In this work we review the reported results as well as the analytical and 3D numerical modeling tools we used to analyze dark current and quantum efficiency data from lattice matched InP/In0.53Ga0.47/InP double layer planar 15 μm pixel pitch focal plane arrays (FPAs). These imaging sensors are designed to operate in the near infrared under overcast “Night-Glow” illumination conditions. A notable finding is that the diffusion dark current component is the dominant current component near and above 300 K and is limited by band-to-band radiative recombination processes. The Shockley-Read-Hall (SRH) recombination through band gap states situated at the intrinsic Fermi level is the dominant component for temperatures below 300 K. 3D numerical simulations consisting of both bulk area and perimeter dependent components explains the dark current component of origin in the space charge region (SCR) with an SRH lifetime of τSRH = 107 μs. Image sensors with extended cut-off wavelength of 2.5 μm at room temperature built on InP/InGaAs are not lattice matched and needed are reductions in the mismatch throughout the InP/InGaAs multilayer epitaxial structure. We analyzed devices with a buffer layer introduced between the InP and the In0.81Ga0.10As absorber and showed that the measured dark current consisted of a diffusion current limited by a back surface interface recombination velocity and a shunt current component expected to be dominant at T ≤ 200 K. At 293 K we calculated a factor of 10 increase in the diffusion current caused by lattice mismatch. Recommended are hole minority carrier lifetime measurements and variable junction diode geometries to assess suspected perimeter/area dependency.
There is great interest in MWIR Ga-free type II strained layer superlattice (T2SLS) nBn detectors for background limited photodetectors (BLIP) operating at temperatures higher than InSb (T ≥ 150K compared to T≈ 80-90K). Recently, Ting et al. [Proc. SPIE, Vol. 10624, 1062410-1 (2018)] reported on measurements of the dark current and quantum efficiency (QE) for e-SWIR (λ>1.7 μm), MWIR, and LWIR Ga-free nBn T2SLS detectors. Of particular interest is the reported MWIR nBn T2SLS data, since the measured dark current and QE provide the opportunity to analyze the measured detector optoelectronic characteristics using optical properties obtained separately from the hole minority carrier lifetime and optical absorption coefficient data, hence no adjustable parameters. The goal is to develop and utilize robust modeling techniques to explain real, measured nBn detector data to understand the technology limitations and the improvements needed to optimize performance and device designs. A notable result is the observation that the dark current data under reverse bias (-100 mV < bias ≤ 0 mV) obeys the ideal diode equation, where the saturation dark current is in agreement with the radiative recombination rate obtained from the measured absorption coefficient and a 9 μs Shockley-Read-Hall (SRH) lifetime obtained from the measured hole lifetime data. Most important is that the expected generation-recombination (G-R) space charge current based on the 9 μs SRH lifetime is not observed as expected for an ideal nBn heterojunction detector and that increasing excess dark currents are observed with decreasing temperature. The in-band external QE measured at 100K is in the range of 30% and is observed to increase with increasing operating temperature which indicates effects influenced by hole mobility anomalies. In contrast to the above listed observations, the Type II Ga-free nBn detector data reported by D. Ting et al. [Appl. Phys. Lett., 113, 021101, 2018] exhibits a hole energy barrier, G-R dark current like characteristics, and both temperature/voltage dependent QE.
In this work we use analytical and 3D numerical modeling tools to analyze data from InP/In0.53Ga0.47/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 τSRH =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 Jdiffusion α to the radiative recombination rate Gr(α) and the radiative minority carrier lifetime τradiative = 1/BNd, where B is the radiative recombination coefficient and Nd the majority carrier concentration.
