The performance demands on uncooled IR imagers are gradually increasing to the point that it appears they will soon be within striking distance of the background radiation limit. This paper addresses the realities facing progress toward that end. A detailed analysis of signal and noise shows that the background-limited performance of thermal detectors is not much different from that of photon detectors, under certain conditions. However, the analysis also shows that these distinct detector types are optimally suited for different types of applications. The barriers to achieving background- limited sensitivity are quite different for thermal detectors. In this paper we quantify the barriers, and discuss their implications.

Sanders IR Imaging Systems (IRIS), a Lockheed Martin Company, has made recent improvements in high performance uncooled IR focal plane arrays and systems. This paper provides performance results for three of these new FPAs and systems. First we discuss a new 320 X 240, 46 micrometer pitch FPA, which when put into a system with a transmission of 74%, will provide a system NETD of < 26 mK (F/0.8, 60 Hz). This FPA has a power of < 250 mW (which includes on-chip 14 bit analog to digital conversion), and virtually no crosstalk from saturation. Second, we discuss the first ever 640 X 480 element uncooled IR camera. This camera, which is based on a 28 micrometer pitch microbolometer staring FPA, produces a system sensitivity of < 150 mK, (F/1, 30 Hz) and has a Minimum Resolvable Temperature Difference of < 0.4 degrees Celsius at the Nyquist frequency. Finally, we have developed a new lightweight thermal weapons sight (TWS). Our TWS, which weighs < 3 lbs. (with battery) and operates over the -37 degrees Celsius to +49 degrees Celsius temperature range, has demonstrated a boresight retention of < 0.2 mrad after 1000's of rounds were fired.

Amorphous semiconducting Y-Ba-Cu-O is attractive as the temperature sensitive element for uncooled IR bolometers and pyroelectric detectors. Thin films can be easily fabricated by RF magnetron sputtering at room temperature from a composite target. It is compatible with micromachining techniques for the fabrication of thermally isolated structures. As a bolometer, Y-Ba-Cu-O possesses a relatively high temperature coefficient of resistance of 3.5% K^-1 near room temperature. This paper will present the IR characteristics of 40-micrometer X 40-micrometer microbolometer arrays fabricated in thermal isolation structures. These detectors are aimed at thermal imaging at 10-micrometer wavelength. Recently, self-supporting YBaCuO pixels have been developed. In this case, the Y-Ba-Cu-O thin film pixel requires no underlying bridge material to provide structural support. The Y-Ba-Cu-O thin film is supported solely by the electrode arms. Responsivity and detectivity greater than 4 X 10³ V/W and 10⁸ cmHz^1/2/W respectively have been measured in these detectors. The development of large area 1.5-mm and 0.4- mm square YBaCuO bolometers for NASA's global warming studies in low-orbiting satellites will also be presented. These large area detectors require large optical bandwidths covering the 0.3-micrometer to 100-micrometer wavelength band.

Polycrystalline Silicon Germanium is a useful material for CMOS compatible uncooled IR bolometer manufacturing due to its excellent material characteristics such as low stress and high TCR. However, for IR imaging applications, fast and yet sensitive detectors are required. We managed to combine these two contrasting characteristics by fabricating very thin devices. This was only possible thanks to a new release technique based on vapor HF at elevated wafer temperatures, and to structural stiffness enhancement of the devices by applying U-profiles. Furthermore, a performant and low cost on-chip vacuum package has been developed. The combination of these features is applied in linear arrays of bolometers which are read-out by a dedicated noise reduction circuit using an MCM board.

The uncooled infrared cameras are now available for both the military and commercial markets. The current camera technology incorporates the fruits of many years of development, focusing on the details of pixel design, novel material processing, and low noise read-out electronics. The rapid insertion of cameras into systems is testimony to the successful completion of this 'first phase' of development. In the military market, the first uncooled infrared cameras will be used for weapon sights, driver's viewers and helmet mounted cameras. Major commercial applications include night driving, security, police and fire fighting, and thermography, primarily for preventive maintenance and process control. The technology for the next generation of cameras is even more demanding, but within reach. The paper outlines the technology program planned for the next generation of cameras, and the approaches to further enhance performance, even to the radiation limit of thermal detectors.

LETI LIR has been involved in Amorphous Silicon uncooled microbolometer development for a few years. This silicon IR detection is now well mastered and matured so that industrial transfer LETI/LIR technology is performed towards Sofradir. Industrial production of 320 X 240 mirobolometer array with 45 micrometer pitch started. After a short description of the technology and the readout circuit architecture we focus on device reliability which is the key point for microbolometer application. Methodology for reliability enhancement is described. First results obtained on amorphous silicon reliability are presented. Electro-optical results obtained from an IRCMOS 320 X 240 with 45 micrometer pitch are presented. NEDT close to 70 mK can be obtained with our standard microbolometer amorphous silicon technology.

Prototypes of VO₂-based bolometric detectors with lateral dimensions of 25 X 25, 30 X 30, 35 X 35, 40 X 40 and finally 50 X 50 micrometers² and fill factors approaching 90% are presented. These detectors are grouped in hardwired linear arrays as large as 512 X 1 pixels. Under DC biasing, the fabricated detectors, even the smallest ones, exhibit responsivities from 48,000 to 120,000 VW^-1, detectivities in the range of 1.5 X 10⁸ cm Hz^1/2W^-1 and response times in the range of 5 ms. These new bolometric detector structures contain hidden-legs placed completely underneath the bolometer platform. Results of simulations of the mechanical, optical and electrical properties of these new detector structures are presented. A complete detector fabrication process flow is described.

The advent of modern photolithography and micromachining techniques has led to the development of many kinds of infrared sensitive focal plane arrays. This paper outlines the history of the development of modern uncooled thermal detector arrays, considerations for reading out those arrays, scaling laws for array design parameters, and ways to improve sensitivity and dynamic range. Future arrays will have smaller pitch (15 micrometer), higher sensitivity (10 - 20 mK, F/1), wider dynamic range (> 100 degrees Celsius, 10,000:1), and better resolution (1280 X 1024). These improvements will come about with better photolithographic resolution, thinner structures, and reduced noise.

