A COMPREHENSIVE ANALYSIS OF SWIFT/X-RAY TELESCOPE DATA. IV. SINGLE POWER-LAW DECAYING LIGHT CURVES VERSUS CANONICAL LIGHT CURVES AND IMPLICATIONS FOR A UNIFIED ORIGIN OF X-RAYS

By systematically analyzing the Swift/XRT light curves detected before 2009 July, we find 19 light curves that monotonously decay as a single power law (SPL) with an index of 1 similar to 1.7 from tens (or hundreds) of seconds to similar to 10(5) s post the gamma-ray burst (GRB) trigger. They are apparently different from the canonical light curves characterized by a shallow-to-normal decay transition. We compare the observations of the prompt gamma rays and the X-rays for these two samples of GRBs (SPL vs. canonical). No statistical difference is found in the prompt gamma-ray properties for the two samples. The X-ray properties of the two samples are also similar, although the SPL sample tends to have a slightly lower neutral hydrogen absorption column for the host galaxies and a slightly larger energy release compared with the canonical sample. The SPL X-ray Telescope (XRT) light curves in the burst frame gradually merge into a conflux, and their luminosities at 10(5) s are normally distributed at log L/ergs s(-1) = 45.6 +/- 0.5. The normal decay segment of the canonical XRT light curves has the same feature. Similar to the normal decay segment, the SPL light curves satisfy the closure relations and therefore can be roughly explained with external shock models. In the scenario that the X-rays are the afterglows of the GRB fireball, our results indicate that the shallow decay would be due to energy injection into the fireball and the total energy budget after injection for both samples of GRBs is comparable. More intriguing, we find that a prior X-ray emission model proposed by Yamazaki is more straightforward to interpret the observed XRT data. We show that the zero times (T-0) of the X-rays are 10(2)-10(5) s prior to the GRB trigger for the canonical sample, and satisfy a log-normal distribution. The negligible T-0's of the SPL sample are consistent with being the tail of T-0 distributions at low end, suggesting that the SPL sample and the canonical sample may be from a same parent sample. Referenced to T-0, the canonical XRT light curves well trace the SPL light curves. The T-0's of the canonical light curves in our analysis are usually much larger than the offsets of the known precursors from the main GRBs. If the prior emission hypothesis is real, the X-ray emission is better interpreted within the external shock models based on the spectral and temporal indices of the X-rays. The lack of detection of a jet-like break in most XRT light curves implies that the opening angle of the prior emission jet would be usually large.