This book is in effect a second edition of a book first published by Wykeham Publications in 1970. The Wykeham series was designed to bridge the gap between school and university science and the mathematical level of the book was designed to be suitable for sixth formers. In fact the main use of the book has been as a university textbook and some of the mathematics which was then taught in schools is now taught at university. In rewriting the book, I have not changed its general level, but I have introduced two appendices containing more mathematical detail relating to topics discussed in the main text.

Many branches of physics such as gravitation, thermodyamics, atomic physics and nuclear physics are combined in determining the structure of stars. As a result the subject provides an ideal example of the application of fundamental physics. Physical conditions in stars are more extreme than on Earth and a successful understanding of their structure should show how valid it is to extrapolate established physical laws to these conditions. Although it is profitable to study stars as isolated objects, an understanding of star formation and stellar evolution is central to the whole study of astronomy.

Significant progress has been made in explaining the observed properties of stars but there is still room for considerable improvement in the relation between theory and observation. In particular the process of star formation is not well understood theoretically or observationally. In fundamental physics there is still need for a reliable theory of fully developed convection.

Introduction

Throughout the book so far I have only discussed the structure of isolated stars, although I have stressed the importance of binary stars in providing information about stellar masses and radii. Although most stars may be partners in binary or multiple systems, for most of them at least for most of their life history the stars are sufficiently far apart that the internal structure of the stars is not affected by the presence of a companion. The mutual gravitational attraction of the two stars causes them to orbit about their common mass centre but otherwise their binary nature can be ignored. This ceases to be the case if binary stars are initially very close or if they become close during their evolution, possibly because one star expands substantially to become a red giant and as a result its surface gets very close to its companion.

When stars are close there are two distinct types of interaction between them. The properties of the surface layers of one star may be affected as a result of irradiation by the other star. This will be particularly true when one component is much more luminous than the other, when the effect on the less luminous star will be very great. In addition both the gravitational attraction of the companion star and the rapid rotation, which must occur because the stars orbit around one another, will cause a star to deviate substantially from spherical symmetry.

Introduction

In the previous discussion of stellar evolution it has frequently been remarked that, so long as the stellar material remains in the form of an ideal classical gas, its central temperature can only increase as it evolves. This result was originally deduced from the Virial Theorem (3.24) on page 55. As I have mentioned on page 201 in Chapter 9, there is at present no completely clear solution to the problem of what happens to a star whose central temperature is still rising at the time that nuclear fusion reactions have converted the central regions to iron, although the association with supernovae of type II seems highly probable; in fact, as I shall explain in the last section of this chapter, the problem can arise even earlier than that. However, if the centre of the star ceases to be an ideal classical gas and becomes a degenerate gas, it is possible that the central temperature may pass through a maximum and that the star may cool down and die. This possibility has already been illustrated for low mass stars in figs. 76 and 86 Such a dying star is likely to have a low luminosity. It is also likely to have a high density. It can only begin to cool down after its central regions have become degenerate and, if the central temperature has previously risen sufficiently for one or more sets of energy-releasing nuclear reactions to occur, a very high density is necessary before degeneracy can occur, as has been seen in Chapter 4 (fig. 45)

Such under-luminous dense stars have been oberved.

Historical introduction

After the main sequence the most prominent group of stars in the HR diagram (fig. 56) is the red giants and supergiants. These stars have larger luminosities and radii than main sequence stars of the same colour. From the discussion in Chapter 5, it appears that red giants are not stars of homogeneous chemical composition and I must now discover how red giants differ from main sequence stars in their internal structure as well as in their surface properties. I have already indicated at the end of Chapter 2 (fig. 30) that stars become red giants when nuclear reactions in their interiors lead to a non-uniformity of chemical composition. Before I discuss this further, I will give a brief historical introduction to the problem of the red giants. Although in this book I mainly discuss the present state of knowledge, it is perhaps instructive in one case to trace the steps by which the present knowledge has been obtained.

When the first theoretical calculations of stellar structure were made, it was very difficult to explain the occurrence of red giants, since at the time it was believed that stars remained chemically homogeneous as they evolved. As will now be described, it was believed that the rotation of stars caused them to be well mixed. Most stars are observed to rotate, even if the rotation of many of them is not sufficiently rapid to distort their structure substantially. Rotation is detected by the Doppler effect.

In this book I have described the methods used in the theoretical study of stellar structure and evolution and I have discussed many of the results obtained. I have tried to discuss the present state of a developing subject and to mention the main uncertainties. As I have stressed, particularly at the end of Chapter 9, some of the detailed theoretical ideas may prove to be wrong, but it is confidently expected that the broad outline of the subject as presented in Chapters 3–5 is correct. In this chapter I discuss further some of the points where important uncertainties remain.

