Beyond Hot Jupiters: Spitzer Observations of the Smallest

Transiting Planets

Heather Knutson, UC Berkeley

Image Credit: ESO/L. Calçada

Good News: Radial Velocity Surveys Indicate Small Planets are Common

Howard  et  al.  (2010)  

Kepler Planet Candidates Follow the Same Trend

Kepler  planet  candidates  Howard  et  al.  (2011)  

Bad News: Most RV planets aren’t transiting, and most Kepler planets are too faint for detailed study.

4 6 8 10 12 14 16V Band Magnitude

0

50

100

150

200

250

Num

ber o

f Sta

rs

Radial Velocity Planets Kepler Planet Candidates

So what can we do with each group?

Project #1: Kepler Planet Candidate Verification

•  Change in Spitzer transit depth wrt Kepler measurement would indicate an astrophysical false positive

•  Valuable for physically-associated triples for which Kepler will not detect shift of photocentroid

O’Donovan, Charbonneau, et al. 2006

Transit  depth  is  color  dependent  

Good  planet  candidate  but…  

800-hour program (PI Charbonneau), work led by Jean-Michel Desert & Francois Fressin  

Warm Spitzer Verification of Kepler Planet Candidates

Jupiter-sized planet in an 18-day orbit around a rapidly rotating F star

Neptune-sized planet in 7-day orbit, Kepler centroids corrupted

Super-earth in an 80-day orbit around a low-mass star

800-hour program (PI Charbonneau), work led by Jean-Michel Desert & Francois Fressin  

600 hours on transits (34 planets), 200 hours on secondary eclipses (4 planets)

Project #2: Search for Transits of Low-Mass RV Planets Orbiting Bright Stars

100 hours with warm Spitzer, PI Gillon

Need continuous, high-precision space-based monitoring for transit search.  

GJ 3634b: 8 Mearth planet in a 2.6 day orbit around a M2.5 dwarf, 6.0 K mag

Bonfils et al. (2011)  

7% a priori transit probability, Spitzer observations reduce this to 0.9% (1σ)  

55 Cancri e: A Super-Earth Transiting a Naked-Eye Star

0 2 4 6 8 10 12 14Time [HJD 2,455,600]

0.997

0.998

0.999

1.000

1.001

1.002

Rel

ativ

e flu

x

Transits of e

Transit of b

0.4 0.2 0.0 0.2 0.4Phase

0.9996

0.9998

1.0000

1.0002

1.0004

Rel

ativ

e flu

x

Winn et al. (2011): Two weeks of monitoring in visible light with MOST

0.02% transit with 0.74-day period, corresponding to 1.6 Rearth

Density consistent with rock/iron composition.

Sun-like star, V mag of 6.0, five known planets

Spitzer Confirmation of 55 Cancri e Transit

Can also do a blind search for additional transits in systems with one known transiting planet (e.g., GJ 436 and GJ 1214).

BJD - 2450000

Flu

x

New result by Demory et al. (2011, astro-ph/1105.0415)

super-M

ercury

pure-Fe

no iron

Earth-like

0.01% H-He

0.1% H-He

50% H2O

U

N

10% H2O

C-7b

K-10b

This study, 55Cnc-e

55Cnc-e Winn et al ‘1110000

8000

6000

20000

Rad

ius

(km

)

Mass (MEarth)1 10

Problem: MOST and Spitzer radius estimates disagree (1.6 vs. 2.3 Rearth).

So you have a small planet transiting a bright, nearby star…

Now what?? Image Credit: ESO/L. Calçada

Project #3: Atmosphere Studies for Small, Cool Planets

HD 209458b Mass: 0.66 MJup Radius: 1.32 RJup Teqil=1360 K

HD 189733b Mass: 1.15 MJup Radius: 1.15 RJup Tequil=1130 K

GJ 1214b Mass: 0.02 MJup Radius: 0.24 RJup Tequil=560 K

GJ 436b Mass: 0.07 MJup Radius: 0.44 RJup Tequil=700 K

Planets drawn to scale.

Smaller radius and lower effective temperature both reduce size of signal for transmission spectroscopy and secondary eclipse observations.

