This is an Open Access document downloaded from ORCA, Cardiff University's institutional

repository: http://orca.cf.ac.uk/102676/

This is the author’s version of a work that was submitted to / accepted for publication.

Citation for final published version:

Jennings, D. E., Flasar, F. M., Kunde, V. G., Nixon, C. A., Segura, M. E., Romani, P. N., Gorius,

N., Albright, S., Brasunas, J. C., Carlson, R. C., Mamoutkine, A. A., Guandique, E., Kaelberer, M.

S., Aslam, S., Achterberg, R. K., Bjoraker, G. L., Anderson, C. M., Cottini, V., Pearl, J. C., Smith,

M. D., Hesman, B. E., Barney, R. D., Calcutt, S., Vellacott, T. J., Spilker, L. J., Edgington, S. G.,

Brooks, S. M., Ade, Peter, Schinder, P. J., Coustenis, A., Courtin, R., Michel, G., Fettig, R., Pilorz,

S. and Ferrari, C. 2017. Composite infrared spectrometer (CIRS) on Cassini. Applied Optics 56

(18) , pp. 5274-5294. 10.1364/AO.56.005274 file

Publishers page: http://dx.doi.org/10.1364/AO.56.005274

<http://dx.doi.org/10.1364/AO.56.005274>

Please note:

Changes made as a result of publishing processes such as copy-editing, formatting and page

numbers may not be reflected in this version. For the definitive version of this publication, please

refer to the published source. You are advised to consult the publisher’s version if you wish to cite

this paper.

This version is being made available in accordance with publisher policies. See

http://orca.cf.ac.uk/policies.html for usage policies. Copyright and moral rights for publications

made available in ORCA are retained by the copyright holders.

TheCompositeInfraredSpectrometer(CIRS)onCassini

D.E.Jennings1,*,F.M.Flasar

1,V.G.Kunde

2,C.A.Nixon

1,M.E.Segura

1,2,P.N.

Romani1,N.J.P.Gorius

1,3,S.Albright

1,4,J.C.Brasunas

1,R.C.Carlson

3,A.A.

Mamoutkine1,5,E.Guandique

1,5,M.S.Kaelberer

1,5,S.Aslam

1,R.K.

Achterberg1,2,G.L.Bjoraker

1,C.M.Anderson

1,V.Cottini

1,2,J.C.Pearl

1,M.D.

Smith1,B.E.Hesman

1,6,R.D.Barney

1,S.Calcutt

7,T.J.Vellacott

8,L.J.Spilker

9,

S.G.Edgington9,S.M.Brooks

9,P.Ade

10,P.J.Schinder

11,A.Coustenis

12,R.

Courtin12,G.Michel

12,R.Fettig

13,S.Pilorz

14,C.Ferrari

15

1GoddardSpaceFlightCenter,Greenbelt,MD,20771,USA

2UniversityofMaryland,CollegePark,MD20742,USA

3TheCatholicUniversityofAmerica,Washington,DC20064,USA

4theHammersCompany,Inc.,Greenbelt,MD20770,USA

5ADNETSystems,Inc.,Bethesda,MD20817,USA

6SpaceTelescopeScienceInstitute,JohnsHopkinsUniversity,Baltimore,MD21218

7DepartmentofPhysics,UniversityofOxford,ParksRoad,Oxford,OX13PU,UK

8ExperianLtd,LondonSW1E5JL,UK

9JetPropulsionLaboratory,Pasadena,CA91109,USA

10SchoolofPhysicsandAstronomy,CardiffUniversity,Cardiff,CF243YB,UK

11CenterforAstrophysicsandPlanetaryScience,CornellUniversity,Ithaca,NY,USA

12Laboratoired’EtudesSpatialesetd’InstrumentationenAstrophysique(LESIA),ObservatoiredeParis,CNRS,UPMCUniv.Paris06,Univ.Paris-

Diderot,5,placeJulesJanssen,92195MeudonCedex,France13IndependentResearcher,Eichfeldstr.3A,76479Steinmauern,Germany

14SETIInstitute,MountainView,CA94043,USA

15CEA/Serviced’Astrophysique,91191Gif-sur-YvetteCedex,France

*CorrespondingAuthor:donald.e.jennings@nasa.gov

Received11April2017;revisedXXMay,2017;acceptedXXMonthXXXX;postedXXMonthXXXX(Doc.IDXXXXX);publishedXXMonthXXXX

The Cassini spacecraft orbiting Saturn carries the Composite Infrared Spectrometer (CIRS) designed to study

thermalemissionfromSaturn,itsringsandmoons.CIRS,aFouriertransformspectrometer,isanindispensable

part of the payload providing uniquemeasurements and important synergies with the other instruments. It

takesfulladvantageofCassini’s13-year-longmissionandsurpassesthecapabilitiesofpreviousspectrometers

on Voyager 1 and 2. The instrument, consisting of two interferometers sharing a telescope and a scan

mechanism, covers over a factor of 100 in wavelength in the mid- and far-infrared. It is used to study

temperature,composition,structureanddynamicsoftheatmospheresofJupiter,SaturnandTitan,theringsof

Saturnandsurfacesoftheicymoons.CIRShasreturnedalargevolumeofscientificresults,theculminationof

overthirtyyearsofinstrumentdevelopment,operation,datacalibrationandanalysis.AsCassiniandCIRSreach

theendof theirmission in2017weexpect thatarchivedspectrawillbeusedby scientists formanyyears to

come.

OCIScodes:(120.3180)Interferometry;(120.6085)Spaceinstrumentation;(120.6200)Spectrometersandspectroscopicinstrumentation;

(300.6300)Spectroscopy,Fouriertransforms.

http://dx.doi.org/10.1364/AO.99.0999

2

1.INTRODUCTIONThe Cassini* spacecraft has been orbiting Saturn since July2004,returningarichvarietyofscientificdataontheplanet,its largest moon Titan, its rings and other moons. Animportant component of this science has come from thethermalinfraredportionofthespectrum,aregioncoveredbythe Composite Infrared Spectrometer (CIRS). Thermalinfrared spectra contain information about composition,temperature, dynamics and structure of atmospheres andsurfaces. CIRS, built byGoddard Space Flight Center [1-3] isthemost advanced of a series of infrared Fourier transformspectrometers (FTSs) flown on NIMBUS 3 and 4, Mariner 9andVoyager1and2[4-7].Thosepreviousinstrumentswereall named IRIS (Infrared Interferometer Spectrometer).Voyager IRIS had a single 4.4 milliradian field of view andcovered the 200 to 2000 cm-1 spectral range with 4.3 cm-1resolution.CIRSwasdesignedtoimproveonIRISbysamplingmany fields of view simultaneously, by extending spectralcoverage to longer wavelengths, and by enhancing bothspatialandspectralresolutions.Thegoalwasalsotoincreasesensitivity and to provide a selection of spectral resolutionsforoptimizingthesensitivityforeachsciencegoal.TheCassini-Huygensmissionconceptwasconceivedinthe

early 1980s as an orbiter follow-on to the VoyagerreconnaissanceflybysofSaturnin1980and1981.Aprimarypurpose of the mission was to perform a comprehensiveexamination of the complex moon Titan. Although a far-infrared thermal emission spectrometer was listed in theinitialversionsoftheCassinistrawmanpayload,developmentof CIRS as a combinedmid- and far-infrared instrument didnot start until 1987. The instrument and investigation wasproposed toNASA in 1990 and selected later that year. Theinstrument was designed, built, tested and delivered to theCassini project in the period 1991-96. Cassiniwas launchedonOctober15,1997,encounteredJupiter inDecember2000andwentintoorbitaroundSaturnonJuly1,2004.Theprimemissionendedin2008,butthegeneralhealthandsuccessofall spacecraft systems justified extending the mission until2017.Duringitsthirteen-yeartour,almosthalfaSaturnyear,CIRShasstudiedseasonalchangesonSaturnandTitan,warmplume-rich regions of Enceladus’ surface, temperaturestructure of the rings, and surfaces of the smaller moons.Close-up emphasis has been on Titan during 127 targetedflybys. In April 2017 Cassini entered the final phase of itsmission,consistingofaseriesofclose“proximal”orbitsinsidetherings.CassiniwillenditslifeonSeptember15,2017whenit undergoes a planned entry into Saturn’s atmosphere. BythattimeCIRSwillhavecollectedover160millionspectra.During the development of CIRS both an engineering unit

and a flight unit were built. Extensive testing of theengineering unit proved to be invaluable and led to designchanges in the flight unit and a better understanding ofperformance prior to flight. The instrument underwent amajor design simplification in 1992 during its earlydevelopmentphaseaspartofanoveralldescopeofthe

*TheCassinimission is a cooperativeproject ofNASA,ESA (theEuropean Space Agency) and the Italian Space Agency. The JetPropulsion Laboratory, a division of the California Institute ofTechnology in Pasadena, CA, manages the mission for NASA'sScienceMissionDirectorate,Washington.

mission and spacecraft. The instrument’s original four focalplanes were reduced to three (afterwards labeled FP1, FP3and FP4) and from 21 detector signal channels to 11. Somescience capability was inevitably lost, but the descope wasbeneficial to CIRS in that the optical, mechanical andelectronics designs were greatly reduced in complexity,makingitpossibletobuildtheinstrumentwithinconstraintsoncostandmass.Development of CIRS was an international collaboration

among several institutions. The University of Oxford, UK,providedthe80Kcooleraswellasintegrationandtestingofthe mid-infrared focal planes. CEA Saclay, France, suppliedthe photovoltaic detector array. Queen Mary & WestfieldCollege, London, UK, contributed the far-infraredbeamsplitter, input polarizer, output polarizer and solarblocker.TheUniversityofKarlsruhe,Germany,developedthefar-infrareddetectors.ObservatoiredeParisatMeudonbuiltandtestedaprototypeofthescanmechanism.2.INSTRUMENTOVERVIEWThe scientific goals of CIRS on the Cassini mission were tostudy:1)thecomposition,thermalstructureanddynamicsofthe atmospheres of Saturn and Titan; 2) the thermalcharacteristicsandcompositionofSaturn’s rings; and3) thecomposition and temperatures of the icy surfaces of themoons. Observations would be global, covering latitude,longitude and altitude, and it was expected that timevariability would be found as Saturn and Titan progressthroughtheirseasons.Theaimwasnotonlyto followupondiscoveries made by Voyager, but also to search for newphenomena.The science requirements of CIRS were driven largely by

the improvements over Voyager IRIS that Cassini wasexpected toachieveatSaturnandTitan(Jupitersciencewasnot a driver of the mission and was not added until afterlaunch). The temperatures of Saturn and Titan place thepeaks of their blackbody radiance spectra near 100 cm-1.Thus,muchof thespectrumwasbelowthe200cm-1 limitofIRIS.IRIShasaCsIbeamsplitter,whichatthetimeitwasbuilthad the longest wavelength cut-off among conventionalsubstrate beamsplitters. It was therefore likely that adifferent technology would be required to reach below 200cm-1. Planned atmospheric studies included examination ofthe vertical structure both in composition and temperature.CIRSwouldneedmid-infrareddetectorsmatchedtothescaleheight in Titan’s atmosphere, about 40 km. From a flybydistanceof100,000kmthiscorrespondstoadetectorfieldofview of 0.4 mrad, about ten times smaller than the field ofview of Voyager IRIS. With the same 50-cm diametertelescope as IRIS, themuch smallerdetectorsofCIRSwoulddemand a greater detector sensitivity, especially since thespectralresolutionwasrequiredtobeimproved.ThismeantthatthermopiledetectorssuchaswereusedinIRISwouldnotbe sufficient for CIRS in the mid-infrared. The search forminor atmospheric constituents drove the requirement onspectral resolution to about an order of magnitude betterthanIRIS’s4.3cm-1.CIRS is a Fourier transform spectrometer, or scanning

interferometer [1,2].The instrument configured for flight isshown in Fig. 1 and its location on the Cassini spacecraft isindicatedinFig.2.InstrumentparametersforCIRSarelisted

3

inTable1.CIRSiscomprisedoftwointerferometersfedbyacommontelescopeandsharingascanmechanism(Fig.3).Inthe focal plane of the telescope the field image is splitbetweenthefar-infraredandmid-infraredinterferometers.Inthe far-infrared a Martin-Puplett type polarizationinterferometer[8,9]coversthe10to600cm-1portionofthespectrum. In the mid-infrared a conventional Michelsoninterferometer [10] covers the spectrum in two spectralsegments,580to1100cm-1and1050to1500cm-1.Thesetwotypes of interferometer work on different principles. In apolarization interferometer thebeamsplitter splits thebeaminto two orthogonal polarization components. The phasebetweenthetwocomponentsisvariedduringthescansothatthepolarizationstateof the recombinedbeam ismodulated.Ina conventionalMichelson interferometer thebeamsplittersplits the amplitude of the electric field into twoapproximatelyequalparts.Whenthephasebetweenthetwoparts is scanned the interference between the two electricfieldsintherecombinedbeammodulatestheoutputintensity.Examples of interferograms recorded in the three spectralsegments of CIRS are shown in Fig. 4. The instrument ismaintainedat170K,except forthemid-infraredfocalplane,which is kept at around 80 K.Warm electronics (18-22 °C)control the operation and handle the data. A block diagramshowingthebasicfunctionalcomponentsofCIRSispresentedinFig.5.The CIRS fields-of-view are shown in Fig. 6. In the far-

infrared the scene is viewed with a single field of view 3.9mrad in diameter [11]. In themid-infrared the two spectralsegmentsbothviewthescenewithalineararrayoftenpixels.Each pixel is 0.27 mrad square full width at half maximum(FWHM)andtheseparationbetweendetectors is0.29mrad.The spectral resolution of CIRS is selected by choosing thescan length. Themost commonly used apodized resolutionshavebeen0.5,1.0,3and15cm-1.Threeof theseresolutionsarepresentedinFig.7.WedefineresolutionasFWHMoftheinstrument lineshape function.Anapodizedresolutionof0.5cm-1 requires the largest optical path difference to be 2 cm.CIRS records interferograms that are up to 97% one-sidedwith 3% of the scan before zero-path difference. The totaltravelofthescanmechanismduringa0.5cm-1interferogramis1.1cmandtakes50seconds.Selectingshorterscanswhenlessresolutionisrequiredallowstheusertooptimizetheuseofobservingtime.

