It has been 17 years since Boyle and Smith of Bell Laboratories announced in March of 1970 the concept of charge coupling. The original form of their invention was a simple 8 bit register that was used to demonstrate a semiconductor equivalent to the magnetic bubble memory. Since then, the charge-coupled device (CCD) has been the subject of active research at laboratories throughout the world, where efforts have been directed toward signal processing and imaging applications (the CCD memory never got off the ground). It is in the field of imaging applications that CCDs have clearly made their greatest contribution, revolutionizing both scientific and commercial optical imaging instrumentation.

Recent developments of backside treatment for the backside-illuminated scientific charge-coupled device (CCD) imagers have shown near-theoretical efficiency even at the short wavelength region of the spectrum. By using a scanning electron microscope (SEM), we report here, for the first time, performance comparisons of backside-treated and untreated CCDs to an electron flux varying from 1 to 100 pA and beam energy ranging from less than 1 keV up to 20 keV. We describe the theoretical analysis, the SEM testing procedure, and the quantum efficiency measurement results. It is shown, for example, that the average quantum efficiency increases from less than 1% for an untreated CCD to nearly 40% for a backside-treated CCD at a beam energy of 1 keV.

The charge-coupled device dominates an ever-increasing variety of scientific imaging and spectroscopy applications. Recent experience indicates, however, that the full potential of CCD performance lies well beyond that realized in devices currently available. Test data suggest that major improvements are feasible in spectral response, charge collection, charge transfer, and readout noise. These properties, their measurement in existing CCDs, and their potential for future improvement are discussed in this paper.

The electro-optical characterization of the first in a new series of Tektronix CCDs is described. This device, the TK512M-011, is a frontside-illuminated CCD with a 512 by 512 format and 27 by 27 um pixels. Electro-optical characteristics measured in this study include linearity, blooming, dark count rate, charge-transfer efficiency (CTE), and quantum efficiency. The results of a detailed study of the noise characteristics of the CCD output FET are reported. The TK512M-011 has excellent photometric lin-earity, high well capacity, and a low dark count rate. Very good low light level CTE is observed in the parallel shift direction; however, CTE problems are observed in the serial direction. The quantum efficiency of the front-side-illuminated CCD over the wavelength range of 400 to 1000 nm is lower than expected based on experience with similar devices. The noise of the output FET of the CCD is equivalent to 5 to 12 electrons, depending on the FET operating conditions and system bandwidth. A very small latent image effect is noted. The conclusion of this evaluation is that despite the problems observed, the frontside-illuminated Tektronix CCD is an excellent sensor for scientific imaging applications.

The improved cosmetic quality, charge-transfer efficiency, and low noise of present-generation CCDs means that modern two-axis star trackers can measure star positions to accuracies of 1/50 of a pixel over the whole chip area. However, CCD selection, testing, screening for enhanced reliability, and careful attention to operating conditions are still necessary. This paper is concerned primarily with the performance, testing, and qualification of English Electric Valve Co. P8604 CCDs for use in star trackers for the German Roentgen-Satellit (ROSAT) x-ray astronomy satellite, although the tests developed should be widely applicable. It was found that the best-quality devices are essentially free of all image blemishes and have a white noise of 15 nV/Hz deferred charge of less than 1 %, response uniformity of 1.5% rms, and dark signal nonuniform ity at 40°C of 30 electrons rms. All of the candidate devices from flight batches passed a burn-in under voltage and temperature stress that are predicted to be equivalent to more than seven years' continuous mission operation.

A double-density (15 µm X15 µm pixels) RCA charge-coupled device has been tested and commissioned for astronomical imaging and spectroscopy at the Canada-France-Hawaii telescope (CFHT). This thinned device has had its support glass removed and has received a broadband antireflection coating to improve the UV and blue response. Presented here are characterization measurements resulting from 10 months of laboratory testing and use at the telescope. Although the read noise is not as low as with other CCDs available, the high pixel density and the permanent quantum efficiency enhancements provided by this RCA device give some advantages over other sensors when used at the CFHT.

