈ v.
The
time
of
the
disk
is
much
shorter
than
the
fallback
time
"
!
"
!
!
−9/2
3/2
Ṁ
≈
M
/t
,
or
acc
d
d (Ohsuga
simulations
etradiation-driven
al. 2005; Sa̧dowski
et al.
accretion
outflow.
InM
our
scenario,
the p
v
R
M
∗an
one
can expect
a strong
outflow
from
the2014;
Here,
rdmax of
≈and
vout,max
tacc
≈ Since
r/t
≈
v.
min
as
time
of the
disk
is
much
shorter
than
the fallback
time
scale.
Thus,
the
accretion
rateisis
given
bymuch
the
the
density
the
outflow
at, The
rrBH
=described
rdensity
the
.vout,mi
(9)
×essentially
0 . the
−5
−1
time
of
the
disk
is
shorter
than
fallback
12
ng, disk.
Stone
1M
10 cm
10 M
outflows
have
alsofrom
confirmed
by numeri⊙ , and
⊙
ṀLuminous
∼& Davis
3 × Blue
10 2014).
M
s
⊙ been
d Such
v̄the
vdescribed
=
fessentially
v̄outgiven
. Inb
v,max
out
out,min
v,min
Fastscale.
Transients
Newborn
BHs
3
isftypically
initially
highly
optically
thick,
the
temperatur
fallback
rate
(Eq.
2),
which
is
larger
than
the
scale.
Thus,
the
accretion
rate
isas
Thus,
the
accretion
rate
is
essentially
given
by
!
"
!
"
!
"
cal
simulations
(Ohsuga
et
al.
2005;
Sa̧dowski
et
al.
2014;
−3/2
1/2
We
model
the
fallback
disk
outflow
as
follows.
First,
1/3
−2/3
Luminous Blue Transients
BHs
3
Here,
rmax
vout,max
tacc
,(Eq.
≈
tacc
=
!
fallback
is,!
typically
larger
choose
fv,max
"
1,
f1which
1,, v#and
>
adiabatically,
Trate
∝
ρrmin
rvout,min
thus
Tout,max
≈ ξT0tha
(r
Md(Eq. from
R∗ Newborn
MBH
Eddington
accretion
rate,
ṀEdd
=≈4πGM
/cκ
∼2),∝
×
v,min
BH
fallback
rate
2),
which
is
typically
larger
than
the
Jiang,
Stone
&
Davis
2014).
×
,
(2)
rmax
accretion
rate,
Ṁ
−15is isotropic,
raction
fṀFinally,
< 1 we
ofM
the
accreting
is⊙ loaded
on cm
the
v̄−1
, and
venergy
f/8πar
v̄02out
. Edd
In
this4πGM
paper,
we∼
′(4π
BH /cκ
out
out,min
v,min
assume
that
outflow
although
in 2fv,max
Kasen
12the
TEddington
≈
(Ṁout
v=Note
)1/4
, or=
0 /10M
10mass
M10
s−1
g−1 )where
(M
that
isfy
the
conservation
1
10
cm
M
BH
⊙ ).
⊙
⊙ (κ/0.2
−15
−1out
2 −1 −1 i.e., ρ r≈ ρ0
We
model
the
fallback
disk
outflow
as
follows.
First,
Finally,
we assume
that
outflowrate,
is isotropic,
although
in 4πGMBH
s (κ/0.2
g ξ )> (M
/10M
Not
min
⊙ ).satEddington
accretion
Ṁ
=∼
/cκ
∼ 2101
×
We
determine
choose
fv,max
" −1
1,Mf⊙v,min
! 1,"cm
so
as
to
Edd
reality
it will
bethe
bipolar.
tflow,
Ṁout
=f fbeṀ
×
Ṁ
or,
!
!0.4 22cm
"BH
dmoderately
# rm
#and
2
−1
2corresponds
the
opacity
κ
0.1,
0.2
and
0.4
cm
g
to
−5/8
1/4
r
2
max
3
1/2
reality
it
will
moderately
bipolar.
the
opacity
κ
∼
0.1,
0.2
and
g
correspon
a
fraction
<
1
of
the
accreting
mass
is
loaded
on
the
−15taccṀ
−1
2 −1
fenergy
v̄conservation
that
the
en
fṀ
out /2.
fnote
the
(4πr
× internal
ρv /2)dr r≈
Next,
let
describe
density
and)−1
temperature
profile isfy
he pror
where
≈
π(R
/8GM
)the
,"or
servation,
8 Wei.e.,
Ṁ Mdthat
∗ us
BH
!
10
M
s
(κ/0.2
cm
g
(M
/10M
).
Note
rmin ionized
⊙
BH
⊙
T
∼
8
×
10
K
m
Next,
let
us
describe
the
density
and
temperature
profile
electron
scattering
for
singly
helium,
fully
i
electron
scattering
for
singly
ionized
helium,
fully
ionized
0
outflow,inṀthe
= fṀ × which
Ṁd!or,are fcrucial
out outflow,
2subdominant at
for !
quantifying
the
electrors have
We
determine
the
norm
"
"
is
r
≈
r
due
to
adiab
10
0.1
2
−1
−6
−1
min
f
M
v̄
/2.
