The Early Years

With the announcement of the discovery of the first pulsar, the Molonglo radio telescope, operated by the University of Sydney, was ideally placed to follow up on this exciting result. The Cambridge pulsars were discovered with an array operating at 81.5 MHz, suggesting that pulsars had steep radio spectra, and the Molonglo telescope operated at the relatively low radio frequency of 408 MHz. It had large collecting area and so had high instantaneous sensitivity, necessary to record the rapidly fluctuating pulsar signals. By November 1968, it had already discovered nine pulsars, more than half of the then world total of 17 (Large, Vaughan & Wielebinski 1968). Included in these was the very important Vela pulsar, the first to be associated with a supernova remnant. Initially, astronomers at Parkes concentrated on detailed studies of the spectrum, polarisation and timing of pulsars, exploiting properties of the telescope such as frequency versatility, and polarisation and tracking capability. These observations were very successful, providing the first spectra of individual pulses (Robinson et al. 1968), the first observation of a period glitch (Radhakrishnan & Manchester 1969) and the genesis of the now widely accepted `magnetic-pole' model for the emission beam (Radhakrishnan et al. 1969). The first succesful search for pulsars at Parkes, reported by Komesaroff et al. (1973), began in 1973 and discovered eight new pulsars. Because of the lower instantaneous sensitivity of the Parkes telescope, this survey was among the first to rely on digital sampling of longer data sets and signal processing techniques to obtain the necessary sensitivity. Because of its higher frequency (750 MHz) and the use of multi-channel receivers, this survey was sensitive to high-DM pulsars. In particular, it discovered the highly luminous pulsar PSR B1641-45, which has a DM of about 480 cm^-3 pc, the highest known at the time. In a good example of synergy, the strengths of the Molonglo telescope and the Parkes telescope were combined to undertake the highly successful Second Molonglo pulsar survey (Manchester et al. 1978). This survey discovered 154 previously unknown pulsars, more than doubling the number of pulsars known at the time. The Molonglo telescope was used in a multibeam mode, giving eight adjacent beams in right ascension which increased the effective integration time to 44/cos $\,\delta$ sec. Improved front-end amplifiers and multi-channel receivers were also constructed specifically for this survey. Candidates from analysis of the Molonglo data were confirmed at Parkes. The Parkes telescope tracked up the $4^{\circ}$ declination width of the Molonglo beam with an effective 300-sec integration at each point. Data were searched in real time about the candidate parameters, thereby giving an improved declination, period and DM for confirmed pulsars. As shown in Fig. 1, the survey covered the whole sky south of declination $+20^{\circ }$ and detected a total of 224 pulsars, giving an excellent sample for statistical studies.

**Figure 1:** Pulsars detected in the Second Molonglo survey (Manchester et al. 1978). The dashed line marks declination $+20^{\circ }$ , the northern limit of the survey.
$\begin{figure} \begin{center} \centerline{\psfig{file=mol2.ps,width=150mm,angle=270}} \end{center} \end{figure}$

The long integration times and wide bandwidths available at Parkes make possible very sensitive surveys. One such survey was that of McConnell et al. (1991) which detected the first known extragalactic pulsars, in the Magellanic Clouds. One of these, PSR J0045-7319, until recently the only known pulsar in the Small Magellanic Cloud, was later shown by Kaspi et al. (1994) to be in a orbit about an optically identified B-star companion. The next major survey to be undertaken at Parkes was the 20-cm survey of Johnston et al. (1992a). This survey covered a strip along the Galactic plane with

