The impact of SN1987a on theoretical studies of the specific mechanism generating Type II supernovae is examined. The explosion energy extracted from analysis of the light curve for SN1987a is on the edge of distinguishing between a prompt explosion from a hydrodynamic shock and a delayed, neutrino-induced, explosion. The detection of neutrinos from 1987a is also reanalysed.

There is now general agreement that the presupernova star for 1987a was indeed SK-69-202, a star that had visual magnitude 12.36 (Walborn et al. 1987), bolometric correction 1.15m (Humphreys and McElroy 1984), and extinction 0.5m (Woosley et al. 1987). The distance modulus to the LMC is presently controversial, but most astronomers prefer a value near 18.5 (e.g., Walker 1985; Chiosi and Pigatto 1986) with an uncertainty of perhaps 0.2m. Putting it together, one arrives at a bolometric magnitude for SK-69-202 of −7.8 with a probable range −7.5 to −8.2, or a luminosity of 4 × 1038 erg s−1 with a range of 3 to 6 × 1038 erg s−1. At the time of the supernova the hydrogen burning shell contributes negligible energy generation so the critical quantity determining the luminosity is the helium core mass. Inspection of a variety of current stellar models indicates that SK-69-202 had a helium core mass of 6 ± 1 M⊙ from which we may infer that, on the main sequence, it had a mass of 19 ± 3 M⊙. A B3-I supergiant has a surface temperature of ∼16,000 K (Humphreys and McElroy 1984). For the above range of luminosities and assuming a temperature in the range 15,000 to 18,000 K one obtains a radius for SK-202-69 of 3 ± 1 × 1012 cm.

We consider here the nebular phase spectra of supernovae whose late time luminosity is provided by the radioactive decay of 56Ni and 56Co synthesized in the explosion. A broad variety of supernovae are known or suspected to fall in this category. This includes all SNIa and SNIb, and at least some SNII, in particular SN1987a. At sufflciendy late times the expanding supernova becomes basically nebular in character due to its decreasing optical depth. The spectra produced during this stage contain information on the density and abundance structure of the entire supernova, as opposed to spectra near maximum light which are affected only by the outermost layers. A numerical model for nebular spectrum formation is therefore potentially very valuable for answering currently outstanding questions about the post-explosion supernova structure. As an example, we can hope to determine the degree of mixing which occurs between the layers of the “onion-skin” abundance structure predicted by current one dimensional explosion calculations.

Since the appearance of the classical papers on stellar nucleosynthesis in the 1950’s most of the observational tests have been through indirect sources of information. Even though this has been rather successful (cf. various contributions in this volume), it represents only an average over all sources, yielding little information about specific stars. The most direct evidence comes from observations of young galactic supernova remnants (age less than ∼103 years). Unfortunately, the analysis of X-ray data are hampered by a lack of understanding of the detailed physics, eg. non-equilibrium and plasma effects, as well as by observational problems (Itoh and Nomoto, 1987). Optical observations (cf. Raymond, 1984) have given some valuable insight of eg. the oxygen-rich remnants. The analysis of these are, however, suffering from the fact that only a small fraction of the mass is seen in the optical.

The observation of X-rays and γ-rays from SN 1987a can provide important constraints on parameters for models of this unique event. We present the results of detailed Monte Carlo calculations of the fluxes to be expected in several X-ray bands and for the strong line at 847 keV associated with the decay of 56Co. Our calculations use Model 10H of Woosley, Pinto, and Ensman(1988), with 0.075M⊙ of radioactive material. If it is assumed that there is no mixing of this material with the layers above, then the X-ray fluxes do not become detectable as early as the observations made by the Ginga team in August, 1987. If these observations correspond to X-rays arising from γ-rays Compton scattered down in energy in the supernova ejecta, rather than the interaction of the ejecta with circumstellar matter, then they can only be explained by mixing outward of radioactive material or an envelope with some combination of less mass or greater kinetic energy per unit mass.

