This PDF file contains the front matter associated with SPIE Proceedings Volume 10624, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.

Attollo Engineering will present results of our research program developing extended SWIR sensors as well as the packaging and camera electronics surrounding it. The 640x512 sensor uses GaInAsSb for the active layer and has a cutoff wavelength of 2.5 m. The unipolar barrier structure enables a higher operating temperature by substantially reducing dark current caused by G-R mechanisms and surface leakage. The material is grown on GaSb and is made up of GaInAsSb absorber and contact layers separated by an AlGaSb barrier. We will present dark current and imaging results from the sensor fabrication at different temperatures. The detector array was hybridized to a 15 m pixel pitch ROIC that has a direct injection unit cell. The hybridized sensor was packaged into a custom 4-stage thermoelectrically cooled package. The package was particularly designed to minimize the heat load and maximize the thermal conduction. We will present the trades that went into designing the package and the internals of the package. The cooler stabilized the sensor temperature at 200K. The electronics used to drive the package have the ability to change biases and timing on the fly using software controls. Attollo designed these electronics to be a low-cost solution for demonstrating sensors in many different modes. We will show information regarding each stage of integration and show the results of the imaging using the eSWIR sensor and supporting equipment.

In the recent years AIM has developed IR modules for imaging applications sensitive in the extended SWIR (eSWIR) spectral range with a cut-off wavelength of 2.5μm. The modules are based on AIM’s state-of-the-art MCT FPA technology with a dedicated ROIC integrated in a low size, weight and power (SWaP) Dewar/Cooler configuration. The different modules are either designated to combine emissive and reflective imaging in one sensor by already detecting thermal radiation in the eSWIR band or to make use of laser illumination including gating capability. Further integration including a lens, image processing electronics and power supply was done resulting in sights for targeting and reconnaissance applications. For targeting under low-light conditions a low SWaP MCT eSWIR module in a 640x512 format with 10μm pitch was developed. The module was integrated to a compact clip-on weapon sight for small arms providing medium range performance up to 1000m. For reconnaissance applications a MCT based SWIR 2D APD array in a 640x512 15μm pitch format was developed providing gated viewing capability. The device was optimized to provide a higher gain for improvement of the signal to noise ratio (SNR). The module was integrated in a camera demonstrator including a 1.5μm laser illuminator for field trials to demonstrate long range identification. The paper will present the latest performance results including field trials of MCT based eSWIR modules and sights for imaging applications.

Extended wavelength InGaAs detectors grown on InP substrates have been generally used only in single element detectors and low resolution linear arrays. The extended wavelength InGaAs is no longer lattice matched to the InP substrate so it requires buffer layers to be used in the epitaxial growth process to accommodate the strain of the mismatched material. These detectors are generally front side illuminated with wire bonded pads. This work describes the results of extended wavelength InGaAs detector arrays that are backside illuminated which presents many more challenges, including imaging through the buffer layers as well as hybridization to Si Readout Integrated Circuits (ROICs). The buffer layers absorb shorter wavelength light making NIR response challenging. The arrays produced in this work are at high resolution, 1.3 megapixels on small pitch of 12 µm. The imagers have response from 700 nm to <2000 nm while imaging via backside illumination. New processing methodologies were developed to extend the short wavelength response to allow for NIR response by removing the substrate and most of the buffer layers from the structure after hybridization. This has produced material with quantum efficiencies <50% across most of its detection range during TEC cooled operation.

In this work we use analytical and 3D numerical modeling tools to analyze data from InP/In_0.53Ga_0.47/InP double layer planar 15 pixel pitch focal plane arrays (FPAs) designed to image in the near infrared to determine array suitability for operation under low-level illumination, including overcast, “Night Glow”- only conditions. Notable is that the diffusion dark current component is the dominant current component near and above 300K and is limited by band-to-band radiative recombination processes. The Shockley- Read- Hall (SRH) minority carrier lifetime is τ_SRH =107μs. Recombination through band gap states in the space charge region (SCR) situated at the intrinsic Fermi level is the dominant component for temperatures below 300K as previously demonstrated using 3-D numerical simulations consisting of both bulk area and perimeter dependent components. 3-D numerical simulations in combination with scanning capacitance microscopy are paramount to characterizing the Zn-diffused p⁺n shallow step homojunction and to identifying technology limitations in small pitch high density (FPAs). Photon recycling effects, i.e., effects caused by repeated trapping of photons, are not observed in the measurements of the minority carrier lifetime and the diffusion dark current component. As a result, the diffusion current component J_diffusion α to the radiative recombination rate G_r(α) and the radiative minority carrier lifetime τ_radiative = 1/BN_d, where B is the radiative recombination coefficient and N_d the majority carrier concentration.

Night Vision Imaging in the Short-Wave Infra-Red (SWIR) has some unique advantages over Visible, Near Infra-Red (NIR) or thermal imaging. It benefits from relatively high irradiance levels and intuitive reflective imaging. InGaAs/InP is the leading technology for two-dimensional (2D) SWIR detector arrays, utilizing low dark current, high efficiency and excellent uniformity. SCD's SWIR Imager is a low Size, Weight and Power (SWaP) video engine based on a low noise 640x512/15μm InGaAs Focal Plane Array (FPA) embedded in a low cost plastic package which includes a Thermo-Electric Cooler (TEC). The SWIR Imager dimensions are 31x31x32 mm³, it weighs 50 gram and has less than 1.4W Power consumption (excluding TEC). It supports conventional video formats, such as Camera Link and BT.656. The video engine image processing algorithms include Non-Uniformity Correction (NUC), Auto Exposure Control (AEC), Auto Gain Control (AGC), Dynamic Range Compression (DRC) and de-noising algorithms. The algorithms are specifically optimized for Low Light Level (LLL) conditions enabling imaging from sub mlux to 100 Klux light levels. In this work we will review the optimized video engine LLL architecture, electro-optical performance and the applicability to night vision systems.

Understanding the failure mechanisms in high performance detector arrays is critical for meeting the demands for a given application. For SWIR sensing, the detector requires the highest sensitivity possible to operate under photon-starved conditions. Using a 3D drift-diffusion model, we simulate the influence of heterointerface traps on the dark current in In_0:53Ga_0:47As / InP p-on-n planar heterojunction photodiodes. We calculate how the dark current changes with junction area, and show that it increases linearly with junction radius when the device is limited by surface recombination. An analytical model is developed to understand the geometric dependence of both bulk and surface generation/recombination currents. Finally, insight on mitigation strategies and possible impact on quantum efficiency are both discussed.

Metal-insulator-metal (MIM) diodes are highly considered in high frequency applications in form of rectennas for energy harvesting applications due to their fast speed, small size, and ease of fabrication and IC compatibility. In these diodes, insulators are integral part of the device, determining performance parameters. In this study, we have evaluated HfO₂ and Al2O₃ based MIM diode structures to compare and determine performance parameters, with conversion efficiency being prioritized. The fabrication processes in physical vapor deposition (PVD) systems for the MIM diodes resulted in the devices having high non-linearity and responsivity. Also, to achieve uniform and very thin insulator layer atomic layer deposition (ALD) was used. We implemented the same MIM structure in 10x10 array form, with active area of 200x325 nm². The efficiency values of same arrays tested with 1200 and 1600 nm wavelength LEDs for 200x325 nm² diode active area without applying bias. The conversion efficiency value of the HfO₂ based structures calculated as 5% for 1200 nm wavelength. These measured values of conversion efficiency are reported for the first time in the literature for MIM diodes in SWIR operation.

The typically used shortwave infrared spectral range (SWIR) between 900 nm and 1700 nm is a spectrally broader wavelengths range than the visible range. Available SWIR cameras generate a gray level image using the intensity over the entire spectral band. However, objects can exhibit completely different spectral behavior in this range. Plants have a high reflection at the lower end of the SWIR range and liquid water has a strong absorption band around 1400 nm, for example. We propose to divide the SWIR range into an appropriate number of spectral channels to extract more details from a captured image.

