The advantages and uses of infrared (IR) imaging continue to grow. As such, a new generation of requirements for IR focal plane arrays (IRFPAs) has emerged that affects the development of both the detectors and readout integrated circuits (ROICs). Because these applications have varying requirements, a universal set of FPAs cannot be made to satisfy all needs, and custom designs are needed. However, the desired capabilities follow a common theme. The industry is receiving more demands in the areas of larger formats, increased sensitivity, smaller pixels, and higher functionality. These must be met in addition to achieving production quantities at a low cost. This paper focuses on two main facets of the complex FPA, the detector and readout integrated circuit (ROIC), to address the evolving requirements. For detectors, we explore both the cooled and uncooled technologies, where HgCdTe grown by molecular beam epitaxy (MBE) and vanadium oxide (VOx) microbolometers are discussed, respectively, for the two areas. Development in ROICs expands in terms of smart features, on-chip signal processing, and on-chip analog-to-digital conversion.

Indium Gallium Arsenide (InGaAs) photodiode arrays have numerous commercial, industrial, and military applications. During the past 10 years, great strides have been made in the development of these devices starting with simple 256-element linear photodiode arrays and progressing to the large 640 x 512 element area arrays now readily available. Linear arrays are offered with 512 elements on a 25 micron pitch with no defective pixels, and are used in spectroscopic monitors for wavelength division multiplexing (WDM) systems as well as in machine vision applications. A 320 x 240 solid-state array operates at room temperature, which allows development of a camera which is smaller than 25 cm³ in volume, weighs less than 100 g and uses less than 750 mW of power. Two dimensional focal plane arrays and cameras have been manufactured with detectivity, D*, greater than 10¹⁴ cm-(root)Hz/W at room temperature and have demonstrated the ability to image at night. Cameras are also critical tools for the assembly and performance monitoring of optical switches and add-drop multiplexers in the telecommunications industry. These same cameras are used for the inspection of silicon wafers and fine art, laser beam profiling, and metals manufacturing. By varying the Indium content, InGaAs photodiode arrays can be tailored to cover the entire short-wave infrared spectrum from 1.0 micron to 2.5 microns. InGaAs focal plane arrays and cameras sensitive to 2.0 micron wavelength light are now available in 320 x 240 formats.

Indigo Systems Corporation has recently developed a line of high performance NIR devices and cameras based upon InGaAs detector arrays. The InGaAs detector arrays are fabricated at Indigo Systems Detector Operations facility and include 640x512 and 320x256 staring focal plane arrays which are utilized in three camera configurations ranging from the miniature alpha camera to the Merlin and Phoenix high performance cameras. The InGaAs detector arrays are very high performance devices with operability routinely exceeding 99.9%. In addition to the staring arrays for imaging applications, two varieties of linear arrays are also being produced at Indigo Systems Detector Operations including a 512 element and 1024 element devices. The linear arrays are intended for use in telecommunications for DWDM applications and are provided in industry standard packages for insertion into DWDM systems. All linear arrays require 100% perfect operability and this is routinely achieved.

The ultimate performance of any remote sensor is ideally governed by the hardware signal-to-noise capability and allowed signal-averaging time. In real-world scenarios, this may not be realizable and the limiting factors may suggest the need for more advanced capabilities. Moving from passive to active remote sensors offers the advantage of control over the illumination source, the laser. Added capabilities may include polarization discrimination, instantaneous imaging, range resolution, simultaneous multi-spectral measurement, or coherent detection. However, most advanced detection technology has been engineered heavily towards the straightforward passive sensor requirements, measuring an integrated photon flux. The need for focal plane array technology designed specifically for laser sensing has been recognized for some time, but advances have only recently made the engineering possible. This paper will present a few concepts for laser sensing receiver architectures, the driving specifications behind those concepts, and test/modeling results of such designs.

DRS (formerly Boeing) has completed the development and demonstration of a 25-micron pixel size 640x480 VO_x microbolometer uncooled IR focal plane product, the U6000. The U6000 incorporates several advanced features to enhance its performance and functional capabilities. A parallel six- bit Smart-Sensor data bus provides external command and data interface capability between the sensor and the focal plane. This includes on chip 6-bit pixel offset correction, detector bias selection and regulation, programmable signal gain, interlaced and non-interlaced output video format selection, signal integration time selection and input referred global offset selection capabilities. The U6000 also includes a high resolution on-chip temperature measurement that is incorporated into the single channel output video during horizontal blanking. This paper describes the U6000's functional capabilities, and provides U6000 functional validation and performance data.

