An unusual stellar death on Christmas Day
ABSTRACT Long Gamma-Ray Bursts (GRBs) are the most dramatic examples of massive
stellar deaths, usually as- sociated with supernovae (Woosley et al. 2006).
They release ultra-relativistic jets producing non-thermal emission through
synchrotron radiation as they interact with the surrounding medium (Zhang et
al. 2004). Here we report observations of the peculiar GRB 101225A (the
"Christmas burst"). Its gamma-ray emission was exceptionally long and followed
by a bright X-ray transient with a hot thermal component and an unusual optical
counterpart. During the first 10 days, the optical emission evolved as an
expanding, cooling blackbody after which an additional component, consistent
with a faint supernova, emerged. We determine its distance to 1.6 Gpc by
fitting the spectral-energy distribution and light curve of the optical
emission with a GRB-supernova template. Deep optical observations may have
revealed a faint, unresolved host galaxy. Our proposed progenitor is a helium
star-neutron star merger that underwent a common envelope phase expelling its
hydrogen envelope. The resulting explosion created a GRB-like jet which gets
thermalized by interacting with the dense, previously ejected material and thus
creating the observed black-body, until finally the emission from the supernova
dominated. An alternative explanation is a minor body falling onto a neutron
star in the Galaxy (Campana et al. 2011).
AN UNUSUAL STELLAR DEATH ON CHRISTMAS DAY
C. C. Th¨ one1,2, A. de Ugarte Postigo3, C. L. Fryer4, K. L. Page5, J. Gorosabel1, D. A. Perley6, C.
Kouveliotou7, H. T. Janka8, M. A. Aloy9, P. Mimica9, J. L. Racusin10, H. Krimm10,11,12, J. Cummings10, S. R.
Oates13, S. T. Holland10,11,12, M. H. Siegel14, M. De Pasquale13, E. Sonbas10,11,15, M. Im16, W.-K. Park16, D. A.
Kann17, S. Guziy1, L. Hern´ andez Garc´ ıa1, K. Bundy6, A. J. Castro-Tirado1, C. Choi16, H. Jeong18, H.
Korhonen19, P. Kubanek1,20, J. Lim21, A. Llorente22, A. Moskvitin23,T. Mu˜ noz Darias24, S. Pak18, I. Parrish6
Draft April 2011
Massive stars can end their lives in many different ways. Long Gamma-Ray Bursts (GRBs) are the
most dramatic examples, releasing ultra-relativistic ejecta that produce non-thermal emission when
interacting with the surrounding medium (Zhang et al. 2004). Usually, those events are accompanied
by a supernova (SN) (Woosley et al. 2006). In a few low-redshift GRB-SNe we could observe the
actual breakout of the shock front from the surface of the star (Campana et al. 2006). Here we
present GRB 101225A, a very peculiar event at a distance of 1.6 Gpc. A bright X-ray transient
with a thermal component and an unusual optical counterpart followed an exceptionally long γ-ray
event detected by the Swift satellite. During the first 10 days, the optical emission evolved as an
expanding, cooling blackbody (BB) with a large initial radius, after which a faint SN was observed.
The absence of a normal GRB afterglow implies that some dense material, likely ejected by the
progenitor star, completely thermalized the high-energy emission. A possible progenitor is a helium
star/neutron star binary which underwent a common envelope phase, expelling its hydrogen envelope
prior to the explosion. The final merging process created a GRB-like event where we observe the shock
breakout of the secondary star before the high-energy emission gets thermalized in the collision with
the previously expelled shell, until finally the emission from the SN itself takes over. GRB 101225A
defines a new, rare type of blackbody-dominated GRB which explodes in a dense environment created
by the progenitor system itself.
Subject headings: Gamma-rays: Bursts: Individual: GRB101225
Based on observations collected at CAHA/Calar Alto, GTC/La
Palma under proposal 105-GTC34/10B, the Liverpool Telescope
at ORM/La Palma, the McDonald Observatory at the Univer-
sity of Texas at Austin, Gemini-North and Keck on Big Is-
1IAA - CSIC, Glorieta de la Astronom´ ıa s/n, 18008 Granada,
2Niels Bohr International Academy, Niels Bohr Institute,
Blegdamsvej 17, 2100 Copenhagen, Denmark
3Dark Cosmology Centre, Niels Bohr Institute, Univ.
Copenhagen, Juliane Maries Vej 30, 2100 Copenhagen, Denmark
4Los Alamos National Laboratory, MS D409, CCS-2, Los
Alamos, NM 87545, USA
5Department of Physics & Astronomy, Univ.
University Road, Leicester LE1 7RH, UK
6UC Berkeley, Astronomy Department, 601 Campbell Hall,
Berkeley CA 94720, USA
7NASA, Marshall Space Flight Centre, NSSTC, 320 Spark-
man Drive, Huntsville, AL 35805, USA
8Max-Planck-Institut f¨ ur Astrophysik, Karl-Schwarzschild-
Str. 1, 85748 Garching, Germany
9Departamento de Astronomia y Astrofisica, Universidad de
Valencia, 46100 Burjassot, Spain
10NASA, Goddard Space Flight Center, Greenbelt, MD
11Universities Space Research Association, 10211 Wincopin
Circle, Suite 500, Columbia, MD 21044-3432, USA
12Center for Research and Exploration in Space Science and
13Mullard Space Science Laboratory, Holmbury St.
Dorking, Surrey RH5 6NT, UK
14Dep. of Astronomy & Astrophysics, Pennsylvania State
Univ., 104 Davey Laboratory, University Park, PA 16802, USA
15University of Adıyaman, Department of Physics, 02040
16Center for the Exploration of the Origin of the Universe,
Dept. of Physics & Astronomy, Seoul National University, 56-1
San, Shillim-dong, Kwanak-gu, Seoul, Korea
17Th¨ uringer Landessternwarte Tautenburg, Sternwarte 5,
07778 Tautenburg, Germany
18School of Space Research,
Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446-701,
Kyung Hee University,1
19Finnish Centre for Astronomy with ESO (FINCA), Univer-
sity of Turku, V¨ ais¨ al¨ antie 20, 21500 Piikki¨ o, Finland
20Institute of Physics, Na Slovance 2, 180 00, Prague 8, Czech
21Dept. of Astronomy and Space Science, Kyung Hee Uni-
versity, 1 Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do
22Herschel Science Operations Centre, INSA, ESAC, Vil-
lafranca del Castillo, PO Box 50727, I-28080 Madrid, Spain
23Special Astrophysical Observatory of the Russian Academy
of Sciences, Nizhnij Arkhyz 369167, Russia
24INAF - Osservatorio Astronomico di Brera, Via E. Bianchi
46, 23807 Merate, Italy
arXiv:1105.3015v2 [astro-ph.HE] 17 May 2011
2C. C. Th¨ one et al.
