Detecting Light From Distant Worlds: What Can We Learn From Secondary Eclipses and Phase Curves?

Heather Knutson Division of Geological and Planetary Sciences California Institute of Technology

Three Questions for Exoplanet Atmospheres Atmospheric Composition  

What do the compositions of exoplanets tell us about their formation histories?

Cloud Physics  

Which exoplanets have clouds, and why?

What’s Next?  

Pushing down towards more Earth-like planets.

What can secondary eclipse and phase curve observations tell us about each of this areas?

We know that hot Jupiters could not have formed at their present locations, but must have migrated inward from beyond the ice line. These planets are large and hot, ideal for detailed characterization. What does their atmospheric composition tell us about the process of planet formation?

Image credit: ESA/C. Carreau

Composition of Planet-Forming Material Varies With Disk Radius Silicates 1500 K Water 130 K CO2 50 K

Condensation Temperatures

Could we detect a super-solar C/O ratio in a hot Jupiter atmosphere? Oberg  et  al.  (2011)  

Secondary Eclipse Interpretation 101 Planet Atmosphere Models

PHOENIX, 3600 K

Wavelength (μm)

Flux (10-9 erg s-1 cm-2 cm)

Flux (erg s-1 cm-2 cm)

Stellar Atmosphere Model

Wavelength (μm)

Fstar + Fplanet

Fstar

depth(%) = €

Fplanet Fstar + Fplanet

≈

Fplanet Fstar

A High C/O Ratio for WASP-12b? Must correct for contamination from nearby M+M binary companions (Crossfield et al. 2012, Bechter et al. 2013)

C/O=0.5 (Solar) Model Data

CFHT J, H, Ks + WFC3 spectrum (new)

C/O=1.0 Model Spitzer 3.6, 4.5, 5.8, 8.0 µm Campo et al. 2011, Croll et al. 2011, Madhusudhan et al. (2011, 2013), Line et al. (2014)

1

2

4

6

8

10

20

Another Way to Constrain C/O Ratios: Doppler Shifts with High Resolution IR Spectroscopy 1

2

Stellar spectrum also contains a component of planetary emission with absorption lines that undergo large radial velocity shifts.

Observer

Detections of CO, water for: HD 189733b: Birkby et al. (2013), de Kok et al. (2013) 51 Peg b: Brogi et al. (2013) τ Boötis b: Brogi et al. (2012), Lockwood et al. (2014)

2. Blueshifted Line

1. Redshifted Line

Detection of RV-Shifted Water Absorption Bands for HD 189733b VLT/CRIRES, Birkby et al. (2013)

R = 100,000 K band spectrum See talks by Birkby, Brogi for more details.

Water absorption detected at 5σ Strength: enables a unique identification of molecular absorbers. Weakness: technique is currently limited to CO, H2O for a subset of the most favorable hot Jupiters

Next Question: Are There Empirical Trends In Atmospheric Composition vs Planet Mass?

Smaller fractional core size, higher H/He abundance C/H ~ 3x solar

Core Accretion Model for Planet Formation

Atmospheric “Metallicity”

Larger fractional core size, lower H/ He abundance C/H ~ 30-40x solar Lodders (2003)

Prediction for GJ 436b (Neptune mass, ~800 K): A Hydrogen-Dominated Atmosphere With Methane and Water

Figure courtesy J. Fortney

GJ 436b: Comparison to Models Equilibrum model (CH4/CO = 100) Best-fit model (CH4/CO < 0.001) VerCcal  mixing   dredges  up  CO-­‐rich   gas  from  interior  

UV  flux  destroys   CH4  molecules  

Can’t match observed abundances with non-equilibrium chemistry models either (Line et al. 2011)

Madhusudhan  &  Seager   (2011)  

Could GJ 436b Have a Hydrogen-Poor Atmosphere? Can explain data if atmosphere has a metallicity of 300x-2000x solar. Neptune’s atmosphere is enriched in carbon by 30-40x relative to the Sun.

Is GJ 436b typical of Neptune-mass exoplanets?

Spitzer secondary eclipses (Stevenson et al. 2010)

300x solar model

Moses et al. (2013)  

Three Questions for Exoplanet Atmospheres Atmospheric Composition  

What do the compositions of exoplanets tell us about their formation histories?

Cloud Physics  

Which exoplanets have clouds, and why?

What’s Next?  

Pushing down towards more Earth-like planets.

What can secondary eclipse and phase curve observations tell us about each of this areas?

