Herschel observations
of cold H2O and its OPR
in planet-forming disks
Magnus Persson (Leiden Observatory)
Michiel Hogerheijde (Leiden), Edwin A. Bergin, Christian Brinch, L. Ilsedore
Cleeves, Jeffrey K. Fogel, Geoffrey A. Blake, José Cernicharo, Carsten Dominik,
Dariusz C. Lis, Gary Melnick, David Neufeld, Olja Panić, John C. Pearson, Lars
Kristensen, Umut A. Yıldız, Ewine F. van Dishoeck, Kees Dullemond
What is the origin of water on Earth?
•
•
•
In the early Solar System
• water vapor in the inner Solar System
(T>100 K)
• condensed as ice on dust grains outside
the snow line at ~3 AU (Hayashi et al.
1981; Abe et al. 2000)
Comets and asteroids may have delivered
large amounts of water from beyond the
snow line to the early Earth (Matsui & Abe
1986; Morbidelli et al. 2000; Raymond et al.
2004)
How large is the ice reservoir?
• 1 ‘Earth Ocean’ = 1.5×1024 g of water
H
~f
rac
tio
n×
Equilibrium between photodesorption
and -dissociation in outer disk
(Dominik et al. 2005):
H2Ogas ~fraction×H2Oice
Evaporation in inner disk (<3 AU)
304
TERADA ET AL.
Freeze out in outer disk (> 3 AU)
H
/H
O
O
≫
Mode
1
Band
Standard Star
H
ic
O
2
Vol. 667
Exposure Time
(s)
Slit Width
(arcsec)
00
Spectral Resolution
(k/!k)
0
1A
?
0
0
00
0
00
0
U
A
1
0
0
0
00
0.
00
00
U Fig.
Subaru detection
of 3µm
water
2. Comparison
of ice
A
1
the observed spectrum of
absorption (Terada
al.combined
2007)
AA Tauriet
to the
2O
ice
×H
~f
ra
cti
on
2O
ga
s
H
ice
O
H2
suggest that substantial molecular synthesis take
place 305
within the disk. The observed region of th
No. 1, 2007
HK Tau B ..................
2001 Dec 25
Spectroscopy
L
HR 1590
1900
0.3
390
2002 Sep 24
Spectroscopy
K
HR 1590
720
0.15
1300
TABLEabundances
2
disk
The
of135HCN0.12.and
C2H
in AA
Tauri,
wavefront
correction
with the adaptive
opticslies
sys- well inside the radius where ices woul
2002 Sep 24
Imaging
H
HD 22686
. . at L after
. .2
.
Magnitudes of HK Tau B and HV Tau C
reached the diffraction limit.
tem.
2002 Sep 24
Imaging
K
HD 22686
67.5
. . . Only the L .images
..
sublimate,
based on the gas temperatures and in
relative
toHDH
higher
than
photometry
was performed
on HK Tau
B and HV
22686
.Aperture
..
...
2002 Sep 24
Imaging
L
2O, are225also substantially
Date
Tau
HV Tau C ..................
2002 Sep 25
Spectroscopy
K
HR 1590
720
0.3 C with a radius
650 of 1.0 and 1.2 , respectively. The photoObject
(UT)
H
K
L
This situation is similar to that of ho
measurements
for1900cometary
volatiles
(fig.2. While
S1),
2002 Sep 25
L
HR 1590
0.3 results
metric
are390
shown in Table
our resultsferred
for HV Tauradii.
C
e 25 Spectroscopy
Sep
Imaging
HD 4461211.54
135
. . . no significant
. . . time variation at the two epochs at K and L ,
show
HK Tau2002
B ....................
2002 Sep 24 H
12.25
11.02
c
i
Sep 25
Imaging
HD
4461212.32
67.5
. .of
. objects
cores, where a warm gas phase follow
whereas
the
abundance
CO
consistent
both
exhibit
variability
of up to 1 mag inmolecular
all the bands
HV Tau2002
C ....................
