Point Spread Function Tests

In order to recover an accurate representation of the sky sampled by AXAF using ACIS, we must understand how the combined HRMA + ACIS spatially redistributes the incoming photons. This spatial redistribution function (the so-called point spread function, or PSF) is a combination of the HRMA PSF and the ACIS instrument's pixelization and detection efficiency. The HRMA + ACIS combined PSF is a function of energy (or input spectrum), off-axis angle, subpixel position, count rate (i.e. pileup), grade selection, ACIS operating mode, and perhaps other quantities.

Point Spread Function (PSF) tests measure the core (PI) and wings (PW) of the PSF on-axis and at several off-axis positions, at the point of ideal focus determined by the Shutter focus/Plate focus measurements (X=0). Two types of data were collected: the core PSF were done by counting single photon arrivals in the core of the PSF; the wings PSF were done with higher fluxes, which produces photon mode in the wings but photons arrive so rapidly that multiple photons are superposed in the core. Measurements were made at medium (Al K), high (Cu K), and low (O K) energy. C K was included for an ultra-low energy point for the BI chips. Tests were also made to estimate the PSF using the ACIS and ACIS-2C continuous readout mode.

Sub-pixel position measurements were also performed as part of the PSF test suite. We moved the PSF by sub-pixel amounts to sample the digitization error resulting from pixel size undersampling of the PSF and to test split X-ray event reconstruction models of sub-pixel interaction location. For FI chips, tests were performed at high (Cu K) and low (O K) energy to test for differing split event fractions. For BI chips, Cu K was also used as the high-energy test, but C K was chosen for the low-energy test for differing split event fractions.

Given the complexity of this problem, the ACIS team proposes to concentrate on the issues specific to ACIS, attempting to assess how ACIS modifies the HRMA PSF. We must, then, rely on the efforts of the Mission Support Team to define the HRMA PSF and stay in close communication with them, reporting our ACIS modeling and data analysis results and receiving updated HRMA model output. We acknowledge that ACIS and ACIS-2C data from XRCF may be useful in refining the HRMA model but we do not propose to take this as the primary goal of our data analysis.

Clearly the goal of PSF data analysis and modeling is to generate a model that predicts the on-orbit HRMA + ACIS PSF. We will use the XRCF calibration data to refine and improve our models of ACIS, then rely on MST's on-orbit HRMA model to predict HRMA + ACIS on-orbit PSFs. We will strive to create a PSF generator, an engine that produces a PSF for a point source falling at any position on any ACIS pixel, given the aimpoint (to determine off-axis angle), the subpixel position of the point source, the source spectrum and energy bandwidth, the count rate, and a list of acceptable event grades. The format of this model PSF should be the same as the dataset being modeled, namely an event list. The ideal system would allow a user to assemble a set of point sources to approximate an extended source.