Abstract

I review several topics in the structure of supernova remnants. Hydro-dynamic instabilities in young remnants may give rise to the cellular structure that is sometimes observed, although structure in the ejecta might also play a role. The presence of ejecta close to the forward shock front of a young remnant can be the result of ejecta clumps or the dynamical effects of cosmic rays. Slower moving ejecta clumps can affect the outer shock structure of older remnants such as Vela. Young remnants typically show a circular structure, but often have a one-sided asymmetry; the likely reasons are an asymmetric circumstellar medium, or pulsar velocities in the case of pulsar wind nebulae. In older remnants, asymmetric pulsar wind nebulae can result from asymmetric reverse shock flows and/or pulsar velocities.

Introduction

Observations of supernova remnants frequently show complex structure that can have its origin in several ways: structure in the freely expanding ejecta, structure in the surrounding medium, and the growth of instabilities that result from the interaction of the supernova with its surroundings. If we are to infer properties of the initial explosion from the supernova remnant, consideration of these various influences is necessary. Pulsar wind nebulae (PWNe) provide an additional probe inside a supernova remnant and can lead to an asymmetry because of a pulsar velocity. Here, I review studies of these phenomena.

Instabilities in young remnants

The basic instability that results from the deceleration of the supernova ejecta by the surrounding medium is related to the Rayleigh-Taylor instability.

Abstract

We briefly review the existing database of supernova spectropolarimetry, concentrating on recent data and on results from our group's research. Spectropolarimetry provides the only direct known probe of early-time supernova geometry. To obtain reliable conclusions, however, it is very important to correctly account for interstellar polarization. We find that Type IIn supernovae (SNe IIn) tend to be highly polarized, perhaps in part because of the interaction of the ejecta with an asymmetric circumstellar medium. In contrast, SNe II-P are not polarized much, at least shortly after the explosion. At later times, however, there is evidence for increasing polarization, as one views deeper into the expanding ejecta. Moreover, core-collapse SNe that have lost part (SN IIb) or all (SN Ib) of their hydrogen (or even helium; SN Ic) layers prior to the explosion tend to show substantial polarization; thus, the deeper we probe into core-collapse events, the greater the asphericity. There is now conclusive evidence that at least some SNe Ia are intrinsically polarized, although only by a small amount. Finally, SN spectropolarimetry provides the opportunity to study the fundamental properties of the interstellar dust in external galaxies. For example, we have found evidence for extremely high polarization efficiency for the dust along the line-of-sight to SN 1999gi in NGC 3184.

Introduction

Since extragalactic supernovae (SNe) are spatially unresolvable during the very early phases of their evolution, explosion geometry has been a difficult question to approach observationally.

Abstract

Recent observations have extended the range of diversity among spectra of Type Ia supernovae. I briefly discuss SN Ia explosion models in the spectroscopic context, the observed diversity, and some recent results obtained with the Synow code for one normal and two peculiar SNe Ia. Relating the observational manifestations of diversity to their physical causes is looking like an ever more challenging problem.

Introduction

“Surprises” refers not only to some recent developments in Type Ia supernova (SN Ia) spectroscopy that will be discussed below, but also to additional recent discoveries that I will be able only to mention, such as the polarization signal in SN 2001el (Wang et al. 2003; see also the chapter by Wang); the unusual properties of SN 2001ay (see the chapter by Howell); and the circumstellar Hα emission of SN 2002ic (Hamuy et al. 2003; see also the chapter by Hamuy). The scope of this chapter is restricted to photospheric—phase optical spectra. For recent results on infrared spectra see, e.g., Marion et al. (2003).

Some background, including mention of the various kinds of SN Ia explosion models in the spectroscopic context, is in §15.2. An overview and update of the SN Ia spectroscopic diversity is in §15.3. Some recent results from direct analysis of the spectra of three events (the normal SN 1998aq and the peculiar SNe 2000cx and 2002cx), obtained with the parameterized, resonance scattering code Synow, are discussed in §15.4. The final section (§15.5) contains more questions than conclusions.

Abstract

We have argued that MHD turbulence in an accretion disk naturally produces hoop-stresses, and that in a geometrically-thick flow these stresses could both drive and collimate an outflow. We based this argument on an analogy of turbulent MHD fluids to viscoelastic fluids, in which azimuthal shear flow creates hoop-stresses that cause a variety of flow phenomena, including the Weissenberg effect in which a fluid climbs a spinning rod.