(Mn, Sb) doped-PZT (PMSZT) thin films have been grown on La_0.5Sr_0.5CoO_3-(delta )/Si and La_0.5Sr_0.5CoO_{3- (delta} )/LaAlO₃ substrates for pyroelectric detector arrays. The La_0.5Sr_0.5CoO_3(delta ) thin films, acting as a bottom electrode and as an atomic template for epitaxial growth of PMSZT, were deposited below 550 degrees Celsius thus allowing for integration with silicon technology. The epitaxial PMSZT thin films was designed to achieve high infrared responsivity using (100) oriented LSCO electrodes. The Ni-Cr/PMSZT/LSCO/Si capacitor-like structures show good ferroelectric properties with a large remnant polarization P_r of 40 (mu) C/cm², a spontaneous polarization P_s of 74 (mu) C/cm², and a coercive field E_c of 115 kV/cm under an electric field of 650 kV/cm. The PMSZT films have an electrical field breakdown strength in excess of 467 kV/cm, which is much higher than the coercive field. Voltage responsivity R_v of 4062 V/W at 2 Hz and current responsivity R_i of 281 (mu) A/W at 25 Hz was achieved under black body illumination. With CO₂ laser illumination at a wavelength equals 10.6 micrometer, an R_v of 4140 V/W at 2 Hz and an R_i of 441 (mu) A/W at 25 Hz was achieved.

An uncooled infrared detector having an optical readout is described. Infrared heat is sensed with bimaterial cantilevers which bend in response to temperature changes. One side of the bimaterial cantilever is an optical reflector. Visible light reflecting off the bent cantilever is detected with a CCD camera to provide the sensor output. We describe this device as a micro-optomechanical infrared receiver with optical readout -- MIRROR. Devices based upon this principle have successfully imaged infrared. Changes to the system are in progress to improve sensitivity.

Infrared imaging based on photoemission in metal- silicide/silicon Schottky barrier arrays is a mature technology that is currently employed in both military and commercial applications. Metal-silicide/silicon Schottky diodes can also be employed in uncooled bolometer arrays. The bolometer detection mechanism is thermionic emission in the Schottky barrier. Schottky bolometer array technology is expected to have both performance and production advantages, when compared with current uncooled sensor technology. In this paper, we compare the physical mechanisms involved in the two Schottky barrier based infrared sensors. We will also present a simplified model for the noise equivalent temperature of each technology.

A polarization-sensitive thermal imager has been assembled using a quantum-well infrared photodetector (QWIP) focal plane array (FPA) with peak responsivity in the long-wave infrared (LWIR) spectral band near 9 micrometer. Polarization-dependent responsivity is achieved by etching linear gratings onto each pixel during QWIP FPA fabrication, with adjacent pixels having orthogonal grating orientation. The direct integration of the gratings with the pixels eliminates all pixel registration errors encountered with previous infrared polarimetry instruments. We present here details of the FPA and thermal imaging system design and performance and show examples of polarization-enhanced imagery.

Full video format focal plane array (FPA) modules with up to 640 X 512 pixels have been developed for high resolution imaging applications in either mercury cadmium telluride (MCT) mid wave (MWIR) infrared (IR) or platinum silicide (PtSi) and quantum well infrared photodetector (QWIP) technology as low cost alternatives to MCT for high performance IR imaging in the MWIR or long wave spectral band (LWIR). For the QWIP's, a new photovoltaic technology was introduced for improved NETD performance and higher dynamic range. MCT units provide fast frame rates > 100 Hz together with state of the art thermal resolution NETD < 20 mK for short snapshot integration times of typically 2 ms. PtSi and QWIP modules are usually operated in a rolling frame integration mode with frame rates of 30 - 60 Hz and provide thermal resolutions of NETD < 80 mK for PtSi and NETD < 20 mK for QWIP, respectively. Due to the lower quantum efficiency compared to MCT, however, the integration time is typically chosen to be as long 10 - 20 ms. The heat load of the integrated detector cooler assemblies (IDCAs) could be reduced to an amount as low, that a 1 W split liner cooler provides sufficient cooling power to operate the modules -- including the QWIP with 60 K operation temperature -- at ambient temperatures up to 65 degrees Celsius. Miniaturized command/control electronics (CCE) available for all modules provide a standardized digital interface, with 14 bit analogue to digital conversion for state to the art correctability, access to highly dynamic scenes without any loss of information and simplified exchangeability of the units. New modular image processing hardware platforms and software for image visualization and nonuniformity correction including scene based self learning algorithms had to be developed to accomplish for the high data rates of up to 18 M pixels/s with 14-bit deep data, allowing to take into account nonlinear effects to access the full NETD by accurate reduction of residual fixed pattern noise. The main features of these modules are summarized together with measured performance data for long range detection systems with moderately fast to slow F-numbers like F/2.0 - F/3.5. An outlook shows most recent activities at AIM, heading for multicolor and faster frame rate detector modules based on MCT devices.

As 2nd and 3rd generation Focal Plane Arrays (FPA) become more complex, the readout integrated circuit (ROIC) has emerged as a major discriminator in system performance. The focus of development and advancement has traditionally involved the detector technology. Early ROICs were simple multiplexers that performed little if any signal processing on the detector diode signal. Advances in silicon fabrication processes for analog integrated circuits have opened a new era in IRFPAs where signal digital functions can be achieved on the focal plane. We present an overview of significant advances in the area of mixed mode ROIC designs that enable greater functionality and performance of the sensor chip assembly. Innovations, continuing progress in CMOS technology, and greater foundry access have allowed enhancements in practically every aspect of the ROIC, from sophisticated unit cells to lower noise and lower power signal paths to highly programmable digital support circuitry. Denser detector input circuits with active amplifiers (FEDI or CTIA) have been implemented in unit cells as small as 27 micrometer X 27 micrometer. In addition, multiple gain, temporal filtering, or spatial filtering capabilities have been incorporated into these small unit cells. Significant reductions in focal plane power have been fabricated and demonstrated enabling a factor of 2 increase in frame rates for very large staring FPAs and a factor of 4 increase in line rates for scanning FPAs. Other developments include, but are not limited to, alternative schemes for time-delayed integration (TDI) and breakthroughs for uncooled applications. As the chip designs increase in capability, greater systems on a chip are feasible, especially with more programmable features provided by the on-chip digital circuitry.

