On a time scale of some days, all pulsars show a remarkable uniformity of rotation rate. This is not surprising, since uniform rotation is exactly what is expected of a spinning body with a large stable moment of inertia and which is isolated in space. The angular momentum of the star can only change through the slowdown torque of the magnetic dipole radiation, or an associated material outflow, or, for the accretion-powered X-ray pulsars, the accelerating torque of in-falling material carrying angular momentum of a binary system. The effects on the radio pulsar are usually smooth and predictable: however, some very interesting irregularities in pulsar rotation have been observed, which are related to changes both within the interior of the neutron star and outside it.

These internal changes in pulsars appear to be spasmodic adjustments towards a slowly changing equilibrium state. For example, the rapidly rotating star will be appreciably oblate, and the equilibrium ellipticity will decrease during the slowdown; the crust is however extremely rigid and can only adjust to the changing ellipticity in a series of steps. The corresponding changes in moment of inertia might be large enough to be observable, since conservation of angular momentum will result in their being seen as steps in rotation rate. This effect is not in fact large enough to account for most of the observed repeated steps in rotation rate, and these are instead attributed to an interchange of angular momentum between the crust and the fluid interior.

The magnetars are a small group of neutron stars with extremely high effective dipole magnetic fields, comprising Anomalous X-ray Pulsars (AXPs) and Soft Gamma-ray Repeaters (SGRs). In both groups the rotation periods are long, and the slowdown rates are high, consistent with a high magnetic field and a short lifetime. The radiated energy is far higher than the rate of loss of rotational energy, and it is commonly assumed that the energy for the radiated X-rays and gamma-rays is derived from a decay of the magnetic field stored in the interior of the neutron star. The magnetars have been regarded as distinct from normal pulsars; it now appears that they may represent the extreme of a continuum.

The Soft Gamma-ray Repeaters (SGRs)

Cosmic gamma-ray bursts (GRBs) were first detected by satellites equipped with gamma-ray detectors, which were intended to monitor man-made nuclear explosions. The discovery was published in 1973 (Klebesadel, Strong & Olson 1973); since then some thousands of GRBs have been observed; many of these were measured from several satellites simultaneously, allowing positions to be determined from relative times of arrival. A typical GRB has a duration of some 10 s, and a rise time less than 1 s, but these are both widely variable; some bursts have structure within 1 ms.

The nature of supernovae

The obvious association between the Crab Pulsar and the remains of the supernova explosion of AD 1054 leads naturally to the suggestion that all pulsars originate in supernova explosions, and even to the speculation that all supernovae might produce neutron stars, which could become pulsars. This turns out to be an over-simplification, and it is necessary to explore the nature of supernovae in some detail before their relation to pulsars can be pursued.

In 1921 Lundmark pointed out that the nova observed by Hartwig in 1885 in the constellation of Andromeda was probably within the Andromeda Nebula itself, and hence very distant and very bright (see a centenary review by de Vaucouleurs & Corwin 1985). He showed that there were many cases of these extremely powerful novae, and he was the first to associate the Crab Nebula with the Chinese records of the bright star that appeared in AD 1054. The physical significance of these enormous outbursts was appreciated by Baade & Zwicky, who first used the word ‘supernova’ in their publication of 1934. They made four very remarkable deductions from the observations:

1. (1) the total energy released was in the range 1051 to 1055 ergs;

2. (2) the remnant could form a neutron star;

3. (3) cosmic rays could have their origin in supernovae;

4. (4) supernova explosions could give rise to expanding shells of ionised gas.

Young pulsars

Although it is now accepted that most neutron stars are born in supernova explosions, only a small number of the known pulsars are clearly associated with visible supernova remnants (SNRs) (see Figure 8.7). This is, of course, entirely consistent with the difference between the lifetimes of a typical pulsar (about 106–107 years) and a supernova remnant (about 104–105 years). Furthermore, most of these associated pulsars are obviously young, as seen from their large period derivatives Ṗ. Most have short periods, with P < 100 ms. Their small characteristic ages, P/2Ṗ, are in sharp contrast with those of the other short-period pulsars, the millisecond pulsars (Chapter 10), which have very small period derivatives, are much older, and are the product of a long evolutionary history.

