2008MNRAS.388L..44L

Millisecond dips in the RXTE/PCA archival data of Sco X-1 taken from 1996 to 2002 were reported recently. Those dips were found to be most likely caused by instrumental dead time but may also contain some true astronomical events, which were interpreted as the occultation of X-rays from Sco X-1 by trans-Neptunian objects (TNOs) of 100 m size. Here, we report the results of search for millisecond dip events with the new RXTE/PCA data of Sco X-1 taken in the year 2007. Adopting the same selection criteria as that in the previous study, we found only three dip events in 72-ks data, much fewer than the 107 events found in the 560-ks data taken from 1996 to 2002 reported earlier. The new data provide more detailed information of individual `very large events' (VLEs), which is not available in the old archival data. Although the number of VLEs does not obviously increase during the occurrence of dip events, all the three dip events are coincident in time with VLEs that have no flags set for any of the propane or the six main xenon anodes. It is a strong indication of instrumental effects. No significant dips which might be real occultation by 60-100 m TNOs were observed. With only 72-ks data, however, the previously proposed possibility that about 10 percent of the dip events might not be instrumental still cannot be strictly excluded. Using the absence of those anomalous VLEs as the criterion for identifying non-instrumental dip events, we found, at a lower confidence level, four dip events of duration 8-10 ms in the 72-ks data. Upper limits to the size distribution of TNOs at the small size end are suggested.