exoplanetcommunityreport.pdf - exoplanet exploration

JPLPublication09‐3

ExoplanetCommunityReportEdited by: P. R. Lawson, W. A. Traub and S. C. Unwin

NationalAeronauticsandSpaceAdministrationJetPropulsionLaboratoryCaliforniaInstituteofTechnologyPasadena,California

March2009

The work described in this publication was performed at a number of organizations, including the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (NASA). Publication was provided by the Jet Propulsion Laboratory. Compiling and publication support was provided by the Jet Propulsion Laboratory, California Institute of Technology under a contract with NASA. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not constitute or imply its endorsement by the United States Government, or the Jet Propulsion Laboratory, California Institute of Technology.

© 2009. All rights reserved.

iii

Theexoplanetcommunity’stoppriorityisthatalineofprobeclassmissionsforexoplanetsbeestablished,leadingtoaflagshipmission

attheearliestopportunity.

v

Contents

1 EXECUTIVESUMMARY.................................................................................................................. 11.1 INTRODUCTION...............................................................................................................................................11.2 EXOPLANETFORUM2008:THEPROCESSOFCONSENSUSBEGINS.....................................................21.3 RECOMMENDATIONSFROMTHEEXOPTF.................................................................................................31.4 CHAPTERSUMMARIES...................................................................................................................................51.5 OVERARCHINGSCIENCEGOALS...................................................................................................................91.6 SPACEMISSIONS.............................................................................................................................................91.7 GROUND‐BASEDANDSUB‐ORBITAL....................................................................................................... 101.8 TECHNOLOGY ............................................................................................................................................... 101.9 CONCLUSIONS .............................................................................................................................................. 101.10 REFERENCES.............................................................................................................................................. 12

2 ASTROMETRY ................................................................................................................................132.1 INTRODUCTION............................................................................................................................................ 132.1.1 AMicroArcsecondAstrometryMission ....................................................................................142.1.2 StronglySupportedRecommendations .....................................................................................212.1.3 RecommendationswithMixedorAmbivalentSupport ......................................................24

2.2 OBSERVATORYCONCEPT........................................................................................................................... 272.2.1 Architecture ...........................................................................................................................................272.2.2 Performance ..........................................................................................................................................29

2.3 TECHNOLOGY ............................................................................................................................................... 312.3.1 PastAccomplishments.......................................................................................................................312.3.2 FutureMilestones ................................................................................................................................38

2.4 RESEARCH&ANALYSISGOALS ................................................................................................................ 392.4.1 Research&Analysisin2010–2020..............................................................................................392.4.2 AstrometryBeyond2020 .................................................................................................................40

2.5 CONTRIBUTORS ........................................................................................................................................... 412.6 REFERENCES ................................................................................................................................................ 42

3 OPTICALIMAGING........................................................................................................................473.1 INTRODUCTION............................................................................................................................................ 473.1.1 ScienceGoalsandRequirements ..................................................................................................48

3.2 OBSERVATORYCONCEPTS......................................................................................................................... 563.2.1 ArchitectureScalingforInternalCoronagraphs...................................................................563.2.2 ArchitectureScalingforExternalandHybridConcepts ....................................................593.2.3 PerformanceScaling..........................................................................................................................603.2.4 CostandRiskDrivers .........................................................................................................................63

3.3 TECHNOLOGY ............................................................................................................................................... 643.3.1 InternalCoronagraphTechnology ..............................................................................................66

vi

3.3.2 ExternalOcculterTechnology........................................................................................................733.3.3 CommonTechnology..........................................................................................................................79

3.4 RESEARCH&ANALYSISGOALS ................................................................................................................ 833.5 CONTRIBUTORS ........................................................................................................................................... 843.6 REFERENCES ................................................................................................................................................ 85

4 INFRAREDIMAGING ....................................................................................................................914.1 INTRODUCTION............................................................................................................................................ 914.1.1 SensitivityandAngularResolution .............................................................................................934.1.2 ProbeScaleMissionScienceGoals...............................................................................................944.1.3 FlagshipMissionScienceGoalsandRequirements ..............................................................96

4.2 OBSERVATORYCONCEPTS.......................................................................................................................1024.2.1 ProbeScaleMissionConcept....................................................................................................... 1024.2.2 FlagshipMissionConcept.............................................................................................................. 107

4.3 TECHNOLOGY .............................................................................................................................................1094.3.1 ExperimentsinNullingInterferometry,1999–2009 ........................................................ 1094.3.2 TechnologyforProbeandFlagshipMissions....................................................................... 1124.3.3 AdditionalTechnologyforaFlagshipMission .................................................................... 1184.3.4 FutureMilestones ............................................................................................................................. 121

4.4 RESEARCH&ANALYSISGOALS ..............................................................................................................1254.4.1 GroundBasedInterferometry .................................................................................................... 1254.4.2 TheorySupport.................................................................................................................................. 1254.4.3 SpaceBasedInterferometry........................................................................................................ 1254.4.4 AgencyCoordinationandProgrammaticSynergies......................................................... 1264.4.5 InternationalCooperation,Collaboration,&Partnership ............................................. 126

4.5 CONTRIBUTORS .........................................................................................................................................1264.6 REFERENCES ..............................................................................................................................................128

5 EXOZODIACALDISKS................................................................................................................ 1355.1 INTRODUCTION..........................................................................................................................................1355.1.1 PropertiesofProtoplanetaryandExozodiacalDisks....................................................... 1365.1.2 TerrestrialPlanetFormationandDeliveryofVolatiles .................................................. 1385.1.3 ScienceGoals ...................................................................................................................................... 1385.1.4 ScienceRequirementsforObservationsofTerrestrialPlanets .................................... 139

5.2 CURRENTSTATEOFTHEFIELD..............................................................................................................1425.2.1 Photometry.......................................................................................................................................... 1425.2.2 DiskImaging....................................................................................................................................... 1435.2.3 SpectroscopyofDustandGas ..................................................................................................... 1475.2.4 TheoryandModeling...................................................................................................................... 148

5.3 TECHNOLOGYDEVELOPMENTANDPERFORMANCEDEMONSTRATION..........................................1525.3.1 PastAccomplishments.................................................................................................................... 1525.3.2 Performance ....................................................................................................................................... 1535.3.3 FutureDevelopmentsandMilestones ..................................................................................... 155

5.4 RESEARCH&ANALYSISGOALS ..............................................................................................................1575.5 CONTRIBUTORS .........................................................................................................................................1585.6 REFERENCES ..............................................................................................................................................159

6 MICROLENSING .......................................................................................................................... 1656.1 INTRODUCTION..........................................................................................................................................1656.1.1 HowMicrolensingFindsPlanets................................................................................................ 1666.1.2 ScienceGoals ...................................................................................................................................... 169

vii

6.1.3 ScienceRequirements..................................................................................................................... 1706.2 OBSERVATORYCONCEPT.........................................................................................................................1726.2.1 Architectures ...................................................................................................................................... 1726.2.2 Performance ....................................................................................................................................... 177

6.3 TECHNOLOGY .............................................................................................................................................1806.3.1 FutureMilestones ............................................................................................................................. 181

6.4 RESEARCH&ANALYSISGOALS ..............................................................................................................1816.4.1 DetailedPredictionsforaNextGenerationGroundBasedSurvey............................ 1816.4.2 SelectionofCandidateFieldsforFutureSurveys ............................................................... 1826.4.3 DetailedTradeStudyforaJointDarkEnergy/MicrolensingMission....................... 182

6.5 CONTRIBUTORS .........................................................................................................................................1826.6 REFERENCES ..............................................................................................................................................183

7 RADIALVELOCITY..................................................................................................................... 1857.1 INTRODUCTION..........................................................................................................................................1857.1.1 RelevancetoSpaceMissions........................................................................................................ 1867.1.2 ScienceGoals ...................................................................................................................................... 1897.1.3 ScienceRequirements..................................................................................................................... 190

7.2 OBSERVATORYCONCEPTS.......................................................................................................................1927.2.1 ArchitectureandPerformance................................................................................................... 192

7.3 TECHNOLOGY .............................................................................................................................................1967.3.1 PastAccomplishments.................................................................................................................... 1967.3.2 FutureMilestones ............................................................................................................................. 197


8 TRANSITS ..................................................................................................................................... 2058.1 INTRODUCTION..........................................................................................................................................2058.1.1 ScienceEnabledbyTransits ........................................................................................................ 2078.1.2 MeasurementRequirements........................................................................................................ 2078.1.3 PreviousandPlannedTransitMissions.................................................................................. 208

8.2 OBSERVATORYCONCEPTS.......................................................................................................................2098.2.1 Architecture ........................................................................................................................................ 2128.2.2 Performance ....................................................................................................................................... 213

8.3 TECHNOLOGY .............................................................................................................................................2168.3.1 PastAccomplishments.................................................................................................................... 2168.3.2 FutureTechnologyDevelopment:Milestones? .................................................................... 217


9 MAGNETOSPHERICEMISSION............................................................................................... 2219.1 INTRODUCTION..........................................................................................................................................2219.1.1 ScienceGoals ...................................................................................................................................... 2219.1.2 ScienceRequirements..................................................................................................................... 225

9.2 OBSERVATORYCONCEPT.........................................................................................................................2269.2.1 Architecture ........................................................................................................................................ 2269.2.2 Performance ....................................................................................................................................... 227

9.3 TECHNOLOGY .............................................................................................................................................2299.3.1 PastAccomplishments.................................................................................................................... 229

viii

9.3.2 FutureMilestones ............................................................................................................................. 2309.4 RESEARCH&ANALYSISGOALS ..............................................................................................................2319.5 CONTRIBUTORS .........................................................................................................................................2329.6 REFERENCES ..............................................................................................................................................232

LISTOF2008EXOPLANETFORUMPARTICIPANTS ............................................................. 235COPYRIGHTPERMISSIONS ............................................................................................................ 241ACRONYMLIST .................................................................................................................................. 245

ExecutiveSummary

1

1 ExecutiveSummary

AlanDressler,CarnegieInstitutionofWashington

WesleyTraub,JetPropulsionLaboratory

1.1 IntroductionAstronomyhasbeenanimportantpreoccupationofhumansforthousandsofyears,butthemost profound questions—“are there otherworlds and other beings?” reach back to theorigin of homo sapiens. Despite centuries of speculation, proof of the existence of otherworldshadtowaituntilGalileoturnedhistelescopetothenightsky400yearsago.GalileowasthefirsttotrulyseetheplanetsandmoonsinourSolarSystemasotherworlds.Yetittook until the end of the 20th Century beforewe developed telescopes and spacecraft toview—upclose—theplanets,theirmoons,andthepersistentdebrisfromwhichtheyhaveformed.AlloftheseholddeepsecretsoftheEarth’soriginsandlikelythebeginningoflifeitself.

Itisironicthatwhatisarguablythemostcompellingsubjectinastronomy—thesearchforother worlds and other life beyond our Solar System—emerges only now, in the 21stCentury. Fourcenturiesofdiscoveryhavebroughtusa remarkableunderstandingof thebirthandevolutionofstars,thehistoryofgalaxies,andevencosmology—thedevelopmentof the entire universe, but now it seems that the first shall be last. Not for a lack ofimaginationormotivation, but simply for thewantof technology, ouroldest anddeepestquestions, theonesmostrelevant toourownoriginsand fate,haveremainedbeyondourgraspforthousandsofyears.

Weareindeedfortunatetoliveinthetimewhenthislastbarriertooursearchisfalling.Itisreasonable to thinkthat thesearch forotherworldsandother life,eventhough limitedfor the foreseeable future to our own corner of the Milky Way galaxy, will dominateastronomicalresearchbeforemid‐century.Thesignsofthisarealloverthefield—theyouthandenthusiasmofleadinginvestigators,thesurgeinresearchpapers,thedeterminationofstudentstochoosethediscoveryandstudyofplanetsastheirlife’swork,thecrowdsdrawnto topical discussions at society meetings, the swell in public interest—all point to astunning growth of what was, only twenty years ago, one of the smallest astronomicalbranches.

In2010theDecadalSurveyofAstronomy&Astrophysicswillreportitsfindings.The1991DecadalSurvey(NationalResearchCouncil1991)recommendeddevelopinganastrometricmission forastrophysics, later called theSpace InterferometryMission (SIM). Exoplanetshad not yet been discovered, and were mentioned only in passing. The 2001 DecadalSurvey (National Research Council 2001) reaffirmed SIM for flight in the decade, notingespecially its planet‐detection capability, and in addition recommended technology

Chapter1

2

development for the Terrestrial Planet Finder (TPF) with launch expected in the 2010decade.Todayweanticipatethatexoplanetswillplayamuchlargerandmorediverseroleas the new Decadal Survey Committee tries to forecast trends in the field and assignscientificpriorities.

Adecadeago,ahandfulofgiantplanets,unlikethoseinourSolarSystembecausetheyorbitvery close to their parent stars, had been discovered by the radial velocity method—arevolutionarystep.TodayJovian‐massplanetsandconsiderablysmalleronesareknownbythe hundreds, and even planetary systems similar to our own have been found.Gravitational lensing has been used to find a true analog to the Solar System, somewhatscaleddown,consistingofbothaJupiter‐likeandSaturn‐likeplanetoncircularorbitsafewastronomical units (AU) from their star. Coronagraphic and related high‐contrast direct‐imagingtechniqueshaveproducedimagesoftwoplanetarysystems.Precisionphotometryhasledtothediscoveryofplanetstransitingtheirparentstars.Thetransitgeometryallowsdirect detection of a planet’s emergent light in some cases, and characterization of itsatmosphere. New methods of finding and characterizing planets—astrometry andoptical/infraredimaging—aredevelopingrapidly.Applicationofthesemethods,whichwillgreatlyincreasenotjusttheinventoryofplanetsbutalsobegintocharacterizethem,isnowlimitedmorebyresourcesthantechnology.

Ten years ago it seemed sensible to “throw the long ball”—build a planet‐finding spacetelescope that could findplanets as small asEarth in the “habitable zone” of neighboringstars,andobtainspectrathatwoulddescribethem,andpossiblyevenfindevidenceforlifethroughthepresenceoffreeoxygenintheiratmospheres.Althoughstillagalvanizinggoalof this rapidly changing field, today’s research directions aremany, focused on studyingplanets of all sizes, characterizing planetary systems, andmapping circumstellar disks ofrockyoricydebrisanddustthatwillbecomenewfamiliesofplanets.Inthecomingdecadecrucial data, for example, the frequency of Earth‐like planets around Sun‐like stars, andspecificallywhichnearbystarshavethem,shouldbeavailabletoinformtheambitiousgoalofdirectimaginganddetailedcharacterizationofextrasolarplanets.

ThegoalofExoplanetForum2008istotakeasnapshotofthisrapidlyemergingfield.Sucha review will be essential to the Committee and panels of the 2010 astronomy andastrophysicsdecadalsurvey—nowbeingorganizedbytheNationalResearchCounciloftheNational Academyof Sciences. The 2010 surveywill chart the prospects for this comingdecade,andwillinturn,forthefirsttime,setacourseforwhatisrapidlybecomingoneoftheleadingfieldsofastronomy.Adoorlongclosed,butforeveronthemindofhumankind,isopeningwide.

1.2 ExoplanetForum2008:TheProcessofConsensusBegins

Theexoplanet scienceand technology community in theUnitedStatesmet face‐to‐faceatthe Exoplanet Forum in May 2008. The community also met electronically from Aprilthrough October 2008, in chapter‐writing sub‐groups, to formulate the individualtechnique‐based chapters in this volume. These meetings were designed to thoroughlyresearch and discuss the many scientific opportunities for studying exoplanets, with thenear‐term goal of providing an integrated plan for exoplanet research for the comingdecade.

ExecutiveSummary

3

Eachof theeightchapterssummarizes the issues, scientificpotential, state‐of‐the‐art,andtechnology needed for the respective observational technique. The broad‐rangingdiscussion at the Forum allowed the scientists and engineers working in each area toprepare their own chapter in the context of the full scope ofmany goals and techniques.Thus, inherent in this document are cross‐references of how the various approachescomplementandcompetewitheachother.Itishopedthatthisdocumentwillbevaluableinputtofuturemeetingsthatwillcontinuetoaddressthefield,andthatthediscussionherewillhelptoformulateoptionsandbranchpointsthatcanguidethestudyofexoplanets inthecomingdecade.Informulatingthisplan,thecommunitywillneedtobalancescientificaspirationswith expectations andhopes for future science, technology, and funding. Ourtemperedoptimismineachareaspringsfromourbeliefthatexoplanetscienceisthemostexciting new area in astrophysics, for scientists and the public alike. We believe thatUSleadership in exoplanet science is essential formajor advances in this field, andwe alsobelievethatthepublicexpectsnoless.

The present community report is outlined below under the following topics:Recommendations from the Exoplanet Task Force (summarized here for continuity),Chapter Summaries, Overarching Science Goals, Space Missions, Ground‐Based and Sub‐Orbital,Technology,andConclusions.

1.3 RecommendationsfromtheExoPTFTheExoplanetTaskForce(ExoPTF)(2008)wasablue‐ribbonpanelofexoplanetscientists,convenedtostudythefieldandmakerecommendations.TheExoPTFwascharteredbytheNational Science Foundation (NSF) and NASA through the Astronomy and AstrophysicsAdvisory Committee (AAAC) “to adviseNSF andNASAon the future of the ground‐basedandspace‐basedsearch forandstudyofexoplanets,planetarysystems,Earth‐likeplanetsandhabitableenvironmentsaroundotherstars.”

TheExoPTFconfrontedtheseprioritizedquestions:

1. What are the physical characteristics of planets in the habitable zones aroundbright,nearbystars?

2. Whatisthearchitectureofplanetarysystems?

3. When,how,andinwhatenvironmentsareplanetsformed?

ThefinalExoPTFReportwasmadepubliconeweekbeforetheExoplanetForummeetinginMay 2008, but preliminary versions of the reportwere already available tomany Forumparticipants well before then. Therefore, the participants were able to discuss scientificgoalsandtechniquesinthecontextoftheExoPTFReport,whenappropriate.

The exoplanet community acknowledges and appreciates the substantial efforts of theExoPTF committee in formulating the ExoPTF Report. The exoplanet community agreeswiththesubstanceoftheExoPTFReport,aswillbeseeninthepresentbook.Howeveronafewissuesthecommunityviewdifferssomewhat,basedonitsunderstandingofthescienceand technology. Specific differences are discussed in the individual chapters and in theconclusionsbelow. TheExoPTFanswered itsprioritizedquestionsby recommending thefollowing list of action items for the coming 15 years (listed here with simplifiednumbering).

Chapter1

4

The1–5yearprogramA.1.Extendradial‐velocity‐detection(RV)tolow‐massplanetsandcontinuesurveyingover

1000stars.

A.2.PrepareforanastrometricspacemissiontodetectEarth‐massplanetsaroundatleast60stars.

A.3.Preparefordirect‐detectionspacemissionstocharacterizeEarth‐massplanets.

A.4.Establishexpertpanelstoexhaustivelyevaluatedirect‐detectionmethods.

A.5.Investinacensusofexo‐zodisystemsaroundnearbystars.

A.6.Searchthenearest1000M‐dwarfsusingtransitsandRV,downtoEarth‐sizes.

A.7.DevelopRVinthenearinfraredandwith1ms–1precision.

A.8.Increasetheground‐basedmicrolensingnetwork.

A.9.IncreaseresearchonAOinthelabandon8‐mtelescopes,andevaluateground‐basedpotential.

A.10.MaintainUS support for the study of accretion anddebris disks, using ground‐ andspace‐basedfacilities,andarchives.

A.11.Supportplanetformationworkviagroundandspaceobservations,andtheory.

The5–10yearprogramB.1.LaunchaspaceastrometricmissiontofindEarth‐massplanetsintheHZofatleast60

stars.

B.2.Formulateadirect‐detectionspacemission,contingentonknowledgeoftheEarth‐likeplanetfrequency(eta‐sub‐Earth)andthezodiacallightofexosolarsystems.

B.3.UseJWSTtocharacterizeEarth‐sizeplanetstransitingM‐dwarfs.

B.4.Continuelong‐termRVtoreachplanetsbeyondthesnowline.

B.5. Launch a Discovery‐class microlensing mission, without impacting the astrometrylaunch.

B.6.Beginconstructionofa30‐mclassground‐basedtelescopefordirectdetectionofgiantplanets.

B.7. Implementground‐basedAOforyounglow‐masscompanions,andinterferometryfordisks.

The11–15yearprogramC.1. Launch a space direct‐detectionmission for Earth‐size planets, if eta‐sub‐Earth and

zodiallow.

C.2.Developfuturespacedirect‐detectionmissionsforlaunchlater.

C.3.Pursueground‐basedsearchesifeta‐sub‐Earthislow(<0.1),exozodisarehigh,orfewEarth‐likeplanetsarefoundthroughastrometry.

C.4.Developtechnologyforadvanceddiskscienceinthefarinfrared.

ExecutiveSummary

5

AllyearsD.1.Supporttheoreticalmodelingofplanetsandplanetsystems.

D.2.Supportdataanalysisandtrainingofyoungscientists.

In the Exoplanet Community Report, we provide important additional background andcontext for the findings of the ExoPTF, and in some cases we express our particularviewpoints.This ispresented in the formof the eight technique‐oriented chapters,whichdetail scientific and technical considerations of different approaches to the study ofexoplanets. Bydebatingtheissuesandwritingthechapters,overaperiodofaboutsevenmonthsinmid‐2008,communityinvolvementandparticipationwasraisedsignificantly.

1.4 ChapterSummariesKeypointsfromeachchapteraresummarizedhere,particularlyastheyrelatetotheoverallcoherentpicture. Theorderingof the chaptersmostly follows theForummeeting,whichwasoriginallyalphabetical.EachchapterischairedbyaUSexpertinthefield,doublingasamember of the Science Organizing Committee of Forum 2008, and co‐chaired by a JPLscientist,doublingasamemberoftheLocalOrganizingCommitteeoftheForum.Thechairand co‐chair led groups of 6 to 50 scientists (median~27) in formulating, debating, andwritingeachchapter.Intotal182scientistscontributedtothechapters.

Weusetheterm“Earth‐twin”asshorthandforaterrestrial‐massplanet(0.3to10.0Earths)inthehabitablezone(HZ)arounditsstar.TheHZisdefinedintheScienceandTechnologyDefinitionTeamreport(Levine,Shaklan&Kasting2006)astherange0.75to1.80AUfromastar,scaledbythesquarerootofthestar’sluminosityrelativetothecurrentSun.Wealsouse“RV”asshorthandforradialvelocity,and“exozodi”forexozodiacal.

Weuse the term “micro‐arcsecastrometry” tomeananastrometric capabilitywithaboutonemicro‐arcsecaccuracypermeaserement,butmore importantlyamissionaccuracyofabout0.2micro‐arcsec,whichissufficient foranEarth‐twinat10pcwith lowfalse‐alarmprobability.

Weusetheterm“flagship‐class”toindicateascaleofmissionwhosecostisgreaterthanabilliondollars.Theterm“probe‐class”isheremeanttoindicateamissionwhosecostislessthanabilliondollars.

We address the issue of mission selection and scheduling, very briefly. The ExoPTFrecommendedthatanastrometricmissionbe launchedbeforezodi‐detectionandimaging(characterization)missions. TheExoplanetCommunityReportrecommendsthat thetypeofmissiontobelaunchedfirstshouldbecompetitivelyselected.Ourreasonsareasfollows.Weagreethatanastrometricmissionwouldbeavaluablefirstlaunch,howeverifitscostistoolargeforanear‐termlaunch,orifitcannotmeetthetechnicalrequirementofbeingableto findEarth‐twinsaroundenoughnearbystars, then itwouldbeworthwhile to launchaprobe‐class zodi or characterization mission first, because these could return valuablescience. Atthepresenttimeit is impossibletodecidetheissuesofcostandperformance,but with expected near‐term advances in these areas, for all mission types, it will bepossibletoselectafirstmissiononthebasisofcompetitiveproposals.

Chapter2:AstrometryMicro‐arcsecondastrometry is theonlyprobe‐scale technique thatcandetectEarth‐twinsaround as many as 100 nearby stars, independent of orbital inclination or exozodi

Chapter1

6

brightness. Becausenoother technique cando this, theExoPTF recommended thenear‐termlaunchofaspacemissionformicro‐arcsecastrometry. Thesciencerationaleforthismission is two‐fold. In a narrow sense, if Earth‐twins aredetectable, then it follows thatmany other planets,with larger astrometric signatures, are detectable, so the Earth‐twingoal stands as a useful benchmark of instrument performance. In a wider sense, thedetectionofpotentialEarth‐twinsisofgreatvalueasexistenceproofsofplanets,asdesignoptimizers for follow‐up characterization missions, and as search optimizers for thosemissions. Combining ground‐based RV observations with space‐based astrometry is anespecially powerful detection strategy, because RV aids astrometry in extracting massesand orbits from the extremes of short‐ and long‐period planet signatures inmulti‐planetsystems. Once nearby planets are identified, and their masses and orbits known, thenfollow‐up can be planned tomeasure exozodi brightness and to characterize the planetsthroughimagingobservations,orpossiblytransits.ThetechnologyforsuchamissionwasdevelopedbyNASAover thepastdecadeand isnowcomplete; all technical aspectshavebeensolved.Itistheopinionofthischapter’sstudygroupthatanastrometricmissioncanbe flown thatmeets the goal of being approximately probe‐class in cost, and that can bereasonablyexpectedtodetectenoughterrestrial‐massplanetstoformthebasisofourfirstexploratorystudyoftheseobjectswithfollow‐oncharacterizationmissions.Amicro‐arcsecastrometric mission is an excellent first step toward understanding and characterizingnearbyexoplanets,andleadsnaturallytodirectopticalorinfraredimaging,giventhatonceaplanet’sorbit isknown,an imagingmissionknowsthataplanet targetexists,and ithasgoodhintsastowhenandwheretolook.

Chapter3:OpticalImagingAvisible‐wavelengthcoronagraphistheonlytechniquecapableofobservingthespectrumofabenchmarkEarth‐twindeepenoughinitsatmospheretoprobethelowertroposphereand surface—to assess habitability and to search for direct signs of life. Here “visible”potentiallyincludesthenear‐ultravioletandthenear‐infrared.“Coronagraph”includestheinternal as well as external occulter types for suppressing the 10‐billion‐times‐brighterparentstar.Asmallcoronagraphcouldmeasurethebrightness,colors,andspectraoflargeplanetsoutsidetheHZaroundnearbystars,possiblyafewterrestrialplanetsintheHZ,andthe zodi brightness distribution in the same regions. A large coronagraph can measurecloser to the parent star, and itwill have better spatial resolution, so it can characterizemore giant aswell as terrestrial planets, includingmore Earth‐twins in theHZ. If eithertypeofcoronagraphisprecededbyasuccessfulastrometrymission,thenthecoronagraphdesign could be potentially optimized, and the efficiency of locating planets would besignificantly improved. Coronagraph technology is well advanced, and the technologicalrequirements are well understood. The key technologies for a large‐scale mission areidentical to those for a small‐scale one, so all such work on a small mission is a soundinvestmentforafuturelargeone.

Chapter4:InfraredImagingAmid‐infraredmissionwouldenablethedetectionofbiosignaturesofEarth‐likeexoplanetsaround more than 150 nearby stars. The mid‐infrared spectral region is attractive forcharacterizing exoplanets because contrast with the parent star brightness is morefavorable than in the visible (10 million vs. 10 billion), and because mid‐infrared lightprobes deep into a planet’s troposphere. Furthermore, themid‐infrared offers access toseveralstrongmolecularfeaturesthatarekeysignsoflife,andalsoprovidesameasureoftheeffectivetemperatureandsizeofaplanet. Takentogether,an infraredmissionplusa

ExecutiveSummary

7

visible onewould provide a nearly full picture of a planet, including signs of life; with ameasureofmassfromanastrometricmission,wewouldhaveavirtuallycompletepicture.A small infraredmissionwouldhave several telescopes thatare rigidly connected,withascience return from the detection and characterization of super‐Earth sized to largerplanetsneartheHZ,plusadirectmeasureof theexozodibrightness in theHZ. Ina largeinfraredmission,withformation‐flyingtelescopes,planetsfromanEarth‐twinandupwardsinmass could be detected and characterized, aswell as the exozodi. If proceeded by anastrometric mission, the detection phase could be skipped and the mission devoted tocharacterization,asinthevisiblecase;lackinganastrometricmission,aninfraredonecouldproceedalone, aswasdiscussed for avisible coronograph, andwith similar caveats. Thetechnology needed for a large formation‐flying mission is similar to that for a smallconnected‐element one (e.g., cryogenics and detectors), with the addition of formation‐flying technology. The technology is now in hand to implement a probe‐scale mission;starlight suppressionhasevenbeendemonstrated tomeet the requirementsofa flagshipmission. However, additional development of formation‐flying technology is needed,particularlyin‐spacetestingofsensorsandguidance,navigation,andcontrolalgorithms.

Chapter5:ExozodiacalDisksFromtheviewpointofdirectimagingofexoplanetsinthevisibleorinfrared,exozodidustdiskscanbebothgoodandbad.Anexozodidiskisgoodifithasstructures(clearedregionsorresonantclumps)thatsuggestthegravitationalpresenceofplanets,howeveritisbadifthedust fillsthe instrumental fieldofviewwithbrightnessthatswampsthesignal fromaplanet.Unfortunately,ittakesverylittledusttocompetewithoroverwhelmthelightfromaplanet:anEarth‐twinsignalisroughlyequaltoa0.1‐AUpatchofSolar‐System‐twinzodi,in thevisibleor infrared. Thus,exozodimeasurementsareextremely important,but theyare also difficult to make. Current limits of detection, in units of the Solar‐Systembrightness,area fewhundredusingtheSpitzerSpaceTelescope,aboutonehundredwiththeKeck Interferometer (KI), and about 10 expected from the LargeBinocularTelescopeInterferometer(LBTI). Asmallcoronagraphorsmall interferometerinspaceisneededinorder to reach the sensitivity required to detect the glow at the level of our own SolarSystem.

Chapter6:MicrolensingGravitationalbendingandfocusingof lightfromadistantstarbyacloserstar(andplanetsystem) is known asmicrolensing, and has already produced several dramatic exoplanetdetections. The statistics of these detections suggest that low‐mass planets are morecommonthanhigh‐massones,anextremelyimportantresultinthecontextofthesearchforEarth analogs. Microlensing is typically sensitive to planets around distant (several kpc)stars,notnearby(severalpc)ones,andisthereforemostvaluableasastatisticalsamplingtool. It is not an effectivemethod for detecting planets around nearby stars, those thatwouldbewithintherangeofdirectimagingandspectroscopiccharacterization. Owingtothescarcenessofmicrolensingevents,moreground‐based telescopesareneeded;aspacetelescopededicatedtothepurposewouldsubstantiallyimprovethestatistics.Anattractivefeatureofthisapproachisthatnonewtechnologyisneeded.

Chapter7:RadialVelocityRadial velocity (RV) is, of course, themethodof detecting planets bymeasuring a parentstar’s periodic line‐of‐sight velocity change due to its orbit around a common center of

Chapter1

8

mass,usingtheDopplershiftofthestar’sspectrallines.Themethodissensitivetowalking‐speedmotions of a star, andpromises to reach crawling‐speed sensitivity in a fewyears.The RVmethod is responsible formost of the 300‐plus exoplanets detected to date, andcontinues to produce results, particularly for long‐period systemswhere at least one fullorbit,andpreferablyseveral,isneededtobesureofadetection.Ground‐basedtelescopesare perfectly adequate for this task, although more telescopes, and additional stablespectrographs, would be very welcome. The main new technology is the laser combspectrum, which can provide many more calibration points across a spectrum thantraditionallaboratoryabsorptionoremissionsources.Incombinationwithaspacemicro‐arcsec astrometry mission, RV holds good promise for helping unscramble long‐periodmulti‐planetsignals,andthusdetectingEarth‐twins.

Chapter8:TransitsThetransit techniquerefers toobservationsofdecreasedbrightnessofastarsystemasaplanetpassesinfrontoforbehindthestar.Specificallyitincludesthespectralinformationthat comes from starlight transmitted through theplanet’s atmospheric annulus during aprimarytransitandreflectedorradiatedlightfromtheplanet’sfulldiskduringasecondarytransit.Itcanalsoincludethephaseeffectfromreflectedorradiatedlightduringitsorbitalmotion,independentofanyprimaryorsecondaryeclipseevents.Themethodworksinthevisibleaswellasthermalinfrared.Thetransitmethoddeliversuniqueinformationaboutaplanet’sdiameter,andabouttheabsorptionspectrumofitsupperatmosphere,indicativeofphysicalconditionsandchemicalcomposition.Thetransittechniqueisespeciallyusefulforshort‐period planets because these are close to their star and have a greater chance oftransiting.Thefullpowerofthetransittechniqueislimitedtoarelativelysmallfractionofplanetarysystems,but similar combined‐light techniquescanbeapplied tomanyclose‐inplanets,eventhosethatdonottransit.Transitsarealsomostusefulforgiantplanets,owingto their large opaque cross sections and relatively large atmospheric annulus areas;conversely,transitsofEarth‐sizeplanetsarerelativelyweakandmaybedetectableonlyforclose orbits of late‐type stars. Ground‐based and space‐based observations are bothvaluable. TheKeplermissionwillbeespecially important forgeneratingameasureof thefrequencyofEarth‐sizeplanetsatseparationsouttotheHZ,andthefrequencyoflarger‐sizeplanets.Themaintechnologyneedisforgreaterstabilityinground‐basedmeasurements.

Chapter9:MagnetosphericEmissionPlanetsintheSolarSystemproduceradio‐frequencywavesbythemechanismofradiationfrom electrons spiraling along the planet’smagnetic field lines, and it is anticipated thatexoplanetswillbroadcastsimilarradiowaves. Theradioluminosityisproportionaltothesolarwindpower incidentonaplanet, so isexpected tobe largest foractivestars. Sincethesestarsaredifficult toobservewithastrometryorRV,magnetosphericemissionsmayoffer a path to detecting planets in these systems. There are a number of ground‐basedinstrumentsunderconstructionthatpromisesignificant improvementsinsensitivity. Themain technology needs are algorithmic, such as the development of improved radiofrequency interference (RFI) avoidance and excision. Groundobservationswill ultimatelybelimitedbyacombinationoftheEarth’sionosphereandRFI,howeverantennasonthefarside of the Moon could have large areas and very low RFI, and are therefore of specialinterest.

ExecutiveSummary

9

1.5 OverarchingScienceGoalsThegoalofexoplanetscienceistoextendourastrophysicalpictureoftheUniverse—theBigBang,stars,andgalaxies—toincludethedevelopmentofdisksaroundstars,theformationofplanets,andtheconditionsthatcanleadtolifeonplanets.

UnderstandingtheconditionsforlifeintheUniverseisarelativelynewscientificendeavor.It is clearly the logical next step for astrophysics, given our natural curiosity about theorigins of life onEarth. We already know thatmore than300 relativelymassive planetsexistaroundnearbystars;thesewerediscoveredusingRV,microlensing,andtransits.

There is broad‐based support in the community for continuing and expanding currentobservationalworkwithtransitphotometryandspectroscopy,microlensing,ground‐baseddirect imaging, and ground‐based interferometric imaging. Beyond these, the Forumdiscussionhighlightedsomeofthemajorstepsthattheexoplanetcommunityhopestotakewithinthenextdecadeortwo,likelyinthisorder:

1. Measurement of the frequency of low‐mass exoplanets, i.e., the mass distributionfunction.Thisworkisalreadyunderwaywithmicrolensing,CoRoT,andKepler.

2. Searches forall typesof exoplanets, from terrestrialmass togasgiants, aroundnearbystars.Probe‐scalemissionscanlikelyfulfillthisneed.Aleadingtechniqueappearstobeaspace‐basedastrometricmission,workingwith ground‐basedRV. However coronagraph,interferometer,transit,andmicrolensingmissionseachcouldmakevaluablecontributions.

3. Measurements of exozodi disk brightnesses to levels approaching that of the SolarSystem.Ground‐basedinterferometerswillpioneerthisarea,butaspacemissionisneededtoreachfaint(Solar‐Systemlevel)exozodis

4.Characterizationandsearchforsignsoflifeonnearbyexoplanets,especiallyEarth‐twins.The only viable techniques are coronagraphs and interferometers in space. Probe‐scalemissions can explore this area, characterizing many planets, including possibly a fewterrestrialones.Butflagship‐scalemissionswillbeabsolutelynecessaryforcharacterizingmostnearbyEarth‐twins.

1.6 SpaceMissionsGiven the discussion at the Forum, it seems that amix of probe‐scale and flagship‐scalemissionsisneededtofullyimplementthesciencestepslistedabove.

There is strong support for a line of probe‐scale exoplanet missions. Kepler, onceoperational, would be the first of these. If a probe‐scale astrometric mission could findnearbyplanetsandmeasuretheirmasses—includingsometerrestrialones,thiswouldhavestrongsupport,asindicatedbytheExoPTFReport,andbythepresentCommunityReport.Additional contenders for the probe‐scale line are a coronagraph or interferometer tomeasure exozodis and accessible planets, as well as a microlensing mission to improveplanetarystatistics.Selectionshouldbecompetitive.

There iswidespread agreement that two flagship‐scale exoplanetmissions are ultimatelyrequired:avisiblecoronagraphandan infrared interferometer. ThesearetheonlyviabletechniquesforfullycharacterizingEarth‐twinsdowntoabouttheirsurfaces,andsearchingfor signsof life. Inprinciple, aprobe‐scalemission couldpartially characterize anEarth‐twin if the target were well‐situated, but in general, basic physics tell us that no smallmissionforexoplanetscouldeverdowhatalargeonecould.

Chapter1

10

1.7 GroundBasedandSubOrbitalGround‐basedwork,includingsub‐orbitalsciencepayloads,inseveralareas,willbeneededinorder toachieve thesciencesteps listedabove. Newdiscoveriesaremade, technologyreadinessisadvanced,studentsaretrained,andnewmethodsperfected,usinggroundandsub‐orbital techniques. The Forum discussion emphasized the importance of continuedresearch using transits, microlensing, direct imaging, and interferometry. New ground‐basedfacilitiesforRVandmicrolensingwouldimprovetheireffectiveness.

1.8 TechnologyChapter 2: Astrometry. Technology development for amicroarcsecond astrometric spacemission is highly advanced, sufficient to consider a near‐term space mission. The onlyremainingworkneededistoengineeraflightinstrumentcapableoffindingnearbyEarth‐twincandidates.

Chapter 3: Optical Imaging. For internal coronagraphs, targeted technology developmentwill be needed in the areas of coronagraph concepts, diffraction modeling, laboratorydemonstrations,deformablemirrors,wavefrontsensingandcontrol,maskfabrication,andasphericpolishing.Forexternaloccultercoronagraphs,targeteddevelopmentisneededfordiffraction modeling, laboratory demonstration, occulter design, occulter deployment,formationflying,scatteredsunlightcontrol,andpropulsion.Forbothtypesofcoronagraph,common technologyneedsare in theareasof telescopeandmirror technology,detectors,precisionthermalcontrol,disturbanceisolation,andverification/validation.

Chapter 4: Infrared Imaging. For infrared interferometers, technology development isneeded in the areas of mid‐infrared spatial filters, adaptive nulling, achromatic phaseshifters, cryocoolers, cryogenic delay lines, thermal shields, detectors, four‐beam nulling,dual‐beamchopping,spectralfiltering,formationflying,andpropulsion.

Chapter 5: Exozodiacal disks. The main technology effort for exozodi disks is getting theLBTIon‐lineandoperatinginsurveymodetocharacterizethedustaroundnearbystars.

Chapter6:Microlensing.Nonewtechnologyisneededforeitheranexpandedground‐basednetworkoraspacemissionformicrolensing.

Chapter7:RadialVelocity.Technologydevelopmentisneededforlaser‐comblightsourcestocalibratethewavelengthscaleinground‐basedspectrometers.

Chapter 8: Transits. The main technology need is for greater stability in ground‐basedmeasurements.

Chapter9:MagnetosphericEmission.Themaintechnologyneedisforlargecollectingarrays,radio‐quietsites,andgoodradio‐frequencyinterference(RFI)rejection.

1.9 ConclusionsItisthesenseoftheexoplanetcommunitythattherearemanypointsonwhichweclearlyagree.Wesummarizethesepointsofagreementhere.

WeagreewiththesubstanceoftherecommendationsoftheExoPTF.Inparticular,thereisaconsensus thatanastrometricmissionwillbea scientificallyvaluable first step,with twoimportantconditions,(a)thatitcanbeexpectedtodiscoverseveralterrestrialplanets,and(b)thatithasanacceptablecost,presumablyintheprobe‐classrange.

ExecutiveSummary

11

Weagreethatalineofcompetedexoplanetprobe‐classmissionsisessentialforexoplanetscience,andweagree thateachof thesecontributescomplementaryscientificknowledge.Effectively,Keplerwillbethefirstofthese,afterit isoperational. Weagreethatpotentialprobe‐class missions include astrometry, microlensing, coronagraph, and interferometertypes.

Weagreethatflagship‐scalemissionsarerequiredtounambiguouslycharacterizeaplanetas Earth‐like, after it is detected. We support the concept of flagship‐class exoplanetmissions because the physics of characterizing a faint planet near a bright star demandsphysicallylargeapertures.

We agree that a flagship‐scale characterizationmission should be launched as soon as ispermittedbyresources,science,andgeneralinterest.

We agree that technical progress is essential for scientific progress. Indeed, thecoronagraphs and interferometers in today’s testbedswere unimagined a decade or twoago—they were invented specifically for exoplanet science. We believe that technologydevelopment is needed in the areas of external occulter coronagraphs, internalcoronagraphs, and formation‐flying interferometers. A more modest level of technologydevelopmentwillcontinuetobeneededintheareasofastrometry,radialvelocity,exozodidisk observations, microlensing, and transits. Significant progress in magnetosphericemissionisinaseparateclass,likelyrequiringlunarbacksideantennas. Weurgeastrongcontinuing program of technology development, guided in part by this and subsequentyearlyForummeetings.

Weagreethatcontinuedground‐basedobservingisimportant,especiallyfortransits,radialvelocity, interferometry,andlargetelescopeobserving. Ongoingdevelopmentsinground‐basedwork, forexampleadaptiveoptics, interferometricmethods,andcoronagraphy,willallcontributesignificantlytoexoplanetscience.

WeagreethattheNSFandNASAshouldworktofindcommongroundintheareaofground‐basedobserving.

We agree that sub‐orbital instruments on balloons and rockets are important forcoronagraph and transit observations. Sub‐orbital conditions provide a near‐spaceenvironment, which is important for eliminating effects of the Earth’s atmosphere, forincreasing technical readiness validation, and for training the next generation of spacescientistsandengineers.

We agree that theory should be adequately supported because it is crucially needed tointerpretandsometimesguideobservations.

Weagree that internationalpartners couldplayavery important role in collaboratingonscienceandinsharingthecostoffuturemissions.WeurgeNASAtofindformulaswherebyinternationalpartners canparticipate scientifically inproportion to their contributions tothecostofmissions.

In addition, there are several issues that require more information or time before thecommunitywilbeabletoexpressitsopinion.Welistthoseissueshere.

Weleaveforfuturediscussiontheissueofdecidingifweshouldprioritizetypesofmissions,forexampleexternaloccultercoronagraphsvs. internal coronagraphsvs. interferometers,andalsoprobevs. flagship. Suchprioritizationcannot replace competitiveproposalsandpeerreviewpanelsformissionselection,butitcouldservetofocusinterest intechnologydevelopmentwherethepayoffwasexpectedtobegreatest.

Chapter1

12

We leave for future discussion the issue of advising on technological priorities andmilestonestobereached,fortheeventualadvancementofexoplanetscience.Similarly,weleavetofuturediscussionhowtobalancerelative investments ingroundvs.space, theoryvs. experiment, testbeds vs. observing, and in short‐ vs. long‐term development projects,and similar issues. These are expected to become focus topics of future annual Forummeetings.

We suggest that, at its next general meeting, the exoplanet community directly addressthese questions: (a) What is the trade‐off between probe‐scale and flagship exoplanetmissions,all factorsconsidered? (b)Should thenextmissionbeastrometric,ordoes thisdepend on whether it would be probe‐scale or flagship‐scale? (c) What is the besttechniquetocharacterizeplanets,andwhenshoulditbecomethefirstpriority? (d)Whatarethemilestonesthatprecedeaflagshipexoplanetmission?(e)Howshouldinternationalpartnerships be pursued? (f) Which technology developments, what ground‐basedobservations,whatsub‐orbitalexperiments,andwhat theoryshouldbemostaggressivelypursued?(g)Whatfurtherstepscanthecommunitytaketodevelopacoherent,integratedplanforexoplanetresearch?

SummaryStatement:

Theexoplanetcommunity’stoppriorityisthatalineofprobeclassmissionsforexoplanetsbeestablished,leadingtoaflagshipmissionattheearliestopportunity.

1.10 ReferencesExoplanetTaskForce2008,WorldsBeyond:AStrategyfortheDetectionand

CharacterizationofExoplanets,Astronomy&AstrophysicsAdvisoryCommittee,NASA,NationalScienceFoundation,andDepartmentofEnergyhttp://www.nsf.gov/mps/ast/aaac/exoplanet_task_force/reports/exoptf_final_report.pdf

Levine,M.,Shaklan,S.,&Kasting,J.,eds.2006,TerrestrialPlanetFinderCoronagraph:ScienceandTechnologyDefinitionTeam(STDT)Report,JetPropulsionLaboratoryDocumentD‐34923,Pasadena,CA,USAhttp://planetquest.jpl.nasa.gov/TPF/STDT_Report_Final_Ex2FF86A.pdf

NationalResearchCouncil1991,TheDecadeofDiscoveryinAstronomyandAstrophysics,(theBacahlReport),NationalAcademyPress,Washington,DC,http://books.nap.edu/catalog.php?record_id=1634#toc

NationalResearchCouncil2001,AstronomyandAstrophysicsintheNewMillenium,(theMcKee–TaylorReport),NationalAcademyPress,Washington,DC,http://www.nap.edu/openbook.php?isbn=0309070317

Astrometry

13

2 Astrometry

MatthewW.Muterspaugh,TennesseeStateUniversity,Chair

AngelleTanner,JetPropulsionLaboratory,CoChairG.FritzBenedict,AlanBoss,MatthewBoyce,RobertBrown,GeoffreyBryden,AdamBurrows,JosephCatanzarite,M.MarkColavita,DavidCole,RolfDanner,DennisEbbets,EricFord,CarlGillmair,SallyHeap,N.JeremyKasdin,MarcKuchner,BenjaminLane,NicholasLaw,CharlesLillie,WalidMajid,GeoffMarcy,JamesMarr,BarbaraMcArthur,StanimirMetchev,JunNishikawa,CharleyNoecker,XiapeiPan,GuyPerrin,MeyerPesenson,StevenPravdo,StephenRinehart,JeanSchneider,StuartShaklan,MichaelShao,WesleyTraub,StephenUnwin,JulienWoillez

2.1 IntroductionAstrometryisapowerfultoolfordetectingandcharacterizingexoplanetsystems.Sofar,itsrolehasbeenprimarilytoprovideunambiguousmassesforplanetarysystemslike55CncandEpsilonEridani,whichwereoriginallydiscoveredwith radial velocity surveys. In thepastfewyears,therehavebeenground‐basedeffortstodetectJovianplanetsaroundlow‐massstarsandbinarysystems.However,ground‐basedradialvelocityprogramsremaintheprimary method for the detection of Jovian planets; the first such system detected byastrometrystilleludesus.Duetothesystematicnatureoftheintrinsicjitterinherenttoallstarsduetostarspots,granulationandfaculae,radialvelocitymeasurementsareunabletoreach thesensitivitiesnecessary todetect terrestrialplanets in thehabitablezonesofournearestsolar‐typestars.This ispossible,however,withmicro‐arcsecondastrometry.Onlymicro‐arcsecondastrometrycandetecthabitableEarth‐likeplanetsaroundnearbySun‐likestarswhilesimultaneouslymeasuringtheplanetmasses,themostfundamentalquantityforcharacterizingexoplanets.

The 2008 Exoplanet Forum Committee on Astrometry strongly supports therecommendations of the Exoplanet Task Force (ExoPTF)Report (Lunine et al. 2008) andthepast threeDecadalSurveys. TheCommittee’shighestpriority is todeploya facilityfor microarcsecond astrometry of nearby stars during the 2010–2020 decade, withprimary goals of finding Earth‐like planets and characterizing the structures of planetarysystems.Amicro‐arcsecondastrometryprogramwillsupportavarietyofscienceprogramsoutsidethefieldofexoplanets;theCommitteestronglysupportsamissiondesignthatwillenablethesenon‐exoplanet‐relatedresearchefforts.Thetechnologydevelopmenteffortsofthepastfewdecadeshaveprovidedasystemarchitecturecapableofthismicro‐arcsecondastrometry goal. The key technology milestones have been achieved. It is now time todeploythefacilityitself.

Chapter2

14

The ExoPTF also made several smaller recommendations for the role of astrometry inexoplanet studies over the next decade. These secondary science programs have moremodest budgets, and theymake smaller contributions to the understanding of planetarysystems.Anumberofthesesecondarysciencegoalsaresupportedbythecommittee,whichfinds that these shouldbepursued,butonly ifdoingsowillnotdelay thedeploymentof thehighestprioritymicroarcsecondastrometrymission.

2.1.1 AMicroArcsecondAstrometryMissionAll the highest priority science programs considered by the astrometry committee areaddressed by a single, ~5 year or longer astrometry mission capable of microarcsecondprecisions.

Onpageviiioftheirdraftreport,thefirstrecommendationtheExoPTFmakesforthe6–10year(2014–2018)timeframeisto“Launchandoperateaspace‐basedastrometricmissioncapableofachieving0.2micro‐arcsecsensitivitytoplanetsignaturesaroundof[sic]60–100nearby stars.” The past three decadal surveys have promoted development of a micro‐arcsecondastrometricprogramfordetectingandcharacterizingexoplanets.Theastrometrycommitteerecommendsthatthe2010decadalsurveycontinuestrongpromotionofamicroarcsecond astrometry mission, to be deployed in the early part of the 2010–2020 decade.Thosepastdecadalsurveysresultedinstrongsupport fortechnologydevelopmenteffortstodesignsuchamission.Allkeydevelopmentmilestonesforanastrometricmissionbasedon theSpace InterferometryMission (SIM)architecture—includingSIM‐Lite (Goullioudetal.2008;Marretal.2008)andPlanetHunter(NASAROSESAstrophysicsStrategicMissionConceptStudy,nowunderway)—havebeenaccomplished,and thecommunity isready tocapitalizeon thissignificant investment. Furtherdelaymaywaste this investment,as thecurrentexpertscouldbelosttootherpursuits.

Table21.AstrometricMissionConcepts

MissionName ScienceTopics PlanetMassSensitivity,10pc

NumberofStarsforTerrestrialPlanetSensitivity

SIM ExoplanetsandOtherAstrophysics 0.5Earths 120

SIM‐Lite ExoplanetsandOtherAstrophysics 0.7Earths 85

PlanetHunter ExoplanetsOnly 0.7Earths 85

ScienceGoals

SupportforMicroarcsecondAstrometryScienceOtherThanExoplanetsApotentiallysurprisingresultofaninternalsurveyofcommitteememberswasunanimoussupport for the nonExoplanet related science that an astrometric mission will address,despite the fact that all the scientists being surveyed work specifically in the field ofexoplanets. Committeemembers found that thiswas important both formaximizing thescienceoutputandtomakeamajormissionattractivetotheastronomicalcommunityasawhole. There has been some misperception that the scope of an astrometry mission islimitedtoexoplanetsorafewothernichetopics.Thecommitteewantstoemphasizethatamicro‐arcsecondastrometrymission isageneralpurposeobservatory thatwillopenupanewdiscoveryspaceforabroadrangeoftopicsinastrophysics.

Astrometry

15

The SIM‐Lite mission is a direct descendent of the Astrometric Interferometry Mission(AIM), the astrophysicsmission endorsedby the1990BahcallReport (NationalResearchCouncil 1991), and re‐confirmed by the 2000 McKee‐Taylor Report (National ResearchCouncil 2001). This mission was originally recommended based on its promise ofaddressingawiderangeofproblemsinstellarandgalacticastrophysics.Planetdetection,in1990,wasposedonlyasaninterestingpossibility,ratherthanamajorsciencedriver.

SIM‐Lite, inadditiontoaddressingtherecommendationsoftheExoPTF, isfullycapableofperformingthesciencelaidoutintheMcKee‐TaylorReport.Indeed,thecriticalimportanceof precision astrometry has not diminished in the intervening years. The expected ESAmission named Gaia will conduct a large astrometric survey, building on the recognizedground‐breaking work of Hipparcos. As a flexibly‐pointed mission capable of very highastrometric precision (4 µas positions) at faintmagnitudes (V < 19), SIM‐Lite occupies averydifferentparameterspace fromGaia. Arecentpeerreviewof theSIMScienceTeamKey Projects (Davidson et al. 2009) showed that with minor exceptions, the broadastrophysicscaseforSIMremainsunsurpassedbyanycurrentorplannedfuturemission.

The science case for SIMPlanetQuest is laidout in considerabledetail in the SIMScienceTeam’s paper (Unwin et al. 2008). SIM‐Lite replaces SIM PlanetQuest as a cost‐effectivewaytodoprecisionastrophysics.IthasessentiallythesameaccuracyasSIMPlanetQuest,but can observe approximately half the number of sources in the faint‐source limit (thisdoesnotapplytotheexoplanetmission,wherethestarsareallbright).Inmostcases,theScienceTeam isable todo the samesciencebutwitha judicious re‐allocationofplannedobservingtime.

FormationandMassDistributionoftheMilkyWay

Tidal tails of dwarf spheroidal galaxies form an exquisitely sensitive tracer of theinteractionhistoryoftheGalaxy,andalsothedetailedshapeofitspotential,butonlyifonehas good3‐space positions and3‐space velocities. SIM‐Lite uniquely provides 2 of the 3spacevelocitycomponentsindistanttails(radii>20kpc).Thistechniquecanbeextendedto the Local Group (out to ~5 Mpc) through astrometry of the brightest supergiants inmember galaxies, providing key data on themass distribution and dynamics of the LocalGroup.

StellarAstrophysics

Themassesofstarsateachendofthemainsequencearenotwellcharacterized. Manyoftheestimatedstellarmasseshave10%errors,whichis insufficienttochallengemodelsofstellar luminosity and evolution. Many ‘exotic’ stars (such as the components of X‐raybinaries) have distances, masses and luminosities crudely estimated to a factor of two.Thesearesimple,butabsolutelyfundamental,measurementsthatSIM‐Litecaneasilymake.

QuasarAstrophysics

SIM‐Litewillprobetheinternalstructureofquasarsatopticalwavelengthstocomplementthehighest‐resolutionVLBIimaging.Thirtyyearsafterthediscoveryofsuperluminalradiosources, many fundamental questions about the formation and propagation of jets stillremain unanswered. Precision optical astrometry, although not able to directly resolve aquasarcore,willsensevariabilityinawavebandordersofmagnitudeawayfromtheradio,and probe a regime where accelerated‐particle lifetimes are very short. Variability anddifferential color shift across theopticalbandarevectorquantities that canbe comparedwithothergeometricandstructuralinformationinquasars.

Chapter2

16

10!2 10!1 100 101 10210!1

100

101

102

103

104

V E

M

J

S

U N

Orbit period (years)

Plan

et m

ass (

M Earth

)

Planets from Ida & Lin simulationsKnown RV!discovered planetsSolar system planets

Figure21.Plotof theexpectedsensitivityofdifferentplanetdetection techniquesasa functionofplanetarymass in Earth‐masses with the period in years. The grey curve represents radial velocity programs with ameasurementaccuracyof1ms–1.ThebrowncurveistheaveragesensitivityoftheGaiamissionat70µas.Thesolid red curve represents the median target sensitivity of the SIM‐Lite mission assuming a sample of 60optimized targets. The blue dots show the properties of our Solar System planets, the purple dots are theproperties of known RV‐discovered planets and the yellow dots are the planets produced through the coreaccretionmodelsofIda&Lin(2004a,2004b,2005).(J.Catanzarite,JPL)

DetectingandCharacterizingLowMassPlanetsThestudyofexoplanetsisdriveninpartbyourdesiretobetterunderstandtheoriginsoflife and our place in the universe. Is the phenomenon of life unique to Earth, or is itcommonplace?Lifeasweknowitrequiresafewcrucialingredientsthatonewouldexpectto findonhabitable exoplanets, foremost beingwater that remains liquid over geologicaltimescales.Onlyasmallrangeofstar‐planetseparationsallowswatertoremainliquid,andasearchforhabitableplanetsrequirestheabilitytofindplanetsinthatregion(near1AUfor Sun‐like stars). While some exoplanets have been found within the habitable zonesaround other stars, they have all been gas giants like Jupiter; current techniques beinginsensitive to the smaller signatures that result from smaller Earth‐like planets. Aftercandidateplanetsarefound,thenextlogicalstepwouldbetoconfirmwhetherwater(andotherlife‐relatedchemical)existonthoseplanets.Thiswillrequireisolatingthespectrumoftheplanet’slightfromthatofthebrightstaritorbits.OnlyforthecloseststarstotheSunwill this be feasible; thus, it is reasonable to prioritize our search for exoplanets to theneareststars.Knowingaplanet’smasswillbeimportanttoreliablyinterpretitsspectrum.Thus, the path to identifying candidate habitable planets requires a new observatory

Astrometry

17

capableofaddressingfourrequirements:itmust(1)searchthestarsclosesttotheSunwithenoughsensitivityto(2)detectEarth‐sizedplanetsat(3)star‐planetseparationsthatallowfor liquidwater and (4)measure theirmasses. Themicro‐arcsecond astrometrymissionmeetseachoftheserequirements.

As is shown in Scargle (1982), to achieve a false alarmprobability of about 1%,within afactoroftwo,foranumberofmeasurementsthatrangesfromtenstothousands,thesignalto noise ratio (SNR) needed is about 5.5 to 6.0, depending slightly on the number ofmeasurements. If σ is the one‐axis RMS noise per differential measurement, N is thenumberofvisits,andα is theamplitudeof theastrometricsignature thatcanbedetectedwithaprobabilityof50%,then

α=SNR×σ/N1/2orN=(SNR×σ/α)2.

Forexample,theEarth’sastrometricsignalamplitudeat10pcisα=0.3µas. ForanRMSmeasurementuncertaintyofσ=1.4µas,N =659measurements areneeded todetect anEarthwithanSNR=5.5. A targetedmission is therefrorerequiredwithmicro‐arcsecondsingle‐measurement precisions, a noise floor below 0.3µas, and the ability to target theclosestandbrighteststars.

Precisionradialvelocitymonitoringofthemostpromisingcandidatestars,beginningnowandextending through theendof theastrometricmission,willmaximizemission successandofferamorecompletedescriptionoftheplanetarysystemswewillexplore.

OperateasaPrecurserMissionforaFutureDirectImagingProgramRecent technological advancementshave ledus to thepointwheremankindmight finallyanswer the ancient question “Arewe alone?” by searching for evidence of life on planetsaroundnearbystars.Terrestrialplanetsof0.3–10M⊕wouldprovidethebestconditionsforhabitability,asheavierplanetscouldhaveunfavorablythickatmospheres,andlighteronesmightnothaveanyatmosphereatall.Orbitsintherange0.85–1.6AUaroundastarofsolarluminosity would offer planetary temperatures compatible with liquid water. For otherstellar luminosities, the “habitable zone” orbits scale roughly as the square root of theluminosity.Mdwarfsarenotfavorabletargets,becausetheirhabitablezonesaretooclosetothestarsfordirect‐imagingmissionsofreasonablesizes,andtheirplanetsmaybetidallylocked,reducingtheirlikelyhabitability. NearbySun‐likestarsarepreferabletargets,andalsomoreamenabletostudybyfollow‐ondirect‐imagingmissions.

Thefirststepistolocatesuitablecandidateplanets.AmicroarcsecondastrometryfacilityisthemostfeasiblemethodforidentifyingcandidateEarthlikeplanetsaroundnearbySunlikestars inthe2010–2020decade. Radialvelocitysurveysare limitedbythe intrinsicsurfacejitterofstarsthatmasksthetinyradialvelocitysignature(~10cms–1)ofahabitableEarth‐likeplanet. Transitsurveysareonlysensitivetoplanetswhoseorbitsareedge‐onasseenfromEarth,andthuswilltransittheirstar;thelikelihoodofwhichisverysmall(<1%)forplanets in the habitable zones of Sun‐like stars, and so many nearby planetary systemswould not be fully explored. Direct‐imaging missions are observatories to measure thespectra of exoplanets, and it would be inefficient to devote a large fraction of theirobserving time solely to finding planets. Other planet detection methods are less welloptimizedforstudyingnearbystars.

The second step is toobtain spectraof theplanetaryatmospheres, looking for signaturesrelatedtolifeonEarth,suchaswaterandoxygen.Thegreatestchallengeforthesuccessofsuch amission is the combinationof the extreme contrast between star andplanet, their

Chapter2

18

small angular separation, and the intrinsic faintness of the planet. Themost reasonabletargets for a direct‐imaging mission are nearby stars where this challenge is minimized(thoughstillquiteformidable).Astrometryiswelloptimizedforidentifyinglowmassplanetsaroundtheneareststars,thebesttargetsfordirectimagingmissions.

The ExoPTF report states on page 55 “The most promising way to mitigate the cost ofspace‐based direct imaging is 1) to identify targets before the direct‐imaging mission isflown.” The astrometry committee agrees with this assessment. By first deploying anastrometricmissiontofindandmeasuretheorbitsandmassesofEarth‐likeplanetsaroundnearbystars,thecostofafuturedirect‐imagingmissioncanbeminimizedandthescientificoutputgreatlyincreased.

Wheretolook:Identifynearbytargets

The best way to search for potentially habitable exoplanets around nearby stars isindirectly—bylookingfortheastrometricwobbleofthestarrespondingtothegravitationaltugoftheorbitingplanet.Theminimumrequiredsingle‐measurementastrometricaccuracyis~1µas.

The alternative—a direct search by coronographic images—is problematic. The firstproblem is the centralobscuration,whichhides large fractionsofmostorbits. Only forafractionof theplanet's orbit is it correctlypositionedwhere a coronographwill detect it.Thesecondproblemisthelowinformationrate,demandingmanylongexposureswithrarepositiveresults,assumingaplanetispresent.Thethirdproblemisthedifficulty—orevenimpossibility—of estimating theplanetary orbit from the small number ofmeasurementsthat canbe obtained in the epochof discovery, typically only6months longdue to solaravoidance. Without knowing the planetary orbit, science operations become amatter ofguesswork,andanyplanetfoundwillprobablybelost.

The ExoPTF recognized these difficulties with the direct planetary search when itrecommended splitting the finding and characterizing tasks between two concatenatedprograms, an indirect finder program and a direct characterizer program. The direct‐imagingmissionwillcharacterizeplanetsfoundbythefinderprogram.Whiletheindirectfinder program can stand alone and still provide valuable science content without thedirect‐imaging program, the direct‐imagingmissionwill not have identified targets if theindirectfinderisnotdeployedfirst.

Whentolook:CharacterizingOrbits

Astrometric observations of stars hosting Earth‐like planets allow the orbits to bedetermined.Thepositionoftheplanetcanbepredictedasafunctionoftime.Thisprovidesasolidbasisforfutureplanningdirect‐imagingprograms,astheobservingschedulecanbesetto lookattargetswhenthestarandplanetaremostoptimallyconfiguredfor isolatinglightfromtheplanet.

The direct‐imaging program will operate more efficiently by only observing when theplanet‐star configuration is likely to produce a positive detection and spectrum of theplanet.Notimeiswastedsearchingforatargetwhentheorbitalconfigurationispoor.Thetechnical requirements of a direct‐imagingmissionmay be reduced because it need onlymeetrequirementsfordetectingplanetswhentheyareinoptimalconfigurations.

Astrometry

19

CharacterizePlanetsTargetedByImaging:Masses

On page 55, the ExoPTF report finds “a capability to determinemass, specifically space‐based astrometry described above, is necessary in conjunction with the direct imagingplatform.” The astrometry committee agrees that without this capability, the sciencereturnsofadirect‐imagingmissionaremarginalized.

Amassestimatederived fromanorbital solutionofastrometricdata fromanastrometric“finder” mission would be the most basic quantity in our understanding a planet'scharacteristics. Knowledge of a planet's mass provides the context for interpreting anyspectrumofaplanetaryatmospheremeasuredbyadirect‐imagingmission,andsowouldbe crucial to determining whether a planet can support life. Only by measuring thegravitational effect of a planet on its host star can the planetmass bemeasured. Directimagingandtransit light‐curveobservablesdonotprovidemeasurementsofplanetmass,and model‐based estimates are a poor substitute, because they require that thecharacteristicsofexoplanetsfollowthepatternsseeninourownSolarSystem.

MultiplanetSystemsandPlanetarySystemArchitectureTheoreticalmodelsofplanetarysystemformationinevitablyproducesystemswithmultipleplanets.CombinedwiththeDopplerdiscoveryofexoplanetsystemswithatleastfour(HD160691)orfive(55Cnc)planets,andtheeightplanetsinourSolarSystem,it isclearthatweshouldbeprepared todiscover thatmultiple‐planetsystemsaremore likely tobe therulethantheexception. WhilethenumberofgasgiantplanetsaroundF,G,andKstarsisbeginningtobewellunderstoodfromDopplerdetections(Cummingetal.2008),thereareonlyweakerconstraintsonNeptune‐massplanets,andnoconstraintsatallonEarth‐massplanets.UsingourSolarSystemasaguide,oneexpectsthatlong‐periodJupitersshouldbeaccompaniedbyinnerterrestrialplanets,andithasevenbeensuggestedthatshort‐period,hotJupitersmightbeorbitedbyhabitableterrestrialplanets,inspiteofthepriormigrationoftheJupiter‐massplanetthroughthehabitablezoneofthestar(e.g.,Raymondetal.2006).The basic theoretical expectation is that terrestrial planets should be commonplace,regardlessofwhetherornotgasgiantplanetshavehadachancetoform(Wetherill1996).Ifmultipleplanetsareknowntoorbitagivenstar,wecanplacestrongconstraintsnotonlyontheorbitalstabilityoftheentiresystem,butalsouponthemechanismsinvolvedintheplanet’s formation and orbital evolution, and so discovering and characterizingmultiple‐planetsystemsareimportantgoalsforexoplanetsearches.

Astrometricobservationsprovidetheopportunitytomeasuresixphase‐spacecoordinatesfor each planet detected. In particular, by measuring the orbit's inclination, astrometryestimates the planet‐star mass ratio with no degeneracies (except for pathologicalgeometries). The precise measurements of planet mass is key to understanding thedynamical state of multiple‐planet systems. Therefore, intensive monitoring of multiple‐planetsystemsshouldbeoneofthekeysciencedriversforanastrometricmission.

If we have only radial velocity data, we can nonetheless, by assuming the system isdynamicallystable,constraintheinclinationstoprovideupperlimitsfortheplanetmasses.Unfortunately, suchanalysesoften leaveuncertaintiesoforder~30degrees in inclinationand a factor of ~2 in the planetmasses. These uncertainties can cause even qualitativeuncertaintiesinthedynamicalstateofthesystem(e.g.,modeofsecularevolution,beinginornearamean‐motionresonance).Withsufficientobservations,astrometricobservationscanresolvesuchdegeneracies,sothatdynamicalmodelingcanconstraintheformationandorbital evolution of multiple‐planet systems. For example, different mechanisms foreccentricity excitation make different predictions for the degree and circumstances of

Chapter2

20

inclinationexcitation.Therefore,dynamicistslookforwardtoastrometricmeasurementsofthe relative inclination between planets and whether this correlates with planet mass,orbitalperiod,eccentricity,and/orpresenceofabinarycompanion.

Both the number and fraction of planets inmultiple‐planet systems is set to increase asplanetsearchesbecomesensitivetoplanetswithlowermassesandlongerorbitalperiods.Sincemany,ifnotmost,planetsaremembersofmultiple‐planetsystems,understandingtheformation and evolution ofmultiple‐planet systems is essential for understanding planetformationingeneral.Withsufficientastrometricobservationsofamultiple‐planetsystem,the current planet masses and orbits can be precisely measured, so that theorists caninvestigate the secular orbital evolution and gain insights into planet formation. Forexample, different models of planet migration and eccentricity excitation make differentpredictionsforthesecularevolutionofplanetaryeccentricitiesandinclinations.Therefore,theorists look forward to astrometric observations that determine if the mode andamplitude of secular inclination evolution correlates with the mode and amplitude ofseculareccentricityevolution. Searchingforsuchacorrelationcouldtestwhethercertainfeatures in the secular eccentricity evolution (previously identified by radial velocityobservations) are reliable fingerprints for recognizing the outcomes of various planet‐formationmodels.

A remarkable feature of our Solar System is its coplanarity. Thanks to the recentredefinitionof“planet,”allplanetsintheSolarSystemarecoplanartowithin7degrees.Ifthisisageneralfeatureinmostplanetarysystems,itisastrongobservableconstraintthatmustbematchedbyplanetary‐systemformationandevolutionmodels.However,withthecoplanarityofonlyonesystem(ours)beingknown,itisimpossibletotellhowcommonthisfeatureisinnature.

Studyingadditionalmulti‐componentSolarSystemswillservetoexpandourunderstandingofwhat a typical Solar Systemgeometrymight look like, andwhether the distribution ofgeometriesevenallows for “typical” tobedefined. It is interesting tonote thatwhile thedistribution of relative inclinations of triple stars systems is slightly biased towardcoplanarity,thesystemstendnottobeaswellalignedastheplanetsinourSolarSystem.

Flexible scheduling can greatly aid the characterization of multiple‐planet systems. Forinstance,aSolarSystem“clone”wouldberecognizedasespeciallyinteresting,butcomplex,after about two years. Such as system would easily warrant an increase in observingcadenceduringyears3–5ofamission.

MicroarcsecondAstrometryintheContextofOtherExoplanetProgramsWhile the astrometry committee strongly supports a direct‐imaging mission aftercompletionofamicro‐arcsecondastrometricfinderprogram,ifitcomestochoosingoneorthe other, an astrometric mission has a higher priority than an imaging mission. Thecommitteefavorsaplanofdeployinganastrometricmissionintheearlypartofthe2010–2020decade,withadirect‐imagingmissionbedeployedattheendofthedecadeorearlyinthe2020–2030timeframe.

Theimagingmissionsciencewouldbecompromisedwithoutanastrometricmission.First,one doesn't know where or when to look. An imaging mission is not a viable “finder”mission;toincorporatethat,theimagingmissionwouldbemuchmoreexpensiveandlessefficient. Second, from imaging data alone, masses are unknown, and it is difficult tointerpretimagingresultswithoutthis.Third,astrometryisreadynow.Thearchitectureissettledandtechnologymilestonescompleted.

Astrometry

21

ScienceRequirements

PlanetFindingTheExoplanetTaskForcereportrecommendedanastrometricmissionwiththecapabilityto achieve a mission minimum detectable astrometric signature of 0.2 µas after manyhundredsofmeasurements, derived from thedesire todetect a one‐Earth‐massplanet atthe inner edge of the habitable zone for a Sun‐like star located at 10 pc, and having athroughput sufficient to survey 60 to 100 nearby stars to this depth during themissionlifetime in order to generate a sufficient number of candidate planets for a later direct‐imagingmission.

These recommendations can bemet by an astrometricmission capable of carrying out aroughly five‐year mission while achieving the recommended astrometric precision fortargetstarsfrom–1.4Vmagthrough7Vmagthateachhavereferencestars(typicallyKgiantsat~1kpc)brighterthan~10Vmaglocatedwithintwodegreesofthetargetstaronthesky.

Wenotethattheactualprogramundertakenbyanastrometricmissionwouldofcoursebeinformed by all prior knowledge, most notably the frequency of terrestrial planetsdiscoveredbyCoRoTandKepler.Aflexiblypointedinstrumentcanbeadaptedtomaximizethesciencereturn.

OtherAreasofAstrophysicsOne of the more general purpose versions of the micro‐arcsecond astrometry missionguaranteesthebestsciencereturnfortheinvestment.Anastrometricmissiononlycapableof meeting the recommendations for planet finding would not be capable of achievingsignificantstridesinotherareasofastrophysicsasrecommendedbythetwopreviousNRCastrophysicsDecadalSurveys,primarilybecausetheinterestingastrometricscienceliesinthe dim‐star regime out to 19 Vmag, which cannot be reached by other ground or spacemissions. Achieving the astrophysics recommendations of the prior Decadal surveysrequires the stability to supportmuch longer integration times than required to achieveonlyplanetfinding.

2.1.2 StronglySupportedRecommendations

ScienceGoals

TheoryAs noted in the Report of the ExoPTF, there is a clear need to combine observationaladvances in detecting and characterizing exoplanets with theoretical work that supportsthesediscoveriesinseveralways. Theastrometrycommitteefocusedontheoreticalworkofmostimportanceforastrometricplanetdiscoveries,asopposedto,e.g.,directdetection,wherethetheoryofexoplanetaryatmospheresbecomesimportant.

First, theoreticalworkonorbitaldynamics isessential fordeterminingthebest fit for theexoplanetorbits,determiningorbitalstabilityoverlongtimeperiods,understandingorbitalresonances and how they formed, and determining the interactions in multiple‐planetsystemswheresomeplanetaryorbitsarepoorlyconstrainedorcompletelyunknown.

Second, theoretical work on planet formation mechanisms across the entire range ofplanets,fromterrestrialplanets,toicegiants,togasgiants,isnecessaryinordertoplacethe

Chapter2

22

exoplanet discoveries in the context of planetary system formation theories. Terrestrialplanetformationisstronglyinfluencedbygasgiantplanetformationandorbitalevolution,soacompletetheoryofplanetformationisneededinordertounderstandtheformationofany one component. Theoretical models provide testable predictions of what might bediscovered,andthereforehelptodefinethenextstepsinalong‐termprogramofexoplanetdiscoveryandexploration.

Third, understanding the extent of habitable zones around astrometric target stars willfocus our attention on the best targets to search for the possible detection of habitableworlds. Recent theoreticalwork, e.g., hashighlighted thepotential of low‐mass,Mdwarfstarsashostsforhabitableworlds(Tarteretal.2007).

TheImpactofGaiaOnpage54oftheExoPTFreport,theExoPTFfinds“TheEuropeanGaiaspacemissionwillbeausefuldemonstratorof theability todospaceborneastrometry to findgiantplanets,and will contribute to the census of Jovian‐mass planets around Sun‐like stars.” ThecommitteeagreesGaia isanexcitingprogram,butfindsthat itsexistencehaslittle impacton setting priorities for other astrometry programs because it operates in acomplementary—ratherthanduplicate—capacitytothoseprograms.

Gaia is a European astrometry mission with planned launch for 2011. It will achieve100µasper‐measurementprecisions(toV=15)on30millionstars(andreducedprecisionforup to abillion fainter stars),with each star revisited1–250 times (typically~90one‐dimensionalmeasurements)overthecourseofthemission.Itisasurveymissionwithoutcapability to be pointed at a particular object,meaning the number of revisits cannot beincreased for a high priority target. Gaia will saturate for very bright stars (V ~ 6),including the stars closest to Earth that are the highest priority targetswhen looking forEarth‐likeplanets.

ThemeasurementprecisionofGaiaisinsufficienttodiscoverEarthlikeplanets;the100timesbetterprecisionofamicro‐arcsecondastrometryprogramisrequiredtoachievethisgoal.Furthermore,thenumberofrepeatedmeasurementsforagiventargetismuchlargerforapointedmission (with perhaps 200–500 two‐dimensionalmeasurements onhighprioritytargets),whichwillbothimprovethesensitivitytolowermassplanets,andprovidebettercoverageofaplanet'sorbit.Goodorbitalcoverageisnecessarytoadequatelycharacterizeaplanet's orbit well enough to predict when the best observing timewill occur for futuremissions.

DetectingLargeNumbersofGiantPlanetsThediscoveryof the first giantexoplanet (Mayor&Queloz1995)occurredonly13yearsago. This discovery opened up the field of exoplanet research. Currently more than 300giant exoplanets have been identified around nearby stars. These have provided initialstatisticsonthepropertiesofgiantplanets.Inordertomakeasubstantialimpactonthesestatistical studies, future efforts will require detection of an order of magnitude moreobjects.

A sub‐milliarcsecond astrometric program can discover and make dynamical massmeasurements forhundredsofgiantexoplanetsdownto themassof Jupiters.Thiswouldsignificantlysupplementtheresultsfromthepastyearsofradialvelocitysearches,transitsearches, andmicrolensing searches. It isworthwhile to greatly increase the numbers ofplanetstoenableclassstudieswithenoughplanetstocometoconclusionsaboutphysical

Astrometry

23

properties and frequencies in eachof thedifferent categories, e.g., spectral type, age, andmetallicity.

MIDEXorDiscoveryClassSpaceSurvey

An astrometric space‐based system that is in the Mid‐sized Explorer (MIDEX) class orDiscovery class can make a significant contribution in the area of giant exoplanets.CompanionsdowntoNeptunemasscanbediscoveredaroundthousandsofstarswiththeadded precision and stability of space observations (Pravdo et al. 2007). An infraredmissioninspacecouldforthefirsttimedoabroadsurveyofexoplanetsaroundlow‐massandyoungstars. ItwouldbecomplementarytoGaiaboth in low‐massstars thatGaiahasdifficultyobserving,andforbrightstarsthatcouldbeobservedwithahigh‐dynamicrangeinstrumentbutnotGaia.

An example of this class of mission is Giant Planets around M, L, and T Dwarfs in theInfrared(GIMLI).GIMLItargetslow‐masssystemsandcanthushelpsettlethedebateoverthedominantformationmechanismforextrasolargiantplanets(EGPs)(e.g.,coreaccretionor disk instability), illuminate the differences between brown dwarfs and high‐massplanets,andprovideacalibrationofthemass‐luminosityrelationshipforthelowerendofthe stellarmain sequence.GIMLI features a1.4‐maperturewith ahigh‐dynamic range IRinstrument that performs narrow‐angle astrometry with 50‐µas precision. It can alsoaccommodate a complementary coronagraph. It is sensitive to exoplanetmassesdown to<0.01Jupiters,i.e.,Uranus‐orNeptune‐massesfornearbyolderstarsandforyoungerstarsinthenearbystar‐formationregions.

GroundBasedOpticalAstrometry

Single‐telescopeground‐basedastrometryhashadsuccessindiscoveringstellarandbrowndwarfcompanionsandmeasuringtheirdynamicalmasses(e.g.,Pravdoetal.2004;Pravdo,Shaklan&Lloyd2005;Pravdoetal.2006),butnoexoplanethasyetbeenastrometricallydiscovered.Thereasonsforthisaresimple: lackofsupportforobservingtimeandlackoffunding for instruments. These indirect observations have demonstrated the requiredsensitivity todetect largeexoplanetsbutneedaminimumofobserving time foradequatetime‐sampling.Amodestfraction,e.g.,10%,ofthetimecurrentlygrantedtoradial‐velocity(RV)observationswouldresultinthefirstastrometricdiscoveriesofexoplanets.

Dynamicalmassmeasurementscouldbemadeof>50of thecurrentlyknownexoplanetswith ambiguous mass measurements (due to unknown inclination angles) with a newground‐based system featuring a detector with high dynamic range. Such an instrumentcould use current technology either in the visible or infrared. An infrared systemwouldhave the added advantage of opening a new region of discovery space, viz. exoplanetsaroundlow‐massstars.Sincelow‐massstarscomprise70%ofallstars,thisactivitywouldbe an important component of the desired census of exoplanets. The discovery ofexoplanets around low‐mass stars is inadequately addressed by current programs. RVprogramshavedemonstratedpoorsensitivity to long‐period(>1yr)planetsaround low‐mass stars, and the detection of such planets with other techniques, e.g., transits andmicrolensing, has very low efficiency. An infrared camera built with existing technology,operatingwith adequateobserving timeon a large‐aperture ground telescope, candetectplanetsaroundmanylow‐massstars,greatlyincreasingthe~10systemsthatarecurrentlyknown.

In a recentpaper (Cameron,Britton,&Kulkarni 2009) itwasdemonstrated that100‐µasastrometry canbeachieved in few‐minuteobservationsusing thePalomar200"PALMAO

Chapter2

24

adaptiveoptics(AO)systeminK‐band.ThepaperarticulatestheproblemofdifferentialAOastrometry in the face of the dominant noise source (which is correlated tiltanisoplanatism),derivesitsexpectedcontributiontotheastrometricuncertainty,developsan optimal estimation algorithm for performing astrometry, and verifies the expectationswith extensive on‐sky tests at Palomar and Keck. The technique achieves ~100‐µasprecision in 2 minutes and ~100‐µas repeatability over 2 months. The technique iscurrentlybeingusedatPalomartosearchforplanetsaroundmid‐M‐dwarfs.

The European Very Large Telescope Interferometer (VLTI) began commissioning thePRIMA narrow‐angle differential astrometry instrument in August 2008. The mostoptimistic estimates predict 10–20‐µas performance. This will be an excellent tool forcharacterizing giant planets around nearby stars, and the expected long lifetime of thisprogramwillallowittocapitalizeonastrometry'sincreasingsensitivitywithplanetorbitalperiod. The committee agrees with the ExoPTF's conclusion that, for ground‐basedobservatories,thereare“nootherfacilitiesthatwouldmatchorexceedtheperformanceofVLTIwithinthe15‐yeartimeframe”(page101).TheVLTIwilloperateinthenear‐IR,andcan study objects obscured at visible wavelengths including protostars and the galacticcenter.

The two 10‐meter Keck Telescopes have been combined as an interferometer. ProjectASTRA, funded by the National Science Foundation (NSF), is building a differentialastrometry module for the Keck Interferometer. The resulting astrometric precision isexpected to be on the level of hundreds ofmicro‐arcseconds. The small amount of timeavailable for interferometry on these large telescopes limits its impact to very specificobservingprograms,suchasRV‐identifiedmultiplanetsystemsorthegalacticcenter.

LiftingtheMass/InclinationAmbiguityforRVIdentifiedExoplanetsThelargemajorityofknownexoplanetshaveambiguousmasses:theirunknowninclinationanglesconstraintheresultsto lowermass limitsonly. Thismassambiguitydisappears inthe astrometric determinations of dynamical mass, because astrometry determines thesystem inclination angles. Specific targets of interest where unambiguous masses areimportant include those with multiple planets where planet‐planet interactions mightbecome significant. The inclination ambiguity does not strongly affect statisticaldistributionsofplanetmasses.However,forsamplesofalargenumberofplanetsitisalsoimportant to know where each of the systems lies in the parameter spaces defined bytheorists. Ida et al. (2004a, see Figures 9 and 12) demonstrate how knowledge ofunambiguous dynamical masses can guide development of theory. The statisticalassignment of inclination angles to many systems degrades the quality of the solutions,especially when the number of planets is limited to a small number of planets in eachcategoryofthemultidimensional(spectraltype,age,metallicity,etc.)parameterspace.

2.1.3 RecommendationswithMixedorAmbivalentSupport

AstrometryatRadioWavelengthsMoststarsarefaintatradiowavelengths.Afewactivestarscanbedetectedinnon‐thermalemissionby theVeryLongBaselineArray (VLBA). Operatingat shorterwavelengths, theAtacama Large Millimeter Array (ALMA) will detect thermal emission from many main‐sequence stars. However, its limited resolutionwill restrict astrometricprecisions to the

Astrometry

25

milli‐arcsecondlevel.Thismaybeusefulforthestudyofgiantplanets,butislimitedinitsscope.TheSquareKilometerArray(SKA)willhavethesensitivityrequiredtodetectmain‐sequencestars,anditscurrentinternationalspecificationsincludeverylongbaselinesthatwould enable astrometry at micro‐arcsecond precisions at centimeter wavelengths.However, there is concern that anynon‐thermal emission,whichmaynotbe centeredonthestaritself,willreducetheastrometricprecision,potentiallytothelevelofhundredsofmicro‐arcseconds.

Thecommitteegenerallysupportsradioastrometryasausefulmethodforexoplanetdetectiononly if it requiresamodestmarginal cost to implementat existing facilities. The impactofradioastrometryonthefieldofexoplanetsisunlikelytobesubstantialenoughtoserveasamajor science driver for major new radio facilities. By itself, astrometry for exoplanetstudiesisnotcompellingenoughtorequirethattheSKAdesigncontinuetoincludethelongbaselines that would enable precision astrometry at centimeter wavelengths. Assumingthat the SKA design continues to include long baselines tomeet other science drivers, aprogramforastrometryofnearbystarswouldbeworthwhileforthesmallmarginalcostitwouldrequire.

Thecommitteedoespromoteradioastrometryasapowerfultoolforscienceinfieldsotherthanexoplanetstudies(e.g.,Honmaetal.2007;Hirotaetal.2007),andbelieves itshouldcontinuetobesupportedforthosepurposes.