Timely technology transition with minimal risk requires an understanding of fundamental and technology limitations of material synthesis, device operation and design controllable parameters. However, this knowledge-based approach requires substantial investment of resources in the Science and Technology (ST) stage of development. For low volume niche semiconductor technologies of Department of Defense (DoD) relevance, there is little drive for industry to expend their limited resources towards basic research simply because there is no significant return on investment. As a result, technology transition from ST to product development is often delayed, expensive and carries risks. The Army Research Laboratory (ARL) is addressing this problem by establishing a Center for Semiconductor Modeling of Materials and Devices (CSM) that brings together government, academia, and industry in a collaborative fashion to address research opportunities through its Open Campus initiative. This Center leverages combined core competencies of partner organizations, which include a broad knowledge base in modeling, and its validation; sharing of computational, characterization, materials growth and device processing resources; project continuity; and ‘extension of the bench’ via exchange of researchers between affiliated entities. A critical DoD technology is sensing in the infrared (IR) spectrum, where understanding of materials, devices and methods for sensing and processing IR information must continually improve to maintain superiority in combat. In this paper we focus on the historical evolution of IR technology and emphasize the need for understanding of material properties and device operation to accelerate innovation and shorten the cycle time, thereby ensuring timely transition of technology to product development and manufacturing. There are currently two competing IR technologies being pursued, namely the incumbent II-VI Hg1- xCdxTe technology and the III-V Type 2 Superlattices (SLs) technology. A goal of the CSM is to develop physics based models for Type 2 SLs with the capability to timely understand the knowledge gap between what is built and what is designed.
In this paper we review the intrinsic and extrinsic technological properties of the incumbent technology, InP/In0.53Ga0.47As/InP, for imaging in the visible- short wavelength spectral band, InSb and HgCdTe for imaging in the mid-wavelength spectral band and HgCdTe for imaging in the long wavelength spectral band. These material systems are in use for a wide range of applications addressing compelling needs in night vision imaging, low light level astronomical applications and defense strategic satellite sensing. These materials systems are direct band gap energy semiconductors hence the internal quantum efficiency η, is near unity over a wide spectral band pass. A key system figure of merit of a shot noise limited detector technology is given by the equation (1+Jdark. /Jphoton), where Jdark is the dark current density and Jphoton ~qηΦ is the photocurrent density; Φ is the photon flux incident on the detector and q is the electronic charge. The capability to maintain this factor for a specific spectral band close to unity for low illumination conditions and low temperature onset of non-ideal dark current components, basically intrinsic diffusion limited performance all the way, is a marker of quality and versatility of a semiconductor detector technology. It also enables the highest temperature of operation for tactical illumination conditions. A purpose of the work reported in this paper is to explore the focal plane array data sets of photodiode detector technologies widely used to bench mark their fundamental and technology properties and identify paths for improvements.
A key design feature of P+-on-n HgCdTe detectors is the depth of the p-type region. Normally, homojunction architectures are utilized where the p-type region extends into the narrow-gap absorber layer. This facilitates the collection of photo-carriers from the absorber layer to the contact; however, this may result in excess generation-recombination (G-R) current if defects are present. Alternatively, properly adopting a heterojunction architecture confines the p-type region (and the majority of the electric field) solely to the wide-gap layer. Junction placement is critical since the detector performance is now dependent on the following sensitivity parameters: p-type region depth, doping, valence band offset, lifetime and detector bias. Understanding the parameter dependence near the hetero-metallurgical interface where the compositional grading occurs and the doping is varied as either a Gaussian or error function is vital to device design. Numerical modeling is now essential to properly engineer the electric field in the device to suppress G-R current while accounting for the aforementioned sensitivity parameters. The simulations reveal that through proper device design the p-type region can be confined to the wide-gap layer, reducing G-R related dark current, without significantly reducing the quantum efficiency at the operating bias V = -0.100V.