We
note
that
the
internal
energy
of
the
shell
Ṁ
3/2
−1/2
out
d
in
for
quantifying
electrohelium
and hydrogen,
respectively.
accretion r
# rThe
Ṁ
!
the
opacity
κare
0.1,
and the
0.4
cm expands
g (4)
corresponds
to
∼the 3outflow,
× 10 which
M
scrucial
helium
and
hydrogen,
respectively.
The
accretion
rate
is
Rf∗0.2
M
ut
⊙
BH
max
2"
4∼
magnetic
emission.
After
the"
launch,
the
outflow
l torus
"
!
"
!
!
Super-Eddington
accretion!
1/4
−3/8
is
subdominant
at
r
≈
r
due
to
adiabatic
cooling.
To
′
obtain
f
,
f
,
and
ξ
consistently
magnetic
emission.
After
the
launch,
the
outflow
expands
t
∼
3
×
10
s
(3)
also
below
the
accretion
rate
at
which
there
is
sign
−6
−1
min
servation,
4πr
0.1
Ṁ
acc
out,max
out,min
RDecember
M
12 cm
ρr0min
×ρ
1
(4)
Ṁ
∼
3the
×
10
M⊙ s 10
∗
Therate
Astrophysical
Journal,
796:106
(14pp), 2014
1 ∼ M4BH
d significant
also
below
the
accretion
at
which
there
is
out
into
surrounding
medium.
For
t
!
t
,
the
accretion
10
M
electron
scattering
for
singly
ionized
helium,
fully
ionized
acc
⊙
into the surrounding
For0.1
t"! tacc , !
the accretion
cooling
(Chen
& Beloborodov
2007).
In
this
obtain fform
,×
fout,min
, and
ξ consistently,
one the
has
to
per"medium.
!
"1/2
!
out,maxneutrino
numerical
simulations,
but
basic
12
−3/2
"
1 expect
MIn
10 cm
10 M!
rate isM
almost
and
the !
outflow
is"approximately
neutrino
cooling
(Chenaccretion
&form
Beloborodov
2007).
case,
⊙ this
⊙
rate
is almost
and
the
approximately
one
can
abut
strong
outflow
fro
M
"constant,
! R
!constant,
dtimescale,
numerical
simulations,
theradiation-driven
basic characteristics
of
helium
hydrogen,
rate
is
1/2 The
−3/2is
is the
free and
fall
R∗∗ outflow
is"respectively.
the
radius
ofBH
the
outermost
. (1)
R
M
M
.
×
the
optical
emission
are
not
so
sensitive
t
steadyThe
wind.
The12
density
structure
can
be
described
as
∗
BH
ddensity
a steadya wind.
structure
can
be
described
as
disk.
Such
outflows
have
been
also
confirmed
by
nu
′
−12
one
can
expect
a. strong
radiation-driven
outflow
from
the
×
We
note
that
the
gas
temperature
in
the
disk
the
optical
emission
are
not
so
sensitive
to
these
paramelayer,
and
M
is
the
total
mass
of
the
disk.
The
torus
is
op−2
2
1
M
10
cm
10
M
ρ
∼
4
×
10
g
c
×
−2
2
⊙
also
below
rate
which
there
is significant
0 =et0.7,
ρ ≈0 )ρ0 (r/r
where
ρcm
v̄M
or
ρ
≈ ρ0 (r/r
,dwhere
≈10
Ṁ12
/4πr
v̄out
, or10
cal simulations
(Ohsuga
al. 2005;
Sa̧dowski
et al.1
0 )⊙ ρ,0accretion
0 ≈ Ṁ
out
1⊙the
M
out
out /4πr
0at
−1/4
⊙,outflows
0disk.
9numeriters.
We
take
f
1.4,
v,min
v,max
Such
have
been
also
confirmed
by
ters.
We
take
f
=
0.7,
f
= this
1.4, f
and
ξ heavy
==
3.75
as
Outflows!
f
T
!
a
few
10
K.
In
case,
nuc
v,min
v,max
0
ion ra-tically
and geometrically
thick,
trapping
the
heat
generated
!
"
!
"
Jiang,
Stone
&
Davis
2014).
!
"
!
"
Ṁ
−3/2
−3/2
neutrino
cooling
(Chen
&
Beloborodov
2007).
In
this
case,
cond,
the
outflow
velocity
is
approximately
the
escape
vefṀ cal fsimulations
fr −3 isfrapproximately
Second,
the outflow
velocity
the escape
vefiducial
choices.
As
long
as
the
is≈!
a
(Ohsuga
et
al. at
2005;
et
al.
2014;
fiducial
choices.
long
as the
thefallback
ejecta
issynthesized
almost
adiabatic,
"
! ejecta
−3material.
r orbit
mostSa̧dowski
O,As
Ne,
and
Mg
can
bedisk
inside
Ṁ
We
model
outflow
asrmax
follows.
