The discovery of the unusual supernova SN1998bw, and its possible association with the ␥-ray burst GRB 980425 1-3 , provide new insights into the explosion mechanism of very massive stars and the origin of some classes of ␥-ray bursts. Optical spectra indicate that SN1998bw is a type Ic supernova 3,4 , but its peak luminosity is unusually high compared with typical type Ic supernovae 3 . Here we report our findings that the optical spectra

Over the past five years evidence has mounted that long-duration (> 2 s) γ-ray bursts (GRBs) the most brilliant of all astronomical explosionssignal the collapse of massive stars in our Universe. This evidence was originally based on the probable association of one unusual GRB with a supernova 1 , but now includes the association of GRBs with regions of massive star formation in distant galaxies 2,3 , the appearance of supernova-like 'bumps' in the optical afterglow light curves of several bursts 4-6 and lines of freshly synthesized elements in the spectra of a few X-ray afterglows 7 . These observations support, but do not yet conclusively demonstrate, the idea that long-duration GRBs are associated with the deaths of massive stars, presumably arising from core collapse. Here we report evidence that

No abstract

Nature © Macmillan Publishers Ltd 19988 amounts of 56 Ni (ϳ0.7 solar masses) have to be synthesized in the explosion 16 ; the large energy and 56 Ni 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 properties 8,9 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 ␥ ¼ ð5:5 Ϯ 0:7Þ ϫ 10 46 erg s −1 (in the 24-1,820 keV band) and its total ␥-ray energy budget is (8:1 ϫ 1:0Þ ϫ 10 47 erg. These values are much smaller than those of 'normal' GRBs which have peak luminosities of up to 10 52 erg s −1 and total energies 5 up to several times 10 53 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. Ⅺ

We present and discuss the photometric and spectroscopic evolution of the peculiar SN 1998bw, associated with GRB 980425, through an analysis of optical and near-IR data collected at ESOÈLa Silla. The spectroscopic data, spanning the period from day [9 to day ]376 (relative to B maximum), have shown that this supernova (SN) was unprecedented, although somewhat similar to SN 1997ef. Maximum expansion velocities as high as 3 ] 104 km s~1 to some extent mask its resemblance to other Type Ic SNe. At intermediate phases, between photospheric and fully nebular, the expansion velocities (D104 km s~1) remained exceptionally high compared to those of other recorded core-collapse SNe at a similar phase. The mild linear polarization detected at early epochs suggests the presence of asymmetry in the emitting material. The degree of asymmetry, however, cannot be decoded from these measurements alone. The He I 1.083 and 2.058 km lines are identiÐed, and He is suggested to lie in an outer region of the envelope. The temporal behavior of the Ñuxes and proÐles of emission lines of Mg I] j4571, [O I] jj6300, 6364, and a feature ascribed to Fe are traced to stimulate future modeling work. The uniqueness of SN 1998bw became less obvious once it entered the fully nebular phase (after 1 yr), when it was very similar to other Type Ib/cÈIIb objects, such as the Type Ib SN 1996N and the Type IIb SN 1993J, even though SN 1998bw was 1.4 mag brighter than SN 1993J and 3 mag brighter than SN 1996N at a comparable phase. The late-phase optical photometry, which extends up to 403 days after B maximum, shows that the SN luminosity declined exponentially but substantially faster than the decay rate of 56Co. The ultraviolet-optical-infrared bolometric light curve, constructed using all available optical data and the early JHK photometry presented in this work, shows a slight Ñattening starting on about day ]300. Since no clear evidence of ejecta-wind interaction was found in the late-time spectroscopy (see also the work of Sollerman and coworkers), this may be due to the contribution of the positrons since most c-rays escape thermalization at this phase. A contribution from the superposed H II region cannot, however, be excluded.