Introduction: the energy problem

The discovery of quasi-stellar objects (QSOs or quasars) in 1963 represents a landmark in observational astronomy. Thanks to a coordination between optical and radio astronomers, it was possible to discover a new and important class of astronomical objects. Because this text book is all about quasars and related phenomena, it will not be out of place to begin at the beginning of the subject and to review briefly how these remarkable objects were first discovered.

The science of radio astronomy really began after the end of World War II, when some of the scientists and engineers engaged in wartime radar projects used their know-how to follow up the pioneering works of Karl Jansky in the 1930s and Grote Reber in the early 1940s. Thus radio dishes and interferometers appeared in England and Australia, at Jodrell Bank, Cambridge, Sydney and Parkes.

The early observations revealed the existence of cosmic radio sources and by the mid- 1950s it became an accepted fact that radio galaxies exist. The nature of their radiation was non-thermal, and its polarization properties indicated that its origin lay in the synchrotron process. As we will discuss in Chapter 3, in this process radiation comes from electrons accelerated by a magnetic field. Thus a typical radio source has as its energy reservoir the dynamical energy of relativistic particles and magnetic field energy.

In 1958 Geoffrey Burbidge drew attention to the enormous size of this energy reservoir.

The year 1963 marks a watershed in extragalactic astronomy. The optical identification of radio sources 3C 273 and 3C 48 and the measurement of their redshifts demonstrated to astronomers the existence of a new class of energy sources that have a star-like appearance, yet produce luminous energy at a rate comparable to a galaxy of a hundred billion (1011) stars.

The quasi-stellar objects (QSOs) or ‘quasars’ as these sources came to be called, arrived on astrophysicists' plates just about when they had digested the long-standing mystery of stellar energy. By the 1960s, the problems of stellar structure and evolution built on the pillars erected by Eddington, Milne, Chandrasekhar, Bethe, Lyttleton, Schwarzschild and Hoyle had been tackled successfully, thanks to the advent of fast electronic computers. The quasars, however, presented challenges of an altogether different nature. How could so much energy come with such rapid variability out of such a compact region and be distributed over such a wide range of wavelengths?

The classic book Quasi-Stellar Objects by Geoffrey and Margaret Burbidge, published in 1967, captured this early excitement and posed the numerous challenges of quasar astronomy very succinctly. Now, three decades later, we have the benefits of vast progress in the techniques of observational extragalactic astronomy and the intricate sophistication of ideas in high energy astrophysics. Yet it is fair to say that the understanding of quasars and the related field of active galactic nuclei (AGN) has not reached the same level of success that stellar studies had attained thirty years ago.

Introduction

The early ideas of Hoyle and Fowler (1963) concerning gravitational collapse to a compact object that would serve as an energy reservoir for a quasar found a modified expression in the black hole accretion disk paradigm, a few years later. This paradigm had been invoked and worked well in the understanding of binary X-ray sources in the Galaxy. In the binary star context the compact member is taken to be either a neutron star or a black hole with mass of stellar order. For quasars and AGN, the compact masses would have to be several orders of magnitude higher, as already pointed out by Hoyle and Fowler (1963). The scenario here had therefore to explain how such objects form in the first place, how they generate an accretion disk and jets, and how and with what efficiency is the gravitational energy converted to the observed radiant energy.

In this brief review of the current thinking on the subject we shall follow the excellent account given by Rees (1984) whose basic tenets have remained more or less the same since then.

The formation of a massive black hole

As first pointed out by Hoyle and Fowler (1963), the energy source of a quasar or AGN is gravitational and could arise from a highly collapsed object or a massive black hole. This much is broadly agreed by most workers in the field. The question is, in the first place how does a collapsed massive object come about?

Introduction

The continuum radiation from AGN stretches over the entire range of the electromagnetic spectrum, from the radio to the high energy γ-ray region, where pair production by photons becomes important. The continuum spectrum has an overall complex shape, but it can often be approximated by a simple power law form over fairly wide wavelength intervals. The radiation is produced in elementary processes like synchrotron emission and bremsstrahlung, and is modified by scattering, absorption and reemission. In this chapter and the next we shall consider some aspects of radiation processes which are important to the basic understanding of the continuum spectrum. The discussion will be brief, and the emphasis will be on developing concepts, summarizing important results and providing them in such a form that they can be directly applied to situations pertinent to quasars and AGN. The subject has been treated in detail in a pedagogic manner by Jackson [J75] and Rybicki and Lightman [RL79]. The more advanced and formal aspects have been covered by Blumenthal and Gould (1970) and an excellent summary, especially of the synchrotron process, with application to AGN, may be found in Moffett (1968). Our treatment and notation owe much to these sources.

In the present chapter we will consider mainly synchrotron radiation and some consequences of relativistic radiation. The other processes important to AGN and quasars will be discussed in Chapter 4.