Here,
ρby
∼ ρfallback
60
0 the
0 g∼cm 60 g cm 1/2
1/2
M
10/r
0.1 Jiang,
locity,
v̄can
≈
(2GM
,10
or
expect
radiation-driven
outflow
from
the
0.1 Stone
profile
described
as
out
BH
0 ) strong
4)
& Davisis2014).
ity,that
v̄one
≈
(2GM
/r
,"
or
aFe
fraction
<be
1 of
theHence,
accreting
mass
loaded
the
temperature
profile
can
bethere
described
ao
out The
BH
0 )a
but
not
groupfcan
elements.
isisv̄dno
radi
disk
accretes
once
the angular
momentum
ef- the temperature
Ṁ
fv,max
out , an
×
"
!
"
!
−3/2 !
−3/2
"
!
"
!
"
!
!
"
!
"
"
!
−3/2
−3/2
outflow,
Ṁ
=
f
×
Ṁ
or,
slowly
out −1Ṁ First,
d
−ξ/3
−1/2model
Md outflows
∗
BH
We
the
fallback disk
outflow
as The
follows.
!nuclear
!1 M
"−
"
!fMrbeen
in the
outflow.
burning
only
occu
disk.
Such
have
also
by
numeriR
M
MdR10
fectively×
transported
by
e.g.,
instabil∗ magnetorotational
BH
−1
−1/2
choose
f⊙v,max
t above
v "
, (6)confirmed
′
−1
"
!
12
,
(6)
×
ated as
t
v
f
M⊙
cmcm s12
M
T ≈ Tdisk,
(10)
v̄1out
∼ 1 ×10
10
. fM < 1(5)of the accreting
r⊙
′thethe
10
0is loaded
fṀ is. likely larger
a10
fraction
mass
on
enthalpy
−6
−1
ity (MRI).
Proga
&M
Begelman
(2003a,b)
simulated
accre⊙ cm10
Ṁ
10
vMout,min
isfy
the
energ
Twhere
≈10T
v̄out ∼
1 × 110
s−1cm
. ⊙ Sa̧dowski
(5) etinner
cal
simulations
(Ohsuga
et
al. 10
2005;
al. most
2014;
Ṁout ∼tacc
3×
⊙ s
0
Here,
r the0.1nuclear
≈vout,min
vout,max
is
the of
density
ofangular
the outflow
at r = r0 . gas
Sinceinthe
outflow
=the
fṀoutflow
× Ṁd or, nuclear energy released,
so tthat
reaction
acc max
out
tion
scenario
quali2
is low
the density
of momentum
the outflow outflow,
at10
r = r0Ṁ
.a Since
Then, what will happen?
"!
"f−3/2
!d v̄out "
!
M
/2.
Jiang,
&000,
Davis
1/
is
initially Stone
highly optically
thick, 2014).
the temperature evolves
ṀMetzger
"
!
f
v̄
,
and
R
Mv
M
ctatively
dynamically
important
(Fernández
&
201
⃝
2015 RAS,
MNRAS
1–7
∗
BH
d
v,max
out
out,m
analogous
to
that
considered
here.
They
showed
is
initially
highly
optically
thick,
the
temperature
evolves
1/3
−2/3
−2/3
×
adiabatically, T ∝ ρ
∝ r 1/3 , thus−2/3
T ≈ T0 (r/r0 )
, −6 −2/3 −1
fṀ First,
subdominan
We
model
the
disk
outflow
as
follows.
1M
1012 rate
cm isdecreases
10 M⊙sign
⊙
2 ρ1/4fallback
At
t
"
t
,
the
accretion
adiabatically,
T
∝
∝
r
,
thus
T
≈
T
(r/r
)
,
(4)
Ṁ
∼
3
×
10
M
s
2015 where
RAS,the
MNRAS
000,
1–7
acc
0
0
that
MRI
redistributes
angular
momentum
during
the
out
⊙
choose
f
"
1, f
T
≈
(
Ṁ
v
/8πar
)
,
or
v,max
0
out out
0
, (2)
obtain
f
2 1/4
0.1
out,ma
where
T
≈
(
Ṁ
v
/8πar
)
,
or
the essentially
outflow velocity
is approximately
the esca
out
out
and theSecond,
outflow
decouples
from the
dis
0 1/4 accreting
acircularization,
fraction 0fṀ
1to!offdissipation
the
the
! <
"−5/8
"
leading
which powers
isfy
the
conser
"is! loaded"on
! v̄out ≈ (2GM
"1/2
!massboth
1/2 energy
fr
form
numerica
−3/2
8
Ṁ−5/8 !
!
"
"
locity,
/r
)
,
or
BH
0
1/4
T
∼ 8 × and
10
Kan =
ejecta
0
R∗ the outflow M
accretion
In ourfṀscenario,
theMviscous
BH will expand in a homologous
d
outflow,
Ṁ
out 10outflow.
r0.1Ṁd or,
8 fṀ f×
2 the
! "optical
.
×
em
−1/2
T0 the
8×
10
K
fṀ Mof
v̄out−1
/2.
12 cm
r/t ≈ v. The10density
profile
homologous
ejec
d
fr We not
"1/4
"−3/8 ! 0.1
"−3/8fallback
!∼ disk
time of
is !much
shorter
than!the
10 the
1
M
M
10
⊙time 10
⊙
"
v̄out ∼ 1 × 10 cm s ters. We . take
R∗
MBH
Md
10
"
!
