A Search for Sub-millisecond Pulsations in Unidentified FIRST and NVSS Radio Sources

We have searched 92 unidentified sources from the FIRST and NVSS radio catalogs for pulsations at 610 MHz. The selected sources are bright, have no identifications, are unresolved and have significant linear polarization. Our search was sensitive to sub-millisecond pulsations from pulsars with a large range of dispersion measures. We have detected no radio pulsations from these sources and conclude that they are unlikely to be a population of previously undetected pulsars. The FIRST and NVSS surveys are recent 1400 MHz VLA radio surveys of the Northern sky. In the published FIRST catalog (White et al. 1997), the positions and fluxes of ∼ 1.4 × 10 discrete radio sources are complete to ∼ 1 mJy. The NVSS survey (Condon et al. 1998) catalogs more than 1.8 × 10 sources complete to ∼ 2.5 mJy and preserves polarization information. Several large-scale pulsar surveys have been conducted at high Galactic latitudes (see Camilo 1997 for a review) which were sensitive to sub-millisecond pulsars with only a small range of dispersion measures (DM <∼ 10 pc cm ). A targeted search of unidentified FIRST and NVSS sources for sub-millisecond pulsars with a wide range of DMs is feasible using small frequency channel widths and a fast sampling rate. We have searched for radio pulsations in 92 bright, point-like unidentified sources from the FIRST and NVSS surveys which are more than 5% linearly polarized at 1400 MHz. Although there is no abrupt cutoff separating the pulsar population from the extragalactic radio source population, a polarization threshold of 5% excludes most (∼ 90%) of the identified non-pulsar population while retaining the majority (∼ 90%) of the identified pulsar population in the NVSS survey (Han & Tian 1999). Each source was observed at a center frequency of 610 MHz in two orthogonal linear polarizations for 420 sec with the Lovell 76meter radio telescope. A bandwidth of 1 MHz was split into 32 channels with signals from each channel recorded as a continuous 1-bit digitized time series sampled at 50 μs. We did not detect any pulsations from the target sources. We argue that our non-detections imply that the sources are unlikely to be pulsars. The selected sources were bright, with the weakest source having a 1400 MHz flux density of 15 mJy. Assuming a typical pulsar spectral index of α = 1.6 (Lorimer et al. 1995), this source would have a flux density 58 mJy at 600 MHz. This is significantly greater than our sensitivity limit for the expected