To extract this information the proposal follows a concept similar to color vision of the human eye. Analog to the three types of color receptors of the eye four spectral channels are defined for the SWIR. Each point of the image is attributed now by four “color values” instead of a single gray level.

For a comprehensive characterization of an object, a special SWIR colorimetry is possible by selecting appropriate filters with suitable band width and spectral overlap. The spectral sensitivity, the algorithms for calculating SWIR-color values, the discrimination of SWIR-color values by Noise Equivalent Wavelength Difference (NEWD) and spectral coded false color image display is discussed and first results with an existing SWIR camera are presented.

In past performance analyses and comparisons of MWIR and LWIR systems, infrared systems scientists and engineers did not have the cumulative technologies that we will soon enjoy. Large format-small pitch, deep wells, and massive processing do not exist in a single focal plane, but they are reality now individually and will exist collectively in the near future. How do we best use these technologies and how do we compare sensors when we use these technologies? From a more fundamental aspect, how do you optimize a system given that practical limits are minimized and theoretical limits apply? Smaller pitch infrared detectors can provide longer range performance for a given aperture and higher photon collection duty cycles (deep wells and faster frame rates) can allow better Modulation Transfer Function correction. Massive processing allows for recovery of resolution by trading surplus signal to noise ratio. Non-uniformity correction becomes an important issue, but there are smart methods using higher duty cycles to address the problems. LWIR can compete with MWIR using the additional photons given an improved photon collection duty cycle. A holistic approach to system design can provide for an extremely high-performance system. It is also worth mentioning that infrared targeting sensor design in the future should be quantified with more than just identification range. Since these technologies provide more than a human can consume, the sensors need to be designed smarter to better utilize human consumption limits. An example is that small pitch high density sensors (solid state imaging) can provide faster target prosecution which allows for faster target engagements. We show these possibilities using a LWIR targeting sensor to demonstrate the concept of optimizing pitch-well-processing (PWP).

In recent years a diversity of new IR-detector formats, mainly in MWIR but as well in LWIR spectral range, has been introduced by leading detector manufacturer. These arise from enormous progress in pitch size reduction, while keeping array size manageable. As a consequence, opportunities on system level for modernization and performance enhancement are manifold. Portable systems will benefit from ultra-compact, low power MWIR solutions, while rotorcraft pilotage or surveillance applications will rely on LWIR detector arrays with highest possible resolution. IR-modules providing HD-format (1280x720) in 12μm pixel pitch were already introduced at AIM in 2016 for both MWIR and more challenging LWIR spectral range. Detector arrays with an even smaller pixel pitch of 10μm with XGA format (1024x768) have so far only been realized in MWIR for usage as system upgrade of 640x512, 15μm pitch IRmodules and for an ultra-compact IR-engine in a low power, high operating temperature (HOT) version. The electrooptical characterization of this, recently presented, HOT IR-engine has been continued and performance and design have now further been optimized. Moreover, next steps in pitch size reduction towards wavelength scale are under development at AIM by introducing an XGA format, 10μm pitch LWIR version using an optimized ROIC design and by preparing required technology and processes for an ultra-small pitch of 7.5μm and beyond. In this paper latest performance results of MWIR and LWIR detector arrays with small pixel pitch will be presented, together with design considerations and optimization of associated cooled IR-modules.

Meteorological satellites have become an irreplaceable weather and ocean observing tool. There are totally 13 meteorological satellites that were launched into both sun-synchronous and geostationary orbit in China. All the satellites have been incorporated into the global constellations of operational meteorological satellites within the WMO framework. As the second-generation geostationary-orbit meteorological satellite, FY-4 was launched on Dec.11, 2016. There are four payloads onboard. Two different type of infrared remote sensors, Advanced Geostationary Radiometric Imager (AGRI) and Geostationary Interfering InfraRed Sounder (GIIRS), are the main payloads on-board FY-4. AGRI has 14 spectral bands located on six focal plane assemblies (FPAs) with the highest spatial resolution of 500 meter, and will scan much faster than the current FY-2 radiometer. AGRI covers from visible band (0.45μm) to thermal infrared band (13.8μm). GIIRS is nadir and limb viewing infrared Fourier transform spectrometers (FTS), which could be able to provide sounding data with 0.625 cm-1 spectral resolution covering wavelengths in the midwave infrared (1650～2250 cm-1) and longwave infrared (700～1130 cm-1) to users, and to get the atmospheric temperature profile and humidity profile. China may be the first country that could get high spectral infrared data from both geostationary and polar-orbiting satellites. Some tradeoffs have been made to build the infrared sensors. In accordance with specification and designing strategy, the following choices of single-pixel detector, linear detector and focal plane detectors, thermal cooler and mechanical cooler, different type of telescope, double-axis gimbaled mirror and single-axis scanning mirror, have been selected respectively. The characteristics of these infrared sensors, design overview conjoined with detector, cooling, optics, pointing, observation strategy, calibration strategy, etc. are introduced in this paper.

In the past decades, hyperspectral imaging technologies was well developed in the whole world. Visible and Near Infrared hyperspectral imagers play an important role in agriculture, land use, forestry, etc. Higher performance airborne hyperspectral imagery is strongly expected these years. Wider Field of View and higher resolution instrument can acquire data more efficiently. A VNIR PHI with 40 degree FOV, 0.125mrad IFOV, 256bands was integrated last year. The system can adapt to the Velocity to Height Ratio lower than 0.04. The system consists of 3 subsystems. Every subsystem consists of TMA fore optics, spectrometer with planar blazed grating and electronics. The 3 subsystems work for left, middle, right FOV, respectively. Thanks to CCD’s pixel binning function, the system can operate in high spectral resolution mode, high spatial resolution mode, and high sensitivity mode for different applications. The integration was finished, and airborne flight validation experiments were conducted.

ABB has recently designed a compact, low-cost and light weight infrared spectral imager. That instrument re-use many building blocks from ABB’s commercial line of products, repackaged in a system adapted for field operations. That instrument incorporates limited imaging capacity in order to improve the fill factor and the signal to clutter ratio when the target is not spatially uniform. The current version operates in the long-wave infrared from 6 μm to 14 μm. Its main applications include chemical detection and identification, environmental monitoring and LWIR infrared signature measurements.

Here we report our recent achievements towards a compact, portable, handheld device for contactless real-time detection and identification of explosives and hazardous substances via reflectance spectroscopy in the 7.5 μm – 10 μm spectral region. The mid-IR spectroscopic measurement principle relies on selective illumination of the target using broadly tunable external cavity quantum cascade lasers (EC-QCLs). A resonant micro-opto-electro-mechanical systems (MOEMS) grating enables fast wavelength tuning in the external cavity, allowing the full spectral scan to be completed in <1 ms. The diffusely backscattered light’s intensity dependence on illumination wavelength provides spectroscopic information to identify threat compounds via our spectral database, containing a large number of materials relevant in a security context. We present a handheld portable, albeit tethered, device capable of real-time identification of hazardous substances at a range of 1 m. We will outline future improvements to increase the system’s usability, such as integrated computing power, automated focusing to that allow use over a range of detection distances and spatial scanning for background subtraction.

We report the results from proton and gamma ray radiation testing of HgCdTe avalanche photodiode (APD) arrays developed by Leonardo DRS for space lidar receivers. The devices were tested with ~60 MeV protons up to 100 krad(Si) without the read-out integrated circuit (ROIC) and 30 krad(Si) with the ROIC. We also measured the transient responses with the device fully powered and the APD gain set to >1000. The detectors output a large current impulse in response to each proton hit, which could saturate the ROIC but recovered within 1 μs. The APD dark currents increased linearly with the proton dose. The quantum efficiency and APD gain decreased slightly with dose. The dark currents due to the radiation damage went up many times after the devices were warmed to room temperature and cooled to 80K again. The radiation damage was found to completely anneal after baking the device at 85°C or higher. These results showed the HgCdTe APD arrays are suitable for use in space lidar for typical Earth orbiting and planetary missions provided that provisions are made to heat the detector chip to 85°C when the system performance is impacted.

