ESAVoyage2050WhitePaperOcculter to Earth: Prospects for studying Earth-likeplanetswiththeE-ELTandaspace-basedocculterDirectdetectionandcharacterizationofEarth-likeplanetsaroundSun-likestarsisacoretaskforevaluatingtheprevalenceofhabitabilityandlifeintheuniverse.Here,wediscussapromisingoption for achieving this goal,which isbasedonplacinganocculterinorbitandhavingitprojectitsshadowontotheE-ELTatthesurface of Earth, thus providing a sufficient contrast for imaging and takingspectraofEarth-likeplanetsinthehabitablezonesofSun-likestars.Doingsoatasensiblefuelbudgetwillrequiretailoredorbits,anocculterwithahigharea-to-mass ratio, and appropriate instrumentation at the E-ELT. We outline thefundamental aspects of the concept, and the most important technicaldevelopmentsthatwillberequiredtodevelopafullmission.Contactscientist:Markus Janson,Dept ofAstronomy, StockholmUniversity, AlbanovaUniversityCenter,SE-10691,Sweden

ExoplanetdetectionThepasttwodecadeshaveseenanexplosioninthefieldofexoplanetresearch,where thescientificconsensushasgone fromnotknowingwhetherexoplanetsexistaltogether,tonowknowingthatexoplanetsarenearlyubiquitousaroundawiderangeofstellarhosttypes(e.g.Burkeetal.2015;Dressing&Charbonneau2015). These insights have largely been acquired through the radial velocity(RV)andtransittechniques,whichhaveprovidedmeanstoefficientlydetectandcharacterize planets in small orbits around quiescent stars (e.g. Howard et al.2013; Coughlin et al. 2016). At increasing planet separation, however, thesetechniques become progressively less efficient, which in practice means thatother methods are necessary to detect and characterize a wider exoplanetdemographic. Methods that can provide the necessary sensitivity at widerseparations (and for less quiescent stars) include gravitational microlensing,astrometry,anddirectimaging.Microlensingconstitutesaspecialcaseinthatithas the potential to provide a statistical census of exoplanet demographics,potentially down to Ganymede-like masses (e.g. Penny et al. 2019); but theplanetarysystemsdetectedwiththistechniquewillgenerallybesodistantthatthey cannot feasibly be followed up past the transient lensing event, so nodetailedcharacterizationoftheplanetsispossibleviamicrolensing.Bycontrast,astrometry and direct imaging favour detections in nearby systems, which isidealfordetailedcharacterization.Astrometry has so far only lead to a few individual exoplanet detections (e.g.Benedict et al. 2002; Snellen & Brown 2018), but once Gaia releases its finalcatalog, this situation will drastically change, with astrometry most likelyproviding more exoplanet detections than RV and transits combined (e.g.Perryman et al. 2014). The astrometric signature of a star-planet systemincreaseswithincreasingsemi-majoraxis,thusresultinginanoppositebiaswithrespect to RV and transits, and an opportunity to study the wider planetpopulation.Itisalsorelativelyinsensitivetotheorientationofthetargetsystem,unlikeRV,whichisinsensitivetonearface-onsystems;andevenmoresounliketransits, which strictly require near edge-on systems for detectability. It alsoyieldsadirectmeasureoftheplanetmass,incontrasttoRV,whichonlyprovidesa lower mass limit unless the system geometry can be independentlyconstrained.To directly image planets themselves is the most intuitively obvious way todetectthem,butitrequirestheabilitytoovercomethevastbrightnesscontrastbetweenaplanetand itsparentstarata smallangularseparation,whichuntilrecently was an excessively difficult technical task. However, technologicaldevelopments in adaptive optics and coronagraphy have been rapid, and nowenable the detection of sufficiently wide, massive, and young planets withground-basedtelescopes(Maroisetal.2008;Lagrangeetal.2009).Planetsareeasiertoimage(intheinfrared)whentheyareyoung,sincetheystartoutwithalotofheatfromtheirformation,andthengraduallycoolovertime.Thisopensupan opportunity to study the formation and early evolution of planets (e.g.Keppleretal.2018),whichisdifficultwithmethodssuchasRVduetothehighactivity levels of young stars. Direct imaging is the most versatile of all