Processing improvements have facilitated manufacturing reduced pixel dimensions for lattice-matched InGaAs on InP short-wave infrared detectors. Due to its technological maturity, this material system continues to garner attention for low-light level imaging applications. With pixel dimensions smaller than minority carrier diffusion lengths, optimizing array performance by reducing crosstalk from lateral carrier diffusion remains an important design issue. Analytical models, however, have provided limited insight on underlying mechanisms limiting device performance in the conventional planar double heterointerface device. Quantitative modeling provides tools to investigate performance sensitivities and their underlying mechanisms. In this work we develop a three-dimensional numerical simulation for dense P+n In0.53Ga0.47As on InP photo detector focal plane arrays using a conventional planar, back-illuminated structure. We evaluate optical generation with finite-difference time-domain analysis, and model carrier transport in a drift diffusion analysis simultaneously solving the carrier continuity and Poisson equations. Using this model we investigate modulation transfer function variations with pixel pitch and diffused junction geometries for small dimension arrays. By accounting for carrier diffusion effects, these results should provide a benchmark against which to evaluate modulation transfer function contributions from other effects, such as crosstalk attributable to photon recycling.
Low light level imaging applications requiring high detectivity demand photon shot noise limited performance at temperatures near 300K. Analytical models, however, have provided limited insight on underlying mechanisms limiting performance in conventional planar double heterointerface In0.53Ga0.47As on InP P+n photodiodes for imaging the visible and short wave infrared. Quantitative modeling provides tools to investigate performance sensitivities and their underlying mechanisms. In this work we use three-dimensional numerical simulation to investigate intrinsically limited diffusion and Shockley-Read-Hall generation recombination dark currents for a planar P+n photodiode situated in a 3×3 mini array. We assess the influence of geometry by varying pitch, junction location, and photodiode size. Modeling shows that SRH generation currents, not including surface effects, vary with both junction perimeter and area, and that the perimeter component dominates small radius junctions. By varying the axial junction placement we show that widegap junctions result in bias-dependent quantum efficiencies that require higher reverse bias, and result in higher dark currents, than shallow homojunctions at comparable efficiencies. Finally, numerical simulation explains lateral diffusion current suppression in dense arrays in terms of suppressed minority carrier density gradients. The analysis demonstrates that the boundary condition applicable to dense arrays requires no lateral diffusion current at symmetry planes bisecting segments connecting uniformly reverse biased nearest neighbor diodes. Following Grimbergen, this leads to radial geometry curves describing dark intrinsic diffusion reductions with pitch. The quantitative modeling provides insight explaining the observation that the ideal diode equation correctly estimates dense array dark diffusion currents.
Well recognized are the potential benefits in camera simplicity, power reduction and increased
cooler life associated with the capability of operating infrared focal plane arrays at or near room
ambient temperatures. Quality imagery in the 3 to 5 μm spectral band at scene temperature of 300K
with focal plane array temperatures up to 175K was demonstrated recently. The array consisted of
640*512, 16 μm pitch N+p(As) detector elements grown by metal organic vapor phase epitaxy on a
GaAs substrate. In this paper, a carrier recombination model is presented that explains the dark
current density data as a function of inverse temperature. Basically the dominant carrier
recombination occurs through ionized donor-like flaws centered in the upper half of the energy gap.
For Hg1-xCd x Te, x=0.3 and 0.2867 materials, the flaw energy level, Eflaw(0K) respectively, is
centered at 0.189 eV and 0.1181 eV above the valence band edge; The shortest possible lifetime τp0
for hole capture respectively is 3.5 and 550 μs. Band to band recombination is not observed to be
dominant even in the temperature region T ≈ 300K, where the radiative and Auger lifetimes are
significantly smaller than τp0. The asymmetry parameter γ = τn0/τp0 <<1.
Mid wave infrared (MWIR) imaging in the 3-5 um spectral band has traditionally been performed by InSb
sensors. InSb technology is presently limited to a near 80K operating temperature and the hunt has been on for a
higher operating temperature (HOT) technology that does as well at 150K as InSb at 80K, but with reduced
power requirements. Amongst these alternative technologies are photovoltaic sensors consisting of heterostructures
of HgCdTe (MCT). In previous work we assessed the device performance of several alternative MWIR
HOT technologies (MCT on Si, MCT on GaAs) as a function of operating temperature. In this work we compare
the NEDT histograms for these alternative technologies with InSb to better understand how their performance can
be improved at higher temperatures. We also present analysis formalism for quantitatively assessing the number
of FPA pixels which reside in the central versus the shoulder portions of the histogram.Begin the Introduction two
lines below the Keywords. The manuscript should not have headers, footers, or page numbers. It should be in a onecolumn
format. References are often noted in the text1 and cited at the end of the paper.
