Search

We report a novel design for bolometric infrared detector operating at the long-wave infrared (LWIR) range for high-resolution human body temperature sensing and monitoring. We propose to incorporate efficient plasmonic absorber and high-responsivity VO2 nanobeam biased at transition temperature in our design to facilitate improvement in both aspects – thermal resolution and spatial resolution. The integration of plasmonic absorber allows efficient and selective radiation absorption. The use of a transducing nanobeam placed in close proximity to the plasmonic local heaters allows the radiation energy to be electronically readable. The use of nanobeam instead of a film in the detector allows over two-orders of magnitude improvement in responsivity owing to the large length to cross-sectional area ratio. Additionally, the use of nanobeam reduces the thermal mass and improves the bandwidth. Our calculation suggests a responsivity of as high as 700 kV/W at 100 Hz for a detector of only $12\,\,\mu \text{m}\,\,\times 12\,\,\mu \text{m}$ pixel size. Also, the theoretical noise equivalent temperature difference is calculated to be as low as 3.67 mK which is almost an order of improvement than the state-of-the-art bolometric LWIR detectors with larger pixel dimensions.

For improving the dynamic range and sensitivity at high power load, we have integrated superconductor-insulator-normal metal-insulator-superconductor (SINIS) bolometers with a frequency selective surface (FSS)-based distributed absorber formed by a series and parallel array consisting of 25 annular antenna elements, each containing two SINIS bolometers. By using a design with 50 bolometers, we reduce incident power load on each bolometer, increase sensitivity and saturation power which is important for ground-based and balloon-borne telescopes with high background power loads. Our main detector pixel is optimized for a frequency band centered at 345GHz. The detectors are matched to incoming telescope beam by a back-to-back horn with a back reflector. Such a configuration improves both the efficiency and the bandwidth of the receiver. Measured voltage responsivity approaches 210(9) VW with an amplifier-limited voltage noise of 20nVHz(12), which corresponds to a NEP 10(-17) WHz(12). The linear voltage response for incoming power is observed for absorbed power of about 5 pW. The current responsivity for parallel array is 210(4) AW and the shot noise limited intrinsic noise equivalent power is NEP 510(-18)WHz(12). The noise equivalent temperature difference is NETD 100 KHz(12) at 2.7-K background radiation temperature.

This paper presents a novel and power efficient readout architecture for uncooled microbolometers that offers adequate noise performance along with a circuit-based method for self-heating compensation. The proposed method employs an event-generation scheme and a time-mode readout chain to achieve dynamic mode of operation resulting in low power. The readout chain consists of a capacitive transimpedance amplifier (CTIA) based current-to-time converter followed by a two stage time-to-digital converter (TDC). The CTIA employs a novel integration scheme for time amplification along with a modified reset mechanism to achieve self-heating compensation. We also present a model to analyze the implications of time-mode readout on imager operation. An experimental readout chip, based on this approach, has been designed using a 130 nm bulk CMOS technology. The proposed architecture offers robust frontend processing and achieves a per channel power consumption of $66~\mu \text{W}$ , which is considerably lower than the most recently reported designs, while maintaining better than 10-mK readout noise equivalent temperature difference (NETD).

This letter presents a device for biasing bolometric infrared imaging sensors to increase the operating temperature range and enhance performance. It addresses the market’s need for uncooled readout circuits that are easier to use. Here, we propose an integrated biasing system capable of managing the bolometric sensor’s high-temperature dependence. We demonstrated that a biasing architecture using thermalized bolometers can automatically offset resistance variations in the active bolometer due to temperature variations of the focal plane array (FPA). Our research concerns a 12- $\mu \text{m}$ pixel pitch QVGA matrix that operates at 50-mK noise equivalent temperature difference (NETD).

We systematically investigate the characteristics of cytochrome c protein, whose absorption rate is 65% at wavelengths of $10~\mu \text{m}$ . We design a chip for an infrared sensor with suspended aluminum electrodes. The protein solution is deposited on sensing pixels using an inkjet printer. The temperature coefficient of resistance, responsivity, 1/f noise, and noise equivalent temperature difference are 29 %/K, 1.5*105 V/W, 5.88*10−5 V/Hz1/2 at 1.78 Hz and 4.29 mK, respectively. Cytochrome c protein thin film, possessing a high-temperature coefficient of resistance, is suitable for microbolometer design. We experimentally demonstrate integrating cytochrome c protein with a CMOS circuit as a sensing pixel for a long-wavelength infrared microbolometer at room temperature.

Mid-wavelength Infrared (MWIR) and long-wavelength infrared (LWIR) Barrier Infrared Detector (BIRD) based on Antimonides Type-II superlattice detector array material are hybridized to 640 × 512, 1024 × 1024 and 1344 × 784 pixels format read out integrated circuits. The focal plane arrays (FPAs) Noise Equivalent Temperature Difference (NEΔT), Quantum Efficiency (η), Dark Current Density (JD), and Modulation Transfer Function (MTF) performance were measured at operating temperatures. High performance 640 × 512 array format MWIR High Operating Temperature (HOT) FPA has typical NEΔT ∼ 19 mK, JD ∼ 6.2 × 10−7 A/cm2, and η ∼ 53% using 3–5 μm band pass filter at 115 K operating temperatures. Best performance 640 × 512 array format LWIR FPA has NEΔT ∼ 25 mK, JD ∼ 2.5 × 10−5 A/cm2 and η ∼ 41% using 8–9.4 μm band pass filter at 60 K operating temperatures. For MTF investigation, a novel direct one-to-one image approach of Line Spread Function (LSF) and Edge Spread Function (ESF) are patterned onto substrate-removed FPAs. The direct LSF and ESF image are produced by 1000 A° thick gold layer deposition. The direct image horizontal and vertical MTFs are compared with projected knife-edge ESF-derived MTFs and results indicate that direct image MTFs have higher values compared to projected knife-edge ESF-derived MTFs at a Nyquist frequency. However, direct image MTF requires no lens MTF correction. The only relevant requirement for the direct image MTF is for the deposited pattern and substrate to be extremely thinned in which the HOT-BIRD FPAs fulfilled for this investigation.