On Dec. 25, 2010, 18:37:45 UT (T0), the BAT (15-
350 keV) instrument onboard the Swift satellite detected
GRB 101225A in an image trigger. It had a soft γ-ray
spectrum and one of the longest durations (T90> 2000
s, the time in which 90% of the energy is released) of any
burst observed by Swift (Sakamoto et al. 2010). Despite
its low fluence, the long duration makes its high-energy
output similar to those of other long-duration GRBs.
At T-T0 = 1400 s after the trigger, XRT, the X-ray
telescope onboard Swift, slewed to the position and de-
tected a bright X-ray counterpart detected up to 2 days
after the burst. An optical counterpart was found by
the Nordic Optical Telescope (NOT) at 1.54 h after the
BAT trigger (Xu et al. 2010a) and followed up both with
UVOT on-board Swift and several optical telescopes in
different ultraviolet, optical and infrared broad-band fil-
ters from 0.38 h to 2 months after the event (see Ap-
pendix).No counterpart in radio frequencies was de-
tected (Frail et al. 2011; Zauderer et al. 2011).
The most remarkable aspect of GRB 101225A is the
spectral energy distribution (SED) and its evolution.
The X-ray SED is best modeled with a combination of
an absorbed power-law and a BB up to 8ks, after which
the signal to noise is too low to disentangle the two
components. The contribution from the BB is around
20% of the total flux and the temperature of the BB is
1 − 1.5 keV with no significant temporal evolution. The
UV/optical/NIR (UVOIR) SED (see Fig. 1) can be mod-
eled with a cooling and expanding BB model up to 10
days after the event. The SEDs of GRBs usually follow
a power-law due to synchrotron emission created in the
external shocks between the GRB jet and the interstel-
lar medium (e.g., Zhang et al. 2004), which is completely
absent for GRB 101225A.
At 10 days, an additional component becomes evident
in the UVOIR SED accompanied with a flattening in the
lightcurve (see Fig. 2). Both the spectrum and lightcurve
are well reproduced with the Type Ic SN 1998bw (con-
nected to GRB 090428 and used as the classic broad-
lined Ic GRB-SN template) (Galama et al. 1998) at a
redshift of z=0.32 stretched by a factor of 1.25 in time
and with 1/12th of the luminosity of SN 1998bw (for a
discussion on the redshift determination, see appendix).
At this distance, the initial expansion velocity of the
UVOIR BB is 90,000 km/s slightly higher than typical
SN ejecta. At z =0.32 the SN has an absolute peak mag-
nitude of only MV= −16.7 mag. This would make it the
faintest SN associated with a long GRB ever detected, 2.3
mag fainter than SN 19988bw/GRB 980425, and signifi-
cantly fainter than the GRB 040924 SN, (Soderberg et al.
2006; Wiersema et al. 2008). In contrast, the isotropic-
equivalent energy release at z=0.32 is > 1.4 × 1051erg
in γ-rays, comparable to the total energy release of other
long-duration GRBs and more luminous than a number
of other low-redshift GRBs connected to SNe (Kann et
An optical spectrum taken 2 nights after the burst does
not show any obvious absorption or emission features (for
a plot of the spectrum see SI). The continuum can be fit
with a power law with Fν∝ ν0.9, consistent with being
on the Rayleigh-Jeans tail of the BB emission. The ab-
sence of characteristic SN lines at 2 d post-burst is not
surprising, in particular considering the late peak of the
SN associated with GRB 101225A. The absence of emis-
sion lines from the host galaxy indicates that it must
be rather faint (see appendix). Pre-imaging of the field
from the 3.5m CFHT (Richardson et al. 2011) revealed
no source at the position of GRB 101225A down to a 3σ
limit of i?> 26.8mag (see appendix). There is a marginal
(2σ) detection of an underlying source at g?= 27.2mag.
If this is the host galaxy at z = 0.32, its absolute magni-
tude would be Mabs= −13.7mag, ∼ 2mag fainter than
the faintest GRB host galaxy detected, the host of XRF
060218 (e.g., (Wiersema et al. 2007)).
3. RESULTS FROM THE BLACK-BODY EVOLUTION
An important clue to the origin of this event comes
from the modeling of the BB component at different
wavelengths. The BB component in X-rays has a radius
of ∼ 2 × 1011cm (1R?) and a temperature of ∼ 1keV
(107K) at 0.07 d. The UVOIR BB starts with a radius
of 2 × 1014cm and a temperature of 8.5 × 104K at the
same time with an initial expansion velocity of 90,000
km s−1. As shown in Fig. 3, the UVOIR BB evolution is
inconsistent with the radius and temperatures of the X-
ray BB component. Therefore, the emission in these two
bands must stem from different processes and regions.
A BB component in X-rays had been suggested for
XRF (X-ray flash) 060218/SN 2006aj (Campana et al.
2006), XRF 100316D/SN 2010dh (Starling et al. 2011)
and GRB 090618 (K. Page et al. in prep.), all GRBs asso-
ciated with Type Ic SNe, this component was attributed
to the shock breakout from the star or the dense cir-
cumstellar wind. A shock breakout was also proposed as
the origin of the X-ray emission observed in XRO (X-
ray outburst) 080109/SN 2008D, a Type Ib SN in NGC
2770 (Soderberg et al. 2008). However, it showed no ther-
mal component in the X-rays due to comptonization of
the BB and was not associated with a GRB. Both XRF
060218/SN 2006aj and XRO 080109/SN 2008D had a
thermal component in the optical over the first few days
(see appendix). For XRO 080109/SN 2008D, the ra-
dius and temperature evolution is steeper than for GRB
101225A, which makes the UVOIR BB emission consis-
tent with the expanding and cooling BB from the shock
breakout. The UVOIR BB of XRF 060218/SN 2006aj is
similar in radius and evolution to GRB 101225A while
the radius inferred from X-rays is much larger (about 20
R?). Despite the large initial radius of the X-ray thermal
component, which was attributed to the shock breakout
from the wind of the progenitor, the UVOIR BB emis-
sion cannot be ascribed to the continuation of the shock
front into the ISM from the initial breakout.
To explain the nature of this unusual event, we need to
explain four different emission processes: 1) The prompt,
very long high-energy emission. 2) The thermal compo-
nent observed in X-rays with a small radius of ∼ 1R?.
