Download Dust in protoplanetary disks

Dust in protoplanetary disks
Dmitry Semenov
Max Planck Institute for Astronomy
Heidelberg, Germany
http://www.alma.inaf.it/attachments/article/118/Lecture12-Semenov.pdf
Suggested literature
•
A. G.G.M. Tielens, "The Physics and Chemistry of the ISM" (2007), CUP
• B. Draine, "Astrophysics of Dust n Cold Clouds" (2003), astro-ph/0304488
• "Protoplanetary Dust" (2010), eds. D. Apai & D. Lauretta, CUP
• "Protostars & Planets V" (2005), Part VI, eds. B. Reipurt et al., Univ. Arizona P.
• N.V. Voshchinnikov, "Optics of Cosmic Dust I/II" (>2004), CUP
Outline
• The role of dust
• Light interaction with grains
• Structure & evolution of protoplanetary disks (PPDs)
• Evolution of dust in PPDs
• Observations of dust in PPDs
• ALMA observations
The role of dust in star-formation
• Dust: ~25Å–1cm
• Widespread, 1% by mass (ISM)
• Opaqueness of matter
• Heating & cooling
• Sink of heavy elements (>Na)
• Provides surface for catalytic reactions & adsorption
I. Light interaction with grains
Basics definitions
• Extinction = absorption + scattering: Qext = Qabs + Qsca
• Cross-section for grain of radius "a":
Cext = πa2 Qext
2πa
• Size parameter for a wavelength "λ": x = λ
• Complex refractive index:
• Single-scattering albedo:
m = n − ik
ω = 1 − Qsca /Qext
• Phase function: p(θ)
• Polarization: full Stokes vector (I,Q, U,V)
• Different theories for various combinations of x and m
• Measured refractive indices are needed
Distinct regimes of light scattering
• Small dielectric grains (Rayleigh limit): x → 0, |mx| → 0
Qabs ∝ a, Qsca ∝ a4 , p(θ) = 3/4(1 + cos2 θ)
x → ∞, 2x|m − 1| → ∞
• Huge grains (geometric optics):
2.2. THE MODEL
Qext = 2,
4kx � 1 : Qabs = 2,
4kx � 1 : Qabs ∼ a
forward scattering
theory
• Intermediate case: need exact
Figure 2.1. In the left panel a typical ballistic particle-cluster aggregate (BPCA) is shown. The
(Mieis presented
1908)in the middle panel. An “onion-like” (multishell) s
cluster-clusterspheres
aggregate (BCCA)
- Mie theory for homogeneous
particle is schematically depicted in the top right panel, whereas a “composite” (many-many layered
et al. 2004)
- Multilayered spheres (Voshchinnikov
is shown in the bottom right panel.
- Infinite/finite cylinders (Bohren & Huffman 1983)
2.2.1.2 1979)
Grain structure and topology
- Spheroidal particles (Asano
In many studies for
of the arbitrary
dust properties in
protoplanetary
discs,
the grains are
3D
particles
(Purcell
&still assumed to be
- Discrete Dipole Approximation
cal particles (e.g., Natta et al., 2001). However, it becomes evident from theoretical investigat
Pennypacker 1973)
laboratory experiments that the dust agglomeration is an efficient process in dense and relativ
environments, like protostellar cores or protoplanetary discs (e.g., Blum et al., 2002; Kemp
1999; Kesselman, 1980; Nuth and Berg, 1994; Ossenkopf and Henning, 1994; Wurm an
22
Light interaction with grains: examples
121
5.2 Physical processes
Q-factors: sphere
Q
m = 1.5
4
3
Qext
2
1
0
m = 1.5 + i·0.02
4
3
Qsca
2
Qext
Qabs
1
0
m = 1.5 + i·0.05
4
B. T. Draine
Scattering efficiency:
tetrahedron
3
2
1
0
Tielens (2007)
0
5
10
x = 2πa /λ
15
20
5.1 The extinction
and scattering
efficiency
calculated using Mie theory
of
optical
constants
&
properties:
• DatabaseFigure
for spherical grains plotted as a function of the size parameter, x = 2!a/".
The adopted optical constants are indicated in the panels. Figure adapted from
H. C. van der Hulst, 1981, Light Scattering by Small Particles, (New York:
Dover).
http://www.mpia-hd.mpg.de/HJPDOC/index.php
Draine (2000)
2
scattering and the absorption efficiency both approach unity and the extinction
efficiency goes to 2. This result in the geometric optics limit may seem, at first
0.11± 0.07
0.36± 0.04
0.45± 0.07
2
0.37± 0.05
0.52± 0.03
0.57± 0.11
"!, cm /
Our IRAM
observations
are crucial
constrain
-2.± 8.
46.7±
3.2Max
9.1±
0.7 to 2.24±
0.20disk 2m
measuring,
or placing
stringent
on the0.09
mass 1o
-1.± 10.
113.5±
4.0
31.3± 1.0limits,1.76±
contained
in
grains.8.0±
Since
the millimeter
emissio1
-2 small
31.± 28.
34.2±
3.1
1.0
1.99±
0.30
10
the disk2 is optically thin,
we can derive themdisk
d dust m
2
Apparent •sizes
and orientations
from/g):a κ(ν)
Gaussian
fit
(Col
2-4)
to
the 1.3 mm
Extinction
per unit massderived
of gas
(cm
∝
πa
Q
(ν,
a)N
(a)
ext
gr
from the observed millimeter fluxes, Sν : mg
n 100 m. Total flux at 1.3 and 2.7 mm (or 3.4 mm for stars with (*) in Col 1) (Col 5-6)
Disks are optically thick: λ ≲ 100 μm -4
•
ved from Gaussian fit to all visibilities. α100 10
(Col 8) is the apparent
index for
Sνspectral
D2
• Dust mass in PPDs at distance D with flux Sv: Mdust = kν Bν (Tdust )
10
100
1000
10000
1
10
100
1000
Grain
sizes, topology, porosities
2.2. •THE
MODEL
!, µm
!, µm
β
where
k
ν = k0 (ν/ν0 ) is the mass absorption coeffic
Conductingisothermal.
materials (Fe,
FeS,
graphite)
The
dust
opacity
as
a
fun
are
thus
vertically
To
allow
a
homoge•
and mean opacity:
mean
opacity:
parameterizesRosseland
the frequency
of kν , Sν is t
2
2 dependence
pletes
the
description.
