The Astrophysical Journal, 542:L89-L93, 2000 October 20
© 2000. The American Astronomical Society. All rights reserved. Printed in U.S.A.
Hubble Space Telescope Space Telescope Imaging Spectrograph Imaging of the Host Galaxy of GRB 980425/SN 1998bw1
J. U. Fynbo ,2 S. Holland ,3 M. I. Andersen ,4 B. Thomsen ,3 J. Hjorth ,2 G. Björnsson ,5 A. O. Jaunsen ,6 P. Natarajan ,7 and N. Tanvir 8
Received 2000 July 12; accepted 2000 August 30; published 2000 October 5
ABSTRACT
We present Hubble Space Telescope Space Telescope Imaging Spectrograph observations of ESO 184-G82, the host galaxy of the gamma-ray burst GRB 980425 associated with the peculiar Type Ic supernova SN 1998bw. ESO 184-G82 is found to be an actively star-forming SBc subluminous galaxy. We detect an object consistent with being a point source within the astrometric uncertainty of 0
018 of the position of the supernova. The object is located inside a star-forming region. The object is 
1 mag brighter than expected for the supernova based on a simple radioactive decay model. This implies either a significant flattening of the light curve or a contribution from an underlying star cluster.
1 Based on observations with the NASA/ESA Hubble Space Telescope, obtained at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555 and on observations made with ESO Telescopes at the Paranal Observatory under program 63.O-0065.
2 Astronomical Observatory, University of Copenhagen, Juliane Maries Vej 30, DK-2100 Copenhagen Ø, Denmark; jfynbo@ifa.au.dk.
3 Institute of Physics and Astronomy, University of Aarhus, DK-8000 Århus C, Denmark; holland@ifa.au.dk.
4 Division of Astronomy, University of Oulu, P.O. Box 3000, FIN-90014, University of Oulu, Finland; manderse@sun3.oulu.fi.
5 Science Institute, Dunhagi 3, University of Iceland, IS-107 Reykjavik, Iceland; gulli@raunvis.hi.is.
6 Institute of Theoretical Astrophysics, University of Oslo, Blindern, N-0315 Oslo, Norway; a.o.jaunsen@astro.uio.no.
7 Institute of Astronomy, Madingley Road, Cambridge CB3 0HA, England, UK; priya@ast.cam.ac.uk; and Department of Astronomy, Yale University, New Haven, CT 06520-8181.
8 Department of Physical Sciences, University of Hertfordshire, College Lane, Hatfield, Hertfordshire AL10 9AB, England, UK; nrt@ast.cam.ac.uk.
1. INTRODUCTION
There is evidence that at least some gamma-ray bursts (GRBs) are related to supernovae (SNe). GRB 980425, GRB 970514, and GRB 980919 are consistent with being in temporal as well as spatial coincidence with SN 1998bw, SN 1997cy, and SN 1999E, respectively (Galama et al. 1998; Germany et al. 2000; Turatto et al. 2000; Thorsett & Hogg 1999). The associations between GRB 970514 and SN 1997cy, and between GRB 980919 and SN 1999E, could be by chance. However, this is very unlikely for GRB 980425 since the temporal and spatial constraints are much tighter than for the two other possible GRB/SN associations (Galama et al. 1998). Furthermore, there is evidence for a contribution from SNe to the light curves of GRB 970228 and GRB 980326 (Dar 1999; Reichart 1999; Bloom et al. 1999; Castro-Tirado & Gorosabel 1999; Galama et al. 2000). However, less than 0.2% of SNe release detectable gamma-ray emission (Kippen et al. 1998; Woosley, Eastman, & Schmidt 1999), and SN 1997cy, SN 1998bw, and SN 1999E were extremely bright and peculiar SNe. This association between GRBs and peculiar SNe suggests that GRB 980425 (and possibly also GRB 970514 and GRB 980919) were members of a rare class of GRBs and may not be indicative of normal GRBs. However, the temporal and spatial properties of the gamma radiation from GRB 980425, except for the low total luminosity, were not unusual (Galama et al. 1998; Pian et al. 2000).
In this Letter we present Space Telescope Imaging Spectrograph (STIS) clear and long-pass images of the host galaxy of SN 1998bw as part of a large survey aimed at studying the morphology of GRB host galaxies and the locations of the GRBs within the host galaxies. SN 1998bw occurred in a spiral arm of the face-on SB galaxy ESO 184-G82 (Holmberg et al. 1977). The redshift of ESO 184-G82 is 0.0085 (Tinney et al. 1998). The aims of this Letter are to characterize the region in ESO 184-G82 where SN 1998bw occurred and to measure or constrain the brightness of SN 1998bw at a late epoch more than 2 yr after the event.