3) The afterglow of the GRB which is fully described by
an expanding, cooling BB with an initial radius of ∼ 30
AU, which starts simultaneously with the second orbit of
the X-ray observations. And 4) a late, faint SN emerg-
ing at around 10 d. The large radius of the UVOIR BB
Fig. 1.— The spectral energy distribution (SED) of the UV/optical/NIR (UVOIR) counterpart of GRB 101225A at different epochs after
the explosion. Filled circles are detections, triangles mark upper limits. Until 10 days, the SED is modeled with a simple BB, the last 3
epochs were modeled with an evolving SN Type Ic, similar to SN 1998bw associated with GRB 980425. The orange line on top of the BB
model at 2.0 days shows our flux-calibrated spectrum taken with the OSIRIS/GTC.
Fig. 2.— Light curves of GRB 101225A in X-rays (black, top panel) and UVOIR (bottom panel). The over plotted lines are the evolution
of the light curves in the different bands as estimated from the temporal evolution of the BB. Observations started almost simultaneously
in X-rays and optical/UV wavelengths. The X-rays reached a peak flux of 4.34 × 10−9erg cm−2s−1. After an initial shallow decay of
slope t−1.108±0.011up to 21ks, the X-rays show a strong decay with a slope of t−5.95±0.20, inconsistent with synchrotron emission. The
UVOIR light curves show a shallow maximum at the beginning, with different peak times for the different bands due to the peak emission
of the BB component passing through the spectrum. The second component emerging at around 10 days post-burst is the contribution of
the underlying SN, modeled with the GRB-SN 1998bw as a template, stretched in time by a factor of 1.25 and decreased in luminosity by
a factor of 12 (see appendix for more details).
emission requires that the material observed cannot be
directly connected to the stellar explosion itself but dense
material has to have been ejected some time before the
An appealing possibility to explain GRB 101225A is a
He-merger model with a common envelope (CE) phase
4C. C. Th¨ one et al.
Fig. 3.— Evolution of the temperature and radius of the BB com-
ponents in X-ray (blue dots) and UVOIR (black dots), excluding
the data beyond 10 d when the SN component becomes dominant.
The red dashed lines show the fit to the evolution of the UVOIR
BB alone which shows that the optical BB emission is not simply
from the cooling of the initial BB component in X-rays. The evo-
lution of the radius of the UVOIR BB follows a power-law with
R ∝ t0.23±0.02, incompatible with a Sedov-Taylor evolution as as-
sumed for late-time SN evolution. The temperature decline does
not follow a simple power-law which is well explained with a gas
sphere cooling due to combination of expansion and radiation loss
that has been proposed as a progenitor for GRBs (Fryer
et al. 1998; Zhang et al. 2001; Barkov et al. 2010). In
this model, a binary system consisting of two massive
stars survives the collapse to a neutron star or black hole
of the more massive component. When the second star
moves off the main sequence and expands, it engulfs the
compact remnant, leading to a CE phase and the ejec-
tion of the hydrogen and part of the helium core as the
compact remnant spirals into the center of the second
star Assuming the inspiral takes about 5 orbits or 1.5
yr and material is ejected at escape velocity, the outer
ejecta are at a few 1014cm when the compact remnant
reaches the center of star, consistent with the radius of
the outer BB we observe in the optical. Although an-
gular momentum will preferentially eject material along
the orbital plane, CE simulations suggest that the ejecta
can form a broad torus which fits to our observations
(see appendix). When the compact companion reaches
the center of the second star, angular momentum forms
a disk around the compact remnant, allowing the forma-
tion of a GRB-like jet. The remnant of this merger might
be a magnetar whose prolonged activity is responsible for
the very long duration of the actual GRB.
The jet and the subsequent breakout of the supernova
can explain the power-law and thermal component in the
X-rays of GRB 101225A. The shock-breakout X-rays ir-
radiate the common envelope ejecta at a few times 1014
cm, where they get thermalized, explaining the UVOIR
emission in the first 10d of the explosion. As the super-
nova shock expands beyond the CE, it provides a final
burst of emission, explaining the bump at ∼30d in the
light curve. The He-NS merger scenario naturally as-
sumes a relatively small Ni-production, leading to a weak
5. COMPARISON WITH OTHER EVENTS
A similar scenario might also explain XRF 060218 with
a dense shell of material having been ejected prior to the
main explosion, probably with a different progenitor sys-
tem producing a brighter SN and a fainter GRB com-
ponent (see appendix). On the other hand, a group of
GRBs exists which feature an initial thermal component
in X-rays (e.g., GRB 090618) due to the shock break-
out, but which are accompanied by a classical, bright
afterglow produced by the interaction of the jet with
a moderately dense circumstellar medium (Cano et al.
2011). We therefore suggest that GRB 101225A was a
member of a new class of “blackbody-dominated” long-
duration GRBs connected to SNe which are arising in a
very dense environment created by the massive-star pro-
genitor or progenitor system itself, fully thermalizing the
high-energy output from the collapsing star. The non-
relativistic and isotropically distributed emission of such
a component makes it difficult to detect such an event at
higher redshifts making GRB 101225A a fortunate coinci-
dence to derive conclusions about the progenitor system
and its environment.
We thank J. S. Bloom for helping with the Keck
observations and the staff of the various observatories
for excellent support. The first author acknowledges a
personal connection to the event 31 years prior to the
burst.The Dark Cosmology Centre is funded by the
DNRF. KLP acknowledges the support of the UK Space
Agency. JG, AJCT, SG and PK are partially supported
by the Spanish programs AYA-2007-63677, AYA-2008-
03467/ESP, and AYA-2009-14000-C03-01. HTJ acknowl-
edges support by DFG grants EXC153, SFB/TR27,
and SFB/TR7. MAA is partly funded through grants
AYA2007-67626-C03-01 and Prometeo-2009-103. PM ac-
knowledges the support from MICINN through grants
AYA2007-67626-C03-01 and CSD2007-00050. SRO and
MDP thank STFC for financial support.
CC, JL, CJ, and SP acknowledge support from the
CRI grant 2009-0063616, funded by the Korean gov-
ernment (MEST). AM acknowledges support from MK-
A. BAT DATA ANALYSIS AND FITTING
GRB101225A was detected by BAT onboard the Swift satellite (Gehrels et al. 2004) on Dec. 25, 2010 at T0 =
18:37:45 UT as an image trigger (Racusin et al. 2010). It was already in progress both when the source entered the
BAT field of view and when it left the field of view due to Swift orbit-constrained slews (Palmer et al. 2010; Cummings
et al. 2010). Therefore, we can only give lower limits on the total burst fluence and the T90duration. The total fluence
of the intervals covered in the observations adds up to (5.6 ± 0.7)× 10−6ergcm−2(implying a total energy release in
γ-rays of Eγ,iso> 1.4 × 1051erg), which is a lower limit to the total gamma-ray emission. No emission was detected
in a previous observation of the field at T − T0= −4950 s. The lower limit on the duration is T90> 2000 s. This
is one of the highest durations ever observed for a Swift GRB, comparable to the longest burst observed by Swift,
GRB090417B (Holland et al. 2010). Fig. 4 shows the BAT light curve. The BAT-observed peak flux of (3.25±0.47) ×
10−9erg cm−2s−1in the 15−150 keV range occurred in the interval T − T0= +1372 to T − T0= +1672 s. No other
γ-ray instrument detected GRB101225A, although the MAXI instrument on board the ISS (2 − 10 keV) reported a
marginal detection at T − T0= +1002 s coincident with the BAT position (Serino et al. 2010).