"−3/8 f
! given
!accretion
described
as
.(7) "−3/8
×
the
is
essentially
by
the
1/4 rate
is subdominant
at r
(3)scale. Thus,
−6
−1
12
Ṁ
1
M
10
cm
10
M
R
M
M
⊙
⊙
∗
BH
d
Second,
the
outflow
velocity
is
approximately
the
escape
ve!
"
!
"
fiducial
choice
(4)
Ṁ
∼
3
×
10
M
s
−3
−ξ
out
⊙
.(7) 1/2
×
fallback rate (Eq.
2),
which
is
typically
larger
than
the
12
t
v
obtain
fout,max
, fout,m
c 2015 RAS, MNRAS
⃝
000, 1–7
1M
10 inlocity,
cm
M≈⊙ BH /r0 )
We note that the gas
temperature
the disk
is 10
Td(2GM
⊙
v̄out0.1
≈
, or
the temperatu
ρ ≈ ρ′0
.
ermost
−1/4
9
Eddington
accretion
rate,
Ṁ
=
4πGM
/cκ
∼
1
×
BH
Edd
fṀ T0We
! anote
few that
10 !
K.
In
this
case,
heavy
nuclei
up
to
t
v
"
!
"
!
"
acc
out,min
form numerical
simula
! "−1/2
1/2
−3/2
the gas
in the disk
is Td ≈
2 temperature
s is op-10−15 M s−1 (κ/0.2
Jiang+14
cm
g−1 )−1 (Minside
/10M
). Note
that
Figure
3.
Snapshot
of
disk
structures
for
density
(left)
and
radiation
energy
⊙Ne, and Mg can
BH
at most O,
be
the ⊙
disk,
−1/4
R
M
M
9d synthesized
f
∗
BH
r
10
fṀ T0 !×a few 10 K. In this case,
heavy
nuclei
to
We
the
theρ0density
nerated
2 radioactivity
−1
v̄out
∼ 1 up
× 10
s−1atdetermine
. 1.13
the
optical
am
densitycm
(right)
time
10.4 ts normalization
. Units
for (5)
ρ and Eof
and ar T04byT
# r×
r are emission
but
Fe group
elements.
Hence,
there
iscm
no
the not
opacity
κ
∼
0.1,
0.2
and
0.4
g
corresponds
to
12
max
2
10
at
most
O,
Ne,
and
Mg
can
be
synthesized
inside
the
disk,
M⊙burning10
cm in the
10
M⊙ servation, r
respectively.
4πr ρdr ≈ fṀ Md , which yields
in the outflow. The above1
nuclear
only occurs
Fast Luminous Blue Transients
2
K. Kashiyama, E. Quataert
KK & Quataert 15
Fast Luminous Blue Transient
Fallback Disk Outflow
Absolute AB magnitude
6
The Inner core is directly swallowed by the central black hole. !
The outermost layers have sufficient angular momentum to form a disk. !
K. Kashiyama, E. Quataert
4
-18
-17.5
2
WR
BSG
DISCUSSION
Figure 1. Schematic picture of failed supernova model for fast luminous blue transients.
DIVERSITY OF BLACK HOLE FORMATION
-17
Stellar-mass BHs are predominately formed in the core
-16.5 of massive stars. For stars
NUVwith zero-age-mainR
collapse
sequence (ZAMS) masses of ! 10M⊙ , the iron core collapses
-16once its mass exceeds the Chandrasekhar limit, forming a proto-NS. The proto-NS cools via intense neutrino
emission of ∼ 1053 erg, which is believed to ultimately
-15.5
power the SN explosion, at least in some progenitors (e.g.,
O’Connor & Ott 2011; Ugliano et al. 2012; Horiuchi et al.
-15
2014;
Pejcha & Thompson 2015). If, however, the accretion
-3 the
-2 proto-NS
-1 0stalls1 and2never
3 reverses
4 5to un6
shock onto
bind the stellar envelope, the
t continued
[day]accretion eventually
leads the proto-NS to collapse into a BH. We are interested
latter case
this paper. of
Even
in thisdisk
case,
however,
Figurein3.the
Absolute
ABinmagnitude
fallback
outflow
emisif the progenitor is a RSG, the change in core mass asso-
Using a simple analytic model, we have calculated the fal
back disk outflow emission from the formation of BHs in oth
emi
erwise
failed
supernova
explosions.
disks power ou
plosions
(e.g.,
Woosley
& Bloom
2006). It Fallback
is not guaranteed,
however,
every,can
or be
evenobserved
the majority
of,
flows
whose that
emission
as a rapidly-evolvin
42−43 momenWR (∼
or BSG
progenitors
with
significant
angular
a few days)41-43
luminous (∼ 10
erg s−1 ) blue (T
-1
tum produce
successful GRBs. For example, the power104bol
K) transient. This outflow can be observed only when
ful jets that
produce GRBs may require large-scale magnotinenshrouded
a quasi-spherical
explosion.
