Chapter 13. Debris disks
Transcript of Chapter 13. Debris disks
1
Chapter 13.Debris disks
Debris disk phasePP phase
10Myr 10GyrEvolutionary Phases of Planetary Systems
2
Outline Observations of debris disks
Physics of debris disks
Size distribution
Radial structure
Azimuthal structure
Long-term evolution
Short-term evolution
Summary
How do we know they exist?
How do they work?
How large are the dust grains?
Where is the material located?
Why are the disks clumpy?
Why do the disks fade away?
Aren� t they all exceptions?..
What was all this about?
5
(1) Infrared excesses (� Vega phenomenon� )
Aumann et al., ApJ 278, L23-L27 (1984)
What is observed?
The first discovery of an infraredexcess over stellar photosphericemission of a main-sequence star:Vega. Evidence for dust!
6
Zuckerman & Song, ApJ 603, 738-743 (2004)
What is observed?
Since then: thanks to IRAS, ISO, ..., Spitzer,the Vega phenomenon has beenobserved around ~1000 nearbystars
About 15% of main-sequence starshave dust around them
Almost the same incidence aroundA,F,G,K stars, may be slightlyhigher for early types, M unclear
(1) Infrared excesses (� Vega phenomenon� )
7
What is observed?
The first image of a dustdisk around amain-sequence star:β Pictoris
Dust is present in the formof a structured disk
Smith & Terrile, Science 226, 1421-1424 (1984)ESO Press photo (1997)
(2) Images
8
(2) Images
What is observed?
Since then: imagingwith many ground-basedinstruments and HST,both in scattered lightand thermal emission
By now, about 20 disks imaged
Vis NIR FIR sub-mm
~1 Gyr age
~10 Myr age
β Pic
ε Eri
α PsA
α Lyr
AU Mic
9
What is the origin of dust?
Ages of systems: 10 Myr ... 10 GyrBut: estimated lifetime of dust particles: < 1Myr
Dust cannot be primordial andneeds continuous replenishment
As growth is impossible,dust must stem from parent bodies, planetesimals
A conceivable mechanism:collisions of planetesimals
11
What is a � debris disk� ?
Dust
Planets
Star
Planetesimals
Thus there are dust and planetesimals around a star.Solids have a continuous size distribution. They all are referred to as a debris disk.
The star may also harbor planets.The debris disk is genetically connected to, and interacts with, them
Debris disk
12
Why should we care?
Dust observed in debris disks tells us about:planetesimals that supply dust
and planets that affect dust (and planetesimals)
Debris disk studies can give insight into
origin and evolution of planetary systems
13
Does our solar system have a debris disk?
EKB dust disk
Zodiacal cloud
The inner part of our debris disk,known for centuries
The main part of our debris disk,yet to be discovered...
15
Comparison with other disk systems
noyesyesyesCollisions
lowmediumlowhighDyn. temperature
yesnoyesnoEmbedded in gas
--~1015~1010~1030Mass distrib. width
~0.1~10-6~0.1~0.1Thickness, h / r
~1~1..21...10>104Toomre Q
~0~10-310-3...-2(gas)~0Shakura’s α
--2...10104...105 104...105Size / Rprimary
~10...100~1012<106106...109Age / orbits
(solids)
16
Physical processes: (a) collisional cascade
Net result
Collisional cascade grinds planetesimals to ever-smaller fragments, down to dust sizes.
Individual collisions
Critical specific energy for fragmentation:strength and gravity regimes
Under debris disk conditions:-- Disruptive collisions- Cratering collisions
W.K. Hartmann
17
Physical processes: (b) gravity and � photogravity�
Stellar � photogravity� planetesimals move in nearly-circular orbits,dust grains in elliptic ones,fine dust leave the system in
hyperbolas
Gravity of planets confines planetesimals and stirs them up, exerts secular and resonant perturbations on planetesimals and dust
18
Physical processes: (c) drag forces
Poynting-Robertson force dust orbits slowly circularize and shrink toward the star (important for early- and solar-type stars)
Stellar wind drag force � corpuscular analog� to the P-R effect (strong for late-type stars)
Wyatt & Whipple, ApJ 111, 134 (1950)Breiter & Jackson, MNRAS 299, 237 (1998)
Burns et al. (1979), Plavchan et al. (2005)
19
(d) Other processes
Sublimation and sputtering eliminate dust close to the star, gradually reduce their size
(important for early-type stars)
Lorentz force (important for small grains onlyand only if appreciable MF is present)
... and many others that are usually not included in debris disk models, but may be important...
