2.1. Supernova Types
Morphologically, Supernovae are distinguished into two main classes, Type I and Type II according to the main criterion of whether their spectra (thus, their ejecta) contain Hydrogen (Type II) or no Hydrogen (Type I).
Type II SNe are produced by the core collapse of massive stars, say, more massive than 8 M and at least as massive as 20 M (SN 1987A) or even 30 or more M (SN 1986J). Thus, the lifetime of a SNII progenitor is shorter than about 100 million years (and can be as short as a FEW million years). Therefore, SNII can be found only in galaxies that are either just formed or that have efficient, ongoing star formation, such as spiral and irregular galaxies.
|Optical||Metal lines||Metal lines||P Cyg lines|
|Spectrum||deep 6150 Å||no 6150 Å||Balmer series|
|Absolute||~ 4 × 109 L||~ 109 L||~ 109 L|
|Luminosity||small dispersion||small disp.?||large disp.|
|at max light||standard candles|
|UV spectrum||very weak||weak||strong|
|Emission||fast decay||slow decay|
|Location||all galaxies||spirals||spirals &|
|Progenitors||white dwarfs||moderately||massive stars|
|in binary systems||massive stars|
The class of Type I supernovae has been recognized (e.g., Panagia 1985) to consist of two subclasses, Type Ia and Type Ib/c that, although sharing the common absence of Hydrogen, are widely apart in other properties and, especially, in their origins. The spectroscopic criterion to discern the two subclasses from each other is the presence (Ia) or absence (Ib/c) (c) of a strong Si+ 6150Å absorption feature which is prominent in their early epoch spectra. The astrophysical difference between Type Ia and Ib/c SNe is that the former are found in all type of galaxies, from ellipticals through spirals to irregulars, whereas the latter are found exclusively in spiral galaxies, mostly associated with spiral arms and frequently in the vicinities of large ionized nebulae (giant HII regions). These characteristics indicate that SNIb/c are the end result of a relatively young population of stars (ages less than 100 million years) while SNIa progenitors must be stellar systems that have considerably longer lifetimes, of the order of 109 years or more.
The progenitors of SNIa are believed to be stars that would not produce a SN explosion if they were single stars but that end up exploding because, after reaching the white dwarf stage, they accrete enough mass from a binary companion to exceed the Chandrasekhar mass, and ignite explosive nucleosynthesis in their cores. This process of "nuclear bomb" is expected to disrupt the entire star while synthetizing about 0.6 M (Ia) of radioactive 56Ni, which will power the SN optical light curves. SNIa are very luminous objects and form a quite homogeneous class of SNe, both in their maximum brightness and their time evolution. Thus, SNIa constitute ideal "standard candles" for distance determinations on cosmological scales (see Sect. 2.4).
Type Ib/c, on the other hand, must be significantly more massive because they are only found in spiral galaxies, and often associated with their spiral arms: this suggests progenitor masses in excess of 5M. Therefore, either they represent the upper end of the SNIa class or they are a subclass of core collapse supernovae, possibly massive stars that occur in binary systems and are able to shed most of their outer H-rich layers before undergoing the explosion.
2.2. Radio Properties
A series of papers published over the past 18 years on radio supernovae (RSNe) has established the radio detection and/or radio evolution for 25 objects: 2 Type Ib supernovae, 5 Type Ic supernovae, and 18 Type II supernovae. A much larger list of almost 80 more SNe have low radio upper limits (e.g., Weiler et al. 1986, 1998). A summary of the radio information can be found at: http://rsd-www.nrl.navy.mil/7214/weiler/sne-home.html.
All known RSNe appear to share common properties of: 1) non-thermal synchrotron emission with high brightness temperature; 2) a decrease in absorption with time, resulting in a smooth, rapid turn-on first at shorter wavelengths and later at longer wavelengths; 3) a power-law decline of the flux density with time at each wavelength after maximum flux density (optical depth 1) is reached at that wavelength; and 4) a final, asymptotic approach of spectral index to an optically thin, non-thermal, constant negative value.
The current model for radio supernovae includes acceleration of relativistic electrons and compression of the magnetic field, necessary for synchrotron emission. These processes occur at the SN shock interface with a relatively high-density circumstellar medium (CSM) which has been ionized and heated by the initial UV/X-ray flash Chevalier (1982a, b). This CSM, which is also the source of the initial absorption, is presumed to have been established by a constant mass-loss () rate, constant velocity (w) wind (i.e., r-2) from a red supergiant (RSG) progenitor or a binary companion.
In our extensive study of the radio emission from SNe, several effects have been noted: 1) Type Ia are not radio emitters to the detection limit of the VLA (d) 2) Type Ib/c are radio luminous with steeper spectral indices and a fast turn-on/turn-off, usually peaking at 6 cm near or before optical maximum; and 3) Type II show a range of radio luminosities with flatter spectral indices and a relatively slow turn-on/turn-off. These results lead to the conclusion that most SNII progenitors were RSGs, SNIb/c result from the explosion of more compact stars, members of relatively massive binary systems, and SNIa progenitors had little or no appreciable mass loss before exploding, excluding scenarios that involve binary systems with red giant companions. In some individual cases, it has also been possible to detect thermal hydrogen along the line of sight (Montes, Weiler & Panagia 1997, Chu et al. 1999), to demonstrate binary properties of the stellar system, and to show clumpiness of the circumstellar material (e.g., Weiler, Sramek & Panagia 1990). More speculatively, it may be possible to provide distance estimates to radio supernovae (Weiler et al. 1998).