One of the more important differences between the Weissenberg effect and astrophysical jets is the source of power. In our previous analysis, we only considered the power due to the spin-down torque on the central object, and thus found that we could only drive an outflow if the central object were maximally rotating. Here we take into account the energy that is liberated by the accreting matter, and describe a scenario in which this energy couples to the outflow to create a thermodynamic engine.

Introduction

We wish to discuss here in simple language some of our ideas regarding jet collimation and acceleration. In this paper, we will concentrate on the basic intuitive notions rather than the mathematics, which we have discussed in print elsewhere (see references below).

Review: turbulence models and jets

We have argued (Williams 2001; see also Ogilvie 2001) that the stress due to magnetohydrodynamic (MHD) turbulence in ionized accretion disks — such as, but not limited to, the turbulence driven by the magnetorotational instability (MRI) — behaves more like the stress in a viscoelastic fluid than the stress in a viscous fluid.

Abstract

The linear stability of stalled accretion shocks is investigated in the context of core collapse of type II supernovae. We focus on a particular instability mechanism based on the coupling of acoustic perturbations with advected ones (vorticity, entropy). This advective-acoustic cycle takes place between the shock and the nascent neutron star. Both adiabatic and non-adiabatic processes may contribute to this coupling, but only adiabatic ones are considered in this first approach. The growth time of the adiabatic instability scales like the advection time, and is dominated by low degree modes l = 0,1,2. Non radial modes (l = 1,2) found unstable by Blondin et al. (2003) can be related to this mechanism.

Introduction

Shocked accretion onto the surface of a compact star is known to be unstable in the context of magnetized white dwarfs, leading to shock oscillations (from Langer, Chanmugam & Shaviv 1981, hereafter LCS81, to Saxton & Wu 2001). Houck & Chevalier (1992, hereafter HC92) made a linear stability analysis of shocked accretion onto a neutron star, and found an instability reminiscent of the instability found by LCS81. HC92 showed specific cases where the cooling occurs mostly in a thin layer at the surface of the neutron star, while the flow is essentially adiabatic above it. The mechanism of the instability was described by LCS81 and subsequent authors as a kind of thermal instability: if the shock surface is moving outwards, the higher incident velocity in the frame of the shock produces a higher temperature blob, which pushes the shock further out if the increased cooling time exceeds the increased advection time.

Abstract

We explore whether the observed variations in the peak luminosities of Type Ia supernovae originate in part from a scatter in metallicity of the main-sequence stars that become white dwarfs. Previous, numerical, studies have not self-consistently explored metallicities greater than solar. One-dimensional, Chandrasekhar mass models of SNe Ia produce most of their 56Ni in a burn to nuclear statistical equilibrium between the mass shells 0.2 M⊙ and 0.8 M⊙, for which the electron to nucleon ratio Ye is constant during the burn. We show analytically that, under these conditions, charge and mass conservation constrain the mass of 56Ni produced to depend linearly on the original metallicity of the white dwarf progenitor. This effect is most evident at metallicities greater than solar. Detailed post-processing of W7-like models confirms this linear dependence, and our calculations are in agreement with previous self-consistent calculations over the metallicity range common to both calculations. The observed scatter in the metallicity (1/3 Z⊙-3 Z⊙) of the solar neighborhood is enough to induce a 25% variation in the mass of 56Ni ejected by Type Ia supernova and is sufficient to vary the peak V-band brightness by |ΔMV| ≈ 0.2. This scatter in metallicity is present out to the limiting redshifts of current observations (z ≲ 1). Sedimentation of 22Ne can possibly amplify the variation in 56Ni mass to ≲50%. Further numerical studies can determine if other metallicity-induced effects, such as a change in the mass of the 56Ni-producing region, offset or enhance the variation we identify.

Abstract

We study the evolution of supernova remnants in the circumstellar medium formed by mass loss from the progenitor star. The properties of this interaction are investigated, and the specific case of a 35 M⊙ star is studied in detail. The evolution of the SN shock wave in this case may have a bearing on other SNRs evolving in wind-blown bubbles, especially SN 1987A.

Introduction

Type II Supernovae are the remnants of massive stars (M > 8 M⊙). As these stars evolve along the main sequence, they lose a considerable amount of mass, mainly in the form of stellar winds. The properties of this mass loss may vary considerably among different evolutionary stages. The net result of the expelled mass is the formation of circumstellar wind-blown cavities, or bubbles, around the star, bordered by a dense shell. When the star ends its life as a supernova, the resulting shock wave will interact with this circumstellar bubble rather than with the interstellar medium. The evolution of the shock wave, and that of the resulting supernova remnant (SNR), will be different from that in a constant density ambient medium.

