Transforming CCD data is difficult. Until recently, the general transformation problem was so poorly understood that there were no established design criteria. Now, the Hilbert-space approach to photometry tells us to match the standard passbands in the least-squares sense.

Unfortunately, the spectral responsivities of CCDs vary much more than do conventional detectors, and must be known to design filters. Each individual chip generally requires its own set of filters, if results are to be better than second-rate. Even so, spatial variations in spectral response limit the accuracy that can be reached, unless color terms are determined for individual pixels. Such CCD calibration problems involve the generally neglected optical properties of CCDs and the systems in which they are used.

Large observatories will have to establish calibration laboratories, and to abandon the common practice of moving “standard” filter sets from one instrument to another. Smaller institutions will need to purchase the services of such calibration laboratories if they are to produce accurate CCD photometry. These are major changes in the way astronomers are accustomed to working; but such changes are necessary to stem the backlash developing against CCD photometry.

The Spacewatch group at the University of Arizona's Lunar and Planetary Laboratory was probably the first (1984) to implement CCD-scanning in a major astronomical program. In the past three years, using a Tektronix 2048 × 2048 CCD, the program has discovered ∼ 45% of the new Earth approaching asteroids, measured astrometric positions for over 50,000 main belt asteroids, discovered two of the three known Centaurs, and found evidence for an unheralded population of small (∼ 10-m) objects in the inner solar system. This success is due to the automated Moving Object Detection Program (MODP) which searches successive scans over the same region for objects showing consistent motion. While visual examination of photographic plates may have a higher efficiency, an automated routine for detecting moving objects does not tire and is repeatable. Our recent work quantifies the efficiency of MODP as a function of the asteroid's magnitude, rate of motion, and orbital parameters. Other work suggests that the observational detection of faint, trailed, Fast Moving Objects may be improved by incorporating linear darkfield subtraction and flatfielding during scans.

CCDs are essentially the only instrument available today for photometry at most observatories; they are also becoming more readily available to amateurs as well. Thus, obtaining good photometric data with these two-dimensional devices is something we all need to understand. The history of and recent developments in CCD time-series photometry will be reviewed with some comments on future directions.

This talk is about the limits to the precision of stellar photometry in comparing one star with another in a single CCD frame. This is concerned with bright stars, and concentrates on three problems at the 0.1 percent level of accuracy: how to flatfield; how to deal with varying point-spread-functions that vary across an image; how to deal with the fact that the response inside a pixel is not uniform. The first is the well-known difficulty of getting a uniform illumination across the CCD to use as a flatfield; the use of a rotatable CCD mounting and of drift-scanning is discussed. The second depends on the ability to detect and define small, but significant, changes in the PSF. The third is the fact that the pixels of optical CCDs can have non-uniformities inside them of ten percent, and these when folded with the PSF produce systematic errors significant at the 0.1 percent level; with infra-red arrays these problems can be much worse. The use of software to model these variations and reduce these errors is described.

CCD photometry is capable of high internal precision, however there are several important requirements necessary to attain high precision in standardized photometry. Firstly, the CCD passbands must match as closely as possible the standard passbands; secondly, new faint standards must be set up in several declination zones and thirdly, for convenience a sufficient number of standards covering a good range in color should be obtained on a single CCD frame so that several different frames should suffice for standardization. Landolt has taken the first steps in defining several such fields. The small systematic differences between different UBVRI systems have been examined and transformations can be applied to the photometry of Landolt and Bessell to place it on the Cape — SAAO system.

In this paper we describe the technique and reduction procedure in use at the SAAO for differential CCD photometry. Much of this work involves the detection of microvariables in galactic and Magellanic Cloud open clusters. The large number of frames requires automated cleaning and reduction software (principally the DoPhot package) which is briefly described. The process of selecting comparison stars and the technique of detecting low-amplitude variables is discussed. The software is well suited to be used for on-line reductions at the telescope and a start has been made in this direction. Finally, some astrophysical results involving precision differential CCD photometry are described.

Array detectors have improved the efficiency of optical polarimetry sufficiently for this technique to become part of the standard arsenal of observational facilities. However, we could gain even more: spatially-differentiating photometry can be implemented as an option of array Polarimeters and low-noise, high-frame-rate array detectors will allow extremely high precision both in polarimetry and in such differentiating photometry. The latter would be valuable for analyzing many kinds of optical or infrared images of very low contrast; the essence of the technique is to use optical (and extremely stable) means to produce the spatial derivative of the flux image, in the form of a polarization image which is then presented to a “standard” array polarimeter. The polarimeter should incorporate a polarization modulator of sufficient quality for the photometric application in mind. If developed properly, using a state-of-the-art array detector and the most sensitive type of polarization modulator (stress-birefringence), optical differencing will allow levels of relative photometric precision not otherwise obtainable. With the optical differencing option taken out of the beam, the same instrument can be used for high-quality polarimetry.

