arXiv:0802.1712v2 [astro-ph] 5 May 2008
An extremely luminous X-ray outburst at the birth of
a supernova
A. M. Soderberg
1,2
, E. Berger
1,2
, K. L. Page
3
, P. Schady
4
, J. Parrent
5
,
D. Pooley
6
, X.-Y. Wang
7
, E. O. Ofek
8
, A. Cucchiar a
9
, A. Rau
8
,
E. Waxman
10
, J. D. Simon
8
, D. C.-J. Bock
11
, P. A. Milne
12
,
M. J. Page
4
, J. C.Bar entine
13
, S. D. Barthelmy
14
, A. P. Beardmore
3
,
M. F. Bietenholz
15, 16
, P. Brown
9
, A. Burr ows
1
, D. N. Burrows
9
,
G. Byrngelson
17
, S. B. Cenko
18
, P. Chandra
19
J. R. Cummings
20
, D. B. Fox
9
,
A. Gal-Yam
10
, N. Gehrels
20
, S. Immler
20
, M. Kasliwal
8
, A. K. H. Kong
21
,
H. A. Krimm
20, 22
, S. R. Kulkarni
8
T. J. Maccarone
23
, P. M´esz´aros
9
,
E. Nakar
24
, P. T. O’Bri en
3
, R. A. Overzier
25
, M. de Pasquale
4
, J. Racusin
9
,
N. Rea
23
, and D. G. York
26
1
Department of Astrophysical Sciences, Princeton University, Ivy L ane, Pr inceton, NJ 08544, USA
2
Carnegie Observa tories, 813 Santa Barbar a St., Pasadena, CA 91101, USA
3
Department of Physics and Astronomy, University of Leicester, Leicester LE1 7RH, UK
4
Mullard Spac e Sci. Lab., Univ. College London, Holmbury St. Mary, Dorking, Surrey RH5 6NT, UK
5
Physics and Astronomy Department Dartmouth College, Hanover, NH 03755, USA
6
Astronomy Department, University of Wisconsin, 475 North Charter Street, Madison, WI 53706, USA
7
Department of Astronomy, Na njing University, Nanjing 210093, China
8
Department of Astronomy, 105-24 , California Institute of Technology, Pasadena, CA 91125, USA
9
Dept. of Astronomy and Astrophys ics, Pennsylvania State University, University Park, PA 16802, USA
10
Faculty of Physics, Weizmann Institute of Science, Rehovot 76100, Israel
11
Radio Astr onomy La boratory, University of California, Berkeley, CA 94720, USA
12
Steward Observator y, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721, USA
13
Department of Astronomy, University of Texas at Austin, Austin, TX 78712, USA
14
NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA
15
Department of Physics and Astronomy, York University, Toronto, ON M3J 1P3, Canada
16
Hartebeestehoek Radio Observatory, PO Box 443, Krugersdorp, 174 0, South Africa
17
Dept. of Physics and Astronomy, Clemson University, Cle mson, South Carolina 29634, USA
18
Space Radiation Labor atory 220-47, California Institute of Technology, Pa sadena, CA 91125, USA
19
Department of Astronomy, University of Virginia, P.O. Box 40032 5, Charlottesville, VA 22904, USA
20
CRESST and NASA Goddard Space Flight Center , Greenbelt, MD 20771, USA
21
Institute of Astronomy and Dept. of Physics, Na tio nal Tsing Hua University, Hsinchu, Taiwan
22
Univer sities Space Research Assoc iation, 10211 Wincopin Circle, # 500, Columbia, MD 21044 , USA
23
Schoo l of Physics and Astr onomy, University of Southampton, Southampton SO17 1BJ, UK
24
Theoretical Astrophysics 130- 33, California Institute of Technology, Pasadena, CA 91125, USA
25
Max-Planck-Institut fur Astrophysik , D-85748 Garching, DE
26
Dept. of Astronomy and Astrophysics , Univ. of Chicago, 5640 S. Ellis Avenue, Chicago, IL 60637, USA
2 Soderberg et al.
Massive stars end their short lives in spectacular explosions, supernovae, that
synthesize new elements and drive galaxy evolution. Throughout history su-
pernovae were discovered chiefly through their delayed optical light, preventing
observations in the first moments (hours to days) following the explosion. As
a result, the progenitors of some supernovae and the events leading up to their
violent demise remain intensely debated. Here we report the serendipitous
discovery of a supernova at the time of explosion, marked by an extremely
luminous X-ray outburst. We attr ibute the outburst to the break-out of the
supernova shock-wave from the progenitor, and show that the inferred rate of
such events agrees with that of all core-collapse supernovae. We forecast that
future wide-field X-ray surveys will catch hundreds of supernovae each year
in the act of explosion, and thereby enable crucial neutrino and gravitational
wave detections that may ultimately unravel the explosion mechanism.
Stars more massive than ab out eight times the mass of the Sun meet their death in
cataclysmic explosions termed supernovae (SNe). These explosions give birth to the most
extreme compact objects – neutron stars and black holes – and enrich their environments
with heavy elements. It is generally accepted that SNe are triggered when the stellar core
runs out of fuel for nuclear burning and thus collapses under its own gravity (see Ref. 1
and references t herein). As the collapsing core rebounds, it generates a shock-wave that
propagates through and explodes the star.