2002
Sep 25 K
13.19
11.37
2.. .. .. is
Sep 25
Imaging
225
...
2005 Jan 18 L
. HD
. . 4461212.31
11.40
compared
to previous
photometry ( Koresko 1998; Woitas &
2 2002
2005 Jan 18
Spectroscopy
K
HR 1152
480 than
0.3
650 & Bouvierin
the
sublimation
of molecules from icy grain man
with or
possibly
less
its 1998;
abundances
both
Leinert
Monin
2000;
White & Ghez
2001).
The
2005 Jan
18 1 Spectroscopy
HR 1152
0.3
390
Note.—The
typical
" uncertainties areL0.02, 0.02,
and 0.1 mag for7680
H, K,
possible
causes
of
the
variability
are
magnetic
activity
on
the
stels and L , respectively.
2005 Jan 18
Imaging
K
FS 122
40
...
...
tles.
The subsequent gas-phase chemistry (19) ma
comets
and
interstellar
ices.
lar. .surface,
variable
mass accretion rate onto the central
star, bright2005 Jan 18
Imaging
L
HD 44612
150
.
...
ga
ness change of scattered light due to geometry change of the
absorptions of the A0 stars were removed
with a method developed
produce
If molecular
cores are
representative
of the
materials,
and extinction change toward
the scattered
light source. the complex organic molecules observe
2
by Vacca et al. (2003). The wavelength calibration was performed
The width
offor
theflat-fielding,
dark lane of HV
Standard data reduction was
applied
sky Tau
sub-C was estimated to be 0.269
was conducted using 9 point dithering with a minimum separation
using the telluric absorption lines in the spectrum.of material at
with
millimeter-wave spectroscopy in both mas
that
is theincorporated
into
H, using
0.258
at K,
andpackages.
0.246 at The
L as
measured by
the distance
an ABBA
traction, and telluric correction
IRAF
of each position of 1.5 . Spectroscopy was made withcomposition
betweenare
thelisted
peakinofTable
the two
components
ice sequence
with 3.0 nodding along to the slit. The slit3.orientation
standard stars for telluric correction
1. HD
22686, of the scattered light from
RESULTS
sive ashot
cores and the corresponding region
disks,
abundances
for
AAatThe
Tauri
HV
TauasC.
The width
is wider
the
shorter wavelengths
exwas set to be orthogonal to the disk with a slit position
angle of then
HD the
44612,higher
and FS 122 were
used
photometric
standards.
Observing Date
(UT)
Object
×
n
o
i
t
c
1
a
r
f
≪
~
gas
ce
O
i
2
H
2O
H
s/
a
g
2O
H
H
ic
O
2
s/ degrees kelvin) regions of moa
(several hundred
g
O
lecular gas
surround
luminous
2WATERthat
ICE IN
PROTOPLANETARY
DISKS protostars.
U
TABLE 1
Observing Log
H2O
1
≫
e
H
s/
a
g
1
≪
e
×
n
o
i
ct
a
2O
r
f
H as ~
g
O
2
H
2O
ga
s
H
th
eo
ry
2O
ice
What we know about H2O in disks
Spitzer detection of hot water
vapor from inner disks (Carr &
Najita 2008; Salyk et al. 2008;
Pontoppidan et al. 2010).
00
0
00
!
1 shows
opticsand
images of
TauL 0Bmagnitudes
on UT
by the
extinction
at the shorter wavelengths.
for HV
Tau C.adaptive
The dithering
ofpected
HD 22686
andgreater
HD 44612
are 7.190,
130.5! for HK Tau B and 20.0Figure
H, HK
K, and
September
24 and
HVinside
Tau Cthe
on UT 2002
September
257.041,
at
We7.05
present
1.93–4.13
!m spectra
of HK Tau B and HV Tau C
nodding were achieved by2002
moving
the
tip-tilt
mirror
7.185,
7.199,
and
7.041,
(Elias
et
al.