3.OPTICSANDDETECTORSTheopticaldesignofCIRShasbeendescribedbyMaymonetal.[12].ThemechanicalandopticallayoutsoftheinstrumentareshowninFig.8.Foreoptics, includingatelescope,afield-dividingmirrorandasolarblocker, feed the light to the far-and mid-infrared interferometers, and a laser referenceinterferometercontrolsthescanninganddatasamplingintheinterferometers.

A.Telescope.ThesmalldetectorfieldsofviewandhighsensitivityrequiredofCIRScalled fora relatively large telescope. Imaging in themid-infrareddemandedthat,inthe7-17µmspectralrange,apointsourceatinfinitywouldbeimagedintoaspotnotmorethan0.3mradFWHMdiameter.Theoriginalplanwastousea

flight back-up telescope left over from theMIRIS (modified-IRIS) development for Voyager IRIS [13]. However, testingshowedthattheMIRIStelescopeprimarymirror,althoughitmet the4.3mradMIRIS imaging requirement, lost its figureuponcoolingto170Kandwouldnotmeetthemorestringent0.3mradCIRSimagingrequirement[14].Intheend,theonlyhardwarefromtheMIRISprotoflightunitthatflewaspartofCIRSwasthetelescopecover,sunshadeandsecondarybaffle(describedbelow).

Fig. 1. Views of CIRS from the side (upper panel) and back(lower panel). In these photos covers are installed on thetelescope (yellow arrows) and cooler (orange arrow). TheCIRS electronics assembly (red arrow) is located below theoptics assembly. These photoswere taken prior to thermal-vacuumtesting.

4

Fig.2.Cassini spacecraftwith the locationofCIRS indicated(arrow).CIRSismountedontheRemoteSensingPallet(RSP)furthestfromthehighgainantenna.(NASAphoto.)A modified primary mirror design was needed, one that

wasmorerobustagainstthermaldistortionbutwhichwouldfit within the same mass and size envelope as the MIRISprimary.Thenewtelescopeprimary[15],showninFig.9,isa50.8 cm diameter paraboloid weighing 2.0 kg. Made fromberyllium, it has a light-weight stiffening structure on itsbackside.Theprimaryfocusesatf/0.7ontoa7.6cmdiameterhyperboloid secondary mirror. After figuring and polishingtheprimarymirrorsurfaceandapplyingathinnickelbarrier,a gold reflecting layer was deposited and protected with adielectric overcoat [16]. By producing an optical qualitysurfaceontheberyllium,onlyminimalnickelandgoldlayerswere required, thereby avoiding the thermal distortion thathad been seen in the MIRIS/IRIS mirror. The secondarymirror is less susceptible to thermal distortion and wastherefore nickel electroplated and polished before applyingthe gold surface. After 19 years in space no degradation intelescope performance has been found, indicating that themirrorcoatingsandfigurehavenotchangedmeasurably.Thiscoating technology, pioneered by CIRS, was subsequentlyusedontheJamesWebbSpaceTelescope[17].

The secondary mirror is mounted at the end of a 33 cmlength,7.8cmdiametertubethatisattachedtoaflangeatthecenter of the primary mirror. Three fins support thesecondarymirrorattheendofthemountingtube.Lightfromtheprimarypassesthefinstoreachthesecondaryandisthensentdowntheinteriorofthetube,atf/6,tothefocalplaneofthetelescope.Atthetelescopefocalplane,11.6cmbehindtheprimary, the beam passes through a field mask that blocksstray light. The image plane is then field-divided at a roof-shaped mirror that directs separate beams to the mid-infrared and far-infrared portions of the spectrometer.Because the detectors are smaller in the mid-infrared and,therefore, the image quality must be higher there, opticalaberrationswereminimizedbyplacingthetelescopeaxisat

Table1.CIRSInstrumentParameters__________________________________________________________________________________________________

Typeofspectrometer: FTS

TelescopeDiameter(cm): 50.8

Totalspectralrange(cm-1): 10to1500

Spectralresolution(cm-1)*: 0.5,1,3&15

Scanspeed(cm/s): 0.0208

Scantime(s): 2to50

DataTelemetryRate(bits/sec) 4000

OpticsTemperature(K) 170

ElectronicsTemperature(°C) 18-22

FAR-INFRARED

PolarizationInterferometer

Spectralrange(cm-1): 10-600

Fieldofview(mrad) 3.9diameter

Focalplane: FP1

Temperature(K) 170

Detectors Twothermocouples

Detectordiameter(µm): 1000

PeakD*(cmHz1/2W

-1) 4x10

9

NEP(W) 3x10-10

MID-INFRARED

MichelsonInterferometer

Spectralrange(cm-1): 580-1500

Detectortemperature(K) 76-87

Detectorarrays twoeach1x10

Detectorsize(µm) 200square

Detectorspacing(µm) 215

DetectorIFOV(mrad) 0.27square

IFOVspacing(mrad) 0.29

Focalplane: FP3

Spectralrange(cm-1): 580-1100

Detectortype PCHgCdTe

PeakD*(cmHz1/2W

-1) 2x10

10

NEP(W) 1x10-11

Focalplane: FP4

Spectralrange(cm-1): 1050-1500

Detectortype PVHgCdTe

PeakD*(cmhz1/2W

-1) 4.5x10

11

NEP(W) 4x10-13

__________________________________________________________________________________________________

*Apodized,seetext

5

Fig. 3. CIRS conceptual diagram. Two interferometers, mid-infrared and far-infrared, share a telescope and a scanningmechanism. A reference interferometer is inserted into themid-infrared beam in front of the moving retroreflector.Calibration is accomplishedwith reference to views of deepspaceandviewsoftheclosedshutter.

Fig. 4. Examples of interferograms from the far-infrared(FP1) and mid-infrared (FP3, FP4) in CIRS. All three arefiltered fornoisespikesanddataquality. In theupperpanelonly the portion of each interferogram around zero–pathdifferenceisshown.Inthelowerpanelthefullinterferogramsare shown with the intensity scale expanded. Theinterferograms, which correspond to 0.5 cm-1 resolutionspectra, were scanned in 50 seconds. Differences in lengthsaredue to theway sampling is performed in the three focalplanes. The slowly varying baseline in FP1 is a repeatableelectricalartifact.All three interferogramswererecordedonSaturnin2011-13andarelargeaveragesofdata.

the mid-point between the two mid-infrared fields of view.The far-infrared field is thus offset as shown in Fig. 6. Weadjusted the position of the telescope field mask duringtestingtoavoidvignettingofthemid-infraredimage.To provide some shielding against solar heating of the

primarymirror,asunshadebafflewasplacedarounditsouterrim.Thesunshadeisisolatedfromtheprimarymirrorexceptat its mount at the center of the back of the mirror. Thesunshade surrounds the back and sides of the primary andblockssunlight fromtheprimarywhenthesun ismorethan77° from the instrument boresight. Sunlight is permitted onthe primary during operation, but the combination of itsreflectingsurfaceandthesunshadeallowsthetemperatureoftheprimarytobecontrolledat170±0.1K.Abaffle,alsomadeof beryllium, surrounds the secondary and itsmount at theend of itsmounting tube. Flight rules require the sun to bekept 15° degrees from the telescope boresight for thermalstability and 5° for damage avoidance. The focus of thetelescopewasadjustedforoptimumperformanceat170K.Atthe telescope focal plane 80% of the energy from a pointsourceat632.8nmwavelengthfellwithina207µmdiameter,correspondingto0.15mradonthesky.SincethisspotsizeissmallerthantheCIRSdetectorsizesinboththefar-andmid-infrared, image resolution was limited by the detector sizeanddiffractionlimit.Theerrorinlocatingtheboresightofthetelescopewasrequiredtobelessthan±0.06mradtocontrolits contribution to the overall 1-mradboresight tolerance oftheinstrument.Inflight,theboresightofCIRS,definedasthecenterbetweenthemid-infrared focalplanes,wasmeasuredby scanning Jupiter and was found to be offset from thenominal spacecraft pointing by 1.7 mrad away from thespacecraft body (spacecraft +X direction) and 0.04 mradtoward the high-gain antenna (spacecraft –Z direction). Adeployable coverprotected the telescopeon thegroundandinflight.Thecoverwasdesignedtopreventcontamination

Fig. 5. Block diagram of CIRS. The electronics assemblycontains all warm electronics boards and provides power,operational control, signal handling, data processing andtemperaturecontrol for the instrument.Theopticsassembly(170K and80K ) contains thedetectors andpreamplifiers,the scanmechanism and the reference interferometer. Data,controlandpowerpathsareindicatedwitharrows.

telescope

KBr

beam-

spli1erlens

reference

interferometer

cubecorner

retroreflector

mid-IR

collimatorfar-IR

collimator

inputpolarizer

andabsorber

shu1er solar

blocker

field

mask

scan

mechanism

dihedral

retroreflector

polarizer

beamspli1er

far-IRfocal

plane

mid-IR

focal

plane

Samplenumber

Signal(arbitraryunits)

processor

bus

interface

unit

OPTICALASSEMBLY

spacecra;

power

scan

mechanism

FP1

reference

interferometer

FP3 FP4

170K

80Kcooler

telescope

powerconverter

assembly

scanmechanism

electronics

&shuKer

reference

interferometer

electronics

instrument

datasystem

temperature

controller

&monitor

front-end

electronics

ELECTRONICSASSEMBLY

housekeeping

18-22C

spacecra;

clock&

commands

6

and to block sunlight when the spacecraft was in the innersolarsystem.FP1,thefar-infraredfocalplane,wasespeciallyvulnerable to damage from direct solar exposure. Having aconicalshape,theberylliumcover(Fig.1)formedasealattherimoftheprimarysunshadeandtaperedtoaclose-fitaroundthe outside of the secondary baffle. A wax-actuatormechanismwith built-in redundancy, located at the apex ofthecoveronthesecondarybaffle,wasactivatedtodeploythecover.Originally theplanwastoeject thecover justprior toarrival at Saturn. However, when the mission obtainedapproval for science operations at Jupiter we analyzed thepossibility of an accidental exposure and decided to accepttherisk.Thecoverwasremovedon20September2000,justbeforetheJupiterencounter.

Fig. 6. CIRS fields-of-view measured using bright stars andrasterscansofJupiterjustpriortoreachingSaturn.TheX,Y-origin is the nominal spacecraft pointing axis. The CIRStelescope axis is offset by X=+1.70 and Y=-0.04 mrad. ThephysicaldiameterofFP1is3.9mrad,butthehalf-peak(50%Amp.)responseismeasuredtobe2.5mrad.Smalldeviationsin position of FP3 and FP4 detectors are within themeasurementaccuracy.

Fig.7.ThreeresolutionsofFP4spectraonSaturn.ThesedataareaveragesofspectraselectedfromthewholediskofSaturnduringthetimeperiod2008-12.CIRSresolutionisselectedbychoosingthelengthoftheinterferometerscan.Spectraat0.5cm-1and3cm-1resolutionareoffsetinradiance.

Fig.8.Mechanical(upper)andoptical(lower) layoutsof theCIRSinstrument.Twointerferometerscoverthemid-infraredand far-infrared and share both the telescope and the scanmechanism

0.0E+00

2.0E-09

4.0E-09

6.0E-09

8.0E-09

1.0E-08

1050 1100 1150 1200 1250 1300 1350 1400 1450 1500

Radiance(W/cm

2/str/cm

-1

Wavenumber(cm-1)

0.5cm-1

3cm-1

15cm-1

0.5 cm-1

3 cm-1

15 cm-1

mid-

infrared

far-infrared

scanning

retroreflectors

25cm

telescope

80Kcooler

7

B.Far-InfraredInterferometer.The requirement to cover 10-600 cm-1 in the far-infrareddrove us to choose a polarization interferometer as thespectrometer for that portion of the spectrum. Wire gridpolarizers are inherentlybroad-band in that theyhave goodefficiencyatallwavelengthslongerthanabouttwicethewirespacing. A drawback is that using only one polarizationintroducesalossofhalfoftheincomingradiationandmakesthe instrument polarization-sensitive. In CIRS wire-gridpolarizersservedasinputpolarizer,beamsplitterandoutputpolarizer in the far-infrared interferometer. The polarizerswere deposited on 1.5-µm thick mylar pellicle substrates.Individualwiresare2µmwideandthewirespacingis4µm(i.e., a2µmgapbetweenwires). In theoriginalCIRSdesignthepolarizerswerefree-standingwire-grids,butinresponseto the 1992 descope we changed to substrate-mountedpolarizers to recover the spectral region of the lost FP2.Because a substrate permitted smaller wires and narrowerspacing,aswellasasmoothersurface,theshortwavelimitofthe polarizer was extended to 600 cm-1. Mylar has spectralfeaturesinthefar-infrared,buttheydonotappearinspectraaftercalibration.Thefar-infraredopticshavebeendescribedbyCrookeand

Hagopian[18,19].Opticaldesignofthefar-infraredportionofCIRSisshowninFig.10.Comingfromthefieldmaskandroofmirror, the far-infrared beam first passes through the solarblocker, a filter that transmits only the far-infrared beyond700cm-1.Thesolarblockerprotectsthefar-infrareddetectoragainstdamagefromaccidentaldirectexposuretothesunatSaturn.Beyondthesolarblockerthebeamiscollimatedatanoff-axisparabolicmirrorandthenstrikestheinputpolarizer.Itswire grid is rotated from vertical (normal to the page inFig.8)suchthattheportionofthebeamthatisreflectedfromthe input polarizer toward the beamsplitter is polarized at45°.Thewiregridofthebeamsplitterisorientedvertically,sothebeamissplitintothetwoarmsoftheinterferometerwiththe reflected beam polarized vertically and the transmittedbeam polarized horizontally. These are sent to dihedralmirrorretroreflectors,oneofwhich is fixed inplaceand theother attached to the scan mechanism. Each dihedralinterchanges the horizontal and vertical orientations of thebeams and sends them back to the beamsplitter. Switchingthepolarizations reverses transmissionand reflectionat thebeamsplitter and causes both beams to be sent to the focalplane. Moving the retroreflector changes the difference inoptical path between the two beams, which changes thepolarization state (ellipticity) of the combined beam. In theFP1focalplaneassembly(Fig.11,Fig.12)aparabolicmirrorfocuses the beam at an output polarizer analyzer. Theanalyzerwiregridisorientedat45°andalternatelytransmitsor reflects the beam as the polarization changes. Thismodulates thesignals fromtwothermocoupledetectors thataresensingthetransmittedandreflectedbeams.Becausethesignalsfromthetwodetectorsare180˚outofphase,theyarewired in series with opposite polarity to double the signal.Duringthemodulation,wheneachdetector isnotseeingtheinput from the telescope, it is instead viewing an absorber

Fig. 9. The CIRS all-beryllium 50.8 cm diameter telescope.Incoming light is focused by the primary mirror onto thesecondarymirror(left),whichthensendsanf/6beamtotheimage plane located 11.6 cm behind the primary (right). Analuminum tube in the form of cut-out fingers makes thetransition from the beryllium telescope to the aluminumopticshousing.