Fig. 4.— Flux in the 15−150 keV band using a fixed power-law index of Γ = 1.867 from a fit to the most intense part of the burst. Note
that the burst started before the beginning of the BAT data at ∼ T − T0 = −100s and probably continued while the source was not in
the BAT field of view from T − T0= +1091 to T − T0= +1372s. The latest upper limit before the burst was 2.65 × 10−9ergs cm−2s−1
at T − T0= −4950 s. Error bars are at 90% confidence.
The time-averaged spectra from T0to T − T0= +963 s and from T − T0= +1372 to T − T0= +1672 s are best
fit by simple power-law models with photon indices of Γ = 1.91 ± 0.35 and 1.87 ± 0.21, respectively. For these fits the
total fluences in each time period in the 15−150 keV band are (1.7±0.4)×10−6and (9.0±0.2) ×10−7erg cm−2. All
the quoted errors are at the 90% confidence level. The BAT spectra are almost equally well parametrised by models
using a cutoff power-law or a blackbody to fit the data due to the low signal-to-noise ratio of the event. Epeakusing a
cutoff model is poorly constrained to 38 ± 20 keV. The blackbody temperature fit gives kT = 10.1 ± 1.1 keV. Errors
are at the 68% confidence level.
We also examined the BAT data to search for persistent emission after the trigger. For this we used the daily sky
image mosaics produced as part of the BAT hard X-ray transient monitor which cover a single energy band of 15−50
keV. We found a 5.3σ excess (0.0048±0.0009 count cm−2s−1) on Dec. 25, 2010 (MJD 55555), the day of the trigger,
and a positive excess in the count rate (≥ 1σ or 0.0011 count cm−2s−1) over the next ten days (until MJD 55565).
We determine the probability that such a sequence of excess rates would occur by chance. To do this, we examine the
light curves of 106 “blank sky” points tracked in the BAT transient monitor. These are points chosen randomly across
the sky at least 10 arcmin from any known X-ray source. Any positive flux from these points is expected to be due to
chance fluctuations. In these 106 light curves (> 200,000 data points), we find only one sequence of six consecutive
days showing a positive excess and none with more than six days. This means that the chance probability of ten days
of excess flux is less than 1/200,000, so the observed prolonged emission is likely real.
6C. C. Th¨ one et al.
B. XRT DATA ANALYSIS AND FITTING
The Swift-XRT data were processed with version 3.7 of the XRT data reduction software (released as part of HEASoft
6.10 on 2010-09-28) and the corresponding calibration files used for subsequent spectral analysis. Data were collected
in Windowed Timing (WT) mode for the first 7.3ks after the trigger followed by Photon Counting (PC) mode for
the rest of the observations. The object was detected by XRT from 1.4ks to 105s after the trigger. The peak flux in
X-rays is 4.3 × 10−9ergcm−2s−1, the total observed fluence 8.2 × 10−6ergcm−2, the unabsorbed fluence 1.1 × 10−5
erg cm−2. At z = 0.32 this corresponds to a total energy release in X-rays of 3.6 × 1051erg. Spectra were extracted
for individual snapshots of data (one snapshot corresponds roughly to one orbit constrained by the observability of
the object during the orbit) and were further timesliced into 100s bins for the initial snapshot (1.4 − 1.8ks after the
? F? (erg cm?2 s?1)
Fig. 5.— Fitting of the X-ray spectra from XRT in the first snapshot (top panel). The dashed line indicates the contribution of the
power-law component, the dotted line shows the BB component. In the bottom panel we show the ratio between the observed data and
the fitted model.
We tried a variety of fits to the X-ray data, using XSPEC version 12.6.0, with the result that an absorbed power-
law plus blackbody component provided a good fit to the data (see Fig. 5). The T¨ ubingen-Boulder absorption model
was used, with the Wilms abundances (Wilms et al. 2000) and Verner absorption cross-section (Verner et al. 1996).
As shown in Fig. 6, there is little spectral evolution within the first snapshot of data, with the best fit for the full 367s
of data being a power-law of photon index Γ = 1.83+0.13
and a total absorbing column of (2.2 ± 0.3)×1021cm−2, for a χ2of 420.7 for 379 degrees of freedom. The Galactic
column density in this direction is 7.9×1020cm−2. The inclusion of the BB is significant at the > 99.9999% level, the
contribution of the BB to the total emission is around 20% (see Fig. 6). The second snapshot of data (also in WT
mode) is again better fit with a BB in addition to the power-law, with Γ = 2.18+0.12
and NH = (2.7 ± 0.2) × 1021cm−2, with χ2/dof = 378/421. This BB is significant at 99.987%. For the X-ray data
after the second snapshot, no BB component is required and a simple absorbed power-law provides an acceptable fit,
likely due to the lower signal at later times.
We also checked for possible periodicity in the X-ray data. To that end, light curves were extracted with 18ms
bins which is the best time resolution available for WT mode. Using the Kronos powspec tool, no significant periodic
signal was identified with a frequency between 0.005 and 28 Hz (0.04 and 200 s) in either the first or second snapshot
−0.10, a blackbody of temperature 0.96 ± 0.13 keV (1.11 × 107K)
−0.09, black-body kT = 0.99+0.15
C. UV, OPTICAL AND IR DATA ANALYSIS
Swift/UVOT began observing GRB101225A 1373 s after the BAT trigger, simultaneous with the XRT observations.
The automatic target sequence did not commence until the end of the BAT image trigger at ∼ 23 minutes. The source
was found to be blue, with strong detections in the UV filters (uvw1, uvm2, uvw2), weak detections in the b and u
BB kT (keV)
Obs. BB %
R @ z=0.3 (cm)
1050 1100 1150120012501300 1350
NH (1021 cm?2)
rest frame time since BAT trigger (s)
Fig. 6.— Results from the fits to the first snapshot of the X-ray data. The panels show from top to bottom: 1. The count rate during
the first snapshot, 2. the photon index Γ, 3. the BB Temperature in keV, 4. the contribution of the BB to the total emission in percent, 5.
the radius of the emitting BB at z = 0.3 and 6. total absorbing column density in X-rays (the Galactic column density in the line-of-sight
is 7.9 × 1020cm−2).
filters, and no detection in the v filter. The data were processed using the standard Swift software tool uvotmaghist
within HEAsoft 6.9 and the latest calibration files (20101231).