$270^{\circ} < l < 20^{\circ}$ and

$\vert b\vert < 4^{\circ}$ , complementing a similar survey of the northern Galactic plane (Clifton et al. 1992). A bandwidth of 320 MHz centred at 1520 MHz was observed with an effective integration time per point of 2.5 min, giving a limiting sensitivity of about 1 mJy for pulsars with period greater than about 50 ms. A total of 100 pulsars were detected by the survey, with 46 being new discoveries. Included in them was the very interesting eclipsing binary pulsar PSR B1259-63 (Johnston et al. 1992b). This pulsar is in a 3.5-year highly eccentric orbit around a 10-M $_{\odot}$ Be star SS 2883, and was the first pulsar known to have a massive non-degenerate companion. Near periastron, the pulsar passes through the circumstellar disk of the Be star and is eclipsed for about 30 days. Significant changes in DM and rotation measure are observed before and after the eclipse, giving information on the properties of the circumstellar disk (Johnston et al. 1996). Although the Johnston et al. (1992a) survey had sensitivity to MSPs at about the 2.5 mJy level, none was detected. The main reasons for this were the high dispersion, scattering and background temperature along the Galactic plane, coupled with the low luminosity of most MSPs. Also, because of their great age, most disk millisecond pulsars are at large Galactic z-distances, comparable to or larger than the reach of most surveys. Consequently they have a nearly isotropic distribution on the sky. These considerations suggested that a lower-frequency search covering a large area of the sky would be more likely to detect a significant number of MSPs. The Parkes 70cm survey (Manchester et al. 1996, Lyne et al. 1998) was designed with these ideas in mind. The survey covered the whole sky south of the equator at a frequency of 436 MHz, with a sampling interval of 300 $\mu$ s and an observation time per point of 157 sec, giving it a limiting sensitivity of about 3 mJy. It detected 298 pulsars of which 101, including 17 MSPs, were previously unknown. Fig. 2 shows the period distribution of these pulsars. This figure highlights the fact that MSPs are a different population, quite distinct from the normal pulsars. As expected, the sky distribution of MSPs was close to isotropic, whereas the normal pulsars were clustered along the Galactic Plane. The large number of pulsars detected and the well defined survey parameters make this an excellent data base for studies of the Galactic distribution and birthrate of both normal pulsars and MSPs. Lyne et al. (1998) estimate that there are about 30,000 potentially observable MSPs with 400 MHz luminosity above 1 mJy kpc² and a similar number of potentially observable normal pulsars above the same luminosity limit in the Galaxy. After taking beaming into account, the corresponding birth rate for normal pulsars is one per 60 to 330 years, and for MSPs, one per 300,000 years.

**Figure 2:** Period distribution of pulsars discovered (full line) and detected (dotted line) by the Parkes Southern pulsar survey (Lyne et al. 1998)
$\begin{figure} \begin{center} \centerline{\psfig{file=sur70_prd.ps,width=120mm}} \end{center} \end{figure}$

Probably the most interesting pulsar discovered by the Parkes Southern survey was PSR J0437-4715, by far the nearest and strongest millisecond pulsar known (Johnston et al. 1993). This pulsar has a period of 5.75 ms, is a 5.74-day binary orbit with a companion of mass $\sim 0.3$ M $_{\odot}$ , and has a mean flux density at 430 MHz of more than 500 mJy. The strength of this pulsar makes possible very precise measurements of its polarisation and timing properties. As shown in Fig. 3, the pulsar has a very wide and complex profile covering more than 80% of the period with at least 12 identifiable pulse components and high polarisation, both linear and circular (Navarro et al. 1997). The position angle variation is complex, suggesting that the usual assumption of dipole magnetic field lines in the pulsar magnetosphere is not valid. Timing observations (Sandhu et al. 1997) have given the pulsar position with a precision of 50 $\mu$ as, the proper motion at the 2000- $\sigma$ level, and a value for the annual parallax, $5.6 \pm 0.8$ mas. This parallax corresponds to a distance for the pulsar about 30% larger than the value derived from the DM. The large proper motion ( $\sim 140$ mas yr^-1) results in an apparent acceleration of the pulsar, accounting for about 80% of the observed period derivative. It also changes the inclination of the pulsar orbit to the line of sight, resulting in a secular change in the projected size of the orbit. This change has been detected at the 20- $\sigma$ level, enabling a limit to be placed on the inclination angle of the orbit,

$i < 43^{\circ}$ , and hence improving the determination of the mass of the companion.

**Figure 3:** Mean pulse profile and polarisation parameters for PSR J0437-4715 (Navarro et al. 1997). The entire pulse period is shown. In the lower part of the figure, the upper line is the total intensity, Stokes I, the other solid line is the linearly polarised intensity,
L=(Q² + U²)^1/2, and the dotted line is the circularly polarised intensity, Stokes V. The line in the upper part is the position angle of the linearly polarised component.
$\begin{figure} \begin{center} \centerline{\psfig{file=0437.ps,width=120mm,angle=270}} \end{center} \end{figure}$

With hindsight, another very interesting pulsar discovered in this survey was PSR J2144-3933, originally believed to have a pulse period of 2.84 s. As part of a study of the pulse-to-pulse fluctuation properties of pulsars, Matthew Young realised that the true period of this pulsar was 8.51 s, by far the longest period known (Young, Manchester & Johnston 1999). This long period places the pulsar beyond the `death line' of most models for the pulse emission mechanism. The pulsar also has a very narrow pulse, less than $1^{\circ}$ of longitude. If this is typical of such long-period pulsars, they could form a large fraction of the total Galactic population.

Next Section: The Parkes Multibeam Pulsar
Title/Abstract Page: Finding Pulsars at Parkes
Previous Section: Introduction
Contents Page: Volume 18, Number 1

ASKAP

Public

Australia Telescope National Facility

The Early Years