Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.eso.org/~bleibund/papers/ASI450.ps Äàòà èçìåíåíèÿ: Sun Jan 7 21:58:56 2001 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 02:15:16 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: photosphere

OBSERVATIONS OF SUPERNOVAE 1
BRUNO LEIBUNDGUT
University of California
Astronomy Department
Berkeley, CA 94720
and
European Southern Observatory
Karl­Schwarzschild­Strasse 2
D­85748 Garching
Germany
Abstract.
The available observations of the various types of supernovae are described and their connection
with the current theoretical understanding outlined. We critically discuss the classification scheme
employed for supernovae and point out its limits.
The observational homogeneity of type Ia supernovae has been challenged by recent, well­
observed events. Nevertheless, the general characteristics are rather well described by models
employing explosive carbon burning in white dwarfs. The most important input which observations
should provide at this point is the mass of the progenitor star and the amount of radioactive nickel
produced in the explosion. With reliable absolute magnitudes and improved model calculations
becoming available such determinations appear in reach.
While the separate character of SNe Ib/c is now firmly established, too few objects have been
observed detailed enough to probe the physical nature of these supernovae. The subclassification
into SNe Ib and SNe Ic has not convincingly been proven. Paradoxically, the most valuable infor­
mation on these supernovae has been provided by the observations of SNe II which switched to
SNe Ib/c at late phases confirming the link between these two classes.
Observations of SNe II have revealed several distinct phases which are governed by different
radiation mechanisms. Many details have been learned from SN 1987A and applied to other SNe II.
The influences of size and mass of the progenitor star as well as the explosion energy on the evolution
of the luminosity have been established. The additional powering of SNe II by conversion of kinetic
energy into electromagnetic radiation in a dense circumstellar medium has revealed a new phase for
these objects which starts several years after the explosion. Few cases have been observed where
this energy source appears to be dominant at all times.
Key words: Supernovae, Formation of Neutron Stars
1. Introduction
The astrophysical significance of supernovae (SNe) is reflected in the many views
astronomers have of them. These explosions mark the end of the nuclear energy
production in massive (?10 M fi ) stars and the start of the life as a neutron star,
or are the disruption of a cooling white dwarf forced into explosive carbon burn­
ing. This for the star catastrophic event also affects the circum­ and the interstellar
medium through ionization and shock heating. The morphology of galaxies is influ­
enced by the deposited energy, the chemistry changed by enrichment in metals, and
1 To appear in The Lives of Neutron Stars, eds. J. van Paradijs and A. M. Alpar, (Dordrecht:
Kluwer)

2 BRUNO LEIBUNDGUT
stellar material recycled. The nucleosynthesis of all elements heavier than oxygen is
provided by supernovae. Finally, supernovae have been employed as light beacons
for cosmological distance measurements.
Observations of supernovae hence bear on many astrophysical issues, like stellar
evolution, stellar mass loss, collapse and explosion physics, radiative hydrodynamics,
shock physics, galactic structure and chemical evolution, and cosmology. There have
been many excellent reviews in recent years. General overviews on supernovae are
provided in the books by Petschek (1990), Wheeler et al. (1990), Woosley (1991),
Danziger & Kj¨ar (1991), and McCray (1994). Woosley and Weaver (1986) have
presented the most complete description of the explosion physics of supernovae.
Reviews dealing with more specific topics include Wheeler & Harkness (1990) on
SNe I, Weiler & Sramek (1988) for radio supernovae, Arnett et al. (1989) and
McCray (1993) for SN 1987A, van den Bergh & Tammann (1991) on supernova rates,
Bartel (1985), Branch & Tammann (1992), and Schmidt (1993) for supernovae as
distance indicators.
The observational tools available for the investigation of supernovae are also very
diverse. Optical photometry and spectroscopy have contributed most information
during the first decades of supernova studies. These data are now supplemented
with observations at other wavelengths and, in the case of SN 1987A, even a neutrino
detection. The different techniques provide insight into the various processes that
occur in the explosion. The evolution of the spectrum (optical and near­IR) carries
information on the chemical composition of the material taking part in the explosion.
