1

April 15, 2004 Ringberg workshop

Chemistry in circumstellardisks: overview

Ewine F. van Dishoeck

Leiden Observatory

OutlineOutline

� Inner disk chemistry (<10 AU)� Thermochemistry� Models with transport and mixing

� Outer disk chemistry (10-400 AU)� Radial transport models� 1+1D, 2D models of static flaring disks� Tenuous disks

� Summary

2

I. Inner disk chemistry (<10 AU)I. Inner disk chemistry (<10 AU)

� Large literature on chemical models of the solar nebula since 1970’s� E.g., Lewis, Prinn, Fegley, Lunine, ….

� Models applied to large range of solar system observations� E.g., CO/CH4 planetary atmospheres, comet abundances,

meteorites

� Chemistry thought to be dominated by thermal equilibrium (3-body) rather than kinetic (2-body) processes� 1 AU: n~1014 cm-3, T=800 K� 15 AU: n~1011 cm-3, T=80 K

ThermochemistryThermochemistry

� Principle chemistryCO + 3H2 ↔↔↔↔ CH4 + H2ON2 + 3H2 ↔↔↔↔ 2NH3

CO + H2O ↔↔↔↔ CO2 + H2

� Reactions proceed more to the right at lower T

� Fischer-Tropsch reactions on metallic iron can also convert CO to CH4

Lewis & Prinn 1980

3

Quench surfaceQuench surface

� Equilibrium region: tchem < tmix

� Disequilibrium region: tchem>tmix� Chemical reactions too slow to be equilibrated; composition

similar to that at quench surface� E.g., CO and N2 dominant in solar nebula, since conversion to

CH4 and NH3 in outer region kinetically inhibited

Other processesOther processes� Ice condensation and evaporation: 4-20 AU (species

and time dependent)� Dust evaporation (<1 AU)� Weak (accretion) shocks

� Q: at what radius does most of the gas and dust enter disk?� Lightning� UV and X-rays from Sun/star� Radial and vertical mixing � Ionization by extinct radionuclides (26Al, 60Fe, …)� …..

4

Modern modelsModern models� 1D models including radial mass transport

� E.g. Bauer et al. 1997, Finocchi & Gail 1997, Gail 2001-2004, Willacy et al. 1998, Aikawa et al. 1997, 1999

� 1+1D and 2D vertical models including UV, X-rays, but dynamics uncoupled� Markwick et al. 2002: no stellar radiation => cold surface� Nomura & Milar 2004: with stellar radiation => hot surface

Markwick et al. 2002

Chemical structuredominated by freeze-outand thermal desorptionfrom grains => T-structure crucial

Three-body reactions do not play a role (?)

Freeze-outCS

� Include mass transport processes in steady 1+1D model, both radial advection and vertical mixing � Vertical mixing much shorter timescale => static

atmosphere� No stellar irradiation in heating => cold upper layers� Move parcel of gas from 10 to 1 AU in 8.5x104 yr� Henning & Ilgner 2003, Ilgner et al. 2004, Gail et al. 2004

1 105 AU0

1

Temperature distribution

5

Models with/without vertical mixingModels with/without vertical mixing

CS

CS

Without

With

- CS more abundant at greater z at 8-10 AU due to vertical mixing- CS destroyed at <5 AU by reactions with O

(O higher because more H2O in gas phase?)

Ilgner et al. 2004=> Large effects on chemistry, but difficult to disentangle various effects

SO2

SO

No observational tests (yet)

II. Outer disk chemistryII. Outer disk chemistry

� Consider chemical evolution of parcel of gas as it moves radially from >100 AU to few AU

� Include large gas-phase chemistry network (few hundred species, few thousand reactions) and gas-grain adsorption/desorption processes

� Chemistry dominated by temperature profile: virtually no gas-phase molecules >10 AU, active gas-phase chemistry <10 AU� E.g., Bauer et al. 1997, Finocchi & Gail 1997, Gail 2001-2004, Willacy

et al. 1998, Aikawa et al. 1997, 1999

1D Radial transport models

6

ExampleExample

Aikawa et al. 1999

Abundances after 3x106 yr

=> Everything frozen out at >10 AU

But this is NOT what is observed!