Improving heat insulation in microelectromechanical system based thermal-type infrared sensor is crucial to increase its performance. The concept of phonon engineering allows us to change material thermal conductivity by several orders of magnitudes, which can be a potential technique for thermal management of such microelectronic devices. Here, we introduced a phononic crystal structure in a prototype thermopile sensor to investigate the effectiveness of phonon engineering. We demonstrate that ultrafine phononic crystal structure, which was consisted of through-holes periodically arranged at 38 nm, resulted in ten-fold reduction in Si thermal conductivity. This led to improvement of the sensitivity of IR detection by a factor of ten. In addition, our phononic crystal could suppress thermal conductivity while maintaining the electrical conductivity, which enable us to increase the noise equivalent temperature difference by 5.6 times. The present study demonstrates great potential of phonon engineering for infrared sensor technology to reach for smaller pixel size and lower device cost.

We applied a simple method to estimate the Minimum Resolvable Temperature Difference (MRTD) of an LWIR and an MWIR camera. A so-called Siemens star, in our case a thin, black aluminum plate framing a circle that is missing (cut out) every other spoke, is mounted in front of a black body whose temperature is relatively close to room temperature. From short recordings of the black body and Siemens star both the Noise Equivalent Temperature Difference (NETD) and the Modulation Transfer Function (MTF) are extracted and a simple estimate of MRTD = NETD/MTF is obtained. The imaged Siemens star almost completely covers the focal plane array; hence, an MRTD curve for the whole array is obtained. We investigated the effect of Non-Uniformity Correction (NUC) and Bad-Pixel Removal (BPX), two often applied pre-processing techniques, on the MRTD estimate. We find that (1) BPX has only limited effect on the result; (2) NUC is required to obtain a good MTF; and (3) NUC is not a prerequisite to obtain a good NETD estimate, but this is contingent on having a proper segmentation tool or template available. Without a segmentation algorithm, NUC together with simple intensity thresholding provides a sufficiently good segmentation and accordingly a good estimate of NETD.

Phase-changing materials are promising due to their sharp temperature dependent characteristics and have high potential of being integrated in optical switching and sensing techniques. Among such materials, vanadium dioxide (VO2) is the most utilitarian because of its transition temperature being close to the room-temperature. VO2-based bolometers utilize the material’s large temperature coefficient of resistivity to detect infrared radiation. However, to achieve large sensitivity, the active radiation absorption area needs to be large enough that allows sufficient temperature buildup from incident radiation absorbed by VO2, thus requiring large pixel dimen- sion and degrading the spatial resolution of bolometric sensing. Moreover, the absorption by the VO2 material is not optimized for a specific frequency band in most of the applications. On the other hand, plasmonic nanos- tructures can be tuned and designed to selectively and efficiently absorb a specific band of the incident radiation for local heating and thermal imaging. In this work, we propose to incorporate plasmonic nanostructures with VO2 nanowires that amplifies the slope of impedance change due to the thermal variations to achieve a higher sensitivity. We present the numerical analysis of the mid-infrared electromagnetic radiation absorption by the proposed detector showing near-perfect absorption by the plasmonic absorbers. Besides, the thermal buildup and the nanowire resistance change is predicted for different substrate, as the substrate plays a big role in heat distribution. We show high sensitivity and ultra-low noise equivalent temperature difference (NEDT) by our novel bolometric detector. We also discuss the fabrication of the VO2.

This paper presents the design of a current-mode CMOS self-amplified imager operating in dark conditions, for thermal imaging, which provides an innovative solution for precision thermal contact mapping. Possible applications of this imager range from 3D CMOS integrated circuits to the study of in-vivo biological samples. It can provide a thermal map, static or dynamic, for the measurement of temperature microgradients. Some adaptations are required for the optimization of this self-amplified image sensor since it responds exclusively to the dark currents of the photodiodes throughout the array. The sensor is designed in a standard CMOS process and requires no post-processing steps. The optimized image sensor operates with integration times as low as one &mu, s and can achieve both SNR and dynamic range compatible to those of sensors available on the market, estimated as 87dB and 75dB, respectively, noise equivalent temperature difference can be as low as 10mK, and detection errors as low as &plusmn, 1%. Furthermore, under optimal conditions the self-amplification process enables a simple form of CDS, enhancing the overall sensor noise performance.

This letter presents the design, simulation, and fabrication of an uncooled infrared micro-bolometer focal plane array (MBFPA) for multiple digital correlated double sampling (MD-CDS) to reduce a large amount of fixed pattern noise (FPN). A hydrogenated amorphous silicon-based MBFPA is fabricated with $80\times 60$ active cell array and $80\times 9$ proposed warm cell array with a 35- $\mu \text{m}$ pitch. The proposed warm cell can be formed simply by using additional thermal paths on the active cell and it closely matches the properties of the active cell, including thermal conductance and capacitance, resistance, and structure. Averaged reference signals generated by the warm cell array during readout have the same FPN as the active cell without reaction from incident infrared. Experimental results show that FPN is 23% reduced by using the proposed warm cell array and the measured noise equivalent temperature difference referred to $f$ /1.0 is 103 mK after the MD-CDS.

Optically immersed bolometer IR detectors were fabricated using electron beam evaporated vanadium oxide as the sensing material. Spin-coated polyimide was used as medium to optically immerse the sensing element to the flat surface of a hemispherical germanium lens. This optical immersion layer also serves as the thermal impedance control layer and decides the performance of the devices in terms of responsivity and noise parameters. The devices were packaged in suitable electro-optical packages and the detector parameters were studied in detail. Thermal time constant varies from 0.57 to 6.0 ms and responsivity from 75 to 757 V W−1 corresponding to polyimide thickness in the range 2 to 70 μm for a detector bias of 9 V in the wavelength region of 14–16 μm. Highest D* obtained was 1.2×108 cmHz1/2 W−1. Noise equivalent temperature difference (NETD) of 20 mK was achieved for devices with polyimide thickness more than 32 μm. The figure of merit, NETD × τ product which describes trade-off between thermal time constant and sensitivity is also extensively studied for devices having different thickness of thermal impedance layers.

