The 3.2 gigapixel LSST camera, an array of 189 thick fully-depleted CCDs, will repeatedly image the southern sky and accomplish a wide variety of science goals. However, its trove of tens of billions of object images implies stringent requirements on systematic biases imprinted during shift-and-stare CCD observation. In order to correct for these biases which, without correction, violate requirements on weak lensing precision, we investigate CCD systematics using both simulations of charge transport as well as with a unique bench-top optical system matched to the LSST’s fast f/1.2 beam. By illuminating single CCDs with realistic scenes of stars and galaxies and then analyzing these images with the LSST data management pipelines, we can characterize the survey’s imaging performance well before the camera’s first light. We present measurements of several CCD systematics under varying conditions in the laboratory, including the brightness-dependent broadening of star and galaxy images, charge transport anomalies in the silicon bulk as well as the edges, and serial deferred charge. Alongside these measurements, we also present the development and testing of physics-based models which inform corrections or mitigation strategies for these systematics. Optimization of the CCD survey operation under a variety of realistic observational conditions, including systematic effects from the optics, clocking, sky brightness, and image analysis, will be critical to achieve the LSST’s goals of precision astronomy and cosmology.
We describe a camera beam simulator for the LSST which is capable of illuminating a 60mm field at f/1.2 with realistic astronomical scenes, enabling studies of CCD astrometric and photometric performance. The goal is to fully simulate LSST observing, in order to characterize charge transport and other features in the thick fully-depleted CCDs and to probe low level systematics under realistic conditions. The automated system simulates the centrally obscured LSST beam and sky scenes, including the spectral shape of the night sky. The doubly telecentric design uses a nearly unit magnification design consisting of a spherical mirror, three BK7 lenses, and one beam-splitter window. To achieve the relatively large field the beam-splitter window is used twice. The motivation for this LSST beam test facility was driven by the need to fully characterize a new generation of thick fully-depleted CCDs, and assess their suitability for the broad range of science which is planned for LSST. Due to the fast beam illumination and the thick silicon design [each pixel is 10 microns wide and over 100 microns deep] at long wavelengths there can be effects of photon transport and charge transport in the high purity silicon. The focal surface covers a field more than sufficient for a 40×40mm LSST CCD. Delivered optical quality meets design goals, with 50% energy within a 5 micron circle. The tests of CCD performance are briefly described.