Ultraviolet and visible imaging and spectrographic imaging instrument

Ultraviolet and visible imagingand spectrographic imaging instrument

J. F. Carbary, E. H. Darlington, T. J. Harris, P. J. McEvaddy, M. J. Mayr,

K. Peacock, and C. 1. Meng

The Ultraviolet and Visible Imaging and Spectrographic Imaging experiment consists of five spectro-graphic imagers and four imagers. These nine sensors provide spectrographic and imaging capabilitiesfrom 110 to 900 nm. The spectrographic imagers share an off-axis design in which selectable slitsalternate fields of view (1.000 x 0.10° or 1.000 x 0.050) and spectral resolutions between 0.5 and 4nm. Image planes of the spectrographic imager have a programmable spectral dimension with 68, 136,or 272 pixels across each individual spectral band, and a programmable spatial dimension with 5, 10, 20,or 40 pixels across the 10 slit length. A scan mirror sweeps the slit through a second spatial dimension togenerate a 10 x 10 spectrographic image once every 5, 10, or 20 s, depending on the scan rate. The fourimagers provide narrow-field (1.280 x 1.590) and wide-field (10.50 x 13.10) viewing. Each imager has asix-position filter wheel that selects various spectral regimes and neutral densities. The nine sensorsutilize intensified CCD detectors that have an intrascene dynamic range of 103 and an interscenedynamic range of - 105; neutral-density filters provide an additional dynamic range of 103. Thedetector uses an automatic gain control that permits the sensors to adjust to scenes of varyingintensity. The sensors have common boresights and can operate separately, simultaneously, orsynchronously. To be launched aboard the Midcourse Space Experiment spacecraft in the mid-1990's,the ultraviolet and visible imaging and spectrographic imaging instrument will investigate a multitude ofcelestial, atmospheric, and point sources during its planned 4-yr life.

Key words: Ultraviolet-visible sensors, atmospheric remote sensing, intensified CCD sensors.

1. IntroductionVarious agencies require scientific information aboutthe terrestrial atmosphere and surface, the celestialscene, and objects appearing against such back-grounds during both day and night conditions.Instruments fulfilling these requirements should pro-vide global or near-global remote sensing over anextended time span of several years. Furthermore,they should conduct observations over as wide aspectral and spatial range as possible. The ultravio-let and visible instrument on the Midcourse SpaceExperiment (MSX) mission' has been designed andconstructed to satisfy these requirements in conjunc-tion with other instruments onboard the spacecraft.The instrument has the acronym UVISI, which signi-fies ultraviolet and visible imaging and spectro-graphic imaging.

The authors are with the Applied Physics Laboratory, The JohnsHopkins University, Laurel, Maryland 20723-6099.

Received 12 July 1993; revised manuscript received 12 October1993.

0003-6935/94/194201-13$06.00/0.© 1994 Optical Society of America.

UVISI will play a unique role in remote sensing.The nine UVISI sensors sample the electromagneticspectrum from 110 to 900 nm with a spectral resolu-tion as good as 0.5 nm, a spatial resolution as good as

100 gLrad, and a time resolution as fast as 0.1 s.The intensified, filtered detectors can accommodate

1011 orders of magnitude in scene brightness.Furthermore, the UVISI instrument includes animage processor that can accept data from any of theimagers, perform a real-time analysis of the scene,and send tracking information to the spacecraft atti-tude system. This gives the instrument and theMSX the ability to perform autonomous acquisitionand tracking of desired elements of a scene.

The five UVISI spectrographic imagers (SPIM's)will provide spatial maps of the spectra of variousobjects and backgrounds. Primary observations forthe SPIM's include terrestrial dayglow and night-glow, auroral radiance, stellar and zodiacal lightspectra, and plume contrails in the upper atmosphere.When suitably inverted, the spectral radiances of theairglow reveal properties of the atmosphere such asconcentrations of major and minor constituents (in-cluding species such as 02, N2, NO, OH, and 03),

1 July 1994 / Vol. 33, No. 19 / APPLIED OPTICS 4201

temperature, and altitude profiles of both.2-4 Be-cause the SPIM's can sweep out a two-dimensionalscene, these observations can also indicate the spatialstructure of the observables. For example, observa-tions of the Herzberg-band 02 nightglow can outlinespatial structures indicating wave motions of theupper atmosphere.5 Spectrographic investigation ofthe aurora provides estimates of the fluxes andenergies of the precipitating charged particles thatcause the aurora, which in turn yield informationabout the ionosphere and magnetosphere. Far-ultraviolet observations can provide such informationeven under sunlit conditions.7 In the realm of as-tronomy, the SPIM's will undertake a measurementof the diffuse ultraviolet background of the sky.8

The four UVISI imagers will provide both wide andnarrow field of view images and, through the use of anonboard image processor, supply information foracquiring and tracking point sources such as starsand extended sources such as aurora. With a highertime and spatial resolution than the SPIM's, theimager observations will complement those of thespectrographs. The use of coaligned imagery pro-vides a precise measure of the position of a source inthe spectrograph slits, which may be important fordetermining the source spectrum. A nominal timeresolution of 0.5 s permits the imagers to makeunprecedented observations of rapidly evolving phe-nomena in the upper atmosphere such as aurora.9"0

Imager filters will permit the UVISI imager to inves-tigate specific atmospheric processes. For example,a cross-polarizer filter on an ultraviolet imager willmake the first polarization measurements of polarmesospheric clouds from space," while the ratio ofradiances from two narrow-band filters (one in thefar-ultraviolet imager and one in the visible imager)can be used in the estimation of auroral particlefluxes and electron densities.'2

Instruments similar to the UVISI instrument forman advanced class of intensified CCD systems thathave recently become prominent in space remotesensing.'3 Early versions of these detector systemsflew on long-term missions such as Voyager' 4 as wellas on short-term missions such as Spacelab.'5 Amongthe newer instruments of this type, the ultravioletimager orbited on Sweden's Viking satellite in 1986and provided an extremely rich sample of ultravioletimages of the dynamic aurora,' 0"16 whereas the Ari-zona Imaging spectrograph provided spectrographyand imagery of airglow and spacecraft glow from theSTS-39 shuttle flight in 1991.17,18