Some of this matter is not burned but merely accelerated, such as the envelope of
a massive star after the core collapse. In other cases, the spectrum is composed of
emission from the ashes of matter processed in the explosion. The line velocities can
be used as diagnostic of the explosion energies. Optical photometry of supernovae
provides insights into the overall energy radiated from the explosion and can disclose
the power source. It is also sensitive to the energy source, the mass of the envelope,
and the size of the progenitor star. Radio observations probe the circumstellar
environment around the supernova and can provide important information on the
newly­born pulsar through the evolution of the spectral index. X­ray observations
have some bearing on the radioactive source powering the emission and indicate
interaction with circumstellar material. Finally, fl­rays are direct measurements
of the radioactive power source, the mass of the envelope in which the radioactive
material is buried, and mixing of the stellar material with the ashes of the explosion.
Additional information has been gathered from supernova statistics and the mor­
phology of, as well as the supernova position in, the parent galaxy. Within the limits
of the spatial resolutions attained it is possible to derive parameters of the class of
progenitor stars for different supernova types. Most recently the direct observations
of the progenitors of SN 1987A and, possibly, SN 1993J have provided very impor­
tant clues to what kind of stars end their lives in these explosive events. These
observations provide much needed constraints on the explosions themselves, not to
mention the surprises they represented.
Next I will describe the current classification scheme for supernovae and some
secondary signatures of the various types (x2). The observational status of each
of the different classes will be outlined separately. The evidence of white dwarfs

OBSERVATIONS OF SUPERNOVAE 3
as progenitors of SNe Ia is presented in x3. In section 4, a brief status on the
observations of SNe Ib and SNe Ic is given. The discussion of SN 1993J then leads
to the observations of SNe II (x4) and the link between the latter two classes. The
summary (x5) discusses open questions.
2. Classification
Classification of supernovae should be based on a simple observational signature to
distinguish different types. The physical implications of such a scheme can, however,
be misleading and might not be too profound. Moreover, it is important to establish
various examples of similar behavior, before a subclass can be created. Another
aspect is to keep the scheme simple and not to involve too many observationally
distinct methods for the definition of a class. The more complete information that
is sought for the understanding of the individual objects or a closely defined group
is stepping beyond any means of the classification scheme. In the above sense,
classification of supernovae is a crude observational instrument for the separation of
individual object groups.
Classification of supernovae was introduced by Minkowski (1941) to distinguish
the two different spectral appearances of supernovae. The two classes were identi­
fied with well­observed supernovae, namely SN 1937C for Type I and SN 1940B for
Type II. The latter had suspected emission of Hff near maximum and later ``resem­
bles'' normal novae. The spectra of SNe I were not well understood but recognizable
by their ``very wide emission bands.'' The classification system was further refined
by Zwicky (1965) with three additional classes. Since these three classes had only
one known representative each, they were dismissed by Minkowski (1964) and Oke
& Searle (1973).
Modern supernova classification has been summarized by Harkness & Wheeler
(1990). It is based on spectroscopic observations near maximum light. Three main
types are distinguished (cf. Figure 1): type Ia, defined by the absence of hydrogen
lines in their spectra and the presence of a Si II absorption near 6100 š A, type Ib
and Ic both lacking H lines and the Si feature but displaying, in the case of SNe Ic
very weak, He lines, and type II which encompass all SNe with hydrogen lines. The
spectral evolution of SNe Ia produces spectra a few weeks past maximum which
resemble SNe Ib near peak light. This has caused confusion in a few cases and
points to the need of careful analysis of the spectra for classification. A separation
between SNe Ib, helium­rich, and SNe Ic, helium­poor, is proposed (Wheeler et al.
1987, Wheeler & Swartz 1994). Whether there is a continuous variation from one
subclass to the other or a clear dichotomy of objects is still undecided.
