1998 - A Good Year for Supernovae

Peter Meikle

Blackett Laboratory, Imperial College, London

Tom Geballe

Head of UKIRT Operations, Joint Astronomy Centre, Hilo

On March 3, Zhou Wan (Beijing Astronomical Observatory) discovered the peculiar type II Supernova 1998S in NGC 3877 (IAU Circ. 6829). The discovery was about 10 days before maximum light and it peaked at about V=+12 making this the brightest type II supernova in the past 15 years observable with UKIRT.

On April 13, Mark Armstrong (Rolvenden, UK) discovered the type Ia Supernova 1998aq in NGC 3982 (IAU Circ. 6875). The discovery was about 3 weeks before maximum light, and it peaked at about V=+12.3. This was the brightest thermonuclear supernova for four years (although it was soon to be superceded by SN 1998bu).

On April 28, T.J. Galama et al. discovered the peculiar type Ic Supernova 1998bw in ESO 184-G82 (IAU Circ. 6895). It was too far south to be observable at UKIRT. The remarkable thing about this supernova is that its discovery was prompted by the observation and location of a gamma-ray burst. After some confusion, the supernova was provisionally classified as a peculiar type Ic. Thus, at last, we appear to have found a connection between some gamma-ray bursts and some supernovae. How a supernova might produce a gamma-ray burst is presently undergoing intensive investigation.

On May 9, Mirko Villi (Forli, Italy) discovered the type Ia Supernova 1998bu in M96 (NGC 3386) (IAU Circ. 6899). The discovery was about two weeks before maximum light and it peaked at about V=+12. Moreover, its host galaxy's distance had already been measured by the HST using Cepheids. Its distance of 11.6 Mpc makes it the nearest thermonuclear supernova ever that has been accessible to UKIRT.

As each accessible supernova was discovered, we began observational programmes involving several telescopes, with UKIRT taking the major role in acquiring infrared data, especially CGS4 spectra. Of course, most of the UKIRT time had already been assigned to specific observers. However, as is usually the case, the scheduled observers were enthusiastic and generous about giving up a little bit of their time to get observations of such remarkable events. The UKIRT Service facility also made a major contribution to this programme. This article is being written at the end of the first season of observations, and much of the data have still to be reduced and analysed. Here, we show some of the beautiful and unique supernova spectra which have been obtained with CGS4 over the  last few months. We shall focus mainly on SN 1998S, with a few additional comments about SN 1998bu.
 
 

FIGURE 1 : Evolution of J-band CGS4 spectra of the peculiar type II SN 1998S. The spectra have been displaced vertically for clarity. The supernova exploded on about March 1, 1998. The spectra on April 16 and 27 are composites of low resolution spectra plus high resolution sections covering the HeI 10830A region.

SN 1998S

FIGURE 2 : High resolution (60 km/s) CGS4 spectra of the peculiar type II SN 1998S. The April 16 spectrum has been displaced vertically for clarity. Its zero flux level is shown by the dotted line on the left hand ordinate. The supernova exploded on about March 1, 1998. The broad emission line is identified with a blend of HeI 10830A and Paschen-gamma emission, and is produced by the rapidly moving ejecta. The narrow P Cygni line is identified with HeI 10830A and is probably caused by radiation from the supernova shock ionising the slow-moving, undisturbed progenitor wind just ahead of the shock.

What is going on in SN 1998S? Salamanca et al. (1998 MNRAS, in press) have reported and discussed similarly narrow P Cygni Balmer lines in SN 1997ab. The narrow P Cygni line tells us a great deal. We must be observing the supernova through a cloud of slow-moving circumstellar material (CSM). The wavelength of the peak, after correction for the redshift of the parent galaxy, leaves us in no doubt that the narrow line is due to HeI 10830Å. This is also quite remarkable, as it is a very high excitation recombination line. What could be driving it? It seems unlikely that it is due to the initial flash-ionisation of the CSM since the recombination time is probably shorter than the observation epochs of several weeks. We believe, instead, that there must be a continuous source of ionisation. The narrowness and P Cygni nature of the line suggests the answer. The line is surely being generated in a dense, slow-moving circumstellar wind released by the progenitor before the explosion. When SN 1998S exploded, the ejecta shock ploughed into the slow-moving wind, causing violent shock ionisation. This would produce EUV and X-ray photons which would move out into the undisturbed CSM and ionise it. The subsequent recombination populates the highly metastable 2s3S level of HeI, which is the lower level of the 10830Å transition. The 2s3S population absorbs radiation from the supernova photosphere, generating the P Cygni profile. Our optical observations show similarly narrow P Cygni lines in Balmer-alpha.