In a fir
parison,
we
used
T
=
T
(100
AU)
=
15
K
and
10
10
100
del
IRS
model
served
flux,
D is the distance to the source, Tdust is th
10
IPS
model,
cept when those parameters can be constrained by the dent of radius and described
T<120
K
temperature,
Bν (Tdust ) is the Planck function. We as
(I)
ns. The validity and impact of this assumption 2will be
k0 = 1 cm
/g at 1.2mm,
β=
= κ1230
and
Tdust =40 K.βm ,
κ(ν)
(ν/230GHz)
Max
n Sec.4.1.
0
0
We
have
detected
5
sources:
10
10 ScoPMS52 and [P Z99] J16
ISM-like
dust
1 (2002)
(V)
2 −1
analyses have been used by Kitamura
et
al.
with
κ
=
2
cm
gβ: (per
230536 (primordial disks),
HD
8907,
HD
104860
and H
230
Dust
emissivity
slope
ws & Williams5-layered
(2007)
for their 2 mm
and
0.8 mm
spheres
5-layered spheres
thethe
mem
(debris
disks).
The 1.2additional
mm fluxesparameter,
measured
and
composite aggregates
ISM
dust
composite
aggregates
ctively. Most previous
studies (Kitamura
etfrom
al. 2002;
• ~2 fortosmall
derived
Eq.1
are
compiled
in
Table
1.
similar
the
Beckwith
et
al
-2
-2
composite spheres
composite spheres
10
10
& Williams
2007;
Isella
et
al.
2009)
used
the
thin
disk
0.1
>0.5
for frequency
large
grains
For the sample sourcesent
millimeter
fluxesto
not
dete
pivot
avoid
•with
homogeneous aggregates
homogeneous
aggregat
tion to compute
visibilities.
our we
samhomogeneous
spheresHere, because
spheres
the 3σ level,
compute
upper
In
diska dust
massmass
onhomogeneous
β.
Thelimits.
dust
m
porousobjects,
5-layered spheres
porous 5-layered spher
(I)
es(I)highly inclined
we assume
that
thethedisks
we
present
derived
dust
masses
and
the upper
limi(
(2007)
and
Andrews
et al.
porous composite spheres
porous composite sphe
2
2
"!, cm /g
2
!R, cm /g
Light interaction with grains: opacities
"!, cm /g
xS
with a -4scale height varying as a power 0.01
law of the with-4a slightly different absor
10 1 100
10 1
−h
1000
100
10000
100
10
100 10
= H100 (r/100AU)
.
For
all
but
the
two
highly
inFinally, we T,
also
T, K
!, µm
!, µm
K assume that
ectsSemenov
(HH 30(2005)
and DG Tau-b), we used H100 = 16 AU equal to 100. In a second ste
I. Light interaction with grains: Summary
• Opacities
• Thermal balance
• Dust emissivity slope varies at >100μm
• Sizes, topology, porosity, conducting materials
• Tools & databases of optical constants & properties
II. Structure and evolution of
protoplanetary disks
Life cycle of dust in Galaxy
• Nucleation and initial growth in
AGB shells
• Mixing & growth in the ISM
• Collapse of a dense cloud
• Rotating accretion disk
• Gas and dust "dispersal"
• Formation of planetary system
• A new AGB star at the end
Credit: Bill Saxton, NRAO/AUI/NSF
Disk structure & evolution
• Conservation & redistribution of angular momentum
• Gas viscosity
• Gravity
• Equation of state
• Initial mass
• Initial angular speed of a cloud
• Characteristics: Rdisk, Mdisk, Tdust, Tgas, surface density, accretion rate, ...
Disk vertical structure
• Equation of state: P = c2s ρ
dP
• Hydrostatic equilibrium: dz
= ρΩ2K z
• Assuming isothermal structure in z-direction:
z2
• Density profile: ρ = ρ0 exp(− 2 )
H
• Disk pressure scale height: H =
• H ~0–5
√
−1
2cs ΩK
Disk shape
• Assume power-law model for temperature: T ∝ r−q , cs ∝ r−q/2
�
• Keplerian rotation: ΩK = GM∗ /r3 ∝ r−3/2
√
• Aspect ratio: H/r = 2cs /rΩK ∝ r(1−q)/2
• Flaring disks when (1-q)/2>0, q<1
• A typical disk with q=1/2: H/r ∝ r1/4
ar mass and accretion rate in young stellar and
more than four orders of magnitude in accretion
ontext of disk evolution models. We note that
sk angular momentum transport efficiency with
n disk initial
conditions.
In this(Shakura
case we find
that 1973): ν = αcs H
viscosity
& Sunyaev
• Anomalous
around more massive objects. By considering
turbulence),
(gravitational instability)
• α ~0.01at(MHD
t-off in accretion
the end
of the disk~1lifetime,
mass-accretion
rate relationship
should increase
& redistribution
of angular momentum
• Conservation
bservations.
sequence — stars: low-mass, brown dwarfs —
• Vertically-integrated quantities
Disk radial structure
• Classical model of 1D viscous evolution (Pringle 1981, Linden-Bell & Pringle
e equation
forShakura-Sunyaev
the surface density
evolution, Σ(R,t), of a
1974,
1973):
us accretion disk is !
$
#
"
∂Σ 3 ∂
1/2 ∂
R
νΣR1/2 ,
(1)
=
∂t
R ∂R
∂R
e R is cylindrical radius, t is time, and ν(R,t) is the kinec viscosity of the disk (Pringle 1981).
