The manufacture of affordable second generation MCT detectors is favorably impacted by the availability of low-cost, large-area (Cd,Zn)Te substrates for LPE. We report here on our results in achieving routine production of high-quality, 4 cm X 6 cm, (Cd,Zn)Te substrates using scaled-up vertical Bridgman technology. Six kilogram ingots with diameters as large as 10 cm were produced. Substrate characteristics such as EPD, IR transmission, and precipitate morphology are reported. Substrate purity, which also impacts detector producibility, was monitored using GDMS of the starting elements, intermediate compounds, and the final ingots. Use of ultra-high purity starting materials enabled the Cu concentration in the substrate to be reduced significantly.

We have successfully validated theoretical models of seeded vertical Bridgman-Stockbarger CdZnTe crystal growth and post-solidification processing, using in-situ thermal monitoring and innovative material characterization techniques. The models predict the thermal gradients, interface shape, fluid flow and solute redistribution during solidification, as well as the distributions of accumulated excess stress that causes defect generation and redistribution. Data from the furnace and ampoule wall have validated predictions from the thermal model. Results are compared to predictions of the thermal and thermo-solutal models. We explain the measured initial, change-of-rate, and terminal compositional transients as well as the macrosegregation. Macro and micro-defect distributions have been imaged on CdZnTe wafers from 40 mm diameter boules. Superposition of topographic defect images and predicted excess stress patterns suggests the origin of some frequently encountered defects, particularly on a macro scale, to result from the applied and accumulated stress fields and the anisotropic nature of the CdZnTe crystal. Implications of these findings with respect to producibility are discussed.

The reproducible growth of large, high-quality CdZnTe crystals is crucial for providing low- cost substrates for IR focal plane arrays. We have grown 3-kg, 6.4-cm diameter Cd_1-xZn_xTe (x 0.04) ingots by the vertical Bridgman method, from which we have obtained large-area wafers that can yield single-crystal, twin-free substrates up to 4 cm X 6 cm. The computer-controlled thermal environment was designed to reduce thermal stresses both on the solidified boule and at the melt/solid interface. The ampoule was constructed to reduce the excess Te concentration without active atmosphere control. FTIR transmission spectra of these wafers exhibited 65% transmission from 2.5 micrometers to 20 micrometers across the entire wafer. Glow discharge mass spectrometry (GDMS) confirmed that the concentrations of detrimental impurities were 3 X 10¹⁴ cm^-3. X-ray synchrotron topography showed that the substrates contained large-area subgrains with minimal residual strain at the boundaries. We discuss the suitability of these substrates for LPE growth of Hg_1-yCdx_yTe (y 0.2) epilayers.

We report the liquid encapsulated Czochralski growth and characterization of large, substrate quality single crystal Ga_1-xIn_xAs, 0 < x < 0.10, in excess of 50 millimeters in diameter and weighing 1000 grams. This unique ability to grow large single crystals of ternary III-V compound semiconductors permits realization of the concept of substrate engineering for both homo- and heteroepitaxial applications. One area of particular interest and importance is the development of short visible wavelength (blue) lasers. Wafers with x equals 0.038 have been used for lattice matched MBE growth of ZnSe and ZnCdSe epilayers for blue emitter applications. Low temperature photoluminescence, WDX and double crystal x-ray diffraction (rocking curve) measurements have been utilized to confirm compositional uniformity and crystal quality. Characterization results for both substrate and epitaxial layer are discussed.

The crystallographic orientation of Cd_1-xZn_xTe (x approximately equals 0.045) grown by molecular beam epitaxy (MBE) on a clean (planar) (100) GaAs surface can be controlled by the proper choice of the GaAs surface stoichiometry. An As-stabilized surface initiates (100) oriented growth, while the Ga-stabilized surface yields (111) oriented growth. Cd_1-xZn_xTe (x approximately equals 0.045) MBE layers grown in recesses of shadow masked patterned (100) GaAs substrates were found to be in the (100) orientation regardless of whether precursor surfaces were stabilized with Ga or As. The epitaxial layer's orientation and optical properties were determined by backscattered electron channeling and low temperature photoluminescence measurements, respectively. CdZnTe layers grown in recesses showed improved optical features as compared to the layers grown on planar substrates.