techniques, opening up opportunities for (e.g.) orbital monitoring, spectralcharacterization, variability characterization, and instantaneous study ofmultipleplanetswithinthesamedataset.Aswewilldetailinthenextsection,itisalso thekey technique for studyingEarth-likeplanets in thehabitablezones(HZs)ofSun-likestars.TowardhabitableexoplanetsThetransitandRVtechniquesarenowatthelevelwhereEarth-sizedplanets(orevensmaller,inthecaseoftransits)canbedetectedinsomecircumstances.Thisis mostly in the cases of small orbits around small stars, since both of thoseaspectsarefavouredinRVandtransitsearches.Coincidentally,smallstarsalsohavetheirHZsclose in,duetotheir lowerbolometric luminosities.Thismeansthat the lower the mass of the star, the easier it is to detect planets in theclassicalHZarounditwithRV/transits,providedthat itremainsbrightenoughfortheinstrumentsthatobserveittoacquireasufficientsignal.Indeed,someofthelowest-massstarsintheSolarneighbourhoodnowhavedetectedplanets(orplanet candidates) in their HZs (e.g. Anglada-Escude et al. 2016; Gillon et al.2017). In the future, there may be a promising option to characterize suchplanets with transmission spectroscopy, in order to map their atmosphericcontentandsearchforbiomarkermolecules.However, this pathway toward studying habitability has several limitations.Firstly,it’snotclearifplanetsaroundlow-massstarscanbehabitableinthefirstplace.TheHZsaroundmid/lateM-typestarsexperienceextremelevelsofX-rayandUVradiationthatdissociatewatermolecules,andstrongstellarwindsthatmay strip away the hydrogen component on rapid timescales. If the planetcannotholdontoitswater,itsequilibriumtemperature(whichdefinestheHZ)is inconsequential. In addition, flare intensities aremuch higher, whichmighterode the atmosphere; planets in the HZ are tidally locked, which impactsatmosphericdynamicsandtheabilitytogenerateplanetarymagneticfields;andlow-mass stars have a very extended pre-main sequence phase, during whichtemperaturesinwhatwilleventuallybecometheHZaremuchtoohighforliquidwater to occur. Secondly, even for the lowest-mass stars, transmissionspectroscopyisverychallenging,becauseitonlyprobestheterminatorregionofaplanet,whichisatinyfractionofitstotalarea.Indeed,eventhefutureESAM-class mission ARIEL, which is specifically designed for transmissionspectroscopy,willnotprobeHZplanets,exceptforveryparticularpurposes(e.g.,probingwhethertheTRAPPISTplanetshaveextendedhydrogenatmospheres).Thirdly, and related to the previous points, transmission spectroscopy scalespoorlyasymptoticallytowardmoreambitiousmissions,andespeciallyintermsof reaching HZ planets around Sun-like stars, which is orders of magnitudebeyondthedifficultyofprobinglower-massstarswiththistechnique.IntermsofdetectingtrulyEarth-equivalentplanets(samesize,stellarhosttype,etc. as the Earth), all of the basic methods described above can contribute indifferentways:TransitscanprovideastatisticalcensusofEarth-likeHZplanetsthroughESA’sfutureM-classPLATOmission(Raueretal.2016).RVinstruments