In this chapter we concentrate on two of the youngest, the Crab Pulsar and the Vela Pulsar, both of which are associated with SNRs and both of which have been studied in great detail. These young pulsars are to be found in the top left of the P–Ṗ diagram (Figure 8.7), and their progress across this diagram is expected to bring them into the upper part of the general population within 106 years. Their youthful characteristic of a large power output makes them easy to observe. Both are detected as radio, optical, X-ray and gamma-ray sources, and the evolution of their rotation rate has been monitored in detail for over 40 years.

More than 40 years after the discovery of pulsars, and despite many attempts to assemble and analyse the very detailed observational data, it is still not possible to give a clear exposition of the processes by which pulsars emit beams of radio waves. The later discovery of intense gamma-ray emission has proved more amenable to analysis, and the observed characteristics can be related to the particle energies and the geometry of the magnetosphere. In both radio and gamma emissions, the source of energy is the enormous electric field which is induced by the rapid rotation of the highly magnetised neutron star. This electric field accelerates electrons and positrons to high relativistic energies; this occurs in two emission regions, involving respectively the open field lines close to the polar cap, and the outer gap. The high-energy photons emitted from the outer gap are curvature or synchrotron radiation from these primary relativistic particles; in contrast, the radio emission from both regions is coherent emission from a plasma created by the primary particle stream.

The two locations

The rapid rotation and strong dipole field of a typical pulsar give a maximum available potential of order 1014–1015 volts. In a fully developed, co-rotating and static magnetosphere this would be shielded by the electrostatic field of the charged particle density nGJ, as envisaged by Goldreich and Julian.

Characteristic ages

The characteristic age of a pulsar, which is derived from its present-day rate of slowdown (Section 5.10), is an unreliable indicator of its actual age since birth. Not only is it unsafe to assume that the rate of rotation at birth was much larger than at present, but also the slowdown law itself may change during the lifetime of the pulsar. Labelling a pulsar as ‘young’ is only secure for the small number actually identifiable with dateable supernovae. The prime example is the Crab Pulsar, but there is a growing list of identifications with supernova remnants (SNRs) which can be dated more or less precisely. Among these is the SNR 3C58, long known as a radio source and tentatively identified with the supernova of AD 1181.

Another opportunity for reasonably accurate dating is through measurement of the proper motion of the pulsar away from the centre of the remains of the supernova explosion, as described in the next section. There are few such examples of reliable dating; even when a pulsar with small characteristic age can be associated with an SNR it may only be possible to assign an upper limit to the age of the SNR itself. In this chapter on young pulsars we therefore use characteristic ages as the best generally available indication of youth, and tentatively consider all characteristic ages below 100 000 years as young.

Apart from the thermal radiation observable in X-rays from the surface of a neutron star, the sources of electromagnetic pulsar emissions are located in two distinct regions of the magnetosphere: the polar cap and the outer gap. These are two regions which are defined by the simple large-scale geometry of the magnetosphere outlined in Chapter 2 (although some further geometry is needed to extend the model for large angles between the rotation and magnetic axes; see Kapoor & Shukre 1998).

Outside these two regions the large electric field induced by the rotating magnetic dipole is shielded by the ambient charged plasma. The outer gap is a surface located adjacent to the open magnetic field lines, and extending to and beyond the light cylinder; electrons and positrons accelerated within this gap generate gamma rays at the highest observable energies. The mechanism for coherent radio emission is not understood, but the location can be deduced from the rich and sometimes complex integrated pulse profiles described in Chapter 15; most radio emission from normal pulsars originates in the polar cap, while some, especially from young pulsars, also originates in the outer gap.

Relativistic aberration has a major effect on the outer gap radiation, which occurs at radii approaching the velocity-of-light cylinder; it has almost no observable effect on polar cap emission, which occurs comparatively close to the star.