Secondary indicators for the classification schemes are becoming more apparent
with more data and better theoretical understanding of the explosions and their
stellar progenitors. So are the spectra several months past maximum rather homo­
geneous for SNe Ia with many broad lines of Fe­peak elements, while SNe II display
distinctive emission lines of Hff, [O I] and [O III]. SNe Ib and SNe Ic, with broad
forbidden lines of oxygen and calcium, are indistinguishable from each other at this
epoch. There have been, however, intriguing cases of SNe II changing to what nor­
mally are typical spectra of SNe Ib/c at this phase. The two well­documented cases

4 BRUNO LEIBUNDGUT
4000
6000
8000
l
(Angstrom)
log
F
l
+
const.
Na
I?
He
I
Fe
II
Fe
II
Ca
II
Type
Ic
(SN
1990B)
Si
II
S
II
Si
II
Si
III
Mg
II
Ca
II
Type
Ia
(SN
1990O) Ca
II
Ha
He
I
Na
I
Hb
Type
II
(SN
1991H)
Fig. 1. Sample spectra near maximum for the main supernova types
are SN 1987K (Filippenko 1988) and SN 1993J (e.g. Wheeler & Filippenko 1994).
These objects represent important links between different classes for the physical
understanding of supernovae.
While the optical light curves of SNe II have been rather distinctive and prompted
a subclassification into linear and plateau­like light curves (Barbon et al. 1979),
SNe Ia and SNe Ib/c display nearly indistinguishable light curve shapes (Ensman
& Woosley 1988, Leibundgut 1988). Although the colors are much redder in the
case of SNe Ib/c, this selection criterion is spoiled by the often unknown extinction
toward the objects. The shape of the near­infrared (JHK) light curves appears to
separate the two subclasses of SNe I (Elias et al. 1985), but some peculiar SNe Ia
have shown intermediate shapes (Frogel et al. 1987).
Radio emission from supernovae is normally associated with circumstellar mate­
rial (Chevalier 1990). No SN Ia has been observed at these wavelengths. A clear sep­
aration between SNe Ib/c and SNe II has been established (Van Dyk et al. 1993a).
There appears to be a subgroup of SNe II with clearly distinct radio light curve
shapes (Weiler et al. 1990, Van Dyk et al. 1993b).
The connection of the classification scheme with the physics of the explosions is
not straightforward. As described below, SNe II are thought to be displays of core

OBSERVATIONS OF SUPERNOVAE 5
4
2
0
Dm
B
SN
1972E
SN
1980N
SN
1981B
SN
1981D
0
50
100
150
4
2
0
t
0
(days
past
B
maximum)
Dm
H
Fig. 2. Optical and near­infrared light curves of SNe Ia (from Leibundgut 1988)
collapses in massive stars, while SNe Ia are more likely the explosive burning of a
white dwarf. The observations of SN 1993J and SN 1987K link the SNe Ib/c with
SNe II and indicate that the former are also core collapses rather than variations of
explosions in white dwarfs, despite the similarities observed in the optical spectra
and the photometry. These cases demonstrate the limitations of the classification
scheme employed for supernovae for the underlying physics.
3. Type Ia Supernovae
Historically these supernovae have been described as extremely uniform in their
appearance. This has been based on the light curve shapes, the observed colors, the
peak luminosities, and the spectral evolution. In the last few years this homogeneity
has cracked under new investigations and peculiar events of the class.
The light curve shapes of SNe Ia are remarkably similar in the optical and near­
infrared broad­band photometry (UBVJHK; Leibundgut 1988). Many events adhere
to a unique light curve shape (Leibundgut et al. 1991b). Peculiar objects, like
SN 1939B (e.g. Minkowski 1964) and SN 1986G (Phillips et al. 1987, Frogel et al.
1987), even though exhibiting clearly discrepant light curves, had been disregarded.
