The design and performance of a third generation 640(H) X 480(V) PtSi focal plane array is presented. The 3 to 5 micron MWIR focal plane array supports interlaced, progressive scan, and subframe readout under control of on-chip digital decoders. The new design utilizes 1.25 micrometers design rules to achieve a 50% fill-factor, a noise equivalent delta temperature of <0.07 C (f/1.5, 30 Hz, 300 K), and a saturation level >1.5 X 10⁶e. The power dissipation is less than 110 mW.

In 1993, FLIR SYStems, Inc. developed and established production of an MWIR staring sensor based upon affordable and reliable platinum silicide staring focal plane array technology. This sensor incorporates many new features in addition to the FPA technology that makes it well suited for a wide variety of applications. The FPA is a 240 X 320 area array of Schottky detectors read out in a snap shot progressive scan mode. The FPA architecture is a hybrid bump bonded detector to a CCD readout for high sensitivity and uniformity. The sensor uses a new miniature linear cooler with an integral dewar; this approach provides unsurpassed reliability and low power at affordable cost. The sensor electronics contain electronically programmable logic to read out the FPA and format the detector gain and offset nonuniformity correction circuitry. The video output can be controlled by manual gain/level or by automatic mode. The optics are designed to be low cost, light weight and easily interchangeable to serve a wide variety of applications. The sensor is designed to support all video formats and be user friendly. The compact size and low power is well suited for portable applications.

With the introduction of the Inframetrics InfraCAM, a new standard is established for small, lightweight, low power, hand- held, high sensitivity, high resolution thermal imaging systems. A unique design approach to video processing as well as the compact and efficient Inframetrics patented Sterling cycle microcooler allow the unit to require less than 5 watts of power during operation. The unit is smaller than most commercially available `palm-corders' with both the sensor and processing electronics housed in the same package. This paper reviews both the architecture and performance of our 256 X 256 platinum silicide array based imager.

Multi-wavelength imaging pyrometer (M-WIP) is presented for remote sensing of temperature profiles of targets with unknown emissivity. Fast algorithm and software package were developed for calibration and real-time M-WIP measurements. Experimental 7- filter line-sensing M-WIP system was implemented with 320 X 122-element PtSi IR-CCD imager. The M-WIP system was calibrated against a commercial blackbody source over temperature range from 450 degree(s)C to 900 degree(s)C. Signal processing included background subtraction and compensation for variation of dark current with detected signal and signal off-set sue to apparent BCCD trapping effect. Initial M-WIP measurement demonstrated temperature resolution (Delta) T of +/- 1 degree(s)C for blackbody target over temperature range from 600 degree(s)C to 900 degree(s)C.

Camera designs and radiometric performance evaluation results are presented for two PtSi IR imagers fabricated by the David Sarnoff Research Center. Measurements on the 640 X 480 IR-MOS imaging radiometer with 25 to 150 degree(s)C background temperatures indicated response non-linearity less than +/- 0.3% over 80% of the full signal range. By operation with variable integration time from 240 microsecond(s) ec to 33 msec a scanned image with a NE(Delta) T of less than 0.1 degree(s)C can be maintained over the full temperature range. The 320 X 122 IR-CCD imaging radiometer was designed for operation with integration times ranging from 0.12 to 133 msec to provide for 12 snapshot image settings. The signals from various integration times were effectively matched and scaled to increase the effective maximum measured signal from 1 X 10⁶ to 50 X 10⁶ electrons/pixel. Correction procedures were developed for achieving radiometric accuracy for achieving radiometric accuracy for operation of the imagers over multiple integration times and taking into account the effects of non- linear response of dark current and charge trapping in the readout BCCD registers. The camera stability was shown to be limited by the stability of the calibration source over a three- hour period.

This paper discusses the design of monolithic thin film thermal detectors, with particular reference to the parameters which influence performance such as thermal conductance. Methods for implementing thermal isolation are described, and critical fabrication technology is outlined. Optional materials technology for thin film resistance bolometers is reviewed, and consideration is given to bias and signal conditioning techniques.