CdTe(111)B layers have been grown on Si(001) and its vicinal surface. Formation of double domains and twins is found to be related to the surface structure of the Si(001) substrate. The tilt angle and tilt direction of the misoriented Si(001) substrate play an important role in suppressing the formation of double domains and twins. A double-domain and twinned CdTe(111)B layer is always obtained, when it is grown on nominal Si(001) substrate. However, by optimizing the tilt parameters, one can consistently obtain single-domain and twin-free CdTe(111)B layers grown on slightly misoriented Si(001) substrates.

Utilizing either the gallium or indium free substrate mounting technique is desirable for producing large area and high quality HgCdTe epitaxial layers. This paper reports that a conventional substrate holder was modified to handle radiational heating. This modification enables substrate rotation to obtain better layer uniformity and realizes reduction in the amount of growth process time. This paper also describes substrate temperature behavior during HgCdTe epitaxy. From the growth initiation to about 2 micrometers -thick HgCdTe growth, the temperature increase was confirmed as being due to absorption of thermal radiation from heated cells and the substrate heater. For further growth, radiation cooling occurred as well. The latter behavior was corrected by the infrared pyrometer. Crystallinity of the epilayer grown by radiational heating was comparable to that of the epilayer grown by conventional thermal conductance heating. Using this technique, both the reduction in process time and the epilayer uniformity of 0.5%((Delta) x/x) over a 2 inch wafer were achieved.

Recent developments in MOCVD growth of Hg_1-xCd_xTe photodiodes using the interdiffused multilayer process are reported. Iodine doping of HgCdTe is described using ethyl iodide. Using ethyl iodide, the iodine doping level can be controlled in the range of 7 X 10¹⁴ - 2 X 10¹⁸ cm^-3 without any memory effect. Activation of the iodine as a singly ionized donor is near 100% at concentrations < 1 X 10¹⁷ cm^-3. Ethyl iodide was not found to react with the other organometallic precursors and abrupt dopant profiles are obtained. The iodine doped HgCdTe films exhibit 80 K electron mobilities >= 1 X 10⁵ cm²/V-s, auger limited lifetimes of approximately 1 microsecond(s) for concentrations of (1-3)X10¹⁵ cm^-3, and x-values approximately 0.22. LWIR p-on-n heterojunctions have been grown in situ using iodine doping for the n-type absorber layer and arsenic doping for the p- type cap layer. Detailed characterization data for the photodiodes are reported.

Cation impurity gettering in Hg_1-xCd_xTe is described in the context of kinetic process models which include the interactions of the impurities and the dominant native point defects. Experimental results are presented using SIMS profiles of Au distributions in Hg_0.8Cd_0.2Te following Hg anneals and ion mills, which are processes known to inject excess Hg interstitials. In either process, Au is shown to redistribute preferentially to high vacancy regions. The junction depth of the low to high Au transition is determined by SIMS. For Hg rich anneals of Au-doped, high vacancy concentration material, the Au junction behavior with respect to anneal temperature, time, and initial vacancy concentration is shown to follow that expected for type converted electrical junctions in vacancy-only material. For milled Au-doped material with a high vacancy concentration, the Au junction depths are found to be approximately proportional to the amount of material removed. Neither the Hg anneals nor the mills form Au junctions in starting material with a low background vacancy concentration.

