THE NEAR INFRARED MAPPING SPECTROMETER EXPERIMENT ON GALILEO by R.W. Carlson, P.R. Weissman, W.D. Smythe, J.C. Mahoney, and the NIMS Science and Engineering Teams* * The Near Infrared Mapping Spectrometer (NIMS) Engineering and Science Teams consist of I. Aptaker (Instrument Manager), G. Bailey (Detectors), K. Baines (Science Coordinator), R. Burns (Digital Electronics), R. Carlson (Principal Investigator), E. Carpenter (Structures), K. Curry (Radiative Cooler), G. Danielson (Co-Investigator), T. Encrenaz (Co-Investigator), H. Enmark (Instrument Engineer), F. Fanale (Co-Investigator), M. Gram (Mechanisms), M. Hernandez (NIMS Orbiter Engineering Team), R. Hickok (Support Equipment Software), G. Jenkins (Support Equipment), T. Johnson (Co-Investigator), S. Jones (Optical-Mechanical Assembly), H. Kieffer (Co-Investigator), C. LaBaw (Spacecraft Calibration Targets), R. Lockhart (Instrument Manager), S. Macenka (Optics), J. Mahoney (Instrument Engineer), J. Marino (Instrument Engineer), H.Masursky (Co-Investigator), D. Matson (Co-Investigator), T. McCord (Co-Investigator), K. Mehaffey (Analog Electronics), A. Ocampo (Science Coordinator), G. Root (Instrument Systems Analysis), R. Salazar (Radiative Cooler and Thermal Design), D. Sevilla (Cover Mechanisms), W. Sleigh (Instrument Engineer), W. Smythe (Co-Investigator and Science Coordinator), L. Soderblom (Co-Investigator), L. Steimle (Optics), R. Steinkraus (Digital Electronics), F. Taylor (Co-Investigator), P. Weissman (Co-Investigator and Science Coordinator), and D. Wilson (Manufacturing Engineer) This paper is dedicated to the memories of Gary Bailey and Hal Masursky. Gary would have been very pleased with the excellent performance of his detectors and Hal would have enjoyed the Galileo flyby of Venus, one of his favorite planets. Their crucial contributions to NIMS and Galileo will continue to be apparent throughout the mission, and will be appreciated in whatever future success we may enjoy. Table of Contents 1. Introduction 2. Scientific Objectives 2.1 Venus 2.2 Earth 2.3 Moon 2.4 Asteroids 2.5 Jupiter Atmosphere 2.6 Jovian Satellites 3. Instrument Description 3.1 General Description and Operation 3.2 Optical Design 3.3 Detectors and the Focal Plane Assembly 3.4 Radiative Cooler 3.5 Mechanisms 3.6 Electronics Design 3.7 Thermal Design 3.8 Instrument Contamination Protection 4. Instrument Calibration 4.1 Introduction 4.2 Spectral Calibration 4.3 Radiometric Calibration 4.4 Spatial Calibration 4.5 Spacecraft Radiometric Calibration Target 4.6 Spacecraft Photometric Calibration Target 5. Operating Modes and Data Acquisition 5.1 Instrument Modes and Operation 5.2 Instrument Commands 5.3 Typical Encounter Operations 6. Spacecraft Interactions 6.1 Thermal Control 6.2 Contamination Control 6.3 Instrument Pointing 6.4 Spacecraft Obscuration 7. NIMS Mission Design Aspects LIST OF ILLUSTRATIONS [Figures are in separate Postscript files named INSTFGnn.PS where nn is the figure number.] Fig. 1 Photograph of the NIMS Instrument. Fig. 2 Schematic Diagram of the Instrument and Scanning Motions. Fig. 3 Schematic of the NIMS Optical Design. Fig. 4 Photographs of the Focal Plane Assembly. Fig. 5 Focal Plane Assembly Components. Fig. 6 Radiative Cooler Assembly Fig. 7 Electronics Block Diagram Fig. 8 Preamplifier and Autobias Circuitry Fig. 9 Analog Signal Chain Electronics Fig. 10 Instrument Timing Fig. 11 Spectral Bandpass Response Fig. 12 Spatial Response Fig. 13 Parameter Table Assignments Fig. 14 Spacecraft Obscuration LIST OF TABLES Table 1 Instrument Parameters Table 2 NIMS Standard Operating Modes Abstract The Galileo Near Infrared Mapping Spectrometer (NIMS) is a combination of imaging and spectroscopic methods. Simultaneous use of these two methods yields a powerful combination, far greater than when used individually. For geological studies of surfaces, it can be used to map morphological features, while simultaneously determining their composition and mineralogy, providing data to investigate the evolution of surface geology. For atmospheres, many of the most interesting phenomena are transitory, with unpredictable locations. With concurrent mapping and spectroscopy, such features can be found and spectroscopically analyzed. In addition, the spatial/compositional aspects of known features can be fully investigated. The NIMS experiment will investigate Jupiter and the Galilean satellites during the two year orbital operation period, commencing December 1995. Prior to that, Galileo will have flown past Venus, the Earth/Moon system (twice), and two asteroids; obtaining scientific measurements for all of these objects. The NIMS instrument covers the spectral range 0.7 to 5.2 microns, which includes the reflected-sunlight and thermal-radiation regimes for many solar system objects. This spectral region contains diagnostic spectral signatures, arising from molecular vibrational transitions (and some electronic transitions) of both solid and gaseous species. Imaging is performed by a combination of one-dimensional instrument spatial scanning, coupled with orthogonal spacecraft scan-platform motion, yielding two-dimensional images for each of the NIMS wavelengths. The instrument consists of a telescope, with one dimension of spatial scanning, and a diffraction grating spectrometer. Both are passively cooled to low temperatures in order to reduce background photon shot noise. The detectors consist of an array of indium antiminide and silicon photovoltaic diodes, contained within a focal-plane-assembly, and cooled to cryogenic temperatures using a radiative cooler. Spectral and spatial scanning is accomplished by electro-mechanical devices, with motions executed using commandable instrument modes. Particular attention was given to the thermal and contamination aspects of the Galileo spacecraft, both of which could profoundly affect NIMS performance. Various protective measures have been implemented, including shades to protect against thruster firings as well as thermal radiation from the spacecraft. 1. Introduction The Near Infrared Mapping Spectrometer (NIMS) experiment on the Galileo Orbiter Spacecraft represents a combination of imaging and spectroscopic methods. The advantage of spectral imaging, as opposed to pure imaging systems or single-field-of-view spectrometers, is the ability to simultaneously find, identify, and map compositional units on planetary surfaces. With such information one can investigate the geochemical evolution of satellite surfaces as well as the dynamical and compositional properties of atmospheres. The NIMS instrument possesses modest spatial and spectral resolution, and operates in the near-infrared range of 0.7 - 5.2 microns. This spectral range is particularly diagnostic of minerals known or suspected to occur on planetary and satellite surfaces, and also includes many observable features of atmospheric species. The spectral resolution was designed for investigating the relatively broad bands seen in surface reflectance, yet is adequate for identifying several major and minor atmospheric constituents. Galileo will be inserted into Jupiter orbit in December of 1995, commencing a nearly two year investigation of the Jovian system, performing eleven orbits around the planet during this period. Prior to Jupiter arrival, Galileo has flown by Venus (February 1990) and will fly by the Earth and Moon twice (December of 1990, 1992) and two asteroids--Gaspra (October 1991) and Ida (August 1993). NIMS will investigate all of these bodies. The scientific objectives for these measurements are briefly presented in the following Section (2). The purpose of this paper is to describe and document the NIMS instrument design and development, which posed many unique design and spacecraft integration challenges, owing largely to the low temperatures required by the detectors and optics. The instrument itself is described in Section 3, followed by calibration aspects (4), instrument operating modes (5), and spacecraft considerations (6). Mission design aspects which relate to NIMS are presented in Section 7. General aspects of the NIMS experiment have been briefly documented by Carlson (1981), and a description of the Galileo mission and its complement of experiments is contained in a volume by Yeates et al. (1985). Although this paper is devoted to a discussion of the instrument - i.e. the hardware - the corresponding analysis tools - the software - must also be mentioned. NIMS generates a great amount of such data in a short amount of time, and this large volume of data can only be digested using highly-efficient, visual, interactive computer methods. These computer tools are therefore an integral part of the NIMS experiment, and a separate discussion is warranted. These aspects are not discussed here. In the interim, refer to Torson (1989) for an overview of NIMS data visualization capabilities. 2. Scientific Objectives 2.1 Venus During the Venus flyby (February, 1990), NIMS measurements concentrated on spectral features which arise from surface and deep-atmosphere thermal emission. These features occur in spectral regions where CO2 is relatively transparent, allowing one to probe far below the cloud region and to measure the spatial variations of the intervening cloud extinction. This infrared radiation is observable on the nightside of Venus and was only recently discovered (Allen and Crawford, 1984). Two types of measurements were obtained from NIMS: (1) multiple spatial images at selected infrared wavelengths for dynamical studies, and (2) detailed spectra at a variety of latitudes and longitudes for chemical abundance information, specifically H2O and CO in the deep atmosphere. The results from this flyby are described by Carlson et al (1991). 2.2 Earth During the first flyby of the Earth-Moon system, NIMS performed both atmospheric and surface spectral-imaging of the Earth and similar geological investigations of the Moon. For the Earth measurements, NIMS investigated mesospheric airglow emission features and obtained geological maps of Australia and Antarctica. Exploratory global mapping of the Earth was also performed; the first time ever in this spectral region. Comprehensive lunar measurements were also obtained by NIMS at multiple phase angles, but with relatively poor spatial resolution. One of the primary goals of NIMS during both the 1990 and 1992 flybys is investigation of mesospheric water, observable through limb scans of infrared fluorescence in the 2.7 micron band. It has been recently proposed (Thomas, et al., 1989) that noctilucent clouds, and upper-atmosphere water in general, has increased over the past century due to a larger amount of biological and anthropogenic emission of methane, a photochemical source of water in the mesosphere. The abundance of water in the upper atmosphere is poorly determined. Microwave measurements are in general agreement with photochemical models, but rocket ion spectroscopy and infrared measurements indicate a greater water abundance, perhaps implying an additional source of mesospheric water (c.f. Garcia, 1989). NIMS may provide an independent measurement of the vertical water profile at several latitudes, extending to the summertime south polar region. Other limb airglow emission features that were investigated are the O2 infrared bands, the well-known infrared hydroxyl bands, the strong CO2 v3 band, ozone emissions at 4.8 microns, and the tail of the NO (1-0) band, a prime cooler of the thermosphere. The NIMS limb measurements occur within roughly +/- 1/2 hour of closest approach, and sample both the night- and day-side airglow. As Galileo receded from the Earth, NIMS investigated specific geographic regions as determined by the timing of the flyby, illumination geometry, and the gain characteristics of the instrument. These gain characteristics were established for our ultimate goal, Jupiter at 5.2 AU from the sun, and tend to be too sensitive for general Earth observations. The Earth at 1 AU is simply too bright. Nevertheless, by choosing favorable geometry, we were able to perform spectral mapping of both Australia and Antarctica during the 1990 pass and plan to investigate other regions in 1992. Using ground-truth measurements, we can corroborate and extend our remote sensing measurements to both continental and planetary scales. 2.3 Moon On the first pass through the Earth-Moon system, the closest approach of Galileo was ~350,000 km, limiting NIMS resolution to ~170 km, in contrast to 10 km typical of spot spectrometer measurements obtained from ground-based observations. This initial pass provided viewing of roughly a quarter of the lunar surface (selenographic longitudes 90 to 180 deg) not previously observed spectroscopically. The phase angle range of observations (30 to 150 deg) extended the range of viewing geometry available from Earth, particularly for sub-Earth meridians, where Earth observations are limited to 90 deg phase angles. The wavelength coverage was also extended by NIMS, terrestrial observations being generally limited to below 2.5 microns by the atmosphere. However, because the NIMS dynamic range is designed for measurements at Jupiter, the lunar observations were saturated at many wavelengths, particularly for small incidence angles; unsaturated spectra to 2.5 microns generally require incidence angles greater than 80 deg. At wavelengths beyond about 3 microns, thermal emission becomes important. Because of this dynamic-range limitation, full NIMS spectra are available for only narrow angular regions. Nonetheless, with multiple observations taken at many geometries, some important lunar questions can be addressed, including: (1) an initial search for hydrated minerals - there is a remote possibility that hydrated minerals may be present near the polar regions where low temperatures may occur. The high sensitivity of NIMS allows a search for the 3 micron hydration feature during the second Earth-Moon encounter which passes over the Moon's north polar region, with a spatial resolution of about 60 km. (2) Spectral characterization of additional lunar areas - with the two passes, about 20% of the Moon not visible from the Earth can be mapped with surface resolution from 200 to 500 km. (3) Obtaining the first spectra in the 2.5 to 5.2 micron region. The NIMS wavelength range will allow an accurate determination of local surface temperatures and a correction for thermal emission for that portion of the spectrum that contains reflected radiation. 