Fig. 10. Optical diagram of the far-infrared section of CIRS.Light from the telescope image (field mask) is sent to acollimating mirror and flat input polarizer and on to theinterferometer. A polarization beam splitter splits the beaminto two orthogonal components. Dihedral retroreflectorssend thebeamsback to thebeamsplitterwhere theyare re-combined. In the focal plane assembly an output polarizersendsmodulatedlighttothethermocoupledetectors.

8

plate located behind the input polarizer. Scanning theretroreflector causes the signal at eachdetector to alternatebetween the telescope and the absorber, creating aninterferogram (Fig. 4). The absorber has an emissivity of98.4%.Itstemperature,near170K,ismonitoredandusedinthe radiometric calibration. Within the interferometer thecollimatedbeamhasadiameterof48mm.Thethermocoupledetectors are each 1 mm diameter with a Winston coneconcentratoracceptinganf/1focusedbeamfromthecameramirror.Thedetectordiameterprojectsto3.9mradinthefieldof view (Fig. 6). In-flight mapping showed that the far-infrared field of view sensitivity pattern is approximatelyGaussianwithaFWHMdiameterof2.5mrad.Thermocouple detectors were chosen for the far-infrared

because they had the highest sensitivity of any type ofdetector operating at 170 K. We used Schwarz-typethermocouplesmadefromgoldfoilattachedtop-andn-typebismuthtellurideposts.ThedetectorsweredevelopedbytheUniversityofKarlsruhe[20].Agold-blacklayercoatsthegoldfoil to enhance infraredabsorption.At the0.02 cm/sec scanspeedoftheretroreflector,modulationonthedetectorsis intherange0.4-25Hz.Atthesefrequenciesourthermocoupleshad detectivities greater than 4x109 cm√Hz/W. The signalfrom the detectors is boosted by a factor of 100 in atransformer and preamplifier that are housed in the warmelectronics assembly. The transformer matches the lowimpedancedetectors,about10ohmseach,tothenoise-levelsof the preamplifiers. The detector circuit providesredundancy in the event of failure of one of the detectors.With the throw of a latching relay, the circuit switchesbetween series and parallel operation. Series is the nominalmode in which the two out-of-phase detector signals areadded. In series mode, failure of one detector would alsoeliminate the second detector, but by switching to parallelmodetheremainingdetectorcouldstillbeused.Althoughtheparallelmodewastestedonthegroundbeforelaunch, ithasnotbeenneededinflight.

C.Mid-InfraredInterferometer.The optical design of the mid-infrared portion of CIRS isshowninFig.13[21].Inthemid-infraredinterferometer,thebeamcomingfromthetelescopefieldmaskandroofmirroriscollimated and sent to the beamsplitter. The dispersionintroducedbythebeamsplittersubststrateisbalancedinthecompensatorbyhavingbothtransmittedandreflectedbeamspass through the same thickness of substrate. A layer ofgermaniumononesurfaceformsthebeamsplitter,whileanti-reflectionandprotectivecoatingsareappliedtoallsurfacesofthe beamsplitter and compensator. Both beamsplitter andcompensator substrates are KBr, and they share a commonmount.Toeliminatefringing,bothsubstratesarewedged0.1°andthegapbetweenthemiswedged2°.Thebeamisdividedapproximately50-50%inamplitudeandpassedtotwocube-cornerhollowretroreflectors,oneofwhichisattachedtothescan mechanism. At the beamsplitter, the return beams arerecombinedwithaphaseshiftandsenttotheinterferometerfocalplane(Fig.14).Asthemovingretroreflector isscannedthe phase changes between the two beams and theinterference causes the intensity at each detector to bemodulated (the combined beam alternates between thedetectorandthetelescope).Thedetectorsignalthereby

Fig. 11. Diagram of the FP1 assembly for the far-infraredportion of CIRS. Polarization-modulated far-infrared lightcomingfromthebeamsplitterisfocusedonasmallpolarizeranalyzer that passes or transmits the two orthogonalpolarization states. Two thermocouples detect theinterferogramsignals.

Fig.12.TheFP1assemblyinCIRS.The48mmdiameterbeamis focussed at f/1 onto a polarizer analyzer and then toWinston cone concentrators in front of two thermocoupledetectors.Aholeandmountingflangeontheopticsassemblyaccepts the FP1 assembly. Cut-outs in the cylindrical wallsdampenacousticdisturbancesduringlaunch.createsaninterferogram(Fig.4).Thebeam,whichis45mmdiameter within the mid-infrared interferometer, is focusedby a lens at f/1.2 onto the focal plane. The anti-reflectioncoated germanium lens forms an image at the focal planewhere two 10-element HgCdTe linear detector arrays, FP3and FP4, are mounted. FP3 has a photoconductive arraycovering 580-1100 cm-1,while FP4 has a photovoltaic arraycovering 1050-1500 cm-1. Each detector element is 200µmsquare.Measurementsprelaunchandin-flightshowthatthe

9

Fig.13.Optical diagramof themid-infrared sectionofCIRS.Light from the telescope image (field mask) is sent to acollimatingmirrorandfoldmirror,whichdirectsthebeamtothe interferometer, where it is split at a KBr beamsplitter.Cube-corner retroreflectors send the beams back to thebeamsplitter for recombining. A lens focuses themodulatedlight on the HgCdTe detector arrays. The referenceinterferometer uses the center of the mid-infrared cube-corner.detectorsareeach0.27mradFWHMinthetelescopefieldofview and are separated center-to-center by 0.29mrad [22].The FP3 and FP4 arrays are attached to the cold finger of apassive radiative cooler described in section 5. The twoarraysarealignedwiththeirdetectorsnexttoeachotherandseparated center-to-center by 0.68 mm. The photovoltaicdetectors were developed by CEA Saclay, France and thephotoconductivedetectorsweredevelopedatGoddardSpaceFlightCenter[23].BothdetectorarraysweredeliveredtotheUniversity of Oxford [24], where they were precisely co-aligned,integratedwiththelens,focusedandboresighted.IntheoriginalCIRSdesigneachmid-infraredarraywasreadby10amplifierchannels.However, thenumberwasreduced tofivechannelsduringthe1992descope.Toretainflexibilityinobservations, a switching circuit apportions the tendetectorsignals among the five channels in three ways: 1) the 5channels are toggled between odd and even detectors forobservations in which high spatial resolution mapping isdesiredwithoutmovingthespacecraft;2)adjacentdetectorsare combined in each of the five channels when full spatialcoveragewith reduced spatial resolution is required; and3)the five center detectors are connected to the five outputchannels for high spatial resolution with reduced spatialcoverage. All three of thesemodes have been used in flight,themostcommonbeingoddandeven.

D.ReferenceInterferometer.The motion of the scanning retroreflectors in CIRS ismonitoredandcontrolledbyareferenceinterferometer(Fig.

Fig. 14. The CIRS mid-infrared focal plane assembly. TwoHgCdTe detector arrays are mounted on a stage that issupportedbyatripodoftitaniumlegs(center)attachedtothelensassembly(right).Aflexiblecopperstrap(right)providesthethermalconnectiontothecoolercoldfinger.Infraredlightfromthebeamsplitter(totheleft)isfocusedbyagermaniumlens(left)ontothefocalplane.15). The reference interferometer uses a solid state diodelaser (SDL model 5601, containing two lasers) to produceinterference fringes thatare regularly spaced inopticalpathdifferenceandareusedtocontrolthescanandtotriggerdatasampling. The reference interferometer has been describedbyMartino,andCornwell[25].Weusedalaserinsteadoftheneon bulb used in previous IRIS instruments to improveoutputintensityandreliability.Thelaseroperatesatthe170Kopticstemperatureandemitsat784.5±0.3nmwavelength.Since thewavenumber scale inCIRS spectradependson thelaser wavelength, we calibrated the wavelength when wereached Jupiter using known, strong lines of atmosphericmethane near 1330 cm-1 [26]. When the laser changedfrequency on occasion later in themissionwe repeated thespectral calibration using strong methane lines in Titan’sspectrum. This process is reviewed further in section 8.B. A“white light” interference signal is also produced by thereference interferometer. A light emitting diode (LED,Opto-Diode model OD880) with a broad emission spectrumbetween800and920nmproducesanarrowburstoffringesatzero-pathdifference(ZPD)ofthereferenceinterferometer.ThewhitelightZPDprecedestheinfraredZPDby347µmandtriggers startofdata collection.Asa result, thebeginningofscan is very repeatable, producing consistent sampling andpermittingon-boardcoaddingofsuccessiveinterferograms.In theearliestdesignofCIRS thereference interferometer

beam followed the same path as themid-infrared, includingthrough the beamsplitter and both cube cornerretroreflectors. In principle this would provide optimaltrackingof thepathdifference inthemid-infrared.However,it was realized that from an engineering perspective therewerereasonstonotsharealloftheopticswiththeinfraredinCIRS. First, two beamsplitter coatingswould be required on

10

theKBrbeamsplitterwithoneinthecenteroptimizedfortheshort-wavelength reference interferometer. Eliminating thecentral coating would simplify the beamsplitter and reducerisk. Second, if the scan mechanism could be operatedseparately from the mid-and far-infrared interferometers,development and testing of the scan mechanism would begreatly simplified.We decided to have a separate referencepath sharing only themovingmid-infrared cube-corner (thealternative of placing the reference interferometer on theoppositeendofthescanmechanism,aswasdoneinVoyagerIRIS,wasnotanoptionbecausethatpositionwasoccupiedbythe far-infrared retroreflector). Although using differentopticalpathsfortheinfraredandreferencebeamsisnotideal,the little relative drift we have seen is corrected duringcalibration by making a phase adjustment. Since the mainrequirementofthereferenceinterferometerwastotrackthemotion of the cube-corner retroreflector during scans, weplaced the reference interferometer directly in front of theretroreflector. Laser and LED beams are first joined at apolarizing beam combiner andpass through a quarter-waveplate to prevent laser light from returning to the laser. Thebeamistheninsertedintothecenteroftheinfraredbeambya foldmirror and sent to thebeamsplitter. Foldmirror andbeamsplitterresideentirelywithinthecentralportionofthetelescope beam, i.e., inside the shadow formed by thesecondary mirror. The glass cube beamsplitter and a 45°prismmirror are glued into a tube to create a single, stableunit. From the beamsplitter half the beam is transmitted tothe scanning retroreflector and half to a small, stationarycube-corner retroreflector. The return beams from theretroreflectorsarerecombinedatthebeamsplitterandsent

Fig.15.Referenceinterferometeropticallayout.Asolidstatelaserprovidesthereferencefringesandalightemittingdiodecreates the broadband “white-light” signal. The referenceinterferometermeasuresthemotionofthemid-infraredcubecornerretroreflector.to the reference detectors. Another beamsplitter separatesbeams between the laser and LED detectors. Pinholes areused in frontof thedetectorsto further isolatethe laserandLEDlight.SiliconPINphotodiodes(EG&GmodelSGS100)areusedforboththelaserandwhitelightdetectors.Apulsetraingenerated by the fringe zero-crossing is sent to the scanmechanismelectronicsforscancontrol.The reference laser has operated flawlessly since the

Cassinilaunchin1997.Wehavenotneededthebackuplaser,but we did verify during the mission that it worked. There

havebeennospectraldrifts,exceptthatthelasersometimeschanges wavelength mode when it is turned on after aninstrument sleep period [27]. Because we know the modespacing we can easily change the spectral calibration tocompensate for mode hops (section 8.B). During the longperiodsbetweenthetimeswhentheinstrumentisturnedoffthe wavelength is very stable. We have seen a long-termdecreaseinthelasersignal,about8%overthemission.Thatmay be due, at least in part, to a slight drift in the opticalalignmentover time.The laserhasbeenusedbeginning justafterlaunch,duringcruiseforinstrumentchecks,atJupiterin2000-01, and at Saturn since 2004. The total amount of on-time isover thirteenyears.Housekeeping information isnotavailable for the LED, but it has shown no signs ofdeteriorationinperformanceoverthemission.