If the viscosity
−1
• "Early"
r
dependent
of timedisk:
andΣ(r)
can be∝expressed
as a power-law
γ
ion of radius
ν
∝
R
then Equation
• "Late" disk: ,Σ(r)
∝ r−3/21 permits exact anc solutions. Here we consider the similarity solution of
mann et al. (1998), after Lynden-Bell & Pringle (1974).
s case the solution for the surface density of the evolving
Evolution of disk: 1D
Dimensionless time
Start
End
Surface density: Power-law with tapered edge
Pringle (1981)
Evolution of disk: 2D
55
The origins of protoplanetary dust and the formation of accretion disks
Age: 131081 yr
30
–10
–14.0
–40
–20
0
x [AU]
20
40
z [AU]
–14.0
–20
–20
0
x [AU]
20
40
–20
0
x [AU]
20
40
Age: 132592 yr
–14
10
.0
0
–12.0
–10
–20
–14
• Accretion flow (arrows)
.0
–12.0
0
–10
–14
.0
3 km/s
–40
–20
0
x [AU]
20
40
• Accretion heating
• Initial disk is decreting disk
• Rapid phase: ~0.1–0.5 Myr
–14.0
–13.0
10
0
–14.0
–10
–40
–20
0
x [AU]
20
40
–30
3 km/s
–40
–20
0
x [AU]
• Shock passage: destruction of
molecules and grains
Age: 132640 yr
–20
3 km/s
• Supersonic to subsonic
transition (black line)
20
z [AU]
z [AU]
–40
30
20
–30
2 km/s
10
–30
Age: 132606 yr
30
.0
–20
2 km/s
–40
–14
20
0
–10
–10
30
–14.0
10
0
–30
Age: 131610 yr
20
10
–20
1 km/s
30
–30
z [AU]
0
–20
z [AU]
20
–14.0
10
–30
30
–14.0
z [AU]
20
Age: 131276 yr
20
40
Gail
Hoppe
(2010)
Figure&2.10
Cylindrically
symmetric hydrodynamical model of accretion flow with
Accretion time
0.18
0.15
Evolution of a star and a disk
Hueso & Guillot (2005)
(Myr)
Mass
accretion
vs Age
Evolution
of protoplanetaryrate
disk structures
Figure
3.3 Left:
measured
accretion
rate astend
a function
age (Sicilia-Aguilar et al.
accretion
ceases
with time:
~10–30ofMyr
• Mass
2006). The line is the prediction of a single Lynden–Bell & Pringle model as a
accretion
at given
age depends
ona initial
angular
momentum
• Mass
function
of time.
Right:rate
measured
accretion
rates as
function
of stellar
mass for
a cluster of a given age, in this case in Ophiucus. The lines are model predictions
for a given initial dimensionless rotation rate of the parent molecular cloud core
Ciesla & Dullemond (2010)
from which these stars were formed. From Dullemond et al. (2006b).
79
Detailed disk models
• 1+1/2D passive models (D'Alessio et al. 1998, Dullemond & Dominik 2004):
• Radiation of a central star
• Radiative transfer:
• plane-parallel/2D,
• frequency-averaged (grey),
• frequency-dependent
• Heating & cooling of gas
• Hydrostatic equilibrium
• 3D MHD models (e.g., Flock et al. 2011):
• ~10–100 orbital periods
• Isothermal or simple power-law for T
• Local disk patches
• Realistic turbulence
Passive steady-state disk: 2D
T Tauri
with PAHs
no PAHs
τross=1
Herbig Ae
with PAHs
τross=1
Tgas = Tdust
Flaring disks, Tgas>Tdust in atmosphere
Kamp et al. 2005; Jonkheid et al. 2004; Nomura & Millar 2005; Woitke et al. 2009
o. 1,
2004
L54
Real disks: 3D!
SPIRAL STRUCTURE
IN DISK AROUND
FUKAGAWA
ET AL. AB AUR
gnificantly affect the derivation of the geometry of the NIR
isk, because the fitting applied to inner and outer regions gave
onsistent results. It should be noted, however, that any intrinsic
rightness distribution asymmetry that makes the northwestern
art darker, as may be the current case caused by an anisotropic
cattering phase function, does affect the inclination. The inination of 30! should be taken as an upper limit in such a
ase.
The derived inclination agrees well with the recent NIR inrferometric measurements for the inner (r ∼ 0.5 AU) disk
Eisner et al. 2003; Millan-Gabet, Schloerb, & Traub 2001)
nd with the constraint (less than 45!) obtained by the optical
maging with the HST/Space Telescope Imaging Spectrograph
STIS; Grady et al. 1999). It is, however, significantly smaller
han 76! estimated from the 13CO observations by Mannings
& Sargent (1997). The position angle of the major axis is also
ifferent from that of the millimeter measurement (P.A. p 79!)
y 20!. Lower spatial resolutions of millimeter observations
may have caused such discrepancies, and higher resolution imgings of the thermal emission are necessary for obtaining more
recise constraints on the disk geometry. On the other hand,
he STIS optical image (Grady et al. 1999) lacks a distinguishble axis, not showing any clear ellipticity. Because the image
hows a nebulosity much more extended and more circularly
Fig. 1.—H-band
image image,
of the circumstellar
structure
aroundbeAB
Aur after
istributed
than the NIR
the optical
flux may
doma
reference
PSF
was
subtracted.
The
surface
brightness
is
multiplied
by the
nated by scattering from the region with a large-scale height
distance squared from the center for display so that the fainter outskirts can
specially
at with
large
radii.
TheBoxcar
STISsmoothing
wedge is
occults
the 5region
be viewed
a high
contrast.
applied with
# 5 pixels.
xactly
along
the
derived
major
axis,
which
also
makes
it inner
dif- area
Directions of the spider patterns are indicated by dashed lines. The
diameter a(r distinguishable
! 120 AU; filled circle)
cultofto1!. 7identify
axisisinphotometrically
the image. unusable and is
AB Aur
AB Aur
Fig. 2.—Azimuthally averaged radial profile of the surface brightness ( filled
circles) after the assumed inclination of i p 30! was corrected. Error bars
show the dispersion of brightness over the azimuth of 360! and radial width
of 10 AU (0!. 07). The dashed line indicates a power-law fit with an index of
!3.0 to the brightness over the radial range between 120 and 580 AU.
RESULTS
Fukagawa et al.3. (2004)
3.1. Scattered Light from the Disk
Fig. 3.—Same as Fig. 1, but the image is deprojected with an assumed
Figure
1 to
is show
the resultant
H-band
after
PSF
subtracinclination
of 30!
the “face-on”
view image
of the AB
Aurthe
disk.