Introduction

The 1993 edition of the catalogue of quasi-stellar objects by Hewitt and Burbidge contains 7315 quasars and BL Lacs, which is an order of magnitude more than the number included in their first catalogue, published in 1977. Many more quasars have appeared in the literature since the publication of the 1993 catalogue, and the number continues to increase as observational and data analysis techniques become more efficient. Another catalogue due to Hewitt and Burbidge (1991) contains 935 galaxies that have nuclei with properties similar to those of quasars. The large numbers of quasars and galaxies with active nuclei (AGN) were discovered because of their peculiar colours, morphology or strong emission in the radio, X-ray or infrared bands, or in one of the surveys specially designed to identify objects that have properties different from Galactic stars and ‘normal’ galaxies. A small number of quasars have also been found through their variability or lack of proper motion and these provide samples independent of the biases inherent in the other techniques.

We shall consider in this chapter the various techniques used in the identification process, but first we provide brief definitions of some of the types of object that appear in the book. The definitions do not provide a water-tight division into physically distinct classes of object, and there are many marginal cases that could belong to more than one class. We shall elaborate on the definition and properties of the different kinds of object in other chapters.

Introduction

We have seen that there are many kinds of quasars and AGN, with observed properties so different that they could belong to fundamentally different populations. However, seemingly different kinds of object sometimes share many properties which clearly set them apart as a class from other kinds of object. It is tempting to think of them as belonging to a single population, the differences arising because of different values taken on by one or more basic parameters.

An example of this kind of behaviour is provided by radio-quiet and radio-loud quasars. We have seen in Section 9.8 that only a small fraction of all quasars are radio-loud. The ratio of radio to optical luminosities has a bimodal distribution, with quasars termed as radio-loud having distinctly higher values of the ratio relative to the radio-quiet variety. However, the emission line properties of these two types of object are so similar that it is very difficult to tell them apart from spectroscopic criteria alone. Radio-loud quasars have higher X-ray luminosity on average than radio-quiet quasars, and flatter X-ray spectra below a few keV, but there is a great deal of overlap in the distribution of these properties. It is tempting therefore to imagine that these radioloud and radio-quiet quasars belong to a single population. It was in fact suggested by Scheuer and Readhead (1979) that all radio-quiet quasars have relativistic radio jets in their cores, and that they appear to be luminous flat spectrum radio sources when the jet makes a small angle with line of sight to the observer.

In the previous chapter we considered how the broad band X-ray emission from quasars and AGN is related to the emission from other bands. The results from the large database on broad band fluxes are important in providing some pointers to the processes by which the X-ray emission is produced, but a detailed understanding requires observation of the spectral shape of the continuum, and any emission lines that may be present. Owing to the small collecting area of the X-ray telescopes that have so far been used, and the high energy per photon, the total number of photons received from a source is rather limited. It is therefore possible to obtain reliable spectra only for nearby AGN, which in spite of their moderate luminosity have a relatively large flux, and for the most luminous quasars. In spite of this limitation, spectral shapes have been sufficiently well established for a detailed comparison to be made with the predictions of the models described in Chapter 4. In this chapter we will consider the observation and modelling of the spectra from discrete sources as well as the diffuse background radiation in the X-ray and γ-ray domains.

X-ray spectra

The X-ray continuum for quasars and AGN in the 2–10 keV energy range can usually be well approximated by a power law. It is found that the spectral index is distributed in a narrow range in the case of AGN, but shows a somewhat wider distribution in the case of quasars.

Introduction

About 10 per cent of giant elliptical galaxies and quasars are radio loud, which means that they have a radio luminosity of ≈ 1041–1046 erg sec-1 in a band extending from ≈ 102 MHz to ≈ 10 GHz. This corresponds to a 5 GHz luminosity ≿ 1025 W Hz-1 and, for the observed range of redshift, a 5 GHz flux ≿ 102 mJy. The ratio of the radio luminosity to the optical luminosity in these objects is ≿ 10. The criteria for separating radio-loud objects from radio-quiet objects are to some extent subjective; they identify a somewhat vague boundary on the brighter side of which rather spectacular manifestations of radio emission from active galaxies become evident.

Radio objects that do not belong to the radio-loud class are not radio quiet. Our Galaxy has a radio luminosity of ≈ 1037 erg sec-1, while Seyfert galaxies show a whole range of radio luminosity up to the boundary of the radio-loud class. The non-thermal emission in normal galaxies can be traced to their centres, supernovae, active stars, energetic particles in the magnetic field of the interstellar medium (ISM) and so on. In the more luminous Seyfert galaxies, the radio emission originates in the supernovae in starburst regions, though in the most luminous cases it is possible that mechanisms similar to those in radio-loud galaxies are in operation.