4.MECHANICALDESCRIPTIONCIRSconsistsoftwoprincipalsubsystems,anopticsassemblyandanelectronicsassembly(seeFig.5).Theopticsassemblycontains thetelescope,scanmechanism,80Kcoolerandthemid-and far-infrared interferometers.Theopticsassembly issupportedonthespacecraftremotesensingpalette(RSP)byathin-walledtitaniumtube,whichallowstheopticsassemblyto be cooled to 170 K (except for the 80 K cooler andmid-infrared focal plane). Aluminum was used throughout theoptics assembly except for the beryllium telescope. Thetelescope was made of beryllium to take advantage of itsthermal uniformity and low thermal expansion. Toaccommodate the difference in thermal expansion betweenberylliumandaluminum,aflexiblealuminumsupportmakesthe transitionbetween the telescopeand theopticshousing.For ease inmachining and toprovide surfaces formountingthe various mirrors, the optics housing is designed with arectangularcross-section,ratherthanthetubulardesignusedinthepreviousIRISinstruments.Externalflatsurfacesonthehousing serve as mounts for mirrors, far- and mid-infraredfocalplaneassemblies,referenceinterferometercomponentsand calibration shutter. External alignment adjustments aremadebyshimmingat theattachmentpoints. Internalopticaladjustments are performed with shims or sliding wedgemechanismswith locks.Removablecoversallowedaccess tothe interior optical components during integration andtesting. Polarizers in the far-infrared, and beamsplitter andcompensator in the mid-infrared, are held with springsagainstprecisionsurfacesintheirmounts.CIRS has two mechanisms, a scan mechanism and a

calibration shutter. The design of the shutter was adaptedfromDIRBE on COBE [28]. The shutter consists of amirrormountedonanarmthatispulledintothemid-infraredbeamby an electromagnet. A failure of the electromagnet wouldleave the shutter in the open position, thereby allowing theinstrumenttocontinuetooperate.Duringgroundtestingandinflightthecalibrationshutterhasperformedflawlessly.Thecalibrationshutterisdescribedfurtherinsection7.We based the CIRS scanmechanism on Voyager IRIS, but

theCIRSdesignneededtoincreasethescanlengthto11mmfromthe1.6mmofIRIS.TheCIRSscanmechanismisshowninFig.16.ItisdescribedindetailbyHakunandBlumenstock[29].AprototypeoftheCIRSscanmechanismwasdescribedbyMichelandCourtin[30].Thescanmechanismmass is1.8

11

Fig. 16. CIRS scan mechanism (engineering unit) preparedfor vibration testing. Paired leaf springs, each consisting ofthreeblades,supportacentralmovingshaftwithamaximumlinear travel of 1.4 cm. A cube corner retroreflector ismounted on themoving shaft. The scanmotion is along theaxisofthecube-cornerandperpendiculartotheplanesoftheleafsprings.kg and it is approximately 15 cm across. Themechanism isbased on leaf flexures arranged in a symmetrical three-leafpattern. At each end of themechanism a flexure leafmakesthetransitionfromthefixedmechanismbodytooneofthreeintermediate “outrigger” rods, and another leaf makes thetransition from the rod to themoving shaft. The leaves arestiffenoughtokeeptheshaftfromsaggingappreciablyduringtesting in1g.Astheshaftmovesall leavesbend,sothattherods “breathe” in unison, and the shaft travels in a straightline. Scaling the scan length up from IRIS presented somechallenges.Onedifficultywasthattheleafflexuresneededtobelongerinordertoachievethescanlengthandavoidover-stress that would distort the springs. But increasing thelengthdecreased the leaf stiffness.Withberyllium-copperasthematerial, an optimum length for the flexure arms is 8.7cm,givinga1.1cmtravelduringtheinterferogramscanand1.4cmtraveltotheendlimits.Toavoidwobbleatthecenteroftravelwheretheleafspringsareinaneutralposition,carewas taken in matching and aligning the leaves duringassembly.Vibrationofthearmsduringlaunchandoperationwas minimized with passive eddy-current dampers on theoutriggerrods.Cube-corner and dihedral retroreflectors are attached to

the ends of the tube-like scanning shaft [31]. Maximum tiltduringmotionwasmeasured before launch to be ±2 arcsecand shear was less than 18 µm. A linear motor moves theshaft.Thepermanentmagnetsofthemotorareplacedonthemoving shaft to eliminate heat conduction throughwires. Awire bobbin provides the changing magnetic field duringscanning. With the bobbin fixed in place, problems withmoving wires are avoided. Differential eddy currentmodulationinarodlocatedinsidethebobbingivesameasureof displacement. This provides coarse scan control as analternative to the reference interferometer. Scan motion iscontrolledby an electronics board located in the electronicsassembly [1]. During data-taking, fine control is maintained

bylockingapulsetrainfromthereferenceinterferometertopulsesgeneratedbyastableoscillator.Withthescanvelocityof 0.0208 cm/sec, digital pulses formed from the zero-crossings of the laser fringes occur at 1066Hz. Scan rate iscontrolled to better than 1%. During flyback following eachscan the mechanism is out of lock and moves at about 0.5cm/sec.For launchtheshaftwasdriventooneendof travelandsecuredwithayokeona lockingmechanism.Unlockingtookplaceoneweekafterlaunch.Themechanismisoperatedmost of the time, even when not taking data, to minimizetemperaturevariations.5.TEMPERATURECONTROLCIRShas four temperaturezones,eachwith itsown thermaloperating range and control. These zones are the: 1) opticsassembly;2)telescope;3)80Kcoolerandfocalplane;and4)warm electronics. Multi-layer insulation covers all surfacesexcept for the 170 K radiator, the 80 K cooler annulus andcoldpatch,andthetelescopeaperture.Thewarmelectronicsassembly is connected to the cold optics assembly by athermally isolated harness that accommodates the ~125 Kdrop between the two assemblies. Thermistor sensors areused in the optics assembly, silicondiodeswereused in thetelescope and 80 K focal plane, and a platinum resistancethermometer(PRT)isusedonthe80Kcoolercoldpatch.Theoptics assembly is cooled to 170 K with a thermal radiatorconsisting of two walls of the housing exposed to space.Temperatures are monitored with thermistor sensors at avariety of locations in the assembly, including a thermistornear the radiator panel heater for temperature control.Another thermistor, attached to the absorber on the far-infraredinputpolarizer,isusedforcalibrationofFP1spectra.Proportional and derivative feedback control of the heatercurrentkeepsthetemperatureoftheopticsassemblywithin170±0.1K.Thetelescopeisalsokeptat170K,butcontrolledseparately from the optics assembly because the transitionmount between the telescope and the optics assemblyreduces the thermal exchange between the two regions. Inaddition,thetelescope’stwomirrors,primaryandsecondary,arecontrolledindependently.Botharekeptat170±0.1Kwithheaters and proportional feedback circuits. Thermistors onthe mirrors provide monitoring and temperature control.Monitoring the telescope sunshade and secondary bafflehelpsmaintainauniformtemperature.BeforelaunchthePRTlocatedonthe80Kcoolercoldpatchfailed,mostlikelyduetoa broken electrical contact. Since it was not needed fortemperature control, and toavoid risk, adecisionwasmadeto flywithout repairing the sensor.However, during aTitanflyby in October 2006 the sensor suddenly returned to fulloperation. It has worked normally during the remainder ofthe mission. CIRS was allocated 9 W of power for thermalcontrol.Near the endof themission theheaterpower camevery close to this limit, but fortunately did not exceed it. Alesson learned from this experience is that an instrumentshould not be limited to the pre-launch power allocation incase it is found that more power is needed in flight. Extrapowersometimesbecomesavailablein-flight.OurchoiceofHgCdTedetectorsforthemid-infraredmade

itnecessarytocoolthatfocalplanetotemperaturesbelow80K. A passive radiative cooler was developed for the mid-

12

Fig. 17. The CIRS 80 K cooler cools the mid-infrared focalplaneintherange75-85K.Leftpanel:Thecoldpatchisblack-painted honeycomb. The white annulus radiator cools thehousing rim to about 130 K. A cover release actuator isattached to theouterrim.Rightpanel:The titaniumhousingsupportsthecoldpatchandsurroundsthefocalplane.Cablesconnecttothefocalplaneelectronics.Atrightthebackofthecube-cornermountcanbeseenattachedtotheoutsideofthemainopticsassembly.infraredbytheUniversityofOxford[24].Thecoolerisshownin Fig. 17. This 80 K coolermaintains the detectors at theiroperating temperature by radiating to space. The coolergenerally keeps the detectors near 76 K, but occasionallywhen solar or planetary heating cannot be avoided thetemperature rises above 80 K. Focal plane temperature iscontrolled to one of a series of setpoints so that thetemperature can be kept stable during observations. Thecooler cold patch is a 40 cm diameter radiator disk aimedaway from the spacecraft toward deep space (along thespacecraft +X axis). The cold patch was protected on thegroundandduringearlycruisebyaspring-deployablecoverthatwasreleasedinApril2000priortoarrivalatJupiter.Thecold patch surface is an aluminum honeycomb coated withconductiveblackpaint formaximumthermalemissivity.Thecold patch is suspended by stainless steel wires from atitaniumringthatinturnisattachedtotherimofatitaniumsupport housing. An aluminum annulus with high thermalemissivity covers the rim of the housing and acts as asecondary radiator to cool the rim. Except for the annulus,multi-layerinsulationcoversthecoolerhousing.Acoldfingerisattachedtotheinsidecenterofthecoldpatchandconnectsvia a flexiblemulti-layer copper foil strap to the focal planemount (Fig. 14). The two 10-element detector arrays arebondedonto carriers that are attachedonto themountwithscrews and positioned using dowels against referencesurfaces. The cold focal planemount is attached to the lensassemblybyathermalisolatorintheformofatripodmadeoflegs of titanium alloy. Cold (170 K) front-end preamplifierelectronicsforbothFP3andFP4aremountedinthehousingadjacent to the focal planes to minimize the cable lengths.FollowingtestingatOxford[22]thefullassembly(coolerplusFP3 & 4) was delivered as a unit to Goddard Space FlightCenterforintegrationwiththeCIRSopticsassembly.

Survivalheatersontheopticsassemblyand80Kcoolerareactivated by the spacecraft when the instrument is off.Decontamination heaters can also be operated by thespacecraft while CIRS is in off mode. The decontaminationheatershavebeen turnedon twiceduring themission, onceprior to mechanism latch release and again just beforeejectionofthetelescopecover.Theheaterswarmedthewaxactuators and drove contaminants onto the cover.Throughoutthemissiontherehasneverbeenanyevidenceofcontamination.6.ELECTRONICS/OPERATIONALMODESThe electronics assembly, housing the warm electronicsboards (see Fig. 5), is mounted directly to the spacecraftremote sensing palette. Circuit boards in the electronicsassembly include a power converter assembly (PCA), scanmechanism electronics (SME) and reference interferometerelectronics (RIE), an optics temperature controller/monitor(TCM), front-end electronics (FEE) and the instrument datasystem(IDS).ThePCA,comprising fourboards,converts theprimary30Vspacecraftbuspowerforsecondarysubsystemsrequiring±15V,±12Vand+5V.ThePCAalsooperated thecoverreleasesforcoolerandtelescope.TheSMEcontrolsthescanmechanismandoperatesthecalibrationshutter.TheRIEprovides a pulse comb for scanmechanism velocity controland for triggering data sampling. The TCM stabilizes theFP3/FP4 focal plane temperature at commandable setpointsbetween 75 K and 87 K. The TCM alsomaintains a uniform170Kintheopticsassembly.TheFEEprovidesconditioning,amplificationanddigitizationoftheinfrareddetectorsignals.The IDS isamicroprocessor thatcontrolsall functionalityofthe instrument, accepts commands, processes data andprovides telemetry. The electronics temperature variedduring themissionas the solardistance increased, changingfrom about 22 °C at Jupiter to about 18 °C at Saturn.Preamplifiers for FP3 and FP4 and for the referenceinterferometer detectors are housed in the 170 K opticsassembly.CIRS has five operating modes: Pre-Sleep, Sleep,

Operational, Initialization, and Science. The various modesaredescribedinFig.18,alongwiththeirapproximatepowerconsumptions. From Off the instrument transitions to Pre-SleepmodeinwhichthePCA,IDSandTCMareonandflightsoftware is loaded. The scan mechanism launch lock wasunlatchedinPre-Sleepmode.Safetywatchdogcommandscantransition any mode (except Off) into Pre-Sleep. When asoftware load is completed and has been validated in Pre-Sleep,atransitionisautomaticallyinitiatedandtheIDSflightsoftware reinitializes itself as it switches to Sleep mode. InSleepthePCA,IDSandTCMareonandthemid-infraredfocalplane temperature is set. Sleepmode serves as the standbymodeof the instrumentand fromSleepanyothermodecanbereached.ThepurposeofInitializationmodeistoallowthetelescope and cooler cover releases to proceed withoutextraneous instrument operations and power usage.Initializationmodehasonlybeenusedduringcoverreleases.In InitializationmodePCA, IDSandTCMcanbeonoroff. InOperational mode engineering data and diagnostics can berun,butnosciencedatacanbeproduced.PCA, IDSandTCMare on in Operational mode and the scan mechanism,reference interferometer and front-end electronics are

13

normally on. In Science mode all electronics are turned onand science datamay be recorded (engineering data cannotbecollected).Housekeepingdata is transmitted inallmodes(except Off). In anymode, in the case of a spacecraft safingevent, the electronics are turned off in stages to protect theinstrument.Datarateandvolumeareminimizedbyoptimallysampling

thedetectors.Allelevendetectorsignalsareinitiallysampledat1066Hz,correspondingtoeveryzero-crossingofthelaserfringesignal.Thesignalsareelectricallyfilteredtolimittheirbandwidths. Focal planes 3 and 4, which are aliased in 2ndand 3rd order, are further bandwidth-limited by numericalfiltering.Numerical filtering isalsoappliedtoFP1.The1066Hzrateisreducedbyunder-samplinginallthreefocalplanes.Thefinalunder-samplingfactorsare18,22.5,and25forfocalplanes1,3and4,respectively.Datavolumeisalsominimizedby spatial averaging using detector pairing, by tailoringspectral resolution, and by onboard coadding of successiveinterferograms. In planning observations it must be kept inmindthateachofthesetypesofobservationrequiresitsownreferencedata,recordedinthesamemanner,forcalibration.

Fig. 18. CIRS operational modes. Use of multiple modespermits efficient management of power consumption andallows the Science mode to be dedicated to data taking.Engineering data are recorded in Operational mode. Sleepmode serves as the standby mode in which power isminimized. Operationalmode permits engineering functionsanddiagnosticstotakeplacewithoutrecordingsciencedata.