Reported is a detailed analysis of the dark current versus voltage versus temperature data
of planar hetero-structure P+n mid wavelength infrared MWIR photodiodes with band
gap energy Eg(78K) = 0.243 eV, λg= 5.1 μm and long wavelength infrared LWIR
photodiodes with Eg(78K) = 0.115 eV, λg= 10.8 μm. The purpose of the investigations is
to identify the dominant carrier recombination mechanisms and in particular to determine
at what temperature and voltage is the onset of Shockley Read Hall (SRH) space charge
currents. The important finding is that the currents can mostly be explained by a
combination of Auger (e-e) and radiative carrier recombination processes with no
evidence of SRH recombination through near mid-gap states; a lower bound estimate of
the SRH lifetime for LWIR photodiode is 100 μs. Intrinsic radiative recombination is
found to be the dominant carrier recombination mechanisms for the MWIR photodiode
with a carrier concentration Nd=1015 cm-3, and Auger (e-e) being dominant for the LWIR
photodiode. The LWIR Auger (e-e) lifetime data is well fitted with the Beattie,
Landsberg and Blakemore (BLB) formulas with a constant overlap integral F1F2= 0.15,
which is in accord with recent electronic band structure calculations. From the analysis
of variable area LWIR photodiodes the minority carrier conductivity mobility and
diffusion length at 80K are calculated to be 350 cm2/V-s and 23 μm respectively. The
LWIR lifetime measured by the photoconductive decay method is in agreement with the
expected intrinsic Auger (e-e) lifetime ≈ 2 μs at 80K and with the lifetimes obtained from
device analysis. For T ≤ 40K, trap assisted tunneling is the dominant current in reversed
bias LWIR photodiodes; forward bias currents are dominated by diffusion currents of
origin in the n- layer. For the MWIR photodiode deviation from diffusion limited
behavior to G-R is observed at T < 80K and, the SRH lifetimes ιn0 and ιp0 are estimated
to be 50 ms. Measured and calculated external quantum efficiencies at the peak
responsivity wavelength λpk for both MWIR and LWIR photodiodes are ≈ 70% at 78K.
For imaging in the 3-5 μm spectral band scene temperature 300K, F/3 optics, the noise
equivalent temperature difference NE▵T of MWIR photodiodes is calculated to be near
background limited performance BLIP =12.4 mK for detector temperatures Td ≤ 150K.
In this work the current versus voltage data of a p-n+ junction is converted into minority
carrier lifetime data. Space charge recombination currents dominate at modest reverse
bias at 80K and taking the dominant recombination centers to be located at the intrinsic
Fermi level, the lowest minority carrier lifetime τ0 is determined to be 35ns. This single
Shockley-Read-Hall carrier recombination parameter provides an excellent fit to the data
over temperature range 40K≤ T≤ 130K; the 35ns minority carrier lifetime also explains
the quantum efficiency data. The transition from diffusion to space charge currents
occurs for temperatures, T ≤ 100K. For T≤ 40K trap assisted tunneling is the dominant
current component. Based on imaging system requirements to be near background
limited for photon flux ≈ 1015 ph/cm2-s and detector temperature of 80K, the minority
carrier lifetime will need to be increased by one order of magnitude.
InAs-GaSb strained layer superlattices (SLSs) form a narrow band gap material whose cut-off wavelength can be tuned from 3 um to beyond 30 um. Theory predicts that in the LWIR and VLWIR, the SLS narrow bandgap layer structures can be engineered to reduce Auger recombination, relative to other narrow bandgap materials, such as HgCdTe. This should result in the SLS diodes having better performance than currently available detectors. A key to achieving this improved performance is knowing the detailed layer structure of the superlattice, and being able to accurately model this layer structure. Having an accurate model to guide the improved performance is essential to optimizing this material system.
