June. 4, 2015 @ NAOJ Probing the Dynamical History of Exoplanets: Spectroscopic Observa ons of Transi ng Systems
Teruyuki Hirano ©NASA Outline
1. Dynamical History of Close-in Planets 2. Measurements of the Stellar Obliquity 3. Observa on of the Rossiter-McLaughlin (RM) Effect 4. Interpreta on of the Past RM Result 5. Towards Smaller Planets 6. Summary Distribu on of Exoplanets I
100
10 B ] J [M 1 A
0.1 anet mass l p Snow Line 0.01 C
0.001 0.001 0.01 0.1 1 10 100 semi-major axis [AU] Distribu on of Exoplanets II
1
0.01 0.01 0.1 1 10 100 semi-major axis [AU] Forma on of Close-in Giant Planets
The Standard Scenario
→ Close-in giants have originally formed at a few AU away from their host stars, then migrated inward by some migra on processes.
ØWhat process can make planets migrate? ØIs there any possibility to observa onally dis nguish among migra on scenarios ?
5 Migra on Processes
1. Interac ons between proto-planets and proto-plantary disk → Planets had gradually migrated inward due to gravita onal interac ons with proto-planetary disk (e.g. Lin et al. 1996, Lubow & Ida 2010)
2. Planet-planet sca ering (or Kozai cycles) + dal interac ons → When mul ple planets form or there exists a companion star, they interact with each other and some mes cause orbital crossing, which can pump up eccentrici es. Subsequent dal interac ons with host stars may cause orbital circulariza on. (e.g. Cha erjee et al. 2008, Wu et al. 2007, Fabrycky & Tremaine 2007)
6 © Weidenschilling and Marzari 1996 Migra on Processes
1. Interac ons between proto-planets and proto-plantary disk → Planets had gradually migrated inward due to gravita onal interac ons with proto-planetary disk (e.g. Lin et al. 1996, Lubow & Ida 2010) → small e and stellar obliquity
2. Planet-planet sca ering (or Kozai cycles) + dal interac ons → When mul ple planets form or there exists a companion star, they interact with each other and some mes cause orbital crossing, which can pump up eccentrici es. Subsequent dal interac ons with host stars may cause orbital circulariza on. (e.g. Cha erjee et al. 2008, Wu et al. 2007, Fabrycky & Tremaine 2007) → large e and stellar obliquity
7 © Weidenschilling and Marzari 1996 Measurements of the Stellar Obliquity Measurement of the stellar obliquity is a unique method to probe the dynamical history of exoplanets! a. The Rossiter-McLaughlin effect -> High dispersion spectroscopy during planetary transits (e.g., Queloz 2002, Winn et al. 2005) b. Spot-crossing Method -> Modeling of anomaly in transit light-curves (e.g., Sanchis-Ojeda, et al. 2011) c. VsinI Method -> Spectroscopic determina on of the rota on velocity + photometric determina on of the rota on period (e.g., Hirano et al. 2012) d. Asteroseismology -> High cadence photometry of pulsa ng stars (e.g., Chaplin et al. 2013) e. Gravity-darkening Method -> Modeling of asymmetry in transit light- curves → Kamiaka’s talk The Rossiter-McLaughlin (RM) Effect λ = 0° λ = 50°
approaching receding
planet’s orbit
during transit ←ver cal axis: radial veliocity anomaly during a transit RV anomaly RV anomaly
time [h] time [h]
Through the RM effect, we can es mate the angle between the stellar spin axis and the planetary orbital axis, which is closely related to the planetary migra on. 9 Results by Subaru/HDS
Subaru RV 40
20 ] -1 0 v [m s !