7.RADIOMETRICCALIBRATIONCIRS is calibrated by comparing with reference sources ofknown temperatures. The procedure is different for the far-infraredandmid-infrared.Inthefar-infraredalloftheoptics,housingand focalplaneareat the same170K temperature.Calibration with uniform temperature requires only onereferenceblackbody.Theconditionofuniformtemperatureissimilar to that of Voyager IRIS [7, 10], with one importantdifference. Since inCIRS the interferometer is apolarizationmodulator, during a scan the input is either reflected ortransmittedattheinputpolarizer.Reflectionatthepolarizerviewstheskythroughthetelescope,whiletransmissiongivesa view of the 170 K absorber behind the polarizer. Theabsorberisablackhoneycombwithemissivityclosetounity.Since it issurroundedbysurfacesallat170K,radiancefromthe absorber is effectively blackbody emission at 170K.Observing deep space, away from sources in the Saturnsystem,isthereforeequivalenttoviewingablackbodyat170K but inverted in modulation phase. Deep spaceinterferograms,Ids,arethususedtoderivetheresponse

R(v) =FT(−I

ds)

B(170 K) (1)

whereFTistheFouriertransform.HereνisthewavenumberandBistheblackbodyradiancespectrum,orPlanck,functionfor 170 K. Writing It for the interferogram from viewing atarget,thecalibratedspectrumfromFP1is

St(v) =

FT(It− I

ds)

R(v) (2)

In themid-infrared there are two temperature zones.The

opticsmoduleisat170Kandinadditionthemid-infraredhasan 80 K focal plane. Views of two reference sources, onewarmandonecold,arerequiredtoobtaintheresponse[10].For the cold reference CIRS observes deep space, whichcorresponds to a source essentially at 0 K. For the warmreference CIRS observes the 170 K insidewall of the opticshousing.Acalibrationmirrorshutterdivertstheviewtowardtheinteriorwall.Theshutterislocatedbetweenthetelescopefocalplaneandthemid-infraredcollimatingmirror.Asensorprovides the temperature of the wall at the shutter viewlocation. The temperature of the housing varies by <0.02 Kduring an ~1 hour observing sequence and by <0.1 K overlong terms. We assume that viewing the housing wall iseffectively looking at ablackbodyat the sensor temperaturewith unit emissivity. This is justified because the shutter iseffectively viewing the inside of a 170K cavity. To calibratethemid-infrared,interferogramsofdeepspaceIdsandshutterIs are recorded at times adjacent to target interferograms It.BecausetheresponsesandnoiseoftheFP3andFP4detectorsdepend on temperature, the target, deep space and shutterinterferograms must be recorded at the same focal planetemperature,oracorrectionmustbeapplied.TheresponsesofFP3andFP4arefoundfrom

14

R(v) =FT(I

s− I

ds)

B(170 K) (3)

andthecalibratedspectraarecalculatedas

St(v) =

FT(It− I

ds)

R(v) (4)

Our technique of subtracting deep space from targetinterferograms before performing the Fourier transformremoves the effect of the instrument background, includingphase,fromthetargetdata[27,32].Thisapproachpresumesthat the instrument phase, particularly the ZPD position, isthe same for It, Idsand Is. In those cases inwhich this is nottrue we introduce a linear phase shift to the interferogrambeforesubtraction.Becausetheirresponsesaredifferent,interferogramsfrom

each of the twenty mid-infrared and one far-infrareddetectors must be calibrated separately. Deep space andshutter reference data are recorded for every detector.Because the shapes of the interferograms vary slightlywithscan length, reference data are recorded for each spectralresolution. In addition, reference data are taken for everydetector configuration, evens, odds, pairs and centers, andalso in coadd mode. Before recorded interferograms arecalibrated theyaresubjected toarigorousqualitycontrol inwhich badly distorted interferograms are rejected, andcorrectionsareapplied toeliminate interference in the formofnoisespikesorspuriousmodulation.Adatabaseoffilteredinterferograms is created that serves as input to thecalibration pipeline. In the calibration pipeline calibratedspectra St(ν) are calculated for each target spectrum. Bothapodizedandunapodizedspectraarecreated.ApodizationisappliedwithaHammingwindowfunction.Theappearanceofline structure in the 615-650 cm-1 band of methylacetylene(CH3CCH) in the unapodized spectrum in Fig. 19demonstrates the value of archiving both apodized andunapodized calibrated spectra. The most commonly usedunapodized resolutions available from CIRS are 0.3, 0.5, 1.5and8cm-1.The first step in the calibration process is to create

averages of reference interferograms, i.e., deep space andshutter-closed. We have two approaches for creating thesereference averages, depending on whether we wish tooptimizetheabsoluteaccuracyofthespectraormaximizethesensitivity toweak spectral features. The goals of these twoapproaches are not always compatible. For best absoluteaccuracy the deep space and shutter-closed interferogramsshould come from the same observation sequence as thetarget interferograms, or from time periods immediatelybefore or after the target observation [33]. This minimizeschangesintheinstrumentbetweentargetandreferencedata,principally in temperature and zero-path difference. ThedatasetcommonlydeliveredtoNASA’sPlanetaryDataSystem(PDS) during most of the mission is calibrated with theseadjacent reference averages. We have called it the “localcalibration”.Whensensitivity is thegoal,however, thereareoften not enough reference interferograms available fromtimesnearthetargetobservation.Thenumberofdeepspace

andshutter-closedspectramustbelargeenoughsothattheirnoise contributions are smaller than the noise in the targetspectra.Overall,CIRScollectsenoughdeepspaceandshutter-closeddatatosatisfy thisrequirement.Theresult is that thesignal-to-noiseintheaveragedspectrumisimproved,sothatweak spectral features can more readily be detected.However, the absolute accuracy of the calibrated radianceoftenworsensasthetimebetweentargetandreferencescansincreases.Whenthereferenceinterferogramsarenot

Fig.19.AportionoftheTitanspectrumshowingtheeffectofapodization. The two spectra are from the same averagedhigh resolution interferogram. Without apodization (blue,offset in radiance) the spectral resolution is 0.3 cm-1, whileafterapplyingHammingapodizationtheresolutionis0.5cm-

1. The most obvious effect is in the 615-650 cm-1 band ofCH3CCH where the line structure is resolved in theunapodized spectrum. Other molecular emissions seen hereare fromC4H2at628cm-1,HC3Nat663cm-1andCO2at667cm-1.recordedunder thesame instrumentconditions(mostlydueto temperature drifts) phase errors can arise. Ordinarily, ifthephasesofthetargetandreferenceinterferogramsarethesame the differences will cancel. However, when we findmismatches between target and reference phaseswe adjustthe phases before subtracting interferograms. Since theerrorsmostoftenmanifest themselvesas shifts in zero-pathdifference, the phase correction is linear. Shifts in zero-pathdifference are predominantly less than 2 microns. Wemaintain two types of databases calibratedwith these largereference averages (in addition to the local calibrationdatabase).Oneofthese,called“globalcalibration,”usesdeepspaceandshutter-closeddatafromperiodsseparateddaysorweeks from the target observations. The number of deepspace scans is typically several thousand and these arecorrected for differences in phase and instrumenttemperature between target and reference times. The other,called“grandaveragecalibration,”usesas largeanumberofdeepspaceandshutter-closeddataaspossibleoverthewholemission (typicallymore than30,000deep space and severalthousand shutter-closed). Investigators can choose thecalibration dataset that is optimum for their science goals.Theglobalcalibrationisbestformostapplicationsandwillbearchived in the PDS at the end of themission. But in caseswheretheuserwishestohavethesamereferencespectrafor

0.E+00

1.E-07

2.E-07

3.E-07

4.E-07

5.E-07

6.E-07

7.E-07

8.E-07

600 610 620 630 640 650 660 670

Radiance(W/cm

2/str/cm

-1

Wavenumber(cm-1)

unapodized

apodized

15

averages of target data taken over a long time period thegrandaveragecalibrationmaybepreferred.8.IN-FLIGHTOPERATIONSANDTROUBLESHOOTINGUnderCassini’sdistributedapproach tooperations, theCIRSscience team is responsible for all aspects of instrumentoperations.CIRSoperationsbeginwiththescienceobjectives,proceed through observation planning, implementation andinstrument commanding, and end with calibration andarchivingofthedata.Theinstrumenthasexecutedmorethan1.4 million commands since April 2001.Aside from acommanderrorinOctober2004,nocommandproblemshaveoccurred in themission.Maintaining the quality of the datawhilemanaging the large rate and volume of data has beenchallenging. To date 166 million raw CIRSinterferogramshave been collected: 11 million on Titan, 29million on Saturn, 39 million on the rings, 7 million on icysatellites,and80milliondeepspaceandshutter.98%of theobservational data have been successfully calibrated.TheCIRS investigation provides a full database, both rawinterferograms and calibrated spectra, to the scientificcommunity via the PDS. We continue to make smallimprovementsandmodificationstoourgroundsystem.Flightsoftware updates were most recently uplinked to thespacecraftinJune2014.MorechangesareinstoreforthelastphaseofthemissionwhenproximalorbitswillexposeCassiniand CIRS to new thermal environment near Saturn and theringsnearperiapsis.Collectionof deep space calibrationdata in allmodes and

focal plane temperatures are performed when CIRS is notpointedatatarget,mostoftenduringdownlinksofdatafromthe spacecraft to Earth. During downlink the spacecraft isoftenrollingandintheseperiodswemustavoiddisturbancescaused by the spacecraft reaction wheels. Changes ininstrumenttemperaturealsocomplicatetheuseofdeepspacedataforcalibration.ElectricaldisturbancesintheinstrumentandspacecraftpresentotherlimitstothequalityoftheCIRSdata. Herewe describe themost important types of troubleandhowwedealwiththem.

A.BIUanomalies.Occasional flaws in communication between the spacecraftbusandCIRSresult in lossof interactionwiththespacecraftand shutdown of the instrument. During these periodsobservationandhousekeepingdatamaybelost.Eightofthesebus interface unit (BIU) upsets have occurred, with the lasteventinApril2014.TheBIUerrorsarebelievedtobecausedbyfaultymessagetiming.Changestotheflightsoftware,morefrequent checksof theBIU andCIRSmemory, in addition toautonomous diagnoses and memory corrections havereducedtherateofoccurrenceoftheseshutdowns.

B.Laserwavelengthchanges.About fifteen timesduring themission,whenCIRShasbeenturnedbackonfollowingatransitionoftheinstrumenttooffor pre-sleep states, the reference laser has changedwavelength. This has generally occurred following 1) a BIUanomaly, 2) an instrument flight software load or 3) aspacecraft safing event. Laser changes appear in Fig. 20 assudden,largejumpsordropsinthereferenceinterferometer

signal [33]. Changes are infrequent, with the laser usuallyremaining stable for a year or more, but some events havebeenseparatedbyonlyweeksormonths.Achangeisusuallyaccompaniedbyarecoveryperiodofseveralweeks.Overthattime the laser is predominantly in a single laserwavelengthmode,butexhibitssomeringingduetomixingwithaweakeradjacent mode. During the recovery the weaker modegradually dies away. As the laser recovers, the principalwavelengthdoesnotchangeenoughtoaffectthecalibration.For calibration, thenewwavelengthmust be determinedbymeasuring the positions of known molecular lines. We useemission lines of methane near 1330 cm-1 that appear inspectraofTitan.Theerrorinthisdeterminationisestimatedtobe less than0.01 cm-1 at1300cm-1.Over themission thelaserhasoperatedatwavelengthsof784.2377+Nx0.1534nm,whereN=0,1,2,3,4and5.Thusthelaseroperatedinoneofaseries of modes spaced by 0.15 nm. Changes tended to betowardshorterwavelengthsas themissionprogressed(N=5occurred in 2006 andN=0 occurred in 2015). This possiblyindicates a long-term trend in laser behavior. Also, a long-termdecreaseinlasersignalofabout8%from2004to2016can be seen in Fig. 20 whichmay have been due to opticalchanges not related to the laser operation. After correctingthelaserwavelengthitisenteredintothecalibrationpipelineand used until the next laser shift. Once the new laserwavelength has been determined it can be appliedretroactively and therefore very little data has been lostduringthemissionbecauseoflasershifts.

Fig. 20. Signal level from the laser detector in the CIRSreference interferometer during the Saturn tour. The lasersignal was typically stable for periods of years, butoccasionallychangedwhenrecoveringfromanoffstateoftheinstrument. Abrupt changes in signal corresponded to laserwavelength changes. The signal voltages were measured 3secondsafterstartofscan.C.Noisespikes.During the first instrument check-out after launch wediscoveredapatternofnoise thathadnotbeen seenduringgroundtesting.Onthegroundthisnoisehadbeenmaskedbyambientnoise.Themaincomponentsof thepatternare tworegular series of spikes in the interferograms, one seriesspacedby8Hzandtheotherspacedby0.5Hz.Theseareseen

16

in FP1 and FP3 but not in FP4. The 8 Hz spikes aresynchronized with the spacecraft clock and correspond toreal-timeinterruptsfromtheBIUortooperationsintheCIRSdata computer that occur at the clock rate. Interferogramscanningisalsolockedtothespacecraftclockandisthereforesynchronizedwith the 8Hz spikes. The 0.5Hz spikes occurwhenCIRSsendsdatatothespacecraft.Wehadsomecontrolover data packetization frequency and found that choosing0.5 Hz simplified the spike removal process. When theinterferograms are transformed into spectra, 8 Hz producesspikesatmultiplesof191cm-1and0.5Hzproducesspikesatmultiples of 12 cm-1. During calibration a deep spacereference interferogram is subtracted from each targetinterferogramand if thespikepatternsarematched inboth,the spikes cancel. Therehavebeen twoother types of noisespikes.1)InMay2007,followingaflightsoftwarechangethatregularized some timing effects, a new 1 Hz spike patternappeared in the interferograms. We eliminated the 1 Hz inJuly2008byreintroducingrandomizationinthetiming.2)InFP3therehasalwaysbeenpresentaspikepatternat8.33Hz.Although this patternwould seem to be related to the 8Hztiming,weneverlearneditssourceandhavenotbeenabletoeliminate it. In general, spike suppression has been mucheasier inFP1than inFP3; inFP3the8and8.33Hzpatternsare mixed and the individual spikes are distorted by theresponsetimeofthedetectors.In Fig. 21 FP1 interferograms are presented from three

periodsduring the Saturnmission. Each interferogram is anaverage of 5000 interferograms recorded on Saturn. Eachinterferogramisshownexpandedsothatthe8Hzand0.5Hzpatterns can be clearly seen. For each period a rawinterferogram and a noise-filtered interferogram are shown.Differencesamongthethreeperiodsareapparent.Inthefirstperiod, 2004-05, the scan times were multiples of twoseconds, so that the positions of the 0.5 Hz spikes repeatedfromscantoscan.Aspikecorrectionalgorithmfitamodeledspike comb to the 8 Hz and 0.5 Hz patterns and subtractedthem from individual interferograms. After 2006 the 0.5 Hzspikeswerefurthersuppressedbychoosingscanlengthsthatshifted the spike positions by 1/16 of their spacing. Inaveragesofmultiplesof16interferogramsthe0.5Hzpatternis converted to an 8 Hz pattern. Large averages of spectrathenhaveastrongerspikeat191cm-1anditsharmonicsbutsuppressed spikes at harmonics of 12 cm-1. In July 2010weintroduced anewmodification to thedata samplingprocessthat randomizes some of the 8 Hz and thereby reduces thespikes in single spectra.We also beganoffsetting the0.5Hzand remaining 8 Hz patterns by 5 msec in successiveinterferograms, so that the spikes are eliminated in largeaverages spectra. In ground calibration we continue tosubtractsyntheticspike“combs”of8Hzand0.5Hzpatternsfromeachinterferogramtominimizeresidualspikes[34,35].As can be seen in Fig. 21 these measures have been highlysuccessful in cleaning up interferograms. A second type ofnoise interference has the form of an almost pure sinewavethatdriftsinfrequencyandphaseintheinterferograms.Thesinewave frequency is strongly correlated with thetemperature of the electronics assembly. The cause issuspectedtobetemperaturedriftsofDC-DCpowerconverterfrequency,but thishasnotbeenconfirmed.Thesinewave inan interferogram is transformed into a single spike in the

corresponding spectrum. As the electronics temperaturevaried from 22 °C to 17 °C between Jupiter and Saturn thesinewavemoved from around20 cm-1 in the FP1 spectra tonear150cm-1(aweakharmonichasalsobeenfoundinFP3).Eventhoughthespikeisnarrowinasinglespectrum,itdriftsovermany scans and broadens in large averages of spectra.Fortunately, its changing locationallows selectionof spectrawhere a given region is clear of the sinewave spike. Also,when large averages of spectra are used the drifting phasetends to reduce the sinewave spike. We have developed aprocess in our noise filter that suppresses the sinewave byfitting and subtracting a synthetic sinewave to each FP1interferogram[34](Carlsonetal.2009).Thisgenerally