Since Jocelyn Bell saw the first record of radio pulses as a mysterious ‘piece of scruff’ on a chart recording, bringing the neutron star from a remote world of theory into reality, we have been able to investigate the structure of the insides and the outsides of neutron stars in astonishing detail. Neutron stars are seen to be an essential component of the Galaxy, both in the young stellar population concentrated towards the plane and in the extended population of older stars. We now have proof of their origin in supernova explosions, and we have good evidence of their evolution in binary star systems. The combination of pulsars and binary X-ray sources has opened a whole new branch of astrophysics concerned with accretion onto condensed stars and the evolution of binary systems.

Pulsars have provided an entirely new means of investigating the interstellar medium, measuring the total electron content through dispersion, irregularities in electron density through scintillation, and the magnetic field through Faraday rotation. The millisecond pulsars have provided celestial clocks of unparalleled regularity, enabling the most sensitive tests of gravitation theory to be made. Building on these successes, the field of pulsar research has widened and become even more active, extending to the whole electromagnetic spectrum. Radio remains the most productive spectral region, but there has been a revolutionary advance in high-energy gamma-ray pulsar astronomy with the advent of the Fermi LAT satellite telescope.

The magnetic fields which are first brought to mind in any discussion of pulsars must surely be the enormous fields at the surface of the neutron star itself. These fields are of the order of 1012 gauss (108 tesla) for most pulsars, and 108 gauss for the older, millisecond pulsars. Pulsar radio waves may, however, be used as a means of measuring the magnetic field on the line of sight to the observer; this interstellar field is, by way of extreme contrast, of order 1 to 10 microgauss. The observational link between these extremes is simple: the high field of the pulsar is responsible for the linear polarisation of the radio emission, while the Faraday rotation of the plane of polarisation in interstellar space is a measure of the interstellar magnetic field along the line of sight.

Optical and radio observations

The existence of an interstellar magnetic field was first demonstrated by the observation of the polarisation of starlight, notably by Hall & Mikesell (1950) and by Hiltner (1949). A small degree of linear polarisation was found, which was common to several stars in the same general direction. The polarisation was due to scattering in interstellar space by particles which are elongated, so that they scatter anisotropically, and which are aligned with their long axes perpendicular to the Galactic magnetic field.

Broadly speaking, when a normal star has exhausted its sources of energy, it collapses under its own gravity. A star with the density of normal matter then ends up in one of three possible states: white dwarf, neutron star or black hole. The extent of the collapse depends on the mass of the progenitor star; the most massive become black holes, and the least massive become white dwarfs. The progenitors of neutron stars have a limited intermediate range of mass, about 8 to 20 solar masses (M⊙ = 2 × 1033 g). It follows from the statistical distribution of stellar masses that more than 95% of stars end their lives as white dwarfs without further collapse.

The formation of a white dwarf is a smooth and continuous process; as nuclear fuel becomes exhausted, a core grows within an expanding outer shell, the total gravitational collapse of the core being prevented by the pressure of electron degeneracy. If and when the mass of the core exceeds 1.4 solar masses (M⊙), this pressure is insufficient to resist the increasing force of self-gravity and to prevent the further collapse of such a degenerate core to a neutron star. At this second stage of collapse it is the pressure of neutron degeneracy that balances gravity. The collapse to a neutron star is catastrophic; within a few seconds a large proportion of the gravitational potential energy of the star is released, and the event is observed as a supernova (Type II; see Chapter 14).

The stream of research and publications in pulsar astronomy has spread to a flood tide, encompassing a wide range of observations and astrophysics. In 1967, when the first pulsar was discovered, digital techniques, wide bandwidth radio receivers, space-based X-ray and gamma-ray telescopes were all unheard of. Observations are now expanding as fast as technical developments allow, and we are already looking forward to another major step forward, the building of the international Square Kilometre Array. Recent years have seen the outstanding success of X-ray and gamma-ray astronomy, extending to energies in the GeV and TeV regions. We have seen spectacular advances in pulsar timing and astrometry, leading to the most stringent tests of relativity theory, while an astonishing range of astrophysics has become accessible through pulsar astronomy, from the cold condensed matter of the neutron star interior and the extremely high energy of the surrounding magnetosphere, to the detailed structure of the interstellar medium.

