2003ApJ...586.1238A

Fast X-ray transients (FXTs) with timescales from seconds to hours have been seen by numerous space instruments. Because they occur at unpredictable locations, they are difficult to observe with narrow-field instruments. Only a few hundred have been detected, although their all-sky rate is in the tens of thousands per year. We have assembled archival data from Ariel-5, HEAO 1 (A-1 and A-2), WATCH, ROSAT, and Einstein to produce a global fluence-frequency relationship for these events. Fitting the logN-logS distribution over several orders of magnitude to a simple power law, we find a slope of -1.0^+0.2_–0.3. The sources of FXTs are undoubtedly heterogeneous, representing several physical phenomena; the α~-1 power law is an approximate result of the summation of these multiple sources. Two major contributions come from gamma-ray bursts and stellar flares. These two types of progenitors are distributed isotropically in the sky; however, their individual luminosity distributions are both flatter than the -3/2 power law that applies to uniformly distributed standard candles. Extrapolating from the BATSE catalog of gamma-ray bursts (GRBs), we find that the fraction of X-ray flashes that can be the X-ray counterparts of gamma-ray bursts is a function of fluence. The exact fraction of GRB-induced X-ray counterparts is sensitive to the R_X/γdistribution, which we estimate from available experimental measurements. Certainly, most FXTs are not counterparts of standard gamma-ray bursts. The fraction of FXTs from non-GRB sources, such as magnetic stars, is greatest for the faintest FXTs. Our understanding of the FXT phenomenon remains limited and would greatly benefit from a large, homogeneous data set, which requires a wide-field, sensitive instrument.