A wider view of planetary
system alignment
Grant Kennedy, IoA, Cambridge
+ Wyatt & Greaves in UK
+ Herschel DEBRIS team
WASP 7b
CoRoT 1b
CV Vel
60
HD80606b
30
HAT P 1b
0
0.01
Spin orbit
misalignment
happens, but
these planets
didn’t form here
TrES 1b
Kepler 16(AB)b
Maximum known misalignment (deg)
90
Can debris disks
around nearby
stars help?
0.1
1.0
10
Binary or Planet semi-major axis (au)
data from www.physics.mcmaster.ca/~rheller
100
“normal” planet formation
debris disk
(Kuiper belt analogue)
has i of PPD at dispersal
NASA/JPL
WASP 7b
CoRoT 1b
CV Vel
Planet migrates
60
HD80606b
HAT P 1b
0
0.01
Kepler 16(AB)b
Ko
za
TrES 1b
i/sc
atte
r
tilts
30
d
n
fi
d
e
t
til
s
k
dis
ing
Planet migrates
?
e
r
he
Disk
Maximum known misalignment (deg)
90
start here
0.1
1.0
10
Binary or Planet semi-major axis (au)
100
Star inclination (deg)
Stellar spin - disk alignment
80
✦
60
✦
40
✦
20
0
0
20
40
60
Disk inclination (deg)
80
No disk “tilting” by
companions
BUT a tilted debris
disk may be easily
destroyed by the
companion
Need to study
implications of
tilting after gas disk
dispersal
Watson et al 2010/11, Greaves, Kennedy et al 2014
WASP 7b
CoRoT 1b
CV Vel
Planet migrates
60
HD80606b
HAT P 1b
0
0.01
Kepler 16(AB)b
Ko
za
TrES 1b
i/sc
atte
r
ing
tilts
30
Disk
Maximum known misalignment (deg)
90
Star+disk
Planet migrates
0.1
1.0
10
Binary or Planet semi-major axis (au)
100
Star-planets-disk alignment
Star–planet–debris disc alignment
901
The Astrophysical Journal, 777:101 (21pp), 2013 November 10
2:1 @ ~1au
χ2
Downloaded from http://mnras.oxfordjournals.org/ at Cambridge Univer
rdisk~70au
inclination
Figure 5. χ 2 of the coplanar 2:1 MMR best fit as a function of inclination.
The open circles represent χ 2 and the filled triangles represent (χ 2 − 2 ln sin i)
2. Herschel 70 µm images of HD 82943, North is up and East is left. The left-hand panel shows the raw image with contours at 3, 5, 10 and
timesof inclination for 2:1 MMR coplanar best fits. The dashed lines
as a20
function
∆χ 2 values of 1.0, 4.0, and 9.0 larger than the minimum χ 2 value
el rms of 1.6 × 10−2 mJy arcsec−2 . The right-hand panel shows the same image and contours after a peak-scaled point source has been represent
subtracted,
around 20◦ , which are the 1σ, 2σ, and 3σ confidence levels. The minimum
near-circular residuals as a clear sign of a near face-on disc.
at about 20◦ is statistically significant, indicating that an inclined solution is
preferred.
✦
>10 years of RV to see planet interactions (~5MJup)
uadratically subtracting the PACS 70 µm beam full-width at
aximum (FWHM) of 5.75 arcsec from the major and minor
nents of the fitted Gaussian FWHM (smaj is also an estimate
characteristic
✦ disc size, about 100 au). To estimate the unty we then added the Gaussian fit image into an off-centre
n in nine other 70 µm observations from our programme (all
ations have the same depth). A Gaussian was then fitted at
osition and the position angle and inclination derived. This
d is a simple way of estimating how the disc geometry can
ue to different realizations of the same noise level. The inclis vary from 25 to 31◦ with a mean of 28◦ , while the position
vary from 133 to 153◦ with a mean of 147◦ .
a second method we fit a physical model for the disc struc-
to the free-fitting parameters when k2 ! 0.025. The right
panel of Figure 9 shows χν2 as a function of inclination with
h2 = 0.01 and different fixed k2 values taken along the arrow
in the left panel of Figure 8. For k2 " 0.025, there is a single
minimum with χν2 ∼ 1.15 at an inclination that increases with
increasing k2 . For k2 ! 0.025, a second minimum with χν2 ! 2.3
appears around i ≈ 15◦ . Because the starting condition of χ 2
minimization is a small inclination along the arrow, the fit is
trapped in the minimum around i ≈ 15◦ when k2 ! 0.035. The
discontinuities in the P1 − P2 grid shown in the right panel of
Figure 8 can be similarly explained.