are being developed with the aim of reaching sensitivity to Earth-equivalentplanets, at least aroundquiet stars (Pasquini et al. 2010).Gaia is not sensitiveenoughtodetectterrestrialplanets,butastrometryconceptssuchasTHEIA(PI:C.Boehm;seee.g.Jansonetal.2018fordetails)existthatwouldbeabletoyieldthis sensitivity, and could therefore provide the invaluable resource of acomplete census of Earth-like (and other) planets in the Solar neighbourhood,withmassmeasurementsineachcase.Directimaging,aswewillseebelow,canalsodetectEarth-equivalentplanetswithconceptualmissionssuchasHabExorLUVOIR.However,thesituationforexoplanetcharacterizationisdifferent:Inthiscontext,directimagingistheonlymethodthatcanfeasiblycharacterizetheatmospheresofEarth-likeplanetsintheHZsofSun-likestars,whichinturnaretheonlykindsofplanetsthatwealreadyknowcanbehabitable.Thisisafundamentalreasonforwhydirectimagingisakeytechnologyinexoplanetresearch,andtheaimofmost large-scale exoplanet-related mission concepts among the main (multi)-nationalspaceagencies.DirectimagingconceptsAt present date, most high-contrast imaging is performed with ground-basedtelescopes. This is primarily due to availability of large aperture telescopes,which is a fundamentally important issue in exoplanet imaging due to theintrinsic faintness of planets in combination with the need to separate themspatiallyfromthePSFoftheparentstar.So-calledExtreme-AO(ExAO)systemsenabletocorrectthewavefrontdistortionsimposedbytheatmospheretoagoodextent,reachingStrehlratiosinexcessof90%inthenear-infrared(where100%would imply a perfectly diffraction-limited PSF), and coronagraphy helps insubstantiallyreducingstellarlightwithoutaffectingplanetarylight.Nonetheless,the fact that the coronagraph needs to operate on a non-perfect wavefrontcurrentlysetsthelimitfortheachievablecontrastfromtheground,whichisoforder10-7atsmallseparationswiththehelpofdifferentialimagingtechniquesinthebestcases.Forcomparison,thefinalcontrastrequiredtoimageanEarth-likeplanetaroundaSun-likestaris10-10(inreflectedlight)Thisfacthasdriventhedevelopmentofspacetelescopeconceptsforexoplanetimaging, which avoids the issues related to the Earth’s atmosphere. Suchconcepts are typically either based on internal coronagraphs or occulters, asdescribed in the next section. WFIRST is planned to have a sophisticatedcoronagraphicmode andwill be of great help for developing such techniques,butitsHubble-classaperture(2.4m)istoosmalltoallowforreachingtrueEarthanalogs. Thus, plans are in motion for larger aperture space missions toaccomplish this, including concepts such as HabEx and LUVOIR. We will notdiscuss these particular concepts in detail in this document, but insteadconcentrateongeneralprinciples surrounding them.Wenote that adedicatedwhite paper on these concepts is being submitted by Ignas Snellen andcoauthors.

Another path in the context of direct imaging in space is the usage of nullinginterferometry for thermal imaging in the infrared, with a formation-flyingsatellitefleet.Wewillnotdiscussthisconceptthoroughlyeither,butrefertothewhitepaperbySaschaP.Quanzandco-authorsformoredetails.OccultersversuscoronagraphsCoronagraphs come in many different forms based on a range of differenttechniques,butthecommonfeatureinthecontextofthisdiscussionisthattheyareinternaltothetelescopedesign,downstreamfromtheprimarymirror.Thismeansthattheircontrastperformanceisnecessarilydiffractionrelated,withaninner working angle (IWA) determined by the telescope size (and observingwavelength),and italsomeansthat theupstreamopticsneedstobeextremelypreciseinordertomaintainaperfectwavefrontforthecoronagraphtooperateon. By contrast, an occulter is external to the telescope, and placed at a largeseparation away from it. The occulter edge is apodized in order to minimizediffractionarounditsedges.Withidealapodizationandasufficientlylargesize,an occulter can produce a 10-10 reduction of incoming starlight in the shadowbehind it, and a telescope placed in the center of this shadow at a sufficientlylargeseparationwillbeabletoseecompanionstothestaroutsideoftheedgesofthe occulter,with the starlight already removed in thewavefront that hits theprimarymirror.Inthisframework,theIWAissetbytheoccultersizerelativetoitsdistancefromthetelescope,independentlyofthetelescopesize.Furthermore,the specifications for the telescopeoptics becomenomoredemanding than inregularnon-coronagraphicimaging,sobasicallyanytelescopecouldbeusedorrelativelyeasilyadaptedtooperateinconjunctionwiththeocculter.