BasicPropertiesofStarsinSupportofExoplanetScienceOnpage47,theExoPTFreportstates“Tooptimizethechoiceoftargetsandmaximizetheeventual scientific interpretation of exoplanet observations, it is important to have anongoing program to measure stellar parameters.” The fractional uncertainty in anastrometricallyderivedexoplanetmassisatbestequaltothefractionaluncertaintyinthehoststar’smass.Astrometrycancontributetoconstrainingstarmassesforarangeofstartypes bymeasuring the orbits of binary stars. Overall, the committee did not place highpriorityondevelopingnewastrometryprogramsinthisarea.Mostofthecommitteefoundthat these quantities are already known well enough across the range of star typescomprisingtheprimarytargetsofasearchforEarth‐likeplanetsaroundnearbystars,andimprovingthisknowledgewouldhaveaminimalimpactonexoplanetscience. Themeritsof stellar astrophysics can be separately evaluated as a science topic on its own,withoutplayingacrucialroleintheoutcomeoftheexoplanetprojects.

Similarly, the distance to a host star is required to properly characterize it and anyastrometrically detected exoplanets. This will be a natural result of any astrometryprogram and a separate devoted effort is unnecessary. Astrometry may play a role indetermining distances to planet host stars with companions detected in other manners,suchasradialvelocityortransits.

SpaceAstrometrySupportForMicrolensingAsdescribedinChapter6,microlensingeventsareapowerfultoolforstudyingexoplanetsaround distant stars. The observable being used in current studies is the overall sourcemagnificationcausebythegravitationalpotentialofthelensingobject(s).Theshapeofthelight‐curveprovidesinformationaboutthenumberofobjectsinalensingsystem,andtheirrelativemasses.

Chapter2

26

Astrometry isa secondobservable that canprovideadditional informationon the lensingsystem. During a microlensing event, the apparent position of the source will vary byamountsmeasurablebythemicro‐arcsecondastrometrymission.

Thecommitteesupportsthisasaninterestingprojecttopursueifthedesignofthemicro‐arcsecond astrometry supports it. However, themission design should not be altered tospecifically address this program, and it should not be allowed to impact negatively theprimarymissionofdetectingEarth‐likeplanetsaroundnearbystars.

AstrometrywithaLargeAperture,WideFieldCamerainSpaceThe committee considered a concept for image‐plane astrometry using a giant 8‐m (or16‐m) space telescope as an alternative for a sub‐microarcsecond‐capable mission.However, it was found to be less capable in terms of astrometric performance. In thisconcept, the telescopewould have a relatively large field of view (FOV) of 2 arcminutes(requiring16k×16kpixelsfor8m,32k×32kfor16m).Thiswouldrepresentareasonableadvanceintechnologyoverthe<3arcminuteFOVofWideFieldCamera3(WFC3),(1k×1kto2k×2kpixelsdependingonchannel)onthe>3×(6×)smallerHubbleSpaceTelescope(HST). (Notethatthe10‐degreeFOVofKepler isnotdiffractionlimitedandnotapplicableowing to the following considerations; pixelization noise in that case limits astrometricprecision,andthedetector‐relatedsystematicerrorsare largeraswell,whileeventhe95megapixelsintheKeplercameraissmallerthanthe250–1000megapixelsrequiredfortheconceptastrometrymission).

Ifoneconsidersonlythephoton‐noiseanddiffractionlimitedSNRofthebrighttargetstar,an8‐m(16‐m)filled‐aperturetelescopeclearlyappearstooffersuperiorperformancetoa6‐mbaselineinterferometer—bothitsangularresolutionandcollectingareaarelarger.

However, the real error budget contains many more terms, and in neither theinterferometer nor the image plane approach does the photon‐noise limited SNR of thebrighttargetstarmakeasignificantcontribution.Lackinginstrumentalorsystematicnoisesources, the dominant termbecomes the photon‐noise and diffraction‐limited SNR of thereferencestars.

Withsmaller individualapertures,theinterferometricapproachhasamuchlargerfieldofviewthanevena“widefieldcamera”onasinglegianttelescope.Innarrow‐anglemode,aSIM‐basedarchitectureusesreferencestarsdistributedovera2‐degreefield,comparedtothehypothetical2‐arcminutefieldofawidefieldimage‐planecamera.Theareaoverwhichthe interferometer can select targets is 3600× that of thewide field camera. By far, thebrightest reference stars in the field contribute themost SNR to establishing a referencegrid, so the number of reference stars on which astrometry is obtained is relativelynegligible as long as the brightest few stars, contributing > 50% of the reference lightavailable,areobserved.Forthemuchwider‐fieldinterferometer,thereferencestarscanberoughly 3600× brighter. The collecting area for an 8‐m (16‐m) filled aperture is 350×(1400×) thatof the2× 30‐cm interferometer telescopes, and the resolvingpower is1.3×(2.7×)thatofthe6‐mbaselineinterferometer.Overall,theinterferometerSNRonreferencestars(proportionaltoF1/2A1/2/R,forsourcebrightnessF,collectingareaA,andresolutionR)has2.4×theSNRofthe8‐mand0.6theSNRofa16‐mwidefieldcamera.

Again, this in not likely to be the dominant term in the error budget of an image planeastrometrycamera.Instead,detectorsystematicsarelikelytobeadominanterrorsource.Assuming the camera is critically sampled (2 pixels/resolution), the focal planewill have

Astrometry

27

32,000×32,000pixels,aCCDmosaic.ThemostpreciseastrometrywithCCDcamerasusessingleCCDsbecauseasinglepieceofsiliconismuchmorestableoveraperiodofyearsthanamosaicdetector.DevelopingamosaicCCDwhosedimensionalstabilityoverseveralyearsisatthe0.1‐nm(single‐atom)levelisfarbeyondcurrenttechnology.CCDmosaicstabilityaside,astrometryat the0.2‐µas level impliespoint spread function (PSF) fitting towithinonepartin75,000(37,500)ofthe15mas(7.5mas)diffractionlimitofthetelescope.ThisisstillamorethantwoordersofmagnitudemoreprecisePSFfittingthantypicallypossibleinmodern imaging detectors such as those on HST. With current technology, the mostoptimisticestimatesoftheimage‐planeastrometryperformanceis~50µas.

Technology obstacles exist requiring a major effort to overcome current PSF fittinglimitationsinimageplanearrays.Firstthereareinhomogeneitiesbetweendetectorpixels,andwithinthepixelsthemselves,introducingnon‐uniformresponsesacrossapixel,makingfittingbelowaboutahundredthofapixelimpossiblewithcurrentdetectorsandcalibrationmethods. New manufacturing processes would be required to make detectors withuniformity100×betterthanthecurrentstateoftheart,andtocalibratethatresponsivity.Second, thevariations inthedifferentialPSFbetweentargetandreferencemustbestableover the timescales of interest (years). These PSF‐fitting issues are mitigated in theinterferometrycase,wheretheastrometryobservableisadifferentialinterferometricdelaytied directly to a laser reference, and the same detector pixels are used for target andreference. Tobecertain,developingtheabilityto fit to farbelowthediffraction limit isachallengingproposition;thistechnologydevelopmentwaspossibleonlythroughadecade‐long effort culminating with the micro‐arcsecond metrology (MAM) testbed, detailed inSection 2.2.2. It is not an efficient use of current resources to duplicate this technologydevelopment effort for the image‐plane astrometry case when the interferometer‐basedapproachisreadyfordeploymentnow.Thisissupportedbythefactthateveninthelimitof systematic errors being less than that of the reference star SNR, the interferometerapproach offers roughly equal astrometric precision as the much more expensive 8‐m(16‐m)filledapertureconcept.

While thecommittee’s fundamental conclusion is to supportdeploymentofany formofasub‐micro‐arcsecond precision, targeted astrometrymission capable of observing nearbystars,thecommitteebelievesthemostrealisticframeworktoachievethis isonebasedontheSIMtechnologyprogram.

2.2 ObservatoryConcept.While many programs for astrometric detection and characterization of exoplanets aresupported by the committee, only the design of the highest priority, micro‐arcsecondastrometrymissionisconsideredhereindepth.

2.2.1 ArchitectureTheSIMbasedarchitecture ismatureandready foramissionnow,and is themost feasibleapproach tomicroarcsecond astrometry. This architecturemay be realized in the form ofSIM,SIMLite,orPlanetHunter.

A space‐basedmission is themost promisingmethod of achievingmicro‐arcsecond levelastrometricprecisions.Ground‐basedastrometrictechniquesmustlookforreferencestarswithintheisoplanaticpath,typically~30arcsecond.Onlyaveryfewofthenearest"best"candidates for an astrometric search for Earth‐like exoplanets are double stars.Backgroundfieldstarsina30‐arcsecondfieldaretypically18–20mag.Atmosphericnoise

Chapter2

28

andphotonnoiseof the18~20mag reference stars limit anygroundbasedastrometricinstrumentto>10µasin1hrofintegration.Whereasaground‐basedastrometricprogramwill require100hoursof integration to reach the1‐µas level, a space‐basedprogramcanreachthis level ina20minutes. Asurveyofthemostpromising60–100targetswill take~40% of a 5 year space‐basedmission, whereas a similar ground based programwouldtake1500years!

Field‐of‐view considerations favor the SIM architecture over filled‐aperture spacetelescopesforastrometry.SIMusesreferencestarsasfaras1degfromthetarget.Aspaceastrometrictelescopewouldhaveamuchsmallerfieldandconsequentlywouldhavetorelyonmuchfainterreferencestars.WhenthephotonnoiseofthefaintreferencestarsistakenintoaccountSIM'sphotonlimitedaccuracyisequivalenttoafilledaperture10‐mtelescopewitha~10arcminfieldofview.

Thescienceinstrumentconsideredhereisaspatialinterferometer,conceptuallysimilartoAlbertMichelson’s interferometeronthe100‐inchtelescopeatMt.Wilsonin1913,andtoother ground‐based amplitude interferometers since that time. It collects twopatches ofthe incoming wavefront from a star, uses a mirror on a movable delay line to apply asuitable timedelaytooneof thewavefrontpatches,andcombinesthepatchesto formaninterferencefringeinthepupilplane.Thedelayisadjustedtogivethemaximumvisibilityof thewhite‐light interference fringepacket. Theoriginof thedelay line isdefinedas itspositionwhentheaxisoftheinstrumentisperpendiculartothepropagationvectoroftheincomingwavefront,wheretheaxisisthelinebetweenthepivotpointsofthetwocollectingmirrorsattheendsoftheinstrument.Thedistancebetweenthesepivotsisthebaselineofthe instrument. Theanglebetweenthe instrumentaxisandthe lineofsight tothestar isthearccosineofthedelaydividedbythebaseline.

The SIM‐based architecture results from twelve years of coordinated technologydevelopmentandmissiondesign.ThearchitectureconsistsofonescienceMichelsonstellarinterferometer (MSI) capable of micro‐arcsecond fringe determination accuracy, a guidestar tracking system to determine the attitude of the instrument at themicro‐arcsecondlevel relative to chosenguide stars (aboutamillion timesbetter thana typical spacecraftstar tracker), and a metrology system to determine the orientation of the scienceinterferometer baseline relative to the guide star tracking system. Supporting missionelements include: (1) an instrument data system to control the instrument and collectinstrument data for return to the ground for processing, (2) a spacecraft bus to providehousekeepingservices,(3)alaunchvehicletoinjecttheflightsegmentintotheproperorbit,(4)amissionoperationssystemtomonitorandcontroltheflightsegment,and(5)ascienceoperationssystemtoplanobservations,reducethesciencedataanddistributesciencedatatothescienceusers.

Anumberofvariantsof theSIM‐basedarchitecturehavebeenstudiedby theSIMprojectteam,aimedatprovidinga full rangeof science/costperformanceoptions for the sciencecommunityandNASAtochoosefrom.Abriefdescriptionoftheseoptionsfollows.

SIMPlanetQuestThisisthemostcapableoftheoptions,consistingofa9‐mscienceinterferometerbaselineandtwo7.2‐mMSIstoformtheguidestartrackingsystem.ThismissionislaunchedintoanEarth‐trailing Solar orbit (ETSO), drifting away from Earth at about 0.1 AU·yr–1. It islaunchedononeofthelargestevolvedexpandablelaunchvehicles(EELVs)(AtlasV551orDeltaIVHeavy)andtrackedbytheNASADeepSpaceNetwork(DSN). Missionoperations

Astrometry

29

areconductedfromJPLandscienceoperationsfromtheNASAExoplanetScienceInstitute(NExScI)atCaltech.

SIMLiteTherearethreeversionsofthisoptionthathitdifferentcost/performancepoints.Allhavea6‐mscienceinterferometerbaseline,allhaveone4.2‐mMSIthatservesasoneofthetwoguide star trackers, andallhaveone30‐cmdiameter telescope that servesas the second,lower accuracy, guide star tracker. All are launched by an EELV (Atlas V 531 class) intoETSO,aretrackedbytheDSN,andareoperatedfromJPLandCaltech,thesameasthefullSIM PlanetQuest. Option variations result from the size of the science interferometersiderostat(50cmor30cm),andthewayscienceoperationsarecarriedout.

PlanetHunterThere are twoversionsof this optionwithdifferent cost/performancepoints. Bothhaveone6‐msciencebaselinethathasarestrictedfieldofregardofonly4°(vs.15°forSIMandSIM‐Lite), both have one 4.2‐m guide MSI, and both have one 30‐cm guide‐2 telescope.OptionsderivefromwhetherthemissionisflowninETSOasforSIMorina900‐kmSun‐synchronous,terminatorlow‐Earthorbit(LEO).LEOoffersalessexpensivelaunchvehicle,less expensive ground communication system (Fairbanks, AK), and the use of magnetictorquerstodumpmomentumallowinglongermissionlifetime,butitcomesattheexpenseofrequiringalargede‐orbitenginetoclearLEOattheendofmissionlifeandtheneedtospend more than 50% of mission time maneuvering to avoid Earth‐shine entering theinstrument.

2.2.2 PerformanceThe accuracy of an angle measurement is determined by the total number of photo‐electrons collected in the interference fringe pattern (mainly set by the brightness of thestar, thecollectingarea,andthe integrationtime), theshapeof thepattern(mainlysetbythe length of the baseline and the wavelength), and knowledge of the baseline vector(directionin3‐spaceandlength).

The relative position of a target star with respect to its group of reference stars isdetermined during a “visit.” Visits can be either short or long, depending on themeasurementaccuracydesired. A shortvisit isabout2200seconds in length. Ina shortvisit, the target star’s two‐dimensional location in the plane of the sky ismeasuredwithrespect to a local framework defined by the average location of 4 to 5 nearby referencestars.Themeasurementalongeachaxistakeshalfofthetotal,or1100s.

Duringthe1100‐sectimeallocatedtoeachaxisinashortvisit,theanglebetweenthetargetstarandthebaselinevectorismeasuredtoa“single‐staruncertainty”ofabout1.0µasforSIM‐Lite and PlanetHunter (and about 0.7 µas for SIMPlanetQuest). Likewise the anglebetweeneachofthereferencestarsandthebaselinevectorismeasured,andtheresultsarecombinedtogivetheaverageofthereferencestars,toasimilar“single‐staruncertainty”ofabout1.0µas(or0.7µasforSIMPlanetQuest).Theanglebetweenthetargetandreferencegroup is the difference of these angles. The uncertainty is the “differential‐measurementaccuracy,”whichis[(1.0µas)2+(1.0µas)2]1/2=1.4µasalongoneaxis(=0.98µasforSIMPlanetQuest).

Chapter2

30

101 102 103 104 105

10!2

10!1

100

Integration time required for N chops, sec

Stan

dard

dev

iation

of N

!cho

p av

erag

es,

!as

MAM, instrument onlyWhite Noise Expectations

Figure22.Astrometricnoisefloordemonstratedinthelaboratory.(B.Nemati,JPL)

The key to micro‐arcsecond precision lies in the measurement process, which is doublydifferential.Differentialinangle,over1degree,anddifferentialintimeoverachopcycleofabout90s.First‐ordererrorsareentirelyeliminated.

SIM’slaboratorymeasurementshaveshownthatmeasurementsarephoton‐noiselimitedtoatleastx=0.035µas.Upto1600(1890)1.4µas(1µas)canbecombinedwithsquare‐rootscaling laws remaining valid for SIM‐Lite or PlanetHunter (SIM). Thenoise floor iswellbelowthe0.3‐µassignatureofanEartharoundaSun‐likestarat10pc.

The MAM results were obtained in a test environment substantially worse than thatexpectedforSIMonorbit. Subsystemtests,suchasMAM,validateSIMthroughintegratedmodelsthatlinkthem.Thismodellinkagewasitselfdefinedasatechnologymilestoneandwassubjectedtorigorouspeerreview.

Theminimumdetectableplanetary astrometric signature (MDAS) is the instrumentnoisefloor(0.035µas forSIM‐LiteandPlanetHunter,0.021µas forSIMPlanetQuest) times theSNRneeded(typically~5.8).Thisis0.20µasforSIM‐LiteandPlanetHunterandis0.12µasforSIMPlanetQuest.

The minimum detectable Earth‐like planet mass is dependent upon the MDAS, stellardistance, stellar mass and the planet’s orbit. For a one Solar mass star at 10 pc, theminimumdetectablehabitable‐zoneplanetmassdependsuponwhere theplanet is in thehabitablezone.SIM‐LiteandPlanetHunter,candetecta0.44M⊕planetattheouteredgeofthehabitable‐zone(0.29M⊕forSIMPlanetQuest);0.7M⊕ata1‐AUpostion(0.47M⊕forSIMPlanetQuest); and 0.85 M⊕ at a the inner edge of the habitable zone (0.57 M⊕ for SIMPlanetQuest).

Astrometry

31

2.3 TechnologyAsintheprevioussection,onlytechnologyrelatedtothehighestpriority,micro‐arcsecondastrometry mission is considered here in depth, though several other programs aresupportedwithlowerpriority.

2.3.1 PastAccomplishments

BackgroundAlthoughopticalandinfraredinterferometershavebeenwelldevelopedforground‐basedobservations (e.g., NPOI, KI, VLTI, CHARA Array, etc.), these facilities are limited by thespatial coherence length, coherence angle, and coherence time imposed by the Earth’satmosphere.Sincetheearly1990s,theconceptsforground‐demonstratedtechniqueshavebeen developed into space‐capable hardware and software with sufficient reliability toenableanunattendedfive‐toten‐yearspace‐basedmission.

Following the success of the Mark III interferometer on Mount Wilson, a program wasbegun at JPL to identify and develop the necessary technologies to enable space‐basedinterferometrymissions. ThistechnologyprogramwaslaterattachedtotheNASAOriginsProgramofficeatJPLandthentotheSpaceInterferometryMission.

From1996to2005aTechnicalAdvisoryCommittee,chairedbyRobertO’Donnell,workedwith the technology development team to complete the focused SIM technologydevelopmentprogram,andcontinuedwith theSIMproject to reviewthe transitionof thetechnologies into flight‐like hardware. Since 2001, a NASA‐Headquarters commissionedIndependentReviewTeam(renamedtheExternalIndependentReadinessBoard,EIRB),hasmonitoredtechnicalprogressforthemission,andnowalsocontinuesinthisroleastheSIMprojecttransitionsthetechnologyintoflight‐likehardware.

PreviousInterferometricAstrometryProgramsA number of smaller‐scale interferometric astrometry programs have been developed onthegroundandinspacetodemonstratethetechnologiesandscienceforamicro‐arcsecondastrometryprogram.Selectedeffortsaredescribedhere.

TheHubbleSpaceTelescopeFineGuidanceSensors(FGS)While spectacularly successful at locating previously unknown companions to stars, thehigh‐precisionradialvelocitiesprovidedbyDopplerspectroscopyonlyprovidesaminimummass,nottheactualcompanionmass.Whilespectacularlysuccessfulatprovidingparallaxesforthousandsofnearbystars,Hipparcoscouldnotreliablyprovide10%parallaxesforstarsmoredistantthan~100pc.AstrometrywiththeHubbleSpaceTelescopehasprovidedboth.

ThemanyresultsfromtheoriginalHSTFGSastrometryobservationsincludeparallaxesofastrophysicallyinterestingstars(Benedictetal.2000,2002a,2002b,2003;McArthuretal.1999,2001),aparallaxfortheHyades(vanAltenaetal.1997),alinkbetweenquasarsandtheHipparcosreferenceframe(Hemenwayetal.1997),adeterminationoflow‐massbinarystar masses (Franz et al. 1998; Benedict et al. 2001), and searches for Jupiter‐masscompanionstoProximaCenandBarnard’sStar(Benedictetal.1999).SubsequentFGSworkcontributedtothestudyofthelowermainsequencemass‐luminosityrelationship(Henryetal. 1999), the inter‐comparison of dwarf novae (Harrison et al. 2004), and parallaxes ofcataclysmicvariables (Beuermannetal.2003,2004;Roelofsetal.2007)and thePleiades

Chapter2

32

(Soderblom et al. 2005). Most recently themes have been the cosmic distance scale andextrasolar planetary systems. This includes the galactic Cepheid Period‐LuminosityRelationship (PLR) (Benedict et al. 2007) and the determination of extrasolar planetarymasses(Benedictetal.2002c;McArthuretal.2004;Benedictetal.2006;Beanetal.2007).

Regarding the cosmic distance scale, FGS researchers have obtained a parallax for thePleiades that supports the validity of modern stellar interiors modeling. This work hasprovided10%parallaxes for tenGalactic, solar‐metallictyCepheids, resulting inaPeriod‐LuminosityrelationthatprovidesdistancestotheLMCandNGC4258.

Currently,fewerthan10%ofthemorethan200candidateexoplanetsorbitingnearbystarshave precisely determinedmasses. Because themost successful technique for detectingcandidate exoplanets, the radial velocitymethod, suffers from a degeneracy between themassandorbital inclinationformostoftheknownexoplanetcandidates,onlyaminimummass,Msini,isknown.FGShasprovidedmassesforplanetsinthesesystems:GJ876,55Cnc,EpsilonEridani,andHD33636.

Apreliminaryindicationisthat,forthefirsttime,thedegreeofcoplanarityofanextrasolarplanetarysystemassociatedwithanormal,main‐sequencestarwillbeestablished.Beyondthat,HSTFGSdataforsimilarcoplanaritytestsonthemultiplanetsystemsassociatedwithHD128311,HD202206,µAra,andgammaCepisbeingcollected.

ThePalomarTestbedInterferometer(PTI)

HandsonTraining

From 2002–2006 SIM sponsored an interferometry training course for Caltech graduatestudentsandpost‐docs,andJPLengineersandscientists. Thiscourseintroducedstudentstothebasicsof interferometrytheoryandpractice,and includedbothsixhoursof lecturematerial, and a two‐night observing session for students at the Palomar TestbedInterferometer (PTI) (Colavita et al. 1999). The school lectures included bothinterferometrytheoryforbothinterferometricimagingandinterferometricastrometry,aswell as a description of preparing for observing and reducing/interpreting data. Theobserving sessions specifically gave studentshands‐onexperience inpreparingobservingplans, preparing the interferometer for observations (students participated in nightlyalignments),and in thereductionofdata (e.g.,homeworkassignments includingreducingandinterpretingdatatakenthenightbefore).

DemonstrationofDualstarAstrometry

PTIhasdemonstratedanastrometricprecisionof<100µasbetweenthecomponentsofabright visual binary (Lane et al. 2000). PTI is a long‐baseline infrared interferometer atPalomarObservatorydevelopedunderNASAfundingtodemonstratetechnologyforfutureground‐andspace‐basedinterferometers. ThespecificexperimentmentionedaboveusedPTI in dual‐starmode,where the field at each telescopewas separated into two narrowfields containing the two components of the bright visual binary 61 Cyg, which wereseparated in declination by 26 arcsec. The interferometer incorporates two separateinterferometer beam trains, including optical delay lines and fringe trackers. Lasermetrology isused tomeasure thedelayrequired todetect fringes,aswellas to referencethetwobeamtrainstocommonfiducialsateachtelescope.

Astrometry

33

Intheexperiment,halfthelightfromonecomponentofthebinarywascontinuallytrackedby one fringe tracker. The other fringe tracker switched, on a several‐minute timescale,between the two components of thebinary. The key observablewas the laser‐measuredchange in delay between the two components, as corrected by residual fringe trackingerrors.

Theexperimentwasconductedoveraseriesof7nightsin1999. ThefinalresultsyieldedRMSresidualsinthemeasurementofdeclinationof97µas,demonstratingthefeasibilityofhighprecisionnarrow‐angleastrometryfromtheground.

PHASES

ThePalomarHigh‐precisionAstrometricSearchforExoplanetSystems(PHASES)atPTIisasearch for giant planets orbiting closely to either star in binary star systems, withastrometric precisions at the 20‐µas level for arcsecond‐separation binaries. PHASES ismotivated both by a desire to study planets in binaries and as an astrometric programdevelopingtoolsandtechniquesforSIM.

Seven refereed journal articles have appeared based on PHASES measurements withanother inpreparation. PHASESscienceresults includereportingtheobservational limitstoplanetarycompanionsinanumberoftargetbinaries,theprecisionorbitsofbinaries,themeasuredphysicalpropertiesofthecomponentstars,andcharacterizationoftwotripleandtwoquadruplestarsystems,includingestablishingthedegreeofsystemcoplanarityandthephysicalpropertiesofthestars.

While the 20‐µas precision achieved by PHASES represents progress in the field ofinterferometric astrometry, it should be stressed that the PHASES hardware can only beapplied to a very specific class of binary stars and thus represents no competition to themuchmoreversatilemicro‐arcsecondastrometrymissionbeingpromotedfor2010–2020.Technology and analysis efforts are shared between the two, and the micro‐arcsecondastrometrymissionwillbenefitfromtheeffortsofthePHASESproject.

TheSpaceInterferometryTechnologyDevelopmentProgramInpreparationforamicro‐arcsecondastrometrymission,atechnologyprogramdevelopedand tested components based on the goal‐level performance requirements set by thepreviousDecadal Surveys: 4‐µaswide‐anglemission accuracy over the full sky and1‐µasnarrow‐anglesingle‐measurementaccuracyovera1°fieldofregard.Thecomponentswerethen integrated into subsystem‐level testbeds thatwereused toverifywholebranchesoftheerrorbudgettoensurethattherewerenomissingterms.Finally,system‐leveltestbedswere developed to fully demonstrate that the components functioned as expected in asystemwithfullcomplexity.

This progressive flow from components, through subsystems, to system‐level testbedsoccurred for two paths: (1) Real‐time optical‐path‐difference (OPD) nanometer‐levelcontrol and (2)Picometer‐level‐knowledge sensing. Thenanometer control path verifiedthatvibrationsandspacecraftattitude‐control‐inducedmotionscouldberejectedtoalevelthat allowed acceptable interference fringe visibility (requires better than 10 nm opticalpathdifferencestability).Thepicometersensingpathverifiedthatdynamicdisplacementsofopticalelementswithintheinstrumentcouldbemeasuredtoarelativeprecisionofafewpicometersandabsolutedistancesbetweenelementscouldbemeasuredtoanaccuracyofapproximatelyonemicronoverdistancestotenmeters.

Chapter2

34

AcompletetechnologyplanfocusingontheSIMreferencemissionwasfirstformallysignedby theOriginsProgramOffice in January1998andupdated in2003and signedbyNASAHeadquarters. Eight key technology developments from this plan were identified as“Technology Gates” with specific objectives, completion dates, and review requirements(TAC and EIRB) as a means for NASA Headquarters to carefully monitor progress. All 8TechnologyGateshavebeencompletedandmettheirobjectives.TheseeightgatesandtheircurrentstatusarelistedinTable2‐2.

Table22.TechnologygatesoftheSIMreferencemission.

TechnologyGate

Description DueDate

CompleteDate

Performance

1Nextgenerationmetrologybeamlauncherperformanceat100‐pmuncompensatedcyclicerror,20‐pm/mKthermalsensitivity

8/01 8/01 Exceededobjective

2Achieve50‐dBfringemotionattenutationonSTB‐3testbed(demonstratessciencestartracking)

12/01 11/01 Exceededobjective

3 DemonstrateMAMTestbedperformanceof150pmoveritsnarrowfieldofregard 7/02 9/02 Exceeded

objective

4 DemonstrateKiteTestbedperformanceat50‐pmnarrowangle,300‐pmwideangle 7/02 10/02 Exceeded

objectives

5 DemonstrateMAMTestbedperformanceat4000‐pmwideangle 2/03 3/03 Exceeded

objective

6 BenchmarkMAMTestbedperformanceagainstnarrow‐anglegoalof24pm 8/03 9/03 Exceeded

objective

7 BenchmarkMAMTestbedperformanceagainstwide‐anglegoalof280pm

2/04,5/04*

6/04 Metobjective

8DemonstrateSIMinstrumentperformanceviatestbedanchoredpredictsagainstsciencerequirements

4/05 7/05 Metobjective

*NASAHQdirectedascopeincrease(byaddinganumericalgoaltowhathadbeenabenchmarkGate)andprovideda3‐monthextensionwhenperformancefellshort

Each of the eight technology Gates developed specific test, modeling, measurement andsuccesscriteriathatwerereviewedbyandagreedtowiththeTACandEIRBpriortotesting.These formed the basis for the post‐test evaluation to determinewhether or not the testwassuccessfulinmeetingtheobjectivesoftheTechnologyGate.

Numerical modeling was a central part of the SIM technology program. Numericaldiffractionmodelingtoolswereverifiedforpicometeraccuracyoverthewholerange(nearfield, mid field and far field) using a testbed specifically developed at Lockheed Martin,Sunnyvale, CA for enabling specific test case comparison tomodeling predictions. Opto‐mechanical modeling tools were verified at the milli‐Kelvin level, again using specialtestbedsdevelopedatLockheedMartin,Sunnyvale,CA,andat JPLinPasadena,CA. Thesetestbed‐modelcomparisonsshowedexcellentagreement(better thana factorof twooverthe full range of test). This experience, coupled with a similar factor of two or betterperformance on the subsystem and system level technology testbeds, has providedconfidenceinthepredictivepowerofthemodelingtoolsusedfordesignandevaluationoftheSIMflightsystem.

Astrometry

35

Table23.TimelineofSIMengineeringmilestones.

EngineeringMilestone

Description DueDate

CompleteDate

Performance

FormulationPhase

EM‐1ExternalMetrologyBeamLauncherBrassboard(meetQualenvironmentalandallocatedpicometerperformance)

5/31/06 6/5/06 Exceededobjective

EM‐2InternalMetrologyBeamLauncherBrassboard(meetQualenvironmentalandallocatedpicometerperformance)

4/30/06 5/3/06** Exceededobjective

EM‐3MetrologySourceAssemblyValidation(meetQualenvironmentalandallocatedperformance)

6/30/06 6/28/06 Exceededobjective

EM‐4

SpectralCalibrationDevelopmentUnit(SCDU)(demoflight‐traceablefringeerrorcalibrationmethodologyandvalidatemodelofwavelength‐dependentmeasurementerrors)

8/30/07 12/10/07 Metobjectives

EM‐5

InstrumentCommunicationH/W&S/WArchitectureDemo(validateSIM’smulti‐processorcommunicationssystemusingtwobrassboardinstrumentflightcomputers,ringbus,andflightsoftwareversion2.0withspecificS/Wfunctionsaslisted)

4/1/07 3/5/07 Metobjectives

ImplementationPhase

EM‐6EngineeringModelsforMetrologyFiducials(doubleandtriplecornercubesfullymeetingSIMflightrequirements)

9/30/07*

EM‐7

MetrologySourceEngineeringModels(opticalbench;fibersplitters;fiberswitches;fiberdistributionassembly;laserpumpmodule:allfullymeetingSIMflightperformancerequirementspertable)

9/30/08*

EM‐8Instrument/MissionPerformancePrediction(updateTechGate#8usinglatesthardwareresults)

9/30/08*

EM‐9IntegrationofS/CFSWbuild‐1withphase‐1oftheS/Cengineeringmodeltestbed(demonstratesspecificS/Wfunctions)

10/1/08*

ImplementationphaseMilestonesdeferred

indefinatelyalong

withmission

*CompletiondatesdeferredindefinatelyduetoFY07NASAdecisiontodelaySIMindefinitely**ActualsignoffbyNASAHQdelayedunil12/12/06duetorequestsforadditionalthermaltestingbytheTACandEIRBboards

Another outcomeof themodeling effortwas the verificationof the full SIMerrorbudget,which showed consistency of the interplay between the terms in the error budget and,perhaps as importantly, showed that there were no missing terms in the error budgets(whichwouldhaveshownupasun‐modelederrorsinthesubsystemandsystemtestbeds).Thisisveryimportantinreducingtheriskthattherewillbefundamentalsurprisesduringtheflightsystemdesign,development,testandoperations.

Chapter2

36

The technology program was completed in July 2005, and the final closeout report wassignedbyNASAHeadquartersinMarch2006afterextensivereviewanddiscussionwiththeTACandEIRB.DetaileddiscussionoftheSIMtechnologyprogramcanbefoundinSPIEandInternationalAstronauticalCongress(IAC)papersbyLaskin(2006aandb).

Figure23.TherelationshipbetweenSIMFormulationPhaseengineeringmilestones.

EngineeringMilestonesSIM continues to transition technology to flight‐qualifiable hardware through a series ofengineering milestones aimed at building flight‐like hardware that is environmentallytested(whenrequired)andperformancetestedtoverifyitscapabilitytoperformtoflightrequirementallocations.Theseflight‐likehardwareassembliesarecalledbrassboardsanddemonstrate that flight hardware for the SIM mission can be built. Each EngineeringMilestoneissubjecttothesameTAC/EIRBestablishmentoftestandsuccesscriteriapriorto testing, and they are subject, post testing, to detailed reviewby the TAC/EIRB againstthese success criteria, following exactly the sameprocess aswas used for the technologyprogram.

ThenineEngineeringMilestonesthatwereestablishedforSIMarepresentedinTable2‐3,showingthefivethatwillbecompletedduringtheFormulationPhase(PhaseA/B)andthefourthatwouldbecompletedduringtheImplementationPhase(PhaseC/D/E)priortotheCriticalDesignReview(CDR).OftheFormulationPhaseEMs,allfivehavebeensuccessfullycompleted.

Manycomponentshavebeenengineered to flight‐readystatus. Theseare summarized inFig.2‐3.

Astrometry

37

DoubleBlindTestofPlanetFindingCapabilityAt first glance, it should not be difficult to extract the astrometric signal of an Earth‐likeplanetfromthecompositesignalofasystemofplanetsaroundastar.Thisisbecauseeachplanet has its own frequency in time, so in a Fourier analysis of the total signal, thesignatureofanygivenplanetshouldstandoutcomparedtoanyotherplanet. Howeverinrealitythecasemightbenotsosimplebecause,forexample,aplanetinaneccentricorbitwithadominatingsignal(e.g.,aJupiter)mighthaveharmonictermsthatarenotrecognizedassuchbutmightlooklikeaseparateplanet,oralong‐periodplanetobservedoveratimeshorter than that period would have noise generated at many frequencies owing to thedifficultyofdistinguishing apropermotionon the sky fromapart of anorbit. For thesereasons, plus the fact that a mission should always be simulated before it is flown, weinitiated a double‐blind simulation to see how well Earth‐like planets (i.e., terrestrialmasses, habitable‐zone periods) inmulti‐planet systems could be detectedwith SIM‐Lite,with the help of RV measurements. An additional goal was to see what measurementaccuracyofSIM‐LiteisnecessarytoachieveagoalofbeingabletodetectEarth‐likeplanets.

The simulation was organized with four teams of scientists, consisting of the planetmodelers (Team A), the data simulators (Team B), the data analyzers (Team C), and theoverall summarizers (Team D). An independent review/reporting team and a NASAappointedExternalIndependentReadinessBoardparticipatedtooverseetheentireprocessandadviseNASA.

If the reliability of detections is taken to be the ratio of correct detections to the total ofcorrect plus false alarms, then the reliability ranged from about 40% to 100%, with 3groupsbeingover80%(onegroupwasnotabletofinishtheexerciseontime).Inprinciple,thisvalueshouldhavebeenabout99%,ifthefalsealarmratehadtrulybeen1%.Howevertheshortamountoftimefortheexercisemeantthatonlyonegrouphadenoughexperienceto fullyweed out false alarms. For this reason the exercise is being repeatedwith extrastatisticaltestsadded.

The accuracy of results also was close to the theoretically expected value, for the keyparameters of period andmass. We calculated the expected accuracy using aminimum‐varianceboundmethod(Gould2008). Comparingthesubjectivelycorrectanswersto theactual answers, and scaling that offset by the expected value of the offset, we found aroughly Gaussian distribution,with approximately the expected number (68%) of valueslyingintherange(–1,+1),nosignificantoffsetfromzero(i.e.,nobias),andveryfewvaluesin the range between plus or minus 1 and 3. However beyond the 3‐σ point, whereessentiallynopointsshouldfall,wefoundahandfulofcases(14%),andtheseappeartobesituationswhere theexpectederrorwasverysmall (say less than1%inmassorperiod),and the actual error was more than 3 times that value, in other words it was a goodmeasurementbutnotasperfectastheoreticallyexpected.

Inoverallsummary,thisfirstphaseofthesimulationshowedthattheanswerstoourinitialquestionsare(1)yes,Earthscanbedetectedinmulti‐planetsystems,and(2)thesensitivityneededisalmostexactlythepositedsituationofa5‐yearSIM‐Litemission,using40%oftheavailable observing time, with the expected noise level and a 6‐m baseline, plus theadditionalhelp(mostlywith long‐periodplanets) from15yearsofRVobservations. Thisfirstphaseofstudyisbeingfollowedupwithphase2,inwhichalltentativedetectionswillbesubjecttoanadditionalstatisticalF‐testandastabilitytest,aswellasadditionaltimefortheanalysis,beforebeingdeclaredasdetections. Inconclusion,wenotethat theexercisegenerated great enthusiasm among the participants, each of whom put in an enormous

Chapter2

38

amount of personal time to find the planets, andwe learned a number of useful lessonsalongtheway.

2.3.2 FutureMilestonesWithwellover$100Minvestedinthedevelopmentofthistechnology,remainingrisksforauserofthistechnologyaretheusualonesthatoccurduringtheimplementationofanylargesystem(manufacturingerrors,interfacemismatches,etc.).

CompletionofEngineeringMilestones6–9The remaining Engineering Milestones will be completed as the mission moves into itsimplementationphase.

DevelopingHardwareandSoftwareintoFlightQualifiableComponentsEngineeringdevelopmentcontinues for turning thedemonstratedcomponents into flight‐qualifiablehardwareandsoftware.Thisefforthasshownthat,evenforthemostsensitivepicometer‐hardware, these components can be built using conventional flight hardwarefabrication techniques with no degradation in performance from that of the technologydemonstrations.

FlightSoftwareDevelopmentTechnology testbedsdeveloped for SIMhavedemonstrated all key algorithms for controland measurement of the instrument. The fifth engineering milestone demonstrated adistributedcomputingenvironment thatsupports thestrict timingrequirements forhigh‐bandwidthcontrolofhardwarethatisdistributedoveraverylargestructure(forexample,SIMorTPF‐I). Processingof theSIMtestbedmeasurementshavetaught theteamhowtoprocesstheinstrumentoutputtoachievetherequiredmeasurementaccuracyandprecisioninways not anticipated early in the technology program, significantly relaxing hardwarerequirements. These data‐processing lessons‐learned should form the basis for groundprocessingforanyflightinterferometer.

CombiningSciencePayloadandSpacecraftInteractions between the spacecraft bus and the instrument are also well understood.These interactions include vibration suppression (simple two‐stage passive vibrationisolationissufficient),attitudestabilizationforbeamwalksuppression(usingthetwoguideinterferometers as a micro‐arcsecond two‐axis star tracker to control the spacecraftattitude control system is about 106 times more accurate than a typical spacecraft startracker), and torque feed‐forward from the instrument to the spacecraft attitude controlsystem to minimize attitude disturbances resulting from the motion of instrumentsiderostatsanddelaylines.

AnalysisofSIMLite’sAstrometricCapabilitiesA Request for Proposals (RFP) was issued for ~20 separate studies for what can beaccomplishedwithSIM‐Lite’sastrometriccapability,someofwhichareexpectedtoaddressplanet finding capability. The RFP can be found on the nexsci.caltech.edu web site.ProposalsweredueJune13,2008andwillresultinone‐yearcontractawards.

Astrometry

39

CompletionofthePlanetHunterStrategicMissionConceptA“PlanetHunter”AstrophysicsStrategicMissionConceptStudywillbecompletebyMarch2009.ThisstudyincludesindependentcostestimatesofthePlanetHuntermissionconcept.

SelectionofMissionConcepttobeDeployedSIM‐based concepts are expected to be reviewed by the upcoming Astrophysics DecadalSurvey.PlanetHunterandaversionofSIM‐Liteareexpectedtobepresentedforreview.

2.4 Research&AnalysisGoals

2.4.1 Research&Analysisin2010–2020Thetopsciencegoalsforastrometryrelatedtoexoplanetscienceinthe2010–2020decadeare:

• DetectEarth‐likeplanetsinthehabitablezonesofnearbystars.

• Characterize Earth‐like planets in the habitable zones of nearby stars, includingimportantphysicalquantitiessuchasmass.

• CharacterizetheorbitsofEarth‐likeplanetsinthehabitablezonesofnearbystars,inpreparationofadirect‐imagingmission,tomakedirectimagingmoreefficient.

• Characterize exoplanet systems with multiple planets, including planet–planetinteractions and orbital coplanarities, quantities important in understanding theformationandevolutionofplanetarysystems.

Allareaddressedbyanastrometricmissioncapableofsingle‐measurementprecisionsonthescaleofamicro‐arcsecondorbetter.Thecommitteefindsthebestwaytoachievethesegoals is todeployaspace‐basedastrometrymissionbasedon theSIMarchitecture, in theearlypartofthedecade.

Previous decades have been spent analyzing the mission design for a micro‐arcsecondastrometrymission. This coming decade should be spent operating such a program andanalyzingthescienceproductsitreturns.

Theastrometry committee recommendsa timeline thatwillobtainenough results fromamicro‐arcsecond astrometry mission by the end of the decade for analyzing resultsimpactingthedesignoffuturedirect‐imagingprograms.

The astrometry committee strongly supports the non‐exoplanet science programs such amissionwouldenable.

No science analysis effort is complete without the support of dedicated theorists. Theastrometry committee recommends supporting theoretical efforts in planet formation,dynamics,andconditionsforhabitability.

Chapter2

40

2.4.2 AstrometryBeyond2020On page 28, the ExoPTF report finds that for the 11–15 year (2019–2023) timeframe“Assuming the space‐borne astrometricmission described above is fielded in the secondtime epoch, no additionalmajor space‐based astrometric effort is envisioned in this timeframe.”Thecommitteeagreeswiththisconclusion.

Beyond the currently envisioned micro‐arcsecond class astrometry mission, what rolesmightastrometryplayinthefieldofexoplanets?First,longer‐termmonitoringofstarscancapitalizeonastrometry’sincreasingsensitivitywithplanetperiod.Amany‐decadesurveyfromgroundorspace,suchasthatbeinginitiatedattheVLTIwithPRIMA,willcontributetothe understanding of outer planets around other stars. Second, one can consider thefundamental astrophysical limits to theastrometric technique, givenanotherwiseperfectinstrument. Asummaryevaluationincomparisontotheradialvelocitytechniqueisgivenasfollows(Catanzariteetal.2008),inregardstosensitivitytolow‐massplanets.

Fromstudiesofourownstar, aswell asnearbyF,G, andKdwarfs,weexpect the fluxofsolar‐typestarstovaryinacomplexwayovermanytimescales.IntensivestudyoftheSun’sfluxvariabilityfromspaceoverthatpastfewdecadeshasrevealedthatitis(downtohourlytimescales) mostly accounted for by magnetic surface features such as sunspots (darkregions)andfaculae(brightregions).Theevolutionofsunspotsandfaculaeintroducejitterinto the Sun’s centroid and radial velocity. Typical sunspot group lifetimes are about 10days;withinthiscoherencetime, the jitter issystematicandcannotbeaverageddownbymultiplemeasurements. Sunspot jitter canbe consideredas ‘noise’ inastrometric andRVmeasurements,andcanbequantified. UsingadynamicsunspotmodeltomatchtheSun’sfluxvariability ithasbeen found that the intrinsic jitterof theSun’s centroid isabout0.7micro‐AUpermeasurementforasignalwithaperiodnearoneyear(suchasEarthina1‐AUorbit),andtheradialvelocityjitterisabout0.6ms–1permeasurement.

TheastrometricjitterinameasurementoftheSun’sangularwobbledependsondistance;at10pc it is~0.07µas.With100measurementsspacedfartherapart thanthecoherencetime,thejitteraveragesdownto0.007µas,afactorof5belowtheinstrumentalsystematicnoise floorof0.035µas,and40 times lower than thesignatureofEarth. Thus, starspots,expectedtobethemajorastrophysicalcontributiontonoiseinastrometricmeasurementsofstars,donotposeasignificantproblemfordetectionofhabitableplanetsaroundsolar‐typestars. Ontheotherhandradialvelocity jitternoiseduetosunspots isabout7 timeshigher than theEarth’sRV signal.AgooddetectionofEarth requires averagingdown thenoise to6 times lower than theRV signal. So sunspot jitter imposesa severe limiton thecapabilityoftheRVmethodtodetectEartharoundastarliketheSun.Evenintheabsenceofotherastrophysicalandinstrumentaleffects,itwouldtake35years(1800measurementsspaced apart by more than a week) to detect Earth via RV. But stellar pulsations andgranulation also introduce RV jitter in measurements of solar‐type stars. The intrinsicastrophysicalnoisefloorforRVmeasurementsduetoalloftheseprocessesisthoughttobeat least 1m s–1. In the long‐term, efforts to improve the astrometric precisions of focalplanearraysmaymakefuturegiantspacetelescopesusefulforprecisionastrometry.

Astrometry

41

2.5 ContributorsMatthewMuterspaugh,UCBerkeleyandTennesseeStateUniversity,Chair

AngelleTanner,JetPropulsionLaboratory,CoChair

G.FritzBenedict,UniversityofTexas,Austin

AlanBoss,CarnegieInstitution

MatthewBoyce,HeliosEnergyPartners

RobertBrown,SpaceTelescopeScienceInstitute

GeoffreyBryden,JetPropulsionLaboratory

AdamBurrows,PrincetonUniversity

JosephCatanzarite,JetPropulsionLaboratory

M.MarkColavita,JetPropulsionLaboratory

DavidCole,JetPropulsionLaboratory

RolfDanner,NorthropGrummanSpaceTechnology

DennisEbbets,BallAerospace&TechnologiesCorp.

EricFord,UniversityofFlorida,Gainsville

CarlGillmair,SpitzerScienceCenter

SallyHeap,NASAGoddardSpaceFlightCenter

N.JeremyKasdin,PrincetonUniversity

MarcKuchner,NASAGoddardSpaceFlightCenter

BenjaminLane,CharlesStarkDraperLaboratory

NicholasLaw,CalTech

CharlesLillie,NorthropGrummanSpaceTechnology

WalidMajid,JetPropulsionLaboratory/CalTech

GeoffMarcy,UniversityofCalifornia,Berkeley

JamesMarr,JetPropulsionLaboratory

BarbaraMcArthur,UniversityofTexas,Austin

StanimirMetchev,UCLA

JunNishikawa,NAOJ

CharleyNoecker,BallAerospace

XiapeiPan,JetPropulsionLaboratory

GuyPerrin,ObservatoiredeParis/LESIA

MeyerPesenson,CalTech

StevenPravdo,JetPropulsionLaboratory/CalTech

StephenRinehart,NASA

Chapter2

42

JeanSchneider,CNRSParisObservatory

StuartShaklan,JetPropulsionLaboratory

MichaelShao,JetPropulsionLaboratory


StephenUnwin,JetPropulsionLaboratory

JulienWoillez,W.M.KeckObservatory

2.6 ReferencesBean,J.L.,MacArthur,B.E.,Benedict,G.F.,etal.2007,“TheMassoftheCandidateExoplanet

CompaniontoHD33636fromHubbleSpaceTelescopeAstrometryandHigh‐PrecisionRadialVelocities,”AJ,134,749–758

Benedict, G. F., McArthur, B., Chappell, D.W., et al. 1999, “Interferometric Astrometry ofProxima Centauri and Barnard's Star Using HST Fine Guidance Sensor 3: DetectionLimitsforSubstellarCompanions,”AJ,118,1086–1100

Benedict,G.F.,McArthur,B.E.,Franz,O.G.,etal.2000,“InterferometricAstrometryoftheDetachedWhiteDwarf‐MDwarf Binary Feige 24UsingHST Fine Guidance Sensor 3:WhiteDwarfRadiusandComponentMassEstimates,”AJ,119,2382–2390

Benedict,G.F.,McArthur,B.E.,Franz,O.G.,etal.2001, “PreciseMasses forWolf1062ABfromHubbleSpaceTelescope InterferometricAstrometryandMcDonaldObservatoryRadialVelocities,”AJ,121,1607–1613

Benedict,G. F.,McArthur,B.E., Fredrick, L.W., et al. 2002a, “Astrometrywith theHubbleSpaceTelescope:AParallaxoftheFundamentalDistanceCalibratorRRLyrae,”AJ,123,473–484

Benedict,G.F.,McArthur,B.E., Fredrick,L.W., et al.2002b, “Astrometrywith theHubbleSpaceTelescope:AParallaxoftheFundamentalDistanceCalibratorδCephei,”AJ,124,1695–1705

Benedict,G.F.,McArthur,B.E.,Forveille,T.,etal.2002c,“AMassfortheExtrasolarPlanetGliese 876b Determined from Hubble Space Telescope Fine Guidance Sensor 3AstrometryandHigh‐PrecisionRadialVelocities,”ApJL,581,L115–L118

Benedict, G. F.,McArthur, B. E., Fredrick, L.W., et al. 2003, “AstrometrywithTheHubbleSpaceTelescope:AParallaxoftheCentralStarofthePlanetaryNebulaNGC6853,”AJ,126,2549–2556

Benedict,G.F.,McArthurB.E.,Gatewood,G.,etal.2006,“TheExtrasolarPlanetɛEridanib:OrbitandMass,”AJ,132,2206–2218

Benedict, G. F., McArthur, B. E., Feast, M. W., et al. 2007, “Hubble Space Telescope FineGuidance Sensor Parallaxes of Galactic Cepheid Variable Stars: Period‐LuminosityRelations,”AJ,133,1810–1827

Beuermann,K.,Harrison,Th.E.,McArthur,B.E.,etal.2003,“ApreciseHSTparallaxofthecataclysmicvariableEXHydrae, itssystemparameters,andaccretionrate,”A&A,412,821–827

Astrometry

43

Beuermann,K.,Harrison,Th.E.,McArthur,B.E.,etal.2004,“AnHSTparallaxofthedistantcataclysmicvariableV1223Sgr,itssystemparameters,andaccretion,”A&A,419,291–299

Cameron,B.P.,Britton,M.C.&Kulkarni, S.R. 2009, “PrecisionAstrometrywithAdaptiveOptics,”AJ,137,83–93

Catanzarite, J., Law N., & Shao, M., 2008, “Astrometric Detection of exo‐Earths in thePresenceofStellarNoise,”Proc.SPIE,7031‐91

Colavita,M.M.,Wallace,J.K.,Hines,B.E.,etal.1999,“ThePalomarTestbedInterferometer,”ApJ,510,505–521

Cumming,A.,Butler,R.P.,Marcy,G.W.,Vogt,S.S.,Wright,J.T.,&Fischer,D.A.,2008,“TheKeck Planet Search: Detectability and the Minimum Mass and Orbital PeriodDistributionofExtrasolarPlanets,”PASP,120,531–554

Davidson, J., Edberg, S., Danner, R., Nemati, B., & Unwin, S. 2009, SIM Lite: AstrometricObservatory,JPL400‐1360,JetPropulsionLaboratory,Pasadena,CA

Franz, O. G., Henry, T. J.,Wasserman, L. H., et al. 1998, “The First Definitive BinaryOrbitDetermined with the Hubble Space Telescope Fine Guidance Sensors: Wolf 1062(Gliese748),”AJ,116,1432–1439

Gould,A.2008,“OntheDifferenceinStatisticalBehaviorBetweenAstrometricandRadial‐VelocityPlanetDetections,”ApJ,submitted,arXiv:0807.4323

Goullioud,R.,Catanzarite,J.H.,Dekens,F.G.,Shao,M.,&Marr,J.C.IV2008,“OverviewoftheSIMPlanetQuestLightmissionconcept,”Proc.SPIE,7013,70134T

Harrison,T.E.,Johnson,J.J.,McArthur,B.E.,etal.2004,“AnAstrometricCalibrationoftheMv‐PorbRelationship forCataclysmicVariablesbasedonHubbleSpaceTelescopeFineGuidanceSensorParallaxes,”AJ,127,460–468

Hemenway, P. D., Duncombe, R. L., Bozyan, E. P., et al. 1997, “The Program to Link theHIPPARCOS Reference France to an Extragalactic Reference System Using the FineGuidanceSensorsoftheHubbleSpaceTelescope,”AJ,114,2796–2810

Henry, T. J., Franz, O. G., Wasserman, L. H., et al. 1999, “The Optical Mass‐LuminosityRelationattheEndoftheMainSequence(0.08–0.20M_solar),”ApJ,512864–873

Hirota,T.,Bushimata,T.,Choi,Y.K.,etal.2007,“DistancetoOrionKLMeasuredwithVERA,”PASJ,59,897–903

Honma, M., Bushimata, T., Choi, Y. K., et al. 2007, “Astrometry of Galactic Star‐FormingRegion Sharpless 269 with VERA: Parallax Measurements and Constraint on OuterRotationCurve,”PASJ,59,889–895

Ida, S. & Lin, D.N.C. 2004a, “Toward a Deterministic Model of Planetary Formation. I. ADesert in theMassandSemimajorAxisDistributionsofExtrasolarPlanets,”ApJ,604,388–413

Ida,S.&Lin,D.N.C.2004b,“TowardaDeterministicModelofPlanetaryFormation. II.TheFormation and Retention of Gas Giant Planets around Stars with a Range ofMetallicities,”ApJ,616,567–572

Chapter2

44

Ida,S.&Lin,D.N.C.2005,“TowardaDeterministicModelofPlanetaryFormation.III.MassDistributionofShort‐PeriodPlanetsaroundStarsofVariousMasses,”ApJ,626,1045–1060

Lane,B.,Kuchner,M. J.,Boden,A. F., Creech‐Eakman,M., andKulkarni, S.R.2000, “DirectdetectionofpulsationsoftheCepheidstarζGemandanindependentcalibrationoftheperiod‐luminosityrelation,”Nature,407,485–487

Laskin, R. A. 2006a, “Successful completion of SIM‐PlanetQuest technology,” Proc. SPIE,6268,626823

Laskin, R. A. 2006b, “SIM‐PlanetQuest technology completion: A retrospective view,”InternationalAstronauticalCongress,Valencia,Spain,IAC‐06‐A3.P.1.05

Lunineetal.2008,WorldsBeyond:AStrategyfortheDetectionandCharacterizationofExoplanets,NationalScienceFoundation,Arlington,VAhttp://www.nsf.gov/mps/ast/aaac/exoplanet_task_force/reports/exoptf_final_report.pdf

Marr, J.C. IV,Shao,M.,&Goullioud,R.2008,“SIM‐Lite:Progressreport,”Proc.SPIE,7013,70132M

Mayor,M.&Queloz,D.1995,“AJupiter‐MassCompaniontoaSolar‐TypeStar,”Nature,378,355–359

McArthur,B.E.,Benedict,G.F.,Lee,J.,etal.1999,“AstrometrywithHubbleSpaceTelescopeFine Guidance Sensor 3: The Parallax of the Cataclysmic Variable RW Triangulum,”ApJL,520,L59–L62

McArthur,B.E,Benedict,G.F.,Lee,J.,etal.2001,“InterferometricAstrometrywithHubbleSpaceTelescopeFineGuidanceSensor3:TheParallaxof theCataclysmicVariableTVColumbae,”ApJ,560,907–911

McArthur,B.E.,Endl,M.,Cochran,W.D.,etal.2004,“DetectionofaNeptune‐MassPlanetintheρ1CancriSystemUsingtheHobby‐EberlyTelescope,”ApJL,614,L81–L84

National Research Council 1991, The Decade of Discovery in Astronomy and Astrophysics,(theBacahlReport),NationalAcademyPress,Washington,DC

National Research Council 2001, Astronomy and Astrophysics in the New Millenium, (theMcKee–TaylorReport),NationalAcademyPress,Washington,DC

Pravdo, S. H., Shaklan, S. B., Henry, T., Benedict, G. F. 2004, “Astrometric Discovery of GJ164B,”ApJ,617,1323–1329

Pravdo, S.H., Shaklan, S.B.,&Lloyd, J.P.2005, “AstrometricDiscoveryofGJ802b: In theBrownDwarfOasis?”ApJ,630,528–534

Pravdo,S.,Shaklan,S.,Redding,D.,Serabyn,E.,&Mennesson,B.2007,“FindingExoplanetsaroundOldandYoungLow‐MassStars,”whitepapertotheAAACExoplanetTaskforce,http://exoplanets.jpl.nasa.gov/documents/Exoplanetnasawhitepaper_final1‐pravdo.pdf

Pravdo,S.H.,Shaklan,S.B.,Wiktorowicz,S.J.,Kulkarni,S.,Lloyd,J.P.,Martinache,F.,Tuthill,P.G.,&Ireland,M.J.2006,“MassesofAstrometricallyDiscoveredandImagedBinaries:G78‐28ABandGJ231.1BC,”ApJ,649,389–398

Astrometry

45

Raymond, S. N., Mandell, A.M., & Sigurdsson, S. 2006, “Exotic Earths: Forming HabitableWorldswithGiantPlanetMigration,”Science,313,1413–1416

Roelofs,G.H.,Groot,P.J.,Benedict,G.F.,etal.2007,“A.HubbleSpaceTelescopeParallaxesofAMCVnStarsandAstrophysicalConsequences,”ApJ,666,1174–1188

Scargle, S.D. 1982, “Studies in astronomical time series analysis. II ‐ Statistical aspectsofspectralanalysisofunevenlyspaceddata,”ApJ,263,835–853

Schaefer,B.E.2008,“AProblemwiththeClusteringofRecentMeasuresoftheDistancetotheLargeMagellanicCloud,”AJ,135,112–119

Soderblom,D.R.,Nelan,E.,Benedict,G.F.,etal.2005,“ConfirmationofErrorsinHipparcosParallaxes from Hubble Space Telescope Fine Guidance Sensor Astrometry of thePleiades,”AJ,129,1616–1624

Tarter,J.C.,Backus,P.R.,Mancinelli,R.L.,etal.2007,“AReappraisaloftheHabitabilityofPlanetsAroundMDwarfStars,”Astrobiology,7,30–65

Unwin,S.,Shao,M.,Tanner,A.M.,etal.2008,“TakingtheMeasureoftheUniverse:PrecisionAstrometrywithSIMPlanetQuest,”PASP,120,38–88

vanAltena,W.F.,Lu,C.‐L.,Lee,J.T.,etal.1997,“TheDistancetotheHyadesClusterBasedonHubbleSpaceTelescopeFineGuidanceSensorParallaxes,”ApJL,486,L123–L127

Wetherill,G.W.1996,“TheFormationandHabitabilityofExtra‐SolarPlanets,”Icarus,119,219–238

OpticalImaging

47

3 OpticalImaging

RémiSoummer,SpaceTelescopeScienceInstitute,Chair

MarieLevine,JetPropulsionLaboratory,CoChairRuslanBelikov,PaulBierden,AnthonyBoccaletti,RobertBrown,ChrisBurrows,EricCady,WebsterCash,MarkClampin,CostasCossapakis,IanCrossfield,LarryDewell,AlanDressler,RobertEgerman,TomGreene,OlivierGuyon,SaraHeap,B.Jaroux,JeremyKasdin,JimKasting,MatthewKenworthy,SteveKilston,AndyKlavins,JohnKrist,BenjaminLane,AmyLo,PatrickJ.Lowrance,RickLyon,BruceMacintosh,SeanMcCully,MarkMarley,ChristianMarois,GaryMatthews,BenMazin,NaoshiMurakami,CharleyNoecker,JunNishikawa,BenR.Oppenheimer,NelsonPedreiro,AkiRoberge,RobertRudd,StephenRidgway,GlennSchneider,JeanSchneider,LaurentPueyo,StuartShaklan,AnandSivaramakrishnan,DavidSpergel,KarlStapelfeldt,MotohideTamura,DomenickTenerelli,JasonTolomeo,VolkerTolls,JohnTrauger,RobertJ.Vanderbei,JeffWynn

3.1 IntroductionThischapterstudiesthedirectdetectionandspectralcharacterizationofextrasolarplanetsusingopticalmethods: internal coronagraphs, external occulters, andhybrid systems.Webuildonpreviousreports,TPF‐CSTDT(Levine,Shaklan&Kasting2006),TPF‐CTechnologyPlan(Dooley&Lawson2005),ExoplanetForum2006(Lawson&Traub2006),ExoplanetTask Force (Lunine et al. 2008),which focusedmainly on detection and characterizationcapabilitiesforaflagshipmissionusinganinternalcoronagraph.

Ourgoalistostudyarangeofpossiblemissionscales,andtorecommendatechnologyplanforthesemissions.TherearetworelevantscalescompatiblewithcurrentNASAplans:alineofnear‐termprobe‐sizemissions($600Mwithoutlaunchvehicle),andlongerrangeflagshipmissions(>$1B).Wedonotdiscussdetailsofproposedconcepts,butratherpresentclassesof mission types, where TPF‐C denotes a large internal coronagraph, and TPF‐O a largeexternalocculter.

The ExoPTF recommended that indirect methods (astrometry and radial velocity) arenecessary to obtain a direct measurement of the masses, orbital parameters, and planet“addresses.” This precursor knowledge would help a follow‐on characterization‐onlymission (i.e., if theplanetsarealreadyknown),byprobably relaxing requirementson theInnerWorking Angle (IWA), and reducing themission duration. However, the precursorinformationwillhavelittleimpactonotherdrivingsystemrequirements,suchasstability,pointing, contrast, spectral coverageandspectral resolution,whichremainessentially thesameasfora fullsearch‐and‐characterizationmission.Thetechnologyplanweproposeistherefore valid for both a characterization‐only and a search‐and‐characterization opticalmission.

Chapter3

48

Observing strategies and mission scales must be designed in tandem between theastrometricandimagingprogram.DesignReferenceMission(DRM)developmentoverthenext several yearswill articulate the tradeoffs in cost and performance between imagingmissionswithandwithoutastrometricprecursors.

We recognize the merits of the strategy developed by the ExoPTF, and expand on itsrecommendationintheshortterm:aprobe‐scaledirect‐imagingmissioncanbecombinedwith existing radial velocity observatories and ground‐based Astrometry (PRIMA,Delplancke2008) forthecharacterizationofmaturegiantplanets,andpossiblyNeptunes,and Super Earths. A probe scale will also detect and characterize exozodiacal disks, aproblemExoPTFidentifiedascriticalforfutureterrestrial‐planetimagingmissions.

This strategy is independent froma space astrometricmissionboth in termsof scientificgoals and timing sequence, and therefore offers an alternate and coherent programespeciallyifanastrometricmissiontodetectEarth‐sizeplanetsisnotrecommended.

Chapterrecommendations:

1) Probescalemission(1.5mclasstelescopeat~$600M)startingwithinthenexttwoyears,forexoplanetarysystemscharacterizationthatisnotaccessiblefromtheground:

a. Exozodical characterization (1 zodi at 1–3 AU) of brightness, structures, andclumpiness. This is critical for future characterization of habitable Earthsizeplanets.

b. Directdetectionandspectralcharacterization(0.4µmto1.0µm)ofplanetsandplanetarysystems,includingatleast13knownradialvelocity(RV)giantplanets,aswellasyetunknownNeptunesandSuperEarths.

2) Longterm flagship mission for characterization of potentially habitable EarthSizeplanets.

3) Aggressiveandsustainedtechnologydevelopmentstartingimmediatelytomeetreadinessbymissionstart,includingnearspaceenvironmentopportunities(rocketsandballoons).

4) R&Asupport fordeveloping scienceandmaturingmissiondesigns for thedefinitionandoptimizationofcombinedAstrometry/RVandimagingmissions.

3.1.1 ScienceGoalsandRequirements

FlagshipscalegoalsThereisnoquestionthatthedirectdetectionofaterrestrialplanetinthehabitablezoneofanearbystarwouldbeofmonumentalimportance.Thevariousstudyreports(e.g.,Levineet al. 2006; Lunine et al. 2008), and a sizeable literature have focused on strategies forcharacterizing suchworlds and searching for signs of life. Earth‐size planets lying in thehabitablezoneofnearbystarsgenerallywillnotexhibitadetectableRVsignal.

Recent studies show that the characterization of atmospheric features in Earth‐analogsaround nearby M Dwarfs using transit spectroscopy with JWST seems very unlikely.However,othertypesofterrestrialplanetsmayexistthatcouldbecharacterizedbyJWST,suchasoceanplanetswithstrongatmosphericwatervaporsignature(Greeneetal.2007;Beckwith 2008). Therefore, an imaging mission will be necessary to observe andcharacterizespectroscopicaly theatmospheresofEarthanalogs.Earth‐sizeplanetswillbetoo small and faint for even low‐resolution spectroscopy with a small telescope, and

OpticalImaging

49

thereforewill require a flagship‐sizemission for their characterization. It is important tokeepinmindthatSuper‐Earths(1–10M⊕)maywellbehabitable,andrecentRVresultswiththeHighAccuracyRadialvelocityPlanetarySearch(HARPS)seemtosuggestthattheyarecommon (Mayor et al. 2009). Such amissionwould be able to detect and characterize alargevarietyofplanetsfromEarth‐sizetoGiantplanets.

ThemaingoalistolookforthepresenceofO2andH2O,whichcanbeidentifiedinEarth’sspectrum with a resolution, R = λ/∆λ of 100 (Kaltenegger et al. 2007). This resolutionshouldbeableto identifythestrongO2bandat0.76µm,andH2Oat0.96µm.Ozone(O3),which is formed photochemicaly from O2 in Earth’s atmosphere is also an interestingbiomarkerbutmoredifficulttodetectinthevisible.Itcanbedetectedintheinfraredat9.6 µm or in the UV at 0.25–0.32 µm. The best combination of biomarkers would be thesimultaneousdetectionofO2orO3 togetherwitha reducedgas likeCH4,whichwouldbeundetectableintoday’sEarthatmospherebutwasmoreabundantontheProterozoicEarth(2.5–0.54Gyrago).CH4wasprobablyevenmoreabundantduringtheArcheanEon,priorto2.5Gyrago,butO2andO3wouldhavebeenabsentatthistime.

FlagshipscalerequirementsThe science requirements for a flagship‐scalemissionwere set by theTPF‐C Science andTechnologyDefinitionTeam(STDT)(Levine,Shaklan&Kasting2006)asa95%confidenceleveloffindingoneEarth‐likeplanetiftheprobabilityofsuchplanets(η⊕)is0.1.Acarefulanalysis of mission completeness requires development of a comprehensive DesignReference Mission (DRM) that includes realistic observing scenarios, time constraints,photometric integrationtimes,andmultiplevisitsfordetectionandorbitcharacterization.Forconservativeestimates,anapproximatemissioncompletenesscanbefoundasthesumof the individual completeness for each star in the observing program. This gives theexpectation value for the total number of planets found and characterized, assuming allstars have one planet drawn from the population of interest and that each visit isindependentandusesnopriorknowledge.

It is the goal of DRM studies to analyze the mission completeness under variousassumptions (single or concatenated programs) and refine the mission parameters(instrumentandobservingrequirements). Inmostcases inconductingtheDRM,themaintradesaretimespentsearchingfornewplanets,timespentformovingformonetargettoanother, timespentdoingspectroscopyonfoundplanets,andnumberofrevisits fororbitdetermination.The initial performance requirements derived from the sciencerequirementsareassumedtobeacontrastof10–10atanIWA=75mas,overthespectralrange,0.5–1.0µm,witharesolutionR~100.

ExoPTF has identified two key parameters for the design of a flagship direct‐imagingmission:theprobabilityoffindingaterrestrialmassplanet(0.3–10.0Earths)intheHZofF,G, or K star (η⊕) and the level of exozodiacal light. These two parameters have a directimpact on the scale of the imaging mission: if η⊕ is high (η⊕ > 0.1), and if the level ofexozodiacallightissmallerthan10zodi(1zodiisthelocalzodiacallightsurfacebrightnessseen from Earth), characterizationwith direct imagingwill bewithin reach of a flagshipmissioninthenearfuture.

Measuring the value of η⊕ is a main goal of the CoRoT and Kepler missions, and thisparameterwillbeknownwithinthenextfewyears.However,knowledgeoftheexozodiatalevelrelevantfordirectimagingiscurrentlylacking,withcurrentground‐basedsensitivityoftheorderof100zodisandfutureinstruments(e.g.,LBTI),whichhopetoreach10zodis

Chapter3

50

in the infrared. A probe‐size high‐contrast mission has a unique capability to detectexozodiacallevelsdownto1zodiinthevisibleandcharacterizetheirstructure.

The ExoPTF recommended that it is essential to combine detection and orbitalcharacterizationbyanindirecttechnique—astrometryorRV—withcharacterizationusingdirectimaging,foratleasttworeasons.First,onlyreflexmotionrevealstheplanetarymass,which is anessentialdatum forunderstandingaplanet inphysical terms, andcan supplythe “addresses” of planets for their characterizationwith direct imaging. Second, indirectdetectionisamorethoroughtechniquefordiscoveringpotentiallyhabitableplanets.Direct‐detection searches, by comparison, typically have lower completeness per searchobservationandseasonalsearchconstraints(Brownetal.2005).

However, a time‐limited mission doing both detection and characterization can beoptimizedtoproduceafewdetectedandcharacterizedplanets,assuminghighoccurrenceη⊕>0.1(Savransky&Kasdin2008).InthisparticularDRMapproach,thenumberofplanetsfound and characterized does not depend strongly on η⊕. This is because most of themissiontimeisspentcharacterizingplanetswithspectroscopyandrevisitingthemtobuildanastrometricorbit.

OtheroptionsexistforDRMscenarioswherespectralcharacterizationiscompletedduringthe initial visit in which the planet was detected, and multiple revisits focus on orbitdetermination(Heap2007;Lindler2007).

Theparticularsensitivityof themissioncompleteness to the IWAcanbeused todeviseaspecific observing scenario for an occultermission (Spergel 2009): a “detectionmode” isused with the starshade at larger distance and smaller wavelengths than for the“characterizationmode,”inordertominimizetheIWAforthedetectionphase.

Thepossibilityofcombinedmissions(astrometry/RVanddirectimaging)requiresspecificDRMstudies includingSNRestimationstodefineamatchedsetofmissions.Weintroducethe problem of concatenated mission design in Figure 3‐1 by showing the astrometricwobblevs.maximumseparationforanEarthattheinneredgeoftheHZaroundaqualifiedsample of stars within 30 pc. This figure updates Figure 11.1 in the ExoPTF report byincludingphotometriclimitsandrestrictionsoftheplanetaryperiod.

Thehorizontallinedefinestheastrometricrequirement(0.2‐µasmissionsensitivity)andavertical line defines the IWA requirement for the imaging mission (94 mas). The otherparametric assumptions in this figure include the definition of the star list (non‐binary,main‐sequenceHipparcosstarswithin30pcwithB–V>0.3),definitionofthecoronagraph(efficiency,bandwidth,etc.),definitionsoftheInnerandOuterHabitableZonesdefinitions(IHZ=0.82AU,OHZ=1.6AU),andthebrightnessoftheexozodi(1zodi).

For stars above the upper red line, the periods of all HZ orbits are less than 5 yrs andtherefore formally detectable by periodogram in a 5‐year astrometric search program;abovethelowerredline,someHZperiodsarelessthan5yearsanddetectable;belowthelowerredline,none.Forstarsabovetheupperblack,dashedline,thecoronagraphobtainstheA‐bandspectrumofO2at760nm,withR=100andSNR=10inlessthan106sec,forallHZEarths;abovethe lowerblackline,some;belowthelowerblackline,none. Inthiscase,some 38 stars are in the first quadrant, where all HZ Earths produce an astrometricsignatureofatleast0.2μasandareresolvedbythe5‐mcoronagraph.For6of38stars,allHZ Earths are discoverable by astrometry (period short enough) and characterizable bycoronagraphy (bright enough). For 10 more stars, some discoverable HZ Earths are

OpticalImaging

51

characterizable. For 22 more stars, some HZ Earths are discoverable, but none of themcharacterizable.OnestarhasnodiscoverableorcharacterizableHZEarths.

Figure31.ParameterspaceforestimatingthesizeofstellarsamplesforanastrometricfinderprogramfeedingHZ“Earths”toacoronagraphiccharacterizerprogram.(R.A.Brown,SpaceTelescopeScienceInstitute)

It is important to note from these results that a small direct‐imagingmissionwould notparticularlybenefitfrompriordetectionofterrestrialplanetswithastrometry,sincetheseplanets would be too faint to be characterized (Figure 3‐1). However a small imagingmission would be able to characterizemoremassive planets detected by an astrometricprecursororradialvelocity.

Reciprocally,alargeimagingcharacterizationmissionfollowingasmallastrometricmissionwould run out of known planets very quickly, and would then spend most its lifetimefollowingasearch‐and‐characterizemissionsimilartoproposedDRMswithoutastrometricprecursor.

Recent improvements in RV techniques allow some hope of finding potentially habitableplanetsdowntoafewtimesEarthmassaroundsolar‐typestars.Also,alternatemethodsforastrometrymightbepossible, forexamplewithafocal‐planewidefieldinstrumentonthedirectdetectionmissionitself.

Adetailedstudyisneededtocomputefractionalcompletenessfortheconcatenatedfinderandcharacterizerprograms,whateverthetechniques.Werecommendcontinuedeffortstodevelop and analyze various DRM approaches, for direct and concatenated missions asinitiatedbyExoPTF.Thesestudiescombinedwithimprovedknowledgeofη⊕andexozodilevelwillhelpdefinethebeststrategyforaconsistentexoplanetprogram.

Chapter3

52

ProbescalegoalsA small mission would be capable of characterizing at least 13 known RV‐discoveredplanets, and could detect large terrestrial planets (~10 M⊕) around the nearest stars,assuminganaperturediameterof1.5mandanIWAof2λ/Dat500nm.Itwouldmeasureexozodiacal brightness down to the level required for Earth‐size planet characterization.TheExoPTFhasidentifiedthisproblemasvital.

Exozodiacal dust: Hundreds of Sun‐like stars have been found to exhibit excess IRemission attributable to dusty circumstellar debris. The vast majority of these systems,however, remain spatially unresolved. High‐resolution scattered‐light images of debrisdiskswillreveal themorphologyof thedisksandtracethe locationof thedust‐producingplanetesimals, but such images today are rare due to the lack of suitable high‐contrastimagingsystems.

Dynamicalinteractionsofplanetswithresidualplanetesimalbeltslikelyplayavitalroleinthearchitecturesofplanetarysystems.Kenyon&Bromley(2004)haveshownthatrecentlyformed terrestrial planets can generate copious quantities of dust in the early phases ofplanetarysystemformation.InourownSolarSystem,Jupiteristhoughttohavedominatedthe dynamics in the early evolution of the asteroid belt. Consequently, volatile‐richplanetesimalsscatteredtowardsourinnerSolarSystemmayhavecontributedsignificantlyto terrestrialwaterabundance,possiblyaffectingorcontributingto thehabitabilityof theearlyEarth.

Table31.Radialvelocityplanetsobservablewithaprobe‐sizemission.

PlanetName Msini Period(d) a(AU) Separation Contrast(10–9)

EpsilonEridanib 1.55 2502 3.39 1.06 1.655Cncd 3.84 5218 5.77 0.43 0.6

HD160691c 3.10 2986 4.17 0.27 1.1Gj849b 0.82 1890 2.35 0.27 3.3HD190360b 1.50 2891 3.92 0.25 1.247Umac 0.46 2190 3.39 0.24 1.6HD154345b 0.95 3340 4.19 0.23 1.0UpsAndd 3.95 1275 2.51

2.0440.19 2.9

GammaCepheib 1.60 903 2.044 0.17 4.4HD62509b 2.90 590 1.69 0.16 6.4HD39091b 10.35 2064 3.29 0.16 1.714Herb 4.64 1773 2.77 0.15 2.447Umab 2.60 1083 2.11 0.15 4.1HD89307b 2.73 3090 4.15 0.13 1.1HD10647b 0.91 1040 2.1 0.12 4.2HD217107c 2.50 3352 4.41 0.12 0.9HD117207b 2.06 2627 3.78 0.12

0.111.3

HD70642b 2.00 2231 3.3 0.11 1.7HD128311c 3.21 919 1.76 0.11 5.9

OpticalImaging

53

High‐resolutionspatialmappingofthedustycircumstellardebris inexoplanetarysystemsisneeded to test thegeneric applicabilityofdynamicalmodels (Morbidelli et al. 2005) inplanetarysystemsotherthanourown.

Currently deployed high‐contrast imaging technologies can detect only the largest, mostmassive, and brightest circumstellar disks, and cannot effectively probe their innermostregions(Oppenheimeretal.2008).Thefewimagesobtainedtodatehaveprovidedcrucialinsights intotheformation,evolution,andarchitecturesofexoplanetarysystems;buttheyare just the tip of the icebergwaiting to be fully revealed. The existing sparse sample ofdebrisdisksimagedwithscatteredstarlightrepresentsonlytheyoungest(10–100Myr),orextremelyanomalous,disksystemsaroundolderstars.

Whileholdingconsiderablescientific interestof theirown,extrasolarzodiacaldiskscouldalso inhibit thedetectionandcharacterizationof terrestrialplanets.TheExoPTFplacedahighpriorityonmeasuringexozodiacallightaroundtargetstarsbecauselevelsgreaterthanabout10 times thatof theSolarSystemwill severelyhamperdetectionofplanets.Futureground‐based systems (LBTI) are expected to approach this level. However, it is unsurehow thesemeasurements at 10 µm can be extrapolated to predict the zodi brightness atvisiblewavelengths.Asmallcoronagraphcouldsurveynearbystarsandprovideconfidencethatalargerobservatorywouldnotbehamperedbyexozodiacaldust.

Giantplanets:Characterizationofextrasolargiantplanetswillfollowahierarchalprocessthat beginswithmass and proceeds through atmospheric temperature, composition, andfurther characterizationof the atmosphere.Herewebriefly highlight threemajor sciencequestionsthataprobe‐sizemissionmightaddress.

What are the masses of RV planets? Measuring the masses of known RV planetscharacterizes stellar system architectures and constrainsmodels of planet formation. ForRV‐detected planets two coronagraphic images spaced by a fraction of the orbital periodwillconstraintheorbital inclinationandthusthemass.Table3‐1liststhemostaccessibleRV planets to direct imaging. More entries for this list will occur as RV observationscontinue.Thisfirstmajorsciencegoalplacesarequirementontheastrometricaccuracyofthecoronagraphicsystem(orientationonthesky,angularseparation,andpositionangleoftheplanetwithrespecttothestar).

Knowledge of the mass is the first step in the characterization of a planet, and itsubstantially narrows the range of possible interpretations of a given spectrum. Well‐characterizedplanetsofknownmasscanalsoserveastemplatesforcomparisonwithotherdirectlydetectedplanetsforwhichtherearenoRVdata.

Whichcloudlayersandmoleculesarepresentintheiratmospheres?Theatmospheresof giantplanets consistof a thickenvelopeofhydrogenandheliumwithanadmixtureofheavy elements. Atmospheric temperature and elemental abundances determine whichmolecules are present in the observable atmosphere, either as gasses or condensed intoclouds.Cloudthickness,atmosphericcomposition,andgravitycontroltheopticalreflectionspectra.

Forexample,the‘icegiants’(UranusandNeptune)arebluebecauseoftheirlowergravities,thinner clouds, and greater methane abundance than the redder Jupiter and Saturn.Changes inanyof thesepropertieswouldchangetheirspectraandcolors. Ifanyof theseplanets were warmer, so that their deep water clouds were instead near the opticalphotosphere, then their continua levelwould bemuch brighter and themolecular bandsless deep. Uranus and Neptune would be almost white instead of blue. Such trends are

Chapter3

54

evident in the model spectra shown in Figure 3‐2, which imagines moving a Jupiterprogressivelyclosertoitsprimarystar.

Thus, reflection spectra of detected planets thus provide the second level ofcharacterization by determining which gasses and clouds are present and consequentlyconstrainingatmospherictemperature.

Forplanetsnotdetectedbyradialvelocity,massmustalsobeinferredfromopticalcolorsand spectra. Since giant planets cool over time, an estimate for the systemage combinedwithtemperatureconstraintsfromtheobservedcompositionandcloudlayersplaceslimitsonplanetmass.

Figure32.ModelgeometricalbedospectraforaJupiter‐massplanetatvariousdistancesfromasolartypestar.Dottedlinesaremodelsforthe3‐and2‐AUplanetsbutwithoutcloudopacity.Whilethesespectraarecastasafunctionoforbitalradius,thesamesequencewouldresultforagiantplanetkeptatafixeddistanceandmodeledwith progressively younger ages, highermasses, or earlier stellar spectral types, all ofwhichwould producewarmereffective temperatures, all elsebeingequal. Forexamplean8‐Jupiter‐massplanetatanageofabout1Gyrwouldhaveasimilarspectrumtotheplanetat1.0AU.A4‐Jupiter‐massplanetatthesameagewouldbesimilartoacloudlessplanet(dottedline)at2AU.ThespectraoftheplanetslistedinTable3‐1willspanmuchoftherangeshownhere.(M.Marley,NASAAmesResearchCenter)

How abundant are key absorbers? The abundance of key absorbers, particularlymethane, are tracers of planetary formation scenarios. For example the core accretiontheorydescribingtheformationofgiantplanetssuggeststhatanymassiveplanetwillhaveathickenvelopeofroughlynebularcompositionsurroundingadensercoreofrockandice.Subsequentprocesses, includingbombardmentbyplanetesimals, canenhance theheavierelements. Jupiter’s atmospheric carbon abundance is enhanced by a factor of three oversolarabundance,whileNeptuneandUranusareenhancedbyabouta factorof30.Directcollapseoftheseplanetsfromnebulargaswouldnotresultinsuchapatternofenrichment.

OpticalImaging

55

Thethirdsciencegoalshouldthusbethedeterminationofatmosphericcompositiontoinferwhether thepatternof enrichment seen in Solar Systemgiantsholds inother systemsaswell.Theenrichmentpatternmightvarywiththemassoftheprimarystar,themassoftheplanet,theorbitalsemi‐majoraxis,orotherparameters.

Although spectroscopy is in principle better for answering these questions, it may beunobtainable in realitywith a smallmission. The utility of very low resolution, low SNRspectra, or filter photometry needs to be studied. One set of possible filters that mightenable the science discussed above is shown in Figure 3‐3. The broadband filterswouldconstraintheoverallslopewhilenarrowerfiltersmightconstrainthedepthsofmolecularbandsforbrightornearbyplanets.

Figure33.Black linesat topshowsomepossible filterchoices foracoronagraphiccamera.Thebroadfiltersdetect planets and constrain their spectral slope. Progressively narrower filters provide increasing levels ofspectralresolutionofkeyspectralfeatures.(M.Marley,NASAAmesResearchCenter)

As shown in Table 3‐1 there is a great diversity in masses and orbits among the mostdetectableextrasolargiantplanets.Byobtainingphotometryinavarietyofwide,medium,and(forthemostfavorableobjects)narrowbandsitwillbepossibletocharacterizearichmenagerieofworlds.Theresultingsubstantialscienceyieldwouldbeanimportantlegacyofanysmallmission.

Super Earths: Recent results of radial velocity surveys with HARPS (Mayor et al. 2009)suggest that theseplanetsarenumerous.Thechallenge foraprobe‐scale telescopewouldbe detecting and recognizing such planets. Color and absolute brightness might help todiscernbetween, forexample,asmallgaseousplanet, likeNeptune,anda largeterrestrialplanet. However, thismay be unreliable, becausemany different processes influence thecolorandalbedoofaplanet,includingsurfacecomposition,altitudeandthicknessofclouds,molecularabsorbers,andphotochemicalprocesses.ThecolorandbrightnessofEarthhascertainlyvariedovertime,forexample.Thus,unlesssuchaplanetisfoundverynearby,itislikely that a small coronagraph could only obtain a “pale blue dot” image which wouldundoubtedly lead to calls for a more capable mission to characterize such an intriguing

Chapter3

56

object.IfindeedasuperEarthisfoundnearby,asmallmissioncouldestablishthepresenceofwatervaporbandsandpossiblyrecognizewaterclouds,whichwouldbeofextraordinaryinterest.Also, ifanorbitcanbeconstructedfromdirect imaging,RVmeasurementsmightbeabletoconstrainthemassknowingtheorbit(Spergel2009).

ProbescalerequirementsA probe‐scale mission can implement the general ExoPTF recommendation to combinedirect and indirect methods, using the overlap between known RV planets and theirdiscovery space. At least 13 of these planets can be characterized with a small spacemission(semi‐majoraxis>150mas).ThisnumberofRVplanetsdetectablewithaprobe‐sizecoronagraphwill increasewithtimeasmoreplanetsarediscovered.Wecanexpectafew tens of RV planets by the time a probe mission could be in operation. Stars withcurrentlyunconfirmed,long‐termradial‐velocitytrendswouldbeparticularlygoodtargetsforsuchamission(Marcy2007).

Ground‐based astrometrywith a few ten µas accuracywith VLT PRIMAwill also overlapwith the probe‐scale imagingmission parameter space, with the detection of Jupiter‐likeplanets fornearbyG stars (<30pc) down toNeptune‐massplanets for the closer lower‐massstars(Delplancke2008).ForGiantplanets,therequirementsstronglydependontheageoftheobjects:

Table32.Coronagraphperformancerequirementsfordetectionandcharacterizationofgiantplanets.