In this work we study the evolution of supernova remnants in circumstellar wind-blown bubbles. The evolution depends primarily on a single parameter, the ratio of the mass of the shell to that of the ejected material. Various values of this parameter are explored.

Abstract

It is a weird and unlikely circumstance that a collapse supernova (Type II) should explode. The peculiar mechanism that facilitates this explosion is the formation and preservation of large scale structures in a high entropy atmosphere residing on the surface of a nearly formed neutron star. The high entropy atmosphere is maintained by two sources: the gravitational energy of initial formation of the neutron star, released by diffusion and transport of neutrinos and secondly and possibly dominantly by the gravitational energy released at the suface by additional low entropy matter falling through to the neutron star surface. The preservation of this entropy contrast between up and down flows requires thermal isolation between the low entropy down flows and the high entropy up flows. This entropy contrast allows an efficient Carnot cycle to operate and thus allows the efficient conversion of thermal energy to mechanical, which in turn drives the explosion. The P-V diagram of various up and down going mass elements in the calculations demonstrates the existence of the cycle and its efficiency. Greater thermal isolation should occur in 3-D as opposed to 2-D calculations because of the difference in relative thickness or surface to mass ratio for the same mass flow in 2 and 3-D. This may explain the observed stronger explosion in 3-D calculations.

Prolog

This paper is written in honor of a long and lasting friendship between Craig Wheeler and the first author for more than half his current life.

Abstract

Current observations stimulate the production of fully three-dimensional explosion models, which in turn motivates three-dimensional spectrum synthesis for supernova atmospheres. We briefly discuss techniques adapted to address the latter problem, and consider some fundamentals of line formation in supernovae without recourse to spherical symmetry. Direct and detailed extensions of the technique are discussed, and future work is outlined.

Introduction

Spectrum synthesis is the acid test of supernova modelling. Unless synthetic spectra calculated from a hydrodynamical stellar explosion model agree with observations, the model is not descriptive. Some explosion modellers contend that only three-dimensional (3-D) models faithfully describe the physics of the real events. If this is so, then the evaluation of those models requires solutions to the 3-D model supernova atmosphere problem. These solutions require full detail, the inclusion of as much radiation transfer physics as possible. Otherwise, a bad fit of a synthetic spectrum to an observed one might have less to do with the accuracy of the hydrodynamical model, and more to do with the shortcomings of the radiation transfer procedure.

On the other hand, solutions (of a sort) to the ill-posed inverse problem constrain parameter space available to hydrodynamical models. Fast, iterative, parameterized fits to observed spectra characterize the ejection velocities and identities of species found in the line forming region. Most importantly, the procedure reveals species that cannot be identified by simply Doppler-shifting line lists on top of observed spectra in search of feature coincidences.

Abstract

Polarization and other observations indicate that supernova explosions are aspherical and often axisymmetric, implying a necessary departure from spherical models. Akiyama et al. investigated the effects of the magneto-rotational instability (MRI) on collapsing and rotating cores. Their results indicate that the MRI dynamo generates magnetic fields of greater than the Q.E.D. limit (4.4 × 1013 G). We present preliminary results of the effects of the super-strong magnetic field on degenerate electron pressure in core collapse.

Introduction

Although core collapse cannot be observed directly, except with neutrinos, observations of explosion ejecta can provide us with information about the explosion mechanism itself. Such observations indicate that explosions of core collapse supernovae are aspherical and often bipolar. HST observations clearly show that 1987A has aspherical ejecta for which the axis aligns roughly with the small axis of the rings (Pun et al. 2001; Wang et al. 2002). Spectropolarimetry is a powerful tool for probing ejecta asphericity, and it reveals that most, if not all, core collapse supernovae possesses asphericity and often times bipolar structure (Wang et al. 1996, 2001). Explosions of Type Ib and Ic are more strongly aspherical, while the asphericity of Type II supernovae increases with time as the ejecta expand and the photosphere recedes (Wang et al. 2001; Leonard et al. 2000, 2001). The indication is that it is the core collapse mechanism itself that is responsible for the asphericity.

The observational evidence of asphericity motivates the inclusion of rotation in core collapse physics.

