Coronagraphy Olivier Guyon ([email protected]) http://www.naoj.org/staff/guyon Center for Astronomical Adaptive Optics University of Arizona Subaru Telescope National Astronomical Obs. of Japan 1 Outline 1. Scientific motivation(1) 2. Introduction to Coronagraphy 3. Coronagraph concepts(2) 4. Coronagraphy theoretical limits 5. Wavefront control for Coronagraphs, PSF calibration 6. Current status. Recent lab results, missions (1) This lecture discusses stellar coronagraphy. Science and techniques for solar coronagraphy, or coronagraphy on other non-stellar sources (Planets, AGNs, etc...) are not discussed. (2) Mostly for reference. 52 slides describe coronagraph concepts. We will go through these slides very quickly. 2 1. Scientific motivation 3 Exoplanets How many planets around other stars ? How do they form, evolve ? Mass, size, composition ? Rocky planets with atmospheres ? What is the atmosphere composition ? Weather, rotation period ? Could have life evolved on other planets ? Intelligent life somewhere else ? What does the general public think about imaging exoplanets ? (quotes from http://www.dailygalaxy.com/my_weblog/2008/02/nasas-new-world.html) “What an amazing idea - to stay at the fuzzy image of another world with oceans and continents and perhaps even clouds! And what if it did discover O2, methane or another bio marker? How would this world be changed to have proof - visual proof at that (best kind for a visual-oriented primate) - that there is other life in the universe? Well worth $3bill. Maybe Bill Gates or another billionaire could be convinced to pay up some! :-)” “Maybe, we could stop wasting money on project like this, and spend it on more important issues? We have already discovered other life forms, countless of documents explain they came and landed, and were shot at, and downed and retrieved...” 5 Key science goals Ground-based imaging (Near-IR, with Adaptive Optics) • Planet Detection: Most sensitive to outer young massive planets (complementary to Radial Velocity, astrometry, transits) -> important for testing planetary formation models • Characterization of planets and disks: • Study planet formation by imaging both disks and planets • Spectroscopy • Achieving high contrast close to the star is key to: • constrain mass/age/luminosity relationship (cooling rate) with overlap with RV • capture reflected light: large sample of “old” planet, many known from RV • increase sample size (currently <10, probably most of them are “exceptions” to the rule) Space-based imaging (Visible, extremely high contrast) • Characterization (spectroscopy) of Earth-mass (and above) planets in habitable zone • Simultaneous imaging of exozodi cloud, massive and rocky planets Ground-based near-IR imaging examples (without coronagraph !) Chauvin et al. 2004 Marois et al. 2008 Lagrange et al. 2009 7 Measurements Habitable exoplanet characterization Modeling / Theory with direct imaging Direct imaging (Visible or IR) Radial Astrometry time planet Spectra / exozodi Velocity photometry position colors map & polarization systems of other of other inclination ? albedo (VIS) ? dynamical Direct imagingDirect Eff. temp (IR) ? model phase function atmosphere variation ? (low mass stars) (low Transit spectra ? spectra Transit Atmosphere Rotation Orbit Mass Radius composition period & structure Asteroid exozodi measurements on systems other belt observations System (nulling, ALMA ...) (nulling, tidal surface temperature, dynamical forces pressure & composition ? impact stability frequency Planet overall structure statistics Transit & Transit (Iron, Rock, Water, Atmosphere) Habitability microlensing microlensing Planet formation volcanism models Spectroscopy : biomarkers • Venus & Mars spectra look very similar, dominated by CO2 • Earth spectra has CO2 + O3 + H2O + O2 + CH4 – Together, these gases indicate biological activity 10 Red edge spectral feature • Red edge in Earth spectra due to plants, and remotely detectable 11 Near-IR Earth spectra • Water is easiest to detect • CO2, CH4, O2 Turnbull et al 2006 Kasting 2004 Earthshine Text Turnbull et al. 2006 Characterization (spectra) of habitable planets Coronagraphy ideally suited for characterizing habitable planets around Sun-like stars (maybe also the best stars for habitability) Coronagraphy / nulling (VIS) M K G F A B - small angular separation - High contrast requires large telescope - few targets, far away - star/planet diameter ratio larger -> planets are faint - planet is faint in visible Coronagraphy / nulling (IR) M K G F A B small angular separation few targets, far away requires large baseline -> planets are faint Transit spectroscopy M K G F A B - insufficient SNR - transit probability < 1 - low transit probability <<1 - limited SNR - large period 14 2. Introduction to