Ideal targets are small planets transiting bright, nearby M dwarfs

GJ 1214b & GJ 436b: A Distinctly Different Breed of (Exo)Planets

•  Metal-rich (large solid rock/ice cores), hydrogen atmospheres(?) •  Methane dominates over CO in equilibrium chemistry. •  Cool enough for clouds? 500-800 K temperatures. •  M dwarfs have high UV fluxes, is photochemistry important?

GJ 1214 to scale.

A Dayside Emission

Spectrum for GJ 436b

Stevenson et al. (2010)

CH4 / CO Ratio:

Predicted: 100  

Measured:

Constraints on Atmospheric Variability

for GJ 436b Knutson et al. (2011)

Time From Eclipse Center [d]

Rela

tive

Flux

Dayside flux varies by

Transmission Spectroscopy of GJ 436b

Spitzer transit observations can distinguish between methane-rich and methane-poor models.

Models from Madhusudhan & Seager (2011)

Eight Spitzer Transit

Observations of GJ 436b

0.992

0.994

0.996

0.998

1.000

Rel

ativ

e Fl

ux

-0.08-0.06-0.04-0.02 0.00 0.02 0.04 0.06-0.002

-0.001

0.000

0.001

0.000

Rel

ativ

e Fl

ux

UT 2009 Jan 9UT 2009 Jan 28 3.6 µm

-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06Time from Measured Transit Center [d]

UT 2009 Jan 17UT 2009 Jan 30 4.5 µm

-0.08-0.06-0.04-0.02 0.00 0.02 0.04 0.06

UT 2009 Jan 25UT 2009 Feb 2 8.0 µm

Problem: Transit depths vary from one epoch to the next!

Knutson et al. (2011), Beaulieu et al. (2011) find the same result

0.68

0.70

0.72

0.74

Tran

sit D

epth

(%)

270 280

0.02

0.04

0.06

0.08

Seco

ndar

y Ec

lipse

Dep

th (%

)

620 630 640 650 660 830 840 850 860 870 880Eclipse Center [BJD - 2454000]

0.68

0.70

0.72

0.74

Tran

sit D

epth

(%)

270 280

0.02

0.04

0.06

0.08

Seco

ndar

y Ec

lipse

Dep

th (%

)

620 630 640 650 660 830 840 850 860 870 880Eclipse Center [BJD - 2454000]

3.6 μm 4.5 μm 8.0 μm

Abrupt transition from shallower to deeper transits… perhaps indicative of increased stellar activity?

Stellar Activity is Bad for Transits

Planet Spot

Scenario 1: Occulted Spot

0.98

0.99

1.00

Relat

ive F

lux

Epoch 5

−1.0 −0.5 0.0 0.5 1.0Time since midtransit [hr]

−202

O−C

[ppt

]

Epoch 260

−1.0 −0.5 0.0 0.5 1.0Time since midtransit [hr]

Carter et al. (2011)

Planet Spot

Scenario 2: Non-Occulted Spot

Time

Rela

tive

Flux

Spot-free

Spotted

Evidence for Stellar Activity Knutson et al. (2011)

0.995

1.000

1.005

Rel

ativ

e Fl

ux, (

b+y)

/2

800 850 900 950 1000Julian Date -2454000

0.55

0.60

0.65

0.70

S HK

40-50 day rotation period w/small amplitude, consistent with relatively old/quiet early M spectral type

0.992

0.994

0.996

0.998

1.000

Rel

ativ

e Fl

ux

-0.0010-0.0005

0.0000

0.0005

0.0000

Rel

ativ

e Fl

ux

0.992

0.994

0.996

0.998

1.000

Rel

ativ

e Fl

ux

-0.10 -0.05 0.00 0.05 0.00-0.0010-0.0005

0.0000

0.0005

0.0000

Rel

ativ

e Fl

ux

-0.10 -0.05 0.00 0.05 0.00 -0.10 -0.05 0.00 0.05 0.00 -0.10 -0.05 0.00 0.05 0.00Time from Predicted Transit Center [d]

UT 2007 Jun 29, 8.0 µm UT 2008 Jul 14, 8.0 µm UT 2009 Jan 9, 3.6 µm UT 2009 Jan 17, 4.5 µm

UT 2009 Jan 25, 8.0 µm UT 2009 Jan 28, 3.6 µm UT 2009 Jan 30, 4.5 µm UT 2009 Feb 2, 8.0 µm

