1. Astronomical Institute "Anton Pannekoek", University of Amsterdam & Center for High Energy Astrophysics, Kruislaan 403, 1098 SJ Amsterdam,The Netherlands
2. Physics Department, University of Alabama in Huntsville , Huntsville, Alabama 35899, USA
3. Universities Space Research Association, Huntsville, Alabama 35812, USA
4. NASA Marshall Space Flight Center, ES-84, Huntsville, Alabama 35812, USA
5. ESO, Casilla 19001, Santiago 19, Chile
6. SRON Laboratory for Space Research, Sorbonnelaan 2, 3584 CA Utrecht, The Netherlands
7. Space Sciences Laboratory, Berkeley, California 94720-7450, USA
8. Netherlands Foundation for Research in Astronomy, Postbus 2, 7990 AA Dwingeloo, The Netherlands
9. Osservatorio Astronomico di Trieste, Via G.B. Tiepolo 11, I-34131 Trieste, Italy
10. Department of Astronomy, School of Science, University of Tokyo, Tokyo 111, Japan
11. Research Center for the Early Universe, School of Science, University of Tokyo, Tokyo 111, Japan
12. Institute of Astronomy, Madingley Road , Cambridge CB3 0HA, UK
13. Department of Physics and Astronomy, SUNY, Stony Brook, New York 11794-3800, USA
14. Astrophysikalisches Institut, Potsdam, Germany
15. Istituto Tecnologie e Studio Radiazioni Extraterrestri, CNR, Bologna, Italy
16. Istituto di Fisica Cosmica e Applicazioni all'Informatica, CNR, Via U. La Malfa 153, I-90146 Palermo , Italy
17. Istituto di Astrofisica Spaziale, CNR, Roma, Italy
18. Mt Stromlo and Siding Spring Observatories, The Australian National University, Weston Creek, ACT 2611, Australia
19. Anglo-Australian Observatory, PO Box 296, Epping, NSW 2121, Australia
20. School of Physics A29, University of Sydney, NSW 2006, Australia
21. Department of Astronomy, PO Box 3818 , University of Virginia, Charlottesville, Virginia 22903, USA.

Correspondence to: T. J. Galama¹ Correspondence and requests for materials should be addressed to T.J.G.
(e-mail: Email: titus@astro.uva.nl).

GRB980425 was detected¹⁰ on 1998 April 25.91  UTwith one of the Wide Field Cameras (WFCs) and the Gamma Ray Burst Monitor on board BeppoSAX, and with the Burst and Transient Source Experiment (BATSE) on board the Compton Gamma Ray Observatory. The BATSE burst profile consisted of a single wide peak. The burst flux rose in +5 s to a maximum flux of (3.0 0.3) 10^-7 erg cm ^-2 s^-1 (24–1,820 keV), at which it remained for 5 s; it decayed steadily to the background in 25 s. The burst fluence E_b is (4.4 0.4) 10^-6 erg cm^-2 ; its duration¹¹ T₉₀ is 23.3 1.4 s. The burst spectrum is well described by a smoothly broken power law, with a constant break energy (148 33 keV) and high-energy power-law photon index (-3.8 0.7); the low-energy power-law photon index varied from -1.0 0.15 during the rise, to -2.6 0.2 during the decay of the burst. Thus, with respect to its -ray properties, GRB980425 was not a remarkable event. In the Beppo SAX WFC no. 2 the burst lasted 30 s, and reached a peak intensity of 3 Crab (2–28 keV)¹⁰. The position derived from the WFC image is right ascension (RA) 19 h 34 min 54 s, declination (dec.) -52° 49.9' (J2000.0), with an error radius of 8' which comprises a 3' statistical error (99% confidence level) and a 5' systematic uncertainty due to incomplete satellite attitude information.

We observed the error box of GRB980425 in R_MACHO and B _MACHO wavebands¹² with the 50-inch telescope at the Australian National University's (ANU) Mt Stromlo Observatory (MSO) starting April 26.60  UT, and in standard U, B, V, R and I bands with the 30-inch telescope at MSO, the 40-inch telescope at the ANU Siding Spring Observatory, the Anglo-Australian Telescope at the Anglo-Australian Observatory, the 3.5-m New Technology Telescope (NTT), and the 1.5-m Danish and the 0.9-m Dutch telescopes at the European Southern Observatory.

Inspection of NTT images obtained on April 28.4 and May 1.3  UTrevealed a point source in the WFC error box, which was not visible in the Digitized Sky Survey. Tying 40-inch V- and R-band images to the Hipparcos Tycho coordinate system, we determined its position at RA 19 h 35 min 03.34 s (0.02 s), dec. = -52° 50' 44.8" (0.2") (J2000.0), 1.6' away from the centre of the WFC error box. The source is within the BATSE/Ulysses Interplanetory Network annulus and coincides with the transient radio source in the WFC error box⁸ to within 0.3". It is located in an H IIregion in a spiral arm of the face-on barred spiral galaxy ESO184-G82 at a redshift⁶ of 2,550 km s^-1, in the DN1931-529 group of galaxies¹³.