In this paper, results from the development of LWIR and VLWIR InAs/GaSb type-II infrared photodetector arrays are presented. Dark currents comparable to the HgCdTe benchmark (Rule07) have been observed and the quantum efficiencies of the detectors exceed 30 %. Bias and temperature dependencies of the QE have been studied showing very low turn on bias (~-25mV) and no variation of the peak QE value with temperature. These results show that there are no unintentional barriers in the detector structures and that the diffusion lengths are long enough to provide efficient collection of carriers. Initial results from the extension of the cut-off wavelength from 11 μm to 14 μm are also presented as well as initial results from photodiodes with thicker absorbers to enhance the QE.

Through the choice of appropriate layer thicknesses, the bandgap of InAs/Ga(As)Sb type II superlattices (T2SLs) can be engineered in a wide range covering the mid-wavelength and long-wavelength infrared (MWIR, 3 μm - 5 μm and LWIR, 8 μm - 12 μm) spectral regions. Using this material system, Fraunhofer IAF develops bi-spectral MWIR image sensors based on homojunction photodiodes for missile warning applications and pursues modern heterojunction approaches as well as heteroepitaxial growth of T2SLs on GaAs. We discuss topics arising from efforts to improve the manufacturability of our bi-spectral arrays and report on the progress of the integration with MWIR heterojunction designs that exhibit reduced dark currents.

The tri-service Vital Infrared Sensor Technology Acceleration (VISTA) program rapidly matured III-V semiconductor epitaxy to produce tactically viable detectors using Type II Superlattice (T2SL) structures. The T2SL material system allows tunable band gaps for creating lattice-matched heterojunction devices. Heterojunction devices are integral to suppressing sources of dark currents, such as internal Shockley Reed Hall (SRH) and device surface currents. Once the VISTA program demonstrated that T2SL detectors offered competitive performance to traditional indium antimonide (InSb) detectors at an operating temperature 40K to 50 K higher, many opportunities emerged. This elevation in operating temperature provides two benefits to infrared (IR) sensors. The first is to miniaturize the integrated Dewar-electronicscooler assembly (IDECA) such that it can support small aerial vehicle and soldier mounted sensors. The second is to increase the mean time to failure (MTTF) of an existing InSb IDECA. To benefit from T2SL higher operating temperature (HOT) detectors, the overall cost of the IDECA must be competitive with InSb. This drives a manufacturing capability that is equivalent to InSb. At the L3 Space and Sensors Technology Center (L3 SSTC), the III-V detector foundry processes 125 mm diameter InSb wafers. The development of 125 mm diameter T2SL detector wafers started with the gallium antimonide substrates. The greater size and weight of these substrates required extra care to avoid breakage. Leveraging the learning reported from the silicon industry, we developed a specification for the substrate thickness and edge bevel to provide a robust platform for wafer processing. Next, we worked with commercial III-V epitaxy suppliers to develop multi-wafer growth capability for 125 mm diameter substrates. The results of this effort, funded by the Office of the Secretary of Defense (OSD) Defense-wide Manufacturing Science and Technology (DMST) program through the Army Night Vision and Electronic Sensor Directorate (NVESD), we were able to improve focal plane array (FPA) yield from virtually zero to InSb manufacturing levels.

Accurate p-type doping of the active region in III-V infrared detectors is essential for optimizing the detector design and overall performance. While most III-V detector absorbers are n-type (e.g., nBn), the minority carrier devices with p-type absorbers would be expected to have relatively higher quantum efficiencies due to the higher mobility of their constituent minority carrier electrons. However, correctly determining the hole carrier concentration in narrow bandgap InAsSb may be challenging due to the potential for electron accumulation at the surface of the material and at its interface with the layer grown directly below it. Electron accumulation layers form high conductance electron channels that can dominate both resistivity and Hall-effect transport measurements. Therefore, to correctly determine the bulk hole concentration and mobility, temperature- and magnetic-field-dependent transport measurements in conjunction with Multi-Carrier Fit analysis were utilized on a series of p-doped InAs_0.91Sb_0.09 samples on GaSb substrates. The resulting hole concentrations and mobilities at 77 K (300 K) were 1.6 x 1018 cm^-3 (2.3 x 1018 cm-3) and 125 cm2 V^-1 s^-1 (60 cm2 V-1 s-1), respectively, compared with the intended Be-doping of ~2 x 10¹⁸ cm^-3.

To improve the performance of photodiodes based on narrow-bandgap InAs/GaSb type-II strained layer superlattices (T2SLs), knowledge of the vertical minority carrier transport is necessary. For this purpose, the key parameters influencing vertical minority-carrier electron transport in an nBp MWIR detector were studied: diffusion length, lifetime, mobility. The detectors were designed with p-type, 10/10 ML, InAs/GaSb T2SL absorbers, targeting a 50% cut-off wavelength of 5.0 µm at 80 K. The nBp structure is attractive because the junction field predominately drops across a relatively wide-gap InAs/AlSb SL barrier, which reduces the expected generation-recombination dark current. Measurements of the electron beam-induced current (EBIC), combined with minority carrier lifetime results from microwave reflectance measurements, enabled the determination of the minority carrier diffusion length (Le) and mobility in the growth direction as a function of temperature. The Le was extracted at each temperature by fitting the EBIC data to analytical expressions for carrier collection efficiency. The EBIC measurements were also repeated at different electron-beam energies to vary the distribution of minority carriers near the surface to gauge the surface recombination velocity. Microwave reflectance allowed for accurate measurement of the minority carrier lifetime over a large dynamic range of excess carrier concentrations, enabling a separation of recombination mechanisms. The lifetime and extracted diffusion length data were then used to estimate the diffusion coefficient and mobility versus temperature by applying the Einstein diffusion relationship.

Recent advances over the last several years in III-V strained-layer superlattice-based infrared detectors have lead this material system to emerge as a solid alternative to HgCdTe for dual-band focal plane arrays (FPAs). Rapid development of superlattice-based detectors has been realized by capitalizing on mature, III-V foundry-compatible processing. Furthermore, superlattice-based epitaxial wafers exhibit a high degree of lateral uniformity with low macroscopic defect densities (< 50 cm^-2) and can achieve dark current levels comparable to HgCdTe detectors. In this paper, we review our recent efforts towards producing HD-format (1280x720, 12 μm pitch) superlattice-based, dual-band MWIR/LWIR FPAs. For a representative FPA, characterization was conducted in a pour-fill dewar at 80K, f/3 and using a blackbody range of 22°C to 32°C. For the MWIR band, the noise equivalent temperature difference (NETD) was 14.9 mK with a 3x median NETD operability of 99.91%. For the LWIR band, the median NETD was 28.1 mK with a 3x median NETD operability of 99.66%. To illustrate the manufacturability of superlattice technology, we will present results on 1280x720, 12 μm pitch MWIR/LWIR FPAs built over the last year at HRL through multiple fabrication lots utilizing 4" epiwafers.

Distribution Statement A: Approved for public release. Abstract III-V Infrared Focal Plane Array Development in US Authors: Meimei Tidrow, Sumith Bandara, Leslie Aitcheson, Lucy Zheng, Neil Baril, Alicia Williams This presentation describes the continuation of work and unprecedented progress stemming from a national effort called Vital Infrared Sensor Technology Acceleration (VISTA). VISTA is a U.S. Department of Defense research and development program that made tremendous strides in the development of III-V antimony-based infrared detectors, pioneering much of the work in this area and transitioning results to industry partners. As part of this success, the program demonstrated very large-format high operating temperature mid-wavelength as well as mid-wavelength/long-wavelength dual band 720x1280 12-m pitch infrared focal plane arrays. While the program ended in 2015, DoD is continuing pursuit of this highly-flexible technology for advanced infrared sensors under numerous programs. Their interest is in capitalizing on this material system’s ability to reduce cost, size, weight, and power requirements of sensor systems, while providing small pixel sizes, very high operability, and large formats—all with high performance. In this presentation, we will outline recent work and results under DoD programs stemming from final results of VISTA.