We extracted counts using a circular aperture with a radius of 5 arcsec where the count rate was above 0.5 countss−1,
and 3 arcsec aperture where the count rate had dropped below 0.5 countss−1, and a source-free background region.
The tool uvotmaghist applies coincidence-loss corrections and aperture corrections. The count rates were converted
to flux density using the standard photometric calibration (Poole et al. 2008; Breeveld et al. 2010).
C.2. McD 2.1m
The CQUEAN instrument (Camera for QUasars in the EArly uNiverse; Park et al. 2011, in preparation) on the 2.1m
Otto-Struve telescope at McDonald Observatory, Texas, USA, observed the optical counterpart starting at 01:16:23
UT, on Dec. 26, 2010, or 6.64h after the burst. Three exposures of 300s were taken in r?, i?, z?, and Y bands each
under photometric conditions. The data were reduced with standard procedures of dark and flat-field corrections. The
afterglow is detected in the r?, i?, and z?-band images, the Y band only give an upper limit.
C.3. CAHA 1.23m
The 1.23m telescope is located at the German-Spanish observatory of Calar Alto (CAHA) in Almer´ ıa, Spain and is
equipped with an optical imaging camera. The optical counterpart was detected in the V RI bands 1.04−1.11d after
the GRB trigger. The 1.23m was also used to calibrate the object field in BV RI bands by observing the Landolt fields
RU149D and SA98 on Dec. 26 and 27, 2010 under photometric conditions.
C.4. LT 2.0m
The Liverpool telescope is a 2.0m fully robotic telescope located at the observatory of Roque de los Muchachos on
La Palma. Observations were carried out with the imaging camera RATCAM. The optical counterpart was detected
in a single i?-band epoch at in a single epoch at 10.09d after the GRB.
8C. C. Th¨ one et al.
We acquired imaging data using OSIRIS at the Gran Telescopio de Canarias (GTC), a 10.4m telescope located at
the observatory of Roque de los Muchachos on La Palma, Canary Islands, Spain.
The observations started in r?band ∼2 days after the burst, exposing for 30 s. A second r?observation was carried
out ∼ 21 days after the gamma-ray event based on 5 exposures of 180 s. We furthermore obtained a late-time SED at
39 days in g?, r?, i?and z?bands and a last image at ∼44 days in the r?band.
The data of our two last epochs (at ∼39 and ∼44 days) were obtained at a considerable airmass (2.14>1.73) since
the object was setting quickly after evening twilight. The data at ∼21 and ∼44 days were acquired with the proximity
of the Moon at ∼54 and ∼36 degrees, when the illumination was 83% and 21%, respectively. The SED at 39 days was
constructed in dark time. The observing conditions were good in our four GTC epochs.
C.6. 3.6. Gemini-North: NIRI and GMOS-N
Late-time imaging of the optical counterpart of GRB101225A was conducted with the Gemini-North observatory
on Mauna Kea/Big Island, Hawaii, on several occasions. On the night of Jan. 23, 2011 we observed the field with the
Near InfraRed Imager (NIRI) in the K?filter for 44×60 s exposures (2×30 s co-adds) before switching to the Gemini
Multi-Object Spectrograph (GMOS-N) for 5 × 180 s exposures each in the r?filter. On the night of Feb. 01, 2011 we
re-observed the field with NIRI in the J band for 32 × 60 s exposures (1 co-add), and finally on the night of Feb. 03,
2011 we imaged the field in all four GMOS broad-band filters (g?r?i?z?). Since the source was setting, all exposures
were taken at moderate to high airmass (1.5 − 2.5), although under relatively good seeing conditions.
C.7. BTA 6m
A final late image was obtained using SCORPIO on the 6.0m BTA telescope, located at the Special Astrophysical
Observatory, in Russia. The observation consisted of 20 × 120 s exposures using an I filter obtained on Feb. 25, 2011
under good weather conditions and a seeing of 1.3 − 2.0 arcsec.
C.8. Photometry of ground-based data
The photometry of V, R and I data was done in a consistent way using a set of 15 comparison stars in the field of
GRB101225A calibrated with the Landolt fields taken on Dec. 26 and 27, 2011 by the 1.23m CAHA. For r?, i?and z?,
photometric calibration was done with observations of the 2.1m Otto-Struve telescope at McDonald Observatory on
Dec. 26, 2010, using standard star data (Feige 34). Finally, g?photometry was derived from the rest of the reference
magnitudes using numerical transformations (Jester et al. 2005). The magnitudes of the comparison stars in the
different filters used for the optical observations are listed in Tab. 1.
We performed aperture photometry using PHOT within IRAF assuming an aperture radius equal to the Full Width
at Half Maximum (FWHM) of the stellar point sources. In a few cases, where the contamination by neighboring
sources was not negligible, PSF photometry within IRAF was carried out. In Tab. ?? we list the final photometry for
all UV, optical and IR data.
Fig. 7.— Secondary standards used for the photometric calibration. The position of the optical counterpart of GRB101225A is indicated
with a red arrow. The photometric magnitudes of each of the reference stars are given in Table 1. The field of view is 6?× 4?.
Magnitudes of calibration stars used for the optical photometry. A finding chart indicating the position of each
reference star is given in Fig. 7. V, R and I are given in Vega system, while the rest are in AB.
15.282 ± 0.009
18.065 ± 0.097
16.794 ± 0.050
18.789 ± 0.079
14.995 ± 0.008
15.174 ± 0.015
19.165 ± 0.180
16.594 ± 0.012
18.796 ± 0.059
18.682 ± 0.132
18.505 ± 0.218
19.253 ± 0.130
18.050 ± 0.066
16.828 ± 0.015
17.516 ± 0.032
14.916 ± 0.049
17.268 ± 0.055
16.421 ± 0.051
18.335 ± 0.133
14.633 ± 0.049
14.838 ± 0.049
18.268 ± 0.090
16.161 ± 0.049
17.924 ± 0.069
17.901 ± 0.107
17.877 ± 0.060
18.893 ± 0.131
17.749 ± 0.085
16.321 ± 0.052
17.024 ± 0.066
14.553 ± 0.052
16.288 ± 0.110
15.972 ± 0.062
18.051 ± 0.162
14.235 ± 0.051
14.456 ± 0.051
17.356 ± 0.188
15.691 ± 0.090
16.694 ± 0.092
17.059 ± 0.182
17.270 ± 0.163
18.460 ± 0.273
17.318 ± 0.108
15.826 ± 0.087
16.321 ± 0.082
Preimaging of the field was obtained from the archive of the 3.5m Canada-France-Hawaii Telescope (CFHT). Ob-
servations were obtained with the MegaPrime/MegaCam for the Pan-Andromeda Archaeological Survey (PAndAS,
Richardson et al. 2011). We combined 3×500 s exposures obtained under very good conditions in g?and i?bands. We
derive 3σ limiting magnitudes for these exposures of i?> 25.5 and g?> 26.9. However, at the position of GRB101225A,
we detect a low-significance object in the g?band, for which we measure r?= 27.2 ± 0.5, which could be the host
galaxy of our source (see Fig. 8).
