An ultra high resolution image sensor has been developed for industrial and scientific applications. The ultra high resolution sensor is a full-frame CCD sensor, consisting of 2048 x 2048 pixels, and its image area measures 18.43 mm x 18.43 mm. The pixel size is 9 x 9 microns. The sensor has dual readout registers to increase the data rate and can be operated in the single or dual readout register mode. The architecture of this imager is suitable for accumulation mode operation, which results in a very low dark current of 10 pA/sq cm at room temperature.

A description is presented of the design and performance of several large-area CCD imagers, including large 1-in. TV imagers, and very large multi-output buttable devices. A photocomposition technique is described which permits the area limitations of the wafer stepper to be surmounted, thus allowing the manufacture of very large devices. The design and performance of a 1-in. format 875-line CCD are reported, as well as the 525- and 625-line variants. A series schematic and design specifications are given for a three-phase, nonantibloomed, buried channel frame transfer CCD. The incorporation of the photocomposition technique, called stitching, in CCD devices is explained. Performance testing results and previous applications are mentioned. The CCDs show low readout noise and may be butted into large arrays, and can handle high amounts of charge at high speed. The products are shown to achieve their design targets, and an extra degree of freedom in image-area limitations is created by applying the autoalign capability of the wafer stepper.

A 1024 x 1024 pixel, interline charge-coupled device (IL CCD) image sensor has been developed that incorporates antiblooming and electronic exposure control while eliminating lag and obtaining a high responsivity. Of the novel features of this device are its noninterlaced, or progressive-scan architecture and dual-horizontal registers that can be used to clock out the image area by one or two lines at a time. These features make it well suited for applications demanding high-resolution-image capture from a single, high-speed scan.

A 2/3-inch format 768(H) x 576 (V) pixel CCD image sensor has been developed, fabricated and tested. It has a frame transfer organization and is incorporating a lateral built in anti-blooming system. A horizontal resolution of more than 500 TV lines has been obtained and photosensitivity reaches 30 mv/lux. By the possible use of two video outputs, an image mirror function is provided. Thanks to its windowing device, the sensor is well suited for tracking applications.

The CCD194 image sensor is described in terms of the high-performance, low-light-level, high-resolution applications (graphics, reconnaissance, and photogrammetry) for which it has been designed. The CCD194 design comprises two monolithic 6000 by 1 CCD image sensors in a body consisting of an aluminum header and an antireflective sapphire window. The photosite arrays are butted end-to-end and linearly aligned, separated by a distance of 20 microns. The peak-to-peak nonlinearity and nonplanarity are less than 15 microns for the entire 120-mm photosite span. A distance of 20 microns separates the arrays. The electrical characteristics of the device are tested; photosite transfer loss is measured. The results show a dynamic range of 15,000 to 1 with respect to rms noise, charge transfer inefficiency of 10 to the -6th/transfer, and a photoresponse nonuniformity of 15 percent. The line-scan imager is shown to function in high-performance, high-resolution, low-light-level applications. A 32 MHz effective data rate is provided by dual outputs on four chips.

The use of charge-coupled devices (CCD) for quantitative microscopy has advantages over standard video detectors. For quantitative imaging in the life sciences the use of CCDs have the following advantages: it is a quantitative photodetector, with a large dynamic range and high quantum efficiency. The advantage of a large dynamic range of the order 16,000:1 (14-bits) is important in confoca]. microscopy where there are regions of extremely low light intensity and regions with high intensity. The linear factor of the dynamic range is important for quantitation of the images. Another important property for biological imaging is sensitivity. The use of back- illuminated, UV enhanced coatings, thinned CCD devices, with anti-reflection coatings all result in higher quantum efficiencies. Slow scan devices can be used in a special mode to capture an image in less than a video frame, however, they are most useful as linear integrating light detectors. In order to demonstrate some of the useful properties of CCD detectors, we have used a CCD detection system to image low light level signals from living transparent biological material. A CCD camera was coupled to a confoca]. microscope and images were collected in reflected light. The samples included the cornea, and the in situ ocular lens. The quality of the images is demonstrated over a wide range of light levels. The images of the eye at submicron resolution clearly demonstrate the advantages of using a solid state charge-coupled device detector for quantitative confocal reflected light microscopy of thick, transparent ocular tissue.