We present our latest results on cooled p-on-n planar mercury cadmium telluride (MCT) photodiode technology. Along with a reduction in dark current for raising the operating temperature (T op), AIM INFRAROT-MODULE GmbH (AIM) has devoted its development efforts to shrinking the pixel size. Both are essential requirements to meet the market demands for reduced size, weight and power and high-operating temperature applications. Detectors based on the p-on-n technology developed at AIM now span the spectrum from the mid-wavelength infrared (MWIR) to the very long wavelength infrared (VLWIR) with cut-off wavelengths from 5 μm to about 13.5 μm at 80 K. The development of the p-on-n technology for VLWIR as well as for MWIR is mainly implemented in a planar photodetector design with a 20-μm pixel pitch. For the VLWIR, dark currents significantly reduced as compared to ‘Tennant’s Rule 07’ are demonstrated for operating temperatures between 30 K and 100 K. This allows for the same dark current performance at a 20 K higher operating temperature than with previous AIM technology. For MWIR detectors with a 20-μm pitch, noise equivalent temperature differences of less than 30 mK are obtained up to 170 K. This technology has been transferred to our small pixel pitch high resolution (XGA) MWIR detector with 1024 × 768 pixels at a 10-μm pitch. Excellent performance at an operating temperature of 160 K is demonstrated.

XBn or XBp barrier detectors exhibit diffusion-limited dark currents comparable with mercury cadmium telluride Rule-07 and high quantum efficiencies. In 2011, SemiConductor Devices (SCD) introduced “HOT Pelican D”, a 640 × 512/15-μm pitch InAsSb/AlSbAs XBn mid-wave infrared (MWIR) detector with a 4.2-μm cut-off and an operating temperature of ∼150 K. Its low power (∼3 W), high pixel operability (>99.5%) and long mean time to failure make HOT Pelican D a highly reliable integrated detector-cooler product with a low size, weight and power. More recently, “HOT Hercules” was launched with a 1280 × 1024/15-μm format and similar advantages. A 3-megapixel, 10-μm pitch version (“HOT Blackbird”) is currently completing development. For long-wave infrared applications, SCD’s 640 × 512/15-μm pitch “Pelican-D LW” XBp type II superlattice (T2SL) detector has a ∼9.3-μm cut-off wavelength. The detector contains InAs/GaSb and InAs/AlSb T2SLs, and is fabricated into focal plane array (FPA) detectors using standard production processes including hybridization to a digital silicon read-out integrated circuit (ROIC), glue underfill and substrate thinning. The ROIC has been designed so that the complete detector closely follows the interfaces of SCD’s MWIR Pelican-D detector family. The Pelican-D LW FPA has a quantum efficiency of ∼50%, and operates at 77 K with a pixel operability of >99% and noise equivalent temperature difference of 13 mK at 30 Hz and F/2.7.

When a space infrared camera has a higher temperature sensitivity , the more tiny temperature information it can record in the imaging data. High sensitive infrared camera takes an important role in observation of urban and sea targets and so on. A new space infrared camera system is designed based on pixel-level digital accumulation technology. The new infrared camera based on this technology realize the analog-to-digital convert of detecting signals into single detector pixel, which can reduce the noise caused by readout of analog signal markedly. This infrared system processes input signal by little capacitor packet digital-counting method，the integral capacitor plays the part of “voltage-to-frequency conversion”, which vary far from the traditional detector. The signal quantization process realized by the capacitor packet digitalcounting, every single charge packet is small, a capacitor of several fF in new system has the equal effect of a capacitor of thousands pF in traditional system, that makes the new infrared system can break the limit of integral capacitor’s physical restriction. So the new camera has the ability of achieving 1mK(0.001K) temperature sensitivity by theory without saturation electrons restriction. This ultra-high temperature sensitivity infrared camera system will break the temperature sensitivity limits of all infrared cameras on hand, and improve infrared system’s ability of distinguishing the tiny temperature a lot. The prototype camera has been manufactured and it has a 5mK NETD (Noise Equivalent Temperature Difference), the engineering feasibility of this camera system is valid. The pixel-level digital accumulation infrared camera will be an important development direction of future high-sensitivity space infrared camera technique.

Utilizing a thermomechanical transduction scheme, we have developed an uncooled, sensitive, and fast THz bolometer by using a doubly clamped GaAs MEMS beam resonator as a sensitive thermistor. Owing to its ultra-high temperature sensitivity (the noise equivalent temperature difference of ~1 µK/ ✓ Hz), the present bolometer achieves not only a high sensitivity but also an operation bandwidth of several kHz, which is ~100 times faster than other uncooled THz thermal sensors. The MEMS bolometers are fabricated by the standard semiconductor fabrication processes and are well suited for making detector arrays for realizing THz cameras.

We present how to optimize efficiently the performance of micro-bolometer array with design parameters such as active area and channel length, and minimize simultaneously the channel resistance of bolometer array with a minimum channel area and length in a given technology and technology window. The results demonstrate that the serpentine is very efficient pattern to minimize the channel since it provides a large area or path through which electron or holes conducts as easy as possible, as well as it maintains a relatively large absorption area in a given pixel area. On the other hand, these approach have also much large effects on the flicker noise reduction because amorphous Si active layer has about 20 M ohm/ which causing an insufficient current level in an array circuits, however, any dopants does not need to reduce resistance of micro-bolometer channel.

Uncooled infrared detectors have enabled the rapid growth of thermal imaging applications. These detectors are predominantly bolometers, reading out a pixel’s temperature change due to infrared radiation as a resistance change. Another uncooled sensing method is to transduce the infrared radiation into the frequency shift of a mechanical resonator. We present here highly sensitive resonant infrared sensors, based on thermo-responsive shape memory polymers. By exploiting the phase-change polymer as transduction mechanism, our approach provides 2 orders of magnitude improvement of the temperature coefficient of frequency. Noise equivalent temperature difference of 22 mK in vacuum and 112 mK in air are obtained using f/2 optics. The noise equivalent temperature difference is further improved to 6 mK in vacuum by using high-Q silicon nitride membranes as substrates for the shape memory polymers. This high performance in air eliminates the need for vacuum packaging, paving a path towards flexible non-hermetically sealed infrared sensors., Though resonant infrared (IR) detectors are an attractive thermal imaging technology owing to its high performance potential, realizing devices with high sensitivity remains a challenge. Here, the authors report high-sensitivity resonant IR sensors based on thermo-responsive shape memory polymers.

This paper describes the design of readout integrated circuits (ROICs) for hybrid infrared focal plane arrays (IR FPAs). This work contains the estimation of the noise equivalent temperature difference (NETD) of IR FPAs based on frame and row integration of pixel signals in the spectral ranges of 8 to 14 and 3 to 5 μm. This paper also describes the development of ROICs for IR FPAs created with the use of mercury—cadmium—telluride (MCT) photodiodes and quantum well infrared photodetectors (QWIPs). The designed ROICs ensure the use of matrix and linear photodetector chips, including those with increased dark currents, in order to produce IR FPAs with temperature resolution corresponding to the world level of array analogs.