The emergence of uncooled infrared detectors has opened new opportunities for IR imaging both for military and civil applications. Infrared imaging sensors that operate without cryogenic cooling have the potential to provide the military or civilian users with infrared vision capabilities packaged in a camera of extremely small size, weight and power. Uncooled infrared sensor technology has advanced rapidly in the past few years. Higher performance sensors, electronics integration at the sensor, and new concepts for signal processing are generating advanced infrared focal plane arrays. This would significantly reduce the cost and accelerate the implementation of sensors for applications such as surveillance or predictive maintenance. We present the uncooled infrared detector operation principle and the development at CEA/LETI from the 256 x 64 with a pitch of 50 micrometers to the 320 x 240 with a pitch of 35 micrometers . LETI has been involved in Amorphous Silicon uncooled microbolometer development since 1992. This silicon IR detection is now well mastered and matured so that industrial transfer of LETI technology was performed in 2000 towards Sofradir. Industrial production of 320 x 240 microbolometer array with 45micrometers pitch is then started., we present the readout circuit architectures designs and its evolution from the 256 x 64 array to the different version of 320 x 240 arrays. Electro-optical results obtained from these IRCMOS are presented. NEDT close to 30 mK is now obtained with our standard microbolometer amorphous silicon technology.

INO has been active in microbolometer and FPA technology development since the early 1990s. Microbolometer detectors based on VO₂ films with TCR above 3% are typically fabricated. VOx films with TCR above 2% have been developed for applications where FPA temperature is not stabilized. INO is continuing its development of high fill factor pixels with sizes down to 25 micrometers and new macro- and micro-packaging technology. All fabrication is done on six inch wafers in INOs newly expanded clean room facility. INO currently offers as standard products 256x1 and 160x120 pixel FPAs with 52 micrometers pixel pitch. Both arrays have simple, robust, and versatile CMOS readout integrated circuits (ROICs) that may be accessed in self-scanning or random access mode, and reference detectors for on-chip coarse offset and temperature drift compensation. Single frame NETDs (f/1, 300 K, 8-12 micrometers ) are on the order of 150 - 250 mK and may be reduced by frame averaging. Prototyping boards have been developed for both arrays, and the 160x120 FPA has been integrated in a number of thermal cameras and instruments. In collaboration with its clients, INO has developed several FPAs for specific space and terrestrial applications. Custom ROICs fabricated in several different CMOS processes from multiple foundries have been used. A 512x3 pixel microbolometer FPA with 39 micrometers pitch is being developed for the European Space Agency. The array is designed for multi-spectral pushbroom imaging applications and features a novel ROIC with very low 1/f noise, pixel by pixel offset and drift compensation, variable integration time, and digital output. Its single frame NETD (f/1, 300 K, 8-12 micrometers ) is nominally 80 mK.

We have been focused on growth of multi-layered LaNiO₃/Pb(Zr,Ti)O₃/LaNiO₃ on bare Si and polymer-coated Si substrates for infrared detector applications. A unique ion-beam assisted pulsed laser deposition (IBAD-PLD) has been employed to address two critical issues related to these thin film ferroelectric (TFFE) devices: to reduce the thermal budget and to enhance the texture of the devices. IBAD has been a well-known technique for deposition of thin films due to the ability to control morphology, adhesion, texture, and stoichiometry of the film by providing extra kinetic energies to, and to generate desired textures in films by preferential sputtering of the growing surface of the film. We have studied the role of several processing parameters of IBAD-PLD process, including ion-beam energy, current density, IBAD time, and substrate temperature in order to identify the best processing window for LaNiO₃/Pb(Zr,Ti)O₃/LaNiO₃.

BAE SYSTEMS has designed and developed MicroIR microbolometer focal plane arrays (FPAs) in three formats (160x120, 320x240, and 640x480) and with two different pixel sizes (46micrometers and 28micrometers ). In addition to successfully demonstrating these FPA technologies, BAE SYSTEMS has produced and delivered thousands of 320x240 (46micrometers pixel) imaging modules and camera cores for military, thermography, firefighting, security and numerous other applications throughout the world. Recently, BAE SYSTEMS has started production deliveries of 160x120 (46micrometers ) systems, demonstrated 320x240 and 640x480 second-generation (28micrometers ) imaging, and demonstrated second-generation thermoelectric cooler-less operation. This paper discusses these recent accomplishments and, when possible, provides quantitative NETD and performance data for our newly developed FPAs and systems. Video will be shown to demonstrate sensor performance capabilities.