The UVISI instrument itself derives from a succes-sion of ultraviolet and visible instruments flown onorbital missions by The Johns Hopkins UniversityApplied Physics Laboratory. The first of these, theauroral ionospheric mapper,19 flew on the Hilat space-craft launched in 1983, and the auroral ionosphericremote sensor20 followed on the Polar Bear satellite in1986. These two instruments operated at ultravio-let wavelengths and provided the first observations ofthe auroral oval in the daytime, confirming its conti-

nuity through 360° Ref. 7). The Hopkins ultravio-let telescope, which was built at the Applied PhysicsLaboratory (APL), flew on the shuttle STS-35 mis-sion on the Astro-1 Observatory in December 1990and returned information on 77 astronomical ultravio-let spectra.21 APL instruments similar to the UVISIinstrument also flew on the Department of Defenseseries of Delta missions,22 all of which carried ultravio-let and visible instruments built at APL. The Ulinstrument on Delta 180 made the exoatmosphericobservations of plume phenomena in 1986.23 Twoyears later, the U2 instrument on Delta 181 observedvarious test objects, characterized plumes, and madedetailed observations of the Earth limb at severallocal times.24 In particular, the U2 sensors demon-strated the robust capabilities of intensified CCD's byobserving the extreme scenes of the day, night, andtwilight Earth limbs.25 26

Each of the UVISI sensors satisfies its designobjectives by using intensified CCD's and small optics(- 10-cm apertures, focal lengths <30 cm). Thesensors achieve a spectral resolution of 1 nm orbetter, a time resolution as fast as 0.25 s, and anangular resolution of 100 jLrad/pixel. With itsimproved sensitivity and spectrographic imaging capa-bilities, the UVISI instrument promises to continuemaking important observations in the field of remotesensing.

2. Instrumentation Overview

The UVISI instrument consists of a suite of fiveSPIM's and four imagers and their associated electron-ics (Fig. 1). An instrument-control processor and atelemetry formatter comprise the data-control system.

POINTINU lZ SEU iIPROCESSOR(IP)

Fig. 1. Block diagram of the UVISI instrument. The solid linesindicate lines of communication (i.e., command, control, data), thedotted lines indicate lines of power, and the vertical dashed lineindicates separation between the spacecraft (S/C) and the instru-ment systems. For redundancy, the instrument has two data-control systems, and each sensor has its own power converter(CONV) and SEM. The image processor has no redundancy.

4202 APPLIED OPTICS / Vol. 33, No. 19 / 1 July 1994

This system sends commands from the spacecraft tothe sensor electronics units, which actually controlthe individual sensors. Normally, the sensor elec-tronics units operate each sensor autonomously, per-forming such functions as gain, gate, and filter control.The data-control system sends commands to thesensors to change modes or to turn on or shut down.The data-control system also parses sensor data tothe spacecraft telemetry stream.

An image processor receives telemetry data fromany one of the imagers and performs an analysis onthe raw digital data. The image processor thencommunicates the results of its analysis with thedata-control system. When programmed to do so,the data-control system can accept commands fromthe image processor to adjust the operation of thesensors. Finally, the image processor relays a list ofpotential targets of interest to the spacecraft pointingsystem, which weights the UVISI information withdata from other sensors to determine spacecraftorientation.

The data-control system, the sensor electronic unit,and the image processor contain multiple processingunits and memory. Many standard commands re-side in read-only memory, and additional commandscan be loaded and stored in electronically erasableprogrammable read-only memory. If desired, all soft-ware for the image processor program can be reloadedduring flight.

The UVISI instrument employs a robust designbased on redundancy and distribution of functions.Thus the data-control system has an identical back upunit that can take over control in the event that theprimary system fails. Each of the nine sensors hasan independent sensor electronics unit and powerconverter that permit the sensor to operate in theevent that other units fail. Therefore, the single-

SIDE VIEW

point failure of one sensor by itself will only partiallydegrade the performance of the instrument as awhole. This design gives the LVISI instrument aprojected operational lifetime of at least 4 years.

The UVISI sensors are mounted on the outside ofthe spacecraft platform, which can be considered acylinder surrounded with a rectangular framework(Fig. 2). All the instruments view forward along thelong axis of the cylinder and share a common bore-sight. The SPIM's occupy positions on a commonpallet on the top of the platform, which is nominallyon the surface pointed away from the Earth; theimagers occupy positions on either side of the cylin-der, with fields of view oriented perpendicular to thenominal Earth horizon. The spectrograph slits lieparallel to and scan perpendicular to the Earth'shorizon. The fields of view are unobstructed, andbaffled sunshades on each unit limit the off-axis lightscattered into the UVISI optics.

3. Spectrographic Imagers

The SPIM's satisfy several science requirements.First, they provide spectral characterization of ob-jects within their fields of view at wavelengths fromthe vacuum ultraviolet (110 nm) to the near infrared(900 nm). Second, they have sufficient spectral reso-lution (- 1 nm) to resolve the atomic and molecularband emissions of stellar and atmospheric sources.Third, they have sufficient spatial resolution to iden-tify the source locations of these emissions to within afew kilometers at ranges of 1000 km. Fourth,they operate sufficiently quickly that they can tempo-rally resolve rapid atmospheric phenomena on timescales of a few seconds. Last, they can measureemissions from local molecular contaminants, particu-larly water (which is detected when solar ultraviolet

TOP VIEW

SPIMs (5)

otherEinstrument

IUN

I UW

Iotherinstrument

spacecraft electronics spacecraft superstructure

Fig. 2. Location of the UVISI sensors on the spacecraft platform. All sensors are coaligned and face in the forward direction. Spacecraftapparatus such as antennas and solar panels are omitted for clarity. In the nominal spacecraft orientation, the Earth would be at thebottom of the figure and the sky would be at the top.


I I I

other. _ < instrument

.-=dd9

causes H20 to dissociate to OH*), as an adjunct to anonboard contamination experiment.

Figure 3 shows a schematic drawing of a UVISISPIM. The SPIM's share a Wadsworth design con-sisting of a sunshade, a scanning telescope mirror, aslit-filter mechanism, a collimating mirror, a spheri-cal grating, a two-dimensional focal plane assembly,and an electronics package called a sensor electronicsunit. The telescope mirror collects incident lightand images the scene on a selectable slit-filter mecha-nism. The slit-filter mechanism provides one of fivepossible configurations: a wide 0.10 slit with noattenuating filter (for low-intensity sources), a nar-row 0.05° slit with no attenuating filter (for betterspectral resolution), a wide slit with an attenuatingneutral-density filter (neutral density 1%), a nar-row slit with an attenuating filter, and an opaque orclosed filter that acts as a shutter. A sphericalgrating disperses the light in wavelength and focusesit onto an array of 272 x 40 pixels on the focal plane.The 272 elements represent a spectral dimensionperpendicular to the slit, whereas the 40 elementsrepresent the spatial dimension along the length ofthe slit. Sensor electronics can sum these pixels andsample the spectral dimension in groups of 272, 136,or 68 pixels and the spatial dimension in groups of 40,20, 10, or 5 pixels (telemetry limitations permit onlycertain combinations of these groups). In conjunc-tion with a particular slit, each sample represents aspectrum over an instantaneous field of view of either1.0° x 0.05° or 1.00 x 0.10°. The sampling can beperformed at a nominal rate of 2 Hz or at a fast rate of4Hz.

