Background Rates and Rejection Efficiency

  Tests were performed at XRCF during Phase I to measure the background rate and rejection efficiency of the ACIS flight CCDs. We have analyzed simultaneous high-speed tap and telemetry data from sequence 1 of science runs 113 and 127 for detectors I3 (FI) and S3 (BI), respectively. To determine the cosmic-ray (CR) flux incident on each detector we plotted the ACIS events frame-by-frame, examined the distribution of events by eye, and decided how many cosmic rays were responsible for creating the events. Using this method the observed cosmic ray flux at XRCF was found to be $\rm 0.026\pm0.001\ CR\ s^{-1}\ cm^{-2}$ in both I3 and S3.

Selecting ACIS events with pulse heights less than or equal to 4095 ADU ($\sim$16.4 keV), we have computed in Table 4.35 the average number of events generated per cosmic ray and the number of unrejected events (i.e. ASCA grades 02346) per cosmic ray. The average number of events per cosmic ray is higher in I3 than in S3 because FI CCDs are thicker than BI CCDs, allowing the charges ejected by an energetic cosmic ray more time to diffuse. The background rejection efficiency was 99.3% in I3 and 78.7% in S3. The unrejected background rate was $\rm 1.11\pm0.21\times10^{-4}\ counts\ s^{-1}\ cm^{-2}\ keV^{-1}$ in I3 and $\rm 3.28\pm0.39\times10^{-4}\ counts\ s^{-1}\ cm^{-2}\ keV^{-1}$ in S3. The spectra of the unrejected events in both CCDs were flat. The cosmic ray flux outside the influence of the Earth's magnetic field is $\rm \sim0.6\ CR\ s^{-1}\ cm^{-2}$ (Allen, ``Astrophysical Quantities'', p. 275). Assuming this were the particle flux incident on the detector, the predicted on-orbit background rates in ASCA grades 02346 for I3 and S3 are $\rm 2.56\pm0.48\times10^{-3}\ counts\ s^{-1}\ cm^{-2}\ keV^{-1}$and $\rm 7.57\pm0.90\times10^{-3}\ counts\ s^{-1}\ cm^{-2}\ keV^{-1}$,respectively.

Analysis of the charge distribution of the rejected events (i.e. ASCA grades 157) in device I3 shows that 91.1% of them were ACIS bit map grade 255 events (i.e. all eight neighboring pixels were above the split threshold). However, only 1.9% of the rejected events in S3 were ACIS grade 255 events. Thus, a large majority of the background events in a FI device could be rejected simply by filtering ACIS grade 255 events, but only a very small percentage of the background events in a BI device would be rejected. If we reject only ACIS grade 255 events, the total on-orbit background rate for six devices should be $\rm \sim 12\ counts\ s^{-1}$, which is about 7% of the telemetry bandwidth.

 

Device No. of CR Events/CR Unrejected Events/CR

$$9.80\pm0.53 0.072\pm0.014$$ I3 375

$$0.95\pm0.07 0.202\pm0.026$$ S3 356