Astron. Astrophys. 364, L93-L96 (2000)

4. Nature of the sources

With luminosities of 1 to W over 4 sr at 500 pc (for E^-2 spectra), the sources cannot be unresolved gas clumps irradiated by the local cosmic-ray flux: the required mass of M at 500 pc cannot have escaped the radio and IR surveys, even when considering radio beam dilution from unresolved clouds (Grenier 1997). Nor can they be slow (20 km/s), old neutron stars, accreting gas from a dense cloud: they are at least times too rare (Grenier 1997) and the maximum Bondi-Hoyle accretion power of W that results from the formation of a surrounding HII region by the neutron star UV radiation, is 10 times too low (Blaes et al. 1995). The accretion power reached for fast (200-400 km/s), highly magnetized (10¹² G) neutron stars with long periods in the intercloud medium (10^-3 H cm^-3), though increased by Kelvin-Helmholtz instabilities in the shocked gas, is also orders of magnitude too low (Harding & Leventhal 1992). Isolated accreting black holes with masses of 10 M (Colpi et al. 1986) and 35 M (Dermer 1997) have been proposed, but they would be too rare in the Belt: for mass progenitors 25 M (Timmes et al. 1996), black holes are 3 to 9 times fewer than neutron stars for indices of -1.1 & -2.0, respectively. Moreover, the luminosity / ratios observed assuming statistically 1 or 2 faint ROSAT source in an EGRET error box, are clearly at variance with p-p interactions in the accretion flow (Colpi et al. 1986), unless non-thermal acceleration is advocated (in micro-quasar jets?). These luminosity ratios are also at variance with standard accreting binary systems. I found no spatial coincidence with a WR star or the numerous O stars at mid latitude despite their highly supersonic winds with kinetic powers of W. Eight pulsars are known in rays, 7 bright young ones at kpc distances in the Galactic disc, and a faint, older one inside the Gould Belt (Geminga). So, the present sample is strongly biased to high luminosity and youth. The stability of most of the Belt sources (Tompkins 1999, Gehrels et al. 2000) is consistent with a pulsar origin. Born with large velocities (Lyne & Lorimer 1994), Galactic pulsars rapidly migrate away from the plane to mid-latitudes. The outer gap model (Yadigaroglu & Romani 1997) for beamed emission predicts that 4-5 Galactic pulsars should be detectable at in contrast with the 40 5 sources associated with the Belt. Similarly, the wide-beam comptonized polar-cap model (Sturner & Dermer 1996) predicts 1-2 Galactic pulsars detectable at as opposed to 25 5 sources linked to the Belt. These discrepancies cannot be resolved by increasing the Galactic pulsar birth rate by more than 30% for fear of overproducing sources at low latitude (Yadigaroglu & Romani 1997), nor by using larger scale heights or velocities which are not supported by the radio data.

Given the enhanced SN rate in the Belt and its inclined geometry, I propose that the sources associated with the Belt be relics of Belt supernovae in the form of few Myr old pulsars. Detecting 20 to 40 Belt collapsed stars as EGRET sources requires the product of the beaming fraction and pulsar age be of order 1-1.5, for instance over 2 or 3 Myr. is predicted for the main polar-cap beam (Thompson et al. 1997), and values of 0.1-0.6 are possible for the outer-gap fan beam depending on pulsar age (Romani 1996). Yet, one should bear in mind the extreme closeness of the Belt objects. The -ray luminosity, , scales with the spin-down power, , as over 4 decades in (Thompson et al. 1997). Extrapolating from the faint Geminga for only half a decade, to W W over 1 sr, suggests that a pulsar 10 times as old as Geminga, i.e. 3-Myr old, remains easily detectable by EGRET out to 500 pc. Furthermore, the recorded -ray lightcurves show that 10 times fainter emission is detected off the main peaks over large phase intervals. This side emission from a 3 Myr old pulsar would remain detectable by EGRET up to 350 pc, thus largely increasing the detection probability . Recent polar cap simulations indicate that 4-5 times as many off-beam sources as on-beam ones would be detectable above 100 MeV at a given distance (Harding & Zhang 2001), therefore up to 350 pc for side emission. In this case, the population of Belt neutron stars born in the last 2-3 Myr may account for the Belt -ray sources. It may also explain the scarcity of bright on-beam-like sources off the Galactic plane. The softness of side emission ( = 1.8 to 2.5) may also explain the soft average spectral index = 2.25 0.03 measured for the Belt sources as opposed to that of = 1.74 0.02 obtained the 5 on-beam pulsars in the Galactic plane. Preliminary simulations show that the Belt spatial signature is preserved over at least 2 Myr despite rapid migration. So, given our present understanding of pulsar -ray emission, the hypothesis that the Gould Belt sources be powered by pulsars, mostly off-beam pulsars, is quite plausible. If true, the Belt pulsars would considerably broaden our understanding of photon-particle cascades inside their magnetospheres to older and lower-luminosity objects at various aspect angles. The radio beam being apparently much narrower than the -ray beam, at least in one dimension, one would expect a majority of radio-silent pulsars among the Belt sources. The spectral criterion 2 often adopted to search for -ray pulsars may not be valid for nearby objects. The next generation telescope, GLAST, to be launched in 2005, will be able to detect their periodicity in the -ray signal, if any. As supernova relics, the Belt sources would bring useful constraints on the initial star mass spectrum at large masses and on the abundance of explosive nucleosynthesis products in supernova remnants.

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001
helpdesk.link@springer.de