Now that commercial infrared is a well-established business with several serious competitors, the pressures for a competitive edge have increased dramatically. Hybrid barium strontium titanate (BST) ferroelectric detectors still provide the basis for the majority of systems being produced today, and tens of thousands of systems have been fielded. The system simplicity of these AC-coupled systems is not matchable by any other current technology, but the complexity of the detector fabrication process limits its potential for further substantial cost and performance improvements. DC-coupled VOx bolometers, currently the most popular technology among manufacturers, offer better sensitivity at somewhat greater cost. Although this technology has been heralded as the technology of the future, it is encumbered by a more complicated system architecture and by spatial noise, which limits the ability to take advantage of its greater sensitivity. Thin-film ferroelectric (TFFE) detectors promise to remove the cost and performance barriers that lie ahead of BST technology, while maintaining the low system cost and low spatial noise characteristic of AC-coupled systems. Until recently the promise has been elusive, but now real-world performance of the best of TFFE systems is competitive with the best of any other technology.

In this paper we present a status of the activity of the LETI infrared laboratory in the field of HgCdTe infrared multispectral detectors. The multilayer doped structures needed to achieve two color pixels are grown by molecular beam epitaxy (MBE) (211)HgCdTe on lattice matched CdZnTe substrates. The device structure is n+ppn and is spatially coherent. The long wavelength layer is a planar like n+/p diode and is made by ion implantation while the shorter wavelength p-n diode is made in-situ during the MBE growth using Indium impurity doping. The last junction is isolated by mesa etch. The detectors are interconnected by indium bumps to a CMOS readout circuit. One or two indium bumps per pixel are used to address sequentially or simultaneously the two wavelengths, the detector pitch being 50micrometers or 60micrometers respectively. Elementary detectors exhibit performances in each band which are very close to those obtained in single color detectors with our standard technology. The Si-CMOS read-out circuits are specially designed to optimize the best performance of the IRCMOS focal plane arrays (FPA) in both wavelengths. The electro-optical performances of a two color IRCMOS FPA with a complexity of 128x128 pixels (pitch of 50micrometers ) operating sequentially within the (3-5micrometers ) middle wavelength infrared range (MWIR) at 77K will be presented.

This paper reports the development of a low-cost, small pixel uncooled infrared detector using a standard CMOS process. The detector is based on a suspended and thermally isolated p⁺-active/n-well diode whose forward voltage changes due to an increase in the pixel temperature with absorbed infrared radiation. The detector is obtained with simple post-CMOS etching steps on dies fabricated using a standard n-well CMOS process. The post-CMOS process steps are achieved without needing any deposition or lithography, therefore, the cost of the detector is almost equal to the cost of the fabricated CMOS chip. Before suspending the pixel using electrochemical etch-stop technique in TMAH, the required etch openings to reach the silicon substrate are created with a simple dry etch process while CMOS metal layers are used as protection mask. Since the etch mask is implemented with available CMOS layers, the etch openings can be reduced significantly, allowing to implement small pixel sizes with reasonable fill factor. This approach is used to implement a 40micrometers x40micrometers diode pixel with a fill factor of 44%, suitable for large format FPAs. The p++)-active/n-well diode has a low 1/f noise, due to its single crystal nature and low bias requirement. Optimum pixel performance is achieved when the pixel is biased at 20(mu) A, where self-heating effect is less than 0.5K. Measurements and calculations show that this new detector has a thermal conductance (G_th) of 1.4E^-7W/K and provides a responsivity (R) of 5800V/W and a detectivity (D^*) value of 1.9x⁹cm(root)Hz/W when scanned at 30fps with an electrical bandwidth of 4kHz. If this detector is used to implement a 64x64 or 128x128 FPA with sufficient number of parallel readout channels, these FPAs will provide an NETD value of 195mK considering only the detector noise. When the readout noise is included, these FPAs are expected to provide NETD value below 300mK. Such FPAs are very suitable for ultra low-cost infrared imaging applications.