This like
neticisflux
the stellar by
progenitor
as well as rapid
rorequires
compact
progenitors,
and
tation
(Tchekhovskoy
& Giannios
2014). like
SinceWRs
large-scale
4 BSGs, whic
magnetic
tendsbound
to slowenvelopes
down the rotation
theneutrino
core
haveflux
tightly
so that ofthe
radiatio
during
stellar
evolution,
it
could
be
that
the
combination
of
of the proto-NS binding energy would not lead to significan
conditions required to produce luminous GRBs is somewhat
mass ejection. Our simplified treatments of e.g., the fallbac
rare. In this paper, we argue that an alternative electrodisk counterpart
formation,associated
the (thermo-)dynamics
the outflow, an
magnetic
with BH formation of
during
the transfer
of the
cooling
radiation,
need tois be
the collapse
of rapidly
rotating
WR or
BSG progenitors
a followed u
by moreUV-optical
detailed numerical
studies.
fast luminous
transient broadly
similar to some
ü t ~ Days to 10 days
ü L
~ 10
erg s
ü blue continua with T ~10 K
events discovered by Pan-STARRs and PTF in the last ∼ 5
The PS1-MDS Transients
Pan-STARRS1 Medium Deep Survey (PS1-MDS) for Rapidly Evolving and Luminous Transients Drout+14
The Astrophysical Journal, 794:23 (23pp), 2014 October 10
Drout et al.
ü t1/2 < 12 day --- rapidly evolving than any SN type
ü Lpeak ~ 1042-43 erg s-1 --- luminous as bright SNe
ü Tpeak ~ a few 104 K --- blue
ü No line blanketing --- not powered by the radioactive decay
ü Host Gal. = star forming Gal. --- related to massive stars
Figure 1. PS1 absolute magnitude, rest-frame, light curves for gold sample transients. Circles represent griz detections and triangles represent 3σ upper limits.
Vertical dashed
indicate epochs
when spectroscopic
acquired. The gray shaded region
the R-band
Typerare
Ibc template from Drout et al. (2011),
ülinesEvent
rate
~ 4-7observations
% of were
core-collapse
SNis --not
normalized to the peak magnitude of the PS1-MDS transient.
P1
(A color version of this figure is available in the online journal.)
Summary and Discussion
How to find newborn black holes
• Vanishing supergiants
• Luminous red novae
• Fast blue transients
üa day to 10 day depending on progenitor structure
üLbol ~ 1041-43 erg s-1
üBlue continua with T ~ 104 K
üThe PS1-MDS transients are from WRs and BSGs?
ümay not be rare (~5% of CCSNe).
• Multi-Messenger Approach
üRadio
üGravitational wave
Back up
Now is the good timing
Annu. Rev. Astro. Astrophys. 2006.44:49-92. Downloaded from www.annualreviews.org
Access provided by University of Nevada - Las Vegas on 10/05/15. For personal use only.
Figure 1
Scale drawings of 16 black-hole binaries in the Milky Way (courtesy of J. Orosz). The
Sun–Mercury distance (0.4 AU) is shown at the top. The estimated binary inclination is
indicated by the tilt of the accretion disk. The color of the companion star roughly indicates
its surface temperature.
Optical Transients
383
Transients in the local Universe
−24
45
SCP06F6
SN2005ap SN2008es PTF09cnd
SN2006gy
PTF09cwl PTF09atu
PTF10cwr
−22
Thermonuclear
Supernovae
43
10
SN2002bj
−18
Core−Collapse
Supernovae
PTF10bhp
.Ia Explosions
−16
PTF11bij
PTF09dav
PTF10iuv
SN2005E
42
10
SN2008ha
−14
Ca−rich
Transients
41
10
SN2008S
PTF10acbp
Peak Luminosity [erg s−1]
V
44
10
SN2007bi
−20
Peak Luminosity [M ]
10
Luminous Supernovae
NGC300OT
−12
PTF10fqs
P60−M82OT−081119
−10
40
10
M85 OT
V838 Mon
Classical Novae
−8
Luminous
Red
Novae
M31 RV
39
10
P60−M81OT−071213
38
−6
10
0
10
1
10
Characteristic Timescale [day]
2
10
Kasliwal 11
Figure 4. Framework of Cosmic Explosions in the Year 2011 (Kasliwal 2011). Note that until 2005 (Fig. 1),
we only knew about three classes (denoted by gray bands). In the past six years, systematic searches,
serendipitous discoveries and archival searches have uncovered a plethora of novel, rare transients. Discov-
Possible Outcomes in a BH Formation
No. 2, 2008
DO SUPERGIANTS VANISH?
1339
Dimer class of SNe ?
(e.g., 1987A)
probably
not rare
Fig. 3.— Possible outcomes in forming a BH. The optical signatures of the ‘‘no explosion’’ scenarios are little explored.
significantly less luminous than before it died. Other known sources
galaxies (axis ratios <0.3). Only 40 galaxies need to be observed
# 5
10
The PS1-MDS Transients
Drout et al.
Figure 7. Phase space of SNe: peak luminosity vs. rest-frame time above halfmaximum for a variety of SNe. The PS1-MDS transients described in this paper
are shown as red stars. They span an order of magnitude in peak luminosity and
significantly increase the number of known transients with short characteristic
Drout+14
The PS1-MDS Transients
The Astrophysical Journal, 794:23 (23pp), 2014 October 10
Figure 11. Best-fit blackbody temperatures (top panel) and radii (bottom panel)
as a function of time. In general, temperatures evolve from around 20,000 K
near maximum to 7000 K at later times. Exceptions include PS1-12brf (whose
initial temperature is around 50,000 K), PS1-10ah (whose temperature rise to
Drout+14
Figure 12. Left panel: p
The PS1-MDS Transients
The Astrophysical Journal, 794:23 (23pp), 2014 October 10
Drout et al.
