MINBAR X-ray burst light curves / analysis

A few checks have been performed on the light curves of the 6793 bursts in MINBAR. The results are shown in three types of plots: 1) histograms per parameter per burster for those bursts which are more or less reproducable, 2) diagrams of paremeters for bursts that were detected with multiple instruments and 3) diagrams of parameters of many PCA bursts, comparing the old method (Galloway et al. 2008) with the one here employed.

Histograms

First, 4 histograms of some basic parameters of most bursts. The duration peaks at about 15 s and the exponential decay time at 4 s. There is a long tail of intermediate duration bursts (nb, superbursts are not included). The peak flux shows a clear peak at 0.8 c/s/cm2 which is about 0.4 Crab units. There is also a peak at lower flux which is due to many weak bursts from IGR J17480. The power law index (without Gauss) peaks between 0.8 and 1.8. The low bound must be due to the rp process and the upper rather sharp bound due to an intrinsic effect (heat capacity due to electrons instead of ions).

Second, histograms have been made per source of peak flux, exponential decay time, duration and fluence, for a few sources that have many bursts:

Diagrams of burst parameters for bursts detected with multiple instruments

There were 25 PCA bursts which were also detected with either WFC (15 times) or JEM-X (10 times). We can compare the peak flux, duration, exponential decay time and fluence in the following plots.

These results do not look so nice, mainly because the PCA measurements are far more accurate than the WFC and JEM-X ones. There were 25 PCA bursts which were also detected with either WFC (15 times) or JEM-X (10 times). We can compare the peak flux, duration, exponential decay time and fluence in the following plots.

Comparing the new with the old method for PCA bursts

Next, we compared the new analysis results with those using the method employed by Duncan in the PCA catalog (Galloway et al. 2008). We can do this with two parameters: the peak count rate per PCU and the time scale which is defined as the fluence divided by the peak flux, both in units of PCU counts. The following two plots show the comparison

There are clusters of peak flux values which do not compare well. Those concern points which were not corrected for the collimator in Duncan's analysis. This concerns bursts at significant off-axis angles, for instance the Rapid Burster and GX 354-0. It also includes a few points from 4U 1636-536.

The other points do not exactly fall onto the 1-1 line. This is probably due to the peak flux being measured over 1-sec bins in the new analysis and over 0.125-s bins in Duncan's method.

The time scales compare reasonably well. The new numbers are generally about 10% larger, probably again because the peak flux is measured over different time scales.

Jean in 't Zand (SRON, April 8, 14:45:00 CET 2016