The scanning motion of the primary mirror sweepsthe instantaneous field of view perpendicular to theslit and measures a second spatial dimension orthogo-nal to the slit's long dimension (see Fig. 4). Depend-ing on the slit size, this scanned dimension effectivelyproduces either 10 or 20 pixels in this direction. Themirror scans the field of regard in discrete incrementsof either 0.05° or 0.10° (i.e., the mirror stops scanningwithin each accumulation interval) so that the instru-ment can sample an unblurred, spatially contiguousscene of up to 1.0° x 1.0°. The mirror can dwell at a

to electronics

Fig. 3. Schematic diagram of a SPIM. All SPIM's have a com-mon dosign and differ only in grating and in photocathode andwindow materials.

40, 20. 10, or 5spatial elements

E

a

2

I

272,136, or68spectral elements

I0.050, 0.100 slit

Fig. 4. Schematic illustration of how a SPIM constructs a two-dimensional image by scanning the slit through a 1 field ofregard. Onboard electronics sums pixels in the focal plane toproduce the combinations shown. Available space in telemetrymay constrain the possible combinations of these array sizes.

given angular position in its scan or dither betweenseveral adjacent scan steps. The limits of scanningcan be adjusted so that the instrument samples any 10region within a 1.20 total scan angle; this featurepermits in-ffight adjustment for small errors in align-ment.

All-aluminum construction minimizes thermal ef-fects that can result from differential solar heating ofthe spacecraft. Mirrors, gratings, housings, fix-tures, and mounting pallets all consist of aluminumand have identical coefficients of thermal expansion.This type of construction permits the SPIM's tooperate at mean temperatures between -30 'C and-10 'C without serious degradation in optical perfor-mance from that at room temperature. That is,within this nominal operating regime, the optics andfocal plane maintain focus and alignment to within0.1 mm. A longitudinal thermal gradient of up to10 C and a lateral gradient of up to 5 C can also existon the pallet without degrading performance. In themost extreme case (for SPIM 1), a gradient of 10 'Cwould cause a spectral shift of less than 0.5 nm.Such a robust thermal design actually permits thesensors to be aligned at room temperature for opera-tion at -30 'C.

Each SPIM shares a similar off-axis design differ-ing only in mirror coating and grating characteristics.The telescope (scan) mirror is parabolic with a vertexradius of 800 nm, whereas the collimating mirror isparabolic with a vertex radius of 666.67 mm. Bothare used off axis with a common optic axis. Theconcave diffraction grating has a 500-mm radius ofcurvature. Each grating is mechanically ruled andhas a different line spacing so that the desireddispersions of the individual spectrographs areachieved. With 2572 lines/mm, the grating for SPIM1 represents a state-of-the-art construction for far-ultraviolet spectroscopy. All gratings have surface


accuracies high enough so that, at their centralwavelengths, 90% of the energy of a parallel inci-dent beam would produce at the focus a blur circle notmore than 0.0250 in angular diameter. All opticalelements are coated with reflective aluminum andhave overcoatings of SiO2 for visible wavelengths orMgF2 for ultraviolet wavelengths. In SPIM 2through SPIM 5, long-pass filters located past the sliteliminate higher spectral orders; in SPIM 1, thedetector's MgF2 coating eliminates higher orders.

Table 1 summarizes SPIM performance character-istics. Each column indicates the performance ofone SPIM. The first row gives the nominal disper-sion, and the second row gives the spectral resolutiondetermined by the Rayleigh criterion. The sensitiv-ity row indicates the minimum flux in one spectralbin (i.e., dispersion/272 pixels) required at the aper-ture to four photocathode events per accumulation.The next two rows indicate the photocathode andwindow materials; the coatings, the window, and thephotocathode materials were optimized for the par-ticular wave band of each SPIM. Here Ad indicatesthe effective aperture area and £4p indicates thegeometric size of one spatial pixel.

The expected scene and the spacecraft recordingrate dictate the various combinations of spatial-spectral pixels and slit sizes. At the nominal 2-Hzdata rate, a single SPIM can transmit a full 1.00 x 1.00scene in 10 s with a spatial resolution of 0.050 and aspectral resolution of 1 nm. Each SPIM operatesindependently, so one SPIM could scan across a full1.00 x 1.00 scene while another could dither within afew individual spatial pixels.

4. Imagers

The UVISI imaging science requirements dictateboth a high spatial resolution and a wide field of view,which can be satisfied by the employment of imagerswith narrow fields of view (NFOV) and imagers withwide fields of view (WFOV). One NFOV-WFOVimager pair operates in the ultraviolet and one oper-ates in the visible; each imager must satisfy differentrequirements. The NFOV imagers must have a field

of view and pixel size similar to those of otherinstruments onboard and must operate with lowinstrument background noise per pixel, so that faintobjects such as stars and other faint point sources canbe detected. The WFOV imagers must provide track-ing of atmospheric and auroral features during bothsunlit and dark conditions, instantaneously view thecomplete Earth limb, and participate in contamina-tion assessment. The NFOV imagers, particularlythe ultraviolet imager, must have sufficiently goodspectral out-of-band rejection to ensure that red leakfrom longer-wavelength radiation does not corruptmeasurements of the scene spatial content. Coinci-dent with this spectral rejection, the NFOV imagersmust have sufficient angular resolution to observespatial variations of scene radiance at a scale of 100prad. Science requirements also dictate that eachset of NFOV-WFOV imagers make observations inboth the visible and ultraviolet. In particular, muchscientific and defense interest will focus on sceneclutter measurements made by the middle ultravioletimager.

In addition, all imagers must share several com-mon features. All must have a fairly wide dynamicrange to accommodate the observation of day limb,night limb, and celestial scenes. All imagers mustsample fast enough to observe target and backgrounddynamics, which can occur on time scales as fast as

1 s. The imagers must also have a good off-axisrejection so that bright objects outside their fields ofview do not overwhelm less bright objects withintheir fields of view.