Age Contrast IWA λ/Δλ Remarks

RecentlyFormed 10–5 0.05" ~50 Moderatenear‐infrared(NIR)contrast,smallIWA:

distantstarformingregions(e.g.,Taurus)

Young(~100MYr) 10–7 0.2" ~40 NIR,nextgenerationinstrumentson8‐mtelescopes

(GPI,SPHERE)

Mature(>2GYr) 10–8–10–10 0.1" ~50 GoaltoreachRVplanets

3.2 ObservatoryConceptsOver the past decade there has been a wide variety of concepts proposed to do high‐contrast imaging, both on the ground and in space. The architectural approaches forstarlight suppression include internal coronagraphs, external occulters, and a hybrid.Weassociatethehybridwiththeexternalocculters,andcarryonlytwocategories,internalandexternal.Foreach,weidentifythemostrelevantmissionscales(Figure3‐4).Weemphasizethe synergy between direct and indirect detection programs, and consider ground‐basedprojects,suborbitalopportunities,andspaceobservatoriesattheprobeandflagshipscales.

3.2.1 ArchitectureScalingforInternalCoronagraphsObservatoriesusing internalsuppressionrelyonsomeversionofacoronagraph(orpupilinterferometer) to reduce the diffraction of starlight by the pupil edges. All the starlightenters the telescope, and they thus requirewavefront sensing and controlwith exquisitestabilitytoeliminatespecklesduetoaberrationsintheoptics.Thetelescopeopticscanuseconventionaldiffraction‐limitedopticssuchastheonethatalreadyexistsforHubbleSpaceTelescope.

OpticalImaging

57

GroundbasedtelescopesAnumberoflargeground‐basedtelescopes(5–8m)arebeingequippedwithcoronagraphsandExtremeAdaptiveOptics(ExAO)toachievecontrastlevelsof~10–7atsmallanglesonthe sky (0.2 arcsec) in the near infrared. Science goals are the detection andcharacterization of young giant exoplanets (10 Myr –1 Gyr) and disks. Planets will bedetectablebytheirthermalemissioninthenear infrared.Theireffectivetemperatureandcompositioncanbeinferredbycomparingtomodels.Theseinstrumentsarelimitedbytheuncorrected atmospheric and staticwavefront error residuals andby their stability. Firstresultsfromtheseprojectsareexpectedaround2011.FutureExtremelyLargeTelescopes(e.g., GSMT) with ExAO coronagraphs will enable similar contrasts at a smaller IWA~30masandhigher contrast (~10–8)at separationsup to~1arcsec.Thevery small IWAwill enable the study of more distant star‐forming regions, and improved contrast willenablethestudyofsomereflected‐light Jovianplanetsaroundnearbystars.High‐contrastinstrumentsonGSMTcanbeexpectedtoproduceresultsaround2020.

Figure34.Rangeofrelevantmissionscalesfordirectimagingtoenableacompletescientificandtechnologicalprogram.(R.Soummer,SpaceTelescopeScienceInstitute,andM.Levine,JPL)

SuborbitalenvironmentsSuborbitalexperimentswithsoundingrocketsorballoonsofferinterestingpossibilitiesfortechnologydevelopmentandriskreduction,inordertohelpadvancelaboratoryconceptstoflight status. Theywould also be able tomake observations of a few exoplanets, planetsbeingformed,andexozodiacaldisks.

Chapter3

58

Acoronagraphinstrumentonaballoon‐bornegondolarequirestheplatformtobeabletopoint at a starwith stabilityof 1 arcsecorbetter (demonstrated in the lab and in flight).Lightweight telescope mirrors, 0.8–3.0 m, already exist (e.g., the 3.5‐m SiC primarydemonstratedonHerschel).Criticalrequirementsincludeinternalfine‐pointingcapability,andminimallydistortingairpathneartoandabovethetelescope(Traubetal.2007).

Sounding rockets can also play a role in the development of direct‐imaging sciencewith500‐kgpayloads,anda600‐s flight timeat400kmaltitude.Thepayload is recoveredviaparachute and canbe re‐flown.NASA currently flies on the order of 10 flights each year.Plansexisttoflya50–70cmtelescopeandpointitto1–2stellartargets(Shaoetal.2007).

ProbesizemissionsThefirstrelevantscaleforspaceimagingmissionsisthe“probemission”scale(~$600M),where it becomes reasonable to consider 1.5‐m class telescopes and very high‐contrastcoronagraphic instruments (10–9–10–10). This would allow measurements of exozodiacaldiskslevels,characterizematuregiantplanetsandplanetarysystemsknownfromRV,andpossibly image a few Super Earths around nearby stars. Because of the small telescopediameter, IWA is a critical parameter for this mission scale, which affects searchcompleteness. Themost promising coronagraph approaches suggest IWA in the 2–3λ/Drange.Activewavefrontcontrolisneededtocorrectopticalaberrationswithtightstabilityrequirementstomaintainthecorrectedimagesqualityduringanobservation.

FlagshipsizemissionsThenextsizescaleisthe"flagship"classmission(>$1B)withatelescopediameterof4to8m. While this introduces new engineering challenges associated with large mirrors inspace,itopensupthepossibilityofimagingterrestrialplanetsinthehabitablezoneoftheparentstar.Anessentialrequirement for themissionscaling is themissioncompleteness.Smallermissions(4–5m)haveasmallersetofstarsavailableandthusasmallernumberofpotential planets and greater sensitivity to η⊕. They require the highest performingcoronagraphs(IWAof2–3λ/D)andtighterstability.

A larger telescope may allow a relaxation in the IWA requirement to about 4 λ/D thuseasing the complexity of the required thermal design for stability. It also opens up thesearchspace,improvestheabilitytoextractplanetsfromtheexozodiacalbackground,andimprovestheprobabilityofdetection.Thiscomes,ofcourse,attheexpenseofamorecostlyandchallengingobservatory.

Mostoftherequirementsonthecoronagraphinstrumentitselfdonotdependonthescaleof the telescope.However, thedifficultyof a largerobservatory comes in the engineeringchallengesoftesting,launching,deploying,andcontrollingthestabilityofthesystem.

An8× 3.5mellipticalmonolithicprimarymirror is the largest size that canbe launchedwithanexistingDeltaIVheavyrocket,oran8‐mcircularwiththeplannedAresV.Thelong‐termdrivetostudyEarth‐sizeplanetsatlargerdistanceswilldemandbreakingthebarrierimposed by the Ares V launch shroud and considering even larger apertures. Largesegmented mirrors are currently a major technology issue for high‐contrast imaging,becauseofthediffractionbythesegments.Specificresearcheffortsareneededtoaddresswhether the characterization of potentially habitable planets can be achieved with largesegmented telescopes. Possibilities may include coronagraph designs less sensitive tosegmentationandtheuseoftwodeformablemirrorsforamplitudeandphasecontrol.Non‐redundantaperturemaskingwithspatialfilteringmayalsoprovideawaytogetaroundthe

OpticalImaging

59

segmentboundary(Lacouretal.2007).Externalocculterswouldnotsufferfromthesamelimitationswitha segmentedaperture.Howeverocculter sizeanddistance scalewith thetelescopediameter,andstudiesareneededtodeterminewhetherthiswouldbeapracticalsolutionintermsofperformance,cost,andscienceoperations.

Yet another alternative to a large telescope is a dilute‐aperture telescope. The potentialadvantageofadilute‐aperturetelescopeistheuseofmultiplesmallertelescopestoachieveaninnerworkingangleequaltothatofalargecontiguous‐aperturetelescope.Thecollectingarea or how dilute the aperture should be configured would be tailored to the exo‐zodilevels detected by existing ground‐based or future suborbital or “probe” class missions.Dilute‐aperture telescopes share many of the properties of a segmented telescope, andwouldrequiretheuseofanullingcoronagraphicarchitecturethatiscompatiblewithnon‐contiguoustelescopeapertures.

3.2.2 ArchitectureScalingforExternalandHybridConcepts

ExternalOccultersObservatories that use external suppression achieve high contrast by preventing thestarlight fromentering the telescopeviaanexternalocculterwhileallowing light fromanoff‐axis planet to pass. By properly designing the shape of the occulter (Cash 2006;Vanderbei et al. 2007), a deep shadow canbe created at the telescope. Sinceno starlightenters, the need formid‐spatial frequency, high‐contrastwavefront control is eliminated.Thetelescopecanbeaconventionaldiffraction‐limitedoneasfortheinternalcoronagraph,howeverthecoronagraphinstrumentcomplexityisremoved.Thisadvantagecomesatthecost of a second formation flying spacecraft with a large deployable screen and theengineering challenges associated with the stringent manufacturing and deploymentrequirementsduetovolumeconstraintsofthelaunchvehiclefairing.

Observatorieswithexternaloccultersobeyverydifferentscalinglaws.Theabilitytoresolveclose‐inplanetsisnolongerafunctionoftelescopesizebutratherthesizeanddistanceofthe occulter (Arenberg et al. 2006). It is also wavelength independent, as long assuppression is achieved at all bands of interest. In principal, an occulter mission can bedesigned at any size scale, but small‐scale missions (1‐ to 3‐m range) are consideredimpractical in general because the cost advantage is offset by the need for the secondspacecraft.Apossibilityhowever,wouldbetouseanexistingtelescope.Afewstudieshavebeencarriedouttotrytoidentifysuchtelescopesandtodesignoccultersforthem(Loetal.2007).Thereareseveralchallengesassociatedwiththispossibility.Theexistingtelescopewould not have a cooperative formation flying system and that burdenwould fall on theexternalocculter.Itmaybedifficulttoobtainenoughsciencetimeontheexistingtelescopetocompleteanadequatesurveyofnearbystars,andspecificsoftheinstrumentationonthetelescopemaybelessthanidealforplanetdetection.

Onceintheflagshipcostrange,allexternalobservatoryconceptsaredesignedlargeenoughto imageEarth‐likeplanets.Thetelescopesize iscurrentlyassumedtobe4m inorder tohavethenecessaryresolvingpowerintheeventof largeexozodiacaldustandtohavethecollectingareaneededforthelargedelta‐magnitudetargets.Muchlargertelescopes(>8m)arenotbeingconsideredhereastheyrequireevenlargeroccultersand,forfindingplanets,theincreasedresolvingpowerisnotbeingused.

Chapter3

60

A4‐m telescopeusing an external occulter to achievehigh contrast at an IWAof 72masrequires a shade at least 50‐m tip‐to‐tip flown at roughly 72,000 km distance. Properlydesigned, sucha systemcanachieve theneeded contrastover abroadband (0.4–1.1µm).The challenges are associated with meeting the tight manufacturing and deploymentrequirements,aswellasmaintainingpositioningandalignment.

Occulter‐basedmissionsneed toslewtheocculter fromtarget to target,which introduceslag‐timebetweenobservations.Efficiencyoftheexternaloccultermustbecarefullystudied,butthismaybeoffsetbythetypicallyhigherthroughputthaninternalcoronagraphs,duetotheverysmallnumberofmirrorsandtheabsenceofmasksimpedingtheplanetlight(thisdepends on the coronagraph design too). The field of regard for observing is limited toavoidsunlightreflectingofftheocculter.Theocculterisrestrictedtoasmallwindowafewdegreesacross,atrightanglestotheSun,andtheobservingseasonis<1month.AddingaSun shield to allow observation to 45 deg would be a major improvement and greatlyincreasesthelikelihoodofrecoveringplanets.Italsogreatlyrelaxestheconstraintsonthestar‐to‐star sequencing, which can save a lot of fuel and increase observationalcompleteness at the same time. However, while the starshade is slewing, there is ampletimeforgeneralastrophysicalobservations.

The starshade spacecraft maneuvers to occult target stars by using a low thrust solarelectric propulsion system to maximize fuel efficiency. The expectation is that for targetstarsspacedonanaverageof15degreesapartonthesky,thestarshadetakes8to13daysto slew to shadow the target star.Once the target starhasbeenshadowed, the starshadeandtelescopemaintainanalignmentwithacross‐tracktoleranceof1mduringthescienceobservationinterval,whichlastsanywherebetween1to7days.

HybridsystemsA recent concept combines an external occulter with an internal coronagraph, termed a“hybrid” system. Here, the occulter is designed to achieve enough suppression so thatopticalaberrationsdonotadverselyimpactperformance,butisnotintendedtocreateallofthecontrast.Thissuppressionisestimatedtobeontheorderof10–6.Acoronagraphisthenusedfortheremainingcontrast.Suchasystemmayhavetheadvantageofstilleliminatingthe need for wavefront control yet requiring a smaller occulter flown closer to thetelescope. Current estimates predict a 36‐m tip‐to‐tip occulter at only 50,000 kmseparation.Currentstudiesareaddressingwhetherproperdesignofsuchhybridsystemscan relax the tight manufacturing and deployment requirements. Nevertheless, hybridscomeat theexpenseofa slightlymorecomplicatedopticaldesign, lower throughput, andtighterpointingandstabilityrequirements.

3.2.3 PerformanceScaling

GroundbasedandspacebasedparameterspaceWediscussbriefly thecomparisonbetweenground‐basedcapabilitiesandaprobe‐size inspace,foraprojectedtimescaleofabout15yearstodemonstratetheircomplementarities.

Ground‐based adaptive optics coronagraphs on large telescopes (VLT, Gemini), will becapable of detecting young giant planets around nearby young stars with observationsstartingin2011.Theseinstrumentswillbemostsensitivetoplanetsatwideseparations(5to50AU)andthuswillnotprobesystemarchitecturesatsmallerorbitalradii.

OpticalImaging

61

ExtremelyLargeTelescopes(ELTs)(e.g.,GSMT,E‐ELT)shouldseefirst lightaround2018,and high‐contrast instruments can be anticipated on these facilities around 2020 at theearliest. InFigure3‐4weprovideastylizedcomparisonoftheparameterspaceaccessibletoground‐basedandspace‐basedinstruments.Aprobe‐sizecoronagraphhassomeoverlapwith an ExAO ELT for mature giant planets in reflected light, therefore extending thewavelengthcoveragewithtypically0.4to1.1µminspaceand1.0to2.4µm(J‐H‐Kbands)ontheground,andalsoaround3.8µm(L’)and4.8µm(Mband).

A30‐mclasstelescopeonthegroundprovidesenoughangularresolutiontoreachdistantplanet‐forming regions to study planet formation at moderate contrast (e.g., Taurus). Aprobe‐size coronagraph in space offers a modest angular resolution due to its smalldiameter, but a very high contrast 10–10 opening the possibilities for characterization offaintreflectedlightJoviansdowntoSuper‐Earths.

Gemini Planet ImagerTMT Planet Formation Instrument

1.5 m Probe (visible)

Figure 35. Complementarity of ground‐based and space‐based parameter spaces for the scales of missionsconsideredinthischapter.(Macintoshetal.2006;R.Soummer,STScI,andM.Levine,JPL)

ScalingofexozodiacalsensitivityInprinciple,abrightexozodibackgrounddoesnotmakedirectobservationsofanexoplanetimpossible; it just makes the exposure time longer to reach a given SNR. In practice,however,itprohibitsobservationsofexoplanetswhoseexposuretimesbecomelongerthanthe observing season for those planets. Also, this is further complicated if the exozodiexhibitsstructuresandclumpiness.

TheparameterEZdefinesthedensityofdustintheexozodi:thetotalexozodiacal+zodiacalbackgroundisthen(2+1)zodis,wherethefactor2accountsforthefactthatweareseeingthroughthewholeexozodidisk,whereaswesee throughonlyhalf thezodiacalcloud.Weassumethemedianofrandominclinations,(60°)foralltargets.

Chapter3

62

Inthisexample,weassumeatelescope4mindiameter,withIWA=72mas,andcapableofdetectingEarth‐likeplanets26magfainterthanthestar(Δm=26mag).Brighterstarstendtohavebrighterexoplanets,andlessluminousstarshavecloser‐inhabitablezones,whereanEarth‐likeplanet receivesandreflectsagreater fractionof starlight.Theplanet count‐raterelationisaffectedbydistance,asshowninFigure3‐6.Incontrast,theexozodicount‐rateisindependentofdistance.

Witha4‐mtelescopeandevenwithnoexozodi,thebackgroundisasbrightasEarth‐sizedplanetsorbitingstarsonly~10parsecsaway.Withanexozodidiskhavingadustdensity3times that of the zodi, the exozodi background is ~8 times brighter than the planet. Anobservation of such a planetary systemwould be background‐limited, so exposure timesscalewithtelescopediameter,D–4.

Theincidenceofexozodiacallightisunknownatalevelbelowhundredsofzodis.Aprobe‐sizecoronagraphwouldsolvethisissuebycharacterizingexozodiacaldisksdownto1zodiat3AUandprovidethecriticalinformationneededtoscaleaflagshipmissionaccordingtothisparameter. Inparticular, itwillbringsome informationabout thedisksstructureandclumpiness,whichcanhamperthedetectioncapabilityfurther.AsconcludedbyExoPTF,amuch larger telescope may be needed if the exozodi level is high (> 10 zodi). As SNRcalculation show (Figure 3‐7), and it is important to keep options open, in particular intermsoftechnologydevelopmentsdiscussedinthischapter.

0 10 20 30Distance (pc)

100

101

102

103

104

105

Coun

t-rat

e (c

nt/h

r)

EZ=0

EZ=0.75EZ=1.50EZ=3EZ=6EZ=12

Figure36.Count‐ratesofEarth‐likeplanetsintheHZaroundnearbystars(coloredcircles)andzodi+exozodibackground (blackdiamonds) for various levels of the exozodi dust density and a4‐m telescope.Blue to redmeansTeff=6600to4000K.(S.R.Heap,NASAGSFC)

OpticalImaging

63

1.5 m 4 m 8 m 16 m

0 10 20 30 402

3

4

5

6

Distance !pc"

logexposure!sec"

forSNR!10

R ! 100 Spectroscopy " 500nm of Earth at Dichotomy around Sun

Figure 37. Exposure time for the range of telescope sizes discussed in this chapter to achieve R = 100spectroscopyat500nmonanEarthatdichotomy(half‐phase)aroundtheSun.Foreachdiameter,threeexozodilevels are plotted, 1 zodi (Dark blue), 10 zodi (lighter blue) and 50 zodi (lightest blue). (R. A. Brown, andR.Soummer,STScI)

ScalingofflightsystemrequirementsandtechnicalmaturityThescalingof flightsystemrequirementsandtechnologymaturity isshown inFigure3‐8for the internal and external coronagraphs. The intent of this figure is to show how thevariousmissionscalesimpacttechnicalriskandreadiness, irrespectiveofthemeritofthescience.Thefirstcolumnliststhemaintechnologiesofinterest,withdemonstrationofthestarlightsuppressionphysicstoflightlevelsbeingthemostimportant.Therawcontrastisdefinedas the ratiobetween thePSFbrightnessat the locationof theplanet and thePSFbrightnessatthePSFcore.Thesubsequentcolumnscorrespondtorepresentativemissionconcepts, and are not an exhaustive list of possibilities. Numerical values, whereappropriate, represent required performance goals for flight and how they scale withmissionsize.Colorsfromgreentoredrepresenttechnologyriskandmaturityfromlowtovery high. These are evaluated as a function of observatory size and are an indication oftechnologydevelopmentneed.Section3.3describesthesetechnologiesandtheirproposeddevelopmentinmoredetail.

3.2.4 CostandRiskDriversBecause of current constraints on theNASA budget and concerns about cost overruns ofmissions currently in development, providing accurate mission costs is a criticalconsideration for future exoplanetmissions. The on‐going Astrophysics StrategicMissionConceptStudies(ASMCS)willofferafirstopportunityforcostingthesemissions,whicharebeingcategorizedaseitherprobe‐scale(inthe$600M–700Mrange,withoutlaunchvehiclecosts) or flagshipmission (for $1B ormore).While the ASMCS costing exercises are notcompleteatthistime,wecandescribetherationaleforthemissionscalesandidentifythemaincostdriversforeachmissionoption.

Telescopesystem:Basedonamultitudeofpastflightobservatorymissions,wecanexpecta 1.5‐m class telescope system at the probe‐scale cost, including instrument and a singlespacecraft,whileflagshipmissionswill includesystemslargerthan2.5morwithmultiple

Chapter3

64

spacecrafts.Thetelescopeopticswillbediffractionlimitedandofthesamequalitywhetherthey will be used for an internal coronagraph or external occulter, and thus not adiscriminator.Theinternalcoronagraphwillcallforanoff‐axismirror.

Systemstability: Internal coronagraphcontrastperformance isvery sensitive to thermalandjitterstabilityduringawavefrontcorrectioncycle,andthissensitivityincreasesastheIWAdecreases, say from4λ/D to2λ/D.Thismeans thatwhilea2λ/D coronagraphcanfindandcharacterizeaboutthesamenumberofplanetswithatelescopeabouthalfthesizeasoneoperatingat4λ/D,itwillrequireatleastanorderofmagnitudetighterrequirementonpointingstabilityandthermaldrift.Onewaytominimizetheimpactofthecostdriveristodevisemethodsthatcansenseandcorrectthewavefrontrapidly.

Coronagraph instrument including wavefront sensing and control: This is a coreelementofinternalcoronagraphs,whichisnotapplicabletotheExternalOcculter(butistotheHybrid).Costneedstoincludethenumberofdeformablemirrorsandthecomplexityofthe instrument optics. Note however that the size and complexity of the coronagraphinstrumentsystemremainessentiallythesameregardlessofmissionsize.

Twoormorespacecraft:This isacoreelementof theExternalOcculterandtheHybrid,which does not apply to internal coronagraphs. This entails a whole separate spacecraftsystem including bus, power, communications, integration & test, software, etc.Furthermore, this particular application requires special propulsion systems such as ionthrusters,whichaddmassandcomplexity,aswellas25‐mto50‐mdeployablemulti‐layermembranestructures,formationflyingmetrologyandalignmentcapabilities.

Endtoend verification and validation (V&V): The cost of final system testing on theground will clearly grow as the size of the mission, and will become prohibitive, if notimpossible, foravery large internalcoronagraph flagshipmissionor foranysizeexternalocculter.V&Vwillthenhavetorelyonveryhighfidelitymodels,whichwillbederivedfromextensivesub‐componentorsub‐scaletestingandmodelvalidationactivities.

Thermalcontrol:Willbeaconcernforeitherinternalcoronagraphsforwavefrontstability(seeabove)ortheexternalocculterformembraneshape.

3.3 TechnologyWe discuss various technologies needed and their technical readiness, and we proposematurationprogramsforinternalcoronagraph,externalocculterandcommontechnologies.PastTPFactivitiesprovidesignificantheritage, includingongoingmissionconceptstudies,infrastructureforcoronagraphtesting,andanalysistools.Thistechnologyheritageremainsdirectlyapplicabletocurrentinterestsinprobe‐scalemissionsandexternalocculters.

ThecurrentstateoftheartforstarlightsuppressiontechnologyisgiveninFigure3‐9andincludesalltheapproachesthatarediscussedherein.Notethatmanytechniqueshavenotyet had the opportunity to be tested in a precision infrastructure such as the JPL HighContrast Imaging Testbed (HCIT) (Trauger et al. 2007) and that future technology fundsshouldbeappliedtosuchtesting.

OpticalImaging

65

Figure 38. Summary of technologies for internal and external coronagraphs as a function of size. Columnscorrespondtorepresentativemissionconcepts.(M.Levine,JPL)

Chapter3

66

Figure39.Currentlaboratoryresultsinstarlightsuppression.(M.Levine,JPL)

A crucial parameter in determining the scope, cost, and capabilities of a coronagraphicplanet‐findingmissionistheinnerworkingangle(IWA).ChangingtheIWAfrom4to2λ/Dreduces the required telescope size by nearly a factor of 2 and improves the mission’sefficiencytofindandcharacterizeplanets.However,therearepracticalconsiderationsforachieving contrast at lower IWA’s, which is a function not just of the coronagrapharchitecturebutcomplexwave‐opticseffects,opticalbandwidth,telescopeandinstrumentstability, etc. An IWA of 4 λ/D has been demonstrated in the laboratory at performancelevels nearly sufficient for detection of Earth‐like planets. Several coronagraphs with anIWAof2–2.5λ/Dhavebeenproposed,buttodatethedemonstratedlaboratoryresultsarestillordersofmagnitudeawayfromthepredictedtheoreticalperformance.OneofthemainreasonsisthattheseaggressivecoronagraphsaremoresensitivetostabilityandthattheyhaveyettobetestedinastablevacuumenvironmentsuchasthatprovidedbytheJPLHCIT.Similar challenges exist with external occulters with respect to their modeling, theirreduced‐scalelaboratorydemonstrations,andtheirtraceabilitytofull‐scaleperformance.

These questions must be resolved prior to selecting any future space direct‐imagingmission. Significant resources should be committed to both laboratory experiments andmodeling—not just simple coronagraph models, but advanced wave‐optics models andintegrated models of entire observatory systems which tie in the contrast capability tosystem thermal and jitter stability. As recommended by the ExoPTF, a blue‐ribbon paneldrawnfromtheengineeringandoptical‐sciencescommunitiesshouldevaluatethestatusofthese models and demonstrations before any mission is allowed to proceed intodevelopment.

3.3.1 InternalCoronagraphTechnologyInternalcoronagraphsconsistofasinglespacecraft.Anexampleofmissiondeploymentofacoronagraphic observatory is given in Figure 3‐10. Specific technologies for internalcoronagraphs include coronagraphic masks, Deformable Mirrors (DM), and Wavefront

OpticalImaging

67

Sensing andControl (WFS&C). It is important to note that these coronagraph instrumenttechnologiesand their requiredperformancegoals remainessentially thesame foranyofthemissionscales.ThemostdetailedcoronagraphicstudyforaspacemissionistheoriginalTerrestrial Planet Finder Coronagraph (TPF‐C) Flight Baseline design 1 (FB‐1), whichinvolvedaneighth‐orderband‐limitedmask.Thisconceptisstillaviableoptionforan8‐mflagshipcoronagraphusingexistingEELVlaunchvehicles.StudiesarenowlookingintotheopportunitiesofferedbytheAresV.

CoronagraphconceptsA large number of coronagraphic concepts have been developed in the past few years.Several studies (e.g., Levine et al. 2006; Guyon et al. 2006) have detailed and comparedtheseconcepts.Coronagraphscanbeorganizedinafewcategories.

Pupilapodization:Apodizationcanbeobtainedusinganamplitudetransmission(ShapedPupils),orPhaseApodizationCoronagraphy(Codona&Angel2004).Anotherapproachforpupil apodization involves pupil remapping, or Phase Induced Amplitude Apodization(PIAA).

Lyot coronagraphs: These coronagraphs use a series ofmasks including at least a focalplanemask followedby a pupilmask (Lyot Stop). Themasks can be based on amplitude(BandLimitedCoronagraphs,ApodizedPupilLyotCoronagraphs)orphase(RoddierPhaseMask,FourQuadrantsPhaseMask,EightOctantPhaseMask,OpticalVortexCoronagraph).

Interferometric coronagraphs: These coronagraphs use an amplitude divisioninterferometer to create a null from the combination of beams (Achromatic Interfero‐Coronagraph,VisibleNullingCoronagraph,TandemCommon‐PathAIC).

Other techniques: In addition, several techniques have been proposed and studied toenhancecoronagraphicperformanceorcombineitwithparticularWFS&Cschemes.

OpticaldiffractionmodelingandlaboratorydemonstrationOptical diffraction technology developments depend in part on the type of coronagraph.Modeling efforts are still on going to refine the current capabilities, and technologydevelopmentinthisareawillcontinueinthefuture.Coronagraphmodelingandvalidationisoneof therequiredmilestonesdefined in theTPF‐C technologyplan fordemonstratingmissionmaturity.

ForLyotcoronagraphs,modelsareexpectedtoworkaccuratelydownto10–10contrastsandbetter,buthavenotbeenfullyvalidated.Polarizationisaconcern.SimulationsforaBand‐LimitedMasksystemwithwavefrontcontrolhavebeendemonstratedattheHighContrastImagingTestbed(HCIT)(Figure3‐11)withabest‐achievedcontrastof5.2×10–10at4λ/Dfor760–840nm(10%band)innaturalunpolarizedlight.

ForShapedpupils,modelsarealsoexpectedtoworkaccuratelydownto10–10contrastsandbetter. Vector propagation analysis comes into playbelow 10–9 contrasts (Ceperley et. al.2006),butdoesnotaffectperformanceforapertureslargerthan25mm.Tinymaskdefectsareimportantatthesecontrasts;however,therearenotheoreticalshowstoppers(Belikovet al. 2007). The best contrast obtained at HCIT is 2.4 × 10–9 at 4 λ/D for 760–840 nm(10%).

Forphasemasks(OVC,EOPM,FQPM)modelingisneededtoverifymodelsto10–10contrastlevels, and to integrate the modelswith wavefront control to investigate broadband

Chapter3

68

performancewithaberrations.Aproofofconcepthasachieved2×10–4at4λ/DwithOVC,and5×10–5at3λ/DwithEOPM(Murakamietal.2008).PhaseApodizationCoronagraphyhas been demonstratedwith on‐sky observations at theMMTObservatorywith amid‐IRcameraandacontrastof10–4at3λ/D(Kenworthyetal.2007).

Figure310.ExampleofdeploymentforaFlagshipcoronagraphmission.(Levine,Shaklan&Kasting2006).

For PIAA, accurate modeling of diffractive effects between the two mirrors even withwavefront errors (Belikov et al. 2006) has been demonstrated numerically for simplegeometries and perfect pre‐and post‐apodizers. Propagation formore realistic cases (off‐axismirrordesigns,focal‐to‐focalPIAA,errorandaberrationsintheapodizers)remainsanopenproblematthe10–10levels.Broadbandwavefrontcontrolat10–10levelsusingtwoDMshasbeendemonstratedinnumericalsimulationsinthecaseofaperfectPIAA(Pueyo2008).Thebestcontrastobtainedinthelaboratoryis6.7×10–7at2λ/Dformonochromaticlight.

For Nulling Coronagraphs (VNC, TCP‐AIC), models need to be verified to 10–10 contrastlevels. The VNC, especially with fibers, cannot be accuratelymodeled using conventionalFourier techniques.A testbedatGSFC incollaborationwith JPLandLockheed‐Martin,hasachievedwhite‐lightnullingto10–4 inthepupilplaneand10–7 inthefocalplaneforshortperiods of time and is now being prepared for vacuum operations. The Null ControlBreadboardisaseparateexperimenttotestsegmenteddeformablemirrorsfrom(IRIS‐AOand Boston Micromachines) coupled to coherent fiber bundles, and to develop,demonstrate, and characterize the null control algorithms for the deformable mirrors.Recently,numericalsimulationsshowedthatTCP‐AICcouldreach1×10–10contrastlevelswith achromatic four‐beam interference (Tavrov et al. 2008), but accurate modeling isneededforpolarizationelements.

OpticalImaging

69

Ground‐basedcoronagraphiceffortsaremainlyfocusingonFQPMandAPLCs,respectivelyonVLT‐SPHEREandGeminiPlanetImager.Bothtestbedresultsachievedsimilarcontrastofabout10–5at6λ/DfortheH‐band(1.5–1.8µm)withoutwavefrontcontrol(Boccalettietal.2008;Macintoshetal.2008)andAPLCachieved5×10–7at5λ/Dat532nm(Thomasetal.2008).

DeformablemirrorsDeformablemirrors(DMs)areacriticalcomponentofspecklesuppressionmethodologiesfor all wavefront‐control architectures under consideration. The coronagraph applicationwill demand more of a DM than previous applications: the DM surface needs to bepositionedtolessthan1ÅRMSofthedesiredshape,andbestableto0.3ÅRMSforanhour.Twocurrent technologiesareviable forwavefrontcontrol:onemadebyXinetics Inc.,andthe other a MEMS device made by Boston Micromachines Corporation. A segmenteddeformablemirrortechnologymadebyIRIS‐AOisalsousedfornullingcontrolwithNullingCoronagraphs.

Figure311.Contrastin760–840nm(10%)bandwidthobtainedatHCIT.(Moody,Gordon&Trauger2008)

XineticsDMsarebuiltup fromelectro‐ceramicblocks, each incorporating1024actuatorsona1‐mmpitch.Theseblockscanbeassembled intomodulescoveredbyasingle‐mirrorfacesheet and driven by a multiplexed voltage supply with 100‐V range and 16‐bitresolutionforamaximumstrokeof0.2µm.

MEMSDMsaremadeofapolysiliconmembranecoatedwithoneormoremetallayersforthereflectivesurfaceandareactuatedby32×32or64×64electrostaticactuatorsonthebackside. The maximum stroke is about 5 µm for 250 V, and fabrication reliability hasincreased to about one out of three devices without any dead actuators. Currentdevelopmentsexisttoimprovesurfacequality(bow,scallopingandprint‐through).

XineticsmirrorsareusedatHCITandarecurrentlyatahigherleveloftechnologymaturitythanMEMSdevices.However,MEMSdevicesareofincreasingimportancebecauseofmassandcostconsiderations.Inbothcasestwoareasrequirefurtherdevelopments.Thefirstismodeling of the DM. This includes characterization of thermal‐elastic properties forhysteresisandothernon‐linearmechanicalphenomena(e.g.,yield limits,yieldproperties,creep, non‐isotropic material elasticity, and thermal expansion). The second are theelectronicsthatpresentlyarelargeandinefficient;amajordevelopmentisrequiredinthistechnologysimilartowhathasbeendoneforspacecraftelectronics.

Chapter3

70

WavefrontsensingandcontrolalgorithmsWedetailbelowtheaspectsofWFS&Cforspaceobservations.Ground‐basedExAOforhigh‐contrastimaginghasdifferentsetsofrequirementsforthecompensationoftheatmosphere(Macintoshetal.2008).TechnologydevelopmentsareneededforthedevelopmentofExAOon large segmented apertures and for speckle‐reduction active methods to compensatequasi‐static speckles (Wallace et al. 2006). These methods are comparable to what isstudiedforspace,butforlowercontrast.

Architecture: In principle, a single DM is sufficient to correct phase errors at allwavelengths.However,propagation‐inducedamplitudeerrors(orotheramplitudeerrors)can only be corrected perfectly at one wavelength. Better results from a single DM andwavefrontsensinginthefocalplanecanbeobtainedbyminimizingthecorrectedintensityfor several wavelengths (Figure 3‐11). In broadband or for larger errors, it may not bepossible tomaintain the contrast performance at the 10–10 levels. A secondDMwill helpdramaticallybroadenthebandandrelaxtherequirements(Shaklanetal.2006).Moreover,asecondDMaddsredundancytotheinstrumentarchitecture.

Control:SeveralalgorithmshavebeendevelopedtocontrolthesurfaceofDMsforagivenestimate of the field in the image plane. Current algorithms solve a simplified linearproblem and iterate to solve the complete non‐linear problem. The speckle‐nullingalgorithm proved the concept, but its convergence is too slow. Faster convergence isobtainedusingNewtonmethods.

Bordé&Traub(2006)showedthemathematicalformalismtotackletheproblemwithoneDM, using energy minimization but limited in practice by the modes hidden behind thecoronagraphicmask.Give’onetal.(2006)generalizedthisformalismbyrecognizingthatallcoronagraphs were linear operators and that the problem could be regularized usingstandardleast‐squaretechniques.ThisisthealgorithmcurrentlyusedatHCIT.

RecentlyPueyo(2008)proposedavariantofthisregularization,withaminimizationoftheDM deformation while constraining the contrast below 10–10. This algorithm was usednumericallytoshowhowtocontroltwosequentialDMswithbothPIAAandShapedPupilsinbroadband.

Nishikawaetal.(2006,2008)proposedanothermethod,whichrelaxestherequirementofwavefront accuracy of the entire optics by an order ofmagnitudewith a combination ofunbalancednullinginterferometerandtwoDMs(UNI‐PAC).

Overall,theproblemofcontrolwithaperfectestimatehassolidtheoreticalfoundations.Werecommendsupport for furtherdevelopment in thisdirectionsothat futurehigh‐contrasttestbedshavethehardwaretooperatewithatleasttwoDMs.

Estimation: In order to avoid non‐common path errors, the estimation has to followsuppressionofthebulkofthestarlight.Somediversityisneededtocalculatethefieldsfromintensitymeasurementsandseveraldifficultiesneedtobeconsidered:First,ifthe“sensing‐exposures”arenotused for science,missionefficiency isdecreased.Algorithms thereforeneed tobeoptimized touse the fewestpossiblediversityexposures tomaximize time forscience. This problem has been partially tackled by Give’on et al. (2006) (pairwiseestimation), Kay et al. (2007) (modified Gerchberg‐Saxton) or Pueyo (2008) (Hermitianproperties and wavelength fit of the estimate). When the target contrast is reached, allexposurescanbeconsideredasscienceexposureasshownbyBelikovetal.(2007).Second,sensing involves solving an inverse problem and is therefore very sensitive to modelingerrors.Themodeloftheforwardtransferfunctionneedstobeknownatalevelconsistent

OpticalImaging

71

with10–10contrast;weneedtoknowtheactualsurfaceoftheDMsandnotjustthevoltagesent.

Inconsequence,algorithmsneedtohavealowsensitivitytosomeerrorinthediversity,orneedtobeself‐calibratingtoadjusttoslowvariationsinthesystem(DMinfluencefunctionsfor example). The next challenge is to design new estimation schemes that are relativelyinsensitive to the system modeling, and minimize the amount of observation downtimeduring “sensing‐exposure.” The newly proposed UNI‐PAC might reduce most of thedowntimeexceptforafirst‐sensingexposure.

CoronagraphopticsmanufacturingWe organize the manufacturing technologies in four classes: binary masks, amplitudemasks,phasemasks,andasphericsurfacepolishing.

BinaryMasks:Thesemasks includeShapedPupils,orhard‐edged focal‐planemasks,andalso the “Lyot stops” used in Lyot‐style coronagraphs. The common technique used forthesemask is deep reactive‐ion etching (DRIE). An example is given in Figure 3‐12 for aseries of shaped pupilsmanufactured at JPLMicro Devices Laboratory (MDL) on silicon‐insulator wafers. Currently, the smallest features that can be manufactured reliably areabout5µmthrougha50‐µmsubstrate.This issufficient fora10–10contrastShapedPupil10mmor larger in size. Formasksof about10mm to30mm in size, the challenge is todeveloptaperedsidewalls,whichhasbeendemonstratedinthemicrofabricationindustrybutnotonshapedpupilsperse.Thisislessofanissueiflargermasksareused(>30mm).

Figure312.ShapedPupilsusingDRIEinSiliconwafers.Theinsideofeveryclosedblackcurveisetchedout.Thesmallestmanufacturableholesare~5µmwide.(CourtesyPrincetonUniversity/JPL)

Amplitudemasks:Thesemasks includesoft‐edgedfocal‐planemasks(e.g.,BLC),orpupilapodizations(APLC).Inmostcasesagray‐amplitudetransmittanceisneeded,andthemaindifficultyistocontroltheamplitudeandinducedphaseshiftinbroadband.ThephaseshiftintroducedbythemaskcanbecorrectedwithaDeformableMirror(DM)atonewavelength,butbroadbandcorrectionisanissue.TheresultsofFigure3‐11havebeenobtainedwithanickel‐deposit mask, shown in Figure 3‐13. However, only 1‐dimensional masks arecurrentlymanufacturable.

Chapter3

72

Mask‐manufacturing methods have been proposed to achieve a mostly phase‐neutralabsorbingoccultermask,forexampleusingametaldepositwithaprofileddielectriclayertocompensateforphasechanges(Moodyetal.2008),orusingafocusedionbeamtoshapedye‐doped acrylic structures embedded in transparent phase‐matching glass (Tolls 2005;Tollsetal.2006).Ion‐beammachiningenables2‐dimensionalmasks.

Another approach for amplitudemasks isbasedonbinarynotch filtermasks (Kuchner&Spergel 2003; Balasubramanian 2008; Crepp et al. 2006), or half‐tone approaches(Vasudevanetal.2005;MacIntoshetal.2008).E‐beamlithographyisusedtowriteabinarypatternofblockingandtransparentareasontoasubstrate.

Figure 313: Nickel is vacuumdeposited on fused silica (left), and comparison of the specified transmissionprofile(red)andmeasured(blue).ThisisthemaskusedintheresultofFigure3‐11.(D.Moodyetal.2008)

Half‐tone methods use the statistical distribution of opaque dots to achieve the desiredextinctionprofile.Thesemasksseemtoworkwellinlaboratorytesting(Thomas,Soummer&Dillon2008).Theirmaximumopticaldensityof5 isan issue forBLC, so theyaremoreappropriate for pupil apodizers. A potential technology issue is due to the presence ofsurfaceplasmonswhenthedotshaveacomparablesizetothewavelength.Thiscanresultin “super‐transmission effects” (Genet & Ebbesen 2007) with wavelength‐dependentamplitudevariations.Theseeffectscanbemitigatedbytheoptimizationofthedotsizes.

Figure314:Half‐toneapodizerusing2‐µmblack‐chromedotsonglass(e‐beamlithography).Theapodizeris12‐mmdiameter.(R.Soummer,STScI,A.Sivaramakrishnan,AMNH,andDarenDillon,UCSC)

OpticalImaging

73

PhaseMasks:Examplesofmanufacturingprocessesarethelift‐offtechniquetoproduce4‐quadrant phase masks (Riaud et al. 2003), and more recently using half‐wave plates toproduce an achromatic FQPM, where four half‐wave plates are assembled side‐by‐side(Mawetetal.2006;Boccalettietal.2008).Phasemasksforvortexcoronagraphshavebeenmanufacturedusinganelectronbeamlithographictechnique(Swartzlander,Ford&Abdul‐Malik2008).

Aspheric surface polishing: aspheric surfaces are needed particularly for the PIAAcoronagraph.Prototypeopticshavebeenmanufacturedusingasphericdiamondturningofbothlensesandmirrors.

3.3.2 ExternalOcculterTechnologyAnexternaloccultermissionconsistsofatleasttwospacecrafts:thetelescopeandatleastoneocculterorstarshade.AnexampleofapossiblemissiondeploymentisgiveninFigure3‐15.Thetelescopedoesnotrequireanyparticularlychallengingtechnology,andwedetailthetechnologyrelatedtothestarshadeitselfandtoformationflying.

TheOcculter technologyassessmenthasbeensubdivided intoeightsub‐technologyareas,which are given below in order of importance. These are modeling, occulter systemmaturation, deployment, formation‐flying and pointing‐control subsystem, precision edgeand scattered light, integrated opto‐mechanical‐thermal performance and analysis,micrometeoroidsimpacts,andpropulsionsystems.

OpticaldiffractionmodelingandlaboratorydemonstrationSeveral modeling and optimization schemes for the shape of the starshade have beendeveloped by groups at Princeton, University of Colorado, Goddard, and NGST, based onFresnelpropagation.Althoughmodelinghasnotbeenvalidated at10–10 contrast andwillneedfurtherstudy,currentmodelsagreewiththetestsatthe5×10–8level.Modelingsmallerrorsontheocculterrequiresnon‐Fouriermethodsthatcomputepiecemealpropagation.

Figure315:ExampleofapossibleTPF‐Omissiondeployment.(A.Lo,NGST)

Chapter3

74

Severallaboratoryexperimentstodemonstratetheconcepthavestartedwith,forexample,small‐scale(centimetersize)starshadesatUniversityofColorado,asshowninFigure3‐16,NGST, and Princeton. Research is still needed to understand how optical and contrastperformancescalefromthesmalltestarticlestothefullsizestarshade. Themainissueisthat a full‐size test of the system will not be possible on the ground. Engineeringdevelopmentsareneededaswelltoenableend‐to‐endsystemverification,includinghigh‐fidelity models for mechanical and thermal behavior from sub‐scale or sub‐componenttests. The effects of gravity on such large systems further complicate ground verification.Modelingisalsonecessarytopredictandverify,priortolaunch,thealignmentcapabilityofthetwospacecraft.

Figure316:Occulterexperimentusingastarshademadewith lithography.Thecontrastobtained is5×10–8withbroadbandsunlight.(Levitonetal.2007;Schindhelmetal.2007)

OccultersystemmaturationSeveral approaches are possible to define the Occulter system maturation. A proposedapproach is summarized in Figure 3‐17. The initial proof of concept is based on thefabricationofasinglesegment.Thisenablesthedefinitionandtestingofthelargerangeofmaterial properties, and thedeployment and shapeaccuracy.A three‐petalmodel is thenfabricatedanddemonstratedfromambientconditionstorelevantconditionsofvacuumandtemperature.Aflowchartrepresentationofaslightlydifferent,butcompatible,maturationscheme is given in Figure 3‐18 detailing a technology development and test approach toreachTRL6by2012forthestarshade.Thisapproachwillvalidateoccultertechnologyandsystemmodelingonasubscaletestarticleinambientconditions.Werecommendcarryingout a detailed study, with further testing in relevant environments, using a microsat ornanosat spacevehicle foron‐orbit demonstrationof the complete subscale systemwith asizeof8–30m(thefullscalecanbeupto50m),orpossiblyusingsecondarypayloadsonAtlasorDeltalaunchers.

DeploymentSinceavailable launchvehicleshavefairingdiametersrestrictedto lessthan5m,externaloccultersmustbefoldedduringlaunchanddeployedaftertheyreachorbit.

Thetechnologychallengeswithdeploymentresideinhowwellthetipandedgefigureoftheocculter can be obtained since it is the outline of the deployed configuration, whichwilldrive theultimate contrast performance.Deployed edge accuracy is required to be about1mmfromnominalconfiguration.Thedeploymentofanexternaloccultercanpossiblybeaccomplished a number of ways as summarized in Figure 3‐19, mainly distinguishingbetweenmonolithicorhybriddeploymentoptions.

OpticalImaging

75

Figure317:Occultersystemmaturation(D.Tenerelli,LM)

Figure318:Flowchartrepresentationoftheoccultermaturation(A.Lo,NGST)

Monolithicdeploymentisthesimplestprocesswheretheopticalelementisdeployedviaasinglemethod as a whole structure. However, there is no simple solution for tensioningmembranesasrequiredbytheshapeoftheoccultertipanditspetals,andaseparateedge‐supportstructuremayberequired.

Hybrid deployments allow separating the deployment of the core structure, which hasrelaxed requirements from the tip andedgeswithhigher accuracy requirements.Optionsincludeacentraldiskusingprovendeploymentmethodsforcircularstructures,followedbyasecondstageforpetaldeployment.

Chapter3

76

Figure319:Occulteddeploymenttradespace(A.Lo,NGST)

AnexampleofamonolithicdeploymentisgiveninFigure3‐20,wherethetipsofthepetalsareattachedtothetipoftheinnermostofseveralnestedtelescopingbooms.Eachpetalhasitsownsetofstemdriveandtelescopingboom.Theboomsarenestedduringlaunch;onceonorbit,acontainmentshieldisreleased,andthestarshadefabricunfoldslikeanumbrella.Motordrivenstemdrivespushouttheboomsegments.

Figure320:Representativestarshadeandspacecraftarchitecture(A.Lo,NGST)

Rigidizableinflatableocculterscouldprovideanalternatedeploymentapproachwithmassadvantages. A 20‐m solar sail demonstrator (Figure 3‐21) can provide a technologypathfinderbenefitting theocculterpetalsystem.Thedeploymentandstructural testingofthe 20‐m solar‐sail systems was conducted in the 30‐m diameter Space Power Facilitythermal‐vacuum chamber at NASA Glenn Plum Brook in April though August 2005. Themassofthissolarsailis0.25kg·m–2.

Figure321:APrototype20‐mSolarSailatNASAPlumBrookStation.(NASA/MSFC)

OpticalImaging

77

FormationFlying&PointingcontrolsubsystemThestarshademustbealignedonthestarforthefullobservation,towithinafewmas.Therequirementsfortherigidbodymotionoftheocculterarecurrentlyworkinprogress.Bestestimationoftherequirementsis1minplaneand±1degreeintilt(Lyonetal.2007).Thistoleranceisanengineeringcompromisebetweenstarshadesize(largershadeequalsloosertolerance and larger IWA) and alignment sensing capability. An angular sensor on onespacecraft can be used to measure this alignment offset. There are a few solutions forocculter acquisition and alignment maintenance during an observation. These include athree‐beaconconceptonthestarshade(R.Lyonetal.2007),oranastrometricsensorandin‐shadowsensors(Noecker2007).Foratypical20,000–100,000kmseparation,thesensorneedsprecision,stability,andultimateaccuracyof~2mas.

The pointing and control subsystem algorithms will have to consider the flexible‐bodydynamicsofthelargeocculterplusthemuchsmallerdeployablesolararrayandhigh‐gainantenna. Included in the control of the systemare thruster firings, reactionwheel forces,micro‐meteroidimpacts,thermaldeformations(whichcausethecenterofmassandcenterofpressuretovaryafterslewing),solarwind,solarpressureandgravitygradient.

PrecisionedgeandscatteredSunlightThe telescope will see the sunlit edges of the starshade as local bright spots, which canpotentiallyoverwhelmanexo‐Earthsignal.A“knife‐edge”treatmentontheperimetercanbe used tominimize the edge surface area to reduce scattering toward the telescope: anedgewitharadiusofcurvatureof100µmwillscatterlessthan30magnitudeequivalentofsunlight towards the telescope. The two associated issues are the knife edge itself andperimetercontrol.Technologydevelopmentsareneeded todemonstratehow tomaintainedge and perimeter quality during deployment and on‐orbit operations with possibletemperature changes through the structure from about 300 K to tens of K. Additionaldevelopment is needed tomitigate scattered light using highly non‐reflectivematerial onthethinedge.

IntegratedopticalmechanicalandthermalanalysisThermal analysis of the system can build on past experience ofworkingwith deployablemembranes(e.g.,JWSTtechnology).Factorsdrivingtheaccuracyofnumericalsolutionsforthermal analysis are well known in general. However, often‐neglected effects (such asmembranewrinklingandmaterialanisotropy)mayhavesignificantimpactontheaccuracyof a thermal,mechanical, and stray‐light prediction. It is important to develop integratedmodeling capabilities that will directly analyze the sensitivities of thermal variations onsystemcontrast tobetterdefinerequirementsonocculter thermalgradientsandstability.Modelingtoolsneedtobefurtherdevelopedtoincludemembranewrinklingandnon‐linearbehaviors. These integrated models are essential for optimizing occulter designs andparameters (e.g., layer separations, material properties, etc.) and for establishing thesensitivityofperformancetothermalandstructuralrequirements.

Thefollowingexample(Figure3‐22)illustrateshowthermalinformationcouldbeusedtoderive design requirements, showing the temperature profiles of the three flat layers ofKapton. In particular large temperature gradients exist for the central layer, which mayinduce large thermal deformations beyond required shape. Non‐uniform deformations oftheoccultercanbemitigated, forexamplebytheuseof low‐CTEgraphite fortheedge,sothat the center of the starshade is allowed to expand and wrinkle without affecting

Chapter3

78

performance.Thisexampleconsiders1‐mseparationbetweenthesheets.Shorterdistanceswould be more favorable for thermal properties, the shortest distance being defined bymicrometeoroidprotection.

MicrometeoroidsMicrometeoroid impactcancreate tinyholes in theocculter,buta triple layerofmaterialcanbeusedtolimittransmissionofstarlight(Cashetal.2006).It isnecessarytoevaluatetheeffectofmicrometeoroidimpactsintermsofstraylightanddynamics.Thefirststepistoquantify theenvironmentanddetermine therateof impactsperyearwithmass largerthan10–7gandvelocity20–50km·s–1.

Figure322:Thermalmodelingofathree‐layer30‐mdiameterOcculter.ThecolorkeyontheleftisinunitsofdegreeKelvin.(A.Klavins,LM)

AnalysisofcandidateoccultermaterialbasedonJWST’sshield(Si‐coatedKapton)hasbeenshowntowithstandtheL2environment.CrackpropagationdoesnotseemtobeanissueforKapton, because there is little material‐induced sputtering after the first layer, and themembrane isnothighlytensioned.However,sometechnologydevelopmentandmodelingareneededformicrometeoroidmitigation,asappliedtotheocculter.

PropulsionsystemsDue to the large distance traveled by the starshade between observations, it is notconceivabletouseconventionalpropulsionsystems,whichwouldrequire largequantitiesofpropellant.Sustainablepropulsionsuchassolarelectricpropulsionisanexcellentoptionforthisapplication.NASAGlennandJPLhavealreadydemonstratedthistechnologyin1998on the Deep Space 1 flightmission and are involved in ongoing technology developmentefforts.

Two types of contamination need to be evaluated: first, the optical effect of amolecularcloud on the seeing of the telescope; second, the effect of hydrazine deposition on theoccultersurfaces,whichcanpotentiallyaffectmechanicalandthermalproperties.

OpticalImaging

79

3.3.3 CommonTechnology

TelescopeandmirrortechnologyFourkeymetricscanbedefinedtoallowtheassessmentofmirrortechnology(Matthews,Egerman,Wynn,ITTSpaceSystemsDivision).TheirmaturityisillustratedinFigure3‐23.

Onaxis versus offaxis PrimaryMirrors: In on‐axis systems, the primarymirror has acenter hole and there is an obstructing structure that holds the secondary mirror withrespect to the primary mirror. Off‐axis telescope configurations are not symmetric withrespect to the center of themirror. The light is reflected to a non‐obstructing secondarymirror.However, the fabrication of off‐axis primarymirrors and telescopes has not beendemonstrated at very large sizes. Themost significant challenge between developing on‐axisandoff‐axistelescopesisinthefabricationandtestingoftheprimarymirror.Modernopticalandopto‐mechanicalmetrologydevices(suchaslasertrackers)reducethedifficultyinaligningoff‐axistelescopesystemssignificantly.

Figure323:TechnologyReadinessLevel(TRL)formirrorsizesandconfigurations.(Egerman,Matthews,ITT)

PassiveversusActivedesigns:Historically,primarymirrorshavebeenverystifftoallowtheobservatorytobepassiveintheoverallimplementationofthesystem.Theprimary‐to‐secondaryalignmentscouldbemadeinfrequentlybymerelymovingthesecondaryonsmallactuators to adjust its position. Sophisticated WFS&C schemes to control higher spatialfrequency errors were not easily implemented in either software or hardwarearchitectures. HST is an excellent example of this type of system. New technology insoftware,computerdesign,dynamiccontrol,andmirrormanufacturingtechniquesenablestheoptical figureof largemirrorstobeadjustedonorbit ifrequired.Theseactivedesignsarenewforspace‐basedapplications,buttheyhavebeendemonstratedontheJamesWebbSpace Telescope (JWST) to be feasible during ground testing. Recent advances inmirror‐processing technologies, used in conjunctionwith advancements in dynamic control,willenable large optics with relatively low stiffness to mass ratio passive or active primarymirrorsforfuturesystemsthatarenotsegmented.

Chapter3

80

Mirror quality: the primary mirror technology for exo‐Earth finding coronagraphs isalready in hand. Earlywork in coronagraph technology emphasized the need for a high‐quality primarymirror withmid‐spatial frequency (5–100 cycles/aperture) performanceexceedingthespecificationoftheHubbleSpaceTelescope.Recentlyithasbeenshownthatthe deformable mirrors in the coronagraph compensate for these errors and relax theoptical design specifications to levels not exceeding HST (Shaklanet al. 2006). Indeed,laboratory results from JPL’s HCIT show broadband performance at levels exceeding1×10‐9contrastusingoff‐the‐shelfqualityoptics(about1/20thwave).Itisnowunderstoodthatdetrimentalpropagationeffectsintheopticaltrainaremostimportantwherethebeamdiameter is reduced, and these effects limit the high‐contrast performance (1 × 10–10contrast)toanglesof20–30cycles/aperture(about2arcsecforaprobe‐scalecoronagraphmissionand0.3arcsecforaflagshipmission).Theexternaloccultersystemdoesnotsufferfromtheseouterworking‐anglelimitations.

Size:Smallopticshavebeenmanufacturedinmanysizesandshapes.Largespace‐qualifiedprimarymirroropticshavetraditionallybeenround.Largeoff‐axis,ellipticalmirrorshavebeenstudied,butnothingofanysizeapproachingthe3‐mby8‐msizehaseverbeenbuiltorqualified for space applications. The largest non‐round off‐axis optics for space‐basedapplicationthathavebeen/arebeingfabricatedarethe~1.4‐moff‐axishexagonalsegmentsmanufacturedduringtheAdvancedMirrorSystemDemonstrator(AMSD)Programandthesimilarly sized and shaped segments currently being fabricated for the JWST primarymirror.

DetectorsDetectorswillbeacriticalcomponentofanyexoplanetdirect‐imagingmission.Whetheritis an exoplanet probe or a flagship mission, it will be straining for photons becauseexoplanetsareveryfaint:about22.5magfainterthanthestarintheVbandforaJupiter‐likeplanet,andabout25magfainterforanEarth‐likeplanet.Theprocessofdispersingthelight of a faint planet overmany pixels of a spectrograph will cause the read noise of aconventional detector to become a limiting factor. As shown in Figure 3‐1, DRM studiessuggestthatforsomestars,thedetectioncanbelimitedbytheplanetflux(andnottheIWAorthecontrast).Detectorshavinghighdetectorquantumefficiency(DQE)intheopticalandnear‐IR(0.4–1.0µm),lowdarkcurrent,andverylowornoreadnoise(photoncounting)arethereforeessential.

PhotonCountingCCDs:Photon‐countingCCDsprovideelectronmultiplicationbeforetheirreadouts,essentiallyeliminatingtheirreadnoise(<1e–).Forexample,a typicalExoPlanetProbewithaphoton‐countingCCD(zeroreadnoise)willtakeonlyhalfaslongtoobtainanR=20spectrumofa Jovianplanet than if theCCDhasareadnoise3e–.Thebenefitsofaphoton‐counting CCD are even greater for a flagship mision, where exposure times forR=70spectroscopyofEarth‐sizeplanetsarecutdownbyafactorof5(Heapetal.2006).Thisgaininultimatesensitivityisequivalenttoincreasingthetelescopeaperturefrom4mto6m.Thetechnologyforthe0.4–0.7µmregionisalreadywellinhand(Wenetal.2006),with high‐DQE photon counting detectors commercially available (e.g., E2V’s L3VisionseriesCCDs).Werecommendthecontinueddevelopment,includingpushingtheirDQEevenhigher(>0.8)andimprovingtheirradiationtolerance.

An opportunity exists to enhance the DQE in the near‐IR by bringing photon‐countingcapabilities to p‐channel CCDs. Regular p‐channel CCDs already exist and have a DQE ofabout 0.8 at the 0.94 µm water line and at the ~0.9 µm CH4 features. We recommend

OpticalImaging

81

technology development to add a charge‐multiplication register to the p‐channel CCD,similartowhatalreadyexistsfortheregular(n‐channel)CCD.

LowTemperature Detectors: Low‐temperature detectors (LTDs) are a class of photon‐countingdetectorsthatareoperatedatverylowtemperatures,typicallylessthan300mK(Verhoeve 2008). In the optical, they allowphoton counting imaging spectrophotometers(thedetectormeasuresalsothewavelengthwithouttheneedforaspectrograph).WithanLTD,everyphotoniscountedindividually,andcosmicrayscanberejectedextremelywellbecause they deposit significantly more energy than an optical photon. Some LTDs areresponsivetolightbeyondthe~1µmcutoffofsiliconCCDs.Severaltechnologiesarebeingdeveloped, including doped semiconductor, Superconductive Tunnel Junctions (STJs),Transition Edge Sensors (TES), Magnetic microcalorimeters, Superconductive QuantumInterferenceDevice(SQUID),andMicrowaveKineticInductanceDetector(MKID).

ThemostadvancedSTJinstrument(S‐Cam,ESA/ESTEC)isa10×12pixelcamerawithR=15 at 500 nm. As long as nomultiplexing scheme exists for STJs, further increase in thenumberofpixelswillbelimited.

TESarrayshavebeendemonstratedforsmallsize(8×8pixels),anddevelopmentsexisttoexpand the wavelength range to 0.3–1.7 µm. However, multiplexed read‐out has not yetbeendemonstratedwiththesearrays.

AnadvantageofMKIDsistheirfrequencydomainmultiplexingallowinglargearraystobebuilt(Mazinetal.2006).However,theircurrentspectralresolutionisstillinsufficient(R=6demonstratedwith1000pixels), butplans exist todevelopa25,000‐pixel arraywithR=20–50.Technologydevelopmentsareneededtodeveloplargerarrayswithhigherspectralresolution.

PrecisionthermalcontrolandanalysisExisting thermaldesignsare capableof achieving sub‐milliKelvin (mK) control inbenignspaceenvironmentssuchas isofferedby theLagrangianorbits.CurrentlyavailablecodesandcomputationalcapabilityshouldallowthepredictionsofrelativetemperaturechangestoseveralmKs,buttheyarenotcapableofabsolutetemperaturepredictionstothoselevels.Factors that drive the accuracy of numerical solutions arewell known, and limitations incommercially available codes force the trading of numerical solution accuracy againstmodel sizeandcomputation time.Modeling toolsareavailable todealwithanisotropyoreffects such as spectral optical properties. Effects of heat‐flow mechanisms that arenormallyneglectedcanbeexamined,e.g.,theworkdonebythermalexpansionorlong‐termchanges in composite properties due to outgassing or radiation. Material propertiesuncertaintiesarereadilyaddressed,evenifthetaskofassessingtheseuncertaintiesgrowsrapidlywhenconsideringnotonlypropertymeasurementerrorsbutalsothemanypossibleways that such errors can be distributed throughout a complex system. Overall, there issufficient understanding of heat transfer physics to evaluate mechanisms, checkassumptions, identify uncertainties, and develop schemes for accommodating these in anend‐to‐endsystemmodel.

Thegreatestriskisinmanagingthesheersizeofthethermalmodels.Inparticular,mappingbetweenthermalandstructuralmodelsisaknownsourceoferrorasitwouldbenecessarytohavethesamehighfidelitydiscretizationbetweenthethermalandmechanicalmodelstopredictthesub‐nanometerwavefrontsthesespaceapplicationsrequire.However,existingcommercial thermal codes are not capable of handling the same number of nodes as thelargenumbertypicallyusedbystructuralmodelsforsub‐nmpredictions,andsosimplifying

Chapter3

82

assumptions are made on the uniformity of temperatures for mapping interpolationpurposes. For example, the SIMprogramdemonstrated±3mKaccuracy inpredicting thethrough‐thickness changes in temperature gradient in a monolithic mirror, but thisrequiredamassiveinsulatedCushroudinthethermalvacuumchambertoprovidespatiallyand temporally uniform boundary conditions that could be accurately described in thethermal model. That problem simplified the configuration to be modeled in comparisonwith thatofa largebuilt‐upprimarymirror, insidea sunshade that isbound tochallengemodelingfidelity,togeneraterelativetemperaturedistributionsaccurateto1mK.

IsolationsystemsActive Dynamic Isolation Systems: Several aerospace companies have demonstratedsystems that allow for active dynamic isolation of an imaging payload from disturbancesthat are generated frommoving components on the spacecraft (such as reactionwheels,gimbaledsolararrays,andgimbaledantennas).OnesuchsystemiscalledtheDisturbance‐Free Payload (DFP). It is a system architecture inwhich payload and spacecraft bus areseparate bodies that fly in close‐proximity formation, allowing precision payload controlandsimultaneousisolationfromspacecraftdisturbances.Thecontrolarchitectureprovidesisolation down to zero frequency, and sensor characteristics do not limit isolationperformance.Atestbedhasdemonstratedbroadbandisolationinexcessof68dB(afactorof 2,512). With the DFP architecture, the payload and spacecraft bus have independentgravityoffloadsystems.Theyarecoupled througha fullycontrollednon‐contact interfacecontaining sensors and actuators. The payload contains fine optical‐pointing sensors; thespacecraft bus contains a star tracker, three‐axis fiber‐optic gyros, reaction wheels, andthrusters. Payload pointing stability is less than 1 micro‐radian in the laboratoryenvironment. Other dynamic isolation system architectures have been shown to havecomparable performance. All known active isolation system technologies are of hightechnologyreadiness,and investmentsbeingmade inthetechnologieswillensurethatallsuchsystemswillbeTRL6bythetimeaflagshipmissionisflown.Similarly,some,ifnotall,ofthearchitectureswillbeatTRL6(somealreadyare)intimeforaprobe‐classmission.

Twostageisolationsystems:Traditionalspacecraftdesignsinvolveisolationofvibrationsources via passive mechanical isolation stages—essentially compliant spring‐dampersystems. In order to achieve sub‐mas stabilitywith such passive systems, two stages arenecessary—one stage separating the vibrating component (such as a reaction wheelassembly) from the spacecraft structure, and a second stage separating the spacecraftstructure from the payload instrument. Such two‐stage isolation systems are complex tointegrateatthesystemlevel,requireextensivemassandvolumeallocations,andtypicallyrequire high sample‐rate track sensors and fast steering mirrors to reject residualdisturbance thatpropagates to thepayload. In addition it is notpossible tomake scienceobservations when dumping momentum using your thruster system. However, theadvantage they have over the DFP system is they have flown before and have beensuccessful.DFP isataTRL levelof5–6(severalof thecomponentshave flown). TheDFParchitectureneedssomeadditional system levelqualification testing tobring itsTRL toasolid6or7.

VerificationandValidationFundamentally theprocessof integrationandtestof largeopticalspace‐bornesystems(>1.1m)isnodifferentthantheprocessthatneedstobeusedonanyspacesystem.Becauseitis impossible to perfectly reproduce space conditions on the ground, a combination of

OpticalImaging

83

techniquesmustbeemployedtoverifyanyspacepayloadpriortoacceptingthehardwarefor launch. Testing is a significant part of the verification process and deserves criticalattention. However, other verification techniques (including analysis, demonstration,inspection,simulationandrecordsvalidation)areallimportantmethodstobeconsidered.Themainpointhereisthattheearlydesignworkofanyspacesystemshouldnotfocusonhowthesystemwillbetestedpriortoflight,butratherhowitwillbeverifiedpriortoflight.

Giventhatanyexoplanetdirect‐imagingmissionwilllikelybemoretechnicallychallengingthan any space missions that have flown to date, a disciplined systems engineeringapproach must be applied to each subsystem in terms of design and verification. It isespeciallycriticaltogiveattentiontotheverificationplanearlyoninthedesignprocesstoreduce technical and programmatic risk. It is certainly preferable to have an end‐to‐endgroundtesttodemonstrateopticalperformanceofpayloadsontheground.However,iftheuncertainties in the test and inground‐to‐orbit changesarenotaccommodatedbydesignand/oroperationallyadjustableparameters(suchasfocusshiftandtemperaturecontrol),thensucha test ismeaningless.Hence, thesystemdesignmustberobustandtunableon‐orbit, and the verification plan must be equally robust to ensure mission success.Furthermore, advanced modeling tools and techniques will play an integral role in theverification of any exoplanet direct‐imaging mission. However, given the complexity werecommend further testbed demonstration to validate verification methodology andmodelingaccuracy/predictabilityforfuturemissionconcepts.

3.4 Research&AnalysisGoalsSciencetargetobservingandmodelingsupportisneededinpreparationforbothprobe‐sizeand flagship‐size missions. In particular, we recommend efforts in the simulation ofexoplanet spectra for a large range of physical parameters (age, temperature, albedo,composition,clouds,etc.).Modelingisnecessarytounderstandbetterwhatkindofspectrawecanexpectandhowtousetheinformationwecanobtainfromlow‐resolutionspectra.Thiswillalsohelpthedisambiguationofplanettypesforarangeofphysicalparameters.

Ground‐basedobservingsupportisneededinparticularformeasurementsofstellarages.Itis generally believed that the exozodi should get fainter with time as the dust disk isgraduallyclearedoutandcollisionsofplanetesimalsleadingtoenhancedproductionofdustbecome less frequent.However, it appears thatmany,perhaps themajorityofTPF targetstarsareyoung(<2Gyr).AccordingtoBarnes(2007),themedianageofhissampleof70nearby,Sun‐likestarsisonly1.3Gyr.Moreresearchonstellaragesisurgentlyneededandmaybepossible through ground‐based astro‐seismology observations.DRM studiesmustbedetailed,inparticularforconcatenatedprogramsthatstudyAstrometry/RVanddirect‐imagingmissions. This is of critical importance to refine the scaling of themissions andmake sure that the combination of techniques offers a consistent programwithmatchedperformance.

Exozodi science is absolutely critical and support will be needed for ground‐basedobservations with Keck Nulling Interferometer and the Large Binocular TelescopeInterferometer.

Chapter3

84

3.5 ContributorsRuslanBelikov,NASAAmesResearchCenter

PaulBierden,BostonMicromachines

AnthonyBoccaletti,ObservatoiredeParis


ChrisBurrows,Metajiva

EricCady,PrincetonUniversity

WebsterCash,UniversityofColorado

MarkClampin,NASAGoddardSpaceFlightCenter

CostasCossapakis,L’Garde

IanCrossfield,UniversityofCaliforniaLosAngeles

LarryDewell,LockheedMartin

AlanDressler,CarnegieInstitutionofWashington

RobertEgerman,ITTSpaceDivision

TomGreene,NASAAmesResearchCenter

OlivierGuyon,Subaru

SaraHeap,NASAGoddardSpaceFlightCenter

B.Jaroux,NASAAmesResearchCenter

JeremyKasdin,PrincetonUniversity

JimKasting,PennStateUniversity

MatthewKenworthy,UniversityofArizona

SteveKilston,BallAerospace

AndyKlavins,LockheedMartin

JohnKrist,JetPropulsionLaboratory

BenjaminLane,DraperLaboratory

AmyLo,NorthropGrummanCorporation

PatrickJ.Lowrance,SpitzerScienceCenter

RickLyon,NASAGoddardSpaceFlightCenter

BruceMacintosh,LawrenceLivermoreNationalLaboratory

SeanMcCully,LockheedMartin

MarkMarley,NASAAmesResearchCenter

ChristianMarois,HertzbergInstituteforAstrophysics

GaryMatthews,ITTSpaceDivision

BenMazin,JetPropulsionLaboratory

OpticalImaging

85

NaoshiMurakami,NationalAstronomicalObservatoryofJapan(NAOJ)


JunNishikawa,NationalAstronomicalObservatoryofJapan(NAOJ)

BenR.Oppenheimer,AmericanMuseumofNaturalHistory

NelsonPedreiro,LockheedMartin

AkiRoberge,NASAGoddardSpaceFlightCenter

RobertRudd,LawrenceLivermoreNationalLaboratory

StephenRidgway,NationalOpticalAstronomyObservatories

GlennSchneider,UniversityofArizona

JeanSchneider,ObservatoiredeParis

LaurentPueyo,PrincetonUniversity


AnandSivaramakrishnan,AmericanMuseumofNaturalHistory,StonyBrookUniversity

DavidSpergel,PrincetonUniversity

KarlStapelfeldt,JetPropulsionLaboratory

MotohideTamura,NationalAstronomicalObservatoryofJapan(NAOJ)

DomenickTenerelli,LockheedMartin

JasonTolomeo,LockeedMartin

VolkerTolls,HavardSmithonianCenterForAstrophysics

JohnTrauger,JetPropulsionLaboratory

RobertJ.Vanderbei,PrincetonUniversity

JeffWynn,ITTSpaceDivision

3.6 ReferencesArenberg, J.W., Lo,A. S., Cash,W. et al. 2006, “NewWorldsOcculter performance: a first

look,”Proc.SPIE,6265,62651W

Balasubramanian,K.,Wilson,D.W.,Muller,R.E.,etal.2007,“Thickness‐dependentopticalproperties of metals and alloys applicable to TPF coronagraph image masks,” Proc.SPIE,6693,66930Z

Balasubramanian, K. 2008, “Band‐limited image plane masks for the Terrestrial PlanetFindercoronagraph:materialsanddesignsforbroadbandperformance,”Appl.Opt.,47,116–125

Barnes, S. A. 2007, “Ages for Illustrative Field Stars Using Gyrochronology: Viability,Limitations,andErrors,”ApJ,669,1167–1189

Beckwith,S.V.W.2008,“Detecting life‐bearingextrasolarplanetswithspacetelescopes, ”ApJ,684,1404–1415

Chapter3

86

Belikov, R., Kasdin N.J., & Vanderbei R.J. 2006, “Diffraction‐based Sensitivity Analysis ofApodizedPupil‐MappingSystems,”ApJ,652,833–844

Belikov, R., Give'on, A., Savransky, D., et al. 2007, “Demonstration Of Synthetic Exo‐earthDetection In The LabWith Speckle Subtraction Techniques,” American AstronomicalSocietyMeetingAbstracts,211,135

Belikov,R.,Kasdin,N. J.,&Vanderbei,R. J.2006, “Diffraction‐basedSensitivityAnalysisofApodizedPupil‐mappingSystems,”ApJ,652,833–844

Boccaletti, A., Abe, L., Baudrand, J., et al. 2008, “Prototyping coronagraphs for exoplanetcharacterizationwithSPHERE,”Proc.SPIE,7015,70151B

Bordé,P. J.,&Traub,W.A.2006, “High‐Contrast Imaging fromSpace:SpeckleNulling inaLow‐AberrationRegime,”ApJ,638,488–498

Brown, R. A. 2005, “Single‐Visit Photometric and Obscurational Completeness,” ApJ, 624,1010–1024

Cash, W., Schindhelm, E., Arenberg, J., et al. 2006, “The New Worlds Observer: usingocculterstodirectlyobserveplanets,”Proc.SPIE,6265,62651V

Cash,W. 2006, “Detection of Earth‐like planets aroundnearby stars using a petal‐shapedocculter,”Nature,442,51–53

Ceperley, D., Neureuther, A., Miller, M., et al. 2006, “Stray‐light sources from pupil maskedgesandmitigationtechniquesfortheTPFCoronagraph,”Proc.SPIE,6271,62711F

Codona, J.L.,&Angel,R.2004, “ImagingExtrasolarPlanetsbyStellarHaloSuppression inSeparatelyCorrectedColorBands,”ApJ,604,L117–L120

Crepp, J. R., Ge, J., Vanden Heuvel, A. D., et al. 2006, “Laboratory Testing of a LyotCoronagraph Equipped with an Eighth‐Order Notch Filter Image Mask,” ApJ, 646,1252–1259

Delplancke, F. 2008, “The PRIMA facility phaser‐references imaging andmicro‐arcsecondastrometry,”NewAstron.Rev.,52,199–207

Dooley, J. A., & Lawson., P. R. 2005, TPFC Technology Plan, (Jet Propulsion Laboratory:Pasadena,CA,),JPLPublication05‐08

Genet,C.,&Ebbesen,T.W.2007,“Lightintinyholes,”Nature,445,39–46

Give'on,A.,Kasdin,N.J.,&Vanderbei,R.J.2006,“Onrepresentingandcorrectingwavefronterrorsinhigh‐contrastimagingsystems,”JOSAA,23,1063–1073

Greene, T., Beichman, C., Eisenstein, D., et al. 2007, “Observing exoplanetswith the JWSTNIRCamgrisms,”Proc.SPIE,6693,66930G

Guyon,O.,Pluzhnik,E.A.,Galicher,R.,etal.2005,“ExoplanetImagingwithaPhase‐inducedAmplitudeApodizationCoronagraph.I.Principle,”ApJ,622,744–758

Guyon, O., Pluzhnik, E. A., Kuchner, M. J. et al. 2006, “Theoretical Limits on ExtrasolarTerrestrialPlanetDetectionwithCoronagraphs,”ApJSup.Ser.,167,81–99

Heap, S. R., Lindler, D. J., & Woodgate, B. 2006, “An Integral Field Spectrograph for theTerrestrialPlanetFinderCoronagraph,”NewAst.Rev.,50,294–296

Heap,S.2007,“TheTerrestrialPlanetFinder‐Occulter(TPF‐O)scienceprogram,”Proc.SPIE,6687,668713

OpticalImaging

87

Kaltenegger, L., Traub, W. A., & Jucks, K. W. 2007, “Spectral Evolution of an Earth‐likePlanet,”ApJ,658,598–616

Kay, J., Kasdin, N. J., & Belikov, R. 2007, “Wavefront correction in a shaped‐pupilcoronagraph using a Gerchberg‐Saxton‐based estimation scheme,” Proc. SPIE, 6691,66910D

Kenworthy,M.A.,Codona,J.L.,Hinz,P.M.,etal.2007,“FirstOn‐SkyHigh‐ContrastImagingwithanApodizingPhasePlate,”ApJ,660,762–769

Kenyon, S. J., & Bromley, B. C. 2004, “Collisional Cascades in Planetesimal Disks. II.EmbeddedPlanets,”AJ,127,513–530

Kuchner,M. J., Spergel,D.N. 2003, “Notch‐filtermasks: Practical imagemasks for planet‐findingcoronagraphs,”ApJ,594,617–626

Lacour,S.,Thiébaut,E.,&Perrin,G.2007,“Highdynamicrangeimagingwithasingle‐modepupil remapping system: a self‐calibration algorithm for redundant interferometricarrays,”MNRAS,374,832–846

Launhardt, R. 2005, “Differential astrometry and astrometric planet searches with theVLTI,”AstronomischeNachrichten,326,563–564

Lawson,P.R.,&Traub,W.A.2006,EarthlikeExoplanets:TheScienceofNASA’sNavigatorProgram(JetPropulsionLaboratory:Pasadena,CA,),JPLPublication06‐0510/06

Levine, M., Shaklan, S., & Kasting, J. 2006, TPFC Science and Technology Definition TeamReport,JPLDocumentD‐34923,JetPropulsionLaboratory,Pasadena,CA

Leviton,D. B., Cash,W. C., Gleason, B., Kaiser,M. J., Levine, S. A., Lo, A. S., Schindhelm, E.,Shipley, A. F. 2007, “White‐light demonstration of one hundred parts per billionirradiancesuppressioninairbynewstarshadeocculters,“Proc.SPIE,6687,6687‐1B

Lindler,D.J.2007,“TPF‐Odesignreferencemission,”Proc.SPIE,6687,668714

Lo,A.S.,Glassman,T.,&Lillie,C.2007, “Newworldsobserveropticalperformance,”Proc.SPIE,6687,668716

Lunineetal.2008,WorldsBeyond:AStrategyfortheDetectionandCharacterizationofExoplanets,NationalScienceFoundation,Arlington,VAhttp://www.nsf.gov/mps/ast/aaac/exoplanet_task_force/reports/exoptf_final_report.pdf

Lyon, R. G., Heap, S., Lo, A., et al. 2007, “Externally occulted terrestrial planet findercoronagraph:simulationsandsensitivities,”Proc.SPIE,6687,668719

Macintosh B., Graham, J., Palmer, D., et al. 2006, “The Gemini Planet Imager,” Proc. SPIE,6272,62720L

Macintosh,B.A.,Graham,J.R.,Palmer,D.W.,etal.2008,“TheGeminiPlanetImager:fromsciencetodesigntoconstruction,”Proc.SPIE,7015,70158

Marcy,G.2007,“ObservedPropertiesofExoplanets,” inProc.Conf.InthespiritofBernardLyot:TheDirectDetectionofPlanetsandCircumstellarDisksinthe21stCentury,Kalas,P.,ed.,4‐8June2007,Univ.California,Berkeley,CA

Mawet,D.,Riaud,P.,Baudrand,J.,etal.2006,“Thefour‐quadrantphase‐maskcoronagraph:whitelightlaboratoryresultswithanachromaticdevice,”A&A,448,801–808

Chapter3

88

Mayor,M.,Udry,S.,Lovis,etal.2009,“TheHARPSsearchforsouthernextra‐solarplanets.XIII.Aplanetarysystemwith3Super‐Earths(4.2,6.9,&9.2Earthmasses),”A&A,493,639–644

Mazin,B.A.,Bumble,B.,Day,P.K.,etal.2006,“Positionsensitivex‐rayspectrophotometerusingmicrowavekineticinductancedetectors,”Appl.Phys.Lett.,89,222507

Moody,D.C.,Gordon,B.L.,&Trauger,J.T.2008,“DesignanddemonstrationofhybridLyotcoronagraph masks for improved spectral bandwidth and throughput,” Proc. SPIE,7010,70103P

Morbidelli, A., Levison, H. F., Tsiganis, K., et al. 2005, “Chaotic capture of Jupiter's TrojanasteroidsintheearlySolarSystem,”Nature,435,462–465

Murakami, N., Uemura, R., Baba, N., Shibuya, H., Nishikawa, J., Abe, L., Tamura, M., &Hashimoto, N. 2008, “Laboratory experiments on the 8‐octant phase‐maskcoronagraph,”Proc.SPIE,7010,70101J

Nishikawa,J.,Murakami,N.,Abe,L.,Kotani,T.,Tamura,M.,Yokochi,K.,&Kurokawa,T.2006,“Nulling and adaptive optics for very high dynamic range coronagraph,” Proc. SPIE,6265,62653Q

Nishikawa,J.,Murakami,N.,Abe,L.,&Kotani,T.2008,“Precisewavefrontcorrectionwithanunbalanced nulling interferometer for exoplanet imaging coronagraph,” A&A, 489,1389–1398

Noecker, M. C. 2007, “Alignment of a terrestrial planet finder starshade at 20–100megameters,”Proc.SPIE,6693,669306

Oppenheimer, B. R., Brenner, D., Hinkley, S., et al. 2008, “The Solar‐System‐Scale DiskaroundABAurigae,”ApJ,679,1574–1581

Pueyo, L. 2008, “Broadband contrast for exo‐planet imaging: The impact of propagationeffects,”PhDThesis,PrincetonUniversity,Princeton,NJ

Riaud, P., Boccaletti, A., Baudrand, J., et al. 2003, “The Four‐Quadrant Phase MaskCoronagraph.III.LaboratoryPerformance,”PASP,115,712–719

Savransky,D.,&Kasdin,J.2008,“Designreferencemissionconstructionforplanetfinders,”Proc.SPIE,7010,70101T

Shaklan,S.B.,Green,J.J.,&Palacios,D.M.2006,“Theterrestrialplanetfindercoronagraphopticalsurfacerequirements,”Proc.SPIE,6265,62651I

Shao,M.,Samuel,R.,Wallace,K.,&Levine,B.2007,“VisibleNullingCoronagraph,” inProc.Conf. In the spirit of Bernard Lyot: The Direct Detection of Planets and CircumstellarDisksinthe21stCentury,Kalas,P.,ed.,4‐8June2007,Univ.California,Berkeley,CA

Schindhelm, E., Shipley, A., Oakley, P., Leviton, D., Cash, W., Card, G., 2007, “Laboratorystudiesofpetalshapedocculters,”Proc.SPIE,6693,669305

Spergel,D.N.,Kasdin,J.,Belikov,R.,etal.2009,“THEIA:TelescopeforHabitableExoplanetsandInterstellar/IntergalacticAstronomy,”AASMeetingAbstracts,213,458.04

Swartzlander,G.Ford,E.,&Abdul‐Malik,R.2008,“Astronomicaldemonstrationofanopticalvortexcoronagraph,”Opt.Exp.,16,10200–10207

OpticalImaging

89

Tavrov, A. V., Nishikawa, J., Tamura, M., Abe, L., Yokochi, K., Kurokawa, T., & Takeda, M.2008, “Achromatic interfero‐coronagraph with two common‐path interferometers intandem,”Appl.Opt.,47,4915–4926

Thomas,S.J.,Soummer,R.,Dillon,D.,Macintosh,B.,Evans,J.W.,Gavel,D.,Sivaramakrishnan,A.,Marois,C.,Oppenheimer,B.R.2008, “TestingtheAPLContheLAOExAOtestbed,”Proc.SPIE,7015,70156I

Tolls, V., Aziz, M., Gonsalves, R. A., Korzennik, S., Labeyrie, A., Lyon, R., Melnick, G.,Somerstein, S., Vasudevan, G., Woodruff, R., 2005, “Study of high‐performancecoronagraphtechniques,”Proc.SPIE,5905,494–501

Tolls, V., Aziz,M., Gonsalves, R. A., et al. 2006, “Study of coronagraphic techniques,” Proc.SPIE,6265,62653K

Traub,W.A.,&Chen,P.2007,“PlanetscopePrecursorExperiment,”AmericanAstronomicalSocietyMeetingAbstracts,211,#30.06

Trauger,J.,Give'on,A.,Gordon,B.etal.2007,“Laboratorydemonstrationsofhighcontrastimagingforspacecoronagraphy,”Proc.SPIE,6693,66930X

Vanderbei, R. J., Cady, E., & Kasdin, N. J. 2007, “Optimal Occulter Design for FindingExtrasolarPlanets,”ApJ,665,794–798

Vasudevan, G., Reale, M., & Somerstein, S. F. 2005, “Some performance results fromNIRCam'scoronagraphicprototypemasks,”Proc.SPIE,5904,66–70

Verhoeve,P.J.2008,“PhotoncountinglowtemperaturedetectorsforvisibletogammarayAstrophysics,”LowTemp.Phys.,151,675–683

Wallace,J.K.,Bartos,R.,Rao,S.,etal.2006,“Laboratoryexperimentfordemonstratingpost‐coronagraphwavefront sensing and control for extreme adaptive optics,” Proc. SPIE,6272,62722L

Wen,Y.,Rauscher,B. J.,Baker,R.G.,etal.2006, “Individualphoton‐countingusinge2vL3CCDsforlowbackgroundastronomicalspectroscopy,”Proc.SPIE,6276,62761H

InfraredImaging

91

4 InfraredImaging

WilliamDanchi,NASAGoddardSpaceFlightCenter,Chair

PeterLawson,JetPropulsionLaboratory,CoChairOlivierAbsil,RachelAkeson,JohnBally,RichardK.Barry,CharlesBeichman,AdrianBelu,MathewBoyce,JamesBreckinridge,AdamBurrows,ChristineChen,DavidCole,DavidCrisp,RolfDanner,PeterDeroo,VincentCoudéduForesto,DenisDefrère,DennisEbbets,PaulFalkowski,RobertGappinger,IsmailD.Haugabook,Sr.,CharlesHanot,ThomasHenning,PhilHinz,JanHollis,SarahHunyadi,DavidHyland,KennethJ.Johnston,LisaKaltenegger,JamesKasting,MattKenworthy,AlexanderKsendzov,BenjaminLane,GregoryLaughlin,OliverLay,RéneLiseau,BrunoLopez,RafaelMillanGabet,StefanMartin,DimitriMawet,BertrandMennesson,JohnMonnier,NaoshiMurakami,M.CharlesNoecker,JunNishikawa,MeyerPesesen,RobertPeters,AliceQuillen,SamRagland,StephenRinehart,HuubRottgering,DanielScharf,EugeneSerabyn,MotohideTamura,MohammedTehrani,WesleyA.Traub,StephenUnwin,DavidWilner,JulienWoillez,NevilleWoolf,andMingZhao

4.1 IntroductionThis chapter describes the science and technology of infrared andmid‐infraredmissionscapable of detecting and characterizing exoplanets through the interferometric nulling ofstarlight.