-20
Subaru (this work) -40 Keck (P08)
30 ]
-1 20 10 0 -10 -20 Residual [m s -30 -0.04 -0.02 0 0.02 0.04 Orbital Phase ↑ HAT-P-16: ↑ HAT-P-7: Teff = 6158 ± 80 K Teff = 6350 ± 80 K λ = -8.4° ± 7.0° λ = -132.6° +10.5° -16.3° VsinIs = 3.9 ± 0.3 km/s VsinIs = 2.3 ± 0.6 km/s T.H. + in prep. Narita et al. 2009 RM Measurement for a Mul ple Transi ng System
Name of Candidate Orbital Period Planet Radius Kepler-89b 3.7 days 0.13 RJ Kepler-89c 10.4 days 0.31 RJ
Kepler-89d 22.3 days 0.83 RJ Kepler-89e 54.3 days 0.49 RJ
40
20 Kepler-89 (KOI-94): ] -1 0 v [m s ∆ Teff = 6116 ± 30 K -20 λ = -6° -11° +13° VsinIs = 8.01 ± 0.73 km/s -40 Kepler-89d
30 T.H.+ 2012b, ApJ Letters ]
-1 20 10 0 -10 This result was later confirmed -20 Residual [m s -30 by Albrecht et al. (2013) -0.01 -0.005 0 0.005 0.01 Orbital Phase Mul ple transi ng systems are generally aligned!
15 2.0•105 ← RM1.5•105 measurements for Kepler-25 10 1.0•105
(multi-planet5.0•104 system):
5 0 −20 0 20 ) − 1 0 Teff = 6190 ± 80 K 12 12 −5 λ = 7°±8° RV (m s −10 VsinIs = 8.7 ± 1.3 km/s )
10 − 1 10 −15
0.23 0.26 10 Obliquity measurements for other multi-
8 ) 8
− 1 5 planetv sin i (km s systems generally show good spin- 0 orbit alignment (Sanchis-Ojeda et al. 2012, −5 6 6 −10 Huber et al. 2013). O − C (m s −15
0 2.0•104 4.0•104 6.0•104 8.0•104 1.0•105 1.2•105 1.4•105
−6 −4 −2 0 2 The only− exception20 0 is 20Kepler-56 (multiple, Time (hr) but misaligned;λ (deg)Huber et al. 2013). Albrecht et al. 2013, ApJ
A companion is present outside of the planetary system, which might be responsible for the misalignment in Kepler-56. Summary of RM Measurements
180
150 WASP-8 b 120 HAT-P-11 b |λ| 90 60 HD 80606 b proj. obliquity [deg] 30 HD 17156 b 0
5000 6000 7000 8000
80 ] -1 60 [km s
star 40
v sin i 20
5000 6000 7000 8000 Teff [K]
Albrecht et al. 2012 Correla on between Stellar Obliquity and Rela ve Tidal Timescale Albrecht et al. 2012
180 |λ| HAT−P−7 b 150 WASP−8 b 120 WASP−33 b HAT−P−11 b 90 HAT−P−32 b
WASP−12 b 60 HD 80606 b XO−3 b
proj. obliquity [deg] 30 KOI−13 b
HD 17156 b 0
100 102 104 106 108 relative tidal−dissipation timescales τCE τRA [yr] ↑ relative tidal damping timescale of stellar obliquity (based on Zahn 1977)
Assuming that planet-planet scatterings are the dominant channel for the formation of HJ
subsequent tidal interactions “realign” the stellar spin and the planet orbital axes. Quiescent v.s. Chao c migra ons
• Quiescent migra ons (i.e., disk-planet interac ons): ü The presence of close-in planets in mean mo on resonance ü Mul ple transi ng systems are generally aligned. ü Misaligned systems are formed by stellar internal gravity waves independently of the planets? (Rogers et al. 2012)
• Chao c migra ons: ü The correla on between the stellar obliquity and rela ve dal damping mescale ü Existence of highly eccentric hot Jupiters (e.g., HD 80606b) To Se le This Issue… Previous Targets of RM Measurements host star Single Transi ng System Mul ple Transi ng System cool (Teff < 6250 K) many a few hot (Teff > 6250 K) many No observa on
180
ü One solution to partly settle the 150 WASP-8 b debate is to observe the RM 120 HAT-P-11 b effect for “hot multiple” systems. 90 60 HD 80606 b ü So far, no hot multiple systems proj. obliquity [deg] 30 HD 17156 b 0 have been investigated for 5000 6000 7000 8000 80
obliquity measurements. ] -1 60 [km s ü There are a few hot multiple star 40
systems detected by Kepler (e.g., v sin i 20 KOI-89). 5000 6000 7000 8000 T [K] Worth a try! eff Towards Smaller Planets Limita on of the RM Measurements a. The RM measurements have been conducted for systems with hot Jupiters (Neptunes). b. When the size of the transi ng planet becomes small (~ Earth- sized planet), the detec on of the RM effect becomes very challenging.