Recently, Phillips (1993a) pointed out that also the optical light curves are not as
uniform as assumed. The decline during the first 15 days after maximum ranges
from 1.5 mag to 2.5 mag for a set of nine supernovae, an effect confirmed in an
independent analysis by Vacca & Leibundgut (1994) for a much larger sample and a
more detailed fitting procedure. Furthermore, Suntzeff (1994) has shown that the R

6 BRUNO LEIBUNDGUT
and I light curves of SNe Ia do not follow a unique shape, but rather are individual
for most SNe Ia. The colors of SNe Ia display some variation, but the scatter
is remarkably small (e.g. oe(B­V)=0.22 mag for 24 supernovae, Leibundgut 1991)
considering extinction which probably affects some of the supernovae. The value of
the intrinsic color (e.g. (B­V) 0 ) is still a matter of debate (Pskovskii 1968, Cadonau
et al. 1985, Capaccioli et al. 1990, Hamuy et al. 1991, Sandage & Tammann 1993)
with the more recent determinations favoring a value of (B­V) 0 ú 0:0 mag.
The spectra of SNe Ia are dominated by many lines of intermediate­mass elements
like Si, S, Mg, and Ca near maximum light (Branch 1990). With more accurate ob­
servations in recent years it has become obvious that significant variations exist in
the line strengths among individual objects near maximum (e.g. Filippenko et al.
1992a, b, Phillips et al. 1992, Leibundgut et al. 1993b, Phillips 1993b, Branch et
al. 1993). Figure 3 illustrates this point. About a month later the continuum emis­
sion disappears and is replaced by broad emission lines of many iron­peak elements
(Axelrod 1980, Pinto 1988). At about seven weeks past peak light the spectra ap­
pear virtually identical (Filippenko et al. 1992b, Leibundgut et al. 1993b). This
is a strong indication that the ashes are very similar and the explosion mechanism
rather unique.
A ``standard'' model was devised to explain most of the above observations and
also account for the fact that SNe Ia were observed in elliptical galaxies and in the
halos of spiral galaxies, which are presumably composed of old stars. In this model
(cf. Woosley & Weaver 1986, Wheeler & Harkness 1990 for reviews) a white dwarf
is incinerated by explosive carbon burning. The triggering of the explosion takes
place when the white dwarf is nudged over the Chandrasekhar mass (¸1.4 M fi )
and the density and temperature in the center of the star become high enough to
start the burning. The optical display of the supernova is then entirely due to
heating from radioactive decays of 56 Ni and 56 Co nuclei. This model has been very
successful in describing most of the available observations. With the basically unique
configuration of the precursor system a satisfactory homogeneity in appearance is
guaranteed. The fact that only ``processed'' material is observed in the spectra
is explained by the burning front converting most material into nuclear statistical
equilibrium, i.e. iron­peak elements. The tracers of intermediate­mass elements
observed in the early­time spectra are due to the dying out of the subsonic flame
and incomplete burning near the surface of the white dwarf. The differences in the
maximum­light spectra can be explained by small variations as the flame peters out
and might arise from only tiny amounts of matter. At late times, when the explosion
becomes transparent, we are truly looking at the ashes. No circumstellar material
is expected around a white dwarf and thus the lack of radio detections (Weiler et
al. 1986) is accounted for naturally. As end­products of stellar evolution white
dwarfs represent an old population which concurs with their non­association with
star forming regions (Maza & van den Bergh 1976).