Cross-sectional scanning tunneling microscopy data will be presented which shows that the actual layer structure differs significantly from the intended layer structure, due to the detailed dynamics of MBE growth and the very thin layers in the superlattice. Specifically, cross-sectional scanning tunneling microscopy demonstrates that the InAs contains excess antimony, and the GaSb excess indium, due to segregation from the underlying arsenide-on-antimonide, or antimonide-on-arsenide, heterojunctions respectively.
These deviations from the intended structure have a significant impact on the predicted properties of the superlattice. The predicted behavior of the intended and actual superlattice structures will be compared to measured performance.
The National Polar-orbiting Operational Environmental Satellite System (NPOESS) Cross-track Infrared Sounder (CrIS) is a Fourier Transform interferometric sensor that measures earth radiances at high spectral resolution. Algorithms use the data to provide pressure, temperature, and moisture profiles of the atmosphere. The CrIS instrument contains photovoltaic detectors with spectral cut-offs denoted by SWIR, MWIR and LWIR. The CrIS instrument requires large-area, photovoltaic detectors with state-of-art detector performance at temperatures attainable with passive cooling. For example, detectors as large as 1 mm in diameter are required. To address these needs, Molecular Beam Epitaxy (MBE) is used to grow the appropriate bandgap n-type Hg1-xCdxTe on lattice matched CdZnTe. The p-side is obtained via arsenic implantation followed by appropriate annealing steps.
This paper investigates 1/f noise performance of Hg1-xCdxTe photovoltaic detectors when detector current is varied by changing detector area, bias, temperature and incident flux. Holding detector bias and temperature constant, measured 1/f noise current is proportional to the detector current. However for all detector areas measured, non-uniformity is observed in the noise current due to the varied quality of the detectors. Even for the λc=16μm , 4-μm-radius, diffusion-limited detectors at 78K held at reverse bias, the average and standard deviation in dark current is Id=9.76+/- 1.59x10-8A while the average and standard deviation in noise current at 1 Hz in a 1 Hz bandwidth is in=1.01+/- 0.63x10-12A. For all detector areas measured at 100 mV reverse bias, the average and standard deviation in dark current to noise current ratio is α D=in/Id=1.39+/- 1.09x10-5. Defects are presumed resident in the detectors that produce greater non- uniformity in the 1/f noise as compared to the dark current at 100 mV reverse bias. Noise was also measured as a function of temperature for two λ c=16 micrometers detectors from 55 K to 100 K. The average and standard deviation in the noise current to dark current ratio is αD=in/Id=2.36+/- 0.83x10-5 for the 26-micrometers -diameter detector and (alpha) D=1.71+/- 0.69x10-5 for the 16-micrometers -diameter detector. Dark and noise current were measured while changing the bias applied to a detector. In the diffusion-limited portion of the detector I-V curve, 1/f noise is independent of bias with α D=in/Id=1.51+/- 0.12x10-5. When tunneling currents dominated, αT=in/Id=5.21+/- 0.83x10-5. The 1/f noise associated with tunneling currents is a factor of three greater than the 1/f noise associated with diffusion currents. In addition, 1/f noise was measured on detectors held at -100 mV and 78 K under dark and illuminated conditions. The average noise to current ratio αD was approximately 1.5 x 10-5 for dark and photon-induced diffusion current. However, detector-to-detector variations exist even within a single chip. The two most important points are that non-uniformities in material/fabrication need to be addressed and that each individual type of current component has an associated 1/f noise current component, the magnitude of the relationship being different depending on the source current.
