The Parkes Multibeam Pulsar Survey

The Parkes multibeam receiver, while primarily designed for HI surveys (Staveley-Smith et al. 1996), is a superb instrument for pulsar surveys. Its 13 beams allow the sky to be covered roughly 13 times as fast, or alternatively, much longer to be spent on a given point. Also, its receivers have excellent sensitivity, with an average system noise of only 21 K. With major contributions from Jodrell Bank Observatory and Osservatorio Astronomica di Bologna, a filterbank system capable of handling the data from the 13 beams was installed at Parkes in early 1997 and the Parkes multibeam pulsar survey commenced in August 1997. The survey is covering the region

$260^{\circ} < l < 50^{\circ}$ with

$\vert b\vert < 5^{\circ}$ . The filterbank has 96 3-MHz channels for each polarisation of each beam and all outputs are one-bit digitised at 250 $\mu$ s intervals and recorded to tape. Each pointing of the 13 beams is of 35 min duration, giving a sensitivity for long-period pulsars away from the hot regions of the Galactic background of about 0.15 mJy. This is about seven times better than the previous best survey of this type (Johnston et al. 1992a), and so a large increase in the number of detected pulsars was expected. These expectations have already been fully realised. With about 80% of the survey completed, more than 570 previously unknown pulsars have been discovered, making this by far the most successful pulsar survey ever. When finished, the survey will come close to doubling the number of known pulsars. Fig. 4 shows the distribution of known pulsars projected on to the Galactic plane, where distances have been computed using the Taylor & Cordes (1993) electron density model. In contrast to the previously known pulsars which are clustered around the Sun, many of the multibeam pulsars are at large distances, with some apparently on the other side of the Galactic Centre. There is some indication of a deficit of detected pulsars within a couple of kpc of the Galactic Centre. The electron density model is not well determined at these large distances though, and the distances may have systematic biases. Also, many of the multibeam pulsars are concentrated in spiral arms, but this may simply be a result of the increased model electron density in the arms. The multibeam sample will be important in helping to refine the electron density model.

**Figure 4:** Distribution of known pulsars projected on to the Galactic plane. The position of the Sun is marked by $\odot$ and the Galactic Centre by +. The distribution of interstellar free electrons according to the Taylor & Cordes (1993) model is shown as a grey-scale. The dark spot near the Sun is the Gum Nebula.
$\begin{figure} \begin{center} \centerline{\psfig{file=pm_xy.ps,width=120mm}} \end{center} \end{figure}$

As shown by Fig. 5, pulsars detected by the multibeam survey are on average much younger than previously known pulsars. They include the three pulsars with the strongest known surface dipole magnetic fields (Camilo et al. 2000a). One of these pulsars, PSR J1119-6127, has a characteristic age of only 1700 years and is associated with what appears to be a previously uncatalogued supernova remnant (Crawford et al. 2000). Another, PSR J1814-1744, has the relatively long period of 3.97 sec, but a very rapid spin-down rate giving it an implied surface field strength of

5.5 x 10¹³ G. These parameters place the pulsar near the so-called `anomalous X-ray pulsars' (AXPs) on the $P - \dot P$ plane (Fig. 5). AXPs are believed to be slowly rotating neutron stars, but they have no detectable radio emission. On the other hand, PSR J1814-1744 has no detectable X-ray emission (Pivovaroff, Kaspi & Camilo 2000). The reason(s) for these very different properties are not well understood.

**Figure:** Distribution of pulsars and anomalous X-ray pulsars (AXPs) in the $P - \dot P$ plane. Binary systems are indicated by a circle around the point. Lines of constant pulsar characteristic age,
$\tau_c = P/(2 \dot P)$ and surface dipole magnetic field strength,

$B_s \propto (P\dot P)^{1/2}$ are indicated. The spin-up line, representing the minimum period attainable by accretion from a binary companion, is also shown.
$\begin{figure} \begin{center} \centerline{\psfig{file=pm_ppdot.ps,width=100mm}} \end{center} \end{figure}$

