Effects of the Flickering Pixels to the SIS Non X-ray Background
Y. Ueda, T. Dotani, Y. Ogasaka,
Institute of Space and Astronautical Science
The background, consisting of the Non X-ray Background (NXB) and the Cosmic X-ray Background (CXB), is very important to extract the spectrum and image of sources obtained from the observation. Frequently, the background data is reconstructed from the data base. The data base maintained at GSFC is the collection of survey observation mainly done in the initial PV phase. Although the initial report on the SIS NXB during the PV phase (Gendreau 1994; 1995) exists, the SIS performance is changing due to the accumulated charged particle damage (Dotani et al. 1995). Thus it is necessary to investigate whether or not such a long-term effect is seen in the SIS NXB as well as its dependence on the clocking mode. We have studied the spectral change of the SIS NXB over two years after launch, in the analysis of the Cosmic X-ray Background (Ueda 1996; Ogasaka 1996).
We accumulated all the dark Earth data from April 1993 to November 1994. The total exposure time obtained is 750 ksec (1 CCD mode), 340 ksec (2 CCD mode), and 750 ksec (4 CCD mode), respectively, after the screening described below. To avoid the optical leakage and the scattered solar X-rays, we only used the data during satellite night. We also discarded intervals of 128 sec after the satellite day-night transition and those after the typical AE-off region to avoid the DFE complexity. Hot pixels are removed by the standard software ("sisclean") applied for each FRF. Faint and Bright mode data are treated together without correcting Echo and DFE. CTI is corrected based on the standard method, assuming all events are produced in the imaging region.
Figures 1, 2, and 4 show the NXB spectra obtained in successive 4 periods, namely (1) April 1993 to August 1993, (2) September 1993 to January 1994, (3) February 1994 to June 1994, and (4) July 1994 to November 1994 for various clocking modes. In 4 CCD mode, a clear increase is noticed below 0.7 keV. Because such increases are not significant in the 1 CCD mode and 2 CCD mode data, the feature below 0.7 keV is suspected to be intrinsic to 4 CCD data. It is known that 4 CCD data has large number of hot or flickering pixels in the lower energy band because of the long exposure. Thus the feature is suspected to be due to the hot and flickering pixels. The number of flickering pixels which cannot be rejected depend on algorithm of the hot pixel rejection. However, even applying the tight constraint that a pixel that registered an event more than twice during the total exposure is regarded as a hot pixel, the result above does not change. This means that some of the flickering pixels are difficult to remove due to their extreme low frequency to exceed the event threshold.
Our analysis has revealed the significant change of the NXB spectrum in 4 CCD mode, mostly in the soft X-ray band below 1 keV. This might affect the analysis such as the soft excess component from the source. We need special care if data is obtained in 4 CCD mode and the energy range below ~ 0.8 keV is used in the analysis. Although the construction of a new data base consisting of the recent data from flat fields would be helpful, it is worth to note that S/N ratio is expected to be very low in the soft energy band.
Figure 1: Long term history of the spectra of SIS non X-ray background in 1CCD mode (grade 0, 2, 3, and 4). All available chips (both S0 and S1) are averaged (unit: per chip). The spectra are accumulated during each five months since 1994 April. Data include those with level discriminator set to ~ 0.5 keV, and exposure corrections for each PH channel are applied.
Figure 2: Long term history of the spectra of SIS non X-ray background in 2CCD mode.
Figure 3: Long term history of the spectra of SIS non X-ray background in 4CCD mode.
Gendreau, K.C. 1994, "Initial Report on the SIS Internal Background", ASCA News No.2
Gendreau, K.C. 1995, PHD, Massachusetts Institute of Technology
Dotani, T. et al. 1995, "Effects of Radiation Damage on SIS Performance", ASCA News No.3
Ueda, Y. 1996, PHD, University of Tokyo
Ogasaka, Y. 1996, PHD, Gakushuin University
Proceed to the next article Return to the previous article