next previous
Up: The DELTA photon counting


1 Introduction

Although a majority of cosmic phenomena is slow enough to allow imaging with long exposures, ground-based high-resolution techniques require very short frame times, to deal with atmospheric seeing and reach the theoretical limit in resolution for large telescope diameters (Roggemann et al. 1997). Stellar speckle interferometry techniques (Labeyrie 1970; Knox & Thompson 1974; Lohmann et al. 1983), use large series of images, each one with a 10 ms average integration time. Adaptive optics wave-front sensors (Hardy 1978; Roddier 1988) require even shorter frame times. Fringe tracking in optical multiple aperture interferometers (Koechlin et al. 1996) also require high spatial and temporal resolution. A new application, requiring high data rate photon counting capabilities, is the "dark speckle'' technique for exo-planet detection (Labeyrie 1995). This single aperture method consists in doing statistics on photons collected in each pixel of the field. It involves a photon flux at the performance limit of existing cameras, as noted from recent experiments (Boccaletti & Labeyrie 1997).

At short integration times, quantum limits are often reached, and photon counting detectors with micro-channel plates (MCP) must be employed. Each frame is then represented by a cluster of detected photo-events. Historically, the first photon detector dedicated to speckle interferometry was an intensified film camera, built by Gezari (1972). Later, video systems were used (Blazit et al. 1977), yielding clipped frame signals where photon positions were marked by a logical 1. Later, detectors were designed, yielding the coordinates of photo-events in the image plane, thus allowing compact data storage and on-line processing. Surprisingly, the earliest device providing such data was dedicated to a long exposure task: spectroscopy (Boksenberg et al. 1972).

Among the cameras currently used for both speckle and multiple aperture interferometry, providing direct photon coordinates, are the CP40 (Blazit 1986), the Resistive Anode Camera, or "Ranicon'' (Clampin et al. 1988), the PAPA camera (Papaliolios & Mertz 1982), the MAMA camera (Timothy & Bybee 1975), the "Wedge-and-Strip'' camera (Martin et al. 1983), and the delay line camera (Lampton et al. 1987). The main problem of the CP40 is its limited maximum output rate of photon coordinates (25000 ph/s -- photons per second --). This limitation, which is due to the photon coordinate determination process causes artifacts in image autocorrelations. The Ranicon is even slower, limited to 10000 ph/s. Beyond this limit, pulse pile-up on the anode causes incorrect photon coordinate measurements. With these cameras the maximal signal-to-noise ratio is often reached on bright objects or with large apertures instruments, such as GI2T (Mourard et al. 1994). Although the PAPA camera can work at higher photon rate, it has field uniformity problems (due to the limited precision of the mask reimaging), as described by Lawson (1994). Wedge-and-Strip cameras are also affected by pulse pile-up, and by external magnetic deflection, requiring careful shielding (Timothy 1983).

Besides the maximum photon rate, another relevant characteristic of a photon counting camera is the temporal resolution (i.e. photo-event dating accuracy). Recently developed tools dealing with spatiotemporal photon coordinates could improve fringe tracking, especially in the case of space-borne interferometers (Koechlin 1985; Vakili & Koechlin 1989), or for moving object recognition (Morel & Koechlin 1997). These techniques require a higher temporal resolution than achievable by existing ICCD matrix photon counting cameras. Although MAMA camera prototypes (Timothy 1985) yield high maximum photon rate (106 ph/s), and high spatial and temporal resolutions (up to $4096 \times 4096$ pixels, with 100 ns event timing accuracy), their cost of realization makes this kind of equipment affordable for major projects only. Improvements of Wedge-and-Strip cameras have lead to hybrid detectors (Lampton et al. 1987) using a delay line approach for the x determination, and the charge partition system of Wedge-and-Strip for the y determination. The delay line system is very promising as it allows maximum count rates and temporal resolutions matching the MAMA performances. Nevertheless, an accurate delay measurement system is required, like the 4 ps resolution time-to-digit converter built by Lampton & Raffanti (1994). With a double delay line (Raffanti & Lampton 1993), the y-resolution of the camera is enhanced by reducing the anode capacitance. A detector using delay lines only for both x and y determination (crossed serpentine delay lines) has been recently built by Friedman et al. (1996).

Building the DELTA camera, our goal is therefore twofold. First, to provide high rates of accurate photon coordinates for speckle interferometry, dark speckle, fringe detection, and wave-front sensing. Second, to achieve high temporal resolution in order to test spatiotemporal methods (Morel & Koechlin 1997) with the best possible accuracy (actually, in the case of the DELTA camera, these two aspects are tied, as we shall see). Our goal is constrained by the requirements of low-cost and reliable technology. We have therefore selected commercially available components, and chosen solutions to avoid using high-precision photo-etching or high-vacuum equipment.


next previous
Up: The DELTA photon counting

Copyright The European Southern Observatory (ESO)