Some
of
the major
are identified.
tion. features
We detected
an extended emission seen from the edge of
• Gravitational instabilities
• Interaction with a nearby star
• Planet-disk interactions
• Large-scale turbulent waves
masked. The field of view is 8!! # 8!!. North is up, and east is to the left.
3.2. The Spiral Structure
vations. The signal-to-noise ratio of the reference star PSF was
greater
than
3 atinrFigure
was similar
to thatspiral
of the
AB Aur
! 6 !!, as
The
H-band
image
1 shows
a remarkable
pattern
data.
FS 111AU.
wasThe
observed
immediately
before
Aur and
rp
200–450
spiral pattern
coincides
with AB
the spiral
usedseen
as ainphotometric
calibrator
al. 2001).
andwas
structure
the optical image
taken(Hawarden
with the HSTet(Grady
The
sky
was
clear,
and
the
natural
seeing
size
was
0!. 5. The
al. 1999). The new image has, however, revealed the entire
piralspatial
patternresolution
located inachieved
the innerwith
part the
(200AO
AUsystem
! r !(Takami
300 AU).et al.
2004)
was
0!
.
10
(FWHM),
which
was
measured
with
the Lyot
n addition, it clearly shows the spirals with a high contrast
stop the
in the
optical light
path toward
an surrounding
unmasked star
in the frames
ecause
scattered
from the
envelope
is
of SAO
egligible;
the57754.
spiral pattern is associated with the circumstellar
At
the
second
run on
11,first
the seeing
isk but not with the observing
envelope. This
is January
indeed the
case insize
anda eventually
the resolution
weregap,
a little
hich
spiral pattern,
not a ring with
or a AO
circular
hasworse
been and
variable,
the sky
clear.
we used
62 deframes
etected
in thealthough
NIR around
a was
young
star,Hence
although
it was
that
resolutions
similar
those obtained
in al.
the2001)
first run.
cted
in had
the optical
image
of HDto 100546
(Grady et
Theastotal
time was 6.2 minutes for the 62 frames,
s well
AB exposure
Aur.
The
southeastern
part is brighter,
whichofsuggests
thiswas
each
taken by co-adding
six exposures
1 s. SAOthat
57393
Schneider et al. (1999, 2009)
L55605
Vol.
HR 4796
the occulting mask (r ∼ 60 AU) out to the radius of 580 AU
(p 4!. 0), where the brightness drops to the detection limit of
4. DISCUSSION
0.3 mJy arcsec!2. Figure
2 shows an azimuthally averaged
radial
brightness
profile
of
image
withspiral
the asWhat is the mechanism to the
excite
anddeprojected
maintain the
sumedininclination
30! Aur,
and major-axis
position
angle
structure
the disk ofofAB
a single star
with the
ageofof58!
(see
below),
showing
that
the
surface
brightness
decreases
4 Myr?
Theoretical calculations show that a forming planet lo- as
!3.0"0.1
r in a disk
withopens
the radius
120 toMiyama,
580 AU.&The
cated
a gapr from
(Takeuchi,
Linpower-law
1996) !2
dependence
revealed
in
this
study
is
steeper
than
of r
that is often associated with a spiral structure extendingthat
inward
the optical
nebulosity
et al.(e.g.,
1999).
Theetsteeper
slope
and for
outward
into the
disk from(Grady
the planet
Bate
al. 2003).
in
the
NIR
suggests
that
the
detected
light
originates
mainly
If the dark lane at r ∼ 300 AU is a gap where an unseen comfromis the
disk itself
without
significantly
panion
located,
its mass
mustbeing
be less
than 10MJcontaminated
, which is
by
the
scattering
emission
in
the
envelope,
which
estimated from the evolutionary tracks given by Burrows shows
et al. a
shallower
slopeet(Grady
et al.in1999).
also justified
(1997)
and Allard
al. (2001)
order This
to be isconsistent
with by
the fact that
theofNIR
has a region.
size similar
the detection
limit
H ∼ scattering
16.5 mag emission
in the interarm
The to
13
CO
disk
(Mannings
&
Sargent
1997).
that
of
the
main structure that we may observe in a disk, however, would
We integrated
thea spiral
scattered
lightsuch
overasthe
range of
be a circular
gap but not
structure
the radial
one revealed
120
AU
≤
r
≤
580
AU
and
calculated
the
ratio
of
the
scattered
in this study if there is an unseen companion, because
matter is
to total fluxes as Fdisk /Ftotal p (1.2 " 0.2) # 10!2, adopting the
cleared away in the gap
while the spiral pattern is merely the
total flux H p 5.1 mag of AB Aur (Hillenbrand 1992). The
density fluctuation of matter. It is therefore not probable that!2the
H-band flux ratio is comparable to those of (2–4) # 10 measpiral structure is produced by an unseen companion.