A pixel array readout integrated circuit (ROIC) containing per-pixel analog-to-digital conversion (ADC) and digital-to- analog conversion (DAC) for infrared detectors is presented with design and test result details. Fabricated in a standard 0.35 micron, 3.3 volt CMOS technology. the prototype consists of a linear array of 64 pixels, containing over 100 transistors per 30 by 30 micron pixel. The 8-bit per-pixel ADC is a Nyquist-rate single-slope design consisting of a three stage comparator and an 8 bit memory. This fully pixel- parallel ADC architecture operates in full-frame 'snapshot' mode and can reach over 1,000 frames per second. Each pixel also contains cascoded current source, globally biased to subtract an identical, fixed amount of current from each pixel in order to remove a common background signal by 'charge skimming.' It operates over more than 3 decades of current cancellation (approximately 10 pA to > 10 nA). As well, each pixel contains a 4 to 6+ bit current-mode DAC, intended to trim-out pixel-to-pixel variations in background current. It consists of 16 unit-cells of switched cascoded current sources per pixel, organized as two separately biased weights and controlled by a 16-bit per-pixel memory. The DAC operates over more than 4 decades of current cancellation (< 10 pA to approximately equals 100 nA) per least significant bit (LSB).

Cryogenic focal plane array (FPA) technology is widely used in both military and commercial applications. Three key factors that must be addressed to meet the future needs of FPA applications and improve their technology are (1) reduction of power dissipation in the FPA; (2) electrical and thermal isolation of the FPA assembly; and (3) simplicity of packaging. Conventional electronic data link packaging, which transmits data received by the FPA detector array to a warm external electronic processing module, does not address the above issues. Therefore, Physical Optics Corporation (POC) has investigated, for the first time, a vertical cavity surface emitting laser (VCSEL) based cryogenic optical link technology. This approach not only addressed the low-cost smart cryogenic FPA sensor, but also was the most viable and practical approach among other optical link technologies. This paper presents our major work in this project, which include (1) design, fabrication, and packaging of cryogenic VCSEL array and VCSEL array transmitter chip; (2) modification of a commercially available dewar to include the optical readout portion; (3) design, fabrication of optic fiber array link subsystem and assemble of entire system; and (4) preliminary test and demonstration.

A 512 X 512 monolithic Platinum Silicide Schottky Barrier detector array with random line selectable operation was proposed. The device modified from an interline CCD configuration by adapt a Random Line selected Charge Accumulation charge-coupled device on the vertical register and four tap readout on the horizontal CCD register to achieve a high frame rate and high fill factor operation. A 9-bit digital decoder is used to select which line of the sensor array that transfer their signals to the vertical CCD register. Accompanied with the vertical reset drain circuitry, either one line or up to 512 lines of the video signals can be selected and transferred to the vertical CCD register. All of the video signals on the unselected lines are then transferred to the vertical CCD channel simultaneously and finally dumped to the vertical reset drain. Since this unique readout structure, a frame rate of up to 240 frames/second can be achieved for 128 X 128 of the SBD array under 5 MHz of clock frequency. A high-speed sub-frame readout format can be easily fulfilled under this architecture. This architecture not only maintains the advantages of line-addressed charge- accumulation structure but also provides the capability to readout any portion of the array.

Noise property is the prime consideration in readout circuit design. The output noise caused by the photon noise, which dominates total noise in BLIP detectors, is limited by the integration time that an element looks at a specific point in the scene. Large integration time leads to a low noise performance. Time-delay integration (TDI) is used to effectively increase the integration time and reduce the photon noise. However, it increases the number of dead pixels and requires large integration capacitors and low noise output stage of the readout circuit. In this paper, to solve these problems, we propose a new concept of readout circuit, which performs background suppression, cell-to-cell background current non-uniformity compensation, and dead pixel correction using memory, ADC, DAC, and current copier cell. In simulation results, comparing with the conventional TDI readout circuit, the integration capacitor size can be reduced to 1/5 and trans-impedance gain can be increased by five times. Therefore, the new TDI readout circuit does not require large area and low noise output stage. And the error of skimming current is less than 2%, and the fixed pattern noise induced by cell-to-cell background current variation is reduced to less than 1%.

Silicon-based hybrid CMOS visible focal plane array (FPA) technology is emerging as a strong contender for scientific applications that require broad spectral response with low noise, highly integrated functionality and radiation hardness. CMOS-based FPAs offer many advantages in high speed, low-noise detection and signal processing. As a high performance alternative to advanced CCD imaging arrays, the hybrid design enables independent optimization of the silicon detector array and silicon readout electronics. Multiplexer commonality with the instrument's IR channels is another attractive feature for integrators of sensor sites such as for hyperspectral spectrometers. In this paper, the technical merits of Rockwell's CMOS-based hybrid visible FPAs are described including key detector performance aspects, interface electronics requirements, radiation hardness and concomitant implications for diverse imaging applications. At this time we have developed 640 X 480 and 1024 X 1024 hybrid imagers with approximately equals 100% optical fill factor, high broadband QE spanning ultraviolet (UV) through near infrared (NIR), wide dynamic range, and high pixel operability. Dark current of approximately equals 0.01e-/sec and read noise approximately equals 6e- have been measured on one prototype 1024 X 1024 FPA that uses Hawaii readout integrated circuit (ROIC). Initial radiation data indicate a total ionization dose (TID) tolerance greater than 35 Krad for our standard CMOS process.