The characteristics of a quadrant-CCD designed for pointing and tracking are discussed with reference to its use in an adaptive optics program. The quadrant-CCD is used to correct stellar image motion proceeding from atmospheric turbulence by means of a system in which a sensor measures the image offset and sends data to a high-speed tip/tilt mirror. The design and control of the device are detailed, including four 100-micron-square pixels, the quadrant architecture, controller electronics, and data acquisition computer and interface. The transfer function is set forth in the x and y directions, and transfer curves are shown. A laboratory simulation of random image motion was conducted to evaluate the performance of the quadrant-CCD as an image motion sensor, and the experimental results are presented. The suitability of the quadrant-CCD for space-based pointing and tracking uses is demonstrated by this test and two earlier theoretical studies. Some future developments which improve performance capabilities are mentioned.

The large physical size of the TK2048 CCD and the high quantum efficiency of the silicon photosensitive surface makes possible X-ray quantum noise limited imaging in both lens coupled and fiber optic coupled configurations that afford advantages over current state-of-the-art X-ray image intensifier based systems. This paper discusses the system tradeoffs and includes examples of X-ray images made using the 2048x2048 pixel, 55mm square format TK2048 CCD.

A large-format CCD imager is described and tested. The CCD imager incorporates floating diffusion as well as floating gate amplifiers on a 2048 by 2048 format which was employed as the design base. The amplifiers are intended to allow repeated nondestructive read operations on individual pixels in the array. The serial register was separated into two independently clocked halves to permit simultaneous readout of all four quadrants of the imager. Extensive schematic layouts of the base model and modification are given. The results of a performance test are presented, showing good results in the cooling curve for average dark current, and for charge transfer characteristics. The amplifiers are intended to reduce net readout noise, and the simultaneous readout capability is intended to reduce total read time, although neither was fully tested. The large-format CCD imager is of interest for astronomical photography and spectroscopic applications.

The paper presents the use of deep level transient spectroscopy to characterize trapping centers in CCD imagers. A discussion is presented regarding the effects of UV illumination and elevated temperature annealing in a hydrogen-rich environment. Two trapping centers are described, and the annealing experiments suggest techniques for extending CCD lifetimes.

This paper discusses recent progress in the understanding of the fabrication of low noise floating diffusion output amplifiers for special purpose charge-coupled devices. Emphasis has been given to reducing the total node capacitance and increasing the output sensitivity. Measurements of noise on experimental devices has yielded noise less than 3 electrons rms at 50 kpixel data rate.

The paper presents the design of a CCD-based codec preprocessor (CP) integrated with an areal imager. 3 x 3 pixel neighborhood blocks, where the center pixel arrives first, are utilized for the lossless compression algorithm in the image coding scheme. A 256 x 256 buried-channel frame transfer device with 15 x 15 square microns pixels is employed as the imager. The neighborhood reconstruction is effected by means of both row regrouping and pixel resequencing operations. Four designs for CCD CP chips are described: two hybridized to the imager array, and two integrated with the imager. The chips operate at a 30 Hz frame rate with power dissipation at less than 10 uW. The size of the hybrid chips is 2.5 x 5.5 sq mm, and the focal-plane chips' CPs require an area of 250 x 400 square microns. The CPs provide differential output appropriate for lossless coding and compression to off-chip electronics when reorganization of the image data into local 3 x 3 neighborhood blocks is complete.

The design of a CCD image half-toner integrated monolithically on the focal plane with a 256 x 256 frame transfer imager is reported and the algorithm used is discussed. The imager/half-toner chip is projected to achieve a throughput of 30 frames per second.

The retina is a smart sensor, but in the sense of intelligent design and not on-chip computing power. It uses a unique layout and elementary charge computing elements to implement in hardware a polar-exponential transform on visual data. The final chip includes a large section of photosites arranged in a circular pattern. Further, the pixels grow m size as radial distance increases. The retina also has a fovea (a high resolution area at the chip's center) and the computational circuitry. The sensor works and will serve as the key component of a real-time imaging system.