Fig.1:Occulterdesignexample.Theocculterconsistsofacentralopaquecircularstructure, surroundedbya ring of petals that gradually narrowdownoutwards.Thispatternhelps tomitigatediffractionaroundtheocculteredges to theextentthataveryhighcontrastdepthisacquiredatthecenteroftheshadowcreatedbytheocculter.Imagecredit:NASA

The two techniques have their respective pros and cons. Operationally,coronagraphsarepreferableoverocculters,becauseitisvastlyeasierandfastertoturnyourtelescopefromthedirectionofyourcurrenttargettothedirectionofthenextonethanitistomovetheocculterthousandsofkilometersfromonetarget to the next. In general, the fact that the occulter requires a separatesatellite unit obviously adds cost and complexity relative to a coronagraphicmissionthat(hypothetically) fulfils thesamesciencerequirements.However,agreat advantage of the occulter is that it is simply technologically easier toacquire a high contrast than it is with a coronagraph. A full occulter-basedmission with sensitivity to a large number of Earth-like HZ planets could belaunchedwithminimaldevelopmentgiventhecurrenttechnologicalreadiness–thefinancialaspectofsuchamissionistheonlylimitationtoitsfeasibility.OccultertoEarthSo far, we have been discussing HabEx-like concepts where a space-basedtelescopeoperatesinconjunctionwithaspace-basedocculter.Inthissetup,thetelescopeitselfobviouslycontributessubstantiallytothemissioncostevenforasmall aperture size, and this cost increases dramatically as the aperture sizeincreases.Meanwhile, telescopesexistandarebeingconstructedonEarth thatarefarlargerthananythingthatexistsorisplannedinspace.Ifanoccultercouldfit a telescope on the ground in its shadowwith respect to a target star for asufficientlylongtime(say,anhour),thiswouldnotonlysavethewholecostoflaunching the telescope, but itwould also yield data ofmuch higher signal-to-noiseratio(S/N)thanwouldbeachievablewiththespace-basedunit.

Fig.2:StarsforwhichEarthsintheHZ(ifpresent)wouldbedetectableinlessthan2hourswithan8m-classtelescopeforanEarth-trailingocculterconcept(Janson2007).Possible targetswould includeeverything fromBA-type (yellow,green) toM-type(cyan)starsandevenpost-MSgiants(magenta),but inparticulara largenumberofFGK-type(blue,black,red)MSstars.

Wehavepreviouslyseen that theatmosphereruins theopportunity toacquirevery high contrast imaging from the groundwith coronagraphy, so onemightintuitivelyexpectthatthesamewouldbetruefortheocculteroption.However,thisisnotthecase.Inthesetupwearedescribing,theocculteroperatesinspace,with no atmosphere around. It thus creates a wavefront with a high-qualityshadow,which,foratelescopeinsideoftheshadow,willbemappedasanimageof the system with the central star already diminished by a factor 1010. Theocculted image will be smeared by atmospheric seeing, but this isinconsequentialwithregardstothecontrast,sincetheseeinghaloofthestarwillanyway be 10-10 of its original brightness. It is however necessary to applyadaptive optics in order to collect light into the PSF core of any planetssurrounding the star, since they would otherwise drown in the backgroundemissionfromzodiacalandexozodiacaldust.Thekeyaspectoftransmittingashadowfromspacetothegroundistoidentifyan orbit for the occulter in which the shadow can be maintained fixed withrespecttoagivencoordinateonthegroundforalongenoughtimetoacquireagoodS/N,atareasonablefuelconsumption.Meanwhile,theEarthisrotatingandanobjectinorbitundergoesconstantmotion,whichisthecentralchallengethatneeds to be overcome. Ideally, the orbit of the occulter during an observationshouldbesuchthattheprojectedvelocityofatelescopeonEarthasseenfromatargetstarisperfectlymatchedforanextendedamountoftime.Whilethiscan’tbe done perfectly due to the fact that the projectedmotion of a point on thegroundchangescontinuouslywiththerotationalphaseoftheEarth,smallactiveboosts can be imposed on the occulter to create a good matching speed at amodestcostinfuel.Forasingletarget,thisisasimpleproblemtosolve,butifwewant to observe several targets,we also need ameans to transition from onetarget to the next, and to time each new target up in phase such that thetelescope and occulter line up with the target star simultaneously for eachobservation.Thisisasignificantlymoredifficultproblem.In Janson (2007),we demonstrated a type of inclined Earth-trailing orbit thatallowsforshadowcontrolontopofaground-basedtelescopefortimescalesoforder hours per telescope, as well as timed target transitions and continuousorbit corrections at a distinctly modest fuel demand. A central trade-off thatallows this to work is that the sky coverage is limited to a band betweendeclinationsofapproximately+/-5degonthesky.Recently,analternativesetoforbits has been developed by JohnMather and collaborators at GSFC and JPLbasedonhighlyellipticalorbits.Thisclassoforbits ismoredirectable,butalsoimposes a challenge for large transitions across the sky. These exercisesdemonstrate that a range of orbits does exist that can fulfil all the basicconstraints for the occulter shadow control, with different trade-offs withregards toskycoverageand targetsamplesize. In theend, thespecificsciencerequirements of the mission would guide the exact choice of orbits for theocculter.