20
Transport-dominated vs collision-dominated disks
Less dusty disks (here: our Zodi) Dustier disks (here: β Pic)
Krivov et al., AAp 362, 1127 (2000)Leinert & Grün (1990)
Collisional timescale can be <> transport timescales:
21
Theoretical methods
Celestial mechanics / N-body simulations- N-body + � inflated spheres� + local PiaB (Thébault, Wyatt)
Accurate dynamics but inaccurate collisions (if any)
Kinetic theory / Statistical codes- Multiannulus PiaB (Kenyon & Bromley, Thébault)
- Kinetics in orbital elements (Dell� Oro, Krivov) Accurate collisions but simplified dynamics
Hybrid methods- � Superparticles� (Grigorieva)
Share (dis)advantages of two previous methods
Main challenge: how to combine accurate dynamics with accurate collisions?
23
Krivov, Löhne, & Sremčevi , ć AAp 455, 509 (2006)
Size distribution
" Tiny blowout grains are present, but in smaller amounts" Grains just above blowout limit dominate cross section" Size distribution is wavy
Blowout limit
24
Size distribution
Thébault & Augereau,AAp 472, 169 (2007)
" Cratering collisions are important" They substantially enhance the � main� maximum
Blowout limit
26
Temperature calculations: Sebastian Müller
Distances from IR excesses
A stars: 24-70µm excesses -> FGK stars: 33-70 µm excesses ->(color) T< ~ 90K (Su et al. 2006) (color) T<~ 45-85K (Hillenbrand et al. 2008)
Compatible with EKB-size disks,extending from several tens AU outward
27
Distances from images
A star (Fomalhaut) K star (e Eri) M star (AU Mic)
150 AU 60 AU
30 AU
Same result: compatible with EKB-size disks,extending from several tens AU outward
29
Planetesimal belts and radial profiles
Krivov, Löhne, & Sremčević, AAp 455, 509-519 (2006)
Surface density index ~-1.5, brightness index ~-3.5
Models
Pla
nete
sim
al b
elt
30
Planetesimal belts and radial profiles
Surface density index ~-1...-2, brightness index ~-3...-4
Liu et al. (2004), Liu (2004), Strubbe & Chiang (2006),Augereau & Beust (2006)
Observations (AU Mic example)
32
Inner gaps: footprints of planets?
Simulations and animations: Alessandro Morbidelli
Inner voids can be naturally explained with two or more planets
34
First attempts to find these of planets
8 young systemswith cold excessessearched with VLT/NACO
no planets found
Apai et al.,ApJ 672, 1196-1201(2008)
36
AU Mic (Liu 2004)
β Pic (press photo)
HD32297 (Kalas 2005 )
α PsA (Holland et al. 2003)
ττ Cet(Greaves et al.2004)
All resolved disks are structured
ε Eri (Greaves et al. 1998)
Radial,azimuthal,
verticalstructure:
●rings,●gaps
●clumps●spirals●offsets●warps
...
Vega (Su et al. 2005 )
HD32297 (Kalas 2005 )
HD139664 (Kalas 2006 )
37
Greaves et al., ApJ 619, L187-L190 (2005)
Example: ε Eri
Theory of clumps
Suggested scenario:resonant (clumpy!) family ofplanetesimals produce dust,which stays in the resonance
Wyatt, ApJ 598, 1321-1340 (2003)
38
Wyatt, ApJ 598, 1321-1340 (2003)
Theory of clumps
Can be captured in resonances by migrating planet
Will dust stay in resonance?
Resonant populationof total mass M
Sizes* from smax (planetesimals)* down to smin (dust)
Collisional cascade
Loss at dust sizes* small grains (RP)* fast grains
Krivov et al., AAp 462, 199-210 (2007)
How to create resonant family?