This paper reports the deposition of (100) GaAs and (111)B CdZnTe layers on silicon substrates up to 4-inch diameter to produce substrates suitable for liquid phase epitaxy (LPE) of high-quality HgCdTe layers. Metalorganic chemical vapor deposition is used for both GaAs and CdZnTe in a reactor capable of deposition onto eighteen 3-inch or ten 4-inch wafers per run. An encapsulation scheme is described that prevents contamination of a Te melt by Si or GaAs during LPE growth. Excellent uniformity of thickness and Zn concentration are achieved in the MOCVD films. The CdZnTe films show only lamellar twins close to the GaAs interface; no twins capable of propagating into the HgCdTe layer are formed. These substrates have been used for the growth of pure HgCdTe films having a dislocation density that is only a factor of 2 to 4 higher than that measured in similar films grown on bulk CdTe substrates.

(100) composition modulation as well as (101) and (101) tweed strain contrast were observed in Zn_1-xMg_xS_ySe_1-y epitaxial films grown on ZnSe buffer layers. Surface buckling of the TEM plan-view specimens along the (110) direction is induced by the modulated structure in the Zn_1-xMg_xS_ySe_1-y films. High quality Zn_1-xMg_xS_ySe_1-y films were obtained by growing a ZnSe buffer layer on As-stabilized GaAs substrates with Zn treatment of the substrate prior to the growth of the film. In these samples, no vacancy-contained Ga₂Se₃ interfacial layer was found at the ZnSe/GaAs interface. Samples with rough ZnSe/GaAs interfaces contained a high density of Frank partial dislocations originating at the ZnSe/GaAs interface. The interface roughness is believed to result from an As-rich GaAs surface after the oxide desorption.

Both Shockley and Frank partial dislocations originating at the ZnS_xSe_1-x/GaAs interface were observed with defect densities in the range less than 5 X 10⁴/cm² to approximately 1 X 10⁸/cm² for samples grown under different conditions. These faulted defects act as heterogeneous nucleation sites for the generation of 60 degree(s) misfit dislocations. A very low density of Shockley partial dislocations was obtained in ZnS_xSe_1-x films grown by the layer-by-layer mode on GaAs substrates. Also, the density of Frank partial dislocations was decreased by exposing the As- stabilized GaAs surface to Zn for 1-2 minutes prior to the growth of the ZnS_xSe_1-x epilayers. These samples contained a defect density lower than 5 X 10⁴/cm². A Ga₂Se₃ compound existing at the ZnS_xSe_1-x/GaAs interface is thought to be the source of the Frank partial dislocations.

In this work, we study the growth parameters of ZnSe on GaAs by MOCVD and characterize the epilayers by various techniques. Epilayers were grown using diethylzinc (DEZn) and diethylselenide (DESe) as source materials. Growth studies were done at 400 degree(s)C under different growth conditions in an atmospheric pressure MOCVD reactor. Different DESe to DEZn ratios (from 0.5 to 5) were used to study the effects of VI/II ratio on ZnSe quality. The as-grown ZnSe epilayers were characterized by double crystal x-ray diffraction, transmission electron microscope (TEM), and scanning electron microscope. The results show excellent surface morphology and crystal quality of ZnSe. The best material was grown on undoped GaAs at the VI/II ratio near unity. The full-width-at-half-maximum of ZnSe (approximately 0.5 micrometers thick) x-ray peak as low as 90 arc seconds was achieved. To our knowledge, this is the narrowest peak among the reported results of ZnSe on GaAs. TEM results also show very low defect density. ZnSe epilayer with low stacking faults density (less than 10⁵/cm²) and large spacing between misfit dislocations (approximately micrometers ) were grown on GaAs.

As-grown bulk ZnSe material prepared by seeded physical vapor transport (SPVT) and melt- grown techniques, and N ion-implanted and heat-treated SPVT material are characterized by room-temperature Raman scattering and low-temperature photoluminescence techniques to evaluate the lattice perfection and to find the impurity and defect levels in the material. The measurements have indicated that the SPVT material is of better quality compared to the melt- grown material. The 200 keV/5 X 10¹³ cm^-2 N ion-implantation and 650 degree(s)C/10 min anneal have resulted in high intensity deep energy level peaks in the photoluminescence spectra recorded on the SPVT samples.