Results of measurement of the dark currents and photocurrents of p-on-n photodiodes (PD) formed in In-doped MBE-grown n-type mercury cadmium telluride (MCT) films are reported. We used 3 in. in diameter (0 1 3) GaAs and (0 1 3) Si substrates. The p-on-n junctions were fabricated using the implantation of As+ ions followed by the thermal annealing activation. It was shown that the regime of background-limited performance (BLIP) of the p-on-n PD with experimentally measured long-wave cutoff wavelength (LW) cutoff (λc) of 11.1 and 10.3 μm at 77 K is observed up to elevated temperatures of 105 K and 117 K, respectively. The temperature dependence of the noise equivalent temperature difference (NETD) for p-on-n PD sized 30 × 30 μm was calculated taking into account only PD shot noise. It was shown that such LW p-on-n photodiodes allows to increase the operating temperature of IR FPA detectors up to 100–105 K with the NETD less than 30 mK.

In this work, we leverage graphene's unique tunable Seebeck coefficient for the demonstration of a graphene-based thermal imaging system. By integrating graphene based photothermo-electric detectors with micromachined silicon nitride membranes, we are able to achieve room temperature responsivities on the order of ~7-9 V/W (at λ = 10.6 μm), with a time constant of ~23 ms. The large responsivities, due to the combination of thermal isolation and broadband infrared absorption from the underlying SiN membrane, have enabled detection as well as stand-off imaging of an incoherent blackbody target (300-500 K). By comparing the fundamental achievable performance of these graphene-based thermopiles with standard thermocouple materials, we extrapolate that graphene's high carrier mobility can enable improved performances with respect to two main figures of merit for infrared detectors: detectivity (>8 × 10(8) cm Hz(1/2) W(-1)) and noise equivalent temperature difference (

Optically immersed detectors with high index of refraction lenses were an effective technique to improve the performance of thermistor bolometer for infrared detection. In this paper we described the design and fabrication of back-incident optically immersed bolometer with germanium hemispherical lens using the Mn1.56Co0.96Ni0.48O4 thin films prepared by chemical solution deposition method on amorphous Al2O3 substrate. The characteristics of back-incident immersed bolometer operated at room temperature were investigated. It was found that the performance of the detector was significantly enhanced by the use of germanium hemispherical lens. The fabricated bolometer exhibited noise equivalent temperature of 1.56 × 10−7 K/Hz1/2, responsivity of 3.2 × 103 V/W, and detectivity of approximate 9.2 × 108 cmHz1/2/W at 30 Hz. The thermal time constant of 7.3 ms was about one third of the non-immersion ones. Field of view of ±35° and optical immersion gain of up to 14 times were also obtained. The results demonstrated that the feasibility of back-incident optically immersed bolometer could be used to uncooled infrared bolometric applications.

This paper presents the physical and technical principles of readout integrated circuit (ROIC) for reading and preprocessing focal plane array signals in the infrared spectral range 8–14 μm. The noise equivalent temperature difference of long-wavelength infrared focal plane arrays based on ROIC with frame integration of signals of multilayer quantum well structures is evaluated. The influence of technological limitations in silicon readout circuits for photo signals on the performance of IR focal plane arrays is analyze for a wide range of parameters of QWIP -based focal plane arrays and CMOS technology design rules.

In this study, we analyzed and showed the enhanced thermal resolution of a lock-in infraredthermography system by employing a blackbody system and micro-register sample. The noise level or thermalresolution of an infrared camera system is usually expressed by a noise equivalent temperature difference (NETD),which is the mean square of the deviation of the different values measured for one pixel from its mean valuesobtained in successive measurements. However, for lock-in thermography, a more convenient quantity in thephase-independent temperature modulation amplitude can be acquired. On the basis of results, it was observed thatthe NETD or thermal resolution of the lock-in thermography system was significantly enhanced, which we considerto have been caused by the averaging and filtering effects of the lock-in technique. Keywords: Lock-in Thermography, Noise Equivalent Temperature Difference, Infrared Microscope, Thermal Resolution[Received: September 23, 2014, Revised: October 20, 2014, Accepted: October 29, 2014] *한국기초과학지원연구원첨단장비개발사업단, ** IT , Corresponding Author: Center for Analytical한국과학기술원스마트 융합시스템연구단 ✝Instrumentation Development, Korea Basic Science Institute, 169-148 Gwahak-ro, Yuseong-gu, Daejeon 305-806, Korea(E-mail: ksc@kbsi.re.kr)

Recently, thermography camera have been using for body-temperature monitoring. We report on fabrication of prototype thermography camera using the chalcogenide-glass lens and the camera test by analysis of thermal image. In this work, it was found out that thermography camera discerned body-temperature between 20 and with noise equivalent temperature difference(NETD) of 87.7mK. It is confirmed that thermography camera using the chalcogenide-glass lens is applicable to the body-temperature monitoring system.

Spaceborne assessment of fire characteristics relies on radiance measurement of fire pixels and non-fire pixels mainly in the midwave infrared (MWIR). Because ambient temperature non-fire pixels have low thermal emission in this spectral range, it remains a challenge to retrieve fire characteristics with the desired accuracy. This paper reports on uncooled microbolometers specially designed with low noise equivalent power (NEP) to enable fire diagnosis at MWIR wavelengths. Each microbolometer forming a 512x3 format array includes a Wheatstone bridge of one active, one blind, and two thermally shunted pixels followed by its own signal chain. Design analyses suggest the conditions for achieving the best NEP performance are: (i) the active, blind, and one shunt pixel have equal electrical resistances while the other shunt pixel has a larger resistance; (ii) the temperature difference between the active pixel and heat sink corresponds to about one-third the heat sink temperature; and (iii) the active and blind pixels have low thermal mass and conductance. Hardwired devices having different structural layouts were prepared for the validation of physical parameters and performance so that the suitable designs could be identified. After this, focal planes of 512x3 microbolometers were fabricated on readout electronics to allow further performance evaluation and development of staggered 1017x3 format arrays for a planned mission. The active pixel designs on the fabricated arrays exhibit a MWIR absorptance as high as 0.83 through implementation of a Salisbury screen absorber, a thermal conductance of ~ 67 nW/K, and a response time shorter than 10 ms. Their responsivities are found to be in good agreement with predictions of the design analysis. The effectiveness of an Al shield platform erected above the blind pixel was investigated, showing that certain designs are capable of attenuating the incident power by up to 24 times. Under optimal operating conditions an NEP of ~ 64 pW was derived from measurements in the spectral range of 3.4-4.2 um, which was corroborated by the probing results obtained on the on-wafer focal plane arrays. When using these arrays with an F/1 telescope to retrieve scenes of 400 K from low Earth orbits, a noise equivalent temperature difference of ~ 320 mK can be achieved.