The UBVRI light curves of the transient are shown in Fig. 1, also see Table 1. These light curves are very different from those of -ray burst (GRB) afterglows—which decay as a power law, F(t) t^-, with in the range¹⁴ 1–2—but are quite similar to those of supernovae. This, and the similarity of its spectrum to that of some supernovae (for example SN1994I, Fig. 2) leads us to conclude that the transient is a very luminous supernova of type Ic.

Figure 1: UBVRI light curves of SN1998bw, corrected for galactic foreground extinction, A_V = 0.20, as inferred from a combination of COBE/DIRBE and IRAS/ISSA maps²⁵.

Time is in days since April 25.91 UT(a full log of the observations and the photometry can be found at http://www.astro.uva.nl/titus ). We determined a photometric (U, B, V, R and I) calibration for a number of reference stars using NTT (May 4.4 UT) and 1.5-m Danish telescope (May 8.3 UT) observations of the Landolt²⁶ fields Mark A and SA110 (stars 496–507) (magnitudes of the reference stars can be found at http://www.astro.uva.nl/titus). We corrected for atmospheric extinction and, for U and B, also for a first-order colour term. By comparison of these two calibration nights we estimate an error of the absolute calibration of 0.10 mag in U and 0.05 mag in B, V, R and I. The R_MACHO and B_MACHO observations have been transformed using ref. 12. We consider a conservative minimum error of 0.03 mag realistic for the differential U, B, V, R and I light curves to account for the effect of seeing on the contribution of the underlying galaxy (<0.01 mag for each band) and the different instruments used. The longer the wavelength is, the later the maximum light occurs (see Table 1). The R-band light curve shows an initial 'plateau', then it rises at a rate of +0.25 mag d^-1 to maximum light on May 12. Lack of early data prevents us from establishing the existence of the plateau in the U, B, V and I light curves. Starting early June 1998 the light curves decay exponentially with 0.025 mag d ^-1. For comparison the V-band light curve of the type Ic SN1994I is shown²⁷. As discussed in ref. 16, the width of the light-curve peak depends on the total ejected mass and the explosion energy.

High resolution image and legend (44K)

Figure 2: Representative spectra near maximum light of SN1998bw, SN1994I (type Ic; ESO supernova archive, courtesy of M.Turatto), and SN1984L (type Ib)²⁸.

Hydrogen lines, characteristic of type II supernovae, and Si II, characteristic of type Ia supernovae, are absent in the spectrum of SN1998bw. The strong He I5876 line which characterizes type Ib supernovae is very weak in SN1994I and absent in SN 1998bw. The overall shape of the spectrum of SN1998bw is similar to that of a type Ic supernova, although the spectral features are less pronounced. The difference is strongest in the 3,500–5,000 å region, where the Ca IIand Fe IIlines are much weaker than in SN1994I. In this respect, SN1998bw appears to represent an extreme case in the odd class of type Ic supernovae.

High resolution image and legend (55K)

Table 1: Times of maximum, and apparent and absolute peak magnitudes of SN1998bw

Full table

Any estimate of the probability that the supernova and the GRB coincided by chance (with respect to both time and direction) suffers from the problem of aposteriori statistics; that is, that the parameters of the problem tend to be set by the observed phenomenon itself. In this case the parameters are the size of the error box, the peak magnitude of the supernova, and the time window within which the events can be considered as possibly related. In our computation we have made generous estimates of these parameters.

The WFC error boxes have 99% confidence level radii varying between 3' and 8' (ref. 15). We conservatively estimate the angular distance, beyond which a connection can be rejected, at 10'. We included all supernovae with peak B-band magnitudes m _B< 16; that is 2 mag below that of SN1998bw. The time of occurrence of the core collapse and the GRB coincide to within (+0.7, -2.0) days (ref. 16). As a GRB which occurred a few days earlier or later would have been considered at least remarkable, we have taken a time window of 10 days.

