1. Sub-mm Interferometry of Protoplanetary Disks (aka “Heterodyne Arrays 101”)
1’’ ?
HD HST ACS
ALMA
Geoffrey A. Blake, Caltech
Ringberg Disks Meeting
16Apr2004

2. Talk Outline: How do heterodyne arrays work?
- Strengths versus weaknesses.
What can we do now?
- Dust versus lines.
III. Will things improve soon?
- SMA and CARMA.
Disks in the ALMA era.
- What will be needed for ALMA to exploit
its full potential?
16Apr04

3. Spectroscopy of “Disk Atmospheres”
G.J. van
Zadelhoff
2002, Ph.D.
thesis K
Chiang &
Goldreich
1997
IR disk surface within several – several tens of AU
(sub)mm disk surface at large radii, disk interior
16Apr04

4. Arrays everywhere! Typically ntel ≤ 6-10. PdBI VLA BIMA SMA ATCA OVRO
16Apr04

5. The 1-Baseline Heterodyne Interferometer:
HST resolution at 1mm D=10 km! Use array.
Can’t directly process 100 – 1000 GHz signals.
Heterodyne receivers detect |V| and f, noise defined by the quantum limit of hn/k.
Positional information is carried by the PHASE.
Spectral coverage depends on the receivers, while the kinematic resolution is determined by the correlator.
Geometrical delay
16Apr04

6. The 1-Baseline Heterodyne Interferometer:
B varies due to earth rotation.
The convolving internal is 2D, and defines the minimum (u,v) “cell size.” Note the minimum baseline is at least D !
Single baseline noise:
in Jy (A = aperture, J = Jy/K for single telescope, hQ = correlator efficiency, Dn = frequency interval, t = on-source time).
16Apr04

7. The n-Element Heterodyne Interferometer:
n(n-1)/2 baselines, imaging performance depends on the array geometry, but
For small to moderate n, the (u,v) plane is sparsely filled.
For a given array, the minimum detectable temperature varies as (resolution = qS)-2 :
qP = primary telescope beam
16Apr04

8. n-Element Imaging: is linear, is not. So, error analysis is
(1)
F.T., convolve
(“Dirty” image)
CLEAN or MEM
deconvolution
(2)
Dirty
Beam
NGC 1333 IRAS2
is linear,
is not.
So, error analysis is
difficult in the image plane!
OVRO
BIMA
Jørgensen et al. (2004)
16Apr04

9. Sub-mm Arrays & Star Formation:
outflow
x1000 in scale
infall
Cloud collapse
Rotating disk
Planet formation
Mature solar system
Sufficiently large sub-mm arrays can examine all phases and radii.
Adapted from McCaughrean

10. “Tailored Imaging:” By varying the (u,v)
coverage, or weighting, the effective resolution (and flux recovery) can be adjusted for large objects. Disks in true protostars?

11. OVRO CO(2-1) Survey of T Tauri stars
(Koerner & Sargent 2003)
stellar ages Myrs
stellar masses ~ 1 M
selection by 1 mm flux, SED characteristics
Taurus 19/19 detections
Ophiuchus 4/6 detections
resolution ~ 2”
20 objects
radii  150 AU
masses  M
(from SEDs)
See also Dutrey, Guilloteau,
& Simon, Ohashi

12. Disk Ionization Structure: CO and Ions
Disk properties vary widely
with radius, height; and depend on accretion rate,
etc. (Aikawa et al. 2002, w/
D’Alessio et al. disk models).
Qi et al. 2004
Currently sensitive only to R>80 AU
in gas tracers, R<80 AU dust.
CO clearly optically thick, other species
likely to be as well.
The surface fractional ionization is >10-8,
easily sufficient for MRI transport.
Disk ages are difficult to determine, but chemistry may be useful in this respect. This figure is from a model of the sulfur chemistry in cores at ~50K. This may be similar to the outer regions of disks studied with OVRO. The dominant species migrates from H2S to SO,SO2 to CS over time. Here the timescale represents the time since the clock was reset, by an influx of H2S from grain surfaces. In the case of disks, this can give us a rough idea of the time since a major reprocessing of the grains. Toward LkCa 15 we observed large amounts of CS, including the isotope, C34S (1/20 CS), but no SO,SO2,H2S or OCS were detected. This indicates that the disk is reasonably old and has been quiescent for a long time, which is consistent with its age ~10Myr and grain chemistry.

13. Chemical Imaging of Outer Disk?
Qi et al. 2004
& in prep
HDO formed via H2D+, possible tracer of H3+?
Kessler et al.
2004, in prep
The effects of UV fields can be measured in the disks with tracers such as CN and HCN. CN is produced photochemically via the destruction of HCN. An increase in the CN/HCN ratio indicates an increase in the UV field. This can be due to stellar UV from a star with high luminosity (as appears to be the case for HD ) or it can be due to interstellar UV which increases as the dust settles to the disk midplane. This is represented by the H/h value (height of the dust in disks over scale height of the gas in an ? Disk as modeled by Chiang et al. 2001). The molecular distribution of CN and HCN is also quite different from that of CO as shown here. It is proposed that this double peaked morphology is a projection effect and that HCN and CN are distributed in a ring around the star, depleted in the center. It is not clear whether this effect is due to UV fields or desorption off grain surfaces.
CO well mixed, while [CN]/[HCN] traces enhanced UV fields.
Is LkCa 15 unusual?
Photodesorption?