Intermezzo: observations

MWC 480OVRO

LkCa 15

MWC 480

CN 3-2 HCN 4-3

13CO 3-2 HCO+ 4-3

JCMT

Dutrey et al. 1997Kastner et al. 1997Qi 2001Thi et al. 2004

7

Some observational findingsSome observational findings

� Simple gas-phase molecules observed� Ion-molecule reactions (HCO+)� Photon processes (high CN/HCN)� High deuterium fractionation (DCO+) � Low abundances complex species (H2CO, CH3OH)

� Data only sensitive to >50 AU� Lines comes from warm 20-40 K layer with n=106-108 cm-3

� Disk-averaged abundances are “depleted” by factor of 5-100 (using mass from dust continuum and assuming gas/dust=100)

� Solid CO and H2O detected in edge-on disks

Detection of DCO+ in a circumstellar disk

TW Hya face-on disk1.1 µµµµm 1.6 µµµµm

HCO+/2

H13CO+

DCO+Scattered light => radius 200 AU

Weinberger et al. 2002

van Dishoeck et al. 2003

DCO+/HCO+=0.035 => gas in disks is cold with heavy depletions

JCMT

Dutrey et al. 1997Kastner et al. 1997

8

Temperature structureTemperature structure

40 K50

602010

12CO 3-2

13CO

- 12CO ττττ=1 surface near top of disk- 13CO emission from entire disk Van Zadelhoff et al. 2001

Dartois et al. 2003

Starting to image the chemistry in disks

Kessler, Qi, Blake et al. 2003

OVRO mm array

Aikawa et al. 2003

CO HDO

HCO+ CN

9

Flaring disk models (>50 AU)Flaring disk models (>50 AU)� Calculate chemistry in vertical and radial

directions in 1+1D static flaring models� Aikawa et al. 1999, 2001:

� Kyoto minimum solar mass disk model � Low temperatures => needed artifically low sticking coefficient

S=0.03 to match observations� Willacy & Langer 2000

� Two-layer Chiang & Goldreich model� All molecules photodissociated in warm layer� All molecules frozen on grains in cold layers => needed high

photodesorption rate to match observations� Aikawa et al. 2002, van Zadelhoff et al. 2003

� D’Alessio et al. models with continuous T,n gradient� Warm molecular layer where molecules stay off the grains even

with S=1

� No radial or vertical mixing included, but models allow important processes and parameters to be identified

Chemical structure of disks- Surface layer: molecules dissociated by UV photons- Warm intermediate layer: molecules not much depleted, rich chemistry- Cold midplane: molecules heavily frozen out

10

Model flowchartModel flowchart

Vertical structure (R=200 AU)Vertical structure (R=200 AU)

midplane surface

PhotodissociationFreeze-out

Tgas is larger

Ionization fraction:- Surface: 10-4 (C+)- Intermediate: 10-9 (HCO+)- Midplane: ~10-11 (H3

+,H2D+)

11

A few words about chemistryA few words about chemistry

� Chemistry in upper layers resembles that of a PDR � Photodissociation H2 and CO only at 912-1100 Å� Important to treat H →→→→H2 and C+ →→→→C→→→→CO transitions accurately

(Leiden PDR benchmark workshop), e.g., sensitivities to self-shielding, mutual shielding, PAH abundance, …

� Photorates are rapid => chemical equilbrium in ~103 yr� High abundances of radicals (CN, C2H, …) in PDR zone

Gas-phase chemistry

0 2 4 6AV

A few words about chemistryA few words about chemistry

� Elemental abundances (e.g. high vs. low metals) and cosmic ray ionization rate important input parameters

� Gas-phase chemistry not very sensitive to temperature in 10-200 K regime

� Many different chemical networks containing a few hundred to a few thousand reactions => reduction?� Wiebe, Semenov et al. 2003, 2004

� Best agreement with well-studied PDRs and dark clouds is a factor of a few – ten => better agreement for disks would be a miracle!