As CMOS technology evolves, more and more functions can be integrated on the infrared focal plane array (FPA). This paper presents a study on the integration of analog to digital (A/D) conversion onto the FPA. A possible application for this is high-resolution (640 X 480 pixels) quantum well infrared photodetector (QWIP) FPAs operating at 70K. Operation at liquid nitrogen temperatures and below gives both advantages and disadvantages. CMOS transistors are performing better at these temperatures: increased transconductance and no leakage currents. On the other hand power dissipation needs to be limited to prevent a high load on the cooling system. The system aspects will ultimately determine the requirements for the A/D converters on the FPA. Some of the most stringent requirements are on: power dissipation, number of bits, die area and throughput. An FPA lends itself very well for the utilization of parallelism, so a trade-off can be made between sample rate per A/D converter and number of converters. With all these parameters in mind, an overview of state-of-the-art A/D converters at room temperature will be given. Trends will be identified and different architectures like delta/sigma and successive approximation will be evaluated. Also different implementation technologies such as switched current and switched capacitor will be reviewed with respect to their applicability in our application.

To meet the demands for high performance infrared imaging systems AIM had developed a family of CMT detector modules with linear focal plane arrays, integrated detector cooler assemblies (IDCA), and command and control electronics (CCE). Common features of these modules are focal plane multiplexers with time delay and integration (TDI) function, pixel deselect, programmable gain for each line, bidirectional scan capability, partitioning and global gain select. The family of IDCA's consists either of single chip focal plane arrays (FPA) directly linked to a read out integrated multiplexer (ROIC) by solder bump technique, or one clip infrared detectors connected to one or more ROIC's using a multichip module (MCM) technique, dewars with optimized thermal heat load, coolers with integrated control electronics, and command and control electronics (CCE). The general design of these modules is outlined. Test results are shown.

We report on the performance of a 640 X 512 pixel, indium gallium arsenide (In₅₃Ga₄₇As) focal plane array (FPA). The device has 25 micrometer pixels and represents the largest and finest pitched imager demonstrated in this material system. The device is sensitive to the 0.9 micrometer-to-1.7 micrometer short wave infrared band and features a room temperature detectivity, D*, greater than 5 X 10¹² cm- (root)Hz/W with greater than 98% of the pixels operable. The performance of the In₅₃Ga₄₇As photodiode array is such that at room temperature the focal plane array is read noise- limited. The presentation will include a description of the FPA fabrication and assembly as well as characterization of dark current versus temperature, spectral response, and resolution. The implications of these results to applications such as passive night vision imaging, active illumination, covert surveillance, target designation using eye safe lasers, and target acquisition and tracking will be discussed.

A high resolution thermal imaging system was developed based on a 384 X 288 mercury cadmium telluride (MCT) mid wave (MWIR) infrared (IR) detection module with a 2 X 2 microscan for improved geometrical resolution. Primary design goal was a long identification range of 3 km and high system performance for adverse weather conditions achieved by a system with small entrance pupil and minimized dimensions to fit for integration in existing apertures of armored vehicles, reconnaissance systems and stabilized platforms. A staring FPA module with its potential for long integration times together with a microscan for improved geometrical resolution provides the answers best fit to these requirements. A robust microscanner was developed to fit for military requirements and integrated with AIM's 384 X 288 MCT MWIR module and data processing. The modules allow for up to 2 ms integration time with 25 Hz frame rate and output a 768 X 576 high resolution CCIR standard image. The video image processing (VIP) provides the calculation power for scene based self learning nonuniformity correction (NUC) algorithms to save calibration sources. This NUC algorithm allows take into account non linear effects for unsurpassed performance in highly dynamic scenes. The detection module and VIP are designed to interface with STN's mature system electronics, used e.g. in hundreds of OPHELIOS thermal camera sets fielded. The system electronics provides a lot of different interface features like double serial control bus (CANBUS) interface, analog and digital outputs as well as different video outputs. The integrated graphic generation part allows to put advanced graphic overlays to the thermal image and also to external video signals via the video input feature. This electronics provides the power supply for the whole thermal imaging system as well as different processor controlled algorithms for field of view or zoom drives, focus drives, athermalization and temperature control of the FLIR. A new zoom lens F/2.0 allows to select field of views from 2 degree to 15 degrees horizontal. This covers a wide area of military and paramilitary applications. The whole camera is miniaturized to fit into existing gunner and commander sights for main battle tanks as well as for infantry fighting vehicles. The overall design is compatible in optical, electrical and mechanical direction with the fielded OPHELIOS cameras and so an easy upgrade for existing fire control, reconnaissance and platform systems. The overall design is made under consideration of mil standard environments and is able to withstand vehicle, airborne and shipborne stress. The presentation gives an overview of the different components of the new camera system. Theoretical range performance data are discussed together with measured NETD, MTF and MRTD data of the unit.

The emergence of new infrared sensor technologies and the availability of powerful, inexpensive computers have made many new imaging applications possible. Researchers working in the area of traditional image processing are showing an increased interest in working with infrared imagery. However, because of the inherent differences between infrared and visible phenomenology, a number of fundamental problems arise when trying to apply traditional processing methods to the infrared. Furthermore, the technologies required to image in the infrared are currently much less mature than comparable camera technologies used in visible imaging. Also infrared sensors need to capture six to eight additional bits of additional dynamic range over the eight normally used for visible imaging. Over the past half-century, visible cameras have become highly developed and can deliver images that meet engineering standards compatible with image displays. Similar image standards are often not possible in the infrared for a number of technical and phenomenological reasons. The purpose of this paper is to describe some of these differences and discuss a related topic known as image preprocessing. This is an area of processing that roughly lies between traditional image processing and image generation; because the camera images are less than ideal, additional processing is needed to perform necessary functions such as dynamic range management, non-uniformity correction, resolution enhancement, or color processing. A long-range goal for the implementation of these algorithms is to move them on-chip as analog retina-like or cortical-like circuits, thus achieving extraordinary reductions in power dissipation, size, and cost. Because this area of research is relatively new and still evolving, the discussion in this paper is limited to only a partial overview of the topic.