Blue Continua
No Line Blanketing
Figure 14. Mg ii absorption in the spectrum of PS1-12bv. Left: vertical lines
mark the systemic velocities of the host galaxy as measured from [O ii] λ3727
and Hβ in emission. The red line signifies the extent of the resolved blue “shelf,”
which may be due to absorption in the CSM surrounding the progenitor. Right:
an example multiple component fit to the absorption feature. The green (blue)
components are blue shifted ∼700 km s−1 (1500 km s−1 ) with respect to the
host galaxy and have a FWHM of 550 km s−1 (900 km s−1 ). The galaxy [O ii]
λ 3727 feature has a FWHM of 300 km s−1 .
(A color version of this figure is available in the online journal.)
5.2. Lack of Nebular Features in PS1-12bb
Figure 13. Explosion spectra for five PS1-MDS transients (colored) in comparison to events from the literature (black). With the exception of PS1-12bb our
events are dominated by a blue continuum, with a lack of strong P Cygni fea-
We obtained two spectra for PS1-12bb. At the time of the
second epoch (+33 days) the i-band light curve of PS1-12bb
had declined by ∼3 mag. Similar to the spectrum obtained near
maximum light, this spectrum is dominated by continuum (with
some contribution from its host galaxy’s light). This is in contrast
to any other “late time” spectra of a rapidly declining SN that has
been obtained to date. Both SN 2005ek (Drout et al. 2013) and
SN 2010X (Kasliwal et al. 2010) were observed between 10
and 35 days post-maximum. Both events (Type Ic)
displayed
Drout+14
a growing emission component in the Ca ii NIR triplet. By
+23 days the spectrum of SN 2010X was dominated by the
Ca ii NIR triplet between 8000 and 9000 Å. This is not the case
r 10
The PS1-MDS Transients
Drout et al.
Table 5
Host Galaxy Properties
Methodb
PP04N2
PP04N2
PP04N2
Z94
PP04N2
···
KD02
KD02
PP04N2
log(Mgal /M⊙ )
SFR
(M⊙ yr−1 )
Figure 16. Cumulative distributions of various host galaxy properties: stellar mass, metallicity, star formation rate, and specific star formation rate. In all panels the
red line(s) represents the rapid evolving SNe presented in this work. In panels 3 and 4 the solid red line represents the six gold/silver objects with host galaxy spectra
and the dashed red line includes lower limits based on the explosion spectra of the remaining four objects. Also shown are distributions from LGRBs and core-collapse
(black and
green Norm.
lines, respectively;
see text for references). The dashed green line in panel 2 is the untargeted Type Ibc SN from Sanders et al. (2012). The
sSFR SNe
Offset
Offset
Offsetc
vertical dashed line in panel 2 signifies solar metallicity.
−1
(Gyr ) (arcsec)
(kpc)
(A color version of this figure is available in the online journal.)
∼0.9
0.42
0.68
0.21
9.12+0.18
−0.21 Journal,∼1.2
The Astrophysical
794:23 (23pp), 2014 October 10
9.12+0.18
∼2.1
∼1.6
0.95
−0.24
10.17+0.20
∼4.3
∼0.3
1.54
−0.37
>0.3
>2.4
0.47
8.01+0.61
−0.70
10.54+0.43
∼2.3
∼0.06
4.93
−0.12
9.89+0.62
>0.3
>0.03
0.97
−0.52
8.73+0.20
>0.3
>0.5
0.42
−0.17
8.78+0.15
>0.1
>0.2
0.33
−0.12
8.96+0.27
∼1.9
∼2.0
1.40
−0.01
+0.07
8.68−0.13
∼4.5
∼9.3
0.14
2.39
0.85
2.90
0.69
Drout et al.
test yields a 14% (37%) probability that the rapidly evolving
12.58
0.98
transients hosts are drawn from the same population as the
8.81
1.05
CC-SN (LGRB) hosts. Thus, while our host galaxies skew
10.99 lower masses,
1.92
slightly toward
there is no statistical evidence for
0.50
a different10.29
parent population
from either LGRBs or CC-SNe.
2.16
6.2.
0.44
Metallicity
the8.28
Markov 0.82
Chain
Using
Monte Carlo method described in
1.04(2012) 0.32
Sanders et al.
we measure the emission line fluxes and
metallicities of nine of the host galaxies. Due to the varying
wavelength coverage and quality of our spectra, it was not
Diagnostics utilized varied between objects.
possible to use the same strong line diagnostic for all of the
Pettini & Pagel (2004); Z94 = Zaritsky et al. (1994); KD02 = Kewley & Dopita (2002); M91
galaxies. In Table 5 we list both the measured value and the
Figure 4. Gold and silver sample explosion environments.