The thermistor infrared detector or bolometer is the detector of choice in many classical remote sensing applications such as horizon sensing, noncontact thermometry, and industrial applications. In recent years, the authors have developed a thin film process where the thermistor material is deposited from a target directly onto the substrate. This is an advance over the labor intensive ceramic technology, where sintered flakes of the thermistor are bonded to the substrate. The thin film technique permits a variety of device constructions and configurations. Detectors fabricated on heat-sunk ceramic substrates can withstand high operating temperatures and large incident optical power, in both pulsed and CW laser measurements. For dc or low frequency measurements, the films can be deposited onto a thermally isolated membrane with applications in motion sensing, gas detection, and temperature measurement. Utilizing advances in micromachining a 2D array of thermally isolated microbolometer sensors, integrated onto a silicon wafer containing readout circuitry may be achieved. This paper describes the construction of the sputtered film thermistor detectors, their operation, and applications.

Several key process developments advance the state-of-the-art in hybrid, uncooled IR detector fabrication. Following a four year producibility roadmap, ion mill reticulation, wafer processing, and standard semiconductor processes are nearing implementation. The discussion describes several proof-of-principle experiments and their results and contrasts the current process approach against an existing, detailed process model. As the thrust continues towards low cost, high volume production, emphasis remains on simultaneous performance improvements. The analyses and preliminary data show that both moderate performance and low detector cost are quite obtainable with the strategies identified.

In many fields of IR technology it is necessary to take a parallel measurement of IR radiation within various defined spectral ranges. And in many cases there is also the demand for an identical measuring dot (moved objects), a high responsivity, a high signal-to-noise ratio, and a minimum cross talk. The paper describes the hybrid arrangement and essential properties of the developed pyroelectric multispectral detectors disposing of up to four channels. The detectors are discrete and handy building bricks fit into a T0 8 case which has an aperture of 2 mm in diameter and an internal spatial and spectral dispersion of the total radiation. It also presents results of these detectors applied especially in a multispectral pyrometer and the possibility of using them in gas analyzers. By using them in a multispectral pyrometer and applying special evaluation algorithms it was possible to reduce the influence of emission rate on the measuring accuracy decisively.

Spectral characteristics of specific detectivity D^*((lambda) _m) and noise equivalent temperature difference NETD^*((lambda) _m) were calculated for background noise limited thermal detectors with restricted spectral range ((lambda) _m - cut-off wavelength). In the actual spectral range ((lambda) _m<15 micrometers ) thermal detectors are able to ensure the same values of D^*((lambda) _m) and NETD^*((lambda) _m) as for quantum detectors. NETD^*((lambda) _m) and D^*((lambda) _m) values are shown to be connected by `kT²-rule': NETD^*((lambda) _m) - (root)2kT²D^*((lambda) _m). It was accentuated that this rule may be applied to quantum photodetectors but with some definitions. The approximations were offered permitting to calculate values of D^*((lambda) _m) and NETD^*((lambda) _m) in all spectral ranges with the accuracy of not worse than 2%. It was shown that as D^*((lambda) _m) of an ideal thermal detector with a cold filter increases the cooling temperature necessary for achievement of theoretical limited values of D^*((lambda) _m) and NETD^*((lambda) _m) also increases. Such a behavior is opposite to that known for quantum photodetectors.

In this paper there are investigated both theoretically and experimentally the possibility of high-sensitive detection and visualization of IR radiation in direct-gap semiconductor crystals due to IR quenching effect for probe visible radiation in the region of relative transparency of the crystal. This comparison allows us to conclude that the IR quenching effect observed by us in CdS crystal is probably the absorption of probe photon with energy deficit with respect to excitonic resonance, which is taken off due to exchange interaction of free exciton (Wannier-Mott type) in an intermediate state with spherical excitons that are localized at many electron impurity atoms.