In the first place it is important to realise that, although this book has been written by a theoretical astrophysicist, who has a particular interest in obtaining a theoretical understanding of the subject, ultimately all of the theoretical work must be related to observations. This has a twofold implication. The theoretical worker must keep the observational results in mind and there is a continuing need for new observations. The subject depends considerably on some of the less glamorous parts of observational astronomy. In these days of quasars, pulsars and the cosmic microwave radiation, the work of measuring parallaxes and proper motions and studying the orbits of binary star systems is often regarded as being very humdrum. However, it is vitally important in supplementing the information possessed about such things as masses, radii and absolute magnitudes.

Introduction

In this chapter I consider what are the main physical processes which determine the structure of stars and what equations must be solved in order to find the details of this structure. At the outset it must be stressed that the theoretical astrophysicist does not usually attempt to calculate the properties of a particular star which has been observed. As we have learnt in the last chapter, the number of stars for which there is sufficiently detailed observational knowledge to make this procedure worthwhile is very small. Instead the theoretician tries to isolate the factors which mainly determine the properties of stars and then tries to calculate the structure of a wide range of possible stars. We shall see that the most important factors are the mass and initial chemical composition of the star and the time that has passed since it was formed. In what follows I shall often refer to the birth of a star, its age and its chemical composition at birth. Once calculations have been made for a range of values of mass, chemical composition and age, the results can be compared with the general properties of stars rather than with the properties of individual stars. I shall consider this comparison in Chapters 5 to 10. For one star, the Sun, we possess extremely detailed observational information and there has been a considerable effort to try to obtain a theoretical understanding of its properties.

Introduction

In most of my previous discussion I have assumed that stars spend their entire life with a constant mass, whose chemical composition changes as a result of nuclear reactions. I have also mentioned that, in fact, mass loss from stars is important and I shall now say more about this subject. Observationally mass loss is apparent in the explosions of supernovae and, to a lesser extent, novae and it is also clear that planetary nebulae are formed of mass ejected by stars. Evidence of mass loss from ordinary stars has only been well-established since the development of space astronomy as will become clear in what follows. In this chapter I shall restrict myself to a discussion of mass loss from single stars or from binary stars whose separation is so great that the two components evolve independently. In the next chapter I shall discuss mass exchange between components in close binary systems, which may also involve mass loss from the entire system.

The solar wind

It has been known since about 1960 that the Sun is losing mass at a rate of between one part in 1014 and one part in 1013 a year. This loss is known as the solar wind because it flows through interplanetary space and past the Earth with a velocity of several hundred kilometres a second.

This book is concerned with the structure and evolution of the stars, that is the life history of the stars. Its aim is to show how observations of the properties of stars and knowledge from many branches of physics have been combined, with the aid of the necessary mathematical techniques, to give us what we believe is a good understanding of the basis of this subject.

Because the stars are so remote from the Earth it may seem surprising that we can learn anything about their physical dimensions. To hope to be able to describe their internal structure and, still more, their evolution appears extremely optimistic. The mass and radius of a few stars can be measured directly, but for most stars the only source of information is in the light that we receive from them. This gives us some idea about the temperature and chemical composition of the surface layers of the star and about the total light output (luminosity) of those stars whose distance from the Earth is known. It also indicates that some stars are rotating rapidly or have strong magnetic fields and that others are losing mass from their surfaces. No direct information is obtained about physical conditions in the interiors of the stars, with the exceptions (discussed in Chapters 4 and 6) that the neutrinos emitted in the solar centre can be detected on Earth and that vibrations of the solar surface can provide information about the interior by techniques similar to seismology.

Introduction

In Chapter 6 I have given an account of calculations of stellar evolution away from the main sequence. These calculations have not followed a star through its entire life history except approximately in the case of those low mass stars which do not burn their hydrogen and helium before their central temperatures cease to rise and the star as a whole then cools down and eventually ceases to be luminous. For more massive stars there are several evolutionary stages after those which have been discussed in Chapter 6.

It is difficult to calculate the evolution of a star all the way from its initial main sequence state to the end of its life history. There are many reasons why this should be so. One important difficulty is that in all calculations errors tend to accumulate. The numerical processes used in solving differential equations can never be completely accurate and over a long period of integration these mathematical errors tend to pile up. In addition, the mathematical expressions for the physical laws are only approximate. In many cases the physical processes which occur in stars cannot be observed directly in the laboratory and in the case of stellar convection there is neither a good theory nor a good experiment. The uncertainties in the internal structure of a star and particularly in its observed properties may be small when it is on or near to the main sequence, but they might lead to the prediction of an incorrect physical process at a later stage.

ABSTRACT

Simulations which explore mergers like those thought responsible for the shells around many elliptical galaxies find little correlation between the distribution of stars and gas in remnants. Mergers of small companion disks consisting of both gas and stars with non-spherical primary potentials produce shell galaxies with gaseous nuclear rings and clumps.