HD 82943 near to pole-on - system-wide alignment
Kennedy et al 2013, Tan et al 2013
Figu
grid, w
to the e
are libr
dashed
dynam
30◦ and
fits in
confide
edge-o
in the 1
region
and the
angles
Figure
grid. T
are sim
Very
their m
familia
in this
the com
(2009)
GJ 876
(2010)
an upd
Sinc
system
to cons
the sys
as an i
reached
sky pla
χ 2 bec
fitting p
the ord
that we
the res
analysi
WASP 7b
CoRoT 1b
CV Vel
Planet migrates
60
HD80606b
HAT P 1b
0
0.01
Kepler 16(AB)b
Ko
za
TrES 1b
i/sc
atte
r
tilts
30
Disk
Maximum known misalignment (deg)
90
ing
HD82943
β Pic
HR8799
Star+disk
Planet migrates
0.1
1.0
10
Binary or Planet semi-major axis (au)
100
Binary-disk alignment
4.08
20
β Tri
15
3.38
α CrB
2.41
10
10
δ offset (")
2.69
δ offset (")
2.91
0
1.91
5
1.99
0
1.41
1.29
-5
0.91
0.59
-10
0.41
-10
Spacecraft Z
-20
-0.10
20
10
0
α cos δ offset (")
-10
-20
-15
-0.10
15
10
5
0
-5
α cos δ offset (")
close binaries: P = weeks/months
circumbinary planets - alignment expected?
Moerchen+2010, Kennedy+2012, Kennedy 21015
-10
-15
0.00
aligned
2.57
10
Secular precession time (yr)
10
original
Radial
distance (")
5.15
7.72
10.29
12.87
tage
109
most of disk
outside ralign
108
primordially
aligned: yes
107
106
rin
ralign
rout
105
0
100
200
300
Semi-major axis (AU)
400
500
Kennedy et al 2012
99 Her, 15au:
2 parameter
model
how? stellar
interaction?
Kennedy et al 2010
WASP 7b
CoRoT 1b
CV Vel
Planet migrates
60
HD80606b
HAT P 1b
0
0.01
α CrB
Kepler 16(AB)b
HD13161
Ko
za
TrES 1b
i/sc
atte
r
tilts
30
Disk
Maximum known misalignment (deg)
90
99 Her
ing
HD82943
β Pic
HR8799
Star+disk
Planet migrates
0.1
1.0
10
Binary or Planet semi-major axis (au)
100
Outlook
✦
Generally infer peaceful history, 99 Her the exception
✦
Alignment the norm so far - but so was early RM work
✦
Need more star inclinations - beyond vsini, P, Rstar
✦
Disk tilting inferences have caveats - models
✦
Circumbinary planets - form in aligned disks?
✦
Planet+disk systems rare - no transit+disk systems yet
✦
Need disk/planet characterisation, planet discovery
✦
Best prospect: GAIA astrometry
Debris disks trace the
plane of primordial disk
90
WASP 7b
CoRoT 1b
✦
Really start there?
✦
Circumbinary case?
✦
What can we test?
✦
What will we test?
CV Vel
Planet migrates
60
HD80606b
HAT P 1b
0
0.01
Kepler 16(AB)b
Ko
zai/
TrES 1b
sca
tter
ing
tilts
30
Disk
Maximum known misalignment (±deg)
How do their geometries and structures help?