Investigation of the performance of GaAs/AlGaAs quantum well IR photoconductors (QWIPs) as compared to HgCdTe photodiodes operated at temperatures below 77 K in the long wavelength IR region are presented. In comparative studies the current standard n⁺-p HgCdTe photodiodes as well as p⁺-n photodiodes are considered. Investigations of fundamental physical limitations of HgCdTe photodiodes indicate better performance of this type of detectors in comparison with QWIPs operated in the range 40 to 77 K. Only at temperature 40 K, QWIPs with a cutoff wavelength of about 8 micrometers indicate higher detectivity. The advantage of QWIPs increases in wider spectral region in lower temperatures - below 40 K. Usually, in the temperature range below 50 K the performance of n⁺-p HgCdTe photodiodes is determined by trap-assisted tunneling. As a result, the advantage of GaAs/AlGaAs QWIPs increases in wider spectral regions ((lambda) _c approximately equals 8 to 12 micrometers ) and at temperatures below 50 K. GaAs/AlGaAs QWIPs at 40 K are background limited in low-background conditions. This observation plus the maturity of GaAs/AlGaAs technology and its radiation hardness characteristics promise that QWIP technology can produce high- quality focal plane arrays for space applications.

Single pixel radiometric measurements and array performance of 256 X 256 miniband-transport multiple quantum well focal plane arrays (FPAs) are presented. The detectors typically have a peak wavelength of 9.1 micrometers with spectral bandwidth of 1.1 micrometers . Laboratory measurements using the NVL 3D noise analysis technique indicate that these first generation 256 X 256 FPAs have a temporal noise equivalent temperature difference of 22 mK and spatial NETD of approximately 40 mK (for an 8 degree(s)C calibration interval) when operated at 60 K in an f/2.36 camera system.

An ultra-low dark current normal incidence p-type strained-layer In_0.3Ga_0.7As/In_0.52Al_0.48As quantum well IR photodetector (PSL-QWIP) grown on semi-insulating (100) InP substrate by MBE technique has been demonstrated for the first time. This PSL-QWIP shows a background limited performance (BLIP) for TBLIP^* for this PSL-QWIP were found to be 7 X 10^-8A/cm² and 5.9 X 10¹⁰ cm - (root)Hz/W, respectively, at (lambda) _p equals 8.1 micrometers , V_b equals -2V, and T equals 77 K.

A novel three wavelength InGaAs focal plane array pixel element for detection at wavelengths from 1.0 to 2.5 micrometers , where each of three wavelength sensitive detectors are individually addressable is described. This device consists of successively smaller bandgap layers in In_xGa_1-xAs (x>=0.53), separated by compositionally graded layers of InAs_yP_1-y to decrease defects induced by lattice mismatch strain with the InP substrate and provide high-bandgap passivation. The various layers were selectively removed so that p-n junctions with different wavelength response could be formed. The three detectors have quantum efficiencies between 55 and 95% for front illumination and 15 and 60% for back illumination, and dark currents from 0.01 to 10 mA/cm².

A Monte Carlo program for simulating multiple quantum well and superlattice IR detectors is described. The model is quasi-2D and uses an effective mass approximation for the superlattice minibands. These approximations make it possible to run the program with the computing power of a workstation. The average velocity of the electrons is computed as function of the bias voltage.

With the increased use of large format detector arrays in IR imaging systems, designers are seeing the limitations that array response nonuniformities place on overall system design, performance, and cost. In many designs where high sensitivities are required and temporal noise has been reduced with the use of high quantum efficiency detectors, response nonuniformity across the array begins to limit the overall system performance. In addition, changes in the pixel response or offset that occur after nonuniformity corrections show up as new nonuniformities that also limit system performance. Because the array response is not adequately stable over time, system designers must include updatable nonuniformity corrections (recorrections) in the hardware or algorithms in the software to remove nonuniformities. We have measured the stability of InSb 2D focal plane arrays including measurements of the residual spatial nonuniformities over time and with temperature cycling. 2D focal plane arrays with stability that allows a one time correction to be used over several days have been measured. To help understand the source of this stability, the sensitivity of the spatial nonuniformity to small changes in the array temperature, detector leakage current, and detector bias was measured. These results are presented here.