Fig. 8.— Pre-imaging exposure in g?band obtained with the 3.5m CFHT. The field of view is 60??× 40??, North is to the top and East
to the left of the image. The blue circle indicates the position of the optical counterpart of GRB101225A, where we see a low significance
detection of what could be the host galaxy.
D. OPTICAL SPECTROSCOPY
We obtained a spectrum of the optical counterpart 51 h after the event using OSIRIS on the 10.4m GTC telescope
on La Palma (Spain). Two spectra of 1800 s exposure time each were taken with grism 300B (R=325, wavelength
range: 3500 − 7000˚ A) under moderate to high airmass (1.26 and 2.05, respectively). The spectra were reduced and
combined with standard tasks in IRAF and flux-calibrated with the spectrophotometric standard G191-2B2 taken
the same night.The continuum is clearly detected, but the spectrum shows no obvious absorption or emission
lines. The limits on the detection of Hα [OIII] and [OII] emission from the host galaxy are < 5×10−18ergcm−2s−1,
< 2.3×10−18ergcm−2s−1and < 3×10−18ergcm−2s−1(3σ) respectively. We can also put a limit on the detection of
Hα at z=0 of < 2×10−18ergcm−2s−1. The flux-calibrated spectrum with the position of typical emission lines from
the host shifted to a redshift of z = 0.32 is show in Fig. 9.
On the night of 2011 Feb. 04 we observed the optical counterpart with the Low-Resolution Imaging Spectrometer
(Oke et al. 1995) on the Keck I telescope during local twilight. Two undithered observations of 600 seconds each
were acquired using the 1.0 arcsec slit and the D500 dichroic at a position angle of 86.5 degrees. On the red arm we
10 C. C. Th¨ one et al.
Fig. 9.— Flux-calibrated spectrum obtained with the GTC 2.1 days after the GRB. The error spectrum is plotted in blue. The red lines
indicate the position of normally strong emission lines from the interstellar medium at a redshift of z = 0.32, none of the lines are detected
in our spectrum.
used the 600/7500 grating and binned the CCD along the spatial direction (2x1 binning); on the blue arm we used
the 600/4000 grism and binned the data along both spatial and spectral axes (2x2 binning). Due to twilight there is
no evidence of a trace in the blue spectrum (and no source is detected in GMOS g?-band imaging from the previous
night). A faint continuum trace at the expected position of the transient is identified on the red side in the second
(less twilight-affected) exposure from 7160 to 8000˚ A with no absorption or emission features visible.
E. MODELING THE UV TO NIR SPECTRAL-ENERGY DISTRIBUTION
Already the early evolution of the UV/optical/IR (UVOIR) counterpart proved to be very unusual for a GRB
afterglow. Instead of a power-law spectrum with a negative spectral slope, it had a very blue counterpart, following
what seemed to be a power-law with a positive spectral slope (Cenko et al. 2010). Furthermore, the counterpart stayed
bright during the first days and then a decayed (Xu et al. 2010a) with a strong color change, transforming into a very
red counterpart two weeks after the trigger (Tanvir et al. 2011). We interpret this early evolution as being produced
by the expansion and cooling of a blackbody (BB), as shown in Section E.1.
The simple BB evolution is not valid any more for the emission beyong ∼ 20 days after the trigger. At that time we
observe a flattening of the light curve, while the very red color is preserved. This late evolution can be well-described
with the presence of a supernova component, as described in Section E.2.
E.1. Early time evolution
For the modeling of the UVOIR spectral energy distribution (SED), we use the photometry presented here together
with some of the data points extracted from the literature (Xu et al. 2010a,b; Wiersema et al. 2010; Cenko et al. 2010;
Xu et al. 2011; Fynbo et al. 2011; Tanvir et al. 2011), all of which we correct of a Galactic extinction of AV = 0.33mag
and transform from magnitudes to flux densities. The data allow us to derive a set of 12 SEDs ranging from 0.07 to
40days after the trigger.
The early optical SEDs are well fitted by using an expanding and cooling blackbody of the following form (in
Here 1026is used to convert W/m2/Hz to Jy. R is the radius of the emitting black body (which we assume to be
spherical), D is the luminosity distance to the object, z the redshift and Tobsis the observed blackbody temperature
(the rest-frame temperature would be Trest= Tobs(1 + z)). The others physical constants: c is the speed of light, h
Planck’s constant and kBBoltzmann’s constant. For simplicity we assume a blackbody with an emissivity of 1.
The blackbody succeeds in reproducing the data up to 10 days, without any intrinsic extinction or additional emission
component, after which another component becomes dominant. We find that this second component can be reproduced
with the spectral templates of SN 1998bw placed at a redshift of z = 0.32+0.06
luminosity of 0.1 of SN 1998bw (see Section E.2).
From the fits to the SED evolution and allowing a second-degree fit, we get the following evolution of the
Fν(Jy) = 1026
?22πν3(1 + z)
−0.07with a stretching factor of 1.25 and a
log1026π(1 + z)
= (0.70 ± 0.04) + (0.46 ± 0.03)log(t) − (0.01 ± 0.05)log10(t)2
where t is the time in days. The temperature evolution (in K) can be described by:
log (Tobs) = (4.342 ± 0.017) − (0.395 ± 0.016)log(t) − (0.11 ± 0.02)log(t)2
Figures 10 and 11 show the temporal evolution of the normalisation constant and the temperature. The normalisation
can be sufficiently described by a linear evolution in log-log space and therefore the second order term in eq. (2) can
be neglected. For the temperature, we need an additional second order term to obtain an reasonable fit to the data.
Measured values for the blackbody evolution. Values in brackets are not be fitted due to a limited amount of data
points in those SEDs.
EpochObserved temperatureNormalisation constant
(1 + z)π1026?
(43 000±8 000
40 000±6 000
35 000±3 950
25 340±5 440
20 900±1 770
15 000±1 090
14 260±1 760
(11 300±2 000
(6 000±2 000
(5 000±1 000
The temperature evolution fits very well to a theroretical model of an expanding and homogeneously radiating gas
sphere (see Fig. 20. Once the redshift is known the normalisation constant can be transformed into physical values in
the rest-frame of the object which is shown in the Fig. 3 in the main paper. For this we assume z = 0.32 (see Section
E.2), or 1661.1 Mpc using a ΛCDM cosmology with H0= 71, ΩM= 0.27 and ΩΛ= 0.73.