Valence band features affecting carrier transport in III-V superlattice nBn detectors. David R. Rhiger and Edward P. Smith, Raytheon Vision Systems. We have investigated non-ideal features occurring in the valence band profile of nBn detectors which affect the transport of minority-carrier holes representing the IR signal. The objectives are to reduce dark currents and improve quantum efficiency. The nBn device consists of an n-type absorber several microns thick, plus a very thin electron barrier B and a thin n-type collector (top contact region). In this investigation, the absorber and collector were built with the InAs/InAsSb superlattice. For normal operation, holes generated by photons in the absorber must flow to the collector. Current is promoted by a combination of diffusion and electric field drift. However, in some cases the transport of holes is limited by (1) absorber-barrier valence band misalignment, (2) bandgap difference between collector and absorber, or (3) possible localization sites in the absorber due to compositional fluctuations. These characteristics, when combined with the known limitations of hole diffusion length, can adversely affect the quantum efficiency, and require the application of an operating bias that is larger than otherwise necessary. We have been able to identify and measure these valence band features by analyzing device characteristics as a function of temperature, bias voltage, photon flux, and wavelength dependence of the response. Examples will be presented. This work was supported by Dr. Meimei Tidrow of NVESD, Contract Number W15P7T-06-D-E402, Task BD30, Agreement No. S08-092256, Purchase Order P000006939.

Long-wavelength (8-14μm) infrared detection ability using third-generation infrared focal plane array (FPAs) is a desideratum for aerography, military and communication. These optical bands contain tremendous information about CO₂ levels, atmospheric quality and biological activity. HgCdTe infrared photodetectors are able to reach high degree of performance even to be background limited. However, the material growth process, doping techniques and capability of defect control become increasingly difficult for the shrinking bandgap. Besides, the dark current characteristic and associated noise behavior are very sensitive to the detector fabrication processes. Thereby, the growth of p-type epitaxial layer is a fundamental and significant subject for long-wavelength HgCdTe infrared photodetector.

Current development efforts in IR-module technology show two major trends: Reduction in size, weight and power dissipation of IR-systems and further increase in system performance by introducing 3rd Gen IR-modules. Concerning 3rd Gen IR-modules AIM is developing SWIR/MWIR and MWIR/MWIR bispectral MCT detectors making use of its established and qualified MBE technology based on the growth of MCT multi-layers on GaAs substrates. The advantage of multispectral versus single color IR-sensors is the ability to combine sensitivity of two different IR wavelengths in one detector. This greatly enhances the ability to gather information from a scene, which is a significant additional benefit for IR-systems, for applications such as seeker heads, missile warners or counter measures against laser-guided beam-rider weapons. In particular, the combination of the SWIR/MWIR or MWIR/MWIR spectral bands promote an enhanced target discrimination and identification via increased identification range, achieved by enabling the target acquisition in front of cluttered backgrounds or of targets with low thermal signature. The information of the SWIR spectral range, which detects mainly the reflected part of the spectrum, and the passive IR-detection in the MWIR spectral range, can be favorably combined for the data acquisition and subsequent image data processing in our bispectral approach due to its temporal and spatial coincidence of the scene image. In this paper results will be presented of AIM’s SWIR/MWIR, as well as MWIR/MWIR bispectral MCT detectors with 320x256 pixels and a 30 μm pitch. The detectors demonstrate very low color cross-talk, and an excellent NETD in conjunction with low defect densities.

Although HgCdTe imagers are a well-established technology, photodetectors fabricated using the same process still yield a large variation in their performance characteristics, largely stemming from hard-to-control pecu- liarities at the interface between the surface passivation and the active region of each photodiode. This work investigates the dark current characteristics of long-wave IR (cutoff wavelength of 10um) Hg_0.774Cd_0.226Te mesa photodiodes, which have been passivated with a CdTe film. We use a 2-D model of a p-on-n device structure to study how interface states and Cadmium diffusion at the passivation interface can influence the photodiode dark current.

SOFRADIR is the worldwide leader on the cooled IR detector market for high-performance space, military and security applications thanks to a well mastered Mercury Cadmium Telluride (MCT) technology, and recently thanks to the acquisition of III-V technology: InSb, InGaAs, and QWIP quantum detectors. This is the result of strong and continuous development efforts to deliver cutting edge products with improved performances in terms of spatial and thermal resolution, dark current, quantum efficiency, low excess noise and high operability. On one hand the advanced performances of Sofradir product rely on a strong partnership with CEA-LETI materialized in a common laboratory named DEFIR.

On the other hand, these cutting edge performances are made possible thanks to Sofradir vertical industrial model. From the CdZnTe (CZT) and HgCdTe (MCT) crystal growth to the last electro-optical characterization recipe before shipping, and all the intermediate steps in between like IDDCA (Integrated Detector Dewar Cooler Assembly) final pumping cycle, all the manufacturing steps are developed, performed and controlled inhouse. This allows direct feedback between IDDCA, system performances and process or material. State of the art relevant performances for IR detection and imaging will be presented, that is to say low excess noise defects, RFPN (Residual Fixed Pattern Noise), NUC (Non Uniformity Correction) table stability for Daphnis product, 10μm pitch XGA extended MW matrix at 110K and HOT (High Operating Temperature) p-on-n technology, VGA format with 15μm pitch MW at 160K.

ASELSAN A.S., the largest defense company in Turkey, initiated research activities on developing Mercury Cadmium Telluride (MCT) detectors in 2014. These research activities include bulk crystal growth and surface preparation of Cadmium Zinc Telluride (CZT) substrates, Molecular Beam Epitaxial (MBE) growth of MCT layers, MCT detector fabrication, Read-Out-Integrated-Circuit (ROIC) design and detector-dewar-cooler (DDCA) assembly development. Focal plane arrays with resolutions/pixel pitches of 320x256/30 μm and 640x512/15 μm are fabricated. Noise Equivalent Temperature Difference (NETD) of 320x256 FPA is 11 mK (f#/1.5, 77K) while the operability is 98.2%. 640x512 FPA provides NETD of 32 mK (f#/1.5, 77K) and the operability is 93.2%.

High-performance infrared detectors based on HgCdTe technology require high quality epilayers, for which bulk CdZnTe is considered as the ideal substrate, thanks to its ability to perfectly match its lattice constant. Reaching very high crystal quality of the material in terms of subgrain boundary absence, low dislocation density, homogeneous zinc distribution, and low micro-defect density is paramount to obtaining excellent image quality. Sofradir takes advantage of growing its own CdZnTe crystals for producing substrates, and thus controlling the quality of HgCdTe epilayers, which allows reaching high-performance imaging. Indeed, mastering the whole manufacturing chain from raw material to Focal Plane Array and throughout all the front-end and back-end steps delivers a unique opportunity for process improvements. This paper shows how the latest process improvements do translate into detector image quality and reliability improvements, focusing on Front End process (substrates and epilayers), showing for the first time correlation between substrate microscopic defects and FPA image quality. This was achieved thanks to the research collaboration between Sofradir and CEA-LETI. This global process optimization is done thanks to a large set of characterizations performed at each process step, such as IR-microscopy for the substrate inspection, chemical revelation of dislocations and x-ray double-crystal rocking curve mappings for the epitaxial layer. Image quality is examined in terms of operability, and excess noise. Finally, in addition to process improvements, knowing how each critical process step impacts the following one and translates into the final image quality allows sorting units at the right process step, which serves yield and product quality. These benefits of the Sofradir’s vertical integration model are illustrated on MWIR and LWIR technologies.