Fig. 3: Matching projected speeds between ground-based telescopes at variouslatitudes (solid lines) and occulters at different separations (dashed lines). Theocculterdriftsinacontrolledmannerfromonetargettothenext,andthenboostsactivelyforafewhoursduringatargetrendezvousinordertokeepinexactpacewiththeEarth’srotation.FromJanson(2007).OccultertotheE-ELTTransmitting an occulter shadow to the ground opens up the opportunity foraperturesizesmuchlargerthanwhatwouldbepossibletolaunchintospace.Inparticular,the~40mE-ELTtelescopeisnowunderconstruction,providingthelargest aperture available anywhere for optical/near-infrared purposes.Therefore,now isagood time toexamine theprospectsof combininga space-based occulter with the E-ELT, potentially providing extremely high S/N forEarth-analog planets relative to what is accessible with any other facility orcombination of facilities. Indeed, the main purpose of this white paper is tomotivatetechnicalstudiesforthispurpose.ThecontrastandIWAprovidedbyagivenocculterisintrinsicallyindependentoftelescopeproperties.TheonlyadditionalconstraintimposedbyusageoftheE-ELTrelativetoanyothertelescopeisthatthe lineardimensionsoftheshadownecessarily have to be large enough to fit the entire aperture. An apodizedocculterconsistsofacentralopaquecircleofdiameterD,withnarrowingpetalsextending out to approximately 2D. The shadow dimensions cannot be largerthanD,andwillgenerallybeabitsmallerforthepartoftheshadowthatreaches10-10contrast,butcanbequiteclosetoD(seee.g.Cash2006).Theexactsizeofthispartoftheshadowdependsonthespecificocculterdesignandbecomespartof an optimization scheme when designing an occulter to provide a givencontrast for a shadowof a given size over a givenwavelength range.Herewesimply note that the ~40 m diameter of E-ELT sets the constraint that theocculterneedstobeatleast80mfromtiptotipofitspetals.Suchasizeiswell

compatible with occulter-to-Earth concepts: The separation between theocculterandtelescopeinsuchconceptsseparatelysetsarequirementofoccultersizesoforder100minordertoreachasufficientcontrast.Thankstothelargeseparation, the IWA in these concepts typically gets very good, in the rangeof~20-60mas,i.e.sufficienttodetectEarthanalogs(1AU)outto~16-50pc.ThisinturngivesaccesstoSun-likestarsinthehundreds;howmanyofthemcanbeobservedisaquestionoffuelbudget,asdiscussedmorebelow.

Fig. 4: Rendering of the E-ELT telescope currently under construction at CerroArmazonesinChile.Imagecredit:ESOTechnicaldevelopmentsThemainpurposeofthisdocumentistomotivatestudiesregardinganumberoftechnological developments that should be performed in order to furtherenhancethereadinessofsuchamissionandevaluatekeycharacteristicsoftheconcept.Welistthemostimportantaspectsbelow.OcculterdesignStudiesregardingocculterdesignandconstructionhavelargelybeenperformedintheUSsofar,butitwouldbebeneficialfortheEuropeancommunityforsuchstudies tobeperformedalso inaEuropean framework.Theocculterhassomenecessary electronic components (such as boosters, power system, possibly alaser for guide star purposes) but is otherwise a potentially fairly simplemechanicalsystem.Challengesincludetheissueofgettingsuchalargestructureinto orbit (the American designs include intricate folding techniques), andensuring a minimal degree of scattered or reflected sunlight, which is animportantconsiderationprimarilyattheedgesoftheocculter.Akeyissueistofindwaysofminimizingmassinthedesign.Themassdirectlysetsthescopeofthemission:Orbital calculations canbeused to calculatedelta-Vvaluesand to