Aprobe‐scalemissionofthistypewouldimagenearbyplanetarysystems,andallowustostudy planet formation, study structure in debris and exozodiacal disks, and detect andcharacterizewarmJupiters,andpossiblySuper‐Earths.Planetsaround~50starswouldbecharacterized.Aprobe‐scalemissionwouldprovide transformativescienceandwould laytheengineeringgroundworkfortheflagshipmissionthatwouldfollow.

AflagshipmissionwouldenablethedetectionofbiosignaturesinthespectraofEarth‐likeplanetsof200ormorenearbytargetstars.Suchamissionwouldhavethehighestangularresolutionforplanet‐findingofanyofthemissionconceptsdescribedinthisReport.

ThemissionsdescribedinthischapterhavegrownfromthelegacyoftheTerrestrialPlanetFinder(TPF),whichwithitscounterpartinEurope,theDarwinmission,havebeenstudiedandreviewedsincethemid‐1990s.TechnologydevelopmentforTPFwasendorsedbytheMcKee‐Taylorreport(NationalResearchCouncil2001)withthegoalofenablingamissionsometimeafter2010.Herewereportonthesuccessofthatinvestment.

Aswillbeshowninthischapter,thetechnologyforaprobe‐scalemissionislargelyinhand.A probe‐scale mission would use a fixed or connected‐structure and operate at infraredwavelengths.Laboratoryexperimentsinstarlightsuppressionhavealreadyexceededflightrequirements for such a mission by a factor of 10, and all of the necessary component

Chapter4

92

hardwarehasbeendemonstratedinthelab.Futuretechnologyeffortsshouldincludesomeadditionalsystem‐leveltestingandmodeling,andcryogenicdemonstrationofsingle‐modefibers.Theprobe‐scalemissionthatispresentedistheFourierKelvinStellarInterferometer(FKSI),asdescribedbyDanchietal.(2003).

Furthersustainedeffort isnonethelessrequired toaddressseveralof therisksassociatedwithaflagshipmission.Themajorriskofsuchamissionisdominatedbytherelianceonasimultaneous and coordinated use of five spacecraft flying in formation; four telescopeseachonseparatespacecraftandanadditionalspacecraft forabeamcombiner. Guidance,navigation,andcontroldemonstrationshaveshownthatsuchcontrolisfeasible,butactualhardwaretests inspacehavenotyetbeenperformed. Ontheotherhand, interferometricnullinghasnowreachedtheperformancelevelrequiredforflight,asdemonstratedthroughthe Adaptive Nuller testbed. Future technology development should include cryogenicsystem‐level tests with mature brassboard designs and space‐based demonstrations offormation flying maneuvers using representative sensors and thrusters. The flagshipmissionthatispresentedistheTerrestrialPlanetFinderInterferometer(TPF‐I),identicaltothe Darwin mission concept that was proposed to the European Space Agency’s CosmicVisionprogramin2007.

Althoughwesupportmostof the long‐termgoalsthattheExoplanetTaskForce(ExoPTF)recommendedforaflagshipinfraredmission,werecommendadifferentpathforwardforthenear‐term. Specifically,wearenotconvinced,as theExoPTFreportsuggests, that theproblemofexozodilevelsanddebrisdiskscanbesolvedwithground‐basedobservationstotheextentnecessaryfortheformulationofaflagshipmission.Wearenotconvinced,inpartdue to our own experience with the Keck Interferometer, for which the lower limit onexozodis is100–200 times thatof theSolar Systemzodi level.Weexpect that thenullinginstrument on Large Binocular Telescope Interferometer (LBTI) will reduce this limitsubstantiallyforarelativelysmallsampleofstars.Wediscusshowaprobe‐classmissionintheinfraredcanmeasuretheexozodilevelsdowntothelevelofoneSolarSystemzodiforessentiallyall of thepotential target stars for theeventual flagshipTPF/Darwinmissions.Thisstep iscrucial,notonlyfortheflagshipcharacterizationmissions,but isalsoofgreatvaluetoanastrometricmissionbecausethatmissioncanthenfocusitssearchesforEarth‐twinsaroundstarswithlowexozodilevels.Moreover,aninfraredprobe‐classmissionhasahigherdegreeoftechnologyleveragingfromJWST,andcouldbeundertakeninarelativelyshort time without undue cost and technology risk. Our summary of recommendationsfromthischapterfollows:

Recommendation: A vigorous technology program, including component development,integrated testbeds, and end‐to‐end modeling, should be carried out in the areas offormation flying andmid‐infrared nulling,with the goal of enabling a probe‐scale nullinginterferometrymissioninthenext2to5yearsandaflagshipmissionwithinthenext10to15years.

Recommendation: The fruitful collaborationwith European groups onmission conceptsandrelevanttechnologyshouldbecontinued.

Recommendation: R&A should be supported for the development of preliminary scienceandmission designs. Ongoing efforts to characterize the typical level of exozodiacal lightaroundSun‐likestarswithground‐basednullinginterferometryshouldbecontinued.

Inthischapter,weoutlinethescientificparameterspaceaffordedbythenextgenerationofmid‐infraredinterferometersinthecontextofground‐basedarrays,includingtheAtacamaLarge Millimeter Array (ALMA), the Keck Interferometer (KI), the Large Binocular

InfraredImaging

93

Telescope Interferometer (LBTI), and theVery LargeTelescope Interferometer (VLTI), aswellascooledspacetelescopes(Spitzer,Herschel,JWST).

The remainder of this chapter is divided into self‐contained sections on Science Goals(§4.1), Observatory Concepts (§4.2), Technology (§4.3), and Research & Analysis Goals(§4.4).

4.1.1 SensitivityandAngularResolutionBoth high‐angular resolution and high sensitivity in the mid‐infrared are key to futureprogress across all major fields of astronomy. Figure 4‐1 shows the parameter space ofspatialresolutionversussensitivity,illustratingthetrade‐offsbetweenthehighsensitivitypossibleinspaceandthehighangularresolutionachievablefromtheground.Thisdiagramhighlights the rich and varied science goals made possible by continued incrementaltechnicaldevelopmentsforadvancedimaginginthemid‐infrared.

Processesoccurringat0.5–5AUinprotoplanetarydisksenabletheformationofgiantandterrestrialplanetsandarerevealedbyobservationsinthemid‐infrared,whicharesensitivetodustemissionfrom100–1000K.

Figure41.Relationshipbetweenresolutionandsensitivityas itpertainstoadvancedimagingintheinfraredforplanetformation,debrisandexozodialdisks,anddirectdetectionofexoplanets.(W.C.Danchi,NASAGSFC)

Chapter4

94

Achieving the necessary angular resolution at infrared and mid‐infrared wavelengths tostudy protoplanetary disks requires long‐baseline interferometry with ~100–250‐mbaselinestoprobe1‐AUscalesfortheneareststar‐formingregionsatadistanceof140pc.Suchbaselinesarealreadypossiblefromtheground.TheKIandtheVLTIarenowabletoprobe the dust and temperature structure around luminousHerbig Ae stars and T Tauridisks. The study of planet formation is being revolutionized by mid‐IR interferometry’ssensitivitytothewarminnerdiskandsoonwillbeaugmentedbyALMA’sprobeofthecoldouter disk. Although thenext‐generationmillimeter‐wave interferometerALMAwill lackthe angular resolution to probe the terrestrial planet formation zone, mid‐infraredinterferometrywith>250meterbaselineswillbesensitivetowarmdustemissionwithintheinnerfewAUsofplanet‐formingdisks.

However, thermal emission from the atmosphere and telescope limits the sensitivity ofground‐basedobservations,drivingmostscienceprogramstowardsspaceplatforms.Evenverymodest sized cooledapertures canhaveordersofmagnitudemore sensitivity in thethermal infrared than the largest ground‐based telescopes currently planned or inoperation.

4.1.2 ProbeScaleMissionScienceGoalsAdvanced Imaging in the Midinfrared: Planet Formation, Debris and Exozodiacal Disks,WarmJupiters,andSuperEarths

Themoremodestbaselinesof~10–20m,attainablewithaprobe‐scalemission,wouldbecapableofmeasuringsolar‐systemfeaturesaroundnearbystars.

Aprobe‐scalemissionwouldfocusonthreemainsciencegoalsforfurtherstudyofgreatestimportance to the direct detection of exoplanets in the infrared, namely (1) measureresonantstructuresinexozodiacaldisks,theirfluxlevels,andthechemicalevolutionofthematerial they contain, (2) detect and characterize the atmospheres of nearby exoplanets,including thosecurrentlyknownandthose thatmaybediscoveredby this technique,and(3)surveyprospectivetargetstarsforlevelsofexozodicaldust.

MeasureResonantStructuresinExozodiacalDisksTheoptically‐thickandgas‐richdisksoutofwhichplanetsformarebright,dominatingoverthe emission from the central star. As the system evolves, the dust grains grow intoprogressivelylargerbodies,andthegasdissipates.Smallparticlesofduststillpervadethesystem, created in collisions between planetesimals, and occasionally large collisionswillproduce dramatic structure in these “debris disks.” Here, mid‐infrared imaging inconjunctionwithnulling(toblockoutthecentralstar),allowsthespatialdistributionandspectralenergydistribution(SED)ofthedebrisdiskstobescrutinized.Ground‐andspace‐based interferometers both have important roles here, since ground systems can probebrighterdebrisdiskswithgreaterangularresolution(necessarytonulloutthestarwithoutnulling out the dust) and the space systemswithmodest~20‐m baseline can sensitivelylookfordebrisaroundnearbystars.

Exozodiacal disks are already under intensive study because the mid‐IR emission andscattered light can mimic or hide the signal of terrestrial planets around nearby stars.Chapter5isdevotedtothiscrucialtopic,andwereferthereadertothatchapterforfurtherdetails.

InfraredImaging

95

DetectandCharacterizeNearbyExoplanetsPlannedmid‐infrared facilitiesarealsobeingdesigned todirectlydetectexoplanetsusingnulling interferometry. Ground‐based systems can contribute at shorter wavelengths(<5µm) but exoplanet detection in the mid‐infrared (6–20 µm) is not expected to bepractical from the ground because thermal emission from the warm sky and telescopeswamps the faintplanetarysignal (exceptperhaps in somerare scenarios).Aprobe‐scalemissioncouldbeusedtocharacterizetheatmospheresofa largenumberof thecurrentlyknownmassiveexoplanetsandpossiblySuperEarths(e.g.,Danchietal.2003a,b),whereasaflagshipformation‐flyinginterferometercouldalsodetectterrestrialplanetsandsearchforatmosphericbiomarkers.

CharacterizetheAtmosphericConstituentsofGiantPlanetsAprobe‐scalemissionwouldhavesufficientsensitivity todetectandcharacterizeabroadrangeofexoplanets.Manymolecularspecies,suchascarbonmonoxide,methane,andwatervapor, have strong spectral features in the 3–8µm region, as can be seen in Figure 4‐2,whichdisplaysmodelatmospheres forplanetsatvariousdistances fromthehoststar, forexoplanets as calculated by Seager et al. (2005). The red curve shows the theoreticalspectrumofaveryhot,close‐inplanetat0.05AU,andthebluecurvedisplaysthespectrumfor amuch cooler planet ten times further out, at 0.5 AU. Also displayed are sensitivitycurves for the IRAC instrument onSpitzer (light blue) and the curve for a representativeprobe‐scalemission,FKSI(purple).

Figure42.TheFKSIsystemcanmeasurethespectraofexoplanetswithawiderangeofsemi‐majoraxes.

Chapter4

96

CharacterizetheAtmospheresofSuperEarthsTheapparentlycommonplacenatureoflowermassplanetssuggestsanintermediateclassofplanets,SuperEarths,maybeobservablewithaprobe‐scalemission.Aprimeexampleofthis is theplanetarysystemofGL581. Theouter twoplanetsof thissystemarenear thehabitable zoneof the star. Selsis et al. (2007)haveanalyzed this systemandsuggest theouterplanet(planetd,a=0.25AU,R~2R⊕,M~8M⊕) is themost likelytobehabitable.Since the star is at 6.3 pc, the planet's angular separation, 0.04 arcsec, is at the firstmaximum of the fringe pattern of FKSI. At four times the flux of an Earth‐size planet,GL581dwouldbewithin the sensitivity limit of FKSI. Thus, there is alreadyoneknown,possibly Earth‐size planet that could be observed with FKSI. As more Super Earths arediscovered, it is likely thatFKSIwillhaveat leasta small sampleofobjects thatmightbecharacterizedfortheatmosphericconstituentsandcouldindicatehabitability.

CharacterizetheDynamicalEnvironmentoftheHabitableZoneInadditiontoobservationsof theatmospheresofgiantplanets,aprobe‐scalemissioncancontribute to our understanding the typical distribution of giant planets in wide orbits(>5‐10AU).Suchplanetscandominatethedynamicalenvironmentofthehabitablezone,andaffectthedeliveryofvolatilestoterrestrialplanets(Lunine2001).Planetsoneccentricorbitsmay disrupt the habitable zone or affect the composition by perturbation of outerplanetesimals into crossingorbitswith thehabitable zone. Understanding theplacementand orbital parameters of outer planets is an important prerequisite to searching forterrestrialplanetsinthesesystems.Suchobservationsaredifficulttoobtainfromground‐basedtelescopes.Massiveoryoungplanetscanbedetected(cf.Lafrenièreetal.2008;Hinzetal.2006),butplanetsthatareolderthan~0.5GyrorarelessmassivethanJupiterarenotdetectablewith8‐mclassground‐basedtelescopes.

SurveytheExozodiacalDustofTargetStarsKnowingwhichnearbymain‐sequencestarshave the lowestamountof zodiacaldustwillallowforamoreefficientstrategyofspectralcharacterizationofEarth‐sizedplanets.Thisinformationcouldalsobeusedtooptimizethesearchstrategyforanastrometricmission.Aprobe‐scale infraredmissionwouldenableanexozodiacaldust survey (<1 zodiat<1 to>4AU) including infrared spectra to elucidate grain chemistry and evolution, diskstructure,andbrightness,forallthelikelytargetstarsofaflagshipmissionwithin30pcoftheSolarSystem.

4.1.3 FlagshipMissionScienceGoalsandRequirements

SearchforHabitabilityandSignsofLifeAflagshipmissionwouldbedesignedtodetectterrestrialexoplanetsaroundnearbystarsandmeasuretheirspectra(e.g.,Beichmanetal.1999,2006;Fridlund2000;Kaltenegger&Fridlund2005).Thesespectrawouldbeanalyzedtoestablishthepresenceandcompositionof the planets’ atmospheres, to investigate their capability to sustain life as we know it(habitability),andtosearchforsignsoflife.Aflagshipmissionwouldalsohavethecapacitytoinvestigatethephysicalpropertiesandcompositionofabroaderdiversityofplanets,tounderstand the formation of planets, and to search for the presence of potentialbiosignature compounds. The range of characteristics of planets is likely to exceed ourexperiencewith the planets and satellites in our own Solar System.Moreover, Earth‐like

InfraredImaging

97

planetsorbitingstarsofdifferentspectral typemayalsobeexpected toevolvedifferently(Selsis2000;Seguraetal.2003,2005).

Biomarkers are detectable species whose presence at significant abundance requires abiologicalorigin(DesMaraisetal.2002).Theyarethechemical ingredientsnecessary forbiosynthesis (e.g., oxygen [O2] and CH4) or are products of biosynthesis (e.g., complexorganic molecules, but also O2 and CH4). Our search for signs of life is based on theassumption thatextraterrestrial lifeshares fundamentalcharacteristicswith lifeonEarth,inthatitrequiresliquidwaterasasolventandhasacarbon‐basedchemistry(Owen1980;DesMaraisetal.2002).Therefore,weassumethatextraterrestrial life issimilarto lifeonEarth in its use of the same input and output gases, that it exists out of thermodynamicequilibrium, and that it has analogs to bacteria, plants, and animals on Earth (Lovelock1975).Figure4‐3displaystheobservedemissionspectrumfromtheEarthintheinfrared.

Figure43.Observedemissionspectrumintheinfrared(Christensen&Pearl1997)oftheintegratedEarth,asdeterminedfromEarthshineandspacerespectively.Thedataisshowninblack,modelandindividualchemicalfeaturecolor.(Kaltenegger,Traub&Jucks2007)

Candidatebiomarkers thatmightbedetectedby a flagshipmissionwith a low‐resolutioninstrument include O2, O3, and CH4. There are good biogeochemical and thermodynamicreasons for believing that these gases should be ubiquitous byproducts of carbon‐basedbiochemistry, even if the details of alien biochemistry are significantly different than thebiochemistry on Earth. Production of O2 by photosynthesis allows terrestrial plants andphotosynthetic bacteria (cyanobacteria) to use abundant H2O as the electron donor toreduce CO2, instead of having to rely on scarce supplies of hydrogen (H2) and hydrogensulfide (H2S). Oxygen and nitrous oxide (N2O) are two very promising bio‐indicators.Oxygen is a chemically reactive gas. Reduced gases and oxygen have to be producedconcurrentlytoproducequantities largeenoughtobedetectable indisk‐averagedspectra

Chapter4

98

ofterrestrialplanetatmospheres,astheyreactrapidlywitheachother.N2OisabiomarkerintheEarth’satmosphere,beingproducedinabundanceby lifebutonly intraceamountsbynaturalprocesses.AlthougharelativelyweakfeatureintheEarth’sspectrum,itmaybemorepronouncedinterrestrialexoplanetatmospheresofdifferentcompositionorhost‐starspectral type (Seguraet al. 2005). Currently, efforts areongoing to explore theplausiblerange of habitable planets and to improve our understanding of the detectable ways inwhichlifemodifiesaplanetonaglobalscale.

Inthemid‐IR,theclassicalsignatureofbiologicalactivityisthecombineddetectionofthe9.6‐μmO3band,the15‐μmCO2band,andthe6.3‐μmH2Obandoritsrotationalbandthatextends from12μmout into themicrowave region (Selsis&Despois2002). Theoxygenand ozone absorption features in the visible and thermal infrared, respectively, couldindicatethepresenceofphotosyntheticbiologicalactivityonEarthanytimeduringthepast50%oftheageoftheSolarSystem.IntheEarth’satmosphere,the9.6‐μmO3bandisapoorquantitative indicator of the O2 amount, but an excellent qualitative indicator for theexistenceofeven tracesofO2. TheO39.6‐μmband isaverynonlinear indicatorofO2 fortwo reasons. First, for thepresentatmosphere, low‐resolution spectraof thisband showlittlechangewiththeO3abundancebecauseitisstronglysaturated. Second,theapparentdepth of this band remains nearly constant as O2 increases from 0.01 times the presentatmospherelevel(PAL)ofO2to1PAL(Seguraetal.2003).Theprimaryreasonforthis isthat thestratosphericozone increases thataccompanied theO2buildup lead toadditionalultraviolet heating of the stratosphere. At these higher temperatures, the stratosphericemissionfromthisbandpartiallymaskstheabsorptionofupwellingthermalradiationfromthesurface.

Methaneisnotreadily identifiedusinglow‐resolutionspectroscopyforpresent‐dayEarth,but the CH4 feature at 7.66 μm in the IR is easily detectable at higher abundances(Kalteneggar, Traub & Jucks 2007). When observed together with molecular oxygen,abundantCH4canindicatebiologicalprocesses(seealsoLovelock1975;Seguraetal.2003).Dependingonthedegreeofoxidationofaplanet'scrustanduppermantle,non‐biologicalmechanismscanalsoproducelargeamountsofCH4undercertaincircumstances.

N2O isalsoacandidatebiomarkerbecause it isproduced inabundanceby lifebutonly intraceamountsbynaturalprocesses.TherearenoN2OfeaturesinthevisibleandthreeweakN2O features in the thermal infraredat7.75μm,8.52μm, and16.89μm.Forpresent‐dayEarth,oneneedsaresolutionof23,23and44,respectively,todetecttheseN2Ofeaturesatthermal infraredwavelengths(Kaltenegger,Traub&Jucks2007).Spectral featuresofN2OalsobecomemoreapparentinatmosphereswithlessH2Ovapor.Methaneandnitrousoxidehave features nearly overlapping in the 7‐μm region, and additionally both lie in the redwingofthe6‐μmwaterband. Thus,N2Oisunlikelytobecomeaprimetargetforthefirstgenerationofspace‐basedmissionssearchingforexoplanets,butitisanexcellenttargetforfollow‐upmissions.Thereareothermoleculesthatcould,undersomecircumstances,actasexcellent biomarkers, e.g., themanufactured chloro‐fluorocarbons (CCl2F2 [Freon 13] andCCL3F[Freon12]) inourcurrentatmosphere in the thermal infraredwaveband,but theirabundancesaretoolowtobespectroscopicallyobservedatlowresolution.

Otherbiogenictracegasesmightalsoproducedetectablebiosignatures.Currentlyidentifiedpotentialcandidatesincludevolatilemethylatedcompounds(likemethylchloride[CH3Cl])and sulfur compounds. It isknown that these compoundsareproducedbymicrobes, andpreliminaryestimatesoftheirlifetimesanddetectabilityinEarth‐likeatmospheresaroundstarsofdifferentspectral typehavebeenmade(Seguraetal.2003,2005). However, it isnot yet fully understood how stable (or detectable) these compounds would be in

InfraredImaging

99

atmospheresofdifferent compositionand for starsofdifferent spectral typeand incidentultraviolet flux. These uncertainties, however, could be addressed by further modelingstudies.

ScienceRequirementsTherequirement fora flagship‐classmissiontodetectdirectlyandcharacterizeEarth‐likeplanetsaroundnearbystars implies that themissionmustdetermine the typeofaplanetand characterize its gross physical properties and its main atmospheric constituents,thereby allowing an assessment of the likelihood that life or habitable conditions existthere.

TypesofStars:Onastrophysicalgrounds,Earth‐likeplanetsarelikelytobefoundaroundstarsthatareroughlysimilartotheSun(Turnbull2004).Therefore,targetstarswillincludemainsequenceF,G,andKstars.However,Mstarsmayalsoharborhabitableplanets,andthenearestofthesecouldbeinvestigated.

TerrestrialPlanets:Considering the radii andalbedosoreffective temperaturesofSolarSystemplanets,themissionmustbeabletodetectterrestrialplanets,downtoaminimumterrestrialplanetdefinedashaving1/2EarthsurfaceareaandEarthalbedo.Intheinfrared,theminimumdetectableplanetwouldbeonewithan infraredemissioncorresponding tothesurfaceareaandopticalalbedo,positionedintheorbitalphasespacestipulatedbelowandillustratedinFig.4‐4.

Habitable Zone: A flagship mission should search the most likely range as well as thecomplete range of temperatures within which life may be possible on a terrestrial‐typeplanet.IntheSolarSystem,themostlikelyzoneisnearthepresentEarth,andthefullzoneistherangebetweenVenusandMars.Thehabitablezone(HZ)isheredefinedastherangeof semi‐major axes from 0.7 to 1.5 AU scaled by the square root of stellar luminosity(Kastingetal.1993;Forget&Pierrehumbert1997).TheminimumterrestrialplanetmustbedetectableattheouteredgeoftheHZ.

Orbital Phase Space: The distribution of orbital elements of terrestrial type planets ispresently unknown, but observations suggest that giant‐planet orbits are distributedroughly equally in semi‐major axis, and in eccentricity up to those of the Solar Systemplanets and larger. Therefore, a flagshipmissionmust be designed to search for planetsdrawn from uniform probability distributions in semi‐major axis over the range 0.7 to1.5AUandineccentricityovertherange0to0.35,withtheorbitpoleuniformlydistributedoverthecelestialspherewithrandomorbitphase.

Giant planets: The occurrence and properties of giant planets may determine theenvironmentsof terrestrialplanets.The fieldofviewandsensitivitymustbe sufficient todetect a giant planet with the radius and geometric albedo or effective temperature ofJupiterat5AU(scaledbythesquarerootofstellar luminosity)aroundat least50%of itstargetstars.Asignal‐to‐noiseratioofatleast5isrequired.

Exozodiacaldust:Emissionfromexozodiacaldustisbothasourceofnoiseandalegitimatetargetofscientificinterest.Aflagshipmissionmustbeabletodetectplanetsinthepresenceofzodiacalcloudsatlevelsuptoamaximumof10timesthebrightnessofthezodiacalcloudintheSolarSystem.Althoughtheaverageamountofexo‐zodiacalemissioninthe“habitablezone”isnotyetknown,weadoptanexpectedlevelofzodiacalemissionaroundtargetstarsof 3 times the level in our own Solar Systemwith the same fractional clumpiness as ourSolar System’s cloud. From a science standpoint, determining and understanding thepropertiesofthezodiacalcloudisessentialtounderstandingtheformation,evolution,and

Chapter4

100

habitabilityofplanetarysystems.Thus,themissionshouldbeabletodeterminethespatialand spectral distribution of zodiacal clouds with at least 0.1 times the brightness of theSolarSystem’szodiacalcloud.

Spectral range: The required spectral range of the mission for characterization ofexoplanetswillemphasizethecharacterizationofEarth‐likeplanetsandisthereforesetto6.5to18µmintheinfrared.Theminimumrangeis6.5to15µm.

Figure44.TheextentoftheHabitableZoneforthesinglemain‐sequencestars(FandGstarsonupperpanel,andM,Kstarsonlowerpanel)within30pcfromtheSun.(Kalteneggeretal.2009)

InfraredImaging

101

Spectrum: Themissionwill use the spectrum of a planet to characterize its surface andatmosphere. The spectrum of the present Earth, scaled for semi‐major axis and starluminosity,isusedasareferenceandsuggestsaminimumspectralresolutionof25withagoalof50. Themissionmustmeasurewater(H2O)andozone(O3)with20%accuracy inthe equivalent width of the spectral feature. Additionally it is highly desirable that themissionbeabletomeasurecarbondioxide(CO2)aswellasmethane(CH4)(if the latter ispresent in high quantities predicted in somemodels of pre‐biotic, or anoxic planets, e.g.,Kasting&Catling2003).

Numberofstarstobesearched:Tosatisfy itsscientificgoals, themissionshoulddetectandcharacterizeastatisticallysignificantsampleofterrestrialplanetsorbitingF,G,andKstars. Although at this time, the fractional occurrence of terrestrial exoplanets in theHabitableZoneisnotknown,asampleof150starswithin30pc(includingasmallnumberofnearbyMstars)shouldsufficebasedonourpresentunderstanding.

Extendednumberof stars: It isdesired to searchasmanystarsaspossible,beyond therequiredcoresample.Weanticipatethatanymissioncapableofsatisfyingtheseobjectiveswill also be capable of searching many more stars if the overall requirements oncompletenessarerelaxed.Itisdesiredthatthemissionbecapableofsearchinganextendedgroupofstarsdefinedasthosesystemsofanytypeinwhichallorpartofthecontinuouslyhabitablezone(seebelow)canbesearched.

Search completeness: Search completeness is defined as that fraction of planets in theorbital phase space that could be foundwithin instrumental andmission constraints.Werequire each of the 150 stars to be searched at the 90% completeness level. For othertargetsinadditiontothe150stars,theavailablehabitablezonewillbesearchedastolimitsinplanet'sorbitalcharacteristics.

Table41.FlagshipInfraredImagingMissionRequirementSummary

Parameter Requirement

StarTypes F,G,K,selected,nearbyM,andothers

HabitableZone 0.7–1.5(1.8)AUscaledasL1/2(Note*)

NumberofStarstoSearch >150

CompletenessforEachCoreStar 90%

MinimumNumberofVisitsperTarget 3

MinimumPlanetSize 0.5–1.0EarthArea

GeometricAlbedo Earth’s

SpectralRangeandResolution 6.5–18μm;R=25[50]

CharacterizationCompleteness Spectraof50%ofDetectedor10Planets

GiantPlanets JupiterFlux,5AU,50%ofStars

MaximumTolerableExozodiacalEmission 10timesSolarSystemZodiacalCloud

*Therearetwodefinitionsintheliteraturefortheouterlimitofthehabitablezone.Thefirstis1.5AUscaledtotheluminositytothe½powerbasedonKastingetal.(1993).Thesecondis1.8AUscaledinthesamewayfromForget&Pierrehumbert(1997).

Chapter4

102

Characterizationcompleteness:Whileitwillbedifficulttoobtainspectraofthefainterorlesswellpositionedplanets,werequirethatthemissionbecapableofmeasuringspectraofatleast10(~50%)ofthedetectedplanets.

Visitations:Multiplevisitsper starwillbe required toachieve requiredcompleteness, todistinguishitfrombackgroundobjects,todetermineitsorbit,andtostudyaplanetalongitsorbit. Themissionmustbecapableofmakingatleastthreevisitstoeachstartomeetthecompletenessandotherrequirements.

MultiplePlanets:Afterthecompletionoftherequirednumberofvisitationsdefinedabove,themissionshouldbeabletocharacterizeaplanetarysystemascomplexasourownwiththree terrestrial‐sized planets assuming each planet is individually bright enough to bedetected.

OrbitDetermination:After thecompletionof therequirednumberofvisitationsdefinedabove,themissionshallbeabletolocalizethepositionofaplanetorbitinginthehabitablezonewithanaccuracyof10%ofthesemi‐majoraxisoftheplanet’sorbit.Thisaccuracymaydegradeto25%inthepresenceofmultipleplanets.

4.2 ObservatoryConceptsAninterferometerisacompellingchoicefortheoveralldesignofaflagshipmission.Atmid‐infrared wavelengths the angular resolution needed to resolve an Earth‐Sun analog at adistanceof10pc,wouldbe~50mas,sothataconventionalsingle‐telescopedesignwouldneed a primarymirror with a diameter larger than 40m. By using separated apertures,baselines of hundreds of meters are possible. Nulling interferometry is then used tosuppress the on‐axis light from the parent star, whose photon noise would otherwiseoverwhelm the light from the planet (Bracewell 1978).Off‐axis light ismodulated by thespatial response of the interferometer: as the array is rotated, a planet produces acharacteristic signal,which canbedeconvolved from the resultant time series (Bracewell1978;Légeretal.1996;Angel&Woolf1997).Imagesoftheplanetarysystemcanbeformedusinganextensionoftechniquesdevelopedforradiointerferometry(Lay2005).

Aprobe‐scalemissionwouldhave collecting aperturesdeployedona single ‘structurally‐connected’ spacecraft, and a flagshipmissionwould have the very large formation‐flyingsystems to achieve the strategic goals of life detection and planet imaging. We brieflydescribe concepts for both a small structurally connected interferometer and a largerformation‐flyingsystem.

4.2.1 ProbeScaleMissionConcept

StructurallyConnectedInterferometerAn example of a small structurally‐connected infrared interferometer concept is theFourier‐Kelvin Stellar Interferometer (FKSI), which is a two‐telescope passively coolednulling interferometer operating with observing wavelengths between 3 and 8 µm (seeDanchietal.2003a,b),showninFigure4‐5.TheFKSIobservatorywilloperateatthesecondSun‐EarthLagrangepoint(L2)inalargeamplitudeLissajious(orhalo)orbit.Thebaselinedesign employs 0.5‐m apertures. The telescopes employ two flat mirrors (siderostats)mounted12.5m apart on composite support booms,minimizing alignment requirementsforthebeamsthatentertheinstrumentpackage.Theboomssupportsunshadesthatallowpassivecoolingofthestructureto60–65K,reducingthermalnoiseinthetelescopesystem

InfraredImaging

103

to a level that is negligible compared to that from the local zodiacal cloud (zodiacalbackgroundlimitedperformance)overmostoftheinstrumentpassband.

ThemechanicaldesignisappropriateforthesmallestofthetheAtlasVlaunchvehicleswitha 4‐m diameter fairing. The design minimizes the number of deployments. Thedeploymentsarelistedasfollows:(1)and(2)thetwosiderostatsmountedontheendsofthe booms, separated by 12.5 m, deployed using the well proven MILSTAR hingetechnology. The booms aremade from composite structures for an optimal strength‐to‐weightratio,andthefixedsunshadesaremountedtothem;(3)ahigh‐gaingimbaledS‐bandantennaforthedatadownlink;(4)asolararrayforpower;and(5)aradiator.

Figure45.Asmallstructually‐connectedinterferometermissionconcept.Twotelescopesandpassivecoolingareusedforobservationsinthe3–8µmband.(W.C.Danchi,NASAGSFC)

A schematic diagram of the optical design is shown in Figure 4‐6. There are five majorsubsystems in the instrument payload, which is mounted through a low‐mechanical‐frequency coupling and gamma‐Al struts to the spacecraft bus. These include: (1) theboom‐truss assembly that holds the siderostatmirrors and the non‐deployed sunshields;(2) themain instrumentmoduleandopticalbenchassembly,which includesapairofoff‐axis parabolic mirrors for afocal beam size reduction, and also a fast steering mirrorassemblyforthecontroloffinepointing,twoopticaldelaylines(onetobeused,theotherisa spare), and two dichroic beam splitters, one for the angle tracker and fringe trackerassemblies (0.8–2.5µm), the other for the science band of 3–8µm; (3) the angle trackerassemblyforfinepointing;(4)thefringetrackerassemblyforstabilizingthefringesforthenullinginstrumentassembly;and(5)thenullinginstrumentassemblyitself.

The nulling instrument is based on a modified Mach‐Zehnder beamsplitter design thatmaximizesthesymmetryofthetwobeams,helpingtoensureadeepnull.Otherelementsinthe system include shutters for alignment and calibration purposes, an assembly foramplitude control of the fringe null, and optical fibers forwavefront cleanup. The fibershelp achieve the desired null depth with reduced tolerances on the preceding optical

Chapter4

104

components(inordertoreducemanufacturingcosts).Thedarkandbrightoutputsaresentto their respective Focal Plane Arrays (FPAs), which could be long wavelength HgCdTearraysfromTeledyne(previouslyRockwell)operatingat35K,orSi:Asarraysoperatingat8K, based on those for theMIRI instrument on JWST. The FPAs for the fringe and angletrackers are operated at 77 K and are based on theHgCdTe arrays used on theNICMOSinstrumentonHST.Cryogenicdelaylinesareusedtoequalizethepathlengthsbetweenthetwosidesoftheinterferometer.

Figure 46. Schematic design of the boom and instrument module subsystem for FKSI. The varioussubassembliesareasnotedinthefigure.(W.C.Danchi,NASAGSFC)

Whentheinstrumentisinoperation,itrotatesslowlyaroundthelineofsighttotheobject.Insodoing,aperiodicsignal isgeneratedat theoutputof thenuller instrument,which isnon‐zeroifthereisaplanetorspatiallyresolvedsourcewithinthetransmissionpatternofthenuller.Datafromthebrightportisalsotakenprovidingmeasurementsofthevisibilityof the object. Simulations have been performed, both for nulling operation as well asimaging (Danchi et al. 2003a; Barry et al. 2005, 2006). Table 4‐2 lists the types ofmeasurements that can bemadewith such as system; the companion figure, Figure 4‐2,shows thatasmall interferometer iscapableofmeasuring theemissionspectraofknownexoplanetswithalargerangeofsemi‐majoraxes.

Theconceptdiscussedabovehasundergonea thorough integratedanalysisandmodelingstudyof thestructureandoptics tovalidate thedesignandensure itmeetsrequirements(Hydeet al. 2004). Amajor concernwas thepropagationof reaction‐wheeldisturbancesfrom the spacecraft bus through the bus‐instrument package coupling up to the boomsystem. Contributionstotheperformanceerrorbudgetfortherequired10–4nullarewell

InfraredImaging

105

understood and have been reported in Hyde et al. (2004). The null depth requirementdrivestheinstrumentperformance.OthersourcesofnoiseincludetheFringeTracker(FT),theOpticalDelayLine(ODL),theAttitudeControlSystem(ACS),andtheboomsthemselves.Theboomsystemhadthelowestordermodesat5.6and7.3Hz.Thismodelingverifiedthatthe closed‐loop performance requirement could be met or exceeded at all wheel speedsfrom0.1to100Hz(Hydeetal.2004).

PerformanceTable 4‐2 lists the planetary characteristics andwhat can be determined using amodestnulling interferometer. This list includes planetary temperature, radius, mass density,albedo,surfacegravity,andatmosphericcomposition,aswellasthepresenceofwater.Todate, progress has been made on the physical characteristics of planets largely throughtransiting systems, but a small planet‐finding interferometer can measure the emissionspectraofalargenumberofthenon‐transitingones,aswellasmoreprecisespectraofthetransitingones.

Table42.CharacteristicsofexoplanetsthatcanbemeasuredusingFKSI

Parameters WhatFKSIDoes

Removessin(i)ambiguity MeasurePlanetCharacteristics

Temperature MeasureTemperaturevariabilityduetodistancechanges MeasurePlanetradius MeasurePlanetmass EstimatePlanetalbedo CooperativeSurfacegravity Cooperative

Atmosphericandsurfacecomposition Measure

Timevariabilityofcomposition Measure

Presenceofwater Measure

PlanetarySystemCharacteristics

Influenceofotherplanets,orbitco‐planarity Estimate

Comets,asteroids,zodiacaldust Measure

Manymolecularspecies(suchascarbonmonoxide,methane,andwatervapor)havestrongspectralfeaturesinthe3–8µmregion,ascanbeseeninFigure4‐2,whichdisplaysmodelatmospheres for exoplanets calculated by Seager (2005), for planets at various distancesfrom the host star. The red curve shows the theoretical spectrumof a very hot, close‐inplanetat0.05AU,whilethebluecurvedisplaysspectrumforamuchcoolerplanettentimesfurther out, at 0.5 AU. Also displayed are sensitivity curves for the IRAC instrument onSpitzer (light blue) and FKSI (purple). Clearly such a mission concept has sufficientsensitivitytodetectandcharacterizeabroadrangeofexoplanets.

Chapter4

106

Figure47.ExpectedperformanceforPegaseandFKSIcomparedtotheground‐basedinstruments,for30minintegrationtimeand1%uncertaintyonthestellarangulardiameters.(Defrèreetal.2008)

Figure48.Skycoverageafter1yearofobservationofGENIE(darkframe),ALADDIN(lightframe)andPegase(shaded area) shown with the Darwin/TPF all sky target catalogue. The blue‐shaded area shows the skycoverageofaspace‐basedinstrumentwithaneclipticlatitudeinthe[–30˚,30˚]range(suchasPegase).TheskycoverageofFKSIissimilartothatofPegasewithanextensionof40˚insteadof60˚.(Defrèreetal.2008)

InfraredImaging

107

Initialstudieswithveryconservativeassumptionsofan8‐mboomlength,140exoplanets(known at that time), 120‐s on‐source integration time, 15‐nm RMS path length error,telescopetemperatureof65K,andasmallsunshieldwitha±20degreefieldofregard;gave~25detections.Currrently,thereareabout250knownplanets,sowiththoseassumptionsandthesamedetectionratioofabout18%,itshouldbepossibletodetectandcharacterizeabout 45 exoplanets, and a larger sunshadewith a field of regard of ± 45 degreeswouldapproximatelydoublethisnumberto~90planets.

Recentstudiesof thedetectabilityofdebrisdisks (Defrèreetal.2008)demonstrated thatthe residualpathlengtherror fora small system likeFKSI shouldbeof theorderof2nmRMS, which would significantly increase the number of exoplanets. As a conservativeestimate,weexpectthatasmallsystemcoulddetect(e.g.,removethesin(i)ambiguity)andcharacterizeabout75–100knownexoplanets.

These recent studies also have shown that a smallmission is ideal for the detection andcharacterizationofexozodialanddebrisdisksaroundallcandidatetargetstarsintheSolarneighborhoodasseeninFigures4‐7and4‐8.(Defrèreetal.2008).Indeed,theperformanceleveloftheFKSIsystemisoftheorderof1SolarSystemZodiin30minutesofintegrationforaG0starat30pc.Withasmallsunshade,thissystemcouldobserveoftheorderof450starsintheSolarneighborhood,andwithalargerfield‐of‐regard,suchas±45degrees,theexozodisanddebrisdisksofabout1000starscouldbestudied.

Although the telescopes are somewhat larger than has been discussed in some of theexistingmissionconcepts (e.g.,1–2m)andaresomewhatcooler (e.g.,<60K)so that thesystem can operate at longerwavelengths, it is possible for a small infrared structurallyconnectedinterferometertodetectandcharacterizeSuperEarthsandeven~50–75Earth‐sized planets around the nearest stars. This is especially important now that there isevidencethatlowermassplanetsmaybeverycommon,basedonthedetectionofthe5.5‐Earth‐massplanetusing themicrolensingapproach(e.g.,Beaulieuetal.2006;Gouldetal.2006).

Furtherstudiesofthecapabilitiesofasmallinfraredstructurally‐connectedinterferometerarenecessarytoimproveuponourestimatesofsystemperformance.

4.2.2 FlagshipMissionConcept

FormationFlyingInterferometerFigure 4‐9 depicts amission concept capable of detecting a large number of Earth‐sizedplanets. This Emma X‐Array formation‐flying design is the culmination of more than adecade of study by NASA and ESA. Light from the target star is reflected from the fourcollectormirrors (~2mdiameter) and focusedonto the input apertures of the combinerspacecraft approximately 1 km away. To observe a target system, the collector array isrotated slowly about the line‐of‐sight, using a combination of centimeter‐precisionformationflyingandcarefulattitudecontroltomaintainthepointingofthebeams.Thelightenteringthecombinerpassesthroughaseriesofoptics(Figure4‐10)thatperformsbeamsteering and compression, delay compensation, intensity and dispersion correction (see“theAdaptiveNuller”below),pairwisebeamcombinationwithπphaseshift(thenullers),cross‐combinationofthenulledbeams,single‐modespatialfiltering,and,finally,dispersionanddetectionof thenulled light on the science camera.The time series of outputdata issynthesized into an image. A key advantage of the X‐Array configuration is its very highangularresolution,asillustratedbythesimulatedimageshowninFigure4‐11.

Chapter4

108

Figure49.‘EmmaX‐Array’formationflyingconcept.

Figure410:Schematicofbeamcombineroptics.

PerformanceThe planet‐finding capability for a formation‐flying Emma X‐Array design is shown inFigure4‐12. Amissionwith2‐mdiametercollectorswouldbecapableofdetectingmorethan100Earth‐sizedplanetsaroundnearbystars,primarilyF,G,andKspectraltype,overaperiodof2years,assumingη⊕=1.TheperformancemodelisdescribedinLayetal.(2007)and includes the dominant sources of noise: photon noise from local zodi, exo‐zodi andstellar leakage, and instability noise (Lay 2004) from fluctuations in the phasing andintensitybalancewithintheinstrumentoptics.Followingacandidateplanetdetectionandfollow‐up confirmation, a longer period of observing time is dedicated to spectroscopiccharacterization,usingthesameobservingprocedureasfortheinitialdetection.DetailsofadesignstudyforthearrayaregivenbyMartinetal.(2008a).

InfraredImaging

109

Figure 411: Simulated ‘dirty’ image from a 400 × 67‐m Emma X‐Array, prior to deconvolution. Angularresolution is 2.5 mas. Planet locations are indicated by red arrows. Simultaneous full resolution (R ~ 100)spectroscopyforallobjectswithinthefieldofviewisanaturalby‐productofinterferometricobserving.Whiledata from the detection and orbit‐determination phases should be sufficient for a coarse spectrum, a deepcharacterization will require significant integration time. Detection of CO2 for an Earth at 5 pc using 2‐mcollectorswill require ~24 hours of integration (SNR of 10 relative to the continuum). The narrower ozoneabsorption line requires 16 days at 5 pc. For ozone at 10 pc, integration times as long as 40 days could beneeded,fallingto~6dayswith4‐mdiametercollectors.(O.P.Lay,JPL)

4.3 Technology

4.3.1 ExperimentsinNullingInterferometry,1999–2009Results obtained in the lab havemet the requirements for infrared nulling needed for aprobe‐scalemissionandareclosetomeetingthoseofaflagshipmission.Laboratoryworkwith the Adaptive Nuller has indicated that mid‐infrared nulls of 1.0 × 10–5 are nowachievablewithabandwidthof34%andameanwavelengthof10µm;thissuggeststhatatthesubsystem‐levelthenullingperformanceforaflagshipmissionisnearTRL4(cryogenictesting would be needed for TRL 5). These results also show that the achromatic phaseshifters,theAdaptiveNuller,andthemid‐infraredsingle‐modefibersthatarecontributingto these results are nowmature technology. A summary of the past decade’s research innullinginterferometryisgivenintheparagraphsthatfollow.

Chapter4

110

Figure 412: Predicted number of Earths detectable by Emma X‐Array architecture as a function of elapsedmissiontimeandcollectordiameter.(O.P.Lay&S.Hunyadi,JPL)

Deepnullsofnarrowbandlighthavebeenachievedusingbothpupil‐planeandfocalplanenullingtechniques,andtheextensionoftheseresultstobroadbandandrandomlypolarizedsources is now the remaining challenge. To achieve deep nulls in an interferometer, theoptical properties of the combined beamsmust be quite similar. The properties includeintensity balance, polarization angle, polarization dispersion, and wavefront differences.Someoftherequirementsforthebalanceinthesepropertiescanbemitigatedbytheuseofsingle‐mode spatial filtering at the detector, for example by using a single‐mode fiber toreceive the light. The components of the optical system will introduce defects into thepristine wavefront received from the stellar source, adding wavefront deformations andpolarizationchanges. Ifthesedefectsaremoreorlessthesameineachinputbeamoftheinterferometer, the system will still be capable of nulling. For a broadband source, theeffects of the defects need to be controlled across the entire nulling waveband, and thisrequirementhasbeenaddressedwiththedevelopmentoftheAdaptiveNuller(Petersetal.2008), which is a system designed to correct intensity and phase differences across thebandforbothpolarizations.

Since 1999whenOllivier (1999) reported laser nulls of~1.7 × 10–4 and Serabyn (1999)reported 5.0 × 10–5, progress has been made in both null depth and bandwidth. Deep,single‐polarization laser nulls in the mid‐infrared waveband of order 3.0 × 10–6 werereportedbyMartinetal.(2003),andSamueleetal.(2007)reportednullsoforder1.1×10–7atvisiblewavelengths. Similarly, sinceWallaceetal. (2000) reportedbroadbandnullsof7.4×10–5at18%bandwidth,Samueleetal.(2007)achieved1.0×10–6at15%bandwidthinthe visible and Peters et al. (2008) have shown through narrowbandmeasurements that1.0× 10–5 nulls at 34% bandwidth in the mid‐infrared are attainable. These results andothers of note are included in Figure 4‐13,which illustrates the experiments undertakensince1999.

InfraredImaging

111

10-7

10-6

10-5

10-4

10-3

0 5 10 15 20 25 30 35 40

Null D

epth

Bandwidth (%)

Brachet 2005

Samuele et al. 2007

Peters 2009

Gappinger 2009

Gappinger 2009

Mennesson et al. 2003

Martin et al. 2003

Martin et al. 2005

Haguenauer and Serabyn 2006

Samuele et al. 2007

Serabyn et al. 1999

Mennesson et al. 2003 Wallace et al. 2000

Ollivier 1999

Serabyn et al. 1999

Martin et al. 2003

Mennesson et al. 2003

Schmidtlin et al. 2006

Buisset et al. 2006

Ergenzinger et al. 2004

Vosteen et al. 2005

MAII breadboards

Visible wavelengths H band K band

N band Results of Laboratory ExperimentsIn Nulling Interferometry (1999-2009)

Figure413.Chartofnulldepthsachievedsince1999.Theblueshadedareahighlightsthenullingperformanceneededtobedemonstratedinthelabinpreparationforaflagshipmission.(P.R.Lawson,JPL)

Serabyn’s early nuller, known as the SIM roof‐top nuller, was used for both laser andbroadband nulling tests near the visible waveband. It used a geometric field inversiontechniquebutwasasymmetricinthatabeamsplitter‐compensatorplatearrangementwasused.AmoresymmetricalnullerwasthenbuiltfortheKecktelescopes,themodifiedMach‐Zehnder (MMZ) nuller (Crawford et al. 2005) in which field inversion was performedoutsidethenullerandthebeamstraversedtwobeamsplitterseach,allowingsymmetricaltreatment(Figure4‐14).

Figure 414. (left) MMZ nuller built for Keck telescopes, (right) single‐polarization laser null at 10.6 µmshowingnulldepthof5×10–7.

Chapter4

112

Figure415.(left)conceptforanullerusingafiberbeamcombiner,(right)unpolarizedbroadbandnullof18%bandwidthat1.65mmshowingnulldepthbetterthan1×10–4.

TheKecknuller (Serabynetal.2004)employs twoMMZnullers tonull four inputbeamsandoperates in themid‐infrared.Morerecently, laboratoryworkhasusedMach‐Zehnderschemes to produce symmetrical nulling systems for the Planet Detection Testbed (PDT)(Martinetal.2006)andtheAchromaticNullingTestbed(ANT)(Gappingeretal.2009),forexample.ThePDThasnulledfourlaserbeamsat10.6µmwavelengthatthe2.5×10–5level.

Asomewhatdifferentinterferometricbeam‐combinationschemehasalsobeenproposedinEurope(Wallneretal.2004;Karlsonetal.2004).Bycombiningbeamsdirectlyonasingle‐mode fiber, the complex interaction of the light with beamsplitter and antireflectioncoatings can be eliminated. Conceptually, such a beamcombiner can be both achromaticandpolarization insensitive,at thecostofsomeoptical throughputefficiency. Small‐scaletestbedsforthistypeofbeamcombinerarebeingdevelopedandhaveachieveddeeplasernulls of 2.0 × 10–4 (Mennesson et al. 2006) in the visible and broadband nulls of order1.0×10–4at18%bandwidthatJ‐band(Martinetal.2008b)(Figure4‐15).

4.3.2 TechnologyforProbeandFlagshipMissions

MidIRSpatialFiltersSpatialfiltersareanessentialtechnologyfornulling interferometry.Theycanbeused toreduce complex optical aberrations in theincomingwavefrontstosimpleintensityandphase differences (which are more readilycorrected), thus making extremely deepnulls possible. Spatial filters may beimplemented inavarietyofways, includingsingle‐mode fiber‐optics made fromchalcogenide glasses, metallized waveguidestructures micro‐machined in silicon, orthrough the use of photonic crystal fibers.Two types of fiber spatial filters have beensuccessfullydeveloped:polycrystalinesilverhalide fibers have been developed by Prof.AbrahamKatziratTelAvivUniversity (TAU) in Israel; andchalcogenide fibershavebeendevelopedbyDr.JasSangheraattheNavalResearchLaboratory(NRL).

Figure416.Single‐modeChalcogenidefibersdevelopedbytheNavalResearchLaboratory

InfraredImaging

113

The20‐cmlongchalcogenidefibersdevelopedbytheNavalResearchLaboratory(showninFigure4‐16)wereshowntodemonstrate30‐dBrejection(afactorof1000)ofhigherordermodes andhave an efficiency of 40%, accounting for both throughput andFresnel losses(Ksendzov et al. 2007). The transmission losses weremeasured to be 8 dB·m–1, and thefibersareusableuptoawavelengthofabout11µm.ThechalcogenidefibersdevelopedattheNavalResearchLaboratorywereused successfully in theAchromaticNullingTestbedandarecurrentlyinuseintheAdaptiveNullertestbed.

The 10–20 cm long silver halide fibers thatwere developed by Tel Aviv Universitywereshown to demonstrate 42‐dB rejection (a factor of 16,000) of higher order modes withtransmissionlossesof12dBm–1(Ksendzovetal.2008).Thishighrejectionofhigher‐ordermodeswasaccomplishedwiththeadditionofaperturingoftheoutputofthefibers,madepossible by the physically large diameter of the fibers. Silver halide fibers should inprinciplebeusableuptoawavelengthofabout18µm,althoughthe labtestsat JPLwereconducted only at 10µm.This is the first time silver halide fiberswere demonstrated tohavesingle‐modebehavior.

Figure417. Schematic of theAdaptiveNuller. Uncompensated light that enters the device is split into twolinearpolarizations,andeachpolarizaionistreatedseparatelythenrecombined.Lightisdispersedandimagedontoadeformablemirror(DM).PistonofelementsintheDMadjustsphaseataparticularwavelength,andtiltofthesamepixeladjustsrelativeintensitybyshearingtheoutputpupilatthatwavelength.

AdaptiveNullingandAchromaticPhaseShiftersThevariationsinamplitudeandphasethatmaybepresentintheincomingwavefrontmustbecorrected inordertoachievedeepnullsateverywavelength.TheAdaptiveNullerwasdesigned to correctphaseand intensityvariationsas a functionofwavelength, in eachoftwo linear polarizations. In principle this should allow high‐performance nullinginterferometry, while at the same time substantially relaxing the requirements on thenullinginterferometer’sopticalcomponents.Thegoalofthetestbedwastodemonstrate,ina3‐μmbandcenteredatawavelengthof10μm,thecorrectionoftheintensitydifferencetolessthan0.2%RMS(1σ)betweentheinterferometer’sarms,andatthesametimecorrect

Chapter4

114

the phase difference across the band to < 5 nm RMS (1 σ). This overall correction isconsistent with a null depth of 1 × 10−5 (1 part in 100,000) if all other sources of nulldegradationcanbeneglected.

TheAdaptiveNullerwasdesigned to correct these variations,matching the intensity andphase between the two arms of the interferometer, as a function ofwavelength, for bothlinear polarizations. A deformable mirror is used to adjust amplitude and phaseindependentlyineachofabout12spectralchannels. Aschematicoftheadaptivenullerisshown inFigure4‐17.Twoprismssplit the incoming light into its two linearpolarizationcomponentsanddivideitintoroughlyadozenspectralchannels.Thelightisthenfocusedontoadeformablemirrorwherethepistonofeachpixelindependentlyadjuststhephaseofeach channel. By adjusting the angle of the reflected beam so that the beam is slightlyvignetted at an exit aperture, the intensity is also controlled. The adaptive nuller uses abroadbandthermalsourcetogeneratelightinthe8–12μmwavelengthband.AphotographofthetestbedisshowninFigure4‐18.

Figure 418. The Adaptive Nuller Testbed (AdN). The Adaptive Nuller demonstrated phase and intensitycompensationofbeamswithinanullinginterferometertoalevelof0.12%RMSinintensityandlessthan5nmRMSinphase.ThisversionoftheAdNoperatesoverawavelengthrangeof8–12µm.Thelong‐focusparabolas,usedinAdNareseenontheupperright.

The initial results of this research have been published by Peters et al. (2008). InMarch/April 2007, the testbed achieved its primary goal of demonstrating phasecompensation to better than 5 nm RMS across the 8–12 μm band and intensitycompensation tobetter than0.2%RMS. Narrow‐bandmeasurementsofnulldepth in thechannels across the passband indicated a null depth of 1.0× 10–5, would be attainable.When demonstrated using the full bandwidth of the Adaptive Nuller, it would be thedeepestbroadbandnulleverachievedbyamid‐infrarednullinginterferometerandwoulddemonstratetheflightrequirementforaflagshipmission.

InfraredImaging

115

CryocoolersWiththeAdvancedCryocoolerTechnologyDevelopmentProgram(ACTDP),theTerrestrialPlanetFinder(TPF)andJWSTprojectshaveproduceddevelopmentmodelcoolersthathavemetorexceededtheirperformancerequirements,whicharetoprovide~30mWofcoolingat 6 K and ~150 mW at 18 K. This demonstrates the approach to cooling the sciencedetectortoatemperaturelowenoughtorevealtheweakplanetsignals.ThisactivityisatTRL6.

CryogenicdelaylinesTheDutchcompanyTNOScienceandIndustryledaconsortiumthatdevelopedacompactcryogenicOpticalDelayLine(ODL)foruseinfuturespaceinterferometrymissionssuchasESA'sDarwinandNASA'sTPF‐I(Figure4‐19).Theprototypedelaylineisrepresentativeofa flightmechanism. The ODL consists of a two‐mirror cat's eyewith amagnetic bearinglinearguidingsystem.TNOanditspartnershavedemonstratedthataccurateopticalpath‐length control is possiblewith the use ofmagnetic bearings and a single‐stage actuationconcept.Activemagneticbearingsarecontactless,haveno frictionorhysteresis,arewearfree,andhavelowpowerdissipation.ThedesignoftheDarwinbroadbandODLmeetstheESA requirements, which are an OPD stroke of 20 mm, stability of 0.9 nm RMS (with adisturbance spectrum of 3000 nm RMS, < 20 mW power dissipation (2 mW with flightcabling), outputbeam tilt<0.24microradians, output lateral shift<10‐µmpeak‐to‐peak,wavefront distortion < 63 nm RMS at 40 K, and wavelength range (0.45–20 µm). TheDarwinODLisrepresentativeofafutureflightmechanism,withallmaterialsandprocessesused being suitable for flight qualification. Positioning is donewith a voice coil down tosubnanometerprecision.Theverificationprogram,includingfunctionaltestingat40K,hasbeencompletedsuccesfully.ThistechnologyisatTRL6(seeFridlundetal.2008).

Figure 419. Overview of the Optical Delay Line with the outer fixed structure holding the active parts formagnetic levitation and axial motion, and the inner moving structure that includes the optical components.(CourtesyofESAandTNO)

Chapter4

116

ThermalshieldsScience requirements for JWST have driven the need for a deployable, low areal‐density,high thermal‐performance efficiency (effective emittance, e* of 10–4 to 10–5) sunshield topassively cool the Observatory Telescope Elements and Integrated Science InstrumentsModule(OTE/ISIM).Thethermalperformancedictatedtheneedforasunshieldconsistingofmultiple,space‐membrane layerssuchthatthe~200kWoftheSun’senergy impingingon the Sun‐facing layerwould be attenuated such that the heat emitted from the rear orOTE‐facing membrane would be < 1 W. This would enable the OTE/ISIM to operate atcryogenictemperaturelevels(<50K)andhaveatemperaturestabilityof0.1to0.2Koverthefield‐of‐regardpointingre‐alignments.

Figure420.TheJamesWebbSpaceTelescope.(CourtesyNASA/JWST)

Basedon these top‐levelmissionrequirementsasunshieldwasdevelopedemploying fivemembrane layers with pitch and dihedral separation angles of 1.6 and 2 degrees,respectively.Throughtheuseofalowsolarabsorptance‐to‐emittanceratio(αS/εH)coatingontheSun‐facingsurface(layer1)andaninfraredhighlyreflectivecoatingonthemajorityoftheothermembranes’surfaces,theSun’sheatloadabsorbedbytheSun‐facingmembraneis reduced and the resulting infrared energy re‐emitted to the other layers would bereflectedlaterallyouttheopensidesofthesunshield,thusgreatlyattenuatingtheresidualinfraredenergyemittedbytheOTE/SIM‐facingmembranesurface(layer5).Fromavarietyof trade studies the Sunshield evolved into a five‐membrane system as shown in Figure4‐20withmembranes140to150m2insizethataresupportedbysixspreaderbar/boomassembliesandarelightlytensionedinadeployedconfigurationbyconstant‐forcespring‐loadedcatenarycableassembliesattachedtotheperipheryofeachmembrane.Thepresentsunshadedesign for JWST isatTRL6andmeets thesimilarrequirements for the flagshipmissionandismorethanadequateforasmallmission(Kurland2007).

InfraredImaging

117

DetectorTechnologyDetectorsorSensorChipAssemblies(SCAs)fromtheJWSTMid‐InfraredInstrument(MIRI)have been developed by Raytheon Vision Systems (RVS) having 1024 × 1024 pixelsfabricated from arsenic‐doped silicon (Si:As) andwhich consist of a detector layer and areadoutmultiplexer.Thedetectorstructureisabackside‐illuminatedsiliconsubstratewithanactive‐detectorlayergrownontopofthat,followedbyablockinglayer(tominimizethedark current), and then indium bumps. The readout multiplexer is a cryogenicallyoptimized electronic circuit chip having a source follower (transistor) for each detectorpixel,whichthenmultiplexesthosepixelsdowntofourvideooutputsthatarereadbytheremaining instrument electronics. The two components of the SCAs aremated togetherthroughmatching indiumbumps. The requiredperformance levels are dark current lessthan 0.03 electrons per second, readout noise less than 19 electrons, and quantumefficiency>50%. Thesedetectorshavebeenrigorously testedandhavemetorexceededthese requirements for JWST, giving aTRLof 6 (Ressler 2007). Thesedetectorsmeet orexceed the requirements for both a small structurally connected interferometer missionandtheflagshipmission.SeeFigure4‐21foranexampleofacompletefocalplanemodulefortheMIRIinstrumentonJWST.

Figure421.TheMIRIFocalPlaneModulecontainstheSCA,afanoutboard,andthemountingstructureneededtopositionandisolatethesensor.(CourtesyNASA/JWST)

IntegratedModelingObservatorySimulation:Demonstrateasimulationoftheflightobservatoryconceptthatmodels the observatory subjected to dynamic disturbances (e.g., from reaction wheels).ValidatethismodelwithexperimentalresultsfromatleastthePlanetDetectionTestbedatdiscrete wavelengths. Use this simulation to show that the depth and stability of thestarlightnullcanbecontrolledovertheentirewavebandtowithinanorderofmagnitudeofthe limits required in flight todetectEarth‐likeplanets, characterize theirproperties,andassesstheirhabitability.ThisactivityisatTRL5(Lawsonetal.2007).

Chapter4

118

4.3.3 AdditionalTechnologyforaFlagshipMission

Fourbeamnulling:DualbeamchoppingandspectralfilteringThe Planet Detection Testbed (PDT) is a breadboard system intended to demonstratetechniquesthatwillberequiredfordetectionofexoplanetsusingamid‐infraredtelescopearray.Asabreadboardsystemoperating inanormal roomenvironment, itdiffers fromaflight‐demonstrationsysteminbothits layoutandinthefluxlevelsthatituses,whicharemuchhigher thanwouldbe encountered in space. It does, however, embody the controlsandsensorsnecessaryforoperationinspace.Furthermore,thedifferencesdonotaffectthekey goal of the testbed, which is to demonstrate faint planet detection using nullinginterferometry,phasechoppingandaveraging.

Figure422.PhotonratesfortheplanetdetectionprocessforflightandPDT.

Key exoplanet‐detection techniques are being tested and demonstrated in the PDT. Toreachtheflightgoalofdetectionofaplanet107timesfainterthantheparentstar,aseriesofsteps(illustratedinFigure4‐22)mustbetaken:

• Thestar’sapparentintensityisreducedrelativetotheplanetbyafactorof100,000.Thisisdonebyinterferometricnulling.

• Theinterferometerarrayisrotatedaroundthelineofsighttothestartosearchthewholeregionaroundthestarforacharacteristicplanetsignature.InthePDTthisisdonebycontrollingtheopticalpathrelationshipsbetweentheplanetsourcebeams,simulatingthephasechangescausedbyrotationofthetelescopearrayaroundthelineofsighttothestar.Duringthisrotation,stablenullsneedtobemaintained;thePDThassystemsandcontrolloopsanalogoustothoseneededforflightinordertocontroltheinstabilitynoiseandmaintaindeepnulls.

InfraredImaging

119

• Theplanetsignalismodulatedagainstthebrightradiationbackground.Thisisdoneusingphasechopping.Thecombinationofrotation,phasechopping,andaveragingovertimereducesthenoiselevelbyafactorof100to10–7ofthestellarintensity.

• Thetechniqueofspectral fittingusescorrelationsbetweennull fluctuationsacrossthespectralbandtoreducetheinstabilitynoise.Thisyieldsafurtherfactorof10inreductionofthenoiselevel.

Thus, the combination of these four techniques yields the necessary performance. Thecurrentobjective is todemonstratethe first threepartsof thisprocess; thestablenulling,thearray(orplanet)rotationandthephasechopping.Atthetimeofwriting,thePDThasdemonstrated detection of a planet 2 million times fainter than the star using somesimplifiedmethods.Thetestbedisnowbeingpreparedforfaintplanetdetectionusingthefirstthreetechniques.Thesuppressioneffectachievedusingthesetechniquesinwhichthenulled starlight at 10–5 is reduced to 10–7 will be representative of flight performance.Validationofthefourthtechnique,spectralfitting,willbelefttoafutureplan.

Figure 423.The Formation Control Testbed (FCT). Shown here are the two robots of the FCT. Each robotcarriescannistersofcompressedairthatallowthemtofloatoffapolishedmetalfloor.Thefloorisflattowithin2one‐thousandthsofaninchandspansamuchlargerareathanshownhere.Therobotsserveasthehardwareinterfaceandtestinggroundforflightsoftwaredevelopedforspaceapplicationsinformationflying.(CourtesyCaltech/JPL/NASA)

Guidance,Navigation,andControlThe Formation Control Testbed (FCT) was built to provide an end‐to‐end autonomousformation‐flying system in a ground‐based laboratory. The FCT provides an environmentfor system‐level demonstration and validation of formation‐control algorithms. Thealgorithms are validated using multiple floating test robots that emulate real spacecraftdynamics. The goal is to demonstrate algorithms for formation acquisition, formation

Chapter4

120

maneuvering, fault‐tolerant operations, as well as collision‐avoidance maneuvering. (SeeLawsonetal.2008andreferencestherein.)

TheFCTiscomprisedoftworobotswithflight‐likehardwareanddynamics,aprecisionflatfloorfortherobotstooperateon,ceiling‐mountedartificialstarsforrobotattitudesensingand navigation, and a “ground control” room for commanding the robots and receivingtelemetry.TherobotsandpartoftheflatfloorareshowninFigure4‐23.ThelayoutoftheFCT emulates the environment of a formation of telescopes that is restricted tomaneuveringinthesameplaneinspace,normaltothedirectionofthetargetstar.

To be as flight‐like as possible, each robot is equipped with a typical single‐spacecraftattitude‐control suite of reaction wheels, gyros, and a star tracker. Thrusters are alsoavailable forattitudecontrol.Eachrobothasa lower translationalplatformandanupperattitudeplatform.Theattitudeplatformisthe“spacecraft”andiscompletelydisconnectedfromthetranslationalplatform.

• Theattitudeplatform/spacecrafthousestheavionics,actuators,sensors,inter‐robotand“ground”‐to‐robotwirelesscommunicationantennas,andspacecraftprocessors.

• Thetranslationalplatformprovidesbothtranslationalandrotationaldegreesoffreedomtotheattitudeplatformvia(i)linearairbearings(theblack,circularpadsatthebaseofeachrobot)thatallowtheentirerobottofloatfreelyontheflatfloor,and(ii)asphericalairbearingatthetopoftheverticalstage(theblack,verticalcylinders).

In2007theFCTdemonstratedprecisionmaneuversusingtworobots,showingautonomousinitialization, maneuvering, and operation in a collision‐free manner. The key maneuverthat was demonstrated was representative of TPF‐I science observations. Repeatedexperimentswiththerobotsdemonstratedformationrotationsthroughgreaterthan90°attentimestheflightrotationratewhilemaintainingarelativepositioncontrolto5cmRMS.Although the achievable resolution in these experiments was limited by the noiseenvironmentofthelaboratory, itwasnonethelessdemonstratedthroughmodelingthat intherelativelynoise‐freeenvironmentofspace,theperformanceofthesealgorithmswouldexceedTPF‐Iflightrequirements.

PropulsionSystemsMissionsemployingPrecisionFormationFlying (PFF) requirepropulsion capabilities thatexhibit low plume contamination, high thrust precision, and high power and propellantefficiency.Ionthrusterstypicallydeliverlow‐contaminationplumesandhighefficiencybyusingnoblegaspropellants,butconventionalthrustersprovidethrustlevelsoveranorderofmagnitude greater than the sub‐milli‐Newton (mN) tomN levels needed forprecision‐controlledformation‐flyingspaceinterferometermissions.

Aminiatureionthruster,theMiniatureXenonIon(MiXI)thrusterdevelopedatJPL,isanewoptionthatmeetsmissionneeds.OneparticularlyusefulcharacteristicoftheMiXIthrusterisitsincrediblylargethrustrangethatprovidessmoothamplitudemodulatedthrustinthe0.1–3.0mN in theamplitudemodulatedmodeand0.001–0.1mNof thrust inpulse‐widthmodulation(PWM)mode(patent‐pending).Withaminimumon‐timeoflessthan1ms,theMiXIthrustercanprovide impulsebitsof lessthan1µN⋅sat lowpower levels. ThePWMmodeoftheMiXIthrusterisachievedbyprecisioncontrolofthethrustervoltages.

InfraredImaging

121

4.3.4 FutureMilestones

ProbeScaleMissionTable4‐3isasummaryofthetechnicalreadinesslevels(TRLs)ofkeytechnologiesusedinthe probe‐scale mission (FKSI). These levels are estimated using NASA standardterminology(Mankin1995).

Many of the key technological hurdles for FKSI, as understood in the original studiesperformed in 2002–2003, have been solved by NASA flagship projects. In particular, theinvestments in JWST and TPF‐I/Darwin technologies can be directly used in the FKSIsystem.Asofearly2007,JWSTpassedallofitskeytechnologymilestones,includingthosefor cryocoolers, precision cryogenic support structures, detectors, cryogenicmirrors, andsunshades.AllofthesetechnologiescanbeadaptedtotheFKSImissionwithlittleeffort.

KeytechnologiesuniquetotheFKSImissionhavealsobenefitedfromNASA’sinvestmentsintheKeckInterferometerNullersystem(nowinoperationattheKeckObservatory)andtheTPF‐ItestbedsatJPL,includingthePlanetDetectionTestbedandtheAchromaticNullerTestbed.ThebroadbandnulldepthsfromthesetestbedsareaboutanorderofmagnitudebetterthantherequirementsfortheFKSImission.

Theoptical fibersneeded forwavefront cleanuphavebeen fabricatedand testedat roomtemperatureaspartof theTPF‐I technologyprogram. Currently two typesof fibershavebeen tested in the laboratory and work sufficiently well for the FKSI mission: these arechalcogenide fibers fabricated at the Naval Research Laboratory (NRL) (Aggarwal &Sanghera 2002) that operate from the near‐ to mid‐infrared, and silver halide fibersdeveloped at Tel Aviv University that operate in the longer wavelength part of themid‐infrared(seeKsendzovetal.2007andreferencestherein). Athirdtypeof fiber,ahollowglasswaveguide fiber developed at Rutgers University (Harrington 2001)may also be ofuse.Furthertestingisplanned,includingcryogenictestingofallfibers.

Table43.TechnicalReadinessLevelsforKeySubsystemComponents

Item Description TRL Notes

1 Cryocoolers 6 Source:JWSTT‐NAR2 Precisioncryogenicstructure(booms) 6 Source:JWSTT‐NAR3 Detectors(near‐infrared) 6 Source:HST,JWSTNircamT‐NAR4 Detectors(mid‐infrared) 6 Source:SpitzerIRAC,JWSTMIRIT‐NAR5 Cryogenicmirrors 6 Source:JWSTT‐NAR6 Opticalfiberformid‐infrared 4 Source:TPF‐I7 Sunshade 6 Source:JWSTT‐NAR

8 NullerInstrument 4–5 Source:KeckInterferometerNuller,TPF‐Iproject,LBTILBTIBLINCinstrument*9 Precisioncryogenicdelayline 6 Source:TNO

*Note:TherequirementfortheFKSIprojectisanulldepthof10–4ina10%bandwidth.LaboratoryresultswiththeTPF‐Itestbedshaveexceededthisrequirementbyanorderofmagnitude(Lawsonetal.2008).

Atthesystemlevel,anullerinstrumentfabricatedatJPLhasbeendeliveredtoandisinuseat the Keck Observatory. The FKSI nuller instrument is actually simpler than the Keckinstrument,asthelatterrequireschoppingoutthebackgroundcreatedbyawarmtelescopeandsky(notneededbyFKSI).IndeedthefirstdatafromtheKeckInterferometerNullerhas

Chapter4

122

recently been published (Barry et al. 2008). The Bracewell Infrared Nulling Cryostat(BLINC, Hinz et al. 2000) is a cryogenic nuller in use on the MMT, which has achievedsignificantresultsinstudiesofaccretiondisksaroundHerbigAe/Bestars.

AlthoughtheBLINCsystemisavacuumcryogenicnullersubsystem,andinmanyrespectsissimilartowhatisrequiredfortheFKSInullinginstrument,ahigh‐fidelityvacuumcryogenicnulling testbed that provides an environment with the expected space‐like signal andthermalnoise levelsandthatalso includesrepresentativedisturbance forces,wouldbeofgreatbenefit to theFKSImission itself,andcouldbeusedasarelatively lowcoststartingpoint for amuchmore complex cryogenic nulling testbed that is needed for the flagshipTPF‐I/Darwinmissions, as discussedbelow. Such a cryogenic testbedwould reduce costandtechnicalriskforboththeProbe‐classmissionaswellastheflagshipmission.

Finally, FKSI requires a precision cryogenic delay line with a stroke of the order of acentimeter. Fortunately, commercial firms in Europe, under contract from ESA havedevelopedsuchaprototypecryogenicdelay line,which isatahighTRL level. Thisdelaylinehasaprecisionof1.3nmanda strokeof10mm,which ismore thanadequate for asmallinterferometer.Also,twoshort‐strokecryogenicdelaylineshavebeenflownaspartof twoMichelson spectral interferometers, bothofwhichweredevelopedat theGoddardSpace Flight Center. One was the FIRAS instrument on the COBEmission, which won aNobel Prize in Physics for Drs. John Mather and George Smoot for cosmic microwavebackgroundresearch(seeforexample,Matheretal.1994,Wrightetal.1994).Asecondhasbeen flown successfully on the CASSINI mission, as part of the CIRS instrument(http://saturn.jpl.nasa.gov/spacecraft/inst‐cassini‐cirs‐details.cfm).

Insummary,wenotethetwoareasforfurthertechnologydevelopmentforasmallinfraredstructurally‐connectedinterferometer:(1)Atthecomponentlevel,cryogenictestingoftheopticalfibersforwavefrontcleanup;and(2)atthesystemlevel,avacuumcryogenicnullingtestbed.

SingleModeFiltersDemonstratedinaCryogenicVacuum,6–20µmThesingle‐modefibersusedintheAdaptiveNulleraremadeofchalcogenidematerialandhave been demonstrated to yield 25 dB ormore rejection of higher‐order spatialmodes,thusmeetingtheflightrequirements(Ksendzovetal.2007).Becausethismaterialbecomesopaque at wavelengths above 12 µm, other materials must be used in the 12–20 µmwavelengthband. Silverhalide single‐mode fibershavebeendemonstrated tomeet flightrequirementsat10µm,andshouldalsoworkwellat longerwavelengths(Ksendzovetal.2008).Althoughthisperformanceissufficientforflight,itwouldbegreatlyadvantageoustoimprovethethroughputofthesedevices,totestthemthroughoutthefullwavelengthrangetheyareintendedfor,andtotestthemcryogenically.SpatialfiltertechnologywouldthenbeatTRL5.Itwould,furthermore,beadvantageoustoimplementmid‐infraredspatialfiltersandbeamcombinersusing integratedoptics, so as to reduce the risk associatedwith thecomplexityofthescienceinstruments.

CryogenicAdaptiveNullingAdaptivenullingisstraightforwardtogeneralizeoverafullscienceband,anditshouldbedemonstratedwithin a cryogenic vacuum, bringing the technology to TRL 5. Thiswouldnecessitatethesuccessfulvalidationofcryogenicspatialfilters(above)andthetestingofacryogenicdeformablemirror.

InfraredImaging

123

PlanetDetectionSysteminaCryogenicVacuumThefinaldemonstrationforthefeasibilityofaprobe‐scalemissionwouldbetointegrateallthenecessarycomponentsinavacuumcryogenictestbed.Thiswoulddemonstratethefullsystem complexity and include flight‐like servo systems and brass‐boards of thecomponentsdescribedpreviously.

FlagshipMissionThe technology program for the Terrestrial Planet Finder Interferometer is now close toachievingallofitsmilestonesinstarlightsuppression.Thebasiccomponenttechnologyforstarlightsuppressionatmid‐infraredwavelengthsisnowatTRL4.TRL5requirestestingina relevant environment,which for TPF/Darwinwould be a cryogenic vacuumnear 40K.Mostresearchsofarhasbeenundertakeninairatroomtemperature.Aswillbediscussedbelow,sufficientprogresshasnowbeenmadethatthegreatestadvanceinthisareawouldbe to proceed to brass‐board designs of already successful components and subsystems,andtoimplementsystem‐levelcryogenictesting.

Ground‐baseddemonstrationswiththeFormationControlTestbed(FCT)haveshownthattheGuidance,Navigation,andControlalgorithmsnowexisttoexecuteprecisionmaneuverswith two telescopes (Scharf 2007; Scharf & Lawson 2008). Additional testing should becarried out to validate collision avoidance and fault tolerant algorithms, but the greatestadvancewould be to transition to space‐based demonstrations in collaborationwith ourEuropeancolleagues.

Futuretechnologydemonstrationsarediscussedasfollows.

PlanetDetectionwithChoppingandAveragingFurther noise suppression is required in addition to starlight suppression. Nulling canreduce the glare of starlight to a level fainter than the warm glow of local zodical dust(surroundingourSun)andexo‐zodiacaldust(aroundthetargetstar),buttheplanet itselfmay still be 100 times fainter. The technique for detecting and characterizing planetsthereforereliesonseveralotherstrategies.

The first step is to suppress the response to any thermal emission that is symmetricallydistributed around the star. In principle this will remove the detected glow of local andexozodiacaldust.Byrotatingthearrayandaveragingtheresponse,theplanetsignalcanbefurther enhanced. The beam combiner for TPF/Darwin therefore combines two pairs ofbeamstonullthestarlight,andthebeamcombiningsystemmodulatestheresponseonthesky(keepingthestarnulled)bychoppingbackandforthbetweenthesenulledpairs.ThismilestonehasbeendetailedintheWhitepaperforthePlanetDetectionTestbed(Martinetal.2008c).

BroadbandSystematicNoiseSuppressionAnadditionalstepisrequiredtosuppressthenoisedownbelowthetypicalbrightnessofanEarth‐like planet. After nulling and chopping, the dominant source of noise is due toresidualinstabilitiesinthenulldepth,whichgeneratesawhite‐noiseofsimilarintensitytothe planet signal. This noise can be suppressed by appropriate choice of interferometerbaselines and by filtering themeasured data from a complete rotation of the array. WeexpectthismilestonewillbecompletedbythePlanetDetectionTestbedin2009.

Additionaltechnologyprogressbeyondthismilestonewoulddependoncryogenictestingofcomponentsandsystems.

Chapter4

124

PlanetDetectionDemonstrationinaFlightlikeEnvironmentLaboratory work is well advanced to demonstrate the detection of planet light with thePlanetDetectionTestbed.Simulatedplanetstwomilliontimesfainterthanastarhavebeendetected in early trials, and the work is ongoing. Although these tests represent the fullcomplexityofbeamcombination,thisresearchisbeingconductedatroom‐temperatureinair,andthenoisepropertiesofthetestbedarethereforeunlikethosethatwouldbemetinflight. Following the completion of this work, the next step would be to build a vacuumcryogenic beam combiner with flight‐like servo systems, including the brass‐boardcomponentsdescribedabove.

Guidance,Navigation,andControl:CollisionAvoidance,FaultDetectionAdditional tests that could be carried out with the FCT include demonstrating newcapabilities suchas (1) reactivecollisionavoidance, (2) formation faultdetection,and (3)autonomousreconfigurationandretargetingmaneuvers.Also,usingareal‐timesimulationenvironment would allow the demonstration of performance with full formation‐flightcomplexity,with five interacting spacecraft showing synchronized rotations, autonomousreconfigurations,faultdetection,andcollisionavoidance.

InflightTestingofFormationAlgorithmsNationalagenciesinEuropeareactivelyadvancingthetechnologyofformationflying,withflight missions starting in 2009–2014. The European Space Agency and national spaceagencies inEuropehaveaprogramofprecursormissionstogainexperience in formationflying.In2009theSwedishSpaceAgencywilllaunchthePrismamission.Thisisprimarilyarendezvousanddockingmission,butitwillalsotestRFmetrologydesignedforDarwin.In2012 ESA plans to launch Proba‐3, which will include optical metrology loops for sub‐millimeter range controlovera30‐mspacecraft separation.TheFrenchand Italian spaceagenciesareplanningtolaunchSimbol‐Xin2015.Simbol‐XisanX‐raysciencemissionwithseparatedetectorand lens spacecraftwitha20‐mspacecraft separation.Simbol‐X shouldenterPhaseBofdevelopmentinearly2009.

Although two‐spacecraft precision maneuvers were demonstrated in 2007, a major goalthat remains is to demonstrate robust algorithms for three or more spacecraft, such aswouldbeusedbyTPF‐I(Lawsonetal.2008).Theopportunitynodoubtexiststoleveragetheexpertisedeveloped forTPF‐I incollaborationwithEuropeancolleagues.Thegreatestadvance in maturing technology for formation flying would be to have a modest‐scaletechnologymission devoted to verifying and validating guidance and control algorithms.Such in‐flight testing would be made even more meaningful if it could include theinterferometriccombinationofstarlightfromseparatedplatforms.

Aground‐based facilitysuchas theFCTshouldcontinuetoprovide themeans to testandimprove real‐time formation‐flying algorithms as the technologymatures, evenwhile thetechnologyisbeingproveninspace.

InfraredImaging

125

4.4 Research&AnalysisGoals4.4.1 GroundBasedInterferometryGround‐basedinterferometryservescriticalrolesinexoplanetstudies.Itprovidesavenuefor development and demonstration of precision techniques including high‐contrastimaging and nulling, it trains the next generation of instrumentalists, and it develops acommunity of scientists expert in their use. With the highest priority, nulling facilitiesshouldberefinedandoperatedforthecharacterizationofdebrisdisksandexozodiacallightandtheimpactoftheseondirectdetectionmissions.

Ground‐based interferometry can also provide unique support to exoplanet studies. RVdetections have been vetted for the possibility of stellar companions on highly inclinedorbits (Baines et al. 2008a). Transit planets can be studied more precisely with directmeasurementof thestellardiameters(Bainesetal.2008b).KeckandVLTIwillsoonofferprecisionnarrow‐angleastrometry,intermediatebetweenHSTandSIM,thatwillopennewareas of exoplanet parameter space, and will improve the characterization of knownexoplanet systems, contributing to construction of well‐understood target lists for directdetectionmissions.Ground‐basedinterferometersmaysoonbecapableofdirectdetectionandstudyofsomeoftheverybrightestexoplanets(Monnieretal.2006).

We endorse the recommendations of the “Future Directions for Ground‐based OpticalInterferometry”Workshop(Akesonetal.2007)andtheReSTARcommitteereport(Bailynet al. 2008) to continuing vigorous refinement and exploitation of existing ground‐basedinterferometricfacilities(Keck,NPOI,CHARA,andMRO),wideningoftheiraccessibilityforexoplanet programs, and continued development of interferometry technology andplanningforafutureadvancedfacility.

ThenatureofAntarcticplateausites,intermediatebetweengroundandspaceinpotential,offers significant opportunities for exoplanet and exozodi studies by interferometry andcoronagraphy(Kenyonetal.2006;CoudéduForestoetal.2006).Balloonaltitudesarealsopromising for high‐contrast imaging (Traub & Chen 2007). Further evaluation of theseopportunitiesshouldbeencouragedandsupported,withtheexpectationthattheymaybecompetitiveforfutureimplementation.

4.4.2 TheorySupportIn order to achieve our most ambitious goals of relating exoplanet atmospherescompositiontoplanet formation,systemevolution,surfacechemistryandbiology,wewillrequiresustainedsupportofstrongastrobiologyandatmosphericchemistryprograms.

4.4.3 SpaceBasedInterferometrySpace‐basedinterferometryservescriticalrolesinexoplanetstudies.Itprovidesaccesstoaspectralrangethatcannotbeachievedfromtheground,anditcancharacterizethedetectedplanets in terms of atmospheric composition and effective temperature. Sensitivetechnology has already been proven formissions like JWST, SIM, and Spitzer, andwithinNASA’s preliminary studies of TPF. With high priority, technology for nullinginterferometers in space should be advanced to make the next generation spacecraft todetectandcharacterizetheclosestplanetsflightready.

Chapter4

126

4.4.4 AgencyCoordinationandProgrammaticSynergiesTheverybroadscientificandpublicinterestinexoplanets,andthedirectrelevancetobothNASAandNSFgoals,makesitanidealtopicforcoordinationbetweentheagencies,andweurgeNASAandNSFstafftoleveragethisrelationshiptocoverthefullbreadthofexoplanetscienceandtechnology.