The Visible Infrared Imaging Radiometer Suite (VIIRS) sensor aboard the Suomi National Polar-orbiting Partnership spacecraft has successfully operated since its launch in October 2011. The VIIRS collects data in 22 spectral bands that are calibrated by a set of onboard calibrators (OBC). In addition, lunar observations are made to independently track VIIRS long-term calibration stability for the reflective solar bands (RSB). This paper provides an overview of VIIRS OBC functions as well as its on-orbit operation and calibration activities. It also describes sensor calibration methodologies and demonstrates VIIRS on-orbit performance from launch to present. Results reported in this paper include on-orbit changes in sensor spectral band responses, detector noise characterization, and key calibration parameters. Issues identified and their potential impacts on sensor calibration are also discussed. Since launch, the VIIRS instrument nominal operation temperature has been stable to within ±1.0 K. The cold focal plane temperatures have been well controlled, with variations of less than 20 mK over a period of 1.5 years. In general, changes in thermal emissive bands (TEB) detector responses have been less than 0.5%. Despite large response degradation in several near-infrared and short-wave infrared bands and large SD degradation at short visible wavelengths, the VIIRS sensor and OBC overall performance has been excellent postlaunch. The degradation caused by the telescope mirror coating contamination has been modeled and its impact addressed through the use of modulated relative spectral response in the improved calibration and the current sensor data record data production. Based on current instrument characteristics and performance, it is expected that the VIIRS calibration will continue to meet its design requirements, including RSB detector signal to noise ratio and TEB detector noise equivalent temperature difference, throughout its 7 year design lifetime.

This paper presents a novel and compact compensation architecture for uncooled microbolometers without substrate temperature stabilization. The substrate temperature-induced nonuniformity and nonlinear resistance characteristics of the microbolometer array elements are investigated in detail. We propose an analytical model to identify the main nonideal effects and dispersion sources caused by the substrate temperature and their respective impacts on the uncooled microbolometers array. This proposed compensation approach is based on a differential current mirror readout architecture, which allows pixel response and offset correction to be performed as a function of the uncooled infrared focal plane array (IRFPA) sensor's operating temperature. An experimental FPA chip based on this approach has been designed and implemented on silicon using a 0.5- μm CMOS technology and surface micromachining process. The proposed architecture allows the microbolometer array nonuniformity control over a wider range of readout integrated circuits substrate temperatures while maintaining better than 63-mK noise equivalent temperature difference with f/1 optics. As a result of the elimination of the temperature stabilization requirement, the IRFPA turn-on time for high performance imaging can be greatly reduced. Power, volume, and weight of the IRFPA are also significantly reduced. This approach is ideally suited for high-volume, low-cost, low-power, and low-weight production applications.

Terahertz (THz) electromagnetic spectrum draws wide attention for nondestructive and/or biocompatible sensing. In order to be widely applicable to the THz sensing, it is of prime importance to develop THz sensors that can be operated at room temperature and have high sensitivity and fast operation speed. However, conventional room-temperature THz thermal sensors fall short of expectations in these characteristics required in various applications of THz sensing, including THz cameras. Utilizing a thermomechanical transduction scheme, we have developed an uncooled, sensitive, and fast THz bolometer by using a doubly clamped GaAs microelectromechanical system (MEMS) beam resonator as a sensitive thermistor. Owing to its ultrahigh temperature sensitivity (the noise equivalent temperature difference of ∼1 μK/√Hz), the present bolometer achieves not only high sensitivity but also an operation bandwidth of several kHz, which is more than 100 times faster than other uncooled THz thermal sensors. The obtained electrical noise equivalent power is as low as ∼90 pW/√Hz, which is close to the limit set by the thermal fluctuation noise. The MEMS bolometers are fabricated by the standard semiconductor fabrication processes and are well suited for making detector arrays for realizing THz cameras.

This paper presents a design model for uncooled microbolometers readout integrated circuits (ROIC). The relation between infrared focal-plane array (FPA) performance and ROIC design specification limits is shown distinctly. A simple way to find optimal operating conditions for a certain ROIC construction is also described. The signal and noise model of the ROIC is investigated in detail. A reasonable estimate of noise equivalent temperature differences for uncooled microbolometer detectors can be calculated with this model. A capacitor feedback transimpedance amplifier readout circuit is treated as an example. An experimental readout chip based on this model has been designed and implemented on silicon using a 0.5-μm CMOS technology. The temporal root mean square output noise voltage is 424 μV, which corresponds well with the measurement result. The result demonstrates the reliable prediction of achievable improvements for FPA by using this model.

This paper reports on advances in the electro-optical characterization of InAs/GaSb short-period superlattice infrared photodetectors with cut-off wavelengths in the mid-wavelength and long-wavelength infrared ranges. To facilitate in-line monitoring of the electro-optical device performance at different processing stages we have integrated a semi-automated cryogenic wafer prober in our process line. The prober is configured for measuring current-voltage characteristics of individual photodiodes at 77 K. We employ it to compile a spatial map of the dark current density of a superlattice sample with a cut-off wavelength around 5 μm patterned into a regular array of 1760 quadratic mesa diodes with a pitch of 370 μm and side lengths varying from 60 to 350 μm. The different perimeter-to-area ratios make it possible to separate bulk current from sidewall current contributions. We find a sidewall contribution to the dark current of 1.2×10-11 A/cm and a corrected bulk dark current density of 1.1×10-7 A/cm2, both at 200 mV reverse bias voltage. An automated data analysis framework can extract bulk and sidewall current contributions for various subsets of the test device grid. With a suitable periodic arrangement of test diode sizes, the spatial distribution of the individual contributions can thus be investigated. We found a relatively homogeneous distribution of both bulk dark current density and sidewall current contribution across the sample. With the help of an improved capacitance-voltage measurement setup developed to complement this technique a residual carrier concentration of 1.3×1015 cm-3 is obtained. The work is motivated by research into high performance superlattice array sensors with demanding processing requirements. A novel long-wavelength infrared imager based on a heterojunction concept is presented as an example for this work. It achieves a noise equivalent temperature difference below 30 mK for realistic operating conditions.