An unusual hard X-ray source was discovered in an error box of 0.2° × 0.3° including SN1987A from the X-ray astronomy satellite Ginga. The energy spectrum is quite unusual for any known classes of X-ray source, and apparently consists of two separate components, a soft component and a very hard component. This source is considered to be identified with SN1987A. The X-ray emergence occurred in July, 1987, or possibly even earlier. The soft component is significantly time-variable and also showed a flare-like increase in January, 1988, while the intensity of the hard component has remained relatively unchanged for more than 200 days.

The image of Sk −69° 202 was scanned and analyzed on eight (of 32 available) blue through near-infrared photographic plates obtained at the prime focus of the Cerro Tololo Inter-American Observatory 4-meter telescope during 1974–1983. Both intensity syntheses of the image and density differences were derived by means of reference stars from the same plates, including the similar nearby object Sk −69° 203. Several of the density differences are shown in Figure 1. The analysis shows that the 12m blue supergiant in Sk −69° 202 (Star 1) has two companions with V magnitudes, position angles, and separations 315°, 3″ (Star 2) and 115°, 1.″5 (Star 3), respectively. Both companions appear to be early-type stars; there is no evidence for a bright red star in the system. The two companions are responsible for the spectra observed by the International Ultraviolet Explorer following the decline of the SN in the far UV, so that Star 1 has disappeared and was probably the progenitor. The most likely interpretation is that it was a post-red supergiant evolving blueward in the HR diagram.

Ideally, the evolutionary models for the precursor of SN 1987 A should account for both the SN properties and the observational constraints for massive stars with relevant mass and composition.

Mass loss is an essential property of massive star evolution. Recent parametrisations of mass loss rates for galactic stars cover the whole HR diagram. There are indications that for given L and Teff values, is lower at lower metallicity and therefore is lower in the LMC than in the Galaxy, thus we take with f < 1. Various models of an intitial 20 M⊙ star with f=0.2, 0.4, 0.6 and 1.0 are constructed (cf. Fig. 1) with a metallicity Z=0.006 and a moderate overshooting dover=0.3 Hp. From these models, we suggest an initial mass on the zero age sequence of 17 to 18 M⊙. The pre-SN location in the HR diagram very much depends on the remaining stellar mass, or more precisely on the mass of the remaining H-rich envelope. A final location at log Teff ≃ 4.2 is obtained for a final mass of about 9.0 M⊙ (cf. Fig.1). Scaled to an initial value of 17 M⊙, this corresponds to a final mass of about 8 M⊙ and a remaining H-rich envelope of a few tenths of a solar mass at most. The stellar surface exhibits CNO equilibrium values with C/N ≃ 0.01 and O/N ≃ 0.1 in mass fraction, and an hydrogen content X (surf) = 0.39. The blue progenitor is obtained for f=0.4, i.e. for -values in the LMC equal to 40% of the galactic values.

The evolution of a 17.5 M⊙ star, chosen to be similar to the precursor of SN1987A, has been studied using the input physics described in Wood and Faulkner (1987). The calculations: use opacities from the Astrophysical Opacity Library of Huebner et al (1977) with H, He, C, N, O and other metals in LMC ratios; treat semi-convection in the manner of Lamb, Iben and Howard (1976); and assume the mass loss rate to be the minimum of (a) α times the rate give in Waldron (1985) (this rate applied in the blue part of the HR diagram), and (b) L/(cv) (this rate applied in the red), where v is the stellar wind expansion velocity which was taken to be 12 km s−1.

Typical evolutionary tracks resulting from these calculations are shown in Figure 1. The track shown as a continuous line is the one shown in Wood and Faulkner (1987) and corresponds to α = 2.5, while the track shown as a dotted line results from a slightly different treatment of semi-convection on the main-sequence and to α = 1.

We have calculated a large grid of hydrogen-rich supernova photospheres, in which radii, effective temperatures, density profiles, and expansion velocities have been varied. Spherical geometry, radiative equilibrium and LTE level populations are assumed. In the quasi-exact radiative transfer, the dilution of the radiation field, and scattering as well as absorption (by all relevant continuous processes and up to 150 000 lines in some models) are accurately considered. Good agreement can be obtained with the UV and IR spectra of supernovae 1979C, 1980K, and 1987A as observed during the coasting phase. Potential methods of parameter determinations for SN II are briefly discussed.