Fig. 21. Averages of raw and filtered interferograms fromFP1. Each interferogram is an average of 5000 scans onSaturn.In2004-05both8Hzand0.5Hzspikeswereremovedbysubtractingsyntheticcombs.Achangewasmade in2006toaverageoutthe0.5Hzspikes,leavingonlythe8Hzspikesinthe2008-09rawdata.In2010afurtherchangewasmadethat randomizes and shifts the spike patterns, suppressingtheminindividualtransformedspectraandeliminatingthemfrom averages. The filter also removes a baseline from eachinterferogrambysubtractinga synthetic shape that includesthe artifact near sample 2200. Burst patterns are due toemissionlinesinSaturn’sspectrum.

0.5Hz

8Hz

2004-05

2011-12

2008-09

raw

filtered

Samplenumber

Signal(arbitraryunits)

8Hz

17

reduces the amplitude of the sinewave spike to below thenoise level. Noise spikes and sinewave interference arediscussedmoreextensivelybyChanetal.[36].D.Reactionwheelinterference.

After release of the Huygens Probe in December 2004 werealized that during some periods the recordedinterferogramshadmoresamplesthannormal.Controlofthescanmechanismwasoften lostduring these interferograms,making thedatauseless.Theproblemwas found tobemostcommonduringdatadownlinkstoEarth,whenthespacecraftwas rolling. The causewas tracked to interference from thespacecraft reaction wheel assemblies (RWA’s). The reactionwheelratesvaryconsiderablyduringrolling,overtherange0to±1600rpm[37,38].RWA1,2and3wereusedatbeginningofmission, but because of excess drag RWA3was swappedwithRWA4inJuly2003andhasbeenusedonlyoccasionallysince then during checkouts. Fig. 22 shows the disturbancedue to eachof the three reactionwheels in eachyearof theSaturntour.Disturbedscanswereidentifiedbyextrasamplesin low resolution (15 cm-1) interferograms beyond thenominal, undisturbed scan length (we used low resolutioninterferograms because that dataset contained the mostscans). Wheel noise causes spurious zero crossings in thereference interferometer, which triggers extra sampling.Besides longer scans, the symptoms included loss of scanmechanism control, premature scan flyback and irregularspacingsofthe0.5Hznoisespikes.Comparingthenumberofbad interferograms with the wheel rates we found that thedisturbances occurred when the wheel rates were above1300 rpm.Each reactionwheelproduces adifferentpatternof disturbance, but the greatest disturbances in all wheelstended to be between 1300 and 1460 rpm. Of the threewheels, theone thathad the greatestdetrimental effectwasRWA2[38].All indicationswerethatthereactionwheelswerecausing

vibrations in the scan mechanism. The CIRS leaf springassemblies on the scanmechanismwere known to resonateabove20Hz, close to the1300 rpm frequency.Althoughwewere able to find some disturbances before 2005, thevibrations became stronger after the release of the probe.(Whilewewerediagnosingtheinterferencefromthereactionwheels,afailureoccurredintheturntableofLEMMS,theLowEnergyMagnetosphericMeasurements System. Operation oftheturntableproducedadditionalvibrationinthescans.TheLEMMSturntablefailedonFebruary1,2005andwasturnedoffthreeweekslater.)Asfuelhasbeenconsumedduringthemission the resonant frequencies have increased by anadditional amount, particularly in RWA 4 by about 10 Hzbetween 2005 and 2016. Additional changes in 2004 fromSaturnorbitinsertionarenotshowninthefigure.Weaddress theproblemof interference fromthereaction

wheels by working with the Cassini project to limit wheelrates during CIRS observations. Careful management of therateshasprovensuccessfulinminimizingthelossofscientificdata.A reactionwheelbiasoptimization toolwasdevelopedat Jet Propulsion Laboratory to provide predictions of thereaction wheel rates, originally to help avoid high-wearrotationspeeds.ThistoolhasproventobecrucialinplanningCIRSoperationsforkeepingtheratesbelow1300rpm.Thereisstilldisturbanceduringrollingdownlinkswhenwecollect

someofourdeepspacereferencedata,butwhenfeasiblewehaveaddedothernon-rollingperiodstopermittherecordingofdeepspacedata.E.Thermalenvironment.

Our main purpose for maintaining thermal stability in theinstrument is to minimize loss of data. Fluctuations intemperature over the duration of an observation causecalibrationerrors[39].Wetrytoavoidpointingthe170Kor80 K radiator surfaces toward the sun to prevent excessivewarming. Even exposing the radiators to a relatively warmplanetcandrivetheinstrumentoutofthermalcontrol,a

Fig.22.Disturbanceofinterferogramsasafunctionofwheelspin rpm for each of the three spacecraft reaction wheels(RWA’s). The ordinate is the year and the abscissa is thewheelrate.Positive(red)andnegative(black)spindirectionsareshownoverlapped.Mostseveredisturbancesoccurredinthe1300-1460rpmrange,predominantlyfromRWA2.Thesedata were derived from low resolution (15 cm-1)interferograms.concernduringSaturnorbitinsertionandtheend-of-missionclose-inorbits.Thisisaparticularconcernforthe80Kcoolerwheretemperaturerisescanbefastandrecoverytimeslong.ThegoalistomaintaintheCIRSmid-infraredfocalplaneataslow a temperature as possible to ensure optimum detectorsensitivity. We typically operate at 76.3 K, but in cases ofextra thermal loading higher temperature setpoints are

18

available (seenextsection).Amodelwasdevelopedearly inthe mission to be used with uplink sequence developmenttoolsthatpredictsthetemperatureexcursionsinCIRSduringplannedspacecraftmaneuvers.Basedonthemodel,thefocalplane setpoint is adjusted and heaters are commanded inanticipation of temperature swings. Predictions havegenerally been within 1 K of actual temperatures duringobservations and have been sufficient to avoid thermalproblems. Themodel has been used successfully during theSaturntourand improvementscontinuetobemade, the lastbeingimplementedin2014.

9.IN-FLIGHTPERFORMANCEDuring flight CIRS has achieved and exceeded the sciencegoals of the investigation. There had been concern duringground testing that the radiance sensitivity was lower thanrequired,andsubstantialeffortwasdevotedtocharacterizingthe noise and throughput of both the engineering and flightunits. To our relief, in flight the sensitivity across thespectrum improvedbymore than a factor of twooverwhathad been seen during ground testing. This was primarilybecause the acoustic noise in the thermal-vacuum chamberenvironmenthadpreventedusfrommeasuringthetruenoisefloor.Fig.4showsexamplesof interferogramsfromallthreefocal planes recorded at Saturn. Responses in the mid-infrared depend on the focal plane temperature. DetectorresponseisdeterminedaspartoftheCIRScalibrationprocessusing observations of deep space and shutter. Underminimum heat loading the mid-infrared detectortemperatures can be as low as 75.6 K. We can usuallymaintain the detector temperature at 76.3±0.1 K, but whenheatfromthesunorSaturndrivesthefocalplanetoahighertemperature we can control at temperatures up to 87 K byselectingoneofanumberofdiscretesetpoints.Fig.23showsFP3responsesforasingledetectoratseveraltemperaturesin2002. The response in FP3 decreases with increasedtemperatureovertheentirebandwidthfrom580to1100cm-

1.At700cm-1theresponsedecreasesbyabout30%between75.6and83.6K.Fig.23alsoshowsthesamecomparisonforFP4.InFP4theresponsesdonotchangeexceptatthelowestwavenumbers.At1100cm-1 the responsedecreasesby25%between75.6 and 83.6K. Sensitivity is characterized by thenoise-equivalent spectral radiance (NESR), which gives thesource radiance equal to the one-sigma noise level as afunction of wavenumber [2, 10]. The temperaturedependence of FP3 NESRs in Fig. 23 is dominated by thedecrease in responsewithhigher temperature.On theotherhand, theFP4NESRsaredominatedbythe increase innoise(the response is unchanged with temperature). Thus,althoughfordifferentreasons,bothmid-infraredfocalplanesdegrade in signal-to-noise when the temperature is raised.When FP3 was raised above 87 K it gradually degraded insensitivity, butwhenFP4was raised above87K it abruptlystopped working. Both focal planes resumed normaloperationwhenthetemperaturewasdecreased.The long duration of the Cassini mission provides a uniqueopportunity to investigate the stability of CIRS’ sensitivityover time. Comparing 2002 with 2015 in Fig. 24 at aparticular focalplane temperature shows that the responsesofFP3andFP4didnotchangemeasurablyduringthethirteenyearperiod[27,39].Weestimatethatchangeswerenotmore

than 1% over this time. Noise levels in the 2002 and 2015spectra are comparable because the numbers of deep spaceand shutter scans used in each were similar. Our overallimpressionisthatnochangeinanyofthedetectorsorsignalchainshasbeenseensincebeforelaunch.Fortheselong-termcomparisons the internal temperature, which is used as theabsolute reference,must be constant throughout theperiod.InCIRS the170Kand80Kportionsof the instrumenthavebeen controlled to within 0.1 K during observations. Ourmeasurementsassume that the temperature sensors inCIRShavenotdriftedandthatemissivitiesoftheinternalsurfaceshave not changed. We believe that the sensors have beenstable because we have seen no long-term change greaterthan0.03Kincomparingreadingsofthecontrolandmonitorsensors.Changesinsurfaceemissivityshouldhavelittleeffectbecausetheinternalstructuresthatareviewedare

Fig. 23. Detector responses (in signal counts perW/cm2/str/cm-1) and noise-equivalent spectral radiances(NESRs in W/cm2/str/cm-1) for FP3 detector 6 and FP4detector 5. Responses are at five focal plane temperaturesbetween 75.6 and 83.6 K derived from data recorded inSeptember2002.Recordedspectralresolutionwas3cm-1andthe responses were smoothed to 15 cm-1. For the FP3responses 99 deep space and 9 closed-shutter scans wereused,whileforFP4thenumberswere100deepspaceand10closed-shutterscans.NESRsarecalculatedfromthestandarddeviationoftheresponses.ThenoisespikesintheFP3NESRsin 2002 were from harmonics of 8 Hz interference. Thesewere suppressed after 2010 by a change in flight software.TheNESRworsensmarkedlywhen the detector is operatedabove80K.