The performance characteristics of polycrystalline SiGe microbolometer arrays are the subject of both design and technological optimizations performed in this work to move the arrays towards the production. An NETD of 90 mK at a time constant of 11 ms is already achievable for the best non-optimized 60 micrometers pixel, 0.26 micrometers thick bolometer design in a linear 128 pixel array according to the results of LWIR characterization. The performance of linear 32, 64 and 128 element arrays of 50-, 60- and 75-micrometers pixel bolometers made with 0.26...0.13 micrometers thin poly-SiGe on several wafer runs was the starting point for the computer simulation of detector features and evolution of its characteristics under reading bias pulses. The material properties and parameters of read-outs are taken into account in the optimization of the design parameters of arrays as well. The typical bolometer characteristics achieved on the latest wafer run if processed with the PC-program accounting for the read-out and heating effects, result in an average NETD of 70 mK at a time constant of 17 ms for 50 micrometers pixels in a 320x240 array. Despite less TCR-to-1/f noise ratio as compared with VO_x arrays, the several advantages make poly-SiGe a very attractive candidate for an uncooled array, i.e. full compatibility with CMOS technology, better characteristics/price ratio, resistance nonuniformity s/mean <0.2%, and a possibility to release extra-thin structures.

A 9 micrometers cutoff 640x512 pixel hand-held quantum well infrared photodetector (QWIP) camera has been demonstrated with excellent imagery. Based on the single pixel test data, a noise equivalent differential temperature (NETD) of 8 mI is expected at 65K operating temperature with f/2 optics at 300K background. This focal plane array has shown background limited performance (BPLI) at 72K operating temperature with the same optics and background conditions. In this paper, we discuss the development of this very sensitive long wavelength infrared (LWIR) camera based on a GaAs/AlGaAs QWIP focal plane array and its performance in quantum efficiency, NETD, uniformity, and operability.

Quantum Well Infrared Photodetectors (QWIPs) afford greater flexibility than the usual extrinsically doped semiconductor IR detectors. The wavelength of the peak response and cutoff can be continuously tailored over any wavelength range between 6-20micrometers . The spectral band width of these detectors can be tuned from narrow ((Delta) (lambda) /(lambda) ~10%) to wide ((Delta) (lambda) /(lambda) ~40%) allowing various applications. Also, QWIP device parameters can be optimized to achieve extremely high performances at high background at lower operating temperatures (~30 K). However, for low- background irradiance levels at low operating temperatures, high resistivity of the active region could lead to non- linear responsivity behavior. A new structure with a photosensitive superlattice and a thick blocking barrier has been tested, and is expected to avoid anomalous behavior under low backgrounds at low temperatures.

Low NETD's, coupled with other improvements in camera design and manufacturing, helps to further enable a new class of very demanding imaging applications in medicine and medical research.. The evolution of QWIP FPA over the past five years, with their low NETD, detector uniformity, and high pixel yield, along with improvements in camera control and processing electronics, represents key technical innovations responsible for the reemergence of medical infrared imaging through the development of a new infrared medical imaging technique called Dynamic Infrared Imaging or DIRI. The QWIP's high thermal and spatial resolution coupled with very fast data acquisition capabilities fill the essential requirements of DIRI. Other features required by DIRI applications are the need for stable operation with drifts in the image below a few mK, which allow longer data collecting time. Longer data collection time provides the camera the capability to detect the functional behavior of the autonomic nervous system which operates on a time scale of 0.1 to 0.2Hz.

We investigated the behavior of the dark current (Id) in quantum well infrared photodetectors (QWIPs) in which the barrier layers were selectively doped instead of the well layers. Because the selective doping bends the conduction band (CB) edge in the portion of the barrier near the interface, the mechanism by which carriers in the wells can be emitted over the barriers, i.e. thermal emission and tunneling through this portion of the barrier, could be emphasized. We first confirmed that selectively doping the barrier layers clearly affects the Id-V characteristics. Then, by evaluating the activation energy obtained from the temperature dependence of Id, we found that the Poole-Frenkel emission (PFE) mechanism and the thermal-assisted tunneling (TAT)-like mechanism are dominant in the lower bias and higher bias regions, respectively.