Autonomous systems have to overcome two major problems in interpreting the data collected by sensors. These problems are due to the fact that multiple real world scenes can produce a given intensity distribution and that sensor data is often corrupted by noise. These problems can be solved by imposing certain a priori knowledge about the world to arrive at a unique solution. This reduces early vision problems to constrained optimization problems, which can be conveniently formulated as energy minimization problems in the framework of Regularization Theory. These energy functions can be implemented in analog hardware. This feature makes this approach very attractive for real time autonomous systems. We are presently developing algorithms and analog CMOS chips based on this approach for focal plane image processing.

The development of technology which integrates a four phase, buried-channel CCD in an existing 1.75 micron CMOS process is described. The four phase clock is employed in the integrated early vision system to minimize process complexity. Signal corruption is minimized and lateral fringing fields are enhanced by burying the channel. The CMOS process for CCD enhancement is described, which highlights a new double-poly process and the buried channel, and the integration is outlined. The functionality and transfer efficiency of the process enhancement were appraised by measuring CCD shift registers at 100 kHz. CMOS measurement results are presented, which include threshold voltages, poly-to-poly capacitor voltage and temperature coefficients, and dark current. A CCD/CMOS processor is described which combines smoothing and segmentation operations. The integration of the CCD and the CMOS processes is found to function due to the enhancement-compatible design of the CMOS process and the thorough employment of CCD module baseline process steps.

Recent research efforts aimed at optimizing charge-coupled devices (CCDs) after their manufacture to achieve maximum quantum efficiency, wide spectral bandpass, and excellent cosmetics and surface flatness are discussed. We present results of a new acid thinning agitation technique which produces very uniform, high quality surfaces on large area square and rectangular CCDs and 4" silicon wafers for back illuminated operation. In particular we present thinning results of Ford Aerospace 2048x2048 pixel CCDs. A method of cleaning thinned CCDs before antireflection coating for increased QE is also discussed. The results of initial experiments with a new packaging method to mount thinned CCDs while maintaining a very flat imaging surface are presented. This bump bonding mounting technique increases yield due to reduced handling and robust packaging and is expandable to tightly packed large area focal plane mosaics.

A structure for the virtual transfer of charge packets across metal wires is described theoretically and is experimentally verified. The structure is a hybrid of charge-coupled device (CCD) and bucket-brigade device (BBD) elements and permits the topological crossing of charge-domain signals in low power signal processing circuits. A test vehicle consisting of 8-, 32- and 96-stage delay lines of various geometries implemented in a double-poly, double-metal foundry process was used to characterize the wire-transfer operation. Transfer efficiency ranging between 0.998 and 0.999 was obtained for surface n-channel devices with clock cycle times in the range from 40 nsec to 0.3 msec. Transfer efficiency as high as 0.9999 was obtained for buried n-channel devices. Good agreement is found between experiment and simulation.

This paper reports on two new advancements in CCD technology. The first area of development has produced a special purpose CCD designed for ultra low-signal level imaging and spectroscopy applications that require sub-electron read noise floors. A nondestructive output circuit operating near its 1/f noise regime is clocked in a special manner to read a single pixel multiple times. Off-chip electronics average the multiple values, reducing the random noise by the square-root of the number of samples taken. Noise floors below 0.5 electrons rms are reported. The second development involves the design and performance of a high resolution imager of 4096 x 4096 pixels, the largest CCD manufactured in terms of pixel count. The device utilizes a 7.5-micron pixel fabricated with three-level poly-silicon to achieve high yield.

It is well demonstrated that the CCD TDI mode of operation provides increased photosensitivity, relative to a linear CCD array, without the sacrifice of spatial resolution. However, in order to utilize the advantages which the TDI mode of operation offers, attention should be given to the CCD TDI design and tradeoffs which exist between the pixel pitch, dead space between pixels in the horizontal direction, high speed, high photosensitivity, high spatial resolution and wide dynamic range. For example, a 2000 pixel ThI with an MTF of 0.5 will have an effective spatial resolution of only 1000 pixels. Other tradeoffs also are present, such as high speed versus power dissipation which exist both in the array and in the on-chip output video amplifier. Further, noise considerations exist at the video on-chip output amplifier. These include high speed which demands high bandwidth, Johnson noise, 1/f noise, reset noise, and output signal charge to voltage sensitivity. In this paper we shall describe the novel approach which we used in the design of a 2048 x 96 TDI CCD imager. We shall show how we addressed successfully these design tradeoffs and issues. Details about the analysis and design of a high speed on-chip output amplifier will also be presented. In addition, we shall explain how the use of buried channel MOSFET's enabled us to address and solve the power dissipation, speed, and noise issues.