Strain parameters of thin films are needed to numerically simulate the optical gain of light amplification for semiconductor devices. We report a complete quantitative treatment of the stress-strain relations for several common (h,h,k) orientations. We find that earlier reports that attributed the presence of tensile elastic strain to both thermal and lattice-mismatches gave strain estimates that were too low. Using our calculations, we present a model to explain the observed shifts in excitonic recombination energies that have been reported by many groups to date.

Zn1 Cd Se ternary thin films were deposited by electron beam evaporation and their c~iition was examined using X-ray photoelectron spectroscopy (XPS). They all indicated a general selenium deficiency. The Auger parameter (~ ) and ionicity (f.) have been estimated fran XPS data. The variat:i.on of d hetween the different c~sitions is not very significant. The f. values are 0.650, 0.680 for 7.nSe and C"..d.Se films respectively.

This paper discusses the recrystallization mechanism, kinetics and morphology of the crystals of different A2B6 powders versus preparation conditions. It is shown that these parameters are mainly determined by the chemical nature and concentration of a flux.

The technology of producing HgCdTe materials, detectors, and arrays is rapidly maturing. In this paper, the performance and producibility of 2.5 - 14 micrometers LPE HgCdTe epilayers, as well as PC and PV detectors, are presented. The diodes show no degradation after a 95 degree(s)C baking for a period of two weeks. Fully operational linear arrays are produced at Fermionics Corporation. Most of the arrays show excellent uniformity. From the cost and quality point of view, these products are currently very competitive. These results demonstrate the producibility of HgCdTe materials, detectors, and arrays.

MBE (molecular beam epitaxy) grown CdTe layers were processed to fabricate a photoconductor array for the diagnosis of short x-ray pulses from synchrotron radiation sources. The MBE (111)B CdTe layers were grown on (100)Si substrates. Photoconductor arrays were fabricated with gaps of 5 - 50 micrometers using conventional photolithography. Electroless Au or sputtered Au/Ni was used as a contact metal. The temporal response of the resulting CdTe photoconductor was measured with mode-locked 100 fsec Ti:Sapphire laser pulses. The FWHM of single crystalline CdTe photoconductor response pulse is as short as 37 psec with a 20 psec risetime. The photoconductor responds linearly to the x-ray tube photon flux with fixed accelerating voltage up to 40 kV. A significant response increase to the x-ray beam is observed for a layer with good crystalline quality. Spatial response of the CdTe photoconductor array was measured using rotating anode and synchrotron x rays for different beam sizes. Excellent spatial resolution was obtained from narrow angular radiation synchrotron x rays. The CdTe photoconductor was exposed to synchrotron x rays for 60 hours without any noticeable degradation.

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 R_oA 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.

The traveling heater method (THM) is usually characterized by crystal defects such as grain boundaries and dislocations. The need for low cost HgCdTe FPA systems requires high photodiode yield. This demands understanding the crystal defect-diode relationships and necessitates a sorting method that is able to sort the as grown THM wafers before process according to the probability of achieving large photodiode arrays. This paper discusses the influence of crystals defects on photodiode performance and presents a sorting method which is under development. Oriented [111] THM HgCdTe crystals were grown and long wave N⁺P photodiode arrays were fabricated on the A (metal) face. It is found that individual or clusters of high current diodes which deviate drastically from their neighbors -- in current magnitude and in their slope on a Weibull distribution -- could be explained by a correspondence of excessive leakage and low angle sub-grain boundaries which cross the diode location. The distribution of single and multiple defects is compared to models based on isolated point defects and line defects. Yield implications of these results as a function of array design are described.