Surface density
Rice et al. (2005)
3D MHD
Photoevaporation of disks
Hubble WFPC 2
Gorti et al. (2009)
• UV/X-ray radiation from the star: 1–10 AU
• Superheated atmosphere gas flows away
• T Tauri star: Mevap ~10-9–10-8 MSun/year
• Clearing of inner regions
II. Structure and evolution of PPDs:
Summary
• Outcome of cloud collapse: ~ 0.1–0.5 Myr
• Anomalous viscosity (turbulence)
• Macc ceases with time
• Macc depends on initial angular momentum
• "Early" disks: accretion heating
• "Late" disks: passively reprocess L*
• 3D structure
• Σ(r, t) can depart from a power law
• Photoevaporation
III. Dust evolution in PPDs
Dust evolution in disks: ~1–10 Myr
(pre-transitional)
ISM dust
Pebbles
Mdisk ~ 1% of Mstar
(transitional)
Protoplanets
Rocks
Williams & Cieza (2011)
Mdisk ~ a few MEarth
Dust evolution in a nutshell
!"#$ %&'($) *+ ,&'$'-./+0$/&1 !*#2#
• <10 cm/s collisions due to Brownian motion
• Sticking
• >1μm–1cm grains sediment
Sticking
• Rain drops-like growth regime
• Fragmentation (V>10–100m/s)
Bouncing
• Erosion
• Turbulence returns small grains upward
Fragmentation
• Inward drift
!"#$ % 78/9-.0# ': ./;'&/$'&1 <'..*#*'+ 08-0&*90+$# ;0$(00+ !"::1 SiO /==&0=/$0#
#".$0@ *+ #$*<2*+= A!"#B> ;'"+<*+= A$%&&'(B> /+@ :&/=90+$/$*'+ A)"!!"$BC D)0 <'..*#*'
• Mostly proved by experiments
(0&0 5C55F m s A*!%+,%-.B> 5C3G m s A)"/-+%-.B> /+@ 6C3 m s A012.$(-!2!%"-B> &
2
−1
−1
−1
H.. 08-0&*90+$# (0&0 -0&:'&90@ "+@0& 9*<&'=&/E*$1 <'+@*$*'+#C I'& <./&*$1> $)0 <'..*@
=&0=/$0# *+ $)0 #$*<2*+= /+@ ;'"+<*+= </#0 /&0 9/&20@ ;1 /+ xC D)0 $*90 *+$0&E/. ;0$(
"E0 *9/=0# *+ $)0 #$*<2*+= </#0 *# 6F 9#> $)0 ./#$ $)&00 *9/=0# )/E0 $*90 *+$0&E/.# ': J65
$)09 $' -&'E0 $)0 @"&/;*.*$1 ': $)0 /@)0#*'+K $)0 $*90 *+$0&E/. ;0$(00+ $)0 *9/=0# *+ $
</#0 *# 3F 9#K $)0 $*90 *+$0&E/. ;0$(00+ $)0 "&#$ $(' *9/=0# /+@ $)0 ./#$ $(' *9/=0# *+
:&/=90+$/$*'+ *# 3 9#> &0#-0<$*E0.1C
Weidenschilling et al. (1993), Blum (2010)
% & '() *+,-.&//0(/&-( 12334,42' 52*(3
experience the largest inward drift velocities (see Weidenschilling 1977
Weidenschilling 2006; and Chapter 10).
How to pass a 1m-barrier?
• Big grains
100% Keplerian rotation
• Radial pressure gradient
gas orbits at
Inward drift
velocities
99% Keplerian velocity
• Head wind: inward drift
• 1m particle: ~104 years from 100 AU
• Vertical settling: ~1 Myr @ 1 AU for 1μm
• Vertical stirring: ~104 years @ 100 AU
Figure 3.5 Plotted are the inward drift velocities of particles of differe
in a disk where the velocity differential between the gas and a Keplerian
70 m s−1 . The kink at ∼10 cm is due to the change in the gas-drag law
• Coagulation: ~0.1–2 Myr
Possible mechanisms:
particles exceed the mean free path of the gas.
• Gravitational instabilities
• Aerodynamic collection of eroded debris
• Restructuring of fluffy aggregates
• Rapid grain growth in pressure bumps
Ciesla & Dullemond (2010)
A plausible growth mechanism >1m
• Corotating patch in midplane
• Weak MHD turbulence
• Density fluctuations
• ~1m-sized "bricks" concentrate in
pressure bumps
• Self-gravitation bounds clumps >10
orbital periods
• Local gravitational instabilities
• Mass concentrations ~10-4 MEarth
Brauer et al. (2008), Johansen et al. (2011)
III. Dust evolution in PPDs: Summary
• Dust growth ~0.1–2 Myr
• Coagulation, sedimentation, fragmentation, stirring, inward drift
• Inward drift: ~103–104 years for 1m-particles @ 100 AU
• 1m-barrier for growth
• Self-gravitation & growth in pressure maxima
IV. Dust in protoplanetary disks and the
early Solar nebula
Observational techniques
• IR features
• Spectral Energy Distribution (SED)
• Emission at mm–cm wavelengths
• Resolved IR-mm images
• Surveys of PPDs of various ages
• Composition of meteorites
• Composition of cometary/IDP grains
• Isotopic dating
• Condensation sequence
Spectral Energy Distribution
Energetic domain
Disk
Star
Wien domain
Dullemond et al. (2007)
Rayleigh−Jeans
domain
follows fro
an SED slo
because m
emitted at
strong emi
flux is wea
at large rad
slopes typi
mann, 199
Ae/Be star
larger spre
determinat
somewhat
divide the
those with
ing slope s
flux (called
but the mo
clearly inc
SEDs: Signatures of dust evolution
Dullemond et al. (2007)
Face-on PPDs: "warm" dust in emission
L. Testi, S. Leurini / New Astronomy Reviews 52 (2008) 105–116
L. Testi, S. Leurini / New Astronomy Reviews 52 (2008) 105–116
PAHs & diamonds
107
Crystalline silicates
107
CONICA VLT observations of the 3.6 lm ‘‘diamonds” and 3.3 lm PAH
the adjacent continuum in the HD 97048 intermediate-mass system
., 2004, 2006). The upper panel shows the intensity profile of the
ubtracted diamond feature as a thick line histogram, the adjacent cofile as a thin histogram and the profile of an unresolved star as dashed
mond emission is clearly resolved. In the bottom panel the PAH and
ofiles are compared with disk model computations for the PAH line
and the continuum (dotted). At the distance of this object, the horiorrespond to ±100 AU from the star.
sphere is the natural explanation for the emission obthis feature in a large variety of circumstellar disks
d Goldreich, 1997). In recent years with high quality
ed spectra becoming available first with ISO and more
th Spitzer, it has been possible to attempt to understand
ty of the profiles observed in various regions.
wed in Natta et al. (2007), the modeling of silicate proks as well as comparison with laboratory measurements
Fig. 3. VLTI/MIDI observations of the silicate profile in the three HAeBe systems
HD163296, HD144432, and HD142527. In the top panel a flared disk is sketched, in
the bottom panels the MIDI observations of the inner disks are compared with the
emission from the outer disk derived by subtracting the interferometric spectrum
from single telescope spectra (adapted from van Boekel et al., 2004).
can give indications on the degree of ‘‘crystallinity” and on the size
of the emitting particles (see Fig. 2).