Fig. 10.— Evolution of the normalization constant.
Fig. 11.— Evolution of the observed temperature requiring a second order term.
E.2. Late evolution and SN template fitting
The late evolution of the light curve requires a component in addition to the evolving BB described in Section E.1
which is best described by a combination of a supernova (SN) and a BB. In the following, we present the late SED
12C. C. Th¨ one et al.
Fig. 12.— Evolution of the velocity of the black body according to the result of our modeling in the UVOIR range. During our observing
period velocities of the range 0.3 − 0.07c are measured, typical of accelerated material during a supernova explosion.
fitting using several SN templates. We then take these fits to estimate the redshift of GRB101225A, which we could
not obtain spectroscopically (Section D).
To determine the redshift of GRB101225A we use the SED at 40 days after the burst where the contamination
from the BB is negligible and where we have 5 detections in different bands. Given the steep slope in the blue part
of the spectrum, we convolve the response of each filter with the spectral shape of the templates. This is particularly
important for the r?-band observations performed from GTC and Gemini at a very similar epoch, which show a
significant difference in flux density. The filter of GTC reaches slightly redder wavelengths, and the difference in flux
densities can be well explained by a very steep slope due to a SN feature as shown in Fig. 13.
We obtain templates for different core-collapse supernovae from the literature25. We exclude SN Ia from the analysis,
as we do not expect a high-energy emission or BB evolution for those events. The template for each SN prototype was
interpolated to the time of the SED for a range of redshifts (see references in Table 3). In the particular case of SN
1988S the templates were created by combining ground-based and HST spectra. For each SN template we evaluate
the best fit for a range of redshifts, allowing a scaling of the flux of the supernova. Table 3 displays the results of the
Fits of the SED with SN templates
best fit z
(Stritzinger et al. 2002)
(Richmond et al. 1996)
(Kulkarni et al. 1998)
(Anupama et al. 2001; Lentz et al. 2001; Fassia et al. 2001)
(Gaskell et al. 1992)
(Leonard et al. 2002)
The absolute best fit is obtained with a SN 1998bw template, a broad-lined Type Ic that is the classical reference for
GRB-related supernovae. For this case we obtain a redshift of z = 0.32+0.06
(excluding the Type II SN 1998S, which clearly does not fit our SED) give redshifts between z = 0.24 and z = 0.40,
with an average value of z = 0.32, confirming our result using SN 1998bw. We therfore use z=0.32 as reference in this
work. Figure 13 shows the fitted SED with the template at z = 0.32+0.06
−0.07. The other core-collapse SNe we tested
E.3. Luminosity and stretching factor of the SN associated with GRB101225A in context of other GRB-SNe
We can undertake a more general comparison to SNe associated with GRBs by following the formalism of (Zeh et
al. 2004). These authors used a SN 1998bw template light curve to fit late bumps in GRB afterglow light curves,
modifying the template by increasing or decreasing the luminosity at peak (the parameter k, with k = 1 implying a
peak luminosity identical to that of SN 1998bw), and stretching or compressing the light curve in time while retaining
the overall shape (the parameter s, again, s = 1 implies the temporal evolution is identical to that of SN 1998bw in
the same band). This procedure also included the creation of synthetic templates by interpolating between the SN
Fig. 13.— Fit of the day 40 SED to a SN 1998bw template. The observations are in black and the best fit, with a redshift of z = 0.32,
in red. The gray lines represent the template at the different redshifts (steps of 0.01) within errors.
1998bw light curve in different filters, and taking into account the cosmological K-correction. Nearly all GRB-SNe
were well-fit by the SN 1998bw light curve template. Using a supernova light-curve template of SN 1998bw redshifted
to z = 0.32 to fit the late light curve of the GRB101225A we find s = 1.25 ± 0.15 and k = 0.08 ± 0.03 using the
designation of (Zeh et al. 2004).
Ferrero et al. (Ferrero et al. 2006) analyzed SN 2006aj associated with XRF 060218, and placed it into the k − s
context. They employed the line-of-sight extinction values derived by (Kann et al. 2006) to derive intrinsic k values.
To place the SN associated with GRB101225A into the k − s context, we fit the light curve analogous to (Zeh et al.
2004), and use the sample of (Ferrero et al. 2006) as well as additional events as a comparison. The complete data is
presented in Table 4.
GRB 990712 has been analysed again with additional data. We find no evidence for host extinction. For GRB
021211, a re-analysis of the afterglow SED finds no evidence for host extinction, the value from (Ferrero et al. 2006)
thus remains unchanged but now counts as extinction-corrected. For GRB 040924, we use the k and s values from
(Wiersema et al. 2008) and correct k with the extinction found by (Kann et al. 2006). For GRB 050525A, we use the
uncorrected k value from (Ferrero et al. 2006), and correct it with the extinction found by (Kann et al. 2010). For
XRF 050824, we use the uncorrected k value from (Sollerman et al. 2007), and correct it with the extinction found by
(Kann et al. 2010). GRB 060729 is analysed in (Kann et al. 2011). GRB 080319B is analysed in (Bloom et al. 2009).
GRB 090618 has been analysed for this work, using the data set of (Cano et al. 2011). We were not able to derive a
good SED for this afterglow, therefore the k value has not been corrected.
As can be seen in Fig. 14, the SN associated with GRB101225A is significantly fainter than any other known GRB
SN (with the SN associated with GRB 040924 being the most similar, but this event is only marginally detected). At
the same time it is similar in temporal evolution, actually being slower than most known GRB-SNe, though not by a
large amount. It is also fainter than two well-studied Type Ic SNe, SN 1994I and SN 2002ap (see (Ferrero et al. 2006)
for discussion), the latter being broad-lined, but not associated with a GRB.
F. REDSHIFT DETERMINATION
Determining the distance scale at which GRB101225A occurred is crucial to understand the energetics and get a
clear picture of the physics involved in this event. The following independent argument are used to strengthen our
redshift estimation. The first strong limit on the redshift comes from the UVOT detection in uvw2 which implies a
redshift lower than z = 1.4 (Campana et al. 2010).
From the SED fit of the first days, we know that the evolution is well-described by a simple BB. Depending on the
distance at which the object is found, we can derive different radii and expansion velocities. For an explosion of this
type, we expect expansion velocities larger than ∼ 103km s−1, which would be barely equivalent to a stellar wind
and, if similar to a SN explosion, of the order of 104km s−1. We cannot, in principle, rule out higher velocities in
the ejecta. However, higher velocities would have further implications. As the velocity of the ejecta is accelerated to
velocities close to c, we would expect to see a broadening in the shape of the blackbody SED due to the fact that the
equal-arrival-time surface would be covering regions of different temperatures.