14. Disk Molecular Distribution Models
LkCa 15
HCN observations
For models:
Using scaled H density distribution
with varying inner radius cutoff
(Kessler 2003, Ph.D. thesis). R0
Rout NT
Ro=50AU
Ro=100AU
Ro=200AU
Ro=300AU
The chemical models of Aikawa & Herbst and Willacy and Langer suggest that HCN may reside in a ring around the star due either to photodissociation by interstellar and stellar UV or just desorption from grain surfaces.
Using a simple model, in which the column density is constant and the volume density is proportional to that of hydrogen, we simulate this ring structure, setting the outer radius and varying inner radius of the annulus. As we vary the inner radius we can see that these models can reproduce the structure of the HCN emission but only with large depletions in the disk center. More complicated models can be used to simulate the conditions required by the effects of desorption from grains and photodissociation of HCN.
Image fidelity presently limited by (u,v) coverage, sensitivity.

15. Future of the U.S. University Arrays – CARMA
CARMA = OVRO (6 10.4m) + BIMA (9 6.1m) + SZ Array (8 3.5m) telescopes.
March 27, 2004
SUP approved!
2004 SZA at OVRO
2004 move 6.1m
2004 move 10.4m
2005 full operations
Cedar Flat
7300 ft.

16. Disk Observations w/CARMA
Dust
With future arrays such as CARMA and of course ALMA, we will be able to observe many more deuterated species at higher resolution. With its improved sensitivity in the 1mm window, CARMA will allow us to detect deuterated species such as DCN and HDO in on the order of 5hours, versus the 40 hours of integration time represented by the HDO observation presented here. Thus we can increase the number of species and disks. ALMA with its much more complete uv coverage will allow us to recover almost all of the disk structure which is washed out with current arrays.
Md=0.01Msun Rout=120AU Rhole=20AU
HDO:
rms (3sigma) = K
(CARMA w/D config. in 5 hrs, many OVRO transits)

17. Disks with inner holes? TW Hydra w/SMA:
CO 3-2
CO 2-1
Qi et al. 2004, ApJL, in press.

18. Transitional/Debris Disks? HD141569 & Vega w/PdBI:
Vega, Wilner et al. 2002
CO 2-1 from HD141569
J.-C. Augereau & A. Dutrey
astro-ph/

19. Planets are born well inside 50-100 AU…
Planetary velocities are considerable, however, should we use spectroscopy or imaging (or both)?
Theory
Jupiter (5 AU):
V_doppler = 13 m/s
V_orbit = 13 km/s
Simulation G. Bryden, JPL
Observation?

20. Enter ALMA: Superb site & large
Dust simulation (L.G. Mundy), unrealistic
phase errors, but no CLEAN/MEM.
Superb site & large
array exceptional performance (64 12m telescopes, by 2012).
Llano de Chajnantor; 5000 m, good for astronomy, tough for humans!

21. Atmospheric Phase Correction (mm Adaptive Optics)
Atm. fluctuations (mostly
H2O) can vary geom. delay.
|V|eif decorrelation
if Ef>p (each baseline).
If the fluctuations vary
systematically across the
array, phase errors ensue.
Problem is NOT solved.
OVRO WLM
System

22. Sub-mm Interferometry of Disks - Conclusions
(Sub)mm-wave instruments can only study the
outer reaches of large disks at present in lines;
even at these wavelengths the disk mid-plane is
largely inaccessible due to molecular depletion.
Can ions, e.g. N2H+ and esp. H2D+, save the day?
New/combined arrays (SMA, CARMA & ALMA)
will provide access to much smaller scales,
lines/dust may selectively highlight regions of
planet accretion/formation.
With present arrays, the nonlinear nature of image reconstruction algorithms means that analysis is best done in the (u,v) plane. Not so with ALMA.
For ALMA to achieve its potential, active phase correction schemes MUST be developed. This is analogous to multi-conjugate AO (HARD!).
16Apr04

23. Poster Summary: IR Spectroscopy & Disks
R=10, ,000 (30-3 km/s) echelles
(ISAAC,NIRSPEC, PHOENIX,TEXES)
on 8-10 m telescopes can now probe
“typical” T Tauri/Herbig Ae stars:
Keck
CO M-band fundamental
AB Aur HD
NIRSPEC
R=25,000

24. What about other species?
NGC 7538 IRS9
Disk ages are difficult to determine, but chemistry may be useful in this respect. This figure is from a model of the sulfur chemistry in cores at ~50K. This may be similar to the outer regions of disks studied with OVRO. The dominant species migrates from H2S to SO,SO2 to CS over time. Here the timescale represents the time since the clock was reset, by an influx of H2S from grain surfaces. In the case of disks, this can give us a rough idea of the time since a major reprocessing of the grains. Toward LkCa 15 we observed large amounts of CS, including the isotope, C34S (1/20 CS), but no SO,SO2,H2S or OCS were detected. This indicates that the disk is reasonably old and has been quiescent for a long time, which is consistent with its age ~10Myr and grain chemistry.
Boogert et al. 2004, ApJ, in press
16Apr04

25. Systematic Line Width Trends:
Objects thought to be ~face on have the narrowest line widths, highly inclined systems the largest.
As the excitation energy increases, so does the line width (small effect).
Consistent with disk emission, radii range from AU at high J.
Low J lines also resonantly scatter 5 mm photons to much larger distances.
Blake & Boogert 2004, ApJL 606, L73.
16Apr04