SIS Grade 6 Events
K. Mukai (ASCA GOF)
K. Weaver (John Hopkins University)
1 Grades of SIS Events
Events detected with the SIS are characterized by "grade", in addition to position, time, and pulse height. Grade is an integer in the range 0-7, which describes the distribution of charges from a single event among the neighboring pixels of the CCD. The distribution of charges from an X-ray event is determined by the 3-dimensional position at which it was absorbed. In addition to the relationship with respect to pixel boundaries, which is pure chance, the depth at which it was absorbed, which is a function of the photon energy in a probabilistic sense, is likely to travel within the CCD before it is absorbed, and therefore the greater the size of the charge cloud becomes.
Grade 0 ("perfect single") is the case where only the central pixel in an array of 3 by 3 pixels contained a significant charge. Soft photons are often detected as grade 0 events. Grades 2, 3 and 4 are cases where significant charges are detected in two (non-diagonally) neighboring pixels, and often caused by harder photons. Some distribution of charges are unlikely to arise from an X-ray event; events with grades 1, 5, and 7 are discarded from analysis for this reason. In particular, a high fraction of Faint mode data (for which grade classification is performed on the ground) are grade 7 ("other") events.
Grade 6 ("L+Q") events contain significant charges in 3 neighboring pixels in an "L" shaped arrangement, or in 4 pixels in a square arrangement. This article demonstrates that a significant fraction of X-ray events above ~5 keV are detected as grade 6 events, and that the particle event rate in grade 6 is not particularly high. Moreover, above 5 keV, the statistical quality of SIS data is often poor; thus for a study of the high energy continuum, and in particular that of the Fe K line(s), there is a strong reason to include grade 6 events in analysis. We finally comment on the current state of calibration for the grade 6 data.
2 Comments on Grade 6 Availability
Prior to 1993 Nov 30 07:30 UT, only grade 0-4 events were placed in telemetry in Bright mode, in which the grade classification is performed on-board. Grade classification for Faint mode data are carried out by the program "faint" which will also discard grade 5-7 events, when used for Faint mode data taken before the above mentioned time to create Bright mode files. However, for more recent data, or in conversion to Bright2 mode, the "maximum grade to output" is a parameter that can be specified by the user. In Rev 1 processing at GSFC, grade 0-6 events are kept in science FITS files except in Bright mode files prior to 1993 Nov 30.
However, it has long been standard practice to select only grades 0, 2, 3 and 4 events. This is the default in ascascreen and in Rev 1 processing. Merged, screened event files from either process probably contain only these 4 standard grades. It is easy to check whether grade 6 events are present in SIS Bright or Bright2 mode file:
example % fstatistic bright.evt GRADE -
This will give statistical information on the grade column in the file, including the maximum value present.
3 The Grade 6 Fraction
We used a 2-day observation of the cataclysmic variable, AO Psc, performed in 1994 June, in order to estimate the fraction of grade 6 events as a function of energy. The observation was made in 2-CCD Faint mode at high bit rate (~40 ksec) and in 2-CCD Bright mode at medium bit rate (~40 ksec). The source is, on average, 0.5 cps.
Figure 1 shows the number of grade 6 events as a fraction of grade 0, 2, 3 and 4 events, as a function of energy, for SIS-0 in Faint mode, after background subtraction. As expected, grade 6 events are predominantly high energy phenomenon. Below 5 keV, the fraction of grade 6 events is a steeply increasing function of energy, except the structure around Si K edge. Above ~6 keV, it is a more-or-less flat function of energy (29% averaged over 7-10 keV range). In bright mode, the situation is worse (40% over the same energy range), because the echo effect leads to misclassification of some events, particularly from grade 2 (vertical split) to grade 6.
The other important consideration is the fraction of particle events among grade 6 events. Although a rigorous measurement is yet to be carried out, it appears that only a small fraction of particle events have a grade 6. Certainly this is not a concern for moderately bright sources such as AO Psc.
4 The Case of MCG-5-23-16
MCG-5-23-16 is a relatively bright (~1.5 cps) extragalactic object, observed in 1-CCD Bright mode on 1994 May 11 for ~40 ksec. One of us [KW], in analyzing the ASCA data, noticed that the Fe K line was broad and had a complex profile. However, using the standard grade selection (0, 2, 3 and 4), there was an insufficient number of photons to characterize the line profile quantitatively. The addition of grade 6 events made a sufficient difference to allow the analysis of the line profile with confidence.
For this, the GIS data with the standard matrix, and the extended (to include grade 6) SIS data and with an experimental SIS matrix that includes grade 6 events (generated using sisrmg v0.8), were analyzed separately and jointly. The SIS and GIS continuum parameters are consistent with each other; however, the GIS data are probably more trustworthy for the continuum shape above 7 keV (both in terms of statistics and systematics). As for the Fe K line, the SIS data clearly show both a narrow and a broad component, and the parameters for the latter are consistent with the broad-line parameters derived from the GIS data. This agreement suggests that the grade 6 response generated by sisrmg, although the program warns that it is poorly calibrated, is good enough to be used with caution. The details of the analysis and results of the MCG-5-23-16 ASCA data can be found in Weaver et al 1996 (submitted to ApJ).
5 Status of Grade 6 Calibration
The SIS team has so far concentrated their resources on calibrating the standard set of grades. Our experiences so far suggest that grade 6 events can be safely used to good advantage for imaging and timing analyses at high energies.
What is involved in calibrating grade 6 events? First, the fraction of grade 6 events (the branching ratio) has to be measured and compared with the model that is currently in the response generator. The current model is probably fairly accurate, otherwise the branching ratios into lower grades cannot be calculated very accurately. Second, the calibration of resolution has to be performed. There are some intrinsic limitations, because the charges are collected from 3 or 4 pixels, which increases the read-out-noise.
We hope that the SIS team will be able to find the resources necessary to calibrate the grade 6 response. In the meantime, observers with statistics-limited observations are encouraged to experiment with grade 6 events.
Figure 1: The ratio of the number of grade 6 events to that of grade 0, 2, 3 and 4 events from SIS Faint mode data on AO Psc are plotted as a function of energy. The events were collected from a circular extraction region with a radius of ~4 arcmin, and local background have been subtracted.
Proceed to the next article Return to the previous article