With peakbsolute magnitudes M_B - 5 log h = -18.28, -16.68 and -15.69 (where h is the Hubble constant in units of 100 km s ^-1 Mpc^-1) for supernovae of types Ia, Ib/c and II, respectively¹⁷, for m _B < 16 such supernovae are detectable out to redshifts of 7,180, 3,440 and 2,180 km s^-1, respectively. (We note that these limiting values are independent of the assumed value of the Hubble constant.) The Shapley–Ames 'fiducial' sample of 342 galaxies within the Virgo circle¹⁷ has a mean B-band luminosity of 6.7h^-2 10 ⁹L(B), and a supernova rate of 3.09h²[100 yr 10 ¹⁰L (B)]^-1. Using galaxy numbers and heliocentric radial velocities from ref. 18, assuming a mean luminosity, galaxy composition, and supernova rate as in the 'fiducial' sample, and taking relative supernova rates¹⁹ for types II:Ib/c:Ia of 4.0:0.8:1.8, we find a total rate of supernovae (with m_B <16 at the peak) of 80 per year. This value includes a correction for absorption¹⁷ within the host galaxy disk. This number should perhaps be increased by a modest factor to account for incompleteness of the radial-velocity distribution¹⁸; we have adopted a final supernova rate (m_B <16) of 120 per year.

With the above parameters, we estimate the probability of catching a supernova in one of the 13 WFC GRB error boxes to be 9 10^-5. In our probability estimate we have included all supernovae with peak magnitudes two magnitudes below that of SN1998bw, and we have ignored the fact that SN1998bw is of a rare type. We therefore believe our estimate is conservative. As a result, the notion that GRB980425 and SN1998bw are physically related becomes difficult to reject purely on the basis of the fact that afterglows observed so far from GRBs are very different from supernovae.

The WFC error box contains two X-ray sources²⁰,²¹, neither of which coincides with SN1998bw. One, 1SAX J1935.0 - 5248 has a constant (2–10 keV) flux of 2 10 ^-13 erg cm^-2 s^-1 . The other, 1SAX J1935.3 - 5252, was detected at (1.6 0.3) 10^-13 erg cm ^-2 s^-1 about 1 day after the burst, and decayed to <1.2 10^-13 erg cm ^-2 s^-1 (3) in 22 hours; it was not detected 6 days after the burst (<1.0 10^-13  erg cm^-2 s^-1 ). This variability is consistent with that of previously observed X-ray afterglows of GRBs, and this object might be a possible counterpart for GRB980425. Comparison of the 50-inch April 26.63 UTand April 28.68 UTimages at the locations of the two X-ray sources shows no sources variable by more than 0.2 mag down to R = 21. However, several GRBs have not shown optical afterglows either, most notably GRB970111²² and GRB970828²³.

The (2–10) keV detection limit (3) for the GRB980425 Narrow-Field Instrument observations was 1.2 10^-13 erg s ^-1 cm^-2. Using the ASCA (2–10 keV) source count distributions²⁴ one expects to find an average of 0.6 X-ray sources above this limit in the WFC error box; the probability of finding two or more sources there by chance coincidence is 12%. The case for a relation between this X-ray source and GRB980425 must therefore be considered tentative at best, in particular because variability is not rare among weak ROSAT sources.

Modelling¹⁶ of the optical light curve of SN1998bw shows that it can be produced with the core collapse of a massive progenitor star composed mainly of carbon and oxygen (a C+ O star); the time of collapse coincides with that of the GRB to within (+0.7, -2.0) days. In the case of the C+ O star core collapse, the kinetic energy was 10 ^52.5 erg. To achieve the observed high luminosity, substantial amounts of ⁵⁶Ni (0.7 solar masses) have to be synthesized in the explosion¹⁶; the large energy and ⁵⁶Ni mass would be unprecedented for a core-collapse supernova.

If one accepts the possibility that GRB980425 and SN1998bw are associated, one must conclude that GRB980425 is a rare type of GRB, and SN1998bw is a rare type of supernova. The radio properties⁸,⁹ of SN1998bw show the peculiar nature of this event independent of whether or not it is associated with GRB980425.

The consequence of an association is that the -ray peak luminosity of GRB980425 is L = (5.5 0.7) 10⁴⁶ erg s^-1 (in the 24–1,820 keV band) and its total -ray energy budget is (8.1 1.0) 10⁴⁷ erg. These values are much smaller than those of 'normal' GRBs which have peak luminosities of up to 10⁵² erg s^-1 and total energies⁵ up to several times 10⁵³ erg. This implies that very different mechanisms can produce GRBs which cannot be distinguished on the basis of their -ray properties, and that models explaining GRB980425/SN1998bw are unlikely to apply to 'normal' GRBs and vice versa.