Hot Jupiters With High Albedos Likely Have Reflective Clouds The presence or absence of a high, reflective cloud deck is the single biggest contributor to variations in albedo.

Measure visible (reflected + thermal) emission with Kepler.

Reflective clouds? Measure infrared (thermal) emission using Spitzer.

Observe the decrease in light as the planet disappears behind the star and then reappears. Visible emission – thermal contribution

= geometric albedo

Kepler-13Ab: Measuring Emission from a Highly Irradiated (~3000 K) Hot Jupiter

Clouds might be high-temperature condensates (corundum, iron oxide, perovskite?)

2500

Occultation depth [ppm]

Excess flux in Kepler band consistent with geometric albedo of 0.33 ± 0.06

Other hot planets have hazes as well (Stevenson et al. 2013, Sing et al. 2013)

200

2000 100

1500 0

0.5 0.6 0.7 0.8

1000

500

=0.2, Pn=0.1 =0.0, Pn=0.1 e e

0

0

Data from Shporer et al. (2014), models from Burrows et al. 1

2

3 [µm]

4

5

The Big Picture: Statistics of Hot Jupiter Albedos Albedo corrected for maximum possible thermal emission. Heng & Demory (2013). Also see Christiansen et al. (2010), Désert et al. (2011ab), Fortney et al. (2011), Kipping & Spiegel (2011).

Kep-13Ab

log1

What Color Is a Hot Jupiter? The Case of HD 189733b

0

−1

Ag Albedo Geometric

1.0 0.8

Quiz: What color is this? Cobalt blue? Azure blue?

Observations constrain composition and location of cloud layer on HD 189733b.

0.6 0.4 0.2 0.0 −0.2 300

350

400

450

500

550

λ (nm)

λ (nm)

Measure wavelength-dependence of visible light secondary eclipse depth with HST/STIS spectroscopy (Evans et al. 2013)

600

Hot Jupiters in 2D Hot Jupiters should be tidally locked, may have large thermal and/or chemical gradients between the two hemispheres. If clouds exist, they may be spatially inhomogeneous (similar to patchy clouds on brown dwarfs?)

Image credit: ESA/C. Carreau

Kepler-7b: Evidence for Spatially Inhomogeneous Clouds from Kepler Q1-14 Kepler phase curve from Demory et al. (2013) Spitzer obs. constrain thermal emission to be negligible.

100

50

Planet

Star

F / F [ppm] Planet-Star Flux Ratio (ppm)

150

0

-50

-100

-0.2

Observer’s View of Planet

0.0

Reflective western 0.2 Orbital phase hemisphere Orbital Phase

NASA press release artist’s Lambertian sphere impression of Kepler-7b. (uniform clouds) Best fit 3-band model 0.4

0.6

Secondary Eclipse Mapping: A New Frontier in Studies of Exoplanetary Climate If the planet is uniform in brightness, the secondary eclipse looks like this:

See talk by Julian de Wit for most Egress Ingress recent eclipse mapping observations of eccentric hot Jupiter HAT-P-2b Non-uniform brightness leads to residuals during ingress and egress after best-fit uniform eclipse is subtracted:

Majeau, Agol, & Cowan (2012)

HD 189733b 8 µm secondary eclipse (Agol et al. 2010, de Wit et al. 2012, Majeau, Agol, & Cowan 2012). Also see Williams et al. (2006).

Three Questions for Exoplanet Atmospheres Atmospheric Composition  

What do the compositions of exoplanets tell us about their formation histories?

Cloud Physics  

Which exoplanets have clouds, and why?

What’s Next?  

Pushing down towards more Earth-like planets.

What can secondary eclipse and phase curve observations tell us about each of this areas?

Pushing Down to Super Earths: 55 Cnc e and GJ 1214b 1.0002

Depth of 131±28 ppm

1.0001

4.5 µm   Depth of 70±35 ppm

1.0000 0.9999 0.9998

4.5 µm   0.35

0.40

3.6 µm   0.45

0.50

0.55

0.60

0.55

0.60

0.0002 0.0001 0.0000 -0.0001 -0.0002

Four Spitzer secondary eclipse observations of 55 0.40 0.45 0.50 Cnc 0.35 e (Demory et al. 2012) Orbital phase Tbright = 2360 ± 300 K

But, GJ 1214b was not detected in secondary eclipse with ~dozen Spitzer observations (Gillon et al. 2014)

Low mass primaries can help boost signal to noise, but relatively cool planets are still very hard to detect at short (