4.4.5 InternationalCooperation,Collaboration,&Partnership

Inpastyears(2002–2006)aNASA/ESALetterofAgreement(LOA)hasallowedtheUSandEuropean communities to develop a joint TPF and Darwin mission concept based on aconvergence of science requirements, design studies, and shared technology goals. TherelationshipsforgedbetweenUSandEuropeancollaboratorsshouldagainbefosteredoverthenextdecade to furtherstudiesofsmallmissionand flagshipmissionconcepts. Anewletterofagreementisnecessarytorenewthesecollaborations.

4.5 ContributorsOlivierAbsil,UniversitédeGrenoble,JosephFourier

RachelAkeson,Caltech,NASAExoplanetScienceInstitute

JohnBally,UniversityofColorado

RichardK.Barry,NASAGoddardSpaceFlightCenter

CharlesBeichman,NASAExoplanetScienceInstitute

AdrianBelu,UniversitédeNiceSophiaAntipolis

MathewBoyce,HeliosEnergyPartners

JamesBreckinridge,Caltech,JetPropulsionLaboratory


ChristineChen,SpaceTelescopeScienceInstitute

DavidCole,Caltech,JetPropulsionLaboratory

DavidCrisp,JetPropulsionLaboratory

WilliamC.Danchi,NASAGoddardSpaceFlightCenter


PeterDeroo,Caltech,JetPropulsionLaboratory

VincentCoudéduForesto,ObservatoiredeParis

DenisDefrère,UniversitédeLiège

DennisEbbets,BallAerospaceandTechnologyCorporation

PaulFalkowski,RutgersUniversity

RobertGappinger,JetPropulsionLaboratory

InfraredImaging

127

IsmailD.Haugabook,Sr.,DigitalTechnicalServices

CharlesHanot,Caltech,JetPropulsionLaboratory,UniversitédeLiège

PhilHinz,StewardObservatory,UniversityofArizona

JanHollis,NASAGoddardSpaceFlightCenter

SarahHunyadi,JetPropulsionLaboratory

DavidHyland,TexasA&MUniversity

KennethJ.Johnston,U.S.NavalObservatory

LisaKaltenegger,HarvardSmithsonianCenterforAstrophysics

JamesKasting,PennsylvaniaStateUniversity

MattKenworthy,StewardObsevatory,UniversityofArizona

AlexanderKsendzov,JetPropulsionLaboratory

BenjaminLane,DraperLabs

GregoryLaughlin,UCOLick

PeterLawson,Caltech,JetPropulsionLaboratory

OliverLay,Caltech,JetPropulsionLaboratory

RéneLiseau,StockholmUniversity

BrunoLopez,ObservatoiredelaCoted’Azur

RafaelMillanGabet,Caltech,NASAExoplanetScienceInstitute

StefanMartin,Caltech,JetPropulsionLaboratory

DimitriMawet,Caltech,JetPropulsionLaboratory

BertrandMennesson,Caltech,JetPropulsionLaboratory

JohnMonnier,UniversityofMichigan

NaoshiMurakami,NationalAstronomicalObservatoryofJapan(NAOJ)

M.CharlesNoecker,BallAerospaceandTechnologyCorporation

JunNishikawa,NationalAstronomicalObservatoryofJapan(NAOJ)

MeyerPesesen,Caltech,JetPropulsionLaboratory

RobertPeters,Caltech,JetPropulsionLaboratory

AliceQuillen,UniversityofRochester

SamRagland,W.M.KeckObservatory

StephenRinehart,NASAGoddardSpaceFlightCenter

HuubRottgering,UniversityofLeiden

DanielScharf,Caltech,JetPropulsionLaboratory

EugeneSerabyn,Caltech,JetPropulsionLaboratory

MotohideTamura,NationalAstronomicalObservatoryofJapan(NAOJ)

Chapter4

128

MohammedTehrani,TheAerospaceCorporation

WesleyA.Traub,Caltech,JetPropulsionLaboratory

StephenUnwin,Caltech,JetPropulsionLaboratory

DavidWilner,HarvardSmithsonianCenterforAstrophysics

JulienWoillez,W.M.KeckObservatory

NevilleWoolf,UniversityofArizona

MingZhao,UniversityofMichigan

4.6 ReferencesAggarwal,I.D.,&Sanghera,J.S.2002,“Developmentandapplicationsofchalcogenideglass

opticalfibersatNRL,”J.Optoelectron.Adv.Matter,4,665–678

Akeson,R.,Allen,R.,Angel,R.,etal.2007,WorkshopontheFutureDirectionsforGroundbasedOpticalInterferometry:FinalReport,NationalOpticalAstronomyObservatory,Tucson,AZ,http://www.noao.edu/meetings/interferometry/Workshop‐report.pdf

Angel, J. R. P.,&Woolf,N. J. 1997, “An imagingnulling interferometer to study extrasolarplanets,”ApJ,475,373–379

Baines, E. K., McAlister, H. A., ten Brummelaar T. A., et al. 2008a, “The search for stellarcompanionstoexoplanethoststarsusingtheCHARAArray,”ApJ,682,577–585

Baines, E. K., McAlister, H. A., ten Brummelaar T. A., et al. 2008b, “CHARA Arraymeasurementsoftheangulardiametersofexoplanethoststars,”ApJ,680,728–733

Bailyn,C.,Barnes,T.,etal.2008,RenewingSmallTelescopesforAstronomicalResearch,ReSTARReport,NationalOpticalAstronomyObservatory,Tucson,AZ,http://www.noao.edu/system/restar/files/ReSTAR_final_14jan08.pdf

Barry,R.K.,Danchi,W.C.,etal.2005,“TheFourier‐KelvinStellarInterferometer(FKSI):aprogress report and preliminary results from our nulling testbed,” Proc. SPIE, 5905,311–321

Barry, R. K., Danchi,W. C., et al. 2006, “The Fourier‐Kelvin Stellar Interferometer: a low‐complexitylow‐costspacemissionforhigh‐resolutionastronomyanddirectexoplanetdetection,”Proc.SPIE,6265,62651L

Barry,R.K.,Danchi,W.C.,Koresko,M.C.,etal.2008,“High‐ResolutionN‐BandObservationsoftheNovaRSOphiuchiwiththeKeckInterferometerNuller,”ApJ,677,1253–1267

Beichman,C.A.,Woolf,N. J.,&Lindensmith,C.A., eds.1999,TheTerrestrialPlanetFinder(TPF):aNASAOriginsProgramtoSearchforHabitablePlanets,JPLPublication99‐3,JetPropulsionLaboratory,Pasadena,California

Beichman, C.A., Fridlund,M., Traub,W.A., Stapelfledt,K.R.,Quirrenbach,A.,& Seager, S.2006, “Comparative planetology and the search for life beyond the Solar System,”ProtostarsandPlanetsV,editorsRiepurth,B.,etal.,UniversityofArizonaPress,Tucson,AZ,915–928

InfraredImaging

129

Booth,A.J.,Martin,S.R.,&Loya,F.2008,“ExoplanetExoplorationProgramPlanetDetectionTestbed:latestresultsofplanetlightdetectioninthepresenceofstarlight,”Proc.SPIE,7013,701320

Beaulieu, J.‐P., Bennett, D. P., et al. 2006, “Discovery of a cool planet of 5.5 Earthmassesthroughgravitationalmicrolensing,”Nature,439,437–440

Bracewell, R. N. 1978, “Detecting nonsolar planets by spinning infrared interferometer,”Nature,274,780–781

Christensen, P.R.&Pearl, J. C. 1997, “Initial data from theMarsGlobal Surveyor thermalemission spectrometer experiment:Observations of the Earth,” J. Geophys. Res., 102,10875–10880

Cockell,C.S.,Herbst,T.,Léger,A.,etal.2009,“Darwin—AMissiontoDetect,andSearchforLifeon,ExtrasolarPlanets,”Astrobiology,inpress,preprint,arXiv:0805.1873

Coudé du Foresto, V., Absil, O., Swain, M., Vakili, F., & Barillot, M. 2006, “ALADDIN: anoptimizednullingground‐baseddemonstratorforDARWIN,”Proc.SPIE,6268,626810

Crawford,S.L.,Colavita,M.M.,Garcia,J.I.,etal.2005,“FinallaboratoryintegrationandtestoftheKeckInterferometernuller,”Proc.SPIE,5905,59050U

Danchi,W.C.,Deming,D.,Kuchner,M.,&Seager,S.2003a,“DetectionofClose‐InExtrasolarGiantPlanetswiththeFourier‐KelvinStellarInterferometer,”ApJL,597,L57

Danchi, W. C., Allen, R., Benford, R. J., et al. 2003b, “The Fourier‐Kelvin StellarInterferometer,” inTowardsOtherEarths:DARWIN/TPFand theSearch forExtrasolarTerrestrial Planets, Heidelberg, Germany, 22–25April 2003 (ESA Publication SP‐539,October2003)

Danchi, W. C., Allen, Barry, R., Benford, R. J., et al. 2004, “The Fourier‐Kelvin StellarInterferometer: A Practical InterferometricMission forDiscovering and InvestigatingExtrasolarGiantPlanets,”Proc.SPIE,5491,236

Danchi, W. C., & Lopez, B., 2007, “The Fourier‐Kelvin Stellar Interferometer (FKSI)— Apractical infrared space interferometer on the path to the discovery andcharacterizationofEarth‐likeplanetsaroundnearbystars,”Comptesrendus‐Physique(C.R.Physique),8,396–407

Defrère,D., Absil, O., Coudédu Foresto, V., Danchi,W. C.,& denHartog, R. 2008, “Nullinginterferometry: performance comparison between space and ground‐based sites forexozodiacaldiscdetection,”A&A,490,435–445

DesMarais,D.J.,Harwit,M.O.,Jucks,K.W.,Kasting,J.F.,Lin,D.N.C.,Lunine,J.I.,Schneider,J.,Seager,S.,Traub,W.A.,&Woolf,N.J.2002,“Remotesensingofplanetarypropertiesandbiosignaturesofextrasolarterrestrialplanets,”Astrobiology,2,153–181

Forget, F.,&Pierrehumbert,R. T. 1997, “Warming earlyMarswith carbondioxide cloudsthatscatterinfraredradiation,”Science,278,1273–1276

Frey,B.J.,Barry,R.K.,Danchi,W.C.,etal.2006,“TheFourier‐KelvinStellarInterferometer(FKSI) nulling testbed II: Closed‐loop pathlength metrology and control subsystem,”Proc.SPIE,6265,62651N

Fridlund, M. 2000, DARWIN: The InfraRed Space Interferometer, ESA‐SCI (2000)12, 47EuropeanSpaceAgency:Noordwijk,TheNetherlands

Chapter4

130

Fridlund,C.V.M.,Gondoin,P.,Bavdaz,M.etal.2008,“Feather‐lighttouchallthat'sneededforDarwin'sfrictionlessoptics,”intheESAwebsite;http://www.esa.int/esaCP/SEMD3J8J50F_index_0.html

Gappinger,R.O.,Diaz,R.T.,Ksendzov,A., Lawson,P.R., Lay,O.P., Liewer,K., Loya, F.M.,Martin,S.R.,Serabyn,E.&Wallace,J.K.2009,“Experimentalevaluationofachromaticphaseshiftersformid‐infraredstarlightsuppression,”Appl.Opt.,48,868‐880

Gould,A.,Udalski,A.,An,D.,etal.2006, “MicrolensOGLE‐2005‐BLG‐169 implies thatcoolNeptune‐likeplanetsarecommon,”ApJ,644,L37–L40

Harrington,J.A.2001,InfraredFiberOptics,M.Bass,J.Enoch,E.VanStryland,&W.Wolfe,eds.,(McGraw‐Hill,NewYork).

Hinz, P.M., Angel, J. R. P., et al. 2000, “BLINC: a testbed for nulling interferometry in thethermalinfrared,”Proc.SPIE,4006,349–353

Hinz,P.M.,Heinze,A.N.,etal.2006,“ThermalinfraredconstrainttoaplanetarycompanionofVegawiththeMMTadaptiveopticssystem,”ApJ,653,1486–1492

Hyde,T.T.,Liu,K‐C.,Blaurock,C.,Bolognese,J.,etal.2004,“RequirementsFormulationandDynamicJitterAnalysisfortheFourier‐KelvinStellarInterferometer,”Proc.SPIE,5491,553

Kaltenegger, L., Eiroa,C., Stankov,A., Fridlund,M.2009, “TheDarwin target star catalog,”Astrophys.&SpaceSci.,inpress

Kaltenegger,L.,&Fridlund,M.2005,“TheDarwinmission:Searchforextra‐solarplanets,”AdvancesinSpaceResearch,36,1114–1122

Kaltenegger,L.,Traub,W.A.,&Jucks,K.2007,“SpectralevolutionofanEarth‐likeplanet,”ApJ,658,598–616

Karlson,A.L.,Wallner,O.,Armengol, J.M.P.,etal.2004, “Threetelescopenullerbasedonmultibeaminjectionintosingle‐modewaveguide,”Proc.SPIE,5491,831–841

Kasting, J. F., Whitmire, D. P., & Reynolds, R. T. 1993, “Habitable zones around mainsequencestars,”Icarus,101,108–128

Kasting,J.F.,&Catling,D.2003,“Evolutionofahabitableplanet,”ARA&A,41,429–463

Kenyon,S.L.,Lawrence, J.S.,etal.2006, “AtmosphericscintillationatDomeC,Antarctica:Implicationsforphotometryandastrometry,”PASP,118,924–932

Ksendzov, A., Lay, O., Martin, S., Sanghera, J. S., Busse, L. E., Kim, W. H., Pureza, P. C., &Nguyen, V. Q. 2007, “Characterization of mid‐infrared single mode fibers as modalfilters,”Appl.Opt.,46,7957–7962

KsendzovA.,LewiT.,Lay,O.P.,Martin,S.R.,Gappinger,R.O.,Lawson,P.R.,Peters,R.D.,Shalem, S., Tsun, A., & Katzir, A. 2008, “Modal filtering for midinfrared nullinginterferometryusingsinglemodesilverhalidefibers,”Appl.Opt.,47,5728–5735

Kurland, R. 2007, “JWST Sunshield Silicon‐Coated Membrane Material: Summary‐LevelReport,” JWST Technical Non‐Advocate Review (T‐NAR), January, 17 2007, GoddardSpaceFlightCenter,Greenbelt,MD

Lafrenière,D.,Jayawardhana,R.,vanKerkwijk,M.H.,“DirectimagingandspectroscopyofaplanetarymasscandidatecompaniontoayoungSolaranalog,”ApJL,689,L153–L156,2008

InfraredImaging

131

Lawson,P.R.,Lay,O.P.,Martin,S.R.,Peters,R.D.,Gappinger,R.O.,Ksendzov,A.,Scharf,D.P., Booth, A. J., Beichman, C. A., Serabyn, E., Johnston, K. J., & Danchi, W. C., 2008,“TerrestrialPlanetFinderInterferometer:2007–2008progressandplans,”Proc.SPIE,7013,7013N

Lawson,P.R.,Lay,O.P.,Martin,S.R.,etal.2007,“TerrestrialPlanetFinderInterferometer:2006–2007progressandplans,”Proc.SPIE,6693,669308

Lay,O.P.2004,“Systematicerrorsinnullinginterferometers,”Appl.Opt.,43,6100–6123

Lay, O. P. 2005, “Imaging properties of rotating nulling interferometers,” Appl. Opt., 44,5859–5871

Lay O. P. 2006, “Removing instability noise in nulling interferometers,” Proc. SPIE, 6268,62681A

Lay, O. P., Martin, S. R., & Hunyadi S. L. 2007, “Planet‐finding performance of the TPF‐IEmmaarchitecture,”Proc.SPIE,6693,66930A

Léger, A.,Mariotti, J.‐M.,Mennesson, B., Ollivier,M., Puget, J. L., Rouan, D., & Schneider, J.1996,“Couldwesearchforprimitivelifeonextrasolarplanetsinthenearfuture?:TheDARWINProject,”Icarus,123,249–255

Léger, A., Herbst, T., et al. 2007, “DARWIN mission proposal to ESA,” 23 Jul 2007.arXiv:0707.3385v1[astro‐ph].

Lopez,B.,S.Wolf,S.Lagarde,P.Abraham,etal.2006,“MATISSE:perspectiveofimaginginthemid‐infraredattheVLTI,”Proc.SPIE,6268,31

Lovelock, J. E., 1975, “Thermodynamics and the recognition of alien biospheres [anddiscussion],”Proc.R.Soc.Lond.B189,167–181

Lunine, J. I. 2001, “The occurrence of Jovian planets and the habitability of planetarysystems,”PNAS,98,809–814

Mankins,J.C.,1995,“TechnologyReadinessLevels,”NASAAdvancedConceptsOffice,OfficeofSpaceAccessandTechnology,NationalAeronauticsandSpaceAdministration,Washington,DC,http://www.hq.nasa.gov/office/codeq/trl/trl.pdf.

Martin, S. R., Serabyn, E., Hardy, G. J. 2003, “Deep nulling of laser light in a rotationalshearinginterferometer,”Proc.SPIE,4838,656–667

Martin, S. R. 2005, “The flight instrument design for the Terrestrial Planet FinderInterferometer,”Proc.SPIE,5905,21–35

Martin,S.R.2006,“Progressinfour‐beamnulling:resultsfromtheTerrestrialPlanetFinderPlanetDetectionTestbed,”in2006IEEEAerospaceConference,BigSky,Montana

Martin, S. R., Szwaykowski, P., Loya, F., Liewer, F. 2006, “Progress in testing exo‐planetsignalextractionontheTPF‐IPlanetdetectionTestbed,”Proc.SPIE,6268,626818

Martin,S.R.,Scharf,D.,Wirz,R.etal.2007,“TPF‐Emma:conceptstudyofaplanet findingspaceinterferometer,”Proc.SPIE,6693,669309

Martin, S.R., Scharf,D. P.,Wirz,R., et al. 2008a, “DesignStudy for aPlanet‐Finding SpaceInterferometer,”IEEEAerospaceConf,BigSky,Mt,USA

Martin,S.R.,Serabyn,E.,Liewer,K.2008b,“Thedevelopmentandapplicationsofaground‐basedfibernullingcoronagraph,”Proc.SPIE,7013,70131Y

Chapter4

132

Martin,S.R.,Booth,A.J.,Lay,O.P.,&Lawson,P.R.2008c,ExoplanetInterferometryTechnologyMilestone#4Whitepaper:PlanetDetectionDemonstration,JetPropulsionLaboratoryDocumentD‐47627,http://planetquest.jpl.nasa.gov/TPF‐I/TPF‐I_M4_Whitepaper_Final.pdf

Mather, J. C., Cheng, E. S., Cottingham, D. A., et al. 1994, “Measurement of the CosmicMicrowaveBackgroundSpectrumbytheCOBEFIRASInstrument,”ApJ,420,439–444

Mennesson,B.,Haguenauer,P., Serabyn,E.,&Liewer,K.2006, “Deepbroad‐band infrarednulling using a single‐mode fiber beam combiner and baseline rotation,” Proc. SPIE,6268,626830

Monnier,J.D.,Pedretti,E.,etal.2006,“MichiganInfraredCombiner(MIRC):commissioningresultsattheCHARAArray,”Proc.SPIE,6268,62681P

NationalResearchCouncil,2001,AstronomyandAstrophysicsintheNewMillennium,http://www.nap.edu/openbook.php?isbn=0309070317

Ollivier,M. 1999,Contribution à la Recherche d'Exoplanète: Coronagraphie InterférentiellepourlaMissionDarwin,Ph.D.thesis,UniversityParisXI

Owen,T.1980,“Thesearchforearlyformsoflifeinotherplanetarysystems,”inStrategiesforSearchforLifeintheUniverse,editorsPapagiannis,M.D.,Springer:Dordrecht,TheNetherlands,p.177

Peters,R.D.,Lay,O.P.,Jeganathan,M.2008,“Broadbandphaseandintensitycompensationwithadeformablemirrorforaninterferometricnuller,”Appl.Opt.,47,3920–3926

Ressler,M.2007,“Mid‐InfraredDetectors,”JWSTTechnicalNon‐AdvocateReview(T‐NAR),30–31January2007,GoddardSpaceFlightCenter,Greenbelt,MD

Samuele, R., Wallace, J. K., Schmidtlin, E., Shao, M., Levine, B. M., Fregoso, S. 2007,"Experimental progress and results of a visible nulling coronagraph," 2007 IEEEAerospaceConference,BigSkyMontana,paper1333

Scharf,D.P.2007,TPFIMilestone#2Whitepaper:FormationControlPerformanceDemonstration,JetPropulsionLaboratory,Pasadena,CA,http://planetquest.jpl.nasa.gov/TPF‐I/TPFI_M2_WhitePaper_Final.pdf

Scharf,D.P.,Lawson,P.R.2008,TPFIMilestone#2Report:FormationControlPerformanceDemonstration,JetPropulsionLaboratory,Pasadena,CA,JPLPublication08‐11,http://planetquest.jpl.nasa.gov/TPF‐I/TPFI_M2_ReportV3.pdf

Scharf, D. P., Hadaegh, F. Y., Keim, J. A., & Lawson, P. R. 2008, “Ground demonstration ofsynchronized formation rotations for precision, multi‐spacecraft interferometers,” inProc. 3rd International Symposium on Formation Flying, Missions and Technologies,Fletcher,K.,ed.,ESASP‐654(CDROM)

Seager,S.,Richardson,L.J.,Hansen,B.M.S.,etal.2005,“OntheDaysideThermalEmissionofHotJupiters,”ApJ,632,1122–1131

Segura, A., Krelove, K., Kasting, J. F., Sommerlatt, D., Meadows V., Crisp D., Cohen, M., &Mlawer, E. 2003, “Ozone concentrations and ultraviolet fluxes on Earth‐like planetsaroundotherstars,”Astrobiology,3,689–708

Segura,A.,Kasting,J.F.,Meadows,V.,Cohen,M.,Scalo,J.,Crisp,D.,Butler,R.A.H.,&Tinetti,G. 2005, “Biosignatures from Earth‐like planets around M Dwarfs,” Astrobiology, 5,706–725

InfraredImaging

133

Selsis,F. 2000, “Physics of Planets I:Darwin and theAtmospheres ofTerrestrial Planets,”DarwinandAstronomy:TheInfraredSpaceInterferometer,ESASP451,EuropeanSpaceAgency,Noordwijk,TheNetherlands,pp.133–140

Selsis, F., Despois, D., & Parisot, J.‐P. 2002, “Signature of life on exoplanets: Can Darwinproducefalsepositivedetections?”A&A,388,985–1003

Selsis, F., Kasting, J. F., Levrard, B., et al. 2007, “Habitable planets around the star Gliese581?,”A&A,476,1373–1387

Serabyn,E.,Wallace,J.K.,Hardy,G.J.,Schmidtlin,E.G.H.,Nguyen,H.T.1999,“Deepnullingofvisiblelaserlight,”Appl.Opt.,38,7128–7132

Serabyn,E.,Booth,A.J.,Colavita,M.M,etal.2004,“TheKeck interferometernuller:systemarchitectureandlaboratoryperformance,”Proc.SPIE,5491,806–815

Traub,W.A.,&Chen,P.,“Planetscopeprecursorexperiment,”BAAS211,30.06(2007).

Turnbull,M.C.,Traub,W.A.,Jucks,K.W.,Woolf,N.J.,Meyer,M.R.,Gorlova,N.,Skrutskie,M.F., & Wilson, J. C. 2006, “Spectrum of a habitable world: Earthshine in the near‐infrared,”ApJ,644,551–559

Turnbull,M.C.,TheSearchforHabitableWorlds:FromtheTerrestrialPlanetFindertoSETI,Ph.D.Thesis,UniversityofArizona,ISBN0496037129

Wallace, K., Hardy, G., Serabyn, E. 2000, “Deep and stable interferometric nulling ofbroadbandlightwith implications forobservingplanetsaroundnearbystars,”Nature406,700–702

Wallner, O., Armengol, J. M. P., & Karlson, A. L. 2004, “Multiaxial single‐mode beamcombiner,”Proc.SPIE,5491,798–805

Wright,E.L.,Mather,J.C.,Fixsen,D.J.,etal.1994,“InterpretationoftheCOBEFIRASCMBRSpectrum,”ApJ,420,450–456

ExozodiacalDisks

135

5 ExozodiacalDisks

PhilHinz,UniversityofArizona,Chair

RafaelMillanGabet,NASAExoplanetScienceInstitute,CoChairOlivierAbsil,RachelAkeson,DavidArdila,JamesBreckenridge,GeoffreyBryden,ChristineChen,DavidCiardi,MarkClampin,VincentCoudéduForesto,WilliamDanchi,DenisDefrère,DennisEbbets,SallyHeap,JohnKrist,MarcKuchner,CharlesLillie,PatrickLowrance,StanimirMetchev,RafaelMillanGabet,MarshallPerrin,MeyerPesenson,PeterPlavchan,SamRagland,StephenRinehart,AkiRoberge,ChrisStark,KarlStapelfeldt,MotohideTamura,AngelleTanner,NealTurner,BruceWoodgate

5.1 IntroductionCharacterization of the debris disks around nearby stars is an important complement toplanetfindingforseveralreasons.Primarily,thedetectionofdebrismaterial,especiallyinorclosetothehabitablezoneisadirectindicationofplanet‐buildingmaterialinthiszone.Inagas‐poordisk,debrismaterialisclearedawaythroughcollisionaldestruction,radiationpressure,orPoynting‐Robertsondragonrelativelyshort timescalescompared to thoseofplanet formationor theplanetarysystem itself. Therefore, thepresenceofdust indicatesthatlargerparentbodiesmustalsobelocatedinthesystem.Thus,thedetectionofdust,intandemwith detection of themassive planets in the system allows the development of amore complete observational picture of the architecture of a planetary system. In fact,informationgleanedfromdiskstructurecanbeusedtodetectplanetsthatwouldotherwisebetoofaintorhaveamassthatistoolowtobedetectedusingothertechniques.Similarly,thepresenceofsmall‐bodypopulationsinahabitablezonecouldaffecthabitabilitythroughthe frequency of large impact events. However, for the ultimate goal of detecting andcharacterizing an Earth‐like exoplanet, debris disks can decrease the significance of thedetectionandpotentiallyobscuretheplanetsignalaltogether.

The recently completed report by the Exoplanet Task Forcemade two recommendationsrelated to dust disk detection. In the near term (1–5 years) they recommended thecommunity“investinacensusofexozodisystemsaroundexoplanettargetstars,”andinthemedium term (6–10 years) that the community should “implement next generation highspatialresolution imagingtechniquesonground‐basedtelescopes:AOfordirectdetectionofyoung,low‐masscompanions,andinterferometryfordiskscience.”

This chapter describes the state of the field and near future plans for exozodiacal dustdetectionandmodeling.Theplansdescribedinthechapteragreesubstantiallywiththoseof the ExoPlanet Task Force. Based on discussion in the May 2008 Exoplanet Forum,followingistheconsensussetofrecommendationsonexozodiacaldust:

Chapter5

136

• NASAshouldcontinueitssupportofKI,LBTI,andsimilareffortsthatwillmeasurezodiacaldustdensityandmorphologycriticalforplanningfuturedirectimagingmissions

• ContinuestudiesofExoplanetProbeclassspacemissionsthatcanaddressthedensity andmorphology of dust in the habitable zone of nearby starswith amuchhighersensitivitylimitandamuchlargersampleofnearbysystemsthancanbedonewithgroundbasedtechnologies.

• Continue the development of collisional debris disk models to complementobservations,aidingininterpretationandextrapolationofsuchresults.

5.1.1 PropertiesofProtoplanetaryandExozodiacalDisksDuring the collapse of dense interstellar cloud cores, conservation of angularmomentumdictatestheformationofarotatingdiskofgasanddustaroundtheyoungstar.Planetsgrowinthiscircumstellardisk,interactwithit,andimplanttheirsignaturesonit.Thediscoveryofsuchdisksaroundnearbystarshasprovidedtheopportunitytostudyplanetformationinreal time, rather than working backward from completed or mature systems. It is nowunderstood that planet formation is a general physical process that operated in our ownyoungSolarSystem, includingthetransformationof interstellarorganicmaterial intopre‐biotic compounds like those that gave rise to life on Earth. The older, dust‐rich disks—debris disks—are also vital for efforts to image and characterize extrasolar terrestrialplanets. Debris dust surrounds our Sun today (the zodiacal dust) and about 10–20% ofnearby stars harbor a currently detectable debris disk, typically detected at a separationsimilartotheSun'sKuiperbelt(e.g.,Beichmanetal.2006;Trillingetal.2008;Hillenbrandet al. 2008) but the extension of these statistics to the habitable zone is not yet known.While exozodiacal dust is a major source of background flux and probably confusion indirectimagingofexoplanets,itmayalsoprovideindirectevidenceofplanetsthroughduststructurescausedbygravitationalperturbations.ThesignatureoftheEarthmaybeseeninclumpswithinthezodiacaldustintheinnerSolarSystem.

The current paradigm for the formation of a mature planetary system has three mainphases:

• Formationofgas‐giantplanetslikeJupiter,• Formationofterrestrialplanets,and• Clearingofmostleftoverplanetesimals(i.e.asteroids&comets)fromthesystem.

ThistimelineisrepresentedinFigure5‐1.Itisduringthelastphasethatmostofthevolatilecontent of Earth’s surface was delivered by the impact of water‐rich planetesimals (e.g.,Morbidellietal.2000).Protoplanetarydisksmaybedividedintothreeclassesthatappeartoroughlycorrespondtoeachphase.

1. Theyoungestclassistheprimordialdisks,whicharecomposedofrelativelyunprocessedinterstellarmaterialleftoverfromstarformation.Primordialdisksaregas‐rich,asevidencedbyobservationsofsub‐mmCOemission(e.g.,Thietal.2000);thebulkofthegas,however,ismolecularhydrogen.Sincegiantplanetsareprimarilycomposedofgas,theymustformingas‐richdisks,andtheprimordialdisksmatchthisdescription.Contrarytoearlierconception,thesedisksareapparentlynotquiescentorstatic.Theyarethesitesofactivechemistryandlarge‐scalemotionsofgas,dust,andplanetarybodiesofallsizesthroughmigration.

ExozodiacalDisks

137

2. Thenextclassisthetransitionaldisks,whichappeartobeclearingmaterialfromtheirinnerdisks,butstillretainopticallythick,gas‐richouterdisks.Thisrelativelynewclasshasbeenthefocusofrecentintensivestudy,astheyprovideawindowontherapidtransformationfromaprimordialdiskintoayoungplanetarysystem.

3. Theoldestclassisthedebrisdisks,whicharedusty,gas‐poordisks.Thesedisksarecomposedofmaterialproducedbythedestructionofplanetesimalsandcontainlittleornounprocessedinterstellarmaterial.Debrisdiskshaveawiderangeofages,fromabout10MyrtoafewGyr.Terrestrialplanetsarelikelyformingintheyoungerdebrisdisks,whiletheolderonescorrespondtothedisk‐clearingphase.DustisremovedfromdebrisdisksthroughPoynting‐Robinsondrag,planetarysweeping,orradiationpressureandreplenishedviacollisionsbetweenandevaporationofplanetesimals.

AcoronagraphicimageofadiskfromeachofthethreeclassesappearsinFigure5‐1,whichimplies a very clean, orderly progression in both planet formation and disk evolution.However, the theoretical stages in planet formation may be occurring simultaneously indifferentpartsofadisk.Inaddition,thedifferentclassesofdiskarenotcleanlyseparated,withapparentoverlap in theagerangesofeachclass, indicatingthatstochasticprocessesmaysignificantlyshapetheevolutionarypathwaysof individualsystems.Therelationshipof this diversity of disk properties with the wide range of extrasolar planetary systemsremainsatopicofcurrentstudy.

Figure51.Timelineforplanetformationandtheevolutionofcircumstellardisks.Abovethearrow,themaintheoreticalphasesinplanetaryformationareshown.Below,theclassesofcircumstellardisksareidentified.Acoronagraphic image of an example from each class is shown; the primordial disk aroundABAurigae at left(Gradyetal.1999),thetransitionaldiskaroundHD141569Ainthecenter(Clampinetal.2003),andthedebrisdiskaroundAUMicroscopiiatright(Kristetal.2005).[Figurecredit:A.Roberge]

Chapter5

138

5.1.2 TerrestrialPlanetFormationandDeliveryofVolatiles

Inorder for terrestrial‐sizedplanets toacquire largequantitiesofvolatilessuchaswater,giantplanetsarelikelyrequiredtogravitationallystirupplanetesimalsandinjectthemintothe inner disk. The gas composition in a debris disk may provide information on thecomposition of the planetesimals producing the disk. For example, recent far‐ultravioletspectroscopy of the gas in the well‐studied β Pic debris disk has shown that carbon isextremelyoverabundantrelativetoeveryothermeasuredelement(e.g.,C/O=18×solar;Robergeetal.2006),despitethefactthatthecentralstarappearstohavesolarmetallicity(Holweger et al. 1997). This overabundance likely reflects the composition of the parentmaterialofthegas,whichwouldhavetobemuchmorecarbon‐richthanexpectedbasedonwhat we know about Solar System asteroids and comets. Either the β Pic planetesimalsproducingthegasareselectivelylosingtheirvolatilecarboncompounds,ortheyaresimplymore carbon‐rich overall than any Solar System planetesimal. The latter possibility hasimportantconsequencesforanyterrestrialplanetsthatmightbeformingaroundβPicandtheirvolatilecomposition.

5.1.3 ScienceGoalsThe extrasolar planetary systems already discovered show a surprising diversity in theirarchitectures compared to our own Solar System. Studies of protoplanetary and debrisdisks are essential to understand the origins of this diversity and to predict what otherphenomena might be observed in the future. They also affect our ability to make newdiscoveries,throughtheimpactofexozodiacaldustondirectimagingandcharacterizationofexoplanets.Theprimarygoalsofprotoplanetaryanddebrisdiskstudiesareasfollows.

ObtainabetterunderstandingofplanetformationasagenericandrobustprocessPlanets can form around normal stars with spectral types ranging from at least late M(Butleretal.2004)toearlyA(Johnsonetal.2007),aroundpulsars(e.g.,Wolszczan&Frail1992), and possibly even around brown dwarfs (Chauvin et al. 2004). Their birthenvironmentsareapparentlyalsodiverse,rangingfromlow‐massstar‐formingregionsliketheTauruscloud toenergetic,high‐massstar‐formingregions like theOrionnebula.Mostplanet formationmodelswere developed to explain the Solar System itself and currentlycannotexplainthediversityofplanetsalreadydetected.Theoreticalimprovementstothesemodels are hampered by a lack of basic information on protoplanetary disks, includinginitialmassesofthedisks,typicallifetimesofdisksaroundstarsofdifferentmassesandindifferent stellar environments, the evolution of the gas‐to‐dust ratio, density andtemperaturestructure,andchemicalevolution. It ishighlydesirablethatmodelsofplanetformationbeimprovedsothattheywouldbeabletopredicttheprevalenceofEarth‐massplanetsarounddifferenttypesofstarsincludingonesofvaryingmetallicityandmass,andthe predicted abundances of volatile chemicals on their surfaces, to motivate and guidefutureeffortstofindandcharacterizepotentiallyhabitableEarth‐likeexoplanets.

CharacterizetheexozodiacaldustaroundtargetsfordirectimagingofexoplanetsWhenviewedfromadistance,themostconspicuousfeatureoftheSun’splanetarysystemisthezodiacaldust,whichistenuousbutcoversalargesurfacearea.Asignificantfractionof

ExozodiacalDisks

139

nearbystarshassimilardustatmuchhigherabundance.Emissionfromexozodiacaldustislikelytobethelargestsourceofastrophysicalnoiseindirectimagingandcharacterizationof exoplanets. If the dust is too bright, the observation time needed to detect andcharacterize a planet becomes prohibitively long. For the TPF‐Cmission, the break pointwasestimatedtobe10timesthe levelofdust intheSolarSystem(10“zodis”). InSection5.1.4,wepresentasimpleprescriptionfortheacceptablelevelsofexozodiasafunctionoftelescope aperture size. In Sections5.2 and5.3,wediscuss the current exozodi detectionlimits andupcomingopportunities topushdown to the levelsofdust relevant fordirect‐imagingmissions.

UsediskstructuretoinferthepresenceofunseenplanetsWhileexozodiacaldusthampersdirect imagingofexoplanets, itmayalsoprovideanothermeans of indirect detection. In the Solar System, the gravitational influence of the Earthtemporarily traps zodiacal dust into a ring at 1 AU, with additional clumps leading andtrailing theplanet(Dermottetal.1994). Rings,clumps,andwarps(or“x‐patterns”)havebeen seen in coronagraphic images of nearby debris disks (e.g., HR4796, Schneider et al.1999;Fomalhaut,Kalasetal.2005).Suchstructuresareoftenattributedtotheinfluenceofanunseenplanetorbiting in thedebrisdisk, andmaybeauniquewayof findingplanetsthatareextremelydifficulttodetectwithotherindirecttechniques,likeyoungplanetsandplanetsonwideorbits.Unfortunately,wehavenotyetbeenabletoassociateanobserveddiskduststructurewithaknownexoplanet.Partoftheproblemisthedifficultyofdetectingplanets in the currently knowndebris disks,which have hundreds to thousands of timesmoredust than the Solar System.Theotherpart of theproblem relates to thedynamicalmodelingused to interpretdust structures. InSection5.2,wediscuss thecurrent stateofthesemodels,near‐termimprovements,andadditionalneededdevelopment.

5.1.4 ScienceRequirementsforObservationsofTerrestrialPlanets

TherequirementsonexozodiacalbrightnessforaTerrestrialPlanetFinderaredeterminedby its impacton the increase in integration time forsuchamission. Missions inboth thevisible(0.5–1µm)andinfrared(6–20µm)aresimilarlyaffected.Itisworthreviewingthefundamental noise parameters for these missions to appreciate the effect of a denserextrasolar zodiacaldust cloud. Bothcalculationsbelowassume thateffects fromresidualstarlight such as varying suppression or similar systematics, as well as limitations ofdetectorperformance,canbekeptbelowthefundamentalPoissonnoisefromlocalzodiacallight,exozodiacallightandresidualstarlight.Similarestimatescanbefound,forexampleinBrown(2005)orBeichmanetal.(2006).

ImpactonanOpticalMissionFForanidealizedopticalmission,thenoise inanopticalobservationisdominatedbythePoisson noise associated with background flux from an exozodiacal disk as well as thebackground flux from local zodiacal dust. The time required to image a planet at opticalwavelengthsinthepresenceofbackgroundfluxfromlocalzodiacaldust,exozodiacaldust,andunsuppressedstarlightisgivenby

Chapter5

140

where ψ is the size of the focal plane photometric aperture in units of λ/D, Δλ thebandwidth,F0 is thephoton flux forzeromagnitude,ms is thestellarmagnitude,mp is themagnitude of the planet, z is the surface brightness of the zodiacal light, λ is the centralwavelength of the bandpass, D is the diameter of the telescope aperture, T is the totalthroughput,SNRisthesignal‐to‐noise, μ istheexozodibrightnessinunitsofzodis,andζisthescatteredanddiffractedlight(the“contrast”)inthephotometricaperturerelativetotheunsuppressedstellar image.Typicalvalues fortheseparametersappear inTable5‐1.Thisequationisappropriateforabackground‐limitedobservationsuchasimaginganEarth‐likeexoplanet.

Table51.Typicalparametervaluesintheexposuretimecalculationfordirect‐imagingmissionsinthevisibleandtheinfrared.

Parameter VisibleValue IRValue Comment

Ψ 1.5 – Image“width”

Δλ 0.11µm 2µm 20%bandwidth

F09.5×1010photonssec–1m–2µm–1

5×107photonss–1m–2µm–1

ms 5 3.5 Sunat10pc

mp 29.9mag 19.6 Earthat1AUfromSunat10pc

Z 23magarcsec–2 14.2magarcsec–2 LocalZodiacalbrightness

Ez – 14.6mag Exozodiacalintegratedflux

λ 5.50×10–7m 1.0×10–5m

D 4m 2m ElementsizeforIRinterferometer

T 0.0864,(0.7087) 0.05 Optical:TPF‐C(TPF‐O)

SNR 10 10

B – 20m

ζ 5×10–11 1×10–5 Contrast

S – 4.6×10–9radians Sun'sdiameterat10pc

ImpactonanIRinterferometermissionForan idealized IR interferometermission (onenot limitedbysystematicerrorsornoiseassociatedwith choppingornulling), the calculation is similar to theopticalmission.Thelimitingnoisesourcesare the localzodiacaldustemission, theexozodiacaldustemission,andresiduallightfromthefinitesizeofthestarandafloortotheperformanceofthestellarsuppression.Thecontributionfromtheexozodiacalemissionisslightlydifferentsince,fortheIRcase, the integrated light fromtheexozodiacalcloudisseenbyeachelementof theinterferometer. This effect results in the interferometer’s sensitivity to exozodiacal dustbeingdependentontheelementaperturesize(D)andthedistancetothestar,whichaffectsthe exozodiacial dust brightness relative to the local zodiacal dust contribution. Theresultingtimefordetectionis

ExozodiacalDisks

141

The first term in the bracket is from local zodiacal dust, the second term from exozodyemission(whereez istheexozodiacalfluxinmagnitudesatNband),andthethirdtermisthenullleakagefromtheinstrument(ζ)andthestar'sangulardiameter,wheresisthesizeofthestellardiscinradians.Thiscalculationassumestheinterferometerisafourelementdesignwithanullingbaselineofbmeters.InstrumentalparametersinTable5‐1arefromLayetal.(2007).Thelight iscombinedusingtwosuccessivestagesofbeamcombiners,sothatthezodiacalandplanetlightisassumedtobesplitevenlyamongfourinterferometricoutputs.

Figure52.TwoestimatesoftheeffectofexozodiacaldustonthedetectionofEarth‐likeexoplanets.Theupperplot shows the effect of exozodiacal dust strength on observing time, based on the equations in the text andusingvaluesfromTable5‐1.ThelowerplotshowstheestimateofSNRdegradationcalculatedbyBeichmanetal.(2006).

Chapter5

142

Thetimeestimatesforbothcasesarenecessarilysimplistic,notincludingtermsorfactorsthatmayrefinetheactualtimeneededforasingleobservation.Inthissensetheequationsmaybeonlyapproximationsofthetruetimeneeded.SimilarlythevaluesinTable5‐1aredebatable,andcansignificantlychangetheresultingplotinFigure5‐2.Thesalientpointofthecalculationsisthatexozodiacalemissionhastheabilitytosignificantlyincreasethetime required for a given observation. This has the cascading effect of reducing thesamplesizeforamissionofafixedduration.Thiseffectcanbecountered,tosomeextentbyusing larger apertures or planning for a longerdurationmission. However, it is clearthat knowledge of average exozodiacal dust strength is a key input for a robust direct‐imaging design referencemission, and indeed knowledge of the exozodiacal levels for allcandidate TPF or Darwin stars will allow for a greatly optimized mission andcharacterizationstrategy(seeWallneretal.2008).

5.2 CurrentStateoftheFieldThe next few years will yield significant progress on the abundance, composition, andmorphologyofexozodiacalmaterialaroundnearbystars.Usingphotometry,spectroscopy,direct imaging, and interferometric observations, the formation and evolution of debrisdisks,theirrelationshiptoplanetarysystems,andtheirdetectabilitywillbeexplored.

5.2.1 PhotometryDisks are typically first detected by observation of infrared‐ to millimeter‐wavelengthemissioninexcessofthatexpectedfromthecentralstar.Thisemissionisproducedwhencircumstellar dust absorbs emission from the star and re‐emits it at longerwavelengths.Models of the dust emission may be fit to the spectral energy distribution (SED) of thesystem,providinganestimateofthefractionalinfraredluminosity(LIR/L∗)andtheeffectivedusttemperature,thecombinationofwhichcanyieldthedustmass.

The most important current mission for the detection and characterization of infraredexcessesaroundsolar‐likestarsistheSpitzerSpaceTelescope.Spitzeriscapableofmakingsensitivemeasurementsofexozodiacalcloudsatwavelengthsof24μm,70μm,and160μm,spanningtypicaltemperaturerangesof40–200Kandcorrespondingtodistancescalesofafew to a fewhundredAU from theparent star. The sensitivity to exozodiacal emission islimitedbySpitzer’sphotometricaccuracyofafewpercentat24μmand7–15%at70and160μm,allowingSpitzer todetectorset limitsonexcesses thataregreater than10–50%abovethestellarphotosphere.

At24μm,thislimitcorrespondstoafactorof100to500timesthelevelofdustemissioninourownSolarSysteminthedistancerangeof1to10AU,where100–200Kdustislocatedforasolar‐typestar(seeFigure5‐3).Atlongerwavelengths,thephotosphereisweaker,andthecorrespondinglimitis30–100timesthelevelofdustinourSolarSystembetween10–100AU(~50Kdust).Spitzerhasalreadyobservedorhasawardedtimeforobservationsofover200likelyTerrestrialPlanetFinder(TPF)targetstars(Beichmanetal.2006).

Herschelwillbeable tosearch fordustatevencooler temperaturesthanSpitzerwiththePhotodetectorArrayCameraandSpectrometer(PACS)instrument.Thelargeraperturewillallowdetectionofdustat loweropticaldepths fortheouterregionsofplanetarysystems.The expected sensitivity level is equivalent to an optical depth of approximately 10–6 forverycolddust(seeFigure5‐3).

ExozodiacalDisks

143

5.2.2 DiskImagingSpectral EnergyDistribution (SED) fitting canprovide some informationon the existenceandspatialdistributionof thedust (forexample, thepresenceofcentralholes in thedustdisk),but,duetocontributionfromthestar,photometricdetection,startingatwavelengthsshorterthanapproximately30µm,isincreasinglydifficultforfaintlevelsofdustemission.Imaging of the disk provides two advantages over photometry: separation of unresolved(starlight)andresolved(disk)emission,anddeterminationofthemorphologyofthedebrisdisk. In principle, imaging should provide fainter levels of detection, compared tophotometriclimits,particularlyforwarmdust,withinthehabitablezonearoundastar.Inaddition,largeplanetarybodiescancreateasymmetriesinthediskstructuresthatwouldgounnoticedwithphotometryalone.

SpitzerSpitzerhasbeenable to imagethe thermalemission fromdust indebrisdisksaroundtheclosest stars such as Fomalhaut and Vega. The thermal and scattering properties of thedisksallowus toplaceconstraintson theamountandsizeof thedustgrains in thedisks.But,perhapsmoreimportantly,thepresenceofasymmetrieswithinthedisksmayindicatethepresenceof largeplanetarybodiesorbiting thehost stars. Forexample, in theSpitzerimagesofFomalhaut(Stapelfeldtetal.2004),thediskisasymmetricatboth24and70μm,andintheHSTimage(Kalasetal.2006),thediskisfoundtobeoff‐centerfromthepositionofthestar.Anunseenplanetmayhavegivenrisetotheseasymmetries.

NullingInterferometryInformationondustintheinnerregionsofdiskswillcomefromtwoground‐basednullinginterferometers planned to come into operation over the next few years: the KeckInterferometer and the LBTI. Keck Interferometer observations are currently capable ofdetectingexozodiacaldiskswitha1‐σnoiselimit,whichcorrespondstoapproximately100zodis. The LBTI, planned for operation beginning in 2010, has a significantly differentarchitecture than KI, which greatly reduces the thermal background and stellarsuppression,enablingdetectionofmuchfainterdisks.LBTIisdesignedtoreacha3‐σlimitof 10 zodies for a sample of nearby stars (see section 5.3.2 for expected performance).Theseobservationswillhelpdeterminewhethermanyorallstarshavemorethan10timesthe level of zodiacal dust in our Solar System, which would likely be a complication foreffortstoimageterrestrialplanetsinthehabitablezonesofnearbystars.

ScatteredLightImagingHST observations of dust disks in scattered light have provided interesting insights intomany nearby systems. Detection of disks such as those around Fomalhaut (Kalas et al.2006) and β Pictoris (Golimowski et al. 2006), has demonstrated the ability to detectmaterialwithopticaldepthsof10–4–10–3(1000–10000×thesolarzodiacallevel),onscalesof10–1000AU.TheinclinationofβPic’sinnerdisk(Heapetal.2000)isconsistentwithaplanet sustaining this structure. These observations are typically sensitive to dust inregionscomparabletotheSun’sKuiperbelt.

Inscatteredlight,aboutfourdozenprimordial(protoplanetary)diskshavebeenresolved,plus a similarnumberof disks seen as silhouettes against brightnebulosity inOrion (so‐calledproplyds).Thecensusofdebrisdisksresolvedinscatteredlightismuchsmaller,lessthan20objects.Anup‐to‐date catalogofall these resolvedobjectsmaybeobtained from

Chapter5

144

http://circumstellardisks.org.Theopticaldepthsoftheseresolveddisksrangefromahighof 3× 10–3 (e.g., BetaPic,HD181327)down to about 10–4 (e.g., Fomalhaut,HD139664).ThemajorityofsuchdiskshavebeenimagedusingtheHubbleSpaceTelescope(e.g.,Gradyet al. 1999; Schneider et al. 1999; Clampin et al. 2003;Krist et al. 2005).Adaptive opticsimaging has shown increasing promise in recent years, particularly through reducing theresidual speckle pattern through either angular differential imaging (e.g., Fitzgerald et al.2007; Kalas et al. 2007) or simultaneous polarimetry (e.g., Apai et al. 2004; Perrin et al.2004).

Figure53:Sensitivitylimitsfordetectionofdustaroundnearbysolar‐typestars.Thetwoplotsshowthelimitsintermsoftemperatureofthedust(top)anddetectionwavelength(bottom).3‐σdetectionlimitsareshowninterms of the dust's fractional luminosity (Ldust/L*, left) or fractional flux (Fdust/F*, right). The assumed 1‐σaccuraciesare20%of thestellar flux forSpitzer/MIPSat70µm,2.5%forSpitzer/IRSat32µm,S/N=10 forHerschel/PACSat100µm,S/N=2forHerschel/PACSat160µm,0.5%nullforKeckInterferometerat10µm,andstarlightremovalto0.01%forLBTIat10µm.(G.Bryden,JPL)

ExozodiacalDisks

145

Though not specifically designed for high‐contrast imaging, HST is extremely stable,allowing for significant reduction of the stellar point‐spread function by subtracting areference imageofanotherstaror thesamestarobservedatadifferentorientation. HSThas three coronagraphs: in theNICMOS camera (0.9–2.2 µm,multiple filters), in the STISspectrograph (0.2–1.1 µm, unfiltered), and in the ACS HRC camera (0.2–1.1 µm,multiplefilters).Noneoftheseareoptimized,however,andwithnowavefront‐controlmechanisms,the bulk of the contrast gain is achieved throughPSF subtraction. ACS andNICMOShavecomplementary strengths in this area, with ACS offering better overall scattered lightrejection,hencegreatersensitivity,whileNICMOShasasmallerinnerworkingangle.Atthemoment the STIS and ACS cameras are nonoperational, though they are expected to berepairedduringtheHSTservicingmissionscheduledforneartheendof2008.

Observationsinscatteredlighthaveallowedustomovebeyondaninitialreconnaissanceofdisks towards the detailed characterization of their compositions and three‐dimensionalstructures. Multiwavelength and polarimetric imaging have provided significant insightsintothedetailedpropertiesofthedustgrainsthatcomprisethesedisks,andthusgivenusbetter understanding of the physical processes, which act upon grains. Because differentwavelengthsaremostsensitivetoparticlesofdifferentsizes,broadwavelengthcoverageisessentialtofullyassesstheoveralldustdistributions(e.g.,Metchevetal.2004;Fitzgeraldetal.2007;Pinteetal.2007).

Polarization of light is a fundamental aspect of the scattering process, depending on thepropertiesofthescatteringparticlesandthescatteringgeometry,andthuspolarizationisan indispensiblediagnosticofgrainproperties.This isparticularly true foredge‐ondisks,where polarization breaks the degeneracy between dust properties and the radialdistributionofdust. Inanedge‐ondisk,withoutpolarizationwecannot tell thedifferencebetweenadiskwithconstantdensityandstrongforwardscatteringandadiskwithasteepradialdensitygradientandisotropicscattering.Measurementoflinearpolarizationbreaksthisdegeneracyandallowsunambiguousmeasurementofgrainpropertiessuchasporosity(Graham et al. 2007). Polarimetric studies of dust grain properties remain currentlyunderway with NICMOS onHST, and will hopefully return with a revitalized ACS in thefutureaswell.

Weemphasizethattogetherthecombinationofvisibleandnear‐infraredimaging(andmid‐infrared, when available) provides significantly greater insight into dust properties thananysinglewavelengthalone.Kristetal.(2007)suggestthatJWSTwillbenobetterthanHSTattheshorterwavelengths(1–2µm)forimagingdebrisdisks;however,Doyonetal.(2008)haveshownthatspectraldifferentialcoronagaphyusingJWST’sTunableFilterImagerwilldeliver better performance. JWST also offers new coronagraphic capabilities with high‐performancecoronagraphicimagingfrom2.5–28µm(Clampinetal.2007).JWSTwilloffersignificantopportunities for imaging thepresent sample in themid‐infraredwherea richselectionof spectraldiagnosticsof grainpropertiesareavailable.Over the last fewyears,newdebrisdiskshavebeenresolvedatarateof2–4peryear,andtheplannedrepairofACSthis fall is a high priority to enable this trend to continue. The currently resolved disksrepresentjustthetipoftheiceberg.LessthanaquarterofthedebrisdisksdetectedbyIRAShavesincebeenresolvedinscatteredlight(cf.Figure5‐4),tosaynothingofthemuchlargersampleofdebrisdisksdetectedbySpitzer.However,thegreatestprogressinthisareawillrequire several orders of magnitude increased contrast compared with present systems.Two orders of magnitude improvement in contrast would allow the entire set of IRAS‐detecteddiskstobeimagedinscatteredlight(thoughstillleavingusunabletoimagediskslessthan100×thesolarzodiacallevel).Sincedisksbroadlytrendtoloweropticaldepthsat

Chapter5

146

greater ages, an increased sample size of resolved disks would allow studies across abroaderrangeofagesandevolutionarystatesthannowavailable.

Figure54:DebrisDisksdetectedfromthermalemissionandscatteredlight.Currently,onlyasmallfractionofthe known debris disks in the Solar neighborhood have been resolved in scattered light. This figure showshistogramsoftheknownnearbydebrisdisksdetectedbyIRAS(fromZuckerman&Song2004),comparedwitha histogram of the debris disks which have been resolved in scattered light (data from Paul Kalas andcircumstellardisks.org).(M.Perrin,UCLA)

In the near term, new high‐contrast adaptive optics systems hope to provide at least anorder of magnitude contrast over present capabilities. The Gemini Planet Imager (GPI),currently under development for first light at Gemini in early 2011, has a sciencerequirement to image disks down to an optical depth of 3 × 10–5, for target stars of I <8mag. GPI will attain this goal using its high‐order adaptive optics system with a dual‐channelinfraredimagingpolarimeterthatusesalensletarraytopixellatetheimageplanepriortopolarizationbeam‐splitting,therebyminimizingnon‐commonpatherrors.SimilarlyESO's SPHERE high‐contrast AO system will include both near‐infrared and visible lightdual‐channelpolarimetry,withcomparablecontrastgoals.Thecharacteristicinnerworkinganglesofthesesystemswillbe~0.2arcsec,andtheircontrastwilldecreaseoutsideofthe~1"‐wide dark hole created by the AO system. Hence for typical target distances of 20–100pc,theywillbemostsensitivetodustonscalesof5–100AU,comparableinscaletotheouter Solar System and Kuiper belt. Dust in or near the habitable zone will remaininaccessibletothesesystemsexceptfortheclosesttargets.

NIRInterferometeryAlthoughmostdebrisdisksdonotshowaclearnear‐infraredexcessintheirspectralenergydistributions, limits set by spatially unresolved broadband photometry are generally notbetterthanafewtoseveralpercent.Asmall,warmdustcomponentcouldbepresentifdustgeneratedbycollisionsmigratedclosetothestarorwasproducedbybodiesincloseorbits.IflocatedwithinafewAUofthecentralstar,thisdustwouldbeattemperaturesthatwouldproducenear‐infraredemission,andsmallgrainswouldproducescatteredlight.Detectionof (or stringent limits on)warmdustwill characterize the inner portions of these debrisdisks.Thespatialresolutionofinfraredinterferometrycanbeexploitedtoprobeforwarmdust in these systems. On long baselines (> 100m) the central star is resolved, and thevisibility is primarily a measure of the stellar photospheric size. On shorter baselines

ExozodiacalDisks

147

(<50m) thephotosphere ismostlyunresolved, and if themeasuredvisibilitieshavehighaccuracy,onecansearchforemissionbylookingfordeviationsfromthevisibilityexpectedfor the stellar photosphere. Any resolved or incoherent emission will decrease themeasuredvisibilityfromthestellarvalue.

Obtainingsufficientobservationalprecisioncurrentlylimitssuchstudiesofnear‐IRexcesstotheclosestandbrightestdebrissystems,andtodate,mostobservationshavebeenofAstar systems. Detections have beenmade of 1–2% near‐infrared excesses for the A starsVega(Ciardietal.2001;Absiletal.2006),ζAql(Absiletal.2008)andβLeo(Akesonetal.2009)andonly1lowermassstar,τCeti(diFolcoetal.2007).Tobeconsistentwithboththenear‐infrared excess and thepreviously knownmid‐ to far‐infrared excess requires aring of small, hot dust near the sublimation radius. Given the small number of objectsobservedsofar,itisnotcleariftheseinnerhotdustringsareduetotransienteventsorarethe product of collisions between larger bodies orbiting close to the star.Wediscuss theperformanceofcurrentandnearfutureinterferometersinmoredetailinsection5.3below.

5.2.3 SpectroscopyofDustandGasInformationonthenature(e.g.,sizeandcomposition)of thedustgrains indebrisdisks isvitally important to our understanding of planetary system formation and evolution. Inparticular,thecompositionofthedustgrainsislikelydirectlyrelatedtothecompositionofanyunseenplanetarybodies.Forexample,SpitzerspectroscopicobservationsofHD69830revealarichspectrumofsilicatesandcarbonatessimilartothatseenoncertainclassesofasteroids intheSolarSystem(Beichmanetal.2005;Lisseetal.2007).Theobserveddustmay be the result of asteroids colliding in amassive belt, producing copious amounts ofsmallgrainsinthehabitablezonearoundHD69830.

Interestingly,HD69830showsnoinfraredexcessatlongerwavelengths(70μm),indicatingthat the presence or extent of a debris disk around potential TPF targets cannot bedeterminedunambiguously fromsinglewavelengthobservations. ASpitzer spectroscopicsurveyforwarmdustaroundTPFtargetstarsfindsonlyafewstarswithsilicateemission,foranoveralldetectionrateofjust~1%(Beichmanetal.2006).

Spectroscopyofcircumstellargasesmayprovideinformationonthecompositionofthediskanditsconstituentbodiesthatislessreadilyavailablefromdustobservations.Inaddition,the gas abundance in debris disks is important formodeling of dust structures that havebeenattributedtothegravitational influenceofunseenyoungplanets(e.g.,clearedzones,dustrings,andclumps).

Detections of gas in debris disks have been achieved with optical and ultravioletspectroscopy.Exceptforonecase,observationofresonantlyscatteredgasemissionfromβPic at opticalwavelengths (Olofsson et al. 2001; Brandeker et al. 2004), these detectionshave always involved absorption spectroscopy of edge‐on disks. Such observations aresensitivetoverysmallamountsofcoldgas.Unfortunately,thegeometricconstraintthattheline of sight to the central starmust pass through thedisk severely limits thenumber ofdisks thatmay be probed for gas in thisway. In the longer‐term, an ideal instrument todetect gas in debris disks and also to image new disks uncovered by Spitzer will be theAtacamaLargeMillimeterArray(ALMA),which isexpectedtobeginoperation late in thisdecade.With 64 antennas of 12‐m aperture spread across baselines extending to 10 km,ALMAwillachieveaspatialresolutionof30mas(0.5AUatβPictoris;4AUinnearbystar‐formingregions).ALMAwillprovidedetailedmapsofthedensity,kinematicstructure,andchemicalstructureofprotoplanetarydisksinmolecularandatomiclineemission,andofthe

Chapter5

148

dustcontinuumemissionofnearbydebrisdisks.Withnostellarcontrastproblematthesewavelengths, ALMA shouldmapmoredisks in greater detail than anyprior astronomicalfacility.

5.2.4 TheoryandModelingObservationsofdebrisdisksoftenrevealstructures(suchasrings,clumps,andwarps)thatare generally attributed to the gravitational influence of unseen planets (e.g., Heap et al.2000;Hollandetal.2003).Analogousstructures inourSolarSystemare theverynarrowrings within Saturn’s ring system confined by shepherding satellites and the apparentcircumsolar resonant ring structure created by Earth (Dermott et al. 1994). Recognizingthesestructuresinimagesofexozodiacalcloudsanddebrisdiskscanallowustoindirectlydetectexoplanetsandmeasuretheirmassesandorbitalparameters.ObservationsofsuchstructuresmaybetheonlywaytodetectNeptune‐andEarth‐massplanetslocatedbeyond~10AUfromtheirhoststars,whereorbitalperiodsaretoolongforradial‐velocity,transitand astrometric techniques, and where these planets are too faint to detect directly inreflected starlight or thermal emission. In addition, unresolved clumps in an exozodiacaldustdiskmightinitiallybemistakenforaterrestrialplanet.

ModelingObservedDebrisDisksandtheOriginsofPlanetesimalBeltsThe most well studied debris disk is that of our own Solar System, which exhibits thezodiacalcloudofdustduetocollisionsbetweenasteroidsandtheoutgassingofcomets,andtheKuiperbeltofsmallicybodies.TheorbitaldistributionofKuiperBeltObjects(KBOs)intheouterSolarSystemsuggeststhatalargefractionofplanetesimalsarescatteredoutwardasaconsequenceofgravitationalperturbationbygiantplanets,andasignificantfractionofKBOsare trapped inmeanmotion resonanceswithNeptune. Theseobservations, amongothers,implythattheKuiperBeltandtheouterplanetsofourSolarSystemarehistoricallylinked.Arecentbreakthroughintheoreticalmodelinghasproducedthefirstmodel,knownastheNiceModel,whichsimultaneouslyexplainsthesefeaturesandtheorbitaldistributionofthegiantplanets,theTrojansofJupiterandNeptune,andtheLateHeavyBombardment(Tsiganisetal.2005;Morbidellietal.2005;Gomesetal.2005).TheNiceModelreproducesall of these signatures by modeling the outward migration of the giant planets in aplanetesimal disk truncated at the current location of Neptune. The Nice Modeldemonstrates the importanceofplanetarymigrationmodelsandgives credence to futuremodels that will attempt to describe the origin and distribution of the residualplanetesimalsindebrisdisks.

Models of planetarymigration have already been applied to extrasolar debris diskswithsomesuccess. Incurrently‐observeddebrisdisks thatexhibitopticaldepthsgreater than~100–1000 zodis, collisions dominate the dynamics. Due to the dust grains’ shortcollisional lifetimes, some modelers have interpreted the observed ring structures astracers of the sources of dust. Models of such disks have focused on capturingplanetesimals, the parent bodies of dust particles, into resonances via planet migration(Wyatt2003;Recheetal.2008).Simulationsthathaveinvestigatedcaptureprobabilitiesofplanetesimalsasafunctionofplanetmigrationparametershavesuggestedthatwemaybeable to learn about the migration history of a system from the appearance of a ringstructure(Wyatt2003).Similarmigrationmodelshavealsoshownthattheeccentricityofthe planet’s orbit and initial eccentricity of the planetesimals’ orbits must be very low

ExozodiacalDisks

149

(<0.1) for the resulting ring structure to exhibit any azimuthal asymmetry (Reche et al.2008).

Othermodelsofcollisionallydominateddiskshavefocusedonthesizedistributionofdustincollisionalequilibriumandtheoutcomeoftransientcollisionalavalancheevents(Krivovetal.2007;Grigorievaetal.2007). Wehavelearnedthatwhilecollisionalavalanchescanlead to brightness asymmetries in an edge‐ondisk similar to those thatmay exist due toresonant trappingofdustparticles, theprobabilityofwitnessinganavalancheevent isontheorderofafewpercentforadiskwithasmuchdustasBetaPic(Grigorievaetal.2007).

Severalobservedcollisionallydominateddisksharbordynamicallysignificantquantitiesofgas. Takeuchi&Artymowicz (2001) showed that a relativelymodest amount of gaswillstronglyaffectthedynamicsofsmalldustgrains,whicharethegrainsmosteasilyseeninoptical and near‐ to mid‐infrared disk images, and this can lead to the formation ofazimuthally symmetric structures, such as cleared zones or dust rings, without thegravitational influence of a planet. This general conclusion is supported by more recentmodeling (Klahr & Lin 2006), which also showed that such gas‐formed structures canpersist after the gas is completely gone. These effects may be a significant source ofconfusionforyoungtransitionaldebrisdisks.

PredictionsforExozodiacalDisksOtherthanresultsfromtheKeckInterferometerorLBTI,mostobservationsofdustpriortoTPF‐CandTPF‐Iwillnotmeasuredustemissionoriginatingfromthehabitablezone.Intheabsenceof thesemeasurements,wemaybe temptedtouseobservationsofdustat largercircumstellar distances to estimate the dust abundance in the habitable zone. However,extrapolatinganobservedabundanceofdustfromlargecircumstellardistancesinwardtothehabitable zoneof agiven system is aprecarious task. Unknownsourcesand sinksofdustmayexist inthesesystemsinteriortotheregionswherethesemissionswillobserve.Forexample,ourSolarSystemhasmanysourcesofdustinteriortotheKuiperbelt:comets,mainbeltasteroids,Trojanasteroids,etc.Exoplanetdiscoverieshaverevealedasurprisingvariety of planetary systems, andwehaveno reason to believe that the diversity of dustproducingbodieswillbeanylesssurprising.Understandingthedustinthehabitablezonesofnearbystarswillprobablyrequirebothobservationsdirectlysensitivetohabitable‐zonedust,andalsomodelingeffortstointerprettheseobservations.

Withfutureobservatories,weexpecttodirectly imageexozodiacaldisksanalogoustoourownzodiacalcloud,withopticaldepthsontheorderof~10–7.Long‐termdynamicaleffects,suchasmigrationduetoPoynting‐Robertsonandcorpusculardrag,areimportantinthesetenuous disks. Models have largely ignored collisions altogether and focused on dust‐captureintoaplanet’sexteriormeanmotionresonancesasthedustmigratesinwardunderthe influence of drag (e.g., Jackson&Zook1989;Dermott et al. 1994; Liou&Zook1999;Moro‐Martín&Malhotra2002;Wyatt2006).

Modelsofthesesteady‐statezodiacalcloudstructureshavecomealongwaysinceDermottetal.(1994)firstrecognizedtheleading/trailingasymmetryinthesurfacebrightnessoftheinfraredskyasacircumsolarringofzodiacaldusttrappedinresonancewiththeEarth.Wenow more‐or‐less understand the geometry of resonant signatures from analyticalcalculations (Kuchner & Holman 2003) and numerical simulations (e.g., Moro‐Martín &Malhotra 2002; Deller & Maddison 2005). Figure 5‐5 qualitatively illustrates the basicgeometries of resonant structures created by a single planet in a collisionless dust cloud(Kuchner & Holman 2003). Basic numerical models have also revealed that planets on

Chapter5

150

eccentricorbitscantrapparticlesinresonanttermsthatproducedensitystructureswhichappear to revolve at mean rates slower than the mean motion of the perturbing planet(Wilneretal.2002).

Figure55.Approximateanalyticdescriptionoftherangeofresonantstructuresasingleplanetcancreateinacollisionlesscloud. Case I,aringwithaco‐rotatinggapat the locationof theplanet isexemplifiedbyEarth'sinteractionwiththesolarzodiacalcloud.Theothercaseshavebeensuggestedasmodelsforotherdebrisdisks,butnotyetverified.(Kuchner&Holman2003)

Figure56.Contrast in optical depth formulti‐particle‐size clouds (solid lines) assuming aDohnanyi (1969)crushinglawcomparedtosingle‐particle‐sizeclouds(dashedlines).Fromtoptobottom,thesixsolidlinesandsixdashedlinescorrespondtosixvaluesofplanetmass:5,2,1,0.25,and0.1M⊕.Thecontributionsofthesmallgrains reduce the contrasts of themulti‐particle‐size clouds compared to single‐particle‐size cloudswith thesameminimumvalueofβ,whereβistheratiooftheforceduetoradiationpressuretothegravitationalforceonadustgrainandisproportionaltos–1,theradiusofthedustgrain.(Stark&Kuchner2008)

ExozodiacalDisks

151

Recent advances in collisionless disk models have allowed for a number of quantitativepredictions regarding the morphology and contrast of resonant structures formed byterrestrial‐massplanets.Thesemodelshaveshownthatmostresonantstructuresfeatureasharpinner‐edgeatacircumstellardistanceof83%ofthesemi‐majoraxisoftheplanet,agapintheringstructureatthelocationoftheplanet,whichvariesinsizelinearlywithringbrightness(fromafewdegreestonearlyninetydegrees),andringwidthsthatvaryfromafewpercentto1.6timesthesemi‐majoraxisof theplanet. Thesamemodelshaveshownthatadegeneracyexists inparticlesize,s,andplanetsemi‐majoraxis,ap; twoexozodiacalring structures can be geometrically identical for a constant value of (sap1/2) (Stark &Kuchner2008).

Modelsofcollisionlessexozodiacalcloudsofasinglegrainsizeestimate that thecontrast,defined as the ratio of optical depth within the ring structure to outside of the ringstructure, of resonant ring structures created by terrestrial‐mass planets can vary fromunityto7:1,asshowninFigure5‐6.Thisfigurealsoillustratesthatadistributionofparticlesizes tends towash out ring structures, reducing their contrast by up to~50% (Stark&Kuchner 2008). Figure 5‐7 shows the minimum detectable planet mass that can beindirectlydetectedviaobservationsof itsresonantringstructure,assumingadistributionofparticlesizesandaminimumdetectableringcontrastof1.5:1. Theseestimatesassumethedustisproducedentirelybyasourceanalogoustoourasteroidbelt;nohighly‐inclinedorhighly‐eccentricsourcescontributetotheproductionofdust,andthusFigure5‐7canbethoughtofasabest‐casescenario.

Figure57. Minimumdetectableplanetmassinamulti‐particle‐sizecollisionlesscloudasafunctionofsemi‐majoraxis,ap, andmaximumgrainsize,smax, assumingaSun‐likestar,aminimumdetectable ringcontrastof1.5:1, anddust produced according to aDohnanyi (1969) crushing law. Earth‐like andMars‐likeplanets aredenotedwith anE andM, respectively. The 5.5M⊕ exoplanet OGLE‐2005‐BLG‐390Lb is denotedwith anO.Listed values for maximum dust size in the ring structure assume perfectly absorbing spherical grains withmeandensityρ=2.0gmcm–3.Theboldlineshowsthecaseofthesolarzodiacalcloud,forwhichtheobservedemission isdominatedby30‐μmgrains. Thedashed linesshowtypical innerandouterdetection limits foramissionsimilartoTPF.(Stark&Kuchner2008)

Chapter5

152

5.3 TechnologyDevelopmentandPerformanceDemonstration

Thetechnologydevelopmentsectionfocusesonthedevelopmentofnullinginterferometryin particular as the key technology needed to advance our knowledge of dust in thehabitable zones around nearby stars. This technique has been developed for the KeckInterferometer and the Large Binocular Telescope Interferometer (LBTI), with prototypetestingforLBTIbeingcarriedoutontheMultipleMirrorTelescope(MMT).

5.3.1 PastAccomplishmentsInterferometricnullingofthermalemissionfromstarswasdemonstratedforthefirsttimein1998usingtheMMT.Byinterferinginfraredradiationfromtwooftheco‐mounted1.8mtelescopes, the stellar image was suppressed by a factor of order 25 and used to mapdirectlythethermalemissionofwarmdustsurroundingBetelgeuse(Hinzetal.1998).

TheMMTwasrebuiltasasingle6.5‐mtelescopeearlyinthedecadeandoutfittedwithanadaptivesecondarymirrorin2002.TofacilitatetestinginpreparationforLBTI,acryogenicnulling interferometer called the Bracewell Infrared Nulling Cryostat (BLINC), wasdeveloped(Hinzetal.2000).Theinstrumentcreatesapseudo‐interferometerwithelliptical2.5×5‐maperturesacrossa4‐mbaseline.Phasevariationsbetweenthebeamsaresensedusing 2.2‐μm light that travels an identical path to the light at 11 μm. A dispersionintroduced into thesystemcreatesanull, that isbothachromaticacross thesciencebandandcreatesasuitablephasingsignalat2.2μm(seeHinzetal.2000formoreinformation).The instrument is mated with a Mid‐Infrared Array Camera (MIRAC) that measures theamountoffluxinthenulledoutputoftheBLINC.

NullinginterferometrictechniqueshavebeendevelopedatJPLandaredescribedindetailinChapter4. ThesetechniqueshavebeenadaptedforuseintheKInullerandarenowinregularuseattheKeckInterferometer.

The Keck Interferometer combines the two 10‐m Keck telescopes as a long baselineinterferometer, funded by NASA, as a joint development among the Jet PropulsionLaboratory,theW.M.KeckObservatory,andtheNASAExoplanetScienceInstitute.TheKInuller (Serabyn et al. 2004; Colavita et al. 2006, 2008) is implemented as a four‐beamsystemoperatingatawavelengthof10µm.ThetwoKecktelescopeaperturesaresplitintoleft (“primary”) and right (“secondary”) halves at a dual‐star module (DSM) at eachtelescope.Accountingfordiffractionandotherdetails,thebeamsizeontheskyat10µmisapproximately0.45''×0.50''.TwomodifiedMach‐Zehnderbeamsplitterscombinethelightfromthelefthalvesandrighthalvesonthelong85‐mbaseline.Theoutputsofthetwolongbaselinesarecombinedinabeamsplitter—thecrosscombiner—withashort4‐meffectivebaseline.Theoutputofthecrosscombinerfeedsamid‐IRcamera,KALI.Modulationonthelong baselines is used to chop the central star to detect surrounding extended emission,whilemodulation on short baseline allows fringe detection in the presence of the strongthermal background. Because of the limited measurement and control bandwidthsachievableintheintegrationtimesrequiredtoobservefaint10‐µmsources,phasingandtiltstabilizationrelyuponfeed‐forwardfromtwo2‐µmfringetrackers(i.e.,phase‐referencing,or cophasing), as well as tilt feed‐forward from the KI angle tracker operating at 1.2 or1.6µm.Lasermetrology and accelerometer feed‐forward is used to stabilize against non‐atmospheric disturbances. A distributed real‐time control system controls the various

ExozodiacalDisks

153

servosand interconnections,aidedbyhigh‐levelsequencers.SeeColavitaetal. (2008) formoredetails.

5.3.2 PerformanceKeck Interferometer is currentlyoperational and isperformingKeyScience surveys. TheLBTI is being completed with operation anticipated for 2010. The performance,demonstratedandexpectedforeachfacility,isdescribedbelow.

KeckInterferometerThecurrentobservationalapproachfortheKeckInterferometeristodividethenightintothreesegments,eachonededicatedtoasinglesciencetarget.Accountingforoverheads,thisiswellmatchedtothe±14‐musablecontiguousdelayrangeforthenullerdelaylines;thisrangeallowsfor2–3hourdelaytracksonthesciencetargets.Theobservationsgoal isfortwoorthreesciencescanswithinterleaved,bracketingcalibratorswithbetween10and15minutes of null/peak data on each target, depending on the source brightness. Includingmorethanonecalibratedscanperclusterallowsforbetterestimatesoftheexternalerrors,rather than relying on formal errors from the internal scatter of a single observation.Repeated observations on subsequent nights provide confidence in the nightly externalerrors.

A series of performance‐validation tests were performed between June and Aug. 2007,whichrevealedaslight,butstatisticallysignificant,night‐to‐nightbias.Thiswasultimatelytracedtoanadditivelong‐wavelengthleakagetermwhoseimpactwasfluxdependent.Themostlikelyexplanationforthebiasappearstoberadiationfromstructureatthetelescopetopendthatisdiffractingintothebeampathandintroducingcorrelatedemissionintothetwo halves of each aperture. By changing the adaptive optics rotator offset angle, whichchangestheorientationofthepupilsplitonthetelescope,andthusthegeometryoftop‐endstructurewithrespect to thecross‐combinerbaseline, this long‐wavelength leakagecouldbe substantially reduced, although not completely eliminated. Observationally, morecarefullymatchingtargetandcalibratorfluxesreducedthesizeoftheeffect,andshiftingthebroadbandnull reductionbandpass slightly toward shorterwavelengthsalso reduced themagnitudeoftheeffect.Withthesethreechanges,thebiasesattributabletothiseffecthavebeen shown to bemuch smaller than the external error per cluster based on 21 scienceclustersmeasured throughApril 2008. From analysis of this data set, theKI achieved anexternalerrorperclustercomputedfromtheerrorofthecalibratedscansofapproximately0.25% RMS in a broadband 8–9 µm channel. This value is equivalent to 100 zodi RMS(where1zodiisthefluxproducedbythezodiacaldustwithinourownSolarSystem)afteraccountingfortransmissionthroughthelong‐baselinefringepattern.

NASAhasallocatedasubstantialportionofits2008KecktimetoKeyScienceprojectswiththenuller.ThreeKeyScienceteamswereselectedinNov.2007;thePIsoftheteamsarePhilHinz,Univ.ofArizona;MarcKuchner,GoddardSpaceFlightCenter;andGeneSerabyn,JPL.Thekeyscienceobservingprogramiscurrentlyunderway,andbythetimeitcompletesinFeb2009,willhaveobservedroughly40nearbymain‐sequencestars,whicharepotentialtargetsforfutureplanet‐findingmissionsandstarswithknowndebrisdisks.Thecontinuedoperation of KI for nulling observations after February 2009 is uncertain; no funding isbeingplannedafterthistimeperiod.

Chapter5

154

LBTITheLargeBinocularTelescope Interferometer is a common‐mount design interferometerwith8.4‐maperturesona14.4‐mbaseline.Thecommon‐mountdesignnegatestheneedforlongdelaylines.Onlythreewarmmirrors(primary,secondaryandNasmythflat)areusedtodirectthebeamstowardthecentralinstrumentplatform.Thisisespeciallyimportantforthermal infrared (> 3 μm) observationswhere telescope emissivity can be the dominantsourceofnoise.ThedeformableelementsoftheAOsystemsoftheLBTareintegratedintothe secondarymirrors of the telescope,which form the aperture stop of the telescope tooptimizeinfraredperformance.ExtrasurfacesinaconventionalAOsystemaddasignificantamount of background light in the thermal infrared, decreasing the system's sensitivity.These several unique aspects of the LBTI allow for efficient high sensitivity beamcombinationthatiswellmatchedtothegoalsofthermalinfrareddetectionofzodiacaldust.

Very deep stellar suppression can be achieved with nulling interferometry. Thefundamentallimittothisapproachistheangularsizeofthestarcomparedtothebaselineoftheinterferometer.Forasolartypestarat10pcthislimitisbelow10–4fortheLBTI.AtthesametimetheLBTIwillbesensitivetodustλ/2bawayorapproximately0.7AUforastarat10 pc. This combination of deep stellar suppression achievable, low background in thethermalinfrared,andgoodspatialresolutionfornearbystarswillenabletheLBTItoprobenearbystarsforzodiacaldusttoanunprecedentedlevel.

PerformanceestimatesfortheLBTIarebasedonobservationaltestsbeingcarriedoutwiththeBracewellInfraredNullingCryostat(BLINC)ontheMMT.Althoughprimarilyintendedasanengineeringprototype,BLINCcanprobetheverynearestandbrighteststarsforwhichtheregionofemissionat11μmiscomparable toλ/2b,or0.25arcsec. Initialresultshavebeen demonstrated by probing for zodiacal dust around Vega (Liu et al. 2004). The 1‐σuncertaintyinthenullisapproximately7×10–3forobservationsofVega,equivalenttoanapproximately 200‐zodi dust disk around Vega. More recent observations (spring 2008)have reduced this uncertainty to approximately 1 × 10–3 within a specific observationsequence,butwithasystematicvariationwhichisthreetimeslarger.

The level of the null uncertainty at the MMT is expected to be limited primarily by theperformanceof theadaptiveoptics correction. For theparametersof theMMTsystem, avariationinthecharacteristicscaleofatmosphericturbulence,r0,by50%overthetimeofobservationwill limit the null to approximately 8× 10–4, in agreementwith the internalerrormeasuredattheMMT.Thesuspectedsourceofthesystematicchangeisvariationsintelescope vibration under differingwind conditions and azimuth angles of the telescope.Morecareful calibrationand improvementsof the telescopepointing shouldalleviate thisaffectinfutureruns,butitremainstobedemonstratedthatthesystematicvariationscanbekeptbelowthestatisticalvariationsintroducedbychangesinatmosphericseeing.

To properly estimate limits to detection of zodiacal dust with LBTI it is important tounderstandphotometricdetectionlimitations,andnulldepthandstabilitylimitations.Theformerlimitsthedetectabledustforlater‐typestarsandstarsfurtheraway,andthelatterlimits dust detection around even the closest, early‐type stars. At 11.1 μm a telescopeemissivityof7% for theLBTgives abackground fluxof1.6× 1010photons·s–1 for a20%passband. The sky background from a HITRAN model atmosphere (Gillett & Mountain1998)isexpectedtobe3.4×109photons·s–1.Ifthetotalthroughputissetat25%,the5‐σlimit to sensitivity is 100 μJy per √hour of integration for the LBTI. This photometricuncertaintyissufficienttodetectdustdisks0.5,2,9,and100timessolarforaF0,G0,K0,

ExozodiacalDisks

155

andM0starat10pc, respectively. Asdiscussedabove, thedetectionofdust isweightedtowardearliertypestars.

For good photometric sensitivity, detection limits are set in practice by the level of nulluncertainty. For the AO performance of the LBT, the null uncertainty in a typicalobservationisexpectedtobeapproximately1.2×10–4.ThisisnearlysixtimesbetterthantheMMTandisduetotwoeffects:thelargeraperturesoftheLBT,relativetotheMMT,andthefasterupdaterateoftheLBTAOsystem(1kHz,versus500HzfortheMMT).Thisnulluncertainty is equivalent to approximately a dust disk that is 3 times solar. At this levelconfirmationofdetectionofasignalwillbeobtainedbycomparingspectralinformationoftheresidualflux.Fluxfromadustdiskwillbereasonablyconstantinfluxacrossthethe8–13μmpassband,while residuals frompoorly suppressed starlightwill followaRayleigh‐Jeans1/λ2spectralslope.

5.3.3 FutureDevelopmentsandMilestonesFutureeffortinstudyingexozodiacaldiskswillfocusonincreasedsensitivityobservationswith theLBTIandHerschel,developing improvedmodelingandtheoreticalpredictionsofdust disk‐planet interactions, studying the use of ELTs for exozodical detection, andexploringconceptsforanexozodiacalprecursormission.

LBTIMilestonesTheLargeBinocularTelescopesaw first lightwith the firstprimary inOctober2005.Thetelescope is currently in use for astronomical observations using prime‐focus wide‐fieldimagers. For interferometric operations adaptive secondaries are required. These areplannedfordeliveryinMarch2009andNovember2009forthefirstandsecondsecondary,respectively.

TheLBTInterferometeriscurrentlybeingassembledandtestedatStewardObservatory.Insummer 2008 it was shipped to the LBT on Mt. Graham for initial integration with thetelescope.Ongoingtestsareplannedastheadaptivesecondariesarecompleted.Oncetheadaptive secondariesare integratedwith the telescope, approximatelyoneyearof testingandcommissioningobservationswillberequiredpriortoroutineexozodiacalobservations.Onthecurrentschedulethiswouldresultinscientificoperationsbeginninginlate2010toearly2011. Futureoperationof theLBTI is still in question, becauseNASA fundingplanscurrentlyonlycoverthecompletionoftheinstrument.

TheLBTIisbeingdevelopedspecificallytocarryoutaNullingInfra‐RedExtrasolarSurveyfor TPF (NIREST, as in the “nearest” stars). This will involve observations of acomprehensive sample (60–80) of nearby, main‐sequence stars. In addition tocharacterizing candidate systems forTPF, the surveywill provideuseful insights into theprevalenceofexozodiacaldustwithvariousstellarparameters.

Thesample,ascurrentlycomposed,hasapproximately15starsineachspectraltypebinA,F,G,andK,with8Mspectraltypestars.Agedeterminationwillbeanadditionalfactorthatwill factor into our final source selection. Each star is anticipated to require 4 hours toachieveareasonablesignaltonoiselevel.Thetotalsurveywillthenrequireontheorderof40–60nights.Asamplewithapproximatelyequalnumberofstarsinthreeagebinsof0–1,1–2,and2–3Gyr,willbeconstructedtoallowaroughdeterminationofzodiacaldustwithbothspectraltypeandageofastar.