Vertically integrated photodiode, VIP^TM, technology is now being used to produce second generation infrared focal plane arrays with high yields and performance. The VIP^TM process employs planar, ion implanted, n on p diodes in HgCdTe which is epoxy hybridized directly to the read out integrated circuits on 100 mm Si wafers. The process parameters that are critical for high performance and yield include: HgCdTe dislocation density and thickness, backside passivation, frontside passivation, and junction formation. Producibility of infrared focal plane arrays (IRFPAs) is also significantly enhanced by read out integrated circuits (ROICs) which have the ability to deselect defective pixels. Cold probe screening before lab dewar assembly reduces costs and improves cycle times. The 240 X 1 and 240 X 2 scanning array formats are used to demonstrate the effect of process optimization, deselect, and cold probe screening on yield and cycle time. The versatility of the VIP^TM technology and its extension to large area arrays is demonstrated using 240/288 X 4 and 480 X 5 TDI formats. Finally, the high performance of VIP^TM IRFPAs is demonstrated by comparing data from a 480 X 5 to the SADA-II specification.

The Sofradir mercury cadmium telluride technology is in production. Among numerous devices one Sofradir product, the 288 X 4 linear array with TDI, is at the top level of the production status. This fact is the result of much effort first at the Infrared Research Laboratory (Lir), then at the Sofradir development center where stabilization of the process, including reliability and industrialization, was done. Today a production pilot line is manufacturing many detector dewar assemblies. The purpose of this paper is to describe this pilot line in terms of process, capacity, and process control at each step of the production. Sofradir's test concept is explained. In order to show the process uniformity and the reproducibility, a statistical study on the main characteristics for a typical quantity of detector dewar assemblies is given.

Sofradir has built an infrared detector dewar assembly (DDA) capable of operating under various environmental conditions corresponding to various applications. In this paper it is shown that Sofradir DDA retain their performance for FLIR applications (ground vehicle, helicopter, or aircraft) as well as seeker applications. In particular, Sofradir DDAs permit the user to meet or to exceed the majority of environmental conditions defined in the US military standards such as MIL STD 81OD. Moreover, it has been shown from studies carried out at Sofradir that for components in production such as the 288 X 4 one, the reliability, thermal cycling, and operating and storage conditions are acceptable for this generation of components. Indeed, for instance, it has been demonstrated by test that the MTTF for standard operating conditions can be higher than 15,000 hours for the 288 X 4 focal plane array.

HgCdTe crystals have a high density of intrinsic material defects generated either during the growth process itself or subsequent annealing steps. An understanding of the relationships between these material imperfections and the electrical performance of long-wavelength infrared (LWIR) HgCdTe photodiodes provides guidance in the choice of the growth technique and subsequent processing to minimize electrically detrimental defects. It also sets limits of material perfection required to achieve a specified performance level. Understandings of these relationships have been gained for subgrain boundaries, twins, dislocations, and Te precipitation in LWIR HgCdTe n-on-p photodiodes. Measurements of bias dependent dark current, responsivity, and noise as functions of defect density were performed on both arrays and test structures. This paper emphasizes the direct relationship between material defects and array performance. Wherever possible, the reader is referred to more detailed discussions. All the diodes reported in this study were planar, ion implanted structures using vacancy doped material. All reported measurements were taken at 77 K.

This paper reports an experimental study of the photoelectron energy distribution in the two-photon photoeffect in CdS under a relatively low excitation. A comparison with the luminescence spectrum obtained on the same sample reveals the energy levels in the forbidden zone. The photoelectron and luminescence spectra were measured under both separate and simultaneous excitation of CdS by He I lamp (21.2 eV) and nitrogen laser (3.7 eV). The maximum laser beam intensity at the sample did not exceed 50 kW/cm² at a pulse length of 8 ns. A detailed description of the experimental arrangement and of the photoelectron spectrum measurement technique used is given elsewhere.

Texas Instruments (TI) validated the feasibility of cryoprobing IRFPA arrays in late 1991. Since then, TI has developed a revolutionary automated cryoprobe for screening four and six inch wafers of IRFPAs. Generic prober automation features include cassette to cassette wafer load and unload, wafer alignment, black body selection, aperture selection, probe tip continuity test, and 77.5 degree(s) to 400 degree(s)K wafer temperature control. Modular construction of the prober enables placement of product specific components such as MWIR or LWIR bandpass filters, coldshield, coldfilter, probe card, and noise suppression circuitry on an easily removable `product specific' tooling plate. Prober operation is controlled through object oriented software. IRFPA specific software modules control array operation, data collection, and data reduction. In addition to describing the prober capabilities and versatility, this paper compares prober test data to lab dewar test data for 240 X 1 IRFPAs and projects benefits in reduced cycle time and labor savings.

