Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26Slide 27Slide 28Slide 29Slide 30Slide 31Slide 32Slide 331Unraveling the Mystery of GJ1214bwith NIRSPECCrossfield, Barman, & Hansen, 2011, ApJ 736:132Keck Science MeetingCIT, Pasadena, 23 Sep 20112345Earth67Earth-like composition8Earth-like0.1% H/He1% H/He9Earth-like50% H20100% H2010Earth-like50% H20100% H200.1% H/He1% H/He11Earth-like50% H20100% H200.1% H/He1% H/He12Transmission spectroscopy: probes atmospheric composition via RP(λ)RP(λ)FluxTimeλ1λ2RP2.00 µm2.41 µm●Observations:●Two half-nights with NIRSPEC●One night has insufficient out-of-transit baseline●One good night covers ~half of the K bandWe get a high S/N spectrum of this K=8.8 M dwarf:Crossfield, Barman, Hansen 2011A full set of observations:Spectrophotometric variations are dominated by a trend common to all wavelength channels Divide out this trend from every wavelength!WAVELENGTH TIMERELATIVE FLUXTIMECrossfield, Barman, Hansen 2011Removing common-mode variations leaves differential transit signal:Observed data: no residual transit spectrum visible by eyeWAVELENGTH TIMERELATIVE FLUXCrossfield, Barman, Hansen 2011Removing common-mode variations leaves differential transit signal:Simulated observation: transit signal 10x stronger than expectedWAVELENGTH TIMERELATIVE FLUXCrossfield, Barman, Hansen 2011Removing common-mode variations leaves differential transit signal:Observed data: no residual transit spectrum visible by eyeWAVELENGTH TIMERELATIVE FLUXCrossfield, Barman, Hansen 2011Initial transmission spectrum of GJ1214b:Transit depth – constantTelluric contaminationWe cross-correlate this spectrum with models to confirm or rule out atmospheric compositionsCrossfield, Barman, Hansen 2011WAVELENGTH [microns]Model cross-correlation can detect the ensemble of lines in a spectrum:Crossfield, Barman, Hansen 2011MODEL RULED OUT:Atmospheric signal expected, but not seenCrossfield, Barman, Hansen 2011MODEL RULED OUT:Atmospheric signal expected, but not seenCrossfield, Barman, Hansen 2011MODEL RULED OUT:Atmospheric signal expected, but not seenCrossfield, Barman, Hansen 2011MODEL UNCONSTRAINED:No significant signal expected; none seen.Crossfield, Barman, Hansen 2011Our Results●We rule out hydrogen-dominated atmospheres in or near chemical equilibrium:●Solar, 10x solar, 30x solar abundances with mild-to-no methane depletion●We cannot constrain atmospheres with flatter spectra:●H-dominated (Low carbon, substantial methane depletion)●Hazes/clouds●Low scale height (high mean molecular weight: e.g., H2O)Crossfield, Barman, Hansen 2011Other results also agree (mostly) on a flat, featureless spectrum:●Bean+2011: 0.8-1.0 um spectrum; flat●Desert+2011: Spitzer/IRAC CH1+2; flat●Croll+2011: NIR photometry, RK > RJ (>4σ)●Bean+(1109.0582): 0.6-1.0 um + JHK; flat.●Berta+(in prep): 1.1-1.65 um spectrum; flat.Conclusions●GJ 1214b has a flat transmission spectrum, meaning the planet either:●--Is covered in opaque clouds, OR●--Has a high mean molecular weight atmosphere (e.g., H2O)●If a 'water' world, GJ 1214b likely formed beyond the snow line and migrated inward without accreting substantial H2/He●NIRSPEC-like instruments can constrain exoatmospheres, and multi-object spectrographs like MOSFIRE (see Bean+2011) are poised to do even better.Backup:●Simulations and time-offset extractions to estimate uncertainties inherent in data●Estimation of non-detection confidenceTransmission spectra can constrain atmospheric composition:GJ 1214b(models by T. Barman)Several techniques can characterize exoplanet atmospheres: CADMethod: Signal scales as:(A) Transit (Rp / Rs)^2(B) Transmission (Tp Rp / Mp) (Rp / Rs)^2(C) Eclipse Tp / Ts (Rp / Rs)^2(D) Phase curve ΔTp/Ts (Rp / Rs)^2BFit a transit model to each wavelength channel:Observed fluxIntrinsic stellar fluxTelluric extinction{{Scaled transit light curvePossible low-level systematicsInitial transmission spectrum of GJ1214b:Transit depth – constantBUT: Tilt should not affect spectral features on the narrowest scalesTelluric