Although ACIS is an immensely capable instrument with a large number of potentially useful configuration options for special purposes, most observers will choose from a small number of broadly useful configurations. We present several ACIS configurations selected for use in the ACIS Guaranteed Time Observations, and explain the rationale based on the astrophysical objectives.
In order to glean some idea of the kinds of objects that observers will commonly select as targets for ACIS, we have categorized the targets selected by the ACIS Guaranteed Time Observations.
Although an enormous range of scientific projects fall into this set of targets, only a small range of technical considerations govern the observing choices for most objects. In particular, many analysis choices can be made on the ground, at the leisure of the observer during the interactive reduction and analysis process. Thus the key considerations are those which irretrieveably affect the data collected in orbit and transmitted to the ground.
Two kinds of irretrievable data corruptions may likely occur, and the observer should take them into account in planning his or her observation. If multiple X-ray photons strike in the same or neighboring pixels during a single CCD frametime, then they will not be recognized as separate events by the on-board CCD signal processing. This effect is called `pile-up'. Other posters at this meeting and the ASC supplied material discuss techniques for correcting for pile-up, but the observation planner can reduce pile-up with suitable modifications to the ACIS instrument set-up.
The other corruption occurs when too many photons are detected by ACIS and the rate of telemetry cannot telemeter all the events to the ground. This effect is called `telemetry saturation'. Again if the observer anticipates this effect, ACIS can be configured in ways to eliminate or ameliorate telemetry saturation.
Thus for these two effects the fundamental considerations divide into a simple 2x2 matrix based on two properties of the targets: Are they point-like or extended? Are they faint or bright?
Faint point sources are the simplest of targets for ACIS to observe. The only two issues to check are the brightness of the brightest source in the field and the total count rate expected from all sources (point-like and diffuse) within the full field of view.
Generally the observer will maximize all ACIS capabilities (field of view, quantum efficiency and energy resolution) by selecting all six imaging chips, and use the least restrictive grade selections for background rejection. Using the full field of view results in a 3.3 second frametime. No detectable effects of pile-up are likely when a source provides less than 0.01 counts during a frame. Most of the effects of pile-up can be corrected if the source brightness is less than 1 count per frame. Above 6 counts per frame substantial data losses are likely. See the attached figure (made by Ann Hornschemeier).
Note: for sources off-axis higher rates are tolerable. See the next section.
If the brightest source exceeds the limits stated above, the observer should take steps to reduce the pile-up effect. The simplest step is to reduce the field of view, and hence to shorten the frametime. For an on-axis point source ACIS can be configured to carry out sub-array readouts. In this case data from most of the CCD is discarded in the clocking process. The smallest sub-array (100 rows) on-axis can be read-out in 0.9 seconds. If the target is placed near to the read-out nodes on the CCD chips this time can be reduced to 0.3 seconds.
If the user can tolerate a low efficiency the readout of an entire full frame can be done in only 0.1 seconds, but the full 3.3 seconds must be spent in processing the frame (yielding an observing efficiency of only 3%).
(If the observer does not need full imaging information ACIS can be operated in Continuous Clocking mode. In this case the image does not appear two-dimensional, but a bright point source becomes a one-dimensional smear across the CCD frame image. Sources can become confused using this technique but the gain in reducing pileup is about 1000.)
If the factor of ten improvement from using sub-frames is still not enough, more severe steps are necessary. See the poster on `How to Observe Really Bright Sources with ACIS' by Bautz et al.
For extended sources it is necessary to consider both the pile-up limitation and the telemetry limitation. The pile-up issue can arise either from a bright localized point or point-like emission within the extended emission, or from a large bright emission which fills a substantial part of the field of view.
In the first case the pile-up limitation is similar to a bright point source, with the only change being converting the extended emission into an equivalent point source on a scale of 0.5 arc seconds.
In the second case the random positional fluctuations of photon arrivals from a diffuse source changes the sensitivity to pile-up. In the MIT CCD calibrations the illumination was diffuse, and the maximum pileup tolerated was set at 0.001 counts/pixel/frametime. At this level about 5% effects on quantum efficiency were detected.
Few astrophysical extended sources are expected to cause problems with pile-up.
When the targets are bright and extended it becomes possible to encounter telemetry saturation. The rate at which saturation occurs is shown in the following table:
Table 1: Telemetry Saturation Limits for each Readout Mode
Because fewer bits are used to describe each event ACIS can handle higher rates in changing from Timed Exposure (TE) to Continuous Clocking (CC), and from faint to graded. The drawbacks are loss of information.
In graded mode only the total charge and ACIS grade are returned, so the ability to reprocess grades on the ground is lost. In Continuous Clocking mode one dimension of spatial information is lost.
Faint+bias saves the time of dumping the bias before the observation (but loses if many events from the same pixels are telemetered). Very faint telemeters a 5x5 region, which allows for more sophisticated treatment of charge splitting and multiple photon interactions.