2001/01/31 by Hiroshi Tsunemi, H. Tsunemi, Koji Mori +9 · 91 citations
Engineering · Physics and Astronomy · #Acis #Adaptive optics and wavefront sensing #Artificial intelligence #Astrophysics #CCD and CMOS Imaging Sensors #Calibration and Measurement Techniques #Computer science #Image (mathematics) #Image quality #Image resolution #Optics #Physics #Pixel #Point spread function #Position (finance) #Resolution (logic) #astro-ph
paper · pdf · doi:10.1086/321338
published in The Astrophysical Journal 554(1), 496-504 (IOP Publishing) · 12 pages, 5 figures. Accepted for publication of ApJ. High resolution and color figures are available at http://wwwxray.ess.sci.osaka-u.ac.jp/~miyata/paper/acis_subpixel.pdf
arxiv created 2001/01/31 · openalex publication_date 2001/06/10 · arxiv updated 2009/12/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
The position accuracy of X-ray photons on a CCD detector is generally believed to be limited by the CCD pixel size. While this is true in general, the position accuracy for X-ray events which deposit charge in more than one pixel can be better than that of the CCD pixel size. Since the position uncertainty for corner events is much better than the pixel size, we can improve the Chandra ACIS spatial resolution by selecting only these events. We have analyzed X-ray images obtained with the Chandra ACIS for six pointlike sources observed near the optical axis. The image quality near the optical axis is characterized by a half-power diameter (HPD) of 0 66 that is a convolution of the point-spread function (PSF) of the high-resolution mirror assembly (HRMA) and the CCD pixel shape (24 μm square). By considering only corner events the image quality is improved to 0 56 (HPD) at the optical axis, which is very close to the image quality of the HRMA alone. We estimated the degradation of the image quality obtained by using all events, compared to that obtained using only corner events, to be 0 33, which coincides with that expected from the pixel size. Since the fraction of the corner events is relatively small, this technique requires correspondingly longer exposure time to achieve good statistics.