2.4 Asteroids Galileo will fly by the main belt asteroids Gaspra (16 km diameter) and Ida (32 km diameter) in October 1991 and August 1993, respectively. Ground based studies have identified both asteroids as S-type. The NIMS goal at each encounter will be to acquire spectral-reflectance and thermal-emission images of the asteroid at maximum possible spatial resolution. Because of the high velocity of the flybys, it will be possible to obtain resolved images of only one hemisphere of each object. Full disc, unresolved spectra of the other hemisphere will be obtained during approach. NIMS will identify and map mineral species on the surface of each asteroid, and will seek to determine if the surfaces are chemically heterogeneous. 2.5 Jupiter Atmosphere Although NIMS was originally conceived for satellite surface spectral reflectance measurements, the experiment is well-suited for a variety of investigations of the Jupiter atmosphere. This is due to: (1) The spectral range available to NIMS, which includes signatures from several minor species such as germane, phosphine, and water, which are produced in the deep atmosphere of Jupiter and may serve as tracers of motions in this unexplored altitude region of the planet. (2) The NIMS spectral range also includes absorption by the more abundant molecules: CH4, NH3, and H2. Due to variations in absorption strength for these molecules and cloud layer variability, NIMS can probe a large altitude range, ranging from the high-altitude polar aerosols, far above the ammonia cloud deck, extending down into the water cloud region at the several-bar-level. (3) The spatial resolution of the experiment, about 300 km, is sufficient to resolve the many dynamical features of the atmosphere and to investigate their temporal changes. Investigation of changes over time scales of hours to years are possible with the repeated observations available from Galileo. Finally (4) a full range of phase angle coverage is available with Galileo, enabling one to optically investigate the microphysical properties of the diverse cloud layers. Prior discussions of the NIMS measurements at Jupiter can be found in Taylor and Calcutt (1984) and Hunten et al. (1986). 2.6 Jovian Satellites For the three large icy Galilean satellites, Callisto, Ganymede, and Europa, the primary NIMS science objectives are to map the various surface compositional units and to identify their elemental and mineralogical composition. An important aspect is to study these three objects as a collection. Voyager multispectral data suggests that there are common compositional units across the three objects. For example, the crust of Callisto, saturated with scars of ancient impact structures, has the same albedo and color as the remnants of the oldest terrain on Ganymede. Likewise, the much younger grooved terrains on Ganymede appear to be similar to the linear markings on Europa. A primary question concerning these units is the composition of the dark components that are mixed with the dominant water-ice crusts. Are these materials silicates or organic-rich materials derived from primitive objects such as comet nuclei? If silicates, NIMS may detect bands due to olivines, pyroxenes, or a range of iron-bearing minerals. C-H features might be present if the satellite surfaces contain dark organic components such as those found on some asteroids. Magnetospheric sources of implanted material (e.g. sulfur) may be an important process, providing chemically reactive species which can modify the surfaces (e.g. generating S-O from the S implanted in H2O). Another set of questions pertain to the formation and history of these compositional units. Are there systematic correlations of dark component abundance or nature with geologic age and setting? Were the dark materials added early during the accretion or late during subsequent impacts? Methods to address these questions involve studying the global distribution of these units, examining leading- versus trailing-hemispheres as well as the geologic setting of such units. For example, dark-rayed craters on Ganymede may expose units of concentrated dark materials that exist as layers and lenses in the subsurface. Ejecta may show compositional correlations with size, age, latitude, terrain type, or longitude that will provide insights into their nature and origin. Another important question concerns other volatile species. These might be involved in active processes of eruption and transport, in particular on Europa, or be found as cold-trapped species, for instance in Ganymede's icy polar caps. The innermost Galilean satellite, Io, is of great interest to the NIMS investigation, exhibiting not only a wide range of volcanic processes, some of which are continuously active, but also an equally bizarre mixture of surface chemistry, composition and mineralogy. Active volcanic processes include: 1) two classes of eruptive plumes thought to be driven by superheated volatiles in the form of sulfur and sulfur dioxide vapor, 2) eruption of icy clouds of material along scarps and fractures that have been proposed to be due to the escape of liquid sulfur dioxide to the surface, where it explosively forms gases and ice, and 3) dark hot spots, ranging in temperature up to 400 kelvins, which have been proposed to be crusted pools of liquid sulfur. All of the processes likely involve a wide range of chemical components other than S and SO2. Other volatiles may be cold-trapped in the polar regions. H2S could exist in regions on the surface; it has been identified in ground-based spectra. Not only is there a plethora of potential spectral reflection features from the exotic compositional components, but gaseous absorption features may be detected over hot volcanic regions and the thermal output of the volcanic activity itself will be monitored and mapped. There is even the possibility that molten silicates will occasionally breach through to the surface. 3. Instrument Description 3.1 General Description and Operation Performing simultaneous spectral and spatial infrared mapping at Jupiter necessitates use of high light-gathering power optics and sensitive detectors, both operating at low temperatures to minimize detector and background noise. The resulting NIMS instrument, shown in a photograph in Fig. 1 and schematically in Fig. 2, consists of the following major elements: (1) a telescope with one dimension of spatial scanning, (2) an optical chopper to provide a dark reference, (3) a diffraction grating spectrometer which disperses the radiation onto the focal plane assembly, (4) a focal plane assembly consisting of multiple detectors, optical filters, and preamplifier circuitry, (5) a passive radiative cooler which cools the focal plane to cryogenic temperatures, and (6) signal processing and control electronics, not shown. The typical operating mode of the instrument can be described with the aid of Fig. 2. One dimension of spatial scanning is provided within the telescope employing a "wobbling" secondary mirror, giving 20 contiguous pixels, each with 0.5 x 0.5 mrad resolution, for a total angular field of 0.5 x 10 mrad. The other dimension of spatial scanning is provided by slowly slewing the spacecraft scan platform in the cone direction. During one half of the up/down mirror scan, the grating is set at a fixed angle, with a corresponding set of 17 wavelengths striking the 17 individual detectors of the focal plane assembly. At the extremes of the mirror scan, the grating is stepped to a new setting, and a new set of wavelengths are measured during the second half of the mirror scan. The chopper frequency is synchronized to the 63 Hz spacecraft timing, and the mirror and grating motions are synchronized to the chopper, with motions taking place during the dark portion of the chopper cycle. The size and number of grating steps can be adjusted for specific encounter conditions and scientific objectives, and the scan platform motion is matched to the resulting spectral scan time. Details of the instrument timing are given in Section 3.6 and the instrument modes are described in Section 5. Additional information on many aspects of the instrument are contained in a series of papers by Aptaker (1982a, 1982b, 1983, 1987). A tabulation of instrument parameters is given below: TABLE 1 Instrument Parameters for the Near Infrared Mapping Spectrometer ============================================================================ Angular Resolution: 0.5 mrad x 0.5 mrad (individual pixel size). Angular Field: 10 mrad x 0.5 mrad (20 pixels, cross-cone direction. Spectral Range: 0.7 - 5.2 u. Spectral Resolution: 0.0250 u (lambda > 1 u), 0.0125 (lambda < 1 u ). No. of Spectral Samples: Variable; 408, 204, 102, ... ,1. Spectral Scan Time: Variable; 8 2/3 sec (408 spectral samples), 4 1/3 sec (204 spectral samples), ........ 1/3 sec (17 spectral samples). Telescope: 22.8 cm Diameter f/3.5 Ritchey-Chretien, 800 mm equivalent focal length, spatial scanning via moving secondary mirror, Operating Temperature ~ 150 kelvins. Telescope Etendue: 1.1 x 10-4 cm2 steradian. Spectrometer: 39 lines/mm plane grating spectrometer, 400 mm focal length, f/3.5, Dall-Kirkham collimator, 200 mm focal length, f/1.8, flat-field camera, Bipartite diffraction grating, 30% blazed for 1.9 u, 70% blazed for 3.8 u. Operating Temperature ~ 150 kelvins. Detectors: Seventeen individual photovoltaic diodes, 15 indium antimonide, 2 silicon, Active area = 0.2 mm x 0.2 mm, Quantum efficiencies > 70 %. Radiative Cooler: Passive radiative cooler, Achieves 64 kelvins, with detectors energized. Noise Equiv. Radiance: 7 x 10-9 W cm-2 sterad-1 per pixel and per spectral resolution element (0.025u), at 3 u and 70 kelvins FPA temperature. Noise Equivalent Albedo: 0.0002 at 5 AU. Mechanisms: Torque motor drives for spatial and spectral scans, Tuning fork chopper, Deployable covers for telescope, radiative cooler. Electronics: Seventeen channel signal chain, Microprocessor controlled (RCA 1802). Gain States: 4 ground-commandable gain states, detectors 1-14, Automatic gain switching, detectors 15-17. Protective Devices: Covers for pre- and post-launch protection, Continuous instrument purging through launch, Heaters for continuing contamination control. Mass: 18 kg. Power: 12 Watts (average), 13 Watts (peak). Dimensions: 83 x 37 x 39 cm (optics), 20 x 25 x 13 cm (electronics). Data Rate: 11.52 kbps Data Encoding: 10 bits (0 - 1023) Mounting: Scan platform, Co-aligned with SSI, UVS, PPR On-board Calibration: Photometric Calibration Target (PCT), (a solar reflectance target), and a Radiometric Calibration Target (RCT-NIMS), a blackbody radiator. 3.2 Optical Design A variety of considerations led to the ultimate design of the NIMS optics, which primarily involved signal and instrument noise aspects. In order to minimize the latter, one must use small-area detectors, as the noise varies as the square root of detector area. At the same time, maximizing the signal on a detector requires large acceptance angles, i.e. use of a low f-number camera system illuminating the detectors. Similarly, adequate signal is obtained through use of a large diameter telescope, consistent with the system etendue' and angular resolution requirements. The resulting optical design of the instrument is illustrated in Fig. 3 and consists of all-reflective telescope and spectrometer sections. Much of the optical design and fabrication were performed by the Perkin-Elmer Corporation. Development and testing of the NIMS optics have been previously described by Macenka (1983). The telescope is a 228 mm diameter, f/3.5 Ritchey-Chretien design, with an equivalent focal length of 800 mm. The secondary mirror steps in 20 equal increments, sweeping the image in the plane of the field stop in 0.5 mrad increments and thereby providing one dimension of spatial scanning. The field stop is a 400 micron wide slit, which defines a 0.5 mrad field-of-view, normal to the mirror scan direction and parallel to the plane of dispersion of the spectrometer. The angular resolution in the other direction is defined by the projection of the detectors at the field stop, and is approximately 0.5 mrad. Spatial response measurements are given in Section 4.3. The slit was made longer than required in order to ensure against misalignments caused by thermal or vibration induced shifts. For a distant point source, the telescope forms an image with 90% of the energy contained within a circle of angular diameter 0.05 mrad, which is quite satisfactory when compared to the aforementioned 0.5 mrad resolution. An InGaAs light emitting diode is mounted on the telescope spider and is used for inflight wavelength verification of the spectrometer. The spectrometer employs a plane diffraction grating, illuminated by a Dahl-Kirkham collimator and followed by a wide-angle flat-field camera which focuses the entrance slit (the telescope field stop) onto the detectors. Stepwise rotation of the grating allows the complete NIMS spectrum to be generated. The collimator has an effective focal length of 400 mm and a focal ratio of f/3.5. The smaller mirror in the collimator has a slightly toroidal surface to correct some system aberrations, primarily astigmatism from grating anamorphism. The grating is a 39 lines/mm dual-blazed grating, with 30% of the area blazed for 1.9 microns and the remainder for 3.8 microns. The first order of the grating is used for wavelengths greater than 1 micron (the InSb