1982;
Leggett
0
. HK
Tau Batand
HV K
Tau
C exhibited
light in
K,fields
and Lfor
in the
left panel of
2.isThe
spectra is scaled to the photoadaptive optics system. TheH,
flat
imaging
H and
and
et al.scattered
2003), respectively.
The
K magnitude
of Figure
FS 122
11.347
0
edge-on
disk,
and two
scatteredetlight
from For spectroscopy
. All of the
spectra show a deep water ice
metric points
K and
spectroscopy in the K and Lanbands
were
produced
bycomponents
combining of the(Leggett
al. 2006).
theatA0
typeL standard
stars
absorption
at 3 !m.
In
addition,
many emission and absorption
HV Tau
C are clearly
all the bands.
The
FWHM
image
images of an integrating sphere
illuminated
by resolved
a halogeninlamp.
HR
1590
(A0
V;
V
¼
5:789)
and
HR
1152
(A0
Vn;
V ¼ 6:430)
00
00
H, 0.10
at K,toand
lines
can beabsorption.
seen in theThe
spectra
of HV Tau C. Since HV Tau C is
sizes of both
The sky flat was used for imaging
at L 0 .objects are found to be 0.13 at
were
observed
correct for
telluric
hydrogen
model spectrum (11).
The observed continuumsubtracted spectrum from
13 to 16 mm (above) is
offset from the model
spectrum (below). The error bar indicates the average uncertainty in the
observed spectrum. The
model is a slab in local
thermodynamic equilibri-
See also Honda et al. (2009)
r
e
s
ob
s
n
o
i
t
a
v
Herschel/HIFI observations
•
Disks ≪ Herschel beam
• HIFI spectrally resolves lines: disk ID
• Beam dilution: long integrations
•
~200 hrs of observing time
• ortho-H2O 110-101, para-H2O 111-000
• o+p: TW Hya, HD100546, DM Tau,
AA Tau
• o: LkCa15, MWC480
• p: HD163296
sequence star TW Hydrae (TW Hya) using the
Heterodyne Instrument for the Far-Infrared (HIFI)
spectrometer (9) on board the Herschel Space
Observatory (10) (Fig. 1) (11, 12). TW Hya is
a 0.6 M⊙ (solar mass), 10-million-year-old T
Tauri star (13) 53.7 T 6.2 pc away from Earth. Its
196-AU-radius disk is the closest protoplanetary
disk to Earth with strong gas emission lines. The
disk’s mass is estimated at 2 × 10−4 to 6 × 10−4 M⊙
in dust and, using different tracers and assumptions, between 4 × 10−5 and 0.06 M⊙ in gas (14–16).
The velocity widths of the H2O lines (0.96 to
1.17 km s−1) (table S1) exceed by ∼40% those of
cold CO (14). These correspond to CO emission
Downloaded from www.sciencemag.org on Dece
Detections and non-detections
•
•
•
Detections: TW Hya, HD100546
Upper limits: AA Tau, DM Tau, LkCa15, HD163296, MWC 480
Lines and upper limits lower than models predict
Fig. 1. Spectra of para-H2O
111-000 (A) and ortho-H2O 110101 (B) obtained with HIFI on
the Herschel Space Observatory
toward the protoplanetary disk
around TW Hya after subtraction
of the continuum emission. The
vertical dotted lines show the
system’s velocity of +2.8 km s−1
relative to the Sun’s local environment (local standard of
rest).