The resulting explosion ejects several solar masses of stellar material with a mean
velocity
2
of about 10
4
km s
−1
, or a kinetic energy of about 10
51
erg. Less than a solar mass
of
56
Ni is synthesized in the explosion, but its subsequent radioactive decay powers
1
the
luminous optical light observed to peak 1 to 3 weeks after the explosion. It is through this
delayed signature that SNe have been discovered both historically and in modern searches.
While the general picture of core-collapse has been recognized for many years, the de-
tails of the explosion remain unclear and most SN simulations fail to produce an explosion.
The gaps in our understanding are due to the absence of detailed observations in the first
days after the explosion, and the related difficulty in detecting the weak neutrino
3
and
gravitational wave signatures of the explosion. These signals offer a direct view of the
explosion mechanism but require the discovery of SNe at the time of explosion.
In t his paper we describe our serendipitous discovery of an extremely luminous X-
ray outburst that marks the birth of a SN of Type Ibc. Prompt bursts of X-ray and/or
ultraviolet (UV) emission have been theorized
4,5
to accompany the break-out of the SN
A SN at the Time of Explosion 3
shock-wave through the stellar surface, but their short durations (just seconds to hours)
and the lack of sensitive wide-field X-ray and UV searches have prevented their discovery
until now.
Our detection enables a n unprecedented early and detailed view of the SN, allowing
us to infer
6
the radius of t he progenitor star, its mass loss in the final hours prior to the
explosion, and the speed of the shock as it explodes the star. Drawing on optical, UV,
radio, and X-ray o bservations we show that the progenitor was compact (R
∗
≈ 10
11
cm)
and stripped of its outer Hydrogen envelope by a strong and steady stellar wind. These
properties are consistent
7
with those of Wolf-Rayet (WR) stars, the favored
8
progenitors
of Type Ibc SNe.
Wolf-Rayet stars are also argued
9
to give rise to gamma-ray bursts (GRBs), a related
but rare class of explosions characterized by highly-collimated relativistic jets. Our ob-
servations, however, indicate an ordinary spherical and non-relativistic explosion and we
firmly rule out a GRB connection.
Most important, the inferred rate of X-ray outbursts indicates that all core-collapse
SNe produce detectable shock break-out emission. Thus, future wide- field X-ray surveys
will uncover hundreds of SNe each year at the time of explosion, providing the long- awaited
temporal and positional triggers for neutrino and gravitational wave searches.
Discovery of the X-ray Outburst
On 2008 Jan 9 at 13:32:49 UT, we serendipitously discovered an extremely bright X-
ray tra nsient during a scheduled Swift X-ray Telescope (XRT) observation of the galaxy
NGC 2770 (d = 27 Mpc). Previous XRT observations of the field just two days earlier
revealed no pre-existing source at this location. The transient, hereafter designated as
X-ray outburst (XRO) 08 0109, lasted about 400 s, and was coincident with one of the
galaxy’s spiral arms (Figure 1). From observations described below we determine that
XRO 080109 is indeed located in NGC 2770, and we thus adopt this association hereafter.
The temporal evolution is characterized by a fast rise and exponential decay, often
observed for a variety of X-ray flare phenomena (Figure 1). We determine the onset of
the X- r ay emission to be 9
+20
−8
s before the beginning of the observation, implying an
outburst start time of t
0
≈ Jan 9.564 UT. The X-ray spectrum is best fit by a power law
[N(E) ∝ E
−Γ
] with a photon index of Γ = 2.3 ± 0.3, and a hydrogen column density
of N
H
= 6.9
+1.8
−1.5
× 10
21
cm
−2
, in excess of the absorption within the Milky Way (see
Suppl. Info.). The inferred unabsorbed peak flux is F
X,p
≈ 6.9 × 10
−10
erg cm
−2
s
−1
(0.3 − 10 keV). We also measure significant spectral softening during the outburst.
The XRO was in the field of view of the Swift Burst Alert Telescope (BAT; 15−150 keV)
4 Soderberg et al.
beginning 30 min before and continuing throughout the outburst but no γ-ray counterpart
was detected. Thus, t he outburst was not a GRB (see also Suppl. Info.). Integrating over
the duration of the outburst, we place a limit on the gamma-ray fluence of f
γ
∼
<
8 × 10
−8
erg cm
−2
(3σ), a factor of three times higher than an extrapolat io n of the X-ray spectrum
to the BAT energy band.
The total energy of the outburst is thus E
X
≈ 2 × 1 0
46
erg, at least three or ders of
magnitude lower
10
than GRBs. The peak luminosity is L
X,p
≈ 6.1 × 10
43
erg s
−1
, several
orders of magnitude la r ger than the Eddington luminosity (the maximum luminosity for a
spherically-accreting source) of a solar mass object, outbursts from Ultra-luminous X-ray
sources and Type I X-ray bursts. In summary, the properties of XRO 08 0109 are distinct
from those of all known X-ray transients.