Making the assumption that the ejecta should not be traveling at velocities larger than 150,000 km s−1(0.5c) and
lower than 1,000 km s−1(0.03c), we can estimate an acceptable redshift range between z = 0.20 and z = 0.60. At
a redshift of z = 0.32, the velocity of the blackbody would have evolved from ∼ 90,000 km s−1at the time of our
14C. C. Th¨ one et al.
0.4 0.60.81.01.2 1.41.6
luminosity factor k
stretch factor s
Fig. 14.— Luminosity factor k and stretch factor s of SNe associated with GRBs. Filled symbols have been corrected for host-galaxy line-
of-sight correction, non-filled symbols have not. We label several well-studied nearby GRB-SNe. We label several well-studied nearby GRB-
SNe, as well as two “canonical” Type Ic SNe, SN 1994I (Ic) and SN 2002ap (broad-lines Ic unassociated with a GRB). The GRB101225A
SN is fainter than all these events.
Luminosity Factor k and Stretch Factor s for GRB SNe
Ferrero et al. (2006)
Ferrero et al. (2006)
Ferrero et al. (2006)
Ferrero et al. (2006)
Ferrero et al. (2006)
Ferrero et al. (2006)
Ferrero et al. (2006), This Work
Ferrero et al. (2006)
Ferrero et al. (2006)
Wiersema et al. (2008), Kann et al. (2006)
Ferrero et al. (2006)
Ferrero et al. (2006), Kann et al. (2010)
Sollerman et al. (2007), Kann et al. (2010)
Ferrero et al. (2006)
Kann et al. (2011)
Bloom et al. (2009)
This Work, Cano et al. (2011)
0.41 ± 0.29
0.60 ± 0.10
0.62 ± 0.09
0.43 ± 0.21
0.74 ± 0.01
1.04 ± 0.03
0.58 ± 0.05
0.08 ± 0.03
1.45 ± 0.93
0.76 ± 0.07
1.40 ± 0.32
0.69 ± 0.25
0.80 ± 0.02
0.97 ± 0.07
0.92 ± 0.08
0.99 ± 0.25
0.85 ± 0.10
1.09 ± 0.07
1.37 ± 0.97
1.38 ± 0.06
0.77 ± 0.04
0.52 ± 0.14
0.69 ± 0.01
0.86 ± 0.02
0.89 ± 0.10
0.99 ± 0.05
1.25 ± 0.15
first SED to ∼ 2,000 km s−1at the time of our last blackbody-dominated epoch, nicely matching the requirements.
The fact that the blackbody detected for XRF 060218/SN 2006aj was very similar to the one found for GRB101225A
when placed at a redshift of z = 0.32 (see Section G) adds additional evidence for the validity of the redshift estimate.
In conclusion, the assumption that GRB101225A is located at a redshift of z = 0.32+0.06
factors, and can be considered as a firm reference when studying the physical processes involved in the event. In any
case we do not expect redshifts significantly smaller than z = 0.2 or higher than z = 0.4.
−0.07is supported by independent
G. COMPARISON BETWEEN GRB101225A AND OTHER GRBS WITH SNE AND BB COMPONENTS
In Table 5, we compare some properties of GRBs which were associated to SNe and did not show a “classical”
afterglow component. All of them are subluminous compared to the average long-duration GRB with Eiso around
1051–1054erg. GRB 101225A, however, lies on the lower end of normal long-duration GRBs. Among those nearby
GRB-SNe without a classical afterglow, there seems to be a class of very long duration GRBs with very low Epeakvalues,
all of them showing a thermal component in X-rays. GRB 060218 and XRO 080109 also had a thermal component at
optical wavelengths during the first few days (Campana et al. 2006; Soderberg et al. 2008). For 100316D, no optical
counterpart was detected before the onset of the actual SN due to high intrinsic extinction in the host galaxy.
XRF 060218 (Campana et al. 2006) shows a similar early behavior to GRB101225A. We compare the early UVOT
lightcurve of XRF 060218, which we obtained from the UVOT catalogue (Roming et al. 2009), to the lightcurve from
GRB 101225A by shifting the XRF 060218 to z=0.3 including a k-correction. To obtain the k-correction, we use
GRB 101225A 15
GRBs with SNe but without afterglows
> 1.4 × 1051
*No γ-rays observed, numbers derived from X-rays.
TC? refers to the early thermal component, mostly attributed to a supernova breakout.
HR is the hardness ratio, defined as the ratio of channels (50 − 100 kev)/(25 − 50 keV)
SN MV is the SN peak absolute magnitude in V
Host MBis the host absolute magnitude in B
XSPEC assuming a blackbody spectrum with kT ∼ 3.7 eV. The temperature was determined from the best-fit model of
a BB to a SED of XRF 060218 taken at 120ks after the trigger (Campana et al. 2006). Using this BB spectrum, we
determined the expected flux density in the observed frame for each filter and at z = 0.32. The ratio of these two flux
densities was taken to be the k-correction for the specific filter, which is ∼ 2.20 for all filters. Figure 15 compares the
flux density light curves of the 3 UV filters of both GRBs. For both GRBs the light curves were corrected for Galactic
Flux Density (erg s?1 cm?2 Hz ?1)
Observed Time (seconds)
Fig. 15.— The 3 UV fliter light curves in flux density for GRB 101225A and XRF 060218. The light curves for XRF 060218 have been
shifted to z=0.3 for direct comparison. The colored solid shapes connected by solid lines are for GRB 101225A, open symbols and dotted
lines for XRF 060218. Circles are uvw1, squares are for uvm2, triangles are uvw2.
To complete this, we performed a fit of the early UVOIR SEDs of XRF060218 and XRO 080109, in a similar way to
what we did for GRB101225A. XRF060218 also seems to follow a black body evolution at early times. However, the
SN starts to dominate already around 3 days after the burst, limiting the possible study of the evolution. As it can
be seen from Fig. 16 and 17, the evolution of the black body is not very different from what we see in GRB101225A
although the radius expansion is slightly steeper. XRO 080109 does not have a measurable thermal component in the
early X-ray data. In the optical, the cooling BB dominates until about 4 days after the event when the onset of the
SN was observed (for the SED fitting see Fig 18). The temperature of the UVOIR BB of XRO 080109 is lower (∼ 30
000 K) than those of XRF 060218 and GRB 101225A. The radius evolution is considerably steeper than for those two
events (Fig 19). The UVOIR BB emission of XRO 080109 is therefore likely due to the cooling of the initial shock
Antonio de Ugarte Postigo