This paper demonstrates that the two premises are interrelated, which is fortunate for us. This means once the scanner step sizes are accurately determined, an identical PSF wave form can be obtained. The paper describes a detector output shifting algorithm for estimating the actual step size to a fraction of a microradian, which also leads to production of an identical wave form measured by all the detectors. Realizing this is actually an image of a small target, which is a convolution of the target, the optics PSF and the detector response, the optics point spread function is determined through the deconvolution of the image with the detector PSF and the ideal geometric image.

The technology division of the Instituto de Astrofisica de Canarias (IAC) is developing a data acquisition system (DAS) for an IR camera to be used at the 1.5m Carlos Sanchez Telescope (TCS) in the Observatorio del Teide (Canary Islands, Spain). This camera will work between a wavelength of 1 and 5 microns and will employ an InSb focal plane array (FPA). The DAS and the user interface are based on a UNIX workstation with a modular transputer based controller. The IGA-256 (Cincinnati Electronics) has been evaluated as a candidate for the focal plane array. The main features related to the potential astronomical performance, such as well depth and dark current are reported. The testing procedures and present status of the camera are discussed.

The InSb metal oxide semiconductor field effect transistor (MOSFET) with three different channel lengths 5, 15, and 30 micrometers were fabricated successfully. The SiO₂ prepared by photoenhanced chemical vapor deposition was used both as the gate insulator and the source/drain passivation layer to reduce the source/drain pn junction surface leakage current. The common sourch current voltage characteristics show a breakdown voltage exceeding 2 V indicating an excellent pn junction reverse characteristics. The capacitance-voltage and the transferred current versus gate voltage characteristics are discussed in detail to explain the geometry effect on the device performance.

Deep levels in p⁺n Hg_0.73Cd_0.27Te/Hg_0.68Cd_0.32Te planar heterostructure diodes, grown by molecular beam epitaxy on CdZnTe wafers, were studied using deep level transient spectroscopy (DLTS). The DLTS spectra showed the presence of at least two hole traps with activation energies close to midgap. The activation energy obtained from the Arrhenius plots showed a strong dependence on the aplied bias, making it difficult to obtain a precise value.

Middle wavelength IR photovoltaic detectors were fabricated with HgCdTe epilayers grown by molecular beam epitaxy. The epilayers were grown on a slightly misoriented (111) B CdZnTe substrate and were twin free. Epitaxy was carried out with substrate rotation and the wavelength uniformity achieved on a 20 X 20 mm² wafer was 0.80%. A planar photovoltaic technology was used to produce the IR photodiodes and the junction formation was obtained by ion implantation. Linear array photodiodes and 128 X 128 2D-arrays interconnected with a CCD readout circuit were achieved. RoA product values of 4.7 10⁵ ohm cm² were measured on diodes with a 4.7 micron cutoff wavelength at 77 K. The good homogeneity of the layer is reflected by the low value, 4.4%, of the standard deviation on the short circuit current histogram of the 2D-array.

A radiometric characterization of a Si:As impurity band conduction hybrid array has been performed. The characterization included measurements of the responsivity and noise as functions of detector bias, photon irradiance, and integration time. Conversion gain, linearity, dynamic range, and NEI were calculated from measured data. The results are discussed in terms of optimal operating parameters for the hybrid array.

Scene generation systems which utilize laser radiation as a source must demonstrate the equivalence of focal plane array (FPA) response to this radiation as compared to the continuous radiation of realistic in-band scene. AEDCs Direct Write Scene Generation technique has been developed to provide mission scenarios for FPA testing. In the course of this development, radiative source issues have been examined in some detail to validate the technique. The purpose of this paper is to report the results of this investigation. In particular the response of a variety of FPAs to pulsed laser radiation is discussed.

We have used the liquid phase epitaxy (LPE) and vapor phase transport epitaxy (VPTE) growth methods to fabricate double heterojunction (DH) CaTe/HgCaTe/CaZnTe and heterojunction CaTe/HgCaTe thin film materials. The detector of zero tunnel leakage current has been developed successfully by using these materials. The advantages of the VPTE method are as follows: (1) There is increased stability of surface of detector; (2) bulk band-to-band tunnel effect in detector will not occur and trap- assisted tunnel effect will be reduced owing to the N⁺-p junction and IR absorption region of the detector separated by using this DH material; (3) Ro is enhanced at the same time, so that sensing circuit of FPA can attain much higher injection efficiency; (4) we overcome mother solution of the Te-rich back melt HgCaTe by LPE growth CaTe on HgCaTe/CaZnTe, and (5) this growth method hardly changes characteristics of HcCgTe.