Planet migrates
0.1
1.0
10
Binary or Planet semi-major axis (au)
100
1
100
HD 38858
10
Minimum planet mass (MEarth)
1
100
61 Vir
10
1
Kennedy et al 2015
0.1
1.0
10.0
Radius (au)
100.0
P
m
er
ht
d
d
rd
la
ht
.2
.2
T
ro
id
ot
ht
ti
ot
Σ
et
.2
-etni yb xim esahp ot destood.
wollTwo
a sageneral
w memechanisms
tsys eht of,osuch
itarplanets
tcepsadistributions (10, 11).
Some of these structures and asymmetries
been identified (1): (i) disk fragmentation
-ro eht egats siht tA .sdhave
o
i
r
e
p
l
a
t
i
b
r
o
r
e
t
u
o
0
2
r
o
f
g
n
i
t
a
r
g
and (ii) accretion of gas onto a solid, typically have been theoretically linked to the presence of
etar lacol eht ta dessecewith
rp saa5wto e10lcEarth-mass
itrap hc(MEarth)
ae fo encore.
alpCurlatibone or more massive planets. An inner warp in
sti deniater elcitrap hcarently,
e eliavailable
hw memodels
tsys do
ehnot
t fooffer
egaadetailed
eht rofthe disk plane (12, 13), in particular, can be
description of all of the physical and dynamical reproduced by detailed models that include a
.niaga dexim esahp sawsteps
meinvolved
tsys einhtthese
yllprocesses.
aniF .eThe
sahlifetime
p latof
ibroplanetary-mass companion (13, 14). In addition,
gnisu stibro eht gnivlovgas-rich
e ot tdisks
nelalimits
viuqthe
e savailability
i erudecofonebular
rp sihTspectroscopic observations over several years
and, thus, defines the time window in which revealed sporadic high-velocity infall of ionized
ytidilav sti dna tenalp egas
h
t
fo laitnetop degareva emit eht
gas giant planets can form. Once formed, giant gas to the star, attributed to the evaporation of
-argetni etelpmoc htiw planets
nosirare
appredicted
moc ytobinteract
dekcwith
ehcthendisk
eebandsahcometlike bodies grazing the star (5, 15, 16). The
lacol ehT .elbisaef erewdistort
esehit,t epossibly
rehw leading
snapsto echaracteristic
mit revo disk
snoitobserved comet infall has been attributed to the
structures that can be used to infer the presence of gravitational perturbations by a giant planet witheb ot nekat saw stibalign
roplanets
raluand
cricto constrain
raen rtheir
of eorbits.
tar Up
noitossnow,
ecerpin the disk (17). Taken together, these data and
most giant planets
:)bhave
799been
1 .ladetected
te telaround
liuoM(models suggest that the b Pictoris disk is populated
L = 7.7 T 0.3) of the source-detected northeas
side of b Pictoris in November 2003 (Fig. 1). Th
source lies at a projected separation of 297.6
16.3 milli–arc second (mas) and at a positio
angle (PA) of 210.6 T 3.6°. Within the error bars
the source is located in the plane of the disk. T
confirm the signal detected southwest of
Pictoris in October 2009, we gathered furthe
data in November and December 2009. Together
these data confirm the detection made in Octobe
2009 (see SOM).
The images show that the source detected i
November 2003 could not have been a backgroun
object (Fig. 2). If it was a background object, give
the star’s proper motion (table S1, SOM), th
November 2003 source would be located an
detectable 5.1 AU away, southeast (PA = 147.5°
of b Pictoris in the fall of 2009. The data do no
show such a source (fig. S2). On the contrary, th
source position in fall 2009 is compatible with th
projected position in November 2003 if the sourc
is gravitationally bound to the star (see below).
Based on the system age, distance, an
apparent brightness of the companion, the widel
used Baraffe et al. (23) evolutionary model
predict a mass of ~9 T 3 Jupiter masses (MJup)
Debris disks feel planets
particles inclined w.r.t. planet
r
(moves out)
warp
stars that are several orders of magnitude older by dust, gas, solid kilometer-sized bodies, and
than the lifetime of gas-rich 2circumstellar
D MG3 disks, possibly one or more planets.
preventing the validation of ·models of disk-planet
= | pω| Near-infrared images of b Pictoris obtained in
5
r
Ω
4
interactions and the final phases of giant planet 2003 (18) show a faint (apparent magnitude L′ =
11.2 mag) pointlike source at ~8 astronomical
! accretion.