findthesolutionsthatminimizetherequireddelta-Vforshadowcontrol(whichgoverns the available amountof observing time for each target) and target-to-targetmotion(whichgovernsthetotalamountoftargetsthatcanbeobserved).Givenanidealorbit,themassoftheocculteristhenthefundamentalparameterthat completely governs the total number of targets and the total hours ofintegrationtimeavailableforthemission.AninterestingdetailinthiscontextisthatNASAisconsideringasystem(IRMA)that could assemble structures in orbit, thus allowing for physically largerstructuresthancanbelaunchedwithexistinglauncherfacilities.Itwouldnotbeable toassemble,e.g., full telescopeunits,due to theprecisionrequirementsofopticalsystems.However,alarge2Dmechanicalsystemsuchasanoccultermaybeanexcellentmatchtosuchafacility. Wealsonotethatthereisapotentiallyfruitful synergy with solar sails that could be utilized, since there are manyoverlappingcriteriabetweenwhatconstitutesagoodocculterandagoodsolarsail,intermsoflargerigid2D-structureswithahighsize-to-massratio.OptimalorbitsAswehavealready touchedupon, theorbital characteristics are crucial to themission design. What constitutes an “optimal” orbit in this regard is contextdependent.Forexample, if thescientificgoal is toobservea limitednumberofvery high-merit targets that are spreadwidely across the sky, elliptical orbitsmay be preferable, while if the goal is to observe a larger number of targetswithoutpriorinformationaboutrelativemerits,anEarth-trailingorbitmaybeabetteroption.Theremayalsobeorbitalsolutionsforagivensetofrequirementsthathaveyet tobediscovered. Since there isan interplaybetween thesciencerequirementsandtheorbitoptimization,thereneedstoexistadialoguebetweenresearchers and engineers on the orbital dynamics side. ESA would be thenaturalcontactpointwithinEuropeanastronomyforthispurpose.E-ELTinstrumentationIt is obviously required that the E-ELT has suitable instrumentation for theobservations that need to be executed once the occulter shadow lands on it.Since theserequirementsoverlap inmanywayswithgeneral requirements forground-basedhigh-contrastimaging,thisshouldberelativelystraightforwardtoimplement, but it needs tobepart of theplan that all of the requirements aredefinitely in place by the time the occulter launches, which in turn requiresinterplay between ESA and ESO in the mission planning phases. The mainrequirementsforasuitableE-ELTinstrumentareasfollows:Firstly, it needs to operate in the visiblewavelength range.Much of the high-contrast imaging that is performed from the ground is done at near-infraredwavelengths, but this is unsuitable for an occulter. The contrast that can beachievedwith an occulter is wavelength dependent and becomes increasinglychallengingatlongerwavelengths.Theexactwavelengthrangethatiscoveredat