The observed systems show a range of properties with grains
similar to those present in the diffuse interstellar medium to grains
Fig. 1. NAOS/CONICA VLT observations of the 3.6 lm ‘‘diamonds” and 3.3 lm PAH
features and the adjacent continuum in the HD 97048 intermediate-mass system
(Habart et al., 2004, 2006). The upper panel shows the intensity profile of the
continuum subtracted diamond feature as a thick line histogram, the adjacent continuum profile as a thin histogram and the profile of an unresolved star as dashed
line, the diamond emission is clearly resolved. In the bottom panel the PAH and
continuum profiles are compared with disk model computations for the PAH line
(dot-dashed) and the continuum (dotted). At the distance of this object, the horizontal scale correspond to ±100 AU from the star.
AOS/CONICA VLT observations of the 3.6 lm ‘‘diamonds” and 3.3 lm PAH
and the adjacent continuum in the HD 97048 intermediate-mass system
Fig. 3. VLTI/MIDI observations of the silicate profile in the three HAeBe systems
t al., 2004, 2006). The upper panel shows the intensity profile of the HD163296, HD144432, and HD142527. In the top panel a flared disk is sketched, in
with atmosphere is the natural explanation for the emission obm subtracted diamond feature as a thick line histogram, the adjacent co- the bottom panels the MIDI observations of the inner disks are compared with the
served in this feature in a large variety of circumstellar disks
profile as a thin histogram and the profile of an unresolved star as dashed emission from the outer disk derived by subtracting the interferometric spectrum
(Chiang and Goldreich, 1997). In recent years with high quality
from single telescope spectra (adapted from van Boekel et al., 2004).
diamond emission is clearly resolved. In the bottom panel the PAH and
mid-infrared spectra becoming available first with ISO and more
m profiles
are compared
with
diskbeen
model
computations
the PAH line can give indications on the degree of ‘‘crystallinity” and on the size
recently
with Spitzer,
it has
possible
to attemptfor
to understand
ed) andthe
thediversity
continuum
(dotted).
the distance
of this
object, the hori- of the emitting particles (see Fig. 2).
of the
profiles At
observed
in various
regions.
ale correspond
to ±100in
AU
frometthe
star. the modeling of silicate proThe observed systems show a range of properties with grains
As reviewed
Natta
al. (2007),
similar
those present in the diffuse interstellar medium to grains
files in disks
asprofiles
wellin HAebe,
as comparison
with
measurements
mpilation of observations
of silicates
TTS and BD systems
and inlaboratory
the laboratory (adapted
from and references therein
Natta et al.,to
2007).
Testi & Leurini (2008)
Fig. 3. VLTI/MIDI observations of the silicate profile in the three HAeBe systems
L 0 . HK Tau B and HV Tau C exhibited scattered light in
n disk, and two components of the scattered light from
are clearly resolved in all the bands. The FWHM image
oth objects are found to be 0.1300 at H, 0.1000 at K, and
in the left panel of Figure 2. The spectra is scaled to the photometric points at K and L 0 . All of the spectra show a deep water ice
absorption at 3 !m. In addition, many emission and absorption
lines can be seen in the spectra of HV Tau C. Since HV Tau C is
Edge-on PPDs: "cold" dust in absorption
Water ice bands at 3μm
306
TERADA ET AL.
Vo
Fig. 3.—Schematic view of geometry for the water ice absorption. The scattered light regions are represented by the gray elliptical areas. The midplane of th
disk is represented by the black elliptical areas. Since the midplane disk is optically thick in the 3 !m band, we can observe only the scattered light through the
above and below the midplane disk.
well known to have a strong outflow activity (Magazzù & Martin
1994), the most prominent emission lines are due to molecular
hydrogen lines excited by the jet. The study of these lines with
0.100 spatial resolution will be discussed in the second paper
( H. Terada et al. 2007, in preparation).
The right panel of Figure 2 gives the optical depth of the 3 !m
water ice absorption for HK Tau B and HV Tau C. The continuum for the spectrum was estimated by fitting the photometric
points at H, K, and L 0 with parameters of a single-temperature
blackbody (T ) and visual extinction (AV ). The fitting results for
HK Tau B were T ¼ 2500 K and AV ¼ 0 to obtain the 3 !m
optical depth of 1.32. This AV does not have any physical meaning; its value was determined only to get a good fit to the continuum. The center of the optical depth profile is at 3.055 !m.
The optical depths of the water ice absorption in two epochs
for HV Tau C were found to be 1.65 for the first epoch (2002
September 25) and 1.09 for the second epoch (2005 January 18)
by assuming a continuum with T ¼ 2200 K and AV ¼ 3:5. The
center wavelength of the optical depth was estimated to be
3.017 !m in the first epoch and 3.026 !m in the second epoch.
35.2 mag for HK Tau B, and 44.1 mag in the first epoc
29.0 mag in the second epoch for HV Tau C.
Although Whittet et al. (1988) found a shallow water i
sorption with an optical depth of only 0.20 toward the pr
star HK Tau A, it is much smaller than the optical depth o
for HK Tau B. The visual extinction (AV ) of 1.91 for the pr
system HV Tau A, HV Tau B of HV Tau C (Woitas & L
1998) is too small to produce any water ice absorption. Th
suming a distance of 140 pc, the detected water ice is loc
within a region of 340 AU (2.400 ) from HK Tau B and 560 AU
from HV Tau C. This excludes the possibility of water i
sorption by the foreground cloud.
The NH corresponding to AV can be estimated using th
tionship of NH ¼ ð1:59 ; 1021 Þ ; AV cm"2 obtained in
Opiuchus dark cloud ( Imanishi et al. 2001). The results of
HK Tau B and HV Tau C in the first and second epochs are
1022 , 7:01 ; 1022 , and 4:61 ; 1022 cm"2. If we assume a sph
4. DISCUSSION
4.1. Origin of Ice Absorption
Stapelfeldt et al. (1998, 2003) showed that the scattered light
from the central star originated within a radius of 105 AU from
HK Tau B and 50 AU from HV Tau C with a scale height of 3.8
and 6.5 AU, respectively, at a distance of 50 AU from the star.
