Probing  the  Dynamical  History  of  Exoplanets:  Spectroscopic  Observa<ons  

of  Transi<ng  Systems

June. 4, 2015 @ NAOJ

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

0.001

0.01

0.1

1

10

100

0.001 0.01 0.1 1 10 100

plan

etm

ass

[MJ]

semi-major axis [AU]

B

C

A

Snow Line

Distribu<on  of  Exoplanets  II

0.01

0.1

1

0.01 0.1 1 10 100

orbi

tal e

ccen

trici

ty

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.  

Ø What  process  can  make  planets  migrate?  Ø Is  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)  

     

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)  

 

7 © Weidenschilling and Marzari 1996

→ small e and stellar obliquity

→ large e and stellar obliquity

Measurements  of  the  Stellar  Obliquity

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    

Measurement of the stellar obliquity is a unique method to probe the dynamical history of exoplanets!

→ Kamiaka’s talk

approaching receding

time [h] time [h]

RV a

nom

aly

RV a

nom

aly

λ = 0° λ = 50°

The  Rossiter-­‐McLaughlin  (RM)  Effect

←ver<cal  axis:  radial  veliocity    anomaly  during  a  transit  

during  transit

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

planet’s  orbit

Results  by  Subaru/HDS

-30

-20

-10

0

10

20

30

-0.04 -0.02 0 0.02 0.04

Resid

ual [m

s-1

]

Orbital Phase

-40

-20

0

20

40

!v [m

s-1

]

Subaru RV

↑ HAT-P-16: Teff = 6158 ± 80 K λ = -8.4° ± 7.0° VsinIs = 3.9 ± 0.3 km/s T.H. + in prep.

Subaru (this work)

Keck (P08)

↑ HAT-P-7: Teff = 6350 ± 80 K λ = -132.6° +10.5° -16.3° VsinIs = 2.3 ± 0.6 km/s Narita et al. 2009

Kepler-89 (KOI-94): Teff = 6116 ± 30 K λ = -6° -11° +13° VsinIs = 8.01 ± 0.73 km/s T.H.+ 2012b, ApJ Letters This result was later confirmed by Albrecht et al. (2013)

RM  Measurement  for  a  Mul<ple  Transi<ng  System  

-30-20-10

0 10 20 30

-0.01 -0.005 0 0.005 0.01

Res

idua

l [m

s-1

]

Orbital Phase

-40

-20

0

20

40

∆v

[m s

-1]

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

Kepler-89d

Mul<ple  transi<ng  systems  are  generally  aligned!

0.23 0.26

−15

−10

−5

0

5

10

15

RV

(m s−1

)

−15−10−5

05

10

O −

C (m

s−1

)

−6 −4 −2 0 2Time (hr)

−20 0 20λ (deg)

6

8

10

12

v si

n i (

km s−1

)

−20 0 200

5.0•104

1.0•105

1.5•105

2.0•105

0 2.0•104 4.0•104 6.0•104 8.0•104 1.0•105 1.2•105 1.4•105

6

8

10

12

← RM measurements for Kepler-25 (multi-planet system):

Teff = 6190 ± 80 K λ = 7°±8° VsinIs = 8.7 ± 1.3 km/s

Albrecht et al. 2013, ApJ

Obliquity measurements for other multi-planet systems generally show good spin-orbit alignment (Sanchis-Ojeda et al. 2012, Huber et al. 2013). The only exception is Kepler-56 (multiple, but misaligned; Huber et al. 2013).

A companion is present outside of the planetary system, which might be responsible for the misalignment in Kepler-56.

5000 6000 7000 8000

0

30

60

90

120

150

180

proj

. obl

iqui

ty [d

eg]

HAT-P-11 b

HD 17156 b

HD 80606 b

WASP-8 b

5000 6000 7000 8000Teff [K]

20

40

60

80

v si

n i sta

r [k

m s

-1]

Albrecht et al. 2012

|λ|

Summary  of  RM  Measurements

Correla<on  between  Stellar  Obliquity  and  Rela<ve  Tidal  Timescale

100 102 104 106 108

relative tidal−dissipation timescales τCE τRA [yr]

0

30

60

90

120

150

180pr

oj. o

bliq

uity

[deg

]

HAT−P−32 b

HAT−P−7 b

KOI−13 b

WASP−12 b

WASP−33 b

XO−3 b

HAT−P−11 b

HD 17156 b

HD 80606 b

WASP−8 b

↑ 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.

Albrecht et al. 2012

|λ|

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…

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

Previous Targets of RM Measurements

ü  One solution to partly settle the debate is to observe the RM effect for “hot multiple” systems.

ü  So far, no hot multiple systems have been investigated for obliquity measurements.

ü  There are a few hot multiple systems detected by Kepler (e.g., KOI-89).

5000 6000 7000 8000

0

30

60

90

120

150

180

proj

. obl

iqui

ty [d

eg]

HAT-P-11 b

HD 17156 b

HD 80606 b

WASP-8 b

5000 6000 7000 8000Teff [K]

20

40

60

80v

sin

i star

[km

s-1]

Worth a try!

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.  

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.  

→　ΔvRM 〜 - f ・ V sin Is・√(1-b^2)

Es<mate  of  Stellar  Rota<on  Periods

ü  Some  of  the  light-­‐curves  taken  by  Kepler  spacecrar  show  periodic  flux  varia<ons  due  to  rota<on  of  starspots.    

ü  From  such  light-­‐curves,  we  can  es<mate  the  rota<onal  periods  of  planet  hos<ng  stars.    

KOI 988

210 220 230 240BJD−2454900 [days]

0.9850.9900.9951.0001.0051.0101.015

Rel

ativ

eflu

x0 5 10 15 20

Period of rotation [days]

0

5.0•103

1.0•104

1.5•104

2.0•104

Pow

er

Es<ma<on  of  Stellar  Inclina<ons  by  Starspots

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.  

!"#$%&'%&($)*+&%,-+./0

1

2

I$

I.

λ$-%334&)$5+/)46+$

534/%-).&#+-43)46+$

7

Results  1

x axis -> y axis ->

rotational velocities estimated by Kepler photometry

projected rotational velocities estimated by spectroscopy

Solid line indicates Veq = V sin Is, suggesting spin-orbit alignments.

KOI-261 may have a possible spin-orbit misalignment along the line-of-sight.

0

5

10

15

20

0 5 10 15 20

V si

n I s

[km

s-1

]

Veq [km s-1]

257

261

262 269

280

367

974

Is=90o

Is=45o

Is=30o

We observed about 10 KOI systems and estimated VsinIs along with stellar radii.

T.H.+ 2012, ApJ, 756, 66

multiple

Results  2

ü  All the systems here have only Earth- or Neptune-sized planets. ü  Some multiple systems show possible spin-orbit misalignment.

We observed additional 25 systems.

T.H.+ 2014, ApJ

0

2

4

6

8

10

0 2 4 6 8 10

Vsi

nI s

[km

s-1]

Veq [km s-1]

(b) multi Is=90o

Is=45o

Is=30o

2261988

304

0

5

10

15

20

0 5 10 15 20

Vsi

nI s

[km

s-1]

Veq [km s-1]

(a) singleIs=90o

Is=45o

Is=30o 18902002

2026285

323

Posterior  distribu<ons  of  Is

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.

0

0.5

1

1.5

2

2.5

0 0.2 0.4 0.6 0.8 1 1.2 1.4

Prob

abilit

y D

ensi

ty

Is [radian]

blue singlered multiblack isotropic

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).