sufficientcontrastdependsonthespecificocculterdesign(e.g.Vanderbeietal.2007),butingeneralterms,achievinggoodcontrastat>1micronwavelengthsisexcessivelychallengingwithoccultertechnology.Secondly, it needs to acquire as high Strehl ratio as possible. The Strehl ratiogovernshowmuchoftheplanetarysignaliscollectedintoitsdiffraction-limitedPSFcore,andhowmuchspillsoutintoitsseeing-limitedhalo.AhighStrehlratioimplies a high fraction of the light collected into the core. This fraction of thelight is theonly light thatwillbeusefullyobservable to the telescope,becausethe halo light will blend together with the zodiacal and exozodiacal lightpollutionandthusnotcontributebeneficiallytotheS/N.Thediffraction-limitedcorefora~40mtelescope,ontheotherhand, is fullydistinct fromthediffusebackground, to the extent that the observation becomes photon noise limited.Thissecondrequirementispartlyattensionwiththefirstone,inthatitismoredifficult to acquire a high Strehl ratio in the visible than in the near-infrared.However, the existing ZIMPOL arm of the SPHERE instrument at VLTdemonstratesthathighStrehlratiosareinfactpossibleinthevisible,routinelyprovidingStrehlratiosof50%andaboveintheredpartofthespectrum.Additionally,itispreferableiftheinstrumenthasIFUcapability.WithanIFU,aspectrumforallpointsourcesinthefieldofviewcanbeacquireduponfirstvisit,without prior knowledge of their locations. This is very useful for an occulterconcept,wherere-visitingagiventargetcanbequitechallenging.Forexample,the Earth-trailing orbit concept described above can only re-visit targets on ayearlybasis,whichisgoodforcommonpropermotionconformationandorbitalcharacterizationsincealotofmotionhasoccurredinbetween;butfortheexactsamereason,itisbadforspectroscopicfollow-upwith,e.g.,aslit,sinceitcannotbe predicted exactly where the companion will re-occur. If a priori orbitalinformation exists, for example from RV or astrometry data, this is less of anissue.An instrument that could potentially deliver on all of these points is PCS,formerlyEPICS,which is foreseenasasecond-generation instrumentontheE-ELTwith first light in the2030s (PI:M.Kasper,ESO). Indesigninganoccultermission,itshouldbecarefullystudiedwhetherPCSissuitablefordeliveringthenecessarydataandifanymodificationscanbedonetobetteraccommodatethispurpose, or whether even a special-purpose instrument may be a desirableoption.SummaryInthisdocument,wehaveoutlinedsomeofthemainaspectsofanocculter-to-Earthconcept,whereanocculterinspaceprojectsadeepshadowontoaspecificlocation on Earth – specifically, the site of E-ELT – and some of the maintechnical considerations thatwe argue should take placewithin the Europeancommunity, with collaborations around theworldwhenever applicable. A keygoal in exoplanet science, as well as in astronomical science as a whole, is toevaluateprevalenceandconditionsforhabitabilityandlifeintheuniverse.This

canonlybeaccomplishedbydetectingandcharacterizingastatisticalsampleofpotentially habitable worlds. The concept presented here constitutes onepromising way of accomplishing this within sensible budgetary constraints.Other possible concepts for doing so exist, and are presented inwhite paperssubmittedbySnellenetal.andbyQuanzetal.Theexistenceofseveraldifferentconcepts for similar scientific purposes should not be seen as a sign ofdivisivenesswithin theexoplanetcommunity.Rather, itdisplays that thereareseveral promising paths toward a very broadly commonly accepted scientificgoal. Depending on scientific and budgetary constraints, it may benefit theEuropeancommunitytochoseonepathoranotheruponcarefulevaluation.Theocculter-to-Earth concept presented here represents a path in which Europecould potentially play a leading role in the detection and characterization ofhabitable planets beyond the Solar system. The E-ELT will be a thoroughlyunique facility, unmatched in size and scope by any other existing or plannedtelescope. By combining it with an occulter in space, it would become aformidabletoolforobservingthehabitableuniverse.

ReferencesAnglada-Escudeetal.2016,Nature536,437 Benedictetal.2002,ApJ581,L115Burkeetal.2015,ApJ809,8Cash2006,Nature442,51Coughlinetal.2016,ApJS224,12Dressing&Charbonneau2015,ApJ807,45Gillonetal.2017,Nature542,456 Janson2007,PASP119,214 Jansonetal.2018,HandbookofExoplanets,SpringerInternationalPublishingAGKeppleretal.2018,A&A617,44Lagrangeetal.2009,A&A495,335Maroisetal.2008,Science322,1348Pasquinietal.2010,Messenger140,20Pennyetal.2019,ApJS241,3Perrymanetal.2014,ApJ797,14Raueretal.2016,AN337,961Snellen&Brown2018,Nat.Ast.2,883Vanderbeietal.2007,ApJ665,794Howardetal.2013,Science340,572

ListofauthorsMarkusJansonStockholmUniversitymarkus.janson@astro.su.seThomasHenningMPIAHeidelberghenning@mpia.deSaschaQuanzETHZurichsascha.quanz@astro.phys.ethz.chRubenAsensio-TorresStockholmUniversityruben.torres@astro.su.seLarsBuchhaveDTUSpacebuchhave@space.dtu.dkOliverKrauseMPIAHeidelbergkrause@mpia.deEnricPalleIACepalle@iac.esAlexisBrandekerStockholmUniversityalexis@astro.su.se