→ ΔvRM 〜 - f ・ V sin Is・√(1-b^2) c. In order to confirm or refute possible hypothesis on the planetary migra on (and relevant dal interac on), we need to expand the parameter space for which the spin-orbit angle is inves gated. Es mate of Stellar Rota on Periods
KOI 988 1.015 ü Some of the light-curves 1.010 1.005 taken by Kepler spacecra 1.000 0.995
show periodic flux varia ons Relative flux 0.990 due to rota on of starspots. 0.985 210 220 230 240 BJD−2454900 [days] 2.0•104 ü From such light-curves, we 1.5•104 1.0•104 can es mate the rota onal ower P periods of planet hos ng 5.0•103 0 stars. 0 5 10 15 20 Period of rotation [days] Es ma on of Stellar Inclina ons by Starspots 1
2
534/%-).+-43)46+$ $-%334&)$5+/)46+$ λ
I$ 7 !"#$%&'%&($)*+&%,-+./0 I.
ü The rotational period Ps of the host star is estimated by Kepler photometry. ü The projected rotational velocity (V sin Is) and stellar radius are estimated by spectroscopy.
We can estimate stellar inclination Is. a. Since the orbital inclina on of a transi ng exoplanet is close to 90°, a small value of Is implies a spin-orbit misalignment.
b. This method can be applied regardless of the size of the planet, enabling us to discuss spin-orbit alignment/misalignment for smaller exoplanets. Results 1
We observed about 10 KOI systems and estimated VsinIs along with stellar radii.
20 x axis -> rotational velocities o Is=90 estimated by Kepler multiple photometry 15 269 262 ] -1 y axis -> projected rotational o 974 Is=45 velocities estimated [km s 10 s by spectroscopy 257 o
V sin I Is=30 5 280 Solid line indicates 367 261 Veq = V sin Is, 0 0 5 10 15 20 suggesting spin-orbit -1 Veq [km s ] alignments. T.H.+ 2012, ApJ, 756, 66 KOI-261 may have a possible spin-orbit misalignment along the line-of-sight. Results 2 We observed additional 25 systems.
20 o 10 I =90 (a) single (b) multi s o Is=90
15 ] 8 ] -1 1 - I =45o 6 s [km s
[km s o
10 I =45 s s s 4 o 2002 Is=30
1890 V sin I V sin I o 2261 323 Is=30 5 988 2026 2 285 304
0 0 0 5 10 15 20 0 2 4 6 8 10 -1 -1 Veq [km s ] Veq [km s ] T.H.+ 2014, ApJ
ü All the systems here have only Earth- or Neptune-sized planets. ü Some multiple systems show possible spin-orbit misalignment. Posterior distribu ons of I s
2.5 blue single red multi 2 black isotropic
1.5
1 Probability Density 0.5
0 0 0.2 0.4 0.6 0.8 1 1.2 1.4
Is [radian]
ü We computed the posterior distribution of Is for each system. ü The solid lines indicate the averaged posteriors for single and multiple systems. ü The two distributions (red and blue) seem slightly different from each other. Summary
1. Measurements of the stellar obliquity have revealed a diversity in the spin-orbit rela on, and yielded new hypotheses on the dynamical origin of exoplanets.
2. Observa onal results of the RM measurement imply that some close- in giant planets have experienced quiescent migra on (e.g., Kepler-89), while others are produced by more chao c processes. The reason for this discrepancy between the two dynamical origins is not known.
3. Measurement of the stellar inclina on is a unique technique to discuss spin-orbit rela ons for systems with smaller planets. Previous observa on showed that transi ng systems with small planets generally have good spin-orbit alignment. The distribu ons of stellar inclina ons for single and mul ple systems show a slight difference (Morton & Winn 2014).