detectors), and the second order for shorter wavelengths (Si detectors). Between the blaze configuration and use of multiple orders, reasonable efficiencies can be obtained over the relatively large wavelength range of NIMS. Measurements of the blaze efficiency of the flight grating are presented by Macenka (1983). Ruling of the master NIMS grating, replication, and measurement of efficiencies were performed by the Perkin-Elmer Corporation. The detectors are widely spaced in the focal plane, requiring a wide-angle, flat-field camera. The two mirrors comprising the camera are both rather extreme aspheric surfaces, with an effective focal length of 200 mm and a focal ratio of f/1.75. The linear dispersion in the focal plane is 8 mm/micron in first order. For the active area of the detectors (0.20 mm square) the spectral width of each is 0.025 micron. This matches the width of the entrance slit, yielding a triangular spectral bandpass, only slightly broadened due to finite spot sizes (see Section 4.2). The grating can be stepped through minimum increments of one-half of a spectral resolution element. The spectrometer exhibits some residual astigmatism and was aligned for the best spectral focus, consequently the vertical spatial resolution profile is somewhat broadened (see Section 4.3). All of the mirrors and the grating were fabricated from fused-silica. Because of the large number of reflections, efficient infrared-reflective surfaces are required; NIMS uses pure gold with no protective overcoats, obtaining reflectivities of R = 96 to 98% in the spectral interval 0.7 to 5.2 microns. Obscuration, mainly by the camera secondary mirror, reduces the incident energy with a transmission factor of Tobs = 60%, giving an etendue' for the optics, exclusive of reflection losses and grating efficiencies, of A Tobs = 6.1x10-05 cm2 steradian. Stray light is reduced with baffles and an interior finish of matte black (Bostik-Finch Catalac Black). The optics and their housing must operate at low temperatures in order to minimize photon shot noise from background thermal emission. Furthermore, initial alignment was performed at room temperature, thus thermal aspects are an important part of the design, and are discussed in Section 3.7. 3.3 Detectors and the Focal Plane Assembly There are seventeen individual detectors (15 InSb and 2 Si) contained within the focal plane assembly (FPA), along with their associated spectral filters and electronic preamplifier components. Shielding from high energy particles, necessary in the Jovian magnetosphere, is provided by a hermetically sealed Tantalum case, which includes a sapphire window for optical input. A platinum resistance thermometer provides a measure of the FPA temperature. Fig. 4 shows photographs of the FPA while Fig. 5 illustrates the various components and their packaging. The FPA was manufactured by Cincinatti Electronics Corporation, and various aspects are described by Bailey (1979) and Smith et al. (1982). Dispersed radiation from the spectrometer enters the FPA through two sapphire windows, the first contained within the radiative cooler (see Section 3.4), while the second is integral with the FPA and is anti-reflection coated for the near infrared. The radiation then passes through optical filters, whose purpose is to reject higher-order radiation and to limit the amount of thermal radiation incident on the detectors. These filters are at the cryogenic temperatures of the FPA and emit negligible amounts of thermal radiation that can be sensed by the detectors. A cold field-of-view limiting aperture is placed just behind the filters, further limiting thermal radiation on the detectors. Each of the seventeen photodiode detectors contained in the FPA has a photo- active area of 0.2 x 0.2 mm , and each is anti-reflection coated for its own individual spectral region. Detector quantum efficiencies of 70% or greater were measured for the coated photodiodes. By operating the detectors at a constant bias voltage, near zero, the detectors act as high impedance current sources, with linear response to incident photon flux. Noise performance of the instrument is determined by the detectors and their preamplifiers, each of which can be characterized by a noise-current spectral density. When the detectors are operated at near-zero bias, only Johnson noise contributes, with a noise-current spectral density of i = sqrt(4kT*Deltaf/RD), where RD is the junction resistance (Hall et al., 1975). RD exhibits a rapidly varying temperature dependence (Kruse et al., 1962), varying as RD oc T^(-1/2)*exp(1/2 eV/kT), V being the detector band gap (0.22 eV for InSb). From this, the InSb detector noise is expected to show roughly a factor of three change for a temperature change of 10 kelvins, and this dependence was indeed observed for the NIMS instrument. Measurements of individual NIMS InSb photodetectors at 77 kelvins show a noise current density of 1.8 X 10^-15 Amps/sqrt(Hz), with 0.9 x 10-15 Amps/sqrt(Hz) for the preamplifiers. This implies that NIMS will be preamplifier-noise limited for FPA temperatures less than about 70 kelvins. NIMS noise levels measured in flight, as well as in the laboratory, are entirely consistent with the above values. A hybrid dual junction field effect transistor (JFET) preamplifier front-end and 1010 feedback resistors are mounted in close proximity to each detector. Locating these components within the FPA minimizes circuit noise contributions by providing minimum capacity at the detector and by providing a low impedance interface to the external electronics assembly. The preamplifier circuitry contains a differential source-follower circuit which was experimentally found to give the best electromagnetic interference (EMI) performance. A commandable automatic bias circuit is incorporated in order to maintain the detectors at the aforementioned constant near-zero bias. This feature allows higher temperature operation of the InSb detectors by enforcing nonsaturation of the preamplifiers due to increased dark current and was included for the unlikely event of high FPA temperatures. Experience in flight has shown that there is no need for this precaution, although possible inflight contamination of the radiative cooler could change this situation. Additional details of the preamplifier electronics are given in Section 3.6 Radiation shielding is provided by a 3 mm thick tantalum enclosure and by shielding the back of the camera secondary mirror, providing complete angular shielding. The shields were designed to limit the integrated exposure to < 10 krad during the nominal Galileo mission, to prevent radiation damage, while simultaneously reducing noise caused by penetrating magnetospheric particles, mainly electrons. Tests were performed on the developmental model using energetic gamma rays; from this one can predict the resulting performance at Jupiter: Jovian radiation noise will be most severe for the observations taken during the Io close encounter just prior to Jupiter orbit insertion, where a predicted signal-to-noise ratio of 10:1 is predicted. Corresponding results for flyby observations of Europa and Ganymede are 30:1 and 100:1 respectively. Measurements from Callisto's distance are unaffected by ambient magnetospheric radiation. High energy magnetospheric protons can also cause displacement damage, with loss of sensitivity for InSb detectors, but such loss can be recovered by annealing the detectors at roughly 300 kelvins. For this reason, the radiative cooler contains a commandable heater which can elevate the detectors to annealing temperatures. 3.4 Radiative Cooler The indium antimonide detectors require cryogenic temperatures for operation and this is achieved with a single-stage passive radiative cooler, illustrated in Fig. 6. The cold stage, containing the FPA, has a 627 cm2 aluminum honeycomb plate which radiates energy to space and cools the detectors down to 64 kelvins. The ultimate temperature is determined by input power to the cold stage and arises from many sources. The FPA dissipates 9 mW of electrical power, and thermal conduction from the cable and support mechanism provides another thermal path. Incident radiation is another source, and can arise from the instrument and the cooler housing and shield, from the spacecraft and other instruments, and from the planet itself. Minimization of these sources is discussed in the following paragraphs and in Section 6.1. The cooler was provided by the Santa Barbara Research Center. Mechanical support of the cold stage is accomplished through a suspension system of fiberglass bands, similar to the mounting of a bicycle wheel hub. This arrangement not only provides an extremely low thermal conductance path, but also very stable mechanical positioning stability ( 0.01 mm). Electrical access to the FPA is provided by a 60-conductor ribbon cable with narrow stainless steel plated conductors in a thin polyimide sandwich, with a Ni coating for shielding. Radiative loading from the instrument is minimized using a surrounding outer shroud with very low-emittance gold surfaces. The outer surfaces of the cold stage are similarly treated giving very low radiative coupling. Radiation from the spacecraft is reduced using a shield around the radiator plate to optically block emissions from the scan platform and portions of the spacecraft, the latter dependent upon the scan platform cone angle. Radiative loading to the honeycomb plate by the inner surface of the shield itself is reduced by use of a low emissivity (<0.03), highly specular (99 % at 7 microns) gold coating. Additionally, the shield is thermally insulated from the cooler body, and radiatively cools to ~ 120 kelvins. The surface of the shield which faces the scan platform is thermally isolated with multiple layers of thermal blanketing, while the outward-facing radiating surface is painted for high emissivity. In order to maintain the cooler performance, it is crucial that the low emissivity of the inner shield surface be maintained. Any contamination would seriously increase the radiative loading; consequently the shield contains a 26 Watt strip heater to elevate the temperature and avoid condensation of spacecraft outgassing and thruster products. This heater is used nearly continuously, except during observation periods, and elevates the shield to about 300 kelvins. Additional contamination protection, for the pre-launch and the early post-launch period, was inclusion of a deployable radiator cover and continuous dry nitrogen purging (see Sections 3.8 and 6.2). A heater is also included for the cold stage, for use in annealing the detectors in the unlikely event of radiation damage (see Section 3.3). The radiator is mounted at 62.5 deg to the telescope optical axis, in the plane of rotation of the scan platform. Its full field-of -view is circular, with an angular radius of 71.5 deg. The total mass is 1.9 kg. The thermal mass of the cold stage is about 250 J deg^-1 (at 90 kelvins,temperature dependent), giving a cool-down time from 300 to 65 kelvins of 30 hours. The initial cooling rate is very rapid, but slows at lower temperatures, requiring 24 hours to cool from 125 kelvins to 65 kelvins. Additional in-flight performance values are given in Section 6.1. 3.5 Mechanisms There are three mechanisms in the NIMS instrument: an optical chopper and two nearly identical, mirror and grating drives. In addition, the telescope and cooler had protective covers, since deployed in flight. The chopper serves to modulate the detected radiation, allowing the dark current level of the detector to be subtracted on a pixel-by-pixel basis. It is a tuning fork with a resonant frequency slightly above the spacecraft 63 Hz timing signal, and contains two moving blades, each mounted on opposing tines, which chop the light. The chopper is located immediately in front of the telescope field stop (the spectrometer entrance slit) and modulates the radiation with an approximate 50 % duty cycle. The back surfaces of the blades are coated with a black finish, while the front is coated with reflective gold. When the instrument is off, the blades are in the open position, and behind a pre-slit. This protects the low-thermal-mass blades from damage by inadvertent sun pointing. The chopper is driven at the spacecraft 63 Hz timing rate. This rate is slightly below the natural resonant frequency; in the event of a drive circuit failure the fork can be operated at its natural frequency, with overall synchronization with spacecraft timing being maintained by NIMS instrument software. Motion sensing and drive power are provided by magnetic pickup coils and a phase comparator feedback loop. The amplitude is kept at a constant level through an amplitude comparison circuit. The tuning fork and electronic design were provided by the American Time Company. Mirror and grating motion is accomplished with direct-current magnetic motors, similar to a loudspeaker drive, but adapted for angular rather than linear motion. The rotor, which undergoes very small, stepwise, angular excursions, is a fork which pivots on low-loss flexures and contains a samarium-cobalt magnet. An enclosing drive solenoid is attached to the stator. Motion is sensed through a pair of linear voltage differential transformers (LVDTs), one for position/velocity feedback control, the second as an independent, telemetered measure of the grating and mirror positions. Each LVDT consists of a permeable core, mounted on the rotor, and three stator-mounted coils - two for the drive current and one for position sensing. Motion of the core within the coils varies their coupling, and the induced EMF is linearly proportional to displacement. In operation, the desired position (mirror or grating) is generated in the microprocessor and converted to an analog voltage. Error signals generate motion drive, which is further controlled by rate information. The mechanism thus steps from one position to the next, settling