We have developed a passive 350 GHz (850 {\mu}m) video-camera to demonstrate lumped element kinetic inductance detectors (LEKIDs) -- designed originally for far-infrared astronomy -- as an option for general purpose terrestrial terahertz imaging applications. The camera currently operates at a quasi-video frame rate of 2 Hz with a noise equivalent temperature difference per frame of $\sim$0.1 K, which is close to the background limit. The 152 element superconducting LEKID array is fabricated from a simple 40 nm aluminum film on a silicon dielectric substrate and is read out through a single microwave feedline with a cryogenic low noise amplifier and room temperature frequency domain multiplexing electronics.

Currently, hardware Trojans have posed a serious threat to the integrated circuit security. A novel approach using chips’ infrared radiation to detect Trojans on a second scale was proposed in 2014. However, the temperature differences can be distinguishable on a microsecond scale between the normal areas in normal chips and the corresponding infected areas in chips with Trojans. As a result, the second-level detection can influence the detection accuracy reversely because of the temperature balance. On the other hand, infrared cameras’ ability to detect Trojans is determined by three parameters. They are the noise equivalent temperature difference (NETD), the pixel size and the frame frequency. It will be of benefit to Trojans detection using infrared cameras, if we determine the influences of the three parameters on detection. In this paper, we utilize finite element analysis to simulate the microsecond-level temperature variations of a fixed pixel-size silicon substrate while logic circuits on this size silicon substrate vary and operate under different challenges. Then, we find that the distinguishable time between different cases is on a microsecond scale according to a normal NETD. Based on our simulation results, an increasing step size (ISS) approach is proposed to capture dies’ microsecond-level infrared maps accurately using the low frame frequency infrared cameras. Finally, we analyze the temperature variations while a fixed logic circuit under a fixed challenge is operating on the different pixel-size silicon substrates. Based on the results, we get the link between NETD and the pixel size on the Trojan detection.

著者人数: 14名, Accepted: 2011-06-09, 資料番号: SA1003050000

In this paper, we considered the issues related to modeling and analysis of uncooled infrared detectors noise analysis utilizing pulse train equalization technique. We first developed a mathematical model appropriate for the noise performance study of the detector and then derive the detectivity and noise equivalent temperature difference (NETD) expressions for the predetection, detection, and postdetection noise sources. Our study of the postdetection noise sources was focused on the capacitive transimpedance amplifier unit cell electronics which provides a tight control over the detector bias current and handles a lower background radiation with a superior noise performance. We developed a parametric tradeoff study which provides a guide for optimizing the bolometer noise performance through the adjustment of detector parameters. We also discussed the use of scanning format and time-delay integration technique for improving the performance of multiple arrays of detectors.

This paper describes the design and modeling of a smart uncooled infrared detector with wavelength selectivity in the long-wavelength infrared (LWIR) band. The objective is to enhance the probability of detecting and identifying objects in a scene. This design takes advantage of the smart properties of vanadium dioxide (VO2): it can switch reversibly from an IR-transparent to an IR-opaque thin film when properly triggered. This optical behavior is exploited here as a smart mirror that can modify the depth of the resonant cavity between the suspended thermistor material and a patterned mirror on the substrate, thereby altering wavelength sensitivity. The thermistor material used in the simulation is vanadium oxide (VOx). The simulation results show that, when VO2 is used in the metallic phase, it reflects IR radiation back to the suspended VOx and enhances IR absorption in the 9.4-10.8-μm band. When the film is switched to the semiconductor phase, it admits most IR radiation, which is then reflected back to the suspended VOχ by a patterned gold thin film under an SiO2 spacer layer. The spacer layer is used to increase the resonant cavity depth underneath the microbolometer pixel. Thus, the peak absorption value is shifted to 8-9.4 μm, creating the second spectral band. The detector is designed with a relatively low thermal conductance of 1.71 X 10-7 W/K to maximize responsivity (Rv) to values as high as 1.27 X 105 W/K and detectivity (D*) to as high as 1.62 x 109 cm-Hz1/2/W, both at 60 Hz. The corresponding thermal time constant is equal to 2.45 ms. Hence, these detectors could be used for 60-Hz frame rate applications. The extrapolated noise equivalent temperature difference is 14 and 16 mK for the 8-9.4- and 9.4-10.8-μm bands, respectively. The calculated absorption coefficients in the two spectral bands were 59% and 65%, respectively.

As an approach for analyzing an uncooled infrared bolometer imager, a method is introduced which supposes that a thermal system is built into the infrared radiation detection mechanism. In this analysis, the bi-directional radiation energy flows between two thermal elements are considered and a radiation thermal conductance between them is defined. Based on the values of the radiation thermal conductances and conventional material thermal conductance, an equivalent thermal circuit for a bolometer detector element is extracted. Calculations were carried out using software written in visual C++. The analysis shows a Noise Equivalent Temperature Difference (NETD) of 50 mK is obtained as a typical value, assuming that no additional noise other than temperature fluctuation noise exists, and that the NETD becomes better as the system time constant built into the equivalent thermal circuit increases.

Third generation thermal imagers with dual/multi-band operation capability are the prominent focus of the current research in the field of infrared detection. Dual band quantum-well infrared photodetector (QWIP) focal plane arrays (FPAs) based on various detection and fabrication approaches have been reported. One of these approaches is the three-contact design allowing simultaneous integration of the signals in both bands. However, this approach requires three In bumps on each pixel leading to a complicated fabrication process and lower fill factor. If the spectral response of a two-stack QWIP structure can effectively be shifted between two spectral bands with the applied bias, dual band sensors can be implemented with the conventional FPA fabrication process requiring only one In bump on each pixel making it possible to fabricate large format dual band FPAs at the cost and yield of single band detectors. While some disadvantages of this technique have been discussed in the literature, the detailed assessment of this approach has not been performed at the FPA level yet. We report the characteristics of a large format (640 × 512) voltage tunable dual-band QWIP FPA constructed through series connection of MWIR AlGaAs–InGaAs and LWIR AlGaAs–GaAs multi-quantum well stacks, and provide a detailed assessment of the potential of this approach at both pixel and FPA levels. The dual band FPA having MWIR and LWIR cut-off wavelengths of 5.1 and 8.9 μm provided noise equivalent temperature differences as low as 14 and 31 mK ( f /1.5) with switching voltages within the limits applicable by commercial read-out integrated circuits. The results demonstrate the promise of the approach for achieving large format low cost dual band FPAs.

