Astron. Astrophys. 337, 207-215 (1998)

3. Results

Fig. 2 shows the late light curves of SN 1996N. We have compiled all our photometry here: five observations in R, four in V and three in I. The magnitudes decline linearly and chi-square fits to the magnitudes and their errors gives for the R lightcurve a slope of 1.720.10 magnitudes per 100 days, between 179 days and 337 days past discovery. The V lightcurve has 1.670.23 mag (100d)^-1 whereas I has 1.930.24. Thus, we see no evidence for a strong color dependence in the decline.

Fig. 2. The late light curves of SN 1996N. The lines are chi-square fits to the magnitudes and their errors. Also indicated is the decay rate for 56 Co , i.e., the lightcurve for full trapping of the -rays. The line labeled "-ray leakage" refers to the model explained in the text. The supernova was discovered on March 12.5 UT, which corresponds to JD 2450155.

An early spectrum of SN 1996N was obtained on March 23.4 at the AAT by L. Germany, B. Schmidt, R. Stathakis and H. Johnston (Germany et al. 1996). We have re-reduced and analyzed that spectrum and confirm the classification of SN 1996N as a Type Ib/c. In fact, we identify three rather weak absorption lines as corresponding to He i 5876, 6678 and 7065, all blueshifted by about 9000 km s^-1. This suggests that SN 1996N was of Type Ib. Apart from the helium lines we identify absorption lines of O i 7774 and the Ca ii near-IR triplet.

The spectra of SN 1996N at five late epochs are shown in Fig. 3. The supernova was already entering its nebular phase at the first observation date, 179 days after discovery, and the lines of intermediate mass elements are clearly seen. The strongest lines are [O i] 6300, 6364, [Ca ii] 7291, 7324, the Ca ii near-IR triplet (probably mixed with [C i] 8727), Mg i] 4571 and Na i 5893. There is also O i at 7774 and perhaps [O i] at 5577. The feature at 5300 Å is a mixture of Fe ii emission. Fe ii is also seen in the absorption dips at 5018, 5169.

Fig. 3. The spectral evolution of SN 1996N. The wavelength scale has been corrected for the redshift of NGC 1398. The flux scale refers to the uppermost spectrum, the first date of observation. The tick marks to the right indicate the zero levels for the fluxes. The strongest emission lines have been identified.

All the emission lines are quite broad, [O i] 6300, 6364 shows a total FWHM of 6000 km s^-1 179 days after discovery, the [Ca ii] 7291, 7324 is somewhat narrower with a FWHM of 4600 km s^-1.

A smaller velocity for calcium than for oxygen is consistent with most explosion models, where Ca is produced further in than O. In SN Ib/c 1985F these lines had about the same velocities (Schlegel & Kirshner 1989) and in SN Ic 1994I, [Ca ii] was even broader than [O i] (Filippenko et al. 1995). Some caution is, however, necessary in interpreting the widths of these lines, as they are both blends. We tried to deconvolve these blends using Gaussian line shapes, and our fits indicate that the [O i] lines were indeed broader. The best fits were obtained for a FWHM of 4500 km s^-1 for [Ca ii] and 5200 km s^-1 for [O i], where we used an optically thin line ratio of 3:1.

For comparison, SN Ib/c 1985F had a FWHM of 4700 km s^-1 for [Ca ii] 7291, 7324 at 300 days (Schlegel & Kirshner 1989). Similarly, SN IIb 1993J had 4600 km s^-1, but SN II 1986I had only 2600 km s^-1 at similar epochs (Filippenko et al. 1994). SNe Ic 1987M and 1994I showed even higher velocities, 6200 km s^-1 and 9200 km s^-1 respectively, at about 4.6 months (Filippenko et al. 1995). The higher velocities in SNe Ib/c are usually attributed to their smaller masses. For any given explosion energy, a smaller mass of the ejecta allows for larger velocities.

In Table 5 we present the evolution of the fluxes for the stronger emission lines. These fluxes are rather uncertain, especially for the later spectra, and the errors could be as large as 50. Line ratios for lines with a large wavelength separation might be affected by additional uncertainties, like reddening and slit losses. Most line ratios seem to stay rather constant during this epoch, although the Ca II near IR triplet appears to be getting weaker compared to [Ca ii] 7291, 7324. This was also noticed in SN Ic 1987M, and interpreted as due to decreasing density (Filippenko et al. 1990). Table 5 also presents the FWHM and central wavelengths of the emission lines. These were simply estimated using Gaussian fits. The observed blueshifts will be discussed in Sect. 4.4.

Table 5. Emission lines
Notes:
The positions and FWHM are simply measured by Gaussian fits, the fluxes are, however, integrated over the whole line. Wavelengths are in the rest frame of NGC 1398, which has a recession velocity of 14076 km s^-1 (de Vaucouleurs et al. 1991).

No corrections for absorption have been applied. We have little handle on any absorption towards SN 1996N. We examined the early phase AAT spectrum for any signs of Na absorption at the redshift of NGC 1398, but the spectrum is of low signal and does not constrain the absorption. Our first spectrum from September 1996 has appreciable flux at the expected Na absorption, but it is also of low resolution and even a fairly strong, narrow Na absorption could not be detected. The HI maps do indicate that there is no absorption from our own Galaxy towards SN 1996N (Burstein & Heiles 1984). Even though absence of evidence is not evidence for absence, we nevertheless have not applied any absorption corrections for what follows.

© European Southern Observatory (ESO) 1998

Online publication: August 6, 1998
helpdesk.link@springer.de