J. Robert Mahan

Optical and electrothermal design of a linear-array thermopile detector for Geostationary Earth Radiation Budget applications

Described is thermal radiation detector conceived for possible deployment on GERB (Geostationary Earth Radiation Budget). It consists of a linear array of 256 elements, each 60 micrometer square and separated by a 3-micrometer gap. Each element is the active junction of a single-junction-pair zinc- antimonide/platinum thermopile. The reference junction is mounted on an isothermal substrate, and the active junction is thermally isolated from the substrate by a thin layer of parylene. The detector is mounted on one wall of a wedge- shaped, mirrored cavity intended to increase the effective absorptivity and improve the spectral flatness of the detector through multiple reflections. A dynamic opto-electrothermal model of the detector/cavity combination has been formulated in order to facilitate its optimal design. The optical part of the model is based on a Monte-Carlo ray trace that takes into account diffraction at the entrance slit as well as the diffuse and specular components of reflectivity of the cavity surfaces. Heat absorption and diffusion through the thermopile structure has been modeled using the finite element method. The model has been used to validate a method for eliminating optical cross-talk among elements of the array through post- processing of data.

Dynamic electrothermal model for the Earth Radiation Budget Experiment nonscanning radiometer with applications to solar observations and evaluation of thermal noise

A dynamic electrothermal model of the Earth Radiation Budget Experiment total, nonscanning channels has been formulated and implemented as a computer program. This model, which is a modification of an earlier version, is used to simulate two types of solar observations: those obtained through the solar port during solar calibration and those obtained during the satellite pitchover maneuver in which the sun is observed by the radiometer in its Earth-viewing configuration. New results of both simulations are compared with actual flight data. These results show an improved agreement between the simulated and observed radiometer response over previous simulations. The improvement in these severe cases justifies the modification to the model and establishes its accuracy. Thermal noise has been studied also, using a separate but related model, to evaluate its contribution to the radiative energy absorbed by the active cavity. This study has revealed that scattering of the collimated solar radiation contributes, on average, 0.071 mW during solar calibration and 0.207 mW during the pitchover maneuver. These values represent, respectively, 0.156% and 0.460% of the peak power that enters the cavity (?45 mW). On the other hand, the maximum amounts of diffuse power due to emission from the field-of-view limiter and the aperture plate are, respectively, 0.120 and 0.01 1 mW, which amount to 0.270% and 0.024% of the peak power. Finally, the cavity self-contamination contributes only 0.034 mW, or 0.071%, of the peak power absorbed by the active cavity radiometer. This study confirms the assumption that, due to the geometry of the radiometer assembly and the optical properties of its components, thermal noise is small and well within the range of previous estimates.

Use of first-principle numerical models to enhance the understanding of the CERES point spread function

NASAs clouds and the Earths radiant energy system (CERES) program is a key component of the Earth observing system (EOS). The CERES proto-flight model (PFM) instrument is to be launched on NASAs tropical rainfall measuring mission (TRMM) platform on 1 November 1997. Each CERES instrument contains three scanning thermistor bolometer radiometers to monitor the longwave and visible components of the Earths radiative energy budget. An integral part of analyzing these measurements will be the use of high-resolution cloud imager data in conjunction with data from the CERES instruments. The use of high-resolution cloud imager data requires that the point spread function (PSF), or the dynamic response of the radiometric channels as they scan across a far-field point source, be well characterized. The PSF is determined by the field-of-view of the radiometric channel, its optical geometry, and the time response of the thermistor bolometer and its associated signal conditioning electronics. The PSF of the CERES instruments is measured in the laboratory using a state of the art radiometric calibration facility (RCF) developed by TRW. Intrinsic difficulties in making this measurement suggest that a better understanding of the data could be obtained by the use of an independent instrument model. High-level first-principle dynamic electrothermal models of the CERES radiometric channels have been completed under NASA sponsorship. These first-principle models consist of optical, thermal and electrical modules. Accurate optical characterization of the channels is assured by Monte-Carlo- based ray-traces in which tens of millions of rays are traced. Accurate thermal and electrical characterization is assured by transient finite-difference formulations involving thousands of nodes to describe thermal and electrical diffusion within the thermistor bolometer sensing elements and the instrument mechanical structure. The signal conditioning electronics are also included in the models. Numerical simulations of the PSFs of the CERES proto-flight model (PFM) radiometric channels have been completed. This paper presents a comparison between the measured PSF and the independent numerically predicted PSF for the CERES proto-flight model total channel. Agreement between the measured and predicted PSFs is excellent. The result of this agreement is a high confidence in the model to predict other aspects of instrument performance. For example, the model may now be used to predict channel PSFs for elevation scan rates different from the nominal Earth scan rate.

Parameter estimation and optimal design of thermal radiation detectors using engineering prototypes and numerical models

Scientists at NASAs Langley Research Center, in collaboration with researchers at Virginia Tech, are developing the next generation of thermal radiation detectors composed of new space-age materials, including carbon-doped Larc-Si and aerogels. In order to accurately model and design these detectors, it is necessary to determine the in situ thermoelectric properties of these new materials, including thin-film effects and contact resistance. The authors present an approach to determine these properties through the use of simultaneous parameter estimation methods in which experimental results obtained from detector prototypes are compared with results predicted from analytical models. Parametric values are varied using an optimization method to minimize the least-squares error between the experimental and model results. A numerical study is presented to validate the use of this approach. Simulated experimental results were produced using a model based on nominal parameter values. These results were then introduced into a parameter estimation algorithm that was able to recover the parameter values without the benefit of a priori knowledge about the material properties. Genetic algorithms, stochastic hill climbers, and a hybrid of the two methods were investigated for use in parameter estimation.