Chapter5

156

TheoryDevelopmentThere is much more numerical and analytical work to be done to provide a soundframeworktounderstandthepossiblerangeoflong‐livedstructuresthatmaypersistinanexozodiacalcloud. Futuremodelsmust focusondisks forwhichbothcollisionsand long‐term dynamical evolution are important. Understanding collisional effects is crucial tounderstanding the dynamics of most currently known debris disks, which have opticaldepths 100–10,000 times that of the solar zodiacal cloud. The number of observedcollisional disks is expected to grow significantly in the next few years with theobservationsofJWST,Keck,ALMA,andLBTI.Interpretingtheseimageshingesonimprovedmodelsofcollisionaldustclouds(e.g.,Wyatt2005;Krivovetal.2007).Atpresent,theonlyexampleofasystemwhereaknownplanetcreatesanobservedstructureinadebrisdiskisthesolarzodiacalcloud.Observationsofotherknownplanet‐diskinteractionswillprovideadirecttestforthepredictionsmadebyourmodels.

A number ofmodels have attempted to include collisions in dynamical calculations (e.g.,Kenyon&Bromley2005;Thébault&Augereau2007),butfewhavebeenabletotreatbothcollisionaleffectsandresonanttrappingsimultaneously.Thisproblemwilllikelyyieldtoastatisticaltreatmentofcollisions(e.g.,Thébault,Augereau&Beust2003;Krivovetal.2007).Several recent codes show promise in solving this issue (e.g., Grigorieva, Artymowicz &Thébault 2007), and dynamical disk models that self‐consistently treat collisions areexpectedtoemergeinthenextfewyears.

ExtremelyLargeTelescopesThenextgenerationoftelescopes,includingtheTMT,GMT,andE‐ELT,potentiallycanreachmuch fainterphotometric sensitivity limits thanKIandLBTIbyhavingsignificantlymorecollectingareaandsuitablediametertoresolvethehabitablezonesofmoststars likelytobeinasamplefordetectionofanEarth‐likeexoplanet. Ifthetelescopesareoptimizedforperformance in the thermal infrared, it is possible thatmany lowermass stars could besearched for debris disks thanwill be possible with KI and LBTI. For example, the GMTexpected photometric sensitivity at N band in 1 hour is approximately 20 µJy. If the AOperformanceissimilartoLBTI, this issufficienttodetectdustdownto3timessolar levelforallstarswithin10pcearlierthanK4V.

ScatteredlightimagingwillalsobeafocusofELTswithsimilarhigh‐contrastAOsystemstoGPI and SPHERE (e.g., the Planet Formation Imager for the Thirty Meter Telescope;Macintosh et al. 2006). Based on current design studies, PFI will reach similar or veryslightly greater contrast than GPI (an optical depth of 3 × 10–5), but its improved innerworking angle (0.03") will allow imaging inwards toward the habitable zone for manytargets,andtoscalescomparabletoourownasteroidbelt(nearthe"snowline")foryoungsystemsasdistantasTaurus.

AnExozodiacalDustMissionGiven the challenging nature of detecting exozodiacal dust from the ground, and theimportanceof it fordirect imaging,amissiontocharacterizeexozodiacaldustexplicitly isanattractiveoption. Suchamissionmightbecarriedout in the infrared,viadetectionofthermal emission, or in the visible via scattered light. In either case itwill be critical tomeasuredustinthehabitablezone,whichrequiresobservingdustat1AUorequivalentforstarsofdifferentluminosity.Todosolikelyrequiresasignificantmission,onthescaleofan

ExozodiacalDisks

157

Exoplanet Exploration Probe, rather than a SMEX‐scopemission, but the scope of such amissionshouldbefurtherexploredasafirststeptowardassessingitsfeasibility.

SeveralcoronagraphicmissionsarecurrentlyunderstudyaspossibleProbeclassExoplanetmissions, including ACCESS and the Visible Light Nulling Coronagraph. A 2‐m classcoronagraph(Stapelfeldtetal.2007)providesthe0.1spatialresolutionneededtoresolvethe rings,warps, andasymmetriesdrivenbyplanetaryperturbations in thesedisks.Withcontrastimproved~1000timesoverHST,italsowillbesensitiveenoughtodetectzodiacaldisks at approximately solar level, enabling comparative studies of dust inventory andpropertiesacrossstellaragesandspectraltypes.

AninfraredmissionthatcanobserveallthenearbystarsofinteresttotheflagshipDarwinand TPFmissions has been under study formany years, and is called the Fourier‐KelvinStellarInterferometer(FKSI)mission(Danchietal.2003).Thepresentdesignisapassivelycooledstructurallyconnectedinterferometer,operatingat60K,from3–8(or10)µm,withtwo50‐cm apertures separated by 12.5mon a boom. This systemhas been extensivelystudiedinthecontextofdetectingandcharacterizingdebrisdisks(seeDefrèreetal.2008)andhasbeenshowntobeabletodetect,withinafewminutes,1SolarSystemzodidisksinthehabitablezoneofanystarwithin30pcthatisaccessiblewithinitsfieldofregard(FoR).With a sunshade having a±20‐degree FoR a survey of 443 stars (28 F, 74 G, 164K, and177Mstars),andthisnumbergoesupto>900starsiftheFoRisincreasedto±40degrees.Thus, an infrared space interferometer is not limited in spectral types or the number ofstarsthatcanbesurveyed,exceptbyitssunshadeconfiguration.FurtherdetailsonasmallstructurallyconnectedinterferometerarediscussedinChapter4inthisvolume.

5.4 Research&AnalysisGoalsFurtheringknowledgeoftheexistence,prevalence,andstrengthofeoxozodiacaldustdisksis critical for planning of future direct‐imaging missions for Earth‐like exoplanets. Thedevelopment of aDesignReferenceMission for any of the direct‐imagingmissions underconsideration, including TPF‐C, TPF‐O, or TPF‐I, requires information about the averagezodiacal dust density around target stars. The information gleaned from detection ofexozodiacal dust may also enhance our understanding of exoplanetary systems, as hasalreadyhappenedwithobservations fromHST andSpitzer. Theprevalenceofdust in thehabitable zones, and the structure of this dust, can provide important clues to thearchitecturesof exoplanetary systems. Thepivotal natureof this information leadsus tothefollowingthreerecommendations.

• Sinceinformationaboutdustinthehabitablezoneiscrucialtoplanningfuturedirectdetectionmissions,thecontinuedsupportofKI,LBTI,aswellasanyadditionaleffortsthatcanaddressthisquestionisneededtomaintainmomentuminthisarea.

• Exoplanet Probe class missions that can address the density and distribution ofdustinotherplanetarysystemsshouldbefurtherstudiedinordertooptimizetheirscientificreturnandcostandtechnicalfeasibility. AnExoplanetProbemissiontodetectandcharacterizedebrisdisksandcircumstellarmaterialaroundallTPFandDarwintargetstars,andtodetectandcharacterizetheatmospheresofSuperEarthtogiantswillreducescientificandtechnicalriskforlaterflagshipmissions.

Chapter5

158

• Arobusttheoryprogramanddevelopmentofcollisionaldebrisdiskmodelswillaidindecodingtheresonantstructuresinducedbyplanetsinobserveddebrisdisks.Thesemodelswillgiveinsightintodusttransportandthedetectabilityofexoplanetsviaobservationsofdiskstructure.

5.5 ContributorsOlivierAbsil,LAOG

RachelAkeson,NASAExoplanetScienceInstitute,Caltech

DavidArdila,SpitzerScienceCenter

JamesBreckenridge,JetPropulsionLaboratory


ChristineChen,STScI

DavidCiardi,NASAExoplanetScienceInstitute,Caltech

MarkClampin,NASAGoddardSpaceFlightCenter

VincentCoudéduForesto,ObservatoiredeParis

WilliamDanchi,NASAGoddardSpaceFlightCenter

DenisDefrère,UniversitédeLiège

DennisEbbets,BallAerospace

SallyHeap,NASAGoddardSpaceFlightCenter



CharlesLillie,NorthropGrumman

PatrickLowrance,SpitzerScienceCenter


RafaelMillanGabet,NASAExoplanetScienceInstitute,Caltech

MarshallPerrin,UCLA

MeyerPesenson,Caltech

PeterPlavchan,NASAExoplanetScienceInstitute,Caltech

SamRagland,KeckObservatory


AkiRoberge,NASAGoddardSpaceFlightCenter

ChrisStark,UniversityofMaryland


MotohideTamura,NAOJ

AngelleTanner,JetPropulsionLaboratory

ExozodiacalDisks

159

NealTurner,JetPropulsionLaboratory/Caltech

BruceWoodgate,NASAGoddardSpaceFlightCenter

5.6 ReferencesAbsil, O., Di Folco, E., Mérand, A., et al. 2008, “A near‐infrared interferometric survey of

debrisdisc stars. II.CHARA/FLUORobservationsof sixearly‐typedwarfs,”A&A,487,1041–1054

Absil, O., Di Folco, E., Mérand, A., et al. 2006, “Circumstellar material in the Vega innersystemrevealedbyCHARA/FLUOR,”A&A,452,237–244

Akeson, R., Ciardi, D. R.,Millan‐Gabet, R., et al. 2009, “Dust in the inner regions of debrisdisksaroundAstars,”ApJ,691,1896–1908

Apai,D., Pascucci, I.,Brandner,W., et al. 2004, “NACOpolarimetricdifferential imagingofTWHya.AsharplookattheclosestTTauridisk,”A&A,415,671–676

Beichman,C.A.,Bryden,G.,Stapelfeldt,K.R,etal.2006, “NewDebrisDisksaroundNearbyMain‐SequenceStars:ImpactontheDirectDetectionofPlanets,”ApJ,652,1674–1693

Beichman,C.A.,Bryden,G.,Gautier,T.N,etal.2005,“AnExcessDuetoSmallGrainsaroundtheNearbyK0VStarHD69830:AsteroidorCometaryDebris?”ApJ,626,1061

Brandeker,A.,Liseau,R.,Olofsson,G.,etal.2004,“ThespatialstructureoftheβPictorisgasdisk,”A&A,413,681

Brown, R. A. 2005, “Single‐Visit Photometric and Obscurational Completeness,” ApJ, 624,1010

Butler, R. P., Vogt, S. S., Marcy, G. W., et al. 2004, “A Neptune‐Mass Planet Orbiting theNearbyMDwarfGJ436,”ApJ,617,580–588

Chauvin,G.,Lagrange,A.‐M.,Dumas,C.,etal.2004,“Agiantplanetcandidatenearayoungbrowndwarf.DirectVLT/NACOobservationsusing IRwavefront sensing,”A&A,425,L29–L32

Ciardi,D.R.,vanBelle,G.T.,Akeson,R.L.,Thompson,R.R.,Lada,E.A.,&Howell,S.B.2001,“Onthenear‐infraredsizeofVega,”ApJ,559,1147–1154

Clampin, M. 2007, “The James Webb Space Telescope (JWST): A Tool for the Study ofPlanetarySystemFormationandEvolution,”AmericanAstronomicalSocietyDivisionofPlanetarySciencesmeeting#39,#12.05,Bull.Am.Astron.Soc.,39,432.

Clampin, M., Krist, J. E., Ardilla, D. R., et al. 2003, "Hubble Space Telescope ACSCoronagraphic Imagingof theCircumstellarDiskaroundHD141569A,”AJ,126,385–392

Colavita, M. M., Serabyn, E., Wizinowich, P. L., et al. 2006, “Nulling at the KeckInterferometer,”Proc.SPIE,6268,626803

Colavita,M.M., Serabyn, E., Booth, A. J., et al. 2008, “Keck Interferometer nuller update,”Proc.SPIE,7013,70130A

Danchi,W.C.,Deming,D.,Kuchner,M.J.,etal.2003,“DetectionofClose‐InExtrasolarGiantPlanetsUsingtheFourier‐KelvinStellarInterferometer,”ApJ,597,57

Chapter5

160

Defrère,D., Absil, O., Coudédu Foresto, V., Danchi,W. C.,& denHartog, R. 2008, “Nullinginterferometry: performance comparison between space and ground‐based sites forexozodiacaldiscdetection,”A&A,490,435–455

Deller,A.T.,&Maddison,S.T.2005, “Numericalmodellingofdustydebrisdisks,”ApJ,625,398–413

Dermott,S.F., Jayaraman,S.,Xu,Y.L.,etal.1994,“Acircumsolarringofasteroidaldust inresonantlockwiththeEarth,”Nature,369,719–723

diFolco,E.,Absil,O.,Augereau,J.‐C.,etal.2007,“Anear‐infraredinterferometricsurveyofdebris disk stars. I. Probing the hot dust content around ε Eridani and τ Ceti withCHARA/FLUOR,”A&A,475,243–250

Dohnanyi,J.S.1969,“Collisionalmodelsofasteroidsandtheirdebris,”J.Geophys.Res.,74,2531–2554

Doyon, R., Rowlands, N,Hutchings, J., et al. 2008, “The JWST tunable filter imager (TFI),”Proc.SPIE,7010,70100X

Fitzgerald, M. P., Kalas, P. G., Duchêne, G., et al. 2007, “The AU Microscopii Debris Disk:MultiwavelengthImagingandModeling,”ApJ,670,536–556

Gillett,F.C.&Mountain,M.1998,“OntheComparativePerformanceofan8‐mNGSTandaGroundBased8‐mOptical/IRTelescope,”inSciencewiththeNGST,editedbySmith,E.P.,andKoratkar,A.,ASPConf.Ser.,133,42

Golimowski, D. A., Ardilla, D. R., Krist, J. E., et al. 2006, “Hubble Space Telescope ACSMultibandCoronagraphicImagingoftheDebrisDiskaroundβPictoris,”AJ,131,3109–3130

Gomes, R., Levison, H. F., Tsiganis, K., et al. 2005, “Origin of the cataclysmic Late HeavyBombardmentperiodoftheterrestrialplanets,”Nature,435,466–469

Grady, C. A., Woodgate, B., Bruhweiler, F. C., et al. 1999, “Hubble Space Telescope SpaceTelescope Imaging Spectrograph Coronagraphic Imaging of the Herbig AE star ABAurigae,”ApJ,523,L151–L154

Graham,J.R.,Kalas,P.G.,&Matthews,B.C.2007,“TheSignatureofPrimordialGrainGrowthinthePolarizedLightoftheAUMicroscopiiDebrisDisk,”ApJ,654,595–605

Grigorieva,A.,Artymowicz,P.,&Thébault,Ph.2007,“Collisionaldustavalanches indebrisdiscs,”A&A,461,537–549

Heap, S. R., Lindler, D. J., Lanz, T.M., et al. 2000, “Space Telescope Imaging SpectrographcoronagraphicobservationsofβPictoris,”ApJ,539,435–444

Hillenbrand,L.A.,Carpenter, J.M.,Kim, J. S., etal.2008, “TheCompleteCensusof70µm‐brightDebrisDiskswithinthe‘FormationandEvolutionofPlanetarySystems’SpitzerLegacySurveyofSun‐likeStars,”ApJ,677,630–656

Hinz,P.M., Angel,J.R.P., Woolf,N.J., et al. 2000, “BLINC: a testbed for nullinginterferometryinthethermalinfrared,”Proc.SPIE,4006,349

Hinz,P.M.,Angel,J.R.P.,Hoffmann,W.F.,etal.1998,“Imagingcircumstellarenvironmentswithanullinginterferometer,”Nature,395,251–253

Holland,W. S.,Greaves, J. S.,Dent,W.R. F., et al. 2003, “Submillimeterobservationsof anasymmetricdustdiskaroundFomalhaut,”ApJ,582,1141–1146

ExozodiacalDisks

161

Holweger,H.,Hempel,M.,vanThiel,T.,etal.1997,“ThesurfacecompositionofβPictoris,”A&A,320,L49–L52

Jackson, A. A.,& Zook,H. A. 1989, “A Solar Systemdust ringwith Earth as its shepherd,”Nature,337,629–631

Johnson,J.A.,Fischer,D.A,Marcy,G.W.,etal.2007,“RetiredAStarsandTheirCompanions:ExoplanetsOrbitingThreeIntermediate‐MassSubgiants,”ApJ,665,785–793

Kalas,P.,Graham,J.R.,Clampin,M.2005, “Aplanetary systemas theoriginof structure inFomalhaut'sdustbelt,”Nature,435,1067

Kalas,P.,Graham,J.R.,Clampin,M.C.,&Fitzgerald,M.P.2006,“FirstScatteredLightImagesofDebrisDisksaroundHD53143andHD139664,”ApJ,637,L57–L60

Kalas,P., Fitzgerald,M.P.,&Graham, J.R.2007, “DiscoveryofExtremeAsymmetry in theDebrisDiskSurroundingHD15115,”ApJ,661,L85–L88

Kenyon, S. J.,&Bromley,B.C.2005, “Prospects forDetectionofCatastrophicCollisions inDebrisDisks,”AJ,130,269–279

Klahr,H.,&Lin,D.N.C.,“DustdistributionaroundHR4796AandHD141569:aselfinducedringformationthroughaclumpinginstability,”inProtostarsandPlanetsV,Reipurth,B.,Jewitt,D.,andKeil,K.,editors,(UniversityofArizonaPress,Tucson,2006),contributionNo.1286,8361.pdf.http://www.lpi.usra.edu/meetings/ppv2005/pdf/8361.pdf

Krist,J.,Beichman,C.A.,Trauger,J.T.,etal.2007,“HuntingplanetsandobservingdiskswiththeJWSTNIRCamcoronagraph,”Proc.SPIE,6693,66930H

Krist, J.E.,Ardila,D.R.,Golimowski,D.A., etal.2005, “HubbleSpaceTelescopeAdvancedCameraforSurveysCoronagraphicImagingoftheAUMicroscopiiDebrisDisk,”AJ,129,1008–1017

Krivov,A.V.,Queck,M.,Lohne,T.,etal.2007,“Onthenatureofclumpsindebrisdisks,”A&A,462,199–210

Kuchner,M.J.,&Holman,M.J.2003“TheGeometryofResonantSignaturesinDebrisDiskswithPlanets,”ApJ,588,1110–1120

Lay, O. P., Martin, S. R., & Hunyadi S. L. 2007, “Planet‐finding performance of the TPF‐IEmmaarchitecture,”Proc.SPIE,6693,66930A

Liou,J.C.,&Zook,H.A.1999,“SignaturesoftheGiantPlanetsImprintedontheEdgeworth‐KuiperBeltDustDisk,”AJ,118,580–590

Lisse,C.M.,Beichman,C.A.,Bryden,G.,&Wyatt,M.C.2007, “On theNatureof theDust intheDebrisDiskaroundHD69830,”ApJ,658,584–592

Liu,W.M.,Hinz,P.M.,Hoffmann,W.F.,etal.2004,“AdaptiveOpticsNullingInterferometricConstraintsontheMid‐InfraredExozodiacalDustEmissionaroundVega,”ApJ,610,125

Macintosh B., Graham, J., Palmer, D., et al. 2006, “The Gemini Planet Imager,” Proc. SPIE,6272,62720L

Metchev,S.A.,Hillenbrand,L.A.,&Meyer,M.R.2004,“TenMicronObservationsofNearbyYoungStars,”ApJ,600,435–450

Chapter5

162

Metchev,S.A.,Eisner, J.A.,Hillenbrand,L.A.,&Wolf,S.2005,“AdaptiveOpticsImagingofthe AUMicroscopii Circumstellar Disk: Evidence for Dynamical Evolution,” ApJ, 622,451–462

Morbidelli, A., Levison, H. F., Tsiganis, K., et al. 2005, “Chaotic capture of Jupiter's TrojanasteroidsintheearlySolarSystem,”Nature,435,462–465

Morbidelli,A.,Chambers,J.,Lunine,J.I.,etal.2000,“SourceregionsandtimescalesforthedeliveryofwatertoEarth,”MeteoriticsandPlanetaryScience,35,1309–1320

Moro‐Martín, A.,&Malhotra, R., 2002, “A Study of theDynamics ofDust from theKuiperBelt:SpatialDistributionandSpectralEnergyDistribution,”AJ,124,2305–2321

Olofsson,G.,Liseau,R.,Brandeker,A.2001, “WidespreadAtomicGasEmissionReveals theRotationoftheβPictorisDisk,”ApJ,563,770

Perrin,M.D.,Graham,J.R.,Kalas,P.,etal.2004,“LaserGuideStarAdaptiveOpticsImagingPolarimetryofHerbigAe/BeStars,”Science,303,1345–1348

Pinte,C.,Fouchet,L.,Ménard,F.,Gonzalez,J.‐F.,&Duchêne,G.2007,“OnthestratifieddustdistributionoftheGGTaucircumbinaryring,”A&A,469,963–971

Reche,R.,Beust,H.,Augereau,J.‐C.,etal.2008,“Ontheobservabilityofresonantstructuresinplanetesimaldisksduetoplanetarymigration,”A&A,480,551–561

Roberge,A.,Feldman,P.D.,Weinberger,A.J.,etal.2006,“StabilizationofthediskaroundβPictorisbyextremelycarbon‐richgas,”Nature,441,724–726

Schneider, G., Smith, B. A., Becklin, E. E., et al. 1999, “NICMOS Imaging of the HR 4796ACircumstellarDisk,”ApJ,513,L127–L130

Serabyn,E., Booth,A.J., Colavita,M.M. 2004, “The Keck interferometer nuller: systemarchitectureandlaboratoryperformance,”Proc.SPIE,5491,806

Stapelfeldt,K.R.,Holmes,E.K.,Chen,C.,etal.2004,“FirstLookattheFomalhautDebrisDiskwiththeSpitzerSpaceTelescope,”ApJSup.Ser.,154,458–462

Stapelfeldt, K. R., Trauger, J., Traub, W., et al. 2007, “First Steps in Direct Imaging ofPlanetary Systems like our Own: The Science Potential of 2‐m Class Optical SpaceTelescope,”WhitepapersubmittedtotheAAACExoplanetTaskForce,arXiv:0707.1886

Stark, C. C. & Kuchner,M. J. 2008, “The Detectability of Exo‐Earths and Super‐Earths ViaResonantSignaturesinExozodiacalClouds,”ApJ,686,637–648

Takeuchi, T., & Artymowicz, P. 2001, “Dust migration and morphology in optically thincircumstellardisks,”ApJ,557,990–1006

Thébault,P.,&Augereau,J.‐C.2007,“CollisionalProcessesandSizeDistributioninSpatiallyExtendedDebrisDiscs,”A&A,472,169–185

Thébault,P.,Augereau,J.C.,&Beust,H.2003,“Dustproductionfromcollisionsinextrasolarplanetarysystems.TheinnerbetaPictorisdisc,”A&A,408,775–788

Thi, W.‐F., van Dishoeck, E. F., Blake, G. A., et al. 2000, “H2 and CO Emission from DisksaroundTTauriandHerbigAePre‐Main‐SequenceStarsandfromDebrisDisksaroundYoungStars:WarmandColdCircumstellarGas,”ApJ,561,1074–1094

Trilling,D.E.,Bryden,G,Beichman,C.A.,etal.2008,“DebrisDisksaroundSun‐likeStars,”ApJ,674,1086–1105

ExozodiacalDisks

163

Tsiganis,K.,Gomes,R.,Morbidelli,A., etal.2005, “Originof theorbitalarchitectureof thegiantplanetsoftheSolarSystem,”Nature,435,459–461

Wallner, O., Ergenzinger, K., Johann, U. 2008, “Terrestrial exo‐planet science by nullinginterferometry: instrument design and scientific performance,” Proc. SPIE, 7013,70132O

Wilner,D.J.,Holman,M.J.,Kuchner,M.J.,etal.2002,“StructureintheDustyDebrisaroundVega,”ApJ,569,L115–L119

Wolszczan,A.,&Frail,D.A.1992, “APlanetarySystemaroundtheMillisecondPulsarPSR1257+12,”Nature,355,145–147

Wyatt,M. C. 2003, “Resonant Trapping of Planetesimals by PlanetMigration:DebrisDiskClumpsandVega'sSimilaritytotheSolarSystem,”ApJ,598,1321–1340

Wyatt, M. C., 2005, “The insignificance of P‐R drag in detectable extrasolar planetesimalbelts,”A&A,433,1007–1012

Wyatt,M.C., 2006, “Dust inResonantExtrasolarKuiperBelts:GrainSizeandWavelengthDependenceofDiskStructure,”ApJ,639,1153–1165

Zuckerman,B.,&Song,I.2004,“YoungStarsNeartheSun,”ARA&A,42,685–721

Microlensing

165

6 Microlensing

B.ScottGaudi,TheOhioStateUniversity,Chair

AndyBoden,NASAExoplanetScienceInstitute,CoChairJ.P.Beaulieu,DavidP.Bennett,EdwardCheng,KemCook,AndrewGould,JohnMather,CharleyNoecker,DomenickTenerelli

6.1 IntroductionMicrolensing isarelativenewcomeramongplanetdetectionmethods. Todateonlyeightplanets have been detected with microlensing (Bond et al. 2004; Udalski et al. 2005;Beaulieuet al. 2006;Gouldet al. 2006;Gaudi et al. 2008;Bennett et al. 2008;Donget al.2008), but these detections have already provided important constraints on planet‐formation theories. Thedetectionof twocool, “SuperEarth”planetsamong the first fourplanetssuggestthattheseplanetsarecommon(Beaulieuetal.2006;Gouldetal.2006).Thedetectionofa Jupiter/Saturnanalogalsosuggests thatSolarSystemanalogsareprobablynotrare(Gaudietal.2008).Thedetectionofalow‐massplanetarycompaniontoabrown‐

Figure61.(left)Exoplanetsdetectedviatransits(triangles),RV(circles),andmicrolensing(pentagons),asafunction of their mass and equilibrium temperature. Blue circles are planets found via RV that weresubsequently found to be transiting. Also shown are the approximate locations of the habitable zone (e.g.,Kasting et al. 1993) and the snow line (e.g., Kennedy andKenyon 2008; Lecar et al. 2006). (right) Ground‐based (red) and space‐based (purple)microlensing surveys are sensitive to planets above the curve in themassvs. semi‐major axisplane.Thegold, greenand cyan regions indicate the sensitivitiesof radial velocitysurveys, SIM andKepler, respectively. The location of our Solar System’s planets andmany exoplanets areindicated,withmicrolensingdiscoveriesinred.(Leftfigure,B.S.Gaudi,OhioStateUniversity;Rightfigure,D.P.Bennett,NotreDameUniversity)

Chapter6

166

dwarf star suggests that suchobjects can formplanetarysystemssimilar to thosearoundsolar‐typemain‐sequencestars(Bennettetal.2008).

Microlensing’sdisproportionatelylargecontributionstemsfromitsabilitytoproberegionsof parameter space that are currently inaccessible to other methods. In particular,microlensing ismostsensitive toplanets in thecold,outerregionsofsystemsbeyondthesnow line, the point in the protoplanetary disk exterior towhich the temperature is lessthanthecondensationtemperatureofwater(Lecaretal.2006;Kennedy&Kenyon2008).Giant planets are thought to form in the region immediately beyond this line,where thesurface density of solids is highest and thus planet formation is most efficient. It is alsothoughtthatthisregionislikelytobetheultimatesourceofthewaterforhabitableplanets.Inaddition,microlensingissensitivetoverylow‐massplanets,potentiallydowntothemassof Mars. Finally, microlensing is the only method capable of detecting old, free‐floatingplanets,hypothesizedtobeacommonby‐productofplanetformationandevolution.

Microlensing planet searches have likely saturated their planet detection rate in theircurrentform.Ontheotherhand,theyhavefarfromsaturatedtheirpotential.Realizingthispotential,however,willrequireafundamentallynew,next‐generationapproachtoground‐based searches, and ultimately will require a space‐based mission. In this Chapter wereview how the microlensing planet‐search method works, and then outline the sciencegoals foranext‐generationground‐basedsurveyandspacemissionandtherequirementsthatmust bemet to achieve these goals. We propose several observatory concepts thatcould meet these requirements, and finally we discuss the relatively few technologydevelopmentsthatareneededtorealizetheseconcepts. OurfinalrecommendationsechothosemadebytheExoplanetTaskForce.

SummaryofRecommendations

1. In the next 1–5 years, construct a small aperture (1–2 m) wide FOV (2–4 deg2)telescopeinSouthernAfricainordertocompleteanetworkofsimilartelescopesinthesouthernhemisphereandsoallowanext‐generationmicrolensingsurvey.Thissurveywoulddeterminethefrequencyofplanetsbeyondthesnowlineanddetect~10Earth‐massplanets ifeverymain‐sequencestar intheGalacticdiskandbulgehassuchaplanetintherange1.5–4AU.

2. In the next 5–10 years, build and launch a 1‐m class wide field‐of‐view spacetelescopethatcanimagethecentralGalacticbulgeinthenear‐IRoropticalalmostcontinuouslyforperiodsofatleastseveralmonthsatatime,foratleastfouryears.Such a space‐based satellite may be realized as an independent, stand‐aloneexoplanet mission, or a joint exoplanet/dark energymission. This survey woulddeterminethedemographicsofplanetswith0.5AU–∞,andmp≥MMars,determinethe frequencyof free‐floatingEarth‐massplanets, anddetermine the frequencyofEarth‐mass in theouterhabitable zoneof solar‐type stars in theGalacticdiskandbulge.

6.1.1 HowMicrolensingFindsPlanetsAmicrolensing event occurswhen a foreground “lens” happens to pass very close to ourlineofsighttoamoredistantbackground“source.” Thelensistypicallyamain‐sequencestar or stellar remnant in the foreground Galactic disk or bulge,whereas the source is amain‐sequencestarorgianttypicallyinthebulge.Thelenssplitsthesourceintotwoimageswhichhaveaseparationnearthetimeofclosestalignmentof~ 2θE ,wheretheEinsteinring

Microlensing

167

radiusisdefinedbyθE = κMπ rel ,whereM isthemassofthelens, π rel = AU(dl−1 − ds

−1) isthe lens‐source relative parallax, dl and ds are the distances to the lens and source, andκ= 4G / (c2 AU). Fortypicallensmassesandlensandsourcedistances,θE~500µas,andsotheimagesarenotresolved.However,thesourceismagnifiedbyanamountthatdependsonlyontheangularseparationofthelensandsourceinunitsofθE.

The transverse motion of the lens, source, and observer results in a time‐variablemagnification and gives rise to a smooth, symmetric microlensing event with acharacteristic form. Typically,observationsofamicrolensingeventcausedbyan isolatedlenscanfitbyasimplefive‐parametermodel.Threespecifythemagnificationasafunctionof time: theminimumangularseparationof the lensandsource inunitsofθE (the impactparameter, which also specifies the maximum magnification), the time of maximummagnification, and finally the Einstein timescale, tE =θE / µ , where µ is the relative lens‐source propermotion. The typical timescales for events toward the Galactic bulge are oforderamonth.Only tE containsanyinformationaboutthephysicalpropertiesofthelens,and then only in a degenerate combination of the lens mass, distance, and transversevelocity.However,asmentionedbelow,ithasproventobethecasethatitisoftenpossibleto obtain additional information, which partially or totally breaks this degeneracy. Theremainingtwoparametersrequiredtofitasingle‐lenseventarethefluxofthesourceandthefluxofanyunresolvedlightthatisnotbeinglensed.Thelattercanincludelightfromacompaniontothesource, lightfromunrelatednearbystars, lightfromacompaniontothelens,and(mostinterestingly)lightfromthelensitself.Inthosecaseswhereitispossibletoisolatethelenslight,thismeasurementcanbeusedtoconstrainthelensmass(Bennettetal.2007).

If the lens star happens to have a planetary companion, the companion can perturb theimages created by the primary, resulting in a short‐timescale deviation from the smooth,symmetricprimarymicrolensingevent (Mao&Paczynski1991;Gould&Loeb1992).ThedurationsoftheseperturbationsrangefromafewhoursforanEarth‐massplanettoadayfor Jovian‐mass planets. Since the planetmust perturb one of the twoprimary images inordertoyieldadetectabledeviation,andtheseimagesarealwayslocatedneartheEinsteinring radiuswhen the source is significantlymagnified, the sensitivity of themicrolensingmethod peaks for planet‐star separations of θEdl ~ 2.5AU(M / 0.3M⊙)

1/2 . Given a planetlocated within 0.6–1.6 of this optimal separation, also called the “lensing zone,” thedetectionprobabilities range from tens of percent for Jovianplanets to a fewpercent forEarth‐massplanets(Gould&Loeb1992;Peale2001;Bennett&Rhie1996).Theperturbedlight curves yield unambiguous planet parameters: for the great majority of events, thebasicplanetparametersthatdescribetheperturbation,theplanet/starmassratio,andtheplanet‐star separation, can be “read off” the planetary deviation (Gould & Loeb 1992;Bennett&Rhie1996).Possibleambiguitiesintheinterpretationofplanetarymicrolensingeventshavebeenstudied indetail (Gaudi&Gould1997;Gaudi1998;Han&Gaudi2008),andthesecanberesolvedwithgoodquality,continuouslightcurves.

Theuniquewayinwhichmicrolensingfindsplanetsleadstosomeusefulfeatures,mostofwhich canbeunderstood simplyas a resultof the fact thatplanetdetection relieson thedirect perturbationof imagesby the gravitational field of theplanet, rather thanon lightfromtheplanet,ortheindirecteffectoftheplanetontheparentstar.

Peak Sensitivity Beyond the Snow Line. The sensitivity of microlensing peaks atequilibriumtemperaturesof

Chapter6

168

Teq = 278K

LL⊙

1/4θEdlAU

−1/2

~ 70K MM⊙

where for the last relation we have assumed L∝M 5 , dl = 4 kpc , and ds = 8 kpc . Thus,microlensingismostsensitivetoplanetsinthecool,outerregionsbeyondthesnowline.

Microlensing is not sensitive to planets with separations very much smaller than theEinstein‐ring radius, as these can only perturb highly demagnified images. Thus,microlensing ismuch less sensitive toplanets in thehabitable zonesof theirparent stars(Park et al. 2006). Nevertheless, a space‐based microlensing survey would still detectdozensofhabitableterrestrialplanets,providedsuchplanetsarecommon(see§6.2.2andFigure6‐6)

Sensitivity toLowmassPlanets.Theamplitudesof theperturbationscausedbyplanetsare typically large, > 10%, Furthermore, although the durations of the perturbations getshorterwithplanetmass(asmp1/2)andtheprobabilityofdetectiondecreases(alsoroughlyasmp1/2),theamplitudeoftheperturbationsareindependentoftheplanetmass.Thisholdsuntilthezoneofinfluenceoftheplanet,whichhasasizeofordertheEinstein‐ringradiusofthe planet θp, is smaller than the angular size of the source θ* . When this happens, theperturbationissmoothedoverthesourcesize.Fortypicalparameters,θp~1µas(mp/M⊕)1/2,and for a turn‐off star in the bulge θ* ~ 1µas , this finite‐source suppression begins tobecome important for planets with the mass of the Earth, but it does not completelysuppress the perturbations until masses below that of Mars, for main‐sequence sources(Bennett&Rhie1996).

Sensitivity to LongPeriod and FreeFloating Planets. Since microlensing can‘instantaneously’ detect planetswithoutwaiting for a full orbital period, it is sensitive toplanetswithverylongperiods.AlthoughtheprobabilityofdetectingaplanetdecreasesforplanetswithseparationslargerthantheEinstein‐ringradiusbecausethemagnificationsoftheimagesdecline,itdoesnotdroptozero.Indeedsincemicrolensingisdirectlysensitiveto the planet mass, planets can be detected even without a primarymicrolensing event.Even free‐floating planets that are not bound to any host star are detectable in thisway.Microlensing is the only method that can detect old, free‐floating planets. A significantpopulation of such planets is a generic prediction ofmost planet‐formationmodels (e.g.,Goldreich et al. 2004), particularly those that invoke strong dynamical interactions toexplaintheobservedeccentricitydistributionofplanets(e.g.,Juric&Tremaine2008).

Sensitivity to Planets Throughout the Galaxy. Because microlensing does not rely onlight fromtheplanetorhoststar,planetscanbedetectedorbitingstarswithdistancesofseveralkiloparsecs.Thehoststarsprobedbymicrolensingaresimplyrepresentativeofthepopulationofmassiveobjectsalongthelineofsighttothebulgesources,weightedbythelensing probability. The lensing probability peaks for lens distances about halfway to thesources in the Galactic bulge, but remains substantial for lens distances in the rangedl =1− 8 kpc .

Sensitivity to Planets Orbiting a Wide Range of Host Stars. The planet‐detectionsensitivityofmicrolensing isweaklydependentonthehoststarmass,andhasessentiallynodependenceonthehoststarluminosity.Thus,microlensingisaboutequallysensitivetoplanetsorbitingstarsallalongthemainsequence,frombrowndwarfstothemain‐sequenceturn‐off,aswellasplanetsorbitingwhitedwarfs,neutronstars,andblackholes.

Microlensing

169

Sensitivity toMultiplePlanet Systems.Microlensing is also sensitive tomultiple‐planetsystems,particularly inhigh‐magnificationevents(Gaudietal.1998).This,alongwiththefact that high‐magnification events are potentially sensitive to very low‐mass planets,makesmicrolensinganexcellentprobeofplanetarysystems.

6.1.2 ScienceGoalsGround‐based microlensing planet surveys, in their current incarnation, have basicallysaturatedtheirdetectionrateatroughlyhalfadozenplanetsperyear. At thisrate, thesesurveyswillbeabletotesttherobustnessofthepreliminaryresults indicatedbythefirsteightdetections,andpushthelowermasslimitofmicrolensingdetectionsdowntoroughlythe mass of the Earth for a handful of rare, very high‐magnification events (Griest &Safizadeh 1998; Rattenbury et al. 2002; Dong et al. 2006). However, these surveys areunlikelytorealizethefullpotentialofthemicrolensingmethod,andareunlikelytoprovideenoughdetectionstoaccuratelydeterminethestatisticsofEarth‐massplanetsbeyondthesnowline,atleastinthenexttenyears.Rather,whatisneededisasurveythatissensitiveenough to detect tens of cool, Earth‐mass planets over the next~5 years, assuming suchplanetsarecommon.Thisrequiresafundamentallynewapproach.

Nextgenerationground‐basedmicrolensingplanetsearcheswilloperateinaverydifferentmode than the currentmodel, inwhich a large area of the bulge ismonitored by surveygroups to identifyongoingmicrolensingevents,whichare then individuallymonitoredbyfollow‐up collaborations to search for planetary perturbations. With a sufficiently largefield‐of‐view (FOV) of 2–4 deg2, it becomes possible to monitor tens of millions of starsevery 10–20 minutes, and so discover thousands of microlensing events per year.Furthermore, these would be simultaneously monitored to search for planetaryperturbations. In order to obtain round‐the‐clock coverage and so catch all of theperturbations,severalsuchtelescopeswouldbeneeded,locatedon3–4continentsroughlyevenly spread in longitude. Detailed simulations of such a next‐generation microlensingsurveyindicatethat,ifEarth‐massplanetswithsemimajoraxesofseveralAUarecommonaroundmain‐sequencestars,itshoulddetectseveralsuchplanetsperyear(Bennett2004).

Ultimately, with a space‐based survey, microlensing is capable of detecting planets withmass as small as that of Mars, with separations greater than ~0.5 AU, orbiting main‐sequence stars with distances between 1–8 kpc. Thus, in combination with Kepler,microlensingcanprovideanearlycompletedeterminationofthedemographicsofplanetsthroughout the Galaxy. This dataset would provide a stringent test of planet ‐ormationmodels.

The exoplanet demographics that a space‐based microlensing survey provides will alsoyielduniqueinsightintoplanetaryhabitability.Thesuitabilityofaplanetforlifedependsona number of factors, such as the average surface temperature, which determines if theplanetresidesinthehabitablezone.However,therearemanyotherfactorsthatalsomaybeimportant,suchasthepresenceofsufficientwaterandothervolatilecompoundsnecessaryfor life (Raymondetal.2004;Lissauer2007).Thus,a reasonableunderstandingofplanetformation isan important foundation for thesearch fornearbyhabitableplanetsand life.Althoughnot itsoptimalregimeofsensitivity,space‐basedmicrolensing isalsocapableofprobingtheouterhabitablezoneforGandKstars.

The particular capabilities of the microlensing method, when compared to the currentsampleofknownexoplanets,andtheexpectedyieldofexoplanetsusingcurrentandfuturedetectionmethods,andwhentakeninconsiderationwithcurrentideasabouttheformation

Chapter6

170

of planetary systems, lead to the following science goals for future microlensing planetsearchexperiments.Wedividetheseinto‘ground‐based’and‘space‐based’goals,asfuturegroundandspace‐basedexperimentsaredistinctinboththetimelinefortheircompletion,aswellastheirintrinsiccapabilities.

SummaryofScienceGoals

1. Ground‐basedSurveyScienceGoals(1–5years):

a. Determine the frequency of planets beyond the snow line. If every main‐sequencestar in theGalacticdiskandbulgehasaplanet in the range1.5–4AU,overthenext5yearsdetect10withmp=M⊕,100withmp=MNep,and500withmp=MJup.

b. Estimatethefrequencyoffree‐floating,Jupiter‐massplanets.IfthereexistsoneEarth‐mass free‐floatingplanetpermain‐sequence star,detectat least200ofthem.

2. Space‐basedSurveyScienceGoals(5–10years):

a. Determinethedemographicsofplanetswith0.5AU–∞,andmp≥MMars. IfeverystarhasananalogtoourSolarSystem,detect150terrestrialplanets(Earth/Venus/Mars), 5000 Jupiters, 1000 Saturns, and 130 ice giants(Uranus/Neptune).

b. Estimate the frequency of free‐floating, Earth‐mass planets. If there existonefree‐floatingplanetpermain‐sequencestar,detectatleast60ofthem.

c. Determine the frequency of Earth‐mass habitable planets (η⊕). Find 30habitableEarth‐massplanetsifthereisanaverageofoneplanetperstarinthehabitablezone.

d. Characterize host‐star masses and distances, and measure star‐planetseparationsandplanetmassesto<20%formostdetectedplanetaryevents.

6.1.3 ScienceRequirementsPlanet detection via microlensing has several general requirements. To start, themicrolensingeventsduetotheprimarystarsmust firstbediscovered. Microlensingisanintrinsicallyrarephenomenon,andthusdetectingmicrolensingeventsatasubstantialratepracticallyrequiresmonitoringstarsintheGalacticbulge,wherethedensityofsourceandlensstarsishigh.TowardtheGalacticbulge,themicrolensingeventrateisΓ~10–5perstarperyear.Next,theseeventsmustbemonitoredwithsufficientprecisionandcadencesuchthattheperturbationsduetotheplanetsofinterest(i.e.,Earth‐massplanets)aredetectableandwell characterized. Finally, a sufficientnumberof sucheventsmustbemonitored inorder to detect the planets of interest. Since the detection probability decreases withdecreasing planet mass, this implies that more events must be monitored for smallerplanets.

If n∗ stars are monitored with a given cadence and precision, the number of planetdetectionsperyearisapproximately

N ≈ n*ΓP ≈ 10 yr-1 n*

108

Γ

10−5 yr-1

P1%

.

Microlensing

171

ForEarth‐massplanetsandmain‐sequencesources,detectionprobabilitiesofP~1%canbe achieved with cadences of 20 minutes and ~1% photometry (Bennett & Rhie 1996).Thus,todetectEarth‐massplanets,roughly108starsmustbemonitoredontimescalesof20minutes with ~1% photometry. Detecting smaller planets requires monitoring a largernumberofsmallerstarswithhighercadenceandbetterphotometricprecision.

Thereareapproximately10deg2intheGalacticbulgeforwhichthemicrolensingeventrateis ≥10–5 yr–1, and the reddening is not extraordinarily high. In this region, there are onaverage107starspersquaredegreedowntoI~21,givingtherequired108main‐sequencesources in theusable 10deg2. Achieving a fewpercent photometry on starswith I~21requiresapertures (orequivalentapertures)of1–2m for reasonableexposure timesofafewminutes.Finally,awide field‐of‐viewof~2deg2 isneeded inorder tosimultaneouslymonitor a sufficient number of stars per pointing to achieve the required ~20 minutecadence.Wenotethereisconsiderableflexibilityintheserequirements.Forexample,itispossible to achieve the same planet detection rate by going deeper on a smaller field ofview,providedthattherequisitephotometrycanbeachievedonthefaintertargetstars.

Figure62.AcomparisonbetweenanimageofthesamestarfieldintheGalacticbulgefromCTIOin1”seeingandasimulatedMPFframe(basedonanHSTimage).Theindicatedstarisamicrolensedmainsequencesourcestar.(Bennett&Rhie2002)

Thefullpotentialofthemicrolensingtechniquecanonlybeachievedfromaspacemission.Ground‐basedmicrolensingsurveysare inherently limited in themassandseparationsoftheplanetstheycandetect.BecauseofthehighdensityofstarsintheGalacticbulge,main‐sequence source stars are not generally resolved in ground‐based images, as Figure 6‐2demonstrates. As a result, the photometry needed to detect Earth‐mass planets cangenerallyonlybeobtainedwhenthesourceissignificantlymagnified.This,inturn,impliesthat ground‐based microlensing is typically only sensitive to terrestrial planets locatedclosetotheEinsteinring(at~1.5–5AU).Furthermore,obtainingtheaccuratephotometryofthesmallerandfaintermain‐sequencestarsthatisrequiredtodetectplanetswithmassmuch smaller than the Earth is effectively impossible from the ground. Ground‐basedmicrolensing surveys also suffer losses indata coverageandqualitydue topoorweatherandseeing.Asaresult,asignificant fractionof theplanetarydeviationsseen inaground‐based microlensing survey will have poor light‐curve coverage and therefore poorlyconstrainedparameters(Peale2003).

Forallbutasmallfractionofplanetarymicrolensingevents,thedetectionoflightfromthehoststarisnecessarytoallowthestarandplanetmassesandseparationinphysicalunitsto

Chapter6

172

be determined. This can be accomplished with HST or ground‐based adaptive opticsobservationsforasmallnumberofplanetarymicrolensingevents(Bennettetal.2007),butspace‐based survey data will be needed for the detection of host stars for hundreds orthousandsofplanetarymicrolensingevents.

Sensitivity to terrestrial planets in all orbits from 0.5 AU to ∞, sensitivity to Mars‐massplanets,andfullcharacterizationofdetectedplanets,onlycomefromaspace‐basedsurvey.

SummaryofScienceRequirements

1. Ground‐basedSurvey:Monitor100millionmain‐sequencestarswithradius~Rin

theGalacticbulgewith2%precisionand20‐minutecadencecontinuously(weatherpermitting)duringthe8months/yearduringwhichtheGalacticbulgeisvisible,foratotaloffiveyears.

2. Space‐based Survey: Monitor 500 million main‐sequence stars with radius ~0.5–1R

with 1% precision and 15‐minute cadence continuously during the 9

months/year during which the Galactic bulge is > 45° from the Sun, for a totalmission lifetimeof fouryears.The resolutionmustbebetter than0.3” inorder toresolvethebrightestmain‐sequencestars.

6.2 ObservatoryConceptWe will discuss two basic observatory concepts. The first is a next‐generation ground‐basedsurvey,whichconsistsofa longitudinallydistributednetworkof1–2‐mclass,wide‐fieldtelescopes.Thesecondisaspace‐based1‐m‐classwide‐fieldimagingsatellite.Suchaspace‐basedsatellitemayberealizedasanindependent,stand‐aloneexoplanetmission,orajointexoplanet/darkenergymission.

6.2.1 Architectures

NextGenerationGroundBasedNetworkTo satisfy science requirement #1 from §6.1.3, a network of 1–2‐m class telescopesequippedwith large‐formatcameras is required. Cameraswith fields‐of‐viewof2–4deg2arerequired inordertomonitortensofmillionsofstarsspreadoutover~10deg2 inthebulgewithcadencesof10–20minutes.Inordertoobtainround‐the‐clockcoverageandsocatchallplanetaryperturbations,severalsuchtelescopeswouldbeneeded,locatedon3–4continentsroughlyevenlyspreadinlongitude.Withthishighercadenceandlargerfield‐of‐view, these telescopeswill find6000eventsperyear insteadof the600 found in currentsurveys.Furthermoretheseeventswillbeautomaticallymonitoredforplanets.Thesetwochangeswillyieldmorethana10‐fold increase in thenumberofeventsprobedandso inthenumberofplanetarydetections.

A next‐generation ground‐based microlensing planet search network is being‘spontaneously’assembledstepbystep.TheJapanese/NewZealandgroupMOAalreadyhasa2.2‐deg2camerainplaceontheir1.8‐mtelescopeinNewZealand,andcurrentlymonitorsabout 4 deg2 every10minutes,while covering amuchwider area everyhour. TheOGLEteam has funds from the Polish government to replace their current 0.4‐deg2 camera ontheir1.3‐mtelescopeinChilewitha1.7‐deg2camera.Whenfinished,theywillalsodenselymonitorseveraldeg2whilemonitoringamuchlargerareaoncepernight.

Microlensing

173

OnceOGLE has upgraded their camera, a network thatwill be able to satisfy the sciencerequirementsforanext‐generationground‐basedsurveywillbenearlyinplace,saveathirdtelescopetofillinthegapinlongitudinalcoveragebetweenChileandNewZealand.Indeed,astronomers from Korea, Germany, and China have considered initiatives, or havemadecomprehensive proposals to their governments, to acquire funding for a 1–2‐m classtelescopewithawideFOVinsouthernAfricaorAntarctica.Todate,noneoftheseinitiativeshavesecuredguaranteedfundingforsuchaproject.Byfundingathirdlegofthenetworkinsouthern Africa, the US has an excellent opportunity to play a pivotal role in thedevelopmentofaburgeoningfieldofexoplanetstudies,afieldinwhichitwillotherwisebeleftbehind. TheExoplanetTaskForcerecognizedthe importanceofcompleting thenext‐generationground‐basednetwork,andoneofitsrecommendationsoverthenextfiveyearswastofundthisthirdfacility.

SpaceBasedMicrolensingMissionThekeymissionrequirements foraspace‐basedmicrolensingmissionaresummarized inTable6‐1.Thebasic requirementsarea1‐mclasswide field‐of‐viewspace telescope thatcan image the central Galactic bulge in the near‐IR or optical almost continuously forperiodsofatleastseveralmonthsatatime.Thewidefield‐of‐viewisrequiredinordertomonitorasufficientnumberofstarswithasufficientcadencetodetecttheshort‐timescaleperturbations from low‐mass planets. The aperture is required both to achieve thephotometricprecisionof~1%on~0.5R

sourcestarsinthebulgerequiredtodetectMars‐

mass planets, aswell as to obtain the angular resolution needed to resolve the crowdedbulgefields.

The proposed MicrolensingPlanet Finder (MPF) mission isan example of a space‐basedmicrolensing survey that canaccomplish these objectives,essentially entirely with proventechnology, and at a cost of~$300M without the launchvehicle.MPFusesa1.1‐mThree‐Mirror Anastigmat (TMA)telescope feeding a 145 MpixelHgCdTefocalplaneresidingonastandard spacecraft bus asshown in Figure 6‐3. The MPFdesign leverages existinghardware and design concepts,many of which are alreadydemonstrated on‐orbit and/orflightqualified. Wedescribe thebasicsofMPFhere.Theexpectedperformance of a four‐yearmission with observing seasonsof9monthseach isdescribed in§6.2.2.

Property Value Units

LaunchVehicle 7920‐9.5 DeltaII

Orbit InclinedGEO28.7 degrees

MissionLifetime 4.0 years

TelescopeAperture 1.1 meters(diam.)

FieldofView 0.95×0.68 degrees

SpatialResolution 0.240 arcsec/pixel

PointingStability 0.048 arcsec

FocalPlaneFormat 145 megapixels

SpectralRange 600–1700 nmin3bands

QuantumEfficiency >75%>55%

900–1400nm700–1600nm

DarkCurrent <1 e‐/pixel/sec

ReadoutNoise <30 e‐/read

PhotometricAccuracy 1%orbetter atJ=20.5

DataRate 50.1 Mbits/sec

Table61.KeySpaceMissionRequirements

Chapter6

174

Orbit:MPFwillcontinuouslyobservetwo1.3sq.degreefieldsinthecentralGalacticbulgewithasamplingintervalof15minutesforeachfield.Thus,MPFrequiresanorbitallowingnearly continuous monitoring of its Galactic‐bulge target fields, interrupted only asnecessary, such as a 3‐month period every year to avoid the Sun. The optimal orbit is ageosynchronousorbitinclinedby28.7°totheEquatorand52°totheEcliptic.ThisallowsacontinuousviewoftheGalacticbulgeforallbut1–2dayspermonthpluscontinuousdatadownlinktoadedicatedgroundstationinWhiteSands,NM.Theorientationofthetelescopeis kept fixed for the first 4.5 months of the 9‐month observing season, and then, thespacecraft is rolled by 180° about the telescope bore‐sight. This prevents sunlight fromfallingonthe“darkside”of thetelescopewherethethermalradiatorsaremounted,sincetheMPFfieldisonly~5°fromtheEclipticplane.

Focal Plane Array (FPA): The MPF focal plane consists of 35 HgCdTe sensor chipassemblies(SCAs). EachSCAisa20482arrayof18‐µmpixels,andsothetotalFPAis147megapixels. The 140million pixels result in a peak raw data rate of 76Mbits·s–1 beforecompression.TheMPFdetectorsallowfornon‐destructivereadouts,whichrequirenodeadtime between exposures at the same pointing, and no shutter. The pixel size of 0.24”correspondstothediffractionlimitFWHMatawavelengthof~1.3µmandissufficienttoresolvemostofthebulgemain‐sequencestars.

Theplatescale is0.24”/pixel, so theFPAwillcoveracontinuousregionof0.95° × 0.68°, or 0.65 sq. degrees.Four pointings yield a total surveyareaof2.6sq.degrees. These fieldswill be selected to maximize themicrolensing event rate and planetdetectionsensitivity. Toachievetherequiredcadence,MPFwill take five33.67‐sec integrations at eachpointing in a 6 × 6 dither patternspanning 2 × 2 pixels for eachsubfield, spending about 3 minutesateachsubfield location.Thisditherpattern is required to order toprovide relative photometry withbetter than 1% precision as well asaccurate astrometry. Allowing forslewing to the next subfield andspacecraft settling, the four‐subfieldsampling rate for microlensingeventswouldbeabout15minutes.

ThebandpassandthesensorsarechosentomaximizethesensitivitytoMPF’stargetstars,which are reddened main‐sequence stars in the Galactic bulge. In order to achieve thedesiredprecisiononthetargetstars,>75%QEisrequiredinthewavelengthregime900–1400nmand>55%QEisrequiredfor700–1600nm.Thisrequirementhasalreadybeenachieved with current‐generation detectors, e.g., for the IR detectors developed forHST/WFC3.TheQEcurvesforthesedetectorsexhibithighQEacrossthefullMPF‐requiredbandpass. The QE operability is typically better than 99%, far exceeding the 75%requirement. These detectors also satisfy the requirements for the dark current and

Figure 63: MPF On‐Orbit Configuration. (Bennettetal.2008)

Microlensing

175

readout noise. Most images will be taken with the high throughput “clear” filter, whichspans 600–1700 nm; but at least once a day, images will be taken through the “visible”(600–900 nm) and “IR” (1300–1700 nm) filters for color determination. Colormeasurementswillhelptoidentifythestellartypeofboththesourceandlensstarsandaidindeterminingtheirproperties.

OpticalTelescopeAssembly (OTA): TheOTAoptical designofMPF is a 1.1‐mapertureconsisting of a three‐mirror anastigmat all‐reflective telescope configuration, providing alargefield‐of‐view(FOV),anddeliveringdiffractionlimitedopticalperformanceoverMPF’swidewavelengthrange. StarlightenteringtheOTAiscollectedbytheprimarymirrorandreflectedoffthesecondarymirrorthroughthecenter‐holeoftheprimarymirrorassemblyintotheaftopticsassembly,wheretheimageisredirectedandfocusedontotheFPA.

Microlensing/DarkEnergyMissionSynergy

The requirements for a space‐based microlensing mission can be accomplished by astandalonemissionatacostofabout$400M(includingthelaunchvehicle),astheexampleof the MPF mission shows; however the telescope requirements are very similar to therequirements foranumberof theproposeddarkenergymissions.Thus, it isattractive toconsider a combineddark energy/planet findingmission that couldbe accomplished at asubstantial savings compared to doing each mission separately. Although detailedsimulationsoftheexpectedplanet‐detectionsensitivityofthevariousproposedjointdark‐energy missions (JDEM) have not been done, several space microlensing planet searchsimulation codes are available (Bennett & Rhie 2002; Gaudi, unpublished) that could beeasilyadaptedtoaccomplishthis.

The technical requirements of a space‐based microlensing mission can be compared tothoseofthethreeJDEMconceptscurrentlyunderstudy,ADEPT,Destiny,andSNAP,aswellas the two dark‐energymissions under study by ESA, DUNE and SPACE. SinceMPF is awell‐exploredexampleofaspacemissionwhosetechnicalspecificationscanaccomplishthescience objective of a space‐based microlensing mission, it serves as a useful point ofreference. All of these JDEM missions are wide field‐of‐view space telescopes with anaperturelargerthanMPF’s1.1m,andallincludenear‐IRobservations,althoughsomealsohave an optical‐imaging capability. As far aswe know, thesemissions have pointing andstabilityrequirementsthatareatleastasstringentasthoseforMPF,andmostalsorequirehigherangularresolutionthanMPF.TheyallproposeanEarth‐SunL2orbit,whichpermitscontinuous viewing of theMPF Galactic bulge field, although aswe explain below,MPF’sinclinedgeosynchronousorbitmightbebetterforbothapplications.

TheDUNEmission,whichisnowunderstudyforESA’sCosmicVisionsprogram,isprobablythe most similar to MPF. In fact, it is so similar that it satisfies all of the sciencerequirements for a space‐basedmission and so canmeet all of the science goals,withoutmodification to the instrument, telescope or spacecraft, provided that the planet‐searchprogram is givena similar timeallocationas theMPFmission.TheDUNEmission isnowbeingcombinedwithSPACEtoformanewmissioncalledEuclid,butEuclidisexpectedtomaintainthesamecapabilitiesrelevanttoamicrolensingplanetsearchasDUNE.

While the hardware requirements for a space‐based microlensing planet search missionand a dark‐energy mission are similar, there are a number of differences in the preciserequirements for each program. In most cases, accommodating both missions willnecessitateadditionalrequirementsbeyondwhatisneededforeachalone,andthisimpliesthat the costwill necessarilybehigher.Wehavenotperformedadetailed trade study in

Chapter6

176

ordertodesignamissionthatisoptimizedforbothscienceapplications,butwelistherethemajorissuesthatsuchatradestudywouldaddress:

1. Observing time.Theplanet‐searchmissionmustobserveverydense fields in theGalactic bulge, which are not useful for dark‐energy science. Thus, the sameobservingfieldscannotbeusedforbothmissions,soajointmissionwouldneedtobedesigned fora longer lifetime thanwouldberequired foronly thedark‐energyscience. Althoughsomeprogresscouldbemadewithasmallamountof telescopetime, detecting habitable planets will require a time commitment at the levelenvisioned for a dedicated space‐based survey, of order four 9‐month seasons.Because the detector requirements for a microlensing survey are generally lessstringent than that for dark‐energy missions, it may be possible to perform themicrolensing surveyafter thedetectorshavedegraded tobelow the requirementsforthedark‐energycomponentofthemission.

2. Pass Bands. The microlensing planet‐search program benefits from detecting asmanyphotonsfromasmanyGalacticbulgestarsaspossible.Duetotheextinctionbydustintheforegroundofthebulge,themicrolenstargetstarsarebrightestinthenear infrared, at wavelengths of 1–1.5 µm. Thus, the microlensing planet‐searchprogramprefersthatmostobservationsbemadewithanextremelywidepassbandcentered in the near‐IR. There is also a requirement that a few percent of themicrolensing observations be made in narrower passbands, although the precisepassbandstobeusedarenotcritical.Thedesireforaverywidepassbandisnotahard and fast requirement, as the losses from using a narrower passband can bemadeupwithalarger(>1.1m)aperturetelescope,ormoreobservingtime.

3. TelescopeFieldofViewandDetectors.TheproposedMPFtelescopecanimageafieldof0.65deg2with3520482HgCdTedetectorsinasingleexposure.Someoftheproposed JDEMmissions have a smaller field‐of‐view. However, the dark‐currentandQErequirements forthemicrolensingplanetsearchmissionareprobably lessstringentthanforJDEM,soitmightbepossibletoaddsomelowergradedetectorstothefocalplane.

4. Orbit.Most JDEMmissionspropose anEarth‐SunL2orbit,whereas theproposedorbitforMPFisaninclinedgeosynchronousorbit.Thelatterorbithastheadvantageof a continuous ground‐link to a dedicated ground station, which makes datatransmissionmucheasierand lessexpensive.Ourpreliminarystudies for theMPFmission indicate that the adverse effects of the trapped electron radiation in thisorbitcanberemovedbyacarefulshieldingdesign.IftheeffectsofscatteredEarth‐shinecanalsobeminimizedbya carefulbaffledesign, itmightbepossible tousethisorbitforthecombinedmission.

5. DataDownlink.Thedatadownlinkrequirementsforamicrolensingplanet‐searchmissionarelikelylargerforthemicrolensingplanetsearchprogramthanforJDEM,duetoshorterexposuresand/oralargerfocalplane.IftheinclinedgeosynchronousorbitproposedforMPFprovestobeunsuitableforthedarkenergymission,usingacosmic ray removal and compression scheme could reduce the data rate for theplanetsearchobservations.

6. Telescope Aperture, PSF Sharpness, Pointing Stability. It is expected that thedark‐energysciencewillimposetighterrequirementsontheseparametersthantheplanetsearchprogram.

Microlensing

177

6.2.2 Performance

GroundBasedNextGenerationNetworkDetailed simulations of ground‐based next‐generation microlensing surveys have beenperformed by several groups (Bennett 2004; Gaudi, unpublished). These simulationsincludemodelsfortheGalacticpopulationoflensesandsourcesthatmatchallconstraints,and account for real‐world effects such as weather, variable seeing, photometrysystematics,Moonandskybackground,andcrowded fields.While these two independentsimulationsdifferindetail,theyreachsimilarconclusions.Suchasurveywouldincreasetheplanetdetectionrateatfixedmassbyatleastanorderofmagnitudeovercurrentsurveys.Figure 6‐4 shows the predictions of one of these simulations for the detection rate ofplanets of various masses, assuming every main‐sequence star has a planet within thespecifiedrangeofseparations. Inparticular, ifEarth‐massplanetswithsemimajoraxesof1.5–5AUarecommonaroundmain‐sequencestars,anextgenerationmicrolensingsurveyshould detect several such planets per year. In four years, such a survey would detectJupiter‐massfree‐floatingplanetsatarateofhundredsperyearifeverystarhasejectedaJupiter‐massplanet.

As mentioned previously, microlensing light curves routinely yield the planet/star massratio and projected separation in units of θE . Space‐based imaging or adaptive optics isneeded to detect the flux of the planetary host stars, which allows the star and planetmassesandseparationinphysicalunitstobedetermined(Bennettetal.2007).Thiscanbeaccomplished for a small number of the most interesting planetary events usingHST oradaptiveopticsimaging.

Figure64.Theexpectednumberofplanetdiscoveriesasafunctionoftheplanetmassifeverymain‐sequencestarhasasingleplanetwithinthespecifiedrangeofseparationsforanext‐generationground‐basedsurvey,anda space‐based survey. The blue curves indicate the sensitivity of a space‐based survey, and the grey curvesrepresentanext‐generationground‐basedprogram.(D.P.Bennett,NotreDameUniversity)

Chapter6

178

Figure 65. The expected number of free‐floating planet discoveries for a space‐basedmicrolensing survey.(D.P.Bennett,NotreDameUniversity)

SpaceBasedMissionBennett&Rhie(2002)andGaudi(unpublished)haveindependentlysimulatedspace‐basedmicrolensing surveys. These simulations included variations in the assumed missioncapabilities that allow us to explore how changes in the mission design will affect thescientificoutput,andtheyformthebasisofourpredictions.Figures6‐4and6‐5showtheexpected number of planets that MPF would detect at three different ranges of orbitalseparations,andat∞(i.e.free‐floatingplanets)assumingonesuchplanetperstar.

Figure6‐6showsthesensitivityofMPFforplanetsofvariousmassesandfixedequivalentorbitalradiusa/L∗0.5,whereaisthesemimajoraxisandL∗isthehoststarluminosity.Thiscorrespondstoafixedequilibriumtemperatureforaconstantalbedo.Thus,theboundariesof thehabitablezone(HZ)are locatedatroughlyconstantvaluesof theequivalentorbitalradiusforhoststarsofdifferentmass.AsFigure6‐6shows,whilebothMPFandKeplerhavesignificantsensitivityintheHZ,MPF’ssensitivitypeaksforplanetsoutsidetheHZ,whereasKepler ismostsensitivetoplanetsinteriortotheHZ. Thus,bycombiningtheresultsfromthesetwomissions,itwillbepossibletorobustlyinterpolatethefrequencyofplanetsintheHZ,evenifthisfrequencyshouldturnouttobesmall.

Aspace‐basedmicrolensingsurveywillprovidenearlycompletedemographicsofplanetarysystems.Thesensitivityofthemicrolensingmethodextendsdowntoverylow‐massplanetswith<0.1M⊕ ,asshowninFigure6‐4.ItisonlyplanetswellinsidetheEinstein‐ringradiusthat are missed because the stellar lens images that would be perturbed by these innerplanetshaveverylowmagnifications,andsocontributelittletothetotalbrightness.ThesefeaturescanbeseeninFigure6‐1,whichcomparesthesensitivityoftheMPFmissionwithexpectationsforotherplannedandcurrentprograms.Otherongoingandplannedprogramscan detect, at most, analogs of two of the Solar System’s planets, while a space‐basedmicrolensingsurveycandetectseven—allbutMercury.Theonlymethodwithcomparablesensitivity to MPF is the Kepler space‐based transit survey, which complements themicrolensingmethodwithsensitivityatsemi‐majoraxes<1AU.

Microlensing

179

Aspace‐basedsurveywillalsodetectthefluxfrommostoftheplanetaryhoststars,whichgenerallyallowsthehost‐starmass,approximatespectraltype,andtheplanetarymassandseparationtobedetermined(Bennettetal.2007).Thedistancetotheplanetarysystemisdeterminedwhenthehoststarisidentified,soaspace‐basedmicrolensingsurveywillalsomeasurehowthepropertiesofexoplanetsystemschangeasafunctionofdistancefromtheGalacticCenter.Thereisusuallysomeredundancyinthemeasurementsthatdeterminesthepropertiesofthehoststars,andsothedeterminationisrobusttocomplicatingfactors,suchasabinarycompaniontothebackgroundsourcestar.Figure6‐7showsthedistributionofplanetaryhost‐starmassesandthepredicteduncertaintiesinthemassesandseparationoftheplanetsand theirhost stars fromsimulationsof theMPFmission.Thehost starswithmassesdeterminedtobetterthan20%areindicatedbytheredhistograminFigure6‐7(a),andtheseareprimarilythehoststarsthatcanbedetectedinMPFimages.

EffectsofDescopesonExpectedPerformanceUnlikemost other planet‐search techniques,microlensing does not rely on flux from thehost star, and as a result the expected yields of ground and spaced‐based microlensingexperiments aremuch less sensitive todescopes. For experiments that rely ondetectinglightfromthehost,thenumberofdetectionsisarelativelysteepfunctionofsignal‐to‐noiseratio,typically∝(S/N)–3,becausethenumberofobjectsinthesampleis∝(flux)–3/2whereasthe S/N∝(flux)1/2. For microlensing, on the other hand, what matters is the brightnessdistributionof the sources,whichare in theGalacticbulgeand soat fixeddistance. As aresult,thescalingofthenumberofdetectionsasafunctionofsignal‐to‐noiseratioismuchshallower,typically∝(S/N)–1(Bennett&Rhie2002).Thus,theexpectationsfortheyieldsof

Figure66:MPFsensitivity(ontheright)iscomparedtothesensitivityofKepler(ontheleft)inthemassvs.equivalentorbitalradiusplaneintwofigurestakenfromtheExoplanetTaskForceReport.Theequivalentorbital radius is the orbital radius (inAU) normalized to the square root of the stellar luminosity, so theHabitableZone (HZ)hasanequivalentorbital radiusof~1.MPFandKepler aremost sensitive toplanetsexteriorandinteriortotheHZ,respectively,butbothMPFandKeplerhavesomesensitivityintheHZ.TheKeplerplotassumesanextendedmission,whereasonlythestandardmissionlifetimeof4yrisassumedforMPF.(B.Macintosh,LawrenceLivermoreNationalLaboratory)

Chapter6

180

microlensingexperimentsarefairlyrobusttodescopes.Foraspace‐basedmissionsuchasMPF,detailedsimulationsindicatethatthenumberofdiscoveriesscalesas

D(FOV)0.6(PSF)–0.26T,

where D is the diameter of the telescope, FOV is the area of the field of view, PSF is thewidthofthepointspreadfunction,andTistheobservingtime(Bennett&Rhie2002).

6.3 TechnologyBecausethemicrolensingplanet‐searchtechniqueisoperationallyrelativelysimple,inthatitrequiresonlyreasonablyprecise(~1%)relativephotometryforalargenumberofstarsover a large field‐of‐view, next‐generation ground‐ and space‐based microlensing planetsearchesrequirealmostnotechnologydevelopment.

Anext‐generationground‐basedsurveyrequiresasingle,1–2‐mclasstelescopewithawideFOV camera (2–4 deg2). Such a system, while not trivial to construct, is certainly notunprecedented. Indeed, theMOA‐II telescope(1.8m,2.18deg2)hasmanyof the featuresthatarerequired. Thisparticulardesign isnotsuitable foraSouthAfricansite,however,because the plate scale is too course. The design of the Pan‐STARSS‐1 telescope is quitewellsuitedforthispurpose. SubstantialsavingsoverthecostofthePS‐1systemcouldbeachievedbyusingtraditionalCCDdetectorsratherthanOrthogonalTransferArrays.

Aspace‐basedmicrolensingsurveyoffersafewadditionalminortechnologicalchallenges;however, these are all manageable. In particular, the MPF design leverages existinghardwareanddesign concepts,manyofwhichare alreadydemonstratedon‐orbit and/orflightqualified.Indeed,essentiallyallelementsareatTechnologyReadinessLevel(TRL)6orbetter.

Spacecraft:

The Spitzer spacecraft, which was developed by Lockheed Martin and has been in flightsinceAugust2003,formsthearchitecturalbasisfortheMPFspacecraftdesign.TheSpitzerspacecrafthasdemonstratedperformancethatmeetstherequirementsforMPF.

Figure 67. (a) The distribution of stellar masses for stars with detected terrestrial planets from asimulationofaspace‐basedmicrolensingsurvey.Theredhistogramindicatesthesubsetofthisdistributionforwhich themasses can be determined to better than 20%. (b) The distribution of uncertainties in theprojected star‐planet separation. (c) The distribution of uncertainties in the star and planet masses.(Bennett,Anderson,&Gaudi2007)

Microlensing

181

OpticalTelescopeAssembly:

The1.1‐mthree‐mirroranastigmatOTAdesignforMPFissimilartothatfortheNextViewprogram. NextView is a next‐generation commercial remote sensing imaging instrumentcurrentlyunderdevelopmentbyITTasasuccessortoITT‐builtIKONOS,andissponsoredbytheNationalGeospatial‐IntelligenceAgency(NGA).Infact,muchoftheNextViewdesignheritagecanappliedtodirectlyMPF.

FocalPlaneArray:

Thefocalplanearray(FPA)isarguablythemostchallengingtechnologicalaspectofaspace‐based microlensing mission. Fortunately, nearly all the needed technology has beendevelopedforothermissions.Forexample,theMPFmissionextensivelyleveragesexistingdetectorandpackagingtechnologydevelopedforJDEM,WPF3,andJWST.

Rockwell Scientific (RSC) has developed infrared HgCdTe detectors, which meet therequirementsofaspace‐basedmicrolensingmission. RSChasdeliveredover1501‐Kand2‐Ksensor‐chipassemblies(SCAs),andinparticular1‐Kand2‐KclassSCAsforspaceflightmissions.Most have detector requirements that are farmore stringent than required forMPF.Also,1.7‐µmcutoff2‐KSCAsoftheexactformatneededforMPFhavebeenproducedfor the JDEM/SNAP project evaluation. The 5.0‐µm cutoff 2‐K SCAs forNIRSpec on JWSThavecompletedflightqualification(TRL6).BoththeJDEM/SNAPandJWSTSCAsmeettheMPFrequirementsforthequantumefficiencyandwavelengthcoverage,darkcurrent,andnoiseproperties(seeTable6‐1).

6.3.1 FutureMilestonesAswehavejustdiscussed,aspace‐basedmicrolensingmissionrequirestechnologythathasbeen either already been demonstrated in‐orbit, or is currently in an advanced state ofdevelopmentforupcomingmissions.FortheMPFmissionspecifically,essentiallyalloftheelementsofthedesignareatTRL6orbetter.Thus,nofuturetechnologymilestonesneedtobeachievedbeforeaspace‐basedmicrolensingmissioncanbelaunched.

TheoneminorexceptiontothisistheFPApackaging.Thekeyissuestodemonstratehereare the cryogenic stability, temperature uniformity, and alignment. The path forward todemonstratingtheseandachievingTRL6isstraightforward. Indeed,apathfinderversionof a FPA for the MPF mission has already been assembled in order to demonstrate theassemblymethodology,andtheprimarycomponentshaveallsuccessfullyundergoneeitherflight qualification tests and/or on‐orbit flight operations on structures of similarcomplexity. Similar largemosaic FPAs are currentlyunderdevelopment forboth ground‐andspace‐basedastronomicalinstruments.

6.4 Research&AnalysisGoals6.4.1 DetailedPredictionsforaNextGenerationGround

BasedSurveyAlthoughrealistic simulationshavebeenperformed toestimate theexpected returnsofanext‐generationground‐basedmicrolensing surveyusing anetworkof ‘generic’ 2‐m‐classwide FOV telescopes, simulations have not been performed that consider the actualcharacteristicsoftheMOA‐IItelescopeandupcomingOGLEIVtelescopes,whichgenerallyhave smaller apertures and/or FOVs than those adopted in the simulations. Because the

Chapter6

182

expectednumberofdetectionsdegradesgracefully for less capableexperiments, it is stillexpected,andsomelimitedsimulationsindicate,thatasurveywiththeseinstrumentswillyieldinterestingscientificreturnswiththeadditionofathirdtelescopeinsouthernAfrica.Nevertheless, it is important to carry these simulations through in order to provide aquantitativeassessmentofthisassertion;thiswill,inturn,helptoinformtherequirementsforthethirdtelescope.

6.4.2 SelectionofCandidateFieldsforFutureSurveysAlthough the currentmicrolensing surveys (OGLEandMOA)havemadeefforts to choosethosefieldswiththehighestrateofobservablemicrolensingevents,aspace‐basedsurveywill generally be going much deeper and so will be operating in a somewhat differentregime. Therefore, it will be important to characterize the potential survey fields usingground‐basedobservations.OGLEandMOAcanmonitorcandidatefields,inordertoselectthefieldswiththehighestrateoffainterevents.Thecandidatefieldsshouldalsobeimagedat thehighest possible resolutionwithoneof the large telescopes located inChile, in theoptical and near‐IR. These images will then provide an accurate map of the stellarpopulations, deep luminosity function, reddening, etc.,withmulticolor characterizationofeach potential microlensing source star. They will also provide accurate multibandphotometriccalibrationsofthousandsofrelativelyuncrowdedstandardstarsinourfields,allowingaccuratecalibrationoftheverywidepassbandsusedinthespaceimageswiththedatafromtheprimarysciencefields.

6.4.3 DetailedTradeStudyforaJointDarkEnergy/MicrolensingMission

Thesynergybetweenadark‐energymissionandaspace‐basedmicrolensingplanetsurveyis quite compelling. A Joint Dark Energy/Microlensing Mission may offer a substantialsavings over stand‐alonemissions, while still allowing both components to achieve theirsciencegoals.However, asoutlinedabove, theremaybe somechallenges to joining thesemissions,andsoadetailedtradestudyneedstobedonetoassessthepotentialimpactsontheexpectedperformance.

6.5 ContributorsJ.P.Beaulieu,Institutd'AstrophysiquedeParis

DavidP.Bennett,UniversityofNotreDame

AndrewBoden,NASAExoplanetScienceInstitute

EdwardCheng,ConceptualAnalytics

KemCook,LawrenceLivermoreNationalLaboratory

B.ScottGaudi,TheOhioStateUniversity

AndrewGould,TheOhioStateUniversity

JohnMather,NASA


DomenickTenerelli,Lockheed

Microlensing

183

6.6 ReferencesBennett, D.P. 2004, “The Detection of Terrestrial Planets via Gravitational Microlensing:

Space vs. Ground‐based Surveys,” Extrasolar Planets: Today and Tomorrow, ASPConference Proceedings, Vol. 321, Eds. J.‐P. Beaulieu, A. Lecavelier des Etangs & C.Terquem,pp.59–67

Bennett, D. P., Anderson, J., & Gaudi, B. S. 2007, “Characterization of GravitationalMicrolensingPlanetaryHostStars,”ApJ,660,781–790

Bennett,D.P.,Bond, I.A.,Udalski,A.,etal.2008,“ALow‐MassPlanetwithaPossibleSub‐Stellar‐MassHostinMicrolensingEventMOA‐2007‐BLG‐192,”ApJ,684,663–683

Bennett, D. P., & Rhie, S. H. 1996, “Detecting Earth‐Mass Planets with GravitationalMicrolensing,”ApJ,472,660–664

Bennett, D. P., & Rhie, S. H. 2002, “Simulation of a Space‐Based Microlensing Survey forTerrestrialExtra‐SolarPlanets,”ApJ,574,985–1003

Beaulieu, J.‐P., Bennett,D. P.; Fouqué,P. 2006, “Discovery of a cool planet of 5.5 Earthmassesthroughgravitationalmicrolensing,”Nature,439,437–440

Bond, I.A.,Udalski,A., Jaroszynski,M., et al.2004, “OGLE2003‐BLG‐235/MOA2003‐BLG‐53:APlanetaryMicrolensingEvent,”ApJL,606,155–158

Dong, S., DePoy, D., Gaudi, B. S., et al. 2006, “Planetary Detection Efficiency of theMagnification3000MicrolensingEventOGLE‐2004‐BLG‐343,”ApJ,642,842–860

Dong, S., Bond, I. A., Gould, A., et al. 2008, “Microlensing Event MOA‐2007‐BLG‐400:Exhuming the Buried Signature of a Cool, Jovian‐Mass Planet,” ApJ, submitted,arXiv:0809.2997

Gaudi, B. S. 1998, “Distinguishing Between Binary‐Source and Planetary MicrolensingPerturbations,”ApJ,506,533–539

Gaudi,B.S.,Bennett,D.P.,Udalski,A.,etal.2008,“DiscoveryofaJupiter/SaturnAnalogwithGravitationalMicrolensing,”Science,319,927–930

Gaudi,B.S.,&Gould,A.,1997,“PlanetParametersinMicrolensingEvents,”ApJ,486,85–99

Gaudi.B. S.,Naber,R.M.,&Sackett,P.D.1998, “MicrolensingbyMultiplePlanets inHigh‐MagnificationEvents,“ApJL,502,33–37

Goldreich,P.,Lithwich,Y.,&Sari,R.2004,“FinalStagesofPlanetFormation,”ApJ,614,497–507

Gould, A. & Loeb, A. 1992, “Discovering planetary systems through gravitationalmicrolenses,”ApJ,396,104–114

Gould,A.,Udalski,A.,An, J.,etal.2006, “MicrolensOGLE‐2005‐BLG‐169ImpliesThatCoolNeptune‐likePlanetsAreCommon,”ApJL,644,37–40

Griest, K., & Safizadeh, N. 1998, “The Use of High‐Magnification Microlensing Events inDiscoveringExtrasolarPlanets,”ApJ,500,37–50

Han, C., & Gaudi, B. S. 2008, “A Characteristic Planetary Feature in Double‐Peaked, High‐MagnificationMicrolensingEvents,”ApJ,689,53–58

Chapter6

184

Hough,J.,Lucas,P.W.,Bailey,J.2006,“Thepolarizationsignatureofextra‐solarplanets,”inThe Scientific Requirements of Extremely Large Telescopes, Proc. IAU Symp. 232,CambridgeUniversityPress,pp.350–355

Juric, M., & Tremaine, S. 2008, “Dynamical Origin of Extrasolar Planet EccentricityDistribution,”ApJ,686,603–620

Kasting, J. F., Whitmire, D. P., & Reynolds, R. T. 1993, “Habitable Zones around MainSequenceStars,”Icarus,101,108–128

Kennedy,G.M.,&Kenyon,S.J.2008,“PlanetFormationaroundStarsofVariousMasses:TheSnowLineandtheFrequencyofGiantPlanets,”ApJ,673,502–512

Lecar, M., Podolak,M., Sasselov,D., et al. 2006, “On the Location of the Snow Line in aProtoplanetaryDisk,”ApJ,640,1115–1118

Lissauer, J. 2007, “Planets Formed in Habitable Zones of M Dwarf Stars Probably AreDeficientinVolatiles,”ApJL,660,149–152

Mao, S. & Paczynski, B. 1991, “Gravitational microlensing by double stars and planetarysystems,”ApJL,374,37–40

Park,B.‐G.,Jeon,Y.‐B.;Lee,C.‐U.,etal.2006,“MicrolensingSensitivitytoEarth‐MassPlanetsintheHabitableZone,”ApJ,643,1233–1238

Peale, S. 2001, “Probability of Detecting a Planetary Companion during a MicrolensingEvent,”ApJ,552,889–911

Peale, S. 2003, “Comparison of a Ground‐based Microlensing Search for Planets with aSearchfromSpace,”AJ,126,1595–1603

Rattenbury, N. J., Bond, I. A., Skuljan, J., et al. 2002, “Planetary microlensing at highmagnification,”MNRAS,355,159–169

Raymond,S.N.,Quinn,T.,&Lunine,J.I.2004,“MakingotherEarths:dynamicalsimulationsofterrestrialplanetformationandwaterdelivery,”Icarus,168,1–17

Udalski,A.,Jaroszynski,M.,Paczynski,B.,etal.2005,“AJovian‐MassPlanetinMicrolensingEventOGLE‐2005‐BLG‐071,”ApJL,628,109–112

RadialVelocity

185

7 RadialVelocity

GuillermoTorres,SmithsonianAstrophysicalObservatory,Chair

DawnM.Gelino,NASAExoplanetScienceInstitute,CoChairPaulButler,WilliamCochran,DebraFischer,JianGe,NaderHaghighipour,ScottHorner,StephenKane,DavidLatham,GregoryLaughlin,JamesLloyd,ChristopheLovis,GeoffMarcy,MichelMayor,ChrisMcCarthy,FrancescoPepe,DidierQueloz,JeanSchneider,StephaneUdry,StevenVogt,KasparvonBraun

7.1 IntroductionHighprecisionstellarradialvelocity(RV)measurementshaveproventobeaspectacularlysuccessfultechniqueforthedetectionofextrasolarplanetarysystems. ThefirstmajorRVsearchwasconductedonthe3.6‐mCanada‐France‐HawaiiTelescope(CFHT)byCampbell&Walker(1979),usingaHFgasabsorptioncellasavelocitymetric.Thispioneeringsurveycametantalizinglyclosetodetectingseveralplanets.ThisCFHTsurvey,alongwithseveralotherearlysearches,wasbasedontheassumptionthatallplanetarysystemswouldhaveanarchitecture similar to our own Solar System. The inner portions would have rockyterrestrial‐massplanets,andgasgiantswouldbefoundintheouterregionsbeyond3–5AU.Thus, whenMayor & Queloz (1995) found a candidate around a solar‐type star, the halfJupiter‐massplanetina4.2‐dayorbitaround51Peg,thediscoverywasmetbyconsiderableastonishment that a Jovian planet could be so close to its parent star. This planet wasquickly followedby thediscoveryofseveralothersorbitingmain‐sequencestars, some insurprisinglyeccentricorbits.Theseearlyexoplanetswerealldiscoveredbypreciseradial‐velocitymeasurements,andthismethodisnowresponsiblefortheinitialdetectionofover80% of the nearly 300 exoplanets known today. As described in this chapter, furtherprogress (particularly toward the detection of Earth‐mass planets) will depend uponimproving the velocity precision to well under 1 m·s–1, and substantially increasing theavailabilityoftelescopetime.

Stellar radial‐velocitymeasurementsdetect the reflexorbitalmotionof a star around thestar‐planetbarycenter. Thesemi‐amplitudeofthestellarradialvelocitysignalKresultingfromanexoplanetarysystemisgivenby

m·s–1,(1)

whereMpisinJupitermasses,M∗isinsolarmasses,Pisinyears,andtheorbitisassumedtobecircular.Fromthisequation,weseeimmediatelythatradialvelocitydetectiongivesthelargestsignal formassiveplanets inshortperiodorbits. Thus, inretrospect, itshouldnothavebeenmuchofasurprisethat51Peghasa4.2‐dayorbitalperiodandMpsiniofabouthalf of a Jupitermass. Its velocity semi‐amplitude isK= 59m s–1,which is significantly

Chapter7

186

largerthanMayor&Queloz’svelocitymeasurementprecisionof13ms–1. Incontrast,theradialvelocitysignaloftheSunduetoJupiter’sorbitisonlyabout13ms–1. Wenotealsothat the radial velocity technique is sensitive toMp sin i, rather thanMp. Thus,RVs allowonlytomeasurealowerlimittothemassofaplanet,unlesstheinclinationiofthesystemcan be determined by some independent means. Shortly after the first planets werediscovered,thissin iambiguity ledsometosuggestthatnoneoftheseobjectswerereallyplanets, but instead that they were brown dwarfs or even very low‐mass stars in orbitsviewednearly pole‐on. Fortunately, the sheer number of these objects discovered in thelate1990slaidthisnotiontorest.

