Thermophotovoltaic (TPV) diodes fabricated from InGaAsSb alloys lattice-matched to GaSb substrates are grown by Metal Organic Vapor Phase Epitaxy (MOVPE). 0.53eV InGaAsSb TPV diodes utilizing front-surface spectral control filters have been tested in a vacuum cavity and a TPV thermal-to-electric conversion efficiency ({eta}{sub TPV}) and a power density (PD) of {eta}{sub TPV} = 19% and PD=0.58 W/cm{sup 2} were measured for T{sub radiator} = 950 C and T{sub diode} = 27 C. Recombination coefficients deduced from minority carrier measurements and the theory reviewed in this article predict a practical limit to the maximum achievable conversion efficiency and power density for 0.53eV InGaAsSb TPV. The limits for the above operating temperatures are projected to be {eta}{sub TPV} = 26% and PD = 0.75 W/cm{sup 2}. These limits are extended to {eta}{sub TPV} = 30% and PD = 0.85W/cm{sup 2} if the diode active region is bounded by a reflective back surface to enable photon recycling and a two-pass optical path length. The internal quantum efficiency of the InGaAsSb TPV diode is close to the theoretically predicted limits, with the exception of short wavelength absorption in GaSb contact layers. Experiments show that the open circuit voltage of the 0.53eV InGaAsSb TPV diodes is not strongly dependent on the device architectures studied in this work where both N/P and P/N double heterostructure diodes have been grown with various acceptor and donor doping levels, having GaSb and AlGaAsSb confinement, and also partial back surface reflectors. Lattice matched InGaAsSb TPV diodes were fabricated with bandgaps ranging from 0.6 to 0.5eV without significant degradation of the open circuit voltage factor, quantum efficiency, or fill factor as the composition approached the miscibility gap. The key diode performance parameter which is limiting efficiency and power density below the theoretical limits in InGaAsSb TPV devices is the open circuit voltage. The open circuit voltages of state-of-the-art 0.53eV InGaAsSb TPV diode are {approx}10% lower than the predicted semi-empirical limit to open circuit voltage for a device having absorbing substrate; the voltages are {approx}17% below that for an Auger-limited device having back surface reflector and two-pass optical design.

A systematic approach to thermophotovoltaic (TPV) array design and fabrication was used to optimize the performance of a 192-cell TPV array. The systematic approach began with cell selection criteria that ranked cells and then matched cell characteristics to maximize power output. Following cell selection, optimization continued with an array packaging design and fabrication techniques that introduced negligible electrical interconnect resistance and minimal parasitic losses while maintaining original cell electrical performance. This paper describes the cell selection and packaging aspects of array optimization as applied to fabrication of a 192-cell array.

In a thennophotovoltaic VPV) energy conversion system, a heated surface radiates range onto photododes which are sensitive at these eneSg;es.Part of the absorbed energy is convetted into electric output.Conversion efficiency is maximized by reducing the absorption of non-convertible energy with some form of spectral control.In a TPV system, many technology options exist.Our development efforts have concentrated on flat-plate geometries with greybody radiators, low bandgap quaternary diodes, fiont surface tandem filters and a multi-chip module (MCM) approach that allows selective fabrication processes to match diode performance. the mid-hfiaredRecently, the authors achieved conversion efficiencies of about 20% (radiator 9 W C , diodes 22°C) for a module in a prototypic cavity test enViranment.These tests employed InGaAsSb diodes w i t h 0.52 eV bandgap and front surface filters for spectral C O R ~~O ~.This papa provides details of the individual system c;omponents and describes the measurement technique used to record these efficiencies. S m YLockheed Martin bas been developing thermophotovoltaic (TPV) direct enedgy conversion f i x about eight years.Significant progress has been achieved i n four key areas:

Simulation has provided valuable quantification of the fundamental behavior of thermophotovoltaic cell networks.The results of simulation studies have supported the design and fabrication of small-scale demonstration networks and are expected to guide assembly of large-scale systems.This paper describes the methodology and software simulator developed to address issues in thermophotovoltaic (TPV) networking, including failure analysis, electrical network design, and nonuniform illumination.Results from simulation studies are given illustrating their application to the design and fabrication of small-scale TPV arrays.

The Seebeck coefficient, electrical resistivity, thermal conductivity, and Hall effect have been studied in γ-phase La3−xS4 (LaSy) in the composition range 0.04≤x≤0.3 (1.35≤y≤1.48) in order to ascertain its suitability for high-temperature (300 to 1400 K) thermoelectric energy conversion. In this temperature and composition range the material behaves as an extrinsic semiconductor whose degenerate carrier concentration is controlled by the stoichiometric ratio of La to S. A maximum figure-of-merit (Z) of ∼5×10−4 K−1 at a composition x=0.3, y=1.48 (LaS1.48) was obtained.

The rare earth chalcogenides are important thermoelectric materials due to their high melting points, self-doping capabilities, and low thermal conductivities. Lanthanum sulfides and lanthanum tellurides have been synthesized in quartz ampules, hot-pressed into samples, and measured. The n-type Seebeck coefficients, electrical resistivities, and power factors generally all increased as the temperature increased from 200 to 1000 C. The figure-of-merit for nonstoichiometric lanthanum telluride was 0.001/deg C at 1000 C, considerably higher than for silicon-germanium. Thermoelectric measurements were made for LaTe(2) and YbS(1.4), and p-type behavior was observed for these compounds from 300 to 1100 C.

The properties of the Zr-O-W(100) emitter and the coadsorption of zirconium and carbon monoxide on W(100) have been studied by Auger electron spectroscopy and work function measurements. The thermionic work function of the Zr-O-W(100) emitter from 1450 to 1800 K is 2.54 eV with a pre-exponential value of 4. The field-emission retarding potential work function is 2.75–2.80 eV. A Zr-O-W(100) emitter free of bulk carbon possesses a longer lifetime than an emitter containing bulk carbon. Heating the emitter in carbon monoxide causes diffusion of zirconium into the bulk, and vacuum annealing restores the zirconium to the surface. The coadsorption of zirconium and carbon monoxide on W(100) followed by vacuum annealing produces a minimum work function of 2.3 eV. The adsorption sequence is important and lower work functions correlate with higher carbon concentrations. The stability of the 2.3 eV surface under background gas adsorption and high-temperature heating is discussed.

The coadsorption of cesium and oxygen on iridium and lanthanum hexaboride has been studied by measurements of elemental surface concentrations, work functions, and electron reflectivities. The elemental surface concentrations were determined by Auger electron spectroscopy and the work functions and reflectivities by the field emission retarding potential technique. Simultaneous deposition of cesium and oxygen on iridium or lanthanum hexaboride substrates produced a thick oxide layer. For both substrates this oxide layer possessed a work function of 1.0±0.05 eV and an electron reflectivity at threshold of 0.45±0.10 eV. Brief stepwise heatings of this layer on LaB6 showed stability to 390 K. The threshold reflectivity for some higher work function surfaces depended strongly on whether the final deposition was cesium or oxygen.