200 300 400 500 600 700 800 900Transit Center [BJD-2454000]

-3

-2

-1

0

1

2

3

O-C

(min

)

0.992

0.994

0.996

0.998

1.000

Rel

ativ

e Fl

ux

-0.0010-0.0005

0.0000

0.0005

0.0000

Rel

ativ

e Fl

ux

0.992

0.994

0.996

0.998

1.000

Rel

ativ

e Fl

ux

-0.10 -0.05 0.00 0.05 0.00-0.0010-0.0005

0.0000

0.0005

0.0000

Rel

ativ

e Fl

ux

-0.10 -0.05 0.00 0.05 0.00 -0.10 -0.05 0.00 0.05 0.00 -0.10 -0.05 0.00 0.05 0.00Time from Predicted Transit Center [d]

UT 2007 Jun 29, 8.0 µm UT 2008 Jul 14, 8.0 µm UT 2009 Jan 9, 3.6 µm UT 2009 Jan 17, 4.5 µm

UT 2009 Jan 25, 8.0 µm UT 2009 Jan 28, 3.6 µm UT 2009 Jan 30, 4.5 µm UT 2009 Feb 2, 8.0 µm

Occultation of bright region on star

Color-dependent timing variations

Visible IR

Conclusion: transit depth variations probably due to occulted spots, star likely viewed near pole-on

Transmission Spectroscopy of GJ 436b

Data from Pont et al. (2008), Alonso et al. (2008), Cáceres et al. (2008), Ballard et al. (2010), Knutson et al. (2011)

Will Stellar Activity Affect All Transiting Planets Orbiting M Stars?

GJ 1214 also has a slow, 52 day rotation period, Teff=3000 K (Berta et al. 2011; 3500 K for GJ 436)

0.98

0.99

1.00

Relat

ive F

lux

Epoch 5

−1.0 −0.5 0.0 0.5 1.0Time since midtransit [hr]

−202

O−C

[ppt

]

Epoch 260

−1.0 −0.5 0.0 0.5 1.0Time since midtransit [hr]

R’ band transits of GJ

1214b, 6.5 m Magellan Clay telescope (Carter et al. 2011)

Spitzer 4.5 μm transit (Desert et al. 2011)

Infrared may be ok for this star.

Warm Spitzer Constraints on the Composition of GJ 1214b’s Atmosphere

Desert et al. (2011)

Based on data from Bean et al. (2010), Croll et al. (2011), Desert et al. (2011)

See poster by Jean-Michel Desert on Spitzer results! Also talks by Eliza Kempton, Bryce Croll, & Jacob Bean, posters by Ian Crossfield.

CFHT:

Secondary Eclipse Constraints on the Composition of GJ 1214b’s Atmosphere

Six secondary eclipse obs. total at 3.6 and 4.5 μm

Madhusudhan et al. (2010)

Methane-rich (equil. chem.) Methane-poor

Non-zero eccentricity?

500, 600, 700 K Blackbody

Conclusions

•  Multi-wavelength photometry with warm Spitzer can be used to confirm planet candidates and search for new transiting systems.

•  For the present, atmosphere studies of super-Earths are focused on planets orbiting bright, nearby M dwarfs.

•  Where is the methane??

•  Devil is in the details– activity on M stars is poorly understood, and problematic for characterization.

Two Commonly Used Methods for Finding & Characterizing Exoplanets

Doppler Method

Determine Planet Mass

Transit Method

Determine Planet Diameter

Calculate Planet Density and Infer Composition:

Gas giant (Jupiter), Ice giant (Neptune), or Rocky planet (Earth)

Transit

Secondary Eclipse See thermal radiation and reflected light from planet disappear and reappear

See radiation from star transmitted through the planet’s atmosphere

Transiting Planets as a Tool for Studying Exoplanetary Atmospheres  

Determining the Composition of GJ 1214b’s Atmosphere with Transmission Spectroscopy  

Compositions: solar 30x solar 50x solar

H20 (steam) 50/50 H20, CO2 CO2 (Venus)

Miller-Ricci & Fortney (2010)

Atmosphere

Star

Planet

Scale Height

€

H = kTgµ Large scale height Small scale height

Spitzer Confirmation of 55 Cancri

e Transit

Demory et al. (2011, astro-ph/1105.0415)