ALFA Finds its First Millisecond Pulsar!

The ALFA Pulsar Survey has discovered and confirmed 38 new pulsars to date. These were found using the "quicklook" processing. This program, developed by Duncan Lorimer and David Champion, searches for new pulsars in real time. This can only be achieved by degrading the data by a factor of 16 in time and resolution, and then using a powerful computer cluster (the Arecibo Signal Processor (ASP), developed by Ingrid Stairs, David Nice, Don Backer and others) to carry out the search. However, by doing this, we have been systematically losing sensitivity to fast pulsars and pulsars at high dispersion measures (DMs). The P-ALFA consortium plans to reduce all the data using its full frequency and time resolution, a far more computer-intensive job.

During the Summer, Patrick Lazarus, an undergraduate student at McGill University in Montreal, worked with David Champion, Jason Hessels and Vicky Kaspi on P-ALFA data reduction. They worked on a series of python scripts developed to automatically process the P-ALFA data with full resolution using Scott Ransom's PRESTO routines. These scripts also load the results into a database, to be hosted by the Cornell Theory Center. Finally he has developed a viewer which connects to the database and allows candidates to be browsed and flagged.

As part of the testing of the scripts and the pipeline several disks worth of data has been processed and pulsars seen in the "quicklook" processing have been strongly detected. In addition one new pulsar has been discovered, PSR J1903+03. It was detected with a S/N of 24.2, it has a spin period of 2.15 ms and a DM of ~300 cm−3 pc. From the confirmation and timing observations made to date, it is clear that this millisecond pulsar (MSP) is in a binary system with an orbital period of several hundred days.


Discovery plot for PSR J1903+03.

Discovery plot for PSR J1903+03, produced using PRESTO. On the left plot, we can see that the signal is persistent, has a constant pulse shape and that the pulse shape is relatively narrow. In the middle plots, we can see that the signal is broadband and, below, that it is dispersed at around 297cm−3 pc. In the right plots, the program searched for the best period and period derivative, the numerical result indicates a very small, but nevertheless significant, period derivative, indicating that the pulsar is a member of a binary system.

Outside globular clusters this is the 5th fastest spinning pulsar known. Including globular cluster pulsars, it is 11th. This object has the highest DM known for any MSP. This is extremely important - it confirms the fact that the P-ALFA survey can see MSPs deep into the disk of the Galaxy, far from the solar system, where the vast majority of them await discovery. MSPs, particularly those in binary systems, are important for many areas of astrophysics (see, for instance, highlight on PSR J1738+0333). Using models of the electron distribution of the Galaxy and the pulsar population, Duncan Lorimer (private communication) predicts that, given the present observing system (i.e. 100 MHz bandwidth, 268-s integrations) we will detect 120 MSPs in the area 32° < l < 77° and | b | < 5°. There are currently only 9 MSPs in this area which are not in globular clusters. These predictions have to be taken with caution, because the effects of scattering are very important in this case, and they are to a large extent unknown. Nevertheless, they agree with the predictions made by Paulo Freire at the 205th AAS meeting in Washington D.C. He used pulsar DM distributions to show that, if the Parkes Multi-beam survey had the same time (64 us) and spectral (0.39 MHz) resolution of the P-ALFA survey, it would have detected between 40 and 60 MSPs in the portion of the Arecibo sky it surveyed. These would probably have a flat distribution of DMs from 0 to 400 cm−3 pc. In reality, it detected four MSPs, all with DMs below 40 cm−3 pc. Pulsars at higher DMs were lost because of dispersive smearing across its 3-MHz filters. Furthermore, because of the relatively small dwell times, the P-ALFA surveys have unprecedented sensitivity to MSPs in binary systems with short orbital periods.

Discovering 120 new MSPs would triple the number of known MSPs in the disk of the Galaxy. Tripling the bandwidth of the system, a feat to be achieved with the new P-ALFA spectrometers, will further increase the number of discoveries. Will this prediction be verified? Patrick Lazarus found 1 new MSP after searching 200 pointings (about 1400 beams). This respresents about 4 square degrees; the full survey is to cover an area of about 440 square degrees. So we might find a large number of MSPs after all.