State-of-the-art large area photovoltaic detectors fabricated in HgCdTe grown by Molecular Beam Epitaxy have been demonstrated for the Crosstrack Infrared Sounder instrument. Large area devices (1 mm in diameter) yielded excellent electrical and optical performance operating at 81K for LWIR band and at 98K for MW and SWIR bands. LWIR and MWIR detectors have near-theoretical electrical performance, and AR-coated quantum efficiency is greater than 0.70. Measured average RoA at 98K is 2.0E7 W-cm2 and near-theoretical quantum efficiencies greater than 0.90 were obtained on SWIR detectors. These state-of-the-art large area photovoltaic detector results reflect high quality HgCdTe grown by Molecular Beam Epitaxy on CdZnTe substrates in all three spectral bands of interest.
The National Polar-orbiting Operational Environmental Satellite System (NPOESS) Cross-track Infrared Sounder (CrIS) is an interferometric sensor that measures earth radiances at high spectral resolution, using the data to provide pressure, temperature and moisture profiles of the atmosphere. The pressure, temperature and moisture sounding data are used in weather prediction models that track storms, predict levels of precipitation etc. The CrIS instrument contains SWIR ((λc approximately 5 μm at 98K), MWIR (λc approximately 9 μm at 98K) LWIR (λc approximately 16 μm at 81K) Focal Plane Array (FPA) modules. A critical CrIS design selection was the use of photovoltaic (PV) detectors in all three spectral bands. PV detectors have the important benefits of high sensitivity and linearity. Each FPA modules consists of nine large (1000 μm diameter) photovoltaic detectors with accompanying cold preamplifiers. This paper describes the performance for all the modules forming the CrIS Detector Preamplifier Module (DPM). Molecular Beam Epitaxy (MBE) is used to grow the appropriate bandgap n-type Hg1-xCdxTe on lattice matched CdZnTe. SWIR, MWIR and LWIR 1000 μm diameter detectors have been manufactured using the Lateral Collection Diode (LCD) architecture. Custom pre-amplifiers have been designed to interface with the large SWIR, MWIR and LWIR detectors. The operating temperature is above 78K, permitting the use of passive radiators in spacecraft to cool the detectors. Recently fabricated 1000 micrometers diameter photovoltaic detectors have the measured performance parameters listed in the Table below. Expected D* performance from the detector/pre-amplifier models are also listed in the table. The D* values are calculated at the CrIS program peak wavelength specified for each spectral band.
Very Long Wavelength InfraRed (VLWIR; (lambda) c approximately equals 15 to 17 micrometer at 78 K) photovoltaic detector operating in the 78 K range are needed for remote sensing applications. This temperature range permits the use of passive radiators in spacecraft to cool the detectors. VLWIR ((lambda) c approximately equals 15 to 17 micrometer at 78 K) photovoltaic detectors in a range of sizes (8 micrometer diameter to 1000 micrometer diameter) have been fabricated and their performance measured as a function of temperature. Molecular Beam Epitaxy (MBE) was used to grow n-type VLWIR Hg1-xCdxTe on lattice matched CdZnTe. Arsenic was implanted and the wafer was annealed to provide the p-type regions. All the material was grown with wider bandgap cap layers and consequently the detector architecture is the Double Layer Planar Heterostructure (DLPH) architecture. Id - Vd versus temperature curves for 8 and 1000 micrometer diameter, (lambda) c equals 17 micrometer at 78 K detectors indicate that the 8 micrometer diameter detector is diffusion limited for temperatures greater than 63 K even at a -200 mV bias. There is no appreciable tunneling at T equals 50 K and at -200 mV applied bias. At T equals 40 K tunneling commences at a bias approximately equals -80 mV. Below T equals 30 K, the diode is tunneling limited. The 1000 micrometer diameter detector is diffusion limited at bias values less than -50 mV at 78 K. At zero bias, the detector impedance is comparable to the series/contact resistance. Interfacing with the low (comparable to the contact and series resistance) junction impedance detector is not feasible. Therefore a custom pre- amplifier was designed to interface with the large VLWIR detectors in reverse bias. The detector is dominated by tunneling currents at temperatures less than 78 K. The 1000 micrometer diameter, (lambda) c approximately equals 17 micrometer at 78 K detectors have dark currents approximately equals 160 (mu) A at a -100 mV bias and at 78 K. Detector non-AR coated quantum efficiency > 60% was measured at -100 mV bias in these large detectors and the response was constant across the (lambda) equals 7 micrometer to 15 micrometer spectral band. With AR- coating the quantum efficiency will be > 70%. Response was measured and non-linearity < 0.15% was calculated for the 1000 micrometer detectors. The flux values were in the 1017 ph/cm2/sec range and were changed by varying the blackbody temperature. In addition, a linear response was measured while varying the spot size incident on the 1000 micrometer detectors. This excellent response uniformity measured as a function of spot size implies that, low frequency spatial response variations are absent, for the 1000 micrometer detectors.