The above model is further supported by the agreement reached from spectral
modeling at early and late epochs (Harkness 1991a, b, Jeffery et al. 1992, Mazzali
et al. 1993, Axelrod 1980, Pinto 1988). The fast decay of the late­time bolometric
light curve is readily explained by the escape of fl­rays from the ejecta, pointing to
a small mass (Leibundgut & Pinto 1992). Detailed calculations of the photospheric

OBSERVATIONS OF SUPERNOVAE 7
3000
4000
5000
6000
7000
8000
9000
l
(Angstrom)
log
F
l
+
const
SN
1991bg
(+1
days)
SN
1991T
(­5
days)
SN
1989B
(0
days)
SN
1986G
(0
days) SN
1981B
(­3
day)
Si
II
Fe
III
Fe
III
Mg
II
Si
III
Ca
II
Ca
II
Mg
II
Si
II
S
II
Si
II
Si
II
Fig. 3. Spectra of SNe Ia near maximum (from Leibundgut et al. 1993b)
peak phase of supernova light curves have recently succeeded in reproducing the
observations (Khokhlov et al. 1993, H¨oflich et al. 1993). A crucial parameter is
the rise time from explosion to maximum light and certain models now can match
the observations (Leibundgut 1994a). The most direct evidence of the radioactive
power source, however, comes from a study of Co and Fe lines several weeks past
maximum (Kuchner et al. 1994). The line ratios follow exactly the predictions of
the decay times determined by nuclear physics.
Although the ``standard'' model can explain many aspects of SNe Ia, it has its
shortfalls. Since the white dwarf has to accrete matter, there must be a companion
star. The lack of radio emission indicates that not much matter is lost by this star,
which excludes giant star companions. Any hydrogen and helium accreted onto the

8 BRUNO LEIBUNDGUT
white dwarf has to be processed quickly enough or it would show up in early time
spectra. Unfortunately, there are no good limits on the H mass on top of a SN Ia from
the observations. Double­degenerate systems, in which two degenerate white dwarfs
merge have been adopted instead as the canonical model (Iben & Tutukov 1984,
Webbink 1984, Tutukov et al. 1992). The dynamical time scales and the formation
of a disk in the merger are not easily explained in these models. Furthermore, there
have been no candidate systems observed despite extensive searches (Bragaglia et al.
1990, Renzini 1994). Another problem with the Chandrasekhar­mass white dwarfs
is the observed mass distribution of white dwarfs (Bergeron et al. 1992). There is
only one white dwarf with a mass above 1M fi , which probably is already a merged
object. Thus, at least 0.4M fi would have to be accreted to build a Chandrasekhar­
mass white dwarf.
Other problems include the IR light curves which display conspicuous minima
about a week after the optical maxima (Fig. 2). Similar dips are also observed in
R and I light curves (Ford et al. 1993, Hamuy et al. 1993, Suntzeff 1994). The
reason for this flux redistribution has not yet been addressed in model calculations.
Direct determinations of the Ni mass from late­time spectra also point to rather
large differences in the Ni production contrary to the predictions from models (Ruiz­
Lapuente & Lucy 1992). Large variations in the expansion velocities as measured
from Doppler displacement of absorption lines (Branch et al. 1988, Branch & van
den Bergh 1993) indicate variations in the explosion energies.
The most severe challenge for the ``standard'' model arouse recently from detailed
observations of SN 1991T (Filippenko et al. 1992a, Ruiz­Lapuente et al. 1992,
Phillips et al. 1992) and SN 1991bg (Filippenko et al. 1992b, Leibundgut et al.
1993b). Both objects displayed differences in the light curve shapes and spectral
peculiarities. In the case of SN 1991bg there are very severe discrepancies in the
late time spectrum as well (Filippenko et al. 1992b, Leibundgut et al. 1993b, Ruiz­
Lapuente et al. 1993). Such variations cannot be explained within the standard
framework of the explosion of a Chandrasekhar­mass white dwarf.
An important conclusion from the observations is that most of the material has
been burned in the explosion and hydrogen and helium are absent, the radioactive
power source appears now well established, and a low ejecta mass is deduced from the
decline in the late­time light curve. All this points toward a white dwarf progenitor
for SNe Ia, but none of the above signatures provides a quantitative measure of
the total mass or the amount of radioactive 56 Ni synthesized. This has prompted
models which employ small­mass white dwarfs in the explosions (Shigeyama et al.
1992, Woosley & Weaver 1994). With the physics available for off­center explosions
(Livne 1990, Livne & Glasner 1991, Woosley & Weaver 1994) it might be possible
to explain simultaneously the apparent homogeneity of the bulk of the objects and
the deviations of a few supernovae from this norm.