Coronagraphy 15 Coronagraphy • Coronagraph role is to block starlight and let as much planet light as possible through the system • Most coronagraph designs are a painful tradeoff between coronagraphic rejection and throughput, inner working angle, angular resolution • R&D in coronagraph has been extremely active in the last 10-15 yrs • Many good coronagraph designs now exist • The theoretical limit imposed by fundamental physics is well understood and approached by a few concepts • Coronagraph performance achieved in labs is already much beyond the requirements of ground-based systems, and at or close to requirements for space mission Coronagraphy language Inner Working Angle (IWA), unit = lambda/D How close can a faint source be imaged from the star. IWA is usually defined as the 50% peak throughput point (where the source throughput is 50% of the maximum throughput) Contrast (example: 1e10 contrast, or 1e-10 contrast) What is usually quoted is the raw contrast (not the detection contrast). Raw contrast = ratio of local surface brightness to peak PSF surface brightness Null order (example: 4th order null coronagraph) Quantifies coronagraph throughput as a function of angular separation close to the optical axis. Higher order = more immune to residual pointing error and stellar angular size (but larger IWA) Why coronagraphy ? Coronagraph can only remove known & static diffraction pattern BUT: - static & known diffraction can be removed in the computer - coronagraphs don’t remove speckles due to WF errors Fundamental reasons: (1) Photon Noise (2) Coherent amplification between speckles and diffraction pattern Practical reasons: (3) Avoid detector saturation / bleeding (4) Limit scattering in optics -> “stop light as soon as you can” (5) Difficult to subtract 2 large quantities to look for small18 residuals Coherent amplification between speckles and diffraction pattern Final image = PSF diffraction (Airy) + speckle halo This equation is true in complex amplitude, not in intensity. I = |A|^2 + 2|AB| + |B|^2 Intensity image will have product term -> speckles are amplified by the PSF diffraction. IF Airy>speckles, then coronagraph is required 19 Aime & Soummer 2004 When is a coronagraph required ? Coronagraphs serve no purpose if dynamic speckle halo is > diffraction -> Very important to keep in mind to avoid over-designing the coronagraph, as this usually would mean giving up something (usually throughput) “Side effects” of coronagraphs : - (Usually) requires very good pointing. Risk of low order aberrations (for example pointing) creating additional scattered light in the region of interest - data interpretation & analysis can be challenging (especially at inner working angle) - Astrometry more difficult (solutions exist) 20 Geometrical optics Coronagraph 21 Geometrical optics coronagraph Geometrical problem: To work at diffraction limit of the telescope, mask needs to be very large and far from telescope (mask should be unresolved by telescope, and needs to be larger than telescope size) + diffraction problem (Arago spot) Solution: put the mask at a plane conjugated to infinity INSIDE the telescope = focal plane BUT: diffraction problem again (Airy pattern) ! Stellar coronagraphy is a diffraction problem, not a geometrical one. 23 Fourier Optics Most coronagraphs place masks in pupil and/or focal planes. Fourier transform links the complex amplitude in these planes. P0 F0 P1 F1 P0 = P x PM0 P = telescope pupil c.a. F0 = FT(P x PM0) x FM PM0 = pupil mask P1 = LM x invFT( FT(P x PM0) x FM ) P0 = coronagraph pupil c.a. F1 = FT( LM x invFT( FT(P x PM0) x FM )) FM = focal plane mask F0 = focal plane c.a. product (x) <-(Fourier Transform)-> convolution (⊗) LM = Lyot mask P1 = Lyot plane c.a. F0 = FT(P) ⊗ FT(PM0) x FM F1 = final focal plane c.a. P1 = LM x ((P x PM0) ⊗ FT(FM)) F1 = FT(LM) ⊗ ((FT(P) ⊗ FT(PM0)) x FM) 24 Fourier Optics & mathematical tools In Lyot-type configuration, the coronagraph design must minimize: P1 = LM x invFT( FT(P x PM0) x FM ) P1 = LM x ((P x PM0) ⊗ FT(FM)) Several mathematical approached: - tune PM0 (pupil mask) Prolate spheroidal functions are key to this optimization, as they are the eigenfunctions of the above equation. Soummer et al. 2003A&A...397.1161S - Tune FM (focal plane mask) Band-limited masks can be used to solve the equation, as their Fourier transform is bounded within a domain the pupil. Kuchner & Traub 2002ApJ...570..900K 25 3. Coronagraph concepts 26 Lyot Coronagraph 27 Lyot Coronagraph figure from Lyot project website Improvement on the Lyot concept Part I: Amplitude masks In the Lyot Coronagraph, some light is leaking. The smaller the focal plane mask, the stronger the light leaks are. ! Apodized Pupil Lyot Coronagraph !!!!!!APLC Soummer et al. 2003, Abe et al. Modify (amplitude apodization)