The ability to hybridize various detector arrays in disparate technologies to an assortment of state-of-the-art silicon readouts has enabled direct comparison of key IR detector technologies including photovoltaic (PV) HgCdTe/Al2O3, PV HgCdTe/CdZnTe, PV InGaAs/InP, and the photoconductive (PC) GaAs/AlGaAs quantum well IR photodetector (QWIP). The staring focal plane arrays range in size from 64 X 64 to 1024 X 1024; we compare these IR detector technologies versus operating temperature and background flux via hybrid FPA test at operating temperatures from 32.5 K to room temperature and photon backgrounds from mid-105 to approximately equals 1017 photons/cm2-s. Several state-of-the-art IR FPAs are included: a 1.7 micrometers 128 X 128 InGaAs hybrid FPA with room temperature D of 1.5 X 1013 cm-Hz1/2/W and 195K D of 1.1 X 1015 cm-Hz1/2/W; a 3.2 micrometers 1024 X 1024 FPA for surveillance; a 4.6 micrometers 256 X 256 HgCdTe/Al2O3 FPA for imaging with BLIP NE(Delta) T of 2.8 mK at 95K; and a 9 micrometers 128 X 128 GaAs QWIP with 32.5 K D > 1014 cm-Hz1/2/W at 32.5K and 8 X 1010 cm-Hz1/2W at 62K.
To achieve the DoD objective of low cost high performance infrared focal plane arrays a manufacturing technique is required which is intrinsically flexible with respect to device configuration and cutoff wavelength and easily scaleable with respect to volume requirements. The approach adopted is to fully develop the technology of molecular beam epitaxy (MBE) to a level where detector array wafers with a variety of configurations can be fabricated with first pass success at a reduced cost. As a vapor phase process, MBE lends itself directly to: (1) the inclusion of real-time monitoring and process control, (2) a single or multiple wafer growth mode, (3) nearly instantaneous changes in growth parameters. A team has been assembled to carry out the program. It is composed of four industrial organizations -- Rockwell International, Hughes Aircraft Company, Texas Instruments, and Lockheed-Martin, and a university -- Georgia Tech Research Institute. Since team members are committed suppliers and users of IRFPAs, technology transfer among team members is accomplished in real-time. The technical approach has been focused on optimizing the processes necessary to fabricate p-on-n HgCdTe double layer heterostructure focal plane arrays, reducing process variance, and on documenting flexibility with respect to cutoff wavelength. Two device structures have been investigated and fabricated -- a 480 by 4 and a 128 by 128.
Extensive material and device statistics of performance and reproducibility are presented to show the maturity of this technology. The demonstration vehicles to monitor yields during this demonstration were long-wavelength infrared (LWIR) HgCdTe multilayer wafers with 128 X 128 detector arrays. The heterostructure photodetectors were of the p-on-n planar configuration. Device data show that MBE LWIR diode test structures have performance that equals that of p-on-n double heterostructure photodiodes made by LPE. Due to the special attention given to understanding and reducing epilayer growth-induced defects, we have achieved improvements in FPA operability values from 92% to 98%. These improvements have resulted in the demonstration of a 128 X 128 FPA hybrid that had detectivity (D*) background limited performance when operating at 80 K in a tactical background environment. Mean D* was 1.28 X 1011 cmHz1/2/W. The corresponding mean NE(Delta) T was an excellent 5.9 mK.
