FlyEyes: A CCD-based Wavefront Sensor for PUEO, the CFHT Curvature AO System

Adaptive optics wavefront sensing imposes stringent requirements on detectors, due to the simultaneous need for extremely low read noise and high frame rates. Curvature wavefront sensing measurements are based on the normalized intensity of the signal in a given subaperture, and avalance photodiodes (APDs) have traditionally been used as detectors in curvature systems such as the Canada-France-Hawaii Telescope (CFHT) adaptive optics (AO) bonnette, called PUEO after the endemic Hawaiian owl. Passively quenched APDs are robust but have low QE (similar to 40%), while actively quenched APDs can have much higher QE, but have been known to fail. Furthermore, curvature systems with large numbers of subapertures are now in operation, and the cost of individual APDs may become prohibitive for such systems. Thus, a CCD-based alternative appears very attractive, and development of a specific chip was initiated at ESO 10 years ago. In this article, we report on the performance of the Fly Eyes camera, a project that was conceived to compare the performance of the backside-illuminated custom-designed CCD detector with an array of APDs, used in an operational and well-characterized curvature wavefront AO system. The on-sky performance is demonstrated to be unaffected on bright guide stars (i.e., negligible latency), and although the faint end suffers from the 2.5 e(-) read noise, the performance can be regained by lowering the frame rate on the wavefront sensor. In this article, we report on results that show that the CCD can be used to replace an array of expensive APDs. This would enable a cost-effective upgrade of PUEO to a higher-order system, as has been proposed at various occasions.