A calibrated tungsten lamp was used to determine the quantum efficiency (QE) of an RCA CCD in the ultraviolet. A set of narrowband interference filters from 280 to 420 nm gave eight absolute QE points, and a double monochromator was used to measure the relative run of the QE in this range. The data show that the CCD is sensitive down to 280 nm; however, the value of the QE is less than 10% below 320 nm.

The salient electro-optic properties of typical Fairchild model 222 interline transfer charge-coupled devices when exposed to single impulses of visible light (100 ps to 10 As) and with subsequent fast single-frame readout (4 ms) are presented. Identification and differentiation between CCD video components arising from the intended image area or photosites and those arising from the vertical registers are discussed, with emphasis on possible ambiguities resulting from improper or random combination of the two components. In addition, CCD/camera signal-to-noise ratio and resolution as functions of readout rate and spectral input are included. Fast readout philosophy, design criteria, and circuitry are discussed.

We describe a model that predicts the performance characteristics of charge-coupled-device (CCD) detectors being developed for use in x-ray imaging. The model accounts for the interactions of both x rays and charged particles with the CCD and simulates the transport and loss of charge in the detector. Predicted performance parameters include detective and net quantum efficiencies, split-event probability, and a parameter characterizing the effective thickness presented by the detector to cosmic-ray protons. The predicted performance of two CCDs of different epitaxial layer thicknesses is compared. The model predicts that in each device incomplete recovery of the charge liberated by a photon of energy between 0.1 and 10 keV is very likely to be accompanied by charge splitting between adjacent pixels. The implications of the model predictions for CCD data processing algorithms are briefly discussed.

Depending on the device structure and hence the electric field present within a CCD, there is often a considerable time for charge clouds generated within the silicon to diffuse and spread outward before being collected at the buried channel. This charge spreading has an important effect on the spatial and energy resolution of CCDs used for soft x-ray, optical, and high energy particle imaging. Further theory on this is presented with regard to the performance of CCDs currently under development, to broaden their spectral response while minimizing the thickness of the field-free region below the depletion layer. It is shown that the charge cloud profiles resulting from field-free diffusion have broad wings that can lead to significant charge splitting between adjacent pixels, even for field-free region thicknesses ~1 um. Formulas are also presented for diffusion in the field region and the substrate.

Most scientific imaging applications of CCDs require low noise and high charge transfer efficiency. These attributes have been exhibited in devices produced by a number of manufacturers, but not all perform well as x-ray imaging spectrometers. This application requires that all charge liberated by an x-ray photon be completely collected within the original pixel. Processes of charge diffusion may prevent this from occurring and degrade the obtainable energy resolution. Measurements of charge diffusion and its effect on x-ray performance are presented for three types of English Electric Valve Co. (EEV) CCDs. These are fabricated on bulk and epitaxial wafers with resistivities of 1000 and 1011 cm. A range of different pixel sizes is employed. Measurements of detection efficiency are compared with predictions of a theoretical charge-diffusion model. Implications for optimizing the CCD structure for use in an x-ray astronomical spectroscopy application are noted.

We describe an astronomical camera for the 200-in. Hale telescope using four 800 X800 Texas Instruments CCDs in an optical arrangement that allows imaging of a contiguous 1600-pixel-square region of sky. The system employs reimaging optics to yield a scale of 0.33 arcsec per pixel, a good match to the best seeing conditions at Palomar Observatory. Modern high-efficiency coatings are used in the complex optical system to yield a throughput at peak efficiency of nearly 50% (including the losses in the telescope), corresponding to a quantum efficiency on the sky of about 30%. The system uses a fifth CCD in a spectroscopic channel, and it is possible to obtain simultaneous imaging and spectroscopic observations with the system. The camera may also be used in a scanning mode, in which the telescope tracking rate is offset, and the charge is clocked in the chips in such a manner as to keep the charge image aligned with the optical image. In this way, a survey for high-redshift quasars has been carried out over a large area of sky. The instrument has produced images for the most distant clusters of galaxies yet discovered as well as spectra of the most distant galaxies yet observed.