We start with a physical model1 of the detector and associated electronics. Using previously derived expressions for the MTF and noise power spectrum, we extend the use of the model to include a signal-to-noise metric, the noise equivalent quanta (NEQ). This approach is then applied in a design example relevant to film and document scanning.

A film digitizer employing a linear CCD imager can provide substantial performance improvements over a flatbed or drum scanner by measuring many film pixels simultaneously. If it is required, however, to retain the full quality of the image on the film, careful design of the illumination system, optics, and electronics is important. This paper will describe the application of a diffuse illumination system and a three-color CCD imager in a high-resolution film digitizer. The imager provides up to 8000 pixels per scan line and dynamic range (peak signal-to-darknoise) of 5000. This technology permits use of a novel means for holding the film, simplifying the task of mounting the film on the scanner. A brief description will be given of the analog electronics, digital electronics, and software procedures used to convert raw CCD output to accurate film density.

A CCD computer-input camera is described which utilizes the virtual phase method for clocking. The camera was designed to acquire images by means of a PC, at low cost and with minimum parts count, and to allow for computer control of imaging parameters. Design components are listed, including ICs and image sensor, a low-resolution monochrome sensor, an 8-bit flash converter, and a programmed I/O. The software system is presented, including enhancements such as a method for avoiding streaking. Camera performance characteristics are listed. The camera resolution is shown to be 192 by 330 pixels with interlaced scanning. The camera permits electronic exposure control, and 8 bits per pixel gray scale. The camera is of interest in astronomical and other scientific applications, primarily due to its S/N and linearity.

A charge coupled device designed for celestial spectroscopy has achieved readout noise as low as 0.6 electrons rms. A nondestructive output circuit was operated in a special manner to read a single pixel multiple times. Off-chip electronics averaged the multiple values, reducing the random noise by the square root of the number of readouts. Charge capacity was measured to be 500,000 electrons. The device format is 1600 pixels horizontal by 64 pixels vertical. Pixel size is 28 microns square. Two output circuits are located at opposite ends of the 1600 bit CCD register. The device was thinned and operated backside illuminated at -110 degrees C. Output circuit design, layout, and operation are described. Presented data includes the photon transfer curve, noise histograms, and bar-target images down to 3 electrons signal. The test electronics are described, and future improvements are discussed.

Three Tektronix 1024 x 1024 multi-pinned-phased (MPP) charge coupled devices were irradiated with protons to obtain data on CCD performance degradation in a proton radiation environment. The devices were irradiated with a spectrum of energies up to 120 MeV, simulating the total radiation dose of a long-term space experiment. Basic parameters such as charge transfer efficiency, dark current, noise, and full well were measured before and after irradiation. A test was also performed to determine the effectiveness of various thicknesses of tantalum shielding in protecting the CCD from damage. Dark current increase and CTE degradation were the most noticeable effects of proton radiation. This paper will present the objectives, test data, and conclusions of the proton testing, and will identify future testing to be performed.

Charge transfer efficiency (CTE) test methods are reviewed, and the results and conclusions of the tests are given. The test methods have been utilized to describe the CTE characteristics of the Tektronix 1024 by 1024 CCD to optimize low dark current, low readout noise, and high CTE at low signal levels. CTE modelling is described, and three test methods are set forth and compared. The Fe-55 X-ray response method utilizes the response of a CCD to X-ray photons from the radioactive source Fe-55. The extended pixel edge response method employs the measurement of the charge lost to successive pixels by a known initial signal as it is shifted through the array. The charge injection method consists of charge injection through the output amplifier reset transistor. These measurements were performed on several devices with known CTEs. The CTEs are found to be in agreement for the three methods, making application and test requirements the principal criteria for their use.

Large screen television displays have been developed for use in stadiums, aircraft simulators, and large auditoriums. Electronic display of these images will become even more widespread with the introduction of high definition television. Consumer applications of HDTV require a large (about 3' x 5 1/3') bright display at a relatively low cost. This paper will discuss the properties of commercially available large screen television displays and evaluate the options for the new technologies that could satisfy the requirements of the future applications.