39
Clumps like ε Eri� s could be produced by a resonant planetesimal family of lunar to martian mass
α =0.833...α =0.87
Theory of clumps
Krivov, Queck, Löhne, & Sremčević, AAp 462, 199-210 (2007)
41
IR surveys indicate a long-term decay of debris disks
Moór et al. ApJ 644, 525-542 (2006)
" On the average, dust luminosity goes down with system� s age" A high dispersion for any given age though
t-1 t-0.5t-2
42
Dominik-Decin (2003) model
Dominik & Decin, AAp 598, 626-635 (2003)
" Equal-sized planetesimals � feed� dust" Dust has a single power-law size distribution
43
Dominik-Decin (2003) model
Dominik & Decin, AAp 598, 626-635 (2003)
" For collision-dominated disks (usually the case),
total disk mass ~ dust mass ~ t -1
" �Delayed stirring� suggested to explain high luminosity in some of the old systems
44
Wyatt et al. (2007) model
Still, an assumption of a quasi-steady state and a single power-law size distribution
Wyatt et al., ApJ 658, 569-583 (2007)
45
Wyatt et al. (2007) model
" For any given age, there is a maximum possible amount of dust" Disk evolution depends on r, e, I, Dc
Wyatt et al., ApJ 658, 569-583 (2007)
46
Löhne, Krivov, & Rodmann, ApJ 673, 1123-1137 (2008)
Löhne et al. (2008) model
" Large planetesimals are not in collisional equilibrium" Strength-gravity transition plays a major role
Krivov, Sremčević, & Spahn,Icarus 174, 105-134 (2005)
48
Löhne et al. (2008) model
" The dust mass decays as ~t ξ
" Index ξ depends on the � primordial� size distribution of planetesimals"Typical values: ξ ~ -0.3...-0.4 and not -1
49
Löhne et al. (2008) model
A synthetic population of debris disks calculated with the modelis in a good agreement with the Spitzer 24 and 70µm statistics
Löhne, Krivov, & Rodmann, ApJ 673, 1123-1137 (2008)
51
Possible transient events
• Major collision between planetesimals
• �Supercomet�
• Passage of a nearby star
• Dynamical instability caused by planets
52
Possible transient events
Grigorieva,Thébault, & Artymowicz, AAp 461, 537-549 (2007)
" A spiral-ilke pattern for ~1000 yrs" Avalanches possible, but only for dustiest disks
• Major collision between planetesimals
• �Supercomet�
• Passage of a nearby star
• Dynamical instability caused by planets
53
Possible transient events
Simulation and animation: Alessandro Morbidelli
Nobody knows whether LHB-type events are common
• Major collision between planetesimals
• �Supercomet�
• Passage of a nearby star
• Dynamical instability caused by planets
54
Su et al., ApJ 628, 487-500 (2005)
Is Vega transient?
Fits to SED & radial brightness profiles @ 24, 70, 160 µm suggest that small blowout grains are overabundant. Resulting size distribution not dynamically justified...
Blowout limit
55
Marsh et al., ApJ 646, L77-L80 (2006)
Is Vega transient?
...so we tried size distribution from our collisional model (Krivov, Löhne, & Sremčević, AAp 455, 509-519, 2006),
with a partial success (Müller, Krivov, Löhne, in prep.).
Blowout limit
56
Is β Pic transient?
The 52 AU blob is warmer than the surrounding disk(T=190 K vs 140 K);gives grain size ~0.1µm;must be blown out on a 1000 yr timescale;therefore transient
Origin:major breakup?supercomet?nearby star?All unlikely...
Telesco et al., Nature 433, 133-136 (2005)
57
Is HD 69830 transient?
Only 2% of FGK starsexhibit hot dust detections (< 24 µm or < 10 AU)
Most of these systems with hot dust (4 / 7) have dust luminosities larger than � maximum for a given age�allowed by stationary collisional models
Wyatt et al., ApJ 658, 569-583 (2007)
Beichman et al., ApJ 626, 1061-1069 (2005),Lovis et al., Nature 441, 305-309 (2006)
Example:HD 69830
1AUPossible parent body locations
58
Beichman et al.,ApJ 626, 1061-1069 (2005)
Is HD 69830 transient?
At a first glance:looks like cometary dust!
Lisse et al., ApJ 658, 584-592 (2007)
A closer look: HD69830 50% crystalline,Hale-Bopp only 7-30%. Asteroidal dust?
59
Summary
• Debris disks are optically-thin, dusty, gas-poor disks around main-sequence stars
• Debris disks are common around those stars
• Debris disks are direct evidence for planetesimals
• Debris disks are indirect evidence for planets
• Debris disks give clues to the history of planetesimal and planet formation and evolution
• There is a bulk of observational data on debris disks; there are elegant theories of debris disks; these observations and theories do not necessarily agree with each other