We have found that many materials that are most commonly used in CCD cryostats are weakly radioactive and therefore contribute to what is rather liberally labelled as Cosmic Ray Events. Some standard optical glasses that are extensively used for lenses and optically flat windows such as UBK-7 contain large amounts of potassium 40, which renders them useless as windows in CCD dewars. Cobalt 60 is sometimes found in excessive amounts in some steel alloys, even in Covar used as a thermal match to silicon. The choice of materials that emit a minimum of ionizing radiation is discussed.

MOCAM is a wide field CCD camera, currently nearing completion, which will be offered to the Canada-France-Hawaii Telescope (CFHT) user community in 1995. The project is a collaboration between the CFHT, the Dominion Astronomical Observatory (DAO, Canada), the Institut des Sciences de l'Univers (INSU, France), Laboratoire d'Astrophysique de Toulouse (LAT, France) and the University of Hawaii (UH). In the interests of producing a reliable and effective camera in the shortest time, it was decided to use existing technologies rather than innovative ones. Two-edge buttable 2048 × 2048 15 μm pixel CCDs were obtained from the LORAL aerospace foundry, based on a mask designed by J. Geary at Smithsonian Astrophysical Observatory (SAO). They are mounted in a dewar designed by G. Luppino (UH); the focal plane mounting keeps the mosaic flat to within two pixels and the CCDs are aligned to within two pixels. A mechanical interface designed and fabricated by the DAO holds a 150 mm shutter and a filter wheel which has a positioning repeatability better than five μm.

The four CCDs are operated in parallel by a San Diego GenIII controller adapted by LAT. The mosaic is read out in seven minutes and a single 33 Mb FITS file is generated to enable convenient on-line preprocessing. The user will control the system through a single CFHT Pegasus environment session. The camera field is 14′ × 14′ with a 0.″2 pixel sampling and the readout noise is less than seven electrons. The scientific goals of the initiators of the project are studies of distant clusters, deep galaxy counts and quasars surveys.

The measurement of some physical parameters of astronomical objects can only be carried out with high resolution spectra. Unfortunately the high dispersion of the light on the detector restricts such observations to relatively bright sources. However, some spectral information can be concentrated into a single spectral “line” by a cross-correlation algorithm, allowing the observation of fainter objects. Such a technique, taken from the CORAVEL optical correlation, is presented. A complete description of the errors of the correlation function parameters is given and the minimum signal-to-noise ratio is also discussed. Finally, a short investigation of the best resolution needed to observe efficiently radial velocities and velocity broadenings is made.

We present some results about the use of an engineering grade NICMOS3 array for medium resolution spectroscopy. All the work described here was performed at the Infrared Laboratory of the Osservatorio Astrofisico di Arcetri in Florence and at the Italian Telescope TIRGO (Citterio 1978). The detector has been mounted on the LonGSp (Longslit Gornergrat Spectrometer), an instrument designed and set into operation in Arcetri and described in Gennari and Vanzi (1994).

This paper presents the principles, the techniques, and some of the astrophysical potential of spectral classification in the era of array detectors.

Targets within the solar system generally fall into one of two types: a) major planets (except Pluto) and our Moon; b) minor planets, comets, Pluto and planetary satellites. The first group is noteworthy for being reasonably bright. Most are also spatially extended. The inner planets never achieve large solar elongation. The second group comprises bodies which are generally faint. Comets are spatially extended. The minor planets and comets may be in orbits which are highly inclined or viewed at small solar elongations. Comets may even be in retrograde orbits. Planetary satellites may be bright or faint but suffer from being in the glare of the parent planet.

Over the last decade, the quantity of scientific results brought by the observations of very faint objects has been quite spectacular. In particular, they concern the photometry of faint galaxies up to B = 27 (Tyson 1988) or K = 22 (Cowie et al. 1994). The consequences of these observations are the detection of a large population of faint galaxies more numerous than any prediction given by standard galaxy evolution and probably a new vision of the distant universe. For faint object spectroscopy, the most recent surveys of field galaxies reach a magnitude range of 23–24 (Colless et al. 1990, 1993, Lilly and Cowie 1993, Tresse et al. 1993) with a reasonable S/N ratio that allows a redshift measurement from absorption-line identification. But in this magnitude range, the sky background flux is dominant with respect to the source, being at least 10 times brighter.