FP3 detector 6

Wavenumber (cm-1)

Wavenumber (cm-1)

Re

sp

on

se

(x1

01

0)

NE

SR

(x1

0-8

)

83.62 K

79.96 K

77.76 K

76.30 K

75.56 K

83.62 K

79.96 K

77.76 K

76.30 K

75.56 K

Re

sp

on

se

(x1

01

0)

NE

SR

(x1

0-9

)

FP4 detector 5

19

Fig. 24. Comparison of the responses (in signal counts perW/cm2/str/cm-1)between2002and2015forFP3detector6andFP4detector5. Spectral resolutionwas3 cm-1 Thedatawere selected to correspond closely to the same focal planetemperature,75.6K.TheFP3andFP4responsesin2002arethesameasinFig.23.In2015forFP3,101deepspaceand30closed-shutter were used, while for FP4 the numbers usedwere54deepspaceand16closed-shutter.TheresponsedidnotchangemeasureablyovertheperiodoftheSaturntour.surrounded by surfaces at the same temperature and sobehaveasnearlyidealblackbodies.CIRSusesDC/DCpowerconvertersthatmightbeexpected

to failafter longexposure toneutrons.Testingofconverterscontainingopticalcouplers[40]showedthattheneutronfluxfrom the radioisotope thermoelectric generators (RTGs) onCassiniwould eventually, after about15 years, cause loss ofpower in the instrument. The RTGs were installed on thespacecraftadjacenttoCIRSin1996,fromwhichfailurewouldbeexpectedtooccuraround2012.Fortunately,asof2017,wehad not seen any significant degradation in performance.Among the various CIRS voltages the largest changewas 10millivoltsina15voltsupply.Thereasonforthelonger-than-predictedlifeofthepowerconvertersisnotknown,butcouldbeduetoextrashieldingoftheCIRSelectronicsbytheopticsassembly.10.EXAMPLESOFSPECTRAThequalityofCIRSdataisbestdemonstratedwithexamples.Figures 25-27 show composite spectra from Jupiter, Saturn,and Titan recorded at 0.5 cm-1 resolution. Each of thesespectraisacombinationofallthreefocalplanes,covering10-1500cm-1.Thesignalisexpressedasbrightnesstemperature,

which is more convenient than radiance for displayingintensityover the full rangeofCIRS.Brightness temperatureindicatesthetemperatureoftheatmosphereorsurfacewherethe thermal emission originates. In the Titan spectrum, forexample(Fig.27),emissionat530cm-1comesfromthe~94Ksurface,whileat30cm-1weseethe~65Ktropopauseandat1100 cm-1 we are detecting the ~110 K stratosphere.Differencesinatmosphericcompositionandtemperatureareclearlydemonstratedbycomparingthesespectra.JupiterandSaturn have hydrogen (H2) atmospheres, while Titan has anitrogen(N2)atmospherewithverylittlehydrogen.Allthreeatmospherescontain thebasichydrocarbonsmethane(CH4),acetylene (C2H2), ethane (C2H6), and ethylene (C2H4).However, ammonia (NH3) is more visible in Jupiter than inSaturn, and Saturn has phosphine (PH3) which does notappear prominently in Jupiter. Neither ammonia norphosphine are found in the Titan spectrum. Nitriles areabundant in Titan’s atmosphere, as is apparent from thestrong bands of hydrogen cyanide (HCN), cyanoacetylene(HC3N) and cyanogen (C2N2). Table 2 lists the molecularbandsandotheremissionfeaturesthatCIRShasobservedinJupiter,SaturnandTitan.CIRSobservationsofobjectswithoutatmospheres,suchas

Saturn’s other moons, its rings and the moons of Jupiter,typically measure thermal emission from surfaces. Thesemeasurements are normally performed at reduced spectralresolution,1,3or15cm-1,becausenarrowgaseous featuresarenotexpected.Asanexample,Spilkeretal. [41] reportedradialmappingbyCIRSofthermalemissionfromtheringsasa function of solar phase angle, revealing that the ringparticlesundergo slow rotation. Similarly,Howettetal. [42]measured thermal inertia and bolometric Bond albedos onthe moons Mimas, Enceladus, Tethys, Dione, Rhea andIapetus. Perhaps the most important and unexpected resultfrom an icy surface was the CIRS discovery, reported byHowett et al. [43], of relatively hot locations along thefissures,or “TigerStripes,”atEnceladus’SouthPole thatareassociated with plumes. We have developed a specializedmethodtotracetheseheatsignatures.Thetechniquehas

Fig. 25. Brightness temperature spectrum of Jupitercombining data from all three focal planes (FP1, FP3 andFP4).Spectralresolutionis0.5cm-1.Forthisdatasetemissionangleswerekeptbelow50°.Prominentatmosphericfeaturesareidentifiedbytheirmolecularspecies.

2002 (75.56 K)

2015 (75.62 K)

FP3 detector 6

Re

sp

on

se

(x1

01

0)

2002 (75.56 K)

2015 (75.62 K)

Wavenumber (cm-1)

Wavenumber (cm-1)

Re

sp

on

se

(x1

01

0)

FP4 detector 5

2002 (75.56 K)

2015 (75.62 K)

100

110

120

130

140

150

160

170

180

0 200 400 600 800 1000 1200 1400

BrightnessTemperature(K)

Wavenumber(cm-1)

C2H2

CH4

C2H6

H2

NH3 NH

3

20

Fig. 26. Brightness temperature spectrum of Saturncombining data from all three focal planes (FP1, FP3 andFP4).Spectralresolutionis0.5cm-1.Forthisdatasetemissionangleswerekeptbelow50°.Prominentatmosphericfeaturesareidentifiedbytheirmolecularspecies.

Fig.27.BrightnesstemperaturespectrumofTitancombiningdata fromall three focalplanes(FP1,FP3andFP4).Spectralresolution is 0.5 cm-1. For this dataset latitudes wererestricted to60-90North latitude and emission angleswerekept below 50°. Prominent atmospheric features areidentifiedbytheirmolecularspecies.been described by Spencer et al. [44] andGorius et al. [45].Duringlow-altitudeCassiniflybysofthewarmfissures,asthepath of CIRS’ field of view crosses the features, the thermalemission rises and falls rapidly. The brief increases ofemission cause transientdistortions in the interferogram. IntheseobservationsCIRSisactingasadifferentialradiometerin that it is measuring the change in emission from eachthermal feature. Fig. 28 shows a sweep of FP1 across theactiveSouthPoleregionduringthecloseflyoverofEnceladusonApril14,2012.Cassinicameascloseto thesurfaceas75kmataspeedof7.5km/sec.Transientsignalscanbeseeninthe interferograms as FP1 crossed onto the dark side ofEnceladus, passed over numerous warm features on themoon’s surface including Baghdad Sulcus (the strongesttransient), and moved off the bright limb of Enceladus. Ateachsurfacefeaturethemagnitudeofthetransientprovidesameasure of temperature change, while the temporal width

indicates spatial extent. CIRS was operating in its normalscanning manner during these observations. We used longscans (corresponding to 0.5 cm-1 resolution) because theyprovided long, uninterrupted timeperiods between carriageflybacks.ThreeinterferogramscomposethesweepinFig.28and two of the ZPD bursts occurredwhile the field of viewwas on themoon. FP3 and FP4 produce similar results andbecause of their much smaller detector sizes they providebetterspatialresolution,assmallas25m.

Fig.28.CIRSusedasadifferentialradiometerduringaclosepassofEnceladus.FP1 longscans(50sec)wererecordedasCassini flew over the active southern region on April 14,2012.Threeconsecutiveinterferogramsareshownherewithtwogapswhere the carriage returned to start of scan.Deepspace reference interferograms have been subtracted fromeach of these. The ingress and egress limb points are seenwheretheradiancechangedabruptly.AstheFP1fieldofviewpassed over narrow, hot features in the vicinity of BaghdadSulcus the sudden changes in radiance created spikes in theinterferogram. Information about the brightnesses andshapes of the emission sources can be derived from thesedata.(NASAphoto.)

11.CONCLUSIONCassini is ending itsmission on September 17, 2017with adive into Saturn’s atmosphere. During the final “proximal”orbits,beginninginDecember2016,thespacecraftwillcomenearer toSaturn thanat anyprevious time.CIRSwill collectdata during these close-ups. The rapidly changing thermalenvironmentduringthesepassespresentsachallengetoCIRSoperations, but we anticipate exciting and unique scienceresults. During the final phase of themission CIRS plans tooperatevariousbackupcomponentsthatwereneverorrarelyactivated during the mission, including the redundant laserand theparalleldetector circuit inFP1.FP3andFP4willbesubjected to high temperatures. Information from end-of-missiontestswillbevaluableasverificationofoperationafter20yearsinspace.Wefacethefinalactwithsadness,butalsowithgratificationattheamazingsuccessofCIRSandCassini.We anticipate that CIRS data will produce new scientificdiscoveriesfordecadestocome.

80

90

100

110

120

130

140

150

160

0 200 400 600 800 1000 1200 1400

C2H2

CH4

C2H6

C2H6

H2

PH3

BrightnessTemperature(K)

Wavenumber(cm-1)

PH3

60

80

100

120

140

160

180

0 200 400 600 800 1000 1200 1400

BrightnessTemperature(K)

Wavenumber(cm-1)

C2H2

C3H4

C4H2 C

2H4

CH3D

CH4

C2H6

C2H6

CO2

C3H4

H2

CH4

HCN

CO

N2

HCN

C4H2

C2N2

HC3N

C3H8

H2

HC3N

C3H6

ingress egressfissure

Enceladusclose-flyby

April14,2012

zpd zpd

Interferogramsignal

21

Table2.MoleculesObservedbyCIRS

Species Formula ν (cm-1) Sources

†[Ref.]

Hydrogen H2 354,587 JST

HD 88,178,265 JS

Nitrogen N2 50-100 T

CarbonMonoxide CO 20-80 T

HydrogenCyanide HCN 30-80,712 JT

H13CN 706 T*[46]

HC15N 711 T*[46]

CarbonDioxide CO2 667 JST

13CO2 649 T*[47]

CO18O 663 T*[47]

Water H2O 140-260 ST

Acetylene C2H2 729 JST

13CCH2 728 JST*[48]

C2HD 519,678 T*[49]

Cyanogen C2N2 234 T

Ammonia NH3 20-300,800-1200 JS

NH3 1060(solid) J*[50]

Phosphine PH3 20-250,900-1200 S

MethylRadical CH3 606 J*[51]

Methane CH4 50-150,1306 JST

13CH4

1297 JST*[48]

CH3D 1156 JT

13CH3D 1148 T*[52]

Ethylene C2H4 949 JST

Ethane C2H6 288,790-850 JST

13CCH6 800-840 JT*[48]

Propene C3H6 912 T*[53]

Diacetylene C4H2 628 JST

13CC3H2 622,627 T*[54]

C13CC2H2 628 T*[54]

Dicyanoacetylene C4N2 477(solid) T

Methylacetylene C3H4 325,628-637 JST

Cyanoacetylene HC3N 499,663 T

HC3N 506(solid) T

H13CC2N 659 T*[55]

HC13CCN 663 T*[55]

HC213CN 663 T*[55]

Propane C3H8 749 ST

Benzene C6H6 674 JST

C6H6 680(solid) T*[56]†J-JupiterS-Saturn T–Titan

*MeasuredbyCIRSforthefirsttime.

Funding. National Aeronautics and Space Administration(NASA), Cassini Mission, CIRS investigation; CNES (CentreNational d’Etudes Spatiales; CNRS (Centre National de laRecherche Scientifique, France); SERC (Science andEngineeringResearchCouncil,UnitedKingdom).

Acknowledgement. The authors wish to thank the manypeoplewhocontributedtothedevelopmentandoperationofCIRS.We acknowledge the support of Goddard Space FlightCenter, NASA and the Cassini Project at Jet PropulsionLaboratory.References1.V.G.Kunde,P.Ade,R.Barney,D.Bergman,J.F.Bonnal,R.Borelli,

D.Boyd,J.Brasunas,G.Brown,S.Calcutt,R.Courtin,J.Cretolle,J.

Crooke, M. Davis, S. Edberg, R. Fettig, M. Flasar, D. Glenar, S.

Graham, J. Hagopian, C. Hakun, P. Hayes, L. Herath, L. Horn, D.

Jennings,G.Karpati,C.Kellebenz,B.Lakew, J. Lindsey, J. Lohr, J.

Lyons,R.Martineau,A.Martino,M.Matsumura,J.McCloskey,T.

Melak, G. Michel, A. Morell, C. Mosier, L. Pack, M. Plants, D.

Robinson, L. Rodriguez, P. Romani, W. Schaefer, S. Schmidt, C.

Trujillo, T. Vellacott, K. Wagner, and D. Yun, “Cassini infrared

Fourier spectroscopic investigation,” Proc. SPIE 2803, 162–177

(1996).

2.F.M.Flasar,V.G.Kunde,M.M.Abbas,R.K.Achterberg,P.Ade,A.

Barucci, B. Bézard, G. L. Bjoraker, J. C. Brasunas, S. Calcutt, R.

Carlson, C. J. Cesarsky, B. J. Conrath, A. Coradini, R. Courtin, A.

Coustenis, S. Edberg, S. Edgington, C. Ferrari, T. Fouchet, D.

Gautier, P. J. Gierasch, K. Grossman, P. Irwin, D. E. Jennings, E.

Lellouch,A.A.Mamoutkine,A.Marten,J.P.Meyer,C.A.Nixon,G.

S.Orton,T.C.Owen,J.C.Pearl,R.Prange,F.Raulin,P.L.Read,P.

N.Romani,R.E.Samuelson,M.E.Segura,M.R.Showalter,A.A.

Simon-Miller,M. D. Smith, J. R. Spencer, L. J. Spilker, and F.W.

Taylor,“Exploring theSaturnsystem in the thermal infrared: the

composite infrared spectrometer,” Space Sci. Rev.115, 169–297

(2004).

3. D. E. Jennings, V. G. Kunde, F. M. Flasar and CIRS Team, “The

Composite Infrared Spectrometer onCassini: 15 years in Flight,”

International Workshop on Instrumentation for Planetary

Missions,GreenbeltMaryland,p.1045,10-12October2012.

4.R.A.Hanel,B.Schlachman,F.D.Clark,C.H.Prokesh,J.B.Taylor,

W. M. Wilson, and L. Chaney, “The Nimbus 3 Michelson

Interferometer,”AppliedOptics9,1767-1774(1970).

5.R.A.Hanel,B.Schlachman,D.RogersandD.Vanous,“Nimbus4

MichelsonInterferometer,”Appl.Opt10,1376-1382(1971).

6. R. A. Hanel, R. A., B. Schlachman, E. Breihan, R. Bywaters, F.

Chapman, M. Rhodes, D. Rodgers, and D. Vanous, “Mariner 9

MichelsonInterferometer,”AppliedOptics11,2626-2634(1972).

7.R.A.Hanel, R.A.,D.Crosby, L.Herath,D.Vanous,D.Collins,H.

Creswick, C. Harris, and M. Rhodes, Applied Optics, “Infrared

spectrometerforVoyager,”AppliedOptics19,1391-1400(1980).

8.D. H. Martin, “Polarizing (Martin-Puplett) interfero-metric

spectrometers for the near- and submillimeter spectra,” in

Infrared and Millimeter Waves, Systems and Components, K.J.

Button(ed.),Vol.6,(AcademicPress,1982),Ch.2.

9.D. H. Martin and E. Puplett, “Polarised interferometric

spectrometry for the millimetre and submillimetre spectrum,”

InfraredPhys.10,105–109(1969).

10.R.A.Hanel,B.J.Conrath,D.E.Jennings,andR.E.Samuelson, in

Exploration of the Solar System by Infrared Remote Sensing

(CambridgeUniversityPress,1992,2003).

11.M. D. Smith, “CIRS Field-of-View Overview,” CIRS internal

document(2003).

12.P.W.Maymon,M.G.Dittman,B.A.Pasquale,D.E.Jennings,K.I.

Mehalick, C J. Trout, "Optical design of the composite infrared

spectrometer(CIRS)fortheCassinimission,"Proc.SPIE1945,pp.