Our intention in this new edition is to provide a guide rather than an encyclopaedia. Both of us are physicists and hands-on observers rather than theorists, and we naturally concentrate on techniques and discoveries, and on the interpretation of observations. Nevertheless we present the basic astrophysics, supplemented by references to papers which will lead to more complete explanations and into the more abstruse physics of, for example, condensed matter and relativity.

The distances of the nearest stars can be obtained from their parallax, which is the apparent annual cyclic movement of position due to the Earth's orbital motion round the Sun. At distances greater than about 1 kiloparsec, stellar distances may be inferred from their apparent brightnesses, since the absolute brightness of a star is generally available from its spectral type. The distances of some of the closest neutron stars and pulsars may similarly be obtained from measurements of parallax, with accuracies comparable to the best optical measurements of other stars. A new and remarkably accurate measurement of kinematic or dynamical distances is available from accurate timing of pulses from some millisecond pulsars, and also from the orbital periods of some binary pulsars.

The observed intensity of a pulsar, in contrast to that of a visible star of known type, is of little use as an indicator of distance, since the intrinsic luminosity is a variable quantity, both from time to time and from pulsar to pulsar. Fortunately, we have available a remarkably direct measurement of distance instead, due to the frequency dispersion of the group velocity of radio waves in the ionised interstellar medium, which leads to an easily measurable frequency-dependent delay in pulse arrival time for all radio pulsars. The magnitude of this delay is directly proportional to the integrated column density of free electrons along the line of sight (see Section 3.11).

A distinct population

The millisecond pulsars are in a different category from the general population of ‘normal’ pulsars. The majority of pulsars are following a simple course of evolution, from a birth in a supernova, through a slowdown from a rotational period at birth of some tens of milliseconds to a death at around one second when the radiation ceases or becomes undetectable. The millisecond pulsars constitute a separate and much longer-lived population, which originate from the general population of normal pulsars as the result of gravitational interactions with binary partners; most still have their companions, but some have lost them or are possibly in the process of losing them, thus becoming solitary millisecond pulsars.

Most millisecond pulsars have rotational periods less than 10 ms, although a more useful definition includes periods up to 30 ms. Figure 11.1 shows their position in the P/Ṗ diagram, which also shows all known binary systems as circles. Although there is a distinct gap between normal and binary pulsars in this diagram, there are some pulsars within the gap; most of these have the characteristics of millisecond pulsars such as a comparatively low slowdown rate, and they have probably experienced the same binary interactions.

Binary stars

It is remarkable that almost all of the neutron stars associated with X-ray sources are members of binary systems, while only one in ten of the radio pulsars is a binary.

Optical scintillation is familiar as the twinkling of stars, and as the shimmer of distant objects seen through a heat haze. At radio wavelengths scintillation is encountered in many different circumstances, because there are many kinds of radio transmission paths which contain the necessary phase irregularities: the solar corona, for example, contains an irregular outflowing gas, which disturbs radio waves passing through it from distant objects to the Earth. The effects of this may be thought of either as refraction or as diffraction; in more general terms the waves are scattered, giving rise to an angular spread of waves and to subsequent fluctuations in wave amplitude, which are seen as intensity variations as the Earth moves relative to a pattern of irregularities. Similar effects are observed in the passage of radio waves through the Earth's ionosphere.

At the time of discovery of pulsars, the known examples of radio scintillation gave a rapid fading pattern, not very different from the visible twinkling stars. The comparatively slow, deep fading of radio signals from the pulsars was an entirely new phenomenon, which was first recognised as a form of scintillation by Lyne & Rickett (1968). The basic analysis of scintillation in terms of random refraction in the interstellar medium was presented by Scheuer (1968). He showed that the fluctuations should have a fairly narrow frequency structure, whose width should depend on the distance of the pulsar.