Cost and weight considerations necessitated threedifferent designs for the imagers. Figure 5 showsthe schematic design of the UVISI NFOV imagers.Both the NFOV ultraviolet imager (IUN) and thevisible NFOV imager (IVN) have the same design,which utilizes reflecting optics. Each imager con-sists of a sunshade, primary and secondary mirrors, afilter-wheel mechanism, a focal plane unit, and anassociated sensor electronics unit (SEU). The IUNoperates in the middle ultraviolet (180-300 nm), andthe IVN actually operates from the near ultraviolet to

Table 1. Spectrographic Imager Performance

SPIM

Unit 1 2 3 4 5

1- X2 (nm) 110-170 165-258 251-387 381-589 581-900AX (nm)

0.100 slit 0.8 1.2 1.8 2.8 4.30.050 slit 0.5 0.9 1.5 2.1 2.9

Sensitivity (photons/cm 2 s) 5 2 3 1 1Photocathode CsI RbTe KCsSb Ext S20 Ext S20Window MgF 2 MgF 2 SiO2 SiO2 SiO2Total dynamic range 242Spatial resolution (sr) 3.81 x 10-7 (0.05° slit) or 7.62 x 10-7 (0.10 slit)Field of view 0.100 x 1.00 or 0.050 x 1.000Collecting area (Ad) 110 cm2

Sampling time 0.50 s (nominal) or 0.25 s (fast)Scan time (1.0) 5 s (at 4 Hz) or 10 s (at 2 Hz)


filter drivewheel -1

to electronics

Fig. 5. Schematic diagram of the two UVISI NFOV imagers. One imager operates in the middle ultraviolet (180-300 nm), and the otheroperates from the near ultraviolet to the far infrared (300-900 nm).

the near infrared (300-900 nm). Both narrow-fieldimagers have a field of view of 1.280 x 1.590.

The NFOV imagers use a Cassegranian telescope tocollect light and direct it through a filter wheel to thesensor array. The parabolic primary mirror has adiameter of 140 mm and a radius of curvature(concave) of 723.26 mm; its inner hole is 44 mm indiameter. The hyperbolic secondary mirror has adiameter of 50 mm and a radius of curvature (convex)of 292 nm. As with the spectrographs, the NFOVimagers are built primarily of aluminum (aluminummirrors, housing, and fixtures). The mirrors consistof an aluminum substrate coated with electrolessnickel. The polished reflecting surface is coated

with high-reflectance aluminum with a MgF2 overcoat-ing. The mirrors have a minimum reflectance of

-85% in the ultraviolet and 90% in the visible.The mirror surfaces have an accuracy of 0.3 waves(visible). All-aluminum construction gives the NFOVimagers robust thermal characteristics. The imag-ers will remain in focus in the normal operating rangeof -1 IO C to - 30 0C; they will also remain in focusagainst thermal gradients of up to 10 'C across anydirection of their housing.

Figures 6 and 7 show the schematic designs for theUVISI WFOV imagers. The far-ultraviolet WFOVimager (IUW, Fig. 6) uses a three-mirror system tocollect and focus light, whereas the visible WFTOV

to electronics

Fig. 6. Schematic diagram of the IUW, which operates in the far ultraviolet (110-180 nm).


Fig. 7. Schematic diagram of the IVW, which operates from thevisible to the near infrared (380-900 nm).

imager (IVW, Fig. 7) uses a five-element refractingtelescope with radiation-resistant glasses. Refrac-tive optics in the visible imager permit a lighter, morecost-effective unit; the ultraviolet imager, however,requires reflective optics because lens transmissivitydegrades severely at shorter wavelengths. The IVWoperates from the visible to the near infrared (380-900 nm), and the IUW operates in the far ultraviolet(110-180 nm). Both wide-field imagers have a fieldof view of 10.50 x 13.10.

Sunshade design, internal baffling, and opticalcleanliness during imager fabrication minimize off-axis light scattering onto the focal planes of all theimagers. Figure 8 shows the expected point-sourcetransmittance for stray light for the NFOV andWFOV ultraviolet imagers. Outside the fields ofview, the transmittances drop 4-5 orders of magni-tude. To achieve this off-axis performance, theNFOV imagers use cylindrical baffles at both mirrorsand a two-stage sunshade with vanes. As a way tomaintain mirror cleanliness, the imagers (and theSPIM's) are assembled in a class 100 environment[i.e., fewer than 100 particles/ft 3 (3500 particles/M 3 )of particles 0.5 ,um in size, which compares to atypical surgical operating room of class 100,000].The design and the cleanliness precautions ensurethat the scatter contribution is approximately noworse than that caused by mirror surface roughness.

-4

-60.

-81

-10

-12

-14L0 5 10 15 20 25

angle (deg)

Fig. 8. Point source transmittance (pst, off-axis rejection) for thenarrow- and wide-field ultraviolet imagers. The large drop-offs at- 90 and - 120 are caused by the sunshades.

Each imager has a six-position filter wheel that canbe rotated so that an appropriate filter is selected.The filters have been chosen so that the observationof expected phenomena is optimized. Each wheelcontains an open, full-wave-band filter for unattenu-ated viewing, a closed filter for protecting the electro-optical detector and for measuring instrument noise,and a neutral-density filter for full-wave-band obser-vation of bright scenes. Rays converging throughWFOV filters make angles less than 120 from thenormal (which may appear to shift some of thenarrower wave bands toward the blue), whereas raysconverging through NFOV filters make angles lessthan 1.50 from the normal (which will have a negli-gible effect on the filter wave bands). Visible imag-ers have a neutral density 4 (X 10-4) attenuatingfilter, whereas ultraviolet imagers have a neutraldensity 3 (x 10-3) filter. The other filters selectparticular wave bands of interest. For example, the200- to 230-nm filter (filter 4) in the IUN selects NOgamma bands for airglow measurements and nadirscene clutter. (The combination of filter cutoff andfalling photocathode response should render themiddle-ultraviolet filters impervious to the red leak ofreflected solar radiation longward of 300 nm.) The145- to 180-nm filter (filter 4) on the IUW screens outthe bright Ly-a- (121.4 nm) and oxygen emissions(130.4 nm) of the dayglow and permits the viewing ofaurora emissions during the daytime. The 426- to429-nm filter of the IVW selects the 428-nm emissionof the N2+ (N) system, which is prominent inairglow.

Two of the imager filters make UVISI uniqueamong space-imaging experiments. A clear lens (fil-ter 6) on the IVW provides a near-field focus of 10 mfor observing small particulates that may drift throughthe field of view. This lens is used in conjunctionwith an experiment that performs on-orbit measure-ments of particulate contaminants having sizes downto - 5 pm. A cross-polarization filter (filter 6) on theIUN permits evaluation of scattering properties ofatmospheric particulates (one half of this filter ispolarized 900 to the other half). The polarizationfilter will be used in the observation of polar meso-spheric clouds, which are composed of ice crystalsfloating in the upper atmosphere. Polarization mea-surements can reveal the sizes and size distributionsof such particles.

The UVISI imagers operate in a starting mode inwhich one full frame of data is accumulated all atonce. No scene scanning takes place that might timebias a measurement. Each imager provides one fullimage of 256 x 244 pixels in a 0.5-s sample time (a2-Hz frame rate). The imagers can also operate in afast mode that provides one partial frame (122 x 256pixels) in a 0.25-s sample time (a 4-Hz frame rate).