100-111(1993).

13.R.A.Hanel,B.J.Conrath,D.Gautier,P.Gierasch,S.Kumar,V.G.

Kunde, P. Lowman, W. MaGuire, J. C. Pearl, J. Pirraglia, C.

Ponnamperuma and R. E. Samuelson, “The Voyager infrared

spectroscopy and radiometry investigation,“ Space Science

Reviews21,129-157(1977).

14.P.Hayes,J.CrookeandB.Perkins,“Cryogenicandambienttesting

oftheCIRSberylliumtelescopes,”Proc.SPIE2227,15-25(1994).

15.P. Losch, J. J. Lyons, and J. Hagopian, “Cryogenic optical

performance of the Cassini Composite InfraRed Spectrometer

(CIRS)flighttelescope,”Proc.SPIE3435,19-29(1998).

16. J.B. Heaney, L. R. Kauder, S. C. Freese and M. A. Quijada,

“Preferred mirror coatings for UV, visible, and IR space optical

instruments,”Proc.SPIE8510,F01-F11(2012).

22

17. R. A. Keski-Kuha, C.W. Bowers,M. A. Quijada, J. B. Heaney, B.

Gallagher,A.McKayandI.Stevenson,“JamesWebbSpaceOptical

Telescope Element Mirror Coatings,” Proc. SPIE 8442, 84422J

(2012).

18.J.A.Crooke,andJ.G.Hagopian,“Alignmentandcryogenictesting

of the Cassini-Composite InfraRed Spectrometer (CIRS) far-

infrared(FIR)focalplane,”Proc.SPIE2814,105-114(1996).

19. J. A Crooke and J. G. Hagopian, “Alignment and polarization

sensitivitystudyfortheCassini-CompositeInfraRedSpectrometer

(CIRS) Far InfraRed (FIR) Interferometer,” Proc. SPIE3435, 30-41

(1998).

20. R. Fettig, B. Lakew, J. Brasunas, J. A. Crooke, C. Hakun, and J.

Orloff, “Thermoelectric infrared detectors with improved

mechanical stability for the Composite Infrared Spectrometer

(CIRS)farinfraredfocalplane,”Proc.SPIE3435,126-135(1998).

21.J.G.Hagopian,P.A.LoschandJ.A.Crooke,B.Mott,A.J.Martino,

“Shear/defocus sensitivity of the mid infrared channel (MIR) of

the Composite infrared spectrometer (CIRS) for the Cassini

missiontoSaturn,”Proc.SPIE3435,42-51(1998).

22.C.A.Nixon,N.A.Teanby,S.B.Calcutt,S.Aslam,D.E.Jennings,V.

G.Kunde,F.M.Flasar,P.G.Irwin,F.W.Taylor,D.A.Glenarand

M. D. Smith, “Infrared limb sounding of Titan with the Cassini

Composite InfraRed Spectrometer: effects of the mid-infrared

detectorspatialresponses,”Appl.Opt.48,1912-1925(2009).

23. R. J. Martineau, Kelley Hu, S. Manthripragada, C. A. Kotecki, S.

Babu,F.A.Peters,AS.Burgess,D.B.Mott,D.J.Krebs,S.Graham,

A. J.Ewin,A.Miles,V.Bly,T.L.Nguyen, J.McCloskey,P.K.Shu,

“HgCdTe Detector Technology and Performance for the

Composite Infrared Spectrometer (CIRS)/Cassini Mission,” Proc.

SPIE2803,178-186(1996).

24.S.Calcutt,F.Taylor,P.Ade,V.G.Kunde,andD.E.Jennings,"The

CompositeInfraredSpectrometer,"J.oftheBritishInterplanetary

Society45,811-816(1992).

25. A. J. Martino, and D. M. Cornwell, “Reference interferometer

usinga semiconductor laser/LED referencesource ina cryogenic

Fourier-transformspectrometer,”Proc.SPIE3435,61-71(1998).

26.R. C. Carlson, D. E. Jennings, J. C. Brasunas, and V. G. Kunde,

“Wavelength Calibration of the Cassini Composite Infrared

Spectrometer (CIRS) Using Spectrum De-Stretching,“ in Fourier

TransformSpectroscopy/Hyperspectral ImagingandSoundingof

theEnvironment,OSATechnicalDigestSeries, (OpticalSocietyof

America2005),paperFWD3.

27.J.Brasunas,A.Mamoutkine,andN.Gorius,“Identifyingsampling

combchangesinFouriertransformspectrometerswithsignificant

self-emissionandbeamsplitterabsorption,”Appl.Opt.54,5461–

5468(2015).

28.A. Tyler, "Shutter Mechanism for Calibration of the Cryogenic

Diffused Infrared Background Experiment (DIRBE) Instrument,"

presentedat20thAerospaceMechanismsSymposium(NASA-CP-

2423-Rev),97-102(1986).

29.C.F.Hakun,andK.A.Blumenstock,“ACryogenicScanMechanism

for use in Fourier Transform Spectrometers,” presented at the

29th Aerospace Mechanisms Symposium, NASA Johnson Space

Center,pp.316-333(1May1995).

30.G. Michel, and R. Courtin, “A Prototype Scan Mechanism for

Composite Infrared Spectrometer – CIRS - of the NASA/ESA

Cassini Mission,” Proceedings of the Sixth European Space

Mechanisms & Tribology Symposium, Zurich, 4-6 October 1995

(ESASP-374,August1995).

31.J.J.LyonsandP.A.Hayes,“High-optical-qualitycryogenichollow

retroreflectors,”Proc.SPIE2540,94-100(1995).

32.H.E.Revercomb,H.Buijs,H.B.Howell,D.D.Laporte,W.L.Smith,

and L. A. Sromovsky, “Radiometric calibration of IR Fourier

transform spectrometers: solution to a problem with the high-

resolution interferometer sounder,” Appl. Opt. 27, 3210–3218

(1988).

33. J.Brasunas,A.Mamoutkine,N.Gorius,“Simpleparametricmodel

for intensity calibration of Cassini composite infrared

spectrometerdata,”Appl.Opt.55,4699-4705(2016).

34.R.C.Carlson,E.A.Guandique,D.E.Jennings,S.H.PilorzandV.G.

Kunde, ”Characterization and Suppression of Electrical

Interference-Spikes,PeriodicWaves,andRipples -FromCassini

Composite InfraredSpectrometer (CIRS)Spectra,” inAdvances in

Imaging,OpticalSocietyofAmericaTechnicalDigest,paperFTuA5

(2009).

35.R.C.Carlson,“RemovingArtifactsintheCalibrationofCassiniCIRS

Spectra of Saturn and Titan,” presented at the EPSC-DPS Joint

Meeting,Nantes,France(2-7October2011).

36.C. Chan, S. A. Albright, N. J. P. Gorius, J. C. Brasunas, D. E.

Jennings,F.M.Flasar,R.C.Carlson,E.GuandiqueandC.A.Nixon,

“Electrical Interferences Observed in the Cassini CIRS

Spectrometer,”ExperimentalAstronomy3,367–386(2015).

37.A. Y. Lee, and E. K. Wang, “In-Flight Performance of Cassini

Reaction Wheel Bearing Drag in 1997–2013,” J. Spacecraft &

Rockets52,470-480(2015).

38.T. S. Brown, “The Cassini Reaction Wheels: Drag and Spin-Rate

Trends from an Aging Interplanetary Spacecraft at Saturn,”

presented at the AIAA Guidance, Navigation, and Control

Conference, AIAA SciTech Forum (AIAA 2016-2085), San Diego,

California,USA,(4-8January2016).

39. J.C. Brasunas and B. Lakew, “Long-term stability of the Cassini

Fourier transform spectrometer en route to Saturn,” in Recent

Research Developments in Optics, A. Gayathri, ed. (Research

Signpost,2004),Vol.4,pp.95–113.

40.R.Reed,K.LaBel,H.Kim,H.LeideckerandJ.Lohr,“TestReportof

Proton and Neutron Exposures of Devices that Utilize Optical

Components and are Contained in the CIRS Instrument,” at

https://www.researchgate.net/publication/237349052,(1997).

41.L. J.Spilker,S.H.Pilorz,B.D.Wallis, J.C.Pearl, J.N.Cuzzi,S.M.

Brooks,N.Altobelli,S.G.Edgington,M.Showalter,F.M.Flasar,C.

Ferrari,C. Leyrat, “Cassini thermalobservationsofSaturn’smain

rings: Implications for particle rotation and vertical mixing,”

PlanetaryandSpaceScience54,1167–1176(2006).

42.C.J.A. Howett, J.R. Spencer, J. Pearl and M. Segura, “Thermal

inertiaandbolometricBondalbedovaluesforMimas,Enceladus,

Tethys, Dione, Rhea and Iapetus as derived from Cassini/CIRS

measurements,”Icarus206,573–593(2010).

43.C. J.A.Howett,

J.R.Spencer,

J.Pearl,andM.Segura,“Highheat

flow fromEnceladus’ southpolar regionmeasuredusing10–600

cm-1 Cassini/CIRS data,” J. Geophys. Res. 116, E03003,

doi:10.1029/2010JE003718(2011).

44.J.R.Spencer,N.J.P.Gorius,C.J.A.Howett,D.E.JenningsandS.

A. Albright, “The Spatial Distribution of Thermal Emission from

Baghdad Sulcus, Enceladus, at 100 meter Scales,” American

AstronomicalSocietyDivisionforPlanetarySciences,meeting#44,

id.104.06(2012).

45.N. J. Gorius, J. R. Spencer, C. J. Howett, D. E. Jennings, J. C.

Brasunas, and S. A. Albright, “Observing the South Pole of

EnceladususingCassini/CIRSasaRadiometer,”OpticalSocietyof

AmericaTopicalMeetingonFourierTransformSpectroscopy,Lake

Arrowhead,CA,1–4March(2015),paperFT3-2.

23

46. S. Vinatier, B. Bézard and C. A. Nixon, “The Titan14N/

15N and

12C/

13CisotopicratiosinHCNfromCassini/CIRS,”Icarus191,712–

721(2007).

47.C. A. Nixon, D. E. Jennings, B. Bézard, N. A. Teanby, R. K.

Achterberg,A.Coustenis,S.Vinatier,P.G.J.Irwin,P.N.Romani,

T. Hewagama, and F. M. Flasar, “Isotopic Ratios in Titan’s

Atmosphere from Cassini CIRS Limb Sounding: CO2 at Low and

Midlatitudes,”Astrophys.J.Lett.681,L101-103(2008).

48.C.A. Nixon, R.K. Achterberg, S. Vinatier, B. Bézard, A. Coustenis,

P.G.J. Irwin,N.A. Teanby, R. de Kok, P.N. Romani, D.E. Jennings,

G.L.BjorakerandF.M. Flasar, “The12C/

13C isotopic ratio inTitan

hydrocarbons from Cassini/CIRS infrared spectra,” Icarus 195,

778–791(2008).

49.A. Coustenis, D. E. Jennings, A. Jolly, Y. Bénilan, C. A. Nixon, S.

Vinatier, D. Gautier, G. L. Bjoraker, P. N. Romani, R. C. Carlson,

“DetectionofC2HDandtheD/HratioonTitan,” Icarus197,539-

548(2008).

50.M. H.Wong, G. L. Bjoraker,M. D. Smith, F.M. Flasar and C. A.

Nixon, “Identification of the 10-μm ammonia ice feature on

Jupiter,”Planet.&Sp.Sci.52,385–395(2004).

51.V.G.Kunde,F.M.Flasar,

D.E.Jennings,

B.Bézard,

D.F.Strobel,

B.

J. Conrath,C. A. Nixon,

G. L. Bjoraker,

P. N. Romani,

R. K.

Achterberg,A.A.Simon-Miller,

P.Irwin,

J.C.Brasunas,

J.C.Pearl,

M.D.Smith,G.S.Orton,P.J.Gierasch,

L.J.Spilker,

R.C.Carlson,

A. A. Mamoutkine,S. B. Calcutt,

P. L. Read,

F. W. Taylor,

T.

Fouchet,P.Parrish,

A.Barucci,

R.Courtin,

A.Coustenis,

D.Gautier,

E.Lellouch,A.Marten,

R.Prangé,

Y.Biraud,

C.Ferrari,

T.C.Owen,

M.M.Abbas,R.E.Samuelson,

F.Raulin,

P.Ade,

C.J.Césarsky,

K.

U.Grossman,andA.Coradini,“Jupiter’sAtmosphericComposition

from the Cassini Thermal Infrared Spectroscopy Experiment,”

Science305,1582-1586(2004).

52.B. Bézard, C. A. Nixon, I. Kleiner, D. E. Jennings, “Detection of13CH3DonTitan,”Icarus191,397–400(2007).

53.C.A.Nixon,D.E.Jennings,B.Bézard,S.Vinatier,N.A.Teanby,K.

Sung,T.M.Ansty,P.G.J.Irwin,N.Gorius,V.Cottini,A.Coustenis,

andF.M.Flasar, “DetectionofPropene inTitan’sStratosphere,”

Astrophys.J.Lett.776,L14-19(2013).

54.A. Jolly,A.Fayt,Y.Benilan,D. Jacquemart,C.A.Nixon,andD.E.

Jennings,“Theν8BendingModeofDiacetylene:FromLaboratory

Spectroscopy to the Detection of13C Isotopologues in Titan’s

Atmosphere,”AstrophysJ.714,852-859(2010).

55.D. E. Jennings, C. A. Nixon,A. Jolly,

B. Bé́zard,

A. Coustenis,

S.

Vinatier,P. G. J. Irwin,

N. A. Teanby,

P. N. Romani,

R. K.

Achterberg,and F. M. Flasar, “Isotopic Ratios in Titan’s

Atmosphere from Cassini/CIRS Limb Sounding: HC3N in the

North,”AstrophysicalJournal681,L109-111(2008).

56.S.Vinatier,B.Bé́zard,S.Lebonnois,N.Teanby,R.K.Achterberg,R.

deKokandD.E. Jennings, “Seasonal variations inTitan’smiddle

atmosphere of Titan during the northern spring derived from

Cassini/CIRS observations,” American Astronomical Society

DivisionforPlanetarySciences,meeting#46,id.211.16(2014).