References

1. Costa, E. et al . Discovery of an X-ray afterglow associated with the -ray burst of 28 February 1997. Nature 387, 783– 785 (1997). | Article | ISI | ChemPort |
2. Van Paradijs, J. et al . Transient optical emission from the error box of the -ray burst of 28 February 1997. Nature 386, 686 –688 (1997). | Article | ChemPort |
3. Frail, D. A. , Kulkarni, S. R. , Nicastro, L. , Feroci, M. & Taylor, G. B. The radio afterglow from the -ray burst of 8 May 1997. Nature 389, 261– 263 (1997). | Article | ISI | ChemPort |
4. Metzger, M. R. et al . Spectral constraints on the redshift of the optical counterpart to the -ray burst of 8 May 1997. Nature 387, 878–880 (1997). | Article | ISI | ChemPort |
5. Kulkarni, S. R. et al . Identification of a host galaxy at redshift z = 3.42 for the -ray burst of December 1997. Nature 393, 35–39 (1998).  | Article | ISI | ChemPort |
6. Tinney, C. , Stathakis, R. , Cannon, R. & Galama, T. J. IAU Circ. No. 6896 (1998).
7. Sadler, E. M. , Stathakis, R. A. , Boyle, B. J. & Ekers, R. D. IAU Circ. No. 6901 (1998).
8. Wieringa, M. et al . IAU Circ.No. 6896 (1998).
9. Kulkarni, S. R. et al . IAU Circ.No. 6903 (1998).
10. Soffitta, P. et al . IAU Circ.No. 6884 (1998).
11. Kouveliotou, C. et al . Identification of two classes of gamma-ray bursts. Astrophys. J. 413, L101–L104 (1993). | Article | ISI | ChemPort |
12. Bessell, M. S. & Germany, L. M. Calibration of the MACHO photometric system. Publ. Astron. Soc. Pacif. (submitted).
13. Duus, A. & Newell, B. Acatalog of Southern groups and clusters of galaxies. Astrophys. J. Suppl. 35, 209 –219 (1977). | Article | ISI |
14. Groot, P. J. et al . The rapid decay of the optical emission from GRB980326 and its possible implications. Astrophys. J. (in the press).
15. Heise, J. et al . in Conf. Proc. 4th Huntsville Symposium on Gamma-Ray Bursts (eds Meegan, C., Preece, R., Koshut, T.) 428, 397– 403 (Am. Inst. Phys., New York, (1998)).
16. Iwamoto, K. et al . Ahypernova model for the supernova associated with the -ray burst of 25 April 1998. Nature 395, 672– 674 (1998). | Article | ISI | ChemPort |
17. Van den Bergh, S. & Tammann, G. A. Galactic and extragalactic supernova rates. Annu. Rev. Astron. Astrophys. 29, 363–407 (1991).  | Article |
18. Giovanelli, R. & Haynes, M. P. Redshift survey of galaxies. Annu. Rev. Astron. Astrophys. 29, 499–541 (1991). | Article | ISI |
19. Cappellaro, E. et al . Astron. Astrophys. 322, 431 –441 (1997).
20. Pian, E. , Frontera, F. , Antonelli, L. A. & Piro, L. GCN MessageNo. 69 (1998).
21. Pian, E. et al . GCN Message No. 61 (1998).
22. Castro-Tirado, A. et al . IAU Circ. No. 6598 ((1997 )).
23. Groot, P. J. et al . Asearch for optical afterglow from GRB970828. Astrophys. J. 493, L27–L30 (1998).  | Article | ISI |
24. Cagnoni, I. , Della Ceca, R. & Maccacaro, T. Amedium survey of the hard X-ray sky with the ASCA Gas Imaging Spectrometer: the (2–10 keV) number counts relationship. Astrophys. J. 493, 54– 61 (1998). | Article | ISI |
25. Schlegel, D. J. , Finkbeiner, D. P. & Davis, M. Maps of dust IR emission for use in estimation of reddening and CMBR foregrounds. Astrophys. J. (in the press); also as preprint astro-ph/910327 available at http://xxx.lanl.gov (1998).
26. Landolt, A. U. UBVRI photometric standard stars in the magnitude range 11.5 < V < 16.0 around the celestial equator. Astron. J. 104, 340–376 (1992).  | Article | ISI |
27. Richmond, M. W. et al . UBVRI photometry of the type Ic SN 1994I in M51. Astron. J. 111, 327–339 (1996). | Article | ISI |
28. Wheeler, J. C. & Levreault, R. The peculiar type I supernova in NGC 991. Astrophys. J. 294, L17–L20 (1985). | Article | ISI | ChemPort |

Acknowledgements

This work is based partly on observations made by the MACHO Project with the 50-inch telescope at the ANU's Mt Stromlo Observatory (ANUMSO), by H. Jerjen with the 40-inch telescope at the ANU's Siding Spring Observatory, and on observations made at the European Southern Observatory, La Silla, Chile. We thank the RAPT Group of amateur astronomers (E. Pozza, A. Brakel, B. Crooke, S. McKeown, G. Wyper, K. Ward, D. Baines, P. Purcell, T. Leach, J. Howard, D. McDowell, M. McDonald, A. Salmon and A. Gurtierrez) for providing images from the 30-inch telescope at ANUMSO, and the SuperCOSMOS team for making a scan of an SERC Survey Plate taken with the UKST. J.v.P., C.K., M.K. and K.H. were supported by NASA.