The B ir th of a Supernova
Simultaneous observations of the field with the co-aligned Ultraviolet/Optical Telescop e
(UVOT) on- board Swift showed no evidence f or a contemporaneous counterpart. How-
ever, UVOT observations j ust 1.4 hr after the outburst revealed
11
a brightening UV/optical
counterpart. Subsequent ground-based optical observations also uncovered
11–13
a coinci-
dent source.
We promptly obtained optical spectroscopy of t he counterpart with the Gemini North
8-m telescope beginning 1.74 d after the outburst (Figure 2). The spectrum is characterized
by a smooth continuum with narrow absorption lines of Na I λλ5890, 5896 at the redshift
of NG C 27 70. More importantly, we note broad absorption features near 5200 and 5700
˚
A
and a drop-off beyond 7000
˚
A, strongly suggestive of a young SN. Subsequent observations
confirmed these sp ectral characteristics,
11,14
and the transient was classified
11,15
as Type
Ibc SN 2008D based on the lack of hydrogen and weak silicon features.
Thanks to the prompt X-ray discovery, the temporal coverage of our optical spectra
exceeds those of most SNe, rivaling even the best-studied GRB-associated SNe (GRB-SNe),
and SN 1987A (Figure 2). We see a clear evo lution from a mostly featureless continuum
to broad absorption lines, and finally to strong absorption features with moderate widths.
Moreover, our spectra reveal the emergence of strong He I features within a few days of
the outburst (see also Ref. 16). Thus, SN 2008D is a He-rich SN Ibc, unlike
17
GRB-SNe.
Observations at high spectral resolution further reveal significant host gala xy extinction,
with A
V
≈ 1.2 − 2.5 mag (see Suppl. Info.).
The well-sampled UV/optical light curves in ten broadband filters (2000 − 10, 000
˚
A)
exhibit two distinct components (Figure 3). First, a UV-dominated component that peaks
about a day after the X-ray outburst, and which is similar to very early observat io ns
18
of
A SN at the Time of Explosion 5
GRB-SN 2006aj . The second component is significantly redder and peaks on a timescale
of about 20 days, consistent with o bservations of all SNe Ibc. Accounting for an extinction
of A
V
= 1.9 mag (Figure 3), the absolute peak brightness of the second component is
M
V
≈ −16.7 mag, at the low end of the distribution
19
for SNe Ibc and GRB-SNe.
A Shock Break-out Origin
Since some SNe Ibc harbor GRB jets, we investigate the possibility that the XRO is
produced by a relativistic outflow. In this scenario, the X-ray flux and simultaneous upper
limits in the UV/optical require the outflow to be ultra-relativistic with a bulk Lorentz
factor, γ ≈ 90, but its radius to be only R ≈ 10
10
cm; here γ ≡ ( 1 − β
2
)
−1/2
and β ≡ v/c,
where v is the outflow velocity and c is the speed of light. However, given the observed
duration of the outburst, we expect
20
R ≈ 4γ
2
ct ≈ 10
17
cm, indicating that the relativistic
outflow scenario is not self-consistent (see Suppl. Info. for details).
We are left with a trans- or non-relativistic origin for the outburst and consider SN
shock break-out as a natural scenario. The break-out is defined by the transition from
a radiation- mediated to a collisional (or collisionless
21
) shock as the optical depth of the
outflow decreases to unity. Such a tr ansition has long been predicted
4,5
to produce strong,
thermal UV/X-ray emission at the time of explosion. A non-thermal component at higher
energies may be produced
22
by multiple scatterings of the photons between the ejecta and
dense circumstellar medium (bulk Comptonization).
We attribute the observed non-thermal o utburst to Comptonized emission from shock
break-out indicating that the associated thermal component must lie below the XRT low
energy cutoff, ∼ 0.1 keV. With the reasonable assumption that the energy in the thermal
(E
th
) and Comptonized components is comparable, we constrain
6
the radius at which
shock break-out occurs to R
sbo
& 7 × 10
11
(T/0.1 keV)
−4/7
(E
X
/2 × 10
46
erg)
3/7
cm. This is
consistent with a simple estimate derived from the rise time of t he outburst, R
sbo
= cδt ∼
10
12
cm, and larger than the typical radii of WR stars,
23
R
∗
∼ 10
11
. We therefore attribute
the delayed shock break-out to the presence o f a dense stellar wind, similar
6,18
to the case
of GRB-SN 2006aj .
The shock velocity at break-out is
6
(γβ)
∼
<
1.1 and the outflow is thus trans-relativistic
as expected
24
for a compact progenitor. Using these constraints, the inferred optical depth
of the ejecta to thermal X-rays is τ
ej
≈ 1.5(E
X
/2×10
46
erg)(R
sbo
/7×10
11
cm)
−2
(γ−1)
−1
≈
3, and Comptonization is thus efficient, confirming our model. Equally important, as the
ejecta expand o utward the optical depth of the stellar wind decreases and the spectrum of
the Comptonized emission is expected
22
to soften, in agreement with the observed trend.
The shock break-out emission traces the wind mass loss rate of the progenitor,
˙
M, in the