Thus, it is likely that the material responsible for the ice absorption is outside the scattered light region, as shown schematically
in Figure 3.
D’Hendecourt & Allamandola (1986) found N ( H2 O) ¼ "H2 O !# /
ð2:0 ; 10"16 Þ, where !# is the FWHM of the feature in cm"1.
For HK Tau B and HV Tau C in the first and second epochs, we
find !# ¼ 502, 509, and 494 cm"1, respectively. Although it was
difficult to estimate the FWHM width of the water ice absorption due to a lack of sufficient data points at the shorter wavelength, we assumed it to be twice as wide as the width between
the center (3.02–3.06 !m) and the half-maximum point at the
longer wavelength (3.25–3.29 !m). The column densities of
H2O for HK Tau B and HV Tau C in the first and second epochs
were found to be 3:31 ; 1018 , 4:20 ; 1018 , and 2:69 ; 1018 cm"2.
Teixeira & Emerson (1999) found that N ( H2 O) ¼ (AV " 2:1) ;
1017 cm"2, from which the visual extinction AV is derived to be
Left: The 1.9–4.2 !m spectra of HK Tau B and HV Tau C. A deep water ice absorption at 3 !m is seen in these spectra. The dotted lines indicate the continua of the
optical depth of the water ice absorption is shown in the right panel. In HV Tau C, the prominent H2 emission lines can be seen at 1.9–2.5, 3.004, and 3.235 !m.
Terada et al. (2007)
Fig. 4.—Comparison of the water ice optical depth. The solid line sh
optical depth of the water ice for Elias 3-16 (Smith et al. 1989), which is
to be a field star behind the Taurus molecular cloud.
IR dust features:
• Vibrational bands (stretching, bending): 3–100μm
• Water ice: 3μm
• PAHs: 3.3, 6.2, 7.7-7.9, 8.6, 11.3 and 12.7μm, etc.
• (Nano)diamonds: 3.43, 3.53μm
• (ISM) Hydrogenated amorphous carbon: 3.4, 6.8, 7.2μm
• Amorphous silicates: 9.8, 18μm
• Crystalline silicates: 10.2, 11.4, 16.5, 19.8, 23.8, 27.9, 33.7, 69μm
• Iron sulfides: ~23μm
• Featureless: Fe, FeO, organics, ...
•
Crystalline silicates requires radial transport
Natta et al. (2006), van den Acker et al. (2004)
Disk masses vs Age
55
Protoplanetary Disks
~5–7 Myr
~2–4 Myr
~1–2 Myr
• M decreases with age
• <M > ~1% of M
• Dust opacities at mm are
Figure 7: Evolution of the disk masses across the empirical sequence defined
by the slope of the infrared
disk SED. The circles show the median disk mass, as
measured at sub-millimeter wavelengths for a sample of 300 YSOs in Taurus
and Ophiuchus. Error bars show the distribution quartiles. Transition disks are
defined here as those YSOs that lack infrared excesses for wavelengths less than
disksub-millimeter emission.Sun
25 µm, but have detectable
With this definition, no
Class III YSO was detected and the stringent limit to their median mass comes
from stacking the non-detections together (Andrews & Williams 2007a).
poorly constrained!
Williams & Cieza (2011)
Mdisk vs Mstar: no correlation
Williams & Cieza (2011)
Large grains in inner PPDs
6
Protoplanetary Disks
•
Silicate band' profiles: up to ~3μm
grains
•
IR SEDs: small dust gaps
•
~10% of 1–3 Myr disks
•
Increase with age
•
Resolved inner dust holes
•
Gas is still around, accretion still
goes on
Williams & Cieza (2011)
Large grains in outer PPDs
Guilloteau et al.: Dual frequency mm imaging of proto-planetary disks
• Dust emissivity β: 0–2
Source
M (10 M ) M (10 M )
Ratio
No
correlation
with
stellar
properties
•
CI Tau
43 ± 4
51 ± 5
1.18± 0.18
18 ± 1
46 ± 2
2.52± 0.09
CY Tau
with
DG•TauNo correlation
36 ± 5
42
± 6 age1.18± 0.27
151 ± 59
179 ± 60
1.19± 0.73
DG Tau-b
with
in0.04
inner disk
51 ± 1
60
± 1 dust
1.18±
DL•TauNo correlation
DM Tau
32 ± 8
192 ± 49
6.07± 0.52
24 ± 1 thick emission?
32 ± 2
1.33± 0.14
UZ•TauOptically
M is the disk mass for Model 2 (tapered edge) for κ(1.3mm) =
2 cm .g ,Free-free
while M is foremission
κ(1.3mm) as in Fig.9. A gas-to-dust ratio
•
of 100 was assumed. Ratio = M /M .
Table 14. Disk masses with variable dust properties
−3
c
c
2
−1
−3
v
"
"
v
c
v
Fig. 13. Characteristic radius Rc (in AU) as a function of estimated stellar ages (in Log10 of 106 years).
Fig. 12. Surface densities of observed sources. Thick lines are
et in
al. which
(2003),
Natta et
(2004),
forTesti
sources
a variation
of al.
β and
thus κRodmann
with radius et
al. (2005), Guilloteau et al. (2011)
Grain evolution in the Solar Nebula
Planet formation and protoplanetary dust
Heterogeneous textures:
chondrules & CAIs in
matrix
32
5
H.-P. Gail and P. Hoppe
Fractal shapes of IDPs
The origins of protoplanetary dust and theIDP
formation
(NASA)of a
High-T condensates in
silicates
Figure 2.1 An IDP composed of very small <0.2 µm subunits. The IDPs show t
fluffy structure of particles that is expected for dust agglomerates in protoplaneta
accretion disks. The IDPs are suspected to be composed partly from particles
interstellar origin. A small fraction of the subunits can be identified as preso
dust grains. (Courtesy of NASA, http://stardust.jpl.nasa.gov/science/sci2.html.
Sometimes, dust models have been considered that contain coated particle
silicate grains with a mantle of carbonaceous material. Such models can pro
be excluded (Zubko et al. 2004); the best fits to observations are obtained fo
grains.
