An international research team, (*1) led by astronomers from the University of Tokyo, Hiroshima University, and the National Astronomical Observatory of Japan, has obtained a spectrum of Supernova 2003jd (*2) using the Faint Object Camera And Spectrograph (FOCAS) instrument on the Subaru telescope in Hawai’i. The supernova was actually a hypernova, the product of catastrophic core collapse in an extremely massive star. The data, taken on September 12, 2004 (about a year after the initial explosion), exhibited some of the expected emission lines of various elements, including oxygen, nitrogen, and magnesium. However, there was something out of the ordinary: double-peaked oxygen emission lines at 630 and 636 nanometers which suggested that the astronomers were witnessing a donut of oxygen rich debris from the side. The debris was created as two jets of material emerged from the explosion near the speed of light, punching a hole through material that had been a star. This matches predictions of what gamma-ray bursts, some of the most energetic outbursts in the universe, would look like when viewed from the side. This observation marks the first time such a line shape has been observed, overcoming the difficulty of taking spectra of a fading supernova this later after the initial explosion. This discovery gives firm evidence supporting a unified theory that gamma-ray bursts are the products of light-speed jets moving out from asymmetric hypernova explosions.

While many stellar explosions are spherically symmetrical, some are not. In these cases, the deviation from spherical symmetry gives a hint as to what is going on during the explosion. A gamma-ray burst (GRB) is among the most energetic explosions in the universe and is an interesting example of an aspherical explosion.

Astronomers were baffled for decades by gamma-ray bursts. Following the successful observation of a number of gamma-ray bursts and their after-glows, the current understanding is that these events are taking place at cosmological (that is, very large) distances – from as far away as a few billion light-years. At those distances, an explosion would need to be extremely large, bright and energetic to be observable from Earth. If these explosions were spherically symmetric, their energy output would exceed the Sun’s total energy output over a lifetime by several times within in a very short period. If the explosion is jet-like rather than spherical, the energy output can be more modest and more realistic. In this case, only gamma-ray bursts whose jet axis happens to point toward Earth would be observable.

Recent research supports the jet hypothesis. A class of long-duration gamma-ray bursts had been linked with hypernova explosions though previous research. For example, astronomers from the University of Tokyo and the National Astronomical Observatory of Japan found that the gamma-ray burst GRB030329 and the hypernova SN 2003dh appeared at the same place and at the same time (*3) in the year 2003. They also discovered possible evidence for high velocity jets associated with the hypernova SN2002ap whose polarized light and unpolarized light have different Doppler shifts. Summarizing these discoveries, the research team proposed a unified model for gamma-ray bursts: a hypernova explosion of a collapsing massive star releasing a pair of high-velocity jets (*4).

This bipolar explosion model predicts that relatively light elements such as oxygen should be ejected in a doughnut-shaped debris ring at the “equator” of an explosion. If the event is viewed from the polar direction, a gamma-ray burst shows up. If it is observed from the side, the gamma-ray burst cannot be seen. Furthermore, oxygen emission lines will show a different appearance that depends on the viewing orientation, because the observed Doppler shift depends on the orientation of the non-spherically symmetric debris. The line should appear as a single-peaked line for a face-on observer in the polar direction. For an edge-on observer placed at the equator, it will appear as a double-peaked line, corresponding to materials moving toward and away from us (*5).

The observation of SN 2003jd (Figure 1) provided astronomers a golden opportunity to link models of such explosions with observed evidence of bipolar structure and its effect on the expected observations of a gamma-ray burst. Earlier observations of the supernova during its peak in brightness had already revealed that it was a hypernova, an explosion caused by the catastrophic core collapse of an extremely massive star. Since there was no known gamma-ray burst accompanying the hypernova, SN2003jd was an excellent observational candidate to test the unified theory of gamma-rays and Hypernovae. Up to the present, the spectra of Hypernovae had shown only single-peaked oxygen emissions lines. This is because astronomers preferentially observed hypernovas associated with gamma-ray bursts, and because such observations are difficult to do and thus few in number. The difficulty arises because hypernovas fade very rapidly. The light gathering and resolving power the Subaru telescope’s large 8.2 meter diameter aperture gave the research team a chance to observe a double-peaked emission line in SN2003jd.

Indeed, when the team observed SN 2003jd on September 12, 2004, about one year after the initial explosion, there were prominent oxygen lines in the red part of the spectrum at 630 and 636 nm matching the double-peaked profile predicted by the theory. This discovery provides strong evidence that the unified model proposed by the research group is correct: a Gamma-ray burst is produced by light-speed jets from a very asymmetric hypernova explosion.

This work was published in the May 27, 2005, edition of the journal Science.

		Figure1 A Subaru/FOCAS image of SN 2003jd and its host galaxy MCG-01-59-21 (about three hundred million light years from the Earth). The image is synthesized from B-band (with 30 seconds exposure) and R-band (5 seconds) images taken by Subaru on 12 September 2004. In the image, top and left correspond to north and east, and the field of view is about 3.2 arcmin x 2.3 arcmin. SN 2003 jd appeared at the left-bottom of the nucleus of the host galaxy, as marked by yellow lines.
		Figure2 The distribution of elements numerically computed for the bipolar hypernova explosion model (left-bottom). Theoretical line profiles computed on the basis of the elements distribution are compared with oxygen lines detected in SNe 1998bw (top) and 2003jd (right-bottom). In the explosion, high-velocity jets are ejected along the polar directions (top and bottom directions). Iron (colored in green and blue) is ejected toward the polar direction, while oxygen (brown) is along the equatorial direction. The latter is confined in a dense, doughnut-like shaped debris (see density contours shown in black). In the panels for the emission lines, the theoretically predicted lines (RED) for an observer placed at the polar direction (top) and at the equatorial direction (right-bottom) are compared to observed oxygen emission line profiles (BLACK) in SN 1998bw and 2003jd, respectively. The emission line profiles are in good accordance with the interpretation, that is, SNe 1998bw and 2003jd are intrinsically similar events, but viewed at the different direction. It is also consistent with the fact that a Gamma-Ray Burst showed up with SN 1998bw but not with SN 2003jd.