We use a Hubble parameter of H0 = 100 h-1 km s-1 Mpc-1 and assume 
m = 0.3 and 
= 0.7 throughout this Letter. For this cosmology, a redshift of z = 0.0085 corresponds to a luminosity distance of 25.65 h-1 Mpc and a scale of 122 h-1 proper parsecs per arcsecond.
2. OBSERVATIONS AND DATA REDUCTIONS
The Hubble Space Telescope (HST) was used to obtain STIS/CCD images of ESO 184-G82 between 21:47 and 23:42 UT on 2000 June 11 (Holland et al. 2000b). A log of the HST observations is given in Table 1. The total exposure times were 1240 s in the 50CCD (clear [CL]) aperture and 1185 s in the F28X50LP (long pass [LP]) aperture.9 The CCD gain was set to 1e- per analog-to-digital converter unit (ADU), and the read-out noise was 4.46 ADU pixel-1. The data were processed through the standard STIS pipeline and combined using the DITHER (v1.2) software (Fruchter & Hook 2000) as implemented in IRAF (v2.11.1)10/STSDAS (v2.0.2). We used "pixfrac" =0.6 and a final output scale of 00254 pixel-1. These observations were taken as part of the Cycle 9 program GO-8640 (Holland et al. 2000a).11
 | Table 1 The Journal of HST/STIS Observations |
9 The 50CCD aperture has a central wavelength of 5850 Å and a width of 4410 Å. The F28X50LP aperture has a central wavelength of 7230 Å and a width of 2720 Å (STIS Handbook v4.1).
10 Image Reduction and Analysis Facility (IRAF), a software system distributed by the National Optical Astronomy Observatories.
11 Survey of the Host Galaxies of Gamma-Ray Bursts. The URL for the survey is http://www.ifa.au.dk/~hst/grb_hosts/index.html.
3. RESULTS
The drizzled CL image is shown as Figure 1. ESO 184-G82 is a barred spiral galaxy of Hubble type SBc (P. Rautiainen 2000, private communication). The magnitude of the galaxy is B = 15.19 ± 0.05, R = 14.28 ± 0.05 (Lauberts & Valentijn 1989), which corresponds to LB = 0.02L
using M =-21.2 + 5 log h (Marinoni et al. 1999).
 | Fig. 1 Drizzled CL image. The total size of the image is 49 |
The galaxy is clearly in a stage of active star formation. Figure 1 shows that the optical appearance of the galaxy is dominated by a large number of high surface brightness star-forming regions, especially in the southern spiral arm where the SN occurred. When comparing the CL and LP images, a large number of red giants can be seen within or near the star-forming regions (two examples are seen in Fig. 2).
 | Fig. 2 Orientation of each panel is the same as in Fig. 1. (a) A region of 40 × 40 drizzled pixels (1 |
As also noted by Holmberg et al. (1977), ESO 184-G82 is member of a group of galaxies consisting of ESO 184-G80, ESO 184-G81, and ESO 184-G82. Furthermore, there is a previously unclassified Sb galaxy 1
1 south of ESO 184-G82 and a fainter bulge-dominated galaxy 17 south of ESO 184-G82. Hence, the star formation activity could be enhanced by interaction (see § 4).
3.1. Astrometry
The most precise determination of the absolute position of SN 1998bw on the sky is based on the radio observations made with the Australian Telescope Compact Array 1998 April 28 (Wieringa et al. 1998). This position is R.A. = 19h35m03
31, decl. = -52
50
447 (J2000.0) with an error of 01. However, SN 1998bw was very faint at the time of the STIS observations, and there are no other sources in the STIS field of view (FOV) for which positions have been determined with equally high precision. In order to determine relative astrometry between SN 1998bw and other objects in the STIS FOV, we therefore retrieved images obtained with the ESO Very Large Telescope (VLT) Antu telescope with the FORS1 instrument from the ESO archive. These images were taken at an epoch (1999 April 18) when the SN was still clearly visible. The journal of archive data is given in Table 2 (Leibundgut et al. 2000; Sollerman et al. 2000).
 | Table 2 Log of FORS1 Observations from the VLT Archive |
The precise location of the SN in the STIS CL image was derived from the ground-based VLT images in the following way. The STIS images revealed several bright blue objects within 05 of the position of the SN. These objects were covered under the point-spread function (PSF) in the VLT images, which had a seeing of about 09. This affected the centroid determination of the SN somewhat, giving rise to a color-dependent position error of 
1.5 drizzled STIS pixels (0038). By smoothing, scaling, and transforming the STIS CL and LP images, the contribution from the host galaxy, including the clusters in the proximity of the SN, could be subtracted completely from the VLT V and I images. Astrometry, based on these two subtracted images, gave accurate and internally consistent positions, based on six and 11 reference stars. The fits are six-parameter affine fits, with weights calculated from a noise model. The errors are the standard deviations of the residuals in the fit, properly normalized by weights and degrees of freedom. The final error of the weighted position was estimated to be about 0.7 drizzled pixels in the CL image, or about 0018. In the drizzled LP image the error is slightly smaller.