Images areutilized.
from PS1 templates.
explosion sitesof
are comparison,
marked with red crosses.
diagnostic
For Transient
the purposes
we then
s of the host galaxy.
(A color version
of thisall
figure
available
in the online
journal.)
convert
ofis the
values
to the
R23 system of Kewley & Dopita
The
Astrophysical Journal, 794:23 (23pp), 2014 October 10
selection
criteria listed
in Section
2.2. The
absolute&
magnitude,
an aperture to encompass all of the visible light.(2002)
Cross-checks
using the
calibration
relations
from
Kewley
Ellison
Figure 17. Mass–metallicity relation for nine gold/silver transients (red stars).
rest-frame, light curves for these events are shown in Figure 1.
show the PS1 and SDSS magnitudes are consistent.
In Figures 4
g to accelerate
(2008).
Our silver sample is composed of four objects. These are
and 6 we show the environments immediately surrounding the
Also plotted are contours representing the 53,000 SDSS star-forming
galaxies follow-up
our spectroscopic
m s−1 ) to these
Many of the host
haveasmetallicities
aretheroughly
eventsgalaxies
that were noted
rapidly evolvingthat
during
normal
transients.
from Tremonti et al. (2004) (shaded region, lines represent discovered
the 2.5, 16, 50,
84, PS1-MDS
in the
s due to CSM
operations
of the PS1-MDS
but that possess
lightIn
curves
solar, with a median
value
of log (O/H)
+ 12sparser
= 8.8.
the
and 97.5 percentile of the distribution in each bin) and the long-duration
GRBtimescale i
Because
their
2.6. Optical Spectroscopy
2).16
In we
all cases
the
observed
light curves
are sufficient
diative driving
hosts from Levesque et al. (2010b). Unlike the LGRB hosts,due
which
are offset
to or maint
second panel of (Figure
Figure
plot
their
cumulative
distribution.
to poor
weather
to
characterize
them
as
rapidly
evolving,
but
sparse
enough
lower metallicities, our sample is consistent with being drawn from the greater
peed found for
Spectroscopic follow-up for the PS1-MDS is carried
out on aare offset to a significantly higher metallicity than
Our hosts
such that they fail the systematic selection criteria described
on the efficiency with which
number of telescopes, with the MMT, Magellan and Gemini
SDSS population.
eas the speeds
above.(Svensson
For instance,et
PS1-13ess
onlySavaglio
has one band
deep
either
LGRB hosts
al. 2010;
etwith
al. a2009;
We therefore use a Monte Ca
bearing most of the load. Spectra are acquired
for roughly
(A color version of this figure is available in the online journal.)
limit in the 9 days prior to observed maximum (as opposed to
of a wind from
in Quimby et al. (2012, 201
10% of the transients identified by the photpipeLevesque
pipeline, withet al. 2010a; Levesque et al. 2010b; Graham &
the requisite two). For a majority of this manuscript, the silver
spectra of five
is one offinal
ourselections left to the observer. We obtainedFruchter
which each PS1-MDS field
2013) objects
or thewill
untargeted
Ibc
hosts
from
Sanders
be analyzed Type
with our
gold
sample,
as they
further
transients while they were active, including two observations
hotometrically
function of distance.
inform
theand
properties
of rapidly
evolving
transients. However,
et
al.
(2012)
(black
line
dashed
green
line,
respectively)
with
of PS1-12bb. The epochs on which these spectra were taken
6.4.
Star
Formation
Rates
in Section 7 when calculating volumetric rates, only objects that
We begin by constructing
uj (which are
was
indicated by a dashed vertical line in the appropriate
panels
a !0.5%
probability
of being drawn from the same population.
pass the well defined set of selection criteria outlined above will
lution
templates based on t
of Figures 1 and 2. Host galaxy spectra were also
obtained
for
m an optically
The entire CC-SN
sample from Kelly & Kirshner (2012) (solid
be considered.
We estimate host galaxy SFRs for the six events
for whicha luminosity
six transients. A summary of our spectroscopic data is given in
then construct
′
′
may therefore
In
Figure
4
we
show
the
25
×
25
region
surrounding
the
green line) is shifted to even higher metallicities than our sample,
Table 3.
we possess galaxy spectra (Figure 15) by measuring
their Hαdistribution
trinsic Gaussian
gold and silver transients, all of which have an associated host.
g wind in the
Initial reduction (overscan correction, flat fielding,
extraction,
although
we caution
fraction
these events
Narrowaemission
andof
absorption
lines werewere
used todiscovered
measure the
line
fluxes
and
applying
the
relation
of
Kennicutt
(1998).
tion
to
accountInfor host gal
wavelength calibration) of all long slit spectra was carried out
by targeted surveys
(which
toward
higher
redshift
to each bias
host. These
range
from z metallicities).
= 0.074 (PS1-10ah)
Quimby fraction
et al. (2013) we ad
each
case
we
apply
a
rough
correction
for
the
covering
using the standard packages in IRAF. Flux calibration and
d by either
a
to z = 0.646 (PS1-11bbq) with a median redshift of z = 0.275.
telluric correction were performed using a set of custom idl
(Hatano
et
al. 1998) for the h
of
our
spectra
by
scaling
to
our
PS1
photometry
of
the
hosts.