The ionization of CaII by the Lα quanta in the envelope of the type II supernova 1970g on 270th day is considered. The ratio CaII/CaIII is found to be very low (<0.1). This results in the low theoretical intensities of the CaII emission lines; they are at least an order of magnitude weaker than the observed ones. The implications are discussed.

The collapse of the iron core of massive stars ( M ≥ 8 MO) is initiated by photodissociation and electron capture. The collapse of the inner core proceeds homologously until it is stopped by the stiffness of the equation of state (hereafter EOS) at nuclear density and it stops or rebounds. A shock forms at the edge of homology. The initial strength of the shock increases with the velocity difference between the inner and outer cores, i.e. it increases with a larger rebound of the inner core. The uniterrupted propagation of this prompt shock through the remainder of the core to the stellar mantle, where it can deliver enough energy to blow off the loosely bound outer layers, has long been proposed as the mechanism of type II supernovae explosions. However most authors did not get an explosion as a result of the prompt mechanism. Recently Baron et al. (1985) reported that the combination of General Relativity (GR) with a relatively soft EOS at nuclear densities leads to a much greater blow off than they got with Newtonian hydrodynamics. In order to see where purely hydrodynamical effects are important, namely for what EOS the GR outburst is greater than the Newtonian, we did a set of pure hydrodynamical adiabatic calculations (complete neutrino trapping) with different EOS above nuclear densities, turning the GR terms on and off. Neutrino leakage, which we do not incorporate, usually leads to harmful energy losses.

Explosion of type II supernova is, in principle, a difficult process: The presupernova star was in gravitationally bound state with negative energy but it has to be divided into two parts, the collapsed core of still lower (negative) energy and the ejected envelope of positive energy. This process is against nature in the sense most of the phenomena in nature proceed towards equipartition of energies. Thus, some finely tuned mechanism should be necessary for successful explosion. Two different mechanisms have been proposed; one is the prompt explosion where the gravitational energy release by the core collapse is transferred to the mantle by a shock wave, and the other is delayed explosion where it is transferred slowly by neutrinos diffusing out of the core. In what follows we shall concentrate in the case of the delayed explosion.

W ehave developed a time-energy correlation method1 to bring forth the mass signature from Supernova 1987a neutrino observations, if the neutrino has any mass at all. This method is particularly effective in analyzing data sets with a small number of events, such as the Kamiokande II2 and the IMB3 observations of neutrino bursts from Supernova 1987A. The time dispersion Δt12 between two simultaneously emitted neutrinos of energies E1, E2 (E1 < E2 ) and a neutrino mass energy mv is given by:

where L is the distance of the source and c is the velocity of light. Conversely, Eq.(1) can also be used to establish time relationships of detected neutrinos and the existence of a mass. Applying Eq.(1) to all pairs for which real values of mv (called the correlation mass) are obtained from the observed Δt12, E1, E2, the existence of a cluster of pairs with essentially the same mass mv will indicate (a) many pairs of neutrinos were emitted within a narrow time window, and (b) the existence of a mass at mv. If a group of neutrinos were emitted within a narrow window, these groups will show a strong time correlation. Thus, this method of analysis does not impose a condition for the emission mechanism - rather, if the result of this analysis indicates the existence of a mass, there must exist a time correlation among the neutrinos.

A massive star has been believed to end his life with the collapsed driven supernova explosion and the formation of the compact object such as a neutron star or a black hole. When the compact object is formed, a large amount of energy corresponding to the binding energy of the object must be released. It has been considered that most of the energy is emitted by neutrinos because of their adequate coupling with the matter. The observation of the neutrino burst from SN1987A by Kamiokande and IMB offered us the first chance to test these scenarios of the collapse driven supernova explosion directly. We began to analyze the data just after their publication and got many important results which are presented below. In our analysis the distance of SN1987A is assumed to be 50kpc.