In this paper we present p-on-n heterostructure HgCdTe photovoltaic device data that illustrates the high performance and flexibility in band gap control of the molecular beam epitaxy (MBE) technology. This flexibility demonstration was carried out by growing material for operation in the following cut-off wavelength ((lambda) co) ranges of interest: LWIR [(lambda) co(77 K) equals 9-11 micrometers ], MLWIR [(lambda) co(77 K) equals 6-7 micrometers ], and VLWIR [(lambda) co(40 K) equals 20 micrometers ]. Detailed analyses of the current-voltage characteristics of these diodes as a function of temperature show that their dark currents are diffusion-limited down to 80 K, 50 K, and 30 K for the MLWIR, LWIR, and VLWIR photodiodes, respectively. In general, the RoA device values were uniform for the three band gap ranges when operating under diffusion limited conditions. The planar MBE HgCdTe technology has been further validated with the successful fabrication and operation of 64 X 64 hybrid FPAs.
We present p-on-n heterostructure HgCdTe photovoltaic device data that illustrate the high performance and flexibility in band-gap control of molecular beam epitaxy technology. This flexibility demonstration was performed by growing material for operation in the following cutoff wavelength (λco) ranges of interest: long wavelength IR (LWIR) [λco(77 K) = 9 to 11 μm], mid-long wavelength IR (MLWIR) [λco(77 K) = 6.8 μm], and very long wavelength IR (VLWIR) [λco(40 K)=20 μm]. Detailed analyses of the current-voltage characteristics of these diodes as a function of temperature show that their dark currents are diffusion limited down to 80, 50, and 30 K for the MLWIR, LWIR, and VLWIR photodiodes, respectively. In general, the R0A device values were uniform for the three band-gap ranges when operating under diffusionlimited conditions. We confirmed this by fabricating a 64 x 64 LWIR (λco = 10.2 μm) hybrid FPA with detectivity (D*) operability greater than 97% when operating at 77 K. The mean D* value for this device was 1.4 x 1011 cm Hz1/2/W and it was background limited at the tested flux of 2.18 x 1016 photons/cm2 s. This device was tested at higher temperatures of operation without changing background conditions, and it remained background limited up to 100 K.
Imagery of long wavelength infrared HgCdTe and GaAs quantum well staring arrays in size 128 X 128 has been demonstrated. In this paper, we compare detector array performance characteristics, discuss the natural and technological limitations of both technologies and identify the improvements likely to be made in the near future. At this stage of feasibility demonstration in the spectral band 8 - 10 micrometers , the effective quantum efficiency in GaAs FPAs is 4% compared to 60% for HgCdTe and the responsivity is 0.08 A/W compared to 4.5 A/W. This value of 0.08 A/W is significantly below the value 2 A/W reported for single quantum well infrared photodetectors (QWIP) detectors. The peak detectivities and NE(Delta) T at 78 K are (5 X 109 cm (root)Hz/W, 0.037 K) and (2 X 1011, 0.005 K) for QWIP and HgCdTe, respectively. The residual nonuniformities after two-point correction are < 0.01% for QWIP arrays and 0.012% for HgCdTe. Crosstalk is currently unsatisfactory in QWIP detector arrays, but design concepts can be used to reduce this effect. For terrestrial imaging, GaAs quantum well detector arrays most likely will need to operate at temperatures below 80 K from fundamental considerations; HgCdTe detector arrays are background limited at operating temperatures <EQ 90 K. Since cooling can drive cost and reliability, and since significant progress has been made in producing high quality HgCdTe detector arrays with good yield, it is unlikely that HgCdTe will be displaced by this technology for terrestrial applications. For low background space applications at (phi) b <EQ 1012 ph/cm2-sec, QWIP detectors at 40 K are background limited. This observation plus their radiation hard characteristics suggest a possible niche in strategic applications.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.