While over 1350 supernovae have now been observed their distances are such that we are unable to retrospectively observe the progenitor directly. (The only significant exception to this was SN 1987A). Yet the progenitor is the starting point in our efforts to understand the supernova phenomenon. However it now seems that, for a few supernovae, nature occasionally provides us with direct clues about the nature of the progenitor. SN 1998S is a beautiful example of this, and detailed analysis of the early time infrared and optical spectra will allow us to constrain the progenitor type for this event.
 
 

FIGURE 3 : CGS4 spectra of SN 1998S at about 115 days post-explosion, compared with spectra of SN 1987A at 110/2 and 192 days. The SN 1987A spectra were obtained at the Anglo-Australian Telescope (Miekle et al. 1989, MNRAS, 238, 193). The spectra have been scaled and verically displaced for clarity. Note the intense Paschen-alpha and first overtone carbon monoxide emission (longward of 22800A) in SN 1998S.

The spectroscopic monitoring of SN 1998S continued throughout the first season, and Figure 3 shows CGS4 spectra taken (by UKIRT Service) at about 115 days post-explosion, just before the target became inaccessible. It is compared with spectra of SN 1987A (at 110/2 and 192 days) taken at the Anglo-Australian Telescope (Meikle et al 1989, MNRAS 238, 193). We can see SN 1998S was somewhat more evolved than was SN 1987A at the same epoch. In particular we can see that the first overtone of CO is already strong in SN 1998S while the broad HeI 10830Å trough is quite weak. SN 1998S is the third supernova for which CO emission has been detected. CO is a powerful coolant in the ejecta and its presence and distribution has a significant influence on the evolution of the supernova, including the condensation of dust. The wider coverage of CGS4 also reveals a very strong, broad Paschen-a line, not covered in the AAT observations of SN 1987A.

Second season CGS4 observations will focus on improving our understanding of the core-collapse explosion mechanism by measuring the extent and nature of mixing in the ejecta and the mass of radioactive nickel released. This will complement our early-time studies of dredge-up of radioactive materials (Fassia et al, 1998, MNRAS in press).
 

SN 1998bu

FIGURE 4 : CGS4 spectrum of the type Ia SN 1998bu at about one month after maximum light. The steep fall in the spectrum between 10000 and 12000A is increasingly thought to be highly characteristic of thermonuclear supernovae. The short wavelength section shows the first ever spectrum obtained using the new I-band coverage of CGS4.

Unfortunately, for the first five weeks after SN 1998bu exploded, CGS4 was off the telescope. However, in late June we acquired (by UKIRT Service) a beautiful spectrum when the supernova was about 55 days old. This is shown in Figure 4. Of particular interest is that this was the first spectrum taken with the new I-band facility. This means we can now obtain CGS4 spectra down to 8500Å, thus at last closing the gap between the optical and infrared. The spectrum shows the well-known calcium triplet feature in the 8500 to 9000Å region. Many of the other features are due to sulphur, iron and cobalt. We also see that, as with the few other type Ia’s we have observed in the IR, SN 1998bu exhibits a dramatic drop in flux in the 10000-12000Å region. It is becoming clear that this deficit, thought to be due to a lack of emission lines in the region, is a powerful identifying feature of thermonuclear supernovae.

Second season CGS4 observations will be used to measure the mass and distribution of iron-group elements ejected and hence constrain thermonuclear explosion mechanisms.

Our thanks go to the dozens of observers who have so enthusiastically gathered data on these supernovae. To a large extent it is through their generosity that observational studies of supernovae are possible.
 
 

Back to Contents