to a stable position within a short interval (less than the dark half of a chopper cycle for the mirror, somewhat more for the grating). Two ejectable covers are used to protect the NIMS optics and radiator (see Section 3.8). Each covers consists of a lightweight aluminum frame with a multi-layer thermal blanket closure. They are mounted onto the telescope and cooler apertures using uncaptured hinges, and locked into place with removable latch pins. Conical dowels ensure against launch-induced translation. Both the 25 cm optics cover and the 41 cm cooler cover are ejected at the same time by a common release device. The cover eject command is performed by an electric signal that fires redundant pyrotechnic squibs. These "bellows actuators" remain hermetically sealed after firing to prevent any contamination to the instrument, and through a lever-and- piston mechanism cause a pair of steel cables to be pulled. Each cable simultaneously unlatches a cover by pulling the constraining latch pin. Cover ejection is accomplished using torsion springs mounted at the hinge, causing a cover to rotate open when unlocked, and to slide free under its own momentum once it has opened approximately 100 deg. 3.6 Electronics Design The electronics assembly is mounted on the spacecraft scan platform near the optical assembly and contains the following circuits: analog, digital, scan mechanism and chopper drivers, and power supplies. A block diagram of the electronics showing the interaction of the various sub-modules and their external connections to the Command and Data Subsystem (CDS) and the Power/Pyro Subsystem (PPS) is presented in Fig. 7. The analog subassembly consists of 17 signal processing circuits (one for each detector), a multiplexer, analog-to-digital converter and miscellaneous circuitry such as engineering telemetry and a calibration lamp driver. To maintain detector bias voltage and dark current stability with temperature and time, each InSb detector amplifier (channels 3-17) incorporates a bias correction servo loop which samples the amplifier output signal during the dark-signal portion of the chopper cycle and holds the amplifier output, and thus the detector bias voltage, at a predetermined level. There are two types of signal processing circuits employed in the NIMS (see Fig. 9). For most of the NIMS wavelength interval, the signals are determined by surface or atmospheric albedo and their range can be accurately predicted. Thus one can use ground commands to accomplish the infrequently required gain changes. On the other hand, there are transient, localized hot spots in the Jovian atmosphere which are due to unpredictable cloud clearings. These features allow one to probe the deep atmosphere, showing variable and often intense thermal emission in the 5 micron region, and require a much larger dynamic range than those required for the lower wavelength channels. Commandable gain state switching is used in channels 1 through 14 (albedo channels). These commandable gain states are achieved by switching resistors at the input to the gated integrators. The instrument may be commanded to one of four possible gain states which then selects the appropriate resistors for each of the 14 albedo channels. The gain of each channel has been set for that detector's predicted response at Jupiter. In particular, the nominal gain state (gain state 2) corresponds at full scale (1023 data numbers, DN) to an albedo of ~ 1.2 at Jupiter's solar distance. Gain states 3 and 4 are each more sensitive by a factors of two and four, respectively. Gain state 1 is similar to gain state 2, except channels 10-14 are each reduced in order to obtain measurements from the spacecraft Radiometric Calibration Target (RCT-NIMS, see Section 4.5). Channels 15 through 17 (the thermal channels) use a dual-gain amplifier which automatically switches gain with input level and achieves a dynamic range of 10,000. The two possible gains (low signal level = high gain, high signal level = low gain) for each of the thermal channels has also been optimized for Jovian system measurements, with brightness temperatures of 230 to 325 kelvins measurable in the low gain configuration. A commandable electronic calibration signal can be introduced to verify the gain (in gain state 2) for each of the seventeen signal chains. The digital subassembly contains the central controller and mode sequencing functions necessary to operate the mechanisms in response to external commands, to acquire and format science and engineering data, and to output engineering telemetry and science data via the spacecraft data bus. Science data is transferred at a rate of 11.52 kilobits/sec and engineering telemetry is transferred at a rate of 36 bits/sec. Commands to the mechanisms are initiated at the closing of the chopper blades. Since the chopper is phase-locked to the instrument timing chain, which in turn is phase-locked to the spacecraft real-time interrupt (RTI), all mechanism and other instrument operations are accurately synchronized to the spacecraft clock. The digital subassembly is based on the RCA 1802 microprocessor. Its operating software is interrupt driven and permits a flexible selection of instrument operating sequences. Mode defining parameter tables are used to control the operation of each sequence. Ground commands permit modification of the parameter tables which allows the instrument operation to be uniquely tailored to a specific science opportunity (see Section 5.3). The chopper driver circuitry operates on commands from the digital subassembly and phase-synchronizes the chopper blades to the instrument and spacecraft clocks. The analog signal chain demodulator signal can be derived directly from the chopper (Chopper Reference mode) or from the 63 Hz drive signal (63 Hz mode). There are slight phase differences between the two; the former is the preferred mode, and is slightly more sensitive. The relative timing of the mechanisms is shown in Fig. 10. The circuitry also regulates the physical amplitude of the chopper vane movement. The chopper's resonant frequency is slightly greater than the 63 Hz drive frequency. This allows a third chopper mode, wherein the chopper oscillates at its own natural frequency, with synchronism maintained by software control, occasionally slipping chopper cycles to maintain lock. Both the scan mirror and the grating are actuated by torque motors. Each torque motor is driven by a closed-loop servo circuit which uses the output of a linear voltage differential transducer (LVDT) as the position sensor. The LVDT output is continuously compared against the position requested by the microprocessor. An error signal is generated which is then applied to the torque motor for position correction. To minimize instantaneous power demand, the software sequences commands to the mirror and grating drivers such that the two will not be in motion simultaneously. The power supply consists of current-limited power converters and regulators necessary to supply all voltages needed by the various subassemblies. It also generates a power-on reset signal which controls the initialization of the instrument. At power-on, the instrument is placed in a quiescent mode, with neither the mirror nor grating scanning; the signal chain is set to the nominal gain state 2 and the instrument is synchronized to the 63 Hz clock instead of the chopper. 3.7 Thermal Design The low temperatures required for the FPA and optics posed many unique problems in both instrument design and thermal loads from external sources; the latter is discussed in Section 6.1. There are two general categories of design problems, the first is in achieving the required low temperatures, the second is ensuring that alignment shifts and mechanical stresses are reduced to tolerable levels. The optical assembly is passively cooled by radiation, attaining ~ 150 kelvins during both laboratory measurements as well as inflight. This is accomplished by having an unblanketed instrument and using narrow stainless steel mounting struts with low thermal conduction. The three struts form a kinematic mount, consisting of a monopod, bipod, and tripod. A prestress was applied during mounting to allow for thermal contraction of the instrument relative to the warm scan platform. Thermal conduction by the electrical cabling was reduced using long, minimum diameter wiring. Radiative coupling to the scan platform and nearby instruments is minimized by their thermal blanketing. The optical assembly was painted with a low solar absorptance, high emissivity white paint (zinc orthotitanate). This material is a moderate electrical conductor, providing protection against electrostatic discharge. Optical alignment was carried out at room temperature and it is important that it be maintained at the much lower operating temperature. Most critical are the primary-secondary mirror spacings, which are athermalized using invar rods. Thermal gradient distortion is avoided by directly mounting the optics, except for the grating, to a central optical "bench" of high-thermal-conductivity material (aluminum). The mirror mount design is crucial to the optical quality, since differential thermal contraction can distort, or even break, the optical elements. NIMS uses an inherently athermal mounting scheme. The back surface of each mirror, in the central mounting region, is ground flat, while the front has a conical shape, with the apex of the cone in the plane defined by the back (c.f. Fig. 3). Flat and conical aluminum retainers were machined with the same geometry. Since there is no change in shape with temperature, clearances and clamping forces remain constant. Only a relative sliding motion results, facilitated by 0.001 inch thick mylar sandwiched between the surfaces. 3.8 Instrument Contamination Protection There are a multitude of contamination sources which can degrade, or destroy, NIMS instrument performance. Ground-based sources include water-vapor absorption and deposition of particulate matter on the optics and radiator. These sources are especially pronounced during the launch phase, where vibration frees many particles, and outgassing of organics and other molecules also occurs. During the interplanetary injection IUS burn, the entire spacecraft is subject to impingement by motor products. During cruise, the spacecraft is a source of outgassing water vapor and organics, which can condense on the cold NIMS instrument. Thruster byproducts from trajectory and attitude correction maneuvers are continuing sources of contamination. For all of these reasons, NIMS has adopted a dual approach for contamination protection: first, to incorporate within the instrument as many protective measures as possible (discussed below), and second, to minimize external contamination sources insofar as possible (see Section 6.2). The first NIMS protective measure was to use only low-outgassing materials and to further subject them to high temperature bakeout, prior to assembly. Upon final assembly, the instrument was kept in a dry environment, either in a dessicated container, or it was purged with dry nitrogen. Purge protection was nearly continuously maintained, even during Shuttle operations and launch, finally terminating two minutes prior to release from the Shuttle bay. There were two separate purge paths within the instrument, dictated by relative contamination sensitivities. The low-emissivity surfaces within the cooler are most sensitive to contamination, consequently the cooler has its own purge path, separate from the optics purge. Protective covers were installed over the telescope and cooler apertures, and remained in place, except during calibration and thermal-vacuum testing. They were deployed 77 days after launch, hopefully after most spacecraft outgassing had occurred. During launch and cruise, NIMS tends to be the coldest object on the spacecraft, and is subject to condensation of water and other volatiles onto sensitive surfaces. Heaters were included in order to minimize any immediate condensation, and to subsequently drive off condensates that might be deposited during unheated periods. The optics assembly has two 40 Watt heaters, operated simultaneously, which produce temperatures of 240 - 250 kelvins. This suffices to remove water, but thruster products may be less easily removed, particularly if they have remained on the surfaces and have reacted to form less-volatile species. A heater was also incorporated in the radiator shield (see Section 3.4). It serves the same purpose, in this case protecting the most contamination-sensitive surface on NIMS - the inner surface of the radiator shield. 4. Instrument Calibration 4.1 Introduction For the NIMS instrument, there are three broad calibration categories: spectral, radiometric, and spatial. The majority of the calibration measurements were performed in the laboratory, however there are several important calibration verification activities that will be performed in flight, and discussions of the relevant spacecraft hardware is included below. Unless otherwise noted, the measurements reported here were obtained in the NIMS thermal-vacuum facility, which is a large stainless steel vacuum chamber, evacuated with a liquid-nitrogen-baffled diffusion pump. The NIMS instrument is mounted on an internal table which simulates thermal properties of the spacecraft scan platform. A liquid-nitrogen-cooled shroud surrounds the instrument, allowing the optical assembly to radiatively cool to its flight temperature - roughly 130 kelvins. An additional space background simulator, cooled with liquid nitrogen or liquid neon, was used to cool the radiator. A large area blackbody source was installed in the vacuum chamber and can be rotated into the field-of-view of the telescope. External optical access is provided by two interchangeable window assemblies, one being a single, large diameter quartz window, the second consisting of a mosaic of small diameter calcium fluoride windows. Calibration measurements were performed for a variety of operating conditions and instrument modes. The focal plane temperature, optics temperature, electronics temperatures, and input power voltages were all varied over expected operating ranges. In addition, all appropriate instrument gain states and modes were investigated. 