A potential source of information on the nature and composition of inter
dust grains comes from IDPs, which are collected in the Earth’s stratospher
Bradley 2003; Alexander et al. 2007). The IDPs with bulk compositions in the
into two broad categories, compact hy
Apai & Lauretta (2010)of chondritic meteorites can be dividedGail
& Hoppe (2010)
particles and porous anhydrous particles. The latter are aggregates of sub
meter building blocks (see Fig. 2.1). They represent a disequilibrium assem
of minerals and amorphous silicates and it has long been suspected that so
Figure 1.2 Components of chondritic meteorites: fluffy CAI (upper left), compact
them are dust particles released during the flyby of comets at the Sun. If true
CAI (upper right), chondrule (lower left), and matrix (lower
right).
Figure
2.2 A GEMS
particle found in an IDP, consisting of a
grains are particles from the formation time of the Solar System that have su
silicate-like composition
embedded
nanometer-sized
inclu
for 4.6 Gyr insideand
extremely
cold (approximately
20 K) cometary bodies.
Alt
Grain evolution in the Solar Nebula
• Most primitive are chondritic meteorites: matrix + chondrules + CAIs
• No signs of alteration & metamorphism (Т<50 K)
• Calcium-Aluminum-rich Inclusions (CAIs), 4.571 Gyr:
• 0.1mm–1cm: mixture of highest-T condensates
• Multiple high-T events
• Chondrules (molten droplets): ~0–2 Myr after CAIs
• Single melting-condensation event
• Peak temperatures: >1800K
• Initial cooling: ~hours (till ~1000K)
• Chondrule densities: >10-5 cm-3
• Formation in "clumps": >1000 km
Courtesy of S. Desch
Presolar grains in meteorites
48 the formation of accretion disksH.-P. Gail
of protoplanetary dust and
45 and P. Hoppe
105
Presolar SiC
AGB stars
104
Mainstream
Type A&B
Type Y&Z
Type X
Nova
Presolar oxides and silicates
in primitive Solar System matter
10–2
Group 1
103
solar
102
17O/16O
Group 2
Group 4
10–3
solar
Group 3
10– 4
10
100
10
102
12C/13C
15M Type II SN
(HeN + HeC zones)
solar
1–1.35 M Novae
100
solar
14N/15N
10–1
10–5
103
104
10–5
10–4
10–3
18
Oxides
Silicates
RGB/AGB/RSG,
PN
10–2
10–1
O/16O
shock passageof&presolar
heating SiC
in the
early Solar Nebula
• Survived compositions
on- and nitrogen-isotopic
grains.
Figure 2.8 Oxygen-isotopic
compositions of presolar oxide and
stellar models are
shownhappened
for comparison.
Solar
Data for
RGB/AGB
stars, redAGB
supergiants (RSG), and planetary n
to ices
& gas?metallicity
• What
ett et al. (2003), Type II SN: shown
Rauscher
al. (2002), novae:
Joséellipses indicate the four isotope
foretcomparison.
The gray
data sources see Lodders & Amari
(2005);
Zinner
(2007).
Notesources see Lodders & Amari (2
for
presolar
O-rich
dust.
For
data
Gail
& Hoppe (2010)
4 N/
15 N
ratio the value inferred for Jupiter’s atmosphere is shown.
Dust composition
• Water & CO2 ices (40%)
• Volatile & refractory organics (30%): "CHON"
• Iron-poor amorphous & crystalline silicates (27%)
• Troilite FeS (~2%)
Chemical and isotopic evolution of the solar nebula and protoplanetary disks
• High-T condensates: Fe, SiC, TiO, Al2O3 (<1%)
Mg2SiO4
Forsterite
Table 4.3 Condensation sequence for the early solar nebula
Component
Elements
Condensation temperature∗
Refractory
Al-, Ca-, Ti-oxides, trace
(Zr, V, Wo, etc.)
1850–1400 K
Main
Mg-rich silicates,
metallic Fe, Ni, Co,
Li, Cr
1350–1250 K
Sulfides, silicates, metals
(Mn, P, Na, Rb, K, F, Zn,
Au, Cu, Ag, etc.)
1250–640 K
Volatile
Highly volatile
MgSiO3
113
Enstatite
(Mg,Fe)Si2O6
(Mg,Fe)Si2O4
C, H, O, N, Cl, Ne, Ar,
Xe, Kr, etc.
Olivine
< 640 K
Al2O3
∗ Condensation temperatures are given for the pressure of 10−4 bar.
Pollack et al. (1994), Semenov et al. (2010)
sulfur is condensed in the form of troilite (FeS). At even lower temperatures other
Corundum
IV. Dust in PPDs & Solar Nebula: Summary
• Disk masses evolve
• Mdisk: no correlation with M*, L*
• Grain sizes evolve
• Gaps & inner holes
• Crystalline silicates in outer disks: transport or shocks?
• Mineralogical composition of meteorites
• ~2 Myr formation of solids in the early Solar System
• Presolar materials in meteorites
VI. The Brave New World: ALMA
• Atacama Large Millimeter Array (2013)
• 50 x 12m + 12x7m + 4x12m
• Spatial resolution: 0.005″
• Spectral resolution: <0.05 km/s
• 8 GHz bandwidth for continuum
• 86 – 950 GHz (250 µm – 1 mm)
• x100 resolution
• x20 sensitivity
Credit: ALMA (ESO)
ALMA imaging of dust in PPDs:
• Dust emissivities: 90GHz–1THz
• Observed fluxes at 1.3mm: 30–1000mJy => 1h to get 0.03mJy @ 0.05"
• Large sample of PPDs: >100
• "Debris" disks
• Spatially resolved images: >0.01"@1THz (~1 AU @ 100 pc)
• Radial gradients of grain sizes
• Total Mdisk
• Transient, small-scale features
• Gaps & inner holes
• Polarization
ALMA imaging of dust in "debris" disks:
Wolf et al. (2002)
Wyatt (2004)
Science Verification observations of β Pictoris at 345 GHz
ALMA imaging of dust in PPDs:
Wolf et al. (2002)
Disk gap at 5 AU opened by 5 MJy planet
Wyatt (2004)
Dust resonances in a "debris" disk