Radialvelocitymeasurementprecisionhascontinuallyimprovedfromtheinitial~15ms–1regime(Campbell&Walker1979)tothecurrentstate‐of‐the‐artof1ms–1orbetter. Theattainment of very high measurement precision has mostly been a matter of controllingsystematic measurement errors. The use of a gas absorption cell or the continualillumination with a separate reference spectrum measures intrinsic spectrograph drifts.Enclosingthespectrograph inavacuumandstabilizingthetemperaturetomKtolerancesresultintheachievementofresidualstoplanetaryorbitfitsaslowas64cms–1(Lovisetal.2006).Reachingthislevelofmeasurementprecisionopensupnewpossibilitiesforplanetdetection.WhiletheradialvelocitysignalofanEarth‐massplanetina1‐AUorbitaroundasolar‐massstarisonly9cms–1,wedohavethemeasurementprecisiontodetecthabitableEarthsaroundlow‐massMstars,andtodetectSuperEarths(planetsofafewEarthmasses)in short‐period orbits around solar‐mass stars. These Super Earths are objects in thetransitionregionbetweenthemassesofice‐giantslikeUranusandNeptuneandthemassesof the rocky terrestrial planets of our inner Solar System. An extremelywide variety ofcompositionsofsuchplanetsispossible,withmixturesofvariousfractionsofmetals(ironandnickel), silicates, “ices” (water,methane, ammonia, etc.) andH2‐He. The detection ofthese objects, and the measurement of their mass and radius (for those that undergotransits),willopenupexcitingnewareasofplanetaryastrophysics.

Once a planet is discovered, the interesting science has only just begun. Thecharacterization of planetary systems via RV observations, among other techniques, isbecoming just as important as the initial discoveries. Planets often occur inmulti‐planetsystems. The first of these found was the system around υ And. Intensive follow‐upobservations often reveal the presence of additional planets in the system. Many times,these planets will be gravitationally interacting with each other. The study of suchinteractions helps us to understand the dynamics of planetary system formation andevolution.

Dopplerstudieshaverevealedthatapproximately10%ofallF,G,andKstarshaveatleastoneplanetinthemassrange0.3−10MJupwithperiodsbetween2and2000days(Cummingetal.2008).Extrapolatingtolongerperiods,itisestimatedthat17–19%ofstarshaveagasgiantwithin20AU.Wehavealso learned thatgasgiant formation ismuchmoreefficientaroundstarswithsuper‐solarmetallicity(e.g.,Santosetal.2004;Fischer&Valenti2005).

7.1.1 RelevancetoSpaceMissionsRadialvelocitiesplayakeysupportingroleinatleasttwotypesofNASAspacemissionstodetectplanets:astrometricmissions,andmissionstofindplanetsbythetransittechnique.Additionally,RVs are a cost‐effectiveway to find targets for futuremissions aimedat thedirectdetectionandcharacterizationofexoplanets.Thesin iambiguityofRVs,mentionedabove,canberemovedwithasingleimageofthesystemifthestellarmassanddistanceareknown, or with just two images otherwise, yielding the true mass of a planet.

RadialVelocity

187

Foreknowledge of the spectroscopic orbit would maximize the returns of such missionsbecauseitwouldenablethetargetstobeobservedatthemostfavorabletimes.

The high potential for detecting low‐mass, even terrestrial, planets via a space‐basedastrometric mission has been promoted in two Decadal Surveys. NASA’s SpaceInterferometry Mission (SIM/PlanetQuest), ESA's Gaia satellite, and ESA's ground‐basedPRIMA projects are examples of astrometric missions capable of detecting low‐massplanets. The astrometric instrumentation measures the side‐to‐side reflex motion of aplanet‐bearingstar,asopposed to the line‐of‐sightmotionmeasuredbyRV. HoweverRVmeasurements are key to establishing a reference frame composed of “fixed” stars, sinceastrometric measurements of greater precision that those missions will obtain areobviouslynotavailabletopre‐selectsuitablereferencestars.

ThesolutionadoptedbySIM‐LiteistoselectdistantKgiantsasreferencestars.Astrometricsignatures of planetary companions to the K giants would be minimized by the stars’distancecomparedtonearbydwarfs.However,stellarandbrowndwarfcompanionscouldstill induce astrometric wobbles in the reference stars above SIM's tolerance of 1 µas.Candidate reference starsorbitedby thesemoremassive companions canbedetectedbypreciseradialvelocitymeasurements,andremovedfromconsideration.

Companions orbiting close to the K giant are not a concern since they induce smallastrometric wobbles. (Note also that the radii of K giants preclude “hot Jupiter” typecompanions). Companions orbiting at very large radii are also not a concern since theastrometricwobbletheyinducewillhaveaperiodlongerthanthemissionlifetimeandwillappearquasi‐linear,andwillprobablybeabsorbedintotheastrometricsolutionforpropermotion.However,companionsatintermediatedistancescorrespondingtoperiodsofafewtoafewtensofyearsarethemostproblematicforSIM.

Frinketal.(2001)havedemonstratedempiricallythatearlyKgiants(K0−K3)arerelativelystable in RV, with a typical scatter of ~20 m s–1. Their numerical simulations adoptingreasonable distributions for the mass and eccentricity of companions to these referencestars based on previous RV work have shown that the expected RV signals of thesecompanionsintheperiodrangeofinterestarenotswampedbytheintrinsicscatterquotedabove,andcanthereforebedetectedefficientlybytheDoppler technique.Hence,SIMhascommencedaprogramofRVobservations foreachof the candidate reference stars. SIMrequires4−5referencestarspertargetstar,eachwithinabout1degreeofthetargetstar,andsoRVmeasurementsareneededforabout10referencestarspertargetstarinordertoensurethat4−5survivethisvettingprocess.Aminimumofthreeobservationsisrequiredto determine that a candidate reference star is stable, and the observation scheduletypicallyspansabout5years. This large‐scaleeffortalreadyunderway isoneexampleofthecriticalimportanceoftheDopplertechniqueforthesuccessofNASAmissionstodetectandcharacterizeexoplanets.

BeyondestablishingastablereferencegridforSIM,preciseradialvelocitymeasurementsofall potential science targetsmaybe viewed as “cheap insurance.”Theywill help leverageSIMobservations,allowing fewerastrometricobservationswith lowercadence,usingRVsto fill ingapsandset constraints that ruleout theeffectsofunknownmultipleplanets inthese systems. An intensive RV study of all nearby potential SIM and Terrestrial PlanetFinder(TPF)targetscouldessentiallyprovidea“treasuremap”forspace‐basedmissionstofollow‐up. RVs could identify all nearby systemswith Jupiter‐mass planets in distant andessentiallycircularorbits(orwithnoeccentricJupiters)wheredynamicalspaceisavailableforhabitableEarthstoresideforbillionsofyears.

Chapter7

188

Radial‐velocity and spectroscopic follow‐up observations are essential for understandingtransitingplanets,bothforconfirmingthatacandidatereallyisaplanetandforderivingthemassandradiusoftheplanet,aswellasforcharacterizingthehoststar.Bothofthesestepsarecriticalforenablingtheinterpretationoftheplanetaryastrophysicsthatcanbelearnedfromtransitingplanets.

Photometric surveys for transiting planets are plagued by stellar systems thatmimic thelightcurvesoftrueplanets(see,e.g.,Mandushevetal.2005;Bouchyetal.2005;Pontetal.2005; O’Donovan et al. 2006). There are several lines of evidence that can be helpful infingeringthestellarimposters.Forexample,thediscoverylightcurvesmayhavemarginalphotometricprecision,goodenoughtosee thatsomething interesting isgoingon,butnotgood enough to see the details of the light curve accurately. High quality follow‐upphotometrycansometimesreveal thestellarnatureof thecompanionresponsible for thelight curve, for example by the detection of shallower secondary eclipses, or subtle lightvariationsout of eclipsedue to gravitational distortionsof theprimary star inducedby amassive secondary, or durations of ingress and egress that are too long to be a planet.Another approach is to look for a slight shift of the centroid of light in the stellar imageduringtransit,butthisrequiresveryaccurateastrometry.Suchapproachesforsortingoutthe stellar imposters are helpful, but the ultimate proof that the companion is a planetrequiresamassdetermination.Inprinciplemassescanbederivedfromeitherastrometricorspectroscopicorbits,butatthepresenttimeweonlyhavethetechnologytodothiswithradialvelocitiesfortheshort‐periodsystemsthatarebeingdiscovered.

The role of quantitative spectroscopy in the study of transiting planets should not beoverlooked. From a single‐lined spectroscopic orbit the mass of the planet can only bedeterminedrelativetothehoststar,soanestimateofthemassofthestarisneeded.High‐quality spectra are essential for such determinations, especially because stellar massesdependonthechemicalcomposition.Indeed,itoftenturnsoutthatthemassesderivedfortransiting planets are limited by the uncertainty in themass estimated for the host star(Torres et al. 2008).Moreover, the characteristics of the host stars are important for theinterpretationof theastrophysicsofpopulationsofplanetsand theoriesof formationandevolution. Fortunately, someof thesamespectra thatareobtained for theradial‐velocityworkcanoftenbeusedforthecharacterizationofthehoststars.

Thus,alltheNASAmissionsthatareaimedatthestudyoftransitingplanetsrelyonground‐based radial‐velocity and spectroscopic follow‐up observations. A good example is theKepler Mission, which is expected to discover hundreds, maybe even thousands, oftransiting planets. However, the transiting‐planet candidates identified byKeplerwill becontaminatedby an even largernumberof stellar imposters. Thus, ground‐based radial‐velocity and spectroscopic follow‐up observations are critical for the scientific success ofthe Kepler Mission. Indeed, the amount of follow‐up work is expected to exceed theresourcesavailable,both inquantityand inquality. Inparticular,wedonotyethave thetechnologicalcapabilitytoderivethemassofatransitingEarth‐likeplanetinanEarth‐likeorbit,especiallyfortherelativelyfaintanddistantsystemsthatKeplerwilldiscover.Similarcomments apply to the Transiting Exoplanet Survey Satellite (TESS), an all‐sky SMEXproject, except that TESS will have the advantage that it will discover the brightest andnearesttransitingplanets.

The EPOChmission of opportunity (Deming et al. 2007) is another example, only in thiscase much of the hard work has already been done, because EPOCh is studying knowntransitingplanetsforwhichradialvelocitiesandspectroscopyhavealreadybeenobserved.Nevertheless,giventhevalueofthenewdatafromEPOCh,thesesystemsdeserveadditional

RadialVelocity

189

attention, forexampletosee ifradial‐velocitymonitoringrevealsadditionalplanets inthesystems. Similar comments apply to the transiting planets being studied by Cold Spitzer(now, the Warm Spitzer) and the James Webb Space Telescope in the future, and othermissionsinthenextdecade.

7.1.2 ScienceGoalsPerhapsoneof themost exciting science goals forNASA inwhichRVswill play a centralrole,andonethatisofcaptivatinginteresttothegeneralpublic,isthesearchforEarth‐typeplanets around nearby stars, particularly those that may harbor life. These planets arelocatedinwhatiscalledthe“habitablezone”(HZ).Wedefinethehabitablezonehereastheregionaroundthestarinwhichliquidwateronthesurfaceoftheplanetisstable.BecausetheRVsignatureon thestardepends inverselyon thestellarmass (seeeq. [1]), late‐typedwarfsaremuchmorefavorablefordetectingsmaller,Earth‐massplanetsbytheDopplertechnique.Thesestarsalsohave lower luminosity,so theHZmovescloser to thestarandthisagainresultsinalargerreflexvelocity,makingthedetectioneasier.Thereisagrowingemphasis in these searches to focus onM dwarfs to achieve the sensitivity to the lowestmassplanets,intheHZoftheirparentstars.ThediscoveryofaSuperEarthorbitingtheM3dwarf GJ 581 that might support liquid water (Udry et al. 2007) and the discovery of atransitingNeptunearoundGJ436(Gillonetal.2007)haveincreasedthefocusonMdwarfs.If the sensitivityof near‐infraredRVmeasurements canbe improved so that theydeliverthesameprecisionasnowachievedintheoptical,thatwouldallowmanyhundredsofmid‐tolate‐MstarstobestudiedforthepresenceofterrestrialplanetsintheHZ,whichwouldcreate synergieswith spacemissions suchasCoRoT,Kepler, andTPF/Darwin.Thecrucialrole of the RV technique herewill be to characterize those planets (i.e., tomeasure theirmasses).ThesynergywithCoRoT/Keplerisparticularlycompelling,sincethecapabilitytoconfirm the radial velocity orbit of planetary candidates detected around low‐mass stars,wherethetransitamplitudeandperiodaremostfavorableforthedetectionofterrestrial‐mass,habitable‐zoneplanetsisespeciallychallenging.CoRoTwassuccessfullylaunchedinDecember2006,andKeplerisscheduledtolaunchinApril2009.

Butanothercompellingquestionofourtimeis:HowcommonareSolarSystemanalogswithSaturn‐andJupiter‐massplanetsbeyond4AUincircularorbits?Howoftenareplanetsinother systems arranged as in our own Solar System? One may argue that wheninterferometric or direct‐imaging techniques reach the capability to detect Earth‐typeplanets, it is the Solar System analogs that will hold the most immediate interest.Understanding the architecture of planetary systems around other stars is a key sciencegoaltoinformtheoriesofplanetformation.

Lookingtothefuture,oneofthehighestprioritiesshouldbetousetheRVtechniquetofindnearbybrightplanetarysystemswithsufficientdynamicalroomtoallowstableEarth‐likeplanets in the habitable zone,which could then be imaged by spacemissions in the nextdecade.ArelatedsciencegoalistoprovidethebesttargetsforJWSTtostudyintheIR,suchasnearbybright transiting systemswith favorable contrast thatwouldallowusaccess totheirphysicalandatmosphericproperties.Knowingthemassesoftheseplanetsisofcrucialimportanceforinterpretingthemeasurements.Forexample,themassfortransitingplanetswillestablishwhethertheobjectismadeprimarilyofrock,ofice,orwhetherithasaH/Heatmosphere.Therefore,measuringthespectroscopicorbitsofallpotentialplanetarytargetsforJWSTandtheimagingmissionsisessential.

Anotherkeyarea thatneedsa targetedstudy ismeasuring the “jitter”ofallpotentialSIMandTPFtargets.Thisistheexcessscatter(overandabovethemeasurementerrors)thatis

Chapter7

190

of astrophysical origin, i.e., that comes from the star itself (see, e.g., Santos et al. 2000;Wright 2005). Sources of jitter include p‐mode oscillations, stellar granulation, activeregions,andmagneticcycles.Stellarradialvelocityjitterisnotverywellunderstood.Somestarsaremuchquieter thanexpected from their spectral characteristics,whileothersarenoisier. Themost stable starsmeasured to date showRMS dispersions below 1m s–1 asmeasuredbytheHARPSinstrumentontheESO3.6‐mtelescopeoverseveralyears,butthemajority have larger scatter. Given a measurement precision now slightly better than1ms‐1,theconclusionisthatwearelimitedinmostcasesbythevarioussourcesofstellarnoise rather than instrumental problems. However, the study of the power spectrum ofstellarnoisesuggeststhat itmaybepossibletoreducetheimpactof jittersignificantlybytargetingthemoststablestarsandbybinningtheobservationsovertimescalessufficientlylongtoaverageoutthesesourcesofastrophysicalnoise.ThisisparticularlyrelevantforthesearchforterrestrialplanetsintheHZ,sincetheperiodsinvolvedarelongenoughtoallowsuchbinning.Theimportanceofjittershouldnotbeunderestimated.Untilweunderstanditbetter, itmaywell be thatwe can only confirm (i.e.,measure themasses of) Earth‐massplanetsaroundlate‐typestarsthatarerelativelyquiet.Onlyafractionofthestarssurveyedmayfulfillthisrequirement,makingtheneedforlargersamplesmorepressing.

7.1.3 ScienceRequirementsThesciencegoalsintheprecedingsectionleadtotwoprimaryrequirements:improvedRVprecision,andincreasedavailabilityoftelescopetime.

Intheoptical,theprecisionoftheDopplermeasurementsshouldbeimprovedtothelevelof~10cms–1,particularlyfornearbyMdwarfs.ThisshouldenablethedetectionofEarth‐likeplanetsinthehabitablezonearoundlate‐typestars.ThisisillustratedinFigure7‐1,whichshows the velocity semi‐amplitude in cm s–1 expected from terrestrial‐type planets withmassesof1M⊕,2M⊕,and5M⊕locatedinthehabitablezone,asafunctionofthemassoftheparent star. It is seen, forexample, thata1‐M⊕planet in theHZaroundasolar‐massstarwillinduceawobbleontheparentstaroftheorderof10cms–1.Sincetheeffectforagivenplanet is inversely proportional to the 2/3 power of themass of the star, and the HZ islocatedcloserinforlessluminousstars(r∝L–0.5),thesameplanetintheHZofa0.1‐M

star

producesa1ms–1effect,whichismuchmoreeasilydetectable.

As identified by the Exoplanet Task Force (ExoPTF) report, it is attractive to apply thistechnique in the NIR region of the spectrumwhere coolM‐type stars emitmost of theirenergy. Themajority of the Sun’s stellar neighbors are cool, low‐mass stars.Of the~150stars lying within 8 parsecs, there are ~120 M dwarfs but only 15 solar‐type G dwarfs.Mdwarfsareintrinsicallyfaintatvisualwavelengths(15to106timesfainterthantheSun),butareconsiderablylessfaintinthenear‐infrared.Thus,anM4dwarfat20parsecshasV~14,whichisbeyondthereachofcurrentopticalinstruments,butwithJ~9.5,itisreadilyaccessibletohigh‐precisioninfraredRVobservations.AlthoughthecatalogsofnearbylateMdwarfsarenotcompleteoverthewholesky,theidentificationofalltheMdwarfswithinat least30parsecs is expected tobe complete in thenear futurewith the combinationof2MASS, proper motion surveys, and synoptic surveys such as Pan‐STARRS and LSST.Although much more difficult than in the optical, it is therefore important to push theprecisionofRVmeasurementswithIRspectrometerstosignificantlybetterthan10ms–1,perhapsdowntothelevelof1ms–1.

In order to characterize the astrophysical limitations to the RV precision, genericallyreferred to as “jitter” (stellar oscillations, granulation, etc.), it is essential to observe asufficiently large sample of stars of different spectral types. These observationswill help

RadialVelocity

191

understandthevariouscomponentsofthejitter,theirvelocityamplitudes,andtheirtypicaltimescales.

Figure71.Radialvelocitysemi‐amplitude(incms–1) inducedontheparentstarbyterrestrialplanets inthehabitablezone,asafunctionofstellarmass.TheHZisindicatedbythedifferentbands,anditsdefinitionfollowstheworkbyKastingetal.(1993).Threedifferentplanetmassesareconsidered,aslabeled.Theorbitalperiodsoftheseplanets(indays)areindicatedbytheverticallines.HZsareshownfortwodifferentstellarages,giventhatthestellarluminositydependsonageandinfluencesthesizeandlocationoftheHZ.(G.Torres,SmithsonianAstrophysicalObservatory)

Perhapsoneofthemostimportantrequirements,however, istohaveaccesstosignificantamounts of telescope time. Present work to survey the latest‐type stars (around whichlower‐mass planets aremore likely to be detected; see above) is severely limited by theavailabilityofobserving time.The leadingRVgroupsareessentially limitedbyshotnoiseandthenumberofvisitstheycanaffordtomaketoagiventarget.Lengtheningtheexposuretimes and/or increasing the number of visits is important to reduce the impact ofastrophysical noise, as described earlier. Longer exposures and more telescope nightswould immediately bring not only higher precision, but also fainterMdwarfs into reach.The present RV surveys of late‐type stars are observing only a few hundred objects.CompletesamplesdowntoV=13wouldincludethousandsofMdwarfs,andextendingthebrightnesslimittoV=14couldbringtensofthousandsmorewithinreach.Exposingsuchfaint stars for long enough to avoid being limited by photonnoise requires an enormousamountofobservingtime,whichisnotpresentlyavailable.

Chapter7

192

7.2 ObservatoryConceptsThelimitedavailabilityoftelescopetimehasemergedasoneofthemostseriouschallengesforreachingmanyofthesciencegoalslistedinSect.1.1.2,asindicatedalsointheExoPTFreport.Intheareaofdetection,presentsurveysarecollectivelyobservingsome3,000brightSun‐like stars systematically,mostlyof spectral typesF,G, andK. It ishighlydesirable toincreasethesamplesize,andinparticulartoincreasethenumberofMdwarfssubjectedtoDopplermonitoring, as those are the onesmost likely to allow the detection of the firstEarth‐massplanetinthehabitablezone.Thecharacterizationofplanets,particularlythosethatundergotransits,isbecomingincreasinglyimportantfortheground‐basedphotometricsurveys, andwillmovecenter‐stagewith the launchof theKeplermission.Theenormousamountof timerequiredon large telescopes to followuprelatively faint transitingplanetcandidatesandmeasuretheirspectroscopicorbits(allowingthemasstobedetermined)isalreadythebottleneckfortheCoRoTmission(see,e.g.,Pontetal.2008).

Althoughgainingaccesstomoretelescopetimedoesnotinitselfrequirenewtechnology,itisstillviewedbythecommunityasamajorobstacleandasignificantcostitemfordetectinglowermassplanets and characterizing all planets, one that policymakerswouldbewell‐advisedtopaycloseattentionto,startingnowandcontinuingintothenextdecade.

Thereareanumberofoptionsforthis,including(inroughincreasingorderofcost):

• Buymoretelescopetimeonexistingtelescopesequippedwith1‐ms–1precisionRVcapability. Examples include the AAT in the southern hemisphere or Keck in theNorth.

• Build instruments with 1‐m s–1 precision capability and place them on existingtelescopesthathaveavailabletimeorareunder‐used.Thiscouldincludeanumberof4‐mclasstelescopes.

• Build dedicated ground‐based large telescopes equipped with 1‐m s–1 capability.This could be either a North‐South pair of 6−8‐m class Magellan‐style clones, orperhaps a global network of 2−4‐m class telescopes with high‐precision RVspectrometers.These facilitiescouldbe fullyautomatedandremotelyoperatedbyexperiencedradial‐velocityteams.

7.2.1 ArchitectureandPerformanceBecause of the large mass of RV spectrometers, launching high‐precision RV planet‐detectioninstrumentsintospaceisnotpracticalintheimmediatefuture.Amuchbetteruseof resources is to support space‐based missions with cost‐effective ground‐based RVprograms (as described above regardingKepler and SIM). Furthermore, focused ground‐basedRVprogramscancarryoutsomeofNASA’splanet‐detectiongoalsintheirownright.Anumberofdesignsforground‐basedfacilitiesthatwilladdressthesciencerequirementsdescribed above are already in the construction stages or will soon be tested. Theyrepresentvalidexamplesofeffortsthatshouldbestronglysupportedoverthenextdecade,andaredescribedbrieflyhere,beginningwiththeoptical.

• HARPSNEF (High Accuracy Radial velocity Planet Searcher—New EarthsFacility):Thisisahigh‐resolution(R=120,000)fiber‐fedopticalspectrographwithbroad wavelength coverage (3780−6910 Å) designed after the very successfulHARPS instrument currently in operation on the 3.6‐m ESO telescope at La Silla,Chile. HARPS‐NEF is a collaboration between the New Earths Facility (NEF)

RadialVelocity

193

scientistsoftheHarvardOriginsofLifeInitiativeandtheHARPSteamattheGenevaObservatory. It is expected to be a workhorse for follow‐up of transit candidatesfrom the Kepler mission and will be operational in the 2010 observing season.HARPS‐NEFisacross‐dispersedechellespectrographthatwillbenefitnotonlyfromupdatesand improvementsover theoriginalHARPS instrument,but inaddition itwill be installed on a larger telescope aperture in the northern hemisphere (the4.2‐mWilliamHerschelTelescopeonLaPalma,Canary Islands). It isdesigned forultra‐high stability (10−20 cm s–1) andwill be placed in a vacuum chamberwithcareful temperature control. Themethod chosen for thewavelength calibration isthe simultaneous Th‐Ar technique, in which light from the comparison lamp isdirected to the CCD through an optical fiber, separate from the fiber carrying thelightfromthestar.However,plansalsocallfortheinstallationofalaserfrequencycombwiththepotentialofincreasingthemeasurementprecisionfurther.

• Magellan PFS (Planet Finder Spectrograph): The design of this instrument isoptimizedtoachieve1‐ms–1RVprecision.Relativelyfewoftheseinstrumentsexisttoday. The motivation is to provide more “air time” for high‐precision velocitymeasurements needed to detect and characterize exoplanets (particularly in theSouth), asdiscussed earlier. TheMagellanPFSwill bemountedon theClay6.5‐mtelescopeattheLasCampanasObservatory,inChile,andisexpectedtoseefirstlightattheendof2008orbeginningof2009.Itisahigh‐resolution(R4grating,givingR~38,000/arcsec)high‐efficiencydouble‐passslitechellespectrographwithpassiveand active temperature control, designed to enable very high quality flat‐fieldingand with careful slit and pupil illumination to minimize velocity errors. Thewavelength fiducial isan iodinegasabsorptioncell.Thegoal is toachieve2‐ms–1precision for a single exposure, and 1‐m s–1 by binning 4 exposures over suitabletimescalestoaverageoutp‐modestellaroscillations.

• MARVELS (Multiobject Apache Point Observatory Radial Velocity ExoplanetLargeareaSurvey):Thisisnotanactualinstrument,butanunprecedentedlarge‐scale Doppler survey of ~11,000 stars (compared to the ~3000 stars currentlybeing observed collectively by other surveys in the brightness range V = 8−12.MARVELS uses new‐generation fiber‐fed multi‐object instruments based on adispersedfixed‐delayinterferometer(seeT‐EDIbelow).Theseinstrumentsconsistof a wide‐angle Michelson interferometer followed by a medium‐resolutionspectrograph. The spectrograph can be thought of as dispersing the light outputfrom the interferometer into a large number of narrowwavelength channels. Theinterferometercreatesinterferencefringeswithineachofthesechannels.ADopplershiftcanbemeasuredasashiftinthephaseofthefringes.Oneofthemainsciencemotivationsistoprovidethelargesthomogeneousgiant‐planetsampleforrevealingthe diversity in giant exoplanet populations, and testingmodels of the formation,migration,anddynamicalevolutionofgiantplanets.Thesurvey,which isslatedtostartinlate2008,isalsoexpectedtofindrareplanets,transitingplanets,aswellassignposts for planetary systems with lower‐mass or more distant planets, to bedetectedwithfollow‐upobservations.

• AutomatedPlanetFinderTelescope:Thisisacompletelyrobotic2.4‐mtelescopeunderconstructionattheLickObservatory(commissioningexpectedinJune2009),equippedwithahigh‐resolution spectrographdesigned togiveanRVprecisionof1ms–1toallowtheDopplerdetectionofplanetswithmassesassmallas5M⊕.Itisoptimized for high efficiency, with a small secondary obscuration and protected

Chapter7

194

silvercoatingsonthesecondaryandtertiarymirrors.Theopticaltrainwillincludean atmospheric dispersion compensator, as well as an iodine cell to serve as thewavelengthreference.Itisadedicatedtelescopethatwillmakeintelligentdecisionseachnightaboutwhichstarstoobserve,whatdataqualityisoptimal,andwhetheraplanet can be considered to have been detected. Given sufficient funding support,one could imagine replicating the facility a number of times to build an efficientnetwork of similar dedicated robotic telescopes thatwould substantially increasethenumberofnightsavailabletothecommunityforRVwork.

Extensiveeffortsover the last fiveyearshave focusedon thedevelopmentof infraredRVcapabilities, toextend the reachofhigh‐precisionmeasurements to the lowestmassstarsand brown dwarfs. Although these objects are faint in the optical where existing RVspectrometers are capable, they are much brighter in the near infrared and haveatmospheresthatarerichmolecularcocktailswithdenseforestsoflinessuitableforpreciseRVmeasurements.Wedescribehereonlythreeofthemoreadvancednear‐infraredefforts.

• TEDI(TripleSpecExternallyDispersedInterferometry):Theprincipleonwhichthis instrument operates is a combination of interferometry and multichanneldispersive spectroscopy that dramatically improves the velocity precision ofmoderate‐resolution,high‐throughputspectrographs(Edelesteinetal.2006).T‐EDIis essentially an interferometer coupled with the existing Cornell TripleSpecinfraredsimultaneousJHK‐bandspectrographatthePalomarObservatory200‐inchtelescope. This scheme effectively multiplies the spectral resolution of thespectrograph by a factor of several to an order of magnitude over its full andsimultaneous bandwidth. The interferometer creates a transmission comb that isperiodic with wavelength, which multiplies the input spectrum to create moiréfringes.Thesefringesprovideaperiodicspectral fiducialcombcoveringtheentirebandwidth of the spectrograph. This comb is analogous to the fiducial lines of aniodine absorption cell, but with lines of exceedingly uniform spacing, shape, andamplitude over the entire bandwidth. The comb, in multiplication with the inputspectrum, heterodynes fine spectral features into a lower spatial‐frequencymoirépattern that is recorded by the spectrograph detector. The EDI fringing signalprovides a precise internal fiducial that can be used to defeat systematicinstrumental noise so that the tolerance to blur or pupil changes is improved byorders of magnitude compared to classical high‐resolution spectrographs thatdirectlymaptelescopeimagequalityandpupilstabilitytospectralperformance.ADopplervelocitychangeinducesaphasechangeinthemoirépatternrelativetothemoiré pattern of a simultaneously measured calibration spectrum, produced byeither a gas absorption cell or an emission lamp that has been simultaneouslyrecorded in the same fringing signal. Tests indicate an expected RV precision of~5−10 m s–1. This instrument was commissioned at the end of 2007, and ispresentlyinthescienceverificationphaseatPalomar.

• PrecisionRadialVelocitySpectrographPathfinder(Ramseyetal.2008):Thisisanin‐planefiber‐fedechellespectrographthatoperatesacrossthespectralrangeof0.9−1.7µmat a resolving power ofR = 50,000, and uses a liquid‐nitrogen‐cooledHAWAII 1‐K detector. In its present implementation, the wavelength reference isprovidedbystandardTh‐Ar,Ar,Kr,Ne,andXelampstoachieveasufficientnumberof lines. Tests in the laboratory have indicated precisions near 10m s–1 over theshortterm,althoughthesystemhasyettobeusedonstarsandlonger‐termstability

RadialVelocity

195

atthislevelhasyettobedemonstrated.Theinstrumentis intendedforuseontheHobby‐EberlyTelescopeattheMcDonaldObservatory.

• UPF (UKIRT Planet Finder): This is a proposed high‐resolution (R ~ 70,000)infrared fiber‐fedwhite‐pupil cross‐dispersed echelle spectrograph for the UKIRT3.8‐mtelescopeonMaunaKea,basedonanearlierdesignoriginallyintendedfortheGeminiObservatory.The instrumentwillwork in theY, J,andHbands,with2×2KHAWAII‐2RGdetectorarraysandsimultaneousarc‐linecalibration,andisdesignedforhighstability.InmanyrespectsitissimilartothehighlysuccessfulHARPSandUVES spectrographs. Extensive simulations indicate the expected instrumentalerrorswillbeunder1.5ms‐1foratypicalM6dwarfat10pc.Thisincludesresidualerrors from telluric‐line contamination, after proper precautions to mask out~30km s‐1 regions around features deeper than about 2%. An instrument of thiskind on UKIRT would be ideal for a planet‐search campaign requiring a largeamountoftelescopetime,anditwouldenablesurveysofhundredsoflate‐typestarswiththepotential fordetectionofseveral terrestrial‐massplanets inthehabitablezone.While originally a UK project, the large cost of funding the instrument andoperations(~$15M)presentsanopportunityforUSinvolvementonacollaborativebasis,alongthelinesoftheoptionspresentedatthebeginningofSection7.2.

Over the next decade significant progress will be made in the development of the nextgenerationof giant (20‐ or 30‐m) segmentedmirror telescopes (GSMT). If equippedwithoptimizedhigh‐precisionRVspectrometers,thesefacilitieswouldhaveuniquecapabilitiestosearchforhabitableEarth‐massplanetsaroundnearbystarslaterthanspectraltypeM3,and brown dwarfs as well. They would also provide the crucial follow‐up capabilityrequiredfordeterminingmassesincurrentandfuturespace‐basedsearchesfortransitingplanets.ThelargerapertureofaGSMTwouldalsomakepossibleveryhighsignal‐to‐noiseratiospectroscopyoftheatmospheresofhotJupiters,enablingstudiesoftheiratmosphericconstituents in transmission (during transits) and in reflection. The following describesbrieflyaninstrumentconceptdesignedfora30‐mclasstelescope(TMT):

• Moderate to High Resolution spectrometer: This concept combines the bestadvantagesofboththeVLT’sUVESinstrumentandKeck’sHIRESspectrometer(seeSect. 7.3.1). The dual‐white‐pupil/dual‐arm configuration of UVES is used to limitthe sizes of the echelle grating, cross disperser, and camera; and a HIRES‐stylecameraisusedtoallowforamuchlargercameraasthespectrometerisscaleduptomatch a TMT’s enormous 30‐m aperture. Moderate resolution could be used toallowmulti‐objectcapability,andthehigh‐resolutionmodewouldbeusedforsingleobjects. The resulting design would work very efficiently on a TMT, alreadyexceeding the throughput (product of slit width and resolution) of any existingspectrometer by about 15%, with even higher throughput possible upon furtherscaling or using a steeper echelle. This is a large instrument, with a footprint ofabout 10 × 11 m. The optics are also large and challenging, though well withintoday’stechnologicalcapabilities.Theechellegratingsthemselvesare~1.0×3.5m,and the collimators are2m indiameter and are off‐axis sections of a 3‐mparentparaboloid.Tolowercosts,theycouldalsobemadesphericalandwarpedintooff‐axisparaboloidsbystressingharnesses.PlacedontheNasmythplatformatanf/15focus, it is anticipated that this queue‐scheduled instrument could be available atanytimeduringthenight,andcouldbebroughtintooperationontheskyinaslittletimeasittakestorotatethetertiarytofeedthelightontoitsslit.Itisexpectedthataprecisionof1ms–1couldbeachievedonaV=13magMdwarfinabout8.3minutes.

Chapter7

196

Thus, larger samples of late‐type stars than presently accessiblewould bewithinreachandcouldbeobservedinareasonableamountoftime.

7.3 Technology

7.3.1 PastAccomplishmentsThe lastdecadehasseentremendoustechnicalprogress inachievinghighprecision inRVmeasurements to search for and characterize exoplanets. In the optical, the two mostproductiveinstruments(HIRESontheKeck10‐mtelescope,Vogtetal.1994;HARPSontheESO3.6‐mtelescope,Mayoretal.2003)nowroutinelyachieve1‐ms–1precisiononslowly‐rotatingsolartypeaswellas late‐typestars,andhavedemonstratedsub‐m/sprecisioninsomecases(Lovisetal.2006).HIRESisacross‐dispersedechellespectrographinoperationsincethemid1990’s,typicallyusedwithslitwidthsgivingresolvingpowersofR∼70,000.Itmakes use of an iodine gas absorption cell to impose a dense forest of I2 lines on thestellarspectrum,providingaverystablewavelengthreferenceandallowingchangesintheinstrumentalprofiletobetrackedveryaccurately(Marcy&Butler1992;Butleretal.1996).HARPS is a R ∼ 110,000 fiber‐fed echelle spectrograph maintained in a temperature‐controlledvacuumchamber,whichhasbeeninoperationsince2003.Insteadofusingagasabsorptioncell,itreliesonthorium‐argonemissionlinesrecordedsimultaneouslywiththestartoprovidethewavelengthreferenceandtrackvelocityshifts.BothoftheseinstrumentshaveallowedthediscoveryofSuperEarths,planetswithmassesafewtimeslargerthanourown.Other instrumentshave achievedprecisionsof a fewm/s, allowing thediscoveryofNeptune‐massplanets(∼20M⊕;McArthuretal.2004).

Based on the experience in operating these spectrographs, the main factors currentlylimiting the precision of the RVs, in addition to photon noise, are stellar noise, thewavelengthcalibration,telescopeguiding,stabilityintheilluminationofthespectrograph,anddetector‐relatedeffects.Theonlysolution tophotonnoise ismorephotons.Thismaycome from either larger telescopes equippedwith high‐precision spectrographs, ormoretelescopetimeonexistingfacilitiespermittinglongerexposureswithoutcompromisingthesize of the samples of stars surveyed for planets. Indications are that the remedy to theproblem of stellar noise may be similar. Longer exposures or binning over suitabletimescales(againimplyingaccesstomoretelescopetime)canreduceastrophysicaljittertosomeextentbyaveragingoutthose intrinsicvariations(see,e.g.,Pepe&Lovis2007).TheinstrumentalfactorsarecurrentlyestimatedtolimittheprecisionoftheRVsobtainedwithHARPSto30−50cms–1onboththeshorttermandthelongterm.Recentprogresshasbeenmadetoimprovethewavelengthcalibrations(e.g.,Lovis&Pepe2007)aswellastostabilizethe spectrograph illumination. All in all, from the technological point of view it appearspossible to achieve ~10 cm s–1 precision in the near future with a new‐generationinstrumentdesignedwiththeabovefactorsinmind.

Recently a new kind of instrument (the Exoplanet Tracker, or ET) entered the arena ofDoppler searches and produced the discovery of a 0.5‐Jupiter mass planet around a K0main‐sequencestar(Geetal.2006).Thismarkedthefirsttimethataplanetwasdiscoveredby a radial‐velocity technique other than the echelle technique. The ET is based on adispersedfixed‐delay interferometer(DFDI) forDopplermeasurements,describedearlier.Itcombinesamoderate‐dispersionspectrographwithatwo‐armedinterferometer,addingagradedphasedelayperpendiculartothedispersiondirectionthatcreatesfringesineachresolution element. Doppler shifts cause the fringes in each resolution element tomove,

RadialVelocity

197

allowingameasurementoftheradialvelocityagainstthereferenceprovidedbyaniodine‐gas absorption cell (Erskine& Ge 2000; Ge 2002; Ge et al. 2002). Although the velocityprecision demonstrated so far on real stars is limited to 10−20m s–1, improvements areunderway.AsignificantadvantageoftheDFDIisthatthesametechniquecanprovidemulti‐object capability on telescopes with large fields of view, although at the cost of fewerphotonsperstar.Testshavebeenconductedonaprototype instrument thatobserved60stars simultaneously in the brightness rangeV = 8−12with the Sloan Digital Sky Survey2.5‐m telescope at the Apache Point Observatory, in New Mexico. This multiplexingadvantageopensthepossibilityofcarryingoutlargesurveysof~10,000starsatmoderateRV precision (seeMARVELS above) that were previously simply not possible due to thelimited amount of telescope time available on telescopes equipped with single‐objectspectrographs.

Efforts to measure precise RVs in the near infrared (0.9−2.5 µm), where M dwarfs arebrighter,havesofarmetwithlimitedsuccess.Inthelate1970’sand1980’sarapidscanningFourier Transform Spectrometer (Hall et al. 1979) was operated on the Kitt Peak 4‐mtelescope to obtain spectra of giant stars in the 2.2−2.5 µm region. Precisions of~30−50ms–1permeasurementwereachievedonαBoousingthetelluricH2Olinesforthezero‐point calibration, and changes of~200m s–1weremeasured over a span of severalyears at the 6−8σ level (Heyer et al. 1988). However, the sensitivity of the instrumentalsetupwasratherlimited,andtheefforthasnotbeencontinued.MorerecentlyWalkeretal.(2003) used UKIRT’s Cooled Grating Spectrometer (CGS4) to measure velocities for GJ229B, a T6 brown dwarf of magnitude J = 14. The precision they achieved was slightlyworse than 1kms–1. Smith et al. (2003) used the PHOENIX high‐resolution infraredspectrometerontheGeminiSouth8‐mtelescopetoobtainradialvelocitiesofEpsilonIndiBaandBb.TheirresolvingpowerwasR~50,000,andtheirstatedradialvelocityprecisionwas only ~700ms–1. Lebzelter et al. (2005) used a NICMASS detector on the 1.9‐m Mt.Stromlocoudéspectrograph toperformaradialvelocitystudyof long‐periodvariables in47Tucat J=7.5−8.5.TheirspectralresolvingpowerwasR~37,000,andtheyachieveda~400ms–1velocityprecision.Forthepast4−5yearsagrouphasbeenattemptingtousethestate‐of‐the‐artNIRSPECnear‐IRspectrometer(R=25,000)ontheKeck10‐mtelescopeforplanet detection around brown dwarfs and very low‐mass stars in the K‐band(Charbonneau 2004). The technique involves using telluric methane absorption featuressuperimposed on a 12COR‐branch band at 2.3µm (c.f., Deming et al. 2005). The velocityprecisionreportedthusfarbythisapproachisabout70ms‐1fortwoM3.5Vstars(GJ725Aand B), although this has only been demonstrated over a 3‐day timescale. A morerepresentative precision obtained over several years for late‐type stars is ~200 m s–1.Stability and calibration issues with the NIRSPEC detector (Raytheon ALADDIN‐III InSb1024×1024)appeartobelimitingtheprecisionofthesemeasurements.

7.3.2 FutureMilestonesOneofthekeyfactorsthatdeterminetheprecisionoftheRVsisthewavelengthreference.Existing technologies in the optical (Th‐Ar technique, iodine‐gas absorption cell) havealreadyachievedsub‐m/sprecisioninsomecases,butfurtherimprovementsareneedediftheDoppler technique is toreachcm/sprecisions.Anewtechnology thathasemerged inthelastfewyearsandthatholdsgreatpromiseforprovidingaverystablereferenceisthatoflaser“frequencycombs.”Asthenamesuggests,afrequencycombgeneratedfrommode‐lockedfemtosecond‐pulsedlasersprovidesaspectrumofverynarrowemissionlineswithaconstantfrequencyseparationgivenbythepulse‐repetitionfrequency,typically1GHzfor

Chapter7

198

this application.This frequency canbe synchronizedwith an extremelyprecise referencesuch as an atomic clock. For example, using the generally available Global PositioningSystem(GPS),thefrequenciesofcomblineshavelong‐termfractionalstabilityandaccuracyofbetterthan10–12.This ismorethanenoughtomeasurevelocityvariationsataphoton‐limitedprecisionlevelof1cms–1inastronomicalobjects(see,e.g.,Murphyetal.2007).Thisdirect link with GPS as the reference allows the comparison of measurements not onlybetweendifferent instruments, but potentially also over long periods of time. To providelines with separations that are well matched to the resolving powers of commonly usedechelle spectrographs, a recent improvement incorporates a Fabry‐Pérot filtering cavitythatincreasesthecomblinespacingto~40GHzoverarangegreaterthan1000Å(Lietal.2008). Prototypesusing a titanium‐doped sapphire solid‐state laser havebeenbuilt thatprovideareferencecenteredaround8500Å.

Inpractice,Dopplermeasurementswillalsobeaffectedbyotherinstrumentalproblemsasdescribed earlier, so that the value of this new technology for highly precise RVmeasurements is still to be demonstrated. For example, it is unclear to what degree thedifferent paths of the laser light and the star light into the spectrographmight affect thevelocitymeasurements.Lasercombsystemshavemanyopticalcomponents(lenses,Fabry‐Pérot elements, mirrors, etc.) whose long‐term stability could affect the resultant RVprecision.Anti‐reflection coatings tend to age, and group‐delay dispersion changes in thecoatings could affect the stability of the system. Non‐linear effects involving power andsystem efficiency variations over time, as well as detector non‐linearities over themuchlargerdynamicrangeinvolved,couldalsopotentiallycompromisethevelocityprecision.

Forthelasercombapproachtobepractical,theseissueswillneedtobeworkedout.Asofthiswriting, tests of this “astro‐comb” are scheduled to begin in late 2008 on telescopesoperated by the Harvard‐Smithsonian Center for Astrophysics onMt. Hopkins (Arizona),and elsewhere. If successful, plans call for the deployment of an astro‐comb at the ultra‐stableHARPS‐NEFspectrographbeingbuiltbytheHarvardOriginsofLifeInitiativeforthe4.2‐mWilliamHerschel telescope onLaPalma (Canary Islands). Achieving thismilestonewouldbringusastepcloser tobeingable toroutinelymeasurethemassesofEarth‐massplanets in the habitable zones of nearby stars, even solar‐type stars, a capability that iscentraltothesuccessoftheKeplermission.

Thelackofasuitablewavelengthfiducialredwardof1µmisconsideredoneofthemajorfactors limiting the precision of RV measurements with IR spectrometers. The currentreliance by some projects on telluric methane‐absorption features that are naturallysuperimposedonthestellarlineshassofarnotproducedprecisionsanywherenearthoseobtainedevenonlate‐typestarsintheoptical.EffectssuchaswindcurrentsintheEarth’satmosphere are expected to limit the precision attainable with this technique to10−30ms‐1.Gasabsorptioncellsthathaveworkedsowellintheopticalhavethepotentialtoovercomethisproblemifasuitablespeciescanbefoundthathassufficientlinedensityandwavelength coverage in the IR. Somepromising candidates have been identified thatproducerichabsorptionfeatures inthe JandHbands, includingthemoleculesCH3I,C2H2,NH3,andHCN(Guelachvili&Rao1993;VanderAuweraetal.2002).Otherspecies(suchasCH3,CO,HCCN,H2S,andHF)havealsobeenproposed.Apracticaldemonstrationthatanyofthese species provide a useful and stable wavelength reference would be a crucial steptowardachievinganRVprecisionof~1ms–1intheIR.Alternatively,itmaybepossibletoextendthelasercombtechniquetotheIR,forexampleusingadifferenttypeoflaser.EitherimplementationofastablewavelengthfiducialwouldbeamajorstepforwardinourabilitytomeasurethemassesofEarth‐typeplanetsorbitinglate‐typestars.

RadialVelocity

199

An additional difficulty in the IR is the lack of high‐resolution IR spectrographs with R∼60,000 or higher and a largewavelength coverage. The lack ofwavelength coverage incurrent instruments is largelydue to the lackof large‐format IRdetectorsand the coarseechellegratingsfabricatedwithcommerciallyavailablegratingrulingtechnology.ThemostobviouswaytoincreasetheIRdispersionistoincreasethegratingsizebyplacingmultiplegratingsinaphasedmosaic,similartowhatwasdonewithHIRESontheKecktelescope.Asignificantdisadvantageisthatthecollimatedbeamsizeislarge,andtheoverallinstrumentbecomesverybulkyandcostly.ItisalsoimpracticaltocoolandmaintainverylargeIRhigh‐resolution spectrographs. Another way of increasing the grating dispersion withoutincreasing the grating size is to use immersion diffraction gratings. Silicon immersiongratings are of special interest for increasing the dispersion power of IR spectroscopybecausesiliconhasoneofthehighestrefractiveindices,n=3.4,andhasgoodtransmissionover the range 1.05−6 µm at cryogenic temperatures (e.g., MacFarlane et al. 1958).Demonstratingthistechnologywouldbeanotherimportantstepforwardtowardachievingm/sprecisionintheIR.

7.4 Research&AnalysisGoalsThefollowingisalistofresearchandanalysisactivitiesthatshouldbesupportedinorderto enable progress in the detection and characterization of exoplanets with the RVtechnique.ManyoftheseprovideindirectsupporttospacemissionssuchasSIMorTPFbyeither helping to improve the precision of the Doppler measurements or allowing moreaccesstotelescopetimesothattargetsofinterestforthosemissionscanbeidentified.

• Supportresearchtosignificantly improvethestabilityofthewavelengthreferenceusedforDopplermeasurementsintheoptical,andtoextendittotheIR.Oneavenuewould be the implementation of femtosecond laser frequency combs, both in theoptical and the IR.Anotherwouldbe to extend theapplicabilityof gas‐absorptioncells already used in the optical to longer wavelengths. Examples include a “hot”iodinecell(~200ºC)thatcouldprovideastablewavelengthreferenceinthe8000–9000 Å range. Recent studies show that there are a number of other promisingcandidate chemicals (CH3I, C2H2, NH3, and HCN; Guelachvili & Rao 1993; VanderAuweraetal.2002)thatproduceabsorptionbandsintheJandHbandsandcanbepotentiallyusedforprovidinganaccuratewavelengthreferenceintheIR.

• Support the development of multi‐object high‐precision optical and IR Dopplerinstruments for wide‐field telescopes (such as the SDSS 2.5‐m telescope with a7‐square‐degree fieldof view,orChina’sLAMOST4‐mwide‐field telescopewitha20‐square‐degree field‐of‐view). These instruments would have an enormousmultiplexingadvantage,andwouldbeamuchmoreefficientwayof increasingthenumberofstarsobservedcomparedtosimplyaddingmoretelescopeswithsingle‐object RV instruments. As an example of this advantage, there are∼50 solar typestarswithV<10withina~20‐square‐degreeoffieldofviewthatcouldbeobservedsimultaneously with such instruments. Multi‐object high‐precision DopplerinstrumentswithR~20,000and4000–6000Åwavelengthcoverageshouldbeabletoreachaphoton‐limitedprecisionof~1ms–1in~0.5hourintegrationsatV=10.

• Support efforts to design Doppler measurement instruments for 20–30‐m classtelescopes,intendedforhighprecision.

Chapter7

200

• Supporteffortstoobtaintheobservationsneededtounderstandstellar jitter.Thiswill require a focused and dedicated effort run by teams that have demonstratedsub‐m/sprecisioncapability.

7.5 ContributorsPaulButler,CarnegieInstitutionofWashington,DepartmentofTerrestrialMagnetism

WilliamCochran,UniversityofTexasatAustin

DebraFischer,SanFranciscoStateUniversity

JianGe,UniversityofFlorida

DawnM.Gelino,NASAExoplanetScienceInstitute,Caltech

NaderHaghighipour,InstituteforAstronomy,UniversityofHawaii

ScottHorner,LockheedMartin

StephenKane,NASAExoplanetScienceInstitute,Caltech

DavidLatham,HarvardSmithsonianCenterforAstrophysics

GregoryLaughlin,UCO/LickObservatory

JamesLloyd,CornellUniversity

ChristopheLovis,GenevaUniversity

GeoffMarcy,UniversityofCalifornia,Berkeley

MichelMayor,GenevaUniversity

ChrisMcCarthy,SanFranciscoStateUniversity

FrancescoPepe,GenevaUniversity

DidierQueloz,GenevaUniversity

JeanSchneider,CNRSParisObservatory

GuillermoTorres,HarvardSmithsonianCenterforAstrophysics

StephaneUdry,GenevaUniversity

StevenVogt,UCO/LickObservatory

KasparvonBraun,NASAExoplanetScienceInstitute,Caltech

RadialVelocity

201

7.6 ReferencesButler,R.P.,Marcy,G.W.,Williams,E.etal.1996, “AttainingDopplerPrecisionof3m/s,”

PASP,108,500–509

Bouchy,F.,Pont,F.,Melo,C.etal.2005,“Dopplerfollow‐upofOGLEtransitingcompanionsintheGalacticbulge,”A&A,431,1105–1121

Campbell,B.&Walker,G.A.H.1979,“PrecisionRadialVelocitieswithanAbsorptionCell,”PASP,91,540–545

Charbonneau,D. 2004, Bok Prize Lecture (shared), “The Brown Dwarf Radial VelocitySurvey,”CenterforAstrophysicsColloquiumLectureSeriesTalk,Cambridge,MA,USA.

Cumming,A.,Butler,R.P.,Marcy,G.W.etal.2008, “TheKeckPlanetSearch:Detectabilityand theMinimumMass andOrbital PeriodDistribution of Extrasolar Planets,” PASP,120,531–554

Deming,D.,A’Hearn,M.F.,Charbonneau,D.etal.2007,“TheEPOXI/EPOChInvestigationofTransiting Extrasolar Planets,” American Astronomical Society, DPS Meeting #39,#22.02

Deming,D., Brown,T., Charbonneau,D. et al. 2005, “A New Search for Carbon MonoxideAbsorption in theTransmissionSpectrumof theExtrasolarPlanetHD209458b,”ApJ,622,1149–1159

Edelstein, J., Erskine, D. J., Lloyd, J. et al. 2006, “The TEDI Instrument for Near‐IR RadialVelocityStudies,”Proc.SPIE,6269,62691E

Erskine,D.J.,&Ge,J.2000,“NovelInterferometerSpectrometerforSensitiveStellarRadialVelocimetry,” in Proc. Imaging the Universe in Three Dimensions, Edited by W. vanBreugelandJ.Bland‐HawthornASPConferenceSeries,195,501–507

Fischer,D.,&Valenti,J.2005,“ThePlanet‐MetallicityCorrelation,”ApJ,622,1102–1117

Frink, S., Quirrenbach,A.,& Fischer,D.A. et al. 2001, “A Strategy for Identifying theGridStarsfortheSpaceInterferometryMission,”PASP,113,173–187

Ge,J.2002,“FixedDelayInterferometryforDopplerExtrasolarPlanetDetection,”ApJ,571,L165–L168

Ge, J., Erskine, D. J., & Rushford, M. 2002, “An Externally Dispersed Interferometer forSensitiveDopplerExtrasolarPlanetSearches,”PASP,114,1016–1028

Ge,J.,VanEyken,J.,Mahadevan,S.etal.2006,“TheFirstExtrasolarPlanetDisoveredwithaNew‐GenerationHigh‐ThroughputDopplerInstrument,”ApJ,648,683–695

Gillon, M., Pont, F., Demory, B.‐O. et al. 2007, “Detection of transits of the nearby hotNeptuneGJ436b,”A&A,472,L13–L16

Guelachvili, G. & Rao, K.N. 1993, Handbook of Infrared Standards—II (Harcourt Brace &Company,Boston)

Hall,D.N.,Ridgway,S.,Bell,E.A.etal.1979,“A1.4meterFourierTransformSpectrometerforAstronomicalObservation,”Proc.SPIE,172,121

Heyer,I.,Hall,D.N.B.,&Hinkle,K.1988,“SearchforDarkCompanionsofKandMGiants.PreliminaryResultsforαBoo,”VistasinAstronomy,31,317–321

Chapter7

202

Kasting,J.F.,Witmire,D.P.,&Reynolds,R.T.1993,“HabitableZonesaroundMainSequenceStars,”Icarus,101,108–128

Lebzelter, T., Wood, P., Hinkle, K. H. et al. 2005, “Long period variables in the globularcluster47Tuc:Radialvelocityvariations,”A&A,432,207–217

Li,Ch‐H.,Benedick,A.J.,Fendel,P.etal.2008,“Alaserfrequencycombthatenablesradialvelocitymeasurementswithaprecisionof1cm/s,”Nature,452,610‐612

Lovis, C., Mayor, M., Pepe, F. et al. 2006, “An extrasolar planetary system with threeNeptune‐massplanets,”Nature,441,305–309

Lovis, C., & Pepe, F. 2007, “A new list of thorium and argon spectral lines in the visible,”A&A,468,1115–1121

MacFarlane,G.G.,McLean,T.P.,Quarrington,J.E.,&Roberts,V.1958,“Finestructureintheabsorption‐edgespectrumofSi,”Phys.Rev.,111,1245

Mandushev,G.,Torres,G.,Latham,D.W.etal.2005, “TheChallengeofWide‐FieldTransitSurveys:TheCaseofGSC01944‐02289,”ApJ,621,1061–1071

Marcy,G.W.&Butler,R.P1992,“Precisionradialvelocitieswithaniodinecell,”PASP,104,270–277

Mayor, M., Pepe, F., Queloz, D. et al. 2003, “Setting New Standards with HARPS,” ESOMessenger,114,20

Mayor,M.&Queloz,D.1995,“AJupiter‐MassCompaniontoaSolar‐TypeStar,”Nature,378,355–359

McArthur,B.E.,Endl,M.,Cochran,W.D.etal.2004,“DetectionofaNeptune‐Massplanetintheρ1CancriSystemUsingtheHobby‐EberlyTelescope,”ApJ,614,L81–L84

Murphy,M.T.,Udem,Th.,Holzwarth,R.etal.2007,“High‐precisionwavelengthcalibrationofastronomicalspectrographswithlaserfrequencycombs,”MNRAS,380,839–847

O’Donovan, F., T., Charbonneau, D., Torres, G. et al. 2006, “Rejecting Astrophysical FalsePositivesfromtheTrESTransitingPlanetSurvey:TheExampleofGSC03885‐00829,”ApJ,644,1237–1245

Pepe, F. A., & Lovis, C. 2007, “From HARPS to CODEX: Exploring the limits of Dopplermeasurements, in Physics of Planetary Systems,” Proceedings of Nobel Symposium135: Physics of Planetary Systems, 18–22 June 2007, Stockholm, Sweden, Phys. Scr.T130,014007

Pont, F., Bouchy, F., Melo, C. et al. 2005, “Doppler follow‐up of OGLE planetary transitcandidatesinCarina,”A&A,438,1123–1140

Pont,F.,Tamuz,O.,Udalski,A.etal.2008,“Atransitingplanetamong23newnear‐thresholdcandidatesfromtheOGLEsurvey—OGLE‐TR‐182,”A&A,487,749–754

Ramsey,L.W.,Barnes, J.,Redman,S.L.etal.2008, “APathfinder Instrument forPrecisionRadialVelocitiesintheNear‐Infrared,”PASP,120,887–894

Santos, N. C., Israelian, G., & Mayor, M. 2004, “Spectroscopic [Fe/H] for 98 extra‐solarplanet‐hoststars.Exploringtheprobabilityofplanetformation,”A&A,415,1153–1166

RadialVelocity

203

Santos,N.C.,Mayor,M.,Naef,D.etal.2000,“TheCORALIEsurveyforSouthernextra‐solarplanets. IV. Intrinsic stellar limitations to planet searches with radial‐velocitytechniques,”A&A,361,265–272

Smith,V.,Tsuji,T.,Hinkle,K.H.etal.2003, “High‐Resolution InfraredSpectroscopyof theBrownDwarfεIndiBa,”ApJ,599,L107–L110

Torres,G.,Winn,J.N.,&Holman,M.J.2008,“Improvedparametersforextrasolartransitingplanets,”ApJ,677,1324–1342

Udry, S., Bonfils, X., Delfosse, X. et al. 2007, “The HARPS search for southern extra‐solarplanets.XI.Super‐Earths(5and8M⊕)ina3‐planetsystem,”A&A,469,L43–L47

VanderAuwera,J.,ElHachtouki,R.,&Brown,L.R.2002,“Absolutelinewavenumbersinthenearinfrared:12C2H2and12C16O2,”Mol.Phys.,100,3563

Vogt, S. S., Allen, S. L., Bigelow, B. C. et al. 1994, “HIRES: the high‐resolution echellespectrometerontheKeck10‐mtelescope,”Proc.SPIE,2198,362–375

Walker,G.,Yang,S.,Puxley,P.,etal.2003,“ADopplerSearchforHabitableSatellitesofεIndiB,” inTowards Other Planets: DARWIN/TPF and The Search for Extrasolar TerrestrialPlanets,22−25April2003,Heidelberg,Germany,eds.M.FriedlundandT.Henning,ESASP‐539,Noordwijk,Netherlands,647

Wright, J. T. 2005, “Radial Velocity Jitter in Stars from the California andCarnegie PlanetSearchatKeckObservatory,”PASP,117,657–664

Transits

205

8 Transits

DrakeDeming,GoddardSpaceFlightCenter,Chair

MarkSwain,JetPropulsionLaboratory,CoChairCharlesBeichman,DavidCiardi,JosephHarrington,SteveKilston

8.1 IntroductionTransits provide enormous astrophysical leverage to reveal the physical nature ofexoplanets.Thetransititself,supplementedbyhigh‐precisionradialvelocityobservations,gives us the mass and radius of the planet. This in turn constrains the planet's bulkcomposition. Absorption of starlight passing through the planet's atmosphere duringtransit(Figure8‐1)tellsusthecompositionandscaleheightof theexoplanetatmosphere(e.g.,Charbonneauetal.2002;Swainetal.2008b).Modulationofthecombinedlightofthesystem during secondary eclipse provides a direct detection of the planet's emergentspectrum(Figure8‐1,andCharbonneauetal.2005;Demingetal.2005,2006;Grillmairetal.2007;Richardsonetal.2007;Swainetal.2008a).Sincetheplanet'semergentradiationpeaks in the infrared (IR) spectral region, secondary eclipse measurements have beenprimarilyfocusedontheIR.

Figure81. Transit geometryprovides theastrophysical leverage that allowsdirectdetectionand in‐depthcharacterizationofclose‐inexoplanets.

Chapter8

206

Figure8‐2showsoneaspectofrecentscientificresultsfromtransitmeasurements,namelythe empiricalmass‐radius relation for exoplanets down to the size of Neptune (GJ 436b,Gillonetal.2007;Demingetal.2007).AsnotedonFigure8‐1,thenatureandoriginofgiantplanets with radii too large for their mass is currently a major open question inexoplanetaryscience.

Figure82. Mass‐radiusrelationfortransitingexoplanetstodate. Thesolidanddashedlinesaretheoreticalrelations from Bodenheimer et al. (2003), for planets with (solid line), and without (dashed line) a heavyelementcore. Notethatseveralgiantplanetshaveradiithatexceedeventhecorelessmodelsbysubstantiallymorethantheobservationalerrors.

Thedifferentgeometriesoftransitandsecondaryeclipseallowlocalizationofatmosphericknowledge: transit spectroscopy probes the atmospheric interface between the day andnight hemispheres of a tidally‐locked planet (Swain et al. 2008b), whereas secondaryeclipse measurements probe the emergent spectrum of the dayside. Moreover,measurementsof transiting systemshavebeenextendedwellbeyond the timesof transitandeclipse,toincludeobservationsinthecombinedlightofstarandplanetatalargerangeoforbitalphases.Ininstanceswheretheplanet'srotationistidallylockedtoitsorbit,thesemeasurements can be inverted to yield the distribution of emergent intensity versuslongitudeontheplanet(Knutsonetal.2007).

The observational techniques used for transiting systems can also be extended to non‐transiting systems, so it is valuable to consider a generalization of the transit technique,namelyexoplanetcharacterization incombined light. Withouta transit, theplanet radiuscannotbemeasureddirectly,andthatisasignificantlimitation.Nevertheless,muchcanbelearned, for example from observing fluctuations in IR intensity that are phased to theplanet'sknownradialvelocityorbit(e.g.,Harringtonetal.2006).

Transits

207

8.1.1 ScienceEnabledbyTransitsThe science enabled by transitmeasurements can be concisely summarized in Table 8‐1.Thequantityofresultsobtainedbythetransittechniqueislarge,andgrowingrapidly(andwillsoonmakethisChapterobsolete!)Virtuallyeverythingweknowaboutthephysicalnature of exoplanets has come from a combination of radial velocity and transitmeasurements.

8.1.2 MeasurementRequirementsThe principal requirement of transit measurements is high photometric andspectrophotometric precision, on relatively bright stars. This need for precision derivesfrom the fact that transitmeasurements aremade in the combined light of the star plusplanet. Hence the planet signal is greatly diluted by the stellar photons, and themeasurement precision must be as high as possible. Moreover, the time scale of thephotometric noise is crucial. Since transits are typically a few hours in duration, theprecision of individual measurements should improve as the inverse square‐root of themeasurement time, for times exceeding several hours. Since a photometric baseline isrequired before and after transit, a reasonable time scale for the required stability is~20hours.SincebothvisibleandIRmeasurementsareimportantfortransits(Figure8‐1),theinstrumentationshouldbedesignedtoreachthefundamentallimitsdeterminedbythestellarphotonnoise(primarily important inthevisible)andthebackgroundnoisecausedby the thermal emission of the instrument, telescope, and zodiacal dust (primarilyimportantintheIR).

Formoreextended(“around‐the‐orbit”)science,thephenomenabeingmeasuredhavetimescales of several days, so ~20 day stability is needed. Since there is also scientificmotivationtostudymuchlonger‐termchangesonexoplanets,dueforexampletoseasonaleffects for transiting planets in longer period orbits, then eventually there may bemotivationtoextendthestabilitytimeto~20months.

OrbitalPhase ScienceEnabled

Transit(primaryeclipse) RadiusmeasurementMassmeasurements(whencombinedwithradialvelocity)BulkcompositioninferredfrommassandradiusAtmosphericabsorptionspectroscopyDetectionofunseenplanetsviatimingvariationsMeasurementoftherelativeinclinationofstellarspinangularmomentumversusplanetaryorbitalangularmomentum,viatheRossiter‐McLaughlineffect

Secondaryeclipse MeasurementoftheemergentspectrumoftheplanetMeasurementoforbiteccentricity(orlimitsthereon)Ultra‐highspatialresolution,mappingthediskoftheplanet(inlongitude,expectedtobeachievedbyJWST)

Otherorbitphases LongitudinaltemperaturemapsandinferencesconcerningzonalwindsSpectroscopyoftheplanetatalllongitudinalaspects(inprinciple,butnotyetachieved)

Table81.Scienceenabledbytransits.

Chapter8

208

Ground‐based photometry has made great strides, and is achieving sub‐milli‐magnitudeprecision in favorable cases (Hartman et al. 2005). However, for most exoplanetcharacterizationscience,space‐borneobservationswillbenecessary.Thebestlocationfora space‐borne transit mission is heliocentric orbit, or placement at a Lagrangian point.Although significant transit science can be done from near‐Earth orbit, those orbits havetwo principal limitations: 1) long uninterrupted observing times, such as are needed forexoplanet around‐the‐orbit observations, are generally not possible from a near‐Earthlocation,and2)proximitytotheEarthresultsintime‐variablescatteredlightandthermalradiationthatcaninterferewithachievingthenecessaryprecision.

Figure83. Exampleof aSpitzer exoplanetdetection, the secondaryeclipseofHD189733b.Bydetecting theeclipseSpitzerisabletomakeadirectmeasurementoftheIRfluxfromtheplanet.(Demingetal.2006)

8.1.3 PreviousandPlannedTransitMissionsMajoradvancesinourknowledgeofthephysicalnatureofexoplanetshavecomefromtheGreat Observatories, principally Hubble and Spitzer. HST observations yielded the firstdetectiononanexoplanetatmosphere(Charbonneauetal.2002),andSpitzerobservationsgaveus"first light"forexoplanetemission(Charbonneauetal.2005;Demingetal.2005).Many follow‐upresultsaresummarizedbelow in thedescriptionofanewcombined lightmission and in reviews by Charbonneau et al. (2007), and Deming (2008), and oneparticular result is illustrated in Figure 8‐3. Although the Great Observatories were notspecificallydesignedforexoplanetobservations, theirgeneralcapabilitiesaresufficienttodeeplyexploitthetransitgeometryandproduceabundantcuttingedgescience.

SeveralNASAandESAmissionshavebeendedicatedtoexoplanettransitsinlargepartorintotal. The Kepler mission (Borucki et al. 2003) aims to determine the frequency ofoccurrenceofEarthanalogs,bymonitoring~100,000solar‐typestars(afieldinCygnus)forfour years. The CoRoT mission is similarly monitoring numerous solar‐type stars todetermine the frequency of occurrence of Super Earths, and to find and characterizeexoplanets ranging in size from giant planets, down to Super Earths. CoRoT has already

Transits

209

identified new giant transiting planets (e.g., Barge et al. 2008), and the CoRoT team iscurrentlyworkingtoextendthedetectionstomuchsmallerplanets.

InadditiontosearchmissionssuchasCoRoTandKepler,severalmissionshavefocusedonfollow‐up of bright transiting systems. TheMOSTmission has observed transits of giantplanetsorbitingbrightstars,andhasbeenparticularlysuccessfulatplacingverystringentlimits on the magnitude of reflected light from HD 209458b, a very low albedo planet.Similarly, theEPOXImission is concentratingon thepropertiesofgiantplanets transitingbrightstars,andissearchingthesesystemsforadditionalplanetsdowntothesizeofEarth(Christiansenetal.2008;Ballardetal.2008).

8.2 ObservatoryConceptsTransit missions fall into two general categories: search and characterization. Here wedescribeanall‐skytransitsearchmissionandtwoexamplesoftransitcharacterization.Inthecaseofthecharacterizationmissions,wefirstdescribeacombined‐lightmissionthatisquite general, and that can make inferences about the emergent spectra of close‐inexoplanets in combined light, even for systems that do not transit. We also describe acharacterization mission that extends transit spectroscopy into the ultraviolet spectralregion,wherethetransitabsorptionsignaturescanbequitelarge.

AnAllskyTransitSurvey:ThepurposeofamissionlikeKepleristoidentifyexoplanetsassmallasEarthinsize,bothinordertofindindividualplanetsandtogetstatisticaldataonthepopulationanddistributionofsuchplanets.Thisinformationcanthenbeusedtoguidefollow‐upstudiesoftheindividualexoplanetsandtoassistindesigningefficientfindingorcharacterizationmissions based on techniques other than transit observations. AlthoughKepler looks for transitingexoplanetsout toconsiderabledistance,about1000parsecs, ithasthelimitationoflookingtowardsonlyonedirectioninthegalaxy,alongtheOrionarmin theconstellationsofCygnusandLyra. It ispossible that terrestrialexoplanetstatisticscould be somewhat different in other regions of the galaxy, and some very interestingindividualplanetscouldbediscoveredbylookingtowardthoseregions.

WearemostinterestedinfindingplanetsrelativelynearbytoourSolarSystembecausetheconsequent greater observed brightness of the star and planet permit more detailedcharacterizationtobeachieved,soanall‐skysurvey,albeitlessdeepthanKepler,isneeded.Kepler observes stars typically of magnitudes 10 to 14 in a field just over 100 squaredegrees in size (about 1/400 of the sky). An all‐sky surveymerely out tomagnitude 12shouldfindatleasteighttimesasmanyplanetsasKeplerdoes,andtheonesfoundwould,onaverage,be four timescloser than theKeplerplanets. Therecent inference thatSuperEarthsarecommonaroundsolar‐typeand lowermain‐sequencestars(Mayoretal.2009)addsgreatlytothemotivationforanall‐skysurvey.Anall‐skysurveywouldfindtheclosesttransitingSuperEarthsor transitingEarthanalogs.TheMayoret al. (2009) results implythattherewillbealargenumberoftransitingSuperEarthsorbitingnearbysolar‐typestars,and theclosestexample is likely tobe theplanet that canbecharacterized to thehighestsignal‐to‐noiseratio.

Asanexampleofsuchapotentialsearchmission,MITscientists,withGoogle’ssupport,aredesigning a satellite‐based observatory to provide a sensitive survey of the entire sky tosearch for transiting exoplanets. The Transiting Exoplanet Survey Satellite (TESS) couldpotentially be launched in 2012 to discover new planetary systems that are nearby andbright.TESSwasrecentlyselectedbyNASAforaPhase‐AstudyundertheSMEXprogram.TESS could rapidly discover hundreds of planets with radii similar to the Earth. The

Chapter8

210

satellitewouldbeequippedwithsixhigh‐resolution,wide‐fielddigitalcameras,nowunderdevelopment. The cameras—whichhave a total resolution of 192megapixels—will coverthewholeskybytheendoftwoyears,observingabouttwomillionstars.Sincetheamountofdatacollectedwillbeenormous,onlyselectedportionswillbetransmittedbacktoEarth.Butremainingdatawillbestoredonthesatellite forabout threemonths,soastronomerscould check images of an unexpected event. The Harvard‐Smithsonian Center forAstrophysicsandtheOriginsofLifeInitiative,NASAGoddardandNASAAmes,andtheLasCumbres Observatory Global Telescope Network are already scientific participants withMITontheTESSprogram.

ACombinedLightCharacterizationMission(CLM):Dramaticmeasurementsmadewiththe Spitzer and Hubble space telescopes have redefined the field of exoplanetcharacterization (e.g.,Barman2007;Beaulieuet al. 2008;Charbonneauet al. 2002,2005;Demingetal.2005,2006,2007;Demoryetal.2007;Gillonetal.2007;Grillmairetal.2007;Harringtonetal.2006,2007;Knutsonetal.2007,2008;Richardsonetal.2007;Swainetal.2008a,b);collectively,thisworkhasdecisivelyestablishedthatdetailedcharacterizationofexoplanetatmospheresisfeasible.Today,duetotheextraordinaryandunforeseensuccessof theSpitzerandHubble spacetelescopes,wecandiscusstheobservationalsignaturesofexoplanet atmospheres includingweather, vertical and longitudinal temperature profiles,molecular abundances (including prebioticmolecules), dayside to night side atmosphericchemistrychanges,and theroleofphotochemistry. Ofequalsignificance to thisscientificadvanceisthemethodbywhichthebreakthroughsweremadewhereinthecombinedlightstar‐planetsystemismeasured. SpitzerandHubblespacetelescopemeasurementsdonotspatiallyresolvetheexoplanets;rather, informationabouttheexoplanet isextractedfromprecisemeasurements in the “light curve”or changes in the spectrophotometric intensityarising fromtheplanetduring itsorbit. Further,observationsofdifferentportionsof thelight curve have demonstrated the ability to localizemolecular abundances to specificregions of the exoplanet atmosphere and explore, for example, the differences betweendaysideandnightsideatmosphericchemistry.TheSpitzerandHubbleresultsdemonstratethatwehave, today, the flight‐proventechnologyrequiredtostudyprebioticchemistryoforganicmolecules in habitable zone exoplanets and to localize this knowledge to specificregions of the atmosphere. This is a completely unexpected yet extraordinaryaccomplishment for the Spitzer andHubble space telescopes, and itwill be an importantpartoftheirlegacy.

Building on the extraordinary successes of Spitzer and Hubble, we advocate that NASAconsideracombined‐lightexoplanetcharacterizationmission.Thismissionwouldfocusonthe detection of molecules in exoplanet atmospheres. By detecting and determining theabundancesofmolecules,theprincipalgoalofacombined‐lightmissionwoulddeterminethe conditions, composition, and chemistry of exoplanet atmospheres. Ahighpriorityforthecombined‐lightmissionisobservinghabitablezoneplanetsfor(1)thedetectionofprebiotic molecules and (2) characterization of the nature of photochemical andthermochemical contributions to atmospheric carbon chemistry. Given an appropriatetarget,suchasanexoplanet ina15‐to25‐dayorbitaroundanearbyMstar,acombined‐lightmissioncoulddetermineadetailedpictureof theatmospheric chemistryof aplanetwhere life could exist. A combined‐light mission could also characterize the surfaces ofrockyplanets.

Transits

211

Figure84.Acomparisonofobservations(blacktriangles)withamodelofwater(blue)andwater+methaneinan absorption spectrum of the planet HD 189733b. These data constitute the first detection of an organicmolecule in an exoplanet atmosphere andwere obtainedwith theHubble Space Telescope using theNICMOSinstrument. A combined‐light exoplanet characterizationmission coulddeliverdataof similarquality for themanyexoplanetsandwouldextendthespectralcoverageintothethermalinfrared.(Swainetal.2008b)

Determining the conditions, composition, and chemistry of exoplanet atmosphereswouldbeaccomplishedbyinterpretingtheobservedspectrausingretrievalmethods(Tinettietal.2007) to extract molecular abundances and the pressure‐temperature profiles; thisknowledge would be localized by repeating the retrieval for different portions of theplanet’sorbit. Examplesof importanttargetmoleculesareH2O,CH4,CO,CO2,andNH3;allthesespecieshavestrongabsorptionbandsinthe2–5‐µmwavelengthrange.Eachofthesemoleculesprobestheatmospheresindifferentways.COandCH4areprimaryreservoirsofcarbon; theCO/CH4ratio issensitive to temperature.CO2probesthevertical temperatureprofileandtheverticalmixingratio(potentiallyadiagnosticofphotochemistry)usingtheV=13–15bands.CO2alsoservesasaproxyforCO,whileCOandCH4abundancesdifferingfrom thermochemical equilibriummight indicate photochemistry. For tidally locked, hot‐Jovian planets, the chemistry of the atmosphere is expected to change, particularly the[CO]/[CH4] ratio, as a function of orbital phase and, in some cases, a dayside hotstratospheremightform(Burrowsetal.2007;Fortneyetal.2008;Knutsonetal.2007);the[CO]/[CH4] ratio at the terminator is also a potential diagnostic of variability, possiblyassociatedwithlarge‐scaleeddiesgeneratedbyzonalwinds.Althoughatmosphericheatingwill be characterized by Spitzer photometry, spectroscopy is required to probe how theatmospherechemistrychangesfromthedaysidetothenightside.Thetargetsaccessibletoa combined‐light mission cover a range of exoplanet sizes, temperatures, and stellarspectral types and will reveal the influences of the stellar primary on the exoplanetatmosphere. Thus, a combined‐lightmissionwould enableus tounderstand the effect ofradiationforcingfromthestellarprimaryonphotochemistryandstratosphereformation.

Chapter8

212

Figure 85. The combined‐light mission concept is an intuitively straightforward, low‐risk design. A 1.4 m‐diameter passively cooled telescope provides light collection for two actively cooled spectrometers, onecovering2–5µm(SWIR)andtheother5–12µm(MWIR).Readilyavailablespacecraftcorespossessthepower,bus‐level pointing, and other capabilities required by a combined‐light mission. A Spitzer‐like solarpanel/thermalshieldcombinationmakesefficientuseofmassandvolumeresources.(Y.‐H.Wu,JPL)

8.2.1 ArchitectureThearchitectureofaCombined‐LightMissioncanbeveryconventional.CLMisenvisionedasa1.2‐to1.4‐mtelescopeinanEarth‐trailingorL2orbit. Thistypeoforbit isprobablyessential forachieving instrumentstability. Thetelescopewouldbepassivelycooled,andmechanical coolers would be used to cool the focal plane. The baseline mission lengthwouldbe5years,butwithnostoredcryogens,anextendedmissionispossible.Thescienceinstrumentwouldbeaspectrometerwithselectabledispersion. Thebaselinewavelengthrangeis2–15μmalthoughmissionswithreducedmid‐IRcoveragearestillvaluable.The2–15μmwavelengthrangewouldbesplitbetweentwofocal‐planearrays(FPA),one for2–5μmandonefor5–15μm.EachFPAwouldhavedispersiveopticsmatchedforthatband.ThedispersionisselectedsuchthattheFPAisphoton‐noiselimitedineverypixelforsolar‐typestarsup tosomedistance(~100pc). Thisconfigurationalsohas theadvantage thatthere are no moving components required to change spectral resolution. Thus, theinstrument couldbedesigned toonlyhaveoneobservingmode (full spectral resolution).Thisincreasestheinstrumentstabilityandsimplifiesthecalibrationprocess.CalibrationofCLMwouldbeaccomplishedbymonitoringofacalibratornetworkthroughoutthemissionlifeand,possibly,byrelyingonanon‐boardcalibrationsource. Thedatarateis low. TheCLMwouldprobably incorporatea fast‐steeringmirror in the science instrument for fineguidance and for precision placement of the science and calibration targets. The fast‐steering mirror control would use visible light imaging with a CCD. At the shortestwavelengths and the lowest spectral resolutions, the CLM has the potential to achieve ameasurementdynamicrangeof~105.Photometrywithadynamicrangeof~5000:1at2μmispossiblefora12thmagnitudeGstar,makingthephotometricstudyofexoplanetsintheKeplerfieldanimportantpartoftheCLMmission.

Transits

213

8.2.2 PerformanceThe performance of all transitmissions, and the CLM in particular, should be specificallydesignedtoapproachthefundamentallimitsimposedbystellarphotonnoiseandintrinsicstellar variability noise. On the time scales of transits, the stellar variability will bedominatedbythephotonnoise.Spitzer,forexample,routinelyachieves~70‐to80‐percentof the photon noise limit for transit and eclipse photometry (e.g., Harrington et al. 2007;Demingetal.2007),sodesigningthemissiontospecificallyachievethisgoalwillgiveevengreaterconsistencyofperformance.

Asidefromtheperformanceofthemissionperse, it isinstructivetoconsiderhowagivenlevel of performance (in terms of signal‐to‐noise for a given brightness) can be usedoptimallybyinsightfulchoiceamongdifferenttransitingsystems.Sincethelightemergentfromatransitingexoplanetismeasuredusingasecondaryeclipsetechnique,considerhowtheamplitudeoftheeclipsescaleswithdifferentplanet‐starcombinations.IntheRayleigh‐Jeans(longwavelength)limit,thedepthofasecondaryeclipse,inunitsofthestellarflux,isgivenasAλ=(Rp/Rs)2(Tp/Ts),whereR isradiusandT isbrightnesstemperature,andtheplanetandstarareindexedbypands,respectively.Inthermalequilibrium,thebrightnesstemperatureoftheplanetisrelatedtothestellartemperatureasTp=α Tsθ1/2,whereαisaconstant that contains theplanet'salbedo, circulationproperties, etc. andθ is theangulardiameterofthestarasseenfromtheplanet. Asweproceeddownthemainsequence,thesmaller sizes of the stars causes the (Rp/Rs)2 to dominate overθ1/2. So planets orbitingsmallstarswillhavemoreprominenteclipses,andthisisonebasisformuchoftherecentinterestinSuperEarthsorbitingM‐dwarfs.TheCLMandothertransitmissionswouldhavealargefocusonM‐dwarfplanetsforthisreason.

Figure86.Modelforthenormalizedin‐transitminusout‐of‐transitspectradisplayingsomeoftheprominentatomic species in theUV andvisible. The twopanels show caseswithdifferent cloud‐topheights. (Seager&Sasselov2000)

Chapter8

214

AnUltravioletTransitMission:Transitingexoplanetshaveallbeendiscoveredatopticalwavelengths and themajority of characterization studies have been performed at opticaland infraredwavelengths. Optical and infraredwavelengths yield information about theatmosphericalbedo(e.g.,Roweetal.2006),atomicabundances(e.g.,Redfieldetal.2008),molecular abundances (e.g., Tinetti et al. 2007) and thermal radiation of the planet (e.g.,Knutson et al. 2007), and shorter wavelengths can yield information about haze, atomicspecies,andcondensedparticles in theatmosphereand the interactionof theplanetwiththestars.

The first detection of an atomic species was made by observing the Na II doublet inabsorption against the stellar host HD 209458 (Charbonneau et al. 2002). However, atultravioletwavelengths, absorption‐line studies of other species can provide informationabout the atomic and ionic species content of the planetary atmosphere. Various atomicspeciesarevisibleinthestellarspectraincludingNa,K,andLi(Figure8‐6)andcanprobedifferent layers and physical conditions of the planetary atmosphere (Seager & Sasselov2000).

Forgiantplanets in shortorbitalperiods (i.e., thehot Jupiters), theobservedatmospheretransitionsfrombeingdominatedbytheplanetarythermalemissiontobeingdominatedbyreflected emission from the star (Figure 8‐7). By studying the emission at shorterwavelengths,wherereflectancedominatestheexoplanetemission,studiesoftheplanetaryhaze and cloud content can be performed. The scattering and polarization signaturesbecomelessdilutedbythermalemissionastheobservationwavelengthsgrowshorter.

Figure87. Predictedspectrum forHD209458b from0.3–30µm,alongwithobservationaldata fromSpitzerandMOST. Solid lines are thermal emission while dashed lines are reflected stellar flux. The colored linesrepresentmodelswithdifferingopacitysources.(Fortneyetal.2008)

High‐precision polarization studies can yield detailed information about the cloudparticulate content, sizes, and cloud‐layer distributions (e.g., Seager,Whitney, & Sasselov2000; Hough et al. 2006). For example, below 0.2 µm, Rayleigh scattering from H2 isimportant. In general, photometric and polarization light curves are a function of the

Transits

215

atmosphericopacityandtheobservationalviewingangleandcan,inprinciple,yielddetailsabouttheday/nightvariationsofthecloudlayersandparticulates(e.g.,Seager,Whitney,&Sasselov 2000). The first polarization detection of a planetary atmospherewas recentlyclaimedbyBerdyuginaetal.(2008)wheretheyconcludedthatHD189733bat(~0.4µm)mayhavea radius that is30% larger than theopticallyopaqueplanetarydisk, indicatingthat scattering material may fill the Roche lobe, possibly escaping from the planetaryatmosphere.

AbsorptionlinestudiesofLyman‐αsuggestthatHD209458bcontainsalargeexosphereofhydrogen (Figure 8‐8). Interpretations of this phenomenon include evaporation of theplanetary atmosphere, where the particles are accelerated by radiation pressure (Vidal‐Madjar et al. 2008). Other interpretations require an interactionwith the stellarwind toachievetheobservedvelocities(Holmstrometal.2007).Ineithercase,itseemsclearthattheplanetandthestarmaybeinteracting,andUVspectroscopyhasenabledthediscovery(andabetterunderstanding)oftheprocess.

Figure88.ObservedHD209458Ly‐αprofiles inandoutof transit, asobservedbyVidal‐Madjaretal.2003,suggestingthathydrogenisevaporatingfromtheplanetaryatmosphere.(Vidal‐Madjaretal.2008)

Theinteractionbetweentheplanetandthestarmayhaveimportantconsequencesfortheboththestellarhostandtheexoplanet.RecentobservationsofTauBoowithMOSTsuggestthatstellarsurfaceactivitymaybecorrelatedwith theplanetaryorbit, indicating that theplanet may be inducing stellar activity on the surface of the star (Matthews, privatecommunication). Further, Penz et al. (2008) suggests high energy stellar flux (X‐ray andUV)maybestrongenoughevaporatetheplanetsataratelargeenoughforthegiantplanetsto turnevolve intohotNeptunes (Penzet al. 2008). Detailed studiesof theexoplanetaryexosphere at the shortwavelengthsmay be crucial to our understanding of how planetsinteractwiththeirstarsandevolveongeologicaltimescales.