A crucial aspect of process control in II-VI based device fabrication is the detailed monitoring of material properties, particularly structural quality. We have developed a method of mapping structural defects over large wafer areas using synchrotron white beam x-ray topography and have used it to characterize large area single crystal CdZnTe substrates and LPE HgCdTe epilayers grown on them. The synchrotron white beam technique produces high resolution topographic images of whole wafers regardless of long range strain, and the multiple images generated as a result of the polychromatic white beam (Laue geometry) provide an automatic defect depth profiling. The topographs reveal various types of defect structure in the CdZnTe substrates, and we have compared these topographic images to IR micrographs and x-ray rocking curve maps. Defect structures as revealed by the x-ray topographs were then followed from the CdZnTe substrates to the LPE grown HgCdTe epilayers. Epilayer topographs were also compared to conventional optical micrographs as well as with x-ray rocking curve maps. Finally, a scanning stage was constructed to topographically image large wafers and boule slabs.

The Microelectronics Manufacturing Science and Technology (MMST) program, a program mutually funded by ARPA, the U.S. Air Force, and Texas Instruments, was recently completed at TI and has had great impact on the fabrication of silicon-based devices. The MMST factory relied heavily on sensors to control processing and increase throughput and yield. It is our intent to utilize this MMST technology for infrared focal plane array (IRFPA) production. Aspects of our program to incorporate MMST technology into IRFPA production are discussed emphasizing the use of automated optical-emission-endpoint detection. Automated optical-emission-endpoint detection was successfully demonstrated on a RIE reactor. Optical emission was used to monitor as well as control our dielectric etch. The operation and application of optical emission endpoint detection for IRFPA fabrication are presented.

A magneto-optical system is described that allows for spatial mapping of Faraday rotation and infrared transmission of HgCdTe thin films. Composition, thickness, and absorption coefficient of HgCdTe samples are determined from analysis of transmission spectra. Carrier concentration is extracted from analysis of Faraday rotation spectra. The system provides noncontact, nondestructive rapid screening or detailed diagnostics of HgCdTe material. We also show that the results of resonant magneto-optical spectroscopy support the observation of Faraday rotation caused by optical transitions from shallow compensating acceptors as well as near-midgap defect levels in material with similar x-value. We show that these magneto- optical methods are powerful tools for the study of impurity and defect levels in HgCdTe as well as for characterizing and screening HgCdTe.

Progress has recently been reported on the innovative technique of noncontact mapping of free carrier concentrations in mercury cadmium telluride by using the Faraday effect at discrete wavelengths. We describe here a modification of the basic setup to extend the range of materials covered. An interferometer output was used as a broad band source that covered the spectral region from 7 - 15 micrometers . This permitted measurements of Faraday rotation to be made in HgCdTe material that could not be characterized with available single wavelength sources. Since Faraday rotation over the entire spectral region of interest was made, data could be extracted at optimum wavelengths for a given x value of the material. Data on HgCdTe is presented that show spectral regions where induced spin and where plasma contributions to the rotation predominate. The ability to do wide band Faraday rotation measurements opens up the possibility of extending the technique as a general analytical mapping tool for other classes of materials with potential applications in other fields.