Thermo-mechanical detector arrays in the 160 x 120 array format are designed, fabricated, and tested. Detectors are composed of SiNx and TiN as absorbing layer, Al to form bimaterial legs, and integrated diffraction grating interferometer underneath each detector. The detector array is a passive component and the optical readout is performed remotely with a laser and CCD camera. All noise sources are considered in making a detailed noise equivalent temperature difference (NETD) estimation, which revealed 12-bit CCD camera for the readout. (C) 2009 Elsevier B.V. All rights reserved.

This paper reports the design, fabrication technology, post-CMOS micromachining and characterization of CMOS-silicon-on-insulator (SOI)-microelectromechanical system (MEMS) transistors. The thermally isolated micromachined CMOS-SOI-MEMS transistor reported here is designed for uncooled infrared (IR) sensing and is dubbed here as ldquoTMOS.rdquo The measured dc and noise electrical characteristics of the as-processed (virgin) transistor as well as those of the post-CMOS-MEMS-processed transistor (TMOS) are reported and compared. In particular, the threshold voltage temperature dependence and the temperature coefficient of current (TCC) at subthreshold are reported. The results indicate that the post-CMOS-MEMS processing does not degrade the performance of the transistors. The electrooptical performance of the TMOS is characterized and reported. With TCC on the order of 4%-10%, depending on the gate voltage, responsivity of 40 mA/W, noise equivalent power on the order of several tens of picowatts, and calculated noise equivalent temperature difference on the order of 64 mK, this uncooled IR sensor in standard CMOS-SOI technology may provide a high performance at a lower cost compared to state-of-the-art uncooled sensors based on bolometers implemented in non-CMOS materials like vanadium oxide or amorphous silicon.

Low manufacturing cost and ease of use favor spreading of 2-D bolometric infrared detector arrays over various application domains such as predictive maintenance, medical imaging, automotive industry, and security. The infrared detector's main figure of merit has long been the noise equivalent temperature difference (NETD), which sets the minimum temperature difference distinguishable from background noise at sensor output. However, while nowadays uncooled detectors have achieved sufficient NETD, fixed pattern noise (FPN) is indeed becoming a crucial figure of merit especially when the focal plane array (FPA) is not regulated by a thermoelectric cooler (TEC). In this paper, we study the various sources of dispersion of infrared bolometric detectors and their respective impact on FPN in the resulting image. We propose an analytical model to identify main sources of nonuniformity, and the confrontation of results with actual measurements leads to the ability of a highly accurate on-chip thermal drift compensation.

Extrinsic p-type doping during molecular-beam epitaxy (MBE) growth represents an essential generic toolbox for advanced heterostructures based on the HgCdTe material system: PiN diodes, mesa avalanche photodiodes (APD) or third-generation multispectral focal-plane arrays. Today, arsenic appears to be the best candidate to fulfill this role and our group is actively working on its incorporation during MBE growth, using an original radio frequency (RF) plasma source for arsenic. Such a cell is supposed to deliver a monatomic As flux, and as expected we observed high As electrical activation rates after annealing short-wave (SW), mid-wave (MW), and long-wave (LW) layers. At last, a couple of technological runs have been carried out in the MW range in order to validate the approach on practical devices. p-on-n focal-plane arrays (FPA) have been fabricated using a mesa delineated technology on an As-on-In doped metallurgical heterojunction layer grown on a lattice-matched CdZnTe layer (320 × 256, 30 μm pitch, 5 μm cutoff at 77 K). Observed diodes exhibit very interesting electro-optical characteristics: large shunt impedance, high quantum efficiency, and no noticeable excess noise. The resulting focal-plane arrays were observed to be very uniform, leading to high operabilities. Noise equivalent temperature difference (NETD) distributions are very similar to those observed with the As ion-implanted p-on-n technology, fabricated in our laboratory as well. In our opinion, those excellent results demonstrate the feasibility of our MBE in situ arsenic doping process. Good electrical activation rates and high-quality layers can be obtained. We believe that such an approach allows precise control of the p-doping profile in the HgCdTe layer, which is necessary for advanced structure designs.

This paper reports the development of an InSb photovoltaic infrared sensor (InSb PVS) operating at room temperature for temperature measurements. The InSb PVS consists of 910 InSb p + –p − –n + photodiodes connected in series on a semi-insulating GaAs (1 0 0) substrate. An Al 0.17 In 0.83 Sb barrier layer was grown between the p + and p − layers to reduce the diffusion of photo-excited electrons. As the InSb PVS operates in a photovoltaic mode, no thermal insulation is required, enabling its miniaturized plastic molding package. The sensitivity of the InSb PVS was 127 μV/K, and a noise equivalent temperature difference (NETD) of 1.0 mK/Hz 1/2 was obtained at room temperature. The results demonstrate the potential for the sensor to be used both in non-contact thermometry, as well as human body detection.

The recent introduction of a M-structure design improved both the dark current and R0 A performances of type-II InAs-GaSb photodiodes. A focal plane array fabricated with this design was characterized at 81 K. The dark current of individual pixels was measured between 1.1 and 1.6 nA, 7 times lower than previous superlattice FPAs. This led to a higher dynamic range and longer integration times. The quantum efficiency of detectors without antireflective coating was 74%. The noise equivalent temperature difference reached 23 mK, limited only by the performance of the testing system and the read out integrated circuit. Background limited performances were demonstrated at 81 K for a 300 K background.

This paper discusses the implementation of a low-cost 128 times 128 uncooled infrared microbolometer detector array together with its integrated readout circuit (ROC) using a standard 0.35 mum n-well CMOS and post-CMOS MEMS processes. The detector array can be created with simple bulk-micromachining processes after the CMOS fabrication, without the need for any complicated lithography or deposition steps. The array detectors are based on suspended p+-active/n-well diode microbolometers with a pixel size of 40 mum times 40 mum and a fill factor of 44%. The p+-active/n-well diode detector has a measured dc responsivity (R) of 4970 V/W and a thermal time constant of 36 ms at 50 mtorr vacuum level. The total measured rms noise of the diode type detector is 0.69 muV for an 8 kHz bandwidth, resulting in a detectivity (D*) of 9.7 times 108 cm ldr Hz1/2/W. The array is scanned by an integrated 32-channel parallel ROC including low-noise differential preamplifiers with an electrical bandwidth of 8 kHz. The 128 times 128 focal plane array (FPA) has one row of infrared-blind reference detectors that reduces the effect of FPA fixed pattern noise and variations in the operating temperature relaxing the requirements for the temperature stabilization. Including the noise of the reference and array detectors together with the ROC noise, the fabricated 128 times 128 FPA has an expected noise equivalent temperature difference (NETD) value of 1 K for f/1 optics at 30 frames/s (fps) scanning rate. This NETD value can be decreased to 350 mK by improving the post-CMOS fabrication steps and increasing the number of readout channels.