Table 2 summarizes the performance of the UVISIimagers. Each column in the table represents theperformance of one particular imager. The first sixrows indicate the nominal (FWHM) passbands of theimagers. The next row shows the expected sensitivi-


Table 2. Imager Performance

Ultraviolet Imager Visible Imager

Unit IUN IUW IVN IVW

Nominal passbands (nm)1, Open 180-300 110-180 300-900 380-9002, Closed - - - -3, Neutral-density filter 180-300 110-180 300-900 380-9004, Wave band 1 filter 200-230 110-123 305-315 426-4295, Wave band 2 filter 230-260 128-138 350-440 529-6316, Wave band 3 filter 260-300 145-180 470-640 380-900

Sensitivity (photons/cm 2 s) 2 5 1 6Photocathode RbTe CsI Ext S20 Ext S20

Window MgF2 MgF2 SiO2 SiO2Resolution

Field of view 1.59° x 1.28° 13.10 x 10.5° 1.59° x 1.280 13.10 x 10.50Pixel size (mrad2 ) 0.108 x 0.092 0.892 x 0.751 0.108 x 0.092 0.892 x 0.751

Collecting area (Ap) 130 cm2 25 cm2 130 cm2 25 cm2

Total dynamic range 241Sampling rate 2 Hz (256 x 244 pixels) or 4 Hz (256 x 122 pixels)

ies calculated when one assumes a square-wave spec-trum with a width of 1 nm centered on the band ofeach of the open sensors; again, the fluxes refer tothose needed to produce four photoevents at thephotocathode. The following two rows give the pho-tocathode material and its window material. Theresolution rows indicate the full field of view and thegeometric size of one pixel; the wide-field imagershave fields of view approximately 100 times as largeas those of the narrow-field imagers. The collectingarea refers to the effective collecting areas. Thedynamic range row breaks down the several factorsthat give a total maximum dynamic range of 241-1012.The final row indicates the time between successiveframes or samples. Note that an increase in thesampling rate reduces the size of the image.

5. Electro-Optical Detectors

Both the imagers and the SPIM's utilize similarelectro-optical detectors or focal plane units. Eachdetector consists of an image intensifier with aV-shaped microchannel plate, a green phosphorscreen, a fiber-optic reducer, and a CCD that convertsphotons to electrical signals. Figure 9 illustrates thebasic configuration of each focal plane unit. Lightfrom the optical system strikes the photocathode atthe focal plane and is converted into electrons. Ahigh voltage across the microchannel plate cascademultiplies the electrons, which accelerate to a phos-phor screen immediately behind the microchannelplate. The phosphor emits green light, which trav-els through a fiber-optic reducer to the CCD, whichconverts the light into an electronic signal for process-ing. Similar intensified devices have previously beenflown.3,26

The Thompson TH7866 CCD used by the UVISIinstrument has an active area of 6.6 mm x 8.8 mm,whereas the microchannel plate measures 40 mm indiameter. This size mismatch necessitates the useof a fiber-optic coupler with a 4.5:1 taper. Further-

more, the coupler has a circular footprint and theCCD has a rectangular shape, so a few CCD pixels atthe corners of the imager arrays obtain no opticalsignal (see Fig. 9). However, the signal from thecorners can be exploited as a way to determine theCCD zero offset, which must be subtracted fromscene pixels. Exposure to radiation in space maycause the coupler transmissivity to degrade slowly

reducer footprint(shaded)

Fig. 9. Schematic diagram of a UVISI focal plane unit. All focalplane units have the same design and differ only in window andphotocathode materials. Focal plane unit electronics report 12 bitsper pixel to the data-control system.


over the lifetime of the instrument. The microchan-nel plate gain can be used for compensation of thiseffect, and periodic in-flight recalibration will monitorthe degradation. A 2-mm-thick aluminum boxhouses the entire image intensifier unit, including thedigitizing electronics, which greatly reduces the ef-fects of energetic charged particles that exist in theterrestrial magnetosphere.

To obtain as wide a dynamic range as possible, thedetectors feature a dual automatic gain control.This control adjusts the voltage across the microchan-nel plate (gain control) and the duration of the voltageacross the photocathode (gate control). One canconceive of this operation by using an analogy with asimple single-lens reflex camera: the gain controlsthe film speed while the gate controls the exposuretime. The voltage control provides a factor of - 104in dynamic range, whereas the gate control provides afactor of 102. Filters provide additional factors of- 1034. The CCD itself provides 12 bits (=4096) of

intrascene dynamic range per pixel. The combina-tion of filter, gain, and CCD produces a total dynamicrange of 1012.

During normal operations, the automatic gain con-trol sets the gain; its circuitry monitors the previoustwo frames to set the gain for the current frame.A sophisticated automatic gain control algorithmdetermines scene brightness while allowing for a(programmable) number of isolated point sources toexceed brightness limits. Selected by command, dif-ferent automatic gain control thresholding schemesor gain tables optimize the gain algorithm for differ-ing situations such as staring at a star field orscanning the daytime limb. The algorithm also pro-vides for a safety shutdown in case the UVISI instru-ment suddenly views a bright scene. Finally, com-mands can override the automatic gain control andallow the user to set the gain manually to a constantlevel or step the gain through a series of levels forin-flight calibration.

The CCD operates at a nominal rate of 2 Hz (0.50-sframe time). At this rate the CCD accumulates dataover 468 ms and uses 32 ms to report data andreset the gain. However, the CCD can also operateat a faster rate of 4 Hz (0.25 s). At this speed theCCD accumulates data over 218 ms and againrequires 32 ms for housekeeping functions. Thespectrograph focal plane units report a full 272 x 40pixels to the data-control system, regardless ofwhether they are operating at the 2- or the 4-Hz rate.However, the imager focal plane units report a full256 x 244 pixels only while operating at the 2-Hzrate; at the 4-Hz rate, the imager focal plane unitsreport a reduced array of the central 256 x 122 pixels.In the reduced scene, however, only approximately256 x 100 pixels contain meaningful counts becausethe CCD cannot be read out fast enough.

A focal plane unit reports 12 bits of data per pixelfor each frame, whether the data are imager data orspectrograph data. The data-control system can com-press these data to 8 bits/pixel to save space. The

data-compression scheme uses a constrained maxi-mum error method, which for 12-to-8-to-12-bit con-version preserves the data within a maximum frac-tional error of less than 1%. The data, whethercompressed or uncompressed, are then forwarded tothe MSX recorder system for later transmission tothe ground. The data arrive at ground receivers inthe same compressed or uncompressed format as theywere recorded onboard. Ground-data processing re-stores the compressed data to their original 12 bitsper pixel.

The Thompson CCD features antiblooming cir-cuitry, which prevents counts in a bright or saturatedpixel from overflowing into adjacent pixels and affect-ing the quality of an extended region of the frame.