HD100546
ionwater
from
main-
9513,
, Union of
Techtitute
Am, and
a, CA
hysics,
d As1218,
, GerTechut für
Tmb (K)
ht to
rains
AU)
ed a
the
or is
0 K)
hase
ssion
from
the
out,
in a
raviyers
cules
cold
adial
water
n ice
results in an equilibrium column of water H2O
vapor throughout the disk (Fig. 2). Consistent
with other studies (19), we find a layer of maximum water vapor abundance of 0.5 × 10−7 to
2 × 10−7 relative to H2 at an intermediate height in
the disk. Above this layer, water is photodissociated; below it, little photodesorption occurs
and water is frozen out, with an ice abundance, set
by available oxygen, of 10−4 relative to H2.
In our model, the 100- to 196-AU region
dominates the line emission, which exceeds observations in strength by factors of 5.3 T 0.2 for
H2O 110-101 and 3.3 T 0.2 for H2O 111-000. A
lower gas mass does not decrease the line intensities, if we assume that the water ice, from
mail:
CTOBER 2011
VOL 334
SCIENCE
www.sciencemag.org
Bergin et al. (2010); Hogerheijde et al. (2011, and in prep)
Disk origin of the H2O emission
•
Confirmed by line width, line profile, and VLSR
SMA: CO
Andrews et al. (2011)
Thi et al. 2004
ALMA Observations of HD 100546
3
Fedele et al. (2013)
Walsh et al. (subm)
Fig. 1.— First-moment map of CO J = 3-2 emission (colour map) overlaid with integrated intensity contours (in white) and 870 µm
continuum emission contours (in black). The intensity contours correspond to 3, 10, 30, 100, and 300 times the calculated rms
(30 mJy beam−1 km s−1 ) and the continuum contours correspond to 3, 10, 30, 100, 300, and 1000 times the rms (0.5 mJy beam−1 ).
The CO integrated intensity reaches 5% of its peak value at ≈ 20 × the rms, whereas the continuum emission reaches 5% at ≈ 60 × the
rms. The synthesized beam (shown in the bottom left-hand corner) is the same for both the continuum and CO observations.
Modeling approach
Disk mass, size, density distribution:
best possible description from literature
beam-convolved
emission lines
Self-consistent temperature distribution
Propagation of stellar UV, X-rays; CR
non-LTE excitation and
radiative transfer
Gas-phase chemistry incl. H2+OH & ion/molecule
Photodissociation, and H2 and CO self-shielding
X-ray and CR ionization
Distribution of gasphase H2O
Adsorption and (photo)desorption on grains
photodesorption rates from Öberg et al. (2007)
Refs: Fogel et al. 2011; Hogerheijde et al. 2011; Cleeves et al. 2013
TW Hya
•
•
Rdisk=196 AU; i=7˚: nearly face-on
Mdisk=2–6×10-4 M⊙ in dust, >0.05 M⊙ in gas (Bergin et al. 2013)
•
•
•
6300 Earth oceans of water ice
See also Calvet et al. 2002; Thi et al. 2010; Hughes et al. 2011)
➝ 0.04 Earth oceans of water vapor, lines 3–5× brighter than observed
Andrews et al. (2011)
Differential settling of icy grains
Large & small grains well mixed
Large grains settle w.r.t. to small grains
UV photons
•
•
•
•
Small grain
Photodesorption
Large grain with ice cover
Water vapor layer
Large grain with ice cover
Water vapor layer
Photodesorption
UV photons
Small grain
Remove 88% of ice from UV-affected layers
Settling of larger, icy grains relative to the small grains which dominate the
UV absorption
Only 12% of ice remains in upper disk
• Gives rise to 0.005 Earth Oceans of water vapor
Underlying ice reservoir unchanged: > thousands of Earth Oceans
• key assumption: elemental oxygen efficiently forms water on grains
• if correct: lines are optically thin, OPR~0.77
Alternative: Radial drift in TW Hya’s disk
CO
•
mm-sized grains have drifted in (R<60 AU) w.r.t. the µm-sized grains and
CO-gas (R~200 AU; Andrews et al. 2011; Debes et al. 2013)
• Expected because mm-sized grains feel a head-wind of the gas that moves
Fig. 1.— (a) Naturally
composite image of the 870 µm continuum emission from the TW
slightlyweighted
sub-Keplerian
Hya disk. Contours start at 10 mJy beam−1 (5 σ) and increase in 30 mJy beam−1 (15 σ) increments.