In this paper, we present a comprehensive calculational model for the noise equivalent temperature difference (NETD) of infrared imaging systems based on uncooled bolometer arrays. The equations ar ...

This paper reports the development and evaluation of an InSb photovoltaic infrared sensor (InSb PVS) operating at room temperature. The InSb PVS consists of 700 InSb p+ - p- - n+ photodiodes connected in series, on a semi-insulating GaAs (100) substrate. An Al0.17rIn0.83Sb barrier layer between p+ and p- layers was used to reduce diffusion of photo-excited electrons. Cutoff wavelength was 6.8 mum and output signal was almost linear with irradiance up to 0.6 mW/cm2. Sensitivity of 67 muV/K and noise equivalent temperature difference of 2.2 mK/Hz1/2 was obtained at room temperature, which shows the sensor to be a suitable for noncontact thermometry.

In this paper we show the latest achievements of HgCdTe-based infrared bispectral focal plane arrays (FPAs) at LETI infrared laboratory. We present and compare the two different pixel architectures that are studied now in our laboratory, named “NPN” and “pseudo-planar”. With these two technologies, a wide range of system applications in dual-band detection can be covered. Advantages of both architectures will be pointed out. We also review performances obtained with these different architectures. The first one has been studied for several years in our laboratory, and we review results obtained on FPAs of size 256 × 256 pixels on a 25 μm pitch, in the MWIR/MWIR (3 μm/5 μm) range. Very high noise equivalent temperature difference (NETD) operability is obtained, at 99.8% for the λc = 3 μm band and 98.7% for the λc = 5 μm band. The second one has been developed more recently, to address other applications that need temporal coherence as well as spatial coherence. We show detailed performances measured on pseudo-planar type FPAs of size 256 × 256 pixels on a 30 μm pitch, in the MWIR/LWIR (5 μm/9 μm) range. The results are also very promising for these prototypes, with NETD as low as 15 mK for an integration time as short as 1 ms, and good operability. The main manufacturing issues are also presented and discussed for both pixel architectures. Challenging process steps are, firstly, molecular beam epitaxy (MBE) HgCdTe heterostructure growth, on large substrates (cadmium zinc telluride) and heterosubstrates (germanium), and, secondly, detector array fabrication on a nonplanar surface. In particular, trenches or hole etching steps, photolithography and hybridization are crucial to improve uniformity, number of defects and performances. Some results of surface, structural and electrical characterizations are shown to illustrate these issues. On the basis of these results, the short-term and long-term objectives and trends for our research and development are presented, in terms of pixel pitch reduction, wavelengths, and dual-band FPA size.

CdxHg1−xTe layers with bandgap in the mid-wavelength infrared (MWIR) and long-wavelength infrared (LWIR) regions were grown by molecular beam epitaxy, and one-dimensional (1D) and two-dimensional (2D) arrays of planar photodiodes were fabricated by ion milling of vacancy-doped layers. The grown layers have varying densities of needle-shaped structures on the surface. The needles are not associated with twins or dislocations in the layers, but could instead be due to (111) facets being reinforced by a preferential Te diffusion direction over steps on the surface. The needles do not seem to affect diode quality. 64 element 1D arrays of 26×26 μm2 or 26×56 μm2 diodes were processed, and zero-bias resistance-times-area values (R0A) at 77 K of 4×106 Ω cm2 at cutoff wavelength λCO=4.5 μm were measured, as well as high quantum efficiencies. To avoid creating a leakage current during ball-bonding to the 1D array diodes, a ZnS layer was deposited on top of the CdTe passivation layer, as well as extra electroplated Au on the bonding pads. The median measured noise equivalent temperature difference (NETD) on a LWIR array was 14 mK for the 42 operable diodes. 2D arrays showed reasonably good uniformity of R0A and zero-bias current (I0) values. The first 64×64 element 2D array of 16×16 μm2 MWIR diodes has been hybridized to read-out electronics and gave median NETD of 60 mK. Images from both a 1D and a 2D array are shown.

This paper describes the modeling and design of two-color microbolometers for uncooled infrared (IR) detection. The goal is to develop a high resolution IR detector array that can measure the actual temperature and color of an object based on two spectral wavelength regions. The microbolometer consists of high temperature amorphous silicon (a-Si:H) thin film layer held above the substrate by Si/sub 3/N/sub 4/ bridge. A thin NiCr absorber with sheet resistance of 377 /spl Omega//sqr is used to enhance the optical absorption in the medium and long IR wavelength windows. A tunable micromachined Al-mirror was suspended underneath the detector. The mirror is switched between two positions by the application of an electrostatic voltage. The switching of the mirror between the two positions enables the creation of two wavelength response windows, 3-5 and 8-12 /spl mu/m. A comparison of the two response wavelength windows enables the determination of the actual temperature of a viewed scene obtained by an IR camera. The microbolometer is designed with a low thermal mass of 1.65/spl times/10/sup -9/ J/K and a low thermal conductance of 2.94/spl times/10/sup -7/ W/K to maximize the responsivity R/sub v/ to a value as high as 5.91/spl times/10/sup 4/ W/K and detectivity D/sup */ to a value as high as 2.34/spl times/10/sup 9/ cm Hz/sup 1/2//W at 30 Hz. The corresponding thermal time constant is equal to 5.62 ms. Hence, these detectors could be used for 30-Hz frame rate applications. The extrapolated noise equivalent temperature difference is 2.34 mK for the 8-12 /spl mu/m window and 23 mK for the 3-5 /spl mu/m window. The calculated absorption coefficients in the medium and long IR wavelength windows before color mixing are 66.7% and 83.7%. However, when the color signals are summed at the output channel, the average achieved absorption was 75%.