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

2. Observations and results

We have used ISO to observe SN 1987A on several occasions. Both the Short Wavelength Spectrograph (SWS; de Graauw et al. 1996) and the Long Wavelength Spectrograph (LWS; Clegg et al. 1996) were used, and Table 1 summarizes our observations. We will concentrate here mainly on the SWS observation from 4 February, 1998 (Sect. 2.1), though we provide a consistency check of this observation against our other SWS observations. The LWS data are discussed in Sect. 2.2.

Table 1. Log of observations.
Notes:
a) Epochs in days past explosion.
b) Time in seconds.

2.1. SWS observations

The SWS observation on 4 February, 1998 was made in the SWS01 mode with speed 4 and was centered on the position of SN 1987A, i.e., R.A. = 5h 35m 28.05s; Decl. = 16´ 1164; (J2000.0). SWS01 provides spectra which together cover the entire wavelength region between . However, as our models predict [Fe I] 24.05 and [Fe II] 25.99 to be by far the strongest emission lines from the supernova at this epoch (day 3 999; see Sect. 3), we have concentrated on measuring the flux in the wavelength range including these two lines. This range, between , is covered by band 3D of SWS.

The reductions were made using the SWS Interactive Analysis software system (SIA) available at ISO Spectrometer Data Center (ISOSDC) at the Max Planck Institut für Extraterrestrische Physik in Garching (MPE). The most recent set of calibration files equivalent to off-line processing (OLP/pipeline) version 7.0 was used. The interactive reduction allows special care to be given to dark subtractions, which is of particular interest when measuring low flux levels. Flat fielding was also applied, but we have not made any fringe corrections since the fringes at low flux levels disappear in the noise.

In Fig. 1 we present a fully reduced spectrum of SN 1987A for band 3D. Although the nominal instrumental resolution of SWS is , the slowest SWS01 mode with speed 4 degrades this by a factor of 2. We have therefore averaged the spectrum with a bin size of 0.052 , corresponding to for the two lines of interest. As the lines could extend to well above this resolution should be sufficient to resolve the lines. However, neither [Fe I] 24.05 nor [Fe II] 25.99 are seen in the spectrum. The `features' that do appear around and are most likely due to instrumental effects, as they are rather robust in the sense that they appear in many detectors and in both up and down scans. They are certainly not due to fringes, nor do we believe they are effects of pure noise. The zero level of the spectrum is well defined and a simple zero-order fit to the spectrum gives an RMS of  Jy over the range .

Fig. 1. ISO SWS/AOT1 Band 3 spectrum of SN 1987A on day 3999. The spectrum was reduced using interactive software (see text). The bin size was set to , in accordance with the instrumental resolution. Overlaid (dashed line) is the modeled line emission from model M2 (without photoionization) described in Table 2 and Sect. 3.2. To obtain the modeled line profiles we have assumed a maximum core velocity of , and that the emission is homogeneously distributed throughout the core (see Sect. 3.3). The modeled peak flux of [Fe II] 25.99 corresponds to what is needed for a 3 detection.

Looking more closely at the spectral regions of the two lines, we have simply measured the RMS between and separately. Again, we used a zero-order fit to the smoothed spectrum. A zero-order fit should provide the most conservative way to estimate the `noise' in our data. The resulting RMS values are  Jy and  Jy for [Fe I] 24.05 and [Fe II] 25.99 , respectively. This gives 3 upper limits of  Jy and  Jy for the peak of the profile of the two lines. These values do not change much even if we double the bin size.

SN 1987A was also observed in the SWS02 mode on three occasions (see Table 1). This mode provides scans over shorter wavelength ranges than the full SWS01 scans. We have reduced also the SWS02 data using SIA. While [Fe II] 25.99 was looked at on all three occasions (see Table 1), [Fe I] 24.05 was only looked at on June 8, 1996. The flux calibration for the SWS02 mode is in general more reliable than for SWS01 at a specific wavelength since dark exposures are taken immediately before and after the short SWS02 scans, and the sampling for the individual data points is better. However, the SWS02 observations in Table 1 cover only the range , which could be just a fraction of the real line width. As the continuum level is therefore unknown, we have to rely on absolute flux calibrations, which are uncertain at the low flux levels of our data. Concentrating on the line, we measure  Jy and  Jy for the two epochs with the longest exposures, May 26, 1996 and July 19, 1997, respectively. The accuracy of absolute fluxes for this wavelength regime at higher flux levels is (ISOSWS Data Users Manual, v5.0), but for our very low flux levels, is probably more fair.

Due to this uncertainty we can only say that our SWS02 flux estimates are consistent with the upper limits we obtain from the SWS01 measurement.

2.2. LWS observations

Our LWS observations were made on 4 February, 1998. The full range of the LWS was scanned, i.e., . Standard reduction of the spectra using the ISO Spectral Analysis Package (ISAP) software resulted in the combined spectrum displayed in Fig. 2. Just as for the SWS observations, no emission was observed from the supernova. However, four unresolved emission lines from neighboring photoexcited gas were identified: [O I] 63.2 , [O III] 51.8 , [O III] 88.4 and [C II] 158 , with the fluxes erg s^{- 1} cm^-2, erg s^{- 1} cm^-2, erg s^{- 1} cm^-2, and erg s^{- 1} cm^-2, respectively. The lines sit on top of a broad continuum which peaks at at a level of  Jy, and which is only marginally weaker in the range . The integrated continuum flux is erg s^{- 1} cm^-2, which for a distance of 50 kpc corresponds to a luminosity of erg. It should be pointed out that the spatial region sampled has a large diameter (), which at a distance of 50 kpc distance corresponds to  pc. We discuss the LWS results briefly in Sect. 4.3.

Fig. 2. ISO LWS spectrum of SN 1987A on day 3999. The continuum up to 140 can be fitted with a spectrum emitted by dust with a temperature of K (solid smooth line). Longward of 140 a second component appears to add in. The temperature characterizing this component is K. The combined emission is shown by the dashed line. See Sects. 2.2 and 4.3 for details.

© European Southern Observatory (ESO) 1999

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