Pulsars and neutron stars/History of observational results

Introduction
Some pulsars have only recently been discovered, but other pulsars have now been observed over many decades. Lyne et al. (2015) described 45 years of rotation of the Crab pulsar, which represents almost 5% of the entire lifetime of the pulsar. Not surprisingly, such data sets lead to many new discoveries.

Observational results can be divided into 1) completely new discoveries that were previously unexpected and 2) gradual improvements that occur as instrumentation becomes more advanced and datasets become longer.

Unexpected discoveries and completely new discoveries
After their discovery, pulsars were thought to be extremely stable rotators. This view was shattered in 1969 when the Vela pulsar suddenly rotated faster than before (Radhakrishnan & Manchester 1969 and Reichley & Downs 1969). This result was so unexpected that the astronomers involved first assumed that this event, now known as a “glitch event”, was caused by an instrumental problem. Now, more than 300 glitch events are known. These occur mostly in young pulsars, but even a millisecond pulsar has been observed to glitch (Cognard & Backer 2004). It is not only radio pulsars that glitch. Galloway, Morgan & Levine (2004) presented evidence for a glitch in a neutron star accreting from a Be companion star. Much more recently, the first anti-glitch event (in which the star suddenly slowed-down, instead of spinning faster) was detected in a magnetar (Archibald et al. 2013).

Even from the original pulsar discovery, it was clear that not all individual pulses from the same pulsar are identical. The first observations of the Crab pulsar were confusing. The pulsar didn’t seem to have a rotational period that could easily be determined. It was later realized that the Crab had been discovered through its giant pulses – individual pulses that are much brighter than the average pulse. When the much weaker, but more common pulses from the Crab were discovered then its rotational period finally became clear. The first giant pulses from a millisecond pulsar, B1937+21, were reported by Cognard et al. (1996).

Some pulses may be much brighter than the average, but other pulses seem to be completely missing. An early description of this “nulling” phenomenon was presented by Backer (1970). The phenomenon of "selective nulling" was presented by Bhat et al. (2007) in which the pulses seem to null in specific observing bands. Rankin & Wright (2008) reported on PSR J1819+1305 whose nulls seem to be periodic.

The individual pulses, or components of individual pulses, do not always perfectly align. Instead, the pulses often seem to be drift in pulse phase. This phenomenon known as “sub-pulse drifting” was described by  Backer et al. (1970) and Cole (1970). The sub-pulse drifting phenomenon is not necessarily constant. Edwards, Stappers & van Leeuwen (2003) detected sudden changes in the subpulse drifting in PSR B0320+39.

It is often thought (or perhaps hoped) that if a sufficient number of individual pulses are averaged together then the resulting pulse profile will be stable. In most cases, this is approximately true, but Lyne (1971) showed that the integrated profiles of some pulsars change abruptly and termed this "mode changing". In a few cases, it is now known that the pulsar emission can be undetectable for long periods of time. Such pulsars are known as “intermittent pulsars”. Kramer et al. (2006) reported on such a pulsar (PSR B1931+24 or J1933+2421) which is detectable for 5 to 10 days and then switches completely off for up to 35 days before switching back on. They showed that the spin-down rate changed between the two states. Hobbs, Lyne & Kramer (2010) and Lyne et al. (2010) studied long-term timing irregularities of a large number of pulsars and suggested that such states in emission and slow-down may be detectable in a large number of pulsars. In some pulsars the emission only changes slightly, whereas in other pulsars it switches completely off. Such changes have been suggested as being caused by an asteroid encountering a pulsar (Brook et al. 2014)

The average pulse profile provides a representation of part of the emission beam of the pulsar. Weisberg & Taylor (2002) demonstrated how they could use the geodetic spin precession of B1913+16 in order to map the beam shape in two dimensions. The pulsed emission then travels through the interstellar medium before being detected at the telescope. Many studies of the interstellar medium came from an analysis of a pulsar dynamic spectrum in which the pulse intensity is shown as a function of time and observing frequency. Stinebring et al. (2001) decided, for the first time, to take the two-dimensional Fourier transform of the dynamic spectrum and to their surprise discovered faint scattering events that showed up as sharply delineated features in the "secondary spectrum".

Pulsars were thought to be either rotation or accretion powered. Standard radio pulsars are accretion powered and X-ray binaries are in the accretion state. Papitto et al. (2014) reported the discovery of the first millisecond pulsar that changed between a rotation-powered radio pulsar and an accretion powered X-ray source.

Are all pulsars neutron stars? The answer depends on your definition of a pulsar. Hallinan (2007) and Hallinan et al. (2007) identified periodic bursts of coherent radio emission from an ultracool dwarf and showed that the star had many properties analogous to a pulsar. Similarly Kellett et al. queried whether CU Virginis was the first stellar pulsar.

Gradual improvements over time
Not much is known about a pulsar when it is first discovered. The discovery observation usually provides an estimate of the pulse period, its dispersion measure, an approximate flux density at the observing frequency and some information about the pulse shape. Over time it is possible to improve these determinations and also to determine other astrometric, pulse or orbital parameters. Manchester, Taylor & Van (1974) showed how pulsar proper motions could be measured. The first determination of a pulsar's parallax was obtained using an interferometer (Salter, Lyne & Anderson 1979). The parallax provides an measurement of the pulsar's distance. From distances and proper motions it is possible to determine a pulsar's two-dimensional velocity. Note that some of the pulsar parameters may be biased. For instance, a discussion of the implications of the Lunz-Kelker bias in pulsar observations was presented by Verbiest et al. (2012).

Usually it is quite straightforward to determine whether a pulsar is in a binary system. For most binary systems the Keplerian parameters can be determined. For some, highly relativistic systems it is also possible to determine post-Keplerian parameters. Pulsars have now been found with interesting companions. The first pulsar binary system is now known to be a double neutron-star system. Fruchter, Stinebring & Taylor (1988) identified that PSR B1957+20 is a millisecond pulsar in an eclipsing binary system (with a low-mass companion star). Thorsett, Arzoumanian & Taylor (1993) showed that the binary radio pulsar, PSR B1620-26 in the globular cluster M4 has a planetary companion making this a triple system. Pulsars are now known orbiting other neutron stars (and, in one case, another pulsar), white dwarfs, main sequence stars and have planetary companions. Recently a triple stellar system was discovered.

Most of the incremental improvements in our knowledge of pulsars get reported in catalogues. A list of the most commonly used catalogues is given below.


 * The second EGRET catalogue of high-energy gamma-ray sources was released by Thompson et al. (1995).
 * The third EGRET catalogue was presented by (Hartman et al. 1999)
 * Muno et al. (2003) presented a deep Chandra catalogue of X-ray point sources towards the Galactic centre.
 * Manchester et al. (2005) advertised the ATNF pulsar catalogue which has now grown to including more than 2500 pulsars.
 * The first Fermi gamma ray source list was presented by Abdo et al. (2009) using three months of observations.
 * Espinoza et al. (2011) presented a study of 315 glitch events.
 * Olausen & Kaspi (2014) - the McGill magetar catalogue
 * Acero et al. (2015) - the Fermi Large Area Third Source catalogue