The radio properties of radio-loud quasars, which are in the minority in the whole quasar population, are quite distinct from those of radio-quiet quasars.

Introduction

The luminosity function of a population of discrete sources describes the distribution of the objects in space as a function of their luminosity. Apart from luminosity, such a function may depend on many other properties, the environment and the evolutionary state of the universe. Surveys for a particular kind of object provide the surface density of the objects in the sky as a function of their magnitude, redshift and perhaps some other attributes. Using redshift as the distance indicator, the surface density can be deprojected to provide their number per unit volume of space, which is the more fundamental quantity. Owing to the limited data available, the deprojection involves a number of techniques, assumptions and models, some of which we will describe below.

The space density of quasars is a fundamental quantity because it could help link the quasar phenomenon to other objects, such normal galaxies, in the universe. On the one hand, the properties of quasars are very similar to those of the active galactic nuclei (AGN) of Seyfert galaxies and radio galaxies. On the other hand, nebulosities which resemble galaxies have been discovered around some low redshift quasars (see Chapter 8). Given these two facts, it is generally believed that quasars represent the extreme end of the active galaxies population, in which the luminosity of the active nucleus overwhelms the luminosity of the rest of the galaxy. Active galaxies similarly are considered to be an extreme subset of all galaxies. From the space density of the different kinds of active objects it will be possible to determine whether this is indeed the correct picture.

Introduction

In the earlier chapters of this book we have covered the present-day understanding of quasars and AGN, based on certain paradigms. We may refocus on these to begin with.

• Quasars and AGN are extragalactic phenomena that form natural steps in the overall scenario of galactic evolution.

• The redshifts of these objects are of cosmological origin.

• The primary source of production of the energy emitted by these objects is the gravity of a highly collapsed supermassive object, which is idealized as a spinning black hole with an accretion disk.

• The ejection of matter from the central region (in the form of jets) is relativistic.

• Relativistic beaming and orientation effects play a key role in explaining certain crucial observed features of these objects.

Astronomy has developed through paradigms, some of which were correct right from the beginning whereas others have had to be corrected, modified or abandoned. From the early days of the geocentric theory to the colliding galaxies hypothesis of radio sources, astronomy also has a history in which a majority have enthusiastically subscribed to a mistaken paradigm, which has thus become dogma. It is against this background that we now look at the above assumptions with the eyes of a sceptic.

Quasars, AGN and galaxies

Unlike stellar evolution, astronomers are still far from piecing together a scenario of galactic evolution in which all the different types of galaxy would find a natural place.

Introduction

The third decade of the twentieth century brought about a significant advance in our perception of the universe. In particular, it became clear to the astronomer that our Milky Way galaxy is but one amongst many such galaxies. And, what is more, the vast collection of galaxies appeared to be boundless (even as it does today). Towards the end of the decade Edwin Hubble came up with the remarkable discovery that the spectra of galaxies appear to be shifted towards the red end of the spectrum, and that the shift in a given galaxy is proportional to the distance of the galaxy from us. Hubble's actual observations showed that redshift increases with increasing faintness of the galaxies. Distance is inferred from the inverse square law of illumination. If we interpret the spectral shift as due to the Doppler effect, the corresponding radial velocity is then proportional to distance. The constant of proportionality is known as Hubble's constant and its value is denoted by H0.

Since the discovery of Hubble's law in 1929, evidence has steadily grown that, barring very few exceptions, the phenomenon of redshift is universally found in all extragalactic objects. The theoreticians have had no difficulty in explaining the phenomenon; in fact, seven years before Hubble, A. Friedmann had found world models as solutions of Einstein's equations of general relativity wherein the property of redshift arose naturally.

Introduction

X-ray emission has been found to be a universal characteristic of all AGN and quasars. The X-ray luminosity of these objects is in the range ≈ 1042-1048 erg sec-1 and accounts for a considerable fraction of their bolometric luminosity. The X-ray flux shows large amplitude variability on a time scale of days, hours and in some cases hundreds of seconds. This implies that the X-ray emission has its origin very close to the central objects. For energies ≿ 1 keV the X-ray spectra appear to have a simple power law shape, which is modified at lower energies owing to absorption at the source. Improved energy resolution has shown the presence of emission and absorption features, which prove to be important diagnostics of conditions close to the active nucleus. At energies below a few keV, AGN and quasars contribute a significant fraction of the observed X-ray background (XRB). The observed spectra of AGN and quasars in the 2–10 keV range appear to be different from the spectrum of the X-ray background and a way has to be found to reconcile AGN spectra with the XRB, if the AGN indeed contribute a substantial fraction to it at high energy.

In the following sections we will discuss some important X-ray missions, elementary steps used in the analysis of data, X-ray surveys, luminosity functions and broad band properties of AGN and quasars. We will consider X-ray spectra in Chapter 11.