On NIR HST Spectrophotometry of Transiting Exoplanets Vasisht

On NIR HST Spectro-photometry of Transiting Exo-planets Vasisht, Swain, Deroo, Chen, Wayne (JPL), Tinetti (UCL), Yung (CIT), Angerhausen (PIUK), Bouwman (MPIA), Deming (GSFC) JPL-CIT Paris 2008

Outline § Motivation for NIR objective mode, time-resolved spectroscopy § Instrumental issues > General issues confronting shot-noise limited spectroscopy > Hubble specific limitations § Modeling and removal of instrumental limitations § Spectroscopy of the emergent flux from HD 189733 b (Swain, Vasisht, Tinetti, Deroo, Yung et al. , accepted Ap. JL) JPL-CIT Paris 2008

Spectroscopy with NICMOS § Slitless Grism spectrograph between 0. 82. 5 microns § Camera 3 – Coverage by 3 grisms located in filter wheel – R = 200 native spectral resolution – Camera 3 is severely under-sampled at 0. 2”/pixel (~λ/D @ 2 um, 52” FOV) § Exoplanet datasets have now been acquired with all grisms – G 141 on HD 209548 b (Brown et al. in 2005, unpublished) JPL-CIT Paris 2008

Scientific Rationale § NIR Emission Spectroscopy (λ ~1 -2. 5 um; Spitzer 3 -30 um) – Observable: Falling but favorable flux contrast (< 3 um) – Energetically Important: Maximum νFν (for emergent flux) – Decreased stellar shot-noise – NIR photosphere at greater pressure depths (0. 1 -1 bar) – Molecular activity: ro-vib bands of major species § Again some of the same advantages apply for transmission spectroscopy – Reduced opacity from small particle scattering JPL-CIT Paris 2008

Hot, Cold or Cloudy Hot T = 1750 K dayside reradiation JPL-CIT Paris 2008 Seager et al. 2005 Homogenous clouds

Active (common) C, N, O molecules Lodders & Fegley 2002 � JPL-CIT Paris 2008

Hubeny & Burrows 2008 JPL-CIT Paris 2008

§ Molecular spectroscopy -> atmospheric physics – Atmospheres are a window to planetary composition, may have clues to evolutionary history – History of the planet can give rise to a range in core sizes, heavy element abundances, and abundance ratios – Relative fractions of refractory and volatile materials should reflect upon § Parent star abundances, history of formation, migration (? ) JPL-CIT Paris 2008

Part II – Photometry with HST 1. Detector anomalies 2. Optical anomalies JPL-CIT Photometric systematic noise Paris 2008

NICMOS Detector Effects § Stress induced structure in the response § Pixel-to-Pixel stochastic response variations § Intrapixel structure in the response § T-dependence Figer et al. 2002 JPL-CIT Paris 2008

Large scale structure NIC-3 is undersampled PAM Defocus provides some “Immunity” This sets R ~ 40 Watch for structure under spectrum. Flats can remove some of this power JPL-CIT Paris 2008

Small-scale structure and MTF Finger et al. 2000 JPL-CIT Paris 2008 Stiavelli et al.

Relative Photometry Evaluate in some statistical fashion JPL-CIT Paris 2008

Relative Photometry k-space § Variance is integral over spatial frequencies of – Power spectrum of the detector response apodised by § 1. Power spectrum of the illumination § 2. 1 -cos() high pass filter JPL-CIT Paris 2008

Diffraction PSF Intrapixel gain Defocused PSF by Ray Tracing: Note this is a PSD 1 -cos(k dx), dx = 0. 1 pix 14/08/2008 Paris 2008

Implications § Significant substructure in the psf (ILS) – At spatial frequencies of D/λ, D/2λ etc – Due to diffraction – D/λ ~ 1/pixel – Mostly preserved in cross-dispersion axis § Varies with wavelength – For shorter λ, higher spatial frequencies § Can interact with sub-pixel structure JPL-CIT Paris 2008

§ Beam wander § In x (spatial) and y (spectral) § Repositioning errors – Filter wheel positioning – Rot. about un-deviated ray § Orbital phase PSF modulation – Proxy (Gaussian FWHM) § Array response variations – QE with temperature – JPL-CIT DISCRETE OFFSETS X, Y, θ, T PERIODIC ~ 1%/K (2. 5 micron), 3%/K (1. 5 micron) σ Paris 2008

dx, dy, dθ Σ d. I d. T dσ § Biggest headache is image motion § Repositioning errors (Monte Carlo) – δx, δy ~ 0. 1 pixel; linear perturbations – δx, δy > 0. 25 pixels; large higher order errors (> 10 -4) § Generally few usable orbits per visit – Adding 2 nd order terms to expansion is problematic JPL-CIT Paris 2008