6. Image Processing for Tracking

The UVISI instrument includes an image-processingsystem that can isolate likely targets in the imagerfields of view and communicate their positions to theMSX flight processor. As used here, target can refereither to a point source such as a star or to anextended source such as an auroral surge. (Oneassumes, of course, that the scale size of a target ismuch less than the scale of a field of view.) Theimage-processing system operates in real time, and, ifdesired, can automatically select operating modes forthe UVISI sensors based on its results.

The system accepts image data from any one of theUVISI imagers and processes them with a 1750Amicroprocessor and a digital signal processing 2100chip, which is used for onboard image processing.Internal software performs initialization and track-ing functions that include filtering, smoothing, thresh-olding, and centroiding. The image processor seekstargets based on an a priori target-description filecontaining weights for a number of target featuressuch as size, shape, brightness, and location. Theimage processor transforms locations of likely targetsfrom UVISI pixel coordinates to spacecraft coordi-nates and passes them to the MSX flight processor,which performs Kalman filtering of other targetinginputs to select true targets for observation.

The image processor can function when the MSXprocessor has no definitive track (i.e., no predictedtarget trajectory) information and can track either apoint source or an extended source. Point-sourcetracking will be exercised when tracking a pointsource such as a star being occulted by the atmo-sphere, whereas extended source tracking is neces-sary for tracking extended sources such as auroralsurges or individual polar mesospheric clouds. Be-cause some types of targets may move rapidly orchange suddenly, the image processor operates onlyon single frames of data.

The image processor offers considerable flexibilityin its operations. The nominal processor programdetermines likely targets based on a number offeatures (e.g., size, shape, brightness, location) con-tained in a target-description file uploaded prior tothe observation. The software maximizes target like-


lihood by weighting these various features. Thus astar source might be weighted to optimize brightness,small size, and position features, whereas an activesegment of the auroral oval would be weighted tooptimize elongated shape and orientation features.The expected target characteristics will determinethe weighting of the various features. Obviously,more than a single target can appear in one frame ofdata, so the image processor can accommodate up to23 different targets at once. If necessary, one canreprogram the entire image processor in orbit byreloading its software modules.

7. Operating Mode Options

The UVISI instrument can operate in a staggeringlylarge number of modes. The use of each modedepends on the phenomena under observation, on thespacecraft power budget, and on the permitted data-recording rates. The spacecraft records data at ei-ther 25 Mbits/s or at 5 Mbits/s. The UVISI instru-ment can output data at rates of up to 5.5 Mbits/s,and this rate is sufficient for recording data from allfive SPIM's and all four imagers at a nominal framerate of 2 Hz. However, the UVISI instrument mustshare the telemetry with several other instruments.In the 5-Mbits/s recording mode, each instrumenthas a dominant mode in which it dominates thetelemetry stream. When not dominant, the UVISIinstrument contributes data at a rate of 1.1 Mbits/s,which permits data from five SPIM's and one imagerin the telemetry. When the UVISI instrument domi-nates, it contributes data at 4.4 Mbit/s, which per-mits two imagers and all five SPIM's. In the 25Mbits/s recording mode, the UVISI instrument con-tributes data at its full 5.5-Mbits/s rate. At any ofthe reduced data rates below 5.5 Mbit/s, the UVISIinstrument can operate in a cyclic mode that permitsdata from multiple imagers to appear in a commu-tated telemetry segment. For example, the telem-etry space for one imager can alternate betweencontaining data from the IUN and containing datafrom the IVW.

Table 3 summarizes the operational options foreach of the UVISI SPIM's. The slit-size optionpermits the selection of either high spectral resolu-tion (0.05) or high sensitivity (0.10). The data-compression option chooses between 8-bits/pixel com-

Table 3. UVISI SPIM Options

Option Choice

Slit size 0.05° or 0.100Data compression 8 or 16 bitsGain control automatic, step, fixed, or lockedNumber of spectral bins (CCD) 68, 136, or 272Number of spatial bins (CCD) 40, 20, 10, or 5Size of mirror scan 0.0O0-1.00°Angle of mirror scan 0.00°-1.20°Scan step size 0.0250 or 0.050Filter selection Closed, open, or neutral densitySample time 0.25 or 0.50 s

On or off Independent

pressed data used for low data rates or 16-bits/pixeluncompressed data used for more accurate countingat high data rates. Gain control features four possi-bilities. The automatic gain control is used whenthe scene intensity is unknown or changing rapidly;the step gain control is used during instrumentcalibration; and the fixed gain control is used whenthe scene intensity is known and fairly stable. Inaddition, the gain can be mirror locked, which meansthe automatic gain control is used but is based ondata from a given mirror position. The number ofspectral and spatial bins permits adjustment of thespectral or spatial resolution. For example, the com-bination of 272 spectral bins with five spatial binsgives a maximum spectral resolution and a minimalspatial resolution, whereas the combination of 68spectral bins and 40 spatial bins yields a maximumspatial resolution and a minimum spectral resolution.Restricted recording bandwidth and formatting donot necessarily permit all possible combinations ofspectral bins, spatial bins, and rates. The mirrorscan angle (the full angle through which the mirrorscans) may be selected anywhere within a 1.00°window in 0.050 increments; the center of this scanangle (a boresight angle) can be selected anywherewithin a 1.200 window. In-flight alignment calibra-tions will determine this center angle to within onepixel (0.05°). Note that scanning with a 0.100 slitpermits ten cross-slit spatial bins per scan, whereasscanning with the 0.050 slit permits 20 cross-slitspatial bins. Filter options include a closed (opaque)filter for protective purposes, an open filter for dimscenes, and an attenuation filter for bright scenes.The attenuation filter effectively extends the dynamicrange of the instrument. For the far-ultravioletspectrograph (SPIM 1), the attenuation filter reducesthe strong Ly-cz emissions of the geocorona and the Oi(1304) emissions of the dayglow (while the dramaticfalloff of the CsI photocathode eliminates solar back-scatter at longer wavelengths). All SPIM's can oper-ate at either a 2- or a 4-Hz frame rate (0.50- or 0.25-ssample time), although data-recording rates mayprevent the use of the faster rate. Finally, each unitcan be independently turned on or off as a way toconserve power, which is severely rationed on theMSX spacecraft.