A new approach to sensing with intersubband absorption is introduced. In contrast to the conventional Quantum Well Infrared Photodetector (QWIP) which is a multi-quantum well device, our structure has 1 - 3 wells and uses resonant enhancement to achieve nearly complete absorption. In the QWIP the dark current is limited by the quantum well barrier in the range of 0.125 ev and thus cryogenic cooling is required in general to achieve BLIP operation. In the new structure, the dark current is limited by the band gap of GaAs/AlGaAs layers (>= 1.4 eV). This difference implies that BLIP operation may be possible near room temperature. The detecting quantum well is used to form the storage well of an active pixel or a CCD device and the intersubband absorption mechanism removes charge from the quantum well starting from the full well condition. The state of depletion of the well is then clocked to the output amplifier as in a conventional CCD using noise reduction techniques such as correlated double sampling. Therefore, the hybrid bump bonding of the Si ROIC is no longer required. In this paper, we describe the concept and its advantage vis-a-vis the existing approach and a preliminary analysis of the sensitivity of the detection.

Short-wave IR and mid-IR photovoltaic detector arrays consisting of In(Ga)As and InAsSb were realized. Maximum array size is 256 X 256 elements on a 25 micron pitch. The layers were grown on 3' semi-insulating GaAs substrates by MBE thereby avoiding the need for substrate removal by wafer thinning after hybridization. A reliable and uniform detector process using improved wet-etching has been developed. The citric-acid based etch has been optimized for minimum underetch such that high fill factor is achieved even with a mesa-type process. Typical R_oA products at room temperature are within a factor of 2 of the theoretical limit for bulk leakage currents. The hybridization with silicon readout circuits consisted of Si-postprocessing by electroless plating or lithographic definition of Ni/Au, indium bump electroplating on the III-V chip and flip-chip integration with individual indium bumps. The indium bump process resulted in 13 micron diameter solderbumps which allows pixel pitches below 20 micron.

We developed a very simple and useful method for observing the optical absorption due to intersubband transition in quantum wells. This new technique based on attenuated total reflection (ATR) is applied to evaluate epitaxial wafers containing multi quantum wells (MQW), used for quantum well infrared photodetectors (QWIP). For the intersubband transition, normal incidence produces only weak absorption because of the quantum selection rules on polarization. ATR has been used to emphasize the intersubband absorption, however, the GaAs substrate including MQW had to be made into a prism. We observed intersubband absorption without such preparation of GaAs substrate by just attaching a Ge prism to it. Attached prism in ATR is normally used for collecting the absorption near the sample surface by utilizing the total reflection at the prism-sample interface. However, we adjusted the angle of incidence to propagate the refracted light into the sample because absorption occurred inside epitaxial layers in the QWIP structure. Using this modified prism attached ATR, we were able to measure the intersubband absorption more clearly than in the case of conventional prism attached ATR. Due to avoid the complicated sample preparation, we could easily compare the absorption of MQW with QWIP device characteristics.

One of the simplest device realizations of the classic particle-in-the-box problem of basic quantum mechanics is the Quantum Well Infrared Photodetector (QWIP). In this paper we discuss the effect of focal plane array non-uniformity on the performance, optimization of the detector design, material growth and processing that has culminated in realization of large format long-wavelength QWIP cameras, holding forth great promise for many applications in 6 - 18 micron wavelength range in science, medicine, defense and industry. In addition, we present the recent developments in long-wavelength/very long-wavelength dualband QWIP imaging camera for various applications.

We report on the results of laboratory and field tests on a pixel-registered, 2-color MWIR/LWIR 256 X 256 QWIP FPA with simultaneous integrating capability. The FPA studied contained stacked QWIP structures with spectral peaks at 5.1 micrometer and 9.0 micrometer. Normally incident radiation was coupled into the devices using a diffraction grating designed to operate in both spectral bands. Each pixel is connected to the read-out integrated circuit by three bumps to permit the application of separate bias levels to each QWIP stack and allow simultaneous integration of the signal current in each band. We found the FPA to have high pixel operability, well balanced response, good imaging performance, high optical fill factor, and low spectral crosstalk. We present data on measurements of the noise-equivalent temperature difference of the FPA in both bands as functions of temperature and bias. The FPA data are compared to single-pixel data taken on devices from the same wafer. We also present data on the sensitivity of this FPA to polarized light. It is found that the LWIR portion of the device is very sensitive to the direction of polarization of the incident light. The MWIR part of the device is relatively insensitive to the polarization. In addition, imagery was taken with this FPA of military targets in the field. Image fusion techniques were applied to the resulting images.

Quantum Well Infrared Photodetectors (QWIPs) have been proposed for use in space based remote sensing applications. These space systems place stringent performance requirements on infrared detectors due to the low irradiance environments and the associated requirement for low temperature operation. This study demonstrates that, under these space conditions, the responsivity of a QWIP detector depends on frequency and that the shape of the frequency response depends on the operational conditions. The non-flat frequency response is empirically similar to dielectric relaxation effects observed in bulk extrinsic silicon and germanium photoconductors under similar operational conditions. Data from four QWIP detectors, obtained from four independent sources, demonstrate how the frequency response of QWIP detectors vary with temperature, photon irradiance, and bias voltage, and how the shape of the frequency response depends on the dynamic resistance of the detector. This QWIP frequency response results in a detector signal that is nonlinear with irradiance for some combinations of detector bias, photon irradiance, and operating temperature.