As an illustration we show that case of SN 1979C that exploded in April 1979 in the spiral galaxy NGC 4321 = M100. This supernova was first detected in the radio in early 1980 (Weiler et al. 1991) and is still bright enough to be accurately measured at different frequencies, thus offering a unique opportunity to do a very thorough study of its radio properties, the nature of the radio emission mechanisms and the late evolution of the SN progenitor. Figure 1 displays the time evolution of SN 1979C radio flux at two frequencies (1.47 and 4.88 GHz). One can recognize the "canonical" properties (non-thermal spectral index, flux peaking at later times for lower frequencies, asymptotic power law decline) that allows one to estimate the circumstellar material distribution, corresponding to a constant velocity pre-SN wind with a mass loss rate of ~ 2 × 10-4 M / year and a probable 20M progenitor. In addition, the almost sinosoidal modulation of the light curves reveals the presence of a 5M binary companion in a slightly elliptical orbit (Weiler et al. 1992). And the marked jump up of the flux about ten years after the explosion (Montes et al.2000) suggests that the progenitor had a rather sudden change in its mass loss rate about 10,000 years before exploding, possibly due to pulsational instability (Bono & Panagia 1999, in preparation).
Figure 1. The 1.47 and 4.88 GHz radio emission of SN 1979C as a function of time.
2.3. Supernova Rates
Determining the rates of SN explosions in galaxies requires knowing how many SNe have exploded in a large number of galaxies over the period of time during which they were monitored. Although it sounds easy, this process is rather tricky because data collected from literature usually do not report the control times over which the searches were conducted. On the other hand, more systematic searches that record all needed information have been started rather recently and the number of events thus recorded is rather limited, so that the statistics is still rather uncertain. In a recent study, Cappellaro et al. (1999) have thoroughly discussed this problem and, from the analysis of all combined data set available, have derived the most reliable SN rates for different types of galaxies. We have taken their rates and, for each galaxy class, we have renormalized them to the appropriate H-band (~ 1.65µm) luminosity rather than the B-band (~ 0.45µm) luminosity as done by Cappellaro et al. (1999). These new rates, displayed in Table 2, are essentially rates per unit galaxy mass because the H-band luminosity of a galaxy is roughly proportional to its mass. We see that SN rates closely reflect the star formation activity of the various classes, not only for type II and Ib/c SNe but also for SNIa. In particular, the rates for SNII-Ib/c are 3-4 times higher in late type spirals (Sbc-d) and irregulars than they are in early type spirals (S0-Sb): this is clear evidence that star formation is considerably more active in the former than it is in the latter group. Also, we notice that late type galaxies (i.e. the ones with most active star formation, Sbc through Irr) have SNIa rates which are 4-10 times higher that the earliest type galaxies (i.e. E-S0). This is a new result (Panagia 1999, in preparation) and implies that SNIa progenitors are intermediate mass stars (say, 8 > M / M > 3) and that early type galaxies are likely to capture and accrete star forming galaxies on a time scale of one to few billion years to replenish their reservoir of SNIa progenitors.
Recent estimates of the global history of star formation in the Universe were used by Madau, Della Valle & Panagia (1998) to compute the theoretical Type Ia and Type II SN rates as a function of cosmic time from the present epoch to high redshifts. They show that accurate measurements of the frequency of SN events already in the range 0 < z < 1, and even more so at higher redshifts, will be valuable probes of the nature of Type Ia progenitors and the evolution of the stellar birthrate in the Universe.
|Galaxy Type||SNIa||SNIb/c||SNII||All SNe|
|E-S0||0.05 ± 0.02||< 0.01||< 0.02||0.05 ± 0.02|
|S0a-Sb||0.10 ± 0.04||0.06 ± 0.03||0.24 ± 0.111||0.40 ± 0.12|
|Sbc-Sd||0.21 ± 0.08||0.14 ± 0.07||0.86 ± 0.35||1.21 ± 0.37|
|Sm, Irr||0.59 ± 0.24||0.33 ± 0.24||0.97 ± 0.60||1.87 ± 0.67|
2.4. Cosmological Applications
As mentioned before, SNIa are virtually ideal standard candles (e.g., Hamuy et al. 1996) to measure distances of truly distant galaxies, currently up to redshift around 1 and, considerably more in the foreseeable future (for a review, see Macchetto and Panagia 1999). In particular, Hubble Space Telescope observations of Cepheids in parent galaxies of SNe Ia (an international project lead by Allan Sandage) have lead to very accurate determinations of their distances and the absolute magnitudes of SNIa at maximum light, i.e. MB = - 19.50 ± 0.06 and MV = - 19.49 ± 0.06 (e.g., Sandage et al. 1996, Saha et al. 1999). Using these calibrations it is possible to determine the distances of much more distant SNe Ia. A direct comparison with the Hubble diagram (i.e. a plot of the observed magnitudes of SNIa versus their cosmological velocities) of distant SNe Ia (30, 000 km s-1 > v > 3, 000 km s-1) gives a Hubble constant (i.e. the expansion rate of the local Universe) of H0 = 60 ± 6 km s-1 Mpc-1 (Saha et al. 1999). Studying more distant SNIa (i.e. z > 0.1) it has benn possible to extend our knowledge to other cosmological parameters. The preliminary results of two competing teams (Riess et al. 1998, Perlmutter et al. 1999) agree in indicating a non-empty inflationary Universe with parameters lying along the line 0.8 - 0.6 = - 0.2 ± 0.1. Correspondingly, the age of the Universe can be bracketed within the interval 12.3-15.3 Gyrs to a 99.7% confidence level (Perlmutter et al. 1999).
c They are classified Ib if strong He lines are present in their spectra, and Ic otherwise. Back.
d The VLA is operated by the NRAO of the AUI under a cooperative agreement with the NSF. Back.