Chapter8

216

8.3 TechnologyThecombined‐lightcharacterizationmission is themost technologicallydemandingof thethree mission concepts described above. Nonetheless, it does not require dauntingtechnologydevelopment todeliverhigh‐impactexoplanetcharacterizationscience; in thisway, it is strongly differentiated from exoplanet characterization missions using othertechniques such as high‐contrast imaging. Over periods of several hours, combined‐lightspectroscopymeasurementshavebeendemonstratedwithadynamicrangeof40,000:1inthevisible(Pontetal.2007)and20,000:1inthenear‐IR(Swainetal.2008b),whilemid‐IRbroad‐bandphotometryhasdemonstrated~30,000:1dynamicrange(Knutsonetal.2007).Thenear‐IRandmid‐IRmeasurementsareachievingperformancethat isapproachingthephoton‐noiselimit.Moreover,thislevelofperformancehasbeenachievedbymissionsthathavenotbeenspecificallydesignedforstabilityonbrightstars.Rather,thesemissions(e.g.,HSTandSpitzer)weredesignedtooptimizethesensitivityforfaintsources.Thus,meetingthe 100,000:1 dynamic range goal for a combined‐light exoplanet characterizationmission—that would be specifically designed for that application—seems quite feasibleusingcarefulapplicationofexistingtechnology.However,thereareafewsimpletechnologyimprovementtasksthatwouldresultinbetterinstrumentperformance.

The primary requirement common to all transit missions is a stable observatory andinstrument configuration (no moving spectral dispersion elements). Determining thestability of the spectral response function (as set by variable thermal conditions) in low‐Earth orbit vs. orbits with better thermal stability (Earth trailing or L2) could have animportant impact on mission design. The design of such a mission should also payparticularattentionto:

• FPA characterization formodeling effects such as long‐term responsivity changes,charge‐trapping, and effects such as intra‐pixel variations in detector quantumefficiencyandcharge‐transferefficiency.

• Incorporation of a fine‐steering mirror (FSM) in the instrument for preciseplacementofthetargetstarinthespectrometerentranceslit.Sincethetargetstarsarerelativelybright,theFSMerrorsignalcouldbegeneratedbyavisible‐lightCCD.

• Ahigh‐stability,on‐board,calibrationsource.Althoughnotessential,theavailabilityof on‐board calibration could relax the requirements on a calibrator starnetworkandimproveobservingefficiency.

8.3.1 PastAccomplishmentsTwocurrenttransitmissionshavespecificallydevelopedtechnologythatwilllikelybeverybeneficialtofuturetransitmissions.Fortransitmeasurements,mathematicaltechniquestoperform high‐precision data analysis (e.g., decorrelation techniques), and techniques toexecute observations in an optimal manner, carry an importance on a par with, orexceeding, the development of new hardware. The Kepler mission has pioneeredmeasurementtechniquestoobtainveryhighprecisionCCDphotometry.Thesetechniquesinclude differential spatial and temporal photometry, and decorrelation methods tocompensate for the effects of image motion. Kepler implemented a photometry testbed(Kochetal.2000)anddemonstratedaphotometricprecisionof10–5forCCDphotometryinthelaboratorypriortodesigningandbuildingflighthardware.

TheMOSTmission(Matthewsetal.1999)hasalsousedspecificmeasurementtechniquestoobtainveryhigh‐photometricprecision. These techniques include imagingof thepupilat

Transits

217

the focal plane to reduce the sensitivity of the photometric signal to spacecraft pointingjitter. Advanced transit spectroscopy missions such as the CLM (described above) willbenefitfromsimilarpupilimaging,butinadispersedmodeforspectroscopy.

8.3.2 FutureTechnologyDevelopment:Milestones?The success of HST and Spitzer for transit science shows in the first place that currenttechnology and careful attention tomission and instrumentdesign go a longway towardachieving excellent transit science. The success of MOST, and the (so far) successfuldevelopmentofKepler, furthersuggestthatinnovativeopticaldesignanddetectortestbeddevelopmentareappropriatevehiclesforachievingtransitmissionsciencegoals.Thistypeofmission‐specifictechnologydevelopmentisthebestmodelfortransitandcombined‐lightmissionscience,asopposedtoaprogramhavingtechnology‐developmentmilestonesthatmight be more broadly applicable. This would, in our opinion, remain true for transitmissiondevelopmentuptothe"ProbeClass,”whichisinfactnotfarremovedinfiscaltermsfromthedevelopmentofKepler.

8.4 Research&AnalysisGoalsTo extract maximum science from transit studies, it will be necessary to implement avigorousResearch&Analysisprogramto:

o Support the Spitzer warm mission and extensive Hubble exoplanet observations,thereby extending the current progress in exoplanet characterization, and thusrefining the science case for a combined‐light exoplanet characterizationmission.Examples of direct scientific support for these missions include ground‐basedtelescopic detections of secondary eclipses (e.g., at 2 µm), and ground‐basedtransmissionspectroscopy(e.g.,Redfieldetal.2008).

o Supportongoingtransitandradial‐velocitysurveysthatfindgianttransitingplanetsfromtheground,andsupportextensionsof thesesurveysto lowermain‐sequencestarssuchasM‐dwarfs(Nutzman&Charbonneau2008).

o Support precursor projects (such as a balloon‐borne instrument) that woulddemonstrateimportantaspectsofthetechnologyandsciencecaseforacombined‐lightexoplanetcharacterizationmission.

o Supportthemodesttechnologydevelopmentefforts(FPAcharacterizationandfine‐steering mirror technology), which are beneficial for a combined‐light exoplanetcharacterizationmission.

o Support a detailed study of the complementary roles of JWST and a dedicated,combined‐lightexoplanetcharacterizationmission.

Chapter8

218

8.5 ContributorsDrakeDeming,NASAGoddardSpaceFlightCenter

MarkSwain,JetPropulsionLaboratory

CharlesBeichman,JetPropulsionLaboratory


JosephHarrington,UniversityofCentralFlorida

SteveKilston,BallAerospaceCorporation

8.6 ReferencesBallard,S.,Charbonneau,D.,A'Hearn,M.,etal.2008“PreliminaryResultsonHAT‐P‐4,TrES‐

3,XO‐2,andGJ436fromtheNASAEPOXIMission,”inProc.IAUSymp.S253,TransitingPlanets(CambridgeUniversityPress:Cambridge,UK),470–473.

Barge,P.,Baglin,A.,Auvergne,M.,etal.2008,“TransitingExoplanetsfromtheCoRoTSpaceMission,I.CoRoT‐Exo‐1b:alow‐density,shortperiodplanetaroundaG0Vstar,”A&A,82,L17–L20

Barman, T. 2007, “Identification of Absorption Features in an Extrasolar PlanetAtmosphere,”ApJ,661,L191–L194

Beaulieu, J. P., Carey, S., Ribas, I., & Tinetti, G. 2008, “Primary Transit of the Planet HD189733bat3.6and5.8Microns,”ApJ,677,1343–1347

Berdyugina, S. V., Berdyugin, A. V., Fluri, D. M., & Piirola, V. 2008, “First Detection ofPolarizedScatteredLightfromanExoplanetAtmosphere,”ApJ,673,L83–L86

Bodenheimer,P.,Laughlin,G.,&Lin,D.2003,“OntheRadiiofExtrasolarGiantPlanets,”ApJ,592,555–563

Borucki, W. J. & 14 co-authors 2003, “The Kepler Mission,” ASP Conf. Ser. 294, 427

Burrows, A., Hubeny, I., Budaj, J., Knutson, H. A., & Charbonneau, D. 2007, “TheoreticalSpectral Models of the Planet HD 209458b with a Thermal Inversion and WaterEmissionBands,”ApJ,668,L171–L174

Charbonneau, D., Brown, T. M., Noyes, R. W., & Gilliland, R. L. 2002, “Detection of anExtrasolarPlanetAtmosphere,”ApJ,568,377–384

Charbonneau, D., Allen, L. E.,Megeath, S. T., et al. 2005, “Detection of Thermal EmissionfromanExtrasolarPlanet,”ApJ,568,377–384

Charbonneau,D.,Brown,T.M.,Burrows,A.,&Laughlin,G.2007,“WhenExtrasolarPlanetsTransitTheirParentStars,”inProtostarsandPlanetsV,Univ.ArizonaPress,Tucson,AZ.

Christiansen, S., Charbonneau, D., A'Hearn, M., et al. 2008, “The NASA EPOXI Mission ofOpportunity to Gather UltraPrecise Photometry of Known Transiting Exoplanets,” inProc.IAUSymp.S253,TransitingPlanets(CambridgeUniversityPress:Cambridge,UK),301–307.

Transits

219

Cowan,N.,Agol,E.,&Charbonneau,D.2007,“HotNightsonExtrasolarPlanets:Mid‐infraredPhaseVariationsofHotJupiters,”MNRAS,379,641–646

Deming, D. 2008, “Emergent Exoplanet Flux: Review of the Spitzer Results,” in Proc. IAUSymp.S253,TransitingPlanets(CambridgeUniversityPress:Cambridge,UK),197–207.

Deming,D.,Seager,S.,Richardson,L. J.,&Harrington, J.2005, “InfraredRadiation fromanExtrasolarPlanet,”Nature,434,740–743

Deming, D., Harrington, J., Seager, S., & Richardson, L. J. 2006, “Strong Infrared EmissionfromtheExtrasolarPlanetHD189733b,”ApJ,644,560–564

Deming, D., Harrington, J., Laughlin, G., et al. 2007, “Transit and Secondary EclipsePhotometryofGJ436b,”ApJ,667,L199–L202

Demory,B.‐O., Gillon,M., Barman,T., et al. 2007, “Characterizationof theHotNeptuneGJ436bwithSpitzerandGround‐basedObservations,”A&A,475,1125–1129

Ehrenreich,D.,Hebrard,G.,desEstangs,L., etal.2007, “ASpitzerSearch forWater in theTransitingExoplanetHD189733b,”ApJ,668,L179–L182

Fortney, J. J., Lodders,K.,Marley,M. S.&Freedman,R. S. 2008, “AUnifiedTheory for theAtmospheres of the Hot and Very Hot Jupiters: Two Classes of IrradiatedAtmospheres,”ApJ,678,1419–1435

Gillon, M., Demory, B. O., Barman, T., et al. 2007, “Accurate Spitzer Infrared RadiusMeasurementfortheHotNeptuneGJ436b,”A&A,471,L51–L54

Grillmair, C. J., Charbonneau, D., Burrows, A., et al. 2007, “A Spitzer Spectrum of theExoplanetHD189733b,”ApJ,658,L115–L118

Harrington, J., Hansen, B. M., Luszcz, S., et al. 2006, “The Phase‐Dependent InfraredBrightnessoftheExtrasolarPlanetUpsilonAndromedaeb,”Science,314,623–626

Harrington,J.,Luszcz,S.,Seager,S.,etal.2007,“TheHottestPlanet,”Nature,447,691–693

Hartman, J.D., Stanek,K. Z.,Gaudi,B. S., et al. 2005, “Pushing theLimitsofGround‐basedPhotometric Precision: Sub‐milli‐magnitude Time‐Series Photometry of the OpenClusterNGC6791,”AJ,130,2241–2251

Holmstrom, M., Ekenback, A., Selsis, F. 2007, “Energetic Neutral Atoms Around theExtrasolarPlanetHD209458b,”AGUFallMeeting,#SH12A‐0857

Knutson,H.A.,Charbonneau,D.,Allen,L.E.,etal.2007,“AMapoftheDay‐NightContrastoftheExtrasolarPlanetHD189733b,”Nature,447,183–186

Knutson,H.A.,Charbonneau,D.,Allen,L.E.,etal.2008,“The3.6–8.0BroadbandEmissionSpectrumofHD209458b: Evidence for anAtmospheric Temperature Inversion,” ApJ,673,526–531

Koch,D.G.,Borucki,W.J.,Dunham,E.W.,etal.2000,“CCDPhotometryTestsforaMissiontoDetectEarth‐sizedPlanets in theExtendedSolarNeighborhood,”Proc. SPIE,4013,508–519

Matthews, J., Kuschnig, R., Walker, G., et al. 1999, “The MOST Space Mission: a 15‐cmTelescopeinthe8‐mClassEra,”J.Astron.Soc.Can.,93,183–184

Chapter8

220

Mayor,M.,Udry,S.,Lovis,C.,etal.2009,“TheHARPSSearchforSouthernExtrasolarPlanets.XIII.Aplanetarysystemwith3Super‐Earths(4.2,6.9,&9.2Earthmasses),”A&A,493,639–644

Nutzman,P.,&Charbonneau,D. 2008,“DesignConsiderationsforaGround‐BasedTransitSearchforHabitablePlanetsOrbitingM‐Dwarfs,”PASP,120,317–327

Penz,T.,Erkaev,N.V.,Kulikov,Yu.N.,etal.2008,“MassLossfromHotJupiters:Implicationsfor CoRoT Discoveries, II. Long Time Thermal Atmospheric Evaporation Modeling,”Plan.&SpaceSci.,56,1260–1272

Pont, F., Gilliland, R. L., Moutou, C., et al. 2007, “Hubble Space Telescope Time‐SeriesPhotometry of the Planetary Transit of HD 189733: No Moon, no Rings, Starspots,”A&A,476,1347–1355

Redfield,S.,Endl,M.,Cochran,W.D.,etal.2008,“SodiumAbsorptionfromtheExoplanetaryAtmosphereofHD189733bDetectedintheOpticalTransmissionSpectrum,”ApJ,673,L87–L90

Richardson,L.J.,Deming,D.,Horning,K.,Seager,S.,&Harrington,J.2007,“ASpectrumofanExtrasolarPlanet,”Nature,445,892–895

Rowe, J. F., Matthews, J. M., Seager, S., et al. 2006, “An Upper Limit on the Albedo of HD209458b:DirectImagingPhotometrywiththeMOSTSatellite,”ApJ,646,1241–1251

Seager,S.,&Sasselov,D.D.2000,“TheoreticalTransmissionSpectraduringExtrasolarGiantPlanetTransits,”ApJ,537,916–921.

Seager,S.,Whitney,B.A.,&Sasselov,D.D.2000,“PhotometricLightCurvesandPolarizationofClose‐inExtrasolarGiantPlanets,”ApJ,540,504–520

Swain,M.R.,Bouwman,J.,Akeson,R.L.,etal.2008a,“TheMid‐InfraredSpectrumoftheTransitingExoplanetHD209458b,”ApJ,674,482–497

Swain,M.R.,Vasisht,G.,&Tinetti,G.,2008b,“ThePresenceofMethaneintheAtmosphereofanExtrasolarPlanet,”Nature,452,329–331

Tinetti,G.,Vidal‐Madjar,A.,Liang,M.‐C., etal.2007, “WaterVapor in theAtmosphereofaTransitingExtrasolarPlanet,”Nature,448,169–171

Vidal‐Madjar, A., Lecavelier des Etangs, A., Desert, J. M., et al. 2003, “An Extended UpperAtmosphereAroundtheExtrasolarPlanetHD209458b,”Nature,422,143–146

Vidal‐Madjar,A.,LecavelierdesEtangs,A.,Desert,J.M.,etal.2008,“ExoplanetHD209458b(Osiris):EvaporationStrengthened,”ApJ,676,L57–L60

MagnetosphericEmission

221

9 MagnetosphericEmission

J.Lazio,NavalResearchLaboratory,Chair

D.Winterhalter,JetPropulsionLaboratory,CoChairT.Bastian,G.Bryden,W.M.Farrell,J.M.Griessmeier,J.Kasper,T.Kuiper,A.Lecacheux,W.Majid,R.Osten,E.Shkolnik,P.Zarka

9.1 IntroductionAlloftheSolarSystemgiantplanetsandtheEarthproduceradio‐wavelengthemissionsasaresultof an interaction between theirmagnetic fields and the solarwind. In the case of the Earth, itsmagnetic field may contribute to its habitability by protecting its atmosphere from solar winderosion andbypreventing energetic particles from reaching its surface. Indirect evidence for atleastsomeextrasolargiantplanetsalsohavingmagneticfieldsincludesthemodulationofcalciumemission lines of their host stars phased with the planetary orbits likely due to magneticreconnection events between the stellar and planetary fields. If magnetic fields are a genericpropertyofgiantplanets,thenextrasolargiantplanetsshouldemitatradiowavelengthsallowingfor theirdirectdetection. In the caseof terrestrial‐massplanets, ifmagnetized, they should alsoemit at radio wavelengths, and detecting this radiation may be a means of assessing theirhabitability. Existingradio‐wavelengthobservationsplace limitscomparable to the fluxdensitiesexpectedfromthestrongestemissions. Futureradio‐wavelengthfacilities,currentlyunderdesignanddevelopmentorconstruction,willoffermorethanoneorderofmagnitudeimprovement.

9.1.1 ScienceGoalsSearchforandexploitextrasolarplanetarymagnetosphericemissionsasameansofdirectlydetectingandcharacterizingthoseplanets.

PlanetaryMagnetosphericEmissionThe Earth and gas giants of our Solar System are described commonly as “magnetic planets”because they contain internal dynamo currents that generate planetary‐scale magnetic fields.Thesemagneticfieldsareimmersedinthesolarwind,asupersonicmagnetizedplasma.Thesolarwinddeformstheplanetarymagneticfield,compressingthefieldonthefrontsideandelongatingiton the back, forming a “tear‐dropped”–shapedmagnetosphere aligned with the solar wind flow(Figure 9‐1) Themagnetopause forms the boundary between themagnetosphere, in which theplanet's magnetic field is dominant, and the solar wind. The stellar wind incident on themagnetopauseisanenergysourcetotheplanetarymagnetosphere.

The magnetospheres of the Solar System's magnetic planets host radio‐wavelength masers,generatedbyelectroncyclotronradiation(so‐called“electroncyclotronmasers”).Themechanism

Chapter9

222

is that the magnetosphere‐solar wind interaction produces energetic (keV) electrons that thenpropagate alongmagnetic field lines into auroral regions, where an electron cyclotronmaser isproduced.

Figure91.AnillustrationoftheinteractionbetweentheSunandEarth'smagnetosphere(nottoscale).Inthisexample,aneruptionfromtheSunhasreachedtheEarthandcompressedthemagnetosphere,injectingenergeticparticlesintoit.(CourtesyofNASA)

Specific details of the electron cyclotron maser emission vary from planet to planet, dependinguponsuchsecondaryeffectsastheplanet'smagneticfieldtopology.Nonetheless,applicabletoallof themagnetic planets is amacroscopic relation relating the incident solarwindpowerPsw, theplanet's magnetic field strength, and the median radio luminosity Lrad (Figure 9‐2). Variousinvestigators(e.g.,Millon&Goertz1988)find

Lrad=εPswx,

withεtheefficiencyatwhichthesolarwindpowerisconvertedtoradioluminosity,andx≈1.Thevalueforεdependsonwhetheroneconsidersthemagneticenergyorkineticenergy,respectively,carried by the stellar wind. The strong solar wind dependence is reflected in the fact that theluminosityoftheEarthislargerthanthatofeitherUranusorNeptune,eventhoughtheirmagneticfieldsare10–50timesstrongerthanthatoftheEarth.

The incident solarwindpowerdependsupon the rampressure of the solarwind and the cross‐sectional areaof themagnetosphere. In general, themagnetopausehas a dynamic configurationdetermined by the instantaneous solar wind flow, and the electron cyclotron maser is anexponentialamplifiersothatmodestchangesinsolarwinddensityorvelocityorboth(e.g.,factorofa few)canchangeplanetarymagnetosphericemissionsby large factors(>100). However, theaveragecross‐sectionalareadependsuponthestrengthoftheplanet'smagneticfield,whichcanbeestimatedfromvariousplanetaryparameters,andontheaveragestellarwindconditions.

Theelectroncyclotronmaseroccursbelowacharacteristicemissionfrequencydeterminedbythecyclotron frequency in the magnetic polar region, which in turn depends upon the planet'smagneticmomentormagneticfieldstrength.UsingsimilarscalinglawsbasedontheSolarSystemplanets,onecanpredictthischaracteristicfrequency.Forreference,thecharacteristicfrequencyofJupiter is approximately 40 MHz (≈ 7.5 m wavelength). Above the characteristic frequency, a


223

planet’s luminosity drops by several orders of magnitude to levels of incoherent synchrotronemission.

Thesescalingrelationsarenotonlydescriptivebutalsopredictive.BeforeVoyager2reachedbothUranusandNeptune,theirluminositieswerepredicted(Desch&Kaiser1984;Desch1988;Millon& Goertz 1988). For both planets, the predictions were in excellent agreement with themeasurements.

Figure92. (Left)MagneticmomentsofSolarSystemplanetsasa functionoftherotationrateωandmassM. (Right)Radio luminosities of Solar System planets as a function of the incident solar wind kinetic power. Even though themagneticfieldstrengthofEarthislessthanthatofUranusorNeptune,itismoreluminousbecauseitisclosertotheSun,andalargersolarwindpowerisincident.(AdaptedfromZarkaetal.2001)

IndirectevidenceforextrasolarplanetarymagneticfieldsisfoundinmodulationsoftheCaIIHandK linesof the starsHD179949andυ And—modulations inphasewithplanetaryorbital periods(Shkolniketal.2005,2008).PhotometricobservationsbytheMOSTspacetelescopeofseveralhot‐Jupiter systems, including HD 179949 and τ Boo, also suggest that the giant planet can inducestellarsurfaceactivityintheformofactivespots(Walkeretal.2008).Inaddition,whiletheCaIIlines of τBoo do not clearly vary with its close‐in giant planet, Catala et al. (2007) find anapparentlycomplexmagnetic field topology for thestar itself,whileDonatietal. (2008)reportapolarityswitchonτBooinanunusuallyshortenedstellaractivitycycle.Bothresultsareconsistentwithapossibleinteractionwiththeplanet'smagneticfield.

TheseSolarSystemscalinglawsenablequantitativepredictionsforanexoplanet'sradioemission(Zarkaetal.1997;Farrelletal.1999;Zarkaetal.2001;Lazioetal.2004;Stevens2005;Zarka2006,2007;Griessmeieretal.2007).Inthecaseofaknownplanet,theseestimatesdependeitheruponmeasured parameters—the planet'smass and orbital semi‐major axis—or upon parameters thatcanbeestimatedreasonably—forexample, therotationperiodcanbeassumedtobeoforder10hr.,orforhotJupiterstakentobetheorbitalperiodundertheassumptionthattheplanetistidallylocked to itshost star. Combinedwith thedistanceof thestar fromtheSun,onecanpredict theplanet'saveragefluxdensity.Inthecaseofstarsnotyetknowntohaveplanetarycompanions,anyradio limitscanbe invertedtoobtainconstraintsonthepresenceofplanets,especiallyuseful foractivestarsforwhichtheradialvelocitymethodislimited.

Implicit insomeoftheearlypredictionsforextrasolarplanetaryradioemissionisthatthestellarwindsofotherstarsarecomparabletothesolarwind.Asnotedabove,changesinthestellarwindparameterscanproducesignificantchanges in theplanet’s fluxdensity(Griessmeieretal.2005a,2007).Frommeasurementsofthesizesofastropauses(i.e.,theboundarybetweenthestellarwindandthelocalinterstellarmedium),Woodetal.(2002,2005)findthemasslossrateasafunctionof

Chapter9

224

age, dM/dt∝ tx,withx≈−2. Thus, the stellarwindarounda1‐Gyrold starmaybe25 times asintenseasthesolarwind.

Accordingly,planetsaroundyoungor“adolescent”starsarelikelytohavestrongercyclotronmaseremissions than the planets in our Solar System. Young stars are often not observed in radialvelocity surveysbecause thehigh stellar activity levels (which are in turn related to their stellarwind strengths) make it problematic to isolate a planetary signal. Thus, a blind survey formagnetosphericemissionspresentsasearchmethodologythatcouldmitigateaselectionbiasinthecurrentexoplanetcensus. If iron‐rich“SuperEarths”exist, theymayalsohavesufficientlystrongmagneticfieldstopowerradioemissionsdetectableoverinterstellardistances.

Asaspecificillustrationoftheeffectofstrongerstellarwinds,earlypredictionsforthefluxdensityoftheplanetorbitingτBoowereoforder1–3mJyatwavelengthsaround10meters(Farrelletal.1999;Lazioetal.2004).Morerecentestimates,thatattempttotakeintoaccountthelikelystellarwind strength of τ Boo, predict flux densities potentially as high as 300 mJy (Stevens 2005;Griessmeier et al. 2007). The former prediction (1–3 mJy) is below the sensitivity of currentinstrumentation;thelatterisnot.

MagneticFieldsandPlanetaryCharacterizationThe dynamo currents generating the planetary magnetic field arise from differential rotation,convection,compositionaldynamics,oracombinationoftheseintheplanet'sinterior.Inanycase,foraplanetarymagneticfieldtobesustained,theremustbeaninteriorregionoftheplanetthatisboth conducting and fluid. Consequently, knowledge of the planetary magnetic field placesconstraintsonthethermalstate,composition,anddynamicsoftheplanetaryinterior,allofwhichwillbedifficulttodeterminereliablybyothermeans.

PlanetaryInteriors:FortheSolarSystemplanets,thecompositionoftheconductingfluidrangesfrom liquid iron in theEarth’s core tometallichydrogen in JupiterandSaturn toperhapsa saltyoceaninUranusandNeptune.Thehigh‐frequencycutoffofJovianemissions(~40MHz)allowedanestimateforthestrengthoftheJovianmagneticfield20yrpriortoinsituobservations.Likewise,radio detection of themagnetic field of an exoplanetwouldprovide an indication of the planet'sinternalcomposition,insofarasitwouldrequiretheplanettohaveapartiallyconductinginterior.Combinedwith an estimate of the planet'smass, one could infer the composition of the fluid byanalogytotheSolarSystemplanets.

Planetary rotation: The rotation of a planet can impose a periodic modulation on the radioemission,as theemission ispreferentiallybeamedinthedirectionof the localmagnetic fieldandwillthereforechangeifthemagneticandspinaxesoftheplanetarenotaligned.Forallofthegas‐giantplanetsintheSolarSystem,thismodulationdefinestherotationperiods.Asthemagneticfieldis presumed to be tied to the interior of the planet, it provides amore accuratemeasure of theplanet's rotation rate than atmospheric phenomena such as clouds. For instance, the rotationperiodofNeptunewasdeterminedinitiallybyobservationsofdifferentiallyrotatingcloudtopsbutthenwasredefinedafterdetectionofitsradioemission(Lecacheuxetal.1993).

Planetary Satellites: In addition to beingmodulated by its rotation, Jupiter's radio emission isaffectedstronglybythepresenceofitssatelliteIo,andmoreweaklybyCallistoandGanymede.AstheJovianmagneticfieldsweepsoveramoon,apotentialisestablishedacrossthemoonbyitsv×BmotioninthestrongJovianmagneticfield.Thispotentialdrivescurrentsalongthemagneticfieldlines, connecting the moon to the Jovian polar regions, where the currents modulate the radioemission.Modulationsofplanetaryradioemissionmaythusalsorevealthepresenceofasatellite.


225

Atmosphericretention:Acommonandsimplemeansofestimatingwhetheraplanetcanretainitsatmosphereovergeologicaltimeistocomparethethermalvelocityofatmosphericmoleculeswiththeplanet'sescapevelocity. Ifthethermalvelocityisasubstantialfractionoftheescapevelocity,the planet will lose its atmosphere. For a planet immersed in a stellar wind, nonthermalatmospheric loss mechanisms can be important (Shizgal & Arkos 1996). These are varied(sputtering, mass loading), but all result from the typical stellar wind particle having a suprathermalvelocityrelativetotheplanet'satmosphere.Ifdirectlyexposedtoastellarwind,aplanet'satmosphere can erodemore quickly than a thermal‐only estimatewould suggest. Based on theMarsGlobalSurveyorobservationsof theremnantMartianmagnetic field, thiserosionprocess isthoughttohavebeenimportantforMars’atmosphereandoceans(Mitchelletal.2001;Lundinetal.2004;Crideretal.2005).

Planetaryalbedo:Cosmicrayscaninducenucleationinwatervapor–saturatedairandstimulatecloudformation.Largercosmicrayfluxescanproducemorecloudcoverandanincreasedalbedo.This effect has already been seen for Galactic cosmic rays (Svensmark 2000), but stellar windparticlesmaybeanimportantsecondaryeffect.

Habitability:Dependingupon its ability todeflect cosmic rays,high‐energy chargedparticles, orstellarwindparticles,themagneticfieldmayinfluencethehabitabilityofaplanet(e.g.,Griessmeieretal.2005b).Inadditiontoitseffectontheatmosphere,ifthecosmicrayfluxatthesurfaceofanotherwise habitable planet is too large, it could cause severe cellular damage and disruption ofgeneticmaterialtoanylifeonitssurfaceorfrustrateitsoriginaltogether.

9.1.2 ScienceRequirementsThe exploitation of extrasolar planetary magnetospheric emissions will require sensitive, longwavelengthradioobservations.

Table 9‐1 summarizes key requirements on observations aimed at detecting and exploitingextrasolar planetary magnetospheric emissions. In the rest of this section we motivate theserequirements.

The characteristic frequency above whichmagnetospheric emissions are no longer generated isdetermined by the strength of the planet’s magnetic field. For reference, at the cloud tops, theJovian magnetic field strength is approximately 4 Gauss, leading to a characteristic (cutoff)frequencyofapproximately40MHz(~7mwavelength).Aplanet’smagneticfieldstrengthislikelytobecorrelatedbyitsmass,assuggestedbythe“magneticplanets”intheSolarSystem.Allowingfor the fact that planets more massive than Jupiter are known, we consider it likely thatobservations below 100 MHz (longward of 3 m) will be required for extrasolar planetarymagnetosphericstudies.

Table91.ScientificRequirementsforExtrasolarPlanetaryMagnetosphericObservations

Parameter Value Comment

Wavelength(Frequency)

>~3m(<~100MHz)

Determinedbyplanetarymagneticfield;Recentbrowndwarfobservationssuggestshorterwavelengths/higherfrequenciespossible.

Sensitivity <25mJyExtrapolationsfromSolarSystemrelations;Constrainedbyexistingradioobservations.

Chapter9

226

ThesensitivityrequirementismotivatedbothbyextrapolationsofmagnetosphericemissionsfromSolarSystemplanetsaswellascurrentobservationallimits(§9.3.1).Thesensitivityrequirementisboth somewhat frequency/wavelength dependent as well as influenced by the planet targeted.Broadly, extrapolations from Solar System planets suggest that at longerwavelengths, planetarymagnetospheric emissions are stronger so that a somewhat less stringent requirement suffices.Also,becausemagnetosphericemissionsaredrivenbythestellarwind,planetsclosertotheirhoststararelikelytobebrighter.

9.2 ObservatoryConceptPlanetary magnetospheric emissions are most likely to be detected with a longwavelength radiointerferometer,aconceptwithwhichtheradioastronomycommunityhasalongheritage(50+yr)ofexperience and forwhich there aremultiple telescopes currently in the design and development orconstructionphases.

9.2.1 ArchitectureTherecognitionthattheUniversecouldbeobservedatwavelengthsotherthanthoseinthevisualspectrumbeganwiththediscoveryofcelestialradioemissionatawavelengthof14.6m(20.5MHz,Jansky 1935). Given the fundamental limitation of λ/D for angular resolution, the largewavelengths implied single apertures too large to be practical and spurred the development, at7.9‐mwavelength (38MHz),ofaperture‐synthesis interferometry (Ryle&Vonberg1946;M.Rylewas awarded a Nobel Prize in Physics for his efforts toward the development of aperturesynthesis). Subsequent work at the relevant wavelengths has been almost exclusively viainterferometers,andthisistheconceptenvisionedforallfuturelong‐wavelengthtelescopes.

Ahigh‐leveldescriptionofaradiointerferometeristhatitconsistsofthefollowingsubsystems:

Antennas/receptors:Attherelevantwavelengths,dipoleantennasareusuallyusedasindividualparabolicantennasareoftenimpracticallylarge.Dipoleshavefurtheradvantages.Theircollectingareascalesas λ2whilethetemperatureoftheGalacticbackgroundscalesapproximatelyas λ2.6,sothattofirstordertheyprovideapproximatewavelength‐independentsensitivity.Moreover,dipolesare intrinsically low cost, suggesting that a substantial collecting area can be obtained relativelyinexpensively. However, the standard thin dipole provides only a relatively narrow frequencyresponse,sovarioustechniquesincludingmakingthedipoleshaveafinitewidthorplacingvariouselectronics at the dipole itself are used to help broaden the frequency response. The spectraldynamicrangeofmoderndipolesisatleastanoctaveinfrequencyandpotentiallyafactorof3.5:1.

Station:Evenwithmodern computing, the requisite number of dipoles required to produce therequired sensitivity (§9.2.2) is often too large to handle individually. Thus, modern radiointerferometersgroupthedipolesintostations.Thesignalsfromallofthedipolesinastationarecombinedtogethertoproduceasingleoutput,conceptuallyequivalenttotheoutputfromasingleparabolicdishforradiointerferometersatshorterwavelengths.

Signal Transmission: Signals from themultiple stations are transmitted to a central processingfacility.Moderninterferometersimplementorplantoimplementthistransmissionoverfiber‐opticlinks.

Centralprocessingfacility/correlator:TheuniquepairsofstationsinaninterferometersampletheFouriertransformoftheskybrightnessdistribution.Acentralprocessingfacility,dominatedbyacorrelator,multiplesandintegratesthesignalsfromuniquepairsofstations,therebyformingthesampledtransform(orvisibility)function. Analternateapproachistoaddtogether,inphase,the


227

signals from the individual stations, effectively forming a single pencil‐beam response in thedesiredpointingdirection.

Postprocessing:Offlinepost‐processingtypicallyisusedtoinvertthemeasuredvisibilityfunctionandmake an image of a part of the sky defined by the beamwidth of a station, aswell as applyvariouscalibrationfactors.

9.2.2 PerformanceThe effective performance of a groundbased longwavelength radio telescope is determined by acombinationoftwodesignparameters(itscollectingareaandangularresolution),thecalibrationofthetelescope,andtheidentificationandexcisionofradiofrequencyinterference(RFI).Modernlongwavelength radio telescopes aim for orderofmagnitude performance improvements by improvingaspectssuchasangularresolution,calibration,andRFIhandling.

Theradiometerequationindicatesthatthetheoreticalsensitivityofaradiotelescopeisdeterminedby its collecting area Aeff, its system temperature Tsys, the observation bandwidth Δν, andintegration timeΔt. Themagnetospheric emission fromSolar‐System‐typeplanets isbroadband,withabandwidthcomparabletotheemissionfrequency,andmodernradioastronomicalsystemscanorwillbeabletoprocessbandwidthsΔνthatarealargefractionoftheobservingfrequency.Atwavelengthslongwardof3meters(<100MHz),thenonthermalemissionfromtheGalaxymakesasignificant, if not entirely dominant, contribution to the system temperature Tsys. Finally, thestandard procedure in radio astronomy is to obtain high sensitivity, in part, via long integrationtimes(Δt).However,magnetosphericemissionfromSolarSystemplanetsisquite“bursty,”sothatlongintegrationsarenotnecessarilyuseful,asalongintegrationmayaveragetogethertimeswhentheplanetisemittingwithtimeswhenitisnot.

Consequently, for long‐wavelength radio telescopes, the only parameter from the radiometerequationavailable for increasing the sensitivityof the telescope is to increase the collectingareaAeff.Alsodeterminingtheeffectivesensitivity,however,istheangularresolution,whichdeterminestheconfusionlimitofthetelescope.Pastgenerationsoflong‐wavelengthradiotelescopeshavehadlargecollectingareas(aslargeasAeff=105m2). Thesecollectingareashavebeenconcentratedinrelativelysmallareas,though,sothattheangularresolutionwaspoorandtheconfusionlimitwasquite large,wellabove thevalue thatwouldbesuggestedbyastraightforwardapplicationof theradiometerequation. Asanindicationoftheimportanceoftheangularresolutionindeterminingthe sensitivity of long‐wavelength radio telescopes, the4‐mwavelength (74‐MHz) systemon theVeryLargeArray,withitssub‐arcminuteangularresolution,hasobtainedfluxdensitiessome1–2ordersofmagnitudelowerthanthoseofpreviousgenerationsoflong‐wavelengthradiotelescopeseven though the collecting area of the VLA is comparable to or smaller than previous long‐wavelengthradiotelescopes.

Many of the modern long‐wavelength radio telescopes under development or in constructioncoupleasignificantcollectingareawithrelativelyhighangularresolution.Thegoalistoapproachorexceedasensitivityofapproximately1mJyatwavelengthsaround5m(60MHz);viz.Table9‐1.

The primary impediment to angular resolution at long wavelengths was the limited maximumbaselinesoftheavailableinstruments.Inallcases,themaximumbaselineswasnomorethanabout5km,whichinturnwaslimitedbythephasedistortionimposedbytheEarth’sionosphereovertheintrinsically wide fields of view and the lack of suitable algorithms to compensate for theseionospherically‐imposed distortions. On baselines longer than a few kilometers, without acalibrationalgorithm,phasedistortionsaresevereenoughtocausedecorrelation,makinghigher‐resolutionimagingdifficulttoimpossible. Inthepastdecade,newcalibrationschemeshavebeendeveloped for the ionospheric phase corruptions at radio wavelengths, often bearing at least a

Chapter9

228

conceptual similarity to adaptive optics techniques for tropospheric phase corruptions at visiblewavelengths.

Modern long‐wavelength radio telescopes plan both to exploit these new algorithms as well asdevelop new algorithms in order to reach the theoretical noise limit implied by the radiometerequation. Indeed, inmany cases, extraction and understanding of ionospheric physics from theradioastronomicaldataisakeypartofthesciencecaseforthenextgenerationoflong‐wavelengthradiotelescopes.

The final impact on performance of long‐wavelength radio telescopes is radio‐frequencyinterference (RFI) produced by civil or military transmitters operating at similar wavelengths.Thesetransmittersareoftenordersofmagnitudestrongerthanthedesiredsignalfromexoplanets.In themostmoderate case,RFI simply restricts the available bandwidthby limiting the rangeofavailable frequencies (and having a concomitant effect on the telescope’s sensitivity). In worstcases,RFIcanbestrongenoughtocausenon‐lineareffects intheelectronicsandrenderthedatauseless.(Inevenmoreextremecases,extremelystrongRFIcandamagetheelectronics.)

TheidentificationandexcisionofRFIremainsasignificantchallengeforthenext‐generationradiotelescopes,particularly for theirutility instudyingexoplanets. RFI that isnotproperly identifiedandexcisedeffectivelyservesasanadditionalnoisesource,limitingthesensitivityofthetelescope.

Figure 93. Images of present and future radio observatories. (Courtesy of JPL; NRAO; NCRA; IRA; UNM/ONR/NRL;ASTRON;SKA;NASAGSFC)


229

Moreover, using Jupiter’s magnetospheric emission as an exemplar of planetary radio emission,therecanbeconsiderablestructureinthetime‐frequencybehaviorofaplanet’semission.CurrentRFI identificationandexcision techniquesarerelativelycrude, largelyconsistingonlyofblankingthe signal either in time or frequency. Further, RFI identification and excision often remains alaborioustask,conductedbyanindividualexaminingthedata. MoresophisticatedalgorithmsfortheidentificationofRFI,e.g.,byemployingbetterstatisticaltests,areamongthekeydevelopmentaspectsforthenext‐generationofradiotelescopes.

9.3 TechnologyA number of longwavelength radio telescopes are in development or construction. Key technicaldevelopmentsareneededintheareasofalgorithmandsoftware(§9.2.2),butnothardware.

9.3.1 PastAccomplishmentsMagnetospheric emissions from the Solar System planets motivated searches for analogousemissions prior to the discovery of exoplanets (Yantis et al. 1977; Winglee et al. 1986). Twosignificantchangesinthepastdecadehavebeenthediscoveryofexoplanetsandthedevelopmentof long‐wavelength telescopes with an order of magnitude more sensitivity than historicaltelescopes.(SeealsoZarkaetal.1997.)

Scaling relations developed from Solar System planets predict that the relevant frequencies arebelow1000MHz(λ>30cm).Thepremierinstrumentsinthisrange,showninthelefthalfofFig.9‐3,aretheVeryLargeArray(VLA),withits74‐MHzsystem,theGiantMetrewaveRadioTelescope(GMRT), with its 150‐MHz system, and the Ukranian T‐shaped Radiotelescope (UTR‐2), whichobserves at 25 MHz. The UTR‐2 observes at low enough frequencies that it can detect Jovianemissions,andthe74‐MHzVLAobservesatafrequencythatiswithinafactoroftwoofthehighestfrequencyJovianemissions(40MHzor7.5‐mwavelength).Therehavebeenahandfulofsearchesatall three telescopes, targetingknownexoplanets,withmanyof thesesearchesstill inprogress.Sensitivities achieved range from a fewmilliJanskys at the GMRT to a few tens or hundreds ofmilliJanskysattheVLAtoafewJanskysattheUTR‐2.

Perhaps themost effort hasbeendirected toward theplanet orbitingτ Boo (Bastian et al. 2000;Farrell et al. 2003; Lazio & Farrell 2007; see also George & Stevens 2007), as various authorspredict that its flux density is comparable to or greater thanwhat can be obtained in amodestduration(fewhours)observation(Farrelletal.1999;Lazioetal.2004;Stevens2005;Griessmeieret al. 2007). Lazio & Farrell (2007) have obtained multi‐epoch VLA observations, asmagnetospheric emissions are expected to be “bursty” and the predicted flux densities are justcomparable to thecurrent sensitivity limits. A singleobservationmightoccuratan inopportunetimewhen the stellarwind fluxwasbelowaverage. Theobservational limitsnowconstrain thisplanet's luminosity to be less than1023 erg s−1, unless its radiation is highly beamed into a solidangle Ω << 1, which would be much smaller than that for any of the Solar System planets.Presumingthattheradiationisnothighlybeamed,thisluminositylimitislowerthansome,butnotall, recent predictions. Although higher sensitivity observations are likely required, the non‐detectionmayalsobehintingthatthemagneticfields,andinternalcompositions,ofexoplanetsareasvariedastheplanetsthemselves.

Chapter9

230

9.3.2 FutureMilestones

FutureGroundBasedTelescopesThere are a number of long‐wavelength telescopes, in various stages of development,whichwillhavesensitivitiespotentiallyasmuchastwoordersofmagnitudebetterthancurrentfacilities.Forsomefacilities,searchingfortheradioemissionfromexoplanetsisrecognizedexplicitlyaspartofthesciencecase.

In the initial constructionphases are theLongWavelengthArray (LWA, inNewMexico) and theLowFrequencyArray(LOFAR,intheNetherlands).TheLWAwilloperateinthe20–80MHzband(3.75–15mwavelength); LOFARwill operate in the30–80MHzand110–240MHzbands (1.25–2.7mwavelength).BothinstrumentscoverthefrequencyrangeexpectedforemissionfromJovian‐masstoseveralJovianmassplanets. Expectedinitialoperationalphasesforbothinstrumentsarethenextfewyears.

TheSquareKilometerArray(SKA)isanext‐generationtelescopethatisexpectedtooperateabove100MHz. Located either inAustralia or SouthAfrica, its design goals are such that it shouldbeeasily capable of detecting the radio emissions from themostmassive exoplanets, for those thathavecutofffrequenciesaboveabout100MHz.TheSKAiscurrentlyinthedesignanddevelopmentphase,withconsiderableinternationalactivity,fundedbyentitiesincludingtheNSF,theEuropeanCommission, various European governments, the Australian government, and the South Africangovernment.

FutureSpaceFacilitiesA significant constraint to all ground‐based facilities is the Earth's ionosphere. While the valuechangeswithtime(e.g.,dayvs.night),theionospheregenerallyisopaquebelowabout10MHz(λ>30m). Jupiter produces themost intense and highest frequency emissions of the Solar Systemplanets,buteventhesecutoffabove40MHz.TheEarth'smagnetosphereemitsauroralkilometricradiation (AKR) below 1 MHz. Thus, the detection of AKR from extrasolar terrestrial‐massplanets—andassessmentsof theirhabitability—canonlybeaccomplished fromspace. ThemostpromisinglocationforatelescopedesignedtodetectAKRfromextrasolarterrestrial‐massplanetsisthefarsideoftheMoon. NotonlyisthefarsidealwaysshieldedfromtheEarth,forhalfoftheMoon'sorbit,anarrayonthefarsidewouldbeshieldedfromsolarradioemissionsaswell.

NASAplansareturntotheMooninthenextdecade(~2019)andisfundingconceptstudiesforalunarradiotelescopeunderboththeLunarSortieScienceOpportunitiesandAstrophysicsStrategicMission Concept Studies programs. In Europe, ESA, the aerospace company EADS, and variousEuropeaninstituteshavebeeninvestigatingthedeploymentofalunarradioarray.Thekeysciencedriver for a large lunar radio telescope is cosmological studies. However, in all cases, extrasolarplanetaryemissionisrecognizedasanimportantsecondarysciencegoal.

Developments required for a lunar radio array include a number of technologies with broadapplicability. These includeultra‐lowpowerelectronics,high‐energy‐density low‐massbatteries,autonomousorsemi‐autonomousrobotics(includingrovers),andhighbandwidthdatadownlinks.


231

9.4 Research&AnalysisGoalsTheconstructionofthenextgenerationradiotelescopesoffersanopportunitytoleverageinvestmentstocharacterizeexoplanets.

The radio telescopes under development and construction (the LWA, LOFAR, and the SKA) arebeing developed by a variety of institutes and funding agencies. Thus, searches for extrasolarplanetaryradioemissioncanleveragetheinvestmentsforlittleornomarginalcost,andextrasolarplanetaryradioemissionisrecognizedasapotentialscienceresultfrommanyofthesetelescopes.Exploitingtheirfullpotentialforextrasolarplanetaryworkwilllikelyrequireadditionalalgorithmdevelopmentorobservations.

KeyalgorithmdevelopmentincludesmoresophisticatedRFIidentificationandexcisionandprocessingofthetime‐frequencynatureoftheradiosignals(§9.2.2).Inparticular,pattern‐matchingalgorithmsthatcoulduseJovianradioemissionasatemplateforhighersignificancedetectionofextrasolarplanetaryradioemissionareunlikelytobedevelopedbyanyoftheseprojects.

Supportforcoordinatedmulti‐wavelengthobservationswithspacemissions,suchastheEuropeanCoRoTmission.SomepreliminaryworkhasalreadybeendonewithCoRoTandGMRTobservations,whichwillmakefurthercoordinationmorefeasible.

Continuedpolarimetricobservationsofthehoststarstodetectandcharacterizeadditionalphenomenarelatedtostar‐planetinteractions(§9.1.1.1).Theseobservationswillallowfundamentalplanetarycharacteristicstobeextracted,particularlywhencombinedwithmodelingefforts,someofwhichareunderwayalready(e.g.,Preusseetal.2006;McIvoretal.2006;Lanza2008;KitiashviliandGusev2008).

Exploitation of the existing ground‐based arrays (VLA, GMRT, and UTR‐2, §9.3.1) as testbeds fordevelopmentofnewalgorithmsaswellasimprovinglimitsonknownexoplanets,whenpossible.

Developmentofalgorithmsthatmaybedeployedinfirmwaretocarryoutactiveadaptivenullingofbrightsourcesinantennaprimarybeams.Theadaptationofsuchtechniquesinroutineobservationsarecurrentlyprohibitivelyexpensiveintermsofcomputingcycles,butwiththeadventofFPGAtechnologiesmaybefeasibleforfutureinstruments.

Finally,thedevelopmentofafuturelunarradioarraywouldbenefitfromanumberoftechnologieswithbroadapplicability.Theseincludeultra‐lowpowerelectronics,high‐energy‐densitylow‐massbatteries, autonomousor semi‐autonomous robotics (including rovers), andhigh‐bandwidthdatadownlinks.

Chapter9

232

9.5 ContributorsJ.Lazio,NavalResearchLaboratory

D.Winterhalter,JetPropulsionLaboratory

T.Bastian,NationalRadioAstronomyObservatory

G.Bryden,JetPropulsionLaboratory

W.M.Farrell,NASAGoddardSpaceFlightCenter

J.M.Griessmeier,ASTRON

J.Kasper,SmithsonianAstrophysicalObservatory

T.Kuiper,JetPropulsionLaboratory

A.Lecacheux,ObservatoiredeParis

W.Majid,JetPropulsionLaboratory

R.Osten,STScI

E.Shkolnik,CarnegieInstituteofWashington

P.Zarka,ObservatoiredeParis

9.6 ReferencesBastian,T.S.,Dulk,G.A.,&Leblanc,Y.2000,“ASearchforRadioEmissionfromExtrasolarPlanets,”

ApJ,545,1058–1063

Catala, C., Donati, J.‐F., Shkolnik, E., Bohlender, D., &Alecian, E. 2007, “TheMagnetic Field of thePlanet‐HostingStarτBootis,”MNRAS,374,L42–L46

Crider,D.,Espley,J.,Brain,D.A.,etal.2005,“MarsGlobalSurveyorObservationsoftheHalloween2003SolarSuperstorm'sEncounterwithMars,”J.Geophys.Res.,110(A9),A09S21

Desch,M.D.,&Kaiser,M.L.1984,“PredictionsforUranusfromaradiometricBode’slaw,”Nature,310,755–757

Desch,M.D.1988,“Neptuneradioemission:Predictionsbasedonplanetaryscalinglaws,”Geophys.Res.Lett.,15,114–117

Donati,J.F.,Moutou,C.,Farés,R.,etal.2008,“MagneticCyclesofthePlanet‐HostingStarτBootis,”MNRAS,385,1179–1185

Farrell, W. M., Lazio, T. J. W., Desch, M. D., Bastian, T., & Zarka, P. 2003, “Radio Emission fromExtrasolarPlanets,”inBioastronomy2002:LifeAmongtheStars,eds.R.Norrisetal.(ASP:SanFrancisco)p.73

Farrell,W.M.,Desch,M.D.,&Zarka,P. 1999, “On thePossibility ofCoherentCyclotronEmissionfromExtrasolarPlanets,”J.Geophys.Res.,104,14025–14032

George,S.J.,&Stevens,I.R.2007,“GiantMetrewaveRadioTelescopeLow‐FrequencyObservationsofExtrasolarPlanetarySystems,”MNRAS,382,455–460


233

Griessmeier,J.‐M.,Zarka,P.,&Spreeuw,H.2007,“PredictingLow‐FrequencyRadioFluxesofKnownExtrasolarPlanets,”A&A,475,359–368

Griessmeier, J.‐M., Preusse, S., Khodachenko,M.,Motschmann,U.,Mann,G.,&Rucker,H.O. 2007,“ExoplanetaryRadioEmissionunderDifferentStellarWindConditions,”Planetary&SpaceSci.,55,618–630

Griessmeier,J.‐M.,Motschmann,U.,Mann,G.,&Rucker,H.O.2005a,“TheInfluenceofStellarWindConditionsontheDetectabilityofPlanetaryRadioEmissions,”A&A,437,717–726

Griessmeier,J.‐M.,Stadelmann,A.,Motschmann,U.,etal.2005b,“CosmicRayImpactonExtrasolarEarth‐LikePlanetsinClose‐inHabitableZones,”Astrobiology,5,587–603

Jansky,K.G.1935,“ANoteontheSourceofInterstellarInterference,”Proc.IRE,23,1158–1163

Kitiashvili, I. N., & Gusev, A, 2008, “Rotational Evolution of Exoplanets under the Action ofGravitational andMagneticPerturbations,”CelestialMech. andDynamicalAstron., 100,121–140

Lanza,A.F.2008,“HotJupitersandStellarMagneticActivity,”A&A,487,1163–1170

Lazio,T.J.W.,&Farrell,W.M.2007,“MagnetosphericEmissionsfromthePlanetOrbitingτBootis:AMulti‐epochSearch,”ApJ,668,1182–1188

Lazio,T.J.W.,Farrell,W.M.,etal.2004,“TheRadiometricBode'sLawandExtrasolarPlanets,”ApJ,612,511–518

Lecacheux, A., Zarka, P., Desch, M. D., & Evans, D. R. 1993, “The Sidereal Rotation Period ofNeptune,”Geophys.Res.Lett.,20,2711–2714

Lundin, R., Barabash, S., Andersson, H., et al. 2004, “SolarWind‐Induced Atmospheric Erosion atMars:FirstResultsfromASPERA‐3onMarsExpress,”Science,305,1933–1936

McIvor,T.,Jardine,M.,&Holzwarth,V.2006,“ExtrasolarPlanets,StellarWindsandChromosphericHotspots,”MNRAS,367,L1–L5

Millon,M.A.,&Goertz,C.K.1988,“PredictionofRadioFrequencyPowerGenerationofNeptune'sMagnetospherefromGeneralizedRadiometricBode'sLaw,”Geophys.Res.Lett.,15,111–113

Mitchell, D. L., Lin, R. P., Mazelle, C., et al. 2001, “Probing Mars' Crustal Magnetic Field andIonospherewiththeMGSElectronReflectometer,”J.Geophys.Res.,106,23419–23428

Preusse,S.,Kopp,A.,Büchner,J.,&Motschmann,U.2006,“AMagneticCommunicationScenarioforHotJupiters,”A&A,460,317–322

Ryle,M.,&Vonberg,D.D.1946,“SolarRadiationon175Mc./s,”Nature,158,339–340

Shkolnik,E.,Bohlender,D.A.,Walker,G.A.H.,&CollierCameron,A.2008, “TheOn/OffNatureofStar‐PlanetInteractions,”ApJ,676,628–638

Shkolnik,E.,Walker,G.A.H.,Bohlender,D.A.,Gu,P.‐G.,&Kuerster,M.2005,“HotJupitersandHotSpots: The Short‐ and Long‐Term Chromospheric Activity on Starswith Giant Planets,” ApJ,622,1075–1090

Shizgal, B. D., &Arkos, G. G. 1996, “Nonthermal Escape of theAtmospheres of Venus, Earth, andMars,”Rev.Geophys.,34,483–505

Stevens,I.R.2005,“MagnetosphericRadioEmissionfromExtrasolarGiantPlanets:TheRoleoftheHostStars,”MNRAS,356,1053–1063

Chapter9

234

Svensmark,H.2000,“CosmicRaysandEarth'sClimate,”SpaceSci.Rev.,93,175–185

Walker,G.A.H.,Croll,B.,Matthews,J.,etal.2008,“MOSTDetectsVariabilityonτBootisAPossiblyInducedbyitsPlanetaryCompanion,”A&A,482,691–697

Winglee, R. M., Dulk, G. A., & Bastian, T. S. 1986, “A Search for CyclotronMaser Radiation fromSubstellarandPlanet‐likeCompanionsofNearbyStars,”ApJ,309,L59–L62

Wood, B. E., Müller, H.‐R., Zank, G. P., Linsky, J. L., & Redfield, S. 2005, “New Mass‐LossMeasurementsfromAstrosphericLyαAbsorption,”ApJ,628,L143–L146

Wood,B.E.,Müller,H.‐R.,Zank,G.P.,&Linsky,J.L.2002,“MeasuredMass‐LossRatesofSolar‐likeStarsasaFunctionofAgeandActivity,”ApJ,574,412–425

Yantis,W.F.,Sullivan,W.T.,III,&Erickson,W.C.1977,“ASearchforExtra‐SolarJovianPlanetsbyRadioTechniques,”Bull.Amer.Astron.Soc.,9,453

Zarka, P. 2007, “Plasma Interactions of Exoplanets with their Parent Star and Associated RadioEmissions,”Planet.SpaceSci.,55,598–617

Zarka,P.2006,“HotJupitersandMagnetizedStars:GiantAnalogsoftheSatellite‐JupiterSystem?”inPlanetaryRadioEmissionsVI,eds.H.O.Ruckeretal.(AustrianAcad.:Vienna)p.543

Zarka,P.,Treumann,R.A.,Ryabov,B.P.,&Ryabov,V.B.2001,“Magnetically‐DrivenPlanetaryRadioEmissionsandApplicationtoExtrasolarPlanets,”Astrophys.&SpaceSci.,277,293–300

Zarka,P.,Queinnec,J.,Ryabov,B.P.,etal.1997,“Ground‐BasedHighSensitivityRadioAstronomyatDecameterWavelengths,” in Planetary Radio Emissions IV, eds. H. O. Rucker et al. (AustrianAcad.:Vienna)p.101.

Listof2008ExoplanetForumParticipants

235


EricAgol,UniversityofWashington

RachelAkeson,NASAExoplanetScienceInstitute

DanielAngerhausen,PhysikalischesInstitut,UniversitätzuKöln

GuillemAngladaEscudé,CarnegieInstitutionofWashington.Dept.ofTerrestrialMagnitism

DavidArdila,SpitzerScienceCenter

JonathanArenberg,NorthropGrummanSpaceTechnology

BalaBalasubramanian,JetPropulsionLaboratory

JeanPhilippeBeaulieu,Institutd'AstrophysiquedeParis

RuslanBelikov,NASAAmes

DavidBennett,UniversityofNotreDame

MichaelBicay,NASAAmes

AndyBoden,USC/Caltech

AndrewBooth,JPL/Caltech

JeffreyBooth,JPL/Caltech

AlanBoss,CarnegieInstitution

MatthewBoyce,HeliosEnergyPartners

JamesBreckinridge,Caltech/JPL



SteveBryson,NASAAmesResearchCenter


PaulButler,CarnegieInstitutionofWashington,DepartmentofTerrestrialMagnetism

JohnCallas,JetPropulsionLaboratory

RichardCapps,JetPropulsionLaboratory

WebsterCash,UniversityofColorado

JosephCatanzarite,JetPropulsionLaboratory

236

ChristineChen,NOAO/STScI

PinChen,JPL/Caltech

MarvinChristensen,NASAAmesResearchCenter


WilliamCochran,McDonaldObservatory,UniversityofTexasatAustin

DavidCole,JetPropulsionLaboratory

VincentCoudéduForesto,LESIAObservatoiredeParis

PhilippeCrane,NASAHQPlanetaryScienceDivision

IanCrossfield,UCLA

WilliamDanchi,NASAGoddardSpaceFlightCenter


JohnDavidson,JetPropulsionLaboratory

PieterDeroo,JetPropulsionLaboratory

MichaelDevirian,JPL/ExoPlanetExplorationProgramOffice

AlanDressler,CarnegieObservatories

AlanDuncan,LockheedMartinATC

DennisEbbets,BallAerospace&TechnologiesCorp.

StephenEdberg,JetPropulsionLaboratory

RobertEgerman,ITTSpaceSystemsDivision

BrianFrank

BruceGary,JetPropulsionLaboratory(retired)

ScottGaudi,TheOhioStateUniversity

ThomasGautier,JetPropulsionLaboratory

JianGe,UniversityofFlorida

DawnGelino,NASAExoplanetScienceInstitute

AmirGiv’on,JetPropulsionLaboratory

TiffanyGlassman,NorthropGrummanSpaceTechnology

TomGreene,NASAAmesResearchCenter

EdwardGreisch

CarlGrillmair,SpitzerScienceCenter

CeciliaGuiar,JetPropulsionLaboratory

OlivierGuyon,SubaruTelescope/UniversityofArizona

NaderHaghighipour,InstituteforAstronomy,UniversityofHawaii

CharlesHanot,JetPropulsionLaboratory


237

JosephHarrington,UCF

IsmailD.Haugabook,Sr.,DigitalTechnicalServices

SarahHeap,NASAGoddardSpaceFlightCenter

PhilHinz,StewardObservatory

ScottHorner,LockheedMartin

S.SonaHosseini,UCDavis

DouglasHudgins,NASAHeadquarters

JasonIbarra,Maexidus

HannahJangCondell,Univ.Maryland/NASAGSFC

MuthuJeganathan,JetPropulsionLaboratory

WilliamJohnson,JetPropulsionLaboratory

LisaKalteneggar,HarvardSmithsonianCenterforAstrophysics

JonathanKamrava,JetPropulsionLaboratory

StephenKane,NASAExoplanetScienceInstitute

N.JeremyKasdin,PrincetonUniversity

RickKendrick,LockheedMartin

MatthewKenworthy,StewardObservatory,UniversityofArizona

BrianKern,JetPropulsionLaboratory

NancyKiang,NASAGoddardInstituteforSpaceStudies

StevenKilston,BallAerospace&TechnologiesCorp.



BenjaminLane,CharlesStarkDraperLaboratory

LiaLaPianaNASAHeadquarters

DavidLatham,HarvardSmithsonianCenterforAstrophysics

GregoryLaughlin,UCO/LickObservatory

NicholasLaw,CaliforniaInstituteofTechnology

PeterLawson,JetPropulsionLaboratory

OliverLay,JetPropulsionLaboratory

JosephLazio,NavalResearchLaboratory

RogerLee,JetPropulsionLaboratory

MarieLevine,JetPropulsionLaboratory

CharlesLillie,NorthropGrummanSpaceTechnology

DougLin,UniversityofCalifornia,SantaCruz

238

JamesLloyd,CornellUniversity

ManojLnu

AmyLo,NorthropGrummanSpaceTechnology

DanittzaLopezBlanco,JetPropulsionLaboratory

PatrickLowrance,SpitzerScienceCenter

RichardLyon,NASAGoddardSpaceFlightCenter

RichardMachuzak,JetPropulsionLaboratory

AmyMainzer,JetPropulsionLaboratory

WalidMajid,JetPropulsionLaboratory

GeoffreyMarcy,UCBerkeley

MarkMarley,NASAARC

JamesMarr,JetPropulsionLaboratory

StefanMartin,JetPropulsionLaboratory

TaroMatsuo,JetPropulsionLaboratory

DimitriMawet,JetPropulsionLaboratory

BenjaminMazin,JetPropulsionLaboratory

ChrisMcCarthy,SanFranciscoStateUniversity

JohnMcCarthy,OrbitalSciences

MarkMcKelvey,NASAAmesResearchCenter

RobertMelisso,HistoryChannel


JulieMikula,CodeVSStrategicManagement&AnalysisDivision,NASAHQ

RafaelMillanGabet,NASAExoplanetScienceInstitute,Caltech

MarkMilman,JetPropulsionLaboratory

DwightMoody,JetPropulsionLaboratory

NaoshiMurakami,NAOJ,Japan

ShawnMurphy,CharlesStarkDraperLaboratory

MatthewMuterspaugh,UniversityofCalifornia,Berkeley

JunNishikawa,NAOJ(NationalAstronomicalObservatoryofJapan)


ChristopherPaine,JetPropulsionLaboratory

XiaopeiPan,JetPropulsionLaboratory

OzhenPananyan,JetPropulsionLaboratory

PeggyPark,JetPropulsionLaboratory


239

AlanPayne,CaliforniaInstituteofTechnology

GuyPerrin,ObservatoiredeParis/LESIA

MarshallPerrin,UCLA

MeyerPesenson,CaliforniaInstituteofTechnology

PeterPlavchan,NASAExoplanetScienceInstitute,Caltech

RonaldPolidan,NorthrupGrummanSpaceTechnology

StevenPravdo,JetPropulsionLaboratory

LaurentPueyo,PrincetonUniversity

SamRagland,W.M.KeckObservatory

DarinRagozzine,CaliforniaInstituteofTechnology

StephenRidgway,NOAO


AkiRobergeNASAGoddardSpaceFlightCenter

LewisRoberts,JetPropulsionLaboratory

KailashSahu,SpaceTelescopeScienceInstitute

RoccoSamuele,NorthropGrummanCorporation

JagmitSandhu,JetPropulsionLaboratory

JeanSchneider,ObservatoiredeParis

ArnoldSchwartz,NHSC/IPAC/Caltech

EugeneSerabyn,JetPropulsionLaboratory

MarthaShaindlin,HomeschoolTeacher

TashaShaindlin,Child


MichaelShao,JetPropulsionLaboratory

EricSmith,NASAHeadquarters

ChristopheSotin,JetPropulsionLaboratory

RémiSoummer,AmericanMuseumofNaturalHistory


ChristopherStark,U.ofMaryland/NASAGoddardSpaceFlightCenter

CallistaSteele,ValparaisoUniversity

RhoadsStephenson,JetPropulsionLaboratory(retired)

MarkSwain,JetPropulsionLaboratory

MotohideTamura,NAOJ,Japan

AngelleTanner,JetPropulsionLaboratory

240

StuartTaylor

MohammadTehrani,TheAerospaceCorp

ShahinTehrani,InnovaPhotonics

DomenickTenerelli,LockheedMartinSpaceSystemsCompany

GiovannaTinetti,UniversityCollegeLondon

GuillermoTorres,HarvardSmithsonianCenterforAstrophysics


JohnTrauger,JetPropulsionLaboratory

MitchellTroy,JetPropulsionLaboratory

ZlatanTsvetanov,NASAHQ

NealTurner,JetPropulsionLaboratory

CarlTuttle,LockheedMartinSpaceSystems

StephenUnwin,JetPropulsionLaboratory

RobertVanderbei,PrincetonUniversity

GautamVasisht,JetPropulsionLaboratory

StevenVogt,UCO/LickObservatory

KasparvonBraun,CaliforniaInstituteofTechnology

JamesWarner,everybodyinspace.com

MichaelWerner,JetPropulsionLaboratory

DanielWinterhalter,JetPropulsionLaboratory

JenniferWiseman,NASAGoddardSpaceFlightCenter

JulienWoillez,W.M.KeckObservatory,Hawaii,USA

BruceWoodgate,NASAGoddardSpaceFlightCenter

RobertWoodruff,LockheedMartin

AlWootten,NationalRadioAstronomyObservatory

JeffWynn,ITT

MingZhao,UniversityofMichigan

CopyrightPermissions

241


Figure2‐1 OriginalbyJ.Catanzarite(JPL),forthispublication.

Figure2‐2 OriginalbyB.Nemati(JPL),forthispublication.

Figure2‐3 OriginalbyJ.Marr(JPL),forthispublication.

Figure3‐1 OriginalbyR.A.Brown(SpaceTelescopeScienceInstitute),forthispublication.

Figure3‐2 OriginalbyM.Marley(NASAAmesResearchCenter),forthispublication.

Figure3‐3 OriginalbyM.Marley(NASAAmesResearchCenter),forthispublication.

Figure3‐4 OriginalbyR.Soummer(SpaceTelescopeScienceInstitute)andM.Levine(JPL),forthispublication.

Figure3‐5 B.A.Macintosh et al.Reprintedbypermissionof the SPIE, fromProc. SPIE, 6272,62720N(2006).

Figure3‐6 OriginalbyS.R.Heap(NASAGSFC)forthispublication.

Figure3‐7 OriginalbyR.A.BrownandR.Soummer(SpaceTelescopeScienceInstitute),forthispublication.

Figure3‐8 OriginalbyM.Levine(JPL),forthispublication.

Figure3‐9 OriginalbyM.Levine(JPL),forthispublication.

Figure3‐10 Original fromM. Levine, S. Shaklan& J.Kasting.Reprinted from the STDTReport,JPLDocumentD‐34923(2006).

Figure3‐11 D.Moody,B.Gordon,&J.Trauger.ReprintedbypermissionoftheSPIE,fromProc.SPIE,7010,7010‐139(2008).

Figure3‐12 CourtesyofPrincetonUniversityandJPL/Caltechforthispublication.

Figure3‐13 D.Moody,B.Gordon,&J.Trauger.ReprintedbypermissionoftheSPIE,fromProc.SPIE,7010,7010‐139(2008).

Figure3‐14 R. Soummer (STScI), A. Sivaramakrishnan (AMNH), and D. Dillon (UCSC) for thispublication.

Figure3‐15 OriginalbyA.Lo(NGST),forthispublication.

Figure3‐16a D. B. Leviton et al. Reprinted by permission of the SPIE, from Proc. SPIE, 668766871B(2007).

Figure3‐16b E. Schindhelm et al. Reprinted by permission of the SPIE, from Proc. SPIE, 6693669305(2007).

Figure3‐17 OriginalbyD.Tenerelli(LockheedMartin),forthispublication.

242




Figure3‐21 CourtesyofNASA/MSFC.

Figure3‐22 OriginalbyA.Klavins(LockheedMartin),forthispublication.

Figure3‐23 OriginalbyEgerman&Matthews(ITT),forthispublication.

Figure4‐1 OriginalbyW.C.Danchi(NASAGSFC),forthispublication.


Figure4‐3 L.Kaltenegger,etal.ReprintedbypermissionoftheAmericanAstronomicalSociety,fromApJ,658,598–616(2007).

Figure4‐4 Reprinted with kind permission of Springer Science and Business Media fromL.Kaltenegger,etal.,Astrophysics&SpaceScience,inpress(2009).



Figure4‐7 D. Defrère, et al. Reprinted by permission of the publisher, from Astronomy &Astrophysics,490,435‐445(2008).

Figure4‐8 D. Defrère, et al. Reprinted by permission of the publisher, from Astronomy &Astrophysics,490,435‐445(2008).

Figure4‐9 OriginalbyT.Herbst(MPIA)&P.R.Lawson(JPL),forthispublication.

Figure4‐10 OriginalbyS.R.Martin(JPL),forthispublication.

Figure4‐11 OriginalbyO.P.Lay(JPL),forthispublication.

Figure4‐12 OriginalbyO.P.Lay&SarahHunyadi(JPL),forthispublication.

Figure4‐13 OriginalbyP.R.Lawson(JPL),forthispublication.



Figure4‐16 OriginalbyA.Ksendzov(JPL),forthispublication.

Figure4‐17 OriginalbyR.D.Peters(JPL),forthispublication.

Figure4‐18 OriginalbyP.R.Lawson(JPL),forthispublication.

Figure4‐19 CourtesyoftheEuropeanSpaceAgencyandTNO.

Figure4‐20 CourtesyofNASA.


Figure4‐22 OriginalbyS.R.MartinandO.P.Lay(JPL),forthispublication.

Figure4‐23 CourtesyofCaltech/JPL/NASA.

Figure5‐1 OriginalbyA.Roberge(NASAGSFC),forthispublication.

Figure5‐2 (Upperpanel)OriginalbyC.A.Beichman(JPL),forthispublication.


243

Figure5‐2 (Lower panel) C. A. Beichman, et al. Reprinted by permission of the AmericanAstronomicalSociety,fromApJ,652,1674–1693(2006).

Figure5‐3 (Upperpanel)OriginalbyG.Bryden(JPL),forthispublication.

Figure5‐3 (Lowerpanel)OriginalbyG.Bryden(JPL),forthispublication.

Figure5‐4 OriginalbyM.Perrin(UCLA),forthispublication.

Figure5‐5 M. J. Kuchner & M. J. Holman. Reprinted by permission of the AmericanAstronomicalSociety,fromApJ,588,1110–1120(2003).

Figure5‐6 C.C.Stark&M.J.Kuchner.ReprintedbypermissionoftheAmericanAstronomicalSociety,fromApJ,686,637–648(2008).

Figure5‐7 C.C.Stark&M.J.Kuchner.ReprintedbypermissionoftheAmericanAstronomicalSociety,fromApJ,686,637–648(2008).

Figure6‐1 Left:OriginalbyB.S.Gaudi(OhioStateUniv.),forthispublication.

Figure6‐1 Right:OriginalbyD.P.Bennett(NotreDameUniv.),updatedfromApJL,644,37–40(2006).ReprintedbypermissionoftheAmericanAstronomicalSociety.

Figure6‐2 D.P.Bennett&S.H.Rhie.Reprintedbypermissionof theAmericanAstronomicalSociety,fromApJ,574,985–1003(2002).

Figure6‐3 OriginalbyD.P.Bennettetal.(2008),fromawhitepapersubmittedtotheExoPlanetTaskForce(arXiv:0704.054).

Figure6‐4 OriginalbyD.P.Bennett(NotreDameUniv.),forthispublication.

Figure6‐5 OriginalbyD.P.Bennett(NotreDameUniv.),forthispublication.

Figure6‐6 Original by B. Macintosh (Lawrence Livermore National Laboratory), for theExoPlanetTaskForceReport.

Figure6‐7 D.P.Bennett, J.Anderson,&B.S.Gaudi. ReprintedbypermissionoftheAmericanAstronomicalSociety,fromApJ,660,781–790(2007).

Figure7‐1 OriginalbyG.Torres(SmithsonianAstrophysicalObservatory),forthispublication.

Figure8‐1 OriginalbyD.Deming(NASAGSFC),forthispublication.

Figure8‐2 OriginalbyD.Deming(NASAGSFC),forthispublication.

Figure8‐3 D.Deming, J.Harrington, S. Seager,&L. J.Richardson.ReprintedbypermissionoftheAmericanAstronomicalSociety,fromApJ,644,560–564(2006).

Figure8‐4 M. R. Swain, G. Vasisht, & G. Tinetti. Reprinted by permission of MacMillanPublishersLtd.:Nature,452,329–331,copyright2008.

Figure8‐5 OriginalbyY.H.Wu(JPL),forthispublication.

Figure8‐6 S. Seager&D.D.Sasselov.Reprintedbypermissionof theAmericanAstronomicalSociety,fromApJ,537,916–921(2000).

Figure8‐7 J.J.Fortney,K.Lodders,M.S.Marley,&R.S.Freedman.ReprintedbypermissionoftheAmericanAstronomicalSociety,fromApJ,678,1419–1435(2008).

Figure8‐8 A. Vidal‐Madjar, A. Lecavalier des Etangs, et al. Reprinted by permission of theAmericanAstronomicalSociety,fromApJ,676,L57–L60(2008).


244

Figure9‐2 Reprinted with kind permission of Springer Science and Business Media fromP.Zarka, R. A. Treumann, B. P. Ryabov, & V. B. Ryabov, Astrophysics & SpaceSciences,277,293(2001).

Figure9‐3 Courtesy of JPL/Caltech/NASA, NRAO, NCRA, IRA, UNM/ONR/NRL, ASTRON, SKA,NASAGSFC.

AcronymList

245

AcronymList

2MASS 2MicronAll‐SkySurvey

AAAC AstronomyandAstrophysicsAdvisoryCommittee

AAT Anglo‐AustralianTelescope

ACCESS Actively‐CorrectedCoronagraphsforExoplanetSystemsStudies

ACS AttitudeControlSystem

ACS AdvancedCameraforSurveys

ACTDP AdvancedCryocoolerTechnologyDevelopmentProgram

ADEPT AdvancedDarkEnergyPhysicsTelescope

ADI AngularDifferentialImaging

AIC AchromaticInterfero‐Coronagraph

AIM AstrometricInterferometryMission

AKR AuroralKilometricRadiation

ALADDIN AntarcticL‐bandAstrophysicsDiscoveryDemonstratorforInterferometricNulling

ALMA AtacamaLargeMillimeterArray

AMSD AdvancedMirrorSystemDemonstrator

ANT AchromaticNullingTestbed

AO AdaptiveOptics

AO AnouncementofOpportunity

APD AstrophysicsDivision

APLC ApodizedPupilLyotCoronagraph

ASMCS AstrophysicsStrategicMissionConceptStudy

ASTRA ASTrometricandphase‐ReferencedAstrometryupgradetoKeckInterferometer

ASTRON NetherlandsInstituteforRadioAstronomy

AU AstronomicalUnit

BLC BandLimitedCoronagraph

BLINC BracewellInfraredNullingCryostat

CCD Charge‐CoupledDevice

246

CDR CriticalDesignReview

CEDL Cats‐EyeDelayLine

CFHT Canada‐France‐HawaiiTelescope

CGS4 CooledGratingSpectrometer

CHARA CenterforHighAngularResolutionAstronomy(GeorgiaStateUniversity)

CIRS CompositeInfraredSpectrometer,instrumentontheCassiniorbiter

CLM Combined‐LightMission

CNRS CentreNationaledelaRechercheScientifique

COBE COsmicBackgroundExplorer

CoRoT ConvectionRotationandplanetaryTransitsmission

CP‐AIC Common‐PathAchromaticInterfero‐Coronagraph

CTE CoefficientofThermalExpansion

DFDI DispersedFixed‐DelayInterferometer

DFP Disturbance‐FreePayload

DM DeformableMirror

DQE DetectorQuantumEfficiency

DRIE DeepReactive‐IonEtching

DRM DesignReferenceMission

DSN DeepSpaceNetwork

DUNE DarkUniverseExplorer

EADS EuropeanAeronauticDefenceandSpacecompany

EELV EvolvedExpendableLaunchVehicle

EGP ExtrasolarGiantPlanet

EIRB ExternalIndependentReadinessBoard

ELT ExtremelyLargeTelescope

E‐ELT EuropeanExtremelyLargeTelescope

EM EngineeringMilestone

EOPM EightOctantPhaseMask

EPOCh ExtrasolarPlanetObservationandCharacterization

EPOXI There‐christenedDeepImpactmission

ESA EuropeanSpaceAgency

ESO EuropeanSouthernObservatory

ESTEC EuropeanSpaceResearchandTechnologyCentre

ET ExoplanetTracker

AcronymList

247

ETSO Earth‐TrailingSolarOrbit

ExAO ExtremeAdaptiveOptics

ExoPTF ExoplanetTaskForce

FB‐1 FlightBaselinedesign1

FCT FormationControlTestbed

FGS FineGuidanceSensor(HubbleSpaceTelescope)

FIRAS FarInfraredAbsoluteSpectrophotometeronCOBE

FKSI Fourer‐KelvinStellarInterferometer

FoR FieldofRegard

FOV FieldofView

FPA FocalPlaneArray

FPGA Field‐ProgrammableGateArray

FQPM FourQuadrantPhaseMask

FSM FineSteeringMirror

FSW FlightSoftware

FT FringeTracker

FWHM FullWidthatHalf‐Maximum

GENIE Ground‐basedEuropeanNullingInstrumentatESO

GIMLI GiantPlanetsaroundM,L,andTDwarfsintheInfrared

GMRT GiantMetrewaveRadioTelescope

GMT GiantMagellanTelescope

GPI GeminiPlanetImager

GPS GlobalPositioningSystem

GSFC GoddardSpaceFlightCenter

GSMT GiantSegmentedMirrorTelescope

HARPS HighAccuracyRadialvelocityPlanetarySearch

HCIT HighContrastImagingTestbed

HD HenryDrapercatalog

HF HighFrequency

HIRES HighResolutionEchelleSpectrometerattheW.M.KeckObservatory

HQ Headquarters

HRC HighResolutionChannel

HST HubbleSpaceTelescope

HZ HabitableZone

248

IAC InternationalAstronauticalCongress

IHZ InnerHabitableZone

IR Infrared

IRAC InfraredArrayCameraontheSpitzerSpaceTelescope

IRAS InfraredAstronomicalSatellite

ISIM IntegratedScienceInstrumentsModule

ITT InternationalTelephone&TelegraphCorporation

IWA InnerWorkingAngle

JDEM JointDarkEnergyMission

JPL JetPropulsionLaboratory

JWST JamesWebbSpaceTelescope

KBO KuiperBeltObjects

KI KeckInterferometer

L2 Lagrangepoint#2

LAOG Laboratoired’Astrophysiquedel’ObservatoiredeGrenoble

LBTI LargeBinocularTelescopeInterferometer

LEO LowEarthOrbit

LESIA Laboratoired’EtudesSpatialesetd’InstrumentationenAstrophysique

LM LockheedMartin

LMC LargeMagellanicCloud

LOA LetterofAgreement

LOFAR LowFrequencyArray

LSST LargeSynopticSurveyTelescope

LTD LowTemperatureDetector

LWA LongWavelengthArray

MAM Micro‐ArcsecondMetrologytestbed

MARVELS Multi‐objectApachePointObservatoryRadialVelocityExoplanetLarge‐areaSurvey

MDAS MinimumDetectablePlanetaryAstrometricSignal

MDL Micro‐DevicesLab

MEMS Microelectromechanicalsystems

MIDEX Mid‐sizedExplorermission

MILSTAR MilitaryStrategicandTacticalRelaysatellite

MIRI Mid‐InfraredInstrumentfortheJamesWebbSpaceTelescope

MIT MassachusettsInstituteofTechnology

AcronymList

249

MiXI MiniatureXenonIonmicro‐thruster

MKID MicrowaveKineticInductanceDetector

MMT Multiple‐MirrorTelescope

MMZ ModifiedMachZehnderinterferometer

MOA MicrolensingObservationsinAstrophysics

MOST Microvariability&OscillationsofSTars

MPF MicrolensingPlanetFinder

MRO MagdalenaRidgeObservatory

MSI MichelsonStellarInterferometer

MWIR Mid‐WavelengthInfrared,3–8µm

NASA NationalAeronauticsandSpaceAdministration

NexScI NASAExoplanetScienceInstitute

NAOJ NationalAstronomicalObservatoryofJapan

NEF NewEarthsFacility

NGA NationalGeospatial‐IntelligenceAgency

NGST NorthrupGrummanSpaceTechnology

NICMASS NearInfraredCamerafromMASSachusettsatMt.StromloObservatory

NICMOS NearInfraredCameraandMulti‐ObjectSpectrometer

NIR NearInfrared,0.75–1.4µm

NIREST NullingInfra‐RedExtrasolarSurveyforTPF

NIRSPEC Near‐InfraredEchelleSpectrographfortheKeckIITelescope

NPOI NavyPrototypeOpticalInterferometer

NRC NationalResearchCouncil

NRL NavalResearchLaboratory

NSF NationalScienceFoundation

NWO NewWorldsObserver

OAP Off‐AxisParabola

ODL OpticalDelayLine

OGLE OpticalGravitationalLensingExperiment

OHZ OuterHabitableZone

OPD OpticalPathDifference

OTA OpticalTelescopeAssembly

OTE ObservatoryTelescopeElements

OVC OpticalVortexCoronagraph

250

PACS PhotodetectorArrayCameraandSpectrometeronHerschel

PAF PayloadAttachFitting

PAL PresentAtmosphereLevel

PALMAO PalomarAdaptiveOpticssystem

Pan‐STARRS PanoramicSurveyTelescopeAndRapidResponseSystem

PASP PublicationsoftheAstronomicalSocietyofthePacific

pc Parsec

PDT PlanetDetectionTestbed

PFF PrecisionFormationFlying

PFS PlanetFinderSpectrograph

PHASES PalomarHigh‐precisionAstrometricSearchforExoplanetSystems

PHOENIX HighspectralresolutionechellespectrometerattheGeminiSouthObservatory

PIAA PhaseInducedAmplitudeApodization

PLR Period‐LuminosityRelationship

PRIMA PhaseReferencedImagingandMicro‐arcsecondAstrometry

PSF PointSpreadFunction

PTI PalomarTestbedInterferometer

PWM Pulse‐WidthModulation

QE QuatumEfficiency

R&A Research&Analysis

ReSTAR RenewingSmallTelescopesforAstronomicalResearch

RFI RadioFrequencyInterferenceRFP RequestForProposal

RMS RootMeanSquared

ROSES ResearchOpportunitiesinSpaceandEarthScience

RSC RockwellScientific

RV RadialVelocity

RVS RaytheonVisionSystems

SC SpaceCraft

SCA SensorChipAssembly

SCDU SpectralCalibrationDemonstrationUnit

SE SystemEngineering

SED SpectralEnergyDistribution

SIM SpaceInterferometryMission

SKA SquareKilometerArray

AcronymList

251

SMD ScienceMissionDirectorate

SMEX SmallExplorermission

SNAP SuperNovaAccelerationProbe

SNR SignaltoNoiseRatio

SPACE SpectroscopicAll‐skyCosmicExplorer,ESACosmicVisionproposal

SPHERE Spectro‐PolarimetricHigh‐constrastExoplanetResearch

SPIE SocietyofPhoto‐opticalInstrumentationEngineers

SQUID SuperconductiveQuantumInterferenceDevice

STDT ScienceandTechnologyDefinitionTeam

STIS SpaceTelescopeImagingSpectrograph

STJ SuperconductiveTunnelJunction

STScI SpaceTelescopeScienceInstitute

SWG ScienceWorkingGroup

SWIR Short‐WavelengthInfrared,1.4–3µm

TAC TechnologyAdvisoryCommittee

TAU TelAvivUniversity

TCP‐AIC TandemCommon‐PathAchromaticInterfero‐Coronagraph

T‐EDI TripleSpecExternallyDispersedInterferometry

TES TransitionEdgeSensors

TESS TransitingExoplanetSurveySatellite

TMA Three‐MirrorAnastigmat

TMT ThirtyMeterTelescope

T‐NAR TechnologyNon‐AdvocateReview

TNO NetherlandsOrganizationforAppliedScientificResearch

TPF TerrestrialPlanetFinder

TPF‐C TerrestrialPlanetFinderCoronagraph

TPF‐I TerrestrialPlanetFinderInterferometer

TPF‐O TerrestrialPlanetFinderOcculter

TRL TechnologyReadinessLevel

UCO UniversityofCaliforniaObservatories

UKIRT UnitedKingdomInfra‐RedTelescope

UNI‐PAC UnbalancedNullingInterferometerPhaseandAmplitudeCorrection

UTR‐2 UkranianT‐shapedRadiotelescope

UV Ultraviolet

252

UVES UltravioletandVisualEchelleSpectrographattheVLT

VLA VeryLargeArray

VLBA VeryLongBaselineArray

VLBI VeryLongBaselineInterferometry

VLT VeryLargeTelescope

VLTI VeryLargeTelescopeInterferometer

VNC VisibleNullingCoronagraph

V&V VerificationandValidation

WFC3 WideFieldCamera3

WFS&C WavefrontSensingandControl

exoplanetcommunityreport.pdf - exoplanet exploration

Documents