Only one millisecond pulsar has been discovered so far. Although this is somewhat surprising given the parameters of the survey, there are several possible contributing factors. The search so far has been concentrated at low Galactic latitudes where dispersion and scattering are large. This limits the maximum distance at which typical MSPs can be detected and so the volume of the Galaxy searched so far is relatively small. Another contributing factor is that only `unaccelerated' searches have been performed so far. Especially with the rather long observation time of this survey, this limits our senstivity to millisecond pulsars, most of which are in binary systems. A third factor is that algorithms for dealing with interference were not optimal in the early stages of processing. This mainly affects MSP detections, since most of the searched frequency space corresponds to millisecond periods. All of these factors are being overcome and we expect to detect more MSPs with future observations and data processing. Eight of the newly discovered pulsars are members of binary systems. Five of these are in near-circular orbits with companions which are probably white dwarfs (Camilo et al. 2000c). These systems differ from most known white dwarf binaries. Except for the one MSP detected, the pulsar periods are relatively long, lying between 45 and 90 ms, and the companions are heavy with minimum mass between 0.15 and 0.9 M $_{\odot}$ . One of the binary pulsars, PSR J1811-1736, is in a highly eccentric 18-day orbit and is very probably a double neutron-star system, the first to be discovered in the southern hemisphere (Lyne et al. 2000). In contrast to PSR J1811-1736 which has a characteristic age $\sim 10^9$ yr, PSR J1141-6545 is a young pulsar (

$\tau_c \sim 1.5 \times 10^6$ yr) in a much tighter ( $P_b \sim 4.1$ h) and eccentric orbit (Kaspi et al. 2000). Precession of the longitude of periastron has been observed for this system, and interpreting this as due to the effects of general relativity gives a value for the total mass of the system of

$2.300 \pm 0.012$ M $_{\odot}$ . The pulsar and orbit properties suggest that the companion is a heavy white dwarf formed before the supernova explosion that created the pulsar. This is unusual. In most binary systems the neutron star is formed from the heavier binary companion which evolves faster. As shown in Fig. 6, PSR J1740-3052 is in a highly eccentric long-period orbit. The interesting thing about this system is that the minimum companion mass is 11 M $_{\odot}$ , implying that the companion is either a massive star or a black hole. Unfortunately the pulsar lies close to the direction of the Galactic Centre and probably at about the same distance, so optical searches for the companion are unlikely to be productive. However, 2.2 $\mu$ m infrared observations with the Siding Spring 2.3-m telescope and the Anglo-Australian Telescope have revealed a K-supergiant star whose position agrees with that of the pulsar to better than 0.3 arcsec (Stairs et al. 2000). The infrared spectrum of this star shows Brackett- $\gamma$ emission, consistent with the presence of a compact binary companion, and the star's colours are consistent with a distance comparable to that of the pulsar. Furthermore, DM and rotation measure changes were observed over the last periastron passage, in February 2000. All of these observations point toward this star being the binary companion. However, there a couple of puzzling features. The pulsar comes to within 1.25 stellar radii of the companion star at periastron. One might expect the radio emission to be eclipsed by the stellar atmosphere or wind, but no eclipses are observed. Also, it should raise large tides on the companion, causing a large precession in the longitude of periastron. This is not observed. Either we do not understand winds and tides in supergiant stars very well, or all the other observations are misleading and the companion is really a black hole. Although the latter is an attractive option (this would be the first known neutron star - black hole system), the former is more likely.

**Figure 6:** Variations in apparent solar-system barycentric period of PSR J1740-3052. The fitted line is for a binary model with orbital period 230.0 days and eccentricity 0.579 (Stairs et al. 2000).
$\begin{figure} \begin{center} \centerline{\psfig{file=1740_per.ps,width=120mm}} \end{center} \end{figure}$

Next Section: Millisecond Pulsars in 47
Title/Abstract Page: Finding Pulsars at Parkes
Previous Section: The Early Years
Contents Page: Volume 18, Number 1

ASKAP

Public

Australia Telescope National Facility

The Parkes Multibeam Pulsar Survey