4.2 Spectral Calibration The goal of the spectral calibration is to establish the wavelengths sensed by a detector for each of the 32 possible grating positions over the range of conditions expected in flight. This involves calibrating the NIMS spectrometer itself, prior to launch, and also characterizing an internal spectral light source, to be used in flight to detect any spectral shifts. In addition, when investigating highly detailed atmospheric spectra, it is necessary to know spectral bandpass profiles in order to convolve theoretical spectra to the NIMS resolution. The grating equation, when expressed for the NIMS optical geometry, reads m(lambda/d) = sin(gamma + phi) - sin(gamma - phi + chi), where m is the order of diffraction, lambda the wavelength, d the grating constant, gamma is half of the angle separating the collimator and camera optical axes, phi is the grating rotation angle, and chi is the angular displacement of each detector from the optical axis, chi-sub-i =arctan(x-sub-i/f). Here, x-sub-i is the linear displacement of detector i and f is the effective focal length, and includes refraction effects by the cooler and FPA windows. The grating rotation angle phi = phi-sub-zero + eps(p - delta p) with p being the grating position ranging from 0 to 31, eps is the grating rotation increment, and phi-sub-zero is the grating offset for grating position zero. delta p is included to account for any launch induced shifts; it is defined as zero for the laboratory calibration. Determination of these constants, and any variation in them due to thermal and electrical effects, constitutes one portion of the spectral calibration. Several of these quantities, in particular chi and x-sub-i were well determined during fabrication and assembly, leaving only eps, f, and phi0 to be determined. In order to accomplish these measurements, an auxiliary monochromator was used, which itself was calibrated using the HgI lambda 5461A green line in various orders. This monochromator, used in conjunction with an incandescent source, a diffuser, and a collimator, was used to illuminate the NIMS telescope, forming a diffuse, spectrally narrow image at the entrance slit of the NIMS spectrometer. For a given setting of the external monochromator, a spectrum was obtained by the NIMS instrument. The wavelength setting of the auxiliary monochromator was then changed by an interval small compared to the NIMS spectral bandpass and the process repeated. Thus, for each detector and grating position, one can determine the spectral position and bandpass profile; an example is shown in Fig. 11. Least squares fitting of such data for a large number of grating positions and detectors allows an accurate determination of the calibration parameters, with an observed variance in wavelength of 0.001 microns, or less than one tenth of a grating step. Wavelength checks were also performed using indene and polystyrene absorption features. Temperature variations of the wavelength positions were determined using Hg lines and by varying the optics assembly temperature using the instrument heaters. Comparing the relative displacement of the positions of atomic lines shows that the grating step value, eps, varies by 0.03 % per degree centigrade, an insignificant amount considering the observed constancy of inflight temperature values. Spectral shifts could arise from launch induced vibration or thermal effects. In order to quantify any such deviations, a spectral lamp is contained within the instrument, mounted on the telescope spider, which can be exercised by ground command. The lamp is an InGaAs light emitting diode, which emits a relatively narrow band of radiation, about 0.025 microns wide, centered at 0.8500 microns (at 130 K). The center wavelength is slightly temperature dependent, and there is a platinum resistance thermometer located nearby on the spider which can be used to accurately determine the temperature and wavelength. Using this internal source, any spectral shifts can be measured to an accuracy of better than 0.05 of a grating step. 4.3 Radiometric Calibration The purpose of the preflight radiometric measurements is to determine the parameters which relate the signal S received from the NIMS instrument (in data numbers, DN, ranging from 0 to 1023) to the radiance I of the target. The response of the instrument was found to be linear, with S = S-sub-0 + sigma I and where the sensitivity sigma depends upon several parameters, including detector, wavelength, detector temperature, instrument gain, chopper mode, and polarization of the source. The dark value offset, S-sub-0, depends upon detector, gain state, and other variables. These dark values were simultaneously determined from laboratory measurements but will also be determined in-flight, before and after an encounter sequence. In order to cover the NIMS wavelength range of 0.7 to 5.2 microns, we used two types of light sources, the first being an incandescent tungsten-filament spectral irradiance standard which allows calibration to 2.5 microns. The second source is the aforementioned extended blackbody source mounted within the NIMS vacuum chamber, and provides useful spectral radiance for wavelengths longer than 2 microns. Some details of the two different sources are given below. The shorter wavelength measurements used a 1 kW filament lamp with a quartz envelope containing halogen gas. The spectral radiance of this source was calibrated by EG&G Inc. and this calibration is directly traceable to the National Institute of Standards and Technology (NIST). The lamp was powered by regulated direct current at the prescribed amperage, measured using NIST-traceable instruments. In order to produce an extended source of known radiance, a large-area Halon target was constructed according to the prescription of Weidner and Hsia (1981), illuminated by the standard lamp. We use Weidner and Hsia's (1981) measured directional/hemispheric spectral reflectance values and their bidirectional reflectance data to find the reflectance for our particular geometry: normal incidence and ~ 20 deg emission angle. This target was viewed by the NIMS instrument through the quartz window, whose transmission was independently measured. The lamp was placed at various distances from the target, allowing a test of the linearity of response and yielding a precise determination of the instrument sensitivity. The precision of the sensitivity determination was found to be a fraction of a percent. The accuracy of the derived values is estimated to be about 10 percent, partly due to uncertainties in the original lamp calibration (< 5%) and partly due to the bidirectional reflectance of the Halon target. This target is currently being compared to a Labsphere Spectralon standard, itself calibrated with traceability to NIST. Measurement of the instrument sensitivity for the longer wavelength region was performed using an extended blackbody source, with a diameter larger than the NIMS telescope aperture. This source is a V-groove radiator, electrically heated and regulated to maintain a constant, preset temperature. The physical temperature of the radiating surface, which exhibits an emissivity of 0.99, is measured by two copper-constantan thermocouples, each relative to the ice point established by a distilled water ice bath. The thermocouples and potentiometer used were all calibrated to NIST-traceable standards. The heated target is loosely thermally-coupled to a liquid nitrogen-cooled heat sink, allowing a controllable temperature range of 200 to 350 kelvins to be achieved. As with the previous lamp, measurements were obtained for a variety of source settings, in this case different temperatures, and the data fit in the sense of least squares to find sensitivity values. The precision in this determination is generally a fraction of one percent. Thermal gradients of ~ 0.5 kelvins occur over the surface, limiting the accuracy of the derived sensitivity to 5 percent. In the overlap region, 2.3 to 2.5 microns, the difference in sensitivities found using the two different sources is 12.5 percent. The instrument uses a diffraction grating, which are known to show efficiency differences for different polarizations. The polarization sensitivity was checked throughout the entire operating range using dichroic and wire grid polarizers. The maximum difference in sensitivity for two orthogonal polarizations was found to be only 5 %, and this occurs only in the vicinity of the grating blaze wavelengths, as predicted by scalar diffraction theory of gratings (Strong, 1958). 4.4 Spatial Calibration In this section we discuss aspects of the spatial calibration, which includes the pointing geometry, the angular resolution and angular sensitivity profile, and scattered light rejection. The NIMS spatial scan pattern is twenty pixels aligned along the cross-cone direction of the scan platform and formed by stepping the telescope secondary mirror through twenty positions. Measurement of the angular location of each pixel was accomplished using an illuminated slit, mounted in the focal plane of an external collimator. By translating this horizontally oriented slit in the vertical (cross-cone) direction one can find the point of maximum response, and thus the angular location for each of the twenty mirror positions. It was found that the angular locations are well represented as a linear function of the mirror position number, with an angular step size of 0.528 mrad. Using a very narrow slit, either horizontally or vertically oriented, and translating it in the orthogonal direction, one can obtain a measure of the angular sensitivity profile in the spacecraft scan platform cone and cross-cone directions, respectively. An example is shown in Fig. 12. It can be seen that the response in cone angle approximates a rectangle, as would be expected since this response is determined by the entrance slit of the NIMS spectrometer which is in the telescope focal plane. The response in the cross-cone direction is more triangular, and is the natural result of spectrometer astigmatism, which gives a slightly defocused vertical image of the detectors at the telescope focal plane. The co-alignment of the Galileo scan platform experiments (NIMS, SSI, UVS, PPR) was measured prior to launch. The co-alignment of the NIMS and SSI experiment was found to be within 0.3 mrad, well within specifications. Simultaneous tests using stellar sources will be used to verify post-launch alignment. For a variety of NIMS measurements, the instrumentally scattered light is of concern. Examples of observations which could be affected by scattered light include limb scans, dark feature measurements, terminator scans, and Jupiter dark side imaging. Although the best time to measure the scattered light properties of the instrument is during flight, at which time any launch and contamination influences will be included, prelaunch measurements of stray light effects were also performed. These tests, reported here, are only upper limits to the actual performance, since additional contributions from the collimator and both surfaces of the quartz window are included. Two different tests were performed to evaluate the scattered light, the first being a "knife edge" test and the second a "black hole" test. In the "knife edge" test, the mirror motion scans across a sharp boundary in the focal plane of the external collimator, the upper half-plane being a brightly illuminated surface while the lower unilluminated portion is made as dark as possible using a black surface oriented to form a cavity. For this geometry, the near-field (1 to 5 mrad from the boundary) scattered light contribution is detector dependent, varying from 0.1 to 0.4 percent. The disadvantage of this test is not knowing just how dark the lower boundary really is, and is partially solved using the following test. In the "black hole" tests, the instrument and collimator are focused at the center of a hole contained within a large, diffusely reflecting plate. An absorbing conical blackbody cavity is placed behind the plate, and the entire assembly is illuminated by a source placed in front and displaced from the axis of the conical absorber. Radiation reflected by the plate, which does not enter the collimator, is absorbed by optically black baffles. By using plates with the same hole size but with varying albedos, radiation emanating from the center of the hole is constant, while the instrumentally produced stray light varies with the surface albedo. Using plates with various albedos and hole sizes, the near-field scattered light response was found to be 30 % less than that for the "knife edge" tests. Again, all of these measurements include extraneous contributions from the three intervening optical elements. Definitive tests will be performed inflight, however preliminary results based upon the Venus and Earth limb scans indicate excellent scattered light rejection. 