A two-dimensional detector consisting of a 40-mm-diameter x-ray fluorescing phosphor, coupled with an image intensifier and lens to a CCD image sensor, was developed for protein crystallography with synchrotron x rays. The intense x radiation from a synchrotron source could, with a suitable detector, provide a complete set of diffraction images from a protein crystal before the crystal is damaged by radiation (2 to 3 min). The CCD-based detector was evaluated in an experiment with a rotating anode x-ray generator to determine its suitability in this application. Diffraction patterns from a lysozyme crystal obtained with this detector were compared to those obtained with film. The two images appear to be virtually identical. The flux of 104 x-ray photons/s was observed on the detector at the rotating anode generator. At the 6-GeV synchrotron being designed at Argonne National Laboratory, the flux on an 80 X80 mm2 detector is expected to be >109 x-ray photons/s. The projected design of such a synchrotron detector shows that with a CCD pixel full-well capacity of 7 X105 diffraction peaks containing 3 X 106 x rays could be recorded. With an exposure time of 0.5 s and an additional 0.5 s readout time of a 512 X512 pixel CCD, the data acquisition time per frame would be 1 s so that ninety 1° diffraction images could be obtained, with approximately 1% precision, in less than 3 min.

The Lick Observatory CCD data acquisition system is described, with some observational results to illustrate the system capability. The electronics for the CCD are subdivided into those attached to the dewar, a "smart" controller near the dewar, and a computer connected by serial link to the smart controller. Software for the controller is in assembler code, while the software for data acquisition and on-line analysis is written in C and uses the UNIX operating system. The computers and controllers are programmed to recognize and operate several different types of CCD. Three separate instruments that use the CCDs are described briefly, together with examples of the data they produce.

A space- and time-resolving EUV spectrometer (SIRS) for fusion plasma impurity diagnostics is in operation at the DIII-D tokamak at GA Technologies in San Diego. An array of three modified Fairchild CCD cameras behind an EUV-to-visible converter/intensifier is used to record spatially resolved spectra in the 50 to 370 A range. Modifications and extensions to the Fairchild camera electronics allow a wide range of spatial binning on the CCDs and 2 to 50 ms time resolution, as well as computer control of the image format and timing. The CCD images are digitized and stored in buffer memories for computer acquisition. This paper briefly describes the spectrometer and the microchannel plate intensifier detector. The design of the imaging electronics is presented, along with the modifications necessary to obtain good flexibility and imaging quality. Results from photometric calibration work are presented.

This paper discusses preliminary investigations of a novel interferometer approach for testing cylindrical optics that makes use of an optical fiber as a cylindrical reference surface.

The design of a processor can vary considerably with the type of technology (optical or electronic, analog or digital), the number system, and the coding scheme used for the number representation. Binary number representation is accepted as the best suited for electronic computers. However, the delay due to carry propagation in binary arithmetic makes the binary number representation a very weak candidate for an optical processor that is inherently parallel. The modified signed digit (MSD) number representation satisfies the requirements of totally parallel addition using modular or identical units and allows the addition of any two numbers in three successive steps. In this paper, we show the design of a parallel optical adder based on MSD number representation using the method of symbolic substitution originally proposed by Karl-Heinz Brenner [Appl. Opt. 25(18), 3061-3064 (1986)] and Alan Huang [Proc. IEEE Int. Optical Computing Conf., pp. 13-17 (1983)]. Polarized light is used to code the inputs and outputs. We also discuss the use of this adder along with barrel shifters to efficiently implement multiplication.

The error rates given in Sec. 3 as 0.06% and 0.04% should actually be 6% and 4%, respectively.

This volume is a collection of individual monographs organized in terms of application areas for passive materials and in terms of the specific optical effect for active materials. In this sense, active optical materials refers to those materials for which the incident electromagnetic energy changes the optical characteristics of the medium. All discussions are with particular reference to laser-related applications.