For 3rd Gen applications, AIM is presently launching its new high speed mercury cadmium telluride (MCT) modules. Two configurations are developed: a 256x256 40μm pitch device in a broadband 3.4-5μm design and a 192x192 56μm pitch device in a dual color mid wave (MWIR) design. The dual color device provides spectral selection with temporal and spatial coincidence for both colors using a new AIM proprietary technology. The spectral bands presently selected are 3.4-4 and 4.2-5μm. In any case, a very high frame rate of 800Hz full frame rate for the broadband and 870Hz for the dual color design are realized. The frame rate is equivalent to a data rate of 80MHz pixel rate. Arbitrary subframes in a step size of 8 pixels can be read out at higher frame rates just limited by the 80MHz pixel rate. The focal plane arrays are integrated in AIM standard dewars with either 1Watt, 1.5Watt or 2Watt split linear coolers to accomplish for various cooldown requirements.

We report on the development and testing of a new dual-band infrared (IR) focal plane array (FPA) specifically designed to detect buried land mines. The detector response spectra were tailored to take advantage of the sharp spectral features associated with disturbed soils. The goal was to have a blue channel with peak response near 9.2 micrometers and a red channel with maximum response at 10.5 micrometers . The quantum well infrared photodetector (QWIP) is particularly suited for this application because of the flexibility available in designing the peak wavelength of the detector and the relatively narrow width of the response spectrum. FPAs were produced and tested under the U. S. Army Research Laboratory's Advanced Sensors Collaborative Research Alliance in co-operation with the Night Vision and Electronic Sensors Directorate. We report on laboratory measurements of the response spectra, the dark current as a function of operating temperature, and the conversion efficiency in both the blue and red channels. Imagery was taken in the field of buried anti-tank mines. The images were analyzed by combining the data from the two channels into single fused images.

Multicolor focal plane arrays are of interest for a variety of applications. We report on a method to create a multicolor detector array of high-performance arsenic-doped silicon Blocked-Impurity-Band (BIB) detectors by using diffractive microlenses. Advantage is taken of the strong chromatic aberration characteristic of diffractive lenses to direct light within a pixel to either a central detector or to second detector concentrically disposed around the first. A theoretical calculation of the efficacy of this approach for spectral separation is presented. Fabrication of diffractive microlenses on the backside (illuminated-side) of thinned specially-designed BIB detector arrays is described. Finally, early initial results and further development plans are discussed.

In this paper we report a high performance 2-stack, 3-color quantum well infrared photodetector (QWIP) composed of InGaAs/AlGaAs QWIP and an InGaAs/AlGaAs/InGaAs triple- coupled (TC-) QWIP grown on the GaAs substrate for the mid- and long-wavelength (MW/LW) infrared (IR) detection. The basic device structure consists of a MWIR QWIP stack with 3 periods of 43 Angstroms In_0.3Ga_0.7As quantum well and an undoped 300 Angstroms Al_0.3Ga_0.7As barrier and a LWIR TC-QWIP stack with 5 periods of 65 Angstroms In_0.18Ga_0.82As quantum well (QW) and two undoped 60 Angstroms In_0.05Ga_0.95As Qws separated by 20 Angstroms Al_0.08Ga_0.92As barriers. The TC-QWIP stack has two response peaks, which are voltage-tunable from 9.2micrometers to 10 micrometers and 12micrometers to 12.2micrometers by the applied bias, respectively. For the LWIR TC-QWIP, a maximum responsivity of 1.96A/W at 12micrometers was obtained at T=40K, and a maximum detectivity of (Formula available in paper) was obtained at V_b=-1.7V, λ_p=12micrometers , and T=20K. As for the MWIR QWIP stack excellent responsivity at the peak wavelength of λ_p=5.1micrometers was obtained up to 120 K.

The past 2 to 3 years has been a period of explosive growth in technology development for imaging sensors at Rockwell Scientific Co. (RSC). The state of the art has been advanced significantly, resulting in a number of unique advanced imaging sensor products. A few key examples are: 2048 x 2048 sensor chip assemblies (SCA) for ground and space-based applications, 4096 x 4096 mosaic close-butted mosaic FPA assemblies, a very high performance 10 x 1024 hybridized linear SCA for optical network monitoring and other applications, the revolutionary CMOS ProCam-HD imaging system-on-a-chip for high definition television (HDTV), and RSC's near-infrared emission microscope camera for VLSI defect detection/analysis. This paper provides selected updates of these products and thereby provides an overview of the ongoing highly fertile period of technology and product development at Rockwell Scientific. A view into future directions for advanced imaging sensors is also provided.