The worldwide application of CCDs in astronomy is concerned mainly with the problems of photometry and spectrophotometry. Astrometry with CCDs is limited by the small fields of view and now is used to solve some special tasks such as parallax determination of faint stars, observations of QSOs, etc. Very interesting results were obtained using a CCD with a meridian circle in a drift scanning mode to solve the problem of the extension of the reference system to stars of 16th–17th mag and the linkage between radio and optical reference frames (Stone 1994).

Based on the preliminary results of our observations, some proposals on astrometry with CCDs are discussed in the paper. The aim is to show the use of short-focus telescopes with CCDs in the mode of traditional astrographs. Numerous tasks may be solved with such instruments more simply and correctly than with meridian instruments in a drift scanning mode.

The problems of the application of CCDs with short-focus optics are also discussed. Some effects of CCD properties on the accuracy of astrometric and photometric measurements are also considered.

The Macho Collaboration uses a dedicated 1.27-m telescope (The Great Melbourne Telescope) at Mount Stromlo to make photometric measurements of tens of millions of stars per night searching for the gravitational microlensing signature of MACHOs in the halo and disk of the Milky Way. A prime focus corrector and dichroic beamsplitter provide red (6300–7800 å) and blue (4500–6300 å) foci with one degree fields. A two by two mosaic of 2048 × 2048 pixel CCDs in each focal plane provides simultaneous images of 0.5 square degrees. By August of 1994, more than 20,000, 32 megapixel images will have been collected of fields in the Large Magellanic Cloud (LMC), Small Magellanic Cloud and the bulge of the Milky Way. We have implemented an online analysis system which produces photometric reductions of a night's data (five Gbyte of images) within 24 hours. This system allows us to identify and follow interesting events in real-time. In this search, we have identified more than 60,000 variable stars, and a preliminary analysis of their types and distribution will be presented. Microlensing events toward the LMC and the bulge have been discovered and detection efficiencies are being calculated to constrain the MACHO content of the Milky Way's halo.

Fiber-fed multi-object systems are now the preferred instruments for gathering spectroscopic data on survey scales. The technique lends itself particularly well to telescopes with fields of 30 arcminutes or more. This paper gives a broad overview of the instrumental considerations involved in its implementation.

TH78997M is a 2048 × 2048 pixel full-frame CCD sensor featuring 15 × 15 micron pixels, MPP operating mode, four parallel outputs and buttability on three sides. This new device makes it possible to build very large focal plane detectors for telescopes (two × n butted CCDs) with less than a 25 dead pixel zone between adjacent sensors. Other features are high full-well capacity (typically 180 Ke−) and very low dark current thanks to four phase MPP clocking which allows a very long exposure time together with a high S/N ratio.

The readout of a complete frame is achieved through four parallel outputs running at frequencies ranging from ten KHz up to five MHz per output. The four on-chip amplifiers have been designed to be completely user controlled and to meet most of terrestrial and spaceborne astronomy requirements: very low noise (less than 4e− at 50 KHz, −40 degrees C), very low power consumption (150 micro-W/amplifier at 10 KHz/output), linearity better than 0.5% together with high conversion factor (4.2 micro-V/e−) over 2.5 V output range. The packaging of the device is compatible with the good flatness of the chip and with the low parasitic “cosmic event” rate thanks to specific care in selecting the packaging material.

We outline the recent developments in CCD imager technology aimed at producing the very large format (8192 × 8192 pixels and larger) detector mosaics required for existing 4-m class and new 8 to 10-m class telescopes. The key technology areas include buttable array design and buttable element packaging, and optimization schemes for QE and readout time. As an example, we highlight the University of Hawaii effort to develop an 8192 × 8192 15 μm pixel CCD mosaic.

The talks at this IAU Symposium have illustrated the spectacular development which has taken place in the last decade in the field of array detectors for astronomy. Just a few years ago it was possible to speak of two-D detectors for the UV-red wavelength range only. At this meeting we have witnessed presentations on array characteristics from the extreme UV (Bonanno 1995), through the blue-visual range (D'Odorico 1995, Jorden and Oates 1995, Iwert 1995 and Luppino et al. 1995); the infrared 1 to 5 μm window (McLean 1995, Finger et al. 1995, Gilmore et al. 1995, Fazio 1995, Glass et al. 1995 and Ueno et al. 1995); the 10–20 μm window (Fazio 1995, Gezari 1995) and finally to an array of bolometers to operate at submillimeter wavelengths (Moseley 1995). Field imaging and spectroscopy are now possible across this entire energy spectrum and some of the first exciting astronomical results obtained with these devices have been presented here.