In order to check this position, DAOPHOT II (Stetson 1987) with extensions (DAOMATCH and DAOMASTER) was applied to the drizzled STIS LP image and the ground-based VLT R image, respectively. Subsequently, the two auxiliary programs DAOMATCH and DAOMASTER were used to derive a four-parameter transformation (offset, scale, and orientation) from the positions of five isolated stars judged to be free of superposed nebulosities. The transformed position of the SN is within 0.5 pixels of the value derived from our more formal astrometric procedure.
3.2. The Environment of the SN
As also noted by Leibundgut et al. (2000) and Sollerman et al. (2000), the SN is superposed on an H II region. The diameter of this region is approximately between 2
and 3, corresponding to a physical diameter of between 250 and 350 h-1 pc at z = 0.0085. Figure 2a shows a region with a size of 10 × 10 centered on the position of the SN as determined in § 3.1. At the exact position of the SN is the object marked by an arrow in Figure 2a. The object is also seen in the LP image shown in Figure 2c. In the following we will refer to this object as the SN. Also seen are six objects (in the following referred to as s1-s6) consistent with being point sources. In the CL image there appears to be an arclike structure extending from the SN toward s6.
3.3. Photometry and PSF Subtraction
We estimated the total AB magnitude of the SN, the other point sources (s1
s6) visible in Figure 2, and the arclike structure on both the CL and LP images in the following way. For the photometry of the point sources we use DAOPHOT II. We determined the PSF in the same way as described in § 4 of Holland & Hjorth (1999). We used ALLSTAR (Stetson & Harris 1988) with a fitting radius of 0071 and a sky annulus of 00760178 to fit a PSF to the object and determined aperture corrections by measuring the magnitudes of several stars in apertures with radii of 1108 (CL) and 0963 (LP). Tables 14.3 and 14.5 of the STIS Instrument Handbook show that these radii correspond to 100% of the encircled energy for a point source. We used the STIS zero points given in Gardner et al. (2000) to convert the count rates to CL and LP magnitudes. We transformed the CL and LP magnitudes to Johnson V and Kron-Cousins I magnitudes using the calibrations of Rejkuba et al. (2000). They used 33 stars with -0.8 < V-I < 2.0 and found that the rms residuals of their calibrations were 
V = 0.24 and I = 0.17.
As seen in Figures 2b and 2d, all seven objects, including the probable SN remnant in the error ellipse, are well subtracted by the PSF. The results of our PSF photometry are given in Table 3. The uncertainties in the V and I magnitudes are dominated by the uncertainties in the transformations from CL and LP. The values of 
X and Y are the coordinates on the drizzled STIS CL image of each object relative to the SN remnant. The colors of s2 and s4 are consistent with red giant stars, whereas the other four point sources have blue colors consistent with massive main-sequence stars.
 | Table 3 Results of PSF Photometry of the Objects Seen in Figure 2 |
We measured the surface brightness of the arclike structure in the following way. We subtracted the point sources in Table 3 and smoothed the resulting images with a Gaussian with an FWHM of 3 pixels. We then fitted by eye a line through the arc and measured the mean count rate along this line. The sky was estimated from empty regions of the CL and LP images. The results of this procedure are 
V = 21.9 ± 0.3 and I = 20.3 ± 0.3, yielding V-I = 1.6 ± 0.4. This structure might be the source of the H
emission seen in late spectra of SN 1998bw (Sollerman et al. 2000).
4. DISCUSSION
4.1. The Afterglow of GRB 980425/SN 1998bw
GRB 980425/SN 1998bw occurred in the southern spiral arm of the subluminous (LB = 0.02L), barred SBc galaxy ESO 184-G82. This galaxy is in a stage of strong star formation. In this way, ESO 184-G82 resembles the LMC, which is also subluminous, barred, and dominated by regions with strong star formation. We note that ESO 184-G82 is 0.3 h-2 as luminous as the LMC. For H0 = 5080 km s-1 Mpc-1 (de Bernardis et al. 2000), ESO 184-G82 has a luminosity of 0.5LLMC1.2LLMC.