Introduction

The assumption commonly made is that Supernovae of Type Ia (SN Ia) are the result of thermonuclear runaways (TNR) in the cores of carbon-oxygen white dwarfs (WD) which are members of binary systems and have accreted material from a companion until their masses exceed the Chandrasekhar Limit (Leibundgut 2000, 2001). However, the binary star systems that end in this explosion are not yet known although there have been numerous proposals. Nevertheless, the importance of SN Ia, both to our understanding of the evolution of the Universe and to the formation of iron in the Galaxy, demands that we determine the progenitors of these explosions.

Originally proposed by Whelan and Iben (1973), virtually every type of close binary which contains a WD has been suggested at one time or another. However, based purely on observational concerns, most of the systems that have been proposed cannot be the progenitors (Starrfield 2003). For example, one of the first suggestions was a classical nova system (CN), but the amount of core material ejected during the outburst implies strongly that the WD is losing mass as a result of the outburst (Gehrz et al. 1998). Other suggestions such as Symbiotic Novae (T CrB or RS Oph, for example) can probably be ruled out.because there is too much hydrogen present in the system (the explosion takes place inside the outer layers of a red giant) and the defining characteristic of a SN Ia outburst is the absence of hydrogen in the spectrum (Filippenko 1997).

Transmission of a polarization state

The word “teleportation” comes from parapsychology and means transportation of persons or things from one place to another using mental power. It was taken over into science fiction literature, where the transport is imagined to take place instantaneously. However this is still to be invented, and is surely nonsense – relativity theory teaches us that the velocity of light is the upper bound for the motion of an object. Nevertheless, teleportation has occupied a firm place in our fantasies, and when renowned quantum physicists (as has happened) use this word, they can be sure to attract attention. So, what is it all about? The basic idea is that it is not necessary to transport material constituents (ultimately the elementary particles). The same particles already exist at other places; we “simply” need to put them together in the right way. To do this, we need a complete set of building instructions, and this is, according to quantum theory, the quantum mechanical wave function representing the maximum information known about an object. We could imagine the wave function measured on the original system, then transmitted via a conventional (classical) information channel to another place and there used for system reconstruction. Unfortunately, the first step, the determination of the wave function on a single system, is impossible (see Section 10.3). However, quantum mechanics offers us another “magic trick.”

Particle properties of radiation

One of the most important properties of macroscopic material systems is their ability to emit radiation spontaneously. According to quantum mechanics, the emission process is realized in the following way: an atom (or a molecule) makes a transition from a higher lying energy level (to which it was brought, for example, by an electron collision) to a lower lying energy level without any noticeable external influence (in the form of an existing electromagnetic field), and the released energy is emitted in the form of electromagnetic radiation. The discrete energy structure of the atom dictated by the laws of quantum mechanics is imprinted also on the emission process (quantization of the emission energy), since the energy conservation law is also valid for single (individual) transitions. Hence, a single photon, in the sense of a well defined energy quantum, is always emitted.

The emitted quanta can be directly detected by a photodetector. (Strictly speaking, identifying a registered photon with an emitted one is possible only when it is guaranteed that the observed volume contains only a single atom. (For details see Sections 6.8 and 8.1.) Under realistic conditions, the experiment can be performed in the following way. First, a beam of ionized atoms is sent through a thin foil; the emerging beam then consists of excited atoms. (This procedure is known as the beam–foil technique.) A detector is placed at a distance d from the foil to detect light emitted sideways by the atomic beam (Fig. 6.1).

Measuring the diameter of stars

As mentioned several times already, the particle character of light is best illustrated by the photoelectric effect. This effect can be exploited in the detection of single photons by photocounting. The analysis of such counting data allows us, as will be discussed in detail in this chapter, to gain a deeper insight into the properties of electromagnetic fields. We can recognize the “fine structure” of the radiation field – in the form of fluctuation processes – which was hidden from us when using previous techniques relying only on the eye or a photographic plate, i.e. techniques limited to time averaged intensity measurements.

The credit for developing the basic technique for intensity fluctuation measurements goes to the British scientists R. Hanbury Brown and R. Q. Twiss, who became the fathers of a new optical discipline which investigates statistical laws valid for photocounting under various physical situations. When we talk of studies of “photon statistics” it is these investigations that we are referring to.

Interestingly enough, it was a practical need, namely the improvement in experimental possibilities of measuring the (apparent) diameters of fixed stars, that gave rise to the pioneering work by Hanbury Brown and Twiss. Because the topic is physically exciting, we will go into more detail.

It is well known that the angular diameters of fixed stars – observed from Earth – appear to be so small that the available telescopes are not able to resolve the stars spatially.