Diffraction models of radiation entering an aperture for use in a Monte Carlo ray-trace environment

This manuscript presents multiple modeling efforts to describe diffraction of monochromatic radiant energy passing through an aperture for use in the Monte-Carlo ray-trace environment. Described is a deterministic model, based upon Heisenbergs uncertainty principle, which predicts the angle at which an approaching ray is diffracted. The result is a curve which approximates the analytical interference pattern, but does not model the side fringes (i.e. secondary maxima). This model is applicable to either Fraunhofer (far-field) or Fresnel (near- field) diffraction situations. In addition to this model, a diffraction model is presented that approximates the interference pattern including the secondary maxima, as produced by radiation passing through a slit or a circular aperture. This model, based on the Huygens-Fresnel principle with a correcting obliquity factor, is useful for predicting Fraunhofer (far-field) diffraction in the Monte-Carlo ray- trace environment. The motivation for this work is the need to properly model the diffraction of radiant energy as it approaches a detector intended for monitoring the Earths radiation budget from a geo-stationary orbit. The proposed detector, a linear-array of thermopile elements, is housed in a wedge-shaped cavity with a 60-

Experimental and theoretical study of uncertainty in pyranometers for surface radiation

micrometer slit through which radiant energy between wavelengths of 0.1 micrometer and 40 micrometer must pass. A radiative model of this cavity which does not account for diffraction effects has already been developed using the Monte Carlo ray-trace method. This detector was originally intended to fly on the Geo-Stationary Earth Radiation Budget (GERB) instrument.

Optical model of a next-generation instrument for monitoring atmospheric energetics from space

The Eppley pyranometer is widely used to measure shortwave irradiances. This instrument consists of a blackened surface in intimate thermal contact with the hot junction of a thermopile. The cold junction of the thermopile is in thermal contact with a heat sink. Shortwave radiation transmitted through two concentric hemispherical domes is absorbed by the blackened surface. The voltage developed by the thermopile is then interpreted in terms of the shortwave irradiance. Measurements obtained using these instruments are known to be influenced by thermal radiation that produces an offset from the signal that would result solely from the incident shortwave radiation. The thermal radiation emitted and reflected by the filters modifies the net radiation at the detector surface. The ongoing efforts to model these exchanges and to use experimental results to verify the model are described. The parallel experimental effort consists of determining the sensitivity of instrument response to thermal radiation effects. In this effort, thermistors are used to characterize the thermal gradients responsible for the instrument offset. The ultimate goal of the work described is to provide reliable protocols, based on an appropriate instrument model, for correcting measured SW irradiance for variable thermal radiation effects.

Numerical simulation of ground calibration of the CERES thermistor bolometer radiometers

A new Monte-Carlo ray-trace (MCRT) environment has been created and used for the conceptual design of a next- generation radiometer for monitoring atmospheric energetics form space. A multi-band, two-mirror reflecting telescope illuminating an array of thermal detectors is under active consideration as a follow-on to the Clouds and the Earths Radiant Energy System instruments. Future instruments must provide narrower spectral resolution without concomitant sacrifices in radiometric accuracy and spatial resolution. Strategies are under study for obtaining tow or more spectral channels from a single telescope without significant optical cross-talk between channels. Differential filtering based on different combinations of interference filters will be used to achieve spectral separation. Filters are potential thermal noise sources because they may absorb and re-radiate varying amounts of power in response to changes in scene spectral radiance. The MCRT design environment is used here to study the optical performance of a candidate instrument.

Bounding uncertainty in Monte Carlo ray-trace models

NASAs clouds and the Earths radiative energy system (CERES) program is a key component of the Earth Observing System (EOS). Under CERES an array of radiometric instruments will be placed in Earth orbit to monitor the longwave and visible components of the Earths radiative energy budget. High-level dynamic electrothermal models of these instruments have been formulated under NASA sponsorship. Accurate optical and thermal-radiative characterization of the instruments is assured by a Monte-Carlo-based ray-trace in which tens of millions of rays are traced, and a transient finite-difference formulation involving hundreds of nodes is used to describe thermal and electrical diffusion within the thermistor bolometer sensing elements. The external electronic circuit is also correctly included in the instrument model. The actual CERES instruments will undergo pre-launch calibration in a unique thermal-vacuum radiometric calibration facility equipped with blackbodies, a cryogenically cooled active-cavity radiometer, and shortwave sources. This ground calibration can also be simulated using the high-level, dynamic electrothermal models of the CERES instruments. This offers a quick and inexpensive means of verifying the calibration procedure and anticipating any problems that may arise. The results obtained from these simulations may then be used to predict the regression coefficients in the count-conversion equation used to convert instrument readings into radiance, and to determine which parameters should be included in the count-conversion equation to maximize its sensitivity. The paper presents results of the simulated ground calibrations of the CERES total channel instrument, including predicted values for instrument accuracy during the ground calibrations.

Uncertainty and confidence intervals in optical design using the Monte Carlo ray-trace method

In a previous effort the authors developed a methodology for describing uncertainty in thermal radiation Monte Carlo ray- trace (MCRT) analyses. An application to radiometric channels used in space-based observations such as those provided by NASAs Clouds and the Earths Radiant Energy System (CERES) was reported.. In the previous study preliminary attempts were presented to confirm the theory. In the current effort extension and modifications of the previous theory are formulated and new examples are presented to confirm the extended theory. A generic MCRT- based computational environment that simulates radiative exchange among surfaces in enclosures is used to obtain performance estimates of a simple cavity-type thermal radiation detector. Standard statistical methods are used to interpret the results as uncertainties and their related confidence intervals. The example problem is used as a vehicle to verify the modified theory. The authors then demonstrate the theory in a more complex situation that of a high-level numerical model used to predict the dynamic opto- electro thermal behavior of the CERES radiometric channels.