H2O dimensions
ice located
only where the mm-sized grains are, smaller source
• If beam
The synthesized
are shown in the lower left. (b) Azimuthally averaged 870 µm
size profile as a function of the deprojected baseline length (real part only; the
continuum visibility
imaginary terms
effectively
zerostrength:
on all baselines).
The observations
uncertainties arewithout
typically smaller
than
line
matches
need to
reduce
• arereduces
the symbol sizes. amount
Note the low-amplitude
oscillations
beyond ∼150 kλ. (c) Complete SED for
of ice subject
to UV-desorption
the TW Hya star+disk system (references are in the text; see §2). The Spitzer IRS spectrum is
increases
the adopted
line width
slightly
(~face
onisdisk):
consistent
with data
• gray
shown as a thick
curve. Our
stellar
photosphere
model
overlaid
as a thin gray
curve (see §4.1).
CO J=3−2
channel
maps
fromthick:
the TW
Hya disk,
the
angular
scale
Fig.
5.—
Moment
maps
ofonthe
CO
J=3−2
emission
the lines
optically
information
on same
OPR
lost
• (d)makes
fro
as the continuum map in panel a. Contours are drawn at 0.4 Jy beam−1 intervals (∼3 σ) in each
structure
models
compiled
in The
Table
The leftmost pan
0.2 km s−1 -wide channel. The
1"" synthesized
beam is shown
in the lower left.
central2.
channel
−1
HD100546
log n(H2O/cm-3)
•
•
•
•
Same modeling philosophy
Disk structure from Bruderer et al. (2012)
Mdisk=0.07 Msun
d=97 pc; M*=2.5 Msun; i=42˚; Rout=400 AU
HD100546
•
➝ With ice extended across the full 400 AU of the disk,
predicted emission lines are again much too strong
HD100546: two-ring structure
•
Recent ALMA data show that the mm-sized grains in the
HD100546 disk are distributed in two rings (Walsh et al.)
• most of them in a narrow ring around 26 AU
• a small fraction in an extended ring 151-226 AU
• shaped by radial drift + a close-in companion + an
ALMA Observations
100546
embedded protoplanet
at ~70of HD
AU
(Quanz et al. 2013) 3
CO gas
mm continuum
Fig. 1.— First-moment map of CO J = 3-2 emission (colour map) overlaid with integrated intensity contours (in white) and 870 µm
HD100546
ice: 151-226
AU only
•
•
•
inner ring too warm for water ice to exist
line profile and strength well matched by water ice residing in (and around)
151-226 AU ring
lines optically thin: OPR~2-3 reproduces ortho/para line strengths
Nature of the weak H2O lines
•
Upper limits to AA Tau, HD163296, LkCa15, MWC 480, and
DM Tau
• consistent with weak lines to TW Hya and HD100546,
taking into account larger distances, disk sizes
•
Water vapor emission lines from disks are weak
• because of the radial drift of the mm-grains
• a small size of photodesorbed water ice
• not clear why ices are preferentially located on mm-sized
grains
• ...can icy grains grow more easily to mm-sizes?
• ...do (more settled) mm-sized grains better preserve ice
reservoir?
Summary
•
•
•
•
•
Herschel detects cold water vapor in the disks from TW
Hya and HD100546
The H2O is likely generated by photodesorption from ices
• co-located with the mm-sized grains that have
undergone significant radial drift
Upper limits to other sources consistent with larger
distances
• or: do radial enhancements of mm-sized grains in TW
Hya and HD100546 protect the water ice?
TW Hya: lines are optically thick; no OPR information
HD100546: lines are optically thin; OPR~2-3
• Both consistent with Solar System comets