There are two possible explanations for why ESO 184-G82 is undergoing strong star formation. These explanations may be related. First, ESO 184-G82 has a bar. Whereas the presence of a bar in spiral galaxies does not, in general, correlate with the global star formation rate (Kennicutt 1998), the bar strength has been found to correlate strongly with the global star formation rate of the galaxy (Aguerri 1999). Second, although we have a measured redshift only for ESO 184-G82 (Tinney et al. 1998), the galaxy is probably member of a group of at least five galaxies simply based on the low likelihood of a chance projection. The nearest neighbor is an Sb galaxy at a projected separation of only 8.3 h-1 kpc. Moreover, ESO 184-G82 shows indications of being morphologically disturbed. Interaction-induced star formation is therefore a likely explanation.
The SN occurred in a star-forming region with several bright, young stars within a projected distance of 
100 pc. In this respect, it is similar to most core-collapse SNe that typically are located within young stellar associations in spiral arms (Barth et al. 1996; Van Dyk et al. 1999).
The magnitude of the emission at the position of the SN 778 days after the discovery of the GRB is 1 mag brighter than one would expect from a simple radioactive decay model (Sollerman et al. 2000). The emission in both the CL and LP image is well fit by a PSF at a position that is consistent with the precise position of the SN determined in § 3.1. The residuals remaining after subtracting the scaled PSF are consistent with noise. This suggests that the majority of the observed emission could in fact be from the SN. We tested our ability to resolve a star cluster in ESO 184-G82 by adding artificial clusters with Michie-King (Michie 1963; King 1966) profiles and the same flux as the SN remnant to the CL image as described in Holland, Côté, & Hesser (1999). We found that star clusters with core radii of rc 0013 (=1.6 h-1 pc) appeared as extended objects and are not well fit by the PSF. This suggests that, if the SN did occur in a star cluster, the underlying star cluster is either physically very small and thus cannot be distinguished from a point source or that the underlying star cluster is very faint. If the emission is really dominated by the fading SN, as can only be confirmed by follow-up observations with the HST, this would imply a significant flattening of the light curve, possibly explained by interaction with circumstellar material or accretion on a black hole (Leibundgut et al. 2000; Patat et al. 2000; Sollerman et al. 2000) or by the effect of the positron decay (Nakamura et al. 2000).
4.2. Implications for GRB Research
These data provide a very solid confirmation of the hypothesis that at least some GRBs are strongly related to star formation (Paczy
ski 1998). However, some GRBs, such as GRB 970508, show evidence that there was no SN like SN 1998bw associated with them. Therefore, either GRBs are associated with nonstandard-candle SNe or only some GRBs are associated with SNe and the others have a different progenitor. Graziani, Lamb, & Marion (1999) have used Bayesian methods to estimate that no more than 5% of the GRBs that have been detected by BATSE were produced by known SNe IbIc. However, Galama et al. (1998) conservatively estimate the probability of the coincidence between GRB 980425 and SN 1998bw to be 10-4. This suggests that GRB 980425/SN 1998bw was a member of an unusual class of GRBs/SNe.
GRB 980425 had a isotropic equivalent energy of 1048 ergs (Pian et al. 2000), making it 104 less energetic than the other GRBs for which distances (and thus isotropic equivalent energies) have been determined. If GRB 980425 had occurred at z = 1 (the median redshift of the GRBs with measured redshifts), it would not have been detected. As of 2000 June 30, redshifts and total isotropic energies have been determined for only 10 GRBs, with nine having energies between 2 × 1051 and 3 × 1054 ergs and one having an energy of 1048 ergs. If we assume that this ratio holds for the entire BATSE 4B catalog (Paciesas et al. 1999), then the observed number of high-energy (E 
1051 ergs) GRBs with known redshifts (nine) implies that a high-energy GRB will occur at z 
0.01 only once every 20,000 yr. Therefore, it is not surprising that we have not observed any high-energy GRBs with z 0.01. GRBs with isotropic equivalent energies of 1048 may be common at z 1 but are undetectable with current instruments.
We thank J. Sollerman and our anonymous referee for several useful comments that improved our manuscript on several important points. We also thank P. Rautiainen for determining the Hubble type of ESO 184-G82. This work was supported by the Danish Natural Science Research Council (SNF).
REFERENCES
. 2000, ApJ, 536, 185 First citation in article | IOPscience | ADS. 2000b, GCN Circ. 704 (http://gcn.gsfc.nasa.gov/gcn/gcn3/704.gcn3) First citation in articleski, B. 1998, ApJ, 494, L45 First citation in article | IOPscience | ADS
25 east of north. The arrow marks the position of the SN (