Table 4 gives the corresponding operating optionsfor the UVISI imagers. As with the SPIM's, theimagers offer two types of data compression andseveral filter positions. With the exception of amirror-locked mode (which is unique to the SPIM's),the imagers have the same gain options as the

Table 4. UVISI Imager Options

Option Choice

Data compression 8 or 16 bitsGain control Automatic, step, or fixedFilter wheel Closed, open, neutral density, or bandsSample time 0.50 or 0.25 s

On or off Independent


spectrographs. The imagers' 8-bit compression op-tion permits the use of a lower data rate at theexpense of a slightly reduced count precision, whereasthe 16-bit option permits the full dynamic range butrequires a fast data-recording rate. Each imager hasa six-position filter wheel (see Table 2). Each wheelhas a closed position for protection, an open positionfor maximum sensitivity, and a neutral-density at-tenuation filter (either 10-3 or 10-4) for viewingbright scenes. The remaining filters permit the ob-servation of various wave bands believed by investiga-tors to have scientific merit.

8. Calibration

The UVISI calibration primarily attempts to deter-mine those factors necessary to convert raw sensorcounts into photometric units such as irradiance(photons/cm 2 s) or spectral radiance (photons/cm 2 srnm). For point sources, which are appropriate forthe imagers, the conversion equation is

E Ci - D(T) (1)

where Eij is the irradiance for pixels i andj, Cij is thedigital counts, D(T) is the dark count offset at tempera-ture T, Ad is the effective aperture area (in squarecentimeters), G, is the gain factor for level n, Qf is thewavelength-intergrated conversion efficiency (counts/photon) for filter f, Uij is the uniformity correction forthe pixel, and Tk is the accumulation time for gatelevel k. One can conceive of Qf as the photocathodequantum efficiency E(X) times filter response functionR(X), integrated over the passband:

Qf= E(X)R(X)dX. (2)

The integral for Qf extends over the characteristicwidth AX of the filter response, nominally the FWHMof the filter transmission curve.

For extended sources, which are appropriate forthe SPIM's, the conversion equation is

Li = 1 Cij - D(T) (3if1P^AX AdG.RijTk

where the terms are the same as in Eq. (1) except thatRij is the response (count/photon) for spectral bin iand spatial pixel j, fup is the spatial size of a pixel(steradians) and AX, is the spectral size of a pixel(nanometers). One can convert the radiance to Ray-leighs per nanometer by multiplying by 47r/106 .

Calibration specifically consists of determining thedark count D(T) as a function of temperature, thegain curve G as a function of level n, the filterresponse curves as functions of wavelength , theflat-field function Uij and the response Rij as func-tions of pixels i andj, and the accumulation time T

k asa function of gate index k and the field of view. Fromthe field of view, one can calculate the pixel size fQp.

For the SPIM's one also requires a wavelength calibra-tion that relates spectral bin to wavelength, whichhas the form

Xi = AO + Ali, (4)

where i is the spectral bin number and AO and Al arecoefficients. The precise metric calibration of theimagers requires characterization of the distortionand alignment.

Measurement of most of the calibration factors isstraightforward. The dark-count function D(T) re-sults from measurements of the residual sensorcounts in a dark environment at various temperatures.The D(T) is also measured at various gain levels, withthe intensifiers both on and off, and as a function ofdetector on time. (A dark-count measurement isautomatically performed at the start and end of anorbital observation, and the addition of dark measure-ments during an observation ensures a good measure-ment of the dark current.) One determines the gainand gate functions Gn and Tk by systematically vary-ing the gain and gate for sources of constant intensity.The wide dynamic range of the sensors requires theuse of devices such as filters and masks to vary sourceintensities. Measurement of the response functionsnecessitates two approaches. For most of the SPIM's,integrating spheres provide sources of known spec-tral radiance that fill the field and the apertures of thesensors and permit radiometric calibration of the fullspectrographic array by using only a few accumula-tions. SPIM 1 and the imagers rely on a calibratedmonochromator beam of known irradiance for deter-mination of response. The latter approach requiresa considerable number of accumulations to samplethe entire spatial-spectral space of the sensor. Boththe integrating spheres and the monochromator beamwere calibrated with detectors traceable to the Na-tional Institute of Standards and Technology. In allcases, the response is measured at several tempera-tures in the operating range and for all filter or slitcombinations. One also measures the response atwavelengths well outside the nominal sensor pass-band to determine the possibility of wavelength leak-age.

One determines the fields of view by locking eachindividual sensor on a precision stage, illuminatingthe sensor with a collimated beam, and moving thestage to known positions of azimuth and elevation.Typically, spatial calibration involves 15-20 pointsequally spaced on the sensor array. This calibrationis carried out at several temperatures within theoperating range as well as at room temperature.Knowledge of the relative azimuth and elevationangles of the stage also permits this calibration to beused in the determination of the total distortion(optical plus focal plane) of the sensors.

Each individual sensor has undergone ground cali-bration at the Optical Calibration Facility at TheJohns Hopkins University APL. The UVISI instru-ment team is now analyzing the ground-calibrationdata. Preliminary results suggest that the sensors


0.100

0.010

SPIM1 - _p iSPIM2 / ,

.. .- / SPIM4) .. I I

I0

PIJ I

SPIM3

0.001 I . . . . .0 200 400 600

wavelength (nm)

Fig. 10. Calculated response curves for the SPINVapply to the wide slit without attenuation at magate. The curves include photocathode quantumdow transmissivity, grating efficiency, mirror reflEfilter transmission effects.

have characteristics that agree well withexpectations. Figures 10 and 11 show tresponse curves of the SPIM's and the in

During the mission, special calibrements will provide data to verify an(calibrations. Because the UVISI instruno calibration lamps, this in-flight calilrely exclusively on natural sources. Ithe known spectral irradiances of certeprovide radiometric recalibration, andairglow of the Earth will provide a sourcethe uniformity calibration. Gain andtion use special observing modes of thethat the full dynamic range of the sensorstars of various brightnesses will have tcThe UVISI instrument team will analycalibration data and provide corrections ttion factors, which will undoubtedly c]mission progresses.

100

UN -V - - -a '~~~ ~~~~~~~VN -

0 1o UW

C 1

0 10-3

o 10

10-60 200 400 600

wavelength (nm)

Fig. 11. Calculated response curves for the imsponse curves apply to the open filter positions of tthe sensors operate at maximum gain and gatinclude photocathode quantum efficiency, windowmirror reflectance, and any filter or lens transmissi

M5

I

The authors thank the large number of supportingengineers and scientists who contributed to the devel-opment and construction of the UVISI instrument.Principal among these were K. Heffernan, instru-ment engineer, and J. Heiss, program manager. TheUVISI instrument receives funding through the Bal-listic Missile Defense Office, formerly the StrategicDefense Initiative Organization.

References1. R. E. Heubschman and C. T. Pardoe, "The Midcourse Space

~L~I Experiment Spacecraft," presented at the American Instituteof Aeronautics and Astronautics 1992 Aerospace Design Con-ference, Irvine, California, February 1992.