In this paper, we discuss the utilities of corrugated quantum well infrared photodetectors (C-QWIPs) in detector material characterization. By measuring the detector responsivity as a function of corrugation period, several important detector parameters, such as the absorption coefficient (alpha) of parallel propagating light and the energy resolved photoconductive gain g, can be directly deduced. For the QWIP material presented, (alpha) at the peak was found to be 0.21 micrometer^-1 under the usual operating condition. This value of (alpha) corresponds to an absorption length of 4.9 micrometer. Instead of being a constant, the value of g also varies significantly across the excitation spectrum, and the peak value is larger than the noise gain at large bias. Our results show that the present characterization technique is capable of providing accurate and detailed information on the intrinsic properties of QWIP materials under actual operating conditions. It is extremely useful in detector optimization. In addition, we also show that the performance of C-QWIPs can be further improved by etching an additional vertical trench at the center of each corrugation.

We summarize the current state of 2-color Quantum Well Infrared Photodetector (QWIP) Focal Plane Array (FPA) technology. Our 2-color FPA architecture features 3 bumps/pixel to permit two vertically-stacked QWIPs to be separately biased and the two photocurrents to be simultaneously integrated. Pixel-registered imagery is simultaneously obtained in two spectral bands. We have successfully applied this architecture to realize 2-color FPAs in three sperate formats: LW/LW, MW/LW, and MW/MW. Fabrication and performance details are presented.

This paper presents a brief overview of recent advances in photovoltaic (PV) HgCdTe infrared detector technology. Many of these recent advances have enabled a new generation of spaceborne multispectral instruments for remote sensing applications. The focus of this paper is on the back- illuminated HgCdTe PV arrays that have made this new generation of spaceborne instruments possible.

The world's first 2048 X 2048 HgCdTe infrared focal plane array (FPA) has been developed by Rockwell Science Center for infrared astronomy. The Hawaii-2 is the largest CMOS multiplexer designed to date, developed to interface with both infrared and visible detector arrays. The 18 micrometer pixel pitch was selected to accommodate both reasonable telescope optics and maximize yield in the fabrication of such a large readout. The fabrication uses world-class submicron photolithography to maximize yield of high quality devices. We will report on the characterization of FPAs using the Hawaii-2 multiplexer mated to SWIR detector arrays with a spectral response of 0.9 micrometer to 2.5 micrometer. These detector arrays have been processed on Liquid Phase Epitaxy (LPE) HgCdTe on sapphire substrates, also known as PACE-1. We also report on characterization of Silicon detectors in terms of their quantum efficiency, spectral response, and dark current.

Very Long Wavelength InfraRed (VLWIR; (lambda) _c approximately equals 15 to 17 micrometer at 78 K) photovoltaic detector operating in the 78 K range are needed for remote sensing applications. This temperature range permits the use of passive radiators in spacecraft to cool the detectors. VLWIR ((lambda) _c approximately equals 15 to 17 micrometer at 78 K) photovoltaic detectors in a range of sizes (8 micrometer diameter to 1000 micrometer diameter) have been fabricated and their performance measured as a function of temperature. Molecular Beam Epitaxy (MBE) was used to grow n-type VLWIR Hg_1-xCd_xTe on lattice matched CdZnTe. Arsenic was implanted and the wafer was annealed to provide the p-type regions. All the material was grown with wider bandgap cap layers and consequently the detector architecture is the Double Layer Planar Heterostructure (DLPH) architecture. I_d - V_d versus temperature curves for 8 and 1000 micrometer diameter, (lambda) _c equals 17 micrometer at 78 K detectors indicate that the 8 micrometer diameter detector is diffusion limited for temperatures greater than 63 K even at a -200 mV bias. There is no appreciable tunneling at T equals 50 K and at -200 mV applied bias. At T equals 40 K tunneling commences at a bias approximately equals -80 mV. Below T equals 30 K, the diode is tunneling limited. The 1000 micrometer diameter detector is diffusion limited at bias values less than -50 mV at 78 K. At zero bias, the detector impedance is comparable to the series/contact resistance. Interfacing with the low (comparable to the contact and series resistance) junction impedance detector is not feasible. Therefore a custom pre- amplifier was designed to interface with the large VLWIR detectors in reverse bias. The detector is dominated by tunneling currents at temperatures less than 78 K. The 1000 micrometer diameter, (lambda) _c approximately equals 17 micrometer at 78 K detectors have dark currents approximately equals 160 (mu) A at a -100 mV bias and at 78 K. Detector non-AR coated quantum efficiency > 60% was measured at -100 mV bias in these large detectors and the response was constant across the (lambda) equals 7 micrometer to 15 micrometer spectral band. With AR- coating the quantum efficiency will be > 70%. Response was measured and non-linearity < 0.15% was calculated for the 1000 micrometer detectors. The flux values were in the 10¹⁷ ph/cm²/sec range and were changed by varying the blackbody temperature. In addition, a linear response was measured while varying the spot size incident on the 1000 micrometer detectors. This excellent response uniformity measured as a function of spot size implies that, low frequency spatial response variations are absent, for the 1000 micrometer detectors.

Researchers have studied the use of double layer heterojunction HgCdTe devices for application in focal plane arrays (FPA's). Such devices are built with a wide bandgap semiconductor on top of a narrow bandgap semiconductor. With a highly doped p-type material at the surface, these devices enhance the ability to make contact with the anode side of the diode with interconnect metal. Optimizing FPA performance with heterojunction detectors has posed serious problems because many of the HgCdTe's material parameters vary as a function of composition, temperature and doping concentration. AET, with funding from the US Army Night Vision Labs, has developed a new system for design of focal plane arrays using heterojunction HgCdTe detectors. By using this new software modeling technique, a double layer heterojunction detector device has been designed with consideration for many of the material and environmental variations. This paper develops the models employed in the simulation program and will compare the simulation results with experimental data.