This paper investigates 1/f noise performance of Hg_1-xCd_xTe photovoltaic detectors when detector current is varied by changing detector area, bias, temperature and incident flux. Holding detector bias and temperature constant, measured 1/f noise current is proportional to the detector current. However for all detector areas measured, non-uniformity is observed in the noise current due to the varied quality of the detectors. Even for the λ_c=16μm , 4-μm-radius, diffusion-limited detectors at 78K held at reverse bias, the average and standard deviation in dark current is I_d=9.76+/- 1.59x10^-8A while the average and standard deviation in noise current at 1 Hz in a 1 Hz bandwidth is i_n=1.01+/- 0.63x10^-12A. For all detector areas measured at 100 mV reverse bias, the average and standard deviation in dark current to noise current ratio is α _D=i_n/I_d=1.39+/- 1.09x10^-5. Defects are presumed resident in the detectors that produce greater non- uniformity in the 1/f noise as compared to the dark current at 100 mV reverse bias. Noise was also measured as a function of temperature for two λ _c=16 micrometers detectors from 55 K to 100 K. The average and standard deviation in the noise current to dark current ratio is α_D=i_n/I_d=2.36+/- 0.83x10^-5 for the 26-micrometers -diameter detector and (alpha) _D=1.71+/- 0.69x10^-5 for the 16-micrometers -diameter detector. Dark and noise current were measured while changing the bias applied to a detector. In the diffusion-limited portion of the detector I-V curve, 1/f noise is independent of bias with α _D=i_n/I_d=1.51+/- 0.12x10^-5. When tunneling currents dominated, α_T=i_n/I_d=5.21+/- 0.83x10^-5. The 1/f noise associated with tunneling currents is a factor of three greater than the 1/f noise associated with diffusion currents. In addition, 1/f noise was measured on detectors held at -100 mV and 78 K under dark and illuminated conditions. The average noise to current ratio α_D was approximately 1.5 x 10^-5 for dark and photon-induced diffusion current. However, detector-to-detector variations exist even within a single chip. The two most important points are that non-uniformities in material/fabrication need to be addressed and that each individual type of current component has an associated 1/f noise current component, the magnitude of the relationship being different depending on the source current.

Shortwave, midwave, and longwave HgCtTe focal plane arrays with a format of 320x256 have been produced by both heterojunction epitaxy and boron implantation techniques. In general, the heterojunction diodes and arrays with a p-on-n polarity have high diode RoAs at high temperatures, while the boron implanted diodes and arrays with an n-on-p polarity have high diode RoAs at lower temperatures and better array operability because of excellent diode surface passivation. Diodes with wavelength longer than 20 micrometers have been produced. The 320x256 HgCdTe arrays have been fabricated and hybridized to readout integrated circuit chips ISC 9705 and ISC 9809 designed by Indigo Systems Inc. Imaging pictures were taken by cameras equipment with these array hybrids. The array operability depends on the hybrid operating temperature. For heterojunction arrays, the best operability of 2.5micrometers arrays at 200K is over 98%, while the best operability of 9.7micrometers arrays at 77K is over 96%. The operability of n-on-p arrays hybridized to ISC9809 cannot be determined because the readout circuit is not specifically designed for arrays with this polarity. However, testing results indicate that with proper readout chips, array operablity over 99% can be achieved with boron- implanted arrays.

To meet the demands for high performance HgCdTe detectors at high yield and producibility, key processes have been optimized and new approaches have been developed. By a superior CdZnTe Bridgman growth process, dislocation densities <1x10⁵cm^-2 in substrate and epitaxial layer are achieved for all substrates, ensuring high performance Focal-Plane-Arrays, particularly for (lambda) _CO=11,5 micrometers arrays. A new guard ring approach for planar diodes, created by a n⁺-region in pixel spacing area reduces pixel crosstalk and improves Modulation Transfer Function. For high thermal cycles of the FPA, the flip-chip- technique has been optimized, leading to >2000 cycles for 640x512-FPA's. Producibility and reliability of AIM's MCT FPA technology are demonstrated.