Cadmium Zinc Telluride (Cd_1-xZn_xTe or CZT) is a ternary II-VI compound semiconductor material that has been widely used in infrared detector applications for many years. Due to its lattice spacing, CZT is the substrate of choice for stabilizing Mercury Cadmium Telluride (Hg_1-xCd_xTe or MCT) crystal layer growth where the lattice matching reduces stress during detector growth processes for high performance infrared detectors and focal plane (FPA) arrays used in guidance systems and a wide array of IR applications. The manufacturing of high performance MCT IR detectors requires CZT substrates of high quality for both bulk and surface conditions thus enabling high quality MCT epitaxial layer crystallinity and low defectivity. In this work, we report on results on bulk CZT material grown using the Travelling Heater Method (THM) that are suitable for infrared focal plane array (IRFPA) detector applications. This proven crystal growth process has been used to manufacture CZT substrates meeting industry requirements of IR transmission, tellurium precipitate size, dislocations and of larger single crystal area. We will present results on chemomechanical (CMP) polishing of CZT substrates of square, rectangular and state-of-the-art round geometries utilizing standard production tool sets that are identical to those used to produce epitaxy-ready surface finishes on related IR compound semiconductor materials such as GaSb and InSb. Surface quality will be assessed by various analytical and microscopy techniques to validate the suitability of this material for epitaxial growth.

An enhanced computer program has been applied to explain in detail the influence of different recombination mechanisms (Auger, radiative and Shockley-Read-Hall) on the performance of high operation temperature long wavelength infrared p-i-n HgCdTe heterojunction photodiodes. The computer program is based on a solution of the carrier transport equations, as well as the photon transport equations for semiconductor heterostructures. We distinguish photons in different energy ranges with unequal band gaps. As a result, both the distribution of thermal carrier generation and recombination rates and spatial photon density distribution in photodiode structures have been obtained. It is shown that photon recycling effect limits the influence of radiative recombination on the performance of small pixel HgCdTe photodiodes. In comparison with two previously published papers in Journal of Electronics Materials (Lee et al., DOI: 10.1007/s11664-016-4566-6 and Schuster et al., DOI: 10.1007/s11664-017-5736-x) our paper indicates an additional insight on ultimate performance of LWIR HOT HgCdTe arrays with pixel densities that are fully consistent with background- and diffraction-limited performance due to system optics.

The main driving force for High Operating Temperature (HOT) detectors is the strong need for low cost, compact IR imaging solution capable of supporting a wide range of military and civilian applications. In the HOT regime where imagers can be cooled with multi-stage thermoelectric coolers, the major portion of the cost is due to the die-level back-end process, from the chip hybridization to final packaging. We present here an approach to achieve significant cost reduction of MWIR imagers by monolithically integrating III-V devices directly on Silicon substrates for wafer-scale fabrication and packaging of focal plane arrays (FPAs). High quality InAs films can be grown on a blanket Silicon wafer by metal-organic chemical vapor deposition (MOCVD) in a low growth temperature regime that complies with the thermal budget of the Si-electronics. High Resolution Transmission Electron Microscopy reveals predominantly oriented, single-crystal-like InAs films, with Σ3(111) twin boundaries, which our band structure calculations predict to be electrically benign. More intriguingly, selective-area growth on SiO₂-masked ROIC-like templates is demonstrated with single-crystal-like InAs film nucleation at small Si(001) openings, together with the suppression of unwanted deposition on the dielectric mask. High crystallinity lateral epitaxial overgrowth of the InAs islands and film coalescence is achieved, enabling the potential to fully cover the entire patterned substrate. MBE-grown MWIR devices (λcut-off = 4.1 μm) on blanket InAs/Si templates exhibit a dark current of 2x10^-5 A/cm² , a specific detectivity of 6x1011 Jones and a quantum efficiency (QE) above 60% at 100K. The QE remains constant at high temperatures (<200K) where the dark current approaches that of baseline single-crystal HOT devices grown on native substrates At 230K, it is 6x10^-2 A/cm², yielding a specific detectivity of 10¹⁰ Jones.

The unipolar barrier photodetector architecture such as the nBn provides an effective means for lowering generationrecombination dark current by suppressing Shockley-Read-Hall processes, and for reducing surface leakage dark current. This has been especially beneficial for III-V semiconductor based infrared photodiodes, which traditionally tend to suffer from excess depletion dark current and the lack of good surface passivation. Advances in bulk and type- II superlattice infrared absorber materials have provided continuously adjustable cutoff wavelength span ranging from 2 to 14 μm and beyond, greatly expanding the limited coverage provided by traditional bulk III-V infrared detectors based on InGaAs and InSb. In this work we discuss recent developments of antimonide-based extended-SWIR, MWIR, and LWIR detectors and focal plane arrays at the NASA Jet Propulsion Laboratory.

Polycrystalline PbSe has the potential to be a V-HOT FPA capable of achieving 50 mK NETD sensitivity with pixel pitch less than 25 μm. However, current approaches using CBD have major reproducibility deficiencies. PVD approaches still struggle with sensitivity and carrier sweep-out. SJOS has explored a variety of PbSe approaches and determined the balance between carrier lifetime, mobility, dark current, and spectral QE that must be achieved in order to produce detectors that have low 1/f noise, high responsivity, and can fabricated reliably. SJOS detectors having these features are to be demonstrated in this presentation We will present data on the microstructure, blackbody performance data, and carrier mobility and lifetime for halogen passivated PbSe that can meet these demands.

Teledyne Judson Technologies (TJT) has developed high operating temperature (HOT) mid-wavelength infrared (MWIR) photodetectors based on InAs/InAsSb type-II superlattice (T2SL) with an electron barrier. Large area discrete detectors of 0.25mm and 1mm diameters were designed and fabricated for front-side illumination. Comprehensive E-O characterization was performed at room temperature and thermo-electric cooled (TEC) temperatures. The unique fabrication process was developed for a quasi-planar structure, enabling simplified fabrication for low-cost large volume production. The detector shows a 50% cut-off wavelength of ~5.5μm at room temperature. Peak responsivity of 2.47 A/W was achieved on 1mm detectors at peak wavelength ~ 4.24μm, -0.3V bias and 295K. Peak quantum efficiency (QE) was 72% with an antireflection coating. The 1mm detectors showed peak detectivity (D*) of 1.9x10⁹ cm-√Hz/W at -0.3V bias, 295K and 10 kHz. Dark current density as low as 1.17 A/cm² was achieved at -0.3V bias and 295K on 1mm detectors. The dark current was diffusion-limited at higher temperatures above ~120K while it was dominated by either tunneling or surface leakage currents at lower temperatures. Similar results were obtained on 0.25mm detectors.

Infrared (IR) polarimetric imaging is drawing significant interest because of its role in the enhancement of object recognition or detection ability. Conventional IR polarimetric imaging requires the use of polarizers or filters with IR cameras, which increases the complexity and cost of such systems, and degenerates performance. If uncooled IR sensors could selectively detect polarization without the need for polarizers or filters, then this would widen their range of applications. We have therefore investigated polarization-selective absorbers based on plasmonic metamaterials. Onedimensional (1D) plasmonic nano-metagrating absorbers (PNMAs) with high aspect ratios (<10) and narrow grooves (ca. 150 nm) are highly promising candidates for this purpose. Numerical calculations indicate that polarization selective absorption of over 90% absorbance is achieved. The incident electromagnetic wave is strongly confined in the narrow grooves and produces plasmonic resonance; the absorption wavelength is defined only by the groove depth and is independent of the incidence angle. Such high aspect ratio gratings with narrow grooves exhibit the optical properties of metamaterials rather than those of conventional metal gratings. We recently developed a top-down fabrication procedure for PNMAs using tapered-sidewall molds with Au deposition, which achieved 100 nm width grooves and an aspect ratio of 15. The absorption wavelengths obtained were larger than the period of the PNMA, and absorption over 90% was achieved. The absorption bandwidth can be controlled according to the groove shape, so narrow and broadband operation can be realized. PNMAs are therefore promising for uncooled IR polarimetric image sensors in terms of both sensor performance and mass production.