The use of in situ process monitoring equipment is becoming increasingly important in the drive to produce focal plan arrays more efficiently and at decreased cost. As part of Texas Instruments' modular approach to microelectronics manufacturing (microelectronics manufacturing science and technology, MMST), a number of internal process monitors and sensors have been developed. We describe the use of one such sensor, a spectral ellipsometer (SE) for in situ, real-time monitoring and control during infrared device fabrication. Using the SE we have demonstrated end-point detection for the removal of the air-contaminated surface layer of HgCdTe during a remote microwave hydrogen plasma cleanup process step. The SE is also being used to monitor the real-time growth of an interlevel MOCVD ZnS insulator. The spectral ellipsometer provides rapid feedback of film thickness, at a rate of about 10 Hz, enabling its use for process control. We describe the use of the SE for both the HgCdTe surface plasma etch cleanup and the ZnS deposition processes.

In this paper we show that scanned photoluminescence can be a useful tool for characterization of III-V materials, particularly CdZnTe and HgCdTe alloys which are preferentially used in high performance infrared detectors. Spatially resolved integrated photoluminescence or spatial-spectral measurements are performed either at room temperature (test of CdZnTe substrates mainly) or at 77 K (test of HgCdTe epilayers detecting in the MWIR range). Firstly, correlation of the results achieved by this technique with those obtained for instance by x-ray topography or cathodoluminescence shows that it can be in the near future an efficient nondestructive and rapid method for IR substrate selection prior to HgCdTe epitaxy. Then results on Te-rich LPE and MBE HgCdTe layers are reported. They demonstrate that such a measuring technique can lead to process optimization and that an accurate material choice before processing can be envisaged according to the required detector performance. Finally, possible future applications of scanned photoluminescence for analysis of II-VI materials for IR detection are briefly proposed.

Production of large quantities of HgCdTe on CdZnTe has been accomplished by using the dipping liquid phase epitaxy (DLPE) process from tellurium-rich solutions. The dipping LPE process has been used to grow n-type indium doped HgCdTe, extrinsically doped and vacancy dominated p-type thick films for photoconductors, and the vertically integrated photodiodes, respectively. The composition is controlled through the on-line utilization of the HgCdTe phase diagram by a constituent weight tracking technique. The composition is controlled to a CdTe mole fraction of 0.225 +/- 0.002 over long periods of time, approximately 2 years. The LPE reactors were designed for high throughput, flexibility, and low material waste.

Focal plane arrays fabricated in Hg_1-xCd_xTe have matured to the point where they are entering pilot production for military and civilian applications. Loral chose the p-on-n heterostructure photodiode architecture produced by a two step LPE process for production scale-up because of its demonstrated high performance over a broad range of cut-off wavelengths from approximately equals 4 micrometers to > 15 micrometers . To meet the manufacturing requirements of high throughput and flexibility, Loral has successfully scaled and transferred into production, both tellurium rich horizontal slider (for growth of the active n-type base layer) and mercury rich vertical dipping (for growth of the p-type cap layer) technologies. This approach capitalizes on the advantages of each process to maximize the producibility of LPE films in a manufacturing environment. This paper discusses the advantages of horizontal Te-rich growth. Production reactors have demonstrated excellent compositional uniformity (+/- 0.005 in x) over 24 cm². Issues related to the scale-up of the equipment are discussed.

Liquid phase epitaxy (LPE) of (Hg,Cd)Te (MCT) is the technique of choice for the preparation of the materials used for high performance focal plane arrays. Its successful development requires the development of advanced sensors and process controls. We detail here progress on the application of four sensor technologies to the LPE process for growth of MCT layers from Te rich melts on CdZnTe substrates. These include: (1) electron beam microprobe/wavelength dispersive x-ray analysis (WDX) for the rapid measurement of film composition immediately after growth; (2) an RTD based precision temperature control system that controls the melt temperature to better than +/- 0.005 degree(s)C and the Hg reservoir temperature to better than +/- 0.020 degree(s)C; (3) UV/visible optical absorption spectroscopy for the determination of the Hg partial pressure over the melt; and (4) CCD imaging for the detection of the liquid temperature of the LPE growth solution. The impact of each of the sensors on process yield is discussed. The application of the CCD camera to Hg rich high pressure LPE growth is also briefly mentioned.