It is well established that for a reduction of clutter and enhancement of desired features/contrast, one will require the use of multispectral focal plane arrays. These arrays will be highly beneficial for a variety of applications such as missile warning and guidance, precision strike, overhead surveillance, target detection, recognition, acquisition and tracking, thermal imaging, navigational aids and night vision, etc. Currently a common approach for the fabrication of multispectral arrays is via epitaxial deposition of layers of mercury cadmium telluride of different compositions on top of each other. Here we start with a bulk wafer of mercury cadmium telluride with a composition that is suitable for the detection of one of the required wavelengths. Next, the compositions suitable for the detection of other wavelengths are achieved via selective deposition and diffusion of CdTe or HgTe. Thus the individual detector elements corresponding to the different colors lie in the same plane and not on top of each other. These can now be interfaced to simple multiplexers, rather than those with switching capabilities. Additional advantages include improvements in cost and producibility.

Photovoltaic mercury cadmium telluride IR detector arrays having spectral response in the medium wavelength IR region out to 5 micrometers have been developed for thermal imaging applications. These detectors require cooling only to 180 K which can be provided by a thermoelectric cooler having six stages. The detectors have achieved noise-equivalent temperature difference performance less than 0.05 degree(s)C for an f/1.6 optical field of view. They are commercially available in a permanently sealed metal package with ceramic feedthroughs and a sapphire window.

An analytical expression for calculating the effective injection efficiency (A.C. injection efficiency) in direct injection readout hybrid focal plane array (FPA) is presented in this paper. The theory takes care of the shunt resistance due to surface leakage currents and the shift of operating point of the pixel with the variation in detector current. A study of effective injection efficiency as a function of RA (product of diode's dynamic impedance at the operating point and the area of diode) product in Hg_1-xCd_xTe hybrid focal plane array is reported and the role of buffered direct injection circuit in improving the effective injection efficiency has been investigated. This effective injection efficiency is then used to calculate the temporal noise. It is concluded that the BDI interface circuit plays an important role in achieving BLIP operation of FPA.

A comprehensive IR seeker imaging simulation (ISIS) model has been developed. ISIS is an open-ended object-oriented environment which supports a myriad of high fidelity modules for seeker simulation. Fractal models are used for focal plane array noise modeling, missing data interpolation, and high-fidelity simulation compression using iterated function system fractal codes. This paper will discuss specifically how various fractal algorithms have been incorporated to aid real-time hardware-in- the-loop simulation activities.

This paper describes an advanced scanning long wavelength IR focal plane array fabricated in HgCdTe. Comprised of an array of front-side illuminated, planar n-on-p homojunction diodes mated to a Si CMOS readout integration circuit, the array has a measured average spectral cutoff wavelength of 10.8 micrometers at 65 K with an average quantum efficiency of 87%. The FPA's peak D^* at temperatures of 65 and 78 K, in the absence of background radiation, exceeds 9.3 X 10¹¹ and 2.6 X 10¹¹ cm(root)Hz/W, respectively. No appreciable 1/f noise is evident even for reverse bias voltages as large as 50 mV. Both the spectral and the spatial responses of the diode are well-behaved; and the measured noise contribution due to background radiation agrees with theory.