In our work, a three-dimensional IRFPA model has been constructed to conduct device simulations for drift-diffusion based Hg_1-xCd_xTe devices. The model uses the finite element method and numerical errors are automatically eliminated during computation. Computational results can thus easily achieve accuracy. The computer model was constructed by using C++ language. We have successfully represented simulation results in three-dimensional graphics. In this paper, a model for analyzing infrared-illuminated p-on-n photodiodes is presented. The computational results were verified analytically and experimentally. Furthermore, an IRFPA device model was built for calculating crosstalk by using uniformly collimated infrared radiation. Devices used for the model were linear FPAs. Ohmic contacts with zero bias were applied on electrodes. Other physics phenomena such as recombinations were also considered in the analyses. This model and simulation approach can provide an efficient way to reduce crosstalk in designing advanced MCT IRFPA devices.

The narrow gap semiconductor HgCdTe is commonly used for IR detection. Conventional HgCdTe IR detectors need significant cooling in order to reduce noise and leakage currents resulting from thermal generation and recombination processes. These cooling requirements considerably increase the cost, size, weight and complexity of infrared systems. The need for cooling to reduce noise and leakage currents resulting from Auger processes has long been thought to be fundamental and inevitable. However, recently, it has been suggested that by means of a steady-state non-equilibrium mode of operation, which holds the carrier densities below their equilibrium values, Auger generation and even radiative generation rates can be reduced. This is possible through the reduction of carrier concentrations because the Auger generation rate depends approximately on the square of the carrier concentration and radiative recombination rate depends linearly on it. This paper reports the modeling of a HgCdTe detector operated in a steady-state non-equilibrium mode at 230 approximately equals 295 K. The device architecture, NvP⁺, which is practical in MBE growth, is suitable for this application. Radiative and Auger lifetimes, zero surface recombination velocities, and zero background photon fluxes are assumed. The dependence of detectivity on minority carrier extraction efficiency is studied in this paper. At 230 and 250 K for N_D equals 1 X 10^{14approximately equals 15} cm^-3, the detectivity appears to become saturated at values in the order of 10¹⁰ cm Hz^1/2/W when the minority carrier extraction efficiency is greater than 3.

To meet the demands for high temperature-cycling reliability of HgCdTe detectors, bonded to a Silicon 'Read-Out-Integrated- Circuit,' AIM has developed a Multi-Chip-Module approach for the infrared Focal-Plane-Array. Bonding of detector array and Si-chips on a sapphire substrate minimizes thermal stress and strain in the FPA, leading to cycle-to-failure of >= 1000. For maximum cycle estimation under varying strain, a correlation was established empirically.

In this paper, two different surface treatments for bulk HgCdTe are compared in the point of surface recombination velocity and diode dynamic resistance-voltage characteristics. One surface treatment, named as standard treatment, is only Br-MeOH etching and the other, named as nitric acid (HNO₃) treatment, is composed of chemical oxidation with nitric acid after Br-MeOH etching and the removal of the oxide with ammonium hydroxide. After such surface treatments, gate- controlled diodes were fabricated and surface recombination velocity were measured to be 170 cm/s and 70 cm/s for the standard treatment and nitric acid treatment, respectively. And, the nitric acid treatment diode satisfied BLIP dynamic resistance characteristics regardless of the variation of the gate voltage. Moreover, from the measurement of capacitance- voltage characteristics, it was found that nitric acid treatment reduced the hysteresis width in the C-V curves of ZnS/HgCdTe MIS capacitor by one tenth compared to standard treatment. It is thought that the nitric acid treatment reduces surface-related defect and charge.

Staring InSb FPAs grown by MBE have been demonstrated. Low growth temperatures have been employed to provide p⁺-n- n⁺ photodiodes with a dark, 80 K R_OA equals 9 X 10⁵(Omega) cm². A degenerately doped substrate has been used to provide transparency in the 3.5 micrometer - 5.5 micrometer spectral region. Free carrier absorption necessitates some thinning of the substrate and an anti- reflection coated external quantum efficiency of 62% has been achieved with a final thickness of approximately equals 40 micrometer. 320 X 256 FPA's operating at 90 K and looking at a 295 K scene in f/2 have a noise equivalent temperature (NE(Delta) T) at half well of 10.4 mK. FPA operability exceeds 99.7%.

An infrared transfer radiometer has been recently developed at the Low-Background Infrared Calibration (LBIR) facility at the National Institute of Standards and Technology (NIST) for the Ballistic Missile Defense Organization (BMDO) program. The BMDO Transfer Radiometer (BXR) is designed to measure the irradiance of a collimated source of infrared light having an angular divergence of less than 1 mrad. It is capable of measuring irradiance levels as low as 10^-15 W/cm² over the spectral range from 2 micrometer to 30 micrometer. The radiometer uses an arsenic-doped silicon blocked impurity band (BIB) detector operated at temperatures below 12 K. Spectral resolution is provided by narrow bandpass interference filters and long-wavelength blocking filters. All the components of the radiometer, which include a mechanical shutter, an internal calibration source and detector, a long baffle section, a spatial filter, two filter wheels and a two- axis detector stage are cooled with an active flow of liquid helium to maintain temperatures below 20 K. A cryogenic vacuum chamber has been built to house the radiometer and to provide mechanical tilt alignment to the source. The radiometer is easily transported to a user site along with its support equipment.

Collimated infrared sources covering the 2 micrometer to 30 micrometer range of wavelengths are necessary to simulate infrared radiation from distant objects. This is important because on-orbit servo and tracking systems make extensive use of infrared radiation for remote sensing. Collimators are used to calibrate infrared detectors in terms of absolute power within a given spectral range. The National Institute of Standards and Technology (NIST) operates and maintains the Low Background Infrared Calibration (LBIR) facility, which uses a 2 K electrical substitution radiometer, the Absolute Cryogenic Radiometer (ACR), that is the primary national standard for broadband and infrared spectral measurements. At this facility, users can calibrate blackbody sources with at most 1% uncertainty. However, users must then rely on optical systems at their own facility to collimate the radiation from the blackbody. The effect of the optics on the output of the beam must then be calculated from models. For this reason, NIST is developing a portable transfer radiometer (BXR) that can be taken onsite to directly measure the spectral output, thus eliminating intermediate steps in the calibration chain. NIST is also developing a source having 1 cm diameter collimated beam, for a preliminary calibration of the BXR at the LBIR facility from 2 micrometer to 8 micrometer. The source must fit into a volume of about 0.03 m³ (1 cubic foot), have an angular divergence of less than 700 (mu) rad, a power output greater than 10 nW, and demonstrate 1% repeatability or better. The development and characterization of this source is the main topic of this paper.