Graphene, which is carbon arranged in atomically thin sheets, has drawn significant attention in many fields due to its unique electronic and optical properties. Photodetectors are particularly strong candidates for graphene applications due to the need for a broadband photoresponse from the ultraviolet to terahertz regions, high-speed operation, and low fabrication costs, which have not been achieved with the present technology. Here, graphene-based transistors were investigated as simple photodetectors for a broad range of wavelength. The photoresponse mechanism was determined to be dependent on factors such as the operation wavelength, the components near the graphene channel of the photodetector, and temperature. Here, we report the detailed mechanism that defines the photoresponse of graphene-based transistors. Graphene transistors were prepared using doped silicon (Si) substrates with a SiO₂ layer, and source and drain electrodes. Single-layer graphene was fabricated by chemical vapor deposition, transferred onto the substrates, and the graphene channel region was then formed. The photoresponse was measured in the visible, near-infrared (NIR), and mid- and long-wavelength IR (MWIR and LWIR) regions. The results indicated that the photoresponse was enhanced by the Si substrate gating at visible wavelengths. Cooling was required at wavelengths longer than NIR due to thermal noise. Enhancement by the thermal effect of the insulator layer becomes dominant in the LWIR region, which indicates that the photoresponse of graphene-based transistors can be controlled by the surrounding materials, depending on the operation wavelength. These results are expected to contribute to the development of high-performance graphenebased photodetectors.

In recent years, plasmonic resonant antennas have seen widespread consideration in many detection and chemistry applications due to their potential for enhancing and confining the emission and polarization of electromagnetic fields. Examples include optical couplers to ultra-compact photodetectors, high-resolution optical microscopy, enablers of single molecule Raman signal detection and heating elements that facilitate nanostructure growth. An asymmetric cross-bowtie antenna is investigated for providing a broad circular polarized frequency response in the long wave infrared (LWIR). The asymmetric cross-bowtie antenna is constructed with two perpendicular bowtie antennas with differing arm lengths. The asymmetric cross-bowtie antenna is numerically analyzed using a finite element method (FEM) solver; Ansys High Frequency Structural Simulator (HFSS). The two perpendicular bowtie antennas, under illumination, provide a wide-band localized circularly-polarized field within a shared antenna feed-gap. At the center frequency of 28.3 THz (10.6μm), a circularly-polarized state over 30% bandwidth is achieved. The antenna is then loaded with a metal-oxide-metal diode in order to design a circularly polarized antenna-coupled detector.

This paper presents the development of a single layer microbolometer pixel fabricated using only ZnO material coated with atomic layer deposition. Due to the stress-free nature and high temperature coefficient of resistance of the ALD coated ZnO material, it can be used both as structural and active layers in microbolometer detectors. The design, simulations, and the fabrication optimization of 35 μm single layer ZnO microbolometers are shown in this study. The designed pixel has a thermal conductance of 3.4x10^-7 W/K and a thermal time constant of 1.34 ms while it has a maximum displacement of 0.43 μm under 1000g acceleration. This structure can be used to decrease the design complexities and fabrication costs and increase the yield of the detectors making them possible to be used in low-cost applications.

In the past few years leaky-wave antennas have been a very active area of antenna research due to their beam-scanning abilities. With the surge in graphene and 2D material research applications, there have been efforts to design graphenebased antennas in the terahertz and infrared spectrum due to graphene’s ability to strongly localize electromagnetic waves which leads to the ability to miniaturize and reconfigure antennas through electrical bias, magnetic bias, acoustic bias or chemical doping. It has also been seen that uniaxial metasurfaces such as graphene strips or phosphorene monolayers demonstrate extreme topological transitions from closed elliptical, quasi-isotropic to open hyperbolic canalization regimes. Here we design a planar long-wave infrared leaky-wave antenna based on periodic graphene strips. The leaky-wave antenna consisting of the graphene strips shall radiate at different angles at long-wave infrared wavelengths but more importantly radiate at 28.3THz at different angles for different values of biasing of the monolayer controlled by the chemical potential of the monolayers. We also explore the anisotropy of ultrathin hyperbolic reconfigurable metasurfaces represented by graphene strips in the context of a leaky-wave antenna. The different canalization regimes of the graphene strip metasurface are explored for applications in the field of flatland optics and planar antenna arrays.

This paper presents the results of a high-performance digital QVGA-IRFPA based on uncooled microbolometers with a pixel-pitch of 17 μm and a chip-scale-package as the vacuum package developed and fabricated by Fraunhofer-IMS. Due to a direct conversion of the microbolometer’s resistance into a 16 bit value by the use of massively parallel on-chip Sigma-Delta-ADCs a high scene temperature dynamic range of more than 300 K and a very low NETD-value below 50 mK is achieved. Due to a broad-band antireflection coating the digital 17 μm QVGA-IRFPA achieves a high sensitivity in the LWIR (wavelength 8 μm to 14 μm) and MWIR (wavelength 3 μm to 5 μm) range. In this paper the microbolometer, the vacuum-packaging, the architecture of the readout electronics, and the electro-optical performance characterization will be presented.

This paper presents the physical device modeling of a Si/Si_1-xGe_x multi-quantum well (MQW) detector to optimize the Ge content in the Si/Si_1-xGe_x well required to enhance thermal sensitivity for a potential microbolometer application. The modeling approach comprises a self-consistent coupled Poisson-Schroedinger solution in series with the thermionic emission theory at the Si/Si_1-xGe_x heterointerface and quantum confinement within the Si/Si_1-xGe_x MQW. The integrated simulation environment developed in Sentauruas WorkBench (SWB) TCAD is employed to investigate the transfer characteristics of the device consisting three stacks of Si/Si_1-xGe_x wells with an active area of 17μm x 17μm were investigated and compared with experiment data.

The presence of 1/f noise in staring imagers is problematic in several ways. It is well-known that in bolometers it sets a limit on the degree to which one can improve NETD by increasing the detector bias, because 1/f noise and responsivity both increase linearly with bias. Perhaps more importantly, at very low frequencies 1/f noise is manifest as drift and appears as spatial noise, which impairs target recognition far more than does an equivalent measure of temporal noise. This results from the fact that the eye filters temporal noise to a far greater extent than it does spatial noise. The perceived spatial noise resulting from 1/f noise leads to a degradation of performance with time after the insertion of a calibration shutter, even if the temperature is perfectly stable. Averaging many samples is not as effective for reduction of 1/f noise as it is for white noise, because the low frequency components are coherent over multiple samples. Additional implications of 1/f noise include uncertainty in measurement of signal and both temporal and spatial noise, as well as inherent errors in nonuniformity correction.

This paper reports the infrared spectral responses of 17 and 35 μm uncooled bolometers fabricated at INO. They are measured by making use of an external readout circuit along with a monochromator. As expected, the spectral absorption strongly depends on the bolometer stack as well as the pixel layout. By proper selection of design parameters, the spectral response can be made flat from 3 to 14 μm without significant deterioration of the detector figure of merit.

The state-of-the-art microbolometers are mainly based on polycrystalline or amorphous materials, typically Vanadium oxide (VO_x) or amorphous-Silicon (a-Si), which only have modest temperature sensitivities and noise characteristics. The properties of single crystalline SiGe/Si multi quantum wells (MQWs) have been proposed as a promising material¹. Particularly, SiGe/Si MQWs structure with high Ge concentration is expected to provide very high temperature coefficient of resistance (TCR) values between 6 to 8% ². Although SiGe/Si MQWs structure as a thermistor material is extremely promising, difficulty of defect free deposition and high sheet resistance of high Ge concentrated SiGe layers are the two main bottlenecks of this approach. In this work, a very high TCR of -5.5 %/K is achieved for SiGe/Si MQWs including 50% Ge with an acceptable noise value of 2.7 x 10^-13 V²/Hz at 10 Hz. The initial pixel resistance of 3 period of SiGe/Si MQWs with 50% Ge concentration is measured as 21 MΩ, which might not be compatible with the ROIC design. By the optimization of insitu Boron (B) doping level in SiGe layers of the MQW stack, 210 kΩ for 25 x 25 μm² pixel size is achieved. The optimized B doping density of ~1 x 10¹⁸ cm^-3 in SiGe wells did not cause any significant change in the TCR value whereas the 1/f noise performance is even enhanced due to the in-situ doping process and measured as 2.9 x 10^-14 V²/Hz at 10 Hz.