As new IR focal plane array (IRFPA) technologies become available, improved methods for coping with array errors must be developed. Traditional methods of nonuniformity correction using simple calibration mode are not adequate to compensate for the inherent nonuniformity and 1/f noise in some arrays. In an effort to compensate for nonuniformity in a HgCdTe IRFPA, and to reduce the effects of 1/f noise over a time interval, a new dynamic neural network (NN) based algorithm was implemented. The algorithm compensates for nonuniformities, and corrects for 1/f noise. A gradient descent algorithm is used with nearest neighbor feedback for training, creating a dynamic model of the IRFPA's gains and offsets, then updating and correcting them continuously. Improvements to the NN include implementation on a IBM 486 computer system, and a close examination of simulated scenes to test the algorithms boundaries. Preliminary designs for a real-time hardware prototype have been developed as well. Simulations were implemented to test the algorithm's ability to correct under a variety of conditions. A wide range of background noise, 1/f noise, object intensities, and background intensities were used. Results indicate that this algorithm can correct efficiently down to the background noise. Our conclusions are that NN based adaptive algorithms will supplement the effectiveness of IRFPA's.

For the Cassini probe, focal plane number four of the Composite IR Spectrometer is equipped with a HgCdTe photovoltaic linear array. This 10 X 1 array detects in the 1100 to 1400 cm^-1 wavenumber range with 200 X 200 micrometers ² diodes at a pitch of 215 micrometers . Although background flux is near zero, peak detectivity, which is set by intrinsic diode noise, reaches 7.2 10¹¹ cmHz^.5W^-1 at -30 mV bias whereas the theoretical limit is 7.8 10¹¹ cmHz^.5W^-1. These performances are obtained by a simple ion implanted n/p planar technology on a liquid phase HgCdTe epilayer grown on lattice matched CdZnTe substrate.

Sofradir has developed many MCT IRFPAs using either the 8 to 12 micron spectral band or the 3 to 5 micron spectral band and using different silicon processor technologies to readout these detector arrays. This paper presents an overview of results coming from different types of arrays, (linear array, TDI array, staring array) using CCD or CMOS readout circuits with different types of signal processors.

The IR detector, which converts the thermal radiation into electrical signal, is key to performance of many optical/IR systems. To achieve adequate thermal sensitivity and to increase the detectivity of the detector, microlenses should be used to focus the incident radiation onto the sensitive area of the detector. But if we look at the integration compatibility of the double refractive surfaces microlenses in a complete device, we find that they have many drawbacks and the technical achievement of these microlenses seems more difficult in overall technological processes. Fresnel zone plate (FZP) lenses present none of these drawbacks, and their technology is quite compatible with the classical processes which are used for other optical microcomponents. In the present paper, a theoretical analysis and numerical calculations are given to verify the achievements of a proposed thin film FZP in decreasing the diffraction losses, focusing the incident radiation, and hence, enhancing the performance of the detector.

PtSi/Si/SiGe/Si Schottky diode IR detectors with extended and tunable cut-off wavelengths have been fabricated. Cut-off wavelengths depend on the SiGe composition and extend up to 10 micrometers for Si₈₀Ge₂₀. The cut-off wavelengths are also tunable by reverse bias. The tunability is due to the SiGe/Si offset serving as an additional potential barrier behind the Schottky barrier that can be varied in energy by a reverse bias. The sensitivity and range of the tunability is controlled by the SiGe thickness and composition. Cut-off wavelengths tunable from 4 micrometers at zero volts to 10 micrometers at 3 volts have been obtained. Quantum efficiency values are normal for operation at the long- wavelength end, but reduced over the rest of tunable range, because of the greater distance from the PtSi to the SiGe/Si offset.

IR systems require more and more performance (high sensitivity, resolution,...) to be adapted to specific system applications (such as surveillance and tracking systems...). To achieve such requirements, IRFPA manufacturers have to perform tradeoffs involving many parameters such as FPA operating temperature, number of TDI elements, cutoff wavelength, and diode area. IRFPA technologies and system limitations must be taken into account for these analyses. Thus, the authors present the general analysis of effects of these parameters on Sofradir IRFPA performance mainly utilizing 8 to 12 micron spectral band mercury cadmium telluride detector arrays. Impacts on electro-optical performance parameters and on thermal characteristics are presented. For example TDI linear scanning arrays are analyzed with emphasis on high IRFPA performance based on existing IRFPA technologies. Advantages of choices of different IRFPA configurations are presented.