SN1987A gave us the first opportunity to study the supernova core directly by providing us the neutrino signal from the core. The observational data of the neutrino flux detected by Kamiokande[1] and IMB[2] show surprisingly good agreements with the theoretical predictions as a whole[3,4]. The fundamental concept of the collapse driven supernova explosion is confirmed for the first time. On the other hand, there are some puzzles. The most peculiar feature of the data is the 7 seconds gap of the Kamiokande data. The first 8 events of Kamiokande were detected in 2 seconds, following the 7 seconds gap and the last 3 events in 4 seconds. Of course just only 7 seconds gap is not unnatural if small neutrino flux come. But there were detected 3 events after the gap. These 3 events may not be produced by the weak flux. We can estimate the time integrated luminosity of corresponding to the last 3 events and get the large value such as 7 · 1052erg [5]. Can we get out of this inconsistency, 3 events after the 7 seconds gap? If not, we may need to consider some nonstandard mechanism of the neutrino emission at the late time. In order to investigate the probability of the case in which there is a 7 seconds gap before 3 events, we have performed Monte Carlo simulations for the simple model of neutrino flux.

Recently it has been suggested that neutrinos have nonvanishing mass from both experimental and theoretical investigations. Accelerator experiments and cosmology, however, impose the constraints on mass mv and lifetime τv of vτ (see Fig. 1). As we showed in the previous paper [1], supernova explosion also imposes the stringent constraint on them. The purpose of this paper is to investigate the constraint imposed on mv and τv of vτ from SN1987A.

Tau neutrinos are emitted from the central part of the iron core together with the electron neutrinos because they are also in thermal equilibrium with matter there.

The neutrino burst from SN1987A were detected by KAMIOKANDE and also IMB in February this year. The neutrino events detected by KAMIOKANDE were clustered into three bunches; event number 1-5, 7-9, 10-12. The first two bunches of the neutrino events can be understood by the standard scenario of supernova explosions.

However, it is very difficult to explain the last bunch of the neutrino events by the standard scenario, especially the total energy of the neutrinos and the radius of the neutrino sphere; it is necessary at least several times of 1052 ergs to explain the total energy of the last three events. In order to explain this energy, we must consider the non-standard mechanism of the neutrino emission.

It is shown that the bunch structure of the Kamiokande neutrino events associated with SN1987a can be naturally understood, if one assumes that the core of the progenitor star was rotating moderately with q(≡ Jc/GM2) ≈ 3 with J the total angular momentum and M the gravitational mass of the core.

We assume that the presence of the observed gap, at least that between the second and the third bunch, is real and consider its implications in the dynamics of the core collapse. Let us define a nondimensional angular momentum q(≡ Jc/GM2) with J the total angular momentum and M the gravitational mass of the core. We assume that the value of q for the core of SN1987a was about 3. Then we expect that the effect of the angular momentum plays an important role when the size of the core r shrinks to < 107cm. In the spherically symmetric collapse model, the size of the unshocked homologous core is of the order of 107cm. Hence we expect that the core at the bounce is essentially governed by the spherically symmetric dynamics. To verify this point we may refer to the simulation by Symbalisty1. Our model almost corresponds to Model ROT2 in which one obtains a neutrinosphere with a roughly spherical radius of approximately 50km. We then expect that the infalling nonhomologous matter onto the core will liberate a gravitational energy of the order of 1052ergs.

The neutrino emission processes play essential roles in stellar evolution as expemplified by the observations of the neutrinos from SN 1987a by the KAMIOKANDE-II and IMB experiments. Recently a very extensive study of the various neutrino emission processes based on the Weinberg-Salam theory has been completed by the present author and his collaborators. The neutrino emission processes calculated by the author’s group include pair, photo-, plasma, and bremsstrahlung neutrino processes. The neutrino energy loss rates due to pair, photo-, and plasma processes in the framework of the Weinberg-Salam theory are found to be substantially lower than the result obtained by Beaudet, Petrosian, and Salpeter. The reduction factor α is in the range 0.35 < α < 0.88 depending on the neutrino masses, density, and temperature. The ionic correlation effects play important roles in the bremsstrahlung neutrino process. The present author and his collaborators recently calculated the bremsstrahlung neutrino energy loss rate taking into account the ionic correlation effects in the crystalline lattice state as well as in the liquid metal state. They found that the ionic correlation effects suppress the bremsstrhlung neutrino energy loss typically by a factor 2-20. The present findings will bear great importance in neutrino astronomy.