+8
3
years
-ol eht si r/ ∗MG = Ω The
dnayoung
suida[~12
r l–4atimillion
bro eh
t si (My)],
r ereHunits (AU) in projected separation from the star,
to the Sun (and, consequently, to Earth) within the northeast side of the dust disk. Howdetpoda metsys eht fo nearby
e
g
a
e
hT .ycneuqerf nairelpeK lac
(19.3 T 0.2 pc), 1.75-solar-mass (MSun) star b ever, these data were not sufficient to determine
, 3−01 = ∗M/ M oitar Pictoris
ssam(2,eh
ot a gwide
nid(several
nopsehundreds
rroc ,eofrehwhether this source was a gravitationally bound
3)thosts
ega detamitse tnecer eastronomical
ht htiw units),
tnetstenuous
isnoc edge-on
y 701 6circum.2 sawcompanion or an unrelated background star,
stellar dust disk (4). It is composed of dust whose projected position in the plane of the sky
dna )9991( .la te s é ucsaparticles
vaN ycontinuously
odarraBreplenished
morf sthrough
irotcicolliP β fo
bodies
-gis eht taht evoba sa etsions
on tofuBlarger
.UAsolid
57 ≃
sdleiy hcihw
wR(planetesimals,
comets) and is referred to as a debris disk (5, 6),
ssam eht dna ega eht foin
tcontrast
cudortopmore
ehtmassive
si retgas-rich
emarcounterparts
ap tnacfiin
.oitar
around younger (ages of a few million years)
inogreat
detail
,ksid ydob tnerap depstars.
raw This
ehtdiskfohasnbeen
oitastudied
rugfin
c eh
T
over the past 25 years. Observations at optical to
disk
inclined planet
ecafrus eht htiw ,serudecorp evoba eht gnisu deniatbo
.1 .giF n1iLaboratoire
detartd’Astrophysique,
sulli si deObservatoire
tpoda edewGrenoble,
ytisned
Université Joseph Fourier, CNRS, BP 53, F-38041 Grenoble,
France. 2Space Telesope Science Institute, 3700 San Martin
Drive, Baltimore, MD 21218, USA. 3Laboratoire d’Etudes Spatiales
et d’Instrumentation en Astrophysique, UMR 8109 CNRS, Observatoire de Paris, UPMC, UniversitéParis-Diderot, 5 place J. Janssen,
92195
Meudon,
France. 4European Southern
Observatory
3 after
Lucy-Richardson
deconvolution
of 1. b
Fig.
(ESO), Karl Schwarzschild Strasse, 2, D-85748 Garching bei
isophotes
are more similar among these
(left) and
München,
Germany.
T 5.— Same images as in Fig.
Fig.
Pictoris imaged at L′ band (3.78 microns) with the VLT/NaCo instrument in November 2003
e roff-spot
PSF. The color-coded
e
the fall of 2009 (right). We used images of the comparison star HR2435 to estimate an
convolved
filter images than among
the
shown
in Fig. remove
3. the stellar halo (see SOM). Similar results are obtained when using angular differential imagin
should
E-mail:
ten*To
alpwhom
gnibconvolved
rcorrespondence
utrep eht fo simages
edon fbeo eaddressed.
nil
m
anne-marie.lagrange@obs.ujf-grenoble.fr
(see SOM).
de
hteffects of the deconvolution process are described in Fig. 4.
Augereau et al 2001
.1 .giF
b gnimussa ssam ksid BP naem eht fo noitisop lacitreVwww.sciencemag.org
citthe
rev ‘‘off-spot’’
ehT .2 .giF fo
noitis
ubespecially
irtsid ytisneimportant
d ecafrus ehfor
t
a -ar dna laof
econvolution
PSF
ksid depraw ehT .erugfi siht ni emas eht ton era selacs laid
Lagrange et al 2010
SCIENCE
VOL 329
2 JULY 2010
Circumbinary dynamics
coplanar
pericenter
e=0.77
polar
Figure 4. Simulation of mm-wave structure for an initially flat 50–100 au