INTRODUCTION

Models which follow the infall of less-massive companion galaxies show that shell galaxies can be formed by accretion. However, it is probable that the sources of material also contain significant amounts of gas. We investigate encounters that produce shells by modeling interactions between non-spherical primary galaxies and companions containing both stars and gas with a three-dimensional code (TREESPH: Hernquist and Katz 1989). Primaries are modeled with rigid elliptical potentials of the form presented by Hernquist (1990) with scale-length a = 1. Physical time t′ is related to the calculation time unit by t′ ≈ 4.3 × 106t. The companion is a rotationally supported disk in which particles are distributed according to an exponential surface density profile. Stars have a total mass 1/10 and gas 1/100 that of the primary. In most interactions, the companion potential is disrupted at a small distance from the primary after which the particles evolve in the solitary primary gravitational field.

ABSTRACT

Formation of first AGN probably follows closely the formation of their host galaxies. Both processes may involve similar astrophysical events or processes, with comparable energetics. Young AGN could have had a profound effect on the host galaxies. High–Z AGN may be used as markers of galaxy formation sites, or probes of early large-scale structure. Some may well be examples of galaxies in the early stages of formation, although the nature of their dominant energy sources and their evolutionary status remain ambiguous. Future observations at IR through sub-mm wavelengths lead to a discovery of possible obscured protogalaxies and nascent AGN, and probe their formation at very high z.

INTRODUCTION: FORMATION OF GALAXIES AND AGN

Probably the two key problems in extragalactic astronomy and cosmology today are the understanding of formation and evolution of galaxies and large-scale structure, and the understanding of formation and evolution of AGN. These problems may be fundamentally connected: it is now generally understood that the same kind of astrophysical processes, dissipative merging and infall, may be central to both formation of galaxies and formation of AGN, and subsequently to be contributing both to the simulation of star forming activity in galaxies and to the feeding of their central engines. AGN may also exert a considerable feedback to their host galaxies, and perhaps even determine some of their global properties.

ABSTRACT

We model a range of flyby galaxy interactions in order to investigate the formation of starbursts in galaxies with induced stellar bars and/or mass loss or accretion. The models indicate that mass transfer is important in triggering radial gas flows and starburst activity if counter-rotating accretion occurs; i.e., accretion from a prograde disk onto a retrograde disk. Such accretion proves effective in shedding rotational angular momentum of the ISM, resulting in the radial gas flow and subsequent nuclear ISM concentration, while leaving behind a relatively unperturbed stellar disk. However, bar formation proves more important under a wider range of interaction scenarios than does mass transfer, and thus bar formation is the dominant process in triggering nuclear activity in interacting systems as a whole.

INTRODUCTION

The link between galaxy interactions and elevated star formation rates has been demonstrated through observations of such star formation tracers as optical emission lines (e.g., Kennicutt et al 1987; Bushouse 1987), strong far infrared emission (Lonsdale, Persson, and Matthews 1984), and radio continuum emission (Hummel 1981). However, while this large body of evidence indicates that interactions can cause starbursts, it is not at all clear that they must do so. A large fraction of interacting galaxies show little or no increased star formation (Kennicutt et al 1987; Bushouse 1987), suggesting that the triggering mechanism for these starbursts must involve some complicated function of interaction geometry and galaxy properties.

ABSTRACT

We have detected a diffuse, continuum knot which may be a reminant nucleus, in the inner regions of Markarian 315. This knot is associated with a complex, ring-like structure in both the continuum and ionized gas emission. The occurance of this feature in combination with an extensive, ionized streamer, or tidal tail, and highly non-circular kinematics in the ionized gas, suggests that this galaxy has suffered a disruptive, tidal interaction whose influence extends well into the inner one kiloparsec region.

INTRODUCTION

Markarian 315 (Markarian and Lipovetskii 1971) (also IIZwl87), is a moderately luminous Seyfert 1.5 galaxy (Koski 1978). It has a redshift of 11,820 km s-1 relative to the galactic center and Mv= –21.6 (Sargent 1970). For this discussion, we adopt a scale of 0.57 h-l kpc arcsec-1.

Radio frequency images of Markarian 315 show that it is a steep spectrum source with a diffuse morphology and a total extent of 2.9 h-l kpc (Wilson and Willis 1980; Ulvestad, et al. 1981). This extended structure is the largest in the sample of Seyfert galaxies studied by these authors. (These typically ranged between 0.4 and 1.0 h-l kpc). It is, however, consistent with an extended starburst in the galaxy and the IRAS fluxes are also consistent with this interpretation (MacKenty 1989).

MacKenty (1986) discovered an extraordinary, 80 kpc, streamer of ionized gas emerging from near the nucleus, extending in a straight line for 60 kpc then bending back in a hook. He suggested two possible origins for this feature: a tidal interaction or a dormant radio jet.