Gas-phase chemistry

12

A few words about chemistryA few words about chemistry

� Freeze-out/thermal desorption depend sensitively on dust temperature profile� Species dependent: CO T>20 K, H2O T>100 K; binding

energies not well known for all species and depend on type of ice or surface

� Fundamental issues with formulation grain surface reactions (diffusion-limited vs. accretion-limited)

� Timescales: tads~2x109/nH yr => strongly dependent on density

Gas-grain interactions

Recent efforts: importance of shape Recent efforts: importance of shape and treatment UV radiation fieldand treatment UV radiation field

Van Zadelhoff et al. 2003

Normal ISRF: ~1x108 ph s-1 cm-2

�Enhanced by factor > 1000 in upper layers

Full 2D calculationincluding absorptionand scattering

Penetration depthdepends on opticalproperties grains

13

Full 2D vs. 1+1DFull 2D vs. 1+1D

1+1D

Full 2D:CO and H2not p.d. by star

Full 2D:CO and H2p.d. by star

Van Zadelhoff et al. 2003

Effect of stellar UVEffect of stellar UV

Green

Red

Blue

=> Molecules extended to greater height if no UVVan Zadelhoff et al. 2003

14

Observational diagnostics stellar UVObservational diagnostics stellar UV

- Difficult to disentangle models with single-dish data- Need ALMA resolution to probe variations- HCN dissociated by Ly αααα, CN not

Importance of Lyman Importance of Lyman αααααααα

-Photodissociation rate kpd =���� σ(λ)σ(λ)σ(λ)σ(λ)I(λλλλ)dλλλλ-Some molecules are dissociated by Lyαααα (e.g., H2O, HCN), others

are not (e.g. CN, CO, H2)Bergin et al. 2003

BP TauTW Hya

Observed FUV spectra of T Tauri stars

ISRFx540

15

Tenuous disks: Tenuous disks: ββββββββ PictorisPictoris

Kamp & Bertoldi 2000, Kamp & Samnar 2004

- Disks are optically thin to UV and IR radiation => analytic solution for Tdust

- Use stellar atmosphere model for stellar radiation- Recent work: extension to G-stars including chromosphere

CO CO vsvs HH22Stellar UV Stellar + interstellar UV

=> Absence of CO does not mean absence of H2!Kamp & Bertoldi 2000

CO p.d.

H2 self-shielding

Note: both CO and H2 are only dissociated at 912-1100 Å

16

CO gas versus iceCO gas versus iceStellar UV Stellar + interstellar UV

CO gas

CO ice

C and CC and C++ as gas tracers of tenuous disksas gas tracers of tenuous disks

0.2 MEarth

2 MEarth

C

C+

ββββPic

2 MEarth 0.2 MEarth

Kamp et al. 2003

Vdrift=max Vdrift=0

17

Summary chemistrySummary chemistry� Inner disks: highly sophisticated models including radial

transport and vertical mixing being developed, but no observational tests yet (need full ALMA with 64x12m)

� Outer disks: 1+1D static flaring disk models reproduce submillimeter lines of simple species within factors of few� Emission comes from intermediate warm layer� Chemistry in warm layer dominated by photodissociation and

thermal desorption => important to know Tdust and UV field accurately

� Molecules strongly depleted in cold midplane => high deuteration fraction

� CO is not a good gas mass tracer (alternatives H2, C, C+ to be determined)

� Importance of vertical mixing to be determined: can chemistry put limits?

Summary chemistrySummary chemistry

� Does chemistry affect disk structure?� Gas temperature in upper disk layers can be

significantly higher than that of dust (see talks tomorrow) => affects vertical structure and line emission; chemistry determines abundances coolants O, C+, C and CO

� Freeze-out and dust destruction modify grain opacities

� Ionization fraction => affects magnetohydrodynamics (but can probably be estimated by small network)