For a low cost approach, thermoelectric coolers have often been thought of because of their solid state structure offering low fabrication cost and zero maintenance for cooling down detectors sensitive in the 3 to 5 micrometers spectral band. The performance of such detection assemblies is highly dependent on the achievable cold temperature due to the contribution of dark current in the 200 K operating temperature range. Drawbacks of such an approach reside in the input power required to maintain the detector temperature in adverse environmental temperature conditions. In the aim of evaluating and improving such detectors, Sofradir has established a program leading to the realization of a thermoelectrically cooled detection assembly for field operation, based on its standard 128 X 128 and 288 X 4 IRCCDs. Results are presented here.

Long wavelength IR detectors need to be cooled at cryogenic temperature to achieve high performances. This specific need makes it difficult to integrate the detector because of high cost of dewar and cooling device designed to fulfill severe vibration conditions. A new era for IR detection could begin with flat structures allowing intrinsic vibration resistance for detectors to be plugged on electronics board. Sofradir has carried out a study about feasibility of detector dewar assembly including a flat Joule-Thomson cooler with porous heat exchanger in cooperation with Air Liquide. The aim of this paper is to put forward the interest of such a product. The very good results achieved demonstrate a promising future for such flat structure detector assembly.

The ultimate signal-to-noise performance of the semiconductor photodetector is limited by the statistical fluctuations of the thermal generation and recombination rates in the semiconductor. Various ways to improve the performance of photodetectors operated at near room temperature including: optimized gap/doping profile, reduced volume of the device, and the use of non- equilibrium depletion of semiconductor have been proposed. Practical 2 to 12 micrometers devices with improved performance are reported.

A high sensitivity thermal video system has been jointly developed by Cincinnati Electronics Corporation and Nippon Avionics Company, Ltd. (Avio) of Japan, the TVS-8000. The system utilizes an Indium Antimonide staring focal plane array sensor, cooled by a Stirling cryocooler. The sensor/cooler assembly is connected to a separate video processor, display, and electronics assembly. The TVS-8000 product series offers the performance advantages of combining a high sensitivity focal plane array sensor with all of the advanced features and quality of an Avio Thermal Video System (TVS) processor. This unique system offers the user a wide variety of options for real time display, analysis, and recording of thermal scenes. The design, performance, and special features of the TVS-8000 Thermal Video System will be discussed.

Diversified Optical Products has designed an integrated lens, blackbody reference sources, and source control electronics for a mid-range IR Radiometric lens/staring focal plane array system. The purpose of the system is to be able to accurately correlate objects within the field of view of the system to two known, calibrated, and controllable blackbody sources also within the field of view. The two internal blackbody sources are thermoelectric cooler based, and their output is optically relayed to an internal image plane of the lens. The optical system also incorporates a neutral density filter wheel which attenuates the scene radiance so that both the blackbodies and the scene radiance can be brought into the dynamic range of the focal plane array. The goal of the design effort was to manufacture a highly accurate, field portable radiometric instrument. The specific design areas which where focused upon were: matching the optical design of the lens system with the camera design; controlling the radiometric properties of the optics, optical design requirements for the projection of the blackbody sources within the system field of view; and calibration requirements and methods for the total radiometric system.

A hybrid 256 X 256 PtSi focal plane array (FPA) with single output CMOS multiplexer readout in 24 micrometers pitch has successfully been developed. The device combines an excellent analog performance like noise equivalent temperature difference NETD < 100 mK (F/1.4, 20 ms, 300 K), dynamics > 12 bit, high output voltage > 2.5 V and reduced blooming with a very flexible digital layout, that enables customer specific applications like interlaced or progressive readout, variable integration time between 20 to 40 ms frametime and ultrashort times < 10 microsecond(s) , or subframe capability by just switching between clock patterns. The development is done in cooperation with DAIMLER BENZ Research and Technology, Munich, and TELEFUNKEN microelectronic TEMIC EZIS, Ulm. Reported are the digital features of the multiplexer results of the hybridization process and test results of our most recent FPA's. An outlook is given on the just started development of a 486 X 640 FPA with identical features, and on the development of integrated dewar cooler assemblies for this detector generation.