Classification of SNe was first introduced by Minkowski (1941) to distinguish two different spectral appearances of SNe. Type I supernovae (SN I) were those that displayed no evidence of hydrogen and type II supernovae (SN II) stood for those showed spectral evidence of hydrogen in their ejecta.
Figure 1: Classification scheme for supernovae based on the early and
late-time spectra and other features. It is reproduced from Wheeler (1990).
Modern supernova classification, which has been summarized by Harkness & Wheeler (1990), can be clearly seen in the scheme shown in figure 1, which is mainly based on spectroscopic observations near maximum as well as at late times when a SN turns into its nebular phase. Briefly there are three main types. First is type Ia, (e.g. SN 1992A; Kirshner et al. 1993), which is defined by the absence of hydrogen lines in their maximum spectra and presence of a Si II absorption near 6100Å. Type Ib (e.g. SN 1983N and SN 1984L), also lacking hydrogen lines, displays helium lines instead of the Si II feature in type Ia. Then the third is Type II (e.g. SN 1979C, SN 1987A), which show hydrogen lines. There are still diversities upon the above basic scheme. One finds, that the two subtypes Ib and Ic (e.g. SN 1983I, 1987M), which are separated in the capacity of He I lines when they are in early phases, are indistinguishable from each other at late times, both displaying broad forbidden lines of oxygen and calcium. In fact, it is now not even clear whether the subtype Ic is appropriate or whether these objects are just variants of SNe Ib (Leibundgut 1994). A list of spectra of these types is given in figure 2.
Figure 2: A quick view of spectra characteristics of the various
types of supernovae near maximum.
The obvious differences in optical light curves of SNe II make a sub-classification into linear and plateau-like type II supernovae (Barbon et al. 1979). SNe II-L, e.g. SN 1979C and 1980K, have the luminosity declined roughly in an exponential fashion (Doggett & Branch 1985, Young & Branch 1989), while the plateau in light curves of SNe II-P , with an example of SN 1983K, tends to last one or two month, as shown schematically in figure 3. The fact that SN 1987A has a second maximum in its light curve, makes it a distinct object other than SNe II-L or II-P.
Figure 3: Schematic light curves for supernovae of Types Ia, Ib,
II-L,II-P and SN 1987A.
SN 1987K (Filippenko 1988), together with SN 1993J (e.g. Wheeler & Filippenko
1996), which we will discuss in detail later, have been a challenge to the
above standard classification scheme of supernovae. They both showed a strong
H 
P-Cygni feature at around 6300Å, so these events were identified as
SNe II. When they were in the nebular phase, there was no evidence of hydrogen,
whereas both SN II-L and II-P show appreciable evidence of hydrogen in
their late-time spectra. In fact, their late time spectra are very similar
to those of SN Ib or Ic. These objects, which are now defined as SN IIb
(e.g. Filippenko 1988; Nomoto et al. 1993), represent important
links between different classes for the physical understanding of supernovae.
The phenomenological classification of supernovae does not necessarily
reflect their different physical mechanisms. Basically,
Two basic mechanisms have been suggested to account for the large energy
release in supernovae. One kind of explosion often occurs in a dense evolved
stellar core composed of carbon and oxygen and the burning of carbon triggers
the thermal runaway. Usually, in a binary stellar system, when the secondary
star evolves and transfers material, the mass of the white dwarf may grow
until it reaches a critical point, where its core has high electron degeneracy
and thermonuclear burning is thermally and dynamically unstable.
This results a supersonic shock wave (detonation) or a rapid but subsonic
combustion (deflagration) or a complex of the two which may tear the star
completely apart. The white dwarf often have had its hydrogen envelope lost.
The second mechanism of supernova
explosion is the collapse of an iron core, which is often a result of
massive stars ( > 8 M 
). Iron is the element that undergoes endoergic
reactions which lead to a stellar collapse. In the core collapse,
the protons capture electrons and therefore neutrons are formed
with neutrinos emitted. Such an electron-capture process gives out large
amount of energy, 
of which is lost through neutrinos, with only 
is responsible for the following explosion.
A neutron star is often left behind in such a case.
The connection of classification scheme with the physics of the explosions
is not straight forward. SNe II are thought to be a phenomenon of core
collapse in massive stars, say, 
M , while SNe Ia are most
likely related with carbon deflagration of a white dwarf. The appearance of
SNe IIb (e.g. SN 1993J and SN 1987K) link the SN Ib/c with SNe II. The
observations of SN 1993J and SN 1987K indicate that they are the displays
of core collapse rather than explosion in white dwarfs (e.g. SN 1993J;
Wheeler & Filippenko 1996). Thus, the classification
system, which gives us a basic impression, though, may has limits
for it no longer reflects the underlying physical processes, and more and
more supernovae are displaying new and unexpected behavior that does not fit
the previously defined classes (Leibundgut 1994).