A frequent application of SNe Ia has been the determination of cosmological
distances and the derivation of the Hubble constant H 0 (Branch & Tammann 1992,
and references therein). To achieve meaningful results two conditions have to be met.
First, the narrow range of peak luminosities of SNe Ia has to be established reliably,
and, second, the absolute value of this peak luminosity has to be determined.
Establishing the standard candle character of supernovae primarily means that

OBSERVATIONS OF SUPERNOVAE 9
they reach the same luminosity at maximum. Further indications of the homogeneity
of the class in its appearance provide additional arguments, but are not essential for
the use as distance indicators. Several groups have measured the observed luminosity
distribution of SNe Ia at maximum (Tammann & Leibundgut 1990, Della Valle &
Panagia 1992, Branch & Miller 1993, Sandage & Tammann 1993) and agree fairly
well with each other. There is little doubt that by appropriate treatment of ``outliers''
a tight distribution around a mean absolute magnitude can be established. There
has been some concern about difficult it is to sort out the ``peculiar'' supernovae,
but Branch et al. (1993) argue that the spectroscopic signatures are clear enough
for a separation. For many SNe Ia observed in the 1960's and 1970's, however, not
enough data are available to assess their status and most of the above derivations
rely on data samples which include many of those SNe. An additional complication
is the often unknown extinction toward SNe Ia. The effect of this has been discussed
in Leibundgut & Pinto (1992). Unfortunately, errors of up to 40% can be incurred in
the determination of the mean of the distribution of the peak luminosity of SNe Ia.
The most reliable measurement of the absolute luminosity of a SN Ia is by Saha
et al. (1994) for SN 1937C. They have measured cepheids within IC 4182 and
established the distance to this galaxy. A different route is the calculation of the
luminosity from models pioneered by Arnett et al. (1985). Branch (1992) has
gathered what he considered the best determinations of this kind. An independent
investigation involving a variety of models has been presented by Leibundgut & Pinto
(1992). They conclude that as long as we are not able to select the ``best'' model,
a large uncertainty has to be accepted. All models considered, however, employed
the explosion of Chandrasekhar­mass white dwarfs. With smaller mass explosions
the luminosities are decreased. A slightly different approach was used by M¨uller
& H¨oflich (1994) who find a suitable model for each individual supernova and thus
individual luminosities to determine H 0 . No assumption on the standard character
of SNe Ia is made.
No value of H 0 is quoted here and the reader referred to the literature. Discussion
of the various error sources is desperately needed to solve some of the outstanding
problems.
4. Type Ib/c Supernovae
Only recognized about a decade ago as a separate subclass among SNe I (Wheeler
& Levreault 1985, Uomoto & Kirshner 1985, Panagia 1985) this group of objects
might represent the ``cross­dressing'' case of supernovae. Around maximum they
resemble SNe Ia in their photometric and spectroscopic appearance, but their late­
time spectra and radio signatures are related to SNe II. With two examples of SNe II
which actually have ``changed type'' (SN 1987K and SN 1993J) the latter connection
has been further strengthened.
The observational data for this class are rather scanty. The general characteristics
have been summarized by Porter & Filippenko (1987). The question whether there
is the need for a further subclassification remains open. Swartz et al. (1993) argue
from model calculations for a clear distinction between SNe Ib (He­rich) and SNe Ic
(He­poor) on the basis of the He mass observed in the spectra. In addition, Wheeler

10 BRUNO LEIBUNDGUT
2
0
SN
1993J
SN
1962L
SN
1984I
Dm
U
4
2
0
Dm
B
­20
0
20
40
60
80
100
4
2
0
Dm
V
t
0
(days)
Fig. 4. Optical light curves of the type Ib/c SN 1962L (Bertola 1964) and SN 1984I (Leibundgut
et al. 1990) compared with SN 1993J (Richmond et al. 1994); the line represents a mean curve for
SNe Ib/c from Leibundgut (1988)
& Swartz (1994) find SNe Ib of higher mass than SNe Ic.