Pixel size reduction in an uncooled infrared detector plays a crucial role in determining significant attributes such as size, weight, and cost. Without a loss in sensitivity, however, it is challenging to develop an uncooled infrared detector that has the pixel pitch of 12 μm and below. Especially, there has been a limitation of an uncooled focal plane array (FPA) with a single-level design, which has to accomplish both functions of absorption and heat conduction in a single layer and thus compromise the sensitivity of FPA. A selective etching process of a resistive material, titanium oxide, on the connecting legs in the microbolometer FPA has been developed to overcome the limitation and increase the sensitivity of the single-level design FPA. Furthermore, structural stress modulation has been applied to ensure the mechanical robustness of the developed FPA. Here, we present a 12 μm single-level design FPA using a titanium oxide as a resistive material of microbolometer FPA. High-resolution detectors with an array size of 640 x 480 pixels (VGA format) and 1024 x 768 pixels (XGA format) have been developed, and the noise equivalent temperature difference (NETD) and time constant for the VGA detector are 40.5 mK and 8.3 msec, respectively. Thermal image obtained by the 12 μm XGA detector shows excellent image quality and fine resolution.

In the handheld optronic device Fusionsight, a smart digital fusion is employed. FusionSight is a night vision device that includes visible and thermal sensors. Our smart digital fusion proceeds in a frequency decomposition of both images coming from the visible and the infrared sensors, in order to obtain N frequency components for both images. The Ith frequency component of the visible image and the Ith frequency of the infrared image are then weighted with an Ith specific coefficient, and then recombined together in order to obtain a synthetic image with a I frequency component. All the synthetic images obtained likewise are then recombined together in order to obtain a fused image.

This article describes new imaging capabilities and technologies developed for infrared focal plane arrays (FPAs) at SCD. One of the new technologies is the patterning of the back surface of the FPA, whose front surface is bonded to a silicon readout integrated circuit (ROIC). Another is the hybridization of a spectral filter to the same back surface.

Increased image resolution has been achieved by using an opaque mask on the backside of the FPA with small central apertures. The reduced fill factor of the sensor leads to lower crosstalk between neighboring pixels and a higher Nyquist frequency. A highly detailed multi-mega pixel image is obtained when the sensor is micro-scanned relative to the imaging optics.

Spectral filtering was achieved by hybridization of a designated filter to the backside of the FPA. The filter was glued to the FPA with high accuracy achieving single pixel resolution. System implementation of these SWIR sensor cameras has been demonstrated at imec and is reported in this paper.

First results are reported for a continuously varying monolithic filter deposited onto the FPA, which has a high spectral dispersion. We report electro-optical measurements on several different sensors and describe some of their key parameters.

In this paper, we propose a data-driven proposal and deep-learning based classification scheme for small target detection. Previous studies have shown feasible performance using conventional computer vision techniques such as spatial and temporal filters. However, those are handcrafted approaches and are not optimized due to the nature of the application fields. Recently, deep learning has shown excellent performance for many computer vision problems, which motivates the deep learning-based small target detection. The proposed data-driven proposal and convolutional neural network (DDP-CNN) can generate possible target locations through the data-driven proposal and final targets are recognized through the classification network. According to the experimental results using drone database, the DDP-CNN shows 91% of train accuracy and 0.85 of average precision (AP) of the target detection.

Thermal imaging is widely used in military applications. The current sensor technology updates make it possible to use on commercial products as well. The quality of the image is significantly important for the final product performance in the field. The main parameters that affect the quality are the sensor resolution, the optical performance and the image processing capability at the final stage. Even though improving final system performance requires development of the optical and the sensor systems, such hardware developments may become a challenging issue for many cases. At that time, FPGA based image processing on the final stage becomes crucial. The noise reduction, the contrast correction and the image optimization can be controlled by the FPGA based image processing techniques. In this study, possible techniques to remove noise from a handheld thermal imaging system via FPGA implementation will be mentioned. Additionally experimental results, regarding the image performance improvement and the source usage will be explained.

In order to satisfy the requirements of short-wave infrared hyperspectral detection, we developed a 1024 x 512 ROIC with 30μm pixel pitch. CTIA with cascode amplifier was utilized as input stage and CDS was used for eliminating KTC noise and 1/f noise in CTIA. For this large chip with sizes up to 30mm x 20mm, it has been found that column circuit was a major bottleneck to achieve high frame frequency. A solution to solve this problem in this work is to pre-establish the signal of the column amplifier and then buffer odd and even column signals to the bus alternatively. In addition, parasitic capacitance of column-level bus was carefully lowered in layout design. The total readout rate reached 120 Mpixels/s with eight parallel output channels which allowed for a frame rate of 250 Hz.

Small pixel, high density arrays have many advantages in terms of SWaP-C and detection range. However, it is a challenge for quantum well infrared photodetectors to make into small pixels. The typical grating on the detector needs a large area to be effective. Recently, we introduced the resonator-QWIP for light coupling. This structure utilizes the active absorption volume as a resonator to trap the incident light until it is absorbed. To determine the size limit of this approach, we optimized the detector at different pixel pitches p (= 30, 12, 6, 3 and 2 microns) using 3-dimensional electromagnetic modeling. We found that their quantum efficiency can be kept relatively constant, and an especially large QE of ~80% appears at p = 3 microns at the wavelength of 9.0 microns for an absorption coefficient of 0.2/micron, indicating a great potential for pixel miniaturization. We conducted experiments on test detectors with p = 30, 12 and 6 microns. The set of wafers have two different active layer thicknesses and three different doping densities to create different detector characteristics. The experimental result is in good agreement with the prediction. We are producing 12-μm and 6-μm pitch detector arrays to confirm these test results. The FPAs will have peak wavelengths at either 8.0 or 9.8 microns, all hybridized to 1280x1024, 12-μm pitch ROICs.

Lately IRnova's research activities have been focused on the development of its next generation Type II Superlattice (T2SL) and Quantum Well Infrared Photodetector (QWIP) based infrared focal plane arrays (FPA) and Integrated Detector Dewar Cooler Assembly (IDDCAs) with 640x512 pixels @ 15μm pixel pitch. Last year we presented the initial results obtained for both of the above mentioned technologies at FPA and single element level. In this paper we will introduce our next generation of fully functional IDDCA (@15μm pixel pitch) solutions for MWIR and LWIR detection based on T2SL and QWIP technology respectively. Novelty of these IDDCAs lies in that fact that both of these products make use of FLIR indigo's ISC0403, and similar Cooler and Dewar assemblies. Performance in terms of picture quality, operability, response uniformity, stability and NETD of these IDDCAs is evaluated using demonstrator cameras developed in-house. This QWIP based LWIR IDDCA is the smallest pixel pitch commercially available IDDCA using this technology.

II-VI colloidal quantum dots (CQDs) have made significant technological advances over the past several years, including the world’s first demonstration of MWIR imaging using CQD-based focal plane arrays. The ultra-low costs associated with synthesis and device fabrication, as well as compatibility with wafer-level focal plane array fabrication, make CQDs a very promising infrared sensing technology. In addition to the benefit of cost, CQD infrared imagers are photon detectors, capable of high performance and fast response at elevated operating temperatures. By adjusting the colloidal synthesis, II-VI CQD photodetectors have demonstrated photoresponse from SWIR through LWIR. We will discuss the synthesis and optoelectronic properties of HgTe CQD films, as well as our recent progress in the development of low cost infrared focal plane arrays and single element detectors fabricated using II-VI CQDs.