4.5 Spacecraft Radiometric Calibration Target Calibration targets are provided on the Galileo spacecraft for inflight calibration verification of the remote sensing instruments, and to monitor the relative response throughout the mission by performing periodic calibration observations. All of the remote sensing instruments can use the Photometric Calibration Target, discussed in the next Section. In addition, there is a Radiometric Calibration Target (RCT-NIMS) which is intended for verification of NIMS performance in the long wavelength region, and is described in the following paragraphs. The NIMS Radiometric Calibration Target is a near-field, extended, blackbody source, mounted on the scan platform sunshade and in front of the NIMS telescope when the scan platform is in the zero degree cone angle position. When used as a calibration source, the target is heated with 25 Watts of electrical power, elevating the target surface to ~315 kelvins. This provides a known radiance which can be used for radiometric calibration for wavelengths longer than 2.5 microns. The emitting surface, slightly larger than the telescope aperture, consists of a mosaic of hexagonal honeycomb cavities, each 0.25 inch in width and 0.50 inch in depth. An infrared black paint (Chemglaze Z004) was applied, which, with the cavity geometry, gives a normal emissivity of greater than 0.98 (Sparrow and Heinisch, 1970). Thermally insulating rods are used to mount the target, and a strip heater is employed to provide uniform heating and temperatures. Temperature differences across the surface were measured using thermocouples, with a root-mean-square deviation of 0.7 kelvins being found. Upon heating, the target reaches equilibrium temperatures, within 0.5 kelvins, in a time period of 6 hours. The physical temperature of the target is measured using the spacecraft Command and Data System (CDS) engineering telemetry. Two temperature sensors are employed, consisting of platinum and nickel resistance thermometers, and both were calibrated with NIST traceability. At the target operating temperature, both sensors exhibit nearly the same resistance, roughly 550 ohms. In order to calibrate the CDS circuitry itself, a temperature insensitive resistance of 562 ohms is also measured. All three measurements use the same current source and measurement circuitry. Consequently, temperatures can be measured to an accuracy limited only by the digitization interval, ~ +- 0.3 kelvins. If we assume a combined temperature error from all sources of +- 1 K, then radiometric accuracies of 7 and 4 percent will be achieved at 2.5 and 5 microns, respectively. Comparison of flight RCT data with pre-launch measurements indicates stable instrument performance. 4.6 Spacecraft Photometric Calibration Target The Photometric Calibration Target (PCT) and an associated optical element, the Photometric Calibration Mirror (PCM) together form a source of diffusely reflected solar radiation which can be used by the remote sensing experiment for intra- and inter-instrument comparisons. These two elements are mounted on the Science Boom; the mirror reflects solar radiation onto the diffusing target surface which is placed outboard from the mirror and in a position that can be viewed from the scan platform. In order that the target be illuminated over the nominal range of solar cone angles, the mirror is convex, with a radius of curvature of 46 cm and located 53 cm from the target. The reflecting surface is vacuum deposited aluminum, with a protective overcoat formed by its natural oxide. The target surface is similar to the Voyager diffuser plate, consisting of sand blasted aluminum. The combination produces a spectrally gray diffuse surface, with an effective albedo of roughly 0.05. They were calibrated over the spectral range of 0.3 to 5.2 microns and for a variety of incidence angles and azimuths. 5. Operating Modes and Data Acquisition 5.1 Instrument Modes and Operation For most Galileo observations, the time available is limited, and one must tailor each observation for specific scientific goals. For the NIMS experiment, this translates into optimization of the spatial and spectral sampling aspects. For example, atmospheric measurements usually require the best available spectral resolution, whereas surface reflectance spectra are generally broader, allowing coarser spectral sampling. In addition, the spatial coverage and resolution demands are quite different for Jupiter and satellite measurements, the latter requiring much more rapid spatial sampling during the short amount of time available. Not only are there internal spatial/spectral tradeoffs to be considered. but, in addition, it has been a longstanding goal among the Galileo remote sensing experiments to perform coordinated and compatible observations through simultaneous use of the scan platform. With these considerations in mind, we have developed a flexible set of instrument modes, described in this Section. The relevant NIMS instrument parameters that can be adjusted for differing observations are mainly spectral, determining the number of wavelengths to be sampled and their relative placement. In one extreme, the entire spectrum is obtained at full resolution, at the other extreme, the grating is fixed and only one wavelength band is sampled for each of the seventeen detectors. Intermediate combinations are possible, each with differing times required to complete a spectrum. Throughout this time, mapping is being accomplished by scan platform motion, with spatial and spectral sampling occurring simultaneously. During the time required to form a complete grating motion cycle, the scan platform will have moved some fraction of a NIMS spatial resolution element (0.5 mrad). In the following, contiguous spatial sampling corresponds to a motion of one spectral sample per spatial resolution element, Nyquist sampling is twice as frequent. NIMS modes are implemented in the instrument software using parameter tables (PTABs). There are two such tables, each describing a specific spectral measurement sequence. Use of two PTAB tables allows for hybrid combinations, giving flexibility to instrument sequencing. The assignment of an individual parameter table (PTAB) is shown below (Fig. 13) and described as follows: N is the number of grating positions per cycle, ranging from 1 to 24, D is the grating angle step size, unity corresponding to a single step of one-half of a spectral resolution element, and S is the grating start position. The number of times to repeat a given spectral sequence is given by the parameter R. Additional parameters include M, mirror scanning (on or off), and A, autobias (on or off). There are a total of sixteen modes available to NIMS, twelve of which are pre-defined in the instrument read-only-memory (ROM). An additional four modes, yet undefined, can be placed in the instrument's random-access-memory (RAM) via uplink commands. All mode parameters can be changed by ground commands. The standard ROM modes are summarized in Table 2. The NIMS instrument modes were designed to be synchronous with spacecraft timing and its various time units. The largest unit is a RIM (or MAJOR FRAME) which is 60 2/3 seconds, subdivided into 91 MINOR FRAMES, having whole fractions of 1/91 (2/3 seconds), 1/26 (2 1/3 seconds), 1/14 (4 1/3 seconds), 1/7 (8 2/3 seconds), and 1/2 (30 1/3 seconds). There are NIMS modes synchronous with all of these fractions except 1/2, which is that used for a compressed imaging mode for the SSI instrument. Table 2. NIMS Standard ROM Modes. The execution time, given in sec, is the time required to complete a grating cycle. The Nyquist slew rate is the scan platform cone angle rate necessary to move through one-half of a NIMS pixel (1/2 of 0.5 mrad) in the execution time. It is given in microradians/sec. Mode Execution Nyquist Number Name N D M Time Slew Rate ====== ======================== === === ======= ========== ========= 0 SAFE 0 0 0ff 1/60 15000 1 FULL MAP 12 2 Scanning 4 1/3 60 2 FULL SPECTROMETER 12 2 Off 4 1/3 60 3 LONG MAP 24 1 Scanning 8 2/3 30 4 LONG SPECTROMETER 24 1 Off 8 2/3 30 5 SHORT MAP 6 4 Scanning 2 1/3 110 6 SHORT SPECTROMETER 6 4 Off 2 1/3 110 7 FIXED MAP 1 0 Scanning 1/3 750 8 BAND EDGE MAP 2 - Scanning 1 1/3 190 9 BAND EDGE SPECTROMETER 2 - Off 1 1/3 190 10 STOP&SLIDE MAP 24,6 1,4 Scanning Variable Variable 11 STOP&SLIDE SPECTROMETER 24,6 1,4 Off Variable Variable The instrument steps through a grating cycle (N positions of the grating) then repeats R times (Mode Repeat Count). When the repeat count is consumed, then the other PTAB controls the instrument mode. This process continues as long as the instrument is powered on. The instrument software has specific branching instructions which depend mainly on the value of N, the number of grating positions. These instructions results in somewhat different timing for the scan mirror, as well as the number of steps required to complete a cycle. N = 12 results in 13 spatial scan cycles (1/3 second each) and a total grating cycle time of 13/3 = 4 1/3 seconds. The scan mirror rests at the starting point during the 13th interval, during which time the grating returns to its starting position. If N = 24, then this number of grating positions are positioned, followed by two scan mirror rest times, during which the grating is reset to the origin, giving a total cycle time of 26/3 = 8 2/3 seconds. The NIMS instrument modes in ROM are named according the the following conventions. The MAP mode indicates the spatial mirror is scanning, SPECTROMETER indicates the spatial mirror is off. LONG, FULL, SHORT (which is sometimes called PARTIAL) and FIXED refer to the number of grating positions in a single grating cycle (24, 12, 6, and 1 respectively). The FIXED mode has been embedded in spacecraft documentation as FIXED GRATING; however the systematic name consistent with the definition in ROM would be FIXED MAP. The BANDEDGE mode has one grating position per PTAB, with a different grating position in each PTAB, alternating between PTABS each 2/3 second. The STOP&SLIDE mode is a hybrid mode, combining the FULL mode in one PTAB with a FIXED mode in the other. In the MAP modes the instrument has the spatial mirror turned on. This yields an effective field of view of 20 pixels, arranged in a linear stripe of 10 mrad x 0.5 mrad in the cross-cone direction. These 20 pixels are measured sequentially, and all within 1/3 sec. The initial scan motion, at the beginning of a RIM, is downward (toward the scan platform), then up for the next grating step. The SPECTROMETER mode, for which there is no internal spatial scanning, is used for certain observations, for example atmospheric limb scans, when only one-dimensional image scanning using the scan platform motion is required, or when redundant sampling of a given location can be used to advantage. In general, the scan platform rates are the same as those for the MAP modes, since the same amount of time is required to complete a grating cycle. The only exception is between the FIXED MAP and SAFE modes. Among the various grating modes, the LONG mode provides the best spectral resolution, with two spectral samples per spectral resolution element (spectral Nyquist sampling). This mode requires 8 2/3 seconds to complete, with seven spectra contained in a RIM. For this mode, Nyquist spatial sampling requires ~30 microrad/sec slew rates, yielding two complete spectra while crossing through a NIMS field of view (0.5 mrad). It may be difficult to match this mode with simultaneous imaging (SSI) observations when time is the limiting factor, since the slew rate for a nominal SSI frame is about 120 microrad/sec. One strategy is to use alternate imaging filters in a sliding mosaic. The FULL mode gives lower spectral resolution, with only one spectral measurement per spectral resolution element, and is useful for diffuse solid surface reflection features. This mode permits a coordinated observation with other scan platform instruments using a slew rate of about 110 microrad/sec, which is twice the preferred spatial Nyquist sampling rate for NIMS. The SHORT mode under-samples the NIMS spectral capability by a factor of 4, using only 6 grating positions. This mode is a compromise which was included to work with the SSI camera's compressed mode and to permit obtaining some spectral information when the time to complete a mosaic is extremely limited. Some optimization is possible by matching the starting grating position for a cycle (and thus the spectral "comb") to spectral bands expected on the observed surface. The FIXED mode reduces the number of grating positions to 1. This mode is spectrally minimal, but ensures excellent spectral registration for the wavelengths that are measured. It proved to be very useful in Venus darkside observations. The BANDEDGE mode alternates between two grating positions, providing two stripes (20 samples at one grating position over the 17 detectors which are approximately evenly spaced over the spectral range of the instrument) at different grating positions. This mode, for instance, alternates sampling between the continuum and the maximum of a spectral feature. It is useful for spatial mapping a particular spectral feature at a high spatial scan rate. This mode samples one stripe every 2/3 second - 1/3 second for the stripe and 1/3 second preparing for the next grating position (defined by the next PTAB). The STOP&SLIDE mode is a combination mode - normally a combination of the FULL mode and the LONG mode. Its purpose is to provide a mode compatible with multi-color sequences for the SSI framing camera. The sequence works as follows: The scan platform is stopped and NIMS enters its highest spectral resolution mode (LONG) while the SSI instrument is acquiring frames in several filters. The scan platform then slews slowly over to the next picture position (one overlapped SSI FOV) with NIMS in a mode compatible with the slew rate (nominally FULL mode), in a time period which corresponds to the readout time for the framing camera. The Mode Repeat Count parameter is set to accomplish this compatible sequence. The result for SSI is a multi-color image. The result for NIMS is a spatially complete spectral map of the target at modest spectral resolution, and spatially sparse high spectral resolution samples with an best spectral resolution. This optimizes the use of scarce scan platform time, characteristic of close encounter geometries, when multi-color framing mosaics are being performed. In addition to the ROM modes discussed above, four of the 16 instrument modes may be loaded from RAM. The PTABs for these modes are defined by ground commands to the instrument and are stored in RAM. The RAM mode may then be loaded into the active area with a single command. The RAM mode remains valid as long as the instrument is powered on and the instrument is not reset. After the two PTABS which define a mode are loaded into the active area, the active area may be modified to change the characteristics of the mode. The appropriate command modifies one of the four PTAB values in each loaded PTAB (independently) with each invocation of the command. This capability is particularly important for the BANDEDGE and STOP&SLIDE modes. The BANDEDGE grating start positions and the STOP&SLIDE duration in the high spectral resolution mode often need to be modified from those defined in the instrument Read-Only Memory (ROM). 