In
Table
1 we list the redshift, Relation
luminosity distance, and Milky
system along
6.3.
Mass–Metallicity
scripts (see, e.g., Matheson et al. 2008; Blondin et al. 2012)
point
for thecases
intrinsic distrib
Way reddening in the direction of each transient (Schlafly &
We
do
not
correct
for
intrinsic
extinction.
In
the
three
andto
standard star observations obtained the same night as the
xy. In order
same
mean and variance as
Finkbeinertransients.
2011). Throughout
this paper we correct only for
Figure
15. Host
galaxy
for
six
of the 17
gold/silver
Nebular
where
both
Hα
and
Hβ
are
detected
the
decrement
is
reasonably
science exposures.
Spectra
obtained
withspectra
the Hectospec
multiIn
Figure
we
plot
the
stellar
mass
versus
metallicity
for
Figure
18.a Projected host offsets for our sample (red) in comparison to
λλ 2796,2803)
Milky Way extinction. All calculations in this paper
assume
of our gold sample after pe
fiber spectrograph
(Fabricant
et al.labeled
2005) by
were
reduced
emission
lines are
dashed
vertical
lines.flat ΛCDM
−1 astronomical
consistent
withphysical
zero extinction.
SFRs
range from
nine
of using
the transients
(redcosmology
stars) versus
∼53,000
SDSS
other
transients.
Left panel:
offsets. Right The
panel:resulting
offsets
with H0the
= 71
km s−1 Mpc
, Ωstarm =
o this portion
the Malmquist bias. As a c
the IRAF package “hectospec” and the CfA pipeline designed
−1
normalized
by the g-band
half-light
radiiand
of theare
hosts.
Our sample
most closely
0.27,from
and ΩΛSDSS
= 0.73. (Tremonti et al. 2004,
1–5
M
yr
listed
in
Table
5.
For
the
four
forming
galaxies
shaded
⊙
for this instrument.
he right panel
Carlo remaining
simulations distributin
distribution of core-collapse SN offsets.
In Figure 5 we plot peak absolute magnitude traces
versusthe
redobjects
we
may
place
lower
limits
on
the
SFR
by
measuring
thein space. Th
regions).
Also
shown
are
the
low
redshift
LGRB
hosts
from
explosion
site
offsets
of
the
host
galaxies
in
comparison
to
the
tribution
evenly
shift in grizP1 for the gold/silver transients. Stars
represent
dividual Mg ii
(A
color
version
of
this
figure
is
available
in
the
online
journal.)
3. SAMPLE OVERVIEW Levesque et al. (2010b, black). The hosts of the PS1-MDS
Hβ
emission
line
flux
from
the
explosion
spectra
(Figure
13)
hosts
of
other
classes
of
transients.
our
observed
magnitudes
corrected
for
distance
and
MW
ex>1.5 mag above the nomina
s their sum. In
magnitudes
been
The 14 rapidly evolving transients we identify
in the PS1and assuming zero extinction. We do not attempt
correct foragree well
rapidly
evolvingtinction.
SNe Circles
appearrepresent
to be absolute
consistent
with that
thehave
bulk
of
forto
detectability)
hat were both
k-corrected to the rest-frame grizP1 bandpasses based
on best- we can draw, there is no statistical evidence that the
conclusions
MDS can be usefully split into three groupsstar-forming
based
on (1) galaxies
6.1. Mass
the
covering
fraction
in
these
cases.
The
lower
limits
range
fromdistance. Th
in
SDSS.
This
is
in
contrast
to
the
hosts
luminosity
and
fit blackbodies (see Section 4). We see that our sample
spans aare drawn from different
M absorptions
the quality of their observed light curves and (2) constraints
samples
−1progenitor populations. The
(i-band)
Gaussian
0.1–0.3
M
yr
.
In
the
second
and
third
panels
of
Figure
16
we distribu
of
LGRBs
and
Type
I
SLSNe,
both
which
have
shown
wide range
of absolute
peakof
magnitude
(−17
> been
M > −20).
⊙
available on their After
distances.
For
the
rest
of
the
manuscript
these
SFRs measured for our hosts are clustered around the median
correcting the host griz-band photometry for Milky
M91
Host Gal. = SF Gal.
Drout+14
The PS1-MDS Transients
The Astrophysical Journal, 794:23 (23pp), 2014 October 10
Figure 19. PS1-MDS rapid-transient detection efficiencies for our rate calculation. Solid lines indicate detection efficiencies for individual PS1-MDS fields
as a function of redshift (bottom axis) or luminosity distance (top axis). A black
solid line indicates the detection efficiencies for the survey as a whole, which
peak around 11% at z = 0.1. The dashed black line indicates the distance at
which we expect our detections to be located. This peaks at z = 0.2. The actual
4%-7% of CCSN@z =0.2
reasonable agre
derived rates to
decrease.
In addition,
examine every p
events. As a by
transients will
will be missed i
timescales. It is
this factor, but
derived rates.
Finally, we a
that no objects
SN 2005ek and
Only PS1-12bb
but its spectra
Ca ii NIR featur
events. Using th
curve and temp
through a simila
Gaussian lumin
Drout+14
find that, due to
SN 2005ek, the
similar transien