JPL-CIT Paris 2008

JPL-CIT Paris 2008

Orbit 1 JPL-CIT Paris 2008

Other Systematics § Optical effects – Flux-migration between grating-orders § § § JPL-CIT Response of interference filter Geometrical shadowing by grooves Woods anomalies Paris 2008

Part III – Observations of HD 189733 b JPL-CIT Paris 2008

State-Variables HD 189733 b angle position defocus temperature Paris 2008

Iterative Multivariate Fits Noise Light curve Design Matrix Model vector JPL-CIT

Raw periodogram Data Modeling-III Post-fit residuals JPL-CIT Paris 2008

Lightcurves Broadband 1. 5 To 2. 5 um K band with Common mode Noise removed Final K band Lightcurve JPL-CIT Paris 2008

HD 189733 (Basic Data) § HD 189733 (K 1 -K 2 V) – T ~ 5000 K – 19. 3 pc – > 0. 6 Gyr – Metallicity -0. 03 +/- 0. 04 § HD 189733 b (Bouchy et al. 2005) – 1. 144 MJ, 1. 138 RJ – Circular 0. 03 AU orbit (2. 22 d) § Secondary eclipse observations – Barnes et al. 2007 (d. C ~ 4 x 10 -4) JPL-CIT Paris 2008 J. Schneider, Ex. Enc.

Spectral Modeling § Retrieval using RT models (Goody & Yung 1989) § Disk-averaged radiative transfer models developed originally for Earthshine, Mars § (Tinetti et al. 2006, 2007) § P-T profiles (Barman et al. 2008, Burrows et al. 2008) § Photochemistry (Yung, Liang) § Layer-by-layer (log P between -6 and 0) – Input T-P profiles – Chemical profiles (simple constant VMR) – Opacities (T, ρ); Cloudless. JPL-CIT Paris 2008

HD 189733 b NIR Contrast Spectrum JPL-CIT Paris 2008

Contrast Spectrum Components 14/08/2008

Comparison with radiation-hydrodynamics models Showman et al. 2008 Planet brightest away from anti-stellar point Knutson et al. 2007 Paris 2008

Retrieval Results § Dayside emission (subsolar) – Water (0. 1 -1 10 -4) – Carbon monoxide (thermochemically very stable at these P, Ts; CO=CH 4 T=1100 K at 1 bar) § Also inferred from IRAC photometry (Charbonneau et al. 2008) § 10 -4 – Carbon dioxide (trace concentration 10 -6) § CO+H 2 O <=> CO 2+H 2 (thermochemical in a CO field; Lodders & Fegley 2002) § CO+OH <=> CO 2+H (photochemical pathway) – Methane upper limit (10 -7) – Significant residuals at the blue end of the spectrum JPL-CIT Paris 2008

Abundances § C/O is high and not well constrained (cloudless model) – 0. 5 to 10 § Solar 0. 48 (Anders & Grevesse 1989) Favor lower values because high C/O implies disappearing water in CO field – Terminator (Swain, Vasisht, Tinetti 2008) § Lower pressure depths § Methane abundance is higher (CO < CH 4) § Water 5. 10 -4 JPL-CIT Paris 2008

In Summary Little evidence for … Hot Jovians not as “hot” as … good hot Curry !. JPL-CIT Paris 2008

Chemistry § Hot less dense atmospheres are more likely to show abundant CO (and CO 2 at lower T), while cooler, denser ones show more abundant methane. § At 1 bar the CO=CH 4 boundary is at T = 1125 K. § C/O atomic ratio is 0. 48 (solar) 14/08/2008 Exeter Exoplanet Workshop

JPL-CIT Paris 2008

JPL-CIT Paris 2008 Pont et al. 2008

JPL-CIT Paris 2008 F. Pont et al. 2008

Carbon & Oxygen Chemistry § Major carbon bearing gases in a solar composition gas of given metallicity are generally CH 4, CO and/or CO 2 depending on T and P. 14/08/2008 Exeter Exoplanet Workshop