5.2 Instrument Commands In addition to the various modes that the NIMS instrument is capable of executing (described above), there are a variety of instrument states which can be used for measurement or calibration. Of prime concern is the instrument gain state, for which there are four (see Section 3.6). In addition, two different calibration sequences can be executed: an OPCAL command, for which an internal electroluminescent diode provides a wavelength reference, and an ECAL command, which injects a known signal into the electronic amplification chain. The commands are labeled by instrument number (NIMS=37) and a mnemonic, which indicates the type of command sent. One NIMS command will be executed in a given spacecraft RIM; in the event of multiple commands, the last command loaded will be executed. The command must be loaded into the instrument command buffer by minor frame 89 of the previous RIM. The following is a brief description of available commands. 37IOP - Instrument Operation: This command loads a NIMS mode (set of PTABS) from ROM or RAM to the active area. It also permits specification of the grating start position (loaded into both active PTABS). 37IST - Instrument Status: This command modifies the gain state (4 available), the chopper state (63 Hz, chopper off, chopper reference, and free run), and can invoke the electronics calibration or the optical calibration. The electronics calibration should be carried out in gain state 2. The calibration lamp (a light emitting diode) measurement should be done in gain state 4 and LONG SPECTROMETER mode. 37MPT - Modify Parameter Table: This command modifies one of the 4 parameters in the PTAB for both of the PTABS in the active area. The value for each PTAB is specified independently in the command (but the same parameter is modified in each PTAB). It is possible to turn off the thermal channel autobias with this command - a capability which is intended for use only where the focal plane is approximately at room temperature. 37GO - Grating Offset: This command sets the grating offset. Acceptable values are 0 - 7; the default is 4 which is entered whenever the instrument is turned on. Note that this grating offset is different than the PTAB grating start position parameter. 37SS - Special Sequence: This command programs the RAM modes. It must be invoked twice, once for each PTAB in a mode. 37IRT - Instrument Reset: This command permits resetting the instrument without cycling instrument power. 37MN - Memory Normal: This sets the instrument CPU to use the normal ROM address space. 37MRL - Memory Reallocate: This sets the instrument CPU to branch to the RAM address space. 37PL - Program Load: This permits loading programs into the spare RAM. 5.3 Typical Encounter Operations During each of the eleven Jupiter orbits, there will be satellite flyby opportunities which occur on time scales of hours, demanding efficient usage of the various instrument capabilities in order to maximize the scientific return. Observations of the Jupiter atmosphere occur over longer time scales, measured in days, but the number of important features to measured, their angular sizes, and temporal measurement frequency also demands efficient instrument mode usage. The following discussion illustrates, in a simplified fashion, the NIMS use of modes during a typical orbit at Jupiter. General sequencing priorities are 1) keep NIMS in LONG mode as much as possible to optimize spectral resolution, 2) maximize spatial coverage and resolution, and 3) include as many scan platform instruments as possible in an observation - to maximize synergistic science return. On a typical Jovian orbit, Jupiter becomes visible over the spacecraft sunshade at about 25-40 Jovian radii (RJ). At this distance it is possible to mosaic the entire planet - at full spectral resolution - in a reasonably short time. At periapsis, occurring at about 10 RJ, the angular size of Jupiter is quite large, and it is not possible to mosaic the entire planetary disc at full spectral resolution. Similarly, for a close Galilean satellite encounter, it is possible to fully map the satellite at highest spectral resolution at about 4 hours out, but at closest approach the angular size and surface-relative smear rates become very high, forcing a choice between spatial and spectral mode coverage. On the inbound portion of a Jupiter orbit, the NIMS instrument would be in LONG MAP mode to mosaic the Jupiter day and night side, occasionally executing the STOP&SLIDE mode for compatibility with SSI multi-color images of the planet. Many of the mosaics will be oversampled vertically, which is required for overlapping fields-of-view for the thermal instrument (PPR). As perijove approaches, the LONG MAP mode is maintained for atmospheric measurements, but the areal coverage is reduced to include only a few specific features. These features will be consistently measured throughout the orbital pass, yielding their temporal and phase-angle variations. It may be useful to develop some RAM modes for these atmospheric measurements, particularly if one is interested in only a portion of the complete spectrum available to a detector. It is possible to maintain high resolution, i.e. using a grating delta parameter of D = 1, but use a lesser number of grating positions N and an appropriate grating start position S to choose the spectral region of interest. In doing so, the observations will encompass less time, allowing more features to be examined. For the satellite encounters, a full-disk mosaic will be acquired each time the spatial resolution changes by a factor of ~2. Thus, a LONG MAP mosaic would be followed by a FULL MAP mosaic, which would be followed by a SHORT MAP mosaic, as spectral resolution is traded for spatial resolution. The STOP&SLIDE mode would be invoked as the framing camera acquires it's multi-color mosaics. At closest approach, the FIXED mode would be utilized to compensate for the high apparent rate of the satellite surface motion. 6. Spacecraft Interactions 6.1 Thermal Control Integration of the NIMS instrument into the Galileo spacecraft design involved a number of new and unique problems for a remote sensing instrument. The greatest problem was minimizing spacecraft thermal loads on the instrument and its radiator, so that it would be able to cool passively to the very low operating temperatures desired. Also, because the Galileo configuration changed several times to accommodate Shuttle and upper-stage delays and different launch opportunities, it was necessary to pay constant attention to changes which might have negative impacts on the instrument's performance. One problem involved the location of NIMS on the despun scan platform, relative to the other remote sensing instruments. The desired location for NIMS was at the right edge of the platform so that the side-mounted radiator had a clear view towards space. In addition, the end position prevented NIMS from being sandwiched between two warm, thermally blanketed instruments, and allowed the instrument body to radiate to space both above and to the side of it. However, this desire conflicted with the Photopolarimeter-Radiometer (PPR) which also wanted the outboard platform position so that it would be able to view around spacecraft structure down to relatively small angles from the Sun. The solution was to mount the PPR on a downward extension of the scan platform so that it was to the right of, and below NIMS, satisfying the requirements of both instruments. During spacecraft assembly it was discovered that the PPR telescope barrel extended into the field-of-view of the NIMS radiator. Located at the end of the Photopolarimeter/Radiometer (PPR) barrel was a thermally actuated hinge mechanism which would have provided a significant thermal source for the NIMS focal plane. The PPR cover was rotated so that its actuating mechanism was out of the FOV of the NIMS radiator. The small fraction of the blanketed telescope barrel still viewed by the radiator was estimated to cause an increase in focal plane temperature by approximately 0.5 kelvins. The greatest thermal problem for NIMS came from the two radio-isotope thermoelectric generators (RTGs) used to produce the spacecraft's electricity. Mounted on booms on the spun side of the spacecraft, each RTG radiates more than 4,000 Watts of heat. The NIMS instrument body and radiator had clear views of the RTGs. Thermal modeling showed that when the radiator viewed the RTG's spinning through its field of view, it would be heated to at least 120 kelvins, well beyond the operating range of InSb detectors. The solution was to implement RTG shades mounted on the booms, blocking direct views to NIMS. The shades had to be carefully designed because the RTG's depended on a clear view to space for cooling; if the shades reflected back too much thermal energy the electrical power of the spacecraft would be degraded. Solar loads on NIMS at Jupiter were not expected to be a significant thermal source. Flight rules required that the NIMS radiator not be illuminated by the Sun prior to observation sequences. The scan platform sunshade provided protection when the platform was stowed at 0 deg cone angle. However, after construction it was discovered that a small slice of the NIMS radiator was not shadowed by either the radiator shield or the sunshade, when the platform was at 0 deg cone. The solution of this problem would nominally have been not to leave the platform parked at low cone angles. However, the spacecraft bus sunshade added for the 1989 VEEGA mission provided shading of the radiator. Early flight experience with NIMS showed that the attention given to minimizing spacecraft thermal loads was successful. However, Venus operations revealed a new problem. Venus darkside observations by the PPR and NIMS required the instruments to "shoot through the booms" of the spinning section, and allowed the NIMS radiator to view a significant amount of warm spacecraft structure. Because of the small heliocentric distance, some spacecraft structures were much warmer than expected. The first PPR observation, 16 hours before Venus closest approach, left the scan platform at a cone angle of 27 deg, giving the radiator a view of the back of the bus sunshade. Apparently, the back of the shade was warmer than predicted and the NIMS focal plane was heated from 64 kelvins to about 95 kelvins, beyond its desired operating temperature. Real-time commands were sent to the spacecraft, which moved the platform to a safe position, enabling the radiator to view deep space. The focal plane cooled to 86 kelvins by the time of the first NIMS observation. Because the NIMS observations of Venus were at larger cone angles than the initial PPR observation, the radiator no longer viewed the bus sunshade, and the focal plane did not heat beyond 88 kelvins during the critical nightside observations of Venus. Flight rules were subsequently changed to require that the scan platform not be left at low cone angles where the radiator could view spacecraft structure, unless science observations were actually underway. Tests were also planned for calibrating the thermal load from the various spacecraft structures, using NIMS itself to measure the temperature of each spacecraft element. 