800 1000 2. S. C. Solomon, P. B. Hays, and V. J. Abreu, "Tomographic

I's. The curves inversion of satellite photometry," Appl. Opt. 19, 3409-3418s.m Th curvesn (1985).imum gain and 3. R. R. Meier, "Ultraviolet spectroscopy and remote sensing of

* efficiency, win- the upper atmosphere," Space Sci. Rev. 58, 1-185 (1991).ectance, and any 4. D. J. Strickland, R. Link, and L. J. Paxton, "Far-UV remote

sensing of thermospheric composition and the solar ELTVflux," in Ultraviolet Technology IV, R. E. Huffman, ed., Proc.

l their design Soc. Photo-Opt. Instrum. Eng. 1764, 117-131 (1992).he calculated 5. M. N. Ross, A. B. Christensen, C. I. Meng, and J. F. Carbary,

aer. a e"Structure in the UV nightglow observed from low earthtagers. orbit," Geophys. Res. Lett. 19, 985-988 (1992).Ltion experi- 6. D. J. Strickland, J. R. Jasperse, and J. A. Whalen, "Depen-

I adjust the dence of auroral FUV emissions on the incident electronment carries spectrum and neutral atmosphere," J. Geophys. Res. 88,)ration must 8051-8062 (1983).For example, 7. C. I. Meng and R. E. Huffman, "Ultraviolet imaging fromtin stars will space of the aurora under full sunlight," Geophys. Res. Lett.the uniform 11,315-318(1984).for checking 8. R. C. Henry, J. Murthy, M. Allen, M. Corbin, and L. J. Paxton,gate calibra- "Spectroscopy and imaging of the cosmic diffuse UV back-gasesor.So ground radiation," in Ultraviolet Technology IV, R. E. Huffman,sensos.m ed., Proc. Soc. Photo-Opt. Instrum. Eng. 1764, 61-71 (1992).rs is sampled, 9. J. D. Craven and L. A. Frank, "The temporal evolution of abe sampled. small auroral substorm as viewed from high altitudes with

ze the flight Dynamics Explorer 1," Geophys. Res. Lett. 12,465-468 (1985).o the calibra- 10. C. D. Anger, J. S. Murphree, A. Vallance Jones, R. A. King,iange as the A. L. Broadfoot, L. L. Cogger, F. Cruetzberg, R. L. Gattinger,

G. Gustafsson, F. R. Harris, J. W. Haslett, E. J. Llewellyn, J. C.McConnel, D. J. McEwen, E. H. Richardson, G. Rostoker, B. R.Sandel, G. G. Shepherd, D. Venkatesan, D. D Wallis, and G.Witt, "Scientific results from the Viking ultraviolet imager:an introduction," Geophys. Res. Lett. 14, 383-386 (1987).

11. G. E. Thomas and J. J. Olivero, "Climatology of polar meso-spheric clouds," J. Geophys. Res. 94, 14,673-14,681 (1989).

12. L. A. Frank and J. D. Craven, "Imaging results from dynamicsExplorer I," Rev. Geophys. 26, 249-283 (1988).

13. A. L. Broadfoot and B. R. Sandel, "Application of the intensi-fied CCD to airglow and auroral measurements," Appl. Opt.31,3097-3108 (1992).

14. A. L. Broadfoot, B. R. Sandel, D. E. Shemansky, J. C.McConnell, G. R. Smith, J. B. Holberg, S. K. Atreya, T. M.Donahue, D. F. Strobel, and J. L. Bertaux, "Overview of theVoyager ultraviolet spectrometry results through Jupiterencounter," J. Geophys. Res. 86, 8259-8284 (1981).

15. D. G. Torr, R. W. Basedow, and G. H. Mount, "An imaging800 1000 spectrometric observatory for Spacelab," Astrophys. Space

Sci. 92, 237-291 (1983).agers. The re- 16. C. D. Anger, S. K. Babey, A. L. Broadfoot, R. G. Brown, L. L.he filter wheels; Cogger, R. Gattinger, J. W. Haslett, R. A. King, D. J. McEwen,;e. The curves J. S. Murphree, E. H. Richardson, B. R. Sandel, K. Smith, andv transmissivity, A. Vallance Jones, "An ultraviolet auroral imager for theLion effects. Viking spacecraft," Geophy. Res. Lett. 14, 387-390 (1987).


S0

0.

,C

0

CL0

'n0

17. A. L. Broadfoot, B. R. Sandel, D. Knecht, R. Viereck, and E.Murad, "Panchromatic spectrograph with supporting mono-chromatic imagers," Appl. Opt. 31, 3083-3096 (1992).

18. B. R. Sandel, A. L. Broadfoot, B. Hatfield, E. Murad, and R. A.Viereck, "The Earth's dayglow spectrum from STS-39," EosSuppl. 29, 360 (1991).

19. F. W. Schenkel and B. S. Orgorzalek, "The Hilat vacuumultraviolet auroral imager," APL Tech. Dig. 5, 131-137 (1984).

20. F. W. Schenkel and B. S. Orgorzalek, "Auroral images fromspace: imagery, spectroscopy, and photometry," APL Tech.Dig. 8, 309-317 (1987).

21. A. F. Davidson, K. S. Long, S. T. Durrance, W. P. Blair, C. W.Bowers, S. J. Conard, P. D. Feldman, H. C. Ferguson, G. H.Fountain, R. A. Kimble, G. A. Kriss, H. W. Moos, K. A. Potocki,"The Hopkins ultraviolet telescope: performance and calibra-tion during the Astro-1 Mission," Astrophys. J. 392, 264-271(1992).

22. T. B. Coughlin, L. J. Crawford, J. Dassoulas, M. D. Griffin,P. E. Partridge, and M. R. Peterson, "Strategic DefenseInitiative," APL Tech. Dig. 13, 200-217 (1992).

23. M. D. Griffin and M. J. Rendine, "Delta 180/Vector Sum: thefirst powered space intercept," presented at the AIAA 26thAerospace Sciences Meeting, Reno, Nev., 11-15 January 1988.

24. B. B. Holland, J. C. Ray, and A. J. Green, "Delta 181, theThrusted Vector Mission," presented at the AIAA 27th Aero-space Sciences Meeting, Reno, Nev., 11-15 January 1989.

25. J. S. Hansen, G. H. Fountain, S. A. Gary, T. J. Harris, K. L.Harrison, S. E. Hawkins III, "Ultraviolet spectrographs usinglinear array technology," in Ultraviolet Technology III, R. E.Huffman, ed., Proc. Soc. Photo-Opt. Instrum. Eng. 1158,155-162 (1989).

26. J. F. Carbary and C. I. Meng, "Limb profiles from low Earthorbit," in Ultraviolet Technology III, R. E. Huffman, ed., Proc.Soc. Photo-Opt. Instrum Eng. 1158, 155-162 (1989).


Documents

Ultraviolet and visible imaging and spectrographic imaging instrument