TITLE:   GCN CIRCULAR
NUMBER:  14895
SUBJECT: GRB 130603B: Detection of possible afterglow/kilonova in HST observations
DATE:    13/06/14 04:35:26 GMT
FROM:    Edo Berger at Harvard  <eberger@cfa.harvard.edu>

E. Berger and W. Fong (Harvard) report:

"We obtained the public Hubble Space Telescope ACS/F606W and WFC3/F160W
images of GRB 130603B (Tanvir et al. GCN #14893) from MAST and performed an
astrometric tie of these images relative to our afterglow images from
Magellan/IMACS (Foley et al. GCN #14745).  The resulting total rms of the
astrometic fit is 33 mas.  At the location of the optical afterglow we
identify an apparent point source in the WFC3/F160W image, with no
corresponding counterpart in the ACS/F606W image (the circles marking the
afterglow position have a radius of 10-sigma):

https://www.cfa.harvard.edu/~eberger/GRB130603B_HST.tif

PSF-matched photometry indicates m(F160W)=25.8+/-0.2 AB mag and
m(F606W)>27.5 mag (3-sigma; the limit is consistent with Tanvir et al. GCN
#14893).  At the redshift of GRB 130603B (z=0.356) these values correspond
to absolute magnitudes of M(F160W)=-15.2 mag and M(F606W)>-13.5 mag.

The red V-H>1.7 mag color is potentially in good agreement with the
afterglow g/r/i colors at early time (8.4 hr), which indicate a spectral
index of beta~-1.7 (Cucchiara et al. arXiv:1306.2028).  Based on this
spectral index and the g/r/i magnitudes from Cucchiara et al., the
interpolated/extrapolated magnitudes in the HST filters at 8.4 hr are
m(F160W)=20.0 mag and m(F606W)=21.9 mag, or V-H~1.9 mag.  Therefore, it is
possible that the source detected in WFC3/F160W is the fading afterglow,
indicating a decline rate of alpha_NIR~-1.6 between 8.4 hr and 9.4 days.
Incidentally, this decline rate is in good agreement with the Swift/XRT
decline rate of alpha_X~-1.6 at about 1 hr to 1 day (
http://www.swift.ac.uk/xrt_live_cat/00557310/)

Alternatively, the red V-H color and the absolute magnitude of
M(F160W)=-15.2 mag can be explained as emission from an r-process powered
"kilonova", along the recent models by Barnes & Kasen arXiv:1303.5787.  For
their fiducial model (with M_ej=0.01 Msun and v_ej=0.1c), the expected
absolute magnitude in the rest-frame J-band (corresponding to observed
H-band) at an observed time of 9.4 days is about -15 mag, while the
expected magnitude in the rest-frame B-band (corresponding to observed
V-band) is exceedingly low (about -3 mag).  Thus, it is possible that the
red source we detected in the WFC3/F160W image represents the first
detection of an r-process powered transient associated with a short GRB,
thereby strengthening their association with NS-NS/NS-BH mergers.

As noted by Tanvir et al. (GCN #14893) additional observations to determine
variability are essential."