6.2 Contamination Control A major concern of the NIMS experimenters was possible contamination of the instrument from various spacecraft sources. Because it was unblanketed and because it operated at such low temperatures, it was feared that NIMS would serve as a cold trap for volatiles outgassed from the spacecraft and from thruster plume byproducts. In addition, the performance of the NIMS radiator shield was closely tied to maintaining the emissivity and reflectivity of the shield surface -- contamination could seriously threaten the ability to cool the NIMS focal plane. An examination of past experience with Viking and Voyager showed several unexpected problems of this type. For example, outgassing of the Viking lander capsule was so severe that it resulted in nongravitational accelerations on the spacecraft orbit for the first six months after launch. On Voyager, thruster plume impingement on spacecraft structures resulted in a 20% inefficiency in spacecraft maneuvers. Most alarming to the NIMS experimenters was data that came from the Infrared Thermal Mapper (IRTM) and Mars Atmospheric Water Detector (MAWD) on Viking. These instruments had a common calibration target which consisted of a small sandblasted aluminum plate, illuminated by the Sun. It was not possible for either instrument to view the target prior to deployment of the Viking landers. When they were finally used, it was found that each target had decreased in albedo by several percent, and continued to decrease over the life of the mission. Spectral data from the two instruments allowed the investigators to determine that the contaminating material was reddish in color. Numerous sources for the contamination were considered, such as dust in Mars orbit, but were rejected. The studies concluded that the most likely contaminant was byproducts from the Viking main propulsion engine. This engine used a bipropellant combination of monomethyl-hydrazine and nitrogen tetroxide, and one of the byproducts was monomethyl-hydrazine nitrate, a dark reddish material with extremely low volatility, particularly after exposure to sunlight. Surprisingly, the IRTM/MAWD cal target did not have a direct view of the Viking main engine, and was located about 160 degrees from the engine thrust centerline. However, plume expansion in vacuum was known to carry contaminants to locations behind the main engine, and even able to expand around spacecraft structures. Galileo used the same type bipropellant system as Viking, not only for the main engine but also for the attitude control thrusters; Viking used a cold gas attitude control system. Extensive studies were undertaken by the Galileo spacecraft team to quantify the expected contamination from thrusters and the main engine. These studies showed that significant contamination could occur on the scan platform instruments and on the calibration targets. As a result, a number of protective measures were instituted. The first protection was shields around the thruster clusters to prevent a direct line-of-sight to the scan platform, and a similar shield between the platform and the 400 newton main engine. Secondly, minimal contamination positions were determined for the platform for either thruster or main engine firings. The platform was commanded to the respective safe positions prior to any thruster use or maneuver. Third, the thrusters on the spinning section were limited to firing during passage through relatively narrow arcs centered approximately 90 deg from the despun scan platform position. Fourth, the NIMS and calibration target heaters were kept on whenever possible to prevent condensation of volatiles on the surfaces, and to sublimate any materials that had condensed. This latter measure also provided protection from spacecraft outgassing early in the mission. The primary outgassing product is water from thermal blankets. Deployable covers were incorporated into the NIMS design for both the telescope and the radiator, fired by a single, redundant pyro device. The covers were retained until 77 days post-launch, providing protection during the initial period of spacecraft outgassing, the first two trajectory correction maneuvers, and during many other early pyro events on the spacecraft. In addition a dry nitrogen purge was provided for the instrument during ground testing and spacecraft assembly. This purge was maintained during the Shuttle launch and in Earth orbit, until a few minutes before deployment of the Galileo/IUS stack from the Shuttle bay. 6.3 Instrument Pointing NIMS is a whisk-broom imager with one spatial dimension being provided by its scanning secondary mirror, and the other dimension provided by either spacecraft motion or movement of the scan platform. The ability to move the scan platform very smoothly at a predetermined rate relative to a target was thus very important for the success of the experiment. This was a somewhat new concept to the attitude control engineers on Galileo whose previous experience was with imaging systems that typically worked in a "stop-and-shoot" mode. Numerous discussions were held with spacecraft engineers to perfect the pointing requirements for NIMS and the other remote sensing instruments. In addition, discussions were held with the other remote sensing science teams to determine compatible instrument modes where all instruments could take data simultaneously. Both the UVS and the PPR were also interested in taking data in a continuous slew mode, whereas the Imaging system still preferred a "stop-and-shoot" system. Scan platform capabilities and instrument modes were developed to accommodate both situations. The nominal FULL MAP operating mode for NIMS called for a continuous Nyquist slew rate of 60 micro rad/sec. Other desired rates were 30 micro rad/sec for LONG MAP, 110 micro rad/sec for SHORT MAP, and 750 micro rad/sec for FIXED MAP. These are all relatively slow rates compared to the maximum scan platform capability of 17500 micro rad/sec (i.e. a degree per second). Thus, the scan platform performance was optimized for slew rates less than 3000 micro rad/sec. Maximum allowed deviation from a desired slew path was set at 125 micro rad, or one-quarter of a NIMS pixel. Since some satellite flybys could result in fairly large target smear rates, target motion compensation was added to the pointing system. This allowed the spacecraft attitude control system to take out relative target motion without the need for extensive ground design of sequences. 6.4 Spacecraft Obscuration The dual spinner design of Galileo led to a situation where much of the sunward hemisphere of the sky is often obscured by spacecraft structure. The obscuration of NIMS by the spacecraft is shown in Fig. 14. It is not possible to synchronize instrument operation with the spinning booms, so observations that view through the spinning structures must be repeated to fill in gaps caused by boom obscuration. Although NIMS can view down to less than 30 deg cone angle, limited by the bus shade, obscuration by other structures is considerable at that angle. Considerable effort was devoted to keeping the anti-sunward hemisphere completely free of obscuring structure. The only spacecraft part that extended into this hemisphere was the ends of the Plasma Wave Spectrometer (PWS) antenna, located at the end of the magnetometer boom. However, the switch to the 1989 VEEGA mission required a second low gain antenna on the aft end of the spacecraft. This was accomplished by hanging a deployable boom off one of the RTG booms, and extending down to about 114 degree cone angle. This additional obscuration was accepted because LGA-2 would not be needed after the second Earth gravity assist flyby, and the boom would be folded out of the way and not used again. However, Venus and Earth observations near 90 degree cone angle were degraded by this additional obscuration. Careful attention was paid to minimizing sources of glint from spacecraft structures, to minimize scattered light during observations. Since most of the spacecraft was expected to be covered in black thermal blankets, this was not a very difficult problem. However, the switch to the 1989 VEEGA mission required the addition of gold foil blankets to many sun-facing surfaces to minimize heating at the expected post-Venus perihelion distance of 0.69 AU. Integration of these changes with science concerns was not as optimal as it had been earlier in the mission, though science suggested changes were accommodated in many cases. During the Venus flyby, glint was observed from the edges of some of the spacecraft booms, however it was weak and from very narrow regions, and did not degrade any of the data. 7. NIMS Mission Design Aspects Design of the NIMS experiment required the opportunity to view targets at optimum geometry for obtaining spectral images. For solid surfaces, this meant viewing at low phase angles, preferably less than 30 deg phase, where shadowing of target surfaces would be minimal. This conflicted sharply with Imaging Team desires to view surfaces at phase angles near 75 deg where shadowing would help to highlight surface topography. Another conflict involved the NIMS team desire to emphasize global mapping of satellite surfaces at resolutions of 25 km/nimsel or better, whereas the Imaging Team was interested in obtaining the highest possible resolution of surface features, at better than 1 km/line-pair. The approach and departure hyperbolae to satellite flybys afforded good opportunities for global mapping by both investigations. However, close flybys, desired by both Imaging and by the fields & particles investigations, made it impossible to view the entire sub-spacecraft hemisphere. In addition, there was not sufficient time to image the sub-spacecraft hemisphere at closest approach because of the high flyby velocity and high smear rates. The solution to many of these problems involved "non-targeted" or "accidental" flybys of satellites. Because of the resonant motion of the inner three Galilean satellites, the spacecraft would often fly close, within about 105 km, to one satellite when being targeted to a gravity assist encounter with another. If these passes occurred on the sunlit side of the satellite they provided good opportunities for global mapping at resolutions of better than 50 km/nimsel, or 2 km/line-pair for the SSI. In some ways these non-targeted flybys were even better than global mapping on the approach hyperbolae to targeted encounters, because the range to the target changed more slowly, and there was a greater possibility for viewing at relatively low phase angles. Again, interaction with the engineers designing the satellite tour at Jupiter was an important factor in optimizing the mission for NIMS. Analysis programs were written to estimate mapping coverage for each satellite prospective tour, and these results were compared and suggestions for improvements made. This was a difficult exercise because of the limits of the trajectory changes possible using gravity assists, and because of the large number of science requirements from the different investigation teams. Jupiter observation geometries were largely dictated by the satellite tour selected, though they were also considered in the tour design. Because the planet covered such a large solid angle near spacecraft periapsis, special sequences needed to be worked out to allow all the remote sensing instruments to cover a maximum area of the planet simultaneously. Flyby trajectories at Venus and the Earth were optimized for gravity assist and had to be used by the science instruments without modification. Even with these restrictions, it was possible to construct valuable science sequences. These sequences proved to be very useful in conducting complete tests of the Galileo sequencing, commanding, downlink, and data reduction systems. Many problems were discovered and solved which might have not been detected until much later in the nominal mission. The value of exercising the spacecraft and instruments on real targets, prior to arrival at Jupiter, cannot be overstated. ACKNOWLEDGEMENTS Development of the NIMS instrument, along with incorporation of necessary spacecraft accommodations, involved the work of many people, whose unique contributions ranged from project management to craftsmanship in soldering. All of these skills are necessary to NIMS' success, and it is unfortunate that we cannot cite each individual for their own contributions. We would, however, like to specifically recognize and thank members of the initial Galileo Project management team who provided encouragement and advice during the development stages of NIMS: John Casani, Ron Draper, Jesse Moore, and Bill Fawcett. This work was supported by NASA Contract NAS 7-100 to the California Institute of Technology, Jet Propulsion Laboratory. REFERENCES Aptaker, I., " Near-infrared mapping spectrometer for investigation of Jupiter and its satellites", SPIE Vol. 331, Instrumentation in Astronomy IV, 182, 1982 Aptaker, I., " A Near-Infrared Mapping Spectrometer for investigation of Jupiter and its satellites", SPIE Vol. 366, Modern Utilization of IR Technology VIII, 96, 1982. Aptaker, I., " A near-infrared mapping spectrometer for investigation of Jupiter and its satellites", SPIE Vol. 395, Advanced IR Sensor Technology, 132, 1983. Aptaker, I., " A Near-Infrared Mapping Spectrometer for investigation of Jupiter and its satellites", SPIE Vol. 834, Imaging Spectroscopy II, 196, 1987. Bailey, G., "Design and test of the Near Infrared Mapping Spectrometer (NIMS) focal plane for the Galileo Jupiter Orbiter Mission", SPIE Vol. 197 Modern Utilization of Infrared Technology, 210, 1979. Carlson, R.W., "Spectral Mapping of Jupiter and the Galilean Satellites in the Near Infrared", SPIE Vol. 268, Imaging Spectroscopy, 1981. Carlson, R.W. et al, "Galileo Infrared Imaging Spectroscopy Measurements at Venus", Science 253, 1541, 1991. Garcia, R.R., "Dynamics, Radiation, and Photochemistry in the Mesosphere: Implications for the Formation of Noctilucent Clouds", J. Geophys. Res. 94, 14605, 1989. Hall, D.N.B., R.S. Aikens, R. Joyce, and T.W. McCurnin, "Johnson Noise Limited Operation of Photovoltaic InSb Detectors", Applied Optics, 14, 450, 1975. Hunten, D.M., L. Colin, and J.E. Hansen, "Atmospheric Science on the Galileo Mission", Sp. Sci. Rev. 44, 191, 1986. Kruse, P. W., L. D. McGlauchlin, and R. B. McQuistan, "Elements of Infrared Technology: generation, transmission, and detection", Wiley & Sons, New York, 1962. Macenka, S.A., " Near-Infrared Mapping Spectrometer Optical Subsystem Development and Testing", SPIE Vol. 430, Infrared Technology IX, 313, 1983. Smith, D.S., J.T. Wimmers, J.A. Hermann, and G.C. Bailey, "Hybrid packaging approach to improved low-noise operation of photovoltaic InSb detectors", SPIE 331, 29 (1982). Sparrow, E.W. and R. P. Heinisch, "The Normal Emittance of Circular Cylindrical Cavities", Appl. Optics 9, 2569 (1970). Strong, J. "Concepts of Classical Optics", W. H. Freeman and Company, San Francisco, 1958. Taylor, F.W. and S.B. Calcutt, "Near-infrared spectroscopy of the atmosphere of Jupiter", J. Quant. Spectrosc. Radiat. Transfer 32, 463, 1984. Thomas, G.E., J.J. Olivero, E.J. Jensen, W. Schroeder, and O.B. Toon, "Relation between increasing methane and the presence of ice clouds at the mesospause", Nature 338, 490, 989. Torson, J.M., "Interactive Image Cube Visualization and Analysis", Proceedings, Chapel Hill Workshop on Volume Visualization, Chapel Hill, North Carolina, May 1989. Weidner, V.R. and J.J Hsia, "Reflection properties of pressed polytetrafluoroethylene powder", J. Opt. Soc. Am. 71, 856, 1981. Yeates, C.M., T.V. Johnson, L. Colin, F.P. Fanale, L. Frank, and D.M. Hunten, "Galileo: Exploration of the Jupiter System", NASA SP-479, National Aeronautics and Space Administration, Washington, D.C., 1985.