The capability of the Low Background Infrared (LBIR) facility at the National Institute of Standards and Technology to spectrally calibrate infrared detectors was demonstrated with the spectral calibration of arsenic doped silicon blocked impurity band (BIB) detectors. The BIB detectors were calibrated over the 2 micrometer to 30 micrometer range, using light from a monochromator with a nominal 2% bandwidth. Photon fluxes used for the calibration ranged from 10¹³ photons/s/cm² to 10¹⁴ photons/s/cm². The large area detectors (10 mm²) calibrated in this paper were very linear up to 2.5 X 10¹⁴ photons/s/cm², where they showed a 1% drop in signal from linearity. The calibrations contained less than 6 1% standard component of random noise uncertainty, and there was about a 6 5% standard component of uncertainty arising from systemic effects that will be discussed in detail. The calibrations were performed in ultra- high vacuum in a 20 K background environment by making direct intercomparisons between the power measured by an absolute cryogenic radiometer and the response measured by a detector irradiated by the same beam. A detailed description of measurement methodology and system apparatus is given. Detector linearity and uniformity are also discussed. The LBIR facility can now provide calibrated BIB detectors as transfer standards as well as evaluate and calibrate customer's large area detectors and detector arrays provided the detectors stay within certain physical limitations.

We report on the fabrication and characterization of interdigitated finger, optical detectors/mixers. These devices are used in an FM/cw ladar system to detect and demodulate low intensity amplitude-modulated optical signals. Three different types of interdigitated finger structure were tested and compared in this study. We also present a theory to explain the asymmetry observed in the devices and discuss its implication in an FM/cw ladar application.

The performance of an InfraRed (IR) system is based on a high spatial resolution and on a high thermal resolution. An increase in spatial resolution means an increase in number of pixels, a decrease in detector pitch and an increase in the detector pixel MTF. Regarding thermal resolution increase, it will be achieved mainly by an increase in the maximum quantity of charges which can be stored in the silicon read-out circuits for 2D staring arrays. At present, only cooled detectors answer this need of high performance detectors, such as 2D arrays with TV format resolution and high NETD. In this paper these trends regarding high performance are discussed and recent IRFPA results at Sofradir are presented. Finally, a comparison with uncooled detectors, also processed at Sofradir, is presented, to outline the remaining gap between both types of detectors.

This paper reports on the design, performance and signal processing of a visible/near infrared (VIS-NIR) chromotomographic hyperspectral imaging sensor. The sensor consists of a telescope, a direct vision prism, and a framing video camera. The direct vision prism is a two-prism set, arranged such that one wavelength passes undeviated, while the other wavelengths are dispersed along a line. The prism is mounted on a bearing so that it can be rotated on the optical axis of the telescope. As the prism is rotated, the projected image is multiplexed on elements of the focal plane array. Computational methods are used to reconstruct the scene at each wavelength; an approach similar to the limited-angle tomography techniques used in medicine. The sensor covers the visible through near infrared spectrum of silicon photodiodes. The sensor weighs less than 6 pounds has under 300 in³ volume and requires 20 watts. It produces image cubes, with 64 spectral bands, at rates up to 10 Hz. By operating in relatively fast framing mode, the sensor allows characterization of transient events. We will describe the sensor configuration and method of operation. We also present examples of sensor spectral image data.

For infrared laser remote sensing, a direct detection receiver may be optimally designed around a high speed, low noise focal plane array (FPA). Short pulse, high repetition rate operation of the laser transmitter makes it beneficial to operate the detector with short integration times, lowering the limiting integrated background photon flux. With this photon signal (which constitutes a noise contribution) made small enough, improved low-noise readout integrated circuits (ROIC) can be used to realize a significantly improved imaging lidar receiver. A 10 by 10 pixel ROIC has recently been designed and fabricated. Demonstrated capabilities include > 100 kHz frame rate, 50 ns integration time, and less than 100 e^- of input-referred readout noise. These ROICs have been mated with long-wavelength HgCdTe infrared detector arrays, with cutoff wavelengths greater than 11 micrometer. Characteristics of the demonstrated ROIC design will be presented, along with testing of the focal plane arrays hybridized to them.

Performance of InSb focal plane array (FPA) detectors depends to a great extent on both the absolute temperature and the temperature fluctuations of the detector. The residual spatial noise, which can be achieved and maintained after a two-point non-uniformity correction (NUC), increases with the FPA temperature changes relative to that at which the NUC procedure was performed. A model is described, which allows prediction of the InSb FPA residual non-uniformity (RNU) as a function of the FPA temperature fluctuations for a given set of the FPA, cold shield and background radiation parameters. The calculated values are confirmed by experimental data. It is demonstrated that, as predicted, RNU degradation is primarily caused by signal offset changes corresponding to the InSb dark current variations, which are induced by the FPA temperature instability. The influence of the FPA temperature variation on NUC can be effectively compensated by a one-point offset correction. When this procedure is impractical, the dark current compensation method is proposed, which allows for a real-time, continuous compensation of the FPA temperature variations, resulting in a low residual non-uniformity.

We report results of experimentation with a new, high- resolution MWIR non-scanning, snapshot imaging spectrometer capable of simultaneously recording spatial and spectral data from a rapidly varying target scene. The instrument is based on computed tomography concepts and operates in a mid-wave infrared band of 3.0 to 5.0 micrometer. High speed spectral imaging was demonstrated by collecting spectro-spatial snapshots of an artificial target in the lab. Raw images were recorded using a 512 X 512 InSb focal plane array in snapshot mode.