In this study, we report high temperature operation of infrared photodetector using p-i-p InAs/GaAs quantum dots. The ground state emission peak at 18 K from photoluminescence spectroscopy was measured at 986 nm. Single pixel detectors were fabricated and device characteristics like temperature dependent dark current, blackbody and spectral response were analyzed. The measured dark current density at 220 K with applied bias of 0.2 V was 2.48×10^-3 A/cm². The spectral response peak (2 μm) was observed in short wave-infrared (SWIR) region. We report an excellent SWIR detection characteristics at 220 K with a responsivity and specific detectivity of 3.81 A/W and 2.18×10¹⁰ cmHz^1/2/W, respectively. The spectral response peak was achieved till 250 K and blackbody signal was observed till 270 K.

A detailed analysis of dark current and noise dependence on capping thickness in vertically coupled quaternary (InAlGaAs) capped InAs/GaAs quantum dot infrared photodetectors is presented. We are investigating the effect of varying capping thickness on device performance with theoretical proposed model. 2.7 ML InAs dots were grown with a combination capping of quaternary InAlGaAs layer (30Å) and GaAs capping thickness varying from 90-180Å (coupled Device A-C) to 500Å (Uncoupled Device D). Photoluminescence (PL) measurement 8 K exhibited multimodal ground state emission peak for device A to C whereas single emission peak was observed from device D. The measured activation energies for dominant peaks using PL were 236.63 meV, 188.92 meV, 164.88 meV and 151.25 meV respectively. The theoretical model gave activation energies of 234.57 meV, 148 meV, 131.69 meV and 144.56 meV respectively, which are on par with experimental values. The theoretical model had two components: tunneling component and the thermionic component. The thermionic component has an exponential dependence on activation energy, electric field and fitting parameter β. Similarly, tunneling component was dependent on electric field and other fitting parameters. Minimum dark current density was observed in device B. Similar trends were observed for noise spectral density and photoconductive gain. For thin capped devices (A-C), maximum photocurrent with narrow spectral response peak around 7 μm was observed. Device A measured highest responsivity of 0.85 A/W while device B measured highest detectivity of 2.48 × 10¹⁰ Jones.

In the present work we are introducing heterogeneously coupled InAs stranski-krastanov and submonolayer quantum dot as an active material for quantum dot based infrared photodetector. Initially, we have optimized the basic SK on SML heterostructure. The thickness of the GaAs barrier layer is varied from 2.5 to 7.5 nm to tune the vertical coupling between seed SML and top SK QDs. PL and PLE response confirms the carrier tunneling between these heterogeneous QDs. The vertical alignment of SML and SK QDs is shown in Cross sectional TEM images. The sample with 7.5 nm barrier layer is incorporated into a N-I-N based quantum dot infrared photodetector, which shows broader spectral response than standard SK QD based IR detectors.

InGaAs detector for SWIR imaging is widely used for remote sensing, medical application, personal identification etc. To reduce the required power for various environmental condition, reducing dark current is crucial. The dark current of InGaAs detector is known to come from defects induced during the growth of wafers and the process to fabricate FPAs. Especially, when high temperature is applied for the diffusion of Zn to form p-type junction on n-type InP/InGaAs substrate, the diffusion barrier of Zn on the substrate experiences large expansion and add stress in the substrate. The induced stress will increase defects and increase dark current. In this work, to reduce the stress of the Zn diffusion barrier, balanced diffusion barrier with multiple layer is applied. By reducing the stress, the dark current density has reduced to below 10 nA/cm², which is suitable for low power operation.

We first reported on a process for on-axis InSb crystal growth in 2014. As we have further developed on-axis (111) crystal growth, we have observed and measured a new distinct regime of interface-controlled dopant segregation. This effect is usually overshadowed by the facet effect and the resulting order of magnitude step change in the carrier concentration profile. When this large step change is eliminated, another interface-controlled effect becomes measurable. We present experimental data showing the magnitude of this effect and the crystal growth techniques used to engineer the interface where this effect is uncovered. We also discuss the atomic scale growth mechanisms that explain it.

This work proves useful in predicting the range of mechanical and electronic properties of wafers cut from ingots that are grown on-axis. More specifically, by understanding the effect of the melt/solid growth interface on the physical properties on the crystal, growth conditions can be optimized to produce more electrically uniform wafers that minimize pixel-to-pixel variation in FPAs.

The infrared focal plane arrays detector is a multilayer structure which is mainly composed of detector chip, Si-ROIC, and fan out layer. In view of the different thermal expansion coefficients between the material layers, considerable thermal stress will be generated in this device among the cooling cycle which could lead to physical breakdown of the chip under extreme circumstances. Models of finite element analysis (FEA) were established to explore the thermal stress of HgCdTe infrared focal plane devices at low temperature. According to the characteristics of the expansion alloy, the two kinds of focal plane device structures were simulated: one is that with invar layer below the Al2O3 piece, the other is that kovar layer between the Si-ROIC and Al2O3 pieces. Both of them can reduce the thermal stress effectively, and improve the reliability of IRFPAs detector.

As a narrow bandgap semiconductor, the preparation of surface passivation layers on HgCdTe film epilayers is essential in the process of device fabrication. Most new infrared detectors use the mesa structure. A stable and reproducible passivation technology which meets the surface uniform cover of the high aspect ratio mesa is particularly important. Atomic layer deposition (ALD) is a new type of accurate surface thin film preparation technique, which has several characteristics such as depositing large-area uniform films, making the film thickness control at nanometer level feasible, and lower deposition temperature. ALD-ZnS film is prepared on the HgCdTe IRFPAs chip at 65°. I-V and R-V curves are similar to that of IRFPAs with CdTe thermal passivation. This shows that ALD ZnS film has a good potential application in the passivation of high aspect ratio mesa-array HgCdTe devices.

We develop a shutter-less method for replacing mechanical shutters. To verify the effectiveness of the proposed method, we fabricated a silicon-on-insulator (SOI) diode uncooled 320 × 240 infrared focal plane array (IRFPA) with 17 μm pixel pitch utilizing a circuit architecture that achieves thermo-electric cooling (TEC)-less operation. Furthermore, we fabricated a prototype uncooled IR camera that implements the proposed method and verified favorable camera operation. The temperature behavior of our proposed SOI diode is highly uniform and predictable, which enables simpler device modeling and consequently simpler TEC-less and shutter-less operation.

A readout IC (ROIC) designed for high temperature coefficient of resistance (TCR) SiGe microbolometers is presented. The ROIC is designed for higher Ge content SiGe microbolometers which have higher detector resistance (~1M Ω) and higher TCR values (~%5.5/K). The ROIC includes column SAR ADCs for on-chip column-parallel analog to digital conversion. SAR ADC architecture is chosen to reduce the overall power consumption. The problem of resistance variation across the bolometers which introduce fixed pattern noise is addressed by setting a tunable reference resistor shared for each column which can be calibrated offline to set the common-mode level. Moreover, column non-uniformity has been reduced through comparator offset compensation in the SAR ADC. The columnwise architecture in this work reduces the number of integrators needed in the architecture and enables 17x17 μm2 pixel sizes. The prototype has been designed and fabricated in 0.25-μm CMOS process.

In this work, we demonstrate two times enhancement in NEDT response of InAs/InGaAs/ GaAs dot-in-the-well (DWELL) structure implanted with hydrogen ions of 3 MeV energy and 5×10¹² ions/cm² fluence. Low temperature photoluminescence study shows emission peak at 1130 nm which corresponds to ground state transition between conduction to valence band. PL enhancement is observed in all hydrogen implanted samples indicating the passivation of defects/dislocations in the vicinity of QDs and surrounding layer. The spectral response peak was observed at 5 μm corresponding to the intersubband transition and measured upto 76 K. Next, 320×256 format infrared FPA was fabricated involving multistage lithography, wet etching, metal stack and bump deposition. The NEDT parameter, which represents minimum temperature difference that FPA based camera can resolve, improved from 239 mK( as-grown) to 129 mK ( implanted sample).