This paper presents a new generation of long linear butted arrays developed at LETI / LIR and dedicated to high resolution IR imaging for space and airborne applications. These IR focal plane arrays are based on modular architecture composed of butted CdHgTe photovoltaic detection circuits and Si CMOS multiplexers. CdHgTe detectors are made with the new planar technology leading to very high performances with low dark current. Specific in house techniques of dicing and butting are used in order to perform linear detection arrays, free of defect. Detectors redundancy has also been implemented in order to compensate bad detectors (deselection) or to improve the sensor uniformity. A presentation of the main electro-optical performances obtained in the two spectral ranges - 3-5 and 8-10 micrometers - is given. Comparison with the previous prototype (first generation) developed at LIR will also be discussed in the paper. Details of involved technologies, FPA architecture, operability and performances (NETD, responsivity, dark current .) of butted sensors working in the 3-5 and 8-10 micrometers spectral ranges will be presented.

Variation in rectification current with ac-bias frequency has recently been observed in metal-semiconductor-metal (MSM) detectors when utilized as optoelectronic mixers in a frequency-modulated continuous-wave (FM/cw) LADAR System. This current degrades the performance of the LADAR System by inducing false targets. In this paper, we present a detailed experimental and theoretical investigation on the origin of this current. We find that MSM detectors exhibit asymmetric current-voltage characteristics that are related to imperfections in device design and processing. We also find that, although the asymmetry is usually small, a rectification current may exist even under zero mean ac bias. Both the dark current and the photocurrent exhibit asymmetric behavior, but have opposite asymmetry with respect to one another. Under transient bias voltage the device shows two transient current responses: (1) a fast one related to the displacement current and (2) a slow one related to the removal of carriers from the device. The asymmetry in current related to the slow process is opposite to the dc asymmetry, while the asymmetry in current related to the fast process is more symmetric. The rectification current varies not only with ac voltage and optical power, but also with ac bias frequency.

Infrared microbolometer thermal detectors have been fabricated on a flexible substrate, a polyimide that shows very similar characteristics to Kapton , when cured. The polyimide is spin-coated on a regular Si wafer with a release layer. Low temperature fabrication techniques are employed to minimize the thermal cycling of the polyimide substrate as well as to maintain compatibility with CMOS circuitry. Infrared microsensors on flexible substrates showed a Temperature Coefficient of Resistance (TCR=(1/R)(dR/dT)) of -3.03%, at room temperature. The microbolometers reached the responsivity and detectivity of 3.5x10³ V/W and 1x10⁷ cm.Hz^1/2 /W, respectively, at 2.88 μA of current bias even though at this time no micromachining has been performed. Some devices were encapsulated between polyimide layers in order to provide protection and passivation. These bolometers demonstrated a responsivity and detectivity of 1.6x10³ V/W and 4.9x10⁶ cm.Hz^1/2/W, respectively at 1 μA of current bias. This is the first attempt towards a smart skin that incorporates flexible microsensors on a cured polyimide-substrate.

Raytheon Infrared Operations (RIO) has achieved a significant technical breakthrough in uncooled FPAs by reducing the pixel size by a factor of two while maintaining state-of-the-art sensitivity. Raytheon has produced the first high-quality 320x240 microbolometer FPAs with 25 micrometers pitch pixels. The 320 x240 FPAs have a sensitivity that is comparable to microbolometer FPAs with 50 micrometers pixels. The average NETD value for these FPAs is about 35 mK with an f/1 aperture and operating at 30 Hz frame rates. Good pixel operability and excellent image quality have been demonstrated. Pixel operability is greater than 99% on some FPAs, and uncorrected responsivity nonuniformity is less than 4% (sigma/mean). The microbolometer detectors also have a relatively fast thermal time constant of approximately 10 msec. This state-of-the-art performance has been achieved as a result of an advanced micromachining fabrication process. The process allows maximization of both the thermal isolation and the optical fill-factor. The reduction in pixel size offers several potential benefits for IR systems. For a given system resolution (IFOV) requirement, the 25 micrometers pixels allow a factor of two reduction in both the focal length and aperture size of the sensor optics. The pixel size reduction facilitates a significant FPA cost reduction since the number of die printed on a wafer can be increased. The pixel size reduction has enabled the development of a large-format 640x480 FPA array. Raytheon has produced arrays with very good sensitivity, operability, and excellent image quality. These FPAs are applicable to wide-field-of-view, long range surveillance and targeting missions. Raytheon is also developing a high performance 160x128 FPA that is designed for applications where miniaturization and temperature invariance are required as well as low cost and low power.