Astron. Astrophys. 347, 500-507 (1999)

1. Introduction

Supernova (SN) 1987A provides a convenient tool to test our understanding of nucleosynthesis in massive stars and during supernova explosions (Thielemann et al. 1996). In particular, the late ( days) emission probes directly the elemental abundances deep in the stellar ejecta. As in other core-collapse SNe without a dense circumstellar medium, the energy production is at this epoch dominated by radioactive energy from the decay of ⁵⁶Co to ⁵⁶Fe, the cobalt itself being the decay product of ⁵⁶Ni. It was quickly realized that the ejected mass of ⁵⁶Ni in SN 1987A was at least (Woosley et al. 1988), the actual number being close to (e.g., Suntzeff & Bouchet 1990). This decay continued to power the bolometric light curve in an undisputed fashion for about days (Bouchet et al. 1996).

However, during the following epoch, radioactive decay seemed to be unable to explain the bolometric flux of the supernova, unless a large amount of ⁵⁷Ni (decaying to ⁵⁷Co and further to ⁵⁷Fe) was included (Suntzeff et al. 1992). Fransson & Kozma (1993) demonstrated that the derived ⁵⁷Ni/⁵⁶Ni ratio could be much closer to the solar ⁵⁷Fe/⁵⁶Fe ratio when time dependence was accounted for. This effect, now known as the "freeze-out" effect, stems from the fact that the energy stored at earlier epochs in the low-density hydrogen gas slowly leaks out and dominates the bolometric light curve after days (Fransson & Kozma 1999). The freeze-out effect becomes less important as the radioactive isotope ⁴⁴Ti starts to dominate. This occurs after days (Woosley et al. 1989; Kumagai et al. 1991; Fransson & Kozma 1999). ⁴⁴Ti decays to ⁴⁴Sc on a time scale of years (Ahmad et al. 1998; Görres et al. 1998), and then quickly further to ⁴⁴Ca.

It is important to determine the ⁵⁶Ni, ⁵⁷Ni and ⁴⁴Ti masses in order to constrain models for the supernova explosion and the explosive nucleosynthesis (e.g., Timmes et al. 1996). In the case of ⁴⁴Ti, models predict that could be synthesized in SN 1987A (e.g., Kumagai et al. 1991; Woosley & Hoffman 1991).

Chugai et al. (1997) estimate that the mass of ejected ⁴⁴Ti should be from the optical line emission at 2875 days. Kozma (1999) and Kozma & Fransson (private communication 1999; hereafter KF99) preliminary find that best explains broad-band photometry of SN 1987A for days. However, nearly all of the emission comes out in the far infrared which is not included in the observed bands. KF99 show that a more definite estimate can be made if one could measure the flux in a few iron lines, mainly [Fe II] 25.99 . There are several reasons for this. First, at days KF99 find that almost half of the luminosity from the supernova is emerging in this line. Second, because the iron lines are only little affected by freeze-out they are good tracers of the instantaneous energy deposition, and the flux of [Fe II] 25.99 is therefore almost proportional to . Third, this line is formed through collisional excitations and is therefore insensitive to uncertainties in atomic data involved in calculating the recombination cascade. In a preliminary analysis, Borkowski et al. (1997) find that their Infrared Space Observatory (ISO; Kessler et al. 1996) observations give an upper limit on the line flux which corresponds to only .

Here we report on observations we have made with ISO. Although we have observed the supernova over the entire wavelength range , we will concentrate our discussion on the region, , where the strongest emission lines are expected to emerge. We compare these observations with theoretical modeling. We also briefly discuss our observations at longer wavelengths where we detect emission originating from elsewhere in the LMC.

© European Southern Observatory (ESO) 1999

Online publication: June 30, 1999
helpdesk.link@springer.de