The spectroscopic evolution is dominated by He lines for the Ib supernovae while
in SNe Ic absorption lines of predominantly C, O, Ca, and Fe are observed (Wheeler
et al. 1987). Faint Balmer lines of H are suspected, but not a clear signal has been
detected (Jeffery et al. 1991, Filippenko 1992). After about 6 months the nebular
spectra show strong Ca and O emission lines (Filippenko et al. 1986, Filippenko et
al. 1989).
The optical light curves are very similar to the ones of SNe Ia (Ensman & Woosley
1988, Leibundgut 1988). The infrared light curves, however, do not exhibit the
second maximum and are rather distinct (Elias et al. 1985).
If SNe Ib/c emerge from massive stars and are caused by a core collapse, as is
suggested by their connection with SNe II, the light curves are expected to display a
sharp early peak from the shock break out at the surface of the star. This has never
been observed in a SN Ib/c proper, but SN 1993J had been detected early enough
(Figure 4). It is clear from this figure that the light curve shapes of SN 1993J are
indeed very similar to the ones of SNe Ib/c. A similar conclusion has been found
for the bolometric light curve of SN 1993J and SN 1983N (Nomoto et al. 1993,

OBSERVATIONS OF SUPERNOVAE 11
Richmond et al. 1994). For the comparison in Fig. 4 the curves have been shifted to
coincide at the maximum of the broad peak. The peak for SN 1962L and SN 1984I
is evidently broader in all bands and the slope of the decline 30 days past maximum
steeper. The occurrence of the maximum in U seems also shifted to later times,
although the data are really not sufficient for a clear result. The first B observations
of SN 1984I and SN 1962L, an estimate of a non­standard band photograph however,
are very interesting as they lie at the bottom of the dip in the SN 1993J light curve.
Both these supernovae might have been caught shortly after the the initial peak of
shock breakout. A duration of a few days is not excluded by the observations; an
upper limit about 2 mag fainter than the B peak brightness was determined 14 days
before the first detection of SN 1984I (Leibundgut et al. 1990). The broader peak of
the SN Ib/c indicates a slower release of the energy diffusing out of the core than for
SN 1993J. The steeper slope at late times on the other hand points toward a faster
increase in fl­ray transparency of the envelope. With similar explosion energies this
implies a smaller mass in the envelopes of SN 1962L and SN 1984I than for SN 1993J.
SNe Ib/c are not clearly stronger associated with star forming regions (Van Dyk
1992). It appears as if they are not the very massive stars (? 30 M fi ) which were
suspected in the beginning (e.g. Ensman & Woosley 1988, Schaefer et al. 1987).
A model where the star has lost its hydrogen envelope due to binary interaction
appears more likely (Shigeyama et al. 1990) and is further supported by the apparent
connection of SN 1993J with SNe Ib/c. Additional support for such a model, as
opposed to one with white dwarfs, is the radio emission from these objects (Van Dyk
et al. 1993a). If the connection to SNe II is confirmed, this means that SNe Ib/c
contribute to the population of neutron stars.
5. Type II Supernovae
This class of supernovae comprises a great many individual examples of objects
displaying hydrogen in their spectra. The additional subclassification into plateau
(II­P) and linear (II­L) events (Barbon et al. 1979, Doggett & Branch 1985, Patat
et al. 1993) is based on photometry rather than spectroscopy. A connection of
these photometric types with spectroscopic features has been proposed and consists
mainly in the shape of the line profiles in maximum­light spectra, but has not yet
been unambiguously proven (Harkness & Wheeler 1990). Also grouped in this class
are objects with rather narrow (¸1000 km/s) emission lines of mainly hydrogen and
helium. These events have been proposed for a separate subtype (Schlegel 1990), but
only a few examples have been observed near maximum light. It might also be that
the emission of these objects is not powered by energy deposited by the explosion
in the envelope, but rather originates in the interaction of the explosion shock with
surrounding material (Chevalier 1990, Fi