Observations and Analysis of

Protoplanetary Disks

Sebastian Wolf Christian-Albrechts University Kiel Institute for Theoretical Physics and Astrophysics [email protected]

The Early Phase of Planet Formation – 18-22 February, 2008 – Bad Honnef

Jupiter

Saturn Uranus

Mercury

Venus

Neptune Earth

Mars

The Solar System: Some striking facts • Planetary orbits are (almost) coplanar • Planets orbit the sun in the same direction • Distribution of Mass and Angular Momentum: Sun Mass: 99.86% but Angular Momentum < 2%

• Age estimation: Sun and Planets have been formed at the same time

Immanuel Kant “Allgemeine Naturgeschichte und Theorie des Himmels” (1755)

Planets formed from a rotating gas disk

since 1995:

> 270 Extrasolar Planetary Candidates identified

• “Hot Jupiters” • High masses • Highly elliptical orbits L.R.Cook

Sources of Information

• Solar System • Observations of other Systems • Extrasolar planets/planetary systems • Protoplanetary and debris disks • Numerical Simulations • Experiments (Laboratory Astrophysics)

Overview

1. Introduction 2. Disk observations 3. Analysis: Basic aspects 4. Tracing the process of planet formation 5. Protoplanets in circumstellar disks

Introduction formation, mass, lifetime

Clouds – Stars – Disks

Disk Evolution

Time

(Waelkens 2001)

(Lada 1987)

Disk Evolution UV

(Waelkens 2001)

visible

mid-infrared

far-infrared

mm

Disk Classification • Classification scheme: Based on spectral index s of the emitted flux in the wavelenth range: 2-50/100micron (Lada & Wilking 1984, Lada 1987):

νFν = λFλ ~ λs • Class I : – s > 0 (SED increases with wavelength) – deeply embedded objects – SED = Reemission of infalling envelope

• Class II: – -4/3 < s < 0 – SED of circumstellar disk (stellar and/or accretion heating)

here: SED depends on disk inclination!

Disk Classification • Class III – s ~ -3 – stellar photosphere (Rayleigh-Jeans Limit) – negligible infrared excess

added: • Class 0 (Andre, Ward-Thompson, & Barsony 1993) : – Emission mainly in the submm wavelength range – Evolutionary stage before Class I

Mass Estimation

τν(r) = κν Σ(r) [Rayleigh-Jeans Limit]

κν − mass opacity coefficient Σ(r) – surface density

- average temperature

= 50K κν ~ 0.02 (1.3mm/λ) cm2/g Mgas/Mdust= 100

Mass Estimation •

Continuum SED: Warm dust (only 1% of total mass, but highly opaque) λ ~ mm wavelength range • Disks optically thin • Typical disk mass: ~ 0.01 Msun comparable to the “Minimum Mass Solar Nebula” (total mass of the original material of solar composition to form the planetary system)

Disk Lifetime: Inner Disk Region

!

Fraction of inner disks traced by near-IR excess dimishes to ~0 over a few Myr (Haisch et al. 2001)

Disk observations

Indirect evidence: Outflows

1. Bipolar Molecular Outflows (weakly focussed)

2. Optical Jets (highly focussed)

Collimating processes on size scales < 100 AU 3. Polarization Maps (scattering in bipolar lobes)

Indirect Evidence: Jets

1. Bipolar Molecular Outflows (weakly focussed)

2. Optical Jets (highly focussed)

Collimating processes on size scales < 100 AU 3. Polarization Maps (scattering in bipolar lobes)

HH30 jet, NASA/Watson 2000

Indirect evidence: Polarization Maps

1. Bipolar Molecular Outflows (weakly focussed)

2. Optical Jets (highly focussed)

Collimating processes on size scales < 100 AU 3. Polarization Maps (scattering in bipolar lobes)

[ Gledhill & Scarrott 1989 ]

Infrared Excess

• Passive Disks: Dust reemission following stellar heating (Adams et al. 1987)

• Active Disks: Accretion (Lynden-Bell & Pringle 1974) Can the reemission SED be considered as “Direct Evidence”? => “Opacity argument”

“Opacity Argument” Problem: Is the emitting material really distributed in form of a disk? 1. Argument: • •

mm observations / mm spectrum mass of the disk (optically thin) Derive optical depth, under the assumption of a spherical dust cloud inconsistency with near-infrared absorption measurements

2. Observed SEDs can be well described by disks but: final proof: IMAGES !!!

McCaughrean et al. 1996

Butterfly Star

HST / OVRO

Resolved Circumstellar Disks circumstellardisks.org

Direct Evidence Large-scale images (entire disk) 1.

Optical / Infrared a.

b.

2.

Space: Hubble Space Telescope •

HH30 (Burrows et al. 1996), further edge-on disks (e.g., Padgett et al. 1999): Size ~ several 100 AU

•

“Silhouette Disks” in the Orion Nebula (McCaughrean & O’Dell 1996) Size ~ 50-1000 AU observed in absorption

Ground: •

Problem: Seeing ~1” : comparable to disk diameters Adaptive Optics

•

Solution: Adaptive Optics (also, e.g., Speckle Imaging/Interferometry)

Millimeter Maps (Continuum / Lines) • •

Subarcsec resolution

Interferometry

(e.g., VLA, CSO + JCMT, OVRO)

Molecular lines: disk velocity structure (possible problems: mass infall, outflows dominate kinematic structure on large scales)

Direct Evidence Small-scale structures 1.

2.

Infrared spectroscopy •

Hot gas + dust at the inner disk radius (~100-5000 K within r < 5AU)

•

High gas density, high temperature vibration-rotational transitions well populated NIR/MIR spectroscopy (disk structure + kinematics)

Near/Mid-Infrared Long-Baseline Interferometry •

disk structure within r < 20 AU (in nearby star forming regions)

The Angular Momentum Problem

Proposed Solutions: [1]

Multiple Stellar Systems

[2]

Angular transport to outer disk regions

[3]

Collimated, fastly rotating gas streams (Jets): Angular momentum transfer from the disk to a small fraction of ionized gas (plasma) Magnetic fields in the disk accelerate the plasma: Bipolar “fountains”

Jets and Outflows

Jet velocities 100-500 km/s

Outflow velocities 10-50 km/s

Jets/Outflows are likely driven by magneto-centrifugal winds from open magnetic field lines anchored on rotating circumstellar accretion disks

Jet rotation in DG Tau Corotation of disk and jet

red blue

Testi et al. (2002)

Bacciotti et al. (2002)

Analysis Basic Aspects

Spectral Energy Distribution Heating of the Dust Stellar Radiation – Absorption + Scattering of stellar radiation (UV – near-IR) Dust temperature ~ 10… >103 K – Reemission: near-IR … mm wavelength range Accretion – during early disk evolution – dominating within the inner ~ 10 R*

Spectral Energy Distribution

Assumption: Geometrically thin disk

L*, R*

Problem: Infrared excess*) of this model weaker than observed. *) near-IR

– mm flux above the stellar photospheric flux

Spectral Energy Distribution

Disk with vertical structure => “Disk flaring”

(Kenyon & Hartmann 1987)

e.g. • T(r) ~ r - 3/4 (flat disk) • Vertical gravitational potential dominated by central star Evert ~ -(z/r) G M* / r ~ k T(r) • Scale height: hscale(r) ~ k / (G M*) r5/4

flaring

M* ~ 100 Mdisk

Spectral Energy Distribution

Warm upper disk layer

Cold “inner” disk (high optical depth)

Flaring => Star can heat the disk more efficiently

Spectral Energy Distribution

Inclination dependence: [A]

[B]

λFλ

[A]

[B]

3 component SED

λ τ >>1

low optical depth

Spectral Energy Distribution

To be considered: •

Structure of a possibly remaining circumstellar envelope

•

Dust Emission / Absorption features Radiative Transfer Simulations (detailed numerical simulations taking into account absorption / scattering / heating / reemission processes)

Result: SEDs can be well reproduced, but not unambigiously Images (Vis … mm) required

Contribution from the Envelope „Butterfly Star“ [Taurus]

SED

Near-Infrared Scattered Light [ Wolf et al. 2003 ]

Spectral Energy Distribution

To be considered: •

Structure of a possibly remaining circumstellar envelope

•

Dust Emission / Absorption features Radiative Transfer Simulations (detailed numerical simulations taking into account absorption / scattering / heating / reemission processes)

Result: SEDs can be well reproduced, but not unambigiously Images (Vis … mm) required

Polarization Maps Polarization mechanisms 1.

Scattering Even spherical grains cause polarization which allows to explain observed polarization patterns without further assumptions

2.

Dichroic Extinction by aligned non-spherical grains or anisotropic particles

[ Gledhill & Scarrott 1989 ]

•

Efficient grain alignment mechanism required to explain observed polarization degrees

•

Important for interstellar polarization (magnetic alignment)

Polarization Maps Observed Polarization Degree •

Grain size: a=5-250nm, n(a)~a-3.5 (Mathis et al. 1977) efficient scattering / polarization in the optical / near-IR wavelength range

•

ISM Pmax at 0.45 … 0.80 micron (“Serkowski law”) YSOs Similar, but also at longer/shorter wavelengths

•

[ Gledhill & Scarrott 1989 ]

Net polarization (optical/near-IR): •

ISM < 5%

•

YSOs – usually larger (e.g. HL Tau 12%, V376Cass: 21%)

Optical / Near-IR Polarization Spatially resolved Polarization Maps

1

3 2 1) Single scattering in the envelope (low optical depth) => centro-symmetric orientation, high polarization degree 2) Multiple scattering => pol.vector parallel to disk plane, low polarization degree 3) “Polarization Null point” Vanishing linear polarization at the disk “edge”

Optical / Near-IR Polarization 1

3 2

Polarization degree depends on • Wavelength • Grain size, chemical composition • Density distribution (geometrical structure / dust opacity)

0°

45°

Intensity maps with superimposed linear polarization pattern resulting from scattering on spherical grains. Grains: Size distribution and composition as in the interstellar medium

90°

Optical / Near-IR Polarization

Polarization filter

Detector

Linear polarization degree: Plin = sqrt(Q2 + U2) / I Circular polarization degree: Pcirc = V / I Orientation tan 2γ = U / Q

[Fischer 1995]

Wollaston prism: Simultaneous measurement of Il and Ir / I+ and I-

Optical / Near-IR Polarization

Goal: High-resolution disk mapping in the near-IR PDI Technique: Relies on the high contrast between the polarized and non-polarized radiation component of the scattered / non-scattered light Stokes Q(TW Hya) observed with NACO/VLT [Apai et al., 2004]

Modeling ‘Guidelines’ 1. Take as many independent constraints as possible from observations into account – – – – – – –

Spectral Energy Distribution (mass, disk structure) Absorption/Emission Features (dust properties) Polarization measurements (dust properties, disk structure) Spatially resolved images in various wavelength ranges (tracing different physical processes) Single dish/telescope + Interferometric measurements (tracing disks on various spatial scales) Characterize embedded source Possible Influence of the environment? (e.g., nearby massive stars?)

Modeling ‘Guidelines’ 2.

Set up a disk model with as few parameters as necessary (which are the parameters do you really want/need to constrain?)

3.

a) Radiative Transfer Modeling if necessary; b) Simple ‘Toy Model Fitting’ if sufficient (Problem here: Resulting model usually not self-consistent)

General Optical /Mid-Infrared Interferometric data: • •

Additional constraints for the structure, flux (and dust + gas properties) in the ~mas scale Most useful if considered in the context of complementary observations

Monte Carlo Radiative Transfer • Monte Carlo method: – Very powerful (e.g., wide range of optical depths) + flexible (model) – Direct Implementation of Physical Processes (e.g., Photon transport, Scattering, Absorption, Reemission)

• Capabilities – Models: • (almost) Arbitrary model geometry / density distribution • Arbitrary heating sources (e.g., Stars, Accretion disks) • Arbitrary optical properties of the absorbing/scattering medium

– Output: • Self-consistent temperature distribution • Spectral energy distribution • Wavelength-dependent Images, Polarization Maps

Monte Carlo Radiative Transfer

Weighted Photon packages:

• Wavelength • Energy • Stokes Vector (Polarization!)

Monte Carlo Radiative Transfer Goal Spatially Resolved Dust Temperature Distribution Solution Local Energy Conservation in each cell

Example Adaptive Grid

Tracing the Process of Planet Formation in Circumstellar Disks

The General Picture t=0

Single isolated low-mass star

outflow

t=105 yr

infall Factor 1000 smaller

Cloud collapse

Formation of planets t=106-107 yr

Protostar with disk

Solar system t>108 yr

Planet Formation in a Nutshell Star Formation Process (sub)µm

particles

Circumstellar Disks

Planets

Core Accretion – Gas Capture • Brownian Motion, Sedimentation, Drift • Inelastic Collision Coagulation

cm/dm grains • Agglomeration; Fragmentation Planetesimales • Gravitational Interaction: Oligarchic Growth Planets (cores) • Gas Accretion (Waelkens 2001)

Alternativ: Gravitational Instability

Giant Planet

Dust Grains

Planets

Beckwith et al. (2000)

0.1 µm

1 mm

particle size

~_

wavelength of observation

1m

1 km

log a

Agglomerates in the Laboratory

Silica Monospheres (1.9µm)

Particles stick and form open (fractal) structures

How to identify grain growth? λFλ

• Spectral Energy Distribution (SED) (sub)mm slope: Fν ~ κν ~ λ−β • Dust emission/absorption features

1µm

100µm

• Scattered light polarization • Multi-wavelength imaging + Radiative Transfer Modelling

O / ~ 10µm

credit: NASA

λ

SED: (sub)mm slope

Fν ∼ κν

−

does not depend on disk structure

Frequency dependence can be observed directly

λ ~ (sub)mm wavelengths:

κν ∼ νβ ∼ λ−β

•

Compact, spherical grains, r > r